A TREATISE ON QUANTITATIVE INORGANIC ANALYSIS. "There is certainly no English work approaching this in com- pleteness." Athenocum. PHYSICO-CHEMICAL TABLES FOR THE USB OF ANALYSTS, PHYSICISTS, CHEMICAL MANU- FACTURERS, AND SCIENTIFIC CHEMISTS. BY JOHN CASTELL- EVANS, F.I.C., F.C.S. IN TWO VOLUMES, EACH COMPLETE IN ITSELF. VOLUME 1. CHEMICAL ENGINEERING, AND PHY- SICAL CHEMISTRY. In Large 8vo, pp. i-xxxii + 1-548. 24s. net. VOLUME II. PHYSICAL AND ANALYTICAL CHEM- ISTRY. In Large 8vo, pp. i-xiv + 549-1 235. 36s. net. Each volume is strongly bound in half -morocco, library style, and may be purchased separately, but the Index is in Volume II. SOME PRESS OPINIONS. "Well arranged and exceedingly comprehensive ... we congratulate the author." Nature. "There is certainly no English work approaching it in completeness . . . the author and compiler must be congratulated." Athenceum. "There is nothing of so comprehensive a character as this." Electrical Review. 1 ' The very tables themselves inspire confidence ... a marvellous compila- tion."- Analyst. " It would be a difficult matter to suggest additions or improvements to make the book more complete or convenient in use." Chemical News. "Pre-eminently a book which should be in every reference library." Chemist and Druggist of Australasia. A TEEATISE ON QUANTITATIVE INORGANIC ANALYSIS WITH SPECIAL REFERENCE TO THE ANALYSIS OF CLAYS, SILICATES, AND RELATE!) MINERALS : , , , BEING VOL. I. OF A TREATISE ON THE CEKAMIC INDUSTRIES. BY J. W. MELLOR, D.Sc. 2 Coloured flMates anD 206 otber 3Hu0tratfons. LONDON: CHARLES GRIFFIN AND COMPANY, LIMITED. EXETER STREET, STRAND. 1913. [All rights reserved.] PREFACE. THIS book is to be considered as the first volume of a treatise on the Ceramic and Silicate Industries. The part dealing with clays and related materials was expanded from a special course of lectures I gave on that subject about three or four years ago ; but, as the work developed, it was thought best to collect all the analytical operations into one volume and thus avoid abrupt changes of subject in the more technical parts, to be published later, and sometimes referred to as " Volume II." The processes here described are those at present used in the Testing Department of the County Pottery Laboratory, Staffordshire, for the analysis of clays, bricks, and related materials ; for glazes and enamels ; for colouring mixtures ; Jor the special refractories introduced in recent years ; etc. The book is divided into five parts. Some general analytical processes are described in Part I. ; the analysis of a typical simple silicate is discussed in Part II. ; the methods to be adopted with more complex silicates glazes, enamels, colours, etc. are indicated in Part III. ; while Part IV. shows the modifications required when some of the more difficult, as well as some of the rarer, elements are present. The last part Part V. indicates methods to be employed when acidic elements have to be determined, and a chapter on the so-called "rational analysis" concludes the volume. Some useful reference tables are included in an Appendix. The methods of testing clays and clay products, as distinct from analysis proper, will be described in Vol. II. The method for the analysis of a typical silicate detailed in Part II. of this book is perhaps that which gives the most accurate results so far as our present knowledge goes for clays, silicate rocks, silicate minerals, and related materials. The work of W. F. Hillebrand of the U.S. Geological Survey has been largely instrumental in developing this branch of analysis to its present condition. These methods do not allow analyses to be conducted with that speed required in certain industrial work, and accordingly I have indicated the short-cuts available to those who have mastered the general methods. So many handbooks have been published which specialise in cement, water, and fuel analyses that these subjects have not been elaborated particularly here. Many of the analytical processes have been described on typed sheets for six or seven years, and employed by the students of the analysis class of the Pottery School, and by the professional staff in the Testing Department. The questions ix X A TREATISE ON CHEMICAL ANALYSIS. which have been asked from time to time have enabled me to strengthen many weak places in the descriptions of the methods. Early proofs have also been used in the laboratory for over a year, and the descriptions of the more common analyses must have been verified at the bench by a score of different workers. During this time, the great mass of literature on the subject has been overhauled anew, and the extensive footnotes indicate where papers bearing on the subjects discussed in the text can be found. The reader must thank the publishers for giving me every facility for elaborating the original proofs after their handling in the laboratory. I am specially indebted to my colleagues Messrs A. D. Hold- croft, J. C. Green, and C. Edwards for the trouble they have taken in reading the proofs ; to Miss M. Chetham for help with the indexing ; and to Mr F. J. Austin for photographing the apparatus and specimens. J. W. M. COUNTY POTTERY LABORATORY, STOKE - ON -T R RNT. CONTENTS. PART I. GENERAL. PAGE Introduction ... . . xxi The Subdivisions of Analytical Chemistry ; The Birth of Analytical Chemistry ; Some Uses of Analytical Chemistry ; Some Limitations of Analytical Chemistry ; The Evolution of Modern Analytical Practice. CHAPTER I. Weighing ..... 1. The Balance ; 2. An Outline of the Theory of the Balance; 3. The Location, Care, and Use of the Balance ; 4. Weighing Hygroscopic and Volatile Liquids and Powders ; 5. Weighing ; 6. The Sensibility of the Balance ; 7. The Accuracy of the Weighing ; 8. Correcting the Weights ; 9. The Influence of Inequalities in the Lengths of the Balance" Arms on the Weighings ; 10. Corrections for the Buoyancy of the Air ; 11. Summary. CHAPTER II. The Measurement of Volumes . . .28 12. Volumetric and Gravimetric Processes of Analysis; 13. The Influence of Variations of Temperature; 14. The Litre and its Subdivisions; 15. Meniscus, Parallax Errors, and Burette Floats ; 16. The Calibration of Standard Flasks ; 17. The Drainage or After-flow from Measuring Vessels ; 18. The Calibration of Pipettes ; 19. The Calibration of Burettes ; 20. Burette Stands, and Burette Cocks. CHAPTER III. Volumetric Analysis . . . .45 21. Normal Solutions ; 22. Standard Solutions of Hydrochloric Acid ; 23. Temperature Correction of Solutions of Hydrochloric Acid ; 24. The Adjustment of Standard Solutions ; 25. The Adjustment of the Specific Gravity of Solutions ; 26. The Calculations of Analytical Chemistry ; 27. The Auto- matic Filling of Burettes and Pipettes ; 28. Indicators ; 29. Standard Solu- tions of Sodium Hydroxide ; 30. The Errors of Experiment in Volumetric Analysis ; 31. Direct Titrations Sodium and Potassium Carbonates ; 32. Back Titrations Calcium Carbonate ; 33. Back Titrations Mixed Hydroxides and Carbonates ; 34. Correction for the Volume of Suspended Solids ; 35. Sodium Silicate Water Glass ; 36. The Volumetric Determination of Silver and Chlorine Volhard's Process ; 37. The Volumetric Determination of Chlorine Mohr's Chromate Process. CHAPTER IV. Colorimetry and Turbidimetry . . .82 38. Colorimetric Methods of Analysis Colorimetry; 39. Duboscq's Dipping Colorimeter; 40. Weller's Colorimeter ; 41. Nessler's Tubes ; 42. Turbidity Methods of Analysis Turbidimetry ; 43. Some Errors in Colorimetry and Turbidimetry. xii A TREATISE ON CHEMICAL ANALYSIS. PAGE CHAPTER V. Filtration and Washing. 8 ? 44. Filter Paper ; 45. Filtration ; 46. Wash-Bottles ; 47. The Theory of Washing Precipitates ; 48. Bunsen's System of Accelerated Filtration ; 49. Tared Filter Papers ; 50. Filtration Tubes ; 51. Filtration through Perforated Discs and Funnels ; 52. Gooch's Filtration Crucibles ; 53. Gooch's Crucibles packed with Soluble or Volatile Felts ; 54. Gooch's Crucibles packed with Platinum Felt Munroe's Crucibles. CHAPTER VI. Heating and Drying . .110 55. Heating ; 56. Platinum Apparatus ; 57. Desiccators ; 58. Labora- tory Hoods, Fume Closets. CHAPTER VII. Pulverisation and Grinding . . . .120 59. Pulverising Large Quantities ; 60. Pulverising Small Quantities ; 61. Grinding in Agate Mortars; 6'2. The Dangers of Fine Grinding; 63. Sieving, Lawning, or Screening. CHAPTER VIII. Sampling . . . . '. 1 127 64. The Problem of Sampling ; 65. Selecting the Sample ; 66. Reducing the Bulk of the Sample ; 67. Machine and Automatic Sampling ; 68. Reduc- ing the Grain Size and Bulk of the Material during Sampling ; 69. Receipt and Dispatch of Samples ; 70. Sampling Beds of Clay ; 71. Commercial Value of Clay Deposits. CHAPTER IX. The Reagents . . . - . . .141 72. Testing the Reagents ; 73. Bottles for Reagents ; 74. The Action of Reagents on Glass and Porcelain ; 75. The Equivalent System of Making Solutions ; 76. Gas Generators ; 77. Distilled Water PART IL TYPICAL SILICATE ANALYSES-CLAYS. CHAPTER X. The Determination of Volatile Matters . .155 78. Hygroscopic Moisture ; 79. Loss on Ignition. CHAPTER XI. Opening up Silicates . 160 80. Summary of the Different Methods; 81. Sodium Carbonate for Silicate Fusions ; 82. Opening Clays and Silicates by Fusion with Sodium Carbonate. CHAPTER XII. The Determination of Silica . 16 7 83. The Determination of Silica ; 84. The Theory of Silica Determinations. CHAPTER XIII. The Ammonia Precipitate . . ^ 85. The Precipitation by means of Ammonia ; 86 The Theorv of th* Ammonia Precipitation ; 87. The Determination of the* Ammonia Precipitate CONTENTS. Xlii PAGE CHAPTER XIV. The Determination of Iron . . . .187 88. The Determination of Iron ; 89. The Reduction of Ferric to Ferrous Salts for the Permanganate Titration ; 90. The Standardisation and Use of Potassium Permanganate ; 91. The Volumetric Determination of Iron Marguerite's Permanganate Process; 92. The Colorimetric Determination of Iron. CHAPTER XV. The Determination of Titanium . . . 201 93. Weller's Colorimetric Process ; 94. The Gravimetric Determination of Titanium Gooch's Process ; 95. Blair's Modification of Gooch's Process ; 96. The Computation of the Results for Alumina. CHAPTER XVI. The Determination of Calcium and Magnesium . 211 97. The Properties of Calcium Oxalate ; 98. The Gravimetric Determina- tion of Calcium ; 99. The Volumetric Determination of Lime Kraut's Process ; 100. The Properties of Ammonium Magnesium Phosphate ; 101. The Gravimetric Determination of Magnesium. CHAPTER XVII. The Determination of the Alkalies . .222 102. Meretricious Methods for Estimating the Alkalies ; 103. The Separa- tion of the Alkalies as Mixed Chlorides Smith's Process ; 104. The Separation of the Alkalies as Mixed Chlorides Berzelius' Process; 105. The Indirect Determination of Potash and Soda ; 106. A General Theory of Indirect Determinations ; 107. The Properties of Potassium Chloroplatinate ; 108. The Separation of Potassium as Potassium Chloroplatinate ; 109. The Separation of Potassium as Potassium Perchl orate ; 110. The Determination of Sodium Oxide; 111. The Preparation of Hydrochloroplatinic Acid from Platinum Residues and Scraps. CHAPTER XVIII. Abbreviated Analyses and Analytical Errors . 242 112. Exacts. Works Analyses; 113. Abbreviated Schemes of Analysis; 114. The Indirect Determination of Lime and Magnesia ; 115. Permitted Errors of Analyses ; 116. The Chief Sources of Error in Analyses Generally ; 117. The Statement of the Results. CHAPTER XIX. Electro-Analysis . . . . .253 118. Some Definitions; 119. Some Factors which Determine Success in Electro- Analysis ; 120. The Apparatus for Electro -Analysis ; 121. The Electrolytic Determination of Copper. PART III. ANALYSIS OF GLASSES, GLAZES, COLOURS, AND COMPLEX SILICATES. CHAPTER XX. The Analysis of Glass, Glazes, Enamels, and Colours . . . . . .265 122. The Selection of the Sample ; 123. " Opening" the Sample ; 124. The Behaviour of Metals of the Hydrogen Sulphide Group in the Silica Determina- tion ; 125. The Theory of Precipitation by Hydrogen Sulphide ; 126. The Separation of the Metals Precipitated by Hydrogen Sulphide in Acid Solution ; 127. The Separation of Tin, Arsenic, and Antimony from the Remaining Metals. xiv A TREATISE ON CHEMICAL ANALYSIS. PAGE CHAPTER XXI. The Determination of Arsenic 281 128. The Separation of Arsenic from Antimony and Tin by Distillation ; 129. The Separation of Arsenic from Antimony and Tin as Magnesium Ammonium Phosphate; 130. Notes on Iodine, Potassium Iodide, Starch, and Sodium Thiosulphate ; 131. Mohr's Iodine Volumetric Process for Arsenic; 132. Pearce's Volumetric Process for Arsenic; 133. The Evaluation of Arsenic Oxide. CHAPTER XXII. The Determination of Antimony . 295 134. Antimony Sulphide ; 135. The Gravimetric Separation of Antimony and Tin Clarke and Henz's Process ; 136. Weller's Volumetric Iodide Process for Antimony ; 137. Gyory's Volumetric Bromate Process for Antimony ; f !38. The Volumetric Determination of Antimony and Arsenic in the Presence of Tin ; 139. Metallic Precipitation ; 140. The Evaluation of Antimony Compounds. CHAPTER XXIII. The Determination of Tin . . . .307 141. The Metallic Precipitation of Tin ; 142. The Precipitation of Tin as Hydroxide LowenthaPs Process ; 143. The Precipitation of Tin as Sulphide ; 144. The Volumetric Determination of Tin Mene's Ferric Chloride Process; 145. Lenssen's Volumetric Iodine Process for Tin ; 146. The Evaluation of Commercial Compounds of Tin ; 147. Henz and Classen's Electrolytic Process for Tin. CHAPTER XXIV. The Determination of Lead . .315 148. The Properties of Lead Sulphate ; 149. The Determination of Lead as Sulphate ; 150. The Separation of Lead from Bismuth, Copper, and Cadmium ; 151. The Analysis of White Lead ; 152. The Analysis of Red Lead ; 153. The Analysis of Lead Chromates ; 154. The Determination of Silver in Lead Compounds by the Turbidity Method ; 155. The Determination of Silver and Gold by Cupellation ; 156. The Analysis of Galena ; 157. The "Government Test" for the Solubility of Lead Frits ; 158. The Gravimetric Deteimination of Lead as Molybdate ; 159. Conversion Factors ; 160. The Volumetric Deter- mination of Lead ; 161. The Electrolytic Process for the Determination of Lead ; 162. The Rapid Deposition of Lead Dioxide by a Rotating Electrode Exner's Process ; 163. The Colorimetric Determination of Lead. CHAPTER XXV. The Determination of Bismuth and Mercury . 341 164. The Separation of Mercury from Lead, Bismuth, Copper, and Cadmium Rath's Process ; 165. The Gravimetric Determination of Mercury as Sulphide Volhard's Process; 166. The Distillation Process for Mercury; 167. The Separation of Bismuth from Lead, Cadmium, and Copper Lowe's Process; 168. The Separation of Bismuth from Copper and Cadmium Jannasch's Process; 169. The Determination of Bismuth Colorimetrically. CHAPTER XXVI. The Determination of Copper and Cadmium . 350 170. lUvot's Thiocyanate Process for Copper : 171. De Haen's Volumetric i, iS e U e ^ S ( r C PP er 5 172 - The Evaluation of Copper Oxide and Carbonate ; rnu ^i Colonm etric Determination of Copper Carnelley's Process; ( 121. Ine Electrolytic Determination of Copper) ; 174. The Gravimetric Determina- tion of Cadmium as Sulphate ; 175. The Volumetric Determination of Cadmium -Bergs Process; 176. The Electrolytic Determination of Cadmium -Beilstein and Jawem's Process. CONTENTS XV PAGE CHAPTER XXVII. The Determination of Zinc . 359 177. The Analysis of Silicates containing Zinc Compounds ; 178. The Theory of the Basic Acetate Separation ; 170. The Basic Acetate Separation ; 180. The Separation of Zinc from Manganese, Cobalt, and Nickel ; 181. The Determination of Zinc as Phosphate ; 182. The Volumetric Ferrocvanide Process for Zinc ; 183. The Evaluation of Zinc Oxide. CHAPTER XXVIII. The Determination of Manganese . .371 184. The Effect of Manganese on Silicate Analyses ; 185. The Precipitation of Manganese by the Bromine Process ; 186. The Precipitation of Manganese by Ammonium Sulphide ; 187. The Gravimetric Determination of Manganese Gibb's Phosphite Process; 188. The Volumetric Determination of Manganese Pattinson's Process ; 189. The Volumetric Determination of Manganese Volhard's Process ; 190. Evaluation of Manganese Oxide Mohr's. Process ; 191. The Colorimetric Determination of Manganese Walter's Pjrocess ; 192. The Analysis of Wads and Manganese Earths. CHAPTER XXIX. The Determination of Cobalt and Nickel . . 386 193. The Detection of Cobalt and Nickel ; 194. The Properties of Cobalt and Nickel Sulphides ; 195. The Separation of Manganese from Cobalt and Nickel ; 196. The Separation of Cobalt and Nickel Fischer's Nitrite Process ; 197. The Separation of Nickel and Cobalt Liebig's Cyanide Process ; 198. The Separation of Small Amounts of Cobalt from Large Amounts of Nickel Ilinsky and Knorre's Nitroso-3-Naphthol Process ; 199. The Separation of Small Amounts of Nickel from Large Amounts of Cobalt Brunck's a-Dimethyl- glyoxime Process ; 200. The Electrolytic Process for Cobalt and Nickel Fresenius and Bergmann's Process ; 201. The Colorimetric Determination of Cobalt ; 202. The Evaluation of Cobalt and Nickel Oxides ; 203. The Volu- metric Determination of Nickel and Cobalt. PART IV. SPECIAL METHODS BASES. CHAPTER XXX. The Determination of Molybdenum, Tungsten, Columbium, and Tantalum . . . 405 . 204. Molybdenum, Tungsten, Columbium, and Tantalum in Silicate Analyses ; 205. The Detection of Tungsten, Tantalum, Molybdenum, Colum- bium ; 206. The Determination of Tungsten as Tungsten Trioxide ; 207. The Gravimetric Determination of Tungsten Berzelius' Process ; 208. The Separa- tion of Tungsten and Silica ; 209. The Separation of Tungsten and Tin ; 210. The Separation of Tungsten from Tin and Antimony Talbot's Process; 211. The Separation of Tungsten from Arsenic and Phosphorus Kehrmann's Process ; 212. The Precipitation of Molybdenum as Sulphide ; 213. The Gravimetric Determination of Molybdenum as Oxide and as Sulphide ; 214. The Gravimetric Determination of Molybdenum as Lead Molybdate ; 215. The Volumetric Determination of Molybdenum by Potassium Permanganate ; 216. The Separa- tion of Tungsten and Molybdenum Hommel's Process ; 217. The Separation of Tungsten and Molybdenum Pechard's Process ; 218. The Determination of Niobium and Tantalum Simpson's Process; 219. The Separation of Tantalum and Niobium- -Marignac's Process ; 220. The Estimation of Niobium and Tantalum Oxides Specific Gravity Method. xvi A TREATISE ON CHEMICAL ANALYSIS. PAGE CHAPTER XXXI. The Determination of Gold and Selenium. . 425 221. The Precipitation of Gold and Platinum ; 222. The Separation of Gold and Platinum from Tin, Arsenic, and Antimony ; 223. The Colorimetric Determination of Gold ; 224. The Analysis of Colours containing Purple of Cassius ; 225. The Determination of Gold and Silver by Cupellation and Parting ; 226. The Determination of Platinum ; 227. The Detection of Selenium ; 228. The Gravimetric Determination of Selenium Sulphurous Acid Process. CHAPTER XXXII. The Determination of Aluminium and Beryllium 444 229. The Gravimetric Determination of Alumina Hess and Campbell's Process; 230. The Analysis of Bauxite; 231. The Analysis of Alumina Hydrated and Calcined ; 232. The Analysis of Cryolite ; 233. The Determina- tion of Beryllium ; 234. The Gravimetric Determination of Beryllium Parsons and Barnes' Process. CHAPTER XXXIII. Special Methods for Iron Compounds . . 451 235. The Volumetric Determination of Iron in Hydrochloric Acid Solutions Reinhardt's Process ; 236. The Volumetric Determination of Iron Penny's Dichromate Process ; 237. The Gravimetric Determination of Iron Ilinsky and Knorre's Process 11 , 238. The Gravimetric Determination of Iron Baudisch's Cupferron Process ; 239. The Separation of Iron Ether Process ; 240. The Analysis of Iron Oxides, Red Earths, and Iron Ores ; 241. The Determination of Ferrous Oxide ; 242. Disturbing Factors in the Determination of Ferrous Oxide. CHAPTER XXXIV. The Determination of Chromium, Vanadium, and Uranium . . . . .467 243. The Errors in Silicate Analyses due to the Presence of Chromium, Vanadium, and Uranium ; 244. The Detection of Chromium, Vanadium, and Uranium ; 245. The Separation of Chromium from Iron, Aluminium, and Titanium ; 246. The Separation of Chromium, in the Analysis of Silicates ; 247. The Colorimetric Determination of Chromium ; 248. The Analysis of Chromites and Chromic Oxide ; 249. The Volumetric Determination of Chromium ; 250. The Gravimetric Determination of Chromium as Barium Chromate ; 251. The Gravimetric Determination of Chromium as Chromic Oxide ; 252. The Volumetric Determination of Vanadium ; 253. The Gravi- metric Separation of Chromium and Vanadium ; 254. The Rapid Determination of Vanadium Cain and Hostetter's Process ; 255. The Simultaneous Determina- tion of Small Quantities of Titanium and Vanadium Colorimetrically ; 256. The Separation of Uranium ; . 257. The Gravimetric Determination of Uranium as Uranium Oxide ; 258. The Gravimetric Determination of Uranium as Uranium Phosphate ; 259. The Separation of Uranium as Uranium Ferrocyanide Fresenius and Hintz's Process ; 260. Belohoubek's Volumetric Process for Uranium ; 261. Operations with Sealed Tubes. CHAPTER XXXV. The Determination of Zirconium, Thorium, and the Rare Earths . . . .495 262. The Separation and Detection of Zirconium ; 263. The Gravimetric Determination of Zirconium as Phosphate ; 264. The Analysis of Zircon 265 The Rare Earths; 266. The Determination of "Rare Earths" in Silicates' 267. The Analysis of a Mixture of Rare Earths ; 268. The Gravimetric Determination of Thorium Metzger's Process ; 269. The Volumetric Determina- tion of Cerium Knorre s Process CONTENTS. xvii PAGE CHAPTER XXXVI. Special Methods for the Determination of Barium, Strontium, Calcium, and Magnesium 513 270. The Influence of Barium and Strontium on the Calcium and Magnesium Precipitates ; 271. The Separation of Calcium from Strontium and Barium Stromeyer and Rose's Process ; 272. The Separation of Barium from Strontium and Calcium Chromate Process ; 273. The Determination of Barium in Insoluble Silicates ; 274. The Complete Analysis of Limestones, Gault Clays, etc. ; 275. The Partial Analysis of Limestones, Dolomites, Magnesites, and Marls ; 276. The Mineralogical Analysis of Limestones and Marls ; 277. The Determination of Free Lime in Quicklime, Mortars, etc. ; 278. The Analysis of Calcium Sulphate, Plaster of Paris, and Gypsum ; 279. The Analysis of Barytes and Witherite. CHAPTER XXXVII. Special Methods for the Determination of Alkalies and their Salts . . .533 280. The Gravimetric Determination of Lithium Kahlenberg's Process; 281. The Gravimetric Determination of Lithium Gooch's Process ; 282. The Evaluation of Potassium and Sodium Salts ; 283. The Colorimetric Determina- tion of Potassium ; 284. The Volumetric Determination of Potassium Cobalt Nitrite Process. PART V. SPECIAL METHODS ACIDS AND NON-METALS. CHAPTER XXXVIII. The Determination of Carbon Free and Combined . . . 545 285. The Direct Determination of Carbon; 286. The Wet Combustion Process for the Determination of Carbon as Carbon Dioxide ; 287. The Errors in Analyses involving the Weighing of Absorption Tubes ; 288. The Detection of Carbon Dioxide ; 289. The Rapid Determination of Carbon Dioxide in Carbonates ; 290. The Gravimetric Determination of Carbon Dioxide ; 291. The Volumetric Determination of Carbon Dioxide in Carbonates Scheibler and Dietrich's Process ; 292. The Volumetric Determination of Carbon Dioxide in Carbonates Lunge and Marchlewski's Process ; 293. The Dry Combustion Process for the Determina- tion of Carbon Shimer's Process; 294. The Analysis of Carborundum ; 295. The Analysis of Siloxicon ; 296. The Analysis of Graphite and Graphite Crucibles. CHAPTER XXXIX. The Determination of Water . . .570 297. Brush and Penfield's Method ; 298. Jannasch's Process for Water ; 299. Fractional Dehydration Water Lost at Different Temperatures. CHAPTER XL. The Determination of Boron Oxide . . 576 300. The Detection of Boric Oxide ; 301. The Determination of Boric Oxide ; 302. The Evaluation of Boric Acid ; 303. The Evaluation of Borax ; 304. The Evaluation of Borocalcite, Boronatrocalcite, Boracite, and Calcium Borate ; 305. The Determination of Boric Oxides in Silicates Wherry's Process ; 306. The Determination of Boric Oxide in Silicates Distillation Process ; 307. The Determination of Silica and Alumina in Borosilicates. I xvili A TREATISE ON CHEMICAL ANALYSIS. PAGE CHAPTER XLI. The Determination of Phosphorus . 308. The Properties of Ammonium Phosphomolybdate ; 309. The Gravi- metric Determination of Phosphorus Woy's Process ; 310. The Reprecipitation as Ammonium Magnesium Phosphate ; 311. The Reprecipitation as Ammonium Phosphomolybdate ; 312. Rapid Processes for Estimating the Ammonium Phosphomolybdate ; 313. The Volumetric Determination of Phosphorus- Uranium Process ; 314. The Volumetric Determination of Phosphorus Joulie's Magnesium Citrate Process ; 3 1 5. The Colorimetric Determination of Phosphorus ; 316. The Simultaneous Determination of Phosphorus and Silica ; 317. The Analysis of Bone-china Bodies. CHAPTER XLII. The Determination of Sulphur . .610 318. The Properties of Barium Sulphate; 319. The Determination of Sulphates in Clays and Insoluble Silicates ; 320. The Determination of Sulphur as Barium Sulphate ; 321. The Filtration of the Barium Sulphate Precipitate ; 322. The Determination, of Sulphur in Pyrites, Limestones, Coals, etc. ; 323. The Volumetric Determination of Sulphates Raschig's Process ; 324. The Determination of the Soluble Salts in Clays ; 325. The Determination of Sulphates by the Turbidity Process ; 326. The Amount of Barium Salt required to make the Soluble Sulphates of Clays Innocuous. CHAPTER XLIII. The Determination of the Halogens . . 634 327. The Detection of Fluorides ; 328. The Determination of Fluorine as Calcium Fluoride ; 329. The Determination of Silica in the Presence of Fluorides ; 330. The Gravimetric Determination of Fluorine as Potassium Fluosilicate ; 331. The Colorimetric Determination of Fluorine Steiger's Process ; 332. The Determination of Fluorine as Gaseous Silicon Fluoride Oettel and Hempel's Process ; 333. The Analysis of Calcium Fluoride ; 334. The Analysis of Sodium Silico- fluoride ; 335. The Properties of Silver Chloride ; 336. The Determina- tion of Chlorides ; 337. The Determination of Silver ; 338. The Determination of Iodine. CHAPTRR XLIV. The Rational Analysis of Clays . . .656 339. Clays ; 340. The Separation of Minerals by Treatment with Chemical Reagents ; 341. The Rational or Mineralogical Analysis of Clays ; 342. The . Effect of the Sulphuric Acid Treatment on some of the Minerals in Clays ; 343. The Composition of the Argillaceous Matter ; 344. Free, Combined, and Colloidal Silica ; 345. The Composition of Felspathic and Quartz Detritus ; 346. The Ultimate and Rational Analysis ; 347. Are the Ultimate and Rational Analyses Consistent ? APPENDIX TABLES ...... 675 THE LIBRARY .... . 732 INDEX OF NAMES . . . ... . 737 INDEX OF SUBJECTS . . 753 LIST OF TABLES. II. Calibration Data for Weights III. Corrections for Weights IV. Effect of Use on a Set of Weights . V. Effect of Buoyancy of Air on Weighings VI. Effect of Humidity of Air on Weighings 3LK PAGE I. Permitted Errors in a Set of Weights . . .17 19 . . .20 21 .23 25 VII. Volume and Density of Water at different Temperatures . . .29 VIII. Reduction for change in Apparent Volume of Water with Temperature . 29 IX. Effect of the Nature of a Liquid on the Volume delivered by a Pipette . 38 X. Burette Corrections ........ 41 XI. Temperature Corrections for N-acid Solutions . . . .49 XII. Conditions under which Indicators can best be used . . .63 XIII. Temperature Corrections for N-alkali Solutions . . . .67 XIV. Indicator Corrections for Mohr's Chromate Process for Chlorine . . 80 XV. Amount of Ash in Munktell's Swedish Filter Papers ... 88 XVI. Relation between Filter Paper Ash and the Liquid Filtered ... 89 XVII. Influence of Grinding Apparatus on the Composition of Powders . . 123 XVIII. Influence of Fineness of a Powder on its Hygroscopicity . . . 124 XIX. The I.M.M. Standard Laboratory Sieve Scale . . .126 XX. Relation between Errors in Sampling and the Size of Sample . . 134 XXI. Relation between Fineness of Bulk Material and Size of Sample . . 135 XXII. Relation between Fineness of Sample and Bulk of Material . .135 XXIII. Action of Reagents on Porcelain and Glass ..... 144 XXIV. Action of Reagents on the Best Types of Glass Apparatus . . .145 XXV. Composition of Chemical Glass Ware. . ... .145 XXVI. Strength of Stock Acids and Ammonia . . . . . 147 XXVII. Composition of Silica Residues .... ,170 XXVIII. Effect of the Dehydration Temperature on the Determination of Silica in Aluminous Clays ....... 173 XXIX. Effect of the Dehydration Temperature on the Determination of Silica in Calcareous Clays ....... 174 XXX. Effect of the Dehydration Temperature on the Determination of Silica in Magnesium Clays . . . . . . 174 XXXI. Effect of Sodium Chloride on Silicic Acid . . . . .175 XXXII. Effect of Time of Drying on Soluble Silica ..... 175 XXXIII. Dehydration of Silicic Acid at 800 ...... 176 XXXIV. Effect of Bismuth Oxide in Inhibiting the effect of Titanium on the Per- manganate Titration . . . . . . .190 XXXV. Conversion Table for Permanganate Solutions .... XXXVI. Temperature Corrections for Potassium Permanganate . . . 198 XXXVII. Comparison of Zinc and Sulphite Reductions in the Permanganate Process for Iron in Clays . . . . . . .199 XXXVIII. Effect of Sulphuric Acid on the Precipitation of Titanic Oxide . . 207 XXXIX. Loss of Alkalies in Smith's Process 225 XL. Test Analyses Indirect Separation of Potassium and Sodium Chlorides . 228 XLI. Relative Errors in the Indirect Processes for the Determination of Potash and Soda ........ 228 XLII. Effect of Drying Potassium Chloroplatinate formed in Dilute and in Con- centrated Solutions ....... 238 XLI 1 1. Accidental and Constant Errors in Silicate Analysis . - . XLIV. Accidental Errors in Eight Clay Analyses . XLV. Comparative Analyses of Argillaceous Limestones . . . 248 XX A TREATISE ON CHEMICAL ANALYSIS. TABLE PAQE XLVI. Influence of Calcination on Magnesium Pyroarsenate XLVII. Test Analyses of Mixtures of Tin and Antimony . 298 XLVIII. Test Analyses of Antimony Ores ... XLIX. Test Analyses with the Electrolytic Process for Tin . L. Solubility of Lead Sulphate in Mineral Acids . . ... 315 LI. Solubility of Lead Sulphate in Ammonium and Sodium Acetates . . 310 LII. Contamination of the Ammonia Precipitate by Zinc . LIII. Contamination of the Ammonia Precipitate by Zinc . . 360 LIV. Effect of Formic Acid on the Separation of Zinc and Nickel . . 365 LV. Distribution of Manganese among the different Constituents of a Clay Analysis . . . . . " 371 LVI. Test Analyses Tungsten and Molybdenum Oxides . . ,. . 416 LVII. Comparison of different Precipitating Agents for Gold . . . 426 LVIII. Effect of Acidity of Solution on the Precipitation of Selenium and Tellurium by Sulphur Dioxide ....... 442 LIX. Relation between the Strength of Acid and the Partition of Iron between Ether and Water ...... 457 LX. Effect of an Excess of Alkali on the Separation of Iron, Aluminium, and Chromium . . . . . ... 470 LXI. Effect of Filter Paper on Chromate Solutions .... 477 LX 1 1. Solubilities of the Rare Earth Oxalates . . . . . ' 500 LXIII. Test Analyses for Lithium ....... 535 LXIV. Liquids for Vapour Baths ....... 574 LXV. Test Analyses for Boric Oxide in Minerals ..... 584 LXVI. Solubility of Ammonium Phosphomolybdate in Nitric Acid . . 591 LXVII. Solubility of Ammonium Phosphomolybdate in Ammonium Nitrate . 591 LXVIII. Effect of Ammonium Nitrate on the Precipitation of Ammonium Phospho- molybdate . .592 LXIX. Composition of Mother Liquid for the Precipitation of Ammonium Phospho- molybdate . 597 LXX. Test Analyses for Sulphur in Pyrites . . . . .610 LXXI. Effect of Foreign Salts on the Precipitation of Barium Sulphate . .611 LXXII. Volatilisation Losses during the Ignition of Barium Sulphate . .616 LXXIII. Test Analyses for Detection of Fluorine . . . . .636 LXX IV. Sulphuric Acid Specific Gravity and Concentration . . . 676 LXXV. Nitric Acid Specific Gravity and Concentration . . . .678 LXX VI. Hydrochloric Acid Specific Gravity and Concentration . . . 679 LXXVII Hydrobromic Acid Specific Gravity and Concentration . . . 680 LXXVIII. Hydrofluoric Acid Specific Gravity and Concentration . . . 680 LXXIX. Phosphoric Acid Specific Gravity and Concentration . . . 680 LXXX. Perchloric Acid Specific Gravity and Concentration . . . 681 LXXXI. Formic Acid Specific Gravity and Concentration .... 682 LXXXII. Acetic Acid Specific Gravity and Concentration .... 683 LXXXIII. Ammonia Specific Gravity and Concentration .... 683 LXXXIV. Atomic Weights ........ 684 LXXXV. The Vapour Pressure of Liquid Water at Different Temperatures in Milli- metres of Mercury . . . . . . .685 LXXXVI. Conversion of Grams of Potassium Chloroplatinate into Grams of Potassium Chloride ........ 687 LXXXVII. Conversion of Grams of Potassium Perchlorate into Grams of Potassium Chloride ........ 695 LXXXVIII. Conversion of Grams of Potassium Chloride into Grams of Potassium Oxide 699 LXXXIX. Conversion of Grams of Sodium Chloride into Grams of Sodium Oxide . 703 XC. Conversion of Grams of Magnesium Pyrophosphate into Grams of Magnesium Oxide ......... 708 XCI. Conversion of Grams of Magnesium Pyrophosphate into Grams of Phos- phorus Pentoxide . . . . . . .715 XCII. Conversion of Grams of Barium Sulphate into Grams of Sulphur Trioxide . 720 XCIII. Conversion of Grams of Lead Sulphate into Grams of Lead Monoxide . 727 XCIV. Solvents for Precipitates in Munroe's Crucible . 730 INTRODUCTION. The Subdivisions of Analytical Chemistry. Qualitative and Quantitative Analysis. The general purpose of analytical chemistry is to find the chemical nature of a given material. With simple substances this is comparatively easy, for the object of the analysis is easily secured by establishing the identity of the given substance with one whose properties are known. A simple inspection of the consistency, colour, or smell may suffice ; in other cases, the specific gravity, boiling point, or freezing point may have to be determined ; and finally, it may be necessary to prepare compounds of the given substance, and compare their properties with those of known compounds. Thus arise the so-called tests of qualitative analysis. On the other hand, if the substance to be analysed be not simple, but is rather a complex mixture of several simple substances, the analytical operations are more intricate, for the properties of the different components in the complex may be masked by the mere presence of the others so as to render identification impossible. In that case, the complex must be resolved into simpler substances which can be separately identified. If the purpose of the analysis be merely identification, the operations are said to be qualitative while if the amount of one or more of the constituents has to be determined, the analysis becomes quantitative. There is a large number of manuals on qualitative analysis in which an elaborate general scheme over fifty years old is described for the identification of the constituents of the most complex inorganic mixtures. In many cases, this scheme is only of pedagogic interest, because, in practice, we generally possess a good idea of the components of a mixture, and only the amounts of the more important of these are industrially important. Hence, although a qualitative analysis should precede the quantitative, yet numerous quantitative analyses are made without the preliminary examination, because the qualitative composition of the given substance is well enough known. Gravimetric Analysis. In the gravimetric analysis, the several constituents are separated and weighed, or, if they cannot be separated, or if they are un- suited for weighing, they are converted into compounds of known composition which can be satisfactorily weighed. The several steps in the operation are : (1) A definite amount of the substance to be analysed is weighed/ (2) The weighed sample is brought into solution. (3) The constituents to be determined are separated from the solution, one by one, in the form of definite insoluble compounds, either by precipitation or by electrolysis electro-analysis. (4) A compound so separated is freed from adherent liquids and solids by nitration and washing. (5) The compound is dried and ignited so as to convert it into a stable pure compound. (6) The compound is weighed. (7) The amount of the substance to be determined is computed by known arithmetical processes from the weight of the ignited compound. Volumetric Analysis. In volumetric analysis, the quantity of a constituent in a given solution is determined by adding a sufficient volume of a standard xxii A TREATISE ON CHEMICAL ANALYSIS. solution containing a known amount of a selected reagent to produce a definite " complete " reaction. The number of operations involved in a gravimetric analysis is here much curtailed ; as a rule, no filtration or washing is needed ; nor is any weighing required after the sample has been brought into solution. The opera- tions are : (1) A definite quantity of the substance to be analysed is weighed. (2) The weighed sample is brought into solution. (3) The standard solution of the selected reagent is added until the whole ot the substance to be determined has been converted into a definite compound. The comple- tion of the reaction is indicated by a change in the colour of the solution, frequently tinted with a selected dye called the indicator. (4) The quantity of the standard solution required to complete the reaction enables the amount of the substance under investigation to be computed by known arithmetical processes. Gasometric Methods of Analysis. If the substance being analysed gives off a gas when it is treated with a selected reagent, such that the volume of gas liberated bears a definite relation to the amount of the constituent in question, the measurement of the volume of the liberated gas, or a determination of the loss of weight which occurs when the gas is all expelled, enables the amount of the constituent to be computed by arithmetic. Physical Methods of Analysis. There are several so-called physical methods of analysis in which the variation of some property say, specific gravity is proportional to the amount of the substance under investigation which is present in a given mixture. A determination of that property enables the amount of substance to be computed. In colorimetry, for instance, the substance under investigation can be made to produce a coloured solution such that the intensity of the colour varies with the concentration. A comparison of the tint of a known volume of the solution with that of another solution containing a known quantity of the same substance as that under investigation, enables the amount of the given constituent to be computed by simple arithmetic. In turbidimetry, instead of comparing the colours, the degree of opacity of the two solutions is compared. This process is of limited application, and is only possible when the solution under investigation can furnish a fine-grained solid precipitate which settles very slowly. The Birth of Analytical Chemistry. The chemists of the phlogiston period Basil Valentine, 0. Tachenius, F. Hoffmann, A. S. Marggraf, C. W. Scheele, etc. 1 made a large number of isolated observations which enabled many inorganic substances to be distinguished from one another. Robert Boyle, 2 about 1661, styled the process of identifica- tion an analysis. About a century later, Bergmann 3 compiled and arranged the different tests in a systematic and methodical way, and thus laid a foundation upon which Berzelius and Rose 4 built a system of qualitative analysis which has persisted with remarkably few changes up to the present day. 1 B. Valentine, Chymische Schriften, Leipzig, 1769 ; 0. Tachenius, Hippocrates Chemicus, qui novissimi salis antiquissima fundamenta ostendit, Venetiis, 1666 (English trans., London' 1677) ; F. Hoffmann, De Methodo Examinandi Aquas Salubres, Leyden, 1708 ; A. S. Marggraf,' Chymische Schriften, Berlin, 1761-7" Method of reducing silver chloride without loss," 1749 "Chemical examination of water,'' 1754 ; C. W. Scheele, Opuscula chemica et physica 'Lipsiae' 1788-9. 2 R. Boyle, The Sceptical Chymist, Oxford, 1661 ; Experiments and Observations uvon Colours, London, 1663. 3 T. Bergmann, Opuscula Physica et Chemica (De minerarum docimasid humidd) Holmise 1780. 4 J. J. Berzelius, De I' Analyse des Corps Inorganiques, Paris, 1827 ; H. Rose Handbuch der analytischen Chemie, Berlin, 1829. INTRODUCTION. XX111 The earliest writings on metallurgy show that from remote antiquity metals have been extracted by heating their ores with appropriate fluxes ; and there can be little doubt that laboratory methods for the dry or fire assay l of ores were modelled upon metallurgical operations conducted upon a large scale. Witness the so-called Cornish process for the assay of copper 2 is conducted very much the same as that which has been employed for the extraction of copper from its ores for thousands of years ; and the origin of the method of separating silver from lead by cupellation cannot be traced, because it is alluded to by early writers 3 Diodorus, Strabo, Suetonius, Pliny, etc. as an old and familiar process. Bergmann's Work. T. Bergmann's brochure upon "the analysis of minerals in the wet way" was published in 1780. In this important essay, Bergmann pointed out the existence of errors in dry or fire assaying processes due to the incomplete attack of the minerals by the fluxes ; to the retention of some of the metal by the slag; etc., and he advocated several wet processes in preference to the dry methods of analysis then in vogue. 4 Bergmann 5 also showed that the amount of a substance in solution was best determined by converting it into a definite compound of known composition, and subsequently deducing the amount of the desired constituent from the weight of the compound so prepared. 6 Thus lime was weighed as oxalate or sulphate ; lead, as sulphate or sulphide ; sulphuric acid, as barium sulphate ; silver, as chloride, etc., although, previous to this, Marggraf 7 had determined the amount of silver in alloys by separating the silver as an insoluble chloride. Bergmann recognised that "the sum of the weights of each of the constituents in an analysis should be equal to the weight of the mineral analysed, allowing for a certain loss during the manipulations," and his analysis of a sample of pure gypsum is singularly good, e.g. : Vitriolic acid. Calcareous earth. Water. Bergmann's analysis . 46 32 22 per cent. Sulphur trioxide, SO.. Lime, CaO. Water, H 2 0. Modern analysis . . 46'5 32*5 21 '0 percent. 1 The term assaying is applied more or less vaguely to methods of chemical analysis which are confined to the determination of the commercially important constituents in given materials. Thus, processes for the assay of vegetable products, drugs, ores, alloys, etc. , are described in current literature. The assay of ores, alloys, and related products can be conducted by treating the material under investigation with suitable solvents so as to get the desired metal in a solution from which it can be subsequently precipitated ivet or solution processes ; or the given material may be fused with a suitable flux so tbat tbe desired substance is reduced to the corre- sponding metal, which collects as a button at the bottom of the crucible, and the associated substances can be separated in the form of a fusible slag dry or fire processes. 2 M. L. Moissenet, Ann. Mines (5), 13. 183, 1858. 3 Thus, in the second century B.C., Agatharchidas of Cnidos described the method employed by the Egyptians for purifying gold, and this resembles the cupellation process for separating silver and lead. 0. S. Pliny, Naturalis Histories, Venetiis, 33. 44, 1472 ; 34. 48, 1472 ; G. Agricola, De Re Metallica, Basiliae, 1546 (English trans., London, 1913) ; A. Libavius, Ars probandi Mineralia, Francofurti, 1597 ; A. Libavius, De Judicio Aquarum Mineralium, Franco- furti, 1597. The two last-named essays are usually stated to be the first books specially devoted to the analysis of minerals. The first-named is considered to be largely compiled from Agricela's book. See also G. Chesneau, Revue Scientifique (5), 3. 321, 357, 1905. 4 A. P. T. Paracelsus (Opera Omnia Medico- Chemico-Chirurgica, Geneva, 1658) is generally credited with having first described an analysis in the wet way. To analyse an alloy of gold and silver, Paracelsus treated the sample with nitiic acid atjua fortis which dissolves the silver and leaves the gold behind in the form of an insoluble black powder. He then precipitated the silver by inserting a plate of copper in the solution. The copper is at tbe same time vigorously attacked by the free acid. 5 T. Bergmann, Opuscula Physica et Chemica (De minerarum docimasid humida}, Holnnae, 349, 399, 1780. 6 In 1 755, J. Black (Experiments on Magnesia alba, Quick-lime, ami other Alcaline Substances, Edinburgh, 1777) estimated the amount of magnesia alba in a solution by the addition of sodium carbonate and weighing the resulting precipitate. 7 A. S. Marggraf, Mem. Akad. Wiss. Berlin, 16, 1749. Xxiv A TREATISE ON CHEMICAL ANALYSIS. Bergmann also introduced the method of decomposing silicates by fusion with alkaline carbonates, and his little treatise, De minerarum docimasid humidd, is thus considered to have inaugurated a new era in analytical practice. Klaproth's and Vauqueliris Work. The next advances were made by Klaproth 1 and by Vauquelin. 2 These workers introduced many improvements in technique and analysed many minerals. Klaproth, for instance, ignited his precipitates before they were weighed in cases where the ignition was not attended by decomposition. Quantitative analysis was greatly stimulated by the need for separating the different constituents of the minerals from one another, during the many investigations on the composition of minerals made about that time. As Klaproth expressed it : " Nature, inexhaustible in her riches, has intended to keep in activity the ardour of the naturalist in the examination of mineral substances." As early as 1808, we can recognise the beginnings of the present-day method of conducting silicate analyses in John's book, 3 which summarised the labours of Bergmann, Marggraf, Klaproth, and Vauquelin. Berzelius' Work. About 1827, Berzelius 4 took up the work, and considerably advanced the art by devising a number of new methods for the determination and separation of many elements, and testing those then in use ; Berzelius also introduced a number of improved methods in manipulation : e.g., the hydro- fluoric acid process for decomposing silicates ; the separation of precipitates by filtration through "filter paper," followed by subsequent ignition ; and, by using small quantities of the material to be analysed, he reduced the errors of mani- pulation, because small precipitates are more readily cleaned than large ones. Berzelius' pupils, F. Wohler and H. Rose, worked up the experience of their teacher with their own in two books : H. Rose's Handbuch der analytischen Chemie, Berlin, 1829, and F. Wohler's Praktische Uebungen in der chemischen Analyse, Gottingen, 1853. The later editions of these books are consulted even to-day. About this time also appeared the first edition of Fresenius' famous book (1846), and this was followed by the founding of the Zeitschrift fur analytischen Chemie in 1862. Thus the modern practice of the art is linked with that of its founders. Some Uses of Analytical Chemistry. The information which the analysis is expected to furnish, as indicated on pages 242-3, determines the method by which it is to be conducted. The in- formation required may be purely of scientific interest, or it may be essentially practical. The utility of analyses for purely scientific purposes need not be here scrutinised. It would be easy to quote many examples where analysis has thrown a new light on the aspect of chemistry. 5 Analysis is an indispensable auxiliary to the mineralogist, for, as C. R. Fresenius (1875) has said, it teaches him the true nature of minerals, and suggests to him principles and rules for their recognition and classification. H. S. Washington (1903), 6 too, says that analyses are no longer ornamental adjuncts, but essential parts of most petro- logical publications, on which most of the discussion hangs, and from which the most important conclusions are drawn. 1 M. H. Klaproth, Beitraye zar chemischen Kenntniss der Miner alkorper, Freiberg, 1795. 2 L. N. Vauquelin, Scherer's Journ., 3. 410, 1799. a J. F. John, Chemische Labor utorium, Berlin, 181, 1808. J. J. Berzelius, De V Analyse des Corps Inorganiques, Paris, 1827. 5 The work of C. F. Wenzel (1777) and of J. B. Richter( 1791-1 802) on chemical equivalents ; the work of H. Cavendish (1784) on the composition of water ; of C. W. Scheele (1779) on the composition of air ; and the 1800-1808 controversy between C. H. Berthollet and J. L. Proust on chemical combination, illustrate the debt chemical theory owes to analytical practice. 6 C. R. Fresenius, Quantitative Chemical Analysis, London, I. 4, 1876 ; H S Washington Prof. Paper U.S. Geol. Sur., 14. 13, 1903. INTRODUCTION. XXV In industrial work, the purpose is essentially utilitarian, for the analysis is generally directed to finding the composition of commercial materials, and to answering such questions as : (1) Does the material to be purchased correspond with the seller's description ? (2) Is the material suitable for the purpose for which it is required 1 (3) Is the material really worth the quoted price ? The Real Value of Rough Working Tests. The practical potter usually tries how a small sample of the material to be purchased behaves in his oven, and if it satisfies that test, he accepts the bulk, believing with blind faith that all is right. It does not need the subtilis diabolus, referred to on page 127, to deceive him, for, when it is worth while, it is comparatively easy to devise adulterations for most potter's materials which will make satisfactory trials. Although this simple test, invaluable in its way, may show (1) whether or not any deleterious impurity is present, or (2) whether the material is or is not capable of doing the work for which it is intended ; yet (a) it gives no information as to the presence or absence of inert admixtures ; otherwise expressed, it does not say whether the material offered for purchase really corresponds with the seller's description ; and (b) it gives no idea whether the material is really worth the price asked. Impurities in the Raiv Materials. An intimation to the vendor that the material may be tested is not sufficient security for the buyer. I am told that "of the samples of materials offered to a Government Department, approximately forty per cent, are either adulterated or of very inferior quality." Price is no criterion of the quality, for it is said that the most inferior samples were some- times quoted at the highest prices. In such cases, analyses and tests are the purchaser's only safeguard, and the cost of the analysis in large deals may be recompensed a thousandfold. Four samples of tin oxide from four different firms were submitted for analysis. The chemist reported that three of the samples were very nearly the same, and contained approximately 99'5 per cent, of stannic oxide, but the fourth sample contained 5 per cent, of " combined water," and ^th per cent, of tungstic oxide as impurity. The potter's trials, however, showed that the cheapest and least pure sample gave the richest opaque glaze ! The first three samples were approximately 228 per ton, and the fourth sample 222 per ton. The potter subsequently obtained a rebate on account of the 5 per cent, of combined water, and the fact that the effective tin oxide in the sample A cost 229'15 per ton, and in the sample B, 240 per ton ; and learned incidentally that he might be able to do with rather less stannic oxide than he had previously employed when tin oxide was not so expensive. The Adulteration of Raiv Materials. So long as human nature is what it is, so long will adulteration be carried on with a view of making large profits, and of cheating the purchaser into paying too high a price for his materials. Caveat emptor let the buyer beware. A glance through the records of any testing laboratory shows that successful adulteration is one of the fine arts. Certificates of analysis of samples selected by the vendor are commonly presented as if some mysterious virtue resided in a material with such* a certificate appended. The sample placed in the analyst's hands may or may not represent the material sold by the dealer. The onus of proof that the certificate really represents the materials received by the purchaser, rests with the vendor (see page 127). The preceding example tin oxide was not a case of deliberate adulteration. The next case is not so clear. A traveller offered a sample, A, of white lead at 28 per ton (5 per cent moisture), another traveller offered another sample, B, at 25 per ton (5 per cent, moisture). The cheaper sample gave the better-looking frit when fused with a given proportion of ground flint in a biscuit cup in the usual way. Analysis showed that the cheaper sample B contained 69'0 per cent, of PbO, and 15 per cent, of barium carbonate, whereas the dearer sample A contained 81 '6 per cent, of PbO, and no barium carbonate. Ground witherite (94-96 XXVI A TREATISE ON CHEMICAL ANALYSIS. per cent, barium carbonate) is worth about 10 per ton. The effective PbO in sample A would cost 34-31 per ton, and in sample B, 36*23 per ton. Hence, the dearer sample is the cheaper. The improved colour of the frit from the cheaper sample B was due to a known secondary effect of the barium salt which is of no significance when the material is used in bulk. Standardising the Products of a Factory. Analysis also furnishes a valuable means of keeping the finished products of a factory up to standard. Suppose that the body and glaze be analysed when everything is at its best, and the results be kept as standards. The standards are then very convenient for reference when substituting fresh clays or other raw materials, in order that the type of the new body may be kept as nearly as possible like the old one, so as to lessen the risk of complications with the glaze, etc. In many cases, too, the cause of those perplexing faults which sometimes crop out, and which are usually attributed to the miller or to wrong mixtures, can be unerringly located and promptly rectified by a comparison of the analysis of the defective body or glaze with the standard. A fresh batch of glaze turned out highly unsatisfactory. The glaze-mixer affirmed that he had mixed the glaze precisely as he had always done before, and it was found that the defective glaze had received no different treatment from that regularly employed on the works. The miller was accordingly blamed for the fault. He sent a sample of the defective glaze along with a sample of good glaze from the previous mixing for analysis. The analyst reported that the good glaze contained the equivalent of 24*5 per cent, of lead oxide, PbO, and the bad glaze but 5'5 per cent. The proportions of the other constituents were the same in both glazes. The fault was obviously due to a wrong mixing, in spite of affirmations to the contrary. When the proper proportion of white lead, computed from the analysis, was added to the defective glaze, it proved quite satisfactory. The Imitation of Commercial Products. In the Sturm und Drang of commercial warfare, it is not only necessary for a firm to test its own materials and products, but the "Intelligence Department" must also keep a watchful eye on the products of rival manufacturers. A manufacturer has for years been supplying the trade with, say, an opaque glaze made from an old recipe a legacy from the founder of the firm. For perhaps half a century there has been no change in quality for better or for worse, and the old recipe is the real master of the works. The traveller reports that he is seriously menaced by the success of a foreign glaze which is considerably better than his own and quite as cheap. Faith in the long-cherished recipe is shattered. After much worry and expense, trial and failure, the manufacturer is inclined to give up ignominiously beaten. As a last resource, he sends a sample of the foreign glaze to the expert analyst, requesting a recipe. The chemist recognises that it is not much use trying to reproduce the unknown This may seem rather a low estimate of the function of an analyst, because a properly equipped staff would not be satisfied with its own old recipes. There would be no standing still. A firm would not wait until its position on the market was jeopardised or made untenable by a rival, but it would be continually experimenting, not directly to imitate its competitors, but rather to beat its own products year by year. Some Limitations of Analytical Chemistry. How to Treat the Analyst. Some have an extraordinary notion of the resources of analytical chemistry, and inquiries are made for all manner of impossible things. A bottle containing about 100 c.c. of water was sent for INTRODUCTION. XXV11 analysis, and the contents smelt strongly of a patent medicine ! The subject of sampling is so important that a special chapter has been devoted to it pages 127 to 140. When sending materials to the analyst, some consider it best to keep back information concerning the nature and origin of the sample so as to prevent a biassed report. The general effect of this procedure is to render the analyst's task more difficult and costly^ and to hamper his usefulness. Owing to the need for completing the analysis in a reasonable time, and at a specified fee, an exhaustive search for every improbable constituent is seldom made. An analyst working under normal conditions would certainly be excused if a clay with an abnormal amount of lithia were reported with soda in place of the lithia. 1 An example is indicated on page 141. This recalls the fact that the presence of arsenic in beer was not suspected until the 1900 epidemic of arsenical poisoning. Consequently, if the amount of any unusual constituent is desired, this should be specified. The Fallibility of the Analyst. Perhaps something of the feeling of awe and wonder which prevailed in the minds of the vulgar towards the alchemists of old, survives to-day in the popular concept of the analytical chemist. Many apparently take it for granted that he is gifted with keener and more occult powers than his fellows. As a matter of fact, no one should be more conscious of his own limitations than the analyst himself, for he is continually humiliated and shamed by the fallibility of his own tests, and by the resulting instability of his opinions. Analyses have been published which emphasise in a remark- able manner how men exceptionally expert in one field of analysis fail ludicrously in an unfamiliar field. Not very long ago, a report on china clay by an excellent county analyst " went the rounds." The report had a highly improbable number for the amount of alkali in the clay, and stated that the alkalies were rather low, but the defect could be remedied by the addition of soda ! In some cases, it is difficult to believe that the " reputed " analyses have been obtained other than by a process of guessing. Not infrequently a manufacturer has sent out samples to different chemists, and had different results returned. He may have asked for explanations, and decided in future to avoid the chemists as much as possible, concluding, as one expressed it, that "chemists are bluffers." Unless two chemists are able to analyse the same sample with results acceptable to both buyer and seller, it is difficult to see what other conclusion the commercial man can draw. The utility of analyses in some of the industries has been discredited by ignorant, slovenly work which would be ludicrous were it not so pathetic. Faulty analyses have been discussed in several parts of this book 2 e.g., pages 222, 248, 365, 610, 1 In this connection it is interesting to note that C. F. Plattner (Pogg. Ann., 69. 443, 1846) was not able to make his analysis of the mineral pollux (from Elba) add up to 100 per cent., and he sought in vain for the missing element. After R. Bunsen and G. Kirchhoff (1860) had discovered caesium, F. Pisiani (Compt. Rend., 58. 714, 1864) showed that Plattner had mistaken csesium (atomic weight 132 '8) for potassium (atomic weight 89*1). By making the correspond- ing correction, Plattner's analysis was found to be quite satisfactory. To make this quite clear, assume that 5 grams of a compound, supposed to be potassium chloride, are obtained. This will be multiplied by 0'631 to get the equivalent amount, 3'16 grams of K 2 ; but if the compound be CsCl, not KC1, then the weight must be multiplied by 0'835 to~get the corre- sponding amount, 4'18 grams of Gs 2 0. The analysis would thus appear to be 4'18 less 3'16, that is 1'02 grams too low if the 5 grams of csesium chloride were mistaken for potassium chloride. This is a remarkable tribute to the accuracy of Plattner's analysis (J. W. Mellor, Modern Inorganic Chemistry, London, 356, 1912). Lithium has a smaller atomic weight than sodium, and accordingly, if much lithium is assumed to be sodium, the analysis will total too high. 2 A recent discussion on the presence of titanic oxide in fireclays can be cited as an illustration of the value to be attached to some brick analyses even by some who say they have analysed " hundreds of bricks," but who have apparently not troubled to find what methods are used by those who have specialised in producing accurate results. A. R. Myhill, Gas World, 58. 299, 364, 1913 ; F. Harvey, t'6., 58. 323, 433, 1913 ; K. C. Orr, 58. 365, 1913 ; Anon., ib. t 58. 402, 1913. XXVlii A TREATISE ON CHEMICAL ANALYSIS. and 674. What H. S. Washington (1903) 1 has said of rock analyses in general applies even more emphatically to some published work on clays : There is a tendency to place implicit confidence in the results of analytical work to accept readily whatever figures the analyst may furnish, with scarcely ever an attempt at a critical estimate of the worth of the analysis. It seems to be taken for granted that the analyst, like the proverbial king, can do no wrong. This applies not to the personal good faith of the analyst, but to the analytical processes which, possibly because they belong to one of the exact sciences, are for the most part tacitly assumed to be infallible. In few cases does there seem to be any recognition of the difficulties and uncertainties of analytical work. The Fallibility of the Analysis. Analytical operations are not performed with an automatic instrument which gives the composition of a substance with machine- like precision. The presence or absence of each constituent has to be established by special tests, and the amount determined by a number of operations each of which introduces a small error into the final result. The limits of error in the more frequently conducted operations are fairly well known, and to those familiar with that type of work, the .precautions needed to reduce the errors to a minimum are a part of their routine practice. With those substances which are not analysed so frequently, the errors are probably somewhat greater, for the disturbing factors are not so well known. At best, however, few analytical operations are altogether free from error and vexation of spirit, and the work is done more or less blindly, quite in ignorance of the why and the wherefore. Standard Methods of Analysis. When different chemists present conflicting results, discord is sure to arise in settlements between buyers and sellers, and in the control of works processes. As a result, many corporations have been driven to insist on the use of certain uniform methods of analysis in testing the materials in which they are interested. The idea has been taken up seriously by a number of chemical societies, and there are now quite a number of so-called " Official or Standard Methods of Analysis " for particular substances e.g. fertilisers, food-stuffs, etc. (vide page 250). The idea is to secure concordant analyses, to reduce errors to a minimum, and to place commercial analytical practice on a higher plane. The proposed methods are not fixed for all time, but are revised periodically so as to substitute improved methods when such are available. Standardised Samples. In 1905, the American Foundry man's Association prepared with great care a set of samples of iron, had them analysed by three or four chemists of recognised ability, and sold portions as " standardised samples " at a reasonable price. The idea has since been further elaborated by the Bureau of Standards, Washington, U.S.A., from which "standardised samples" with detailed certificates of analysis can be obtained. The work has up to the present been largely restricted to alloys and ores of various kinds. 2 Standardised samples of clay can be obtained from the County Pottery Laboratory, Stoke-on-Trent, at 10s. 6d. per 100 grams. Some of the uses of standardised samples are as follows : (1) An analyst unfamiliar with a particular analytical process can practise with the analysed samples until he is perfected. The standardised samples also furnish useful checks for advanced students of quantitative analysis. (2) The applicability and accuracy of a new or "improved" method of analysis can be determined with comparatively little effort by working with standardised samples. 1 H. S. Washington, Prof. Paper U.S. Geol. Sur., 14. 14, 1903. ' 2 Circular Bur. Standards 25. 3, 1912; 26. 3, 1910; 40. 3 1912- W F Hillebrand Journ. Ind. Eng. Chem., I. 41, 1909; W. C. Ebaugh, ib., i. 63 1909- L C Jones ib i' 269, 1909 ; W. D. Richardson, ib., i. 5, 1909. INTRODUCTION. xxix (3) In cases of disputed results, owing to the use of faulty methods of analysis by one of the chemists, both parties can analyse a standardised sample, and find who is at fault before the case comes into court. This seems more satisfactory than submitting the original sample to a third party as umpire, because -the umpire may be no more capable than the contending analysts. The Evolution of Modern Analytical Practice. How Analytical Processes have Grown. A comparison of the classical text- books of H. Rose (1829), C. F. Rammelsberg (1845), C. R. Fresenius (1841-6), C. F. Mohr (1855), and L. E. Rivot (1861-6) with those published in more recent years shows that the development of the art has been painfully slow. Many, perhaps most, of the standard processes are purely empirical or rule-of- thumb. The original methods devised by the fathers of analytical chemistry gave more or less approximate results ; with increasing experience, these processes were modified now liere, now there until methods were evolved which would furnish results accurate within the limits of experimental error tolerated in practice. The methods now in vogue for dealing with the so-called rare earths are in the earlier stages of the evolutionary process ; while the elaboration of the regular methods for the determination of phosphorus, magnesia, sulphur, and potassium ; the chromate process for the separation of barium and strontium ; the sulphide processes for the separation of antimony and tin, and for the separation of zinc ; and the well-known basic acetate process, all bear eloquent testimony to the adventitious and empirical way the art has developed. Much work has been directed to developing speed without sacrificing accuracy, and towards determining one or two constituents in a mixture, and ignoring the others. Hence, some industrial routine processes have been abridged (mechanicalised) into a rigid code of instructions such that boys or girls "testers," as they are called, of no special education or preparation, can be taught in a short time to perform the necessary operations and get good results. The Substitution of New Processes. Owing to the curious evolution of the processes used in analytical chemistry, analysts are reasonably reluctant in introducing new methods. It is not wise to substitute comparatively untried processes however promising they may appear without very careful considera- tion. The dread of incorporating unknown errors in our work begets caution, And makes us rather bear those ills we have Than fly to others that we know not of. In an old, long-tried process, the disturbing factors and the precautions necessary for accurate work are fairly well known, whereas with a novel process much has to be learned in humouring its little idiosyncrasies in order that it may furnish the best results. In spite of this conservatism, several new processes seem to be gradually ousting many of the older methods, and winning tbeir way into general practice. The reagent list is accordingly extending to include hydro- xylamine, hydrazine, nitroso-/?-naphthol, a-dimethylglyoxime, nitrosophenyl- hydroxylamine, etc. Indeed, many separations can be expedited by some of the newer methods, which are accordingly often recommended in preference to the older processes. Class Exercises. The need for humouring the different analytical processes is seldom taught in the schools. The beginner, working through routine class exercises with fairly pure substances, or with commercial materials whose com- position has not been checked, gets little or no inkling of the difficulties which dog the footsteps of those who apply these methods to industrial products under industrial conditions, working against time. The early simple exercises are of course indispensable for training the hands of the neophyte in the operations of XXX A TREATISE ON CHEMICAL ANALYSIS. weighing, filtration, washing precipitates, etc. ; and it is here assumed that these exercises have done the work which is expected from them, so that we can take up our studies presupposing a certain amount of knowledge of general and analytical chemistry, and more particularly of qualitative analysis. In Part 1. of this book, however, the instructions for the fundamental operations have been revised in order to introduce hints and " tips " which are often left to be " picked up," if at all, in a random, haphazard way. A study of experimental errors is also of fundamental importance to an analyst who takes an interest in his work, and who is something more than a mere tester. The Variety and Diversity of the Work in the Silicate Industries. The methods of analysis for clays and related silicates are fairly general, and require little or no modification from sample to sample ; it has therefore been possible to give a detailed scheme for their analysis in Part II., and this is used as a standard of reference for the remainder of the book. The ^case is very different with glazes and enamels, for it is not often that two consecutive analyses can be conducted by exactly the same method ; the change in method "from sample to sample is still more pronounced in dealing with colours. Some of the materials sub- mitted for analysis involve extremely difficult separations, and a special method must be devised to suit each case. Sometimes, too, the report of the analysis has to be accompanied by a working recipe, and, with the more delicate colours, this is a severe test of the accuracy of the work. In our laboratory, for example, an enamel containing arsenic, antimony, tin, and fluorine was once followed by a black colour containing iron, chromium, manganese, cobalt, and zinc ; and this, in turn, by a pink colour containing gold, silver, tin, and lead. In each case, these elements were accompanied by the ubiquitous silica, alumina, alkalies, and alkaline earths, and, in three cases, by boric oxide. There was also an inquiry for an examination of a faulty oleaginous platinum lustre for elements of the platinum group other than platinum. Unfamiliar Operations. Although the proportion of analyses of complex mixtures of the kind just indicated to analyses of clays, bricks, and related materials, which are required from a laboratory which specialises in this kind of work, averages over one in thirty, yet it would be a mistake to cut down the directions for the analysis of the less frequently occurring substances because of their rarity. On the contrary, fuller minutife are then required, because, the less familiar the road, the greater the need for guide-posts and danger signals, and the greater the probability of "accidental errors." So many cases have arisen where schemes for the analysis of unfamiliar mixtures are not included among those given as types in text-books on quantitative analysis, that much time must have been wasted in devising feasible processes, even when a good library is available. It is quite impracticable to describe methods suited to every possible or likely case which might be required in practice ; nor did it seem to me so expedient to give detailed schemes for a few mixtures as to take an imaginary mixture, more complex than would obtain in practice, and indicate the short-cuts to be -made when the qualitative analysis shows that the conditions are favourable. It is thus nearly always possible to curtail the general schemes pages 319, 387, 439, 509 to suit particular cases, of course bearing in mind the hints given on page 278. The Theory of Analytical Operations. The technique of analytical chemistry has been worked out without much aid from theory, for the classical text-books have merely elaborated details of manipulation necessary for exact results. In 1894, W. Ostwald demonstrated the important bearing which the theories of physical chemistry have upon the practice of analytical chemistry. Opinions may differ very much as to the function of the ionic hypothesis in analytical chemistry, and whether anything is really gained by describing the facts of an INTRODUCTION. XXXI essentially practical art in the language of a very hypothetical doctrine. Ex- cluding this hypothesis, perhaps the most important ideas derived from physical chemistry which have tempered analytical operations are : (1) The theory of adsorption (pages 97, 179, 611, etc.). It seems to be impossible to wash a precipitate perfectly free from adsorbed liquids or solids, although it is usually possible to make the error so introduced negligibly small. (2) The laws of the <(>//(>/( l (>/' iiHitfcr (pages 96, 212, 275, etc.). When an "insoluble" pre- cipitate separates from a solution in the absence of dissolved electrolytes, it is frequently in the colloidal condition, and it cannot then be isolated by filtra- tion and washing. An electrolyte must be added in order to get the substance in a condition suited for treatment. (3) The theory of incomplete reactions (pauvs 181, 272, etc.). The reactions which result in precipitation and neutralisation are not usually complete. Instead of running to an end, the system takes up a state of equilibrium between the initial (or solution) stage and the final (or precipitate) stage, and a certain amount of the element under investigation escapes precipitation. Perfect separations by precipitation are not therefore possible, and consequently it is desirable to study each process with the object of finding the conditions necessary to make the precipitation complete enough for practical requirements. This is partly what is meant above by 11 humouring an analytical process." PART I. GENERAL. CHAPTER I. WEIGHING. i. The Balance. THE balance is one of the most important instruments used by the analytical chemist. It is essentially an instrument for comparing weights. The object of weighing is to compare the quantity of matter in a given substance with the quantity of matter in a standard substance the weight or weights. The com- parison is made on the balance by suspending the object to be weighed at one end of a beam supported at the middle by a pillar. The weights are suspended at the opposite end of the beam. The beam is virtually a kind of lever, and the mechanical theory of the balance is mainly founded on the properties of levers. 1 The beam of the balance rests on a central knife-edge, generally of agate, which presses against a plane agate bearing fitted to the beam. This arrange ment acts as the fulcrum of a lever of the first class. Two pans for supporting the masses to be compared are suspended from stirrups, each of which has an agate plate which rests on a knife-edge fixed at each extremity of the beam. Each arm of the balance is so graduated that pieces of wire* 2 /jd^rs ,of knovyii weight can be placed on the beam at any required distance' from, tlie cential knife-edge. The rider is used instead of weights below a centigram 0*01 grm. If the three knife-edges press continually against their agate bearings, they soon become blunted, and wear furrows in the bearings. When this occurs, the balance is inaccurate, and must be laid up for repairs. In order to prolong the life of the knife-edges and bearings as much as possible, every balance should be provided with an "arrest," which separates the three knife-edges from their bearings when the balance is not in use. The "arrest" is usually worked from the front or left side of the balance. Great pains are taken by the makers to reduce the friction upon the knife-edges to a minimum. When the balance shows signs of stiffness in the motions of beam and pans, the fault should be investigated at once. The defect may be due to an accumulation of dust between the knife-edges and their bearings ; to the rusting of the knife-edges, if made of steel ; to the blunting of the knife-edges ; or to the wearing of furrows in the bearings. To prevent dust accumulating on the knife-edges, etc., the balance is enclosed in a glass case. Small movements of the beam are scarcely perceptible, and consequently the beam is provided with a long pointer which multiplies the rotational displace- ment. When equilibrium is established, the pointer rests in front of the zero of a scale cut in a piece of ivory. Instead of taking the centre of the scale as 1 The theory of the balance is usually discussed in text- books on physics. The student is referred to J. Walker, The Theory and Use of a Physical Balance, Oxford, 1887 ; E. Brauer, The Construc- tion of the Balance, London, 1909 ; W. S. Aldis, Trans. Newcastle Chem. Soc., 3. 151, 161, 1876. 2 The rider is said to have beendevised by J. J. Berzelius P. Schwirkus, Zeit. Instr., 6. 42, 1887. 3 A TREATISE ON CHEMICAL ANALYSIS. "zero," it will be found more convenient, later on, to number the scale to 20, starting from the left, as indicated in fig. 6. In that case, "10" will be the zero" the position of rest of the balance. There are several other adjusting devices and conveniences, which vary with different balances made by the same or different manufacturers. 1 Whatever balance be selected, the operator must make himself pertectly familiar with its peculiarities. The conditions which must be satisfied by a good balance are : (1) The balance must be consistent. It must give the same result in FIG. 1. Sartorius' balance. successive weighings of the same body. This condition depends upon the accuracy of the knife-edges. (2) The balance must be accurate. The beam must be horizontal when the pans are empty, and when equal masses are placed on the pans. This condition depends upon the equality of the two arms. To eliminate any error arising from the inequality of the arms of the beam, see page 21. (3) The balance must be stable. The beam after being displaced from its horizontal position must return to its horizontal position. This condition depends on the adjustment of the centre of gravity ; see below. 1 A balance suitable for analytical work may be obtained from 8 upwards. 10 will purchase an excellent balance. The eight-guinea balance of Sartorius, represented in fig. 1, is quite satisfactory for general analytical work, but other makers' balances, as cheap, may be equally good. The mechanism under the floor of the balance can be protected from dust, etc., by a wooden cover at a cost of two or three shillings more. Among the more important makers are F. E. Becker, Rotterdam ; P. Bunge, Hamburg ; L. Oertling, London ; A. Rueprecht, Vienna ; F. Sartorius, Gottingen. These and other makers' catalogues might be consulted before purchasing. If possible, get a balance of nearly constant sensibility. The weights will cost 25s. or more. A cheaper balance and set of weights are also necessary for weighing reagents, etc. , where great accuracy is not needed. WEIGHING. 5 (4) The balance must be sensitive. The balance must show any inequality in the two masses on the scale pans even when the differences are small. (5) The balance beam must oscillate quickly. The time taken for an oscillation of the beam, as indicated by the pointer, should be as small as possible, in order that the weighing may be done quickly. 2. An Outline of the Theory of the Balance. Let AOB, fig. 2, represent the beam of a balance, 1 and let the beam be so fixed that it moves freely about the axis or central knife-edge 0. Let A and B represent terminal knife-edges. Let the centre of gravity G of the system be below the axis O, 2 and let W denote the weight of the beam, and let the pans be loaded with equal weights P. Suppose an excess of weighty be placed on the right pan. The balance will then turn through a certain angle a so that A passes to A', B to B\ and G to G'. The weight of the beam now acts at G', and tends to bring the beam back to its horizontal position. In consequence of the resistance of the axis 0, the two weights P neutralise one another, and the weight p applied at B' and the weight W applied at G' remain. These two forces are parallel and in equilibrium FIG. 2. Theory of the balance. Then, from the about the axis 0, so that their resultant passes through 0. principle of levers, W.G'R=p.B'L. Let I denote the length of each arm, AO and OB. Then, B'L = l cos a. Let OG = OG' = r. Then G'R = r sin a. Hence, Wr . sin a =pl . cos a. Or, Wr tan a.=pl. This formula embodies the theory of the sensibility of the balance. When the deflections are small, a may be written in place of tan a. 3 Hence, the Angle of deflection = -^ Wr 1 It is here assumed that the central knife-edge is in the same straight line as the terminal knife-edges. If not, it can be shown (1) that the sensibility of the balance is dependent on the load ; and (2) that the balance is liable to take up a position of unstable equilibrium. Some of the more elaborate and expensive balances provide for the adjustment of the three points of sus- pension terminal and central knife-edges. M. Thiesen, Zeit. Instr., 2. 359, 1882 ; 3. 81, 1883 ; F. Sartorius, ib., 2. 385, 1882. 2 If the centre of gravity G be above the central knife-edge 0, the beam will be in unstable equilibrium, and the slightest addition of weight to either scale pan will lead the beam to turn upside down. The balance is then said to be "cranky."'' If G coincides with 0, the beam will be in a state of neutral equilibrium and it will rest in any position. 3 It is convenient to measure the angle in terms of the divisions of the scale (fig. 6). This is sufficient for practical requirements M. Thiesen, Trav. Bur. Int. Poids. Mes., 5. 2, 1886 : G. Schwirkus, Zeit. Instr., 7. 41, 83, 412, 1887. For a discussion on this subject, see A. Bachinskii, Zeit. phys. chem. Unterrickt., 24. 24, 1911 ; 0. Hartmann and Pforzheim, ib t 24. 93, 1911. 6 A TREATISE ON CHEMICAL ANALYSIS. From this relation it follows that the angle through which the balance turns for a given difference of weight p depends upon (1) The length of the beam, I. The longer the beam, the greater the sensibility. (2) The weight of the beam W. The lighter the beam, the greater the sensibility. (3) The distance, r, between the central knife-edge and the centre of gravity. The closer the centre of gravity to the central knife-edge or point of suspension of the balance, the greater the sensibility. The position of the centre of gravity is regulated by the " gravity bob " (page 14). Obviously, it is necessary to make a compromise between these opposing qualities. For instance, if the arms are lengthened, the beam is made heavier. 1 From the theory of the compound pendulum, it can be shown_that the time required for a single small oscillation of the beam is equal to C*JS, where C is a constant for a given balance carrying a particular load, and S represents the sensibility of the balance. Hence, for a given load, the time of oscillation varies directly as the square root of the sensibility. It is important to make the period of vibration as short as possible, in order that too much time may not be consumed in weighing. The time of oscillation may be taken to represent the stability of the balance. Consequently, high stability can only be obtained by sacrificing sensibility. 2 Hence, the greater the distance of the centre of gravity G from the axis 0, and the longer and heavier the beam, the longer the time qf oscillation and the less the stability of the balance. Here again we are confronted with opposing qualities. The balance must be adjusted so that the length, strength, weight, and location of the Centre of gravity of the beam may produce the maximum sensibility, stability, rapidity, and strength requisite for a specific purpose. This is a problem for the manufacturer, and need not be discussed further. The solution of the problem has led to the manufacture of two classes of balance short-beam and long-beam. 3. The Location, Care, and Use of the Balance. If the balance be in danger of vibrations, shocks, or jars while in use, it should be placed on a firm foundation either on solid masonry built from the ground, or isolated from the floor vibrations by resting it on a shelf fixed to heavy brackets against the walls, and not with legs resting on the floor. The balance should be adjusted perfectly horizontal by means of levelling screws outside the balance case and a spirit level or plumb bob inside the case. The balance should be located so that it is not likely to be heated unequally, 3 1 And the beam also loses rigidity. If the terminal knife-edges are above the central knife- edge, and the beam is slightly bent by a load, there will be an increased sensibility with increase of load -until the three knife-edges are in the same straight line. After that any further lower- ing of the terminal knife-edges will diminish the sensibility (B. S. Proctor, Trans. Newcastle Chem. Soc., 3. 183, 1876 ; Chem. News, 34. 14, 1876). The beam must be made as light and as inflexible as possible. Hence, magnalium and aluminium beams, made in the form of light girders, are common. 2 Time is wasted in protracted weighings, and there is a great danger of an absorption of moisture by certain powders during the operation. 3 If one arm be 1 hotter than the other, the error introduced in a weighing amounts to about O'OOl grm. on a 50-grm. load, and proportionally less for a smaller load. W. H. Miller (Phil. Trans., 146. 753, 1856) detected a difference of '00001 mm. in the thermal expansion of the two arms arising from a change in the temperature of the room. The arms expanded unequally owing to a difference in the quality of the metal forming the beam. T. Middel, Drude's Ann., 2. 115, 1900 ; P. Schwirkus, Zeit. Instr., 7. 42, 1887 : J. J. Manlev Trans Roy >Sor., 210. A, 387, 1910. WEIGHING. 7 and therefore it should not be placed near a door frequently opened, nor adjacent to a stove, hot-water pipe, window, or ventilating shaft. If the location be not suitable, the " zero point " of the balance will be continually changing owing to the unequal expansion of the arms. If possible, the balance should be kept in a separate room. If a small room cannot be partitioned off from the laboratory, a second glass case covering the balance case proper may be needed. Keep the mechanism free from dust. Hence, the balance case must not be opened in a dusty room. If artificial light be employed, it should be above and at the back if possible, over the right shoulder of the operator. If there be any likelihood of the atmosphere in the vicinity of the balance being contaminated with acid fumes, lime and alkaline carbonates should be kept in the balance case to neutralise their effects. Calcium chloride is frequently kept inside the balance case in glass vessels made for the purpose, 1 or in a funnel, which in turn rests in an Erlenmeyer's flask. The idea is to keep the atmosphere inside the case dry. The remedy is by no means effective, as will be shown later. A mixture of fragments of quicklime and granulated calcium chloride is good. Some object to the use of concentrated sulphuric acid as a desiccating agent in the balance case owing to the fumes (sulphur dioxide) which are given off when organic dust collects in the acid jar. 2 The operator should sit in front of the balance case, so as to avoid errors due to parallax in the reading of the pointer. Move the release and arrest screw with a slow, steady motion, not in jerks, since a jerky motion injures the knife- edges. The mechanism should work smoothly, and the "pan arrests" should touch the pans when the beam is lowered. When the beam is raised, the pointer should swing at equal distances, or very nearly equal distances, on each side of the zero. If not, the balance pans should be brushed free from dust, etc., by means of a camel-hair brush. 3 It may, however, be necessary to restore the equilibrium of the beam by adjusting the proper screws. 4 The beam and pans must always be arrested before any change is made in the load or weights on the pans. The object to be weighed and the heavy weights should be placed in the middle of their respective pans, since a heavy load near the edge of the pan sometimes causes troublesome oscillations not easily stilled. The balance case should be closed while weighing with the rider, so as to avoid currents of air. If the beam does not swing when released, it may be set in motion by using the hand as a fan to waft a gentle puff of air on to one of the pans. 5 There is a " trick " in lowering the beam supports so that the oscillations of the pointer will have the required amplitude. The object to be weighed should neither be (1) hotter, nor (2) colder than the atmosphere in the balance case. Currents of hot air may impinge on the arms of the balance and cause one arm to expand unequally, or buoy up the beam. For instance, a platinum crucible which appeared to weigh 20*649 grms. when warm, really weighed 20*692 grms. when cold. Hence, the crucible weighed 0'2 per cent, less when hot than cold. If the crucible be colder than the atmosphere of the balance case, moisture may condense on the object to be weighed. If the object to be weighed is likely to be electrified, it should be allowed to stand some time after it has been wiped, before it is weighed. 6 1 Two are shown in fig. 1. 2 G. S. Johnson, Chem. News, 68. 211, 1893. :! If anything solid or liquid be spilt on the pans or in the balance case, clean it off at once. 4 Usually, a small pennant is placed near the top of the pointer, or screws at or near the ends of the beam, for this adjustment. These screws should be interfered with as seldom as possible. 5 If the movements of the beam seem to be erratic, the fault is probably due to (1) the beam touching the rider or the carrier of the rider ; or (2) the pan touching some object in the balance case ; or (3) a displaced stirrup. 6 J. A. R. Newlands, Chem. News, u. 107, 1865. 8 A TREATISE ON CHEMICAL ANALYSIS. The electrification of the weights 1 or object causes erratic movements of the pointer. The balance should not be cleaned too frequently. With proper care, and the balance in a suitable position, a good cleaning every three or four months should suffice. This may occupy from a quarter to an hour. All the loose parts should be carefully removed and dusted ; the movable parts, cleaned and oiled. Wipe off any excess of oil. Polish the suspensions with a piece of chamois leather. Restore all the parts. Adjust the equilibrium screws and gravity bob as indi- cated later in this chapter. The weights can be tested at the same time. 4. Weighing Hygroscopic or Volatile Liquids and Powders. Never place solid reagents in direct contact with the pans. Small crucibles, watch-glasses, basins, capsules, glazed paper, 2 etc., are convenient supports for substances not affected by exposure to the air while being weighed. Hygro- scopic, efflorescent, and volatile substances, and substances liable to absorb FIG. 3. Weighing bottles and weighing tubes. carbon dioxide, etc., from the air, must, if possible, be weighed in closed vessels, clipped watch-glasses with ground edges, etc. A selection of such vessels is shown in the diagram, fig. 3. The weighing bottle or tube must be adapted for the material under investigation. Weighing tubes must be suitably supported. The stand shown in the diagram, fig. 3a, is suitable for supporting a tube either vertically or horizontally. The weighing tube may be permanently closed at one end, and fitted with a ground stopper at the other fig. 3a ; or it may have a ground cap at both ends fig. 3/. Weighing tubes for boats may have two legs to prevent rolling fig. 3g. 3 Watch-glasses with ground close-fitting edges are 1 Quartz weights (H. Butf, Dingler's Journ., 222. 159, 1878; S. Stein, Zeit. anal. Chem., 17. 471, 1878 ; glass weights R. Ulbricht, Ber., 10. 129, 1877) in a velvet-lined box are liable to become electrified as they are removed from their bed (Chem. Ztg., 12. 494, 1888). See page 551. R. Hennig (Zeit. Instr., 5. 161, 1886) discusses the errors due to air currents which are set up when liquids are weighed in open vessels. 2 Paper, horn capsules, and similar substances are somewhat hygroscopic and vary in weight with the humidity of the air. ' ' Xylonite paper " has many advantages over ordinary glazed paper. It is less hygroscopic ; it can be washed with water ; it is not attacked by ordinary acids and alkalies. It is attacked by organic solvents (H. Schweitzer, Chem. Ztg. , 14. 698, 1890). H. F. von Jiiptner, Die Einfuhren einheitlicher Analysenmethodtn, Stuttgart, 261, 1896. I keep weighing bottles, brushes, etc., in a glass "catgut" box beside the balance. * A. Gawalovski, Chem. Centr. (3), 16. 369, 1886. WEIGHING. clamped together by a suitable clip as shown in the diagram fig. '3e. 1 Weighing bottles may have different shapes and sizes figs. 36, 3c, 3d. They may have ordinary ground stoppers, or ground caps as in Guttmann's weighing bottle fig. 3c. The latter are best for powders, 2 in weighing by difference, since (1) several portions can be weighed out successively without the joint having to be cleaned, as is necessary with the ordinary ground stopper, and (2) dust does not accumu- late between the stopper and the ground surfaces. A modification with ground- in tubes is useful for drying substances in a current of gas, for determining water of crystallisation, etc. fig. 182. In using weighing bottles or tubes, the vessel plus powder may be weighed. Some powder is removed from the vessel, and the vessel and powder weighed again. The loss in weight represents the powder removed. The capped weighing bottle may also be used for weighing liquids fig. 4. The tube with the liquid has a small pipette inside. All is weighed. Some liquid is withdrawn by means of the pipette. 3 The pipette is returned to the bottle, and all is weighed again. The loss in weight represents the amount of liquid removed. Berl 4 has a convenient pipette for weighing cor- rosive liquids fig. 5. The apparatus is weighed. The tap C fits closely without lubrication. To fill the pipette, connect A with K ; apply suction at A, close C. Let the point S dip in the liquid under examination ; connect P and K. The liquid runs into the pipette. Close the FIG. 4. Weighing FIG. 5. Berl's tap before the liquid has reached the tap bottle and pipette. weighing pipette. C. Clean the end of the tube S. Return the pipette to the tube F and weigh. The increase in weight represents the amount of liquid in the pipette. Now let the tip S dip under water. Let the contents gradually run from the pipette by opening the cocks. Run water through the apparatus. This will ensure the removal of all the liquid from the pipette. 5 This procedure may be obviously modified to suit particular conditions. 6 The instrument is rather lighter than Lunge and Rey's well-known pipette for a similar purpose. Large glass and porcelain vessels, platinum crucibles, basins, etc., dried and cooled in desiccators are particularly liable to condense moisture on their surface while being weighed. 7 Many powdered substances also begin to absorb 1 J. J. Griffin, Chem. News, 101. 71, 1910. 2 C. Mangold, Zeit. angew.-Chem., 4. 441, 1891 ; L. F. Guttrnann, Journ. Amer. Chem. Soc., 28. 1667, 1906 ; A. Breneman, Chem. News, 48. 168, 1883 ; T. Zohren, Chem. Ztg., 36. 824, 1912. 3 For a weighing bottle with a pipette stopper, see L. E. Levi, Journ. Amer. Chem. Soc., 27. 614, 1905. 4 G. Lunge and H. Rey, Zeit. angew. Chem., 4. 702, 1891 ; H. Rosenlecher, Zeit. anal. Chem., 37. 209, 1898 ; E. Berl, Chem.' Ztg., 34. 428, 1910. 5 For strong fuming acids, a drop of liquid may escape from S into the tube F while it is being weighed. In that case, a drop of water is placed in F, and all is weighed without allowing the tip of the tube S to be wetted. The pipette is then iilled as described in the text. Berl has modified bulb for liquids with a high vapour tension (i.e. volatile liquids). 6 When weighing liquids which spoil on exposure to the air, Holde's pipette may be used. D. Holde, Zeit. angew. Chem., 12, 711, 1899 ; H. Schweitzer, Journ. Amer. Chem. Soc., 15. 190, 1893 ; E. Reichardt, Zeit. anal. Chem., 7. 187, 1868. 7 Vessels full of air have a uniform weight as soon as their surfaces are in equilibrium with the atmosphere. When these conditions are changed by a variation of temperature for 10 A TREATISE ON CHEMICAL ANALYSIS. moisture and possibly also other gases quite energetically immediately they have left the desiccator, and Pagasogli l mentions an example where a powder dried over sulphuric acid increased in weight from O'OOl to 0*003 grm. during the weighing, which occupied between 3 and 6 minutes. The weight of a body will increase until the vapour pressure of the absorbed moisture is equal to the vapour pressure of the moisture of the air. A dish of concentrated sulphuric acid, or a vessel 2 of ground calcium chloride, is usually kept in the balance case with the idea of retarding the absorption of moisture while the substance is being weighed. The remedy, however, is not particularly efficacious. This is well illustrated by the following observations : 3 A hair hygrometer was placed in a balance case. Some fresh granulated calcium chloride was placed in a funnel resting over a flask. The hygrometer showed a humidity of 60*5 to 61 per cent. In an hour, the humidity fell to 55'5 per cent.; in 15 hours, to 53'5 per cent, The humidity then gradually rose during the next three days, to its former value, 60-61 per cent. The calcium chloride was then appreciably moist, and liquid began to drop from the funnel containing the salt in question. Ostwald 4 recommends weighing the substance as rapidly as possible after remov- ing the vessel from the desiccator. The idea is to make the error due to adsorption as small as possible. In igniting, drying, etc., to a constant weight the same time may not always be occupied in making the control weighings, and different results may accordingly be obtained with the same body. For instance, three control weighings of the same body gave 17'4334, 17'4332, and 17'4331. The idea is best applied by getting the approximate weight of the dried body in a covered crucible. Repeat the ignition and cooling, place the necessary weights on the pan, so that the rider alone is needed to complete the weighing. This will usually suffice, but a third ignition and drying will enable the third weighing to be made in a few seconds. On the other hand, Smith 5 recommends leaving the body 20 minutes in the balance case after it has been removed from the desiccator, in order that the body may adsorb its normal film of moisture, i.e. until its weight is constant. The objection to this plan is the long time required for the control weighings, and there is nothing to show that the atmosphere has remained of a constant humidity during the experiment. However, it is not likely to change appreciably under ordinary conditions. 6 5. Weighing. Weighings smaller than O'Ol grm. are made with the rider. The milligram weights are not used. When each arm is divided into ten divisions, use a example, it takes a long time to bring the vessels to their original condition. Thus, a hard glass tube, 60-80 cm. long and 2-2 '5 cm. diameter, must be kept two or three hours in a balance case before its weight is constant. Glass flasks, 1 to 4 litres capacity, after having been heated or rubbed, do not attain a constant weight until they have remained in the balance case five or six hours. Polished platinum rapidly regains its primitive weight, but when the surface is more or less roughened, it attains its proper weight more slowly than glass or porcelain. J. S. Stas, (Euvres Completes, Bruxelles, i. 317, 1894 ; Chem. News, 4. 206, 1861 J. L. Smith, ib., 31. 55, 1875. 1 G. Pagasogli, VOrosi, 10. 109, 1888 ; J. L. Smith, Chem. News, 31. 55, 1875. Two calcium chloride jars are shown in the back of the balance case, fio-. 1. 3 O. Kuhn, Chem. .%., 34. 1097, 1108, 1910. 4 W. Ostwald and R. Luther, Hand- und Hulfsbuch zur Ausfuhrung physiko-chemischer Messungen, Leipzig, 51, 1902 ; London, 38, 1899. 5 J. L. Smith, Amer. Chemist, 5. 212, 1874 ; F. P. Treadwell, Kurzes Lehrbuch der analy- tische Chemie, Leipzig, 2. 21, 1911 ; New York, 22, 1904. 6 R. Hottinger (Zeit. anal. Chem., 48. 73, 1909) recommends plotting the increases in weight as ordmates, with time as abscissae. The curve will be horizontal after the elapse of a few minutes. The first weighing will not be zero. The time required for the first weighing is longest ; subsequent weighings occupy less time. Hottinger extrapolates for zero by continuing the curve. F. Richarz, Verhandl. phys. Ges. Berlin 83 1886 WEIGHING. I I centigram rider; if each arm be divided into twelve divisions, use a 12-milligram rider. 1 Each division on the beam then corresponds with a milligram, each of the n subdivisions with the n~ l ih milligram. The 12-mgrm. rider enables fractions below 12 mgrm. to be weighed with the rider; the 10-mgrm. rider, fractions below 10 mgrm. 2 Some balances require a half-centigram rider. This is for use on a light scale attached to the front of the beam of the balance. The effect pro- duced by the half-centigram rider on the auxiliary beam is the same as would be obtained with a centigram rider on the beam proper. This device is only used with the very short-beam balances where the graduations on the beam with a centigram rider would be inconveniently close. In attempting to weigh to the tenth of a milligram, it might be thought best, at first sight, to add weights and move the rider until the pointer of the balance swings equally on both sides of the zero of the scale. It is assumed, quite correctly, that when the exact weight has been added to a properly adjusted balance, the pointer will swing the same number of divisions to the right and the left of the zero, provided it does so when the pans are empty. As a matter of fact, this method is perhaps more often in error than otherwise. It entails a frequent adjustment of the zero point of the balance, owing to unavoidable variations in the zero point. 3 Even if the position of equilibrium of the pointer with an unloaded balance be adjusted with the pointer at the zero of the scale, the zero point may change in a short time. The arms may be unequally heated, etc. By ignoring this fact, appreciable errors may creep into the work. This method of weighing, in which the rider is adjusted until the swing of the pointer of the loaded balance is nearly the same as with the un- loaded balance, is very common, and it is accurate enough for most analytical work. We shall see later that the errors incidental to the methods of preparing precipitates for weighing, mask the small errors introduced by the effects just indicated. In special cases, when weighing to the tenth of a milligram, it is important to be able to make the weighing independent of any temporary inequality of the arms, etc. Hence we have : (1) Gauss' method of double weighing. The object is weighed first in one pan, then in the other. The square root of the product is supposed to represent the weight of the object. (2) Borda's method of weighing by tares. Here the object is balanced by a suitable tare (wire, weights, shot, potash bulb, etc.). The object is removed, and weights are added in its place until equilibrium is restored. (3) Method of iveighing by double vibrations (or swings).^ The following description may make the process of weighing by double vibrations appear somewhat laborious. The labour is, however, more apparent than real. Practice with the method is also an excellent way of mastering the manipulation of the balance. First, find the zero position of the pointer, that is, the position the pointer will occupy on the scale where the balance, swinging without a load, will come to rest. 1 The rider enables all weights below O'Ol grm. to be discarded, and this (1) saves time and trouble ; (2) ensures greater accuracy owing to the tendency of the small weights to collect dirt, etc. ; (2) decreases chances of error in reading the weights on the pans ; and (4) renders it possible to estimate weights to a greater degree of accuracy. 2 The weight obtained by adjusting the rider is usually treated as if it were a weight on the pan. :? T. E. Thorpe, Journ. Ckem. Soc., 47. 116, 1885 ; J. J. Manley, Proc. Roy. Soc., A, 86. 591, 1912 ; Trans Roy. Soc., A, 212. 227, 1912. 4 For details see text-books on physics. J. H. Poynting, Chem. News, 39. 45, 1879 ; Proc. Manchester Lit. Phil. Soc., 18. 33, 1879. ilmilniili 12 A TREATISE ON CHEMICAL ANALYSIS. It would be too great a waste of time to wait until the oscillations of the pointer cease when it is possible to deduce the position directly from the distances the pointer swings to the left and right of the scale. 1 Following F. Kohlrausch, 2 I prefer to number the scale with the zero on the extreme left, not in the middle of the scale. This prevents any confusion of signs later on. Neglect the first two or three swings on account of the shock, and air currents set up when the door of the balance FIG. 6.-Scale of pointer of is closed ' Take 3 > \ r 7 consecutive readings of the balance. turning points of the swinging pointer. Take the average of the odd, and also the average of the even numbered readings. Add the two results and divide by 2. In illustration, suppose the turning points read 7-2; 13-0; 7*2. The mean of the first and third readings is 7 '2 ; the mean of both the even and odd readings is 7-2 + 13-0 divided by 2, that is, lO'l. This number represents the required zero point of the balance. The needle will come to rest when the pointer is at lO'l. 8 Second, find the deviation of the scale per milligram, that is, the sensibility of the balance. The object to be weighed is placed on the left pan, the weights on the right pan. 4 When the weights 5 are so far adjusted 6 that another centigram weight would be too much, close the door of the balance case, and move the centigram rider on the divided beam until the pointer moves to the right and left of the 10th division. Find the position of rest, e.g. Weight on pan. Turning points. Averages. Zero point. 11-216 8-2; 13-3; 8-4 8'3;13'3 10'8 Move the rider another milligram division to the right. . Weight on pan. Turning points. Averages. Zero point. 11-217 4-1; 11-6; 4-3 4'2;ll-6 7'9 Hence, the zero point is displaced 10'8 7*9 = 2*9 divisions by increasing the weight 1 milligram ; or 2*9 scale divisions correspond with 1 milligram. This number, 2 -9, represents the required sensibility of the balance for the given load. The sensibility of a balance for a given load is therefore defined 1 The reading of the pointer in accurate work is greatly facilitated by the use of a lens or a magnifying mirror made for the purpose. The lens or mirror is so arranged that an enlarged image of the pointer and scale meets the eye when the head is in its natural position before the balance. The lens is shown in position in front of the scale, fig. 1. 2 F. Kohlrausch, Leitfaden der praktischen Physik, Leipzig, 44, 1896. 3 Scales with red line's instead of the black ones are claimed by C. M. Clark (Journ. Amer. Chem. Soc., 32. 884, 1910) to be more easily read. 4 Keep rigorously to this rule : object on left pan, weights on right pan. The reason will appear later. 5 The weights are not usually kept in the box, but rather on the "front floor" of the balance case. F. Clowes and J. B. Coleman (Quantitative Chemical Analysis, London, 4, 1909) recommend keeping the weights on a piece of cardboard ruled in squares and numbered to correspond with the weights to be kept thereon. Small porcelain slabs with suitable depressions, and properly lettered, are more convenient. Weights over, say, 20 grins, may not be needed very frequently ; in that case, keep them in the box. The milligram weights are redundant, and may also be left in the box. For the condensation of "moist air ''on the weights see T. Jlimori ( Wied. Ann., 31. 1006, 1887) and E. Warburg (ib., 27. 481, 1886). 6 The weights are not to be handled with the fingers, but always lifted with the ivory-tipped forceps provided for that purpose. Never alter the load on either pan without first arresting the balance, so that the beam or stirrups no longer rest on the knife-edges. WEIGHING. 1 3 as the displacement of the position of rest of the beam produced by an excess of 1 milligram weight on either pan. Third, calculate the weight of the load on the pan. From the preceding results, it follows that the load weighs ll"216 + #. The zero point of this load is dis- placed 10-8-10-1 = 0-7 scale divisions. Since 2 -9 scale divisions correspond with 1 milligram, 07 scale divisions will correspond with 07 -=- 2*9 = 0'24 scale divisions or 0-24 mgrm. Hence the weight of the body is 1 1 '216 + 0'00024 = 11-21624 grms. These calculations may be summarised in the formula ^ , a- 2 Correction = + mgrm., a o where z represents the zero point of the unloaded balance ; a, the zero point with not quite sufficient weight on the right pan ; and 6, the zero point with a milligram more on the right pan than corresponds with a. Analytical balances will rarely indicate with certainty smaller weights than O'OOOl grm. Hence, although the weight has been stated to the fifth decimal, in future weights will generally be rounded off by dropping the fifth decimal and raising the fourth decimal one unit when the dropped figure exceeds 5. It is well to keep rigidly to one routine process, so that the preceding operations may become mechanical. With a little practice, time is saved owing to the fact that no useless trials are made in the final adjustment of the rider. The result is also more exact than the method of weighing by equal deviations of the zero point. The following may be taken as representing the operations involved in weighing a porcelain crucible: Add 10 grms. too little; add 2 grms. too much ; add 1 grm too little ; add 0'5 grm. too much ; add 0*2 grm. too little; add O'l grm. too much; add 0*05 grm. too much; add 0*02 grm. too much ; add 0-01 grm. too little. Close the door of the balance case. Place the rider on the 5th division too little ; on the 6th division too little, but nearly right. Read the turning points as indicated below. Place the rider on the 7th division and read the turning points. Remove the loads on the pans, and read the turning points. We thus obtain Turning points. Zero points. Unloaded balance .... 7'2;13-0;7-2 2 = 10'1 11-216 ...... 8-2; 13-3; 8-4 a = 10'8 11-217 4-1; 11-6; 4'3 6= 7'9 Consequently, q-z_10-8-10-l = 0-7 m rm Hence the weight is 11 '2 162 grms. After a certain amount of experience has been obtained with a given balance, it is possible, before the weighing is completed, to estimate the approximate weight of the load from the rate at which the pointer moves over the scale, and thus skip some of the steps indicated for adjusting the exact weight by the process just outlined. Thus, the 0'5 grm., indicated above, was not placed on the pan during the actual weighing. This experience, however, takes some time to acquire, and it is better to work by system rather than at random. 6. The Sensibility of the Balance. Balances may now be obtained which show practically no change in sensitiveness (a - b) between full and empty loads. The mere statement that "a balance is sensitive to y^th mgrm." is not sufficient unless the corresponding 14 A TREATISE ON CHEMICAL ANALYSIS. load be specified. A balance carrying 5 grms. might be sensitive to T Vth mgrm., whereas y^th mgrm. with a load of 100 grms. would have no perceptible influence on the movements of the beam. Most balances have a " gravity bob " screw attached to the central portion of the beam, or a sliding weight on the pointer (fig. 1). The gravity bob provides a means of regulating the value of (a-b\ the sensitiveness of the balance, within , certain limits. 1 By its means, the distance OG or r (fig. 2) between the centre of gravity and the central knife-edge can be altered. We have seen (page 6) that the nearer the centre of gravity to the central knife- edge, the greater the sensibility, and the greater the period of vibration of the balance. If the gravity bob be too near the central knife-edge, the time of vibration is too long, and the balance will be too sensitive ; if the gravity bob be too far away from the central knife-edge, the balance may be unstable, either with the empty pans or with a full load. It is generally possible, by adjusting the gravity bob, to make the numerical value (a - b) a single figure. Suppose, for example, we have a balance of constant sensibility so adjusted that its sensibility is 0'7, the work of weighing is much simplified. For instance, in weighing a platinum crucible, we simply find Turning points. Zero points. Unloaded ..... 5'4 ; 17'6 ; 5'6 s = ll'55 Loaded 20 '023 grms. . . . 6'0;187;6'2 = 12'40 Displacement of zero point . . . . . a-z= 0'85 Hence, 0'85^0'7 = 0'12. The required weight is 20-02312 grms. With such a balance, properly adjusted, a weighing is very quickly and accurately performed. - V = *-, ^ s h* mm mm mm mm mm* ^ ^ S ta. - mm mm mm* ^ = MK = *. . , * * 30 40 50 60 70 Load in grms. FIG. 7. Variation of sensibility of Balance No. 3 with load. If the sensibility of the balance varies a little with the load, 2 as is usual with common analytical balances, owing to a slight bending of the loaded beam, it is simplest to determine the value of this constant for the different loads once for all. Suppose that we find BALANCE No. 3. Load . . 0, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100 grms. Deflection . 076, 074, 072, 072, 072, 070, 070, 070, 0'68, 0'66, 0'65 divs. The value of a - b can then be read at a glance, for any given load, from the graph of these numbers (fig. 7). When a - b is so fixed, a and z are alone determined each weighing. Analytical balances are now usually made with a short, light beam which has a period of vibration of about 10 seconds. The old-fashioned long-beam balances are seldom used in analytical practice. The long-beam balances sometimes have a period of vibration approaching 60 seconds ; and, in consequence, a lot of valuable time may be lost in watching the oscillations of the beam, with an additional risk of error due to the absorption of moisture, etc., by the object being weighed (page 1 For a device for adjusting the centre of gravity, see C. Bunge, Chem. Ztq., 37. 280 1913 a H. R. Proctor, Chem. News, 30. 255, 1874. WEIGHING. 1 5 9). Short-beam balances are therefore used when rapid weighing is needed. The unavoidable errors in setting the knife-edges are of greater significance in short-beam balances; and the oscillations being small, minute differences are not so readily detected as in long-beam balances. The latter objection is partly overcome by using a lens to magnify the motions of the pointer (page 12). 7. The Accuracy of the Weighing. The finest weighing cannot compensate for the lack of purity of substance, or the absence of perfection in chemical operations. G. D. HINRICHS. Accurate measurement is the foundation of real science. Many students of chemistry are introduced to the art of exact measurement through quantitative analysis. It is therefore necessary to get clear ideas about the principles of measurement in analytical chemistry. It would obviously be an absurd waste of time to weigh to an accuracy of 0*001 per cent, at one stage of the work, if at another stage of the work the method of preparing the substance for weighing introduced an error of 0'5 per cent. Although many errors are small enough to be neglected, yet cases often arise where errors, trivial under one set of conditions, assume serious dimensions under another set. Hence, it is necessary for every analyst to have clear ideas of the character of the different errors which may affect his work, in order that corrections may be introduced whenever necessary. A chemist will not consider any time lost which is necessary for obtaining accurate results ; but it is certainly a waste of time to carry the accuracy of any single operation (e.g. weighing) much beyond the limits of experimental error incidental to the process of analysis. Errors in weighing and measuring should fall well within the limits of the ex- perimental error due to the analytical operations. There is no advantage in making one of a series of interdependent measurements much more accurate than the others. A chain is no stronger than its weakest link. Under ordinary conditions, ten consecutive weighings of the same substance should not differ much more than 0*0005 grm., that is, by 0'05 per cent, working with one-gram samples ; whereas in other cases the error in weighing should not exceed O'Ol per cent., that is, O'OOOl grm. EXAMPLE. If an error no greater than O'OOl grm. were made in the initial weighing of a gram of a clay which contained 0*2 per cent, of MgO, the greatest error in the determination of the magnesia could be no greater than 0*1 per cent, of its value ; and the final result could not be affected by more than 0*1 per cent, of 0'2 per cent., that is, 0*0002 gram. This is negligibly small. The weighing is even more accurate than is necessary. Suppose further that an error of 0*001 be possible in the weighing of the 0'0055 grm. of Mg 2 P 2 7 , an error of O'OOl grm. would represent an error of 1 00 x O'OOl -7-0*0055 = 18 per cent. This is too great. The weighing must be more exact. Clay analyses are usually made upon one-gram samples. The final result is represented as a percentage taken to the second decimal. This seems to imply that the weighings are exact to the 0*000 1th of a gram. Possibly they are. It would not be very difficult to make the weighings exact to '0000 1th grm. This would enable us to represent the final analysis, as beginners frequently do, with three decimals. But we shall see that the appearance of accuracy conveyed by these decimal figures the second and third is quite illusory. 1 Calculations should stop ivhen the limit of precision of the analytical process is reached. Pre- cision inform does not necessarily represent accurate work. 1 In illustration, see the analyses of tincal by H. Vohl in Dingier 's Journ., 2OI. 165, 1871 ; and of iron ores by J. C. Welch in the Chem. News, 52. 32, 1885, to four places (per cent.)! This discussion is resumed later. 1 6 A TREATISE ON CHEMICAL ANALYSIS. Altogether apart from the accuracy of the sampling, analytical operations involved in preparing the different substances for weighing are seldom more exact than is represented by Washington's l statement : " For silica or other constituents which amount to 30 per cent, or over, the allowable limits of error are 0'2 to 0'3 per cent, of the whole material ; for alumina and other constituents which amount to from 10 to 30 per cent., O'l to 0-2 per cent.; and for constituents which amount to 1 to 10 per cent., 0'05 to O'l per cent." If we accept these numbers as representing the errors incidental to the best processes available for the analysis of silicates, clays, etc., it will be obvious that little, if anything, is gained by increasing the accuracy of the weighings beyond the tenth of a milligram, until the operations generally used in analytical chemistry admit of greater refinement. Indeed, our analyses would approximate more closely to the truth if we represented the percentage results with one decimal instead of two. As a matter of fact, two decimals are generally used in technical analyses because we have grown accustomed to the plan, not because it represents the accuracy of the work. The first decimal is supposed to be nearly right; the second is retained in order to keep the first decimal as nearly accurate as possible in calculations made with the analytical data. Under ordinary conditions, therefore, a third decimal is quite out of perspective,, while a fourth is an abomination. 2 Few analytical processes can be depended upon beyond a limit of accuracy represented by 1 in 500; still fewer beyond 1 in 1000. In fact, a method which would enable a real distinction to be drawn between 50 '0 and 50 - 1 per cent, would be considered excellent. As a matter of fact, this degree of accuracy is seldom, if ever, attained in a complex separation. Hence, the statement that 50-13 per cent, of a constituent is present can rarely represent more accurate work than if the result had been simply expressed by the number 50*1 per cent. To quote an illustration by Hinrichs, 3 Berzelius (1826) worked with a balance accurate to O'Ol grm., Ramsay and Aston (1893) worked with a balance accurate to '0000001 grm. In three determinations of the water of crystallisation of borax, Ramsay and Aston found 0'471677 grm., and Berzelius, 0*4710 grm. The individual determinations in the former case varied from '4 7 109 9 to 0-472026. Here then, with a variation in the third decimal, it follows that the fourth and succeeding decimals have no more value than the results determined on the less sensitive balance used by Berzelius. In some cases, working under special conditions with large amounts of raw materials, it is possible to determine quantitatively certain constituents occurring in amounts below O'Ol per cent. In that case the first significant figure might appear after the second place in the final statement of the results. The determination of these constituents may be useful and important for special purposes, which are not usually industrial. 8. Correcting the Weights. When working to the tenth of a milligram, it is necessary to test an ordinary box of analytical weights in order to find if the weights are consistent amongst themselves. The errors usually permitted by a standard testing laboratory, before a first-class box of weights can receive its imprimatur, are indicated in 1 H. S. Washington, Manual of the. Chemical Analysis of Rocks, New York 24 1904 M. Dittrich, Neues Jahrb. Min., 2. 69, 1903. 2 These remarks, of course, have no reference to the analysis of special materials under special conditions where three decimals may be needed to represent the accuracy of the work for instance, in certain atomic weight determinations ; but even here, the value of the result is frequently overrated. See C. Molar, Rass. Min. Met. Chim., 36. 101, 1912. 3 G. D. Hinrichs, The Absolute Weights of the Chemical Elements, St Louis, Mo., 42, 1901. WEIGHING. the second column of the following table. The third column represents the degree of accuracy of the calibration. Table I. Permitted Errors in a Set of Weights. Weight. Tolerance. Values given to Weight. Tolerance. Values given to 200 O'OOl o-ooooi 1 to 2 0-00005 o-oooooi 100 0-0005 o-ooooi 0-5 0-00004 O'OOOOOl 50 0-0003 '00001 0-2 0-00003 o-oooooi 20 o-oooi o-ooooi 0-05 to 0-01 0-00002 o-oooooi 5 o-oooi o-ooooi A set of weights which has passed this test will be more accurate than is needed for most analytical work. C^d^*/ **" ^Fairly accurate_weights can be purchased for a reasonable sum, and for most" gravimetric work the inaccuracies of the better class of weights are negligibly small in comparison with the errors of experiment, and the imper- fections in the processes of analysis. An analyst, however, will not be satisfied with having his weights probably exact enough. He must know that the weights actually have the required degree of accuracy. This certainty can only be established by calibrating the weights. The errors introduced by the imperfections of the weights can easily be made less than O'OOOl grm., or O'Ol per cent. The weights should be tested at periodic intervals, say every three or four months. All depends upon the frequency with which the weights are used, and who uses them. Faraday J has observed : " The balance and weights should be carefully examined at intervals, to ascertain their accuracy, for if they involve unnoticed errors, the experiments made with them may be worse than useless. Some curious conclusions, tending to subvert most important chemical truths, might be quoted as having arisen solely in this way." The weights should be cleaned by wiping them with a camel-hair brush, or silk hand- kerchief. They must not be rubbed. Two hours will suffice for the calibration. In special cases, e.g. where volumetric apparatus is being calibrated, absolute weights may be required, but for general analytical work absolute weights are not necessary. For instance, suppose that 1 grm. of clay be weighed with a perfect one-gram weight, and exactly O'l grm. of silica is obtained. The clay contains 10 per cent. Si0 2 . Suppose that the clay be weighed with a one-gram weight which is really 0'9 grm. (called 1 grm.), and that the other weights are consistent. The silica obtained would then weigh 0-09 grm. (called O'l grm.). Here, the amount of silica reported in the clay will be the same as before, because 100x0-09 0-9 = 10 per cent. Si(X Hence, if the weights are consistent with one another, their absolute values have no influence on the accuracy of a quantitative analysis. Most books on analysis and physics 2 give the modus operandi for calibrating a set of weights. The following plan will be found to differ but little from 1 M. Faraday, Chemical Manipulation, London, 29, 1842. 2 W. H. Miller, Phil. Trans., 146. 753, 1856; K. L. Bauer, Zeit. anal. Cfiem., 8. 390. 1869 ; Pogg. Ann., 137. 103, 1869 ; R. Bunsen, Zeit. anal. Chem., 6. 1, 1867 ; T. W. Richards, 1 8 A TREATISE ON CHEMICAL ANALYSIS. those usually recommended. Every student beginning analytical work should be drilled by standardising a box of weights and using the method of weighing described in the preceding section. The weights are usually made in sets 1, 2, 5 ; and, to get the 4 and 9, it is necessary to duplicate either the 1 or the 2. German and English weights are arranged on the former system, and French weights on the latter. There is no marked difference in rapidity of manipulation with either system, i.e. in the average number of times the weights are removed and returned to the box during a series of weighings. With the French system, it is easier to check the weights against one another, since the 1 + 2 + 2 should equal the 5 weight. Hence, as Pickering 1 says : " if a mistake is suspected, a test can be applied by using an additional weight instead of putting back all the small weights, and adding a larger one." To meet this objection, an extra one-gram weight is often added by the English and German makers. Weights below 0-01 grm. are not needed. The Centigram Weights. The rider should be rubbed with fine emery paper until it weighs exactly 10 (or 12) milligrams. 2 Suppose the box contains I \\J1 / U**UUB6*MJUyB UU.JJJJUQ' 0-05, 0-02, 0-01, 0-01 gram weights. Take a centigram weight from another box, or make one from a piece of wire or foil the weight need not be absolutely correct. It is used temporarily as a tare. Call its weight A. It is necessary to have some distinguishing mark for the two 0'01-grm. weights; for the two 0'1-grm. weights; for the three 1-grm. weights; and for the two 10-grm. weights. The weights may be stamped in different ways, or advantage may be taken of accidental markings ; the foil weights may be distinguished by turning up different corners. 3 Place one of the O'Ol-grm. weights say [0'01] ft on the left pan of the balance, and balance this weight with the centigram tare A and the rider. Sup- pose that we find the weights on the right A + "00000 grm. We write down [0-011, = 4 + 0-00000 (1) Replace the weight on the left with the [0'01] & weight. Suppose we find [0-01] 6 = 4 + 0-00001 (2) Since the two weights [0'01] and [0'01] 6 should together make up the [0'02] grm., place the two former on the right pan, and the [0 02] grm. on the left pan. We find that [0-02] = [0-01] + [0-01] 5 + 0-00002. Hence, from (1) and (2), we deduce that [0-02] = 2A + 0-00002 . . . . . (3) Similarly, [0 05] = [0-02] + [0-01] + [0-01], + A + 00004, Journ. Amer. Chem. Sbc., 22. 194, 1900; A. T. H. Verbeck, Dingier ' Journ., 227. 400, 1878 ; W. Dittmar, Exercises in Quantitative Analysis, Glasgow, 1, 18*87 ; W. Ostwald and R. Luther, Hand- und Hiilfsbuch zur Ausfiihrung physiko-chemischer Messungen, Leipzig, 55, 1902 ; F. Kohlrausch, Leitfaden der praktischen Physik, Leipzig, 50 1896 W Crookes' Phil Trans., 162. 277, 1873 ; Ghrnn. News, 15. 191, 1867; Select Methods in Chemical Analysis, London, 687, 1905 ; H. N. Morse, Exercises in Quantitative Analysis, Boston 26 1909 For Benoit s method of calibration, see W. Watson, A Text-book of Practical Physics, London, 74, 1906. 1 E. C. Pickering, Elements of Physical Manipulation, London, I. 48, 1878. 2 Or a standard centigram rider can be purchased with, say, a National Physical Laboratory certificate. 3 It is best to get the maker to stamp the weights of the same denomination a b c before purchasing. WEIGHING. which, by substitution from (1), (2), (3), furnishes [0-05J- 54 + 0-00006 . . . . (4) The Decigram and the Gram Weights. The decigram weights [0'5], [0*2], [0-l] a , [0*1 ] b are treated in a similar manner. We thus obtain [0-l] a = [0-05] + [0-02] + [0-01] a + [0-01] 6 + A + O'OOOO. From (1), (2), (3), (4), we get [0-l] rt = 104 + 0-00005 (5) Similarly, we obtain the data for the remainder of the decigram weights, and also for the gram weights, indicated in the second column of the subjoined table : Table II. Calibration Data for Weights (Student's Box No. 8). Nominal or face value. Data obtained in calibration. Preliminary values assuming Corrected, assuming 10-00040 to be Correction. [10]a = standard. A = 0*01 grm. standard. [0 01J, ^4+0-00000 o-oiooo o-oiooo +0-00 [0 01 ^4+ 0-00001 0-01001 o-oiooo + 0-01 [0-02J 2^+0-00002 0-02002 0-02000 + 0-02 [0-05] 5/^+0-00006 0-05006 0-05000 + 0-06 [0- ij a 10^4+0-00005 0-10005 o-ioooo + 0-05 [0-lja 10^+0-00007 0-10007 0-10000 + 0-07 [0 2] 20^4 + 0-00010 0-20010 0-20001 + 0-09 [0 5) 50^ + 0-00013 0-50013 0-50001 + 0-12 [IV 100^4+0-00005 1-00005 1-00004 + 0-01 [i; 6 100^4 + 0-00007 1-00007 1-00004 + 0-03 [it [2] 100^ + 0-00007 200^4+0-00009 1-00007 1-00004 2-00009 2-00008 + 0-03 + 0-01 [5; 500^+0-00014 5-00014 5-00020 -0-06 [10 a 1000^4+0-00040 10-00040 10-00040 +0-00 ho b 1000^4+0-00039 10-00039 10-00040 -o-oi [20 2000^4 + 0-00070 20-00070 I 20-00080 -o-io [50; 5000^4+0-00169 50-00169 50-00200 -0-31 We can assume pro tern, that A is exactly O'Ol grm., and draw up a table of corrections referring the weights, as is usually done, to the [10] a weight as a basis. This has been done in Table III., where the first column represents the normal or face value of the weights ; the second column, the results expressed in terms of A ; the third column, the results on the preliminary assumption that .4 = '01 grm.; the fourth column, results on the assumption that the [10] a -grm. weight is the standard ; and the last column represents the deviations of the weights from the standard of reference [10] tt = 10-00040 grm. If desired, we can find the exact value of the [10] a -grm. weight in terms of a standard 10-grm. weight. It is by no means necessary to do this. All that is necessary is that the corrected weights be consistent amongst themselves. If the [10] rt and the standard 10-grm. weights are equal, the corrections in the above table need not be altered. If it were found that [101, = 10 grm. +0-00037, it follows that our standard was 0*00003 grra. too high, and a corresponding change is therefore made in the table of corrections. This has been done in the following table : 2O A TREATISE ON CHEMICAL ANALYSIS. Table III. Corrections for Weights (Student's Box No. 8). Nominal or face value ' of weight. Corrections. [10] a = 10 grms. + -00037 grm. Nominal or face value of weight. Corrections. [10] a = 10 grms. + JO -00037 grm. [0'01]a [0'01] 6 [0-02] [0-05] [0'1] [O'l], [0-2] [0-5] [1] + 0-03 + 0-04 + 0-05 + 0-09 + 0-04 + 0-06 + 0-05 + 0-12 + 0-04 [1> [lie [2] [5] [lOJa [io> [20] [50] [100] + 0-06 + 0'02 -0-08 -0-23 + 0-03 -0-02 -0-07 -0-28 -0'42 It will be observed that, if the weight be not heavy enough, the correc- tion must be subtracted from the apparent weight of the substance to be weighed. Such a correction is indicated in the table with a minus sign. Conversely, if the weight be too heavy, the correction must be added to the apparent weight of the substance. Beginners most frequently trip over this. If the corrections anywhere exceed the- values indicated in the table, page 17, return the box to the makers. A good set. of weights is a treasure ; a bad set, an incubus. The way to use the table of corrections will be obvious from the following example : c )bserved weights. Corrections. Corrected weights. Crucible and substance . Empty crucible . 29-7459 . 27-6429 / -0-30\ I +0-21 / / -0-381 \ + 0-39 / 29 '7450 grms. 27-6426 grms. Weight of substance 2-1024 ffnns. The method of calculating the corrections is as follows : The 29 grms. is made up from [20] + [5] + [2] + [1] + [!]& which have respectively the correction factors -0-07, -0-23, -O08, -t-0'06, + 0'02. The 0'74 grm. was made up from the [0'5] + [0-2] + [0'02] + [0'01] + [O01] 6 with the tabulated corrections: + 0-04, +0-05, +0-05, +0-03, +0'04= +0-21. Hence, 297459 is to be increased by - 0'30 + O21, that is, by - 0'09 ; etc. In the course of time the weights gradually decrease in weight. This is due to " wear and tear." Highly polished weights are less liable to loss by wear than weights with a dull surface. 1 The weights particularly the smallest some- times become heavier. This can usually be traced to the adhesion of dirt to the metal. The following example 2 illustrates the effect of wear on a set of weights which had been in regular use for some years in a laboratory where the need for calibrating the weights was considered superfluous : 1 The Berichte der Kaiserlichen Normal-Echungskommission (Chem. Ztg., 10. 1481, 1886) mentions a change, due to internal oxidation within the blowholes of the castings, which was accelerated by the introduction of salt solutions when the weights were gilded. 2 H. D. Richmond, Analyst, 19. 99, 1894 ; B. Weinstein, Handbuch der physikalischen Maassbestimmungen, Berlin, 2. 363, 1888 ; R. C. Benner, Mining Sci. Press, 100. 492, 1911 ; T. K. Rose, The Metallurgy of Gold, London, 460, ]906. WEIGHING. 21 Table IV. Effect of Use on a Set of Weights. Face value. Actual value. Face value. Actual value. Face value. Actual value. [10]a Standard [1> 0-9989 [O'l ] 6 0-0982 [10] b 9-9997 [l]c 1-0008 [0-05] 0-0485 [5] 4-9990 [0-5] 0-4982 [0-02] 0-0191 [2] 1-9991 [0-2] 01993 [0'01] a 0-0092 [it 0-9987 [0-lja , 0-0971 [O'OIJ, 0-0089 These numbers speak eloquently of the need for checking the weights from time to time. They also show how two analysts might get different results with the same analytical process when possibly one is using faulty weights. 1 9. The Influence of Inequalities in the Lengths of the Arms of the Balance on the Weighing. In what precedes, it has been assumed that the two arms of the balance are equal in length. This is not really the case. It is a mechanical impossibility to ensure perfect equality. To find the difference in the lengths of the two arms, add weights of the same nominal value to each pan. The (corrected) weights should be about half the maximum load say, 50 grms. Bring the balance into equilibrium by means of a rider. Interchange the weights on the pan, and again bring the balance into equilibrium by means of the rider. Call the two weights W and w, and let I and r respectively denote the additional weights required for equilibrium on the left and right sides. Then, on the first weighing, w + l = W ; and W=w + r on the second weighing. Let L and R respectively denote the lengths of the left and right arms. Then, from the law of levers, L(w + l)=R.W', and L.W=R(w + r). Multiply these two expressions together, and reduce by the usual method of approximation for small quantities. We obtain : l-r I 2w * L , and =1 - (i) Suppose that the weighings with corrected weights for an ordinary analytical balance were found to be : [50] Left. Right. [20] + [10] a + [10] 6 + [10] + 0-13 mgrm. [10] 6 + [10] [00]+ 0-19 mgrm. Here then I 0*13, and r= +0'19 mgrm. Consequently, from the second equation, (1), above, L:R= 1:1-0000032. 1 The consistency of the riders can easily be checked from the 0'01-grm. weights, or a standard rider. In the above case the adjustment of the rider has been effected in terms of the given rider. T. K. Rose (I.e.) reports that he found some gilded brass riders at the Royal Mint increased in weight from 5'0 to 5 '025 mgrms. in six months. 22 A TREATISE ON CHEMICAL ANALYSIS. It is then necessary to find the magnitude of the error introduced in a weighing for a known deviation from equality in the lengths of the two arms of the balance. The inequality may be allowed for by multiplying the apparent iveight by the ratio of the lengths of the two arms ; using the ratio R : L or L : R according to the arm with which the weight is connected. This is necessary if the true weight of the object is to be determined. With the preceding ratio, L:R= 1-0000032, a weight w on the left pan will be equivalent to a weight w x 1 '0000032 on the right pan. Hence, if a substance on the left pan balances the weight 50 + 0-0001 grm. on the right pan, the weight of the load is w= 1-0000032 x 50-0001 = 50-0016 grm. an error of about 0-003 per cent. There is therefore no need to apply the correction. Each balance has its own constant R : L for a given load. The numerical value of the ratio varies with different loads. Most analytical balances do not need an allowance for variations due to in- equalities in the lengths of the arms; the lengths are usually made sufficiently exact. When the weights are always placed, say, in the right pan, the correction need not be applied in analytical work, because, although there may be a difference between the apparent and the true weight of the substances which are weighed with the faulty balance, yet this difference is constant and does not affect the value of the ratio. For instance, if R and L respectively denote the lengths of the right and left arm, and the true weights w v w 2 , W B , . . . are represented on the balance by the apparent weights p } , p 2 , p 3 , . . ., it follows, from the law of levers, that w^L =p^R ; w 2 Z =p%R ; Hence, This means that the value of the ratio of the true weights is the same as the value of the ratio of the apparent weights. That is, L L L w-, : Wn : w = -5-^ : -^ -w* : -T>W<>=P-I : p% ' p*- & K J& Consequently, when two quantities of matter are increased or decreased in the same proportion, 1 they will equally balance one another. Hence, if the iveights be always confined to one pan, the apparent weights will always be increased or decreased in the same ratio, and the results of an analysis ivill be as accurate on the defective balance as on a balance with arms perfectly equal. This method of eliminating the error due to the inequality of the arms of the balance pre- supposes that no weights are placed on the left pan for the purpose of making up a given weight in what may sometimes appear to be the easiest manner, by subtraction. 10. Corrections for the Buoyancy of Air. It is assumed that, if two things are equal in weight at the same time and place, they contain the same mass or quantity of matter. The mass of a 100-grm. weight is supposed to be 100 times the mass of a 1-grm. weight. This assumption is only true if the two substances have the same volume, or if the comparison be made in vacuo. A body weighed in air is partly buoyed up by a pressure equivalent to the weight of a volume of air equal to the volume of the body (Archimedes' principle). Suppose that a 100-grm. platinum weight (sp. gr. 21-55) be balanced against a 1 00-grm. brass weight (sp. gr. 8 -4). It is easy to show that the weight of dry air at and 760 mm. displaced by the platinum weight is 1 It is assumed that the same kind of matter is in question, or that the effect of the buoyancy of the air is eliminated. See R. Kempf, C'hem. Zty., 36. 1349, 1912. WEIGHING. 23 equal to 0-0054 grm. 1 ; while the weight of the air displaced by the 100-grm. brass weight is 0'0143 grm. Hence, if the platinum weight be exactly 100 grms., the brass weight which exactly counterpoises it will not be 100 grms., but 100 + (0*0143 -0-0054) = 100-0089 grms. Hence, the buoyancy of the air thus produces a sensible effect whenever the volume of the load differs materially from the volume of the weight. 12 The arithmetic of the above calculation is summarised in the formula : Corrected weight = w + o>( (i) where w represents the apparent weight of the object ; s, the specific gravity of the object being weighed; s p the specific gravity of the weights; and w, the weight of a cubic centimetre of air at the temperature, pressure, and humidity prevailing at the time of observation. EXAMPLE. A porcelain crucible (sp. gr. 2'3) weighs w?= 7*5392 grms. when the 7-grm. weights are brass (Sj = 8'4) and the 0-5392-grm. weights are platinum (s 2 = 21-55). What is the corrected weight of the body when = 0-0012 ? With the 7-grm. brass weights, Correction = 7 x 0-0012 -i --- ~ =0 '00245 grm. ; \2"3 8"4/ and with the 0'5392-grm. platinum weights, Correction = 0'05392 xQ'QQuf - --- 1\ = 0'00030 grm. \2'3 21'55/ Consequently, the corrected weight is 7'5392 + 0*0026 -f 0'0003 = 7'5421 grms. To illustrate the effect of the buoyancy of air on the different precipitates usually weighed in clay analyses, the following table may be quoted : Table V. Effect of Buoyancy of Air on the Weighings of a Clay Analysis. Error per gram of substance weighed. Substance weighed. Specific gravity. Brass weights. Platinum weights. Clay . 2-55 0-00033 0-00041 Silica .... 2-23 0-00039 0-00048 Alumina .... 3'85 0-00017 0-00028 Ferric oxide ... 5-12 o-oooio 0-00018 Magnesium pyrophosphate . Calcium oxide ... 2-40 2-90 0-00036 0-00027 0-00044 0-00035 Potassium chloride 1-99 0-00046 0-00065 Sodium chloride 2-13 0-00042 00054 Potassium chloroplatinate . 3'34 0-00021 0-00027 1 Thus, the 100-grm. platinum weight has weighs 0-0012. Hence 0*0012 x 4 '5 0*0054 grm volume 100 -"-21-55 = 4 -5. One c.c. of air ^ Of course, we ought to take the weight in vacuo, but the error introduced by taking the apparent weight in air is negligibly small. If the correction be O'OOl grm., the error only amounts to '000001 grm. 2 The buoyancy correction appears to have been first used by E. Turner about 1830 in determinations of specific gravities and atomic weights (Phil. Trans., 119. 291,1829; 123. 523, 1833). J. J. Berzelius used the correction in some early work, but later regarded the correction as trifling. W. Crookes, Chem. News, 15. 191, 1867 ; 29. 29, 1874 ; C. W. Folkard, ib., 29. 30, 1874 ; F. C. Cloud, ib., 35. 133, 1874 ; H. R. Proctor, ib., 30, 255, 1874 ; Trans. Newcastle Chem. Soc., 2. 188, 1873; L. L. de Koninck, Chem. Ztg., 18. 1816, 1894; J. P. Cooke, Zeit. anal. Chem., 23. 187, 1884 ; Chem. News, 48. 39, 1883 ; W. H. Miller, Phil. Trans., 146. 753, 1856; G. F. Becker, Liebig's Ann., 195. 222, 1879. To eliminate the buoyancy correction in weighing bulky glass apparatus use a similar apparatus as a tare see page 550. 24 A TREATISE ON CHEMICAL ANALYSIS. The table shows that when the amount of a precipitate is determined from the difference in the weight of an empty crucible and of the crucible plus residue, and the atmospheric conditions are the same when the two weighings are made, the buoyancy correction is not needed for small precipitates, nor for precipitates with a specific gravity not sensibly different from that of the substance under- going analysis. In ordinary analytical operations we have to deal with differences in weight, not with absolute weights. For instance, an empty platinum dish weighs 63 grms. It is weighed with brass weights (sp. gr. 8-4). A litre of water is evaporated to dryness in the dish, and the dish weighed again. The difference in the two weighings represents the solid matter derived from the water. If there be a considerable difference between the specific gravity of the sub- stance undergoing analysis and the weight of the ignited precipitate, an allowance must sometimes be made for the different effects of the buoyancy of the air on the different substances. For instance, in determining the amount of sulphur in a sample of pyrites by weighing the precipitated barium sulphate, the error due to the buoyancy of air will be negligibly small. This arises from the fact that the pyrites and the barium sulphate have nearly the same specific gravity, and, con- sequently, nearly the same buoyancy correction. On the contrary, in standardis- ing a solution of silver nitrate by precipitating silver chloride from a given volume of the solution, the buoyancy of air may affect the result by y^th per cent. Effect of Variations of Temperature and Pressure on the Buoyancy of Air. In the experiment cited above, we assumed that the weight of the empty platinum dish was affected by the air to the same extent before and after the evaporation. It is necessary to examine the validity of this hypothesis. The evaporation takes so long that we cannot reasonably assume that the temperature and barometric pressure have remained constant. The weight of a given volume of air, of course, depends upon the temperature and pressure. Hence the buoyancy of the air is affected by these two factors. The brass weights, 63 grms. in the above example, occupy a volume of 7 '5 c.c., and the material of the platinum dish occupies a volume of 2*99 c.c. Hence, 7*5 - 2'92 = 4'58 c.c. more air is buoying up the brass than the platinum weights. The weight of 4-5 c.c. of air at 760 mm. and 15 is 4-58 x 001227 = 0*00552 grm. Consequently, the 63-grm. brass weights really represent 63 '00552 grms. Suppose that the barometer falls to 740 mm. and the temperature rises to 25, the weight of the 4*5 c.c. of air 1 will then be 4-5 x 0-001173 = 0-00528 grm. Hence, 0'00552 -0*00528 = 0'00024 grm. represents the decreased buoyancy of air during the second weighing. Hence, if the dish weighed just 63 grms. the first weighing, it would weigh 63-00024 grms. the second weighing. Assuming the two atmospheric conditions just indicated, and assuming that the observed weights were : Dish plus residue 63 '00048 grms. Empty dish . 63 '00000 grms. Residue . . . . . . . 0-00048 grm. we might be led to assume that the litre of water contained the equivalent of 0-00048 grm. of solid matter, when in reality it only contained 0*00024 grm. The apparent weight of the dish has increased 0-00024 grm. owing to the decreased 1 The weight in grams of a litre of dry air at any temperature and pressure is readily com- puted by dividing the pressure expressed in mm. of mercury by 273 added to the temperature in C. and multiplying the quotient by 0-4644. The arithmetic is summarised in the formula : Density of dry air at t and p mm. pressure = 1293 . JL . The product of the volume of the substance and the weight of 1 c.c. of dry air represents the weight of a volume of dry air equal to the volume of the substance. WEIGHING. buoyancy of the air. If, therefore, we make any pretence to accuracy in the fourth place, we must not only allow for the buoyancy of the air, but we must also allow for the changes in the buoyancy of the air with variations of temperature and pressure. 1 Potash bulbs and calcium chloride tubes, in virtue of their relatively large bulk, might easily lead to errors of 1 per cent., owing to the neglect of the variations in the buoyancy of air under changing atmospheric conditions. Fresenius, in his classical Anleitung zur quantitativen chemischen Analyse, after showing 2 that, in analytical work weighings are usually carried to the y^th mgrm., refers to the correction for the buoyancy of air in these terms : " This defect is so very insignificant, owing to the trifling specific gravity of air in proportion to that of the solid substance, that we generally disregard it altogether in analytical experiments." Fresenius' remark is also true of analytical practice at the present day, and it is equally true that weighings below 0*0005 grm. are then of no particular value, and this altogether apart from the inaccuracies mentioned on page 15 arising from the particular character of the analytical process. Effect of Variations of Humidity on the Buoyancy of the Air. The humidity of the air is also constantly changing. We must therefore inquire what effect this has on the buoyancy of the air. 3 The preceding calculations were referred to dry air at the temperatures and pressures indicated. As a matter of fact, the buoyancy of moist air is less than that of dry air ; hence, the correction for moist air will be less than for dry air. In illustration, 4 a crucible kept in the balance case during the summer months July and August showed the following variations in weight : Table VI. Effect of Humidity of Air on Weighings. Weight in air. Weight corrected for the buoyancy of dry air. Weight corrected for buoyancy of moist ak. Maximum ..... grms. 39-35659 39-35601 grms. 39-37046 39-37040 grms. 39-37061 39*37052 Mean ...... Deviations from mean . . . \ 39-356362 + 0-000228 -0-000352 39-370437 + 0*000023 -0-000037 39-370562 + 0-000048 -0-000042 Hence, the average effect of variations in moisture is somewhere near 0-00013 grm. In one of the experiments, a variation of 10 per cent, in the humidity of the air only altered the corrected weight 0-00003 grm. That is outside the range of weighings conducted to T ^th milligram. Applying our test the errors in weighing should fall within the limits of experimental error due to the 1 THE EFFECT OF GRAVITATION ON THE BUOYANCY OF AIR. Strictly speaking, the weight of a litre of dry air at 760 mm. differs in different localities. This is not merely because the action of gravity on a litre of air is different in different places, but because a pressure of 760 mm. of mercury varies with the intensity of gravitation. Thus, the intensity of gravity at Greenwich is to that at Paris as 3457 : 3456. Hence, if a litre of dry air at Paris weighs 1-2932 grms., a litre of air at Greenwich will weigh 1*2936 grms. under the same conditions of pressure and temperature. Hence, the effect of gravitation on the buoyancy correction is negligible. The small difference in the buoyancy of air in the two places has no appreciable effect when applied to particular cases. 2 R. Fresenius, Quantitative Chemical Analysis, London, I. 18, 19, 1876. 3 The fallacy of assuming that the air inside a balance case is dry simply because a couple of jars of calcium chloride or of sulphuric acid are in the case will appear from page 10. 4 0. Kuhn, Chem. Ztg., 34. 1097, 1108, 1910 ; R. Kempf, ib., 36. 1349, 1912. 26 A TREATISE ON CHEMICAL ANALYSIS. analytical process it follows that it is necessary to allow for variations in the buoyancy of air due to variations of temperature and pressure when the weighings are conducted to the T Vth of a milligram, but further variations due to humidity can be neglected under these conditions. It is generally convenient to assume that the average weight of a cubic centimetre of moist air is 0'0012 grm. 1 Remembering that for general purposes 0-0002 0-0004- 0-0006 FIG. 8. Buoyancy correction for substances of different specific gravity. 0-0008 0-0010 000/2 Correction Factor the weights are calibrated in air, and that we assume the buoyancy is a constant corresponding with the weight 0'0012 grm. per c.c., we are only concerned with variations in the buoyancy of the object being weighed. 2 In that case, we use, in place of (1), Corrected weight = (2) 1 If / denotes the actual pressure (in mm. mercury) of the vapour present in a unit volume of moist air (saturated or unsaturated), the weight of the moist air is determined from the formula page 24 by substituting^? - f/ in place of the barometric pressure p. We thus obtain : Density of moist air at t and p mm. pressure = j Remember that the mass of vapour present in a given volume of moist air saturated at a given temperature depends upon the temperature, and is independent of the pressure ; and that the pressure of water vapour at the dew-point is nearly the same as the pressure / at the given tem- perature. The value of / is the maximum pressure of the water vapour at the dew-point, and this can be obtained from a table of the vapour pressures of water at different temperatures (Table LXXXV.). Hence, the relative humidity of air is the ratio of the maximum pressure of the aqueous vapour, /, at the dew-point, to the maximum pressure, F, at the actual temperature of the air ; or Relative humidity = ^ . EXAMPLES. (1) What is the relative humidity of air at 15 when the dew-point is 10? From Table LXXXV., the maximum pressures of the aqueous vapour at these two temperatures are respectively 1279 and 9 -21. Hence, 9 '21 -=-1279=: 072, or 72 per cent, humidity (2) What is the density (or the weight of 1 c.c.) of moist air at 15 when the dew-point is 10, and the barometer 759 '14 mm. ? From Table LXXXV., the maximum pressure of water vapour at 10 'is 9'21 mm. Hence, by substituting/^ 9 "21, * = 15, and ^ = 759'14 in the above formula, it follows that 1 c.c. of moist air at 15 weighs '001219 grm. 2 W. Dittmar, Exercises in Quantitative Chemical Analysis. Glasgow 8 1887 P Schott lander, Zeit. phys. Chem., 16. 458, 1895. WEIGHING. 27 By making w = 1 grm. and plotting the corrected weight for different values of s (specific gravity), we get the curve shown in fig. 8. This enables the correction for a given weight of substance to be seen at a glance. EXAMPLE. A silica "precipitate" weights O6021 grm. What is the corrected weight if the specific gravity of calcined silica is 2*25 ? From the diagram (dotted lines), we see that the correction factor is 1'0005 per grm. Hence, 1-0005 x 0-6021 = 0'6024 grm. is the corrected weight. II. Summary. The important lesson we learn from this chapter is that in analytical chemistry, as well as in other work involving delicate measurements, it is always advisable to make, at least, a rough estimate of the influence of the various sources of error on the final result. These errors can only be neglected when their effect is small in comparison with the error derived from other sources. The chief sources of error commonly introduced in the balance room are those arising from : (1) Variations in the zero point of the balance; (2) Inconsistent weights ; (3) Inequalities in the lengths of the arms of the balance ; and (4) The buoyancy of the air. We have seen that in weighings making any pretence to accuracy to the " second decimal " or to the " T Vth milligram" (1) The zero point of the unloaded balance should be determined and the weighing made by the method indicated in the text 1 (pages 10 to 13). (2) The weights should be calibrated, and periodically checked (pages 16 to 21) for consistency amongst themselves. (3) The error due to the inequality in the arms of the balance can be neglected in ordinary analytical work (page 21). (4) The correction of the weighings for the buoyancy of air is necessary when the determination involves the weighing of substances with appreciably different specific gravities (page 23). In general analytical work this correction can be neglected, since the resulting error is overshadowed by the errors associated with the preparation of the precipitates for the balance. 1 Or, of course, some of the other methods indicated by name on page 1 1. CHAPTER II. THE MEASUREMENT OF VOLUMES. 12. Volumetric and Gravimetric Processes of Analysis. MEASUREMENTS with a good balance and weights can readily be conducted with a precision far greater than is needed for general analytical work. It has been pointed out that the errors involved in the preparation of a troublesome preci- pitate for weighing will impair the value of an exact weighing. The measure- ment of volume, in volumetric analysis, may not be so precise and reliable as the measurement of weight, yet the results of volumetric processes, based on suitable reactions, are frequently more trustworthy than gravimetric processes, because the method of preparing the substance for the measurement by volu- metric processes is less liable to error. There is a prejudice in some minds against volumetric processes. It is claimed that in general "there is a lack of precision in volumetric analysis." In many cases the prejudice arises from the fact that some essential precautions have been neglected ; in other cases there may be an intrinsic weakness in the method itself. There is also some confusion possible in the different systems of measurement. With proper precautions many volumetric processes are excellent, and, for technical analyses, invaluable. Indeed, in technical work, where time is an essential factor, volumetric processes are used in preference to gravimetric wherever expedient. 1 It is important to have a clear idea of the precautions necessary for a high degree of accuracy, in order that errors of vital importance maybe eliminated. The need for the "analytical perspective," mentioned on page 16, is here of great moment. 13. The Influence of Variations of Temperature. The volume of a solution depends on the temperature. The higher the temperature, the greater the volume of a given weight of a solution. A glass measuring vessel also has a greater capacity the higher the temperature. Although the level of a solution in, say, a litre flask may rise above the litre mark as the temperature rises, this naturally occurs because the expansion of the water more than counterbalances the increased capacity of the flask. If the expansion of the glass were greater than the expansion of the solution, with rise of temperature, the level of the liquid in a standard flask would sink below the mark on the neck of the flask. 2 The liquid would appear to contract. The volume measured at any temperature, different from the standard temperature, is the joint effect of the changed capacity of the flask and the changed volume of the liquid. 1 R. Hasenclever, Ber., 33. 3827, 1900. 2 Text-books on Heat describe experiments illustrating this point. 28 THE MEASUREMENT OF VOLUMES. 29 Let us take the coefficient of cubical expansion of glass l to be 0-000025 ; and Thiesen, Scheel, and Disselhorst's 2 values for the volume of water at different temperatures indicated in the subjoined table (Table VII.). The Table VII. Volume and Density of Water at Different Temperatures. Temp. - Volume. Density. Temp. Volume. Density. 1-0001324 0-9998676 20 1-0017728 0-9982303 1 1-0000734 0-9999266 21 1-0019853 0-9980186 2 T0000320 0-9999680 22 1 -0022083 0-9977966 3 1-0000078 0-9999922 23 1-0024414 0-9975645 4 1-0000000 1 -0000000 24 1-0026847 0-9973225 5 1-0000082 0-9999918 25 1-0029378 0-9970708 6 1-0000320 0-9999680 26 1-0032006 0-9968097 7 1-0000707 0-9999293 27 1-0034729 09965391 8 1-0001241 0-9998759 28 1-0037546 0-9962594 9 1-0001917 0-9998084 29 1 -0040455 0-9959708 10 1-0002730 0-9997271 30 1-0043456 0-9956732 11 1-0003678 0-9996324 31 1-0046546 0-9953670 12 1-0004756 0-9995246 32 1-0049724 0-9950522 13 1-0005962 0-9994041 33 1-0052989 0-9947290 14 1-0007292 0-9992713 34 1-0056341 0-9943975 j 15 1-0008744 0-9991264 35 1-0059777 0-9940578 16 1-0010314 0-9989697 36 1-0063297 0-9937101 17 1-0012000 0-9988014 37 1-0066899 0-9933545 18 1-0013799 0-9986220 38 1-0070584 0-9929911 19 1-0015709 0-9984315 39 1-0074349 0-9926200 number of cubic centimetres of water which must be subtracted from or added to 1000 c.c. of water to give the volume of distilled water at t, in order to fill a litre flask calibrated at 15, is indicated in the following table 3 (Table VIII.). The data have been calculated from the preceding Table VII. 4 Table VIII. Reduction for the Change of the Apparent Volume of a Litre of Water with Temperature. (Standard temperature = 15.) Temp. 1 2 3 4 5 6 7 8 9 + 0-63 4-0-62 40'60 40-57 + 0-53 1 2 3 + 0-48 -077 -3-10 + 0-41 -0-96 -3-38 40-32 -1-16 -3-67 + 0-22 -1-37 -3-97 + 0-12 -1-58 -4-28 Unit. -1-81 -4-60 - 0-13 -2-05 -4-93 -0-28 -2-30 -5-27 -0-43 -2-57 -5-61 -0-60 -2-82 -5-96 1 The value for glass used in volumetric work ranges from 0-000023 to 0-000028, according to the nature of the glass. The error introduced by these variations is negligibly small. 2 M. Thiesen, K. Scheel, H. Disselhorst, Wiss. Abh. phys. tech. Reich., 3. 1, 1901. 3 Fora table based upon some older data, see P. Casamajor, Journ. Amer. Chem. Soc., I. 188, 1876 ; 2. 19, 1877 ; Chem. News, 35. 160, 170, 1877 ; 38. 137, 1879. See also W. Schlosser, Zeit. angew. Chem., 17. 953, 977, 1004, 1904; 21. 2161, 1908; Chem. Ztg., 29. 509, 1904; Zeit. anal. Chem. , -46. 392, 1907 ; H. L. Payne, Journ. Anal. App. Chem., 6. 326, 1892. 4 If a " temperature of reference" other than 15 be adopted, another table can be easily compiled from this. For instance, suppose 20 be adopted, 15 becomes 40 '77, 25 becomes - 1 -04, etc. A TREATISE ON CHEMICAL ANALYSIS. The use of Table VIII. may be illustrated by the following examples : EXAMPLES. (1) 25 c.c. of water are measured in a burette at 28 : what is this volume at the standard temperature, 15 ? From Table VIII. it follows that since 1000 c.c. requires a correction of - 2'57 at 28, 25 c.c. will require a correction of - 0'06. Hence, 25 0-06 = 24-94 c.c. is the required volume. (2) What is the volume at 15 of a litre of water measured in a flask at 10, and also at 18? From the table, the correction for 1000 c.c. at 10 is + 0'48, and at 19, -0'43. Hence the required volumes are 1000 + 0*48 = 1000' 48 c.c. at 15, and 1000-0-43 = 999-57 c.c. at 18. Table VIII. thus represents the volume which a litre of water alters in a glass vessel when its temperature changes from the standard 15. If the pre- vailing temperature does not differ by more than, say, 3 from the standard 15, the changes may be ignored. These numbers differ so little from the results with iN- and ^N-solutions that they may be also used for such solutions. Normal and still more concentrated solu- tions require special tables, since the deviations are then greater see, for example, Tables XIII. and XXXV. In weighing a litre of water in a glass vessel, the most important source of error arises from the variation of temperature. A difference of 1 will give a difference of 0'02 c.c., correspond- ing with 0'002 per cent. A variation of the barometer of 15 mm. of mercury will cause an error of two units in the second significant figure, and this will affect the measurement by 0-014 c.c. Hence, barometer variations can gener- ally be neglected. In titrating hot liquids, there is a danger of heating the burette and its contents. In such cases, burettes with a delivery jet at the side (fig. 9) are recommended. Koninck's burette 1 is an ordinary burette with a , '-shaped jet. The idea is to prevent clouding the burette with condensed steam, and the heating of the liquid in the burette, which would cause an expansion of the contained solution. 14. The Litre and its Subdivisions. Owing to the great difficulty in measuring directly the relation between cubic capacity and the unit of length, the International Committee of Weights and Measures, 2 in 1880, recommended that the word " litre " be used to represent " the volume of a kilogram of pure water at its maximum density"; and in 1901 this definition was amplified to read : " The litre is the volume occupied by the mass 1 L. L. de Koninck, Zeit. angew. Chem., I. 187, 1888. 2 Procts-Verbaux Oomite Internat. Poids et Mesures, 30, 1880 ; 175 1901 P Chappius ib (2), 2 .72 1903 ; J. At de Lepinay Jour*. Phys., (3), 5. 477, 1897 ; D. Mendel'eeff, Proc. Roy. Soc., 59. 143, 189o ; W. Schlosser, Zeit. angew. Chem., 16. 960, 1903. FIG. 9. L. L. de Koninck's burette. THE MEASUREMENT OF VOLUMES. 31 of one kilogram of pure water at its maximum density and under normal atmo- spheric pressure." l The litre so denned is a little larger about 50 c.mm. than a cubic decimetre. Evidently, then, the cubic centimetre is not really the thousandth part of a litre, but a little smaller than the thousandth part. Hence, if it be desirable to distinguish between a "millilitre" and a "cubic centimetre," ,it would be necessary to abandon the latter term in analytical chemistry, where the unit of volume is based on the definition indicated above. 2 Unfortunately, there are a number of different litres used in laboratories. The more important are the following : (1) Normal litre. The volume of 1000 grms. of air-free water (weighed in vacuo} at + 4. If, therefore, the litre flask is marked correctly at + 15, this means that at a temperature of 15 the capacity of the flask is the same as the volume of a kilogram of water at +4. This is the normal, true, or international litre, used in this work. 3 (2) Mohr's litre. The flask contains the same number of grams of water (free from air) at the specified temperature as the number of cubic centimetres or grams stated on the flask. For instance, the Mohr's litre is " the volume occupied by a kilogram of water, at 17*5, when weighed in air with brass weights." 4 It matters very little, in analytical work, which litre be adopted, but it is necessary to keep rigorously consistent to the one system in all volumetric apparatus burettes, pipettes, etc. This matter is serious enough to require emphasis, since apparatus, if not specifically ordered, may be supplied by the dealers, at different times, graduated according to different systems, and mixed graduations may thus be introduced on one bench. We can get some idea of the magnitude of the error which would be introduced in confusing the different systems, by noting that the two "litres " indicated above are related as (1) 1000 ; (2) 1002-3 c.c. Hence the normal litre is to the Mohr's litre 5 as 1 : 1-0023. Standard Temperature. It would be exceedingly inconvenient, if not im- possible, to weigh exactly a kilogram of water at + 4 in vacuo ; nor is it practicable to work at + 4. Some other more convenient temperature must be selected, and the necessary corrections made. In this work, 15 is supposed to be the standard temperature. 6 This temperature is recommended by the Kaiserliche Normal-Echungskommission (K.N.E.K.), 7 Berlin ; 17'5 was recom- mended by Mohr, 8 and by Tread well ; 18, by Lunde'n ; 9 and 20 is recommended 1 The temperature of maximum density is not denned it appears to be a shade under 4 C. nor does the definition make any reference to weighing in vacuo. In the calculations it is assumed the weighings are in vacuo. 2 T. W. Richards, Journ. Amer. Chem. Soc., 26. 413, 1904. 3 At 15, therefore, this litre does not represent 1000 grms. of water, but rather 993 '069 grms. 4 F. Mohr, Zeit. anal. Chem., 7. 285, 1868 ; H. Beckurts, Die Methoden der Massanalyse, Braunschweig, 42, 1910. Mohr made no allowance for variations in the atmospheric conditions. 5 First show (page 23) that 1000 grms. of water weighed in air with brass weights at 17 "5 are equivalent to 1001 *1 grms. if weighed in vacuo. Then, since the density of water at 17 "5 is 0-99875 (Table VII.), it follows that 99875 grms. of water at 17-5 will represent a normal litre at 17*5, for it will occupy the same volume as 1000 grms. of water at 4. Hence, if 99875 grms. represent 1000 normal c.c., 1001*1 grms. will represent 1002'3 normal c.c. 6 The tables can, of course, be modified for other appreciably different climates. If the flasks, etc. , were cooled for weighing, there would be a danger of the deposition of moisture on their surface. 7 Zeit. angew. Chem., 6. 557, 1893 ; L. A. Fischer, Journ. Amer. Chem. Soc., 20. 912, 1898 ; C. Laurent, Bull. Soc. Chim. Nord France, I. 20, 1891 ; H. Ost, Chem. Ztg., 14. 1747, 1899 ; W. Fresenius, Zeit. anal. Chem., 30. 461, 1891 ; A. Classen, Theorie und Praxis der Mass- analyse, Leipzig, 64, 1912. 8 F. Mohr,LehrbucJiderchemisch-analytischen Titriermethoden t Braunschweig (H. Beckurts' edition), 42, ]910; F. P. Treadwell, Kurzes Lehrbuch der analytischen Chemie, Leipzig, 2. 432, 1911. 9 H. Lunden, Svensk Kern. Tids., 24. 96, 1912. 32 A TREATISE ON CHEMICAL ANALYSIS. by Wagner, 1 and the U.S. Testing Laboratory. 2 The right choice should be determined by the prevailing temperature of the laboratory. Apparatus in- tended for working in warm climates may be graduated for 20, 22, or 25 ; and 15 or 17*5 for cooler climes. From the discussion on page 31 it is easy to find what amount of water must be weighed at any given temperature to have a volume equal to the volume of a kilogram of water at + 4. Given the specific gravity of air, say, 0*0012 kilogram per litre ; the specific gravity of the brass weights, 8*4 ; the coefficient of cubical expansion of glass, '000025. Let S denote the specific gravity of water at t Table VII. To find what amount of water at any specified temperature must be iveighed in air with brass weights to give a normal litre. Since a kilogram of water at + 4 weighed in vacuo occupies 1 litre, it follows from the discussion on page 22 that a litre of water will weigh w = S- 0-001 2l - kilogram . . . (1) when weighed in air with brass weights. But a flask exactly 1 litre at 15 has a capacity of 1 + 0'000025( - 15) at t. Hence, it is necessary to weigh W=[l +0-000025($- 15)] \S- O'OOl 2^1 - \\ kilogram . (2) at t in order to find what the capacity of the flask would be if the water had been weighed in vacuo at 4 with brass weights. EXAMPLE. What amount of water must be weighed in a flask at 15 in order that a mark can be made on the neck of the flask to represent the volume of a kilogram of water weighed at 4 in vacuot From Table VII., page 29, /S=0'999126, and = 15. Substitute these numbers in the preceding formula (2). We get a weight 0*99806 kilo- grams. In practice, therefore, the flask is counterpoised on the balance, a kilogram weight is placed on the right pan, and 1000 - 998*069 = 1 '931 grms. on the left pan. When the flask has sufficient water to restore equilibrium, the volume of the water in the flask represents the volume of a kilogram of water at 4 when weighed in vacuo. In order to avoid the labour of calculation, Schlosser 3 has computed a table of values ranging from 5 to 30 '9, referred to air and water at 15, and the barometer at 760 mm. 15. Meniscus, Parallax Errors, and Burette Floats. Parallax Errors. In reading the level of liquids in burettes, pipettes, etc., the lower boundary of the dark meniscus is taken as the normal reading. Parallax errors must be carefully watched. If the eye be not in the same horizontal plane as the meniscus, BB' (fig. 10), the reading may be either too high or too low. For instance, if the eye be looking in the direction A A' the reading will be too high ; and if in the direction CC', the reading will be too low. The direction BB gives a correct reading free from parallax. In the diagram, the ^correct reading is 30 '0 c.c; the two readings affected by parallax errors are AA', 30-3 c.c., and CC', 29 '85 c.c. If the mark to be read encircled the burette, the eye would see only one horizontal line when at the proper level for reading. The marks rarely encircle the burette, and some auxiliary is therefore needed. 4 Meniscus Screens. The level of the meniscus is rendered more distinct by placing the burette so that it has a white wall as background and shading the 1 J. Wagner, Zeit. phys. Chem., 28. 193, 1899. 2 Circular Bureau of Standards, 9. 2, 1904 ; N. S. Osborne and B. H. Veazev. Bull. Bur. Standards, 4. 553, 1908. 3 W. Schlosser, Zeit. angew. Chem., 16. 960, 1903 ; Chem. Ztg., 27. 4, 1904. 4 A reading lens is often a great help in exact work. THE MEASUREMENT OF VOLUMES. 33 light from below by holding two fingers behind the burette just below the meniscus, 1 or Mohr's simple and effective plan of holding a piece of paper, C ^ B< FIG. 10. Meniscus errors. partly blackened, with the blackened portion just below the meniscus, as illustrated in fig. II. 2 FIG. 11. Mohr's shaded screen. FIG. 12. Rohrbeck's burette. Burettes made to Accentuate the Meniscus. Rohrbeck's burettes (fig. 12), made 1 B. Reinit/er, Zeit. angew. Ohem., 7. 547, 573, 643, 1894. 2 J. H. M'Mahon (Chem. News, 45. 109, 1882; F. A. Gooch, Amer. J. Science (3), 44. 239, 1892 ; P. Kusnetzotf, Zeit. anal. Chem., 46. 515, 1907 ; A. Prinzl, Dent. Amer. apoth. Ztg., 4. 637, 1885; F. Kb'hler, Brit. Pat. No. 10936, 1903) recommends a small mirror on the surface of which are pasted two parallel strips of black paper each, say, 2 in. long and J in. wide, with a space about \ in. wide between them. When reading, place the mirror behind and in contact with the burette, in such a position that the reflection of the eye can be seen in the space between the strips of paper. This ensures the eye being at right angles to the burette; J. F. Sacher, Chem. Ztg., 35. 622, 1911. 3 34 A TREATISE ON CHEMICAL ANALYSIS. with a flat instead of a cylindrical tube, make errors due to parallax less likely. The meniscus with these burettes comes out very clearly. Schellbach l re- commends the use of burettes with two narrow white enamel longitudinal strips Eye too low. Eye correct. FIG. 13. Meniscus errors Eye too high. separated by a dark-coloured (blue) strip down the back of the burette. A fine point appears symmetrically placed between the strips, and at the boundary of the meniscus (fig. 13) 2 when the eye is in the right position for a reading. If FIG. 14. Gockel's burette screen. the eye be too high or too low, the point does not appear sharp, but is more or less ill-defined and blunted. A few minutes' trial will soon show that there is a fairly wide range of height and depth within which the point would be adjudged correct for a reading. In spite of this, these burettes are much used. 3 1 P. Schellbach, Chem. Ztg., 8. 1515, 1885 ; J. Milbauer, ib., 35. 419, 1911 ; J. F. Sacher ib 35. 622, 1911 ; Goetze, Zeit. anal. Chem., 50, 373, 1911. It will be noticed that Schellbach's burette does not read the lowest level of the meniscus. 2 Vernier burettes C. Meinecke, Chem. Ztg., 16. 792, 1892; A. F. Reid, Chem. News, 65. 125, 1892. 3 Some testing stations refuse their imprimatur on these burettes. THE MEASUREMENT OF VOLUMES. 35 Gbckel's screens 1 (fig. 14) are quite satisfactory for reading the level of opaque liquids, and for avoiding the parallax errors in other solutions. The Gockel screen is a blackened clamp which grips the burette two or three milli- metres below the level of the meniscus. The blackened part cuts off extraneous light from below, and leaves a sharp boundary-line at the meniscus. A sheet of white paper held behind the burette or fixed on to Gockel's screens as shown in the diagram still further emphasises the meniscus. A reading is made with the eye placed so that the front and back edge of the screen coincide. The Gockel's screen is also useful for avoiding parallax errors with opaque liquids where the meniscus is not visible. Burette Floats. Cylindrical floats arranged to swim on the surface of the liquid in the burette are marked so that the juxtaposition of the mark on the float with the graduations of the burette gives the required reading. The floats are supposed to eliminate dangers from parallax errors, and troubles with the meniscus. The ordinary Erdmann's float 2 is a plain cylinder. It has a tendency to stick to the inner walls of the burette, and is also inclined to assume an oblique position in the burette if it is not perfectly weighted. The simple float has therefore been condemned. Volhard placed a number of glass spines over the outer wall of Erdmann's float, with the idea of preventing the float sticking to the inner walls of the burette. Beutell's spherical floats 3 furnish good results provided they are properly weighted so as to swim vertically in the liquid. It is possible to read to O'Ol c.c. in a good burette with a good Beutell's float provided the bore of the tube is not too wide, and the burette is suitably graduated. This float sticks to the side of the burette when the burette is emptied, and therefore, when the burette is to be refilled, the float must be taken out, carefully cleaned, and returned to the liquid. Diethelm 4 added an extra bulb to Beutell's float with the idea of making it unnecessary to remove the float every time the burette is refilled. In Key's float, a small second bulb projects above the surface of the liquid and carries a mark. Hence, this float can be used for opaque liquids and dark-coloured solutions. The mark on the bulb above the surface of the liquid is used for reading differences of level of the liquid in the burette. The principal objection to floats is the difficulty in getting them properly weighted. A float which works satisfactorily with one liquid may be quite unsuited for another. There is an element of chance in purchasing a good float. 16. The Calibration of Standard Flasks. Just as in gravimetric work the accuracy of the weights is not to be taken on trust, so must the exactness of the measuring apparatus in volumetric analysis be established by trial. It is not at all uncommon to find that the weights in a laboratory have been the subject of wise scepticism, while the volumetric instruments have been accepted with blind faith, and conversely. 1 H. Gbckel, Chem. Ztg., 27. 1036, 1903; J. Bergmann, ib., n. 853, 1898; G. Lunge, Zcit. angew. Chem., 17. 198, 1904 ; G. Kottmayer, Zeit. anal. Chem., 30. 327, 1891. 2 0. L. Erdmann, Journ. prakt. Chem. (1), 71. 193, 1857 ; J. Volhard, Liebicfs Ann., 176. 240, 1875 ; A. Schulze, Zeit. anal. Chem., 21. 167, 1882 ; 26. 626, 1887 ; A. Gawalovski, ib., 38. 237, 1899 ; 22. 240, 1883 ; R. Benedikt, Chem. Ztg., 16. 217, 1892 ; N. Wolff, ib., 13. 389, 1889 ; F. Musset, Pharm. Centr. (3), 16. 459, 1885. C. Meissner (Chem. Centr. (3), 18. 135, 1887) has a float fitted with a thermometer. 3 A. Beutell, Zeit. angew. Chem., 2. 8, 1889. 4 B. Diethelm, Chem. Ztg., 26. 607, 1902 ; H. Key, Ber., 24. 2098, 1891 ; L. W. Andrews, Zeit. anorg. Chem., 26. 179, 1901 ; J. Wagner, Zeit. phys. Chem., 28. 193, 1899 ; P. Kreitling, Zeit. angew. Chem., 13. 829, 990, 1900; 15. 4, 1902; W. Schlosser, ib., 16. 953, 962, 977, 1004, 1903; G. Lunge, ib. t 13. 936, 1900 ; H. Thiele, Zeit. o/ent. Chem., 6. 172, 1902. 36 A TREATISE ON CHEMICAL ANALYSIS. It matters little what litre be adopted as a standard, but it is of the greatest importance to have the pipettes, burettes, and measuring flasks rigorously consistent with one another. The litre flask, for example, must hold 10 times as much as the 100-c.c. flask; 20 times as much as the 50-c.c. pipette; 200 times as much as the 5-c.c. pipette; etc. Quite satisfactory calibrations are made by the National Physical Laboratory and other recognised testing stations, although some chemists prefer to test or calibrate their own apparatus. True, this may not be done with the precision exercised by officials accustomed to the work with every facility close at hand ; but pipettes and burettes may be quite satisfactorily calibrated with the regular laboratory outfit. 1 The calibration of litre flasks requires a balance large enough to weigh a kilogram of water, and sensitive enough to indicate 0'02 grm. An ordinary analytical balance will suffice for small flasks. 2 A litre flask with a long cylindrical neck, between 14 and 20 mm. in diameter, is thoroughly cleaned 3 and dried. The empty dry flask is balanced by a suit- able tare. Fill the flask with distilled water until the lowest point of the meniscus is in the same horizontal plane as the mark on the neck. Let the flask stand in the balance case until the temperatures of the air and water are within 1. Adjust the level of the water in the flask. Remove any excess of water by a piece of capillary tube. Do not allow drops of water to adhere to the interior neck of the flask ; if any should be present, they may be removed by means of a piece of filter paper tied round the end of a glass rod. Weigh the flask and contents. Take the temperature of the water with a thermometer reading to 0*1- The mean of three weighings may be taken as the correct value. Suppose the mean of three weighings to be 998 '95 grins, (brass weights in air) ; temperature of the water 15 ; barometer, 760 mm. Since a litre of air at 15 and 760 mm. weighs 1*13 grms., the flask by formula (2), page 32 will weigh 998-95 + 1-13 = 1000*08 grms. The brass weights have probably been adjusted for weighing in air. In that case, we neglect the air displaced by the weights. If the weights have been reduced to weight in vacuo, we must make an allowance as indicated page 23. 4 It is not necessary to have the temperature of the water just 15. When we know the weight of a given volume of water under any conditions of temperature, it is easy to calculate the corresponding amount of water for any standard temperature from Table VII., page 29. 1 As a rule, the analytical chemist cannot spare much time for calibrating his apparatus, and the cost in time will be greater than the charge made by the National Physical Laboratory, which has every facility for doing the work quickly. 2 H. N. Morse and T. L. Blalock, Amer. Chem. Journ., 16. 479, 1894; A. Gawalovski, Chem. Cenlr. (3), 10. 236, 1879 ; A. Demichel, Rev. Anal. Chim., 5. 1897 ; A. Mulder, Chem. Weekblad, 5. 830, 1908. 3 CLEANING VOLUMETRIC APPARATUS. Grease is the great foe of accuracy in volumetric work. Thorough freedom from grease is very important. The smallest trace of grease may distort the meniscus and alter the amount of water retained as a film on the glass. The error so introduced may amount to O'Ol to 0'04 per cent., according to the method used in cleaning the vessel. To clean flasks, pipettes, or burettes, wash the vessel thoroughly with a concentrated aqueous solution of caustic soda ; rinse with water ; wash with a solution of chromic acid or potassium bichromate in sulphuric acid ; and thoroughly rinse with distilled water. Flasks, etc. , seem to gradually acquire a greasy film on their inner surface while in use. In some cases, the grease can be traced to the contamination of the distilled water by the oily packing between the "worm" and "still-head" of the still. If grease be absent, the surface of the glass will remain covered with a continuous film of water until all is removed by evaporation. If the water collects in drops instead of adhering as a continuous film, the vessels are greasy and want cleaning. W. 0. Atwater and C. D. Woods, Journ. Anal. App. Chem., I. 373, 1887 ; C. Mohr, Chem. Ztg., II. 561, 1887; W. Glenn, Journ. Amer. Chem. Soc.,22. 302, 1900; J. Walz, Dingler's Journ., 207. 427, 1873. M. Vandevyver (Rev. Anal. Chim., 5. 1897) shows that the errors due to variations in surface tension are reduced to a minimum if perfectly clean vessels are used. 4 An error of O'l in the density of the brass weights (8 "4) will produce an error of about 0-0017 grm., that is, 0-00017 per cent. THE MEASUREMENT OF VOLUMES. 37 The above procedure can be followed for graduating a flask to contain (" Einguss ") a specified volume. A narrow strip of paper is gummed longi- tudinally on the neck, the flask is weighed empty, and then filled part way up the neck with water, weighed, and a light mark made with a lead pencil on the strip of paper corresponding with the horizontal plane at the lowest point of the meniscus. If the first trial be not correct, add more or less water and make another mark ; if this be inaccurate, make another trial, and repeat, if necessary, until it is correct. In this way the right volume can soon be obtained. A distinct line is then made on the strip of paper. The flask is then emptied. The paper above the last mark is cut away, and the neck of the flask, where the mark is to be etched, is warmed, and coated with a uni- form layer of wax. 1 Starting from the top of the paper, a mark is carried uniformly round the neck of the flask with the blade of a pocket-knife, so as to expose a line of glass not protected by the wax. The "ringing" of the neck is best per- formed by placing the mouth of the flask on a conical plug so placed that the left hand can rotate the flask while the right hand is used for cutting the ring (fig. 15). The hand with the knife 2 is supported and guided by blocks of wood at a suitable height. The operation is shown by photo- graph, fig. 15. A drop of hydrofluoric acid is then placed on the line while the neck of the flask is held horizontally. The drop is allowed to flow all round the neck on the exposed line by rotating the flask. In two or three minutes, the acid is washed off, the wax removed by wiping the hot wax with a cloth, and cleaning off the last traces of wax by means of methylated spirit. When the flask is to be graduated "to deliver" ("Ausguss"), the contents are poured into a vessel of known weight and weighed. The temperature is taken just after the weighing. The entire neck should be wetted and drained into a flask at least one minute before pouring. With flasks marked " Contain," the neck should be dry above the meniscus, except for the wetting necessary to secure a constant meniscus of normal shape. The flask is held in an inclined position, so that the bottom and the parts near the neck may drain. At the end of a minute, the suspended drop is removed with a glass rod and included in the amount delivered. The "error" allowed by most of the testing stations for measuring flasks is as .follows : FIG. 15. " Ringing" flasks. Capacity . 50 100-250 300-500 500-1000 2000 c.c. Permitted error (Contain) Permitted error (Deliver) . . 0-05 . O'l O'l 0'2 0-15 0-3 3 0-6 0-5 i-o c.c. c.c. 1 A mixture of equal parts of paraffin wax and carnauba wax, or beeswax alone page 634. 2 Or engraving tool. For a more elaborate apparatus, see C. B. Williams, Journ. Amer. Chem. Soc., 24. 246, 1902. A TREATISE ON CHEMICAL ANALYSIS. 17. The Drainage or Afterflow from Measuring Vessels. The drainage of measuring apparatus is of great importance. Schlosser l has established a relation between the time the discharge is stopped and the amount afterflow subsequently drained from the sides. The drainage is greater the more rapid the discharge of the liquid. When the discharge from a burette is 0'7 cm. per second, the afterflow during the first two minutes after stopping the outflow was found to be about O05 c.c. When the time of the discharge from the burette was changed from 65 to 125 seconds, an increase of 0*05 c.c. in the volume of the afterflow was obtained. Hence, if 0*05 c.c. of the standard solution will appreciably affect the result, it is necessary to work under standard conditions, and pay attention to the time of outflow of all apparatus marked "Deliver" ("Ausguss") burettes, pipettes, flasks, etc. The rate of outflow for burettes and measuring pipettes is limited by the size of the tip, and it is usual to make the tip so that the time of outflow is not greater than three minutes, nor less than the numbers indicated in the following table : Length graduated . . 65 60 55 50 45 40 35 30 25 20 15 c.c. Time of outflow . . 140 120 105 90 80 70 60 50 40 35 30 sees. When titrating and standardising burettes, it is well to wait a couple of minutes 2 before every reading in order to give the sides time to drain. The amount of liquid which adheres to the walls of a burette, etc., is dependent upon (1) the viscosity of the fluid ; (2) the adhesion of the liquid to the walls of the vessel ; which in turn depends partly upon (3) the cleanliness of the walls ; (4) on the surface exposed to the liquid ; (5) on the rate of outflow from the jet of the burette, 3 etc. In illustration, contrast alcohol with concentrated sulphuric acid. Again, a pipette which delivered 100 c.c. of water at 15-5, with a drainage of 10 seconds, and then touching the delivered liquid with the point of the pipette, 4 delivered the volumes indicated in the accompanying Table IX. when the solutions indicated were employed in place of water : Table IX. Effect of the Nature of the Liquid on the Volume delivered by a Pipette. Solution. Specific gravity. Volume delivered. Zinc sulphate Sodium chloride . 1-4246 1-2072 99-52 99-73 Ammonium sulphate Potassium chloride 1 -2500 1-1812 99-75 99-82 Ammonium chloride 1-0724 99-95 Water . roooo 100-00 Hence it is obvious that a pipette graduated for water delivers a sensibly different volume when applied to other liquids. Pipettes may be used inter- 1 W. Schlosser, Zeit. angew. Chem., 16. 953, 977, 1904, 1903 ; N. S. Osborne and B. H. Veazey, Bull. Bur. Standards, 4. 553, 1908 ; T. Milobendzki, Zeit. anal. Chem., 46. 20 1907* 2 H. von Jiiptner (Chem. Ztg., 8. 1766, 1885) recommends three minutes' drainage for alcohol and water. See also T. Milobendzki, Zeit. anal. Chem., 46. 20, 1907. 3 J. Mulder, Die Silberprobirmethode, Leipzig, 1859. 4 R. R. Tatlock, Chem. News, 23. 14, 1871 ; A. Schulze, Zeit. anal. Chem. 21 167 1882 THE MEASUREMENT OF VOLUMES. 39 changeably for solutions not exceeding normal concentration provided the time of outflow be the same as that allowed for pure water. But a pipette or burette which delivers a volume of liquid which differs appreciably from water, requires a special calibration with that particular liquid in place of water. According to Schlosser, 1 the effect of temperature on the drainage is negligibly small between 15 and 30. 18. The Calibration of Pipettes. It is generally cheaper to purchase ungraduated pipettes and mark them according to the following scheme. 2 First thoroughly clean the pipette by sucking caustic soda into it by means of a rubber tube and pinchcock. Follow by rinsing with water, etc., as indicated page 36. Fasten a narrow strip of paper longitudinally to the upper part of the tube. Tare a small beaker of water on the balance and suck water from the beaker into the pipette until the beaker has lost the necessary amount of water for the capacity of the pipette. Mark the position of the lower level of the meniscus with a lead pencil on the paper strip. Run the liquid into a weighed stoppered flask. Hold the pipette vertically during the outflow and let it drain 1 5 seconds. Remove the suspended drop by contact with a wet surface (side of flask) and include it with the delivered amount. 3 Weigh the flask and its contents. If the amount of liquid delivered from the pipette be too large or too small, mark the paper below or above the first mark, and repeat the filling and weighing until finally the weight of liquid discharged from the pipette is exactly the amount desired at the stated temperature. The pipette may be marked with a writing diamond, the paper removed and the pipette marked all round with the diamond, or the mark may be etched with hydrofluoric acid as indicated for "ringing" flasks. 4 Now verify the accuracy of the graduation by filling the pipette as before. 5 The " error " in pipettes allowed by testing stations is usually as follows : Capacity . . . . . 1 to 2 10 25 50 100-200 c.c. Permitted error .... O'Ol 0'02 0'03 ' 0'05 O'l c.c. It is easy to buy pipettes far more accurate than this. 19. The Calibration of Burettes. The burette is a most important instrument used in volumetric analysis. Burettes can be purchased accurate to O02 c.c. A bad burette is a nuisance, and is best discarded. To calibrate a burette, fill the burette with distilled 1 W. Schlosser, Zeit. anal. Chem., 46. 392, 1907 ; W. Schlosser and C. Grimm, Chem. Ztg., 30. 1071, 1906. 2 F. Clowes, Journ. Soc. Chem. Ind., II. 327, 1892. For the graduation of glass vessels, C. Foord, Chem. News, 30. 191, 1874 ; M. C. Lea, Amer. J. Science (2), 42. 373, 1866 ; R. Bunsen, Gasometrische Methoden, Braunschweig, 28, 1877. 3 EMPTYING PIPETTES. Pipettes may be emptied in at least four ways: (1) the liquid is allowed to run from the pipette without touching the pipette against the sides of the vessel : (2) by blowing out the last drop (Ostwald). This method is useful in dealing with capillary pipettes ; (3) by touching the point of the pipette against the surface of the solution ; and (4) by touching the point of the pipette against the side of the vessel (a) all the time the pipette is emptying, (6) at the end of the discharge (Fresenius ; Treadwell). The same method of work must be employed when the pipette is in use as is employed during standardisation. H. Gockel, Chem. Ztg.. 25. 1084, 1901 ; 26. 159, 1902 ; Zeit. angeiv. Chem., 15. 707, 1902 ; 16. 49, 562, 1903 ; W. Schlosser, ib. t 16. 977, 1903 ; L. A. Fischer, Journ. Amer. Chem. Soc., 20. 912, 1898 ; N. S. Osborne and B. H. Veazey, Bull. Bur. Standards, 4. 553, 1908. 4 A simple modification of the apparatus, fig. 15, enables this to be done neatly and quickly. 5 C. B. Williams, Journ. Amer. Chem. Soc., 24. 246, 1902. 40 A TREATISE ON CHEMICAL ANALYSIS. water at the normal temperature, and withdraw successive portions ranging from 2 to 5 c.c. into a tared flask. By weighing these quantities accurate to a milligram the volume delivered by the burette between any two divisions can be readily determined. If a number of burettes have to be calibrated, the following procedure, based upon Scheibler's and Ostwald's methods, 1 is sufficiently exact, and it will save time. A glass pipette is cut from an old pipette (about 2 c.c.), as shown, and mounted with the burette to be tested by means of pressure tubing and stopcocks (or pinchcocks), as shown in the diagram, fig. 16. Two marks, a and , respectively FIG. 16. Calibration of a burette. below and above the bulb, are cut or etched on the stem of the pipette. The pipette and the burette are thoroughly cleaned. The whole is mounted as 1 C. Scheibler, Journ. prakt. Chem. (1), 76. 177, 1859 ; W. Ostwald, ib. (2), 25. 453, 1882 ; W. Ostwald and R. Luther, Ausfilhrung physiko-chemischer Messungen, Leipzig, 87, 1894 ; T. W.Richards, Journ. Amer. Chem. Sac., 22. 144, 1900 ; H. N. Morse and T. L. Blalock, Amer. Chem. Journ., 16. 479, 1894 ; D. W. Horn and E. M. van Wagener, ib., 30. 196, 1903 ; 0. von Spindler, Schweiz. Woch. Chem. Pharm., 46. 145, 1908 ; L. W. Andrews, ib., 28. 491, 1902 ; D. W. Horn and E. M. van Wagener, ib., 30. 96, 1903 ; H. L. Payne, Journ. Anal. App. Chem., 6. 326, 1892 ; M. J. C. Boot, Bee. Trav. Pays Bas, 13. 417, 1894 ; 1). Carnegie, Chem. News, 64. 42, 1891. W. Ostwald (I.e.}; and A. S. Cushman (Journ. Amer. Chem. Soc., 23. 482, 1901) use a pipette with the upper stem graduated. The error is read directly on the graduated stem of the pipette every time the pipette is filled with, say, 2 c.c. of liquid from the burette. This pipette is easily obtained by using one of Cobbett's pipettes instead of that indicated in the text. For calibrations with mercury, C. A. Bell, Proc. Chem. Soc., 17. 179, 1901 ; Chem. News, 83. 21, 1901. THE MEASUREMENT OF VOLUMES. 41 shown in the diagram. Take care that the tubes are quite vertical. The burette, the tubes between the burette and the pipette, and the jet of the pipette are filled with pure water. Water is run into the pipette from the burette until the lower level of the meniscus is in the same horizontal plane as the lower mark of the pipette. The burette is filled to zero. Water is run from the burette until the pipette is filled to the mark b. Read the burette. This water is then discharged from the pipette. The filling and emptying of the pipette is 'continued until insufficient water remains in the graduated portion of the burette to fill the pipette. The burette is read each time the pipette is filled. Two minutes are, of course, allowed for drainage, as indicated above. Suppose that the 50-c.c. burette has filled the pipette 16 times, and, as a mean of three comparisons, the burette reads 49*64 c.c. The burette is then refilled, and exactly the same amount of water is run from the burette into a weighed flask. The weight of water gives the necessary data for calculating the volume of the pipette. Suppose we get a weight of 49*64 grms. at 18 -5. From the table, page 29, we see that the density of water at this temperature is 0*9985, and its volume is therefore 49'64 -j- 0-9985 = 49-71 c.c. Hence, 49*71 Volume of water in pipette (18 '5) = = 3*107 C.C. 16 Suppose that the numbers in the second column of Table X. represent the successive readings of the burette under the conditions just indicated. The numbers in the third column of Table X. will represent the corresponding volumes, that is, the product of 3-107 with the corresponding number in the first column. The fourth column gives the error of the readings indicated in the second column. Table X. Burette Corrections Burette No. 5. Number of reading. Burette reading. Corrected burette reading. Corrections. c.c. c.c. c.c. 1 3-11 3-11 o-oo 2 6-20 6-21 +0-01 3 9-28 9-32 + 0-04 4 12-39 12-43 + 0-04 5 15-51 15-53 + 0-02 6 18 61 18-64 + 0-03 7 2172 21-75 + 0-03 8 24'53 24-57 + 0-04 9 '27-92 27-96 + 0-04 10 31-02 31-07 + 0-05 11 34-13 34-18 + 0-05 12 37 24 37-28 + 0-04 13 40 34 40-39 + 0-05 14 43-44 43-50 + 0-06 15 46-63 46-69 + 0-06 16 49 64 49-71 + 0-07 These numbers can be plotted on squared paper (fig. 17): ordinates, corrections (column 4) ; abscissae, volumes (column 2). Hence the correction for any reading of the burette can be seen at a glance. In the present 42 A TREATISE ON CHEMICAL ANALYSIS. example, the burette reads 49*64 when the corresponding volume is 49*71 c.c. Hence that burette reading is to be corrected by +0*07 c.c. Had the corrected reading for another burette been 49*59 c.c., the actual burette reading 49*64 c.c. O-iO 0-075 0-05 0-025 10 20 30 FIG. 17. Corrections for burette No. 5. 50 would have been corrected by - 0*05 c.c. The sign of the correction wants thinking over. This burette is unsatisfactory probably the bore of the tube is not uniform. EXAMPLE. At the beginning of a titration, the burette read 3 '69 c.c. ; at the end, 20*06 c.c. From the graph (tig. 17) of the above table, 3'69 c.c. represents 3*70 c.c. and 20*06 represents 20*09 c.c. The difference 20*09-3*70 = 16'39 c.c. represents the number of cubic centimetres of fluid run from the burette at 18*5. If the solution be standardised at 15, another correction must be made (page 29). The following " errors " are usually " permitted " at a standard testing station : Capacity Permitted error 1 to 2 0*01 2 to 10 0*02 10 to 30 0*03 30 to 50 0-05 50 to 100 c.c. 0-10 c.c. For parts of the volume less than half the total content, the limits of error are one-half the above. Burettes ar usually tested at five points at the National Physical Laboratory. Charge : 2s. to 2s. 6d., and 6d. for each additional point. The burette nozzle should be such that the time of delivery of the burette does not exceed 10 seconds. Wagner (I.e.) thinks that 70 seconds is sufficient for the time of outflow. It might also be added that Forster J noticed that burettes which have been used with alkaline solutions increase perceptibly in volume. Thus, one is mentioned which changed from 50 c.c. to 51 c.c. after being in use for a year. Such burettes must be frequently calibrated. 2 20. Burette Stands, and Burette Cocks. The burette, when in use, should be clamped vertically to a firm stand. The simpler types of stand are usually the best. It is, however, important to select those clamps which hide as little of the scale as possible. Kaehler's 3 clamp, 1 0. Forster, Chem. Ztg., 28. 147, 1904. 2 For general reports on volumetric apparatus, see Mitt, der K.N.E.K., 3. 5 1908- Internal, Kong, angeic. Chem., 7. 41, 1909. 3 M. Kaehler, Zeit. anal. Chem., u. 190, 1872 ; J. Dietl, ib., 15. 186, 1876 ; C. Kippenber^er ib., 43. 233, 1904; A. Eiloart, ib., 33. 205, 1894; A. Gawalovski, ib., 38. 237 1899 -V Pribram, ib., 12. 299, 1873 ; F. Allihn, Chem. Ztg., 10. 647, 1886; J. Vosatka ib 28 795* 1904; F. Oettel, ib., 19. 1384, 1895; L. Weinstein, ib., 8. 1870, 1885; A. T. Lincoln Journ Amer. Chem. Soc., 27. 1442, 1905 ; F. J. P. van Calker, Dingier s Journ., 225 84 1878 R Muencke, ib., 229. 336, 1879 ; G. Mtiller, Zeit. angew. Chem., 21. 2318, 1908. ~ THE MEASUREMENT OF VOLUMES. 43 shown in fig. 9, is excellent. Vosatka's stands (fig. 18) hide none of the scale. Other clamps are shown in different parts of this work. The points to be kept in view when selecting a burette stand are: (1) rigidity; (2) clear scale; (3) simple mechanism ; and (4) vertical adjustment. Burettes are supplied with funnel-like tops 1 (fig. 18) fora few pence extra. This enlargement of the tube is convenient for filling. The contents of the burette must be protected from dust. Loose glass caps like inverted test-tubes, FIG. 18. Vosatka's burette stand. or glass spheres, are convenient for keeping out dust, etc., when the burette is in use. Burettes are now usually fitted with glass stopcocks the plain straight cock, fig. 16, left; Geissler's cock, fig. 18, right; Meissner's cock 2 ; or Greiner and Friedrichs' cock, fig. 24. The latter are also arranged with three ways to connect the burette either with the outflow jet or with the stock of standard solution. Glass cocks are a nuisance when they begin to leak, as they often do, after they have been in use some time. 3 The glass between the two openings 1 A. C. D. Poole, Brit. Pat. No. 470129. 2 C. Meissner, Zeit. anal. Chem., 26. 625, 1887; E. Greiner and Friedrichs, ib., 26. 49, 1887 ; 27. 470, 1888. 3 LUBRICANTS FOR GLASS STOPCOCKS. F. C. Phillips, Journ. Amer. Chem. Soc., 20. 149, 1898 ; Chem. Nacs, 78. 311, 1898. According to L. L. de Koninck and M. Lejeune (Bull. Soc. Chim. Bdg., 23. 79, 1909) stopcock burettes for iodine and permanganate solutions may be safely greased with vaseline. 44 A TREATISE ON CHEMICAL ANALYSIS. of the bored plug gets worn. Hugershoffs cocks l (fig. 19) overcome this difficulty. Burettes with glass cocks are nearly always used for permanganate solutions (see also page 197). Permanganate solutions are not used with rubber tube and pinchcock (fig. 18, left). Glass stopcocks are not used for caustic alkali solutions on account of the tendency of the stopcock to stick, but they are used for iodine, and permanganate solutions. 2 The rapid wear of glass cocks with caustic alkaline solutions also causes the cock to leak very soon. 3 Rubber tube and glass jet with (old-fashioned) metal pinchcock may be used for caustic alkali solutions and solutions which do not attack rubber. 4 The plug cocks are better than the metal pinchcocks. 5 Plug cocks are made by plugging a piece of rubber tube FIG. 19. Hugershoff's glass cock. FIG. 20. Plug cock. FIG. 21. Binder's jet. with a glass bead, or bit of glass rod (fig. 18, left, and fig. 20) about 8 mm. long. By squeezing the tube around the plug, a channel is made for the passage of the liquid. Binder's nozzle, shown in fig. 2 1 , is made from a piece of capillary tube by sealing up one end, and blowing a hole at the side. It too is opened by pinching or squeezing the rubber tube. Both these cocks are excellent. To avoid the inconvenience which arises when glass stopcocks fall out and break, C. E. Munroe 6 has designed a simple clip made of hard rubber sheet shaped so that it will encircle some part of the apparatus and grip a 2-pronged fork. The prongs grip the thinnest part of the stopper so as not to impede the motion of the stopper, and yet they hold it so securely as to prevent its falling out in whatever position the apparatus is held. Snelling recommends stoppers with a small groove cut in the extended portion. A rubber ring is fitted in the groove and affords complete security from the slipping of the cock from its place. 1 F. Hugershotf, Chem. Gentr. (3), 18. 135, 1887 ; C. Meissner, Chem. Ztg. Rep., 12. 65, 1888. 2 L. L. de Koninckand M. Lejeune, Bull. Soc. Chim. Bclg., 23. 79, 1909. 3 P. Rudnick (Jowrn. Amer. Chem. Soc., 32. 971, 1910) proposes a silver stopcock for caustic solution. P. Fiebag (Archiv Pharm. (3), 22. 544, 1885) used a tin stopcock. 4 H. F. von Jiiptner, Oester. Zeit. Berg, Hutt., 28. 33, 1880. 5 F. Mylius, Archiv Pharm. (3), 3. 151, 1874 ; F. Mohv, Lehrbuch der Titriermethode, Braunschweig. 1874 ; A. Lebrasseur, Analyst, 10. 194, 1885 ; F. Sutton, ib., 10. 214, 1885 ; E. Thiele, Zeit. anal. Chem., 40. 405, 1901 ; J. Wallensteiner, ib., 25. 547, 1886 ; A. Gawalovski, ib., 38. 237, 1899 ; W. Leybold, ib., 26. 230, 1887 ; 0. Binder, ib., 27. 178, 1888 ; C Kippen- ber#er, Chem. Ztg., 27. 1255, 1903. 6 W. 0. Snelling, Journ. Ind. Eng. Chem., 4. 613, 1912. CHAPTER III. VOLUMETRIC ANALYSIS. 21. Normal Solutions. IN volumetric analysis l a standard solution of known strength is gradually added to a solution of the substance under investigation until the reaction between the two is completed. The volume of the standard solution required for this purpose is proportional to the amount of the substance in the solution under investigation. The two solutions must be of such a kind that the reaction takes place quickly and quantitatively. It must also be possible to recognise the end of the reaction. This is frequently indicated by a change in colour which occurs in the presence of a third substance, the indicator, immediately a small excess of the standard solution has been added. The indicator is usually indifferent to the main reaction. Since a definite amount of the standard solution can react with a definite amount of the substance under investigation ; given the volume of the standard solution, the amount of the substance under investigation can be calculated by simple rule-of-three. The standard solutions may be any strength, but it is convenient to make them of such a strength that the weight of the dissolved substance, per litre, bears a simple numerical relation to the molecular weight of the compound. A solution ivhich contains the hydrogen equivalent weight in grams of an element or compound per litre is called a normal solution, written N-solution ; a semi-normal solution is one-half the strength of a normal solution JN-solution ; a deci-normal solution y^N-solution is one-tenth the strength of a normal solution ; and a centi-normal solution- T ^N-solution is one-hundredth the strength of a normal solution. The "equivalent weight" of a substance is the number of grams of the substance which brings into reaction combination or replacement the equivalent of 1-008 grms. of hydrogen; 8 grms. of oxygen, etc. To illustrate further, the molecular and equivalent weights of HC1, NaOH, NaCl, AgN0 3 , NH 4 CNS, and Na. 2 S 2 3 , 5H 2 are identical ; the equivalent weights of H 2 C 2 4 , H 2 S0 4 , Ba(OH) 2 , "and CaC0 3 are half the molecular weights; the equivalent weight of As 2 3 is one-fourth the molecular weight ; the equivalent weight of KMn0 4 in acid solution is one-fifth the molecular weight ; and the equivalent weights of KI0 3 and K 2 Cr 2 7 are one-sixth the molecular weights. It is not always easy to see how the equivalent weight of some of these substances is determined, hence a few illustrations may be cited. In titrating iodine solutions with sodium thiosulphate, the reaction is represented by the symbols: 2Na. 2 S 2 3 + 1 2 -> 2NaI + Na 2 S 4 6 . Here one molecule of sodium 1 Following up some early suggestions by E. A. H. Descroizilles (1802) and L. N. Vauquelin (1812), J. L. Gay Lussac demonstrated the applicability and utility of volumetric methods in analytical chemistry for chlorimetry, 1824 ; for alkalimetry, 1828 ; silver, 1832. See L. L. de Koninck, Bull. Assoc. Belg. Chim., 19. 28, 73, 422, 1904 ; Histoire de la Methode Titrimetrique, Brussels, 1901. 45 46 A TREATISE ON CHEMICAL ANALYSIS. thiosulphate is equivalent to one atom of iodine, which, in turn, is equivalent to one atom of hydrogen. Hence the equivalent of Na 2 S 2 3 . 5H 2 is numerically equal to the molecular weight. With potassium bichromate, in titrating ferrous salts, we have the reaction : K 2 Cr 2 7 + 6FeSO< + 7H 2 -> etc. This reduces to 30 + 6FeO - 3Fe 2 3 . Hence the potassium bichromate furnishes three atoms of oxygen, which, in turn, correspond with six atoms of hydrogen. Hence the molecule of potassium bichromate is equivalent to six atoms of hydrogen, or the equivalent is one-sixth the molecular weight. The case of potassium permanganate is a little curious. When ferrous sulphate is oxidised by potassium permanganate in acid solution, the equation 2KMn0 4 + 10FeS0 4 + 8H 2 S0 4 -> K 2 S0 4 + 2MnS0 4 + 5Fe 2 (S0 4 ) 3 + 8H 2 reduces to 50 + 10FeO->5Fe 2 8 . Or, two molecules of potassium permanganate furnish five atoms of oxygen, which are equivalent to ten atoms of hydrogen. Hence the normal solution of KMn0 4 will contain one-fifth of the molecular weight, that is, T ^ of 2 x 158 = 31 '6 grms. per litre. In Volhard's process for manganese, the titration is made in neutral solution. The reaction is 2KMn0 4 + 3MnS0 4 + 2ZnO -> 5Mn0 2 + K 2 S0 4 f 2ZnS0 4 , which reduces to Mn 2 7 + 3MnO -> 5Mn0 2 . Or, two molecules of potassium permanganate give three atoms of oxygen, so that the normal solution contains one-sixth of twice 158 = 52'66 grms. per litre, or the equivalent weight is one-third the molecular weight. Hence there are two different normal solutions, depending 011 the particular reaction under consideration. But potassium permanganate is generally employed in acid solution, and in consequence, unless otherwise expressed, a normal solution of KMn0 4 is understood to contain 31 '63 grms. per litre. 1 Again, in titrating stannous chloride with ferric chloride, two molecules of the latter give two molecules of chlorine to one molecule of stannous chloride : 2FeCl 3 + SnCl 2 -> 2FeCl 2 + SnCl 4 . Hence one molecule of ferric chloride is equivalent to one atom of chlorine, which, in turn, is equivalent to one atom of hydrogen. The molecular and equivalent weights are identical. A normal solution of stannous chloride, on the other hand, will have half the molecular weight of SnCl 2 expressed in grams per litre, because one molecule of stannous chloride takes two atoms of chlorine from two molecules of ferric chloride. The preceding definition of normal solutions is virtually that given by F. Mohr in 1859. 2 This definition has been adopted by the leading writers on volumetric analysis (Fresenius, F. Button, E. F. Fischer, A. Classen), and most analysts. Other definitions have been proposed, and a few writers (N. Menschutkin, M. M. P. Muir, C. Winkler, J. Muter) have defined a normal 1 For "the available oxygen" standard, see R. Abegg, Anleitung zur Berechnung volu- metrischer Analysen, Breslau, 1900. 2 F. Mohr, Lehrbiich der chemisch-analytischen Titriermethode, Braunschweig, 1859 ; W. Fresenius, Zeit. anal. Chem., 25. 205, 1886 ; B. Tollens, ib., 25. 3fi3, 1886 ; C. Marx, ib. t 26. 217, 1887 ; L. L. de Koninck, ib., 25. 487, 1886 ; A. H. Allen, Chem. News, 40. 239, 1887; C. Winkler, Die Maassanalyse nach neuen titrimetrischen System, Freiburg, 1883; Ber. t 18. 2527, 1885 ; Zeit. anal. Chem. 25. 484, 1886. . VOLUMETRIC ANALYSIS. 47 solution to be a " solution containing the molecular weight of the salt in grams per litre." This definition is advantageously used in physical chemistry, but not in analytical chemistry. It might be noted en passant that Gay-Lussac's "normal solution of sodium chloride" is "a solution of sodium chloride such that 100 c.c. will exactly precipitate 1 gram of silver." This special definition is used in many assaying laboratories for these particular solutions. 22. Standard Solutions of Hydrochloric Acid. Hydrochloric acid HC1. Molecular weight : 36 '468 ; equivalent weight : 36'468. The most commonly employed solutions are : Griii. HC1 per c.c. N-Hydrochloric acid '0364(58 iN- Hydrochloric acid 0'018234 |N-Hydrochloric acid . . . . . . . 007291 i^N-Hydrochloric acid '003647 To prepare a semi-normal solution of hydrochloric acid, that is, a solution of hydrochloric acid containing half the equivalent weight (viz. 18*25 grms.) per litre, make a solution rather more concentrated than is finally needed ; determine the exact strength by comparison with another solution of known titre, or by analysis; and then dilute to the required strength. Never start with a solution more dilute than that finally required, because of the difficulties involved in measuring accurately small quantities of material. Suppose that the hydrometer shows that the stock of concentrated acid has a specific gravity 1'14. From Table LXXVI., an acid of this strength has 28 per cent, of HC1 by weight. Hence, 77 c.c. of this acid will have between 20 and 27 grms. of HC1. Make 77 c.c. of this acid up to 1100 c.c. in a Giles' flask l (fig. 22r;) with recently boiled distilled water. Let the solution stand a few hours so that it may attain the temperature of the room. If the temperature of the room be not that indicated on the flask, an allowance must be made as indicated, Table VII., XL, or XIII. Compact fragments of Iceland spar 2 calcium carbonate weighing from 2 to 3 grms. each, are broken from a large crystal by pressing the edge of the blade of a knife against a large clear crystal and tapping the back of the blade sharply with a hammer. These fragments are rinsed with dilute hydrochloric acid or nitric acid to free them from adhering powder, and to round their edges and corners. They are then well washed with distilled water, and dried in an oven at 110. The pieces are preserved in a stoppered bottle for use. Two or three pieces of the Iceland spar are placed in a clean dry beaker or 1 VV. B. Giles, Chem. News, 69. 99, 1894 fig. 22c. ; W. Wislicenus, er., 29. 2442, 1896 fig. 22&. These flasks are convenient when standard solutions are to be made up from reagents which cannot be exactly weighed. Ordinary litre flasks fig. 22e may of course be used, but the subsequent dilution is not so easily done. In Biltz's or Pfliiger's flask (E. Ffliiger, Zeit. anal. Chem., 25. 91, 1886) fig. 22^ there is a bulb above the mark on the neck to facilitate shaking. With the latter flasks, the amount of concentrated acid taken must be adjusted for 1000 c.c. In the present case, 70 c.c. would have been taken for 1000 c.c. in place of 77 c.c. for 1100 c.c. A. Goske (Chem. Ztg., 28. 795, 1904) and E. Hirschsohn (Zeit. anal. Chem., 30. 9o, 1891) describe a flask with a wide neck graduated in cubic centimetres between 920 and 1000 c.c. fig. 22a. The use of this flask will be obvious, and it would have been as convenient as Giles' flask for our purpose. The wide neck, however, makes the "error of experiment" somewhat greater. 2 Selected crystals are remarkably pure more pure indeed than the regular " sodium carbonate purissimus" usually employed for standardising hydrochloric acid. This excellent method of standardising hydrochloric acid was recommended by G. Duerr, Chem. News, 28. 156, 1873 ; 0. Masson, ?&., 81. 73, 1900 ; W. H. Green, ib., 87. 5, 1903. 4 8 A TREATISE ON CHEMICAL ANALYSIS. Erlenmeyer's flask. Weigh the beaker and contents, all dried at 110 and cooled in a desiccator. Cover the beaker with a watch-glass, and add 20 c c. of the hydrochloric acid from the flask l by means of a clean pipette in such a way that there is no loss by spurting. When the evolution of gas has almost ceased (3-4 hours) rinse the side of the beaker with distilled water. Restore the cover, and heat the solution to near the boiling point on an asbestos pad or quartz plate for about an hour. This is to ensure that the reaction between the acid and the spar is complete. Bumping must not be permitted, or minute fragments may be detached from the undissolved spar which would cause high results. Decant off the solution, wash the beaker and the contained spar thoroughly with distilled water, dry the beaker and contents at 110, cool in a FIG. 22. Standard flasks. desiccator, and weigh. This furnishes the data necessary for calculating the strength of the acid. For example : First weighing of beaker and spar . Second weighing of beaker and spar Iceland spar dissolved 21-6327 grms 21-1016 grms. 0'5311 grms. Since the molecular weight of Iceland spar is 100 (more exactly lOO'OT), and 2HC1 + CaC0 3 -> CaCL + 2H 9 0, o J 2 * it follows that 1 litre of normal acid will dissolve 50 grms. of the Iceland spar, or 20 c.c. will dissolve 1 gram. Had the acid been exactly half normal, 0'5 grm. of the spar would have dissolved. 2 The acid we have prepared is too concen- trated. Pipette acid from the Giles' flask until the level of the liquid is at the 1000-c.c. mark. From the proportion 0-5311 : 0'5 = 20 : x, where 1 If ordinary flasks are used figs. 22^ or 22e the volume withdrawn for testing, etc., should be noted, so as to avoid the use of measuring cylinders later on when the solution is diluted. 2 It is here assumed that 100 is the molecular weight of the Iceland spar. If 100'07 be taken, the corresponding change must be made in the other numbers. VOLUMETRIC ANALYSIS. 49 #=18'83 c.c. It follows, therefore, in order to make the acid exactly half* normal, 18 '83 c.c. must be made up to 20 c.c., or 18 '83 c.c. must be diluted with 1'17 c.c. of water. Hence, the 1000 c.c. now in the Giles' flask must be diluted with 62 c.c. of water in order to make the solution exactly half normal. The accuracy of the dilution should be established by repeating the experiment with the Iceland spar. 1 The most important source of error in this method of standardising the hydrochloric acid arises from the fact that the reaction is some- what tardy towards the end, hence there is a danger of stopping before the reaction between the acid and spar is completed. The advantages of hydrochloric acid as the standard acid for general work are : (1) it forms readily soluble salts with the alkalies and the alkaline earths (hence preferable to sulphuric and oxalic acids) ; (2) it is easily obtained pure (nitric acid often contains a little nitrous acid) ; (3) the strength of the acid is easily verified by means of silver nitrate ; (4) the dilute solutions employed suffer no perceptible loss when boiled. For instance, a JN-solution boiled 10 minutes did not colour blue litmus paper held in the vapour, hence the steam which passes off is almost free from hydrochloric acid. (5) The strength of the acid does not change on keeping. 23. Temperature Correction for Solutions of Hydrochloric Acid. It has been assumed that the temperature of the acid solution did not differ more than 3 from the standard temperature, 15. If the actual temperature deviates appreciably from this number, a correction must be applied in order to obtain the strength the acid would be at the standard temperature. The following correction table Table XL is to be used in accurate work when allowance is to be made for variations of temperature of a normal solution of hydrochloric acid at a temperature different from that at which the substance was standardised, namely, 15. 2 For solutions more dilute than one-tenth normal the correction table, page 29, for water may be employed. Table XI. Temperature Correction for N-HCl Solutions. (Standard temperature, 15.) Temp. 1 2 3 4 5 6 7 8 9 + 1-3 + 1-2 + 1-1 + 1-0 + 0-9 1 + 0-8 + 07 + 0'5 + 3 + 0-2 Unit. -0-2 -0-4 -0-6 -0-8 2 -I'l -1-3 -T5 -1-8 -2-1 -2'3 -2'5 -2 8 -3-1 -3'4 1 It should be remembered that in diluting the more concentrated solutions of salts, acids, etc. , a change of volume may occur such that the volume of the mixture is less or greater than the sum of the separate volumes of the two components. The effect is seldom appreciable with dilute solutions, but with more concentrated solutions the effect may be serious if the dilution be not checked. Thus, with a solution of ammonium sulphate containing 56 grms. per 100 c.c., the contraction was found to be : Ammonium sulphate taken . . . . .50*0 25 '0 12*5 c.c. Water taken 50'0 75'0 87'5 c.c. Volume of resulting solution 98'83 99'64 99'89c.c. Changes of temperature on mixing sometimes cause a temporary expansion. Sufficient time must be allowed for the mixture to assume the temperature of the room before measuring liquid or the control experiment. 2 The table also applies for N-oxalic acid. It is based on A. Schulze (Zeit. anal. Chem. 21. 167, 1882). The table can be readily transposed into any other standard temperature different from 15, as indicated on page 29. 50 A TREATISE ON CHEMICAL ANALYSIS. EXAMPLES. (1) 21'30 c.c. of J^N-HCl have been used in a titration at 22. The solution was standardised for 15. The actual volume at 15 would be less than 21-30 c.c. From Table XL the correction for 1 c.c. is -0*0015 c.c. Hence, 21-3x0-0015 = 0-032 c.c. or 21'30 c.c. at 22 corresponds with 21'30- 003 = 21*27 c.c. at 15. (2) A titration is made at 12, and later on, at 20. In the former case, 19'99 c.c. were employed, and in the latter case, 20'022 c.c. Suppose that no temperature correction be made, it follows that the two determinations give a difference of 100(20-022- 19-99)^19-99 = 0-16 per cent. As a matter of fact, both really correspond with 20 c.c. at the standard temperature, 15. These two examples give an idea of the error to which we are liable when the temperature is neglected. The temperature factors will be given for other solutions as occasion demands. In order to facilitate the compounding of standard solutions at temperatures different from the standard, Heygendorff 1 has designed measuring flasks gra- duated for, say, 1000 c.c. at temperatures between 9 and 25. These flasks resemble a, fig. 22. 24. The Adjustment of Standard Solutions. When large volumes of a solution are in question, the adjustment can be made without measuring the volume of the liquid litre by litre. 2 Suppose that 40-50 litres of a T ^N-soda solution are to be prepared. Make sufficient solution, about double the required strength, to about half fill the carboy in which the solution is to be stored. Pipette, say, three lots of 25 c.c. and titrate with ^N-acid as indicated later on. Suppose that 25 c.c., as a mean of three titrations, require 31-25 c.c. of the standard acid. In that case, 80 c.c. of the given solution correspond with 100 c.c. of the y^N-acid. This means that 250 c.c. of water must be added per litre of the solution in order to make it exactly T ^N. Now transfer, say, 200 c.c. of the solution into a 500-c.c. flask. Add a litre of water to the solution ; stir ; withdraw, say, 200 c.c. for titration. Suppose that 8'33 c.c. now correspond with 100 c.c. of the y^N-acid. The addition of the water has made a difference of 50 c.c. per litre in the amount of water needed to convert the given solution to the required strength. Hence, the carboy contained 1000-^50 = 20 litres before the second sample was withdrawn for titration. The original solution required 250 c.c. of water per 1000 c.c. to make it exactly deci-normal. Hence, the original 20 litres will require 20 x 250 1000 =51ltres of water in order to make it exactly deci-normal. But 1 litre has been added. Consequently, 4 more litres of water must be added. A slight error is caused by withdrawing the 200 c.c. for the second trial. To correct this, dilute the 200 c.c. of the original solution standing in the 500-c.c. flask by adding 50 c.c. of water per litre ; that is, 200 c.c. required 10 c.c. of water. The solution in the flask is the same strength as the 200 c.c. withdrawn for the second titration before adding the 4 litres of water. Hence, restore the 200 c.c. to the liquid in the carboy, and, if there has been no change of volume on mixing, the solution should be exactly decinormal, and occupy 25 litres. Some prefer to leave the acid approximately half-normal, instead of adjust- ing the solution exactly, as described in the preceding section, by means of 1 M. von Heygendorff, Chem. Ztg., 35. 382, 1911. 2 F. M. Lyte, Chem. Neius, 26. 159, 1872 ; 29. 23, 1874 ; W. Smith, ib., 30. 220, 1869. VOLUMETRIC ANALYSIS. 51 the Giles' flask. In that case, 20 c.c. of the acid are equivalent to 0*05311 grm. of calcium carbonate. But, from the above equation, 100 grins, of Iceland spar correspond with 79-936 grms. of hydrochloric acid, and therefore 0-5311 grm. of Iceland spar will correspond with 07294 x 0*5311 =0*3874 grm. HC1. Hence, 20 c.c. of the acid has 0'3874 grm. of HC1 ; or 1 c.c. has 0*01937 grm. HC1. The bottle containing the solution may be labelled accordingly. If but a small volume of acid is to be standardised, it may be advisable to follow the plan just described. In other cases a factor may be used. In the above example, the acid has 0*3874 grm. HC1 per 20 c.c. If it were exactly normal, 20 c.c. would have 0*3647 grm. HC1. Or 18*83 c.c. is equivalent to 20 c.c. of half-normal acid ; or the acid is 0*531 1-normal when it should be 0*5-N. Dividing any pair of these numbers, say, ' 5 =0*9415, 0*5311 the resulting number 0*9415 represents the fraction of a cubic centimetre which is equivalent to 1 cubic centimetre of exactly |N-acid. Hence, in any given titration, the number of cubic centimetres multiplied by 0*9415 will give the number of cubic centimetres which would have been used with exactly JN-acid. Similarly, had 0*4311 grm. of spar been dissolved, the acid would have been too dilute, and the conversion factor would have been 0*5-0*4311 = 1*16. 1 25. The Adjustment of the Specific Gravity of Solutions. Solutions of a definite specific gravity have sometimes to be made by diluting solutions of known specific gravity. The dilution is not usually required to be of very great accuracy. 2 When tables connecting specific gravity with concentration are available, the method can also be applied to the dilution of solutions of a given percentage composition to solutions of another percentage composition. In the absence of a knowledge of the law of contraction or expansion on dilution, the problem can only be solved approximately, because the methods of calculation assume that "the total volume of a mixture is equal to the sum of the volumes of its parts," and this assumption is not always justified. The following represent three typical cases : 1 . It is required to prepare a volume V of a solution of specific gravity S, by diluting a volume v of a solution whose specific gravity is s. Here F, S, and s are known, v is to be determined. It follows that Vv will be the volume of the water needed, and we must have vs+ V -v VS. Hence, it follows 1 E. Peterson (Zeit. anal. Chem., 45, 14, 439, 1906 ; G. Bruhns, ib. t 45. 204, 1906) recom- mends the adjustment of standard solutions by representing the approximately normal solution in terms of "equivalent volumes," that is, the number of cubic centimetres of the solutions which contain an equivalent weight of the substance in solution. Thus, in standardising hydrochloric acid, we have found 0'5311 grm. of Iceland spar to be equivalent to 20 c.c. of acid ; hence, half of one equivalent of Iceland spar, namely, 25 grms., would be equivalent to half an equivalent of hydrochloric acid, and equivalent to 941 '5 c.c. of the solution. Hence, the equivalent volume of the solution was 941*5 c.c. 2 A. Vogel, Dent. Illustr. Gewerbezeit., 202, 1865 ; C. D. Howard, Journ. Amer. Chem. Soc., 19. 587, 1897. See almost any text-book on the elements of hydrostatics. 52 A TREATISE ON CHEMICAL ANALYSIS. EXAMPLE. Soda lye of specific gravity 1*34 is to be diluted with water to make 50 c.c. of a solution of specific gravity I'll. How much soda lye must be taken? Here, F=50 S=M1, and s = l'34. Arisr. v = 16-18 c.c. 2. It is required to dilute a volume v of a solution of specific gravity s, so as to make a solution of specific gravity S to find the amount of water needed. Here, v, s, and S are known to find V. When V has been determined, the required volume is given by V -v. From the preceding relation, we see that r-= ..... (2) EXAMPLE. What volume of water is needed to reduce 50 c.c. of a solution of caustic soda of specific gravity T34 to a specific gravity I'll ? Here, v = 50, S=l'll, * = 1'34. Ansr. 154*55 c.c. 3. It is required to dilute v c.c. of a solution of specific gravity s with another liquid of specific gravity z in order to get a solution of specific gravity S to find ivhat volume of liquid of specific gravity z is necessary for the purpose. Here v, s, z, and S are known. Let u denote the required volume of the solution of specific gravity z, then . -*? ...... (3) EXAMPLE. 50 c.c. of soda lye, specific gravity 1'34, is to be diluted with soda lye, specific gravity I'll, in order to get a solution of specific gravity 1-2. What volume of the latter is needed ? Here, v = 5Q,z = l'Il,s=l "34, S= 1 '2. Ansr. 77'8 - 50 c.c. = 27'8 c.c. The problem of dilution is usually very easily solved by means of the specific gravity tables. EXAMPLE. A solution of sulphuric acid, specific gravity 1'2, is to be prepared from the stock acid, specific gravity T84. From Table LXXV., 100 c.c. of acid, specific gravity 1-84, has 175*9 grins. H 2 S0 4 ; 100 c.c. of an acid specific gravity 1'2 has 32'8 grms. H 2 S0 4 . If 100 c.c. of the concentrated acid has 175'9 grms. H 2 S0 4 , 18'64 c.c. will have 32*8 grms. of acid. Hence, if 18*64 c.c. of the concentrated acid be transferred to a 100-c.c. flask, and the solution made up to 100 c.c. (temperature constant), the result will be an acid of the desired concentration. 26. The Calculations of Analytical Chemistry. Slide Rule. The slide rule is a convenient check on the accuracy of the arithmetic. Most of the calculations in general analytical chemistry are simple proportion. Simple proportion can be done very rapidly with the slide rule. For instance, take the proportion 2 : 3 = 4 : x. Place 2 on one scale over 3 on the other, and the value of x will be found over 4. Again, suppose a standard solution is such that 984 c.c. is equivalent to 1000 c.c. of a solution exactly normal; and 21 -2 c.c. of the solution have been used in the titration. Set 984 on one scale over 1000 on the other scale, and the number of c.c. of an exactly normal solution will be found over 21 '2. Again, what weight of S0 3 is equi- valent to 0-321 grm. BaSO^ Since 233 (BaS0 4 ) : 8(H (S0 3 ) = 0'321 : x, set 80 '1 over 233, and read the value of x over 0*321. The regular slide rule of moderate length does not read to more than two figures, and a third by approximation, that is, approximately to the first decimal of one per cent. It is customary to write two decimals, in spite of the fact that the second figure is of no real value. Chemists have not therefore taken up the slide rule, owing to the uncertainty in the third and fourth figures. It is a general principle in chemical arithmetic that the magnitude of the error due to VOLUMETRIC ANALYSIS. 53 the method of calculation must be well ivithin the limits of the experimental error. The ordinary slide rule does not satisfy this criterion. Hence, the chief function of the ordinary slide rule is to check the accuracy of calculations. 1 Errors which might escape detection by simply going over the figures may be thus revealed. G. Fuller's slide rule, however, is accurate enough for analytical requirements. It deals with four-figure numbers, and a fifth by an easy approximation. Its first cost (3) is rather high, but it is the slide rule for chemical calculations. 2 I find it excellent. Squared Paper. z Let 1000 be taken as an abscissa on squared paper, and as ordinate lay off a length corresponding with the reduction factor of a given determination, say 736 for PbO in lead sulphate; 247 for chlorine in silver eoo 600 400 200 1000 800 600 400 FIG. 23. Charts to facilitate calculations. 200 chloride, etc. Join, say, the ordinate 736 with the abscissa, then any particular weight of PbS0 4 , found by analysis, is located on the abscissa axis, arid the corresponding amount of PbO is read on the ordinate axis. This method is useful for checking calculations, and if the curve be drawn on a larger scale, it will be found as convenient as a table for saving time in analytical calculations. EXAMPLE. Suppose a precipitate of lead sulphate weighs 64 grm. The ordinate fig. 23 (dotted lines) corresponding with the abscissa 0'64 is 0'47 ; hence 0'47 grin, is the equivalent weight of PbO, corresponding with 0'64 grm. of PbS0 4 . 1 "A. Nestler's slide rale for chemists" (D.K.G.M. 409844, 1910) is the ordinary type of slide rule with marks on the scale corresponding with some of the commoner factors used in certain analyses. The scale may be obtained from 25 to 50 cm. long, the price 10s. 6d. As a matter of fact, W. H. Wollaston invented a similar slide rule in 1814 ! W. H. Wollaston, Phil. Trans., 104. 1, 1814. 2 A pamphlet explaining how the rule is used is sold with the instrument. ., ^. r -^v , . T j .. ^tr ... C1~. ~- KOT 1 flAQ F. P. Dunnington, Journ. Amer. Chem. Soc., 25. 537, 1903. 54 A TREATISE ON CHEMICAL ANALYSIS. Analytical Tables. Logarithms are in common use. Five-figure logarithmic tables suffice for the calculations in most analytical work. 1 In order to lessen the labour involved in calculating oft-repeated determinations, conversion tables may be used with advantage. This prevents wasting time and energy on a repetition of old operations, and conduces to more accurate work. Once accurate tables have been compiled, there is less liability to error. For instance, the amount of MgO in a given weight of magnesium pyrophosphate can be read from Table XC. at a glance ; similar tables can be used for the amount of K 2 or KC1 in a given weight of potassium chloroplatinate ; PbO in PbS0 4 ; etc. Some of these tables are given in the Appendix. Rational Weighing. Another artifice for abbreviating calculations is to weigh an amount of substance for the analysis which bears a simple relation to the conversion factor. 2 When the amount of substance taken for analysis is the same as the conversion factor, the weight of the precipitate directly corresponds with the percentage value sought ; and when the amount of substance taken for analysis is a simple multiple or sub-multiple of the conversion factor, the weight of the precipitate is a corresponding multiple or sub-multiple of the percentage value sought. For instance, if PbO is to be determined in a given substance, and if 0*7359 grm. of the sample be taken, then, if 0*5 grm. of lead sulphate be obtained, it follows that the sample contains 50 per cent, of PbO. It is easy to see this. Since 0*5 x 07359 grm. of PbO has been obtained from 0'7359 grm. of the sample, 100 grms. of the sample will have 50 grms. of PbO. It must be remembered that time is lost in weighing a definite amount of a substance to such a degree of accuracy, and it is sometimes questionable if the time consumed in weighing is not greater than the time spent in solving the proportion, in this particular case : Weight of sample : 0'7359 = Weight of lead sulphate : x. At any rate, the risk of error, owing to the hygroscopic properties of fine powders, is sometimes greater in the former case. Similarly in volumetric work. A little thought spent in "designing" routine operations may save a great deal of work. To find the relation between the weight of the sample to be taken for the analysis and the concentration of the standard solution employed so that the number of cubic centimetres of the standard solution may directly represent the percentage amount of the constituent sought. The strength of the standard solution must be adjusted so that a weight iv of the sample is equivalent to 100 c.c. In other words, 100 c.c. of the standard solution must be equivalent to a weight w of the (pure) constituent sought. For example, if 100 c.c. of the standard solution be equivalent to 3*942 grms. of the constituent under investigation, we must weigh exactly 3*942 grms. of the sample for the titration. In that case, if 41 c.c. of the standard solution be used in the titration, the sample contains 4 1 per cent, of the constituent in question. EXAMPLE. Suppose that it be required to determine the amount of sodium carbonate in a given sample of soda ash. For convenience, let w=l grm. The standard acid solution must be so made that 100 c.c. contains the equivalent of 1 gram of sodium carbonate. By weighing 1 gram of the sample for each determination, every c.c. of the acid used in the titration will represent one per cent, of sodium carbonate in the sample. 1 For example, F. D. Kiister, Logarithmische Rechnentafeln fur Chemiker Leipzig 1910 C. J. Woodward, Five- figure Logarithms for Chemists, London, 1910. a E. A. Uehling, Journ. Anal. App. Chem., i. 402, 1887. K. B. and L. A. Voorkees (ib 7, 121, 1893) recommend specially graduated pipettes. VOLUMETRIC ANALYSIS. 55 It may be advisable to use a more dilute solution. The strength of the standard solution can be made such that the percentage amount of the constituent sought in the given sample is one-tenth the number of cubic centimetres used in titration. Then, 1000 c.c. of the standard solution must be made equivalent to iv grms. of the constituent (pure) under investigation. EXAMPLE. Suppose that it be required to find the percentage amount of sodium chloride in a sample of soda ash. For convenience, let 10 = 2. The standard solution of silver nitrate must be made such that 1000 c.c. corresponds with 2 grms. of sodium chloride. But 2 grms. of sodium chloride are equivalent to 5'81 grms. of silver nitrate. Hence, when the standard silver nitrate solution contains 5 -81 grms. of silver nitrate per litre, and 2 grms. of the sample are taken, every 10 c.c. of the silver nitrate solution will represent one per cent, of sodium chloride in the given sample. 27. The Automatic Filling of Burettes and Pipettes. When titrations with a given solution have to be made at frequent intervals, one of the many * ingenious forms of burette available for automatic filling will save time. The four forms indicated below are typical, and useful in special cases, but there are scores of others. The chemist has here to consider a number of factors : convenience in filling, fragility, simplicity, first cost, repairs, etc. Is there a sufficient number of determinations to justify the installation ? Automatic burettes may reduce the risk of error as well as save time and frequently also material. Knofler's Burette. The burette shown in fig. 24 is a modification of Knofler's. It is filled by opening the cock A, and forcing the liquid into the burette from the stock solution S by means of the rubber blower B. When the burette is full, close the cock A, and adjust the level of the liquid in the burette by means of the cock A or C. The burette is then used in the ordinary manner. If an excess of liquid remains in the burette, it can be run back into the main solution by opening the cock A. In Krawczynski's burette, there is a central tube pass- ing from the cock A to the zero mark. The burette is graduated with this tube in position. If the burette should be filled above the zero mark, the excess runs back into the bottle when the blower stops. Hence the burette is auto- matically filled to the zero mark. There is then no provision for returning unused liquid to the main bulk. I prefer the burette without the central tube. The blower should have a tube packed with glass wool 2 to remove dust. The 1 0. Knofler, Chem. Ztg., 12. 1142, 1888 ; A. Thilmany, ib., 24. 115, 1900 ; W. Fleming. ib., 28. 818, 1904 ; W. Schmidt, ib., 28. 154, 1904 ; S. Schiff, ib., 14. 233, 1890 ; A. Knauer, ib., 9. 231, 1884 ; A. Stein, ib., II. 786, 1886 ; von der Heide, ib., 35. 568, 1911 ; F. Eisner, ib., 7. 1396, 1882 ; F. W. Dufert, ib., 10. 340, 1885 ; L. Hartmann, ib., 8. 418, 1883 ; H. R. Proctor, ib., 16. 17-nitrophenol is made by dissolving 1 grm. of />-nitrophenol in 75 c.c. of absolute alcohol, and making the solution up to 1 litre, that is, 1000 c.c., with water. 3 jo-Nitrophenol is yellow in alkaline solution, and colourless in neutral or acid solutions. It is rather more sensitive to carbon dioxide than methyl orange. The first appearance or disappearance of the yellow colour during a titration is so gradual that it requires a sensitive eye in daylight to recognise the change 1 H. N. and C. Draper, Glum. News, 55. 133, 143, 1887 ; J. II. Long, ib., 51. 160, 1885. 2 R. T. Thomson (vide infra} recommends a solution containing 0*15 grm. per litre, and uses 0*5 c.c. of this solution per iOO c.c. of the solution to be titrated. 3 L. Spiegel, Ber., 33. 2640, 1900 ; Zeit. angew. Chem., 17. 715, 1903 ; G. Lunge, ib., 16. 560, 1903 ; 17. 202, 1904; A. Goldberg and K. Naumann, ib., 16. 644, 1903 ; H. W. Langbeck Zeit. anal. Chem., 21. 100, 1882. VOLUMETRIC ANALYSIS. 63 accurately. Hence, methyl orange is preferable to jo-nitrophenol in most cases. jo-Nitrophenol is somewhat sensitive to boric acid in the presence of glycerol or mannitol, but riot very sensitive to boric acid without these auxiliaries. 1 Hildebrandt 2 summarises what he calls the " usability " of a number of indicators in the form of a table, from which Table XII. has been taken : Table XII. Conditions under which Indicators can be best used (Hildebrandt). Titration. In Ammonia presence Indicator. and of Cold. Hot, Alkali to acid. Acid to alkali. NH 4 C1. sodium acetate. 1. Litmus Usable only Usable Usable only Usable only Usable Not on exclu- on exclu- on exclu- usable. sion of C0 2 sion of C0 2 sion of C0 2 2. Phenolphthalein Usable on Usable Usable Usable Not Usable. exclusion of usable C0 2 3. Methyl orange . Usable for Usable Usable Usable for Usable Not N-and T VN- baryta ; for usable. Na 2 C0 3 N - Na 2 C0 3 with N- and only if hot r^N - HC1 and Ba(OH) 2 solutions 4. ^-Nitrophenol . Usable ex- Not Usable Usable only Usable Not cept with usable for baryta usable. N - Na 2 C0 3 and hot N-Na 2 C0 3 5. Cochineal. This substance occurs in commerce in the form of rounded grains which are the dried bodies of female hemiptera which lived on certain species of cacti. The so-called "silver cochineal" is the best. To prepare a solution, macerate 3 grms. of coarsely powered 3 cochineal in 250 c.c. of a mixture of 3 vols. of water, and 1 vol. of alcohol. Decant off the clear deep ruby red solution through a filter paper. The solution is coloured violet by alkalies, carbonates of the alkalies and alkaline earths, sodium phosphate, and salts of weak acids generally. Acids restore the yejlowish-red colour. The indicator is less affected by carbon dioxide than litmus, and it gives better results when used in artificial light. It is useless for titrating organic acids. Salts of iron, alu- minium, and copper yield a pink colour with this indicator, and hence these salts should be absent from the solution to be titrated. 6. Phenacetolin. This substance occurs in commerce as a yellowish-brown powder. It is used as an indicator in aqueous or alcoholic solution. The aqueous solution is 1 : 500 ; the alcoholic solution 1 : 200. The green-coloured solution is turned yellow with acids and caustic alkalies, and red with carbonates 1 A. Goldberg aud K. Naumann, Zeit. angew. Chem., 16. 644, 1903. - H. Hildebrandt, Wochschr. Brauerei, 22. 69, 1906 ; R. T. Thomson's papers on the indicators should be studied. R. T. Thomson, Chem. News, 47. 123, 184, 1883; 49. 32, 119, 1884 ; 52. 18, 29, 1885 ; A. H. Allen, Analyst, 17. 186, 215, 1892. 3 If finely powdered, the decoction is difficult to filter. 6 4 A TREATISE ON CHEMICAL ANALYSIS. of the alkalies and alkaline earths, alkaline sulphides, and ammonia. It is therefore useful for estimating calcium hydroxide or calcium oxide in the presence of calcium carbonate; sodium hydroxide in the presence of sodium carbonate, etc., provided sufficient hydroxide be present, and ammonia be absent. 1 If an excess of the carbonate be present, the colour change is not well denned. It requires some practice to get reliable results. The titration is made by adding standard sulphuric or hydrochloric acid rapidly until the yellow colour is succeeded by a faint rose colour. All the caustic alkali is now neutralised, and the burette is read. Any further addition of acid intensifies the red colour until all the carbonate is decomposed, when the colour passes to a yellowish red, and finally yellow. The burette is read, and the results furnish data for calculating the amount of hydroxide and carbonate in the given sample. It will not do to change indicators except for a specific reason. Working with Y^N-solutions of sodium hydroxide and hydrochloric acid, Scholtz 2 found that 10 c.c. of acid and alkali required respectively for neutralisation Litmus. Phenol phthalein. Methyl orange. y-Nitrophenol. 10 c.c. alkali . 9'83 9'68 1070 lO'OO c.c. acid. 10c.c. acid. . 9'92 10'08 9'28 9'99 c.c. alkali. The inference is obvious i in using centi-normal solutions, the indicator, and the direction of the titration acid to alkali or conversely employed in standardisa- tion should be the same as is used in the actual determination. Indicators are conveniently kept in bottles fitted with a pipette, in order that FIG. 32. Nest of pipette bottles fur indicators. the indicator may be removed without contact with the stain which usually collects about the necks of the bottles when the solutions are transferred in the ordinary manner. This stain interferes with the sharpness of the end point e.g., methyl orange in the bromate process for antimony. A nest of four pipette bottles for the most commonly used indicators is illustrated in fig. 32. The indicators, when kept in this way, are easily carried from bench to bench. 1 G. Lunge, Chem. Ind., 4. 349, 1881 ; Jo urn. Soc. Chem. Ind., I. 56, 1882. See page 72 for Winkler's barium carbonate process. 2 M. Scholtz, Arch. Pharm., 242. 575, 1904. VOLUMETRIC ANALYSIS. 29. Standard Solutions of Sodium Hydroxide. Sodium hydroxide NaOH. Molecular weight : 40'01 ; equivalent weight : 40'01. The most commonly employed solutions are : Grm. NaOH per c.c. N-Sodium hydroxide 0*04001 ^N-Sodium hydroxide ....... '020005 iN-Sodium hydroxide '008002 T VN-Sodium hydroxide '004001 Sodium hydroxide does not make an ideal standard solution, since it is liable to absorb carbon dioxide from the air. 1 A solution of sodium hydroxide is also particularly liable to attack glass, and to gradually lose its strength. In consequence, it is not usually advisable to make up a large stock of the solution. Kiister recommends storing sodium hydroxide in the nickel bottles made by Krupp of Berndorf. Glass stoppers and glass stopcocks are also objectionable because they are liable to become immovably fixed. It is therefore advis- able to use rubber stoppers and rubber plug cocks with the burettes working caustic lyes. To prepare a half-normal solution of sodium hydroxide, select a number of clear sticks of sodium hydroxide, pure by alcohol. Scrape off the opaque parts ; dissolve 27 grms. in water and make the solution up to 1100 c.c. in a Giles' flask. Recently boiled distilled water must be used in order to keep out the carbon dioxide. When the solution has attained a temperature of 15, pipette 25 c.c. into an Erlenmeyer's flask ; 2 add two or three drops of methyl orange solution as indicator ; run acid from the burette until the yellow colour of the methyl orange changes to an orange colour. Read the burette. Continue the titration, drop by drop, until the addition of a single drop of acid turns the solution pink. 3 Again read the burette. Repeat the titration with two more portions of 25 c.c. Let 26'42 c.c. denote the mean volume of the acid used in the titration, then 25 c.c. of the sodium hydroxide solution contains an amount of alkali corresponding with 26*42 c.c. of |N- acid. If the alkali be exactly JN, like the acid, 25 c.c. of acid would have been used. Hence, 25 c.c. of alkali must be increased by 26-42 less 25= 1-42 c.c,, or 1000 c.c. by FIG. 33. Burette with guard tube. 56-8 c.c. of water. Hence, pipette alkali solution from the flask until the liquid stands at the 1000-c.c. mark, and add 56'8 c.c. of water in order that the alkali may be exactly half-normal. Verify the accuracy of the dilution by repeating the titration. 1 E. Fleischer (Chem. News, 19. 203, 1869), inconsequence, prefers a iN-ammoma solution as a standard alkali. If solutions over this strength be used, they are liable to change, owing to the loss of ammonia ; and even ^N-solutions are liable to lose ammonia on a hot summer s day, when the temperature is, say, 25. There is also a difficulty with some of the indicators when ammonia solutions are used. R. Rempel, Chem. Ztg. , 9. 1906, 1884 ; G. T. Gerlach, Chem. Ind. t 12. 97. 1889. 2 The liquid should be thoroughly agitated during the titration. An Erlenmeyer's flask is generally preferable to ordinary narrow-necked flasks, to beaker and glass rod, or to basin and glass rod, because its contents are so accessible, and so easily agitated without stirring. 3 At first, it is best to work with comparison flasks, as described page 60. 5 66 A TREATISE ON CHEMICAL ANALYSIS. In the case of boric oxide titrations special care must be taken to exclude carbon dioxide. The burette and stock solution should be well protected from atmospheric carbon dioxide. The burette may be mounted as described in figs. 24 and 27 ; or, for small quantities, a soda-lime l tube may be fitted to the top of the burette as indicated in fig. 33. To test if the solution is really free from carbonates, make two parallel titrations with phenolphthalein as indicator. If the one solution be hot, and the other cold, carbonates will be absent if the two results agree. If otherwise, carbonates are present. Kuster's Process. In special cases Kiister's method 2 of preparation may be used. Boil distilled water ih a vessel through which the air is aspirated. The air is freed from carbon dioxide by passing it through a potash wash-bottle, and through soda-lime. This removes any carbon dioxide which would be dissolved by the water. Heat 200 c.c. of absolute alcohol in a round-bottom flask on a water bath until the alcohol boils. Meanwhile, weigh out about 12 grms. of bright, perfectly dry, metallic sodium freed from naphtha by pressing between folds of blotting paper. As soon as the alcohol begins to boil, add sodium, cut into small pieces, one piece at a time, to the alcohol. The reaction is violent; large volumes of hydrogen and alcohol vapour escape. Hence keep the flask covered with a watch-glass or small funnel to avoid loss by spurt- ing. The violence of the reaction gradually diminishes. When the reaction is over, sodium alcoholate, small bits of metallic sodium, and alcohol remain in the flask. Add the hot water, freed from carbon dioxide, in small quantities at a time to the flask containing the sodium alcoholate, etc. Boil off the alcohol by aspirating air through the flask containing the hot liquid. When the solution no longer smells of alcohol, cool the flask quickly ; add cold water freed from carbon dioxide. Transfer the mixture to a litre flask and dilute with COjrfree water to the mark on the neck. Shake. The solution quickly absorbs carbon dioxide from the air. Hence, let the neck of the flask be plugged with a rubber stopper fitted with a soda-lime tube as shown in fig. 33. The strength of the solution must be determined by titration with standard hydrochloric acid. It is best to work -with the solution as it stands, approximately half-normal, to avoid risk of contamination by exposure to the air. If the temperature of the soda solution be more than 3 different from the standard, an allowance must be made for expansion or contraction. The following correction table (Table XIII.) is used, as has been indicated by example, Tables VIII. and XL (pages 29 and 49), and when the temperature of the soda solution differs more than 3 from the standard 15. 3 For solutions more dilute, say y^N, the table for water (page 29) is used. 1 E. Fleischer (Ckem. News, 19. 303, 1869) prefers a mixture of Glauber's salt and caustic lime. A. Beutell, Chem. Ztg., 12. 86, 1887. 2 F. W. Kuster, Zeit. anorg. Chem , 13. 134, 1897 ; 41. 474, 1904 ; W. A. Smith, Zeit. phys. Chem., 25. 155, 1898 ; H. Ley, ib , 30. 205, 1899 ; W. R. Bousfield and T. M. Lowry, Phil. Trans , 204. 253, 1905; E. Ncitzel, Zeit. anal. Chem., 32. 422, 1893; W. N. Hartley, Journ. Chem. Soc., 26. 123, 1873; C. P. Hopkins, Journ. Amer. Chem. Soc., 23. 727, 1901. ' For explosions during the preparation of standard sodium hydroxide from metallic sodium, see A. Harpfand H. Fleissner, Zeit. chem. Apparaten/cunde, I. 534, 1906; F. W. Kuster, ib., 2. 535, 1906. For the preparation of colourless alcoholic solutions of potassium hydroxide, H. Thiele and R. Marc (Zeit. offent. Chem., 10. 386, 1904) recommend the use of potassium sulphate and barium hydroxide. See R. Gaze, Apoth. Ztg., 25. 668, 1910. 3 This table may also be used for N-H 2 S0 4 ; N-HN0 3 ; N-Na 2 C0 3 ; N-NaOH. It is based on A. Schulze, Zeit. anal. Chem., 21. 167 1882, and it can be easily adapted to other standard temperatures. VOLUMETRIC ANALYSIS. 6 7 Table XIII. Temperature Corrections for Normal Soda Solutions. (Standard temperature, 15.) Temp, | 1 2 3 4 5 6 7 8 9 + 2'0 + 1-9 + 17 + 1 5 + 1-3 1 +*!! + 0-9 + 0-7 + 0-5 +0-3 Unit. -0-2 -0-5 -07 -I'O 2 -1-3 -1-5 -1-8 -2-1 l -2-4 -27 -3-0 -3'4 -37 -4 ; 30. The Errors of Experiment in Volumetric Analysis. The principal errors incidental to volumetric work are : (1) unrecognised changes in the strength of the standard solutions ; (2) inaccurate measuring instruments (page 28) with care, these can usually be kept below 0*1 per cent. ; (3) weighing the substances used in the analysis (page 27) the errors in weighing need not exceed O'l per cent. ; (4) the use of dirty burettes, and drainage errors previously discussed ; and (5) a small excess of the standard solution is needed before the indicator will show the end of the reaction this error is somewhat variable in magnitude, and depends upon a number of factors. For instance, it depends upon the amount and concentration of the standard solution used in the titration, and on the magnitude of the drops falling from the burette. In the regular types of burette, one drop is nearly 0*05 c.c. Hence, if the reaction be not completed on the addition of the last drop, and completed with the addition of one drop more, it is often assumed that half a drop, namely, 0*025 c.c., completed the reaction. 1 The smaller the amount of substance used in the analysis, the smaller the volume of the standard solution required for the titration, and the greater the error of experiment. If 4 c.c. of the standard solution be used, half a drop, namely, 0*025 c.c., may lead to an error of 0*62 per cent. ; whereas, if 40 c.c. of the standard solution be needed for the titration, the corresponding error would be '06 2 per cent. Again, suppose three independent determinations of the ferric oxide be made on a given clay by titration with permanganate solution (1 c.c. representing 0*002527 grm. Fe 2 3 ), and 1*0, 1*01, and I'l c.c. were respectively required. The corresponding percentage amounts of ferric oxide would be respectively 0*2527, 0*2552, and 0*2780 grm. Suppose, further, that with another sample of clay 10*0, 10*01, and 10*1 c.c. of permanganate were needed in three independent titrations. The corresponding percentage amounts of ferric oxide would be 2*527, 2*529, and 2*552 grms. respectively. Again, suppose that 100*0, 100*01, arid 100*1 c.c. of permanganate were needed with a third sample of clay. The corresponding amounts of ferric oxide would be 25*270, 25*272, and 25*297 grms. Hence, the greater the volume of the standard solution used in the titration, the less the effect of small deviations in the measurement of the volume of the standard solution on the final result. Consequently, it is advisable to arrange the quantity of substance to be analysed, and the strength of the standard solution, so that a comparatively large volume oj the standard solution is used. Since 50-c.c. burettes are generally employed, something rather less than 50 c.c. is a convenient amount. For instance, 5 c.c. 1 In some processes, the ' : excess" of standard solution needed to produce a colour with the indicator is specially determined. This is done, for instance, in the uranium process for phos- phorus (page 600); the ferrocyanide process for zinc (page 367); the chromate process for chlorides (page 79) ; etc. 68 A TREATISE ON CHEMICAL ANALYSIS. of concentrated hydrochloric acid (sp. gr. 1*14) will require over 40 c.c. of N-NaOH solution. 1 If an acid of sp. gr. 1/015 be in question, this has 32 grms. of HC1 per 1000 c.c. Hence, 5 c.c. of this acid has 0'16 grms. of HC1. This would only take about 4 c.c. of N-NaOH. Hence, it is advisable to use a more dilute soda solution, say y^N-NaOH, in which case, about 40 c.c. of the standard solution will be needed. For the analysis of more dilute solutions of acid, more dilute solutions of "alkali will be needed say yi^N-NaOH. 2 It does not follow that the weaker the standard solution the more accurate the titration, because, if the standard solution be too weak, a relatively large volume will be required to colour the indicator after the end of the reaction, and nothing is really gained in accuracy. The above principles can be discussed from another point of view. Suppose 2N-acid is used for the titration of a gram of sodium hydroxide, then 1 c.c. of the 2N-acid will represent 8 per cent, of NaOH, and O'l c.c. will represent 0'8 per cent, of NaOH. Unavoidable variations in the amounts of standard acid required for the titration of duplicates will represent an error of 0*2 to 0'3 per cent. NaOH. This error is too great, and it is considerably reduced by using a standard acid but one-fourth the strength, namely, JN-acid. It is not advisable to make, say, y^N-NaOH by diluting 10 c.c. of T ^N-NaOH to 100 c.c. unless the solution so obtained is standardised. This arises from the fact that a small error in the measurement of the volume of the concentrated solution is multiplied into a relatively large error by the process of dilution. For a similar reason, if a concentrated solution of, say, sp. gr. 1*14 is under investigation, and it is convenient to work with a dilute standard solution, say, y^N-NaOH, it is not advisable to take, say, 0*5 c.c. of the concentrated acid for the titration. Rather should 20 c.c. of the concentrated acid be made up to a litre, and 25 c.c. of this solution be used for the titration. The 25 c.c. of the diluted acid would represent 0'5 c.c. of the concentrated acid. The reading is obviously the sharper, the narrower the burette. If the burette be narrow, it must have an inconvenient length if it is to hold enough fluid for a titration. A 50-c.c. burette reading to y 1 ^- c.c. will be about 60 cm. long, and the tube will have a bore of approximately 11 or 12 mm. In special cases, as in the colorimetric determination of iron, where a burette reading to -g 1 ^ c.c. is used, the burette is about 0*5 cm. bore, with a capacity of 10 c.c., and yet is 75 cm. long. These preliminary remarks may be used as a guide in the succeeding problems. For instance, to determine the amount of calcium carbonate in a sample of whiting, and the amount of calcium carbonate in a given sample of ground flint or ground Cornish stone. In the former case, the sample may have 98 per cent, of calcium carbonate, and in the latter, 2 per cent. Since 50 grms. of calcium carbonate will correspond with 1000 c.c. of N-HC1, 1 grm. of calcium carbonate will correspond with 20 c.c. of N-HCL Hence, if 50 c.c. of N-HC1 be added to the gram of whiting, and the excess of HC1 be titrated with N-NaOH, we shall require about 30 c.c. of the standard alkali. Again, 5 grms. of calcium carbonate will correspond with 1000 c.c. of y^N-HCl, and 1 grm. of calcium carbonate will correspond with 200 c.c. of y^N-HCl. Hence, if the ground flint contains 2 per cent, of calcium carbonate, 50 grms. of flint 1 From Table LXXVI., page 679, 100 c.c. of acid, sp. gr. 1'14, has 315 grms. HC1. Hence, since 1000 c.c. of N-HC1 has 36 '5 grms. of HC1, this is equivalent to 1000 c.c. of N-NaOH. Hence, 40 c.c. of NaOH will be equivalent to 1-46 grms. of HC1. But 315 grms. HC1 are in 1000 c.c. of the acid in question ; hence, 1*46 grms. HC1 will be found in approxi- mately 5 c.c. of the given acid. 2 To avoid frequently refilling the burette when many titrations have to be made, see the "automatic burettes," page 55. VOLUMETRIC ANALYSIS. 69 will want 200 c.c. of y^N-hydrochloric acid, or 5 grms. will want 20 c.c. of jijN-HCl. Hence, if 50 c.c. of y^N-HCl be mixed with 5 grms. of flint, the titration of the excess of acid will require about 30 c.c. of y^ 31. Direct Titrations Sodium and Potassium Carbonates. Suppose that it be required to determine the amount of sodium carbonate in a given sample of soda ash. A quantity of the powdered soda ash is placed in a weighing bottle, and dried at 100. When cold, the bottle and contents are weighed. A portion, approximately 1 grm., is transferred to an Erlenmeyer's flask, say 400 c.c. The weighing bottle and contents are again weighed. The difference between the two weights gives the amount of soda ash transferred to the flask : Weighing bottle and powder (before) . . . . 25*2931 grms. Weighing bottle and powder (after) . . . . . . 24'1110grms. Amount of soda ash . . ... . . 1 '1821 grms. Add about 150 c.c. of water. When all is dissolved, add about two drops of a solution of methyl orange, and titrate the solution with ^N-HC1 as indicated on page 65. Suppose that the mean of three experiments shows that 43 '6 c.c. of the standard solution have been used. Then, since 2HC1 -> 2NaCl + H 2 + C0 2 , 26*5 grms. of sodium carbonate correspond with 18*235 grms. of hydrogen chloride ; and since 1000 c.c. of the half-normal acid are equivalent to 26*5 grms. of sodium carbonate; or, 1 c.c. of acid represents 0*0265 grm. of Na 2 C0 3 ; or, 43*6 c.c. of the semi-normal acid represent 0*0265 x 43*6 = 1*1554 grms. Na 2 C0 3 per 1*1821 grms. of sample; hence, 100 grms. of the sample have 97*74 grms. of Na 2 C0 8 , i.e. 97*74 per cent, of sodium carbonate. All this arithmetic is summarised in the expression : = per cent, of Na 2 C0 3 , w where n represents the number of cubic centimetres of the standard JN-HC1 used in the titration, and w is the weight of powder used for the titration. It will be obvious that if w be exactly 2*65 grms., the number of c.c. used in the titration will represent directly the percentage amount of Na 2 C0 3 in the given sample. 1 It is sometimes most convenient to work according to the latter system (page 54). This question has to be solved : Is it quicker to weigh exactly the required amount, say, 1 *325 grms., or to take an approximate weight and calculate the corresponding amount of sodium carbonate as indicated above 1 The answer will largely depend upon the number of determinations to be made, and whether the weight of the powder is likely to change by the absorption of moisture, etc., while being weighed. Washing-soda and pear-lash may be treated by a similar method to that described above. 2 1 In that case we should have to use a 100-c.c. burette, or a normal solution. If 1*325 grms. of powder be weighed, twice n will represent the percentage amount of Na 2 C0 3 in the sample. 2 See page 72 seq. for a more detailed analysis of soda ash. A TREATISE ON CHEMICAL ANALYSIS. 32. Back Titrations l Calcium Carbonate. Suppose that whiting is to be investigated, weigh about 1 grm. of the dried sample in a weighing bottle, or on a piece of glazed paper. 2 Brush 3 every trace of powder from the paper into a dry 400-c.c. flask. Add, say, 50 c.c. of N-HC1. Take care that no powder sticks to the neck of the flask, and so escapes the action of the acid. 4 When all action has subsided, shake the flask vigorously. Add two drops of methyl orange, 5 and titrate the solution with N-NaOH until the yellow colour of the methyl orange appears. 6 The following results were obtained in an experiment on whiting : 1 F. Molir, Liebig's Ann., 86. 129, 1853. 2 Cut into the form of an elongated v. The narrow end is not too wide to pass into the neck of the flask. The paper is folded lengthwise so as to form a kind of gutter. The powder is placed near the broad end, and the narrow end of the paper is placed in the neck of the flask. The flask is then placed upright, and the powder is transferred to the flask by tapping and brushing. 3 Good camel-hair brushes, with the hair cut rather shorter than the brushes used by painters, are convenient for this purpose. A selected tail-feather of the Gallus domesticus FIG. 34. Dust brushes. (barnyard fowl), trimmed as indicated in fig. 34, is usually more effective than the brush. The brushes and feathers are best kept in the glass box mentioned on page 8. 4 Erlenmeyer's flasks (fig. 35a) with a broad neck are recommended for titrations. The neck and bottom should be such as to permit easy cleaning with the "bottle brush" and " policeman," page 92. When liquids are effervescing or boiling, the neck should be stopped with a bulb made for the purpose, or a funnel. The stoppers are afterwards washed. With the idea of preventing loss by spurting, Bolton's flasks (A. Dettloff. Chem. Ztg., 31. 181, 1907 ; German Pat. D.R.P. No. 183222, 1907; A. Gawalovski, Zeit. anal. Chem., 14. 170, 1875) FIG. 35. Modified flasks. may be used. These are unsymmetrical Erlenmeyer's flasks (fig. 356). Bolton's flasks are also convenient for decanting a liquid from a solid. The solid is allowed to settle while the flask is resting on a suitable stand (fig. 35c). The form shown in fig. 35d is due to E. Herzka (German Pat. D.R.G.M. No. 4357f>3, 1910). This is an improvement on Bolton's flask for preventing loss by spurting, because the fluid particles are projected against the walls of the flask when a liquid is boiling or effervescing, or when a gas is being passed rapidly through the contents of the flask (fig. 35e). 5 If phenolphthalein be used as indicator, the solution must be boiled, before the titration, in order to get rid of the carbon dioxide in the solution. 6 The ' ' transition tint " is often taken as the end point. VOLUMETRIC ANALYSIS. 71 Glazed paper and powder ....... 1 "9573 grms. Glazed paper after removal of powder ..... 0*5363 grras. Whiting 1-4210 grms. 50 c.c. of N-HC1 containing 35*2 grms. of HC1 per 1000 c.c. were added as indicated above. The acid remaining after the reaction required 25*5 c.c. of N-NaOH containing 39*1 grms. of NaOH per 1000 c.c. From the equation NaOH + HC1 -> NaCl + H 2 it follows that 40 '01 grms. of NaOH represent 36 -47 grms. of HC1. Hence, 0-0391 grm. of NaOH represents 0-03564 grm. HC1 ; and 1 c.c. of the NaOH solution represents 0*03564 grm. HC1, or 1*01 c.c. of the acid used for decom- posing the whiting. Consequently, 25'5 c.c. of alkali used in the titration represent 25'75 c.c. of the same hydrochloric acid. Hence, 50 - 25'75 = 24*25 c.c. of the same hydrochloric acid were taken up by the whiting. From the equation CaC0 3 + 2HC1 -> CaCl 2 + C0 2 + H 2 it follows that 72;94 grms. of HC1 represent 100*09 grm. of CaC0 3 . Hence, 1 c.c. of the given acid (or 0-0352 grm.) represents 0*0483 grm. CaC0 3 . Hence, the 24-25 c.c. of hydrochloric acid which reacted with the whiting represent 24*25 x 0*0483 = 1*17 grms. of CaC0 3 in the given sample, that is, 82'4 per cent. If solutions of exact normality had been employed, and 50 c.c. of N-HC1 and 25*7 c.c. of N-NaOH were required, it follows that 50 - 25'7 = 24-3 c.c. of N-HC1 would be taken up by the whiting. But 1 c.c. of the N-HC1 repre- sents 0-050 CaC0 3 ; consequently, 24*3 c.c. of the N-HC1 represent 1*215 grms. of CaC0 3 in the 1*421 grms. of sample, or the sample has 85 -5 per cent. CaC0 3 . Summarising this arithmetic, if n represents the number of cubic centimetres of N-NaOH used in the titration, 50 n represents the number of cubic centi- metres of N-HC1 " consumed " by the whiting. Hence, if w represents the weight of the sample in grams, the sample has w If we make w = 1*6667 grms., and use normal solutions of acid and alkali, it follows that 3(50 - n) will give the percentage amount of CaC0 3 in the sample. A similar method can be used for evaluating BaC0 3 , ZnO, MgC0 3 , MgO, etc. The method can be employed for determinations of the amount of calcium carbonate in samples of ground flint, ground stone, etc. A given weight of these materials is treated with deci-normal acid and alkali as indicated above. For instance, 5 grms. of flint (dried at 110) were digested with 50 c.c. of J^N-HCl on a water bath. The cold solution was titrated with ^N-NaOH. The mean of three experiments gave 39*51 c.c. of T VN-NaOH. Hence, 50-39*51 = 10*49 c.c. of y^N-HCl were taken up by the calcium carbonate in the flint. But 1000 c.c. of T VN-HC1 correspond with 5*0 grms. of CaC0 3 ; hence, 10*49 c.c. correspond with 10*49x0*005 = 0*05245 grm. of CaC0 3 per 5 grms. of flint that is, 1 *05 per cent. CaC0 3 . Summarising the arithmetic, it follows that T V(50 - n) here represents the percentage amount of CaC0 3 in the given sample. It will be observed that if yJ^N-NaOH had been used for the back titration with the idea of obtaining more exact results, it will be obvious that the increased accuracy is illusory, for the accuracy of the process depends on the exactness of the measurement of the more concentrated solution. A TREATISE ON CHEMICAL ANALYSIS. 33. Back Titration Mixed Hydroxides and Carbonates. Suppose it be required to determine the amount of caustic soda in a given sample of soda ash. 1 Dissolve, say, 50 grms. in warm water, filter and wash the insoluble matter, if any be present. 2 Make the washings up to 1 litre. Titrate 20 c.c., equivalent to 1 grm., as indicated above with JN-acid and methyl orange as indicator. This titration includes sodium carbonate as well as caustic soda and sodium sulphide, if these be present. To determine the caustic soda, pipette, say, 40 c.c. into a 200-c.c. flask, add a slight excess 3 of barium chloride (20 c.c. of a 10 per cent, solution will suffice). Barium carbonate is precipitated. Make the liquid with the solid in suspension up to- the mark with cold distilled water. Shake the corked flask. When the solution is cold, 4 make it up to the mark on the neck with cold distilled water. Shake the flask. Either let the precipitate settle, and pipette, off 50 c.c. from the clear, or filter a portion of the solution, and take 50 c.c. from the clear filtrate. 5 The standard flask shown in fig. 36 will be found very con- venient for withdrawing aliquot portions from, say, a litre of solution. It is an ordinary litre flask with a side tap sealed on to the flask before graduation, as shown in the diagram. Titrate the 50 c.c. with, say, y^N-HCl, using methyl orange as indicator. Each cubic centimetre of the solution will represent '004006 grm. of NaOH per gram of soda ash. 6 FIG. 36. Measuring liquids. 1 For another method, see page 64. The method in the text is due to C. Winkler. A. C. Andersen, Tidskr. Kem. Farm. Tempi, n. 161, 1908 ; Journ. Pharm. Chim., 28. 370, 1908 ; R. B. Warder, Chem. News, 43. 228, 1881 ; G. Lunge and W. Lohhofer, Zeit. angew. Chem., 14. 1125, 1901 ; G. Lunge, ib., 10. 169, 1897 ; J. Tillmans and 0. Heublein, ib., 24. 874, 1911. (There is an error due to the escape of carbon dioxide obviated by the method indicated in the text.) F. W. Kiister, Zeit. anorg. Chem., 13. 142, 1897; C. F. Cross and E. J. Bevan, Zeit. anal. Chem., 37. 685, 1898 ; K. Novolny, Zeit. K. Novolny, Zeit. anorg. Chem., 51. 181. 1909 ; M. le Blanc, ib., 53. 344, 1907 ; F. W. Kiister, , Electrochem. , u. 453, 1905 ; M. le Blanc and ib., 13. 127, 1896 ; S. P. L. Sorensen and A C. Andersen, Zeit. anal. Chem., 47. 279, 1908 ; H. ScKerand E. Cramer, Tonind. Ztg., 18. 593, 1894. 2 The amount can be estimated by drying and weighing, if needed (page 155). 3 According to W. Smith (Journ. Soc. Chem. Ind., I. 85, 1882), if a large excess of barium chloride be used, some alkali will be lost by absorption. Exact precipitation is necessary to secure accuracy. S. P. L. Sorensen, Zeit. anal. Chem., 45. 220, 1906 ; 47. 279, 1908. 4 A COOLING BOX, in cases like this, where speed is essential, will be found a great convenience. This can be made of wood lined with sheet lead or sheet copper, and provided with entrance and overflow pipes for the cold water. The flask or flasks to be cooled are placed in the box, and a current of water allowed to flow through the box. To prevent the flasks overturning, either a series of lead rings can be provided to fit on the necks of the flasks, or a shelf with sockets may be fitted to the box. 5 According to A. Miiller (Journ. prakt. Chem. (1), 83. 384, 1861) the filter paper absorbs an appreciable quantity of the barium salt, and gives low results. But here we are now only concerned with the sodium hydroxide. 6 Sodium sulphide, if present, will be included in this result, and a deduction must be made after the sulphides have been determined. VOLUMETRIC ANALYSIS. 73 Sulphides in the presence of chlorides can be determined volumetrically by Lestelle's process, 1 in which 100 c.c. of a solution containing the equivalent of, say, 5 grms. of soda ash are titrated with an ammoniacal solution of silver nitrate, 2 until no black precipitate of silver sulphide Ag 2 S is formed on adding another drop of the silver nitrate solution. To observe the end of the titration accurately, the solution is filtered towards the end of the operation, 3 and the titration is continued if necessary. It may be necessary to repeat the filtration a number of times. With the above proportions, 1 c.c. of the standard solution represents O'l per cent, of sodium sulphide in the given sample. 4 Sulphites are determined by acidulating 100 c.c. of a solution containing the equivalent of 5 grms. of the sample with acetic acid. Add starch solution, and titrate with a standard iodine solution 6 until the blue colour of starch iodide permanently appears. For every 1 c.c. of the silver solution used in the sulphide test, deduct 1'3 c.c. from the iodine solution used in the titration. Each c.c. then represents 0'000001613 grrn. Na 2 S0 3 , or 0-0000323 per cent, of Na 2 SO ? . Sodium sulphate is best determined gravimetrically as barium sulphate (page 618) ; sodium chloride is determined by Mohr's process (page 79) ; iron, by the permanganate process (page 198) ; and sodium silicate, by evaporation . with hydrochloric acid (page 167). 34. Correction for the Volume of Suspended Solids. It is here necessary to examine the nature of the error due to the assumption that the solid barium carbonate in suspension occupies no appreciable volume. The error will obviously be negligibly small when the volume of the solid is small in comparison with the volume of the solution, but the error may be appreci- able when the volume of the precipitate is large in comparison with the volume of the flask. 6 It is therefore necessary to look into this subject more closely. By a previous titration we found the amount of "sodium carbonate" per 20 c.c. of solution. Let a represent the "sodium carbonate" in the 100-c.c. flask. One gram of sodium carbonate is equivalent to 1'862 grms. of barium 1 H. Lestelle, Compt. Rend., 55. 739, 1862. 2 AMMONIACAL SILVER NITRATE. Dissolve 13'810 grms. of silver in pure nitric acid, add 250 c.c. of concentrated ammonia, and dilute the solution to a litre. 1 c.c. represents 0'005 grm. Na 2 S. The ammonia keeps the silver chloride in solution. 3 Some recommend a Beale's tube for this purpose. A piece of filter paper is tied over the Iqjver end A, fig. 37, and a piece of muslin is tied over the paper to prevent it breaking. When A] FIG. 37. Beale's filter pipette. the end A is dipped in the mixture, clear liquid rises in the cylinder. This is poured from the little spout B and tested. If the titration be not completed, the liquid withdrawn must be returned to the main solution. 4 If no sulphites are present, the titration for sulphides is best made by the process used for sulphites. 5 Iodine solution : 3 '249 grms. of iodine per litre. See page 288 for details preparation, precautions, etc. 6 0. Eberhard, Zeit. offent. Chem., 4. 867, 1898 ; M. Ruoss, Zeit. anal Chem., 37. 422, 1898 ; M. Bicard and H. p ellet, Zeit. anal. Chem., 24. 460, 1885 ; Bull. VAssoc. Chim. Sue., I. 230, 1885 ; E. Lenoble, Bull. Soc. Chim. (3), n. 336, 1895; W. R. Smith, Journ. Amer. Chim. Soc., 31. 935, 1909. Advantage may be taken of the fact that dilute solutions of oxalic acid exert no appreciable action on alkaline earthy carbonates, to titrate with standard oxalic acid all the liquid with the precipitate instead of an aliquot portion. Phenolphthalein as indicator. 74 A TREATISE ON CHEMICAL ANALYSIS. carbonate. Hence, a grms. of sodium carbonate, found by titration, will be equivalent to l*862a grms. of barium carbonate. The specific gravity of dry barium carbonate is 4 -3. Hence, the volume of the barium carbonate will be 0*433a c.c. If the soda ash contained 98 per cent, of sodium carbonate, a = 0*98. Hence, the barium carbonate in the 100-c.c. flask occupied 0*98 x 0*433 = 0*424 c.c. Hence, the 100-c.c. flask contained 100 - 0*424, or 99*6 c.c. By drawing off 50 c.c. for the titration we drew, not the equivalent of 0'5 grm., but of 0'504 grm. Hence, each c.c. of the y^N-acid represents 0*004006 grm. of NaOH per 1*04 grms. of the soda ash, or 0*003974 grms. NaOH per grm. of soda ash. The error may be somewhat serious when appreciable amounts of NaOH are in ques- tion, although the correction is usually neglected in practice. 1 The process just indicated may be a little more complicated. In determining soluble sulphates, a known amount of barium chloride may be added to the slightly acidified solution. The solution is boiled to get rid of carbon dioxide and then exactly neutralised. An excess of standard solution of sodium carbonate is then added. The solution is made up to a definite volume and the excess of sodium carbonate in an aliquot portion of the clear titrated with standard acid. This gives data sufficient to calculate the amount of barium chloride which reacted with the sodium carbonate. The remaining barium chloride must have reacted with the soluble sulphates. We have both barium sulphate and barium carbonate in suspension. For the sake of illustration, assume that it is required to calculate the amount of barium chloride in a given solution. Add an excess of sodium carbonate, say, N or 1000 mgrms. (i.e. 1 grm.), and make the solution with the precipitate in suspension up to the V c.c. mark in a standard flask, say, F=50 c.c. Either pipette or filter v c.c. of the clear solution, suppose v = 25 c.c. The sodium carbonate is determined by titration with standard acid. Suppose that 22*35 c.c. of iN-acid are needed. Hence, 50 c.c. needed 22*35 x 2 = 44*7 c.c. of iN-acid ; then the excess of sodium carbonate, w, was 44*7 x J x 53 = 473*8 mgrms. But 1 grm. of sodium carbonate is equivalent to 1*862 grms. barium carbonate; and 1000-473*8 = 526*2 grms. of sodium carbonate will represent a = 1*862 x 526*2 = 979*8 mgrms. barium carbonate, or a = 1*862(1000 -w) (1) This number a = 979*8 mgrms. is not quite right, because we did not allow for the volume of the precipitate in drawing off the 25 c.c. of clear solution. Let A be the right value which would have been obtained had the precipitate been filtered off, washed, and the filtrate made up to the 100 c.c. Then, 4 = 1-862(1000- TF) (2) where W represents the correct excess of sodium carbonate which would have been obtained in the filtrate if the precipitate occupied a negligibly small volume. Let s denote the specific gravity of the precipitate (sp. gr. BaC0 3 = 4*3). The precipitate then occupies A/lOOQs c.c., and , wA wA W = w - ; w - W = WOQsV ~ 1000s F The error is then, from (1) and (2), 1 Say, checking the composition of black ash ; vide Lunge (I.e.). The neglect may be some- times justified when comparable results are alone required. In some cases the error might even be greater than the percentage amount of the constituent in question. A different method of analysis, of course, is then used. VOLUMETRIC ANALYSIS. 75 Consequently, = 1-SWAW ~ __ lOOOsF- A ~ 1000s 7- l-862w ' The terms A and 1 -8G2w in the denominators are very small in comparison with lOOOsF, so that these terms can be omitted without materially affecting the value of the fractions. Hence, the error 1-862 x 473-8x979-8 1000x4-3x50 Hence the total weight of the barium carbonate is 979-8 + 4-0 = 983-8 mgrms. The amount of barium chloride corresponding with this barium carbonate is readily calculated. All this arithmetic may be summarised in the formula : Weight of precipitate = r(N - w)(\ + (3) where N denotes the weight of the precipitating agent in mgrms. (1000) ; s, the specific gravity of the precipitate (4'3) ; F, the volume of the measuring flask (50) ; iv, the excess of the precipitating agent (47 3 -8) ; and r, the ratio of the equi- valent weight of the precipitate and precipitating agent (98 -685 -f- 53 -00 = 1-862). A great many " rapid " processes involve the use of one or other of these two methods of correction. Once the reasoning is understood, there is no need to go through all the steps. The numbers corresponding with the letters can be substituted directly in formula (3). 1 A mathematical formula is not the end, but the means of attaining the end. Students sometimes get the opposite idea, just as if a builder erected a mansion for the sake of showing off the ladders and scaffolding employed in its construction. 35. Sodium Silicate Water-Glass. In commercial water-glass sodium silicate the ratio Si0 2 : Na 2 generally lies between 2:1 and 4:1. In analysis, it is generally assumed 2 that free caustic alkali is present when this ratio falls below 1:1, corresponding with about 50-8 per cent, of Na 2 0, or with Na 2 O.Si0 2 . Water-glass is found on the market in the form of a solid yellowish or greenish glass, or in a more or less viscid aqueous solution. For analysis, 10 grms. of the solid are ground to a very fine powder and dissolved in hot water, or 20 grms. of the liquid are digested in water. In both cases the solution is made up to 500 c.c. The solution is allowed to stand some time in order to allow the insoluble matters to settle as a sediment, 3 and to allow air-bubbles to rise to the surface. 1. Combined and Free Alkali* Titrate 100 c.c. of the clear solution with N- or JN-HC1, using phenolphthalein as indicator. The action of the acid is represented by the equation Na 2 Si0 3 + 2HC1 = 2NaCl -f H 2 Si0 3 . 1 It is also possible to compute the volume of the solid matter in suspension by an applica- tion of the principle indicated in 33, page 72. 2 There are some reasons for doubting if the assumption is valid. 3 The sediment can be filtered off. washed, and determined later. J. Ordway, Chem. News, 9. 61, 1864. 4 The term " combined alkali " does not here include the alkali which may be present in the form of neutral salts sodium chloride, etc. There is usually so little sediment that the cor- rection indicated on page 73 is not needed. 76 A TREATISE ON CHEMICAL ANALYSIS. The liberated silicic acid does not affect the indicator. The reaction towards the end is very slow, and the indicator may appear to be permanently decolorised before the reaction is complete. There is therefore some danger of under- titra- tion. If, however, a large excess of sodium chloride be added before the titration, this difficulty does not give much trouble. 1 From the preceding equation, we see that every c.c. of N-hydrochloric acid corresponds with O031 grm. of Na 2 0. Hence, if w grms. of the sample were made up to 500 c.c., and 100 c.c. taken for the titration, and if n c.c. of the N-hydrochloric acid were consumed in the titra- tion, the sample has 0'155w grm. of Na 2 per w grms. of sample. 2. Free Alkali. To 100 c.c. of the solution, 2 add 100 c.c. of a solution of 10 grms. crystalline barium chloride gradually with constant stirring. Make the solution up to 250 c.c. Shake well. Filter through dry filter paper. Reject the first 20-30 c.c., and titrate the next 100 c.c. which passes through the paper with T yS"-hydrochloric acid and phenolphthalein. 3 Since 1 c.c. of T T ^N-hydro- chloric acid corresponds with 0*0031 grm. Na. 2 0, if n c.c. of this acid are used in the titration, n x 0'0031 grm. of Na 2 is present in 100 c.c. Hence, n x 0*0031 x 2'5 = > 00775ft grm. of Na 2 is present in the 250 c.c. Hence, the sample has 0'03875w grm. Na 2 per w grms. of the sample taken. The difference between the results of this and the preceding, titration represents the amount of combined alkali. 4 Calculation. Suppose that 20 grms. of the powder were made up to 500 c.c., and that 42 '4 c.c. of N-hydrochloric acid were required for the first titration, and 3 '7 c.c. of y^N-hydrochloric acid for the second titration. The results are expressed : Free alkali 072 per cent. Combined alkali 32 '11 ,, Silica 63'81 ,, Alkaline chlorides, etc. Insoluble matters Water (loss on ignition, page 157) .... The basicity 5 of the sample, that is, the ratio Si0 2 : Na 2 0, or the molecular proportion of the silica to the free and combined alkalies, regarded as soda, is nearly 2. 6 36. The Volumetric Determination of Silver and Chlorine Volhard's Process. When potassium or ammonium thiocyanate is added to a solution of silver nitrate, a white precipitate of silver thiocyanate is formed : AgN0 3 + KCNS = AgCNS + NH 4 N0 3 . 1 R. T. Thomson (see page 63) ; C. Lunge and W. T. Lohbfer, Zeit. angew. Chem., 14. 2 The solution should be as concentrated as possible. Hence, some prefer to take 10 grms. made up to 100 c.c. for this determination. 3 P. Heermann. Chem. Ztg. t 28. 883, 1904. If sodium carbonate be present, it will be pre- cipitated as barium carbonate by the barium chloride 4 The term '_' alkali" used here is supposed to represent soda, but potash may be present. If a distinction is necessary, a special analysis may be needed to determine the ratio K 2 : Na 2 0. The silica is determined by evaporating 100 c.c. with hydrochloric acid, as indicated on page 167. The filtrate is treated with ammonia, ammonium carbonate, and ammonium oxalate. Filter. The filtrate is evaporated with hydrochloric acid, ignited to drive off ammonium salts, and the residue is supposed to represent the " alkalies." See page 226 for details. The " total alkalies " so obtained, less the free and combined alkalies, are regarded as alkali belonging to the neutral salts alkaline chloride, etc. 5 The term "acidity " might perhaps be more appropriately applied for the ratio Si0 2 . Na 2 0. 6 Basicity = (63 '81 * 60)/(82 "S3 -f 62)= 1 -96. VOLUMETRIC ANALYSIS. 77 The precipitate is insoluble in nitric acid. The reaction occurs even in presence of ferric salts, and only when all the silver has been precipitated does the thiocyanate react with the ferric salt, forming the characteristic blood-red colour of ferric thiocyanate. The persistent appearance of this coloration shows that the titration is finished. Ferric nitrate or sulphate can be used as indicator, but not the chloride. Similarly, the water and other reagents must be free from chlorides. 1 This method is due to Charpentier, but usually called Volhard's process. 2 Standardisation of the Solutions. A standard solution of, say T VN, silver nitrate is made by dissolving 16*989 grms. of pure silver nitrate 3 in water, and making the solution up to a litre with distilled water at 15. A solution of potas- sium thiocyanate is made by dissolving 10 grms. of the salt in water and making the solution up to a litre. 4 This latter solution is standardised by pipetting, say, 25 c.c. of the silver nitrate solution into a white porcelain dish. 5 Run the potassium thiocyanate solution from the burette until most of the silver is precipitated 6 as white silver thiocyanate, AgCNS, and then add about 5 c.c. of solution of ferric alum 7 to act as indicator. Each drop of the thiocyanate solution produces a white cloud of silver thiocyanate coloured with a reddish halo of ferric thiocyanate. The colour disappears on shaking, provided any silver nitrate remains in the solution. When all the silver has been precipitated as insoluble silver thiocyanate, one drop of the potassium thiocyanate will colour the solution reddish brown. The colour persists after a vigorous shaking. The reddish-brown colour produced when the titration is not quite finished is but slowly removed when the solution is vigorously agitated. Hence, it is necessary to guard against "under-titration." 8 Suppose that 24 '3 c.c. of the thiocyanate be required for the titration, it follows that 24 '3 c.c. of the potassium thiocyanate corresponds with 25 c.c. of silver nitrate, which in turn corresponds with 0*4247 grm. of silver. If the 1 Chlorides interfere not only by removing silver as insoluble chloride, but they also interfere with the end point, since silver chloride removes part of the red ferric thiocyanate from the solution. 2 J. Volhard, Journ. prakt. Chem. (2), 9. 217, 1874; Ltebig's Ann., 190. 1, 1878; Zeit. anal. Chem., 13. 171, 1874 ; 17. 482, 1878 ; J. B. Schober, ib. t 17. 467, 1878 ; G. Briigelmann, ib., 16. 1, 1877 ; F. A. Falk, Ber., 8. 12, 1875; B. G. Gentil, Wochenschr. Pharm., 30. 133, 1903 : P. Charpentier, Bull. Soc. Ing. Gin. France, 325, 1870 ; E.Drechsel, Journ. prakt. Chem. (2), 15. 191, 1877 ; C. Mann, Oester. Zeit. Berg. Hutt., 26. 426, 1878; H. F. von Jiiptner, ib., 28. 33, 51, 1880 ; 0. Lindemann, Berg. Hiitt. Ztg., 35. 333, 1877 ; F. T. Shutt and H. W. Charlton, Chem. News, 94. 258, 1906 ; Trans. Roij. Soc. Canada, n. 67, 1906 ; A. Dubosc, Ann. Cliim. Anal., 9. 45,1904; C. Hoitsema, Zeit. angew. Chem., 17. 647, 1904; V. Rothmund and A. Burgstaller, Zeit. anorg. Chem., 63. 330, 1909 ; E. M. Hamilton, Min. Scientific Press, IO2. 364, 1911 ; A. T. Stuart, Journ. Amer. Chem. Soc., 33. 1344, 1911 ; A. E. Knorr, ib., 19. 814, 1897 ; T. K. Rose, Journ. Chem. Soc., 77. 232, 1900. 3 Powdered and dried at 130 for about an hour. The silver nitrate solution may also be prepared by dissolving 10'8 grms. of pure metallic silver in 50 c.c. of dilute nitric acid. Boil off the nitrous fumes, since they interfere with the indicator later on. Dilute the solution to a litre. 4 It is very difficult to weigh out exactly the required quantity of the potassium thiocyanate, owing to its deliquescent character. Hence, the solution must be standardised. If a solution of definite strength be required, it is best to take rather more salt than is necessary, and, when the strength of the solution has been determined, dilute the solution as indicated on page 48. 5 Or in an Erlenmeyer's flask or beaker, and then titrate with a white background as described on page 60. 6 The results are low if the temperature be much higher than 25. The nitric acid then " bleaches" the ferric thiocyanate. Sufficient nitric acid should be present to remove the colour produced by the indicator before the titration is finished. An excess of nitric acid does no particular harm, but a large excess gives low results, since it retards the formation of the ferric thiocyanate at the end of the titration. 7 FERRIC ALUM SOLUTION : Dissolve 10 grms. of ferric alum in 90 c.c. of water. 8 In case of " over-titration " a known volume of the standard silver nitrate solution can be added to the solution, and the titration continued Due allowance must, of coarse, be made for the additional silver. 7 8 A TREATISE ON CHEMICAL ANALYSIS. potassium thiocyanate is to be exactly ^th normal, the 24 '3 c.c. must be diluted to 25 c.c. Hence, proceed as indicated on page 48. Determination of Silver. The titration of the nitric acid solution of the sample x under investigation is conducted in the same manner as that described for the standardisation of the potassium thiocyanate solution. The method is valuable provided an accuracy of y^th per cent, will suffice. The results are a little low when large quantities of silver are present, owing possibly to (1) the adsorption of some of the silver nitrate by the precipitated silver thiocyanate ; and (2) the action of tne precipitate on the solution. The process can also be used in the presence of barium, bismuth, antimony, arsenic, lead, iron, manganese, zinc, etc. Mercury interferes with the titration, since it is precipitated by the thiocyanate. Copper, cobalt, and nickel form coloured solutions and thus interfere with the indicator, although it is claimed that 70 per cent, of copper will not spoil the determination. Determination of Chlorides. By this method silver can be estimated in the presence of nitric acid, whereas Mohr's chromate process requires a neutral solution. Hence, the thiocyanate process can be used to estimate the haloids chlorine, bromine, and iodine in the presence of phosphates and other salts which give precipitates with silver in neutral solutions, but not in acid solutions. Chlorides, bromides, and iodides in solution are determined by adding an excess of a standard solution of silver nitrate. The excess of silver which has not reacted to form silver chloride is determined by titration with the standard thiocyanate. 2 Thus, W. Dittmar determined chlorides in sea-water by precipi- tating the chloride with an excess of standard silver nitrate, filtering off the silver chloride, and titrating the excess of silver nitrate as indicated above. It is important to filter off 3 the precipitate of silver chloride, 4 owing to the fact that the silver chloride reacts with the soluble thiocyanate in such a way that 165 times as much excess T yST -ammonium thiocyanate per 200 c.c. is needed to produce the red coloration in the presence of silver chloride as is needed when the chloride is absent. 5 The results in the presence of the precipitated chloride are more variable, and some 2 per cent, low, under conditions where the results in absence of the precipitated chloride are satisfactorily constant ; and but 0'15 per cent, high, when the silver chloride has been filtered off. The presence of sulphates interferes with Volhard's process, because appreciable quantities of silver sulphate are precipitated with the chloride and thiocyanate. 6 EXAMPLE. An aliquot portion, say, 20 c.c., of the solution of soda ash indicated on page 72, or an aqueous solution of a gram of the dry salt, is treated with an excess of nitric acid. Then add an excess of, say, 25 c.c. of j^N-silver nitrate solution, and titrate with j^N-potassium thiocyanate solution as indicated above. Suppose that 24*2 c.c. of the JpN-potassium thiocyanate solution are needed. This is equivalent to 24*2 c.c. of the silver nitrate solution. Hence, 25- 24'2 = 0'8 c.c. of the silver nitrate solution has formed silver chloride. But 169*89 grms. of silver nitrate represent 58'46 grins, of sodium chloride. Each c.c. of the silver nitrate solution contains 0'016989 grm. of silver nitrate, which is equivalent to 0*005846 grm. of sodium chloride. Hence, 0-8x0-005846 = 0-0046768 grm. of sodium chloride. Hence, the sample has the equivalent of 0'47 per cent, of sodium chloride. 1 For example, a silver coin dissolved in nitric acid. >2 For iodides, the nitric acid is added after the standard silver solution. 3 Say, in a filter tube containing glass-wool (page 102). 4 M. A. Rosanoff and A. E. Hill, Journ. Amer. Chem. Soc., 29. 269, 1407, 1907 ; Chem. News, 96. 264, 274, 299, 1907 ; G. A. Sanger, Proc. Amer. Acad., 26. 34, 1879 ; G. Briigel- mann, Zeit. anal. Chem., 16. 1, 1877 ; E. Dreschel, ib., 16. 351, 1877 ; Journ. prakt. Chem., (2), 15. 191, 1877 ; 0. Knupfer, Zeit. phys. Chem., 25. 266, 1898 ; L. L. de Koninck, Chem. Ztg., 15. 1558, 1891. 5 With bromides and iodides, the interference of the precipitated salt is negligibly small. 6 L. W. Andrews, Journ. Amer. Chem. Soc., 29. 275, 1907. VOLUMETRIC ANALYSIS. 79 37. The Volumetric Determination of Chlorides Mohr's Chromate Process. If a solution of silver nitrate be gradually added to a neutral solution of an alkaline chloride containing a little potassium chromate, silver chloride will be precipitated. The reaction is represented by the equation AgN0 3 + KC1 = AgCl + KN0 3 . When the reaction is complete, any further addition of -the silver nitrate reacts with the potassium chromate : 2AgN0 8 + K 2 O0 4 - Ag 2 Cr0 4 + 2KN0 3 . The permanent red tint of the silver chromate so developed indicates that the reaction between the silver nitrate and the silver chloride is at an end. By using a standard solution of silver nitrate, the reaction is quantitative, and was used by Mohr for the volumetric determination of chlorides. 1 If acids be present, the chromate may form dichromate, which interferes with the recognition of the end point, and, moreover, acids dissolve appreciable quantities of the silver chromate, and the end point is then indistinct. Hence, any acids which may be present should be neutralised by sodium carbonate, or, better, by pure calcium carbonate. If the solution be feebly alkaline, owing to the presence of a slight excess of carbonate, a little silver carbonate will be precipitated with the chromate. Alkaline solutions should accordingly be neutralised with acetic or nitric acid. A slight alkalinity is not so baneful as a slight acidity. The Titration. Dissolve a gram of the dried sample under investigation in water and make the solution up to 100 c.c. Pipette 25 c.c. into an Erlenmeyer's flask and add 1 c.c. of potassium chromate solution. 2 Gradually add y^N -silver nitrate solution 3 with constant stirring or agitation. When the chloride is all precipitated a red coloration of silver chromate will appear, which does not disappear on agitation of the solution. The silver nitrate must be added very slowly towards the end of the titration, so as to run no risk of over-titration. The titration is made with the flask over a sheet of white paper, or a white tile. A certain amount of silver nitrate is needed to colour the indicator. Hence, Mohr recommends adding, drop by drop, y^N-sodium chloride solution until the red colour of the indicator gives way to the yellow colour of the alkaline chromate. Deduct the volume of sodium chloride so employed from the silver nitrate used in the titration. The result represents the amount of silver nitrate corresponding with the chlorides undergoing titration. If the precipitate of silver chloride is very large, the end point will be some- what masked. In that case, add a slight excess, 1-2 c.c., of a dilute standard solution of sodium chloride ( T VN). Filter and wash the precipitate twice with 1 F. Mohr, Liebig's Ann., 97. 335, 1856. Mohr first used potassium arsenate as indicator ; A. Levol (Bull. Soc. JSncour., $2. 220, 1853; Journ. prakt. Chem. (1), 60. 384, 1853) used sodium phosphate ; F. Stolba (Zeit. anal. Chem., 13. 65, 1874), potassium calcium chromate. 2 POTASSIUM CHROMATE SOLUTION. Dissolve 1 grm. of the salt in 100 c.c of water. The chromate must be free from chlorides. To test for chlorides, dissolve the chromate in water, add a little silver nitrate and then some nitric acid. If the red precipitate which forms dissolves completely, forming a clear solution, chlorides are absent. 3 SILVER NITRATE SOLUTION. Dissolve 16 '869 grms. of pure silver nitrate in a litre of water. 1 c.c. corresponds with 0'003518 grm. of chlorine. If this solution be standardised by dissolving 5 '806 grms. of pure sodium chloride in a litre of water, so as to form a T VN-sodium chloride solution, 25 c.c. of the jVN-sodium chloride will require 25'1 c.c., instead of 25 c.c., of the rVN -silver nitrate solution. 8o A TREATISE ON CHEMICAL ANALYSIS. water, and again titrate the clear solution, making allowance for the extra sodium chloride added. Disturbing Agents. Salts of lead, bismuth, barium, and iron will form insoluble chromates, and should therefore be absent; coloured salts cobalt, nickel, and copper obscure the end point and are therefore objectionable. Pellet 1 has shown that arsenates, arsenites, phosphates, and fluorides do not interfere, since silver chromate forms earlier than silver fluoride, phosphate, etc. The presence of any salt which augments the solubility of the silver chromate gives high results, since more silver nitrate must be added to produce the red chromate than corresponds with the end of the reaction. Hence, also, salts which augment the solubility of silver chromate should be absent for instance, the nitrates of the alkalies and alkaline earths, 2 and ammonium salts. The solubility of the silver chromate is also augmented by raising the temperature of the solution, 3 and, in consequence, it is best to titrate cold solutions. An excess of potassium chromate diminishes the. solubility of the silver chromate, and, in consequence, a larger quantity of the indicator is employed than would normally be the case. If too little potassium chromate be employed, more silver nitrate is needed to produce the red-coloured chromate. 4 Difficulty with the Indicator. In spite of these ^precautions, a measurable excess of silver nitrate solution must be added after the reaction with the chlorides is ended, before the permanent red colour of the silver chromate is developed. This is particularly noticeable when small quantities of chloride are present in large volumes of solution, and Winkler 5 has accordingly drawn up an empirical correction table (Table XIV.) representing the excess of silver nitrate solution (1 c.c. equivalent to O001 grm. chlorine) required to develop the colour of the indicator in 100 c.c. of solution, when 1 c.c. of a one per cent, solution of potassium chromate is employed. Table XIV. Correction for Indicator in Mohr's Chromate Process for Chlorides. Silver Silver Silver solution Deduction. solution Deduction. solution Deduction. used. used. used. c.c. c.c. c.c. c.c. c.c. c.c. 0-2 0-20 0-8 0-39 5-0 0'50 0-3 0-25 09 0'40 6-0 0-52 0-4 0-30 1-0 0-41 7-0 0-54 0-5 0-33 2-0 0-44 8-0 , 0-56 0-6 0-36 3-0 0-46 9-0 0-58 07 0-38 4-0 0-48 10-0 0-60 Of course this table only refers to solutions of the concentration stated. Lunge deducts 0'2 c.c. from the amount of y^N-silver nitrate solution used both in titrating a solution of sodium chloride which requires 50 c.c. of the standard solution, and in titrating sodium sulphate which contains a little sodium chloride. Winkler's experiments show that Lunge's correction is prob- 1 H. Pellet, Bull. Soc. Chim. (2), 28. 68, 1877. 2 F. C. Carpenter, Journ. Soc. Chem. Ind., 5. 286, 1886 sodium and calcium nitrates have least effect ; ammonium, potassium, and magnesium the greatest effect. G. Biscaro, Ohem. News, 53. 67, 1886. 3 One part of silver chromate is soluble in 16,666 parts of water at 15'5, according to W. G. Young, Analyst, 18. 124, 1893. 4 L. L. de Koninck and E. Nihoul, Rev. Univ. Mines (3), 16. 42, 1891 ; Zeit. angew. Chem., 4. 295, 1891. 5 L. W. Winkler, Zeit. anal. Chem, 40. 596, 1901. VOLUMETRIC ANALYSIS. 8 1 ably about right. Fresenius l recommends making the solution to be titrated approximately the same volume and strength as the solution used in standardising the silver solution. Hazeu 2 recommends correcting the solution used in titrating by deducting 003 F + 0'02 c.c. from the standard silver solution, when V denotes the volume of the solution which has just been titrated. If the correction be neglected when, say, a y^N-silver nitrate solution is used, the resulting error is about 0'4 per cent. 3 Modification for Small Amounts of Chlorides. When very small amounts of chlorides are in question, more dilute solutions of silver nitrate are employed say, 4 '7 95 grms. of silver nitrate per litre ; 1 c.c. of this solution represents O'OOl grin, of chlorine ; or 2'3975 grms. of silver nitrate per litre; 1 c.c. of this solution represents O'OOOo grm. of chlorine. Suppose that the mixed chlorides obtained in the determination of the alkalies in clays be in question. Add 1 c.c. of the potassium chromate solution. Make the solution up to 50 c.c. in an Erlenmeyer's flask. Add 50 c.c. of distilled water to the same amount of potassium chromate solution in a similar flask ; this is used for comparing with the contents of the flask which are being titrated, so that a change of colour of the indicator can be quickly detected. The correction for the amount of silver nitrate solution required to colour the indicator can be made in the same flask. 4 The titration is made over a white porcelain tile, or sheet of white paper. Use a burette reading to ^ths or -g^ths c 2 = d l :d 2 . (1) Hence, if any three of these magnitudes be known, the fourth can be calculated by simple proportion. The comparison of the colours of the solutions is done in glass cylinders called test glasses made from clear, colourless glass. The instrument used for the comparison is called a colorimeter, or tintometer. There are three types of colorimeter used in this work. 39. Duboscq's Dipping Colorimeter. In this instrument 2 the thickness of the liquid in each test glass through which the light passes can be varied until each liquid appears to have the same 1 G. and H. Kriiss, Kolorimetrie und quantitative Spectralanalyse, Hamburg, 1909. 2 H. Morton, Chem. News, 21. 31, 1870. See also W. G. Smeaton, Journ. Amer. Chem. Soc., 28. 1433, 1906 ; 0. Schreiner, ib., 27. 1192, 1905 ; G. Steiger, ib., 30. 215, 1908 ; C. H. Wolff, Pharm. Ztg. t 24. 587, 1879. 82 COLORIMETRY AND TURBIDIMETRY. 83 colour tint. The thickness of the liquid in each tube is measured by a vernier and scale, and since the concentration w 2 of one solution is known, the concen- tration w l of the unknown solution follows at once from (1) above, where w,=^A /ox 1 d 2 Fig. 38 represents a photograph of the front of the apparatus. The operator stands on the right. Fig. 39 represents the path of the light inside the instru- ment. The diffused light of a lamp or of a monochromatic burner is the source of illumination. This light is reflected from the mirror A, and two beams pass FIG. 39. Course followed by light in the colorimeter. FIG. 38. Duboscq's colorimeter. respectively through the tubes B. The beam of the right is reflected twice in the right half of the prism K, and passes through lenses M N of the eyepiece, where it only illuminates the right half of the field. The beam on the left follows a similar path, and finally illuminates the left side of the field. Hence, the two halves of the field P are only illuminated by the beams of light (dotted in the diagram) passing through the two sets of tubes B and D. 1 To illustrate by example : the solution of unknown concentration was made up to 250 c.c. and placed in the left test glass A of Duboscq's colorimeter; the standard solution containing O'Ol grm. of titanic oxide per 100 c.c., that is, 0-025 grm. per 250 c.c. This solution was placed in the right test glass B. 1 To clean the apparatus, raise the tubes D, remove the tubes , and take off the glass at the bottom. The tubes B and D can then be cleaned. Polish the rest of the apparatus with chamois leather. 8 4 A TREATISE ON CHEMICAL ANALYSIS. The amount of liquid placed in the tubes must not be so great that liquid over- flows when the tubes are lowered to the bottom of B. Lower the tubes D until they touch the bottom of the tubes B. The verniers now mark zero, and the two halves of the field in the eyepiece should be uniformly bright (fig. 40a) ; if not, the lamp illuminating the mirror A must be moved until the desired result is attained. Now raise the tube D in the standard liquid until it reaches a point convenient for estimating the tint of the liquid. Then raise D in the other tube until equal tints are obtained in both fields (fig. 40c). Read the two scales. Suppose that for w 2 = 0*025, d 2 = 2*1 cm. ; and ^ = 4-37. Hence, FIG. 40. Appearance in eyepiece of colorimeter. 0-025 x 4-37 AKO ^ = 0-052 grm. per 250 c.c. Fig. 406 represents the appearance of the field of view before the final adjust- ment (fig. 40c). 40. Weller's Colorimeter. Duboscq's colorimeter and similar instruments are expensive, and are only bought when it is probable that they will save in time what they cost in money. A cheaper instrument, quite efficient for most comparisons, can be made as described below. In this type of colorimeter the two test glasses have the same thickness, the light passes horizontally through each, hence c? x = d 2 . Consequently, w^w 2 . . ... (3) Hence, if t0 2 , in the case of titanic oxide, be 0'052 grm. per 250 c.c. of the standard solution, the same number will also represent the concentration w^ of the titanic oxide in the unknown solution per 250 c.c., when the tints of the two solutions are the same. Weller's l and Hillebrand's colorimeters may be cited in illustration. These are easily made. A wooden box C, about 35 cm. long and 12 cm. square, is stained dead black 2 inside and out. The ends are open. A ground-glass partition is placed at B (fig. 41), and a blackened shutter S slides stiffly up and down 3 to 3 '5 cm. behind the glass partition. Two test glasses are selected, of square section, 8 to 12 cm. high, and 3 to 3 '5 cm. side. These glasses are placed between the two partitions when in use. 3 Two methods of conducting the determination are indicated on pages 201 and 205. J. W. Lovibond 4 has devised sets of " standard " coloured glasses ; each set has the same colour, but the glasses are regularly graded in depth of tint. By superimposing glasses from the same and different 1 A. Weller. Ber., 15. 2599, 1882 ; W. F. Hillebrand, Bull. U.S. Geol. Sur., 422. 34, 1910 ; J. W. Mellor, Trans. Eng. Cer. Soc., 8. 123, 1909. 2 DEAD BLACK PAINT. Dissolve shellac in methylated spirit ; add French polish, and add lampblack. If the mixture dries a dead black which rubs off, more French polish or shellac is needed ; if it dries light, more lampblack is needed. 3 Sold by dealers in chemical apparatus. 4 J. W. Lovibond, Journ. Soc. Chem. 2nd., 7. 424, 1888 ; Measurement of Light and Colour Sensations, London ; An Introduction to the Study of Colour Phenomena, London, 1905. . 41. Weller's colori- meter (modified). COLORIMETRY AND TURBIDIMETRY. 85 sets red, blue, yellow composite colours can be obtained. In this way the tint of a liquid placed in a test glass with parallel sides at a definite distance apart can be matched, and the exact proportion of each component of the colour read off. By finding the tints of a series of standard solutions of known strength once for all, the standard tints so obtained render the subsequent preparation of a standard solution unnecessary provided the same method of preparation is employed. The instrument " Lovibond's tintometer " used for the com- parison is similar in principle to the instrument indicated in fig. 41, but Lovibond has introduced several improvements which render the results much more accurate see page 486. 41. Nessler's Tubes. A third method of comparing the tints is commonly used in water analysis. It is rather more tedious than the two methods which precede. The colour is developed in a solution of unknown strength, and in a series of solutions of equal volume but of known strength. The two solutions of known strength which come near to the tint of the test solution are selected, and the colour developed in another series of solutions within narrower limits. This method of trial and failure is repeated until the colour in the test cylinder matches the colour of the standard solution. The comparison tubes should be selected with great care, and the bottoms should be regular. The bores of the test glasses should be uniform and not tapered. As a matter of fact, the principle of the system is the same as that employed in Weller's colorimeter. Although the cylinders are usually marked "25 c.c.," "50 c.c.," "100 c.c.," it is really the lengths of the columns, not the volumes of the solutions, which are compared. Hence, as in Weller's colorimeter, w l = Wy It is obvious that in properly matched tubes the volume marks and the heights of the liquids should coincide. 42. Turbidity Methods of Analysis Turbidimetry. Turbidity methods resemble colorimetric methods, but instead of comparing the different intensities of the colours of two solutions, the degrees of opalescence of two solutions are compared. The opalescence is produced by the fine-grained solid matter in suspension, which settles very very slowly, and which is not usually removed by filtration in the ordinary manner. For example, traces of silver chloride, barium sulphate, and calcium oxalate. The action of light on a turbid solution is twofold. (1) The light is dispersed. This gives the solution its characteristic colour. (2) The light is absorbed, so that an object viewed through the liquid appears with diminished distinctness. This latter is the quality compared. It is well to prevent dispersion in the comparison of opalescent solutions as much as possible, by covering the cylinder with, say, a black cloth, and admitting the light from a circular aperture at the bottom. The subject is discussed in more detail, pages 631 and 653. 43. Some Errors in Colorimetry and Turbidimetry. The most important errors in colorimetry arise from : (1) Imperfect perception of the colour. Colour is a subjective phenomenon due to the action of light on the retina of the eye itself. Possibly no two individuals see exactly the same colour tints in the same substance. A person may be afflicted with a form of colour-blindness and be sensitive to one colour but not to another. Hence, the operator should test himself for any given colorimetric process by matching a standard colour of different intensities against 86 A TREATISE ON CHEMICAL ANALYSIS. itself. If concordant results cannot be obtained, the operator must abandon that particular process. It is here assumed that the operator is normally sensitive to colour impres- sions. If a person of normal colour sensitiveness examines a coloured liquid with, say, the right eye over the cylinder, and then turns the head suddenly so that the left eye is over the cylinder, the colour seems less intense ; in another moment, the right eye may be brought over the cylinder again the colour seems less intense. This shows that first impressions are the stronger, and that the impression imperceptibly weakens in intensity as the eye gets fatigued. The eye quickly becomes fatigued and less sensitive to changes in the intensities of the tints. The eye should therefore be frequently rested by looking on the floor, or into the dark corner of a room. There is also an appreciable difference in the sensitiveness of the two eyes for changes of colour. Hence, many shield or close the least sensitive eye, and use the most sensitive eye for comparing tints. The left eye, for example, is usually more sensitive than the right eye for variations in red tints, and hence the left eye is used for the colorimetric determination of iron. A similar remark applies to the titanium yellow. The eye requires more experience in detecting differences of shade with yellow tints, e.g., in the titanium determination, than with many other colours. 1 (2) Unfavourable light. The colours of the solutions might be compared with the light from a blue sky ; light reflected from grey clouds, or clouds tinged with yellow or orange ; light reflected from coloured objects; artificial light; or light which has been transmitted through a yellow fog. In each case, the effect of the incident light on the colour of the transparent solution is different. The source of the illumination is therefore of great importance in colorimetric work. In general, avoid artificial light and light tinged with yellow in, say, the titanium determination. (3) Differences in the concentration of the test and standard solutions. The colour of many solutions is quite different in concentrated and in dilute solution. The solvent absorbs a certain amount of light ; consequently, the concentration of the standard and test solutions should be the same. In the determination of iron there is a reaction between the coloring agent Fe(CNS) 3 and the solvent water, so that the amounts of iron with different amounts of solvent will not be represented by the relation (1) above. Hence Duboscq's type of colorimeter is not suitable for the colorimetric determination of iron. (4) Changes in the standard solutions. The standard solutions may de- teriorate in strength, and may also be contaminated with impurities derived from, say, the glass bottles in which the solutions are kept, or from the atmosphere of the laboratory when the solutions are temporarily exposed. Such impurities may be fatal to the accuracy of determinations of ammonia, nitrites, and phosphates. 1 Constant errors in the comparison of colours can be eliminated by the method of substitu- tion in which a solution of known strength is substituted in the test glass, after the unknown solution has been compared with the standard. CHAPTER V. FILTRATION AND WASHING. A tremendous amount of time is consumed, often wasted, in the filtration and washing of precipitates for analytical work. Every device which will accelerate these fundamental operations, without interfering with their efficacy, must raise the efficiency of the laboratory. 44. Filter Paper. THE purpose of filtration is to separate a solid from the liquid in which it is suspended. Filtration is usually effected by causing the liquid to pass through a medium which is porous enough to permit the passage of the liquid, and compact enough to retain the solid. Among the different filtering media in use are cotton- wool, sand, charcoal, crushed glass, glass-wool, asbestos pulp, paper pulp, plati- num felt, asbestos paper, 1 and lastly, but most important of all, unsized paper filter paper. Filter paper should give no residue when distilled water or dilute hydro- chloric acid is repeatedly passed through the paper, and the liquid then evapor- ated to dryness; a 10 per cent, aqueous solution of salicylic acid should give no perceptible coloration (showing the absence of iron) ; ammonium sulphide should not blacken the paper ; nor should a dilute soda solution, which has been repeatedly passed through the paper, give a turbidity when neutralised with an acid (showing the absence of oils and fats). 2 Analysts, however, are not much troubled about the quality of their filter paper, because filter papers of remark- ably good quality almost pure cellulose can be easily obtained. Filter paper for analytical work should be kept in air-tight vessels. If the paper be exposed to laboratory fumes, the paper may absorb acid, ammoniacal or other fumes, and disturb certain determinations, 3 e.g., determinations of the hardness of water. Paper which has been kept in an exposed place may subse- quently liberate iodine from a solution of potassium iodide (free from iodate). Andrews too detected nitrites in filter paper which had been kept some time in a laboratory. After filtration and washing, the paper is usually heated in a weighed crucible together with the solid until nothing but the ash of the paper remains, mixed with the calcined precipitate. The ash is weighed with the precipitate, and an allow- ance is made for the additional weight due to the ash of the filter paper. Owing to the fact that the precipitate has sometimes to be further investigated, it is an 1 A. Convert, Chem. Centr. (3), 16. 850, 1885 ; A. Gruner, Zeit. anal. Chem., 9. 68, 1870 ; W. Johnstone, Analyst, 12. 234, 1887. 2 Pharm. Post, 252, 1896 ; L. Fade, Bull. Soc. Chim. (2), 47. 243, 1887 ; W. Wicke, Liebig's Ann., 112. 127, 1859. 3 H. R. Proctor, Journ. Soc. Chem. Ind., 23. 8, 1904 ; M. Mansier, Journ. Pharm. Chim. (6), 16. 60, 1902 ; E. Mallinclcrodt and W. N. Stull, Journ. Amer. Chem. Soc., 26. 1031, 1904 ; L. L. de Koninck, Bull. Soc. Chim. Belg., 23. 221, 1909, 87 88 A TREATISE ON CHEMICAL ANALYSIS. advantage to use filter papers with as little ash as possible. By treating the best filter paper with suitable acids, the manufacturer is able to reduce the amount of ash to a negligibly small quantity. 1 Indeed, the best filter papers behave as if they were almost pure cellulose i.e. almost ashless. Both Nos. and 00, J. H. MunktelPs Swedish papers, are excellent. My experience in quantitative work is confined to these, to C. Schleicher and SchiilPs, and to M. Dreverhoff's filter papers. These manufacturers all supply papers of excellent quality. If not otherwise stated, Munktell's papers are recommended for the routine analyses in this book. The Swedish filter paper No. has been washed in hydrochloric acid 2 to remove traces of iron, alumina, lime, and magnesia ; No. 00 has been washed both in hydrochloric and hydrofluoric acids 3 to remove traces of silica, 4 as well as traces of iron, alumina, etc. Both brands of paper are quite satisfactory in general analytical work. No. 00 is rather expensive. It is suitable for the alumina nitrations where the silica must be kept as low as possible. Neither No. nor No. 00 is suited for calcium oxalate and barium sulphate, which generally pass through the pores of these papers. 5 No. OB has also been treated with hydrochloric acid, and is made rather thicker than No. 0. Consequently, it gives less trouble in the filtration of, say, calcium oxalate. A thin film of calcium oxalate will sometimes be found on the walls of a dried funnel which has been used with Nos. or 00 papers for the filtration of the lime precipitate. No. IF and No. 2 are unwashed Swedish filter papers which are suited for general work where a large number of filter papers of good quality are needed, and expenses must be kept down. No. IF has rather a closer texture than No. 2, and does not filter so rapidly. The following Table XV. shows the amount of ash generally found in the different brands of unwashed paper, expressed in grams per paper : Table XV. Amounts of Ash in Munktell's /Swedish Filter Papers. (Grams of ash per paper.) Brand. 5*5 cm. 7 cm. 9 cm. 11 cm. 12-5 cm. 15 cm. 00 O'OOOOl Q'00002 0-00003 0-00004 0-00006 0-00008 0-00006 0-00010 0-00017 0-00025 0-00033 0-00046 OB 0-00008 0-00014 0-00024 0-00035 0-00047 0-00065 IF 0-00014 0-00023 0-00038 0-00056 0-00073 0-00105 2 0-00018 0-00030 0-00051 0-00074 - 0-00095 0-00138 The amount of ash is so small that with, say, the No. 00 papers, and the smaller sizes of the other papers, the amount of ash can be neglected in routine work. If these papers have been used in a filtration, the amounts of ash will not be those given in the preceding table. The amount of ash furnished on burning a filter depends upon the nature of the liquid which has passed through the paper. A certain amount of fixed alkali, alkaline earth, and some salts are retained by the cellulose of the paper very tenaciously. If a neutral solution be passed through 1 For slight acidity of acid-washed filter paper, M. de la Source, Ann. Chiin. Anal, 2. 82, 2 U. Krensler, Zeit. anal. Chem., 24. 81, 1885 ; Chem. Ztg., 8. 1323 1884 3 P. T. Austin, Chem. News, 37. 149, 1878. 4 Not quite all W. Lange, er., n. 323, 1878. 5 The smaller sizes of papers vary considerably in compactness. Some will retain these precipitates. FILTRATION AND WASHING. 8 9 the paper, the salts may be retained by the paper, as in the case of the fixed alkalies. 1 Acids remove traces of mineral substances from the unwashed papers. Consequently the ash of unwashed papers may be reduced below the numbers given in the above table if they have been used for filtering acid solutions. In illustration, the following results (Table XVI.) were obtained for the ash of No. 00 filter papers which had been treated respectively with dilute sulphuric acid, and potassium hydroxide, and subsequently washed until the wash-water gave no reaction with neutral litmus solution. Ten of the 11-cm. papers were then incinerated, and the results divided by 10. The other sizes were obtained from these results by calculation. Table XVI. Relation between the Filter Paper Ash and Liquid filtered. (Grams of ash per paper.) 5*5 cm. 7 cm. 9 cm. 11 cm. 12*5 cm. Acid Alkali . 0-000011 0-000044 0-000019 0-000071 0-000031 0-000118 0-000047 0-000176 0*000061 0-000219 The inference is obvious. It is not sufficient to adopt the weight of the ash of the unused papers for a precipitate which has been isolated by the nitration of solu- tions containing salts which may be adsorbed by the filter paper. In exact work, the amount of ash in papers which have been treated with solutions resembling those used in the given filtration must be determined for a given precipitate. How- ever, the ash of these papers in neither case need be considered in most analyses. 2 C. Schleicher and Schiill's Nos. 589 and 590 correspond approximately with Munktell's Nos. and 00 respectively. No. 589,, " white ribbon," is an excellent paper with an ash but slightly greater than Munktell's No. 00, and hence negligible. It filters rapidly. No. 589, "black ribbon," has a more open texture than the corresponding "white ribbon," and filters more rapidly, but it is hope- less with finely divided precipitates like calcium oxalate and barium sulphate. The No. 589, "blue ribbon," is closer in texture than the corresponding "white ribbon." It filters more slowly, but will retain the finely divided precipitates better. It is useful for filtrations accelerated by gentle suction. 3 To summarise, the three desiderata in a filter paper are : (1) porosity, to ensure rapid filtration ; (2) sufficient compactness, to ensure complete retention of the solid ; and (3) low amount of ash. The first and second qualities are antagonistic. It is important to adapt the size of the filter paper to suit the precipitate to be washed. Use " rapid " (porous) papers for all precipitates ivhich do not readily pass through the paper, and use the slow, compact papers only when absolutely necessary.* 1 M. Mansier, Journ. Pharm. Chim. (6), 16. 60, 1902. 2 L. L. de Koninck (Lehrbucli der qualitaiiven und quantitativen chemischen Analyse, Berlin, i. 39, 1904) gives the following analysis, in round numbers, of the ash of Munktell's filter papers: Si0 2 , 32; A1 2 3 , 15; Fe 2 3 , 9; MnO, 8; MgO, 11 ; CaO, 21 ; CuO, trace; alkalies, 2 ; S0 3 , 2 ; P 2 5 , : 2 ; Cl, trace. 3 C. Schleicher and Schull's No. 551 is black in colour, and used for filtering light-coloured sediments, minute traces of which can be easily detected on the black background. T. Fleitmann, Zeit. anal. Chem., 14. 77, 1875 ; W. Hempel, ib., 14. 308, 1875. 4 A. Gawalovski, Zeit. anal. Chem., 16. 59, 1877; 18. 246, 1879; 37. 377, 1898; R. Fresenius and Caspar i, ib., 22. 241, 1883 ; K. Kraut, ib., 18. 543, 1879 ; F. Mohr, ib., 12. 148, 1873 ; A. von Wich, Viertel. prakt. Pharm., 8, 187, 3859; J. J. Berzelius, Lehrbuch der Chemie, Dresden, 10. 260, 1835 ; H. Uelsmann, Dingier 's Journ., 220. 534, 1876. A TREATISE ON CHEMICAL ANALYSIS. 45- Filtration. The art of quantitative analysis is largely dependent upon the skill exercised in transferring substances from one vessel to another without gain or loss. Selecting the Funnel. Do not use ribbed or guttered funnels for quantitative work. These often give trouble in washing. Finkener's funnel, 1 fig. 42, has a long capillary stem, 2 and it is supposed to have smooth walls inclined at an angle of 60, so that when the paper is folded four- ply and opened out in the regular manner, it will fit close to the walls of the funnel. 3 The folding of papers with creases 4 to expose a large surface and prevent the paper lying close to the wall of the funnel is not recommended, 5 because (1) precipitate and paper are more difficult to wash ; (2) there is a greater risk of breaking the paper ; and (3) filtration is not so. rapid as when the stem of the funnel is kept filled with liquid. Fitting the Filter Paper in the Funnel. When the paper is placed in the funnel it should not come nearer than half a centimetre to the top edge of the funnel, and on no account should the paper project beyond the funnel. The funnels best suited for the filter papers of different size are indicated below : FIG. 42. Long-stemmed funnels. Filter paper . Funnel . 11 15 18 cm. 9 10 -cm. Place the paper in the funnel, wet the paper, and carefully bed it against the walls of the funnel. When the paper is filled with water, the stem of the funnel should fill with a column of water, 6 and air should not pass between the walls of the funnel and the paper as the paper empties. When the filter paper is properly bedded, water will flow through the paper quickly, 7 and filtra- tion will usually proceed quite rapidly ; at any rate, the paper is doing its best under the given conditions. The liquid in the filter paper is under a pressure 1 A. Gwiggnes, Chem. Ztg. t 27. 889, 1903. 2 F. Weil, Zeit. anal. Chem. 2. 359, 1863 ; I. B. Cook, Chem. News, 27. 261, 1873 ; 29. 81, 1874 ; J. F Kerr, ib., 29. 71, 1874 ; J. Picc&rd (Zeit. anal. Chem., 4. 45, 1865 ; G. Lunge, Chem. News, 13. 23, 1866) proposed joining a narrow tube about 32 cm. long to an ordinary funnel by means of a piece of rubber tubing, as shown in fig. 42. Piccard's knot, as the tube is called, is much used in qualitative analysis, but is less suited for quantitative work. I have given the different types of funnel on the market a three months' trial in my laboratory, and found those indicated in fig. 42A best for quantitative work. "Funnel hangers" for supporting funnels on the sides of beakers are useful for short-stemmed funnels, but not for funnels with a long stem. 3 If the slope of the funnel be not quite right, it may be necessary to alter the crease of the filter paper a little, so that the paper lies close to the walls of the funnel when opened out. If a very bad funnel is found in a batch, it may give more trouble than it is worth. 4 C. E. Avery, Chem. News, 17. 294, 1868. 5 P. Hart (Chem. Ztg., 32. 1228, 1908) recommends the folded papers for fine sand and colloidal precipitates. 6 The filling of the stem of the funnel is facilitated by keeping the stem free from grease by frequent treatment with the cleaning mixture of page 36 ; and also by making two constrictions in the stem. P. E. Raaschou, Zeit. anal. Chem., 49. 759, 1910. 7 If the water does not flow through sufficiently rapidly, discard the paper, and use another. FILTRATION AND WASHING. 91 equal to the weight of a column of liquid of the same diameter as the bore of the stem, and a height h, fig. 42, where a represents the level of liquid in the filter paper. Hence, in filtering, keep the paper filled to its greatest capacity in order to get a maximum value for h, which, in turn, corresponds with maximum velocity of filtration. If the bore of the tube be too wide, liquid will run down the sides of the tube and no pressure will be produced ; while if the tube be too narrow, surface tension may prevent the flow of the liquid even when the funnel is full. A sharp tap on the filter stand will often start the flow. 1 Filter Funnel Stands. The funnel should rest in a suitable stand with its stem vertical. Filter stands are generally made of hard wood mahogany or teak. Stands are also made of metal and porcelain, with a metal rod, etc. The bevel of the opening through which the funnel passes should be the same as the funnel. For general work, the arm holding the funnel or funnels should be fitted with a screw so arranged that the funnels can be fixed at any desired height. In routine work, where a great number of filtrates may be in progress, with funnels all the same size, a larger stand with or without an arrangement for adjusting the height can be used with advantage. 2 Inexperienced students are inclined to adapt the size of the filter paper to the bulk of the liquid to be filtered. In quantitative work, the size of the filter paper must be determined by the magnitude of the precipitate, not by the bulk of the liquid to be filtered. If too large a paper be selected, time is wasted in washing the paper, and it has just been pointed out that paper as well as precipitate has the property of retaining certain salts very tenaciously. Transfer of Precipitate to Filter Paper. The precipitate is usually allowed to settle before filtration. 3 The clear liquid should not be poured directly on to the funnel, but down a glass rod, where the stream is being directed towards the side, not the centre of the filter paper. 4 The receiving vessel for the filtrate should be so placed that the liquid running from the funnel does not fall into the centre, but down the side of the beaker so as to avoid any danger of loss by splashing. If desired, the receiving vessel for the filtrate can be covered by a glass plate with a hole at the side so as not to interfere with the position of the stem of the funnel. The funnel can also be covered with a clock-glass while the filtration is in progress. The object is to keep out dust. If it be noticed while a filtration is in progress that too small a paper has been selected, so that the precipitate is likely to fill the paper more than half full, it is better to use another paper, and distribute the precipitate between the two papers. Ample room should be left for washing the precipitate. If the precipitate occupies more than two-thirds of the paper, difficulties will be encountered in the subsequent washing. Policeman. When the liquid has been transferred to the paper, it will generally be found that small portions of the precipitate remain adhering to the 1 Any dirt in the bore of these tubes is easily removed with a " tube cleaner" or a tobacco- pipe cleaner. E. Bauer (Chem. Ztg., 12. 789. 1888) recommends a funnel with no stem at all for precipitates which filter with difficulty ! His idea is to let the funnel dip in a vessel of water. By renewing the water occasionally, the precipitate is washed by diffusion. There is no advantage in this suggestion for general work. See also P. Blackman, Chem. News, -104. 30, 211, 312, 1911. 2 F. Julian, Journ. Anal. App. Chem., 3. 41, 1889; C. Mueneke, Zeit. anal. Chem., 16. 228, 1877 ; G. H. Bostock, Chem. News, 57. 213, 1888 ; C. Simon, Chem. Ztg., 9. 1870, 1885 ; A. A. Besson, ib., 35. 408, 1911 ; P. Blackman, Chem. News, 104. 30, 211, 1911. 3 (1) Fine precipitates are then not so likely to run through the filter paper ; and (2) the time occupied in standing ensures more complete precipitation with precipitates which form slowly, e.g. , magnesium ammonium phosphate, ammonium phosphomolybdate, potassium platinichloride, potassium cobaltinitrite. 4 Do not stir up the liquid when the rod is returned to the liquid. 92 A TREATISE ON CHEMICAL ANALYSIS. vessel. 1 These can usually be loosened by means of & policeman, which is made by covering the end of a glass rod with a tight-fitting piece of rubber tube. Pieces of rubber tube with solid ends are sold for the purpose, or they can be made from a piece of rubber tubing 3-4 cm. long by placing a little solution of caoutchouc in chloroform (or naphtha) within the rubber tubing at one end, and pressing the sides together between the jaws of a clamp 2 for a couple of days. The sealed end is then trimmed with a pair of scissors, and the open end is slipped on to a piece of glass rod, a, fig. 44, which fits the rubber tightly. 3 Specially rounded rubber cones, screwed to the end of ebonite rods, are sold for the purpose (6, fig. 44). The form fig. 44a is the better. A rubber finger-stall is usually more effective and quicker than the " policeman " in removing adhering precipitates from beakers particularly if shallow beakers be in use. Care should be taken to select the Erlenmeyer's flasks used for precipitations with a bottom easily accessible to the "policeman." Washing the Precipitate. The "policeman " is washed, and then the particles loosened by the "policeman" are washed from the beaker or flask into the filter F IG> 43. Clamp for making "policeman." FIG. 44. " Policemen. paper by a jet of water blown from a wash-bottle. The nozzle of the wash- bottle is directed round and round the inside of the beaker. The final washing of the precipitate is effected by means of a jet of, say, water from a wash-bottle. The stream is first directed round the upper edge of the filter paper so as to wash downwards towards the apex of the cone ; and the precipitate is collected as nearly as possible at the apex of the cone at the last washing. The paper is never filled with water, and the precipitate is nearly always allowed to drain completely before adding fresh wash-ivater. The washing is continued until a few drops of the filtrate collected in a test tube show the absence of the salts which are being washed from the precipitate. 4 The filtrate is usually required for the determina- tion of another constituent, and, in consequence, as little of the filtrate as convenient must be employed in making the test, otherwise some of the filtrate will be lost. Six washings may suffice, but there is no fixed rule. Some precipitates retain the mother liquid more tenaciously than others, partly attracted by some kind of surface action, and partly entangled mechanically in the solid. 1 If the particles cannot be scraped off, dissolve them in a suitable solvent and reprecipitate in a small beaker. In some cases, the particles need not be loosened, but left in the beaker. This is usually done when the precipitate is to be dissolved and reprecipitated, as is the case with the ' ' alumina " precipitate in clay analyses. The beaker and the particles are washed as well as possible, and the solution of the precipitate on the filter paper is allowed to return into the beaker where the first precipitation was made. 2 A clamp for the purpose is easily made from two pieces of wood, say, 30 cm. long, and 5 cm. by 2 cm. cross section. Holes are bored through each end to fit a pair of "sash screws," fixed as shown in fig. 43. The rubber tips are clamped by turning the thumb-screws. 3 A. A. Blair, The Chemical Analysis of Iron, Philadelphia, 31, 1908. 4 In some cases, until a few drops evaporated on a piece of clean platinum foil give no residue. FILTRATION AND WASHING. 93 It may seem that undue stress is laid upon manipulation. The great accuracy and skill with which a competent analyst does his work can only be acquired by a repetition of the same operations a great number of times. The student will probably notice that, as he gets more and more expert, a less volume of wash- water will be needed to wash a precipitate (page 97). There are some troublesome precipitates which at best require inconveniently large volumes of wash-water. When another constituent is to be separated from the nitrate, time is lost by the necessary evaporation, and the risk of loss during transfer from vessel to vessel is increased. In such cases it is best to change the receiver as soon as the liquid is filtered and washing commences. Collect the washings in an evaporating basin, and concentrate the washings by evaporation to a convenient volume. Mix the concentrated washings with the first filtrate. It is a good plan to regularly change the receiver as soon as washing commences, so that if the filtrate begins to get turbid towards the end of a washing, there will be no need to refilter all the liquid. Experience will soon teach what filtrates and washings may be safely collected in one receiving vessel. 46. Wash-Bottles. A great many types of ' wash-bottle have been suggested. are generally best. Fig. 45 shows the forms most useful. The simpler types A represents the FIG. 45. Wash-bottles. ordinary cold-water bottle ; B, the hot- water bottle. Both are fitted with rubber stoppers l and glass tubes. The neck of the hot- water bottle is wrapped round with thick string, flannel, cork, or some suitable non-conductor of heat. 2 1 There is a danger of contaminating the water with sulphuric acid derived from the sulphides antimony used in vulcanising the rubber. J. Pattinson and J. T. Dunn, Journ. Soc. Chem. Ind., 24. 16, 1905. 2 Asbestos pape*r is objectionable, because it is liable to flake off, and particles may thus get into the filter paper or filtrate. 94 A TREATISE ON CHEMICAL ANALYSIS. The jets shown in the diagram are joined with pressure tubing. The jets can then be easily turned in any direction. 1 A small Ostwald's heater, 2 shown at B, fig. 45, is useful for keeping the contents of the bottle warm. The stopper should rest on the neck while the bottle is being heated, in order to prevent steam escaping from the mouthpiece, and risking scalded lips or tongue. If the arrangement indicated below is used, there is no need for removing the stopper, etc., while the wash-bottle is being heated. It requires a little practice to regulate the k ' breath " so as to ensure hot steam is not sucked into the mouth. The student ought to learn to breathe through the nose while blowing from the mouth. The same action is used in working with the mouth blowpipe. Bottles for cold liquids may be made from thick glass, 3 but the hot-water bottles must not be too thick, or they may be fractured on heating. The " R " or the Jena glass type of flask is generally best. FIG. 46. Valves for wash-bottle tubes. Some prefer Erlenmeyer's flasks for wash-bottles. The wash-bottle C, with stoppered tubes, is used for alcohol and other volatile liquids, or for liquids which spoil on exposure to the atmosphere. 4 Hot-water bottles, and wash-bottles containing ammonia water, ammonium carbonate, hydrogen, or ammonium sulphide solutions, and other unpleasant liquids, are best fitted with a special attachment, as shown in fig. 46. The part A is closed by the finger or thumb while blowing. The excess pressure closes the valve at V. The jet of liquid is stopped by withdrawing the thumb from A. The valve may be Griffin's (fig. 46, U), Bunsen's (fig. 46, V), Waters' (fig. 46, Tf), etc. 5 With Griffin's valve, a two-holed stopper suffices, since closing a, fig. 46, serves the same function as closing A in Bunsen's or Waters' valves. 1 G. Foord, C/iem. News, 30. 191, 1874 ; A. R. Leeds, ib., 20. 236, 1870; W. J. Land, Amer. Chemist, 3. 221, 1874 ; A. Gawalovski, Zeit. anal. Chem., 14. 170, 1875. 2 W. Ostwald, Zeit. anal. Chem., 31. 180, 1890 ; F. Muck, ib.. 28. 611, 1889 ; J. Volhard, Liebig's Ann., 285. 330, 1895. 3 W. Dittmar (Chem. Ztg., 15 1521, 1891) recommended nickel and copper bottles. 4 For compressed-air wash-bottles, see W. C. Ferguson, Journ. Amer. Chem. Soc., 16. 149, 1894. 5 For Bunsen's valve and some modifications, see page 188. J. J. Griffin, Brit. Pat. No. 464801, 1905 ; M. Stuhl, Chem. Ztg., 21. 396, 1897 ; E. Stroschein, ib., 13. 464, 1889 ; C. E. Waters, Journ. Amer. Chem. Soc., 27. 298, 1905 ; R. K. Meade, ib., 19. 581, 1897; F. M. Haldeman, Journ. Anal. App. Chem., 2. 301, 1888 ; I,. M. Dennis, Amer. Chem. Journ , II. 218, 1889; E. Jacob, Zeit. anal. Chem., 5. 168, 1866; T. Bayley, ib.,-iS. 295, 1879: E. Borgmann, ib., 22. 60, 1883 (stoppered tube on A) ; H. Dubovitz, Chem. News, 91. 147 1905. FILTRATION AND WASHING. 95 A Beutell's valve is easily made for this purpose, and it works better than Bunsen's. A piece of glass tubing 15 cm. long and 4-5 mm. bore is slightly contracted about 3 mm. from one end. Another piece of thin glass tubing, which fits loosely into the larger tube, with a free space about 0'5 mm. all round, is sealed and rounded off at one end. If the end be too tapered, it will be inclined to wedge tightly (fig. 46, W). The closed end is then fitted to the contracted part by grinding with emery and water, all the while rotating the tubes in one direction. If the joint appears tight after washing off the emery and sucking at A, fig. 47, dry the tube. Push the closed end of the narrow tube into a cork, and seal it off as near the closed end as possible, say 1 cm. away. The cork serves as a support for the closed end. The end must not be sealed too symmetrically, or it will be difficult, later on, to blow into the bottle. Bend the tube in the usual manner for the mouthpiece of a wash-bottle, as at W, FIG. 47. Making a valve for a wash-bottle. fig. 46. With these instructions it will be easy to place the valve nearer the mouthpiece, and make it serve the same function as Griffin's valve. Washing a Large Number of Precipitates. When a number of precipitates have to be washed, the wash-bottle can be almost discarded, and a water supply "bottle and syphon" placed high above the working bench. The syphon tube is connected with a glass jet by a piece of rubber tubing sufficiently long to reach all the precipitates to be washed, and with a suitable tap for arresting or regulating the flow of the liquid used for washing the precipitates. If water be the washing fluid, arrangements can easily be devised for heating the reservoir. This arrangement is much more convenient than wash-bottles. 1 47. The Theory of Washing Precipitates. Under ideal conditions the reagent employed for precipitations the precipi- tant should form a precipitate of definite composition, which is very sparingly soluble, or rather insoluble, in the mother liquid and washing fluids. Any excess of the precipitating agent should be readily removed from the precipitate by washing (or by ignition). The precipitating agent should not introduce any substance likely to interfere with the precipitation of other constituents from the filtrate later on. Again, the precipitate should be compact, easily filtered and washed. Few processes satisfy all these desiderata, and an important part of analytical chemistry is to know what conditions favour and what conditions hinder the separation and purification of a given precipitate. There are, however, a few general principles of such wide applicability that they should be carefully studied. Colloidal and Fine-grained- Precipitates. In general, the finer the grain of the precipitate, the greater the quantity of contaminating salts retained by the wet precipitate. Fine-grained precipitates expose a large surface of separation between the solid and solution. The salts appear to be retained by a kind of surface attraction which is called, for convenience, adsorption. Hence, fine- grained precipitates are more difficult to wash clean than coarse-grained precipi- tates. 2 Colloidal gelatinous precipitates like ferric and aluminium hydroxides * G. E. Boltz, Journ. Amer. Ckem. Soc., 33-514, 1911. 2 The term " coarse-grained precipitate" of course does not mean aggregates formed by the clotting of a number of fine grains. 96 A TREATISE ON CHEMICAL ANALYSIS. are in an extremely fine state of subdivision, and, in consequence, they are most difficult to wash clean. Such precipitates may require ten to twenty washings, and then not be so clean as a coarse-grained crystalline precipitate after two or three washings. Again, fine-grained precipitates, like newly precipitated barium sulphate, lead sulphite, calcium oxalate, and silver chloride, are particularly liable to pass through the filter paper, while coarse-grained precipitates give no trouble. Hence, the analyst employs various artifices in order to coagulate or to crystallise gelatinous and amorphous precipitates. Among the more important means of effecting these changes are : (1) The grain size can be frequently increased by allowing the precipitate to stand in the mother liquid for some time. This, for instance, is the case with lead chromate, antimony oxychloride, manganese ammonium phosphate, sodium antimoniate, stannic hydroxide, calcium oxalate, silver chloride, metallic sulphides, barium sulphate, etc. 1 (2) Precipitates produced in hot solutions are often coarser-grained than precipitates produced in cold solutions. The boiling of precipitates in a fine state of subdivision may lead to the flocculation of a large number of the fine particles into a relatively small number of coarse grains. Zinc su-lphide, barium sulphate, and manganese ammonium phosphate may be cited in illustration. (3) The flocculation of a precipitate which separates in a colloidal condition is frequently effected by the salts present in the mother liquid. When these salts have been almost removed, during the later stages of the washing, the coagulated precipitate sometimes passes to a colloidal or gelatinous condition, and it may then give a turbid filtrate, or become so slimy as to be almost impermeable to the washing liquid. In such cases it is necessary to w T ash the precipitate, not with pure water, but with a solution of an acid or salt which will prevent the deflocculation of the precipitate, and which can be easily removed by drying or ignition. Acids can only be used with precipitates of an acid nature, e.g., washing titanium hydroxide with acetic acid (page 208) ; and with precipitates insoluble in even strong acids, e.g., silver chloride, which is washed with dilute nitric acid (page 652). Usually, we have to depend upon the ammonium salts nitrate, sulphate, chloride, and acetate. 2 For instance, washing the "alumina" precipitate with ammonium nitrate (page 183). The ammonium salt is volatilised when the precipitate is calcined. 3 Adsorption of Salts by Precipitates. On account of the adsorption of a certain amount of salt with colloidal precipitates, such as occurs, for instance, in the ammonia precipitate, the adsorbed salts cannot be all removed by washing, and there are many reasons for supposing that all precipitates carry down with them, that is, adsorb, substances from the solution in which they are formed. The adsorbed salts cannot always be removed by washing. The wash-water may show no indication of the impurities so retained by the precipitate. Thus, 1 G. Watson, Glum. News, 63. 109, 1891. 2 R. Bunsen, LieUg's Ann., 106. 13, 1858. Salts of polyvalent metals and alkaline earths give the best results, but these, with the exception of mercury, are excluded because they would remain with the precipitate after ignition. Mercury salts can only be used in a limited number of cases on account of secondary reactions. Ammonium salts are perhaps least effective, but they are usually the best we can do. See R. G. Smith, Journ. Soc. Chem. IncL, 16. 872, 1897 ; N. Pappada, Zeit. Chem. Ind. Kolloide, 9. 233, 1911. 3 See page 363. Sometimes precipitates which are difficult to filter clear can be satisfactorily filtered if a little recently ignited kieselguhr ( J. P. Ogilvie, Journ. Soc. Chem. Ind., 30. 62, 1911), or china clay (F. Watts and H. A. Tempany, ib., 27. 53, 1908), be added to the mixture e.g., for lead sulphite. The device is particularly useful when the filtrate is alone wanted, or when the precipitate is to be afterwards dissolved in a solvent which does not attack the clay or silica. FILTRATION AND WASHING. 97 Warington l found that ferric hydroxide precipitated by potassium carbonate only lost t\vo : thirds of this salt by a washing which would have removed 0'0069th part of^a salt not adsorbed by the precipitate. The same observer records the adsorption of various salts by aluminium and ferric hydroxides. It is therefore frequently advisable to dissolve the precipitate in a suitable solvent, and reprecipitate. The objectionable impurity divides itself in a definite ratio between the precipitate and mother liquid. A relatively large amount may be retained by the precipitate during the first precipitation, but on a second precipitation, when only that amount of salt which was retained by the first precipitate is in solution, the partition of the undesirable salt between the precipitate and the solution in the given ratio means that a relatively small amount of impurity will be retained by the second precipitate. Two, possibly three, precipitations will generally remove appreciable amounts of the objec- tionable impurity from the precipitate. For instance, aluminium and ferric hydroxides, manganese ammonium phosphate (page 374), zinc sulphide (page 364), nickel and cobalt sulphides (page 388), etc. Suppose that the first precipitate retains O'Olth part of the objectionable salts, while O99th part is removed by filtration. The second precipitate will contain O'Olth of O'Olth of the salt; the third precipitate, OOOOOOlth of the salt. Thus, repeated reprecipitation will soon carry the amount of impurity outside the range of the balance. 2 These facts also lead us to conclude that, if a small quantity of a substance A is to be separated from a large quantity of a substance 13, it is generally better to precipitate A rather than precipitate B. The loss of A through absorption by B makes the " error of experiment " greater than if some of B be lost through absorption by A. Amount of Fluid required for Washing Precipitates. A rather important question has to be decided in washing precipitates. Is a relatively small number of washings with large volumes of liquid more effective than a relatively large number of washings with small volumes of liquid ? Suppose that a precipitate be allowed to drain on a filter paper, and that the precipitate exercises no physical or chemical action on the salts in solution in the mother liquid. 3 Let the volume of the mother liquid retained by the moist precipitate be represented by v c.c. Add V c.c. of water to wash' the precipitate. The total volume of liquid will be (v + V) c.c. Let this drain on the filter paper. The moist precipitate retains v c.c. of the v+ Fc.c Otherwise expressed, the precipitate on the first draining retains ~ - c.c. of the mother liquid. After the first washing, again add V c.c. of water and let the precipitate drain. The moist precipitate retains, on the second draining, a volume of the - c.c. of the mother liquid. Hence, the precipitate retains : First draining [ - ) c.c. of the mother liquid ; Second draining ( j c.c. of the mother liquid ; nth draining ( ~^ \ c.c. of the mother liquid. 1 R. Warington, Journ. Chem. Soc., 21. 1, 1868 ; P. Jannasch and T. W. Richards, Journ. prakt. Chem. (2), 39. 321, 1889 ; Ed. Schneider, Zeit. anal. Chem., IO. 425, 1882 ; T. W. Richards, ^.,46. 189, 1903 ; Proc. Amer. Acad., 35. 377, 1900 ; K. Scheringa, Pharm. Weekblad, 48. 674, 1911. 2 In the case of some "rare earth " separations, where the ratio is relatively large, a great number of precipitations fractional precipitations may be required to effect the separation. W. Crookes, Chem. News, 54. 131, 1886 ; K. Scheringa, Pharm. WeekUad, 48. 674, 1911. 3 The assumption is rarely justified, but this does not affect the principle under discussion. 7 98 A TREATISE ON CHEMICAL ANALYSIS. In order to understand what these expressions symbolise, it is advisable to take a numerical example. Suppose that a precipitate, which retains each draining, v=l c.c. of the mother liquid, is to be washed until it retains but O'OOOOOl c.c. of the mother liquid. In that case, after the nth washing, we have n =0-000001. If n = 6, V is nearly 6 c.c., and if n = 4, V is nearly 18 c.c. This means that the precipitate will be washed as clean with six washings (n = 6), using 6 c.c. of water (F= 6) each time, as with four washings (n = 4), using 18 c.c. ( V= 18) each time. In the former case, 36 c.c. of wash-water pass through the filter paper, and in the latter case, 72 c.c. of water pass through the filter paper. Hence, in this particular case, half as much water is needed to wash the precipitate six times with 6 c.c. of water each time as is needed to wash the precipitate four times with 18 c.c. of water each time. Hence, the washing of a precipitate is more efficiently performed by the frequent application of a small volume of water than by using a relatively large volume of water applied a small number of times. It is here assumed that the precipitate is allowed to drain before it is refilled with washing liquid. 1 This theory of washing precipitates was developed by Bunsen, in 1868. 2 Horsley deduces from his experiments : 1. To keep down the volume of the wash-water, keep the quantity of liquid on the filter paper small throughout. 2. The time required for filtration cannot be quickened or delayed by any change in the method of adding the wash liquid, provided the upper edge of the filter paper be attended to. 3. To minimise the drudgery of filtration, make each addition as large as possible, and allow the precipitate to drain ; but the time of washing will not be affected. 48. Bunsen's System of Accelerated Filtration. The long-stemmed funnels I have just recommended give very good results in routine work, but a considerable amount of time can be saved by using suction, as recommended by Bunsen. 3 The Swedish filter papers recommended above will not bear much suction. They will burst near the tip. In consequence, a small perforated platinum filter cone, fig. 48, is placed in the funnel to help the paper to bear the pressure. 4 The filter cone is first placed in the funnel, and the paper is bedded on to cone and funnel. FIG. 48. Filter ^ the paper be not properly bedded to both cone and funnel, cone. it will inevitably be torn as soon as strong suction sets in. The funnel is fixed in the neck of a stout-walled filtration flask, F, figs. 49 and 56, by means of a rubber stopper. The filtration flask is 1 In washing by decantation with boiling water, T. W. Das (Chem. News, 101. 169, 1910) re- commends trapsferring the precipitate, after each decantation, to the filter, washing out the precipitate into a beaker, adding fresh water, and repeat. This ensures the better removal of the mother liquid, and more rapid washing. 2 R. Bunsen, Liebig's Ann., 148. 269, 1868 ; G. F. Horsley, Chem. News, 87. 237, 1903. 3 R. Bunsen, Liebigs Ann., 148. 269, 1868 ; Zeit. anal. Chem., 8. 174, 1869 ; E. C. Hildebrand, Chem. News, 34. 57, 1876 ; C.Jones, Journ. Anal. App. Chem., I. 383, 1887 ; V. Kreusler, Chem. Ztg., 8. 1324, 1885. For filtration under pressure, T. Feller, Zeit. anal. Chem., 3. 325, 1864. 4 A. Gawalovski, Zeit. anal. Chem., 23. 372, 1884. R. S. Dale, Chem. News, 2O. 128, 1869, for copper cone. See, and for funnels with a perforated inner wall, C. Nickles, Journ. Soc. Chem. Ind., 6. 134, 1887 ; J. de Mollins, Zeit. anal. Chem., 19. 334, 1880. FILTRATION AND WASHING. 99 FIG. 49. Filtration with platinum cone and Witt's flask. FIG. 50. Zopfchen's filtering tube. IOO A TREATISE ON CHEMICAL ANALYSIS. connected with a WoulfFs bottle or similar flask W, and this in turn with the suction pump. The bottle between the filtration flask and the pump prevents a back-rush of water into the filtration flask when the pump is stopped. The con- nections are made with stout-walled rubber tubing which will not collapse under the diminished pressure. If the filtrate is to be used afterwards, I prefer Witt's filtration jar, F, fig. 49. l In this, the filtrate is collected in a beaker as indicated in the diagram. Zopfchen's 2 filtration tube, fig. 50, is useful when small quantities are under investigation, and in other special cases. Instead of a platinum cone, filter papers with toughened tips are useful when the pressure is not too great. 3 These are made by immersing the tip of the folded paper in nitric acid (sp. gr. 1-42) for a few moments, and then thoroughly washing the paper with water. These papers are C. Schleicher and Schiill's No. 580. The same firm makes toughened .filter papers, No. 575. The small sizes (4 or 5'5 cm.) can be used in place of the platinum cone. The toughened paper cone can be pierced in a number of places near the tip, with a fine needle. 4 In filtering and washing precipitates under diminished pressure, as a rule, do not allow the precipitate to drain, but add more liquid before the former liquid has run through. If the precipitate be allowed to drain, as recommended for ordinary filtration, channels or fissures sometimes form in the precipitate. The wash-water then simply runs through the channels without coming in contact with the bulk of the precipitate. 49. Tared Filter Papers. Some precipitates, when collected on a filter paper, cannot be ignited with the paper on account of volatilisation, decomposition, etc. These cases require special treatment. In some cases the filter paper is separated from the precipitate and ignited separately. In other cases the filter paper is dried at, say, 100 or 120 and weighed. 5 The precipitate is then collected and washed on the filter paper in the usual manner. The paper and contents are then partially dried in the funnel (so that the paper and contents can be removed without tearing the paper), placed in the weigh- ing bottle, and later dried at the desired temperature, say, 110 or 120". The paper is very hygroscopic, and it should be protected from the atmosphere while being weighed. Consequently, the paper may be folded and FIG 51 -^Reinhardt's P laced between a pair of weighed watch-glasses, e, fig. 3 ; in weighing bottle. weighing tubes, /, fig. 3 ; or in Reinhardt's weighing bottle, 6 fig. 51. Some papers, after use, are liable to disintegrate and break if folded; in that case, either Reinhardt's bottle or Koninck's weighing funnel, B, fig. 52, can be employed. The latter is made specially thin, and it fits on to a ground tube G in the stopper of the filtration flask F. After the funnel has been used for filtering, it is placed on the tripod A, which acts as a ground stopper. The funnel and contents are then dried, the ground cover C placed on the funnel, and the whole cooled and weighed. Instead of the funnel, 1 0. N. Witt, Chem. 2nd., 510, 1899; A. R. Leeds, Chem. News, 23. 177, 1871; 21. 236, 1870 ; A. Burgemeister, Zeit. anal. Chem., 28. 676, 1889 ; F. Allihn, ib., 26. 721, 1887. 2 H Zopfchen, Chem. Ztg., 25. 1008, 1901 ; H. S. Bailey, Journ. Amer. Chem. Soc., 31. 1144, 1909. 3 E. E. Francis, Journ. Chem. Soc., 47. 183, 1885 ; J. F. Stoffart, Journ. Anal. App. Chem., 4. 1, 1890 ; C. R. C. Tichborne, Pharm. Journ. (3). 2. 881, 1871. 4 M. H. Cochrane, Chem. News, 32. 80, 1875. 5 C. Gilbert, Rep. anal. Chem., I. 264, 1882 ; J. L. Smith, Chem. Neivs, 31. 55, 1875. 6 C. Reinhardt, Zeit. angew. Chem., 2. 61, 1889 ; L. L. de Koninck, ib., I. 689, 1888. FILTRATION AND WA IOI a filter tube b also can be fitted on to the same jon^ s ^5 fc 45 = ^fie"stopper of the filtration flask F. The filter tube b has a glass cap c and plug a for use during the weighing. If an allowance is to be made for the action of the liquid being filtered on the paper, it will be necessary to treat a second paper with the clear filtrate and find the effect on the weight of the paper after the paper has been washed and dried. The loss or gain in the weight of the empty paper, by the treat- ment, represents the effect of the solution on the filter paper containing the precipitate, and an allowance must be made accordingly. 1 The following is the best plan: Two filter papers are cut the same size, and one is weighed against the other. The difference in weight is marked with a pencil on the FIG. 52. Koninck's filtration apparatus. heavier. Each filter paper is placed on a funnel. The solution to be filtered is poured through the one, and the filtrate is poured through the second paper. Similarly, duplicate the washings. The two filter papers are dried in the ordinary manner. In weighing, the empty paper is placed on the right pan as a tare to the other, and due allowance made for the excess in the weight of the one paper over the other. 2 A certain amount of judgment is required in using the different weighing tubes. For example, the Koninck's apparatus weighs about 50 grms. The precipitate may weigh 0*005 grm. The apparatus is thus 10,000 times heavier than the precipitate ! The large surface exposed by the glass may introduce a small error which is relatively large in comparison with the substance being weighed. In general, the greater the difference between the weight of the pre- cipitate and the weight of, or rather the surface exposed by, the apparatus, the greater the percentage error affecting the determination. It is therefore best to 1 If any solid matter should escape the first paper, and be retained by the second, the work would probably be stultified. This is the objection to placing the second paper below the first in the funnel while the liquid is being filtered, and separating the papers, later on. for weighing. a C. Riidorff, Zeit. angew. Chem. t 3. 633, 1890 ; Chem. News, 66. 25, 1892. IO2 A TREATISE ON CHEMICAL ANALYSIS. keep the weighing apparatus as small as possible. Processes involving the use of tared filter papers are rapidly falling out of use. The weighing of precipitates by means of tared filter papers will soon be obsolete. 50. Filtration Tubes. Instead of using tared filter papers, substances which are injured by burning filter papers may be filtered through cotton- wool, 1 asbestos, 2 felt cloth, 3 glass- wool, 4 gun-cotton, etc. Thus, a tube shaped as indicated in fig. 53, A E. Allihn's tube may be packed first with glass-wool (a), then with, say, asbestos (b). Some precipitates are liable to clog if the asbestos be packed in too thick a layer for instance, manganese oxide precipitated by potassium chlorate. In that case, some prefer a small plug of glass-wool (a), followed by a (V6 to 0*7 cm. layer of calcined Calais sand (6), 5 with or without a thin layer of asbestos over the sand. Several different ways of packing have been suggested, and several different forms of filter tube. 6 For instance, the tube may be stoppered at each end Jannasch's filter tube, fig. 204, page 653 or the tube may be capped at one end and stoppered at the other, as a b c, Koninck's tube, fig. 52. In Fresenius' filter tube there is a con- striction a, fig. 53, B ; and the tube may have a bulb in the region of the packing, fig. 53, (7, which is Gibbs and Taylor's tube. In Mason's filter tube, D, fig. 53, the stem is separate from the body of the tube, so that, when in position, it forms a ledge .on which a /perforated "filter plate" can rest. The stem is rather longer than the body of the tube, so that it can be used to eject the filter bed, etc. In the analysis of. iron* where the carbon is separated as insoluble matter, the carbon is filtered in a tube of various forms e.g., a platinum tube with a per- forated removable disc which is covered with asbestos before filtration ; a glass tube with a platinum spiral, 7 fig. 53, E, covered with long fibre and ignited asbestos ; etc. The carbonaceous matter is afterwards dried and burnt in a current of oxygen, or by the wet combustion process. Several other modifica- tions are in use. 8 Thus, the carbon may be collected directly in a platinum boat fitted with a suitable bottom, and holder for the filtration flask. FIG. 53. Filtration tubes. 1 A. B. Clemence, Journ. Anal. App. Chem., I. 273, 1887. 2 See Gooch's crucibles, page 104. 3 R. Friihling and J. Schuz, Zeit. anal. Chem., 13. 146, 1874. 4 It is well to remember that the glass-wool is often made from lead glass, and may give up some lead to acid solutions. F. Muck, Zeit. anal. Chem., 19. 140, 1880 ; L. Blum, ib., 31. 292, 1892 ; F. Stolba, ib., 17. 79, 1878 ; R. Bbttger, NotizMatt, 34. 3191, 1884 ; M. Battandier, Journ. Pharm. Chem. (4), 30. 55, 1880. 5 C. M. Sargent and J. K. Faust, Journ. Amer. Chem. Soc., 21. 287, 1899 ; Chem. New*, 79. 158, 1899. S. Kern, ib., 38. 157, 1878, platinum tul.es with asbestos packing. P. Weisskopf, Dingier' s Journ., 206. 243, 1872, felted glass-wool packing. 6 R. Fresenius, Zeit. anal. Chem., 8. 154, 1869 ; 0. Lohse and P. Thomaschewski, ib., 39. 158, 1900 ; W. Gibbs and E. R. Taylor, Amer. J. Science (2), 44. 215, 1867 ; F. Allihn. Journ. prakt. Chem. (2), 22. 56, 1880 ; F. Soxhlet, ib. (2), 21, 231, 1880 ; A. Goske, Chem. Ztg., 22. 21, 1898 ; Ber., 32. 2142, 1899 ; J. S. C. Gray, Chem. News, 22. 165, 1895 ; H. P. Mason, ib. 91. 180, 1905 ; T. Macfarlane, Analyst, 18. 73, 1895. 7 T. M. Drown, Tech. Quart., 20. 552, 1891. 8 See Shimer's filter tube, page 620. FILTRATION AND WASHING. 103 51. Filtration through Perforated Discs and Funnels. In 1870, Carniichael 1 proposed a different system of nitration from those which precede. Carmichael's plan was to keep the solution in a platinum dish in which the precipitate is to be finally weighed. A disc of filter paper was placed against the lower perforated surface of the vessel, whose interior was connected with the suction apparatus. The liquid, wash- water, etc., was sucked from the vessel, and the precipitate was prevented from leaving the vessel by the filter paper. Casamajor 2 modified this by using a kind of perforated false bottom with an aspirator, and cover- ing the upper surface of the perforated plate with a disc of filter paper or asbestos pulp. Any liquid poured into the dish could then be filtered by suction through the perforated bottom. Sauer 3 suggested the use of flat discs of platinum gauze ; and Grosjean, perforated platinum discs, in an ordinary FIG. 54. Hirsch's funnel. A piece of filter paper, rather larger than the disc, was funnel, supposed to be placed on the upper surface of the perforated disc. The paper was wetted and pressed close against the disc and funnel, so as to form a watertight joint all round the disc. The chief objection to the use of these discs for quantitative work arises from the fact that only part of the precipitate collects on the disc ; part collects on the funnel, whence it can only be removed with difficulty. In order to keep the discs horizontal, Smith 4 soldered a platinum rod to the under side of the disc. When the rod was dropped into the stem of the funnel, the disc was in a horizontal position. The so-called Witt's filter plates 5 are made of glass or porcelain, rather larger in diameter on the upper side ; otherwise they resemble Grosjean's discs, and Kaehler added the vertical rod as previously suggested by Smith. Kaehler's filtration discs are also made with a groove to take a rubber band between the disc and the funnel, thus ensuring a better joint between plate and funnel. Kaehler's discs are also made of FIG. 55. Biichner's funnel, porous earthenware, asbestos, or alundum (fused alumina) without perforations and of different porosi- ties. 6 In Hirsch's 7 funnels (fig. 54) the disc is an integral part of the funnel. 1 H. Carmichael, Zeit. Chem. (2), 6. 481, 1870 ; Chem. News, 24. 213, 1871 ; W. Jago, ib., 33. 54, 1876 ; G. Ville, ib., 30. 200, 1874 ; J. P. Cooke, Proc. Amer. Acad., 12. 125, 1877 ; W- A. Puckner, Journ. Amer. Chem. Soc., 15. 710, 1893 ; R. Friililing and J. Schulz, Zeit. anal. Chem., 13. 146, 1874 ; L. W. Bahney, Met. Chem. Eng., 10. 737, 1912. 2 P. Casamajor, Amer. Chem., 5. 440, 1875 ; 6. 124, 1876 ; Chem. News, 32. 33, 45, 183, 1875 ; 45. 148, 1882 ; 46. 8, 1882 ; 53. 194, 248, 1886 ; Journ. Amer. Chem. Soc., 3. 125. 1881 ; 8. 17, 1886 ; K. Zulkowsky, Zeit. anal. Chem., 17. 198, 1878 ; 18. 459, 1879 ; 22. 173, 1883. For reversed filtration, see also A. Wildenstein, ib., I. 432, 1862 ; M. C. Lea, Amer. J. Science (2), 42. 379, 1866 ; E. Fleischer, Chem. News. 19. 169, 1869. 3 A. Sauer, Zeit. anal. Chem., 14. 312, 1875 ; B. J. Grosjean, Joum. Chem. Soc., 16. 341, 1879 ; Chem. News, 39. 182, 1879 ; 45. 107, 1882. 4 J. C. Smith, Amer. Chem. Journ., I. 368, 1879; S. L. Penfield and W. M. Bradley, Amer. J. Science (4), 21. 453, 1906 ; Chem. News, 94, 293, 1906. 5 0. N. Witt, Ber., 19. 918, 1886 ; M, Kaehler, Zeit. anal. Chem., 33. 63, 1894 ; Journ. Amer. Chem. Soc., 16. 58, 1894 ; C. H. Piesse, Chem. News, 28. 198, 1873 ; R. Warington, Zeit. anal. Chem., 25. 354, 1886, used lead plates for tartaric acid. A. Borntrager, Ber., 19. 1690, 1886. 6 R. C. Benner and W. H. Ross, Journ. Amer. Chem. Soc., 34. 51, 1912. 7 R. Hirsch, Chem. Ztg., 12. 340, 1888 ; E. Biichner, ib., 12. 1277, 1888 ; 13. 95, 1889 ; E. Murmann, Zeit. anal. Chem., 50. 742, 1911. IO4 A TREATISE ON CHEMICAL ANALYSIS. In the different forms of the so-called Biichner's funnels (fig. 55) the disc may or may not be permanently fixed, but the walls of the funnel above the disc are vertical. The sloping sides of the funnel below the disc narrow rapidly. The disc presents a more extended surface than in Hirsch's funnel. These are all different modifica- tions of the one principle. Some of these funnels are valuable auxiliaries in preparation work. They are used with perforated rubber stoppers, filtering flask, and suction, as illustrated in fig. 56. For other forms see page 630. 52. Gooch's Filtration Crucibles. In 1878, F. A. Gooch 1 proposed to filter certain liquids by suction through a bed of asbestos resting on the perforated bottom of a crucible, Gooch's crucible. The crucible, c, is fitted into a large rubber ring, b, fig. 56, which, in turn, fits tightly into the top of a glass adapter, a, resembling a " thistle- headed funnel." The glass adapter is fitted into a filtration flask. F. Gooch's crucibles are used so much in analytical work that their preparation and use must be described in some detail. 1. Preparation of the Asbestos? There are several varieties of asbestos on the market. The long-fibre "silky" crysolite asbestos gives the best results, since it does not pack so closely as many of the less suitable varieties, and, in consequence, it filters most rapidly. Place a 10's brass wire sieve, bottom upwards, on a sheet of paper. Rub a handful of asbestos roughly over the sieve so as to break the asbestos into smaller fragments which pass through the sieve. 3 Collect that which remains on the sieve into another bundle, and repeat the operation until sufficient fibre has passed through. Next remove the dust and the fine powder by placing the material on a 30's lawn and agitating the mass while water is poured over the sieve. When the water which passes through has lost its " milky " appearance, and seems quite clear, transfer the washed asbestos to a beaker or flask and boil for about half an hour with hydrochloric acid (1 : 4). Now wash the asbestos by pouring it over a perforated funnel Hirsch's or Biichner's fitted with a piece of filter paper. Mix the washed asbestos with water, and the asbestos emulsion so prepared may be kept in a bottle ready for use. 2. Packing the Crucible with Asbestos Felt. A platinum or glazed porcelain crucible with a perforated bottom of the required size is fitted into an adapter shaped as shown in fig. 56, a. 4 This is mounted in a perforated rubber stopper which fits the neck of the filtration flask F. A piece of rubber b, made for the purpose, is placed over the mouth of the funnel, and the crucible c fitted into the rubber collar. J. J. Griffin & Sons (Kingsway) make perforated rubber stoppers which enable Gooch's crucible to be fitted directly into the filtration flask (fig. 57). 5 Pour enough asbestos emulsion into the crucible to give a layer of asbestos 1 to 2 mm. thick. The suction must not be too great, or the asbestos will pack too 1 F. A. Gooch, J'roc. Amer. Acad., 13. 342, 1878; Chem. News, 37. 181, 1878; G. C. Caldwell, Journ. Amer. Chem. Soc., 13. 105, 1891 ; T. Paul, Zeit. anal. Chem.. 31. 537, 1892 ; Chem. News, 67. 8, 1893. 2 P. Casaraajor, Chem. News, 47. 17, 1883 ; T. Paul, Zeii. anal. Chem., 31. 543, 1902; W. P. Barba, Journ. Anal. App. Chem., 5. 596, 1891 ; 6. 35, 1892 ; Chem. Neivs, 65. 101, 1892 ; P. A. Kober, Amer. Chem. Journ., 41. 430, 1909. 3 Or scrape the fibrous asbestos down with a knife. 4 T. W. Richards (Journ. Amer. Chem. Soc., 31. 1146, 1909) has added a funnel-like flange the top of the ordinary Gooch's crucible in order to facilitate washing. H. Vollers (Chem. \, 29. 1088, 1905) has a raised cylinder with - * J ' " ' ordinary Gooch's crucible. The designer cla blocked than the ordinary flat-bottomed cru< 5 W, R. Forbes, Chem. Neivs, 105. 27, 1912. to Ztg., 29."l088, 1905) has" a raised cylinder with perforated walls in placeof the flat bottom of the ordinary Gooch's crucible. The designer claims that this renders the crucible less liable to get blocked than the ordinary flat-bottomed crucibles. Neivs. I FILTRATION AttD WASHING. 105 tightly and the subsequent filtration will be slow. 1 Place the perforated disc filter plate upon the asbestos, and pour more emulsion into the crucible, so as to make a layer 0'5 to 1 mm. thick, as indicated in section, fig. 58. Run water through the crucible until no asbestos fibres run through. If the water which has run through the crucible be held in a good light, the floating asbestos fibres can be readily seen. Usually, J to | litre of water suffices. Dry the asbestos FIG. 56. Filtration with Gooch's crucible. and weigh. Again pass- J to J litre of water through the crucible, dry, and weigh again. If the weight be the same as before, the -crucible is ready for use. 3. How to use the Gooch Crucible. The dry and weighed crucible is mounted over a filter flask or Witt's jar (figs. 49 and 56). The liquid to be filtered is passed through the crucible, and the precipitate washed as if the crucible were a filter paper and funnel. When the precipitate has been washed and the crucible D/si ^Asbestos FIG. 57. Fitting Gooch's crucible in filtration flask. FIG. 58. Section of Gooch's crucible. dried, the whole is weighed. The increase in weight represents the weight of the precipitate. If the precipitate is to be ignited, it must be placed within a larger crucible or in a Gooch's saucer (fig. 59) and heated, at first gently, and then at the desired temperature. The ignited crucible is cooled and weighed in the usual manner. The same crucible can be used for a number of determinations of the same 1 It is often a good plan to let the joint between the crucible and rubber leak a little, or insert a regulating tap ' between F and W, or W and the pump. F. H. Wolff, Neues Jahrb. Pharm., 36. 65, 1871. IO6 A TREATISE ON CHEMICAL ANALYSIS. substance. When the collection of precipitates in the crucible becomes too large, the upper part can be removed, or the crucible recharged, and used as indicated above. 4. Sources of Error. Asbestos dried at 100-110 to a constant weight may lose between 6 '001 and 0'002 grm. of moisture on ignition. It is therefore necessary, when the weight of an ignited precipitate is to be determined, to dry or ignite the packed crucible the same length of time and at the same tempera- ture as that intended for the precipitate. 3 It is important, also, to remember that calcined asbestos may absorb appreciable amounts of alkali not removed by FIG. 59. Ignition of Gooch's crucible. washing. Most of the asbestos of commerce is also slightly attacked by water and feebly acid solutions. If the crucible is prepared as indicated above, there is no danger of losing asbestos during the washing of a precipitate. The filter tube is usually preferable to the Gooch's crucible when it is necessary to heat the precipitate in a current of gas. 2 53. Gooch's Crucibles packed with Soluble or Volatile Felts. The Gooch's crucible has deservedly won a permanent place in general analytical practice. It is exceptionally valuable when precipitates are to be re- dissolved from the felt after washing (e.g., sodium titanate in hydrochloric acid) ; when a mixed precipitate is to be separated into parts by appropriate solvents (e.g., washing antimony sulphide free from sulphur by carbon disulphide) ; separation of sulphides soluble and insoluble in alkaline sulphides, and generally when the desired solvent attacks filter paper and not asbestos. In some cases, the felt, after the action of the solvent, has to be separated from the solution 1 Or find the amount of moisture retained by the dry Gooch's crucible, and make the necessary allowance when the precipitate is weighed. G. Auchy, Journ. Amer. Chem. Soc., 22. 46, 1900; H. Theile, Zeit. offent. Chem., 7. 388, 1901 ; K. Windisch, Chem. Centr., 75. ii. 1621, 1905 ; M. Austin, Amer. J. Science (4), 14, 156, 1902. 2 E. Murmann (Monats. Chem. , 19. 403, 1 898) has described a modified Gooch's crucible in which the precipitate can be dried and heated in a current of gas e.g., lead sulphate, barium sulphate, zinc, manganese, copper, and antimony sulphides, etc. They are rather expensive and fragile. FILTRATION AND WASHING. 1 07 by filtration and washing. In that case, Gooch l proposed packing the crucible with anthracene instead of asbestos. The anthracene emulsion 2 is made by moistening anthracene with alcohol and then shaking the mass with water. The emulsion is used as if it were asbestos, 3 with due regard to its special properties. To remove the anthracene from a precipitate, stand the crucible in a small beaker, 4 add enough of the solvent benzene for preference and gently warm the vessel by immersing the beaker in hot water. 5 Wash the precipitate with the solvent, then with alcohol, and finally with water if necessary. The volatility of anthracene enables it to be removed by volatilisation in some cases where it is not wanted say, when a barium sulphate precipitate is to be treated with concentrated sulphuric acid. Crucibles with anthracene felt are not much used, and they must give way to the more convenient Munroe's crucible. 6 54. Gooch's Crucibles packed with Platinum Felt Munroe's Crucible. C. E. Munroe 7 proposed platinum sponge in place of the asbestos in Gooch's crucible, and crucibles so packed have so many applications, not possible with the asbestos-packed crucibles, that they are bound to win a permanent place in general practice. Although Munroe's crucible was suggested in 1888, its merits have only recently attracted serious attention. The main objection to Munroe's crucible is its* cost the platinum felt is packed in a platinum crucible ; but Brunck 8 has made a similar crucible with Royal Berlin porcelain, and "burned" the felt into the glaze so that it cannot be detached without the application of force, or overfiring. Brunck's crucible will bear a red heat if the temperature be raised gradually. 1. Preparation of the Platinum Felt. A concentrated solution of chloro- platinic acid is treated with ammonium chloride in slight excess. The resulting precipitate of ammonium platinichloride is washed several times with water, and finally with alcohol. The excess of alcohol is poured off as soon as the ammonium platinichloride has settled. 2. Packing the Crucible. A perforated platinum crucible, with small and numerous perforations, is placed upon several layers of filter paper, and pressed firmly on the paper while the alcohol-moist mass of ammonium platinichloride is 1 F. A. Gooch, Proc. Amer. Acad., 20. 390, 1885 ; Chem. News, 53. 234, 1886. 2 Anthracene is insoluble in water and aqueous solutions of salts, alkalies, and dilute acids ; but soluble in benzene, carbon disulphide, etc. Anthracene volatilises at 213. Naphthalene might be used in place of anthracene. 3 If the anthracene of commerce be too coarse, dissolve it in hot water, and precipitate by cooling with water. 4 Wagner's filter tube (M. Wagner, Pharm. Ztg., 52. 766, 1907) has a stopcock and perforated glass disc fused into a filter tube. By closing the stopcock, and pouring in the solvent, a precipitate can be kept immersed in a small quantity of the solvent as long as desired. 5 Note the inflammability of the solvent. No naked flame must be near. 6 In some special cases, an emulsion of filter paper or gun-cotton is used The paper pulp filters very slowly. Bottger recommended gun-cotton in place of asbestos in cases where the precipitate is to be ignited in an oxidising atmosphere (H. N. Warren, Chem. News, 6l. 63, 1890 ; F. Stolba, Zeit. anal. Chem., 17. 79, 1878). 7 C. E. Munroe, Journ. Anal. App. Chem., 2. 241, 1888 ; Chem. News, 58. 101, 1888 ; H. Neubauer, Zeit. anal. Chem., 39. 501, 1900 ; W. C. Heraeus, Zeit. angew. Chem., 13. 745, 1900 ; P. Bernhardt, Chem. Ztg., $2. 1227, 1908 ; 0. Brunck, ib. , 33. 649, 1909 ; W. 0. Snelling, Chem. News, 99. 229, 1909 ; Journ. Amer. Chem. Soc., 31. 456, 1909 ; 0. D. Swett, ib., 31. 928, 1909 ; T. W. Richards and A. Staehler, ib., 29. 623, 1907 ; F. Zerban and W. P. Naquin, ib. t 30. 1456, 1908; F. A. Gooch and F. B. Bayer, Ainer. J. Science (4), 25. 249, 1908; M. M. Brewer, George Washington Univ. Bull., 5. 79, 1906; Chem. Eng., 5. 261, 1906; H. J. F. de Vries, Chem. Weekblad, 6. 816, 1909. 8 0. Brunck, Chem. Ztg., 33. 649, 1909. IO8 A TREATISE ON CHEMICAL ANALYSIS. poured in until it is filled, say, to a height of 0'25 or 0'5 cm. 1 The alcohol will be rapidly absorbed by the filter paper upon which the crucible is being pressed, and at the end of a few moments the surface of the salt in the crucible will suddenly appear " dry," showing that the excess of alcohol has been absorbed by the filter paper. The filter paper will prevent the salt running through the perforations at the bottom. The crucible is then removed from the filter paper, and a uniform layer of ammonium platinichloride should cover the bottom of the crucible. Warm the crucible in a steam oven until the alcohol has all volatilised. Place the cap and cover on to the crucible, and gradually warm the crucible until the ammonium platinichloride has all decomposed. The temperature may then be raised to dull redness. Let the crucible cool. The bottom of the crucible should be covered with a continuous and uniform layer of platinum sponge, but generally the sponge will be found to have with- drawn, more or less, from the sides of the crucible, and to be interrupted by several cracks. Press the mass gently with the finger, or with a glass rod with a flattened end, so that the cracks are pressed together and a continuous layer of the platinum sponge is spread over the bottom of the crucible. If, however, the layer is not quite continuous, use fresh chloroplatinic acid to fill up the cracks, and to " patch " the layer of platinum sponge. The cracks, etc., may be so bad that it is advisable to remove the platinum sponge by means of a spatula, and start again. When the crucible is covered with a uniform layer of platinum sponge, it is ready for use. 2 Munroe recommends rubbing the platinum sponge with a glass rod at this stage in order to smooth out the mat and remove cracks. This compresses the mat and burnishes its surface. By gently rubbing all parts of the mat until the surface becomes considerably burnished, the smooth, polished surface so obtained has many advantages, even though the porosity of the felt is greatly reduced. The felt prepared by the above directions, due to Snelling, is said to be 1 00 times more porous than one of asbestos. It also retains the finest precipitates without the difficulties of turbid filtrates or clogging of the felt presented by asbestos felt in filtering barium sulphate and calcium oxalate. There is also -no danger of contaminating the filtrate with silica, alumina, iron oxide, magnesia, and lime, as is the case with asbestos felt. The crucible is used for filtering and washing, like the regular Gooch's crucible, page 104. One special advantage of Munroe's crucible is that the precipitate can be removed by means of appropriate solvents, and the crucible dried and weighed ready for the next determination. In illustration, the following results may be quoted from Munroe's paper : grms. Platinum crucible alone - . . 30*6744 Crucible and felt, first ignition . . . 31 '0607 Crucible and felt, second ignition . . . 31 -0607 Crucible and felt, third ignition . . . 31 -0607 Crucible, felt, and calcium oxalate . . . 31 '5385 The same after treatment with HC1, etc. . 31 '0609 Second treatment with hydrochloric acid . . . 31 '0608 Crucible felt, and barium sulphate . . . . 31-4403 After treatment with cone. H 2 S0 4 31-0640 Second treatment with cone. H 2 S0 4 ...... 31 '0609 1 Snelling recommends placing a circular piece of fine platinum wire gauze of suitable size in the bottom of the crucible before pouring in the ammonium platinichloride. The felt is then tougher and less easily injured, and also less liable to crack and curl. 2 Snelling recommends the addition of two or three drops of chloroplatinic acid to the platinum felt in the crucible. The porosity of the felt will cause the solution to distribute itself throughout the whole mass. On ignition, the chloroplatinic acid will be decomposed with the separation of platinum. The platinum which is so formed appears to cement the particles of platinum sponge together and to the walls of the crucible, so that the felted mass is toughened. FILTRATION AND WASHING. The crucible can be used over and over again for hundreds of analyses, unless careless handling makes the introduction of a new felt necessary. After weighing, the crucible is inverted to remove part of the precipitate, and the remainder is removed by placing the crucible on a pipeclay triangle over a porcelain dish so that the bottom of the crucible is immersed in the given solvent. By gradually heating the solvent, not to boiling, the precipitate will soon dissolve. Care must be taken with carbonates, or effervescence may cause particles of the felt to be dislodged. 1 A list of solvents for various precipitates, taken from Swett's paper, is given in the Appendix Table XCIV., page 730. 1 Precipitates sometimes cling tenaciously to the surface of the platinum felt, and if they are removed mechanically, will often take particles of felt with them. In cases where it may be necessary to mechanically remove precipitate, a perforated disc of platinum of suitable size may be placed on top of the platinum felt, as recommended by Richards and Staehler (I.e. ). CHAPTER VI. HEATING AND DRYING. 55. Heating. Bunsen's Burner. The burner, fig. 136, devised by R. Bunsen about 1855 1 is the best type for general purposes. It is too well known to need any descrip- tion. 2 There are a number of modifications. In Marshall's type of Bunsen's burner 3 the burner tube passes straight through the base, and the gas enters at the side. These are the so-called " drip-proof " burners, since there is no central gas jet to become choked with matter falling down the tube. The regulator works beneath the base. Teclu's burner 4 (figs. 109 and 145) is a useful modification of Bunsen's burner. It can give a rather higher temperature than the original form of Bunsen's burner. The gas and air can be regulated so as to give anything between a luminous smoky flame and a "blowpipe" flame. The latter will decompose calcium carbonate. Porcelain burners (figs. 100 and 1 60) are convenient for use in hoods, etc., where metal burners would be corroded by the fumes. Accessories for Bunsen's Burners. The burners may be provided with supports for holding the frustum of a copper cone to protect the flame from draughts (figs. 138 and 184) ; with Lendrich's 5 jets for splitting the flame into three, four (fig. 60), or five smaller flames. The multiple flame so obtained is better than the ordinary single flame for heating basins, beakers, etc., over wire gauze. For instance, there is less risk of spitting at the later stages of an evaporation ; 6 with a mushroom top for spreading the flame (fig. 162) ; with a jet for giving a flat flame (fig. 113) ; an attachment for heating tubes, e.g., Weston's jet (fig. 122) or Ramsay's jet (fig. 145); with a ring top 7 for heating a crucible from above downwards (fig. 170). Some burners are supplied with a rod permanently fixed to the base. The rod is intended to support a retort ring carrying a triangle, which can be adjusted at any height for heating crucibles, basins, etc. ; or a similar support may be temporarily attached to the outside of the tube of the 1 P. Desaga, Dingier' s Journ., 113. 340, 1857. For the efficiency of Bunsen's burners, see O'Connor, Ckem. Trade Journ., 51. 461, 1912. 2 The burners are usually attached to the supply nozzle by means of rubber tubing. A drop H. O'Connor, Ckem. Trade Journ., 51. 461, 1912. led to the supply of glycerol between the rubber and the metal will prevent the rubber sticking fast A. H. Church, Ckem. News, 35. 1, 1877. Flexible copper tubing is ultimately cheaper and safer than rubber tubing. There is a variety of "flexible" copper tubing on the market which is a nuisance on the bench, since it is too rigid. 3 H. Marshall, Journ. Soc. Ckem. Ind., 16. 395, 1897 ; Brit. Pat. No. 238, 1897 ; W. P. Evans, ib., 16. 863, 1897 ; F. Allihn, Ckem. Ztg., 23. 996, 1899. 4 N. Teclu, Zeit. anal. Ckem., 31. 429, 1892; 33. 450, 1894 ; Journ. prakt. Ckem. (2), 47. 535, 1893. 5 K. Lendrich, Zeit. Nakr. Genuss., 12. 593, 1906 ; D.R.G.M. 279398 to 279400, 1903. 6 A. A. Blair (The Chemical Analysis of Iron, Philadelphia, 20, 1908) recommends placing a thin circular disc of asbestos paper, about 2 cm. diameter, on the wire gauze so as to cover the point of the Bunsen's flame, and throw the heat more to the sides of the dish being heated on the gauze. 7 L. Hormuth, Zeit. anal. Chem., 43. 231, 1904. HEATING AND DRYING. I I I burner, as shown in fig. 60. 1 The burner may also be provided with two slots which fit a fork sliding up and down a retort stand. The burner can then be adjusted at any required height (figs. 125, 162). The burner may also have a by- pass for a " pilot " light (fig. 137). This is a convenience when a burner has to be frequently lighted and extinguished. There are several types of burner in which air and gas are simultaneously regulated so that, by merely turning a stopcock, the flame can be turned quite low without being extinguished, or "striking back." There is then no need for a pilot light e.g., Finkener's and Pethybridge's burners. 2 The burner can also be provided with a jacket to prevent losses by the radiation of heat e.g., fig. 109. Argand Burners. The Bunsen's burner can be fitted with a ring for supporting a mica chimney. This is convenient for low-temperature ignitions. The flame FIG. 60. Stand for Bunsen's burner. is not affected by air-currents. This modification imitates the Argand burner, which is coming into general use for charring filter papers, low -temperature ignitions, etc., particularly where uniformity of temperature is desirable. The Argand burner is provided with a steatite jet, and with a metal or a mica chimney (fig. 112). The so-called " micro-burners " are useful sources of heat when a bath is to be heated to a low temperature for a long time, and when a comparatively low uniform temperature is desired. See fig. 198, page 635. Meker^s Burners. G. Meker's burners (figs. 61 and 94) 3 are probably the best high-temperature burners on the market. Their virtues are usually extolled 1 See L. Quennessen, Chem. News, 88. 66, 1903. 2 Chem. Eng. Works Chem., I. 187, 1911. For stands to enable crucible to be rotated while it is being heated, see Hodes and Gobel, Chem. Ztg., 35. 488, 1911. 3 G. Meker, Chem. News, 99. 88, 1909 ; C. G. Fischer, Met. Chem. Eng., 9. 222, 1911. I I 2 A TREATISE ON CHEMICAL ANALYSIS. in the dealers' catalogues. 1 They can be obtained with or without a blast (figs. 61 and 92) attachment. The blast Meker's burner is quite satisfactory for silica and other ignitions requiring a blast. 2 Meker's burners are also adapted for heating muffles, and for crucible furnaces. There are many advantages in making ignitions in muffles when a large number have to be made. The muffle furnace shown in fig. 61 has a fireclay muffle, and is heated by a blast Meker's burner ; quartz muffles are not suited for the blast furnaces. Hot Plates. Sand baths are convenient for heating round-bottomed vessels, but they are dirty at best, and they should be banished from analytical laboratories. Quartz plates or plates of asbestos millboard are frequently used, but the asbestos soon deteriorates where the flame strikes. Quartz plates are more economical, since they last a long time. 3 Plates of cast iron, or boiler plate, 4 about O5 cm. thick, rest- ing on a tripod or quadripod, and heated by a ring or a Fletcher's burner, are preferable to the sand bath. 5 Radiator. When a liquid is to be evaporated or a moist solid is to be heated in a crucible, to avoid spattering, it is generally safer to heat the crucible in a so-called "radiator," rather than on a hot plate. The "radiator" corre- sponds with the "nickel beaker" of Jannasch. 6 The most convenient form of radiator is a nickel crucible with a triangle placed inside, so arranged that the crucible to be heated is ap- proximately equidistant from the sides and bottom of the nickel crucible. The nickel crucible and contents are heated over a small flame, when the liquid in the inner crucible is soon eva- porated. The lid can then be placed on the nickel crucible, and the contents of the inner crucible baked at a higher temperature. Several modifications naturally suggest themselves. The ring burner, fig. 96, page 170, serves a similar purpose. Electric Heating. If an electrical current is available, a great deal of the heating may be done more cheaply and cleanly than with gas. The gas supplied at the present day is very dirty. Baths and utensils soon become covered with thick FIG. 61. Blast Meker's burner with muffle. 1 But not exaggerated in, say, the catalogue of the Cambridge Scientific Instrument Co. 2 The blast for the burner photographed in fig. 61 is supplied by a Root's blower. It gives excellent results. A double water- injector pump, fixed on to a water tap, was found quite satisfactory with a 13-cm. Meker's blast burner. This latter type of blower has the advantage that it can also be used for the exhaustion of filtration flasks, desiccators, etc. Certain types of crucible furnaces which will give the necessary temperature, are to be avoided because they dent the platinum crucible when it is heated while resting between the fireclay projections. 3 E. Erlenmeyer, Zeit. Chem. Pharm., 8. 639, 1864 ; C. Weigelt, Rep. anal. Chem., I. 9, 1881. 4 Cast iron is less inclined to warp than wrought iron. 5 Nickel wire gauze, it may be added, does not "rust or perish" to the extent that other gauzes do. H. L. Robinson, Chem. News, 76. 253, 1897. 6 P. Jannasch, Praktischer Leitfaden der Gewichtsanalyse, Leipzig, 37, 1904 ; W. M. Thornton, Journ. Ind. Eng. Chem., 3. 418, 1911; F. P. Treadwell, Kurzes Lehrbuch der analytischen Chemie, Berlin, 24, 1911 ; W. F. Hillebrand, Butt. U.S. Geol. Sur.,^22. 31, 1910. HEATING AND DRYING. I I 3 crusts of naphthalene, etc., when heated by gas. Electricity is the ideal heating agent for a laboratory. Hot plates, air and water baths, hot-jacketed funnels, muffles, etc., can all be conveniently heated by this agent. With a little ingenuity, too, the cost of the necessary apparatus for electrical heating is less than for gas. Improvements are being published nearly every month, so that published descriptions of apparatus are soon obsolete. Triangles. Plain pipeclay, porcelain, or quartz triangles, fig. 62&, with iron, or, better, nickel or nichrome 1 wire are employed for general work. Quartz triangles are best for high-temperature work ; they are far more durable than porcelain or pipeclay, and better resist sudden changes of temperature. Coleman 2 has suggested placing a projection in the middle of the pipeclay tubes as indicated in fig. 626. These projections allow the flame to play about the crucible, 3 and less heat is thus lost by conduction ; but they are liable to dent platinum crucibles. The same remark applies to Schmelck's triangles. 4 Many modifications have been suggested. Some types of triangle can be adjusted to fit crucibles of different sizes, but generally these triangles are something of a nuisance after they have been in use some time. The springs and screws get FIG. 62. Triangles. out of order. 5 Lienau's triangle, fig. 62c, is supposed to allow expansion without breaking the pipeclay. 6 Plain platinum triangles, fig. 62d, are generally used for platinum crucibles. They must be made of stout platinum wire, and this is costly. If a platinum crucible be blasted on a platinum triangle, the triangle may stick to the crucible. The two can only be separated by tearing a piece out of the crucible. Hence, use quartz triangles for platinum crucibles which are to be "blasted." Hebebrand 7 made triangles of nickel with platinum buttons at the points where the triangle would come in contact with the platinum crucible. The plain platinum triangle gives best results for general work, but Hebebrand's triangles work all right for low-temperature work. The triangle shown in fig. 60 can be screwed tight when the wires have begun to sag. Hot platinum crucibles must not be lifted with iron, nickel, or brass tongs in such a way that these metals come in contact with the hot platinum. Brass tongs with a pair of platinum shoes (fig. 94) over the tips are generally used. Another form recommended by Blair 8 is illustrated by fig. 63. The body of the tongs is made of iron. The part a to b is of platinum, and fits over or into (fig. 63) the iron ends of the tongs. By means of these tongs the crucible can 1 K. C. Benner (Eng. Min. Journ., 91. 360, 1911) prefers nichrome wire that is, an alloy of chromium and nickel since it does not oxidise so readily in the blast lamp. 2 J. B. Coleman, Journ. Soc. Chem. Ind., II. 326, 1892 ; 0. Winkler, Zeit. anal. Chem., 18. 259, 1879 ; E. Murmann, Oester. Chem. Ztg., 12. 145, 1911. 3 Von Heygendorff (Chem. Ztg., 35. 523, 1911) uses pipeclay beads instead of clay tubes. 4 L. Schmelck, Chem. Ztg.,2O. 407, 1895. 5 L. Martius, Chem. Ztg., 24. 15, 1900 ; H. Lienau, ib., 29. 991, 1905. 6 The triangle should be so placed on the tripod that it is least likely to let the crucible fall should the wires be burnt through during an ignition, and those triangles least likely to let the crucible fall under these conditions are generally best. 7 A. Hebebrand, Chem. Ztg., 24. 37, 1900. 8 A. A. Blair, The Chemical Analysis of Iron, Philadelphia, 34, 1908. 8 114 A TREATISE ON CHEMICAL ANALYSIS. be clasped and held firmly without any danger of bending the crucible. These tongs are recommended for holding crucibles containing liquid fusions. In view of the high price of platinum, I prefer to cover the " bow " of a pair of ordinary brass tongs with platinum foil. This is cheaper and quite as effective. If the FIG. 63. Platinum- tipped tongs. tongs be placed on the bench points downwards after holding a hot crucible, the hot points are liable to pick up matter from the bench, which is afterwards transferred to the crucible. 56. Platinum Apparatus. Modern platinum Ware is inferior in quality to that on the market some years ago, and the cause has been the subject of special inquiry by a committee of the American Chemical Society. 1 The main objections are: "(1) undue loss of weight on ignition ; (2) undue loss in weight on acid treatment, especially after strong ignition; (3) unsightly appearance of the surface after strong ignition, especially after the initial stages of heating ; (4) adhesion of crucibles and dishes to triangles sometimes to such an extent as to leave an indentation on the vessel at the points of contact with the triangle, even when complete cooling has been reached before the two are separated ; (5) alkalinity of the surface of the ware after strong ignition ; (6) blistering ; and (7) development of cracks after continued heating." It is the general opinion that the trouble arises from the working of scrap platinum into chemical ware. The main difficulties here mentioned are not characteristic of platinum ware from some of the best manufacturers. Many new and old platinum crucibles lose weight on long blasting. This is generally thought to be due to the distillation of iridium from the platinum alloy. 2 If a crucible suffers from this disease, the rate of loss per five minutes' blasting must be determined, and an allowance made. Beilstein 3 states that the loss in weight becomes less after repeated ignition. On the other hand, Hall 4 found the loss on the twentieth heating even greater than the first. The following numbers represent the loss in weight which a new platinum crucible suffered after half-hour ignitions on a blast. The weight of the new crucible was 22-5299 grms. : 0-0005; 0-0004; 0-0003; 0-0002; O'OOOl grm. After a month's use, the crucible remained practically constant in weight during 1 " Preliminary Report of the Committee on the Quality of Platinum Utensils," Journ. Ind. Eng. Chem. ,3. 986, 1911. 2 L. L. de Koninck, Zeit. anal. Chem., 18. 569, 1879. H. Kayser (Wicd. Ann., 34. 607, 1888) noticed that platinum lost weight when heated in tubes in a current of air, and assumed that the platinum gave up fine particles to the air. L. Eisner (Journ. prakt. Chem. (1), 99. 257, 1866) found platinum crucibles lost weight when heated in porcelain ovens. E. Sonstadt, Chem. News, 13. 145, 1866; G. A. Hulett and H. W. Berger, Journ. Amer. Chem. Soc., 26. 1512, 1904 ; E. Goldstein, Ber., 37. 4147, 1904. G. C. Wittstein (Dingier s Journ., 179. 299, 1866) referred the loss to osmium, but F. Stolba (ib., 198. 177, 1870) pointed out that the loss of weight is greater than the amount of osmium in the platinum ore. 3 F. Beilstein, Pharm. Zeit. Russland, 19. 630, 1880. 4 R. W. Hall, Journ. Amer. Chem. Soc., 22. 494, 1900. HEATING AND DRYING. I I 5 the same period of blasting. 1 Another crucible did not show this phenomenon, but seemed to steadily lose about 0*0003 grm. per half an hour's blasting. The only inference to be drawn from these discordant observations is that the alloys used for making different crucibles are not the same. With luck, crucibles can be obtained which undergo practically no loss when heated in the blast. Crucibles which suffer a change in the rate of loss with time must have the correction determined at frequent intervals. If a crucible loses 0'2 mgrm. on ten minutes' blasting, the necessary correction is obvious. It is sometimes convenient to take the weight of the crucible after the ignition. Although platinum is not oxidised nor seriously attacked by the single acids used in analytical chemistry, many substances attack and combine with platinum at comparatively low temperatures. 2 The more important precautions to be considered in using platinum vessels may be conveniently summarised in the following decalogue : 1. Do not heat "unknown" substances in platinum vessels. 2. Always keep the platinum ware bright and clean. For this purpose use fine sea sand, with rounded grains and free from grit, as a polishing powder. 3 This removes most impurities, and polishes without serious loss. Apply the sand with the finger. Gmelin's recommendation to fuse a little potassium bisulphate in the dish is a good method for cleaning the inside of badly tarnished crucibles. 4 The bisulphate is poured out of the crucible while fluid, and the crucible then cleaned with water and sand. 3. Never heat the platinum crucible or dish in contact with iron or metals other than platinum; nor let hot platinum be placed in contact with foreign metals. Use nothing but pipeclay, quartz, or platinum triangles, and platinum- shod tongs. 5 4. Do not fuse metals in a platinum crucible, nor heat in the crucible salts of the easily reducible metals, particularly salts of lead, 6 silver, arsenic, antimony, copper, bismuth, tin, etc. See page 266. 5. Do not heat phosphides nor the sulphides of the metals in platinum crucibles. Sulphur alone has no action. 7 Phosphorus attacks platinum. 1 The lid of this crucible contained about 1*5 per cent, of iridium. 2 For the corrosion of platinum by ammonium and potassium bromoplatinates, see G. Meker, Compt. Rend., 125. 1029, 1897 ; Chem. News, 77. 19, 1898 ; by potassium bisulphate, see page 186; by borax and boric acid; by ferric chloride (4FeCl 3 + Pt = PtCl 4 + 4FeCl 2 ), see A. Bechamp and C. Saint Pierre, Compt. Rend,, Chem. News, 4. 284, 1861 ; by ignition of ammonium chloroplatinate, see F. P. Treadwell, Kurzes Lehrbuch der analytischen Chemie, Leipzig, 2. 225, 1910 ; by solutions of titanic hydroxide in the presence of nitric or sulphuric acid, see H. Rose, Traite Complet de Chimie Analytique, Paris, I. 1028, 1859 ; W. B. Giles, Chem. News, 99. 4, 1909 ; by sulphuric acid (A. Scheurer-Kestner, Compt. Send., 91. 59, 1880 ; 86. 1082, 1878 ; Bull. Soc. Chim. (2), 24. 501, 1875 ; (2), 30. 28, 1898 ; M. Delepine, Compt. Rend., 141. 886, 1013, 1905; Chem. Neivs, 93. 108, 120, 1906; L. Quennessen, ib., 93. 271, 1906; J. T. Conroy, Journ. Soc. Chem. Ind., 22. 465, 1903; L. R. W. M'Cay, Inter. Cong. App. Chem., 8. 351, 1912) ; by sodium carbonate, see page 163 ; and by selenium compounds, page 441. 3 F. Stolba (Dingier' 's Journ., 198. 178, 1870) says good sand for the purpose may be obtained by shaking dried sponges. 4 If outside, put the bisulphate in a dish of suitable size, and allow the flux to completely envelop the vessel to be cleaned. F. Winkle (Chem. Ztg., 23. 803, 1899) recommends borax instead of potassium bisulphate for crucibles stained by the ignition of plant ashes. F. Stolba (Dingler's Journ., 221. 135, 1877) cleans dirty crucibles by melting equal parts of boron fluoride and boric acid in the crucible. 5 Sodium amalgam is useful for cleaning platinum stained with foreign metals (Chem. News, 2. 286, 1860). The amalgam is gently rubbed on the platinum with a cloth, and then moistened with water. This oxidises the sodium, and leaves the mercury free to amalgamate with the foreign metals. The mercury is wiped off, and the crucible cleaned with sand. 'Unfortunately, we sometimes have to take risks ; but never if it can be avoided. 7 W. C. Heraeus and W. Geibel, Zeit. angew. Chem., 20. 1892, 1907. 1 1 6 A TREATISE ON CHEMICAL ANALYSIS. Phosphates, under reducing conditions, 1 may form phosphides and phosphorus, which attack the metal. 6. Alkaline carbonates do not attack platinum very much, but the attack by caustic alkalies and alkaline earths is serious particularly lithium and barium oxides. Hence, alkaline nitrates, nitrites, hydroxides, and cyanides should not be heated in contact with platinum, if it can be avoided. 7. Do not use aqua regia or free chlorine in contact with platinum. Watch for reactions (e.g. nitrates, chlorides, etc.) in which chlorine might be liberated. For instance, hydrochloric acid and oxidising agents. A moment of thoughtlessness may render it necessary to purchase a new crucible. 2 8. Direct contact of platinum with burning charcoal, coal, or coke must be avoided. The platinum becomes brittle and liable to fracture if this precaution be neglected. 3 9. Never permit a smoky flame to be used in contact with platinum, since a grey film 4 is formed on the platinum in mild cases, and the platinum is quite disintegrated in bad cases. For a similar reason, do not let the inner cone of a Bunsen's flame come in contact with the metal. 5 10. Platinum is easily bent, 6 and care must be taken not to bend the crucible in handling, gripping with the tongs, etc. 7 57. Desiccators. Fig. 64 represents convenient desiccators. That illustrated in fig. 64.A is intended for general use the so-called Scheibler's desiccator. The desiccating agent is placed in the bottom of the desiccator ; a porcelain plate, with suitable holes, serves as a support for the crucibles and dishes. An aluminium plate is sometimes preferred, because it cools the crucibles, etc., rather more rapidly than porcelain. The crucible support rests on a ledge formed by a constriction in the walls of the desiccator. If the substance is to be dried in vacuo, the form fig. 64s can be used. This resembles fig. 64A, but it has a glass stopcock fixed in the lid. The desiccator can thus be connected with an air-pump. To open an exhausted desiccator, turn the stopcock very slowly, or a current of air may dis- place the powder under investigation. The rim of the desiccator which comes in contact with the lid, as well as the stopcock, should be smeared with vaseline, or, better, with a mixture of vaseline and beeswax melted together. There is a fault with these desiccators. Moist air is lighter than dry air. Hence, it is bad on principle to place the desiccating agent at the bottom of the desiccator, beneath the substance to be dried. The action is then dependent 1 S. Kern (Chem. News, 35. 77, 1877) refers to the attack of platinum by the sulphur and carbon compounds in coal-gas. 2 Old crucibles are purchased by dealers in "scrap" platinum. 3 P. Schiitzenberger and A. Colson, Compt. Rend., 94. 26, 1882 ; A. Colson, ib. 94. 1710, 1882 ; 82. 591, 1876 ; J. Boussingault, Ann. Chim. Phys. (2), 16. 5, 1821 ; C. L. Berthollet, ib. (1) 67. 88, 1808 ; C. G. Memminger, Amer. Chem. Journ., 7. 172, 1886. V. Meyer, Chem. News, 73. 235, 1896 ; A. B. Griffiths (Chem. News, 51. 97, 1885) says "platinum fuses at a comparatively low temparature when heated in contact with carbon." 4 Another type of grey film is formed even when the platinum is only heated in an oxidising flame. 0. L. Eramann (./burro, prakt. Chem. (1), 79-117, 1860 ; Chem. Neivs, 2. 256, 1860) showed that this film is due to a surface molecular change (crystallisation W. Rosenhain, Proc. Roy. Soc., 7 252, 1902), and is not attended by a gain or loss of weight. The stain should always be scoured off as soon as it is formed, otherwise the devitrification is inclined to spread. 5 A. Remont, Bull. Soc. Chim. (2), 35. 486, 1881. H. Petrzilka (Zeit. angew. Chem., 7. 255, 1894) recommended a gilded platinum protective capsule to resist attack by smoky names. 6 A. A. Blair (The Chemical Analysis of Iron, Philadelphia, 33, 1908) recommends the use of a wooden plug the same shape as the inside of the crucible, to be used for straightening the crucible when it is bent. 7 For soldering and repairing crucibles, see J. W. Pratt, Chem. News, 51. 181, 1885 ; H. J. Seaman, ib., 49. 274, 1884 ; T. Garside, ib., 38. 65, 1878. HEATING AND DRYING. I i 7 upon the slow diffusion of the lighter moist air downwards. Hempel 1 devised a desiccator in which the drying agent is placed above the substance to be dried. The action is said to be more rapid because of the circulation induced by the natural tendency of the light moist air to accumulate in the upper part of the desiccator. 2 This argument does not apply to drying in vacuo, though it appears a valid objection to the form shown in fig. 64A. Fig. 64c represents Hempel's improved desiccator. The main objection to this type is that the acid is apt to spill, particularly when the desiccator is being opened. When the loss on ignition of certain clays is to be determined, and the clays, after ignition, are cooled in the usual manner, it is very difficult to get constant results. This is particularly noticeable when the blast is not very powerful. FIG. 64. Desiccators. The difficulty is eliminated if the ignited clay be cooled in vacuo. The trouble arises from the fact that clays, dehydrated at a comparatively low temperature, absorb the dry gases in an ordinary desiccator in a remarkable manner. As much as 1*5 per cent, may be so absorbed. 3 Granulated calcium chloride, concentrated sulphuric acid, and phosphorus pentoxide are employed as desiccating agents. The drying agents require atten- tion. The sulphuric acid should be frequently renewed, since that which has stood a long time in a desiccator is ineffective ; 4 and, if it has become discoloured, owing to contamination with dust, etc., sulphur dioxide may be given off (page 7), and this can sometimes be detected by its odour. Pumice soaked in concentrated sulphuric acid is excellent. Phosphorus pentoxide soon becomes ineffective owing to the glazing of the surface with metaphosphoric acid, and hence this agent is not used except under special circumstances. Fresh 1 W. Hempel, er., 23. 3566, 1890 ; Zeit. angew. Chem., 4. 84, 200, 1891 ; F. Cochius, Chem. Ztg., 24. 266, 1900 ; E. Biltz, Pharm. Centrb., 12. 353, 453, 1890. For the "ether" vacuum desiccator, F. G. Benedict and C. R. Manning, Amer. Chem. Journ., 27. 340, 1902; H. C. Gore, Journ. Amer. Chem. Soc., 28. 835, 1906 ; E. Douzard, Amer. Journ. Pharm , 80. 588, 1908. 2 There is also the objection that the air in the interior expands when a hot crucible is intro- duced. On cooling, a partial vacuum may be produced, which renders the removal of the lid difficult. Hence, some prefer a desiccator with a side tubulure in which is fitted a calcium chloride tube which provides for the free circulation of dry air. A. R. Leeds, Chem. News. 23. 177, 1871. 3 C. Friedel, Compt. Rend., 122. 1006, 1896 ; Bull. Soc. Min., 19. 14, 94, 1896 ; 22. 5. 84, 1899 ; J. W. Mellor and A. D. Holdcroft. Trans. Eng. Cer. Soc., 10. 94, 1911. 4 W. F. Hillebrand, Amer. Chem. Journ. 14. 6, 1892. I 1 8 A TREATISE ON CHEMICAL ANALYSIS. granulated calcium chloride is most useful for general work. Krai l prefers a mixture of calcium chloride and quicklime in place of calcium chloride alone. 2 58. Laboratory Hoods, Fume Chambers. In quantitative analysis, where a great many evaporations are conducted on a water bath, it is necessary to protect the liquids from dust during the evaporation. The arrangement shown in fig. 95, page 168, is convenient for single evaporations, 3 but where many evaporations have to be made, special closets are advisable. In silicate analyses, great volumes of hydrochloric acid, steam, and ammoniacal vapours are delivered into the atmosphere of the laboratory by evaporating liquids, and these vapours condense on the metal work ; this sets up rapid and serious corrosion. Hence, an efficient fume closet is necessary for accurate work, for protecting the health of the worker, and for the preservation of apparatus in the laboratory. The comfort of the laboratory is largely deter- mined by the efficiency, of the hood. The fume chamber must have an efficient draught independent of atmospheric conditions ; and it must protect the exposed surface of evaporating liquids from dust. The need for a vigorous draught is most emphasised in dealing with opera- tions in which hydrofluoric acid vapours, hydrogen sulphide, and other noxious vapours are evolved ; while the need for protection from dust is most pronounced when liquids are evaporating in shallow basins. If a large volume of air be passing through the hood, the risk of contamination from dust may be greater than when the draught is less. Hence, it may be advisable to install two types of hood : one specially designed for vigorous draught, and the other specially designed to prevent contamination from falling dust. In the former type of hood the fume closet the chamber should be relatively small. A number of small hoods are more efficient than one large hood. The carrying capacity of the air for moisture, and the draught of the hood, are largely dependent upon the slight heating which the air receives in the chamber. In a large hood the warmed air is chilled and its carrying capacity for moisture, etc., is reduced, so that the fumes are not carried through the outlet provided, but work their way into the laboratory. The upper roof of the chamber should not taper too abruptly, and, if possible, the exit flue should be vertical throughout its entire length, so that the rising air may be baffled as little as possible. The sectional area of the exit flue should be about one-third the floor area. Thus, in a chamber with a floor 3 x 1 J sq. ft. the vertical flue should be about 4x4 sq. ins. For structural reasons it is not always convenient to provide a vertical flue. When the exit leads into a 1 H. Krai (Pharm. Centr., 37. 105, 1895) also recommends fused potassium bisulphate in place of sulphuric acid. The " spent" bisulphate can be regenerated by fusion. 2 For the amount of moisture left in gases by desiccating agents sulphuric acid, zinc and calcium bromides and chlorides, phosphorus pentoxide, and anhydrous copper sulphate, see R. Fresenius, Zeit. anal. Chem., 4. 177, 1865 ; C. Voit, ib., 15. 432, 1876 ; H. C. Dibbits, ib., 15. 121, 1876 ; P. A. Favre, Ann. Chim. Phys., (3), 12. 223, 1844 ; V. Regnault, ib. (3), 15. 129, 1845 ; J. D. van der Plaats, Rec. Trav. Chim., 6. 45, 1889 ; E. G. Morley, Airier. J. Science (3), 30. 140, 1885; (3), 34. 199, 1887; Journ. Amer. Chem. Soc., 26. 1171, 1904; G. P. Baxter and R. D. Warren, ib., 33. 340, 1911. From this work it appears that the weight of residual water (in grams) left in a litre of gas dried by these different agents is : Calcium chloride . . 0'0021 Calcium bromide . . . 0'0002 Zinc bromide . . O'OOll Sulphuric acid (1 '838) . . 0'0000025 Zinc chloride . . O'OOOS Phosphorus pentoxide . . '000000025 For alumina as a desiccating agent, see F. M. G. Johnson, Journ. Amer. Chem. Soc., 34. 911, 1912. 3 C. Winkler, Eer.,21. 3563, 1888 ; W. Hempel, ib., 18. 1434, 1885; C. M. Stewart, Chem. yews, 52. 208, 1885; C. R. McCabe, Chem. Eng., 12. 182, 1910; Spectator, Chem. Ztg., 35. 254, 1911. HEATING AND DRYING. 119 chimney, the hood usually draws well, but if the exit flue in the ground floor of a building be led through the outer wall, the hood may be quite useless against the " back draught " set up by an unfavourable wind. A fan is then a valuable auxiliary, since it makes the action of the hood independent of the weather. Sometimes a jet of compressed air blown into the hood just below the base and directed into the exit flue, on the principle of a steam ejector, is very efficient, and there is then very little trouble with corrosion, as is the case when a fan is used. A gas burner placed just below the base of the exit flue will sometimes provide sufficient draught. The hood illustrated in fig. 65 is typical of a good fume closet. For evaporating chambers a particularly vigorous draught is not needed, because the passage of a large volume of air over the surface of an evaporating liquid may lead to the deposition of dust from the air on to the liquid. If there FIG. 65. Front of fume chamber. FIG. 66. Treadwell's flue. is too little draught, the crystallising salts may " creep " badly, and the vapours will work back into the laboratory. There is also a risk of contamination from dirt falling down the exit flue, particularly on windy days. Hence, the exit flue of an evaporating chamber should not lead directly into a chimney. 1 Fresenius recommends leading the exit flue into a separate Russian chimney. " No fire must ever be made under this chimney, but it is most desirable to have this chimney placed close to another chimney kept constantly warm." Treadwell recommends the arrangement illustrated in fig. 66, where a kind of "dust trap " prevents the passage of dust into the evaporating chamber when a "back draught " occurs on a windy day. The hood has a glass roof A, and about 15 cm. below is a second glass plate I>, which comes within 3 cm. of the inner wall of the hood throughout its whole length. Between the two glass plates is a clay pipe (7, 15 cm. long and 5 cm. diameter, placed above the inner edge of the lower glass plate, and leading into the chimney in which a small gas flame D is burning. Dust, sand, etc., from the chimney "falls upon the plate B ; none gets into the hood." 1 R. Fresenius, Quantitative Chemical Analysis, London, I. 62, 1876; F. P. Treadwell, Kurzes Lehrbuch der analytischen Chemie, Leipzig, 2. 26, 1911 ; H. S. Sherman, Chem. Eng., II. 21, 1910 ; E. Keller, Journ. 2nd. Eng. Chem., 3. 246, 1910 ; C. R. McCabe, ib., 12. 183, 1910 ; C. F. Mabery, Journ. Anal. App. Chem., 6. 247, 1892 ; J. K. Meade, The Design and Equipment of Small Chemical Laboratories, Chicago, 1908 ; T. H. Russell, The Planning and Fitting -up of Chemical and Physical Laboratories, London, 41, 1903. For contamination by paint from inside of hood, see A. Schaeffer, Zeit. Nahr. Gemiss., 24. 403, 1912. CHAPTER VII. PULVERISATION AND GRINDING. 59. Pulverising Large Quantities. THERE is no difficulty in reducing friable materials to powder in, say, porce- lain mortars, etc., but harder substances present some difficulty. It is almost impossible to pulverise hard, tough substances without some contamination from the crushing or grinding apparatus. For sampling, and for other experimental purposes, the breaking of stone, in bulk, down to particles, say 10's lawn, is com- monly done in a crusher. Braun's jaw crusher 1 fig. 67 is useful for crushing a large quantity of material. This machine reduces the material between a pair of jaws, one of which is stationary, and the other vibratory. The fineness of the grinding is regulated by opening or closing the jaws by a regulating screw. The jaws can be adjusted to crush material from a diameter of 2J in. to that which will pass a 10's lawn. For fine crushing, however, the material is best run, at least, through twice once with the jaws comparatively wide apart ; the second time with the jaws close together. The machine is run at about 200 revolutions per minute. It can be operated by power or hand. It should be perfectly clean before it is used. In Braun's disc pulveriser the material is pulverised between a pair of iron discs. A small machine will reduce 80 to 90 Ibs. of material from J in. mesh to 10's lawn per hour. It may be run by power or hand. There is always a tendency to contamination from the iron disc of the pulveriser or the steel jaws of the crusher. Hence the sample for analysis should be very carefully magneted if it has been passed through the jaw or disc crushers. The worn parts of either of the above machines are made interchangeable, and consequently repairs are easy. 60. Pulverising Small Quantities. Large lumps maybe broken into coarse powder 2 by wrapping a lump in several folds of writing paper and striking it sharply with a hammer while it is FIG. 67. Braun's jaw crusher. 1 F. W. Braun, U.S. Pat. No. 497802, 1902 ; K. Zulkowsky, Zeit. anal. Chem , 27. 24, 1888. 2 If there are no objections, many minerals (felspar, flint, etc.). difficult to pulverise, become quite brittle if calcined and suddenly quenched in cold water 120 PULVERISATION AND GRINDING. 121 resting on a steel plate placed upon a firm, solid support. The coarse fragments so obtained can be still further reduced on a hard steel plate (a, fig. 68) on which is placed a steel cylindrical ring b. A hard steel pestle c fits into the ring. The steel pestle and plate are made from the hardest "diamond" steel hence the term "diamond mortar." 1 The fragment of rock to be pulverised is placed on the steel plate within the ring. The plate is placed on a firm support. The pestle is depressed by a few sharp blows with a mallet. Care must be taken not to grind or rub the metals against the fragments, otherwise the risk of contamina- tion is greater. This is the case, for example, with the mortar of the type shown in fig. 69, which is made from the very best hardened tool steel, and is useful for breaking hard rocks. Most materials can be magneted after crushing so as to remove steel particles introduced during crushing. The magneting of a ground powder to remove fragments of iron from the crusher is conveniently done by means of an electro-magnet. A certain amount of discretion is needed, because, not only is iron derived from the crushing FIG. 68. Abich's mortar and pestle. FIG. 69. Steel mortar and pestle. apparatus withdrawn, but certain magnetic minerals magnetite, for example may be removed at the same time. With the more powerful electro-magnets, minerals which are sometimes considered to be non-magnetic 2 limonite, hema- tite, chromite, iron and copper pyrites, etc. may be removed. This occurs with the more powerful electro-magnets which have their poles tapering to blunt points, and with the points close together. Hence, false results would be obtained in the analysis of certain materials if the powdered sample were first magneted by such instruments. When the substance under investigation itself contains magnetic materials magnetic oxide of iron, etc. the magneting is not admissible, even though it is practically impossible to crush large or small quantities of hard material in iron vessels without contamination of iron. Hence the iron determination must necessarily be high. The material should be crushed fine enough to allow it to be ground in an agate mortar without the particles flying off during grinding. Care must also be taken to grind all the sample in this way, otherwise the harder fragments may be unconsciously rejected and the softer particles alone retained for examination. 3 1 Also called Abich's steel mortar, although it was not devised by H. Abich (Pogg. Ann., 23. 309, 1831). 2 But which are in reality feebly magnetic M. Faraday, Phil. Trans., 153. 1, 43, 1846; C. G. Gunther, Electro -magnetic Ore Separation, New York, 1909. 3 For errors in the analysis of iron pyrites due to grinding with Wedgwood-ware mortars and pestles, see N. Glendinning and A. J. M. Edgar, Chem. Aews, 27. 1873. 122 A TREATISE ON CHEMICAL ANALYSIS. 61. Grinding in Agate Mortars. An agate mortar and pestle are usually employed for fine grinding. The mortar and pestle should be selected with care. They should be free from cracks and indentations. It is always necessary to be on guard against contamination from the grinding apparatus. Cracks and soft seams in an agate pestle and mortar are a source of danger. The small agate mortars and pestles are usually too short to hold with comfort. They should be fitted in wooden handles (fig. 70). Beginners most frequently err in charging the mortar with too much material. It is best to grind the material in small portions at a time until sufficient is obtained for the analysis. When large quantities of material have to be ground, an agate mortar and pestle driven by power can be used. Carling's instrument is superior to M'Kenna's. 1 These machines work automatically and with but little attention. The motions of the agate pestle and mortar simulate hand-grinding. In Carling's instrument, fig. 71, the table holding the mortar is rotated by a worm- and-worin wheel in the pedestal. This is driven by a chain band in the base. The pestle is worked by an eccentric movement on the rotating mortar. The FIG. 70. Agate pestle and mortar. FIG. 71. Carling's power-driven agate pestle and mortar. machine is easily cleaned. To remove the mortar, depress the feeder lever, raise the pestle clear by pushing the spindle through its bearing, and loosen the 1 K. Zulkowsky, Ber., 20. 2664, 1887 ; C. T. M'Kenna, Eng. Min. Journ., 70. 462, 1900 ; Iron Age, 66. 9, 1900; W. H. Hillebrand, Butt. U.S. GeoL Sur., 422. 55 1910; J. W. Mellor, Trans. Eng. Cer. Soc., 10. 94, 1911. PULVERISATION AND GRINDING. I2 3 knurled brass screws, one of which is attached to a loose pillar which, on being withdrawn, allows the mortar to be removed. The driving wheel should run from 100 to 150 revolutions per minute. As a rule, a power-driven agate mortar and pestle wears more rapidly than one used for hand-grinding. The speed must not be great enough to project the powder over the sides of the mortar. The materials in the mortar must be protected from contamination by oil, etc., from the wheels and gearing. A copper cup may be attached to the handle of the pestle, and a rubber sheet wrapped round the axle to reduce the risk of contamination from oil, etc., on axles and wheels above the mortar. The machine can also be advantageously run in a glass case to protect the contents of the mortar from dust. 62. The Dangers of Fine Grinding. There are a few risks incidental to the operation of grinding which must be noticed. Fine grinding is not to be applied indiscriminately. It must not be assumed that the powders ground specially fine for the Smith's process for alkalies, for the determination of ferrous iron, for carbonate fusions, etc., have the same properties as coarse powders. The more important effects of fine grinding on the composition of a powder are as follows : 1. Contamination from the Grinding Apparatus. The finer the grinding the greater the risk of contamination from the grinding vessels, and, in consequence, the greater the variation in the amount of silica, etc., derived from the pestle and mortar. Hempel l has investigated the effect of grinding 10 grms. of glass in mortars and pestles made of agate, steel, and cast iron until the glass passed through a given sieve. The results are given in Table XVII. Table XVII. Influence of Grinding Apparatus on the Composition of a Powder. Material. Mortar (Reibschale). Pestle. Weight. Loss in weight. Weight. Loss in weight. Agate .... Steel (new) Steel (old) Cast iron (new) Cast iron (old) . Green bottle glass 371-741 295-078 295-049 884-917 884-825 195-584 0-041 0-029 0-005 0-041 0-014 0-027 44-242 134-647 134-645 . 144-383 144-382 0-011 0-0021 0-0004 0-0009 o-oooo This table shows that the loss in weight, and consequently also the contamina- tion of the materials being ground, is greater with agate mortars and pestles. Hempel supposes that steel mortars and pestles are the best for grinding purposes ; there is then less contamination. The nature of the contamination, however, presents a difficulty. Iron is one of the important impurities in clays, etc., and the analytical numbers for this impurity would be too uncertain if iron mortars and pestles were used for the grinding. The effect of the impurity derived from the agate has an inappreciable effect on the analytical numbers. It is therefore 1 W. Hempel, Zcit. angeiv. Chem., 14. 843, 1901; G. A. James, Chem. Eng., 14. 380, 1911 ; V. Lenher, Journ. Ind. Eng. Chem., 4. 471, 1912. I2 4 A TREATISE ON CHEMICAL ANALYSIS. best to avoid, as much as possible, the use of the metal mortars and pestles, and keep to agate. 1 2. Gain and Loss of Hygroscopic Moisture. The fact that powders in " a very fine state of subdivision greedily absorb moisture from the air during the operation of weighing " 2 is well known. The speed of absorption is greater the finer the grinding. Thus, Day and Allan 3 have shown that coarsely ground felspar, containing little or no hygroscopic moisture, takes up progressively increasing amounts of moisture as the powder becomes finer, and the quantity so taken up may amount to 1 per cent, of the weight of the sample. Again, Hillebrand 4 found for a noritic rock the results indicated in Table XVIII. Table XV II I. Influence of Fineness of a Powder on Hygroscopicity. Water. Condition of rock Below 100. Over 1 00. 60 apertures per cm. After 30 min. grinding 0-03 o-io 0-66 0-66 After 120 min. grinding 074 i-oo This all agrees with an old observation of Descharmes, 5 who showed that rock crystal (quartz) increased in weight on pulverisation to such an extent that the original weight might be doubled. Furthermore, Thugutt 6 has shown that some minerals take up water during fine grinding, while others e.g. apophyllite lose water. Hence, he infers that the hygroscopic moisture is best determined on a coarsely ground sample. 3. Gain or Loss of Combined Water, Water of Crystallisation, etc. Bleeker 7 has noticed that in grinding crystallised magnesium sulphate 2-55 per cent, of water was lost during two hours' grinding with a Wedgwood-ware pestle and mortar; disodium hydrogen phosphate lost 1'85 per cent.; potash alum, CK9 per cent. With barium chloride there was an increase of 2*11 per cent, in weight. Steiger obtained similar results with calcium sulphate. 4. Oxidation of Ferrous Compounds. Mauzelius 8 has shown that with minerals containing ferrous compounds quite an appreciable amount is oxidised to ferric oxide during fine grinding. Thus, 5'13 per cent, of ferrous oxide was obtained on a coarsely ground sample, and 3 '13 per cent, on the same sample finely ground. It is therefore obvious that the powder must be ground as coarsely as possible say 120's to 150's lawn when the ferrous oxide is to be 1 H. Wurtz (Amer. J. Science (2), 26. 190, 1858) suggested removing the iron by digesting the powder with iodine water. a R. W. Atkinson, Chem. News, 49. 217, 1884. 3 A. L. Day and E. T. Allen, Amer. J. Science (4), 19. 93, 1909. 4 W. F. Hillebrand, Journ. Amer. Chem. Soc., 30. 1120, 1908. 5 P. Descharmes, Recueil Jndustriel, i. 64, 1828. 6 S. J. Thugutt, Centr. Min., 677, 1909 ; F. von Kobell, Journ. praJct. Chem. (1), 107. 150, 1869. 7 J. B. Bleeker, Chem. News, 101. 30, 1910; C. E. Gillette, ib., 104. 313, 1911. 8 R. Mauzelius, Sveriges Geol. Und. ArsboTc, 3. 1, 1907 ; N. Knight, Chem. News, 97. 122, 190 (oxidation of siderite) ; G. Steiger, Bull. U.S. Geol. Sur., 413. 33, 1910 ; E. T. Allen and J. Johnston, Journ. Ind. Eng. Chem., 2. 196, 1910 (oxidation of pyrites). PULVERISATION AND GRINDING. 125 determined. The degree of coarseness is determined by the decomposability of the powder in 15 or 20 minutes by digestion with boiling hydrofluoric acid. 1 Hillebrand has confirmed these observations, and tried grinding the hard mineral under water and under alcohol, with the idea of cutting off oxygen from the powder during the grinding. Alcohol gave the better results. He showed that it was not merely increased surface which gave rise to the oxidation, but local heating of the grains in contact with the air at the moment of crushing. He there- fore recommended the following procedure: "The coarse powder is covered in the mortar with absolute alcohol that leaves no residue whatsoever on evaporation. The amount used should not be more than enough to form a very liquid emulsion during the operation of grinding. One or two additions of alcohol may be needed if the grinding has to be of long duration. The grinding should be continued in any case until the mass becomes somewhat thick, then the mortar and pestle are left to dry by unaided evaporation of the excess of alcohol ; a wide glass tube should be supported above the mortar to keep out the dust without too much impeding air circulation. When supposed to be thoroughly dry, the powder is transferred in its entirety to a watch-glass and placed on the balance pan, where it is to be kept till it is quite evident that no further loss in weight takes place. It is then transferred to the sample tube." But even then, adds Hillebrand, "if a very fine powder must be employed, there seems no way known at present of correcting for whatever oxidation may have taken place during the grinding." 63. Sieving, Lawning, or Screening. Ground materials are sieved or lawned to make sure that no particles above a certain maximum size are present. Great care must be taken that there be no contamination from the material of which the screens are made. Screens made of the same metal as one of the constituents to be determined are forbidden. Lawns made from brass, phosphor bronze, and silk are in common use in analytical laboratories. Silk lawns are used whenever metallic contaminations are prohibited. Silk lawns are made from the best Italian silk mounted in wooden or metal frames, with or without collecting box and lid (fig. 72). The following are some commercial sizes of silk lawns : 23, 30, 46, 61, 76, 86, 117, 132, 147 apertures per linear inch. The powder should be gently shaken through these lawns in order to avoid contamination. The FIG. 72. sieving of powders with silk lawns for the determination of ferrous oxide is condemned by many analysts (page 465), because of the possibility of contamination with silk fibres. In sieving, all the powder must pass through the lawn. If only a portion be allowed to pass through, the "knottings" (i.e. the residue on the lawn) might have a different composition from that which has passed through. The more brittle constituents of a mixture e.g., quartz and felspar will be pulverised first. Thus, Zaleski 2 showed that the dust produced in pulverising granites is richer in felspar relative to quartz and the darker-coloured minerals than the coarser grains of the powder ; and with glassy frits, the parts which pulverise first are richer in alkali. 3 Hence, the residue on the lawns must be reground until all passes through. 1 If the single treatment does not effect complete decomposition, "which will often be the case 1 ' (Hillebrand), the solution, after titration, is allowed to settle, the residue washed once by decantation with water, and again treated with hydrofluoric acid, etc. 2 S. Zaleski, Tschermak's Mitt., 14. 350, 1895. 3 T. E. Thorpe and C. Simmonds, Proc. Manchester Lit. Phil. Soc., 45- 1, 1901 ; Journ. Chem. Soc., 79. 791, 1901. 126 A TREATISE ON CHEMICAL ANALYSIS. It is generally advisable to state the size of the mesh of the lawn used in analytical and experimental work. There is some confusion owing to differences in the size of wire, method of weaving, etc. In order that there might be as little confusion as possible, the Institute of Mining and Metallurgy have arranged a set of sieves with a definite size of aperture. 1 Their table follows (Table XIX.). Table XIX. The I.M.M. Standard Laboratory Sieve Scale. Mesh, Diameter of wires. Diameter of aperture. Screening area i.e. apertures per cent. per linear inch. inch. mm. inch. mm. holes. 5 o-i 2-540 o-i 2-540 25-00 8 0-063 1-063 0-062 1-574 24-60 10 0-05 1-270 0-05 1-270 25-00 12 0-0417 1-059 0-0416 1-056 24-92 16 0-0313 0-795 0-0312 0-792 24-92 20 0-025 0-635 0-025 0-635 25-00 30 0-0167 0-424 0-0166 421 24-80 40 0-0125 0-317 0-0125 0-317 25-00 50 0-01 0-254 o-oi 0-254 25-00 60 0*0083 0-211 0-0083 0-211 25-00 70 0-0071 0-180 0-0071 0-180 25-00 80 0-0063 0-160 0-0062 0-157 24-60 90 0-0055 0-139 0-0055 0-139 24-50 100 0-005 0-127 0-005 0-127 25-00 120 0-0041 0-104 0-0042 0-107 25-40 150 0-0033 0-084 0-0033 0-084 24-50 200 0-0025 0-063 0-0025 . 0-063 25-00 This subject is discussed in more detail in Vol. II. of this work. CHAPTER VIII. SAMPLING. When conscious choice is permitted to enter into the operation of sampling, a fair sample will not result, except by a miracle. W. GLENN. 64. The Problem of Sampling. A FEW grams of material are sufficient for a chemical analysis, although tons of material may be bought or sold on the result. Large sums of money, too, may be involved in the results of experiments made with but a minute fraction of the total bulk of the material. In such cases it is of fundamental importance to work with samples typical of the whole, since a small error on a gram of material would be multiplied a miliionfold when calculated up to tons of material. If the material be perfectly homogeneous, a " grab " sample, taken at random from any part, will be quite representative. In other cases for instance, a clay bank, or a prospective clay field, or a cargo of bone for moisture, where the amount of moisture might vary from 3 to 13 per cent. analyses and tests made on "grab" samples may be quite misleading. Sampling is the art of extracting from a large bulk of material a small portion which shall fairly represent the character of the ivhole bulk. It would be difficult to over-emphasise the importance of accurate sampling. Indeed, the errors incidental to the process of chemical analysis are sometimes insignificant when compared with the errors involved in bad sampling. It is surprising, to one who realises the importance of this subject, to contrast the care taken in conducting a chemical analysis with the carelessness which prevails in sampling. The analysis of a carelessly selected sample is worth nothing. Many complaints of the failure of chemical analyses can be traced to faulty sampling. In sampling commercial materials it is necessary to be "on guard " against unscrupulous vendors or manufacturers. They sometimes exercise a diabolical ingenuity in getting a good sample into the analyst's hands when perhaps the bulk of the material is indifferent or bad. Accurate sampling is difficult enough when we have to deal with the natural vagaries of raw materials, but the subject is still more difficult when we are confronted with ingenious artifices meant to deceive. Examples soon accumulate in a laboratory, but inventors are fertile in resource, and the field is always verdant. The chance or probability of a sample answering expectations might be compared with the chance of drawing the ace of diamonds from a new pack of cards ; of throwing a six with a dice, or of tossing " heads " with a penny. The probability of these events is very different, and the probability of selecting the best representative value for material in bulk by means of a grab sample depends upon the nature of the material. The less homogeneous the material the less likely is the grab sample to represent the character of the material in bulk (see size ratio, page 134). Samples selected by a vendor for advertising purposes are 127 128 A TREATISE ON CHEMICAL ANALYSIS. here excluded from consideration, because the chance of getting a " first-class result " can sometimes be compared with the probability of throwing a six with a dice loaded for that number. 65. Selecting the Sample. If the material be in slip for instance, in checking the composition of a glaze in the dip tub the material must obviously be thoroughly agitated before the sample is taken, or difficulties will arise owing to settling. A device like that indicated in fig. 73 is useful for withdrawing samples from different parts of a tub or an ark of slip. The end of the tube A with the plug B pulled tightly into its socket by means of the wire and handle C is sunk to the required depth in the FIG. 73. Slip sampler. slip. The plug is then pushed downwards for a few moments. Slip runs into the lower end of the tube. The plug is then pulled back into its socket, and the tube with the slip inside withdrawn. The slip can be removed in an obvious manner. Metzger l has a much more complicated device by which several samples may be taken simultaneously at several different depths. The sample of " slip " must be dried and then thoroughly mixed. This arises from the fact that clay and glaze slips, on standing, are liable to separate into layers of different composition. With powdered manufactured or natural products, " grab " samples can be selected from a number of different parts of the stuff in bulk. A cheese scoop is useful for drawing portions from stuff in bins, sacks, etc. Metzger (I.e.) has devised a good sampler for this purpose (fig. 74). 2 In the diagram (a) it is repre- FIG. 74. Metzger's sampler. sen ted ready for insertion. When inserted it is screwed so that an opening is made for collecting the material to be sampled (b). Another screw will close the apparatus ready for withdrawal (a). The number and size of the different "grab" samples is determined by the nature and size of the material under investigation. A ship-load of bone, for instance, wants different treatment from a truck-load of clay, a boat-load of felspar, or a load of borocalcite. Accurate sampling can only be performed when the materials can be pulverised. A perfect method of sampling must eliminate the personal factor, and work with machine-like precision. This is quite easy with powdered materials 1 P. Metzger, Zeit. anal. Chem., 39, 791, 1900 ; 0, Steinle, Zeit. angew. Chem.,$. 584, 1892 ; A. Gawalovski, Allgem. Osterr. Chem. Tech. Ztg., 6. 197, 1889. 2 R. Kroupa, Berg. Hiltt. Ztg., 48. 301, 1887 ; J. T. Stoddard, Journ. Anal. App. Chem., 4. 34, 1890. SAMPLING. 129 intimately mixed, and the " time-sampling machines " which grind and periodi- cally select a fraction of the whole do this work very well. This kind of sampling is seldom practicable in dealing with clays and potter's materials. In sampling many ores, 5 per cent., that is, one-twentieth of the whole, is supposed to be selected in portions of equal weight and at frequent intervals. If the material is being delivered in shovels, every twentieth shovel is sometimes put on one side ; if delivered in barrows, every twentieth barrow ; if in sacks, every twentieth sack, and so on, according to the nature of the material to be sampled. With cargoes of some materials " one-in-fifty " will be ample. 66. Reducing the Bulk of the Sample. A large sample must be reduced in bulk before it can be treated in the laboratory. The reduction may be done by hand or by machine. There are at least five methods available for reducing the sample to a convenient bulk for treatment in the laboratory: (1) fractional selection; (2) quartering; (3) channelling; (4) split shovelling; and (5) riffling. In machine sampling the material may be (1) divided into two or more unequal streams, and one or more streams reserved for the sample (split-stream samplers) ; (2) the whole of the stream may be taken off intermittently (time samplers). (1) Sampling by Fractional Selection. The selection of 5 per cent, of the total bulk of the material (heap A) to be sampled as indicated above is an illustration of this method of sampling. The first sample (heap B) is ground, if necessary, into grains on average, say, one-twentieth the size of the first sample, arid every twentieth shovel or barrow-load collected into another heap, C ; this is again ground and subdivided until one or two hundredweights are obtained in a heap, say D. The operations are summarised as follows : A= 1000 tons of a cargo of material to be sampled ; B = 50 tons in first fraction ; C = 2 '5 tons in second fraction ; D = 0-125 tons or 280 Ibs. in the third fraction. Hence, if w represents the weight of material in the mass to be sampled, the nth fraction will have w(0*05) n th part of the original bulk. (2) Sampling by Coning or Quartering. 1 A homogeneous pile of the material is built up as a cone about a real or imaginary rod as centre. Every shovelful of the material is thrown directly on top of the growing cone, so that it runs and spreads more or less evenly down the sides. As a matter of fact, the fine material has a tendency to accumulate about the centre of the cone, while the coarse material rolls down the sides and an almost imperceptible sorting occurs (see fig. 78). The operator, in building up the cone, drops the shovel of material on top of the cone by giving the shovel a jerk upwards (just over the apex of the cone), and then another jerk downwards and outwards as indicated by the arrow, a, fig. 75. The material falls downwards on the apex of the cone, 6, fig. 75. With experienced men, the circumference of the cone, when finished, will be nearly circular. The cone is then spread out into a " pancake " by means of shovels working from the centre to the periphery round and round the cone (fig. 76). The pancake is then separated into quarters by means of boards or steel blades (fig. 77), pointing, say, north and south, and east arid west. If the N.E. and S.W. quarters are rejected, the N.W. and S.E. quarters are piled into another cone, and again quartered. Thus 200 Ibs. would be reduced by 1 W. Glenn, Trans. Amer. List. Min. Eng. t 20. 155, 1891 ; C. M. Roberts, ib., 28. 413, 1898 ; P. Johnson, Eng. Min. Journ., 53. Ill, 132, 1892. 9 3 o A TREATISE ON CHEMICAL ANALYSIS. the first cut to 100 Ibs. A second cut would reduce this to 50 Ibs. And generally, starting with a weight w of material, the nth cut gives by quartering FIG. 75. The cone almost ready for flattening. Thus, starting with 200 Ibs., the a weight w(0'5) n th of the original weight w. tenth cut will give 200 x (0'5) 10 = 0'195 Ib. FIG. 76. The cone partially flattened. Two men can reduce a ton of material down to a few pounds in two or three hours. Perfect mixing, accurate subdivision, thorough cleanliness are essential to success in all the methods for reduction. It will be noticed that owing to the accumulation of the fine material to- wards the centre, if the apex of the cone be deliberately or accidentally deflected while being built, it is possible and probable that in levelling, the first layers, at the bottom of the cone, will lie wholly in one quarter, and thus the proportion of fine material in that quarter will be abnormally high. The effect of imperceptibly drawing the centre of the cone on the relative proportion of fine and coarse material will be evident from fig. 78 (after Brunton), which is a photograph of a cone built up in actual FIG. 77. " Pancake " cutting. SAMPLING. sampling practice, bisected by a sheet of glass, and one-half of the cone removed. The diagram shows well the structure of a cone with a drawn centre. 1 (3) Sampling by Channelling? The material to be sampled is spread out in the form of a square about 4 inches thick. Parallel grooves say 1 foot apart are cut FIG. 78. " Drawn" cone. (After J. D. Branton.) across the cake ; and then another set of grooves at right angles to these. The process is repeated with the material so cut from the square. The method is slow and inaccurate, for coarse pieces may fall into the grooves as they are being cut. (4) Split-shovel Sampling. One workman (left, fig. 79) throws the material FIG. 79. Sampling by the U-shovel or the split shovel. from a broad shovel upon a narrow U-shaped scoop-like shovel held by another workman (right, fig. 79) over a car or wheelbarrow (fig. 79). The material which remains on the shovel is retained. When a pile of this has accumulated in the w r heelbarrow, it may be further reduced in bulk by repetitions of the process. 1 J. D. Brunton, Trans, Amer. Inst. Min. Eng., 40. 675, 1909. 2 F. A. Lowe, Eng. Min. Journ., 31. 203, 1881; W. Glenn, Trans. Amer. Inst. Min. Eng., 20. 115, 1891. For a mechanical device for cone-sampling, see C. W. Kneff, Journ. Ind. Eng. Chem., 4. 682, 1912. 132 A TREATISE ON CHEMICAL ANALYSIS. There are several different types of split shovel. 1 Some have a series of such scoops on one shovel, with spaces and scoops equal in width. (5) Riffle Sampling. The riffle consists of a series of grids arranged to form alternate series of troughs and spaces. When the material to be sampled is spread by means of a shovel over the riffle, supported over, say, a wheelbarrow, part of the material passes into the barrow, and part remains in the troughs. When the troughs are full, the contents are put on one side for further trial. That which passes through the grids is rejected. There are several modifications. In Jones' riffle the troughs are inclined, and slope in opposite directions. There are no spaces. The material when spread over the troughs (2 in. wide) is broken into two streams passing in opposite directions, and each stream is collected in FIG. 80. Taylor and Brunton's split-stream sampler. a suitable receptacle. The riffle is usually employed for cutting down samples for the laboratory when the amount is too small for the larger mechanical samplers. Taylor and Brunton's splitter (fig. 80) is a modification in which numerous small spouts are arranged across the entire width of a large gutter, so that the main stream of the material, poured into the gutter, is divided into a number of smaller streams. The odd-numbered streams may be deflected to, say, the right, the even-numbered streams to the left. The material which collects on one side is rejected ; the other the sample side is retained. The method of pouring- will be obvious from fig. 80. The operations are repeated on the material which has passed through the divider as often as desired. These samplers reduce the volume 50 per cent, each time the material is treated, as is the case in quartering. It is important to thoroughly mix the material collected on the sample side before it is. passed through again. 2 1 J. D. Hrunton, Eng. Min. Journ., 51. 718, 1891 ; S. A. Reed, School Min. Quart., 3 253, 1882. 3 J. D. Brunton, Trams. Amer. Inst. Min. ng., 40. 567, 1909. SAMPLING. 133 67. Machine and Automatic Sampling. p Split-stream Samplers. In Clarkson's divider, 1 the material is poured into a funnel where it is split into six similar streams (fig. 81). The material FIG. 81. Clarkson's split-stream sampler. FIG. 82. Bridgman's time sampler. before it is placed in the funnel is supposed to have been ground to pass at least an 8's lawn. It is fundamentally important in all systems of sampling to make sure that the " sample " and " reject " are of the same composition. 1 T. Clarkson, Journ. Hoc. Chem. Ind., 12. 214, 1894. 134 A TREATISE ON CHEMICAL ANALYSIS. There are many other types on the same principle. In some, a series of funnels are superposed one above the other, so that a fraction of the stream is rejected and another fraction is further fractioned as it passes on to the subjacent funnel, e.g. F. W. Braun's " Umpire " sampler, 1 which is excellent. Time Samplers. In the time-sampling machine, the whole of a falling stream of the material is periodically deflected during a portion of the time. This eliminates one fault with split-stream samplers where the sizes of the grains are never evenly distributed across the stream. In some, on a large scale, the material is alternately subdivided and crushed from the coarsest down to the finest powder. The machines are represented by Snyder's time sampler ; 2 the Vezin time sampler ; 3 and the Brunton time sampler. H. L. Bridgman's time- sampling machine 4 is a compact little instrument occupying but 14 square inches. It has a divider D set in motion by clockwork or any other power (fig. 82). F is a funnel with a receptacle R for the fraction separated from the material being sampled. The portion rejected passes through the discharge orifice 0. The material runs from the hopper in a continuous stream to the divider, which cuts the stream eight times during one revolution. Four of these are delivered into the discharge orifice, and four are delivered into the receptacle. 68. Reducing the Grain Size and Bulk of the Material during Sampling. The process of sampling involves two operations : (a) reducing the bulk of the material as indicated in the preceding sections ; (b) reducing the size of the grains of the material. The important factors in accurate sampling are : (1) the relation between the amount of material selected as the sample and the original bulk of the material to be sampled the bulk ratio ; (2) the relation between the coarseness of the grain and the amount the size ratio (sampling); (3) the relation between the size of the grain and the amount isolated at each stage of quartering, etc. sizo ratio (quartering). (1) The Relation between the Bulk of the Sample and of the Material to be sampled. The greater the ratio between the weight of the sample and the weight of the whole of material sampled ; that is, the greater the ratio, Weight of samgle Bulk ratio, Weight of material in bulk the greater the probability of the sample representing the true character of the material under investigation. This is illustrated by the following table, due to Bailey : Table XX. Relation between Errors in Sampling and Size of Sample. Bulk ratio. Percentage error sample and bulk. 1 : 100 2 : 100 5 : 100 1-4 0-8 - 0-6 1 F. W. Braun, U.S. Pat. No. 682528, 1901. 2 F. T. Snyder, Canada Min. Rev., 17. 43, 1898. 3 H. A. Vezin, Trans. Amer. Inst. Min. Eng., 26, 1098, 1896. 4 H. L. Bridgman, Trans. Amer. Inst. Min. Enq., 17. 639, 1884: 2O. 416, 1891 U.S Pat. No. 553508, 1896. SAMPLING. 135 (2) The Relation between the Grain Size and the Weight of the Sample. The smaller the ratio between the weight of the largest pieces and the total weight of the sample ; that is, the smaller the fraction, Weight of largest pieces in sample Q . _ _ _ _ ^ olZ6 r&LlO. lotal weight of sample the more likely is the sample to represent the^character of the material in bulk. Bailey's experiments ! on sampling coal (with 5 per cent, ash) illustrate this point very well. He found that if the error is to be less than 1 per cent., the relation between the grain size of the sample and the weight of the largest pieces in the coal must be near that indicated in the following table, due to Bailey : Table XXI. Relation betiveen Fineness of Bulk Material and Size of Sample. Weight of largest Amount of sample pieces in Ibs. not less than (Ibs.) 6-7 39,000 2'5 12,500 075 3,800 0-38 1,900 0-24 1,200 0-12 600 0-046 230 0-018 90 Under these conditions the sample is less likely to be disturbed by an accumulation of abnormally fine dust or coarse material. (3) The Grain Size of each Fraction (Quartering). The relation between the grain size and the amount to be isolated at each stage of the quartering is an important factor, as will be obvious from the preceding notes on the relation between the grain size and the amount to be selected as a sample. Continuing the experiments just described, Bailey found that in order to get the deviations in the composition of the ash of coal within the limits stated, it was necessary to reduce the sample before quartering or fractioning to the grain size -indicated in the following table : Table XXII. Relation between Fineness of Sample and Bulk of Material. Weight of fraction in Ibs. Size to be crushed. Diam. in inches. 7500 2 3800 1-5 1200 1 460 075 180 0-5 40 5 0-5 0-25 0-425 (2'slawn) 0'20 (4'slawn) O'l (8's lawn) 0-076 (10's lawn) 1 E. G. Bailey, Joum. Ind. Eng. Chem., I. 161, 1909 ; 2. 543, 1910 ; F. C. Weld, ib., 2. 426, 1910 ; F. Fischer, Stahl Eisen, 32. 1408, 1912 ; W. B. Blyth, Min. Eng. World, 37. 613, 1912. 136 A TREATISE ON CHEMICAL ANALYSIS. According to Bailey, if the coal be crushed to a size 2-inch grain, do not subdivide to a fraction less than 8300 Ibs. without further grinding; 4-inch mesh for not less than 1100 Ibs., etc., as indicated below: Size of lawn .... 2 4 8 10 20's Minimum amount . . . 8300 1100 120 55 3 Ibs. Of course the best value for the bulk and size ratios indicated above will vary with the nature of the material and the degree of accuracy demanded in the sampling. In the absence of data for particular materials, the preceding data will serve as a guide in determining these factors. 1 To summarise, good sampling involves : (1) Adequate mixing ; (2) Impartial selection ; (3) Proper grinding. There is always a danger in applying rules blindly. The exercise of common sense, guided by these three principles, is necessary to cope with the different problems which arise from time to time. All the preceding discussion would be misleading if the object of the investigation were to determine the composi- tion of a specific mineral. In that case the purpose would probably be best served by very carefully isolating as clean and as pure a crystalline fragment as possible. All associated impurities, foreign matter, and all altered fragments should be excluded. Neglect of this precaution would render it necessary to admit the existence of innumerable varieties of a particular mineral. 2 69. The Receipt and Dispatch of Samples. In discharging a cargo of bone ash, the "English Form of the Hamburg Contract " specifies : Regular samples shall be drawn from each day's quantity of Ash discharged, and after these samples have been sufficiently mixed, four large bottles shall be filled and sealed, which shall serve as MOISTURE SAMPLES for the - . The rest of the landing cargo sample shall be ground sufficiently fine to pass through a sieve of 50 holes per square centimetre, of which ground sample there shall be filled and sealed four large glass bottles which shall serve as QUALITY SAMPLES. All this to be done under the superintendence of Sellers and Buyers or their repre- sentatives. Each appointed Chemist shall receive one MOISTURE SAMPLE for the determination of MOISTURE, and one QUALITY SAMPLE for the determination of MOISTURE AND PHOSPHATE OF LIMK, and, in ascertaining the percentage of Phosphate of Lime in the QUALITY sample, the calculation shall be based on the percentage of Moisture in the as shown by the MOISTURE SAMPLE. cargo Samples for analysis or testing should be clearly and distinctly marked ; 1 For papers on the theory of sampling see P. Argall, Trans. Amer. Inst. Min. Eng., 31. 235, 1902; D. W. Branton, ib. t 25. 826, 1895; S. H. Pearce, Chem. Met. Soc. S A., 2. 155, 1898 ; S. A. Reed, School Min. Quart., 3. 253, 1882 ; 6. 351, 1885 ; A. D. Hodges, Eng. Min. Journ., 52. 264, 1891; V. Samter, Chem. Ztg.,%2. 1209, 1224, 1250, 1908; C. Geissler, ib., 23. 43, 1058, 1899; German Pat. No. D.R.P. 100067, 1898; 100516, 1898; 0. Bender, Zeit. anal.- Chem., 48. 32, 1909 ; Zeit. Berg. Hutt. Sal., 55. 1, 1907; F. Janda, Oester. Zeit. Berg. Hutt., 52. 547, 561, 577, 1904 ; J. A. Barral and R. Duval, Dingle? a Journ., 217. 246, 1876; M. L. Griffin, Journ. Soc. Chem. Ind., 24. 183, 1905; W. Glenn, ib., 17. 123, 1898; C. E. Stromeyer, Pract. Eng., 44. 781, 1911; L. T. Wright, Chem. Eng., 13. 30, 1911; A. C. Fieldner, ib., 17. 50, 1913 ; M. Webber, Min. Scientific Press, 102. 846, 1911. 2 For the methods available for isolating minerals from clays, see Vol. II. of this work. SAMPLING. 137 and be securely packed so that there is no possibility of contamination during transit. 1 In disputes on the quality of materials, the sampling should be performed in the presence of the seller's and buyer's representatives. Three 8-oz. bottles should be filled with, say, the powder in question, and both representatives should put their seals on the bottles and sign an account of the taking of the sample. The bottles should have a label stating what the contents are, the name and address of the senders, and a distinguishing mark number or letter. 2 One bottle may be sent to an analyst agreed upon by both representatives ; or one bottle may be sent to the analyst selected by the buyer, and one bottle to the analyst selected by the seller. The remaining bottle or bottles are reserved for verification in case the two analysts differ appreciably in their results. In some cases, e.g. the purchase of iron ore, bone ash, etc., it is the practice to refer a sample to the " buyer's chemist " and a sample to the " seller's chemist." If these differ by less than, say, 1 per cent., the material is bought on the mean of the two results ; but if the results differ by more than the agreed 1 per cent., it is the custom to send the sample to a third chemist as referee, and to take the mean of the two nearest results. For bone ash, the " English Form of the Hamburg Contract " specifies that the third chemist shall only be called in when the results by the seller's and the buyer's chemists have " a difference of over 2 per cent, of tribasic phosphate of lime." The chemist who receives the sample registers the date of arrival, the pro- fessed nature of the sample, and the marks on the seals. A formal acknowledg- ment of the receipt of the sample should be sent, say : I hereby acknowledge the receipt of ... specimens of Packed in Nature of seal Marked Laboratory number. Purpose Charge Date received... Director of Laboratory. The whole mass is then emptied from the bottle on to a sheet of white paper 3 and sifted. 4 If any particles do not pass through the sieve, they must be ground until all passes through. The mass is then mixed with a spatula and re-sifted through a coarser sieve. A portion is then taken for analysis, and the -remainder 1 I have known a batch of firebricks which (so far as I could see) must have been saturated with salt during transit from the sender to the receiving chemist. 2 In case of a seizure of, say, a glaze sample by the Inspector of Factories, the material should be thoroughly agitated and three nearly equal portions removed in the presence of the proprietor or his representative. The corks of the bottles should be pressed level with the glass ; sealing wax placed on the top of the cork (or string if the sample should be a powder wrapped in paper), so that the package cannot be opened without breaking the seal of wax. A distinctive seal should then be impressed on the wax, and each package labelled with distinctive and similar labels showing the name of the firm, where taken, date of sampling, date sent to analyst, distinctive number or letter, and the nature of the sample. 3 The analyst should inquire if the sample has been divided into two or three portions when sampled. If not, the sample should be divided into two parts nearly equal, and one part should be sealed and labelled as indicated in the text. This sample may be reserved or returned either on receipt of the sample or when the certificate is supplied. Slop glazes are dried as indicated on page 330. 4 The mesh of the sieve must be determined by the nature of the powder. 138 A TREATISE ON CHEMICAL ANALYSIS. is returned to the bottle. The result of the analysis is recorded on a certificate somewhat as follows : Keferences to Laboratory Number I hereby report that the specimen of , marked , sampled by , and received from on , has been duly I analysed in this laboratory I tested with the following result : Director of Laboratory. No responsibility is taken for the accuracy of the sampling unless it has been done under our own supervision. - 70. Sampling Beds of Clay. Another type of sampling is sometimes required. To find if it is advisable to install plant for working a natural bed of clay, it is necessary to make a somewhat close approximation as to the amount of clay available, as well as to determine the nature of the clay, possible market, cost of transport, fuel, labour, etc., etc. Here we are alone concerned with making an estimate of the amount of clay and the collection of samples for investigation. In sampling an unworked bed of clay near the surface, a hole may be sunk through the different strata well into or through the clay under investigation, and 28-lb. grab samples taken from the exposed face. It is well to keep for reference small samples of the different superincumbent strata. Care must be taken in digging and sampling the hole not to mix fragments from one stratum with those of another. The distance between the original locus of the different samples and the surface should be measured and recorded on and in the boxes or bags in which the different samples are preserved. If the clay under investigation is any great distance from the surface, this procedure would be laborious if it were required to explore the thickness and extent of the clay. Many of the well- known types of borer bring a core to the surface with much less labour than is involved in digging holes, as described above. The type of borer to be employed is largely determined by the hardness of the different strata and the depth of the required boring. Two men can conveniently work one of H. Meyer's borers to a depth of 30-50 feet in a few hours. 1 Fraenkel's borers can be sunk about 12 feet by one man in a few hours in ordinary surface clays. Bore-holes have been driven by power over 7350 feet. 2 The cost for the deep-seated clays would generally be prohibitive, although bores ranging up to 500 feet are common enough. When a series of cores properly measured has been obtained, the nature of the clay can be investigated by methods to be described later. Much naturally depends on the geological character of the bed in question. The continuity, thickness, and nature of a bed of fireclay in, say, the coal measures in a given area can often be predicted with a high degree of proba- bility by a geologist provided with particulars of the same stratum in part of the prescribed area. Hence a boring within the prescribed area can only demon- strate what was almost a certainty. 3 A skilled geologist may be able to correlate 1 A. Fauck, Anleitung zum Gebrauche des Erdbohrers, Leipzig, 1877; Suppl., 1899; H. Bansen, Das Tiejbohrwesen, Berlin, 1912. 2 With rotary diamond drills 2003 '34 metres at Paruschowitz (Upper Silesia) B.A. Rep., 67, 1901 ; with percussion drills 5570 feet at Sperenberg (about 25 miles south of Berlin) H. W. Bristow, Oeol. Mag., 4. 95, 1875 ; B. Jager, Zeit. Berg. Hull. Sal., 59. 89, 1911. 3 The process of boring may then be compared with interpolation in mathematics. The possibility of "faults," etc., may render an extrapolation futile and stultify a prediction. SAMPLING. 139 the stratum under investigation with another known stratum elsewhere. The nature of certain strata frequently persists throughout wide areas ; for instance, the peacock marl (N. Staffs.), the Old Mine clay (Stourbridge), Oxford clay, London clay, etc. Other strata, and even the clays just mentioned, may vary in an apparently erratic manner in certain localities. A prudent man, therefore, will try to estimate what amount and what quality of clay is available. Sampling and boring is somewhat costly work, but, when properly done, it gives a clear idea of the extent and nature of a hidden bed of clay, and it prevents awkward mistakes, such as the erection of the kilns, machine sheds, etc., over what might be the best part of the clay. 1 Exposures and cuttings in a given neighbourhood may furnish enough information as to the amount of clay available in a given area without any need for an elaborate system of boring. When a series of borings have been obtained, they may be recorded in the Bo re 7 FIG. 83. Section constructed from borings. following manner. For convenience we apply the method to a particular problem. A few years ago, a valuable plastic clay 15 feet thick gave signs of tapering off. A series of borings 20 feet deep were made, and it was proved that the clay seam "petered out" between 40 and 50 feet away from the working face. A new seam of a somewhat similar plastic clay was "struck" below a 10-20 feet seam of a sandy clay. In boring systematically to find the extent of the new seam, what appeared to be the old seam was taken up not far from where it had tapered off. In boring, the " field " was divided into a series of imaginary rectangles resembling a chess-board. . A boring was made at every corner. With the data so obtained diagrams were drawn to scale representing sections through the different strata in particular directions. One such section is illustrated in fig. 83. This is drawn from the borings Nos. 5, 6, 7, 8, which were : No. 5. No. 6. No. 7. No. 8. ft. in. ft. in. ft. in. ft. in. Overburden 2 8 2 10 3 2 5 Plastic clay 9 7 2 11 2 6 Sandy clay Plastic clay B. 7 8 + 6 7 10 + 4 8 5 4 5 6 2 Sand ... ... 6 + 3 2 + The distance apart and depth of the borings for each particular " field " must be determined by the circumstances of the case, and the advice of a geological expert may be required. An instructive example occurs at an abandoned clay pit near Cobridge (Staff's.). 140 A TREATISE ON CHEMICAL ANALYSIS. 71. The Commercial Value of Clay Deposits. In order to evaluate a clay deposit, it is necessary first of all to determine the commercial value of the goods to be made from the clay, which, after all, reduces to the present market value of a cubic foot of the clay. Due considera- tion is of course paid to the district, facilities for transport, cost of fuel, etc. In some inaccessible districts land covering a first-class clay might have no more value than ordinary agricultural land, owing to lack of facilities for transport, cost of fuel, etc. In the second place, it is necessary to determine how much clay is available. In the third place, an approximate estimate must be made of the yearly consumption of the clay, and consequently also how many years the seam of clay will last at the estimated rate of consumption. We are then in a position to deal with the problem by arithmetic. 1 Let R represent the estimated value of the clay consumed per annum ; r, the rate of compound interest for the money were it invested at the current rate ; and n, the estimated number of years the clay is likely to last before the seam, or the given portion of a seam, is exhausted. The regular rules for compound interest show that the present value, P, of the given field of clay can be repre- sented by the expression : where, for convenience, we have written _ Too" To illustrate by example, suppose that 10,000 cubic feet of clay be removed each year ; that 1,000,000 cubic feet of clay are still available. 2 The clay is worth |d. per cubic foot. The rate of interest is 3J per cent. What is the present value of the field of clay 1 Here, r = 3'5 ; hence, from (2), p = 1 -035. The 1,000,000 cubic feet of clay will last 100 years if 10,000 cubic feet be removed per annum. Hence, R = 10,000 x \ = 2500d. Consequently, from (1), P-2500 - _PQQQ l 6d 3 1 -035100(1.035 -1)~ It is then possible to decide whether it is better to pay an annual royalty for the clay, to pay a proposed rental, or to purchase the field at a stated price. 1 E. Tschenschmer, Tonind. Ztg., 14. 121, 139, 1890 ; Brit. Clayworker, 18. 279, 1910. 2 The estimation of the cubic contents of a given seam is simple arithmetic solid mensura- tion when the borings are known. 3 Log p* = n log p. Hence 100 log 1*035 = 100 x "01 49403 = 1-49403 = log 31 -191, or 1-035 100 = 3M91. CHAPTER IX. THE REAGENTS. 72. Testing the Reagents. THE purity of the reagents and of the distilled water for analytical laboratories should not be taken on trust. 1 The glass containers may contaminate liquid reagents with iron, potassium, sodium, silica, calcium, and alumina by the dis- solution of the glass ; and zinc may be derived similarly from Jena glass and " nonsol " glass vessels. Iron is common in acids transported in carboys. Small flakes of iron may scale from the blowpipes used by the glass-blower, and this may not be removed or noticed in washing these vessels. Reagents may be con- taminated from their contact with the vessels tanks, stills, condensers, crystal- lising dishes, etc., made from lead, silver, copper, iron, aluminium, nickel, zinc, tin, porcelain, glass, etc. used in the process of manufacture. Thus, riot only may the ordinary constituents of glass be present, but nickel may be found in caustic alkalies, copper and aluminium in acetic acid, zinc in barium carbonate, 2 selenium and iron in hydrochloric acid, 3 and lead in organic acids and in hydro- fluoric acid. The latter has also been known to contain potassium sulphate and ammonium and sodium chlorides. Impurities may also be derived from the raw materials used in the process of manufacture. In fact, it is almost impossible to free some reagents from certain impurities except at a very high cost. For example, iron, alumina, and silica are found in all but the highest-priced grades of caustic alkalies and alkaline carbonates ; calcium is very difficult to remove from ammonium oxalate ; nickel from cobalt ; iron and lead from copper ; free acid from ferric salts ; ammonia from magnesium chloride ; etc. 4 All this is mentioned to emphasise the fact that constant vigilance is the price of successful work. Many impurities might easily escape detection. For instance, selenium in hydrochloric acid ; sulphates in platinum chloride ; lead in ammonia ; phosphorus ' in ammonium chloride and nitrate; and calcium in ammonium oxalate. 5 Hence, 1 The "analysed" chemicals supplied by many of the better-class makers are generally very good. 2 R. Wegscheider, Zeit. anal. Chem., 29. 20, 1890. 3 W. B. Hart, Chem. News, 48. 193, 1883. 4 J. W. Shade, Journ. Amer. Chem. Soc., 28. 1422, 1906 ; J. T. Baker, Journ. Ind. Eng. Chem., I. 464, 1909. 5 C. Krauch gives the regulation tests for impurities in reagents in his Die Priifung der chemischen Reagentien aufReinheit, Berlin, 1896 (J. A. Williamson and L. W. Dupre's transla- tion, New York, 1902) ; E. Merck, Prufung der chemischen Reagenzien auf Reinheit, Darmstadt, 1912 ; or E. White, Analytical Reagents, London, 1911. I am here reminded that one of the county councillors, with the idea of "testing the analyst," introduced a "poison" into some beer and sent it to the county analyst to report on adulterations (Chem. News, 54. 254, 1886). The analyst reported the sample to be "genuine unadulterated beer." The analyst was right. The poison was not an adulteration. Had the analyst been asked to test for every conceivable admixture, the "bill of costs" would have startled the councillor. I mention this because it illustrates what grotesque notions are rife as to the functions of an analyst working for commercial purposes at a specified (minimum) fee. 141 142 A TREATISE ON CHEMICAL ANALYSIS. in quantitative analysis particularly, the reagents are best tested by blank or dummy analyses. 1 The blank analysis does not necessarily give absolute cer- tainty, because the reactions without the substance are not necessarily the same as with the substance. In most cases, however, the blank analysis is an excellent and reliable method of testing the purity of reagents, and it frequently furnishes a means of correcting the results of the analysis proper. 73. Bottles for Reagents. In a laboratory where a number of workers use the same bottles, particular care must be exercised. Errors might easily creep into the work from an inter- change of stoppers, placing stoppers on the working benches, mistakes in filling and weighing out the reagents, etc. The necks and mouths of all the bottles should be kept clean. The bottles and the stoppers should be numbered so that no confusion is possible. The "mushroom" stopper (fig. 84) is not sufficient to protect the mouth and lip of the bottle from dust, and when these flat stoppers are " stuck " they are unusually difficult to loosen. 2 F. F. Jewitt has designed FIG. 84. FIG. 85. FIG. 86. FIG. 87. a stopper with a pendant flange (fig. 85), with the idea of protecting the mouth and lip of the bottle from dust. These bottles are at present made in two sizes 125 c.c. and 250 c.c., with enamelled labels. The 250-c.c. bottles cost 1 6s. 6d. per dozen. These stoppers are also very difficult to loosen when stuck. 1 W. A. Dixon, Chem. News, 55. 228, 1887 ; ib. t 78. 294, 1898. 2 LOOSENING FIXED GLASS STOPPERS AND STOPCOCKS. Usually the stoppers are not ground to fit the bottles well enough to prevent the slow evaporation of volatile liquids like ether, chloroform, etc. The better the stopper is ground into the bottle, the more it is liable to " stick." When a stopper is fixed, it can sometimes be loosened by tapping with the handle of a pocket-knife, followed by a twisting wrench with the fingers or a pair of pliers with a cloth between the pliers and the glass. Too powerful a wrench will break the stopper. If the bottle still resists, warm the neck with a cloth wet with hot water, or better, slowly warm the bottle by turning the neck rapidly in a Bunsen's flame at intervals of about a minute. The tempera- ture of the neck is thus slowly raised. The stopper can be given a tap and wrench immediately after the neck has passed through the flame. Too great or too sudden heating will fracture the bottle. An obstinate stopper can sometimes be removed by leaving a little oil or glycerine above round the neck of the bottle, say overnight. A tap and wrench may then loosen the stopper. If the stopper be still fast, invert the bottle in a vessel of water so that the water reaches to the shoulder of the bottle. Let the bottle be inclined so as to prevent a bubble of air being entrapped in the gutter between the stopper and the bottle. This would prevent the water gradually working its way between the stopper and the bottle. The stopper may be tried with the tap and wrench after standing one, two, three, or four days. If it still resists, warm the water. As the water is warming, the bottle may be removed every now and again, to find if the stopper will yield to the tap and wrench. Time and patience are generally effective. Sometimes the bottle is wanted at once, and the tapping and wrench, or the heating over the Bunsen's burner, are continued until either the stopper loosens or the bottle breaks. The means taken to prevent accidents will naturally depend on the contents of the bottle. The naked flame would not be tried with a bottle of ether, nor would such a bottle be heated very much. THE REAGENTS. 143 Swarts l has designed a glass stopper (fig. 86) which has the advantage of not stick- ing with solutions of caustic alkali, etc. The cone of the stopper fits the neck of the bottle loosely, but the flat part of the stopper on the under side, as well as the corresponding neck of the bottle, are polished flat, so that a close joint is obtained. These bottles cost about 25 per cent, more than ordinary stoppered bottles. I prefer to keep the regular reagents in 250- and 500-c.c. resistance or Jena glass bottles with stoppers shaped as indicated in fig. 87, and with a rubber or loose glass cap placed over the neck as illustrated in the diagram. The 250-c.c. bottles (10s. 9d., unlabelled, per dozen) are recommended when only small quantities of solution are needed ; the 500-c.c. bottles cost 13s. 6d. per dozen. The caps cost Is. per dozen. When the bottles are in use, the stoppers can either be placed on a watch-glass or similar surface, or held between the fingers, or gripped in the palm of the hand, in such a manner as not to interfere with the using of the bottle for pouring, etc. For standard solutions 1000- or 2000- c.c. bottles are recommended (16s. 3d. and 23s. per dozen). Winchester bottles are usually excellent for this purpose, and are much cheaper. The acid bottles, if convenient, are best kept on sheets of glass or rubber which rest on wooden shelves. 2 The labels of the acid bottles should be sand- blasted (4s. per dozen) or enamelled (6s. per dozen). Books (Is.) of printed labels of the more common reagents are sold by the dealers. Special paper labels for solutions, made up according to the equivalent system described below, are available where enamelled or sand-blasted labels are not used. Blank labels are sold in packets and gummed ready for use. 3 The proper designation should be written in Indian ink. When the ink is dry, the label is gummed to the bottle and then coated with size. 4 When the size is dry, the whole is varnished 5 one or two coats. Solids and liquids are usually supplied in corked or stoppered bottles. Their labels should be sized and varnished, as indicated above, before the bottles are shelved. Otherwise a bottle may be found with a label missing at an inconvenient time. Unprotected labels have a way of dropping off, and the writing is liable to fade in humid climates. When an analysis qualitative or quantitative is in progress, the beakers, basins, flasks, and funnels should be labelled without delay. This will prevent confusing filtrates, precipitates, and samples with one another. A circular or oval patch can be sand-blasted on glass beakers, etc., so that they can be easily marked with a lead pencil. Faber manufactures a pencil (No. 2251) which can be used for writing (blue) on glass and porcelain. 6 It writes best if the surface is clean and warm. The writing can be removed by wiping with a damp cloth. 1 T. Swarts, Chem. Ztg. t 14. 836, 1890; A. Gawalovski, RundsJiau, 1131, 1890; Oel- Fett- Industrie, 9. 114, 1892. 2 Painting the side of the lip of the bottle not the top of the bottle with melted paraffin prevents solutions trickling down the outside, and incidentally facilitates the delivery of the reagent in drops Chem. News, 65. 179, 1892. 3 CEMENT FOR FIXING PAPER TO WOOD, GLASS, OR METAL. Either of the following mixtures may be used : (1) Dissolve 40 grms. of gum arable and 10 grms. of gum tragacanth in 200 c.c. of water. Strain the solution through linen, and add 40 grms. of glycerine. Make the solution up to 400 c.c., and stir in 30 drops of oil of cloves to prevent putrefaction. (2) Rub 40 grms. of rye flour with 5 grms. of alum to a paste with 80 c.c. of water. Mix this with 200 c.c. of boiling water, and heat the solution until the paste is stiff when cold. Stir in 10 grms. of glycerine and 30 drops of oil of cloves. 4 PAPER SIZE. This can be made from isinglass or gelatine, or purchased from the dealers. 5 PAPER VARNISH. Quick-drying "paper" or "copal" varnish, from dealers in artists' materials, gives good results the copal varnish for preference. If the size or varnish be unsuitable, or if the label has been imperfectly sized, unsightly greasy-looking patches will appear on the label. The first coat of varnish must be " set hard " before the second is applied. 6 "Vitro ink" is sometimes useful for writing on glass. The writing will resist damp atmospheres and concentrated acids, but not boiling water or alkalies. 144 A TREATISE ON CHEMICAL ANALYSIS. 74. The Action of Reagents on Glass and Porcelain. In his study of the alleged transformation of water into earth, Lavoisier l (1770) showed that water dissolves glass vessels. This subject is of great importance in analytical chemistry. Errors may arise owing to the action of the acid on the glaze of the porcelain basins during the silica evaporation, etc., and some glazes are more readily attacked than others. Even the best of dishes are much eroded after they have been in use some time. Some of the poorer types of porcelain may lose 0'005 grm. in weight after a four-hours' evaporation of the hydrochloric acid solution, while the better types will not lose O0005 grm. by a similar treatment. The former number means that in a clay analysis with two evaporations, in that type of porcelain, the silica precipitate may be augmented 0*006 grm., the alumina precipitate 0*002 grm., and the lime pre- cipitate 0*0003 grm. This matter therefore requires some attention. Of course the difficulty is overcome by evaporating in platinum basins, but that involves an outlay of over .20. It is therefore necessary under ordinary circumstances to use porcelain glazed with a resistant glaze. The action of different solutions on glass is far more vigorous than on porcelain, but some types of glass resist better than porcelain with an unsuitable glaze. It is therefore bad, on principle, to allow filtrates, etc., to lie any length of time in porcelain or glass vessels particularly the latter. In some cases the con- tamination from the glass may be serious. Thus, if glass contains arsenic 2 (page 285), an error may arise in toxicological work with serious consequences. A great many investigations have been made on the action of water and various solutions on glass and porcelain vessels. 3 The following may be taken to represent the amount in milligrams dissolved by 100 c.c. of the given reagents in four hours : Table XXIII. Action of Reagents on Porcelain and Glass. Water HC1 H 2 S0 4 KOH Na 2 C0 3 (20 per cent.). (24 per cent.). (1 per cent.). (6 per cent.). Berlin porcelain evap. basin . 0-06 5-6 3-3 German porcelain evap. basin 1 1 '8 0-4 0*6 1 1 9-2 Bohemian glass flask 1-8 o-i 0-6 1-2 175 Common glass flask 1-4 27 8'0 5-3 4'8 From Cowper's work it appears that solutions of ammonium chloride and of ammonium sulphide usually attack glass rather more vigorously than sodium carbonate. It is important to bear these facts in mind when precipitations are made in ammonium sulphide solutions and left for some time, as is often done, to ensure the complete precipitation of the sulphides 1 A. L. Lavoisier, (Euvres, Paris, 2. 1, 1864. 2 W. Fresenius, Zeit. anal. Chem., 22. 397, 1883 ; S. R. Scholes, Journ. Ind. Enq. Chem 4. 16, 1912. 3 R. Fresenius, Anleitung zur quantitativen chemischen Analyse, Braunschweig, 2. 767, 1905 ; Eng. trans., London, 2. 626, 1900 ; A. Emmerling, Liebigs Ann., 150. 257, 1869; R. Weber and E. Sauer, Zeit. angew. Chem., 4. 662, 1891 ; Ber., 2$. 70, 1814, 1892 ; F. Foerster, ib.,22. 1092, 1889 ; 25. 2494, 1892; F. Mylius and F. Foerster, ib., 22. 1092, 1889; 24. 1482, 1891; Zeit. anal. Chem., 31. 141, 1892; F. Foerster, ib., 33. 299, 322, 1894; C. Bunge, ib., 52. 15, 1913 ; P. Truchot, Compt. Rend., 78. 1022, 1874 ; R. Cowper, Journ. Chem. Soc., 41. 254, 1882. For a lengthy discussion on this subject, see H. Hovestadt, Jena Glass, London, 319, 1902. For toughened glass, see R. J. Friswell, Chem. News, 52. 5, 1885 ; J. S. Stas, ib.. 17. 1, 1868 ; F. Siemens, Journ. Soc. Arts. 33. 386, 1885. THE REAGENTS. 145 Mylius and Foerster give the following data for the action of water, 2N-NaOH and N-H 2 S0 4 solutions at 100 for six hours., and a 2N"-Na 2 C0 3 solution at 100 for three hours. The numbers represent milligrams per square decimetre. Table XXIV. Action of Reagents on the Best Types of Glass Apparatus. Beakers. Flasks. Type of Glass. Water. TT o/-\ \r_r\TT "Ma on Water. 20. 80. i 2 bU 4 . JN a 2 UU 3 . 20. 80. H 2 S0 4 . NaOH. NaaGOg. "R" . 0-0054 00144 41 23 0-0128 0-0128 51 26 Jena 0-0071 0-0035 53 19 0-0063 0-0057 63 24 Bohemian 0-118 0-219 5 37 49 0-093 0-255 11 52 70 We are thus in a position to form some idea of what is taking place when nitrates, etc., are allowed to stand any length of time in porcelain or glass vessels. To illustrate the possible constituents to be found in the product of the reaction between the reagents and different types of glass, the following analyses are quoted from Walker's investigation ] on chemical glassware : Table XXV. Composition of Chemical Glassware. A1 2 3 Trade Name. Si0 2 . and MnO. ZnO. MgO. CaO. K 2 0. Na 2 0. B 2 3 . As a 3 . Fe 2 3 . J. Kavalier's Bohemian 76-02 0-64 tr. 0-30 7-38 770 7'60 Resistance " R" 68-00 2-32 014 2-40 5-04 4-80 1-82 10-17 5 V 53 0'24 "Wiener Normal Gerathe 74-00 0-66 o-oi 0-24 016 776 5. -51 9-69 2-15 ... Glas" Thiiringen 74-36 0-90 tr. ... 0'16 9-40 0-14 14-83 ... "Schott & Gen. , Jena" . 6674 277 0-65 8-28 4-50 0-28 0-08 8'99 7-18 ... "Nonsol W.T. Co." 65-04 3-78 0'04 8-88 1-44 1-75 08 1272 6-23 Bohemian 70-80 1-00 1-04 ... 0-08 7-88 7-67 8-59 ... ... The glaze of porcelain basins contains silica, alumina, alkalies, and alkaline earths. The glaze is of a felspathic type, with or without lime. In work with silicates, therefore, the best type -of glass to use is obvious. It is unfortunate that the impurities introduced from the glass and porcelain vessels are those very constituents which have to be specially determined. A few rules may now be indicated : (1) If possible, do not allow the solutions to stand in glass or porcelain vessels for any length of time, particularly in glass. (2) If the work be interrupted, so that solutions must stand over, if there be no other objections, acidify the alkaline solutions before they are placed on one side. (3) If solutions must stand over, it is better to use good porcelain than glass vessels. (4) One of the three types of glass indicated in Table XXIV., or some other resisting glass, should be employed for general work in place of the more soluble types of glass. (5) For exact work, a blank experiment should be made, using nothing but the regulation reagents, in similar quantities, and under R. H. Walker, Journ. Amer. Chem. Soc., 27. 865, 1905. IO 146 A TREATISE ON CHEMICAL ANALYSIS. similar conditions to those actually employed in the analysis proper. This plan not only corrects impurities in the reagents, but it also enables us to practically eliminate the sources of error now under discussion. This has been done in the example given on page 185. (6) Precipitations made in alkaline solutions which have stood some time in glass vessels are almost certain to be contamin- ated with silica, and in exact work a correction must be made e.g., cobalt and nickel precipitations, page 391. (7) Remember that solutions of reagents are usually kept in glass vessels. Such solutions may accordingly be contaminated with silica, etc. See, for instance, the determination of phosphorus by the colorimetric process, page 603. l 75. The Equivalent System of making Solutions. There are many advantages in making the concentrations of the liquids and solutions used in analytical work follow 'a definite system. Some make the solutions, as nearly as possible, on the " 1, 2, 5, 10, 20, . . . parts of reagent per 100 parts of water" system. Wollny 2 recommends using multiples or sub- multiples of the equivalent weight (page 45) employed in volumetric analysis, but the solutions need not be made up with such a degree of accuracy as obtains in volumetric analysis. Hence, the amount of reagent in a given volume of solution is approximately known, and equal volumes of the reagents bear a fixed relation one to another. For instance, barium chloride (molecular weight 244) will have 122 grins, per litre. This is called an " E " solution ; concentrated hydro- chloric acid (molecular weight 36 *5), of specific gravity 1'1565, has 365 grms. per litre, and is therefore a " 10E" solution ; potassium chromate (molecular weight 194-5) has 486 grms. per litre, and is therefore a " 5E " solution. The same system was suggested by Reddrop in 1890. Reddrop (1890) recommended the designation " E," and Wollny (1885) the German equivalent " Aeq." The differ- ence between an "E" solution and a "N" solution rests on the fact that a " normal " solution is supposed to be exact ; the " equivalent " solution is approxi- mate, and "rounded" atomic weights may be used in the calculations. With this system, the amount of reagent required for a specific purpose can be readily calculated, and the addition of a great excess avoided. This means a saving in time, labour, and material ; and it indirectly leads to more accurate and reliable work. Thus, 1 c.c. of an" E" solution of sulphuric acid or any sulphate will very nearly suffice for the quantitative precipitation of the barium from 1 c.c. of an "E" solution of barium chloride. Again, suppose that 1 grm. of clay be fused with 7 '5 grms. of sodium carbonate (molecular weight 106), and the cooled mass taken up with water and an excess of hydrochloric acid. Since 7 '5 grms. of sodium carbonate correspond with a little less than one-seventh of an equivalent (53), the addition of 14'3 c.c. of 10 E-HC1 will suffice to neutralise the sodium carbonate. Hence, say, 15 c.c. of the acid will be an excess. If the acid be evaporated down to dryness twice, in the regular manner, and taken up the second time with 2 c.c. of the 10E-HC1, it will be obvious that 2 c.c. of 10E-NH 3 will suffice for the neutralisation of the acid, and 3 c.c. will be an excess of ammonia. This example shows how easily we can learn just how we stand with the quantities of the reagents present in our solutions. Measuring cylinders are useful for delivering definite volumes of the "E "solutions. 1 Ammonia and many other solutions are best kept in cerasine bottles or bottles lined with cerasine (cerasine is a white wax prepared from ozokerite) 0. Schreiner and G. H. Failyer, Bull. U.S. Dept. Agric. (Soils], 31. 20, 1906. 2 R. Wollny, Zeit. anal. Chem., 24. 402, 1885; R. Blochmann, Ber., 23. 31, 1890; J. Reddrop, Chem. News, 61. 245, 256, 1890. THE REAGENTS. Wollny recommends the use of pipette bottles (fig. 30), l by means of which definite volumes can be quickly measured. These bottles, however, are rather expensive (4s. 6d each), though they are handy for a few of the more common reagents. The strengths of the acids and ammonia are easily obtained by means of the hydrometer and the tables in the Appendix. For example, take sulphuric acid. From Table LXX1V. and the hydrometer we find that an acid of specific gravity 1-84 has 1759 grms. of H 2 S0 4 per litre. Taking 49 as the equivalent of H 2 S0 4 , it follows that this acid is T *y of 1759 = 36E strength. Obviously, 1 : 1 acid requires 500 c.c. of the concentrated acid per litre. The litre of the cold solution thus corresponds with 18E acid. If 5E acid be needed, the solution must contain 5 x 49 = 245 grms. of H 2 S0 4 per litre. If 1759 grms. of H 2 S0 4 correspond with a litre of the stock acid, 245 grms. of H 2 S0 4 will correspond with 139*3 c.c. of the concentrated acid. This must be made up to a litre. In further illustration : Table XXVI. Strengths of Stock Acids and Ammonia. Sulphuric acid, sp. gr. 1-84 = 36E. Hydrochloric acid, sp. gr. 1*16 = 10E. Nitric acid, sp. gr. 1*50 = 22 4E. Ammonium hydroxide, sp. gr. 0'88 = 18E. E. c.c. acid per litre. E. c.c. acid per litre. E. c.c. acid per litre. E. c.c. per litre. 20 555-6 10 277-8 io 1000 10 446-5 10 555-6 5 139-3 5 500 5 223-2 5 277-8 1 28-0 1 100 1 447 1 55 6 The concentrations of the reagents used in this work are indicated in the footnotes where the reagents are first mentioned. These can easily be located by reference to the Index. 76. Gas Generators. The arrangements for the supply of hydrogen sulphide in a laboratory naturally depend upon the quantity of gas to be used at one time ; the frequency with which the gas is wanted, etc. An enormous number of generators have been devised with the object of giving a continuous stream of gas for inter- mittent periods. Many of the artifices designed for supplying the hydrogen sulphide partake of the nature of toys. They are not suited for serious work. Preparation of Small Quantities of Hydrogen Sulphide. The old-fashioned " Kipp," still much in vogue (fig. 153 2 ), has many advantages and many dis- advantages. The parts are not replaceable, so that if one part be fractured, as not infrequently happens while recharging, a new " Kipp" must be purchased. The acid is not well agitated, so that when that portion of the acid which comes in contact with the ferrous sulphide has lost its activity, the reaction becomes sluggish, and that frequently at a most inconvenient time. The " Kipp " must then be recharged, even though the acid be not exhausted. Several modifica- tions of the original " Kipp " have been devised to rectify some of these faults. Wartha's (fig. 138) and FriswelPs (fig. 122) modifications are improvements, and either of these forms will be found preferable to the ordinary "Kipp." 1 R. Wollny, I.e. ; Warmbrunn und Quilitz, Chcm. Ztg., 17. 454, 1893. 2 The surface of the glass "Kipp," photographed in fig. 153, is much corroded by hydro- fluoric acid. 148 A TREATISE ON CHEMICAL ANALYSIS. If a great amount of gas be not required I recommend one of these forms. 1 Their advantages over the plain " Kipp " are : (1) the freshest acid is brought in contact with the solid (zinc, ferrous sulphide, marble, etc.) ; (2) the emptying and refilling is simple ; and (3) a great over-pressure is avoided.' 2 Preparation of Large Quantities of Hydrogen Sulphide. I prefer an apparatus of the type of de Koninck's 3 when large quantities of hydrogen sulphide are wanted at frequent intervals. L. L. de Koninck's apparatus (modified) is shown in fig. 88. The bottle A (1 to 2 litres) has two tubes on opposite sides near the base. This bottle is filled with glass marbles to a height of about 2 cm. above the side tubes. The remainder of the bottle is filled with ferrous sulphide (lumps or preferably sticks). By means of glass tubing, rubber joints, and perforated rubber stoppers, the side tubes are connected with the two side necks of a three-necked WoulfF s bottle B (about 4 litres). One tube passes to the bottom of the bottle, the other passes just through the stopper. This latter tube has a side branch by means of which it is con- nected, by rubber tubing, with a side tube blown on to a bottle (7(1 to 2 litres). The bottle A is connected with a Muencke's 4 gas washer D containing water. A three- way Greiner-Friedrichs' stopcock 5 is sealed on to the gas washer, and connected, by means of a rubber joint, with a tube passing through a rubber stopper in one of the necks of a three-necked Woulff's bottle F (2 to 4 litres), which has a tubulure fitted with a stopcock J ground and clamped to the side neck of the tubulure. The middle neck of the Woulff's bottle is fitted with a stoppered funnel /, which is convenient when the bottle is to be filled with water. The other neck of the bottle is fitted with a trap G. 6 The tube H leading from the trap passes to the stink closet, where it is fitted with two to four stopcocks which supply the gas to test tubes 7 FIG. 88. Hydrogen sulphide apparatus. 1 V. Wartha, Ber., 151, 1872; Zeit. anal. Chem. II. 430, 1872; J. Meister, ib., 25. 373, 1886 ; R. J. Friswell, Chem. News, go. 154, 1904 ; 94. 106, 1906 ; G. Thiele, Chem. Ztg., 25. 468, 1901 ; C. Arnold, ib., 26. 229, 1902. 2 The drying and washing bottles can be fixed to the apparatus itself by suitable clamps H. Reckleben, Chem. Ztg., 35. 279, 1911. 3 L. L. de Koninck, Chem. Ztg., 17. 1099, 1893 ; Journ. Amer. Chem. Soc., 16. 63, 1894 ; F. M. Perkin, Journ. Chem. Soc. 2nd., 20. 438, 1901. Perkin added the bottle for storing " H 2 S water " and also improved the joints. 4 R. Muencke, Zeit. anal. Chem., 15. 62, 1876. 5 E. Greiner and Friedrichs, Zeit. anal. Chem., 26. 50, 1887. 6 To prevent contamination by sucking back if, by accident, H, etc., be opened before E. 7 Blocks of wood, 5 cm. thick, with a hole 4 cm . deep and 2 cm. diam. , are useful for support- ing test tubes while their contents are being subjected to the action of the gas. There are several ways of delivering the gas in the fume chamber C. L. Parson, Journ. Amer. Chem. Soc., 25. 231, 1903. THE REAGENTS. 149 or flasks. The free way to the tap E is convenient if a current of gas is to be led elsewhere, say, to an apparatus on a working bench. All the rubber connections should be wired on to prevent leakage. The whole apparatus may be mounted on a wooden stand as shown in the diagram. The bottles should be prevented moving from their position by screwing three pieces of wood to their part of the wooden base. The middle neck of B has a perforated stopper fitted with a stoppered tube passing to the bottom of the bottle. To charge the apparatus : open all the stopcocks, pour dilute hydrochloric acid l into C. When the acid has risen to L, close this cock. When the acid commences to rise in A, close E, and the other stopcocks. Pour the remainder of the acid into C until C is between half and three-fourths filled. Fill the bottle F with cold, recently boiled distilled water vid the stoppered funnel 7. 2 Connect D with F by means of the stopcock E, and the apparatus will be ready for use in a short time. The absence of all but a trace of air from the atmo- sphere in the bottle F when in use retards the decomposition of the solution in f t and only a trace of sulphur will be deposited in F by the oxidation of the " H 2 S water." To empty the apparatus for cleaning : connect the tube L with a rubber tube, open the stopcock, and the acid will soon be syphoned into a suitable receiver say the sink. To recharge the apparatus : close the stopcock L, connect E with the atmosphere, and proceed as before to pour fresh acid into C. To introduce new ferrous sulphide : remove D by loosening the rubber joint and the rubber stopper. Although this apparatus costs 45s., and a large-size "Kipp" 19s., it is worth the difference. (1) The acid and ferrous sulphide do their work more efficiently and economically ; (2) when the apparatus is not in use it need not stand under pressure, for C can be lifted on to the base, and the cock E closed ; 3 (3) the gas can be obtained under great pressure by using a longer connection for C and lifting C to a higher level ; and (4) each part can be easily replaced, and the cost for breakage is very small. Hydrogen Sulphide free from Arsenic. In special work, toxicological investiga- tions, etc., it is necessary to employ hydrogen sulphide quite free from arsenic. It is then necessary either to purify the gas by a suitable train of wash-bottles, etc., 4 or to employ materials free from arsenic in the gas generator. The former is the more troublesome process ; the latter is usually the most convenient. Various mixtures have been recommended for this purpose. Fresenius 5 used calcium sulphide in a "Kipp" with dilute hydrochloric acid (1 volume acid, sp.gr. 1*12; 1 volume water) ; Winkler 6 recommends barium sulphide; others, 7 1 Concentrated hydrochloric acid, 2 volumes, water 3 volumes. With bottles the sizes mentioned above, a Winchester of acid mixed with 1| Winchesters of water will suffice for a charge. Sulphuric acid is objectionable because the ferrous sulphate is liable to crystallise and choke the tubes. 2 Covered by a lid to keep out dust. 3 Although E is usually left connecting A with F t and as shown in the diagram, E must always be closed before C is lifted down, and C must be lifted, as shown in the diagram, before E is opened. Otherwise, water will flow from F to A. C can be corked when not in use, or it can be fitted with a perforated rubber stopper and tube leading into the stink closet or outside the building. 4 0. von der Pfordten (Ber., 17. 2897, 1884) purifies the gas from ferrous sulphide and hydrochloric acid by passing it over a layer of potassium monosulphide at 300-350. 5 R. Fresenius, Zeit. anal. Chem., 26. 339, 1887. 6 C. Winkler, Zeit. anal. Chem., 27. 26, 1887. 7 R. Otto and W. Reuss, Archiv Pharm. (3), 20. 919, 1884; Ber., 16. 2947, 1883; H. Klosmann, Ber., 17. 209, 1884 ; B. Kosmann, Chem. Ztg., 8. 138, 1884 ; W. Lenz, Zeit. anal. Chem.. 22. 393, 1883 ; H. Hager, Pharm. Centr. (2), 5. 212, 1860. A TREATISE ON CHEMICAL ANALYSIS. mixtures of these two sulphides ; Divers and Schimidzu l recommend magnesium sulphide ; Hampe 2 uses sodium sulphide with dilute sulphuric acid (1 : 10) ; etc. 3 The sulphides here named are relatively easy to prepare free from arsenic. Washing Carbon Dioxide. See page 192. Washing Hydrogen Gas. One way of cleaning hydrogen is indicated on page 390 ; another, more thorough process, is to attach the following train to the auto- matic generator : (1) A saturated solution of mercuric chloride to remove hydro- carbons ; (2) potassium hydroxide bulbs ; (3) silver nitrate solution ; (4) pieces of solid potassium hydroxide; (5) granulated calcium chloride. See also page 391. Washing Air. The air is forced or drawn through the following train : (1) Concentrated sulphuric acid; (2) concentrated potassium hydroxide; (3) con- centrated sulphuric acid. 77. Distilled Water. The water used in quantitative analysis should be distilled from a tin-lined copper still connected with a tin-lined copper or a block-tin condenser. 4 The still is usually fitted with an automatic filler, so that the overflow water from the water jacket of the condenser can run into the still. When a supply of steam is available, the still can be replaced by a tin-lined copper cylinder. The cylinder is fitted with a couple of perforated shelves supporting layers of calcined quartz pebbles, which serve to filter the steam and prevent dirt being mechanically carried over to the condenser. In some cases the distilled water is obtained by affixing the condenser to the steam apparatus used for drying, heating, etc. 5 All depends upon the local conditions the quantity of distilled water required per diem, etc. Stills of the type illustrated in fig. 89 are compact and convenient where but small quantities are needed. The still can be fixed to a wall in any convenient place. These stills are made of different sizes capacity from 2 to 10 litres per hour ; price from 50s. If the water is to be freed from carbon dioxide and ammonium carbonate, the first portions which pass from the still (not an automatic still) are re- FIG. 89. Jewell's water still, jected, and the water should be boiled some time to drive off the adsorbed gases air, carbon dioxide, etc. If the water is to be free from organic matter, it can be distilled from an alkaline solution of potassium permanganate. This, however, does not destroy 1 E. Divers and T. Schimidzu, Journ. Chem. Soc., 45. 270, 1884. 2 W. Hampe, Chem. Ztg., 14. 1777, 1890 ; F. Gerhard, Archiv Pharm. (3), 23. 384, 1885. * J. Habermann (Verhandl. naturf. Ver. Brunn, 17. 14, 1891) heats a mixture of one part calcium sulphide with two parts of magnesium chloride and enough water to make the whole into a thin slurry. 4 Distilled water is almost without action on block tin. Water condensed in glass tubes is objectionable because the water dissolves appreciable quantities of glass. This can be demon- strated (a) by evaporating such water to dry ness in a platinum dish ; and (6) by the fact that the glass tubes soon become deeply corroded. 5 L. W. Hoffmann and R. W. Hochstetter. Journ. Amer. Chem. Soc., 17. 122, 1895; C. E. Wait, ib. t 17. 917, 1895 ; H. M. Hall, Journ. Anal. App. Chem., 6. 190, 1892 ; R. Seligman, Chem. News, 93. 27, 1906. THE REAGENTS. 151 all the ammonia. To remove the ammonia, Stas 1 distilled the water from aluminium sulphate. Receiver for Distilled Water. The receiver for the distilled water depends upon the amount of water required per diem, etc. Owing to the fact that distilled water attacks glass, the water should not be kept any great length of time in glass vessels. Tanks linked and soldered with block tin are best for laboratories where inorganic gravimetric determinations are made, since the water is then free from silica, etc., which lead to high results. The tank can be fitted with a glass syphon gauge. Some use a couple of glass bottles with glass stopcocks at the side. One bottle is filled in the morning and used for the next day's supply. Each bottle is emptied the night after it has been filled with distilled water. The distilled water should be tested every now and again by evaporation in a weighed platinum dish to make sure that it is free from contaminations. 1 J. S. Stas, Chem. News, 4. 207, 1861 ; 15. 204, 1867 ; Zeit. anal. Chem., 6. 417, 1867 ; (Euvres Completes, Bruxelles, i. 101, 536, 1894. PART II. TYPICAL SILICATE ANALYSES-CLAYS. CHAPTER X THE DETERMINATION OF VOLATILE MATTERS. 78. Hygroscopic Moisture. THE term "hygroscopic moisture" generally includes all the volatile con- stituents, principally moisture, which are given off when the substance is heated to a certain standard temperature, say, 110 . 1 The heating is continued until the substance ceases to lose weight. The ground 2 sample is placed in a stoppered weighing bottle of known weight. The bottle and contents are then weighed, and placed in an air bath 3 at the desired temperature. The stopper is removed, and a piece of filter paper is placed over the mouth of the weighing bottle to keep out the dust. The bottle and contents can be left in the oven overnight ; the stopper of the weighing bottle is placed in the desic- cator with the bottle; and, when cold, the stopper is inserted and the whole weighed. About four hours usually suffices for the drying, but, after weighing, the bottle is then returned to the air bath for an hour, and, when cold, re-weighed. If no further loss of weight occurs, the hygroscopic moisture of the clay in the bottle is represented by the total loss of weight which the clay suffers when dried at 109-110. The finely ground powder is dried in another stoppered bottle, in a similar manner, for subsequent analysis. One purpose of removing the hygroscopic moisture before the 'analysis is to secure a uniform hygroscopic condition as a basis for comparing the analytical data. 4 If an analysis were made on two samples of one clay, one sample with 5 per cent, less hygroscopic moisture than the other, the silica in one might appear, 1 Some heat the clay to 100, but this is scarcely high enough ; others use 105, 120, etc. 2 Most finely ground powders absorb moisture so tenaciously that some is retained even after the powder has been heated to 110 an indefinitely long time. In fact, R. Bunsen (IVied. Ann., 24. 327, 1885 ; A. L. Day and E. T. Allen, Amer. J. Science (4), 19. 93, 1905 ; J. T. Bottomley, Chem. News, 51. 85, 1885) has shown that a red heat is required for the expulsion of all the adsorbed water. 3 Either a toluene bath, fig. 90, or a bath fitted with a thermostat, fig. 116 both are shown with the door open. Instead of toluene, a mixture of water and glycerol may be used : 6 parts of water to 1 of glycerol boils nearly at 105 ; and a mixture of 1 part of water to 6 parts of glycerol, at 120. Intermediate temperatures can be obtained by the use of intermediate mixtures. If the proportion of glycerol be increased much beyond the limit here mentioned, acrid fumes are given off E. J. Reynolds, Chem. News, 4. 319, 1861. An aqueous solution of calcium chloride, which boils at the desired temperature, can also be used, e.g., 100 grrns. of water with 25 grms. of calcium chloride boils at 105 ; with 41 '5 gi-ms. calcium chloride, at 110 ; with 69 grms. of calcium chloride, at 120 G. T. Gerlach, Zelt. anal. Chem., 26. 413, 1887. I prefer toluene. Concentrated solutions of calcium chloride sometimes cause trouble by attacking the joints of the bath, and require more attention owing to the gradual loss of water. To prevent any escape of inflammable toluene vapours into the room should the water of the condenser be turned off while the. bath is going, say over- night, I have a tube (a, fig. 90) leading from the top of the condenser to the outside of the laboratory. 4 See Z. A. James, Eng. Min. Journ., go. 1047, 1910. 155 156 A TREATISE ON CHEMICAL ANALYSIS. in the statement of the analysis, as 60 per cent., and in the other as 57 per cent. This all in the same clay. Hence, the analytical results will appear different according as the sample is collected in moist or in dry weather. It is generally advisable, with commercial materials, to determine the amount of " hygroscopic moisture lost at 100 " or " at 110." l For clays, it is best to use 109-110. Some materials would be decomposed at this temperature. For FIG. 90. Toluene bath at 109- 110. example, superphosphates are liable to decompose at 110. This is illustrated by the following determinations of the weights of acid calcium phosphate 2 CaH 4 (P0 4 ) 2 when heated one hour at the temperatures stated : Temperature . Decomposition 100 105 2-0 110 4-3 120 C 4-9 130 140 8'4 150 13-3 160 177 200 50 "4 per cent. If the samples for analysis are received in paper or in canvas bags, it is not much use determining the hygroscopic moisture, because the result may be greater or less than when the sample was originally packed for the analyst. If the hygroscopic moisture is to be determined, the sample should be hermetically sealed in a vessel which will not allow an absorption or evolution of moisture. There are some " tricks " in sampling when the moisture question is left open. 3 1 For the moisture retained by many salts dried in a steam oven 97-98 see H. W. Hake, Proc. Chem. Soc., 13. 147, 1897 ; F. W. Smither, Amer. Chem. Journ., 19. 227, 1897. 0-2864 grin, of calcium chloride, CaCl 2 , after the 30th day, weighed 0'3200 grm. ; 0'4349 grm. calcium nitrate, Ca(N0 3 ) 2 , 0-4550 grm. ; 0'1907 grm. magnesium chloride, MgCl 2 , 0'2136 grm. The numbers varied a little according to the hygroscopic condition of the atmosphere. Obviously, therefore, we cannot consider drying in the steam oven,Und possibly air baths, a sufficient means of standardising the hygroscopic condition of all materials. 2 J. Stoklasa, Zeit. anal. Chem., 29. 390, 1890 : H. Birnbaum, Zeit. Chem. (2), 14. 137, 1871. For white lead, see p. 319. 3 Spooner and Bailey, Chem. News, 21. 21, 1870. THE DETERMINATION OF VOLATILE MATTERS. 157 In order to distinguish between the moisture which has been absorbed by a given substance from its surroundings, and that which is given off by the de- composition of hydrated aluminium silicates, salts with water of crystallisation, etc., at 110, the substance is dried, not at 110, in an air bath, but at atmospheric temperatures in a desiccator over concentrated sulphuric acid preferably in vacua. This may occupy days or even weeks. The time can be shortened by drying the powder in a tube, in a current of air, dried by passage through, say, sulphuric acid, followed by phosphorus pentoxide. The loss in weight indicates the amount of moisture lost. 1 The moisture can be reabsorbed in a drying tube charged with phosphorus pentoxide, and the increase in its weight represents the moisture removed from the clay. This gives the hygroscopic moisture less the trace which the clay retains with remarkable tenacity, even at elevated temperatures. The atmosphere in some air baths, steam ovens, etc., is quite humid, and contaminated with sulphurous oxides, etc., arising from the products of the com- bustion of the coal gas beneath the oven. These gases find their way into the oven between the door and the front of the oven. In the diagrams, figs. 90 and 116, these gases are shown deflected up the back of the oven, where they do no harm. A ledge vertically downwards is fixed on the front and on two sides, as shown in the diagrams. There are also certain risks attending the use of ovens lined inside with metal. Dust from the corroded metal may fall on to the materials being dried. It is not difficult to line an oven with glazed porcelain or earthenware slabs, which can be obtained to fit the oven. 2 79. Loss on Ignition. This term is employed to represent the total loss in weight which occurs when the clay is calcined at a bright red heat. The "loss on ignition" may thus include the loss in weight which attends the expulsion of water (page 573) formed during the decomposition of the clay ; carbon dioxide from the carbon- ates (page 553), sulphurous gases from the sulphates, and the oxidation of sulphides ; 3 the combustion of carbon and organic matter (page 546) ; the volatilisa- tion of ammonium compounds, alkalies, etc. The loss in weight on ignition is usually great enough to mask the gain in weight which occurs during the more or less incomplete oxidation of ferrous to ferric and magnetic oxides ; manganese compounds to Mn 3 4 , etc. However, it is not uncommon to find firebricks which gain up to 0'5 or even I'O per cent, on ignition. This is probably due to the oxidation of ferrous and metallic iron. Sometimes these bricks, when powdered and treated with acids dilute hydrochloric acid, for instance give off hydrogen gas, showing the presence of metallic iron. 4 The loss on ignition is not always satisfactory for exact investigations, 5 but for industrial purposes it some- times indicates what is likely to happen when a clay is fired in the ovens or kilns. To determine the loss on ignition, about 1 grm. 6 of the clay is calcined in 1 When this experiment is done carefully, some very curious curves are obtained. Part of the water, usually thought to be "combined water," behaves as if it were "hygroscopic moisture" E. Lb'wenstein, Zeit. anorg. Chem., 63. 69. 1909. Seepage 575. 2 F. P. Tread well, Kurzes Lehrbuch der analytischen Chemie, Leipzig, 2. 21, 1911 ; Eng. trans., 22, 1904. 3 The sulphur may be retained by free lime, etc., if present in the form of sulphates. 4 F. Stolba (Zeit. anal. Chem. , 7. 93, 1868) shows that a correction can be made if the amount of ferrous oxide be determined, because one part of this oxide increases by 0*1111 grm. on passing to the ferric condition. 5 For the direct determination of water, carbon, carbon dioxide, etc., see later chapters. 6 If the clay be not very hygroscopic, exactly 1 grm. may be weighed out. This saves trouble in calculations, but it is necessary to guard against the error mentioned on page 54. 158 A TREATISE ON CHEMICAL ANALYSIS. a weighed platinum crucible recently heated red-hot, and cooled in a desiccator. The crucible, during the earlier stages of the ignition, is inclined 30-45 on a platinum or quartz triangle, as indicated in fig. 91. The flame is so arranged that it does not envelop the whole crucible, but impinges on one side, so that the air can circulate freely about the mouth of the crucible. Some clays, ball clays in particular, must be heated very slowly at first, or fine particles of clay will be carried away from the crucible with currents of gas and steam. A blast is usually needed for the later stages of the ignition. 1 A crucible is shown in FIG. 91. Ignition over a Meker's burner. fig. 92 being- heated over a blast. After the clay has been blasted about 10 minutes, the crucible is allowed to cool until down to a temperature below red heat. It is then lifted into the desiccator by means of a pair of platinum-pointed tongs, allowed to cool, and weighed. Again blast the crucible and contents 5 minutes, cool, and weigh as before. If a further loss in weight takes place, the blasting must be repeated until two successive weighings are approximately constant. Highly carbonaceous clays are sometimes advantageously mixed with a known weight of ignited magnesia to prevent the fusion or agglomeration of the residue during ignition. 2 Errors. Determinations of the loss on ignition of eight portions of one sample of powdered clay, each weighing 1 grm., gave the following numbers : 0-0927, 0-0933, 0-0927, 0-0931, 0-0931, 0-0930, 0-0929, 0-0930 grm., with a 1 Alkalies begin to escape at a red heat. Potassium usually comes off faster than sodium. The alkalies, after a long blasting, may be found partly condensed on the lid. If the blasting belong continued, most of the alkalies will be volatilised in this way. If the ignition tempera- ture be properly adjusted, the loss of sulphur and alkalies is inappreciable. 2 According to A. Gutbier (Chem. Ztg., 34. 211, 1910), W. C. Heraeus has designed a per- forated crucible lid with a partition extending into the crucible. By heating one side of the crucible a circulation of air is induced which is said to facilitate the combustion of organic matter. THE DETERMINATION OF VOLATILE MATTERS. 159 mean of 0'0930 grm., that is, 9'30 per cent., with a maximum and minimum deviation of 0'03. These numbers give us an idea of the differences which might be expected in duplicate determinations. If another analyst made a determination, his result ought not to differ by much more than 0*03 from FIG. 92. Ignition over a blast Meker's burner. the 9 '30 per cent. If it did, one of the results is wrong, or the sampling is faulty, or the "drying temperature" is different from 110. The chief sources of error arise from : (1) Imperfect drying (hygroscopic moisture) ; (2) Oxidation or reduction of ferrous iron, etc. ; (3) Imperfect expul- sion of water owing to a too low temperature of ignition ; 1 (4) More or less imperfect decomposition or volatilisation of some of the constituents fluorides, 2 sulphates, 3 etc. ; and (5) Mechanical loss of the substance by the transport of fine particles along with the steam when the temperature is raised too rapidly. 4 1 Some minerals e.g. talc, steatite, etc. require a high blasting temperature to drive off all the water. T. Sclmrer,' Pogg. Ann., 84. 321, 1851. 2 K. List, Liebig's Ann.',%i. 189, 1852. 3 C. R. Fresenius (Quantitative Chemical Analysis, London, I. 55, 1876) says the loss of sulphuric acid from sulphates can often be guarded against by adding about six times the weight of the substances of finely divided, recently ignited lead monoxide. 4 G. Cesaro, Mem. Soc. Roy. Sciences Lieges (3), 5. 1, 1906. CHAPTER XL OPENING UP SILICATES. 80. Summary of the Different Methods. SOME silicates are easily decomposed by more or less prolonged digestion in acids ; others require a more drastic treatment. Some, apparently insoluble silicates, decompose in acids after a preliminary calcination at 500- 600. A great variety of methods have been proposed and used. 1 Many of these methods are now obsolete. Although the sodium carbonate fusion is generally employed, some of the other methods are invaluable in special cases. A number of methods are indicated below ; others will be described later in connection with glazes and glasses, also chromite and uranium. I. Silicates decomposed by Digestion ivith Mineral Acids. Sulphuric, hydro- chloric, or nitric acid, or mixtures of these acids in open vessels. (a) Natural silicates. Used for the so-called " rational analysis " ; analysis of slags, some lead frits, etc. See pages 460, 657 et seq. (I) After the silicate has been calcined at a dull red heat, below the melting point of the silicate. 2 E.g., tourmaline in hydrofluoric acid, etc. See page 460. (c) After exposure to reducing gases at a dull red heat. See page 269. (d) Digestion with hydrochloric or sulphuric acids in sealed tubes, under pressure. 3 See page 493. II. Silicates decomposed by Hydrofluoric Acid alone or by a Fluoride mixed ivith a Mineral Acid. Sulphuric, hydrochloric, or nitric acid. (a) Vapour at a red heat. 4 (6) Aqueous solution alone or with mineral acids. 5 This process is much used for the determination of ferrous iron and alkalies, etc. (c) Digestion with a fluoride mixed with an acid or some salt which decom- poses the fluoride. 1 For the decomposition of chromite, see page 474 ; and for ferruginous minerals, page 461. a F. Mohr, Zeit. anal. Chem , 7. 293, 1868 ; F. Rocholl, ib., 20. 289, 1881 ; C. Rammelsberg, Pogg. Ann., 80. 457, 1850 ; H. Rose, ib., 108. 1, 1859 ; see page 461, FeO. 3 A. Mitscherlich, Journ. prakt. Chem. (1), 81. 108, I860; 83. 455, 1861; P. Jannasch, Ber., 24. 2734, 3206, 1891 ; Zeit. anorg. Chem., 6. 72, 1894 ; F. C. Phillips, Zeit. anal. Chem., 12. 189, 1873. 4 A. Brunner, Pogg Ann., 44. 134, 1838; F. Kuhlmann, Compt. Rend, 58. 545, 1864; A. H. Allen, Analyst, 21. 87, 1896 ; A. Miiller, Journ. prakt. Chem. (1), 95. 51, 1865. 5 J. J. Berzelius, Pogg. Ann., i. 169, 1824; H. Rose, Liebig's Ann., 72. 324, 1849 ; D. Craig, Chem. News, 60. 227, 1889; E. Linnemann, ib., 52. 220, 233, 240, 1885; N. S. Maskelyne, ib., 21. 27, 1870 ; Proc. Roy. Soc. t 18. 147, 1869 ; F. Hinden, Zeit. anal. Chem., 25. 332, 1906. 1 60 OPENING UP SILICATES. l6l (i) Ammonium fluoride and sulphuric acid. 1 (ii) Potassium hydrogen fluoride. 2 (iii) Calcium fluoride and sulphuric acid. 3 (iv) Sodium fluoride with an acid or with potassium bisulphate. 4 (v) Barium fluoride with an acid or with barium nitrate. 5 III. Silicates decomposed by Fusion with Alkaline Oxides or Salts. This method is not usually applied if the silicate is decomposed by simple treatment with acids, in open vessels. (a) Fusion with alkaline hydroxides. Fused caustic potash and soda ash were among the earliest fluxes in use for decomposing silicates. 6 Pure sodium carbonate and caustic soda are now much used for opening up silicates. 7 (b) Fusion with sodium peroxide. This is much used for opening up chromites hsematites, etc., as described on pages 266, 461, etc. The violence of the action may be tempered by mixing the peroxide with sodium hydroxide or lime. (c) Fusion with alkaline carbonates. (i) Sodium carbonate. 8 (ii) Fusion mixture. 9 (iii) Sodium bicarbonate. 10 (iv) Potassium carbonate. 11 This salt is sometimes preferred to sodium carbonate e.g., with tungstates, niobates, tantalates, etc. on account of the greater solubility of the potassium salt. (d) Potassium or sodium bisulphate or pyrosulphate. 12 Smith prefers the sodium salt in certain cases e.g., in the decomposition of corundum since the double salt of aluminium and potassium is dissolved with greater difficulty under conditions where the sodium salt is freely soluble. 1 L. von Babo, Amt. Ber. deut. Naturforscher. Aertze Mainz, 20. 103, 1842 ; J. Potyka, Vntersuchungen uber einige Mineralun, Berlin, 38, 1859 ; H. Rose, Pogg. Ann., 108. 20, 1859 ; R. Hoffmann, Zeit. anal. Chem., 6. 367, 1867 ; Chem. News, 17. 24, 1868. The ammonium fluoride should leave no residue when evaporated to dryness. P. T. Austen and F. Wilber (Chem. News, 48. 274, 1883) purify by dissolving commercial ammonium fluoride in as little water as possible in a platinum dish (or, better, place hydrofluoric acid in a platinum dish), add an excess of concentrated ammonia slowly from a pipette to avoid loss by spurting. A voluminous precipitate may separate. Decant the clear liquid for use. In silicate decom- positions, acidify the solution with sulphuric acid and evaporate to dryness. 2 C. Marignac, Ann. Chim. Phys. (4), 8. 115, 1866 ; W. Gibbs, Amer. J. Science (2), 37. 357, 1864 ; Chem. News, 10. 37, 49, 1864 ; F. A. Clarke, Zeit. anal. Chem., 7. 463, 1868. 3 C. E. Avery, Chem. News, 19. 270, 1860 ; C. A. Wilbur and W. Whittlesey, ib., 22. 2, 1870. 4 F. W. Clarke, Amer. J. Science (2), 45. 173, 1868 ; (3), 18. 290, 1877 ; Chem. News, 17. 232, 1868 ; C. E. Avery, ib., 19. 270, 1860. 5 G. Gore, Journ. Chem. Soc., 15. 104, 1863 ; C. E. Avery, Chem. News, 19. 270, 1869. 6 M. H. Klaproth, Analytical Essays, London, I. 50, 1801. 7 M. W. lies, Eng. Min. Journey.. 58, 1881 ; Chem,. News, 43. 78, 1881 ; C. A. Burghardt, ib., 61. 260, 1890 ; Proc. Manchester Lit. Phil. Soc., (4), 3. 171, 1890. See pages 266, etc. 8 S. Kern, Chem. News, 35. 203, 1877 ; D. Lindo, ib., 60. 14, 33, 41, 1889 ; L. H. Freid- burg, ib., 62. 22, 32, 1890 ; Journ. Amer. Chem. Soc., 12. 134, 1890 ; W. Knopp, Ber. kb'nig. Sachs. Ges. Wiss., 33, 1882 ; Zeit. anal. Chem., 22, 558, 1883 ; Chem. News, 49, 62, 1884 ; E. Mallard, Ann. Chim. Phys. (4), 28. 86, 1873. F. Stolba (Sitzber. konig. Bohm. Ges. Wiss. t 1885 ; Zeit. anal. Chem., 25. 378, 1886) recommends placing a layer of sodium chloride over the charge in the crucible before the mixture is heated. 9 See page 163. C. F. Chandler, Pogg. Ann., 102. 446, 1857. 10 C. Holthof, Zeit. anal. Chem., 23. 498, 1884 ; Chem. News, 51. 18, 1885. 11 H. Rose, Ausfuhrliches Handbuch der analytischen Chemie, Braunschweig, 2. 630, 1852 ; J. J. Berzelius, Pogg. Ann., 4. 152, 1824 ; W. B. Giles, Chem. News, 99. 25, 1909 ; M. H. Bedford, Journ. Amer. Chem. Soc., 27. 1216, 1905. 12 J. L. Smith, Amer. J. Science (2), 40. 248, 1865; Chem. News, 12. 220, 1865 ; W. H. Worthington, Min. Science, 63. 521, 1911. II 1 62 A TREATISE Ott CHEMICAL ANALYSIS. IV. Calcining the Silicate with an Alkaline Earth. Oxide, carbonate, etc. (a) Calcium carbonate or oxide, alone or with calcium chloride. 1 (b) Calcium carbonate with ammonium chloride. 2 (c) Calcium sulphate. 3 This was used by von Hauer for lepidolite. (d) Barium oxide or carbonate or nitrate. 4 V. Fusion with Lead or Bismuth Compounds. (a) Lead oxide. 5 This method is sometimes used in determining the total water in clay by Jannasch's process. The lead oxide retains fluorine, boron, etc. (b) Lead carbonate. 6 (c) Lead chromate. 7 (d) Fusion with bismuth oxide or nitrate. 8 VI. Fusion with Boron Compounds. (a) Boric oxide, or boric acid. 9 This process promises to be useful for the determination of silica in the presence of fluorides, since, it is stated, the fluorine is evolved as boron fluoride BF 3 not as silicon fluoride SiF 4 . (b) Borax. 10 According to Wells, borax and boric anhydride attack platinum quite appreciably during high-temperature fusions. (c) Potassium borofluoride mixed with potassium carbonate. 11 This mixture was used by Stolba for opening up zircons (4 of flux, 1 of powdered zircon). VII. Electrical Current. Several attempts have been made to decompose certain silicates and minerals while exposed to the joint effect of acids and an electric current, but with success only in special cases. 12 1 F. Glinka, Journ. Russ. Chem. Soc., 24. 456, 1892; L. R. von Fellenberg-Rivier, Ze.it. anal. Chem., 5. 179, 1866; H. Cormimbeuf, Ann. Chim. Anal. 15. 295, 1910; H. St C. Deville, Ann. Chim. Phys. (5), 6l. 309, 1861 ; Chem. News, 4. 255, 1861 ; E. Bonjean, ib., 80. 240, 1899 ; H. Rocholl, ib., 41. 234, 1880. 2 J. L. Smith, Amer. J. Science (2), 50. 269, 1871. 3 F. Stolba, Zeit. anal. Chem., 16. 99, 1877. 4 G. Worthier, Journ. praU. Chem. (1), 91. 321, 1864 ; H. Abich, Pogg. Ann., 50. 128, 341, 1840 ; P. Berthier, ib., 14. 100, 1828 ; J. W. Dobereiner, ib., 14. 100, 1828 ; L. R. Fellenberg- Rivier, Zeit. anal. Chem., 9. 459, 1870; W. Hempel, ib. t 52. 86*, 1913 ; G. Gore, Journ. Chem. Soc., 15. 104, 1862 ; J. J. Berzelius, De V Analyse des Corps Inorganiques, Paris, 72, 1827; M. H. Klaproth, Beitrdge zur Kenntniss der Miner alkorper, Berlin, 3. 240, 1802. 5 P. Berthier, Ann. Chim. Phys. (2), 17. 28, 1821 ; G. Bong, Zeit. anal. Chem. 18. 270, 1878; P. Jannasch and H. J. Locke, Zeit. anorg. Chem., 6. 168, 321, 1894 (for topaz) ; A. Leclerc, Compt. Rend., 125. 893, 1897 ; Chem. News, 77. 27, 1898. 6 P. Jannasch, Ber., 26. 1497, 2909, 1893 ; 27. 2228, 1894 ; Zeit. anorg. Chem. 8. 364, 1893 ; Chem. News, 71. 51, 1895. 7 P. Jannasch, Ber., 22. 221, 1889. 8 W. Hempel and R. F. Koch, Zeit. anal. Chem., 20. 496, 1881 ; Chem. News, 45. 81, 1882 (nitrate); T. M. Chatard, Amer. J. Science (3), 29. 379, 1889; Chem. News, 50. 279, 1884. (oxide). 9 H. Davy, Phil. Trans., 85. 231, 1805 ; P. Jannasch, Ber., 28. 2822, 1896 ; P. Jannasch and 0. Heidenreich, Zeit. anorg. Chem., 12. 208, 219, 1896 ; P. Jannasch and H. A. Weber, Ber. t 32. 1670, 1899 ; H. A. Weber, Ueber die Aufschliessung der Silikate durch Borsdure- anhydrid, Heidelberg, 1900; K. Pfeil, Ueber die Aufschliessung der Silikate und anderer schwer zersetzbar Mineralien mit Borsdureanhydrid, Heidelberg, 1901 ; E. Rupp and F. Lehman, Chem. Ztg., 35. 565, 1911. 10 C. Rammelsberg, Zeit. deut. geol. Ges., 24. 69, 1872; W. Suida, Tschermatfs Mitt. (1), 5. 176, 1876 ; Zeit. anal. Chem. 17. 212, 1878 ; J. S. C. Wells, School Mines Quart., 12. 295, 1891 ; Chem. News, 64. 294, 1891. 11 F. Stolba, Chem. News, 49. 174, 1884. 12 E. F. Smith, Ber., 23. 2276, 1890 ; 130. 152, 1892; L. F. Frankel, Chem. News, 65. 54, 66, 1892; M. Mayen^on, ib. t 76. 24, 1897. OPENING UP SILICATES. 81. Sodium Carbonate for Silicate Fusions. In most silicate analyses, the mineral is broken down by fusion with anhydrous sodium carbonate a plan first used by T. Bergmann in the eigh- teenth century. Obviously, the sodium carbonate must be free from impurities which are going to be determined in the fused silicate. 1 For instance, if sulphur is going to be determined, the carbonate must be free from sulphur compounds. The total impurities silica, alumina, etc. should not exceed 0*01 per cent. If 10 grrns. of such a carbonate be taken for the analysis, the maximum error for the impurities in the sodium carbonate will not then exceed Ol per cent. The impurities can be determined by a blank analysis. Potassium carbonate fuses at about 885, and sodium carbonate at 810, 2 while a mixture of equimolecular proportions fuses at about 690. The fusion curve for all possible mixtures is indicated in fig. 93. The so-called "fusion mixture " (sodium carbonate, 4 parts by weight ; potassium carbonate, 5 parts) /ooo god 8OO 70O O 26 3O 75 /OO% FIG. 93. Melting points of mixtures of sodium and potassium carbonates. is the most fusible mixture of these two carbonates. Hence, this mixture is sometimes recommended for opening up silicates in preference to sodium car- bonate alone, because of its greater fusibility. This point is, however, of little importance, because the temperature of silicate fusions is generally greater than the melting point of either salt alone ; and there is no difficulty in maintaining the necessary temperature. Dittrich 3 has pointed out that potassium salts are washed from the different precipitates, later on, with greater difficulty than sodium salts. In illustration, Reidenbach shows that a "magnesia" precipitate made in a 3N-solution of potassium chloride retains nearly twice the weight of occluded salt as a pre- cipitate made under similar conditions in a 3N-solution of sodium chloride. Still further, Smith has shown that the resulting cake with potassium carbonate dissolves in water with greater difficulty than when sodium carbonate is used. Platinum crucibles are slightly attacked by the fusion with sodium carbonate. Koninck 4 says that 6 grms. of fusion mixture when melted in a platinum crucible 1 J. L. Smith (Chem. News, 30. 234, 1874) on the preparation of pure potassium and sodium carbonates for silicate fusions. Excellent sodium carbonate can now be purchased with a guaranteed composition. 2 H. le Chatelier, Bull. Soc. Chim. (2), 47. 300, 1887. 3 M. Dittrich, Anleitung zur Gesteinanalyse, Leipzig, 5, 1905 ; R. Reidenbach, Ueber die quantitative Bestimmung des Magnesiums als Magnesiumpyrophosphat, Kusel, 69, 1910. 4 L. L. de Koninck, Zeit. anal. Chem., 18. 569, 1879. 164 A TREATISE ON CHEMICAL ANALYSIS. over a Bunsen's burner, and then over a blast, caused the platinum crucible to lose in weight O'OOIO grm. Hence, Koninck says that in exact analyses the platinum should be removed by treatment with hydrogen sulphide in acid solution. Koninck's numbers are higher than those 1 obtained in my own experience with sodium carbonate fusions. There is also no need to remove the platinum until after the potassium pyrosulphate fusion, as indicated on page 186. 82. Opening Clays and Silicates by Fusion with Sodium Carbonate. Charging the Crucible. Either 1 grm. of the powdered clay dried at 110, or the clay remaining in the crucible after the loss on ignition 1 has been determined, is mixed with six to eight times its weight of anhydrous sodium carbonate, by adding the carbonate to the crucible and thoroughly mixing the contents of the crucible by means of a spatula. Take care to get plenty of sodium carbonate at the bottom of the crucible. Two more grams of sodium carbonate are introduced and levelled down with the spatula. The spatula is also cleaned by the sodium carbonate at the same operation. The Fusion. The crucible is placed slightly inclined on a platinum or pipe- clay triangle, so that the flame does not completely envelop the crucible (fig. 94). FIG. 94. Fusion of silicate with sodium carbonate. The object is to keep the atmosphere inside the crucible as oxidising as possible. The lid is placed on the crucible, and the latter is heated over a low flame with the bottom of the crucible at a dull red for a quarter of its length. This drives off the carbon dioxide without spluttering. In about 15 minutes the crucible is heated to bright redness by means of, say, a Teclu's or a Meker's burner for about 15 minutes. The contents of the crucible will then probably be in a state 1 If the clay sinters into clots when the loss on ignition is determined, it is best to work with an uncalcined sample, because sintered masses are decomposed with difficulty by the fused carbonate. OPENING UP SILICATES. 165 of quiet fusion, 1 and gas bubbles will have ceased to come off. Little if any spluttering of the contents of the crucible on to the inside of the lid will have taken place if the operation has been properly performed. Cooling the Crucible. The crucible is then removed from the flame by the platinum-pointed tongs (fig. 63 is the better type), and, while still red-hot, placed under the water tap so that a gentle stream of water flows from the tap on the sides of the crucible ; 2 or else dip the crucible in a dish of cold water. Not a drop of water must be permitted to splash inside the crucible. By giving the contents of the crucible a rotary motion as the contents begin to solidify, it is possible to spread the contents over the sides and the bottom and render the subsequent removal of the cake easier. Let the crucible cool on a slab of granite, marble, or an unglazed tile. A green-coloured cake indicates the presence of manganese sodium manganate but manganese may be present without showing the green colour if the inside of the crucible, during the fusion, was not sufficiently oxidising. The dirty brown colour of highly ferruginous clays might also obscure the manganese green. Transfer of the Cake to the Basin. When cold, the crucible is half filled with water, and a gentle heat applied without allowing the contents to boil. If the crucible be not dented, and has been allowed to cool a sufficient length of time, the cake will quickly loosen, 3 and may be tipped into a 250-c.c. evaporating basin (better a platinum basin). 4 The cake is dissolved in 100-200 c.c. of water, 5 and about 20 c.c. of concentrated hydrochloric acid are poured down the inside of the evaporating basin, a few drops at a time, so as not to lose any of the liquid by effervescence ; or better, cover the basin with a clock glass, lift up the cover a little, and add the acid from a pipette. Any portion of the 1 Highly siliceous clays, Hint, and felspar usually give clear transparent fusions ; while Cornish stone, clays, and pottery bodies, even though completely decomposed, form more or less turbid liquids. Highly aluminous clays form viscid liquids ; siliceous clays, limpid liquids. It is a good plan to use rather more sodium carbonate for the former type of silicates. 2 Some object to sudden cooling by a blast of air ; dipping in cold water, etc. The sudden cooling is said to shorten the life of the crucible (C. Stockmann, Zeit. anal. Chem., 15. 283, 1876). I have not noticed any deterioration in a crucible which has been used for over 200 fusions, and cooled as described in the text. It promises to do an indefinite number more. A similar remark might be applied to several other crucibles I have used. 3 Do not "squeeze the platinum crucible between the finger and thumb to loosen the fused mass," as one writer recommends. 4 To mark porcelain vessels, scratch a mark on the vessel with a diamond. Smear platinum chloride over the mark, and when it is almost dry, wipe (not wash it off). When fired in a muffle, the vessel is marked with metallic platinum (B. Blount, Chem. Neivs, 56. 66, 1887 ; A. A. Kelly, ib., 83. 95, 1901). H. Jervis (Chem. News, 83. 118, 1901) recommends marking the crucibles with ink, and W. C. Kriescher (Chem. Mews, 83. 130, 1901) with blue pencils and then firing. C. Reinhardt (Zeit. anal. Chem., 23. 42, 1884) recommends marking in coloured enamel, finely powdered, and ground in aniseed or lavender oil. This is marked on the porcelain, and fired in a muffle. P. A. Yoder, Chem. J?ng., 15. 102, 1911 ; Journ. Ind. Eng. Chem., 4. 567, 1912 ; C. D. Mason, ib., 4. 691, 1912. 5 Analysts are frequently puzzled why the silica from a given silicate fusion is sometimes gelatinous and difficult to filter and wash, and at other times it appears more or less pulverulent and easy to wash. In many cases this is due to the way the fused cake in the crucible is treated. The pulverulent form is obtained by dissolving the cake in a large volume of water, say 200 c.c., and then acidifying the solution. This large volume may take longer to evaporate than when the acid is added to a concentrated solution ; but if the concentrated solution be acidified, the silica is more likely to separate in the gelatinous form. In the latter case, a considerable amount of time will be spent in filtering and washing, and the result will be less satisfactory. The silica which separates from the dilute solutions also appears to be of greater purity. When the silica separates from concentrated solutions, it probably encloses particles of liquid not easily removed by washing. These results are not so marked when considerable amounts of alumina are present, but in any case the different effects are noticeable. D. Lindo, Chem. News, 60. 14, 33, 41, 1889. Some consider that the addition of alcohol to the solution makes the silica separate during evaporation in the granular form, which is easily filtered and washed. It is also claimed that two evaporations are not then necessary. 1 66 A TREATISE ON CHEMICAL ANALYSIS. cake adhering inside the crucible must be removed by means of a " policeman " and hydrochloric acid. 1 Any white spots which appear on the crucible after it is supposed to have been cleaned must be transferred to the basin. The crucible lid is also cleaned and the washings transferred to the basin. The basin will now be half or three-quarters full of liquid. It is placed on the water bath and the disintegration of the cake is assisted from time to time by means of the spatula. The cake will not dissolve very readily if the fusion has been conducted at a very high temperature. Flakes of silicic acid may now be observed floating about the liquid. If any gritty particles remain in the basin, the decomposition by the sodium carbonate was not complete. 2 In that case it is best to start again. 3 1 The less dented and the smoother the crucible, the more easily is the cake removed. A very obstinate cake is dissolved in the crucible on the water bath, with or without the addition of hydrochloric acid, and the contents emptied from time to time into the evaporating basin. Fresh water or acid is introduced into the crucible. If the acid be used with a green cake, chlorine may be evolved and attack the crucible. In that case, a few drops of alcohol will destroy the manganate, and prevent the generation of chlorine. F. Stolba (Zeit. anal. Chem. , 25. 378, 1886) pours the melted mass on a suitable slab ; J. Herman ( West. Chem. Met., 5. 476, 1909) pours the melted mass into a beaker containing 50 c.c. of water. In that case, the mass should be a limpid fluid when poured, or an explosion might result. Bisulphate and caustic alkali fusions must not be poured into water. L. L. de Koninck (Zeit. angew. Chem., I. 569, 1888) recommends inserting a piece of platinum wire, bent at one end in the form of a spiral, into the fused mass when fusion is complete ; when cold, the other end of the platinum wire is sus- pended from a glass rod, or a retort clamp on a retort stand, so that the crucible hangs a few millimetres above the triangle. On heating the crucible quickly, the crucible falls into the triangle and the "melt" remains suspended on the platinum wire separated from the crucible. Both it and the residue in the crucible can then be dissolved in the regular manner. E. R. E. Muller, Chem. Ztg., 32. 880, 1908. 2 Magnetite and ilmenite may escape decomposition by the fusing carbonate, and yet be subsequently dissolved by the hot hydrochloric acid. According to P. AV. Shinier (Journ. Amer. Chem. Soc., 16. 501, 1894) in the analysis of blast-furnace slag, spinel magnesium silicate or aluminate is not decomposed by fusion with alkali carbonates. By treating the slag with hydrochloric and hydrofluoric acids, and afterwards fusing the residue with sodium carbonate, the spinel can be isolated as a crystalline powder. Treatment with concentrated sulphuric and hydrofluoric acids decomposes the spinel. This substance, therefore, if present, will be broken down at a later stage of the analysis correction for silica. 3 If the amount of sample available be small, the undecomposed material can be collected in a filter paper, ignited, and re-fused with sodium carbonate. CHAPTER XII. THE DETERMINATION OF THE SILICA. 83. The Determination of Silica. First Evaporation, The hydrochloric acid solution of the cake contains the various constituents of the clay in solution principally as chlorides. These are subsequently removed one by one. First the silica. Evaporate the contents of the basin to dryness on the water bath, protected from dust. 1 There is no appreciable loss of the constituents to be determined in the solution during the evaporation. 2 Heat the dry residue in an oven (fig. 9QV between 1QQ a-^fl 1 1 ft T Q^ leave it on the water bath 1 until t.hft sme]l of the hydrocmoric acid bas disappeared/* Four about 5 c.c. of concentrated hydrochloric acid and 30-40 c.c. of water on tte residue; warm on a water bath 10 to 15 minutes; break up any coarse lumps with the spatula or " policeman " ; decant the clear into, say, a 9-cm. filter paper, and collect the nitrate in a 400-c.c. beaker. Add more water and acid to the basin, warm again ; again decant on to the filter paper, repeat the treatment, and finally transfer the contents of the dish to the filter paper. 4 Wash with cold water until all the chlorides have been washed out. Test a drop of the filtrate with a drop of silver nitrate solution ; if there be no turbidity, the washing is complete. Second Evaporation. The filtrate is transferred to the same basin as was employed in the first evaporation, and the solution evaporated to dryness as 1 A. V. Meyer's evaporating funnel, fig. 95, is suitable (Ber., 16. 3000, 1883). The evaporat- ing funnel has been raised a little in the diagram. 2 F. Kehrmann and B. Fliirscheim (Zeit. anorg. Chem., 39. 105, 1904) refer to a loss of silica by evaporation with acids; but C. Freidheira and A. Pinagel (ib.,^$. 410, 1905) show that the loss probably occurred during the (careless) filtration, not during the evaporation. K. F. Fohr (Berg. Hillt. Ztg., 41. 252, 1882; Chem. News, 46. 40, 1882; W. Skey, ib., 16. 207, 1867; D. H.Browne, Journ. Anal. App. Chem., 5. 342, 1891; A. Vogel, Neites Rept. Pharm., 18. 157, 1869) has stated that some ferric chloride is lost during the evaporation of the acidified solution, but R. Fresenius (Zeit. anal. Chem., 6. 92, 1867) and L. L. de Koninck (Zeit. angew. Chem , II. 258, 1899 ; H. P. Talbot, Amer. Chem. Journ., 19. 52, 1897 ; R. W. Atkinson, Chem. News, 49. 217, 1884 ; H. Seward, ib., 16. 219, 233, 1867) have shown that the statement is erroneous. There is no appreciable loss either during the evaporation or during the drying of the residue at 130. If, however, residues containing ammonium chloride or aqua regia be so treated, there is, according to Talbot, an appreciable loss of iron. 3 For the influence of the drying temperature on the final result, see page 173 et seq. 4 Washing with hot water is inclined to precipitate basic chlorides of iron with the silica. Basic iron salts, for instance, are precipitated on boiling slightly acid solutions in the presence of much alkali salts (U. S. Pickering, Journ. Chem. Soc., 37. 807, 1880). In that case, the ignited silica is not white, but dirty brown. Cold water is therefore best for washing the silica of ferruginous clays, and this the more as sodium chloride is not much more soluble in hot water (39 '8 grms. per 100 c.c. at 100) than in cold water (35 '8 grms. in 100 c.c. of water at 10). But, other things being equal, hot solutions filter more rapidly than cold solutions. W. F. Hildebrand (Bull. U. S. Geol. Sur., 422. 92, 1910) does not recommend washing with dilute hydrochloric acid as advocated by C. Freidheim and A. Pringel (Zeit. anorg. Chem., 45. 411, 1905). 167 1 68 A TREATISE ON CHEMICAL ANALYSIS. before. Digest in dilute hydrochloric acid ; filter through a 7-cm. filter paper, and collect the filtrate in a 400-c.c. beaker. Take special care to remove all particles of silica which may adhere to the basin by rubbing with a " policeman," etc. 1 Wash with cold water as before. The filtrate will occupy 150 to 200 c.c. 2 Ignition of the Silica. The free edges of the filter paper are folded down on the silica, and both filter papers, while still moist, are placed in a weighed platinum crucible. The paper containing the bulk of the silica is placed in the crucible first. The triple folds of the filter paper are placed uppermost to facilitate oxidation. Heat the crucible and contents slowly some distance away FIG. 95. Evaporation for silica. from the small flame of the Bunsen's burner until the paper is thoroughly charred. An Argand burner and chimney (fig. 112) is useful for this purpose. Then gradually raise the temperature, but not high enough to ignite the vapours issuing from the mouth of the crucible, or there will be a danger of loss owing to the draught set up by the escaping gases whirling the light powdered silica out of the crucible. When the paper has all burned off, put the cover on the crucible, and blast the silica for 20 to 30 minutes in order to dehydrate the silica and render it less hygroscopic. Cool in a desiccator and weigh. 3 The silica 1 Diy the porcelain basin, when any traces of silica not transferred to the filter paper can be easily detected as white spots. F. L. Kortright, Chem. Eng., 5. 19, 1905 ; 0. Knofler (Zeit. anal. Chem., 28. 673, 1889) recommends a dark underglaze colour in the interior of evaporating basins to show up any traces of residues or precipitates like aluminium hydroxide, etc. F. Moldenhauer (Zeit. anal. Chem., 50. 754, 1911) recommends platinum basins. 2 A trace of silica passes into the second nitrate, and is partly recovered at a later stage of the analysis, as " extra silica." 3 The lower the temperature at which the silica is calcined, the more hygroscopic the resulting powder. A. Souchay (Zeit. anal. Chem., 8. 423, 1869) states that precipitated silica calcined at a low temperature absorbed 14 '38 grms. of hygroscopic moisture per 100 grms. of THE DETERMINATION OF THE SILICA. 169 should of course be ignited until no further loss in weight occurs, but experience shows that the above period is ample with a good blast. 1 Record the weight as indicated below : " Unconnected silica plus crucible." Impurities in the Silica. The silica so obtained is not quite pure. It probably contains small quantities of titanic oxide, phosphoric oxide, alumina, and ferric oxide. 2 Add 5 c.c. of water 3 and about 5 drops of concentrated sul- phuric acid. The latter is necessary to prevent the volatilisation of the titanium fluoride 4 TiF 4 (and possibly also some aluminium fluoride A1F 3 ). Add hydrofluoric acid 5 carefully, a few drops at a time, to prevent loss by the violent effervescence which sometimes occurs. The crucible is filled not quite half full with hydrofluoric acid. Warm the crucible on a hot plate over a small flame in the fume chamber until the contents are almost dry. Add 2 or 3 c.c. more hydrofluoric acid, and take the contents of the crucible to dryness on the hot plate. The radiator (page 112), or the burner illustrated in fig. 96, is useful in taking sulphuric acid to dryness without loss by spurting. The crucible is heated from above downwards, and the flame can be so adjusted that evaporation proceeds quietly. Heat the crucible to bright redness, and blast the silica ; silica calcined at the full temperature of a Bunsen's burner, 2*00 grins, per 100 grms. of silica ; and silica calcined in a blast absorbed 0'09 grin, of hygroscopic moisture per 100 grms. of silica. Powdered quartz under similar conditions absorbed no hygroscopic moisture J. W. Mellor and A. D. Holdcroft, Trans. Eng. Cer. Soc., 9. 94, 1911. 1 This point should, however, be verified. See also T. Bauer (Tonind. Ztg., 37. 89, 1913) for the difficulty in getting rid of the last traces of hydrochloric acid. 2 And, in special cases, lead sulphate, calcium and barium sulphates, tin oxide, antimony oxide, tungsten oxide, and basic salts. Silica is very liable to retain phosphoric and tungstic acids W. Skey, Chem. Neivs, 16. 187, 1867 ; and also arsenic and tin oxides if present. 3 Be careful in adding liquids to dry powders (especially if hot). There is a tendency for some of the powder to be dissipated as a "cloud of dust" E. Jordis and W. Ludewig, Zeit. anorg. Chem., 47. 180, 1905. 4 E. Riley, Journ. Chem. Soc., 12. 13, 1860 ; P. Holland, Chem. News, 59. 27, 1889 ; P. Truchot, Rev. G6n. Chim., 18. 173, 1905. Thus, Holland has shown that titanic oxide is lost by fuming with hydrofluoric acid when sulphuric acid is absent, but not when it is present: Sulphuric acid. Ti0 2 present. Ti0 2 found. None present Present Present 0-0466 0-0414 0-0520 0-0340 0-0413 0-0520 E. Wedekind (Ber. , 44. 1753, 191 1 ) has also shown that the presence of sulphuric acid considerably reduces the loss of zirconium when fuming with hydrofluoric acid to drive off the silica. 5 HYDROFLUORIC ACID. The acid must be free from non-volatile impurities K. F. Stahl, Zeit. offent. Chem., 3. 13, 1896 ; A. H. Allen, Analyst, 21. 87, 1896 ; B. Blount, ib. t 21. 87, 1896. Hydrofluosilicic acid, derived from the silica in the fluorspar used in the manufacture of the hydrofluoric acid, is a common impurity. Sulphuric acid, as well as hydrofluosilicic acid, may distil over with the hydrofluoric acid. The acid may also contain substances which reduce potassium permanganate (pages 195, 226, and 463). For apparatus for working with hydrofluoric acid, see G. Foord, Chem. News, 30. 191, 1874 (apparatus for pouring) ; E. Cohen, Natur. Ver. Neuvorpommern PMgen, 20. 1, 1889 ; H. C. Andersch, Chem. Ztg., 12. 1475, "1888 ; R. Benedict, ib.-, 15. 881, 1891 ; E. Hart, Journ. Anal. App. Chem. 3., 372, 1889 (bottles for) ; G. P. Vanier, ib., 4. 48, 1890 (pipette). Preparation of pure acid : R. Hamilton, Chem. News, 60. 252, 1889 ; W. Hempel, Ber., 18. 1434, 1885; A. P. Stuart, Amer. Chem., 2. 384, 1871. A drop of hydrofluoric acid on the skin produces a serious "burn." According to M. Kessler (Chem. Neivs, 8. 17, 1863), the " first aid " treatment for hydrofluoric acid on the hands, etc., is to use lint wetted with ammonium acetate and inject the same solution into the blisters. If the acid has touched parts of the skin difficult to moisten, e.g., under the nails, concentrated ammonia will give the best results. The excruciating pain which attends the first application of the ammonia is transient. A TREATISE ON CHEMICAL ANALYSTS. residue five minutes. 1 Weigh the crucible and contents, 2 and enter the result as "crucible plus residue." The difference between this weight and the preced- ing represents the silica. 3 A minute quantity of silica has yet to be added to the silica so determined the "extra silica" of page 185. FIG. 96. Ring burner. It is here assumed that the residue in the crucible is free from lime, magnesia, and alkaline salts, and that the crucible only contains constituents belonging to the ammonia precipitate. Bloor 4 has investigated the validity of this assumption by analysing the residues. He found among other determinations that : Table XXVII. Composition of the Silica Residues. Total residue. Alumina. Ferric oxide. Magnesia. Lime. 0-0072 0-0023 0-0008 0-0022 0-0007 0-0062 0-0006 0-0020 0-0014 0-0005 .0-0080 0-0018 0-0020 0-0007 0-0020 0-0036 o-oooo o-ooio 0-0014 0-0004 0-0074 0-0028 0-0018 o-oooo 0-0020 0-0046 0-0006 0-0018 o-oooo 0-0023 0-0080 0-0025 0-0031 0-0012 0-0008 0-0013 o-oooo o-oooo 0-0005 0-0012 1 It is important to drive off all the hydrofluoric acid, or troubles with the alumina will follow later. 3 If a sand bath is used, see that the outside of the crucible is free from sand grains before the crucible is weighed. The burner shown in fig. 96 can frequently be used instead of the hot plate or sand bath. The Bunsen's flame, if adjusted very low, is also satisfactory W. Gibbs, Chem. News, 28. 30, 1873. 3 S. V. Peppel reports that with low-grade limestones he has found that the loss in weight obtained on treating the insoluble residue with hydrofluoric and sulphuric acids is sometimes greater than the amount of silica actually present H. E. Ashley, Chem. News, go. 274, 1904. 4 W. R. Bloor, Journ. Amer. Chem. Soc., 29. 1603, 1907 ; C. Meineke, Rep. anal. Chem., 7. 214, 1887 ; E. Jordis, Zeit. anorg. Chem., 45. 362, 1905, THE DETERMINATION OF THE SILICA. Hence, Bloor concludes that the residue "is contaminated to some degree by the main constituents of the clay. The amount of contamination by sub- stances other than ferric oxide and alumina is, however, except in extreme cases, 1 so small that it may be neglected unless extreme accuracy is required '' ; in that case the residue can be taken up with a little sodium carbonate, and the acid solution of the fused mass added to the main solution. Weighings. The results of the silica weighings in the above operations, together with the " extra silica " from a later operation, will be entered in the notebook somewhat as follows : Silica and crucible 21 '0603 grms. Crucible empty 2 20*4530 ,, Uncorrected silica ........ '6073 grm. Residue and crucible 20*4552 grms. Crucible empty 20*4530 ,, Residue 0*0022 grm. Extra silica and crucible ........ 22*9531 grms. Crucible, etc., after action of HF 22*9522 Extra silica. 0*0009 grm. Silica uncorrected 0*6073 grm. Residue 0*0022 Silica e-6051 ,, Extra silica 0*0009 ,, Silica found 0'6060 ,, Correction for silica in reagents ....... O'OOOS ,, Total silica 0*6052 ,, Errors. The following results were obtained in eight independent deter- minations with one sample of clay : 0*6052 ; 0'6048; 0*6047; 0'6046 ; 0*6053; 0*6044; 0'6044; 0*6043 grm. The mean is 0*6047 grm., or 60*47 per cent., with a deviation of about 0*07. The deviations will be different with clays containing different amounts of silica. But if this particular sample be analysed by another analyst, we should expect the silica to come somewhere between 60*40 and 60'54 per cent. The variations obtained with a number of analyses as indicated above give an idea of the errors liable to affect particular determinations, but they tell nothing about the presence or absence of constant errors say, errors due to the solubility of silica in the mother liquid, etc. The case of lithium, page 537, might be cited as an instructive example. These constant errors can rarely be checked, although an approximate idea can sometimes be obtained by control analyses with artificial mixtures containing known amounts of the constituents under investigation page 247. 3 Every method of analysis has its own peculiar sources of error. The principal sources of error in silica determinations are : (1) Imperfect decomposition of the 1 E.g., when the amount of magnesium or calcium in the original sample is high, as was the case with the third sample in the above table. 2 The loss in weight by the blasting of the crucible was negligibly small for the small crucible employed. If the crucible lost appreciably in weight during the blasting, write : " Empty crucible plus loss in weight during blasting." See page 114. 3 M. Stoermer (Tonind. Ztg., 35. 453, 1911) considers that alumina is volatilised when the silica is heated with sulphuric and hydrofluoric acid. No appreciable loss can be detected under the conditions of the experiment described in the text. A TREATISE ON CHEMICAL ANALYSIS. silicate ; (2) Loss by spurting when the acid is added to the carbonate fusion ; (3) Imperfect transfer of the silica from dish to filter paper; (4) Tendency of silica to remain in a soluble condition after the baking at 109 ; (5) Mechanical l<5ss of fine particles of silica transported with the gases from the burning filter paper, and during the dehydration of the silica; (6) Contamination by the reagents, and by the porcelain vessel during evaporation; (7) And the loss of weight of the crucible itself during the prolonged blasting. Determination of Silica in the Presence of Fluorides. If fluorine be present, as, say, calcium fluoride, part of the silica will be lost by volatilisation as silicon tetrafluoride SiF 4 when the solution is evaporated to dryness. Supposing, in the extreme case, that all the fluorine be so evolved, the maximum possible error will be nearly equal to three-eighths of the calcium fluoride. The error becomes appreciable in the case of Cornish stone, which contains from 2 to 5 per cent, of calcium fluoride. In this case the silica and fluorine are to be separately determined, as indicated on page 640. Determination of Silica in the Presence of Boric Oxide. When borates are present, it is necessary to get rid of the boric oxide by adding methyl alcohol, saturated with hydrogen chloride, during the evaporation for silica, as indicated on page 589. 84. The Theory of Silica Determinations. When an aqueous solution of sodium silicate is treated with hydrochloric acid, 1 part of the silicate will be decomposed, forming sodium chloride and silicic acid, say, H 2n SiO n+2 . Conversely, when a solution of silicic acid is treated with sodium chloride, hydrochloric acid and sodium silicate are produced. These reactions are symbolised : 2rcNaCl + H 2n SiO w+2 ^= 2rcHCl + Na 2 SiO B+s . The reversed arrows are intended to represent the fact that the reaction proceeds in each direction. When the speeds of the two reactions are the same, the system is in equilibrium, and the solution contains all four substances in certain definite proportions. If not, the system is not in equilibrium, and it will be doing its best to attain that condition. Whatever be the condition of the system after it has been evaporated to dryness, when the residue is taken up with dilute acid, a definite proportion of the silica in solution will be present as colloidal silicic acid, 2 and the rest as sodium silicate. If all the soluble silicic acid in the dried residue could be converted into insoluble silicic acid by baking at a high temperature, the sodium silicate, when taken up with more hydrochloric acid, would be hydrolysed and furnish the same relative proportions of soluble silicic acid and sodium silicate as before. If the silicic acid so formed could be rendered insoluble by another evaporation, it follows that repeated evaporation to dryness and soaking the dried residue would transform practically all the silica into the insoluble condition, and the 1 Gelatinous orthosilicic acid H 4 Si0 4 is precipitated. This, when heated, or digested with sulphuric acid, forms metasilicic acid H 2 Si0 3 (R. Meldrum, Chem. News, 78. 235, 1898) ; and this at 100"-110 forms the trisilicic acid H 2 Si 3 7 which is insoluble in water and acids : at a high temperature the latter is converted into anhydrous silica. 2 For the solubility of silica, see R. Bunsen, Pogg. Ann., 61. 265, 1847; E. Ludwig, Ze.it. anal. Chem., g. 321, 1870; C. Meineke, Hep. anal. Chem., 7. 214, 757, 1887; P. Jannasch and 0. Heidenreich, Zeit. anorg. Chem., 12. 214, 1896; C. Winkler, Chem. Centr., 4. 673, 1859 ; J. W. Mellorand A. D. Holdcroft, Trans. Bug. Cer. Soc., IO. 1, 1911. For the alleged relatively greater solubility of silica in hydrochloric acid than in aqua regia, see G. C. Wittstein, Zeit. anal. Chem., 7. 433, 1868. For the solubility of silica in aq. ammonia, see page 183. THE DETERMINATION OF THE SILICA. 173 filtrate would be practically free from silica. Some analysts are under the impression that the action does occur, and recommend this procedure. 1 The percentage amount of water lost per hour in the earlier stages of the drying (110) is very much greater than in the later stages. Indeed, the last stage of the dehydration requires an indefinitely long time for its completion. 2 This is illustrated by the gradual approach of the curve (silicic acid containing 23 -9 per cent, water), fig. 97, to a horizontal line. If the temperature be raised, the 10 14- FIG. 97. Effect of time occupied by drying at 110 on the dehydration of silicic acid. transformation takes place much more rapidly and completely ; but experience shows that the higher the desiccation temperature, the greater the amount of foreign matter associated with the silica. " This," says Gilbert, 3 " is probably due to the alumina being rendered insoluble in acids." For example : Table XXVIII. Effect of the Dehydration Temperature on the Determination of Silica in Aluminous Clays. Total Si0 2 per cent. Temperature. Residue after first evaporation (per cent. ). Residue after second evaporation (per cent). Si0 2 in filtrate (per cent. ). 64-15 64-28 64-55 100 120 280 64-46 64-19 64-90 0-0018 0-0023 0-0070 0-0017 0-0032 0-0035 If the amount of alumina be high, the residue left after the removal of silica is greater the higher the drying temperature. In the case of calcareous slags (46 per cent. CaO), the calcium chloride in the residue seems to facilitate the dehydration of the silica and reduce the amount of silica soluble in the filtrate. 1 E.g., G. Lunge, Technical Methods of Chemical Analysis, London, I. 581, 1908. M. Stoermer (Tonind. Ztg., 35. 453, 1911) considers that all the silica is converted into an insoluble form by evaporating the hydrochloric acid solution ; baking at 130 ; and soaking two hours with hydrochloric acid. He considers a second evaporation to be unnecessary, for it makes no difference whether or not the silica be filtered off before the second evaporation. The experi- mental results quoted in this chapter show that two evaporations with an intervening filtration are necessary for exact work ; but see the section on abbreviated systems of analysis, page 242. 2 J. M. van Bemmelen, Zeit. anorg. Chem., 13. 233, 1897 ; Die Adsorption, Dresden, 196, 1910. According to P. H. Walker and J. B. Wilson (Circ. U.S. Dept. Agric., 101. 1, 1912), a two-hours' ignition at the highest temperature obtainable with a Bunsen's burner suffices for the dehydration of silica, and three hours for alumina ; there is then no risk of error owing to a change in weight of the platinum crucible. 3 J. P. Gilbert, Tech. Quart., 3. 61, 1890 ; Chem. News, 6l. 270, 281, 1890 ; G. Craig, ib., 60. 227, 1889. 174 A TREATISE ON CHEMICAL ANALYSIS. Table XXIX. Effect of the Dehydration Temperature on the Determination of Silica in Calcareous Clays. Total Si0 2 . Temperature. Residue after first evaporation (per cent. ). Residue after second evaporation (per cent. ). Si0 2 in filtrate (per cent. ). 41-28 41-20 41-52 100 120 280 41-33 41-32 4178 0-0013 0-0018 0-0029 0-0008 0006 0-0003 On the other hand, magnesium chloride seems to retard the dehydration at 100; and at 280 it increases the amount of soluble silica in the filtrate. Thus, with a slag containing 35 per cent, of CaO and 15 per cent, of MgO : Table XXX.Effe of the Dehydration Temperature on the Determination of Silica in Magnesian Clays. Total Si0 2 . Temperature. Residue after first evaporation (per cent.). Residue after second evaporation (per cent. ). Si0 2 in filtrate (per cent). 3370 100 33-67 0-0020 0-0023 33-80 120 33-63 0-0028 0-0008 33-94 280 33-81 0-0065 0-0052 Hillebrand x considers that magnesia begins to recombine with silica to form a magnesium silicate at temperatures over 120. This silicate is subsequently decomposed by hydrochloric acid, with the separation of silica. It is not, therefore, possible to separate all the silica by the evaporation of magnesian clays if the residue be dried at temperatures over 120. A certain amount of soluble silica will always be formed by the decomposition of the resulting magnesian silicate. If magnesium be absent, the drying of calcareous silicates can be conducted at a higher temperature than 110 for instance, 280 when " it would seem that there is no tendency for silica to recombine with lime and alumina " (Gilbert). 2 So long as any silica escapes transformation into the insoluble condition, so long will a certain proportion pass into solution, probably as sodium silicate, by the reaction between the sodium chloride and silicic acid. In illustration, Mr A. B. Trickett mixed Kahlbaum's " Kieselsaure " silicic acid with different proportions of sodium chloride solution (25 c.c. of the solution contained 5 grms.) ; evaporated the mixture to dryness, and baked the residue four hours in a steam oven. The insoluble silica was filtered off, washed, and the soluble silica determined in the filtrate by the molybdate colorimetric process (page 605). 1 W. F. Hillebrand, Journ. Amer. Chem. Soc., 24. 262, 1904. 2 If zinc be present, see page 359. THE DETERMINATION OF THE SILICA. Table XXXI. Effect of Sodium Chloride on "Soluble" Silica. 175 NaCl per gram Si0 2 (grms. ). Si0 2 in the filtrate (per cent). 18-5 26-9 45-8 62-4 078 0'81 1-06 1'07 This shows that the greater the amount of sodium chloride in the solution, the greater the amount of silica in the filtrate, presumably owing to the transforma- tion of silicic acid into soluble sodium silicate. This re-solution of sodium silicate is one of the troubles attending the determination of silica. If the silicic acids could be completely dehydrated with baking below the temperature of recombination, the difficulty would be overcome. It is important to remove as much silica as possible at this stage of the analysis, since any silica which escapes in the filtrate will contaminate the different pre- cipitates later on: 1 The possibility of a recombination of the silica with the bases during the baking of the silica limits the safe temperature of desiccation. 2 The complete drying, that is, the complete conversion of the silica into the insoluble form, might possibly be effected at 110 if an indefinite period of time were available. The longer the time of baking, the more perfect the drying, and the less the amount of silica which passes into solution in the filtrate. This is illustrated by the following experiments due to Hillebrand : Table XXXII. Effect of Time of Drying on the "Soluble" Silica. Time drying in steam oven (hours). Silica in filtrate (per cent. ). 4 24 48 2-11 1-63 1-48 In agreement with fig. 97, it follows that the complete "dehydration" of the silicic acid is excessively slow too slow to be of any practical use. 3 It will therefore be obvious that taking up the dry residue with hydrochloric acid and re-drying the residue a number of times has very little influence on the "silica in filtrate." But if the silica which has separated be removed, the amount of silica remaining when the filtrate is evaporated to dryness is comparatively small. The drying curve for this silica resembles that indicated in fig. 97 ; and although 1 R. Bunseri, Liebig's Ann., 61. 265, 1847 ; E. Ludwig, Pogg. Ann., 141. 149, 1870 ; Zeit. anal. Chem., 9. 321, 1870 ; C. Meineke, Rep. anal. Chem., 7. 214, 1887 ; A. Cameron, Chem. News, 69. 171, 1894 ; E. Jordis and W. Ludewig, Zeit. anorg. Chem., 45. 362, 1905 ; 47. 180, 1905 ; N. Knight and F. A. Menneke, Chem. News, 94. 165, 1906 ; J. A. Phillips, Phil. Mag. (4), 41. 87, 1871. 2 B. Blount, Journ. Amer. Chem. Soc., 26. 995, 1904. W. H. Stanger and B. Blount (Journ. Soc. Chem. Ind., 21. 1216, 1902) consider the temperature of baking Portland cements should not be lower than 200. This high temperature is not safe for general work, although with calcareous cements, under industrial conditions, the results are satisfactory (Table XXVIII.). T. Bauer, Tonind. Ztg., 37. 89, 1913. 3 It is probable that some volatile reagent say alcohol, page 169 might be found to do the work efficiently, but this has not yet been fully investigated. 7 6 A TREATISE ON CHEMICAL ANALYSIS. a similar percentage of the silica will remain untransformed at the end of a certain period of drying, yet the actual amount is small in comparison with the total silica. Indeed, after the second evaporation and drying, less than Ol per cent, of the total silica will be found in the nitrate. This is very well illustrated by the following numbers : Filtrate . . . First. Second. Third. Fourth. Silica ... .-0-00141 O'OOOSl 0-00023 '00006 grm. Silica .... 0-0307 0'0035 0'0024 0'0015 grm. Cameron gives the third evaporation, thus reducing the percentage silica in the third nitrate to a negligibly small amount, and after the fourth evaporation, the amount escaping dehydration is almost out of range of the balance. 1 Two evaporations suffice for technical work. It follows from Table XXIX. that when the silicate under investigation is soluble, or almost wholly soluble in acid, say hydrochloric acid, and, in conse- quence, comparatively little sodium salt is present, one evaporation will be almost as effective as two. This explains how some discordant statements are rife. For instance, with Portland cement, which is nearly all soluble in hydrochloric acid, one evaporation and baking at 180 will suffice. The relatively large proportion of lime in these silicates also facilitates the dehydration of the silica, as indicated in Table XXIX., p. 174. 2 Evaporation with sulphuric acid, 3 in place of hydrochloric acid, gives a silica which is comparatively easily dehydrated ; 4 but the presence of this acid is a source of danger on account of the risk of forming sparingly soluble sulphates which may contaminate the silica, and the escape of alumina from precipitation in the presence of sulphates page 1 80. The rate at which the silicic acid is dehydrated is connected with the con- centration of the solution from which the silicic acid was precipitated by the hydrochloric acid. For instance, Mr J. C. Green has measured the amount of water driven off from silicic acid precipitated from concentrated and from dilute solutions of sodium silicate, and afterwards dried by heating up to 450. His results are : Table XXXIII. Dehydration of Silicic Acid at 800. Time heated at 800. Minutes. Amount of moisture lost per cent. Silica from concentrated solutions. Silica from dilute solutions. 5 10 20 30 1-6 4-25 5-72 6-18 4-49 6-45 8-72 10-04 1 In these experiments the residues were not baked above 110. 2 B. Blount, Journ. Amer. Chem. Soc., 26. 995, 1904 ; S. F. Peckham, ib., 26. 1636, 1904. 3 T. N. Drown, Chem. News, 40. 40, 1879 ; Trans. Amer. Inst. Mm. Eng., 7. 346, 1879 ; V. M. Goldschmidt (Tids. Kemie Pharm. Terap., 325, 1910) recommends making the silica in- soluble by evaporation with nitric acid and hydrogen peroxide instead of with hydrochloric acid. 4 Nitric acid is used in special cases for instance, when lead is present. CHAPTER XIII. THE AMMONIA PRECIPITATE. 85. The Precipitation by Means of Ammonia. VV^E have now to deal with the filtrate, acidified with hydrochloric acid, from the silica. This solution contains aluminium, iron, titanium, phosphorus, manganese, calcium, magnesium, and alkaline salts. 1 Assume that the manganese is either absent, or present in inappreciable quantities. 2 The addition of ammonia, in the presence of ammonium chloride, 3 precipitates the aluminium, iron, 4 phosphorus, titanium, 5 and part of the magnesium. The precipitate carries down a little lime and alkalies. 6 The ammonia should be added drop by drop with constant stirring. " If an excess of ammonia be poured into the solution rather quickly, a considerable quantity of magnesia will be precipitated with the alumina " ; whereas, "if the ammonia be added, drop by drop, with constant stirring, to the 1 If lead, bismuth, copper, cadmium, tin, arsenic, molybdenum, antimony, selenium, gold, and the platinum metals be present, they should be removed by hydrogen sulphide, etc. as indicated later on page 186. 2 If manganese be present, and time is of great importance, add a few cubic centimetres of bromine water, or a few drops of bromine (page 372) to the boiling solution containing the equivalent of 10 per cent, of concentrated hydrochloric acid by volume. Make the solution alkaline with ammonia, or add hydrogen peroxide with the ammonium chloride and ammonia M. Dittrich, Ber. , 35. 4072. 1902. The manganese will be precipitated with the alumina, iron, etc. The presence of one part of iron in 80,000 parts of solution gives a visible precipitate with aqueous ammonia ; for aluminium the numbers are 1100,000; for chromium, 1 170,000; zinc, 180,000 ; manganese, 1 170,000. L. J. Curtmann and A. D. St John, Journ. Amer. Chem. Soc., 34. 1679, 1912. 3 A. Mitscherlich, Journ. prakt. Chem. (1), 81. 108, 1860 ; (1), 83. 458, 1861. 4 Ferric, not ferrous oxide. But the iron after the preceding treatment will generally be all ferric. If much manganese be present (as will be indicated by the colour of the sodium carbonate fusion), the nitrate from the silica may contain ferrous iron. " The tendency of manganese salts to exert a reducing effect on ferric iron in solution is probably the cause of this phenomenon" G. C. Stone (W. G. Waring, Journ. Amer. Chem. Soc., 26. 4, 1904). If ferrous salts be present, a few drops of hydrogen peroxide will effect the conversion (A. Carnot, Compt. Rend., 107. 948, 997, 1888). If hydrogen sulphide has been employed to remove the metals indicated in a pre- ceding footnote, it will be necessary to remove the hydrogen sulphide by boiling, filter off the sulphur, and oxidise the iron. Note that commercial hydrogen peroxide sometimes contains fluorine, chlorine, sulphuric acid, hydrofluosilicic acid, sugar, glycerine, calcium, magnesium, barium and aluminium salts, etc. H. P. Talbot and H. R. Moody, Tech. Quart., 5. 123. 1893 ; G. Arth, Monit. Sclent. (4), 15. 715, 1901 ; Chem. News, 85. 184, 1902. To purify the hydrogen peroxide : (1) Add one-tenth the volume of alcohol ; add powdered barium hydroxide free from carbonate ; filter on the pump ; precipitate barium with a slight excess of sulphuric acid ; evaporate off the alcohol. (2) Precipitate the silica and hydrofluosilicic acid with potash lye; filter ; and distil under reduced pressure. See page 205. 5 If chromium, beryllium, vanadium, and the rare earths be present, they too will be pre- cipitated as hydroxides ; uranium will be precipitated as ammonium diuranate (NH 4 ) 2 U 2 7 . 6 C. F. Cross, Proc. Manchester Lit. Phil. Soc., 17. 49, 1877 ; P. Oornette, Ann. Pharm., 4. 1, 1900; J. Thoulet, Compt. Rend., 99. 1072, 1884 ; R. Warrington, Journ. Chem. Soc., 21. 1, 1868. Some consider that the precipitated hydroxide forms a true chemical compound with the substances in question. E.g., A. V. E. Young, Amer. Chem. Journ., 8. 23, 1886 (Al) ; V. J. Hall, ib., 19. 512, 1897 (Fe) ; 19. 901, 1897 (Zn) ; H. P. Patten, ib., 18. 608, 1896 (Or) ; Journ. Amer. Chem. Soc., 25. 186, 1903 (Mg, Mn, BaS0 4 ). 177 I2 178 A TREATISE ON CHEMICAL ANALYSIS. hot solution, until the ammonia is in slight excess, the precipitated alumina will be either free from magnesia, or retain but slight traces of that element." 1 In illustration, mixtures of 1'23 grms. of magnesium sulphate, with 2 '2675 grms. of ammonia alum with 10 grms. of ammonium chloride in 750 c.c. of boiling water, gave precipitates which weighed, after ignition Slow precipitation . . . 0-2581 0'4910 grm. Rapid precipitation . . . 0'2641 0'4965 grm. The precipitate should be dissolved in hot hydrochloric acid, and reprecipitated, washed, ignited, and weighed. The ignited mass is then fused with potassium pyrosulphate, and the resulting cake digested with water. The insoluble silica 2 is filtered off, washed, and weighed as " extra silica." The iron, titanium, phos- phorus, and manganese, if present in the filtrate, are determined separately. The amounts of these constituents are added together, and the difference between the sum and the weight of the total ammonia precipitate is the alumina. Before taking up details of the process, a few special points may be considered. Ammonia for the "Alumina" Precipitation. The ammonia used should be free from carbonates; otherwise, calcium carbonate will be precipitated with the alumina, etc. 3 Ammonia 4 absorbs carbon dioxide from the atmosphere, and hence some analysts make a fresh solution of ammonia periodically for this work. If calcium carbonate be formed, H. Rose 5 thinks that it will be decomposed by the ammonium salts present in the solution, when the mixture is boiled. The ammonia should be kept in cerasine bottles, or glass bottles lined on the inside with cerasine. Freshly prepared ammonia kept in common glass bottles for one month gave a residue of 0*7 grm. per 25 c.c. ; in a Jena glass bottle, 0'4 grm. ; and in cerasine bottles, no residue. 6 Influence of Soiling and Long Standing on the Alumina Precipitate. There are some objections to the boiling of solutions containing the alumina precipitate. Rose and Fresenius 7 recommend boiling off the excess of ammonia from the solution. Lunge 8 considers the boiling unnecessary. The objections to the boiling are : (1) Some alumina may be redissolved by prolonged boiling owing to the decomposition of the ammonium chloride, with the formation of hydro- chloric acid and volatile ammonia. 9 Hence, it is well to make sure that the solution is alkaline before filtering. (2) Prolonged boiling to drive off the excess of ammonia may lead to a contamination of the precipitate by dissolution 1 L. F. J. Wrinkle, Chem. News, 22. 4, 1870; H. Abich, Pogg. Ann., 23. 352, 1831 ; W. R Nichols, Amer. J. Science (2), 47. 16, 1869. 2 The trace which escapes in the second nitrate mentioned in the preceding chapter, as well as any dissolved from the glass vessels, reagents, etc., will be, for the most part, recovered later. 3 For the action of ammonia on silica, see pages 172 and 183. 4 The ammonia is tested for carbon dioxide by an aqueous solution of calcium chloride. There should be no opalescence. For the analysis of commercial "aqua ammonia," see J. D. Pennock and D. A. Morton, Journ. Amer. Chem. Soc., 24. 377, 1902. For pyridine in ammonia, see H. Ost, Journ. prakt. Chem. (2), 28. 271, 1883 ; for lead, see W. F. Lowe, Journ. Soc. Chem. Ind., u. 133, 1892. The latter impurity is introduced when the manufacturer places the ammonia in leaden vessels for dilution to the required specific gravity. 5 H. Rose, Chem. News, 2. 291, 1860 ; Pogg. Ann., no. 292, 1860. 6 E. T. Allen and J. Johnston, Journ. Ind. Eng. Chem., 2. 196, 1910. 7 R. Fresenius, Anleitung zur quantitaliven chcmischen Analyse, Braunschweig, I. 160, 1903; London, i. 192, 1900; L. Blum, Zeit. anal. Chem., 27. 19, 1888; H. Rose, I.e. For the action of ammonium chloride on metallic sulphides, see P. de Clermont, Compt. Rend. , 88. 972, 1879. 8 G. Lunge, Zeit. angew. Chem., 12. 635, 1889 ; G. Lunge and H. von Keler, 7. 670, 1894 ; C. Meineke, Rep. anal. Chem., 7. 214, 757, 1888. 9 R. Fresenius, Anleitung zur quantitativen chemischen Analyse, Braunschweig, 2. 807, 1905; R. Fittig, Zeit. anal. Chem., 27. 29, 1888; Liebig's Ann., 128. 189, 1863; H. C. Dibbits, Zeit. anal. Chem., 15. 245, 1876 ; Ber., 5. 820, 1872; A. R. Leeds, Chem. News, 29. 256, 1874; D. Gernez, Compt. Rend., 64. 606, 1867. THE AMMONIA PRECIPITATE. 179 of silica, etc., from the glass vessels. (3) Prolonged boiling also tends to make the precipitate slimy and difficult to filter and wash. The same objection applies to precipitates which have stood some time. (4) Calcium carbonate is also precipitated under these conditions owing to the absorption of carbon dioxide from the air. The important thing is to filter rapidly. If the filtration and washing of the precipitate from a gram of clay occupy much over half an hour, it will be almost impossible to remove the adsorbed salts, in a reasonable time, by washing hence some use a hot funnel (fig. 130). If the precipitate has reached the slimy stage 1 before the washing is completed, it is generally advisable to redissolve it in hydrochloric acid, and reprecipitate. For the reasons stated above, if the filtrate from the silica has to be left standing some days, let the solution be acidified by hydrochloric acid before "shelving." Do not let the alumina precipitate stand in its mother liquid unfiltered. The effect of temperature, etc., on the precipitation of colloids, of which the ammonia precipitate is a typical example, has been discussed on page 96. Filtration of Gelatinous Precipitates. Before adding the ammonia, Dittrich 2 mixes the solution with macerated filter paper pulp. 3 This facilitates the washing of the gelatinous precipitate, and also the oxidation of the precipitate during the ignition. The process gives good results. The precipitate is rather bulky, and accordingly a larger filter paper must be used, and also a larger crucible for the ignition. A certain amount of salt is absorbed by the filter paper, although Mansier 4 says that sodium chloride is not retained by the washed paper. Do not use paper pulp for highly ferruginous clays, for reasons which will appear later (page 184). In order to coagulate gelatinous precipitates, and render them easy to filter, Divine, 5 recommends the addition of 6 c.c. of a 2 '5 per cent, solution of tannin the addition of 2 c.c. of the tannin solution, before adding the ammonia, works well. Guyard 6 recommended the addition of glycerol ; Palmer, stirring with 1 The precipitated aluminium and ferric hydroxides sometimes appear to dissolve and pass through the filter paper, particularly towards the end of a washing C. F. Cross, Chem. News, 39. 161, 1879 ; E. Schirm, Collegium, 99, 1911. G. P. Baxter and R. A. Hubbard (Journ. Amer. Chem. Soc., 28. 1208, 1906) state "that the solubility of ferric oxide in ammonia is caused by the presence of some organic impurity in the ammonia," and they tried, without success, to reproduce the phenomenon by mixing methyl-, ethyl-, diethyl-, isoamyl-amines, ethylenediamine, aniline, and phenylhydrazine with the ammonia. I have always attributed the phenomenon to a deflocculation of the colloidal precipitate (page 96). 2 H. Jervis, Chem. News, 78. 257, 1898; F. Ibbotson, Technics, 2. 357, 1904 ; M. Dittrich, Ber., 37. 1840,1904. A. C. F. M'Kenna (Jour. Amer. Chem. Soc., 21. 125, 1899; Chem. Netvs, 79. 184, 18-99) recommended the paper pulp process for zinc sulphide in 1899 ; E. Murmann (Zeit. anal. Chem., 50. 742, 1911) recommends the addition of a little of the finest starch, mercuric sulphide, or shredded paper for gelatinous precipitates like aluminium hydroxide, manganese and zinc sulphides, etc. 3 PAPER PULP FOR FILTRATION. The pulp may be prepared for filtration in the following manner : Crush Swedish filter paper into small balls, put the balls into a large empty cerasine bottle. Add concentrated hydrochloric acid, and a little hydrofluoric acid. Seal up the stopper of the bottle with wax, and shake in a shaking machine for a couple of hours. Wash the resulting pulp free from acid by decantation. Keep the emulsion of pulp and water in a bottle for use. 4 M. Mansier, Rev. Internal. Fal.nf., 15. 115, 1903. 5 R. E. Divine, Journ. Soc. Chem. Ind., 24. 11, 1905. 6 A. Guyard, Zeit. anal. Chem., 22. 426, 1883; L. Liebermann, ib., 14. 359, 1875; K. Zulkowsky, Chem. Ztg., 8. 772, 1885 ; H. N. Warren, Chem. News, 61. 63, 1890; C. S. Palmer, Eng. Min. Journ., 80. 582, 1906. T. M. Chatard (Amer. J. Science (2), 50. 247, 1870 ; (3), 2. 416, 1871; Chem. News, 22. 246, 1870; 24. 270, 1871) recommends the evaporation of gelatinous precipitates to dryness on a water bath, when it is claimed that chromic, feme, aluminium, beryllium, titanium hydroxides; nickel carbonate; cerium, lanthanium, and didymium oxalates become granular and easy to filter. The washing of these precipitates free from salts is then a difficulty. i8o A TREATISE ON CHEMICAL ANALYSIS. one or two drops of albumen (white of egg), and heating to boiling ; Liebermai uses starch in a similar manner ; Warren, a few drops of an ethereal solution pyroxyline ; and Zulkowsky recommended shaking the liquid with one-third its volume of ether. The ether entangles the precipitate and carries it to tht surface. Some of these recommendations must be adopted with caution, on account of the tendency of organic substances to retard the precipitation of alumina. Glycerol, for instance, is under certain conditions highly objectionable. 1 Influence of Fluorides on the Precipitation of Alumina. The presence of fluorides hinders the precipitation of aluminium hydroxide by ammonia. 2 A soluble aluminium fluoride is produced which is not completely decomposed by the ammonia. The reaction is represented by the equation : A1F 3 + 3NH 4 OH;=^A1(OH) 3 + 3NH 4 F. The reaction therefore proceeds in both directions, and in all probability a soluble salt, (NH 4 ) 3 A1F 6 , is produced when much ammonium fluoride is present. The effect of different proportions of ammonium fluoride on the amount of alumina which is "lost," owing to its remaining in solution, is shown by the graph, fig. 98. This was worked out by Hinrichsen. Hinrichsen further showed 100 50 1 ^r/775 NH^ 10 0'2 0'4 0'6 0-8 FIG. 98. Action of fluorides on the pre- cipitation of alumina (Hinrichsen). (j ^ ( ^ fc ' s y ^ ^ r\i ^ ^ * y A 6 Uj f / ^ "Q^ ^ J y . jz T c **;,|i ^ /I 0-8 1-6 2-4 3-2 FIG. 99. Action of sulphates on the pre- cipitation of alumina (Trickett). the danger of introducing fluorine into the solution when the clays, etc., are opened up with sulphuric and hydrofluoric acids, and when the silica is corrected by treatment with the same acids, because of the difficulty in driving off the last traces of the fluorine. In these cases the solution should be evaporated to dryness, and the residue calcined in order to transform the fluorides into oxides. There is then no difficulty with the fluorine. Influence of Sulphates on the Precipitation of Alumina. There is also a risk of alumina escaping precipitation in the presence of sulphates. Mr A. B. Trickett, in my laboratory, has measured the amount of alumina which escapes precipita- tion in the presence of different amounts of ammonium sulphate, and his results are illustrated by the graph, fig. 99. Starting with the equivalent of 0'0077 grm. of A1 2 3 in solution, and increasing the amount of ammonium sulphate from O8 to 3*2 grms. (abscissae, fig. 99), the corresponding amounts of alumina in solution when an excess of ammonia is added are represented in the diagram. 1 A. Guyard, Bull. Soc. Chim. (2), 31. 354, 1879. 2 F. P. Veitch, Journ. Amer. Chem. Soc., 22. 246, 1900 ; "W. R. Bloor, ib., 29. 1603, 1907 ; L. J. Curtman and H. Dubin, ib., 34. 1485, 1912 ; F. W. Hinrichsen, Ber., 40. 1497, 1907 ; Zeit. anorg. Chem., 58. 83, 1908. THE AMMONIA PRECIPITATE. l8l he presence of a large excess of ammonium chloride appears to reduce the lount of alumina which escapes precipitation in this manner. According to Wrinkle, 1 magnesia is much more liable to be precipitated with 6 he alumina if sulphates be present. 86. The Theory of the Ammonia Precipitation. Many salt solutions are decomposed by the action of water. The phenomenon is termed hydrolysis. Thus, bismuth or antimony chlorides with water form insoluble oxychlorides ; mercuric sulphate forms a basic sulphate ; and boiling solutions of ferric chloride form a mixture of ferric hydroxide and a basic ferric chloride, together with free hydrochloric acid. It is possible to draw up a list of salts where the phenomenon is marked and well-defined at one end of the series, and at the other end the hydrolysis is but ill-defined and feeble. In the former case the reaction is quantitative, and may be employed in analytical separations ; in the latter case the separations are incomplete. It will be observed that an acid is usually one product of the hydrolysis. For instance, the hydrolysis of titanic sulphate is represented in symbols Ti(S0 4 ) 2 + 4H 2 0^^2H 2 S0 4 + Ti(OH) 4 . Titanium hydroxide Ti(OH) 4 is soluble in the free acid, but at first the rate at which the sulphate is decomposed is much greater than the rate of dis- solution of the hydroxide by the acid. Hence, acid accumulates in the solution. As the acid accumulates in the solution, its effects become more and more marked, and finally, when the free acid has attained a certain concentration, the speeds of the two reactions will be equal, and no further separation of the hydroxide will be observed, because it will be dissolved by the free acid as fast as it is formed. When weak bases aniline, ammonia, phenylhydrazine, etc. are present, some of the liberated acid is converted into a neutral salt of the base. The concentration of the free acid is thus diminished, and a much greater proportion of the hydroxide will separate. If a sufficient quantity of the base be present to prevent the acid accumulating in the system, all the salt may be hydrolysed. The salt say, ammonium chloride, NH 4 C1 formed by the union of the base with the free acid may itself be hydrolysed by the water H 2 + NH 4 C1^NH 4 OH + HC1. Hence, there is a limit to the work the base can do in neutralising the free acid. In the case of the alumina precipitation, there is nearly always enough free acid present to prevent the hydrolysis of the magnesium salts, whereas with aluminium, titanium, zirconium, chromium, beryllium, thorium, cerium, and ferric salts, the liberated hydroxide can exist in the presence of the small amount of free acid produced by the hydrolysis of the ammonium salt. The complete separation of these hydroxides thus depends upon the amount of free acid which is liberated by the hydrolysis of the salt formed by the com- bination of the freed acid with the base. With aniline more free acid will be formed in the system than with phenylhydrazine, and with phenylhydrazine, more than with ammonia. Hence, hydroxides completely precipitated by 1 L. F. J. Wrinkle, Chem. News, 22. 4, 1870. According to H. Bley (Journ. prakt. Chem. (1)> 39- 1> 1846), a certain amount of sulphuric acid or sulphate may also be dragged down with the precipitate. 1 82 A TREATISE ON CHEMICAL ANALYSIS. ammonia may be only partially precipitated by phenylhydrazine, and not at all by aniline. With beryllium chloride, for example, the hydroxide is completely precipitated by phenylhydrazine, but not by aniline, and ferric iron is precipi- tated by ammonia, but not by phenylhydrazine. The nature of the acid is of great importance, since the salts of one acid may be more susceptible to hydrolysis than the salts of another. Thus, complete precipitation may occur with beryl- lium chloride, while but a trace is precipitated with the nitrate. 1 The general principle here indicated may, of course, be modified in special cases, because of the formation of double salts, basic salts, etc. Some authors attribute the effect of ammonium salts, in retarding the pre- cipitation of magnesium hydroxide by ammonia, to the formation of the complex salt Mg(NH 4 ) 2 Cl 4 which is not decomposed by ammonia. As a matter of fact, ammonia, in the absence of ammonium salts, precipitates half the magnesium as hydroxide, and the other half as the complex salt. It is then supposed that the addition of an ammonium salt dissolves the hydroxide with the formation of more Mg(NH 4 ) 2 Cl 4 . The accumulation of ammonium chloride 2 in the system will also lessen the amount of the hydrolysed salt in accord with the view indicated in the text. 3 Similar remarks might be applied to manganese salts. With zinc, 4 cobalt, 5 and nickel 6 complex salts may be formed, so that both influences indicated above come into play. 87. The Determination of the Ammonia Precipitate. First Precipitation. The filtrate from the silica, in a 400-c.c. beaker, is mixed with about 10 c.c. of ammonium chloride solution 7 and heated to boiling. 8 Aqueous ammonia is then added in slight excess. 9 The ammonia is added 1 A. M. Jefferson, Journ. Amer. Chem. Soc., 24. 540, 1902 ; B. L. Hartwell, ib., 25. 1128, 1903 ; E. T. Allen, ib., 25. 421, 1903 ; W. H. Hess and E. D. Campbell, ib., 21. 776, 1889. 2 J. M. Loven, Zeit. anorg. Chem., II. 404, 1896; F. P. Treadwell, ib., 37. 326, 1904; W. Herz and G. Muhs, ib., 38. 138, 1904. 3 Usually the solution contains sufficient hydrochloric acid to form enough ammonium chloride with the ammonia to prevent the precipitation of the magnesium hydroxide. A. A. Noyes, W. C. Bray, and E. B. Spears (Tech. Quart., 21. 14, 1908 ; Journ. Amer. Chem. Soc., 30. 481, 1908) state that 0'005 grm. of magnesium chloride made up with 5 c.c. of hydrochloric acid (sp. gr. 1'12) to 100 c.c. gave no precipitate with 40 c.c. of ammonia (sp. gr. 0'96), but a precipitate appeared with 50 c.c. of ammonia. 4 W. Herz, Zeit. anorg. Chem., 25. 225, 1900 ; W.Gaus, ib., 25. 236, 1900 ; W. Bonsdorff, ib., 41. 132, 1904 ; H. Euler, Ber., 36. 3400, 1903. 5 A. Werner, Ber., 40. 15, 1907. H. M. Dawson and J. M'Crae, Journ. Chem. Soc., 77. 1239, 1900 ; W. Bonsdorff, Zeit. anorg. Chem., 41. 132, 1904 ; M. Konowaloff. Chem. Centr., i. 646, 1900. 7 AMMONIUM CHLORIDE SOLUTION. 107 grms. of the salt per litre. The object of the ammonium chloride is to retard the precipitation of magnesium and manganese hydroxides, as indicated above. 8 W. E. Taylor (Chem. News, 103. 169, 1911) claims that adding the ammonia to the solution at 66, and then raising the temperature to the boiling point, gives a more granular precipitate than adding ammonia to the boiling solution. 9 The use of the term "excess" in analytical chemistry often misleads beginners. It is comparatively rare to find reactions in which complete precipitation is effected by adding the theoretical amount of the precipitating agent. When such reactions are known, they generally make useful volumetric processes. In most cases, more of the precipitating agent must be added than is satisfied by the regular type of equation representing the reaction. The term " excess " means that enough precipitating agent must be added to ensure complete precipitation and no more. As J. W. Mallet used to say, ad maxima per minima. Analytical chemistry has not yet reached that stage where it can answer, for each reaction, the following types of question : What amount constitutes an excess ? How does this excess differ when different salts are present ? Does the presence of certain salts diminish the amount of the ' ' excess " needed ? etc. THE AMMONIA PRECIPITATE. 183 slowly in order to prevent bubbles of gas violently projecting some of the hot liquid from the beaker. While the solution is still boiling, filter promptly through, say, a 12'5-cm. filter paper, 1 and wash four times by decantation 2 with a hot solution of ammonium nitrate 3 in order to coagulate the gelatinous precipi- tate. The filter paper is washed at the same time. Second Precipitation. Hot^dilute hydrochloric acid (1 : IHs then run through the filter paper into the beaker in which fo precipitation was first made. The 'precipitate dissolves.* me precipitation with ammonium chloride and ammonia is repeated. The precipitate is washed as before four times by decantation, with a hot solution of ammonium nitrate. The precipitate is then transferred to the filter paper, 5 and washed with the hot solution of ammonium nitrate until the washing liquid is free from chlorides, when a few drops of the filtrate are tested with silver nitrate. Recovery of Alumina and Silica from the Filtrate. Towards the end of the washing, a re-solution of some of the alumina precipitate sometimes takes place. 6 In any case some alumina and silica 7 is generally found in the filtrate. This must be recovered. Evaporate the filtrate nearly to dryness. 8 Ammonia is added to the solution and the evaporation continued. The solution is kept alkaline to coagulate any iron and aluminium hydroxides. Collect and wash these on a 5-cm. filter paper, and run the filtrate in, say, a 250-c.c. beaker. The filtrate is used for the determination of lime and magnesia. 9 Ignition of the Precipitate. Let the precipitate drain, and transfer the filter paper and contents, while still moist, to the platinum crucible containing the residue from the silica. Put the filter paper 10 in the crucible with the triple 1 With many clays the precipitate settles too slowly for washing by decantation. 2 For china clays use a 12*5-cm. paper ; for Cornish stone, a 11-cm. paper; and for glazes and siliceous clays, a 9-cm. or even a 7 -cm. paper will be ample ; see page 89. 3 AMMONIUM NITRATE SOLUTION. Neutralise 20 c.c. of concentrated nitric acid with ammonia, and dilute to a litre (R. Bunsen, Liebig's Ann., 106. 13, 1858 ; S. L. Penfield and D. N. Harper, Amer. J. Science (3), 32. 112, 1886 ; Chem. News, 54. 90, 102, 1886). A few drops of litmus will indicate whether the solution be acid. An acid solution is, of course, fatal to success. 4 Freshly precipitated aluminium and ferric hydroxides are readily soluble in dilute acids, but after standing a short time they take a long time to dissolve M. Jeannel, Compt. Rend., 66. 799, 1868 ; J. Attfield, Chem. News, 17. 303, 1868. Hence the need for speedy work. To facilitate the solution of gelatinous precipitates, F. A. Gooch (ZeU. anorg. Chem., 46. 208, 1906 ; Chem. News, 92. 64, 1905) uses a cone of platinum gauze between the filter paper and the- precipitate. Most of the precipitate can then be lifted with the cone from the paper and trans- ferred to the beaker for solution P. T. Austen, Chem. News, 38. 88, 1878. 5 Instead of the " policeman," a swab made from a quarter of a 7-cm. paper may be rubbed against the sides of the beaker and transferred to the filter paper. The swabbing is repeated with each of the other three quarters. 6 When water alone is used for the washing, or the ammonium nitrate solution becomes acid. 7 For the solubility of silica in aqueous ammonia, and water, see W. Skey, Chem. Neivs, 17. 165, 1868 ; A. M. Edwards, ib., 73. 13, 1896 ; R. Pribram, Wittstein's Viertel., 16. 30, 1867 ; Chem. News, 17. 227, 1868 ; A. Souchay, Zeit. anal. Chem., II. 187, 1872 ; G. Karsten, Pogg. Ann., 6. 357, 1826. 8 It is best to evaporate the main filtrate and washings separately. The washings are taken to dryness, and the ammonium salts driven off. 9 F. M.uck(Zeit. anal. Chem., 19. 140, 1880) recommends removing the sodium and potassium chlorides which accumulate in the solution particularly after the " basic acetate separation" (page 362) by evaporating the solution to dryness ; dissolving the residue in concentrated hydrochloric acid ; and washing the residual salts with concentrated acid on a glass-wool filter. There is, however, rarely any occasion for this operation. Ammonium salts are usually removed by evaporation to dryness and direct volatilisation, or by heating with nitric acid (H. Jervis, Chem. News, 86. 271, 1902), or with some nitrous acid (P. Jannasch, Journ. prakt. Chem. (2), 72. 38, 1905). 10 Also the filter paper containing the residual alumina. 184 A TREATISE ON CHEMICAL ANALYSIS. fold uppermost, and leave a passage for the exit of steam. 1 Ignite the precipitate, slowly at first, in order to prevent the bubbling mass from carrying a portion of the precipitate to the upper side of the crucible, where it is liable to stick, and subsequently escape the solvent action of the potassium pyrosulphate. Heat the crucible until the paper is charred. The crucible 2 is of course laid on its side on the platinum triangle to permit free access of air, and heated red-hot about 20 minutes. This is followed by 5 to 10 minutes 5 blasting, or heating over a Meker's burner. The alumina is not properly dehydrated if heated on the ordinary Bunsen's burner. 3 Let the crucible cool in a desiccator. Moisten the cold mass with a drop of concentrated nitric acid, and heat gently until no more fumes are evolved. Re-ignite, cool, and weigh. The number remaining when the weight of the empty crucible has been subtracted, gives the weight of the alumina, ferric oxide, 4 titanic oxide, and phosphoric oxide (if present) in the precipitate. 5 Dissolution of the Ignited Precipitate. The precipitate dissolves so slowly in acids that it is advisable to fuse the ignited precipitate with about six times its weight of potassium pyrosulphate. 6 The preliminary heating, especially if potassium bisulphate be employed in place of the pyrosulphate, must be very gradual, in order to avoid loss by the spattering of the fused mass. The covered crucible is heated over a small flame until the contents are melted. The 1 With highly ferruginous clays it is best to dry the precipitate in an air bath at about 110, and ignite the paper separately in order to prevent the reduction of the ferric oxide by the carbon of the paper. The magnetic oxide of iron formed by the reduction cannot easily be re- oxidised to Fe 2 3 , since it is protected from the air when buried in the alumina. Alumina, however, "decolorises " the ferric oxide and retards the reduction to magnetic oxide (H. Wartha, Chem. News, 84. 305, 1901) ; indeed, a mixture of alumina with 6'8 per cent, of Fe 3 4 after ignition contained nothing but ferric oxide. If the alumina has been precipitated with paper pulp, this precaution is not required, since the precipitate is then fine enough and open enough to readily reoxidise, even if it be partly reduced W. Suida (Tschermak's Mitt. (1), 5. 176, 1876) has shown that calcined ferric oxide is not reduced to ferrous oxide if reducing agents be excluded (H. Rose, Pharm. Centr. (1), 19. 483, 1848). C. Bodewig (Zeit. Kryst., 7. 176, 1883) states that there is always a certain amount of reduction in platinum crucibles, even when the crucible is but half covered with a sloping lid, since the ignited precipitate, when taken up with hydrochloric acid, gives a blue coloration with potassium ferricyanide. H. St C. Deville and L. Troost (Compt. Rend., 56. 977, 1863) have shown that platinum is permeable to the flame gases at high temperatures (see page 372). Bodewig prefers to ignite the precipitate in a porcelain crucible, moisten with nitric acid, dry, ignite, and weigh. This sequence of operations is repeated a second time. The oxide thus obtained is said to be free from ferrous oxide. 8 H. von Juptner (Chem. Ztg., 13. 1303, 1889) has devised a little asbestos cover for crucibles which accelerates the combustion of filter papers, and of organic matter during " ignitions." 3 A. Mitscherlich, Zeit. anal. Chem., I. 67, 1862 ; E. T. Allen and V. H. Gottschalk, Amer. Chem. Journ., 24. 292, 1900 ; E. T. Allen and H. F. Rogers, ib., 24. 304, 1900 ; but see page 173. 4 Ferric oxide sometimes stains the platinum crucible badly (see page 115). This stain can generally be removed by letting the crucible stand overnight in contact with concentrated hydrochloric acid, and then warming it for a short time. Fused potassium bisulphate or pyrosulphate will also clean off the iron stain. 5 Here, too, will be found, if present : niobium (columbium), tantalium, tungsten, zirconium, beryllium, chromium, thorium, and the rare earths. In the rare event of a determination of these constituents being required, see the later pages of this work. 6 Some prefer potassium bisulphate, but with this salt the rise of temperature must be very, very gradual in order to prevent frothing over. This means an expenditure of time. The pyro- sulphate gives less trouble in this respect. It is necessary to examine the commercially pure potassium pyrosulphate. I have found a number of packages of the best potassium pyro- sulphate, sold by a celebrated maker of pure chemicals, containing bits of glass. This impurity is fatal to successful work. It is derived from the glass vessels in which the salt is fused. The glass vessels are broken to get the pyrosulphate out. To convert ordinary bisulphate to the pyrosulphate, melt 'the bisulphate in a platinum dish, and when the spluttering has ceased, and white fumes begin to come off freely, pour the fused mass on to another dish or plate. When cold, the pyrosulphate can be easily broken into pieces and bottled. J. L. Smith (Amer. J. Science (2), 40. 248, 1865) preferred the sodium salt. It acts more rapidly than the potassium salt, and also forms a more suitable cake. It is, however, liable to crust over during THE AMMONIA PRECIPITATE. crucible is then raised six or nine inches above the flame, as illustrated in fig. 100. The moisture can then be driven off without danger. The crucible can be lowered nearer the flame in a short time. 1 If any particles of the precipitate adhere to the sides, wash them down by imparting a rotary motion to the contents of the crucible, or tilt the crucible a little to permit the fused salt to act on the grains. Slowly raise the temperature until the bottom of the crucible shows faint redness, but watch carefully to prevent frothing. If the crucible be lifted away with the tongs, and the contents of the cooling crucible be watched in a good light, it will be easy to see through the transparent mass if all has dissolved. Heat the crucible a few minutes more, even if all has dissolved. 2 Cool by placing the crucible on a cold slab. When cold, pour cold water into the crucible. The cake soon comes away from the crucible. It is then dissolved in water. 3 The cake dissolves quicker in warm water, but if the solution be boiled, titanium oxide may be pre- cipitated. The solution is mixed with about 10 c.c. of dilute sulphuric acid. It is well to know how much acid and pyrosulphate have been em- ployed, so that an allowance can be made later on when dealing with the titanium. Correction for Silica. The solution formed by the dissolution of the cake is evaporated on a water bath to a small volume and then heated to a higher temperature until fumes of sulphuric acid come off copiously. 4 Sufficient sulphuric acid should be present to form a pasty mass when the dish is cold. Water is added to the cold mass. The dish is placed on a water bath, and the silica which separates is filtered off, washed, ignited, and weighed. 5 The residue is treated with hydrofluoric acid, as indicated on page 169. The residue is ignited and weighed. The difference in the two weighings is called the " extra silica " ; its weight is to be subtracted from the weight of the precipitate, the fusion ; and the cold cake is not so easily detached from the crucible as when potassium pyrosulphate is employed. It can be adopted with advantage. E. Deussen (Zeit. anorg. Chem., 44. 423, 1905) recommended acid potassium fluoride with the idea of eliminating the solvent action of the pyrosulphate on the platinum. This salt is not to be recommended in accurate silicate analyses, because (1) the " extra silica" is lost ; and (2) the fluorine, unless expelled, later on interferes with the permanganate titration and the titanium determination. 1 Half an hour, if potassium bisulphate be used instead of the pyrosulphate. 2 If the mass be heated too strongly and normal potassium sulphate be formed, the small contraction of the cooling sulphate and the relatively large contraction of the crucible may cause the crucible to burst. 3 E. Hart (Journ. Anal. App. Chem., 2. 410, 1888) adds concentrated sulphuric acid in excess, and warms the mixture until the mass is dissolved. When cold, dilute with water and neutralise the excess of acid with sodium carbonate. 4 W. F. Hillebrand, Hull. U.S. Geol. Sur., 422. 106, 1910. If the mass darkens, platinum, derived from the crucible during the bisulphate fusion, is separating. 5 See also J. A. Phillips, Phil. Mag. (4), 41. 87, 1871. FIG. 100. Pyrosulphate fusion. 1 86 A TREATISE ON CHEMICAL ANALYSIS. as indicated on page 184, and added to the total silica of page 169. The residue in the crucible may be treated for barium, as indicated below ; or fused with a little more pyrosulphate, and the cold mass, when dissolved in water, added to the main solution, 1 which is made up to 250 or 300 c.c. in a measuring flask. Correction for Barium. If traces of barium be retained by the silica and the alumina precipitates, it will be found in the residue in the crucible left after determining the "extra silica." It can be recovered by floating off the barium sulphate from any platinum by means of water. Fuse with potassium pyro- sulphate, dissolve the cold cake in dilute sulphuric acid, collect the barium sulphate on a 5-cm. filter paper, ignite, and weigh (page 618). 2 Correction for Platinum from the Crucible. In exact analyses provision for the removal of platinum should always be made after a pyrosulphate or a bisulphate fusion. Platinum is readily precipitated from hot sulphate solutions by hydrogen sulphide. Some platinum may be found in the filtrate from the silica, but it is not. usually necessary to remove it at this stage of the work. Hence all the platinum is not necessarily derived from the crucible during the pyrosulphate fusion, and, in consequence, if the platinum be weighed in the same crucible as that in which the bisulphate fusion was made, any excess over that lost by the crucible should be deducted from the alumina. Errors. The same clay, mentioned at the end of the chapter on the silica determination, was treated for the alumina, etc. The following numbers were obtained with eight independent determinations for the " ammonia " precipitate : 0-2460; 0-2452; 0-2474; 0'2458 ; 0-2447; 0-2461; 0'2465 ; 0-2457, with a mean value of 24*60 per cent., and a deviation of approximately 0'13. The chief sources of error are: (1) Imperfect precipitation of the alumina and iron ; (2) Imperfect washing of the precipitate ; (3) Precipitation of lime owing to the use of "old" ammonia; (4) Imperfect dehydration ; (5) Variable state of oxidation of the iron on ignition ; (6) Contamination with silica subsequently dissolved in the bisulphate fusion. The imperfect washing is a most serious error, and it is probably here that beginners usually go wrong. The alumina determination is the pons asinorum of clay and silicate analysis. The iron (page 187), manganese (page 372), phosphoric oxide (page 595), and titanium (page 203) are now determined in aliquot portions of the acid solution of the pyrosulphate fusion. The corrections for vanadium, rare earths, beryllium, uranium, chromium, etc., are discussed in the chapters dealing with these elements. 1 According to W. F. Hillebrand (Journ. Amer. Chem. Soc., 24. 369, -1902; Chem. Neivs, 86. 90, 1902), from one to two milligrams of silica still remain in solution and escape recovery. 2 If an appreciable amount be present, it is well to prove that it really is barium sulphate by fusion with sodium carbonate, etc. If phosphoric acid be present, some titanium phosphate may be precipitated, and this may be mistaken for barium sulphate. CHAPTER XIV. THE DETERMINATION OF IRON. 88. The Determination of Iron. THERE is a wide choice of methods for the determination of iron depending upon volumetric, colorimetric, or gravimetric processes. In volumetric processes the ferric iron may be reduced to the ferrous condition by stannous chloride, metallic zinc, hydrogen sulphide, etc. The special features of the different processes of reduction here recommended are indicated in the text. The ferrous salt is re- oxidised to the ferric condition by a standard solution of potassium permanganate or potassium dichromate. The .iron can also be determined volume trically while in the ferric condition by titration with titanous chloride, 1 which reduces the ferric to the ferrous condition during the titration. Stormer 2 has compared per- manganate and dichromate titrations with Rothe's ether process, and pronounces in favour of the permanganate titration. The permanganate process has long been a favourite on account of its simplicity, accuracy, and elegance, and it is the best process to use when the conditions are favourable but unfortunately, as we shall soon see, it is not always satisfactory in clay analyses. The colorimetric process (page 200) has the advantage that the iron is directly determined in the ferric condition, without reduction, and it is quite satisfactory, both in speed and accuracy, for general work on clays containing less than about 2 per cent, of ferric oxide. For gravimetric processes, see pages 455 et seq. 89. The Reduction of the Ferric to Ferrous Salts for the Permanganate Titration. Reduction by Metallic Zinc or Magnesium. Zinc, in spite of many objections, is the favourite method of reduction. The filtrate from the bisulphate fusion is treated with iron-free zinc 3 and dilate sulphuric acid until a drop removed on a 1 E. Kneclit and E. Hibhert, New Reduction Methods in Volumetric Analysis, London, 11, 1910; E. Knecht and E. Hibbert, Ber., 36. 1549, 1903; E. Knecht, ib., 36. 166, 1903. A. Purgotti titrates ferric solutions with an acid solution of molybdenum oxide Mo 3 8 (Gaz. Chim. Ital., 26. ii. 197, 1896). 2 M. Stormer, Tonind. Ztg., 32. 1609, 1908; R. Davidson, Journ. Soc. Chem. Ind., 6. 421, 1887; P. Lehnkering, Zeit. offent. Chem., 4. 459, 1898. 3 It is not easy to get zinc free from iron and carbon. G. F. Rodwell (Chem. Neivs, 3. 4, 1861) found that zinc contained about 1'3 per cent, of a black insoluble residue consisting of about 0'5 per cent, of carbon along with some lead and iron. Carbon, if present, decolorises a certain amount of permanganate, and gives high results (J. B. Mackintosh, Chem. News, 50. 75, 1884). It is generally advisable to make a blank test with, say, 10 grms. of the zinc, and thus find a correction for the zinc used in the reduction. If the zinc be pure, it dissolves very slowly in acids, and even when the pure zinc is in contact with platinum a reduction may occupy 24 hours. L. Moyaux (Rev. Univ. Mines (1), 25. 148, 1869) recommends amalgamated zinc. Amalgamated zinc in contact with platinum is better, but the action may stop owing 187 1 88 A TREATISE ON CHEMICAL ANALYSIS. glass rod gives no reddish-brown coloration with a drop of ammonium thio- cyanate. ' This method is convenient and extensively employed when the titanium is ignored. The solution under investigation is placed in an Erlenmeyer's flask. The FIG. 101. Bunsen's valve. FIG. 102. Binder's valve. solution should not contain more than about O'l grm. of Fe 2 3 per 100 c.c. 1 Add some thin flakes of granulated zinc, or, better still, two or three portions of stick magnesium. 2 Add sufficient sulphuric acid to make a solution containing about 17 per cent, of H 2 S0 4 . 3 The flask is closed with a one-hole rubber stopper fitted with a Bunsen's valve (fig. 101) or a Binder's (fig. 102), 4 or, better, with a to the amalgamation of the platinum during the action (A. L. Beebe, Chem. News, 53. 269, 1886). To amalgamate zinc, shake it in a flask with a solution of mercuric sulphate in 2-5 per cent, sulphuric acid (1 grm. metallic mercury per 100 grms. of zinc). Wash the metal several times with 2*5 per cent, sulphuric acid, and finally with water. G. T. Morgan (A nalyst, 26. 225, 1901 ; J. H. Gladstone, Chem. News, 32. 75, 195, 1875 ; J. H. Gladstone and A. Tribe, ib., 32. 150, 1875) recommends a zinc-copper couple in a 3 per cent, solution of sulphuric acid. The zinc-copper couple is made by immersing, say, 8 grms. of granulated zinc in 200 c.c. of a 10 per cent, solution of copper sulphate. The action is said to be much more rapid than with zinc alone. For the aluminium reduction in hot solutions, see F. J. R. Carulla, Journ. Soc. Chem. Ind., 27. 1049, 1908 ; W. H. Seamon, Chem. Eng., 8. 124, 1908. Aluminium foil does its work more rapidly than zinc, but it too must be corrected for iron. Magnesium ribbon is generally free from iron, phosphorus, and sulphur. It reduces much more rapidly than zinc (S. Kern, Chem. News, 33. 112, 1876 ; H. N. Warren, ib. t 60. 187, 1889), but is liable to float on the surface of the solution. This latter objection does not apply to sticks of metallic magnesium, which is in many ways preferable to zinc but the cost is a little greater. A. Gemmell (Analyst, 35. 65, 1910) recommends zinc-aluminium and magnesium- zinc (magnalium) alloys for some reductions. For reductions with palladium-hydrogen, see L. Kritschewsky, Ueber die Anwendung des metallischen Wasserstoffs in der analytischen Chemie, Bern, 1885 ; W. H. Gintl, Zeit. angew. Chem., 15. 424, 1902 ; A.C. Chapman, Analyst, 29. 346, 1904. For the efficiency of different reducing agents, see A. C. Chapman and H. D. Law, ib., 31. 3, 1906. For reduction with metallic copper, see L. Storch, Ber. Oester. Ges. Forder. Chem. Ind., 15. 9, 1893 ; W. C. Birch, Chem. News, 99. 273, 1909. 1 A. Mitscherlich (Zeit. anal. Chem., 2. 72, 1863) says that there is a danger of loss owing to the precipitation of some metallic iron on the zinc N. W. Fischer, Pogg. Ann., 9. 266, 1827. The iron dissolves when the last trace of zinc dissolves. There is, however, no danger under this head when the solutions are diluted and acidified as indicated in the text. E. Miiller and G. Wegelin (Zeit. anal. Chem., 50. 615, 1911) note the danger indicated by Mitscherlich, and recommend reducing the ferric iron by adding 5-10 drops of N-CuS0 4 and warming the solution with amalgamated zinc rods for a couple of hours. 2 10 grms. of metallic zinc, or 4 grms. of metallic magnesium, usually suffice for reducing O'l grm. of ferric oxide. 3 That is, about 20 c.c. of concentrated sulphuric acid per 100 c.c. of the solution (sp. gr. 1 '8). Note the amount of acid already present in the solution. 4 For a Bunsen's valve (fig. 101), fit the hole in the stopper with a piece of short glass tubing (a, fig. 101), to which is attached a piece of indiarubber tubing plugged at the other end with a piece of glass rod (b}. The rubber tubing has a longitudinal slit (cd). It readily permits the outflow of gas, but offers some opposition to the back flow of air. The valve will sometimes hold so well that the flask will break before the valve gives way, particularly when the contents of the flask have been heated and allowed to cool with the Bunsen's valve in position. R. Fresenius, Anleitung zur quantitativen chemischen Analyse, Braunschweig, 278, 1853 ; A. Krbnig, Pogg. Ann., 122. 170, 1864 ; C. H. Bostock, Chem. News, 57. 213, 1890 ; U. Kreusler, Zeit. anal. Chem., 24. 393, 1885 ; 0. Binder, ib., 27. 178, 1888 ; O. Reitmar and A. Stiitzer, Eep. anal. Chem., 5. 232, 1885 ; L. R. Milford, Journ. Ind. Eng. Chem., 4. 845, 1912. The construction of Binder's valve will be obvious from fig. 102, where c represents a small aperture in the side of a glass tube closed at one end ; and ab, a piece of rubber tubing. THE DETERMINATION OF IRON. 189 Kempf s or a Schiebler's gas- washer as guard tube. 1 These vessels are glass bulbs fitted inside with a syphon tube as shown in fig. 103. Water is poured intp the bulb until the free end of the inner tube just dips below the surface. When the iron is reduced, pour an aqueous solution of sodium bicarbonate into the bulb. As the flask cools, the carbonate solution is drawn into the flask, and carbon dioxide is given off. When equili- brium is established, the contents of the flask are well protected from the air by the carbon dioxide. The flask may be warmed to facilitate the reduction. The reduction is complete when a drop of the solution gives no red coloration with a drop of ammonium thio- cyanate. 2 The flask may be heated until all the zinc is dissolved ; or the undissolved zinc and carbon may be filtered off through glass-wool. There is a great risk of re-oxidation during the filtration. 3 If the titanium is to_be neglected, the solution can be titra.f.pH a.f, vnt,h sta.rirkrrj permanganate as described belqw. Effect of Titanium. When titanium is present, 4 the zinc reduces, more or less completely, 5 the titanic oxide to titanium sesquioxide : 2Ti0 2 ->Ti 2 3 , etc. The FIG. 103. Reduction in flask fitted with guard tube. latter is re-oxidised to titanic oxide, Ti0 2 , during the permanganate titration. 6 In consequence, more or less 7 titanium is estimated as if it were iron. It is by no means uncommon to find clays with ^ to 4 per cent, of titanic oxide, and such clays would be reported with J to 4 per cent, of ferric oxide more than that actually present. This is a serious matter. 8 If, therefore, the zinc reduction be adopted, the titanium should be re-oxidised by adding a little bismuth oxide to the reduced solution. According to Gooch and Newton, 9 this 1 T. Kempf, Zeit. anal. Chem., 7. 442, 1868 ; W. T. K. Stock, Chem. News, 39. 46, 1879 ; R. Jahoda, Zeit. angew. Chem., 12. 87, 1890 ; H. Gockel, ib., 12. 620, 1899 ; Chem. Ztg., 37. 235, 1913 ; A. Contat, ib., 22. 298, 1895 ; M. Spang, ib., 36. 1465, 1912 ; M. Mittenzwey, Journ. praJct. Chem. (1), 91. 86, 1864. Several other devices have been used. 2 This test will indicate one part of iron in 1,600,000 parts of water A. Wagner, Zeit. anal. Chem., 20. 349, 1881; E. F. Smith (ib., 19. 350, 1880) says 1 in 80,000,000. A. Ebeling (Zeit. offent. Chem., 8. 144, 1901) adds the potassium thiocyanate to the solution being reduced; J. Volhard (Zeit. angew. Chem., 14. 609, 1901) says Ebeling's method is not trust- worthy, because (1) nascent hydrogen reduces the KCNS and leads to low results; (2) any excess of KCNS leads to an increased consumption of the permanganate. 3 According to M. M. P. Muir (Chem. Neivs, 97. 50, 1908), the addition of 100 c.c. of a saturated solution of mercuric sulphate stops the reaction. MERCURIC SULPHATE SOLUTION. Mix 20 grms. of mercuric sulphate with 8 c.c. of concentrated sulphuric acid, and stir up the pasty mass with 80 c.c. of water. If a yellow precipitate separates, add more sulphuric acid. Note the danger suggested by Mitscherlich, page 188. 4 The solution reduces more quickly apparently owing to the catalytic action of titanic oxide. 5 See page 199. 6 F. 0. von der Pfordten, Liebig's Ann., 234. 257, 1886; 237. 201, 1887; F. Pisani, Compt. Rend., 59. 289, 1864 ; A. Gemmel, Analyst, 35. 198, 1910 ; C. Marignac, Zeit. anal. Chem., 7. 112, 1868; E. Wiegand, ib., 21. 510, 1882; H. A. Wells and W. L. Mitchell, Journ. Amer. Chem. Soc., 17. 78, 1895; G. Gallo, Atti Accad. Lincei (5), 16. i. 525, 1907 ; H. D. Newton, Amer. J. Science (4), 25. 130, 1907 ; Chem. News, 98. 134, 1908 ; F. W. Hinrichsen, Chem. Ztg., 31. 738, 1907 ; E. Knecht and E. Hibbert. Analyst, 36. 96, 1911. 7 " More or less" because, while acid solutions of ferrous sulphate oxidise slowly on exposure to air, titanous sulphate oxidises very quickly, and this oxidation is the principal difficulty in the volumetric determination of titanium by this reaction. 8 For the effect of vanadium, see pages 467 and 480. 9 F. A. Gooch and H. D. Newton, Amer. J. Science (4), 23. 365, 1907 ; Zeit. anorg. Chem., 54. 213, 1907; Chem. News, 96. 148, 1907; H. D. Newton, ib., 98. 218, 1908; Amer. J. 190 A TREATISE ON CHEMICAL ANALYSIS. treatment will oxidise the Ti 2 3 without affecting the ferrous salt. The reduced solution should be filtered in order to remove the unreduced bismuth oxide and bismuth.' In illustration of the effect of bismuth oxide, Gooch and Newton quote the following test experiments : Table XXXIV. Effect of Bismuth Oxide in inhibiting the Effect of Titanium in the Permanganate Process for Iron. Grm. Ti0 2 taken. Grm. Fe 2 3 taken. Grm. Fe 2 3 found. Error. 0-04 0-0993 0-0992 -0-0001 0-06 0-0993 0-0993 o-oooo 0'08 0-0993 0-0997 + 0-0004 o-i 0-0993 0-0997 + 0-0004 0-2 0-0993 0-0997 + 0-0004 o-i 0-1986 0-1986 o-oooo It is difficult to filter a ferrous solution without re-oxidation of a portion of the ferrous salt. A filtration flask can be fitted with a filter tube packed with a layer of broken glass beneath a layer of calcined quartz sand. A couple of fragments of magnesite (iron-free) are placed in the filtration flask along with about 2 c.c. of dilute sulphuric acid ; another couple of fragments of magnesite are placed in the filter tube, and in the flask containing the reduced solution. The filter flask is connected with the pump, and the solution under investigation is poured into the filter tube. The flask is washed with dilute sulphuric acid with a couple of fragments of magnesite placed in the flask at each washing. In this way, an atmosphere of carbon dioxide may be kept between the solution con- taining the reduced iron and the atmosphere. 1 The solution in the filtration flask may then be titrated with permanganate without interference from the titanium. The Reductor. The apparatus suggested by Jones 2 can be employed for rapid zinc reductions. A modification is illustrated in fig. 104. A piece of glass tubing 40 to 45 cm. long and 12 to 15 mm. internal diameter is drawn out at one end, and there fitted into the neck of a filtration flask by means of a one-hole rubber stopper. The other end of the tube is fitted with a stoppered funnel, or rubber tube and pinchcock. Put a filter plate or a few pieces of broken glass into the lower portion of the tube, then about 2J cm. of coarse, clean, calcined quartz sand. Fill the tube with about 200 to 300 grms. of amalgamated or ordinary zinc powder 3 granulated to pass a 20's lawn, and remain on a 30's lawn. The zinc should be as free from iron as possible. 4 To Science (4), 25. 343, 1908 ; R. Rieke and R. Betzel, Sprech. Archiv, I. 45, 1912. For copper sulphate and copper oxide instead of bismuth oxide, see W. C. Birch, Chem. News, 99. 272, 1909 ; A. Storch, Ber. Oester. Ges. Forder. Chem. 2nd., 15. 9, 1893. 1 For more elaborate schemes, see W. Bachmeyer, Zeit. anal. Chem., 24. 59, 1885; N. von Klobukow, ib., 24. 395, 1885. 2 D. J. Carnegie, Journ. Chem. Soc., 53. 468, 1889; C. Jones, Trans. Amer. Inst. Min. Eng., 17. 414, 1889 ; Chem. News, 60. 93, 1889 ; P. W. Shinier, Journ. Amer. Chem. Soc., 21. 723, 1899 ; E. H. Miller and H. Frank, ib., 25. 919, 1903 ; D. L. Randall, ib., 28. 389, 1900 ; Amer. J. Science (4), 24. 313, 1907 ; Chem. News, 97. 113, 1908; C. B. Dudley and F. N. Pease, Journ. Anal. App. Chem., 7. 108, 1893 ; F. L. Crobough, ib. t 6. 366, 1892. 3 The use of amalgamated zinc (proposed by A. J. McKenna) in place of ordinary zinc, enables a shorter reductor to be used. This is due to the more vigorous action of the amalgamated zinc. 4 Commercial zinc dust is generally less pure than the granulated metal. It contains cadmium, lead, and zinc oxide, together with iron, etc. Zinc, at 210, is very brittle, and it THE DETERMINATION OF IRON. use the apparatus, close the stopcock ; pour 100 c.c. of cold dilute sulphuric acid (1 : 20) l into the funnel. Apply suction at the nitration flask, since the hydrogen evolved retards the percolation of the acid through the tube. Open the stopcock to allow the acid to pass slowly through the column of granulated zinc. Wash the zinc free from acid with about five rinsings of distilled water. Repeat the treatment with acid and water. Close the stopcock so that the tube remains full of water. Disconnect the nitration flask, wash and restore it to its place. The tube is now ready for use. Pour the iron solution 2 to be re- duced into the funnel. Connect the filtration flask with the purnp. Open the stopcock cautiously, since the rate of flow of the solution to be reduced is regulated by the stopcock. As the solution flows through the column of zinc, complete reduction occurs, and the solution collects in the filtration flask ready for titration or for treat- ment with bismuth oxide. Just before the funnel is emptied, spirt water round the sides, and also rinse the beaker with water. Let the washings collect in the funnel, and ultimately pass them through the column of zinc. 3 Wash the column of zinc with distilled water as before. The re- ductor is then ready for another reduction. The reductor should be tested from time to time with blank tests. A correction for the iron in the zinc can be made by re-reducing the solution used in the permanganate titration. The amalgamated zinc in the tube charged as indicated above w 7 ill suffice for thirty to forty reductions. Reduction by Ammonium Bisulphite. Hydrogen sulphide, sulphur dioxide, can be easily reduced to fine grains by trituration at this temperature T. AI. Brown, Iron, 12. 361, 1878; Dingler's Jonrn., 22$. 378, 1879. 1 The ratio of free sulphuric acid (sp. gr. 1*84) to the total solution should be between 1 : 5 and 1:6. If more acid be present than 1 : 5, zinc sulphate is inclined to crystallise in the reductor ; and if less than 1 : 7 be present, the reduction may not be complete in the case of uranium solutions E. F. Kern, Journ. Amer. Chem. Soc., 23. 685, 1901 ; F. Ibbotson and S. G. Clark, Chem. News, 103. 146, 1911. A. A. Blair (The Chemical Analysis of Iron, Philadelphia, 92, 1908) used the reductor for phosphorus determination ; F. A. Gooch and G. Edgar (Chem. Neius, 87. 265, 1903 ; Amer. J. Science (4), 25. 233, 1908) for vanadium ; G. Edgar (ib. (4), 25. 332, 1908) for vanadium and molybdenum ; ' and by E. F. Kern (Journ. Amer. Chem. Soc., 23. 685, 1901 ; F. Ibbotson and S. G. Clark, Chem. News, 103. 146, 1911) for uranium ; see phosphorus, page 599. 2 The iron solution should not contain more than about 15 c.c. of concentrated sulphuric acid per 100 c. c. of solution. 3 In washing the reductor free from iron the water should be kept above the level of the zinc, so as not to allow auy air spaces between the successive additions of water. There is other- wise a possible formation of hydrogen peroxide which might spoil the results. FIG. 104. Reductor. 192 A TREATISE ON CHEMICAL ANALYSIS. ammonium bisulphite, or sodium sulphite can be used for reducing the iron. These agents have the advantage of leaving the titanic oxide unaffected. Ammonium bisulphite, recommended by Austin and Hurff, 1 is convenient. It can be either made or purchased. 2 The solution under investigation is con- centrated by evaporation to 40 or 50 c.c. Gradually add about 10 c.c. of dilute sulphuric acid (1:1) to the solution in an Erlenmeyer's flask, so as to make the solution distinctly acid. 3 Add a concentrated solution of ammonium bisulphite ; 4 and agitate the mixture thoroughly ; or pass a current of sulphur dioxide through the acid solution. Gradually raise the temperature of the solution to boiling, and when a drop no longer gives a brownish-red coloration with ammonium thiocyanate, the ferric oxide is all reduced. To remove the excess of sulphurous acid, place the flask on a sheet of asbestos on a tripod. Add 15 c.c. of dilute sulphuric acid (1 :1). Cover the flask with the perforated lid of a Rose's crucible. 5 Pass a current of carbon dioxide at the rate of 3 or 4 bubbles per second through the solution. 6 The delivery tube passes through the hole in the cover of the flask. Meanwhile the flask is heated to the boiling point of the solution. The velocity of the carbon dioxide is then reduced to about 1 bubble per second. After 20 or 30 minutes' boiling, the escaping steam will probably be free from sulphur dioxide, as shown by its failure to discolour a mercurous nitrate test paper. 7 In that case, place the flask in a dish of cold water to cool, while the current of carbon dioxide still bubbles through the solution 1 bubble per second. 8 1 P. T. Austen and G. B. Hurff, Chem. News, 46. 287, 1882 ; Amer. Chem. Journ., 4. 282, 1882 ; T. W. Hogg, Chem. News, 59. 207, 1889 ; R. W. Atkinson, ib., 49. 217, 1884. According to B. Glasmann (Zeit. anal. Chem. , 43. 506, 1904), the ammonium sulphite does not reduce bichromic sulphate and hence the presence of chromium does not interfere. 2 AMMONIUM BISULPHITE. Pass sulphur dioxide into a concentrated solution of ammonia until the solution becomes yellow and smells strongly of sulphur dioxide. If the solution be kept cool during the passage of the gas, white crystals of neutral sulphite are formed. These are gradually dissolved by the excess of sulphur dioxide, and the solution becomes clear yellow. For the sulphur dioxide, use either a syphon of the liquid gas, or boil copper with concentrated sulphuric acid in the proportion 250 grms. of copper and 500 c.c. of concentrated sulphuric acid. 18 c.c. of this solution will reduce about 10 grms. of ferric oxide. G. Neumann (Ber. t 20. 1584, 1881) makes sulphur dioxide in a Kipp's apparatus by the action of concentrated sulphuric acid on cubes made from three parts calcium sulphite and one part plaster of Paris. This method is not suited for very large quantities of the gas. Sulphur dioxide is best purchased in "syphons" of the liquefied gas. The "syphons" are a convenient source of the gas for analytical work. 3 The reduction of ferric solutions by sulphur dioxide proceeds rapidly if the solution contains a little free acid, but not if the solution is alkaline to litmus A. (J. Gumming and E. W. Hamilton, Proc. Roy. Soc. Edin., 32. 12, 1912. 4 1 c.c. of the solution per 0'5 grm. of Fe 2 3 . 5 Or bend down the corners of a piece of platinum foil with two holes : one hole permits the escape of steam and gas ; the other is for the gas delivery tube. 6 The carbon dioxide is washed by passing it through a column of pumice soaked in copper sulphate and a water wash-bottle. 7 MERCUROUS NITRATE TEST PAPER. A piece of mercurous nitrate test paper held in the vapour becomes brownish black if sulphur dioxide be in the escaping steam. The paper is made by soaking No. 00 Swedish filter paper in a solution of mercurous nitrate (0'5 to 5 per cent.) and drying. The paper is cut in strips and preserved for use in dry glass tubes, corked and sealed. H. Jervis (Chem. News, 77. 133, 189S) closes the flask with a stopper fitted with a syphon-like tube, one end of which is dipped into an acid, very dilute solution of permanganate. If the solution be completely decolorised (without a brown precipitate), sulphur dioxide is still being evolved. 8 As a matter of fact, it is rather difficult to remove the last trace of sulphur dioxide, and if all be not removed the iron determination will be high. The condensation of water in the neck of the flask and on the stopper is the main source of the trouble, for the condensed water absorbs some sulphur dioxide and, on dropping back into the solution, returns some sulphur dioxide. For the retention of sulphur dioxide by rubber stoppers and rubber tubing, see E. W. Hamburger, Zeit. physiol. Chem., 2. 191, 1878 ; 4. 249, 1880 ; K. H. Huppert, ib., 17. 87, 1893. THE DETERMINATION OF IRON. 193 When cold, the contents of the flask can be titrated with a standard solution of potassium permanganate, as indicated below. The titanium oxide is not reduced by the treatment with ammonium bisulphite. 90. The Standardisation and Use of Potassium Permanganate Solution. The "commercially pure" potassium permanganate picks up dust, etc., and is rather poorer in oxygen than theory requires, 1 although, by repeated crystallisation from boiling water and careful drying the salt can be prepared to give a solution of exactly theoretical strength. 2 It is, however, the invariable custom to find the strength, that is, to standardise the solution of potassium permanganate. Standard Solution of Sodium Oxalate. Dissolve about a gram of potassium permanganate (KMn0 4 , molecular weight 158-03) in a litre of water, and let the solution stand two or three weeks in order that the permanganate may oxidise any impurities in the water before the solution is standardised by titra- tion with, say, a solution of sodium oxalate of known strength. Other salts are in common use, but sodium oxalate gives exact results with very little trouble. Whatever salt be selected, the standard solution must be made up with the greatest care, since all the subsequent results depend upon the accuracy of the work at this stage. Sorensen's sodium oxalate 3 (Na 2 C 2 4 , molecular weight 134) is the best "brand" for the purpose. This salt is dried in a water oven at about 200 and cooled in a desiccator over calcium chloride. Transfer 2 -13 grms. of the salt to the litre flask. Dissolve the salt in a little water, make the solution up to a litre mark. One cubic centimetre will be equivalent to 0-001 grm. of KMn0 4 , or to 0-00253 grm. Fe ? 8 . This solution can be used for standardising the permanganate. The equivalents of other substances can easily be expressed in terms of sodium oxalate by means of the usual chemical equation (page 194). The following table of conversions may be useful for reference : Table XXXV. Conversion Table for Permanganate Solutions. is equivalent to One gram of FeS0 4 Fe. FeO. K 2 2 7 . Fe 2 3 . Na 2 C 2 4 . KMn0 4 . (NH 4 ) 2 S0 4 CaO. 6H 2 0. Iron 1-0000 1-28613 0-8782 1-42919 1-1992 0-56555 7-01514 0-50198 Sodium oxalate 0-8336 1-0724 0-7324 1-1918 1-0000 0-4716 5-8369 0-4186 Potassium 176819 2-27412 1-5528 2 52708 2-1204 1-0000 12-404 0-8876 permanganate Ferrous 0-14255 0-18334 0-1249 0-20373 0-17055 0-08062 1-0000 0-07139 ammonium sulphate Potassium 1-1383 1-4644 1-0000 1-6274 1-3655 64388 8-0064 0-5716 bichromate 1 According to F. Raschig (Zeit. angew. Chem., 16. 585, 1904), a solution prepared from commercial permanganate was about "8 per cent, too weak. 2 W. M. Gardner, B. North, and A. R. Naylor, Journ. Soc. Chem. Ind., 22. 731, 1903. 3 S. P. L. Sorensen, Zeit. anal. Chem., 36. 639, 1897 ; 42. 333, 512, 1903; 44. 141, 1905 ; S. P. L. Sorensen and A. C. Andersen, ib., 44. 156, 1905 ; L. Vanino and E. Seitter, ib., 41. 13 194 A TREATISE ON CHEMICAL ANALYSIS. Standardisation of Potassium Permanganate with Sodium Oxalate. Pipette 25 c.c. of the standard sodium oxalate solution into each of three Erlenmeyer's flasks; add 10 c.c. of dilute sulphuric acid (1 : 4). Place the flask under a glass- stoppered burette filled with the permanganate solution to be standardised, and warm the oxalate solution to between 60 and 70. Run permanganate from the burette into the flask, and keep the contents of the flask in a state of rotary agita- tion while the permanganate is being added. The colour of the permanganate is discharged slowly at first. It is an advantage to work with the solution at about 60 or 70 in order to accelerate the action. When the reaction first starts, it proceeds very slowly ; afterwards it progresses more rapidly. When the colour of the permanganate begins to disappear slowly, add the permanganate very cautiously, drop by drop. Any drops of permanganate adhering to the sides of the flask must be washed into the bulk of the liquid. When a drop of the permanganate produces a permanent pink blush throughout the liquid in the flask, the titration is finished. Read the burette to the nearest tenth. Take the mean of three such titrations. Let n denote the number of cubic centimetres of the permanganate solution required for the titration of the 25 c.c. of sodium oxalate solution. The reaction is symbolised : 5Na 2 C 2 4 + 2KMn0 4 + 8H 2 S0 4 = K 2 S0 4 + 5Na 2 S0 4 + 2MnS0 4 + 10C0 2 + 8H 2 0. Since 1000 c.c. of the sodium oxalate solution are equivalent to 1 grm. of KMn0 4 , 25 c.c. will be equivalent to 0*025 grm. of KMn0 4 . Hence, n c.c. of the per- manganate solution has the equivalent of 0*025 grm. KMn0 4 , and this is equivalent to 0*06317 grm. of Fe 2 3 . Or, 1 c.c. of the permanganate solution is equivalent to 0*0632 -n grm. Fe 2 3 . The error with careful work is about one part in a thousand. Influence of Acids. Sulphuric acid is required to decompose the oxalate, and in titrating with permanganate, free acid is needed to keep the manganese oxides in solution. Otherwise, the brown precipitate which separates will obscure the end point. Too much sulphuric acid may lead to high results, since concentrated sulphuric acid reduces potassium permanganate. This is illustrated by the following experiments, 1 with 20 c.c. of an acidified solution of ferrous sulphate containing the equivalent of 0*080 grm. Fe 2 3 , and with various pro- portions of sulphuric acid (1 : 4). The volume of the permanganate solution re- quired with different quantities of the acid is indicated by the following numbers : H 2 S0 4 ... 10 20 25 40 50 c.c. KMn0 4 . . . 17'89 17*85 17*89 18'07 18*20 18*62 c.c. To show that the increased consumption of permanganate was not due to impurities in the sulphuric acid, 150 c.c. of water and 50 c.c. of the same sulphuric acid were added to another 50 c.c. of the ferrous solution. The consumption of permanganate was practically the same as when a solution of the ferrous salt diluted with water only was employed. Hydrochloric acid is baneful because some permanganate is used in the side reaction, possibly : 16HC1 + 2KMn0 4 = 5C1 2 + 2MnCl 2 + 2KC1 + 8H 2 0, 141, 1902; A. Skrabal, ib., 42. 359, 1903 ; C. Meineke, ib., 39. 322, 1900 ; G. Lunge, Zeit. angew. Chem., 17. 230, 269, 1904; 18. 1520, 1905; F. Dupre and E. Miiller, ib., 15. 1244, 1902; F. Dupre, ib., 17. 815, 1904; F. Dupre and A. von Kiipffer, ib., 15. 352, 1902; W. Schranz, Bull. Soc. Chim*. (3), 18. 89, 1899; H. Kinder, Chem. Ztg., 30. 631, 814, 1906; 31. 69, 1907 ; P. Lehnkering, ib., 30. 723, 1906 ; H. von Jiiptner, Oester. Zeit. Berg. Hiitt., 44. 14, 1896 ; S. M 'Bride, Journ. Amer. Chem. Soc., 34. 393, 1912 ; Circ. Bur. Standards, 40. 3, 1912. 1 E. Waitz, Zeit. anal. Chem., 10. 158, 1870 ; ,T. P. Blunt, Chem. News, 8. 54, 1863. THE DETERMINATION OF IRON. 195 which proceeds slowly in dilute solutions, but is accelerated by the presence of iron salts. Similar experiments to those indicated in the last paragraph were made with hydrochloric acid (sp. gr. 1-13). The results were : HC1 . . . 5 10 15 20 30 c.c. KMn0 4 . . 17-89 19'25 1975 20'12 20'50 20'50 c.c. The side reaction is inhibited by the addition of manganese sulphate, as indi- cated on page 451. Nitric acid should be absent, because it is reduced to nitrous acid by the zinc in the reduction of the ferric oxide, and nitrous acid reduces the per- manganate. 1 Standardisation of Potassium Permanganate with Metallic Iron. It is nearly always advisable, if convenient, to standardise a volumetric solution against a solution similar to that which is going to be titrated. Hence, many prefer to standardise the permanganate solution for iron determinations against a solution containing a known amount of metallic iron dissolved in, say, sulphuric acid. The difficulty is with the iron. Perfectly pure iron is exceedingly difficult to procure. The trace of impurity present in most specimens of "pure " iron acts as a reducing agent on the permanganate. Even electrolytic "iron" is not quite free from objection. 2 It is frequently assumed that iron wire contains on an average 0'3 per cent, of foreign impurities the variations lie usually between 0*1 and 0'4 per cent. The amount of iron actually weighed out is accordingly multiplied by O997 in order to estimate the strength of the solu- tion. It is then supposed that the impurities in the iron wire diminish the amount of metal acted upon by the permanganate and exert no other reducing action. This assumption is not justified. The impurities, carbon, sulphur, etc., are present as carbides, sulphides, phosphides, silicides, etc., and these substances develop hydrocarbons, hydrogen sulphide, hydrogen phosphide, etc., when the iron wire is dissolved in sulphuric acid. These products are easily oxidised by permanganate. The consequence is that more permanganate is used than corresponds with the amount of iron in the wire, and the permanganate titre is accordingly too low. An error ranging from 1 to 2 per cent, may be committed in standardising potassium permanganate by means of iron wire assumed to represent pure iron, and the error is increased instead of diminished by assuming that the wire contains less than its own weight of iron. 3 Samples of "analysed" iron can be purchased, but these are not satisfactory for standardising the permanganate unless their iron value has been determined in terms of the permanganate. This is a troublesome determination in a technical laboratory. 4 Satisfactory samples with a known permanganate 1 A. Terreil, ZeiL anal. Chem., 6. 116, 1867. 2 C. F. Roberts, Amer. J. Science (3), 48. 290, 1894 ; Chem. News, 70. 189, 1894 ; F. P. Treadwell, Kurzes Lehrbuch der analytischen Chemie, Leipzig, 2. 499, 1911 ; A. Classen, Zeit. anal. Chem., 242. 516, 1903; F. Mohr, Lehrbuch der chemisch. analyt. Titrier methods, Braun- schweig, 215, 1896 ; A. Skrabel, ZeiL anal. Chem., 42. 395, 741, 1903 ; 43. 97, 1904 ; G. Lunge, L. Brandt, ib., 32. 812, 830, 840, 851, 1908 ; C. A. Kohn, F. J. Brislee, and H. H. Froysell, B. A. Rep., 174, 1900; L. Moyaux, Rev. Univ. Mines (1), 25. 148, 1869; H. Cantoni and 'M. Basadonna, Ann. Chim. Anal., g. 365, 1904 ; A. Ledebur, Stghl Eisen, 22. 1242, 1902 ; H. Kinder, ib., 30. 411, 1910 ; L. Brandt, ib. t 30. 1844, 1910 ; W. M. Gardner, B. North, arid A. R. Naylor, Journ. Sac. Chem. Ind., 22. 731, 1903. 3 J. R. M. Irby, Chem. News, 30. 142, 1874 ; M. Berthelot, Bull. Soc. Chim. (2), 21. 58, 1874. 4 This is made by using electrolytic iron or sodium oxalate. 196 A TREATISE ON CHEMICAL ANALYSIS. titre can be purchased. 1 The iron wire is cleaned, if necessary, with sand- paper to remove grease or oil, and dissolved in sulphuric acid at a gentle heat in a 200-c.c. flask fitted like tig. 103, and from which the air has been previously expelled by a stream of carbon dioxide. The cold solution is titrated with permanganate, as previously described. Higher results are obtained if the solution is made on a water bath in a flask fitted with a Bunsen's valve than if the iron be dissolved in a similar flask kept for a long time "on the boil." In the former case, the solution contains more of the impurities than in the latter case. Similarly, if the dissolution be conducted in a flask through which a current of carbon dioxide is passing, the results depend upon the duration of the boiling and the speed at which the carbon dioxide is passed. But if the carbon dioxide be passed for a long time, the results are the same whether the iron be dissolved on the water bath or at a boiling temperature. Suppose 1 grm. of " analysed " iron wire contains the equivalent of 0'9991 grm. iron so far as the permanganate solution is concerned, this amount of iron is equivalent to 0'9991 x 1-4292 = 1-4279 grms. of Fe 2 3 . If w gram of the sample of iron be taken, and n c.c. of permanganate are required for the titration, obviously, 1 c.c. of the permanganate is equivalent to 1*4279 w-^-n grm. of ferric oxide. However, the manipulations with Sorensen's sodium oxalate are a little easier than with metallic iron, and the results appear quite as satisfactory. 2 Preservation of Permanganate Solutions. Permanganate solutions should be stored in a bottle with a well-fitting glass stopper. The stopper must not be greased. The bottle must be protected from direct sunlight. Solutions con- taining suspended oxide are liable to lose strength, and this with increasing rapidity with the lapse of time owing to the increase in the quantity of the suspended oxide. 3 Solutions of permanganate, freed from suspended oxide by filtration through asbestos, may have a lower " titre " after filtration than before, since asbestos exerts a slight reducing action on solutions of potassium permanganate. 4 The solution will keep a long time if it be kept out of contact with organic matter, dust, etc. 5 For the best work the solution should be standardised about every two months, but after keeping through the summer months, about nine months, a solution with a titre 1 c.c. = 0*002576 grm. Fe 2 3 only changed to 1 c.c. = 0-002522 grm. Fe 2 3 , that is, about 2 per cent. The stock bottle should be shaken before use in order to remove any water condensed in the upper part of the bottle by distillation from below. Burettes for the Permanganate Titration. Glass-stoppered burettes must 1 A. Miiller, Stahl Eisen, 26. 1477, 1906. A sample recently purchased contained : carbon, G'027; silicon, 0'013 ; phosphorus, 0'034 ; copper, 0'024 ; manganese, 0'005 ; sulphur, O'OOS total impurities, O'lll per cent. The "iron value" of the sample in terms of permanganate was equivalent to 99 '91 per cent, metallic iron. 2 Note that if the ' ' permanganate titre " of the iron wire be determined by titration with sodium oxalate, the use of the iron wire appears redundant. H. N. Morse, A. J. Hopkins, and M. S. Walker, Amer.' Chem. Journ., 18. 401, 1896 ; er., 26. 267, 1893 ; W. Blum, Journ. Amer. Chem. Soc., 34. 1379, 1912. 4 P. A. Tscheihwile, Journ. tiuss. Phys. Chem. Soc., 42. 856, 1910. 5 R. W. Oddy and J. B. Cohen, Journ, Soc. Chem. Ind., g. 17, 1890; W. M. Gardner and B. North, ib. 23. 599, 1904 ; B. Collitt, Pharm. Journ. (4), 27. 724, 1908 ; R. Luboldt, Journ. prakt. Chem. (1), 77. 315, 1859 : G. Bruhns, Zentr. Zuckerind., 14. 968, 1906 ; H. von Jiiptner, Oester. Zeit. Berg. Hiitt., 31. 502, 1883; C. Meineke and K. Schroder, Zeit. dffent. Chem., 3. 5, 1897; F. Simond (Dingier* 8 Journ., 248. 518, 1883) coats the store bottle with black varnish on the outside in order to protect the contents from light. THE DETERMINATION OF IRON. 197 be employed ; for regular work that shown in fig. 105 or in fig. 18 (right) is excellent. The solution of permanganate must not come in contact with organic matter from, say, rubber jets, lubricant of stopcock, etc. Scheibler 1 recommends a burette shaped like fig. 106. The flow of liquid from the burette is regulated by the stopcock or pinchcock A. The tube B is stuffed with glass- or cotton-wool to filter the air from dust before entering the burette. The burette is filled by dipping the nozzle at C in the fluid and applying suction at D. The nozzle is protected by a ground cap C when the burette FIG. 1 05. Burette for permanganate solutions. FIG. 106. Scheibler's burette for permanganate solutions. is not in use. These burettes work well provided changes in temperature or pressure do not occur while the solution is in the burette, otherwise the solution may be forced from the jet by the expansion of the imprisoned air on a rising temperature or a diminishing pressure, and there is then an objectionable "sloppiness" about the nozzle of the burette. The advantage claimed is that 1 C. Scheibler, Journ. prakt. Chem. (1), 71. 245, 1857 ; F. Stober, ib. (1), 59. 599, 1853 ; N. H. Barton, Journ. Amer. Chem. Soc.,3. 124, 1881 ; M. Kleinert, Zeit. anal. Chem., 17. 183, 1878; K. Abraham, ib., 22. 28, 1883; W. Alexandron, ib., 49. 436, 1910; T. H. Garret, Chem. News, 56. 185, 1887 ; M. Heriot and R. Biggs, ib., 26. 189, 1872 ; E. Collens, ib., 9. 94, 1864; 26. 203, 227, 1872; A. V. Harcourt, ib., 26. 239, 1872; C. Jones, Journ. anal. Chem., i. 179, 1887 ; E. Hubner, Kept. anal. Chem., 4. 273, 1884 ; M. Vogther, Archiv Pharm. (3), 22. 539, 189. IQO A TREATISE ON CHEMICAL ANALYSIS. there is no danger of contamination by the lubricant of the stopcock. I prefer the ordinary glass-stoppered burette. Burettes and other glass apparatus which have been used for permanganate solutions should be cleaned and rinsed immediately after use. Brown stains of manganese oxide can be removed by means of hydrochloric acid. Calculations. Obviously, if n c.c. of KMn0 4 are consumed in a titration, and w grms. of the clay are being investigated, the clay will contain 0'0632?i -t-w per cent, of Fe 2 3 . The reaction between the permanganate and the ferrous salt is : 10FeS0 4 + 2KMn0 4 + 8H 2 S0 4 = K 2 S0 4 + 5Fe 2 (S0 4 ) 3 + 2MnS0 4 + 8H 2 0. EXAMPLE. The results of the permanganate titration will be treated in the note- book somewhat as follows : 1 c.c. of the permanganate used represented 0'001510 grin. Fe 2 O 3 ; 1 grm. of clay was treated ; the reduction was complete after 10 grms. of zinc had been dissolved in the solution, and the solution required 13'84 c.c. of permanganate at 16. In a blank experiment, 10 grms. of zinc dissolved in sulphuric acid required 0'51 c.c. of the permanganate solution. Hence, 13"84- 0'51 = 13'33 c.c. represents the permanganate consumption by the iron salts in the gram of clay. This is equivalent to 13-33 x 0-00151 =0-0201 grm. Fe 2 3 ; or the clay has 2'01 per cent, of Fe 2 3 . There was no need to apply the temperature correction. 1 Temperature Corrections. It is necessary to allow for the effect of variations of temperature exceeding 2 or + 3 from the standard. Assume that the solution is standardised at 15 ; then, if the solution be more concentrated than T J^N-KMn0 4 , the correction table for water, page 29, may be applied. If the solution be between YfrN- and j-J^N-KMnO^ the following table may be employed : 2 Table XXXVI. Temperature Corrections for Potassium Permanganate Solutions. (Standard temperature 15.) 1 2 3 4 5 6 7 8 9 + 0-6 + 0-6 + 0-5 + 0'5 + 0'5 1 + 0-4 + 0-4 + 0-3 + 0-2 + 0-1 -0-2 -0-3 -0-5 -0-6 2 -0-8 -ro -1-2 -1-4 -1-6 -1-8 -2-0 -2-3 -2-5 -2'8 The method of using the table will be obvious from the examples on page 30. 91. The Volumetric Determination of Iron Marguerite's Permanganate Process. 3 Starting with the solution reduced by the ammonium bisulphite through which carbon dioxide is passing (page 192), remove and wash the carbon dioxide delivery tube inside and out into the flask with recently boiled distilled 1 After the titration, the ferric oxide may be again reduced and the manganic oxide, if present, removed by filtration. The titration can thus be repeated until the volume of the liquid becomes unwieldy B. Godwin, Amer. J. Science (2), 50. 249, 1870 ; Chem. News, 22. 269, 1870. 2 The table may be also employed for N-, T VN" , and T ^N-solutions of sodium chloride ; T VN- and Y^N-solutions of silver nitrate ; ammonium thiocyanate (1000 c.c. equivalent to 10 grms. Ag). The table is based on the work of A. Schulze (Zeit. anal. Chem., 22. 167, 1882), and may be modified for standard temperatures of reference other than 15. 3 F. Marguerite, Compt. Rend., 22. 587, 1846 ; Ann. Chim. Phys. (3), 18. 244, 1846. THE DETERMINATION OF IRON. 199 water. 1 Run in the standardised permanganate solution exactly as described in the preceding section, until a permanent pink blush is suffused throughout the liquid being titrated. 2 If desired, an allowance can be made if the vanadium be afterwards determined. Errors. The most frequent sources of error arise from the imperfect reduction of the ferric oxide and the deterioration of the permanganate. The latter difficulty is easily overcome by frequently standardising the permanganate ; and the former by testing the solution for the presence of ferric, salts before the reduction is stopped. Care must also be taken that no re-oxidation occurs before the titration. There is also a slight error due to the incomplete reduction of the permanganate, which was pointed out by Bray. 3 This occurs more particularly with low-temperature titrations, high acid concentration, and with a large volume of dilute solution. The conditions here recommended give most favourable results. In illustration of the effect of titanium on the iron determination, and also in illustration of the magnitude of the errors which may be expected in the iron determination both by the ammonium bisulphite and the zinc reductions, the following data were obtained with eight independent analyses on one sample of clay. One gram of clay was taken for each determination, and the bisulphate fusion was divided into two equal portions, one reduced with zinc, and the other with ammonium bisulphite. Table XXXVII. Comparison of Zinc and Sulphite Reductions in the Permanganate Process for Iron in Clays. Zinc reduction. Ammonium bisulphite. c.c. KMn0 4 . Fe 2 3 . c.c. KMn0 4 . Fe 2 3 - 6'92 o-oioo 5-15 0-0078- 6-85 0-0099 470 0-0069 6-84 0-0099 5-27 0-0079 6-32 0-0091 5-34 0-0076 6'61 0-0096 4-89 0-0073 6-09 0-0088 4'43 0-0068 6'49 0-0094 5-13 0-0077 6-19 0-0089 5-01 0-0075 The mean value for the ferric oxide with the zinc reduction is 0-0094, or 1-88 per cent., and with the bisulphite reduction, 0*0075, or 1'50 per cent., with a deviation of about 0'06. 4 The clay contained 1 -2 per cent, of titanium. If the clay contains vanadium, 1 If the solution has been reduced with zinc, with or without the bismuth oxide treatment, proceed with the titration as indicated in the text. It must be added that particles of filter paper may also reduce both permanganate and hot dichromate solutions, and thus lead to high results. The filter paper may get into the solution when the ammonia precipitate is dissolved in dilute acid, and titrated directly. 2 The solution must not be too hot, or an excess of permanganate will be required. Some prefer to work with cold solutions ; others with solutions between 60 and 70. (Probably best W. C. Bray, Journ. Atner. Ghem. Soc., 32. 1204, 1910.) Whatever be the temperature employed, the conditions must be the same as those adopted in standardising the permanganate. 3 W. C. Bray, Journ. Amer. Chem. Soc., 32. 1204, 1910 ; F. P. Treadwell, Kurzes Lehrbuch der analytischen Chemie, Leipzig, 2. 498, 1911. 4 The error is no greater, indeed less, if the whole of the sample be operated upon. 200 A TREATISE ON CHEMICAL ANALYSIS. the V 2 5 will also be reduced by the zinc to V 2 3 ; and to V^ by sulphur dioxide (and possibly by hydrogen sulphide). The reduced oxides will be oxidised to V 2 5 during the permanganate titration. The vanadium is, however, usually present in such small quantities that its influence can be neglected ; see page 467. Correction with very Dilute Solutions. If the amount of iron be small, per- manganate containing half a gram per litre may be used. An error may now creep into the work owing to the fact that a certain quantity of permanganate is required to impart a perceptible colour to pure acidulated water. In such cases, the volume of the sodium oxalate used for the standardisation of the per- manganate must be diluted to approximately the same volume as the solution con- taining the iron to be titrated, or the amount of permanganate required to impart the desired pink coloration to an equivalent volume of acidulated water must be determined, and the result deducted from the permanganate used in the titration. The solution remaining after the permanganate titration can be used for the titanium determination, and afterwards for the determination of phosphorus. 92. The Colorimetric Determination of Iron. In 1853, Herapath 1 proposed to utilise the red colour produced when potassium or ammonium thiocyanate is mixed with a ferric salt for the deter- mination of small quantities of iron. 2 According to Wagner, 3 this reaction enables 1 part of iron to be recognised in 1,600,000 parts of water. The colour is due to the formation of a double salt Fe(CNS) 3 .9KCNS.4H 2 0. 4 A large excess of potassium thiocyanate is needed to transform all the ferric salt into the thiocyanate and so produce the maximum coloration. 5 The intensity of the coloration is proportional to the amount of ferric thiocyanate in solution ; hence, if a given solution containing a known amount of iron has the same colour as a solution containing an unknown amount of iron, it is inferred that both solutions have the same concentration, as indicated in discussing the principles of colorimetry, page 82. Unfortunately, the intensity of the coloration of solutions of ferric thio cyanate is very sensitive to the presence of other salts in the solution. In the 1 T. L. Herapath, Journ. Chem. Soc., 5. 27, 1853 ; A. Thomson, ib., 47. 493, 1885 ; J. Davies, Chem. News, 8. 163, 1863; A. Zega, Chem. Ztg. t 17. 1564, 1862; F. Seller and A. Verda, ib., 26. 803, 1906 ; L. Lapique, Bull. Soc. Chim. (3), 2. 295, 1890 ; J. W. Leather, Journ. Soc. Chem. Ind., 24. 385, 1905 ; A. Jolles, Arch. Hygiene, 13. 402, 1901 ; N. Damaskin, Arbeit Pharm. Inst. Dorpat, 7. 40, 1891 ; J. W. Mellor, Trans. Eng. Cer. Soc., 8. 123, 1908 ; Hadank, Sprech., 42. 445, 1909 ; T. T. Morrell, Amer. Chem., 4. 287, 1874 ; H. P. T. Oerum, Zeit. anal. Chem., 43. 147, 1904 ; A. Jolles, ib. } 43. 537, 1904 ; 44. 6, 1905. 2 H. B. Pulsifer (Journ. Amer. Chem. Soc., 26. 967, 1906) used the red colour of ferric acetylacetonate with excellent results ; and A. Vogel (N. Rep. Pharm., 25. 180, 1876 ; S. Pagliani, Oaz. Chim. Ital., g. 23, 1879 ; E. E. Smith, Journ. Amer. Chem. Soc., I. 335, 1879), the red colour of ferric salts with salicylic acid. For the colours with ammonium sulphide, and with potassium ferrocyanide, see J. W. Mellor, Trans. Eng. Cer. Soc., 8. 123, 1909. 3 A. Wagner, Zeit. anal Chem., 20. 349, 1881. E. E. Smith (ib., 19. 350, 1880) says 1 part in 8,000,000. 4 G. Kriiss and H. Moraht, Liebig's Ann., 209. 98, 1889 ; Zeit. anorg. Chem., i. 399, 1893 ; A. Rosenheim and R. Colin, ib., 2J. 295, 1901. 5 J. H. Gladstone, Phil. Trans., 145. 179, 1885 ; Journ. Chem. Soc., g. 54, 1857. 6 J. Pelouse (Ann. Chim. Phys. (1), 44. 214, 1830) has shown that organic acids reduce the red colour very rapidly ; also oxalates, citrates, tartrates, acetates, iodates, arsenates, phosphates, fluorides, and sulphates. Barium, strontium, and calcium chlorides exert a decolorising action (M. Vernon, Chem. News, 66. 177, 191, 202, 214, 1892 ; J. H. Gladstone, ib., 67. 1, 1893 ; A. Dupre, ib., 32. 15, 1875 ; A. J. Shilton, ib., 50. 234, 1884 ; H. Werner, Zeit. anal. Chem., 22. 44, 1883) ; alumina also retards the development of the colour (R. R. Tatlock, Journ. Soc. Chem. Ind., 6. 276, 352, 1887); and generally, the colour obtained with a given proportion of iron depends upon the nature of the substances present in the same solution. THE DETERMINATION OF IRON. 2OI case of dilute aqueous solutions the intensity of the coloration is not quite proportional to the amount of ferric thiocyanate present in the solution, 1 because some is hydrolysed by the solvent: Fe(CNS) 3 + 3H 2 0:^=Fe(OH) 3 + 3HCNS. Hence, when aqueous solutions are compared, the concentration of the solutions must be approximately the same. With the pyrosulphate fusion, in silicate analyses, the solutions have so nearly the same composition that the error from this cause can be regarded as negligibly small. Consequently, when but small quantities of iron are in question up to 2 per cent. the determination can be made much more conveniently by the colorimetric process than by gravimetric or volumetric methods. Test Solution. Make the pyrosulphate fusion up to 250 c.c. with water. Pipette 25 c.c. into a 250-c.c. 2 flask, and make the solution up to the mark with water. Pipette 5 c.c. of this solution into the test cylinder; add 5 c.c. of potassium thiocyanate solution, 3 and add 20 c.c. water. The cylinder 4 then contains 30 c.c. of solution. Standard Solution. Pipette 5 c.c. of potassium aluminium sulphate solu- tion 5 into the second test glass of the colorimeter; add 5 c.c. of potassium thiocyanate solution ; and pipette 20 c.c. distilled water into the cylinder. The Comparison. Arrange the colorimeter (Weller's), burette, etc., as indicated in fig. 107. Fill a 10-c.c. burette, reading to -?-$ c.c., with the standard ferric oxide solution, and add the solution drop by drop to the standard 1 J. Riban, Bull. Soc. Chim. (3), 6. 916, 1892 ; G. Kriiss and H. Moraht, I.e. ; A. Rosen- lieim and A. Cohn, Zeit. anorg. Chem., 27. 280, 1901 ; H. Schultze, Chem. Ztg., 17. 2, 1893 ; H. Ley, Zeit. phys. Chem., 30. 193, 1899 ; G. Magnanini, ib., 8. 1, 1891 ; L. Andrews, Proc. Iowa Acad. Sciences, I. 4, 1894; C. Fery and E. Tassilly, Bull. Sci. Pharm., 19. 11, 1912; E. Tassilly, Bull. Soc. Chim. (4), 13. 34, 1913 ; 0. Mayer, Monit. Scient. (5), 3. 81, 1913. In order to eliminate the variable effects produced by the presence of other substances in the same solution, Tatlock (I.e. ) proposed to compare the tints of the ethereal extract of the ferric thiocyanate, since J. Natanson (Liebig's Ann. , 130. 246, 1864; C. Glaus, ib , 99. 51, 1856; W. Skey, Chem. News, 16. 201, 324, 1867) has shown that, under these conditions, the reaction is more sensitive and less affected by the composition of the aqueous solution G. Lunge, Zeit. angew. Chem., 10. 3, 1896; H. von Keler and G. Lunge, ib. , 8. 669, 1894; A. Seyda, Chem. Ztg., 22. 1086, 1891 ; H. Lachs and H. Friedenthal, Biochem. Zeit., 32. 130, 1911. For similar reasons, W. M'Kim Marriott and C. G. L. Wolf (Journ. Biol. Chem., i. 451, 1905 ; J. W. Gregory, Proc. Chem. Soc., 23. 306, 1907) used acetone ; and H. N. Stokes and J. R. Cain (Bull. Bur. Standards, 3. 115, 1907), a mixture of ether and isoamyl alcohol 2 : 5. 2 The dilutions must be modified to suit the iron content of the clay, etc. For white burning clays, make 25 c.c. up to 250 c.c., that is, equivalent to making the whole clay up to 2500 c.c. Cream-coloured calcined clays may require 10 c.c., making up to 250 c.c. ; this is equivalent to making the whole clay up to 6250 c.c. Red burning clays may require 5 c.c., making up to 250, 500, or 1000 c.c. 3 POTASSIUM THIOCYANATE SOLUTION. Dissolve 97 grms. of the recrystallised salt, free from iron, in a litre of water. For potassium chloride impurity,' see J. Hendrick, Chem. News, 63. 130, 1891. 4 Smaller cylinders may be used than those indicated for Weller's colorimeter, page 84. Cylinders with a square cross section, 2 '5 cm. side and 8 cm. high, are convenient. 5 POTASSIUM ALUMINIUM SULPHATE SOLUTION. Fuse '05 grm. of iron- free alumina with 5 grms. of potassium pyrosulphate until all is dissolved. Treat the solution as indicated for the pyrosulphate fusion, page 184. When the solution is made up to a litre, the amount used here should be such as to make the concentration of the solution of potassium aluminium sulphate in the two test cylinders of the colorimeter nearly the same. 6 FERRIC OXIDE SOLUTION. Dissolve 0-6303 grm. of ferric potassium alum Fe 2 (S0 4 ) 3 . K 2 S0 4 . 24H 2 (molecular weight 1006 '56) in water, add 5 c.c. concentrated sulphuric acid, and, when cold, make the solution up to a litre. One c.c. of this solution represents O'OOOl grm. of Fe 2 3 . This solution will keep indefinitely, under conditions where more dilute solutions would hydrolyse and deposit a brown oxide on the glass. For use, pipette 5 c.c. into a 100- c.c. flask, and make the solution up to the mark with distilled water; 1 c.c. has O'OOOOOS grm. Fe 2 3 , or 0'005 mgrm. Fe 2 3 W. French, Chem. News, 60. 235, 1889 ; L. L. de Koninck, Bull. Soc. Chim. (3), 23. 261, 1909. 2O2 A TREATISE ON CHEMICAL ANALYSIS. cylinder in the colorimeter until the two tints are the same. The contents of the cylinder are stirred after every addition of O'l c.c. added from the burette, and for every addition from the burette, add an equal amount of water to the test cylinder, so that the conditions for the distribution of the red ferric thio- cyanate may be approximately the same in both cylinders. If more than 3 or 4 c.c. of the standard ferric oxide solution, are required the results will be inexact, because it is difficult to determine the changes of tint with concentrated solu- tions of ferric thiocyanate. In that case, the test solution was not sufficiently diluted, and another start must be made with a more dilute solution. Two cubic centimetres of the standard iron solution may be run into the FIG. 107. Colorimetric determination of iron. test cylinder and the standardisation repeated ; or, another determination may be made de novo. It is well to take the mean of at least three determinations. EXAMPLES. (1) The bisulphate fusion was made up to 250 c.c. ; 25 c.c. of this were made up to 250 c.c., that is, equivalent to making the iron in the whole clay up to 2500 c.c. 5 c.c. of this solution required 2'24 c.c. of standard ferric oxide solution to give uniformity of tint. Hence, since 2 '24 c.c. of the standard iron solution have the equivalent of 0'005 x 2-24 = 0-0112 mgrm. per 5 c.c., the 5 c.c. of the test solution had 0-0112 mgrm. of ferric oxide; hence, 2500 c.c., or 1 grm., of clay had 01)112x500 = 5-6 mgrms. =0-0056 grm. that is, 0*56 per cent. Fe 2 O 3 . (2) The bisulphate fusion was diluted to 250 c.c., and 5 c.c. were made up to 500 c.c. This is equivalent to diluting the "iron" in the gram of clay to 25,000 c.c. 1 '74 c.c. of the ferric oxide were required to match the tint. This corresponds with 0-005 x 1'74 =0-0087 mgrm. of Fe 2 3 per 5 c.c. of test solution. Consequently, 25,000 of test solution would have 43'5 mgrms., or 0'0435 grm., of Fe 2 3 . Hence, the clay has 4'35 per cent. Fe 2 3 . CHAPTER XV. THE DETERMINATION OF TITANIUM. 93. Waller's Colorimetric Process. ScHEERER 1 and Riley have emphasised the almost "ubiquitous occurrence of titanium in silicate rocks. The quantitative separation of titanium from ferric oxide and alumina is not particularly difficult, but the process is somewhat tedious and laborious. The most important sources of error are due to (1) the tendency of the ferric oxide and alumina to separate with the titanium ; (2) the presence of phosphoric acid hindering the complete precipitation of the titanium ; (3) mechanical losses owing to the adhesion of the precipitated titanium oxide to the sides of the beaker ; 2 and (4) the escape of some of the very finely divided precipitate through the pores of the filter paper. The colorimetric process is usually more convenient. 3 A solution of titanium sulphate produces an orange-yellow colour when oxidised with hydrogen peroxide. 4 The intensity of the coloration depends upon the proportion of titanium present. If, therefore, a solution containing a known amount of titanium has the same tint as another solution of equal depth of liquid, it is assumed that both solutions contain the same proportion of titanium. Hydrogen peroxide is therefore added to the solution under investigation, and both the test and the standard solutions containing a known amount of titanium are systematically diluted until both solutions have the same tint. 5 The presence of fluorine leads to low results, because fluorine partially bleaches the yellow colour produced by hydrogen peroxide. For instance, with solutions- containing O'Ol grm. of titanic oxide : Hydrofluoric acid ... '00039 '00194 0'0039 grm. Titanic oxide .... O'Ol 0'0093 O'OOSO 0'0068 grm. Steiger 6 has proposed a colorimetric process for fluorine based on this property. 1 T. Scheerer, Liebig's Ann., 92. 178, 1854 ; Chem. News, I. 143, 1860 ; E. Riley, Journ. Chem. Soc., 12. 13, 1860. a From which it can only be removed by the action of a solvent say, cone, potash lye. 3 Note, if the filtrate from the silica was treated with hydrogen sulphide, some titanium will be precipitated by the hydrogen sulphide. 4 H. Schbnn, Zeit. anal. Chem., 9. 41, 330, 1870 ; 8. 380, 1869 ; V. Lehner and W. JGL Crawford (Jour. Amer. Chem. Soc., 35. 138, 1913) recommend thymol ; H. J. H. Fenton (Journ. Chem. Soc., 93. 1064, 1908) dihydroxymaleic acid page 468. 5 A. Weller, Ber., 15. 2599, 1882 ; C. Baskerville, Journ. Soc. Chem. Ind., 19. 419, 1900 ; J. Brakes, ib., 20. 23, 1901 ; H. M. Ullmann and J. W. Boyer, Chem. Eng., IO. 163, 1909 ; A. Gautier, Chim., I. 177, 1910 ; Rev. gen. Chim., 14. 14, 1910 ; G. P. Pamfil, Monit. Scient. (4), 24. 643, 1911. 6 T. B. Osborne, Amer. J. Science (3), 30. 329, 1885 ; Chem. News, 53. 43, 1886 ; W. F. Hillebrand, ib., 72. 158, 1895; Journ. Amer. Chem. Soc., 17. 718, 189f> ; J. H. Walton, ib., 29. 481, 1907 ; G. Steiger, ib., 30. 219, 1908 ; H. E. Menvin, Amer. J. Science (4), 28. 119, 1909; E. Jackson, Chem. News, 47. 157, 1883; L. Levy, Compt. Rend., 105. 754, 1888; C. Reichard, Chem. Ztg., 28. 16, 1904. 203 204 A TREATISE ON CHEMICAL ANALYSIS. The presence of phosphoric acid bleaches the colour and leads to low results. For instance, with solutions containing the same amount of titanium oxide, O'Ol grm., and Phosphoric acid O'OO 0*13 0'26 0*52 078 1*04 1-30 grm. Titanic oxide O'OIOO 0'0090 0'0083 0*0074 0*0069 0*0066 '0064 grm. Potassium sulphate also weakens the tint produced by the titanium oxide, unless an excess of sulphuric acid be present. 1 For instance, with 6 grms. of potassium sulphate in each of three solutions containing the same amounts of titanium and hydrogen peroxide, Merwin found that with H 2 S0 4 .0-4 2-9 8-0 c.c. Bleaching . . . . . . .21 14 5 per cent. From which it follows that, the greater the excess of sulphuric acid, the less the bleaching action. Owing to the fact that the test solution obtained in the above analytical scheme contains the potassium sulphate used in the pyrosulphate fusion, an equivalent amount of potassium sulphate can be added to the standard solution in order that the comparison of tints can be made under similar conditions. In the same way, if the test solution contains appreciable quantities of phosphoric acid, an equivalent amount of phosphoric acid can be added to the standard solution. A solution of ammonium molybdate in nitric acid, 2 as well as uranium and vanadium salts, 3 gives somewhat similar tints, and hence these salts, with the chromates, should be absent. The influence of iron salts will be discussed later. The colorimetric process works well with quantities of titanium up to about 4 per cent. There is a wide range over which the colorimetric process is accurate. For titanium this ranges between concentrations represented by 0*0015 and - 0200 grm. per 100 c.c. According to Wells, 4 the change in concentration required to produce a perceptible difference in the intensity of the colour of two solutions is about 6*5 per cent. The error is greater with weaker solutions, although by increasing the thickness of the layer of liquid 5 satisfactory com- parisons can be made with solutions containing less than 0*0015 grm. of titanium per 100 c.c. Solutions more concentrated than 0*0200 grm. per 100 c.c. are not suited for colorimetric work. With the exercise of the greatest care, the accuracy of the colorimetric process for titanium is about 2 per cent. If more than 4 per cent, of titanic oxide be present, which is rarely the case with clays, the gravimetric process is best employed. More accurate results are obtained with colours not too intense. Colours approximating a deep straw yellow give the best results. The eye is not so sensitive to the small differences of tint in concentrated solutions. The left eye is usually rather more sensitive than the right eye. Better results are obtained after the eye has had a little practice with test solutions of known strength. Cf. p. 85. Preparation of the Test Solution. If necessary, evaporate the aqueous 1 According to G. Steiger (I.e.}, sodium salts give too high results. 2 H. Schonn, Zeit. anal. Chem.,g. 41, 330, 1870 ; H. Berwald, Ber., 18. 1206, 1885 ; M. Fairley, Journ. Chem. Soc., 31. 127, 1877 ; J. Aloy, Bull. Soc. Chim. (3), 27. 734, 1903 ; Chem. News, 87. 102, 1903. 3 G. Werthier, Jonrn. prakt. Chem. (1), 83. 195, 1861 ; E. Jackson, Chem. News, 47. 157, 1883. Uranium salts (M. Fairley, Journ. Chem. Soc., .31. 127, 1877) give a yellow colour with hydrogen peroxide. 4 R. C. Wells, Journ. Amer. Chem. Soc., 33. 504, 1911. 5 In colorimeters of the type of Duboscq's (page 82). THE DETERMINATION OF TITANIUM. 205 solution of the pyrosulphate fusion, acidified with sufficient sulphuric acid l to make the solution between 15 and 17 per cent, of H 2 S0 4 , that is, 9 c.c. of sulphuric acid per 100 c.c. If insufficient sulphuric acid be present the results are generally low. When the solution occupies about 150 or 200 c.c., cool. Transfer to, say, a 250-c.c. flask. Add enough hydrogen peroxide 2 to oxidise all the titanium in the solution. In general, 5 to 10 c.c. of hydrogen peroxide (20 vols.) suffice for 0*5 to 2 per cent, of titanium oxide. Make the solution up to 250 c.c. with distilled water. 3 Pour a part of this solution into the left cylinder of the colorimeter. Preparation of the Standard Solution. Pipette, say, 5 c.c. of the standard solution 4 of titanium sulphate, containing, say, 1 grm. per litre, into a 100- c.c. flask ; add 5 c.c. of hydrogen peroxide, 5 and make the solution up to 100 c.c. 6 Each cubic centimetre is then equivalent to O'OOOl grm. of Ti0 2 . Put, say, 10 c.c. of this standard solution into the right glass cylinder of the colori- meter, and 50 c.c. of water into a burette. 7 Add water to the solution until the colour of the standard is the same as the colour of the test solution in the left cylinder of the colorimeter. Note the amount of water added. The disposition of the colorimeter will be apparent from fig. 103. The sliding door of the colorimeter, glass stirring rod, burette with water, etc., are shown in the diagram. Readings and Calculations. The test solution was made up to 250 c.c. This was derived from 1 grm. of clay. 10 c.c. of the standard titanium solution containing 1 grm. of Ti0 2 per litre were made up to 100 c.c. The 10 c.c. required 11*2, 11 -3, 11 '5, mean 11 '33 c.c. of water to match the test solution. The 10 c.c. of the standard solution contained 0*01 grm. of Ti0 2 , and when 1 To transform any basic titanium sulphates formed during the fusion into normal sulphate, and to counteract bleaching by potassium sulphate. P. Faber, Ghem. Zty., 31. 263, 1906 ; Zeit. anal. Chem., 46. 277, 1907 ; F. P. Dunnington, Journ. Amer. Chem. Soc., 13. 210, 1801. 2 HYDROGEN PEROXIDE. This must be free from all traces of fluorides and phosphoric acid. Merck's perhydrol, containing 30 per cent. H 2 2 by weight, is excellent when diluted. Make, say, 20 c.c. of perhydrol up to 100 c.c. with distilled water for a 6 per cent, (or 20 vol.) solution. The solution keeps better if acidified with, say, sulphuric acid. The general rule for dilution is : n volumes of perhydrol made up with water to 10 volumes furnishes, approximately, a 3?i per cent. H 2 O 2 , or a lOn volume, solution. See pages 177 and 322. 3 If the colorimetric process for iron has been used, this simply means adding the hydrogen peroxide, etc., to an aliquot portion of the pyrosulphate fusion. 4 STANDARD SOLUTION OF TITANIUM OXIDE. Digest between 0'6 and 0'7 grm. of feebly calcined potassium titanofluoride K 2 TiF 6 in a platinum dish, with concentrated sulphuric acid, and evaporate until white fumes are evolved. Repeat the treatment three times, so as to drive off all the fluorine. Take up the residue with a little concentrated sulphuric acid and dilute until the solution has between 5 and 10 per cent, of sulphuric acid. Make up the solution to about a litre. Take two aliquot portions, between 50 and 100 c.c., dilute with water, heat to boiling, and precipitate the titanium hydroxide by the addition of ammonia. Wash the precipitate with hot water until free from alkalies, ignite for Ti0 2 , blast, and weigh. Duplicate determina- tions should be concordant. The solution may be diluted until it contains exactly 1 grm. of Ti0 2 per litre. The solution should be kept in a bottle with a glass stopper coated with vaseline. Withdraw the amount needed for a determination by means of a pipette. Do not pour the solution from the bottle. See page 644 for an alternative process. 5 The solutions should always be freshly prepared for an observation in both the test and standard cylinders. 6 As much potassium sulphate as is present in the test solution may be added to the flask before making up with water to 100 c.c. Suppose that the 250 c.c. of the test solution was derived from the fusion of the ammonia precipitate with 6 grms. of potassium pyrosulphate K 2 S 2 7 , the 250 c.c. will contain the equivalent of 07x6 = 4 grms. of K 2 S0 4 per 250 c.c. Hence, approximately 20 c.c. of a solution of potassium sulphate containing 80 grms. K 2 S0 4 per litre will be needed. The same quantity of potassium sulphate is added per 100 c.c. ot: water from the burette. 7 The test solution should have a paler tint than the standard before the latter is diluted from the burette. If the clay contained excessive amounts of Ti0 2 , it may be necessary to further dilute the test solution say, 25 c.c. to 100 c.c., or the standard solution may be made more concentrated. 2O6 A TREATISE ON CHEMICAL ANALYSIS. diluted to 100 c.c., 10 c.c. of the diluted solution contained O'OOl grm. TiO l? . Hence, 10 + 11-33 = 21-33 c.c. contained 0-001 grm. Ti0 2 ; hence, 250 c.c. would have Consequently, 1 grm. of the clay has 0*0117 grm. Ti0 2 , or the clay contains 1*2 per cent. Correction for Iron. According to Faber, if much iron be present, the results may be too high, because the iron intensifies the colour of the test solution. Hillebrand deducts Q-Q2 per cent, from the final result for every JO pp/Tfipmt. ot' iron oxine present..! Faber 2 recommends the addition ot phosphoric acid to both the standard and the test solutions in order to neutralise the effect of iron. It is necessary to add the phosphoric acid to both solutions, because phosphoric weakens the yellow tint, and both solutions are then affected in the same way. 50 c.c. of phosphoric acid ^sp. gr. 1*3) per 250 c.c. will usually suffice. W. A. Noyes A adds approximately an amount 6t ternc" ammonium alum 4 ' to the standard, equivalent to the amount of ferric oxide in the test solution. To allow for the presence of iron, therefore, first make an approximate determination of the titanium in the clay, and let a denote the final number of c.c. in the standard after dilution ; b the number of c.c. of water used in diluting the standard; and p the amount of iron in the test solution in terms of Fe 2 3 per c.c. ; and x the amount of Fe 2 3 to be added per c.c. of water used in diluting the standard in the final test. Hence, bx denotes the amount of ferric iron in the water used for diluting the standard solution ; and ap + 250 denotes the amount of ferric oxide in the a c.c. of test solution. Consequently, ap ap = or, * = ofFe 2 3 perc.c. The Fe 2 3 is added as ferric ammonium sulphate. Since 160 of Fe 2 3 is equivalent to 964 of ferric ammonium alum, every gram of ferric oxide corresponds with 6 grms. of ferric ammonium alum. Gautier 5 recommends making up a permanent set of standard comparison tints from methyl orange dissolved in water ; as a matter of fact this method is somewhat risky. Lovibond's tintometer offers some advantages in making up a standard comparison scale ; see page 486. . JResults. To show the deviations which might be expected in titanium determinations, the following results were obtained in eight independent determinations of titanium in one sample of clay : 0-0120; 0-0125; 0-0125; 0*0113; 0-0114; 0'0126; 0'0125 ; 0-0113, with a mean value 0*0120, or 1'20 per cent. The deviations from the mean are approximately 0*07. 1 W. F. Hillebrand, Bull. U.S. Geol. Sur., 422. 133, 1910. 2 P. Faber, Chem. Ztg., 31. 263, 1906 ; J. H. Walton, Journ. Amer. Chem. Soc., 29. 481, 1907. 3 W. A. Noyes, Journ. Anal. App. Chem., 5. 39, 1891. 4 STANDARD IRON SOLUTION. Dissolve 27 '6 grms. of ferric potassium sulphate iron alum in 500 c.c. of water. Add 100 c.c. of concentrated sulphuric acid, and make the solution up to a litre. The solution may be standardised, if necessary, by titration (page 198). Ferrous salts may exert a bleaching action on the titanium colour. 5 A. Gautier, Chim., 2. 2, 1911 ; Rev. ge"n. Chim., 14. 16, 1911. 1 grm. of methyl orange is dissolved in 500 c.c. of water ; 10 c.c. of this solution are diluted to 200 c.c. ; and the solution is matched with solutions containing known amounts of titanium. The standards can afterwards be preserved, for, according to Gautier, they do not fade. THE DETERMINATION OF TITANIUM. 2O7 94. The Gravimetric Determination of Titanium Gooch's Process. Titanium can be separated from aluminium by the prolonged boiling of a slightly acid solution of the sulphate. Thus, Levy found the amount of hydrolysis (cf. page 181) in solutions containing pure titanium sulphate partly " neutralised " with potassium hydroxide, and boiled for six hours : Table XXXVIII. Effect of Sulphuric Acid on the Precipitation of Titanic Oxide. Titanic oxide. Free H S0 4 per lOOo.c. Used. Precipitated. o-ooo* 0-086 0-108 o-ooo 0-086 o-ioo o-ooo 0-036 0-047 0-083 0-036 0-036 0-500 0-086 0-085 1-000 0-082 0-086 5-766 0-030 o-ooo * Slight excess of potassium hydroxide. These results clearly show that if too much sulphuric acid be present, the titanium will be but imperfectly .precipitated. If too little acid be present, some aluminium will be precipitated. Hence, the adjustment of the acidity of the solution is so difficult that the process is not at all satisfactory for general work. 1 Gooch's 2 method is based upon the solubility of alumina, and the insolubility of titanium hydroxide in solutions containing more than 5 per cent, of acetic acid by volume. The reaction is so delicate that Gooch obtained a distinct opalescence when 0*0005 grm. of titanic oxide was present in 500 c.c. of liquid containing in solution 10 grins, of alum and 15 grins, of sodium acetate, and 7 per cent, of acetic acid by volume. The process gives excellent separations of aluminium and titanium, but it does not work well in the presence of iron. The precipitation of basic ferric acetate will be prevented by the presence of 11 per cent, of acetic acid, yet "in the presence of a solution of ferric acetate, titanium shows a very marked tendency to remain dissolved." Thus, 400 c.c. of a solution of 10 grms. of sodium acetate, 17 per cent, of acetic acid, and the equivalent of 0'2 grm. of ferric oxide retained 0'06 grm. of 1 L. Levy, Journ. Pharm. Chem. (5), 16. 56, 1887 ; Ann. Chim. Phys. (6), 25. 433, 1892 ; H. Pellet and C. Fribourg, Ann. Agron. (2), 2. 20, 1905; P. Holland, Chem. News, 59. 27, 1889 ; D. Forbes, ib., 19. 3, 1869 ; J. Brakes, Journ. Soc. Chem. Ind., 18. 1097, 1899 ; C. Baskerville, ib., 19. 419, 1900. A. Leclerc (Compt. Rend., 137. 50, 1904) adds to an aqueous solution of the potassium bisulphate fusion enough formic acid to make the solution 5 per cent, acid. On standing two days at 100 all the titanic acid and silica are said to be precipitated. 2 F. A. Gooch, Proc. Amer. Acad. Science (2), 12. 435, 1885; Amer. Chem. Journ., 7. 283, 1885 ; Chem. Neivs, 52. 55, 68, 1885 ; T. M. Chatard, ib., 63. 269, 1891 ; Amer. Chem. Journ., 13. 106, 1891; B. Neumann, Stahl Eisen, 30. 457, 1910; H. L. Vogt, Zeit. prakt. Geol., 8. 379, 1900 ; W. F. Hillebrand, Bull. U.S. Geol. Sur., 422. 135, 1910 ; P. T. Austen and F. A. Wilber, Amer. Chem. Journ., 5. 389, 1883 ; Chem. News, 48. 113, 1883 ; G. Werther, Journ. prakt. Chem. (1), 91. 321, 18fi4; C. Baskerville, Journ. Amer. Chem. Soc., 16. 427, 1894; J. W. Bain, ib., 25. 1073, 1903 ; G. W. Wdowiszewski, Eng. Min. Journ., 185.. 1200, 1908 ; W. M. Thornton, Eng. Min. Journ., 94. 353, 1912. 208 A TREATISE ON CHEMICAL ANALYSIS. titanium oxide in solution. 1 It is advisable to separate the iron as sulphide 2 from the mixture containing aluminium, iron, and titanium sulphates. The gravimetric determination involves four operations : 1. Separation of Iron from Titanium and Aluminium? The iron is best precipitated as ferrous sulphide from the feebly ammoniacal solution containing sufficient tartaric acid to keep the aluminium and titanium in solution. An amount of tartaric acid equal to three times the weight of the oxides to be held in solution is sufficient unless the ammonium salts be present in great excess. Hence, add, say, 0*7 grm. of tartaric acid to the solution; 4 reduce the iron to ferrous sulphide by means of hydrogen sulphide. 5 Add ammonia until the solution has cleared and the ammonia is in slight excess. Again pass hydrogen sulphide through the solution. 6 The clear supernatant liquid should not be tinted green, 7 though it might be slightly yellow. Filter the solution and wash the precipitate as quickly as possible in water containing ammonium sulphide in solution. 8 The precipitated iron may be dissolved in hydrochloric acid, boiled to expel hydrogen sulphide, oxidised with hydrogen peroxide, precipitated with ammonia, ignited, and weighed as Fe 2 3 . 2. Decomposition of the Tartaric Acid. 9 The nitrate and washings may 10 be concentrated by evaporation, the solution acidified with sulphuric acid, and sufficient potassium permanganate added to leave the solution distinctly coloured after all the tartaric acid has been oxidised. 11 Generally, about 2'5 times the weight of the tartaric acid is needed. 12 If there be a deposit of manganese hydroxide formed, add sulphurous acid until it is dissolved. 3. Separation of Titanium from Aluminium. 3 The solution may now contain manganese, aluminium, and titanium, together with some potassium sulphate. Add ammonia until the precipitate first formed dissolves with difficulty on stirring. If the precipitate does not dissolve, a drop or two of hydrochloric acid may be added. Add 20 grms. of sodium acetate and 7 to 10 c.c. of glacial acetic acid for every 100 c.c. of the solution under treatment (that is, about one-tenth the volume of the solution). Heat the solution to boiling and, after boiling one minute, let it stand a few more minutes to allow the flocculent precipitate of titanium hydroxide to subside. Decant the solution through a porous filter paper (e.g., No. 589, Schlecher and Schull), and wash the precipitate first with 7 per cent, acetic acid, and finally with hot water. . Dry the precipitate. Ignite 15 to 20 minutes over a Meker's burner. Cool, and weigh as Ti0 2 . 4. Purification of the Titanium Oxide. The titanium oxide carries down J G. Streitand B. Franz, Journ. prakt. Chem. (1), 108. 65, 1869. 2 F. Reich, Journ. prakt. Chem. (1), 83. 266, 1861; R.Fresenius, Zeit. anal. Chem., i. 69, 1862. 3 Also chromium. J. J. Berzelius, Pogg. Ann., 4. 3, 1825. 4 Be careful to see that the tartaric acid is free from alumina. 5 A. Cathrein (Zeit. Kryst., 6. 244, 1882 ; 7. 250, 1883) recommends a repetition of process to recover traces of titanium precipitated with iron. 6 Some platinum sulphide may separate. If so, filter and wash. The platinum crucible is slightly attacked during the bisulphate fusion. 7 Showing that " ferrous " iron is still in solution. 8 Keep the funnel covered with a clock-glass to prevent oxidation as much as possible, otherwise soluble ferrous sulphate may be formed. 9 Enough sulphuric acid should be present to leave an excess after all the permanganate subsequently added has formed manganese sulphate. M. Dittrich and R. Pohl (Zeit. anorg. Chem., 43. 236, 1905) prefer oxidising the tartaric acid by evaporating to dryness, and digesting the residue with dilute sulphuric acid and potassium persulphate ; W. M. Thornton (Amer. J. Science (4), 34. 214, 1912) uses a mixture of sulphuric and nitric acids. See page 309. 10 A. Fleischer, Ber., 5. 350, 1872 ; J. Hetper, Bull. Acad. Science Cracow, 601, 1910 ; L. Lindet, Chem. News, 76. 212, 1897. See page 309. 11 To carbonic and formic acids. 12 Hence, in the example under consideration, about '7x2 '5 = 1'8 grms. of potassium permanganate will be enough. THE DETERMINATION OF TITANIUM. 2 09 some manganese, alumina, alkaline sulphates, vanadic acid, etc. For exact work, therefore, the precipitate must be fused with about 10 times its weight of sodium carbonate 1 for about an hour over a blast. A residue of sodium titanate, 2 insoluble in water, remains; sodium phosphate, vanadate, and aluminate pass into solution. Filter and wash with water, containing a little sodium carbonate. Dry. Transfer the mass to a watch-glass. Ignite the filter paper in the crucible, and add the powder in the watch-glass to the ash. Fuse the contents with a little sodium carbonate. Cool. Dissolve the mass in 100 to 150 c.c. of water and add 20 c.c. of sulphuric acid. Nearly neutralise with ammonia as before and treat the solution with 5 grms. of sodium acetate and one-tenth its volume of glacial acetic acid. Boil one minute, filter, and wash as before. The fusions, precipitations, and ignitions should be repeated until the " titanium oxide " obtained is white in colour and constant in weight. Usually the two precipitations indicated above suffice. The purification of the precipi- tate does not take so long, because the precipitated titanium hydroxide is flocculent and filters easily. In illustration, the following results represent the weights obtained with a titaniferous bauxite : 1st 2nd 3rd 4th precipitation. 0-0752 0-0699 0'0694 0'0696 grm. Ti0 2 obtained. If zirconium be present, the titanium does not precipitate satisfactorily by this method, since, as Gooch has pointed out, the zirconium oxide acts in a similar manner to the ferric oxide mentioned above. Hillebrand 3 showed that 0'2 per cent of zirconia in a solution prevented the precipitation of 0-3 per cent. of Ti0 2 . Hence, Hillebrand recommends the removal of the zirconium as phos- phate by the method indicated on page 498, before the titanium is determined. 95. Blair's Modification of Gooch's Gravimetric Process. If the amount of titanium is large, time can sometimes be saved by de- termining the amount of titanium in the original sample by Blair's process. 4 Separation of Iron, etc. Fuse, say, 1 grm. of the sample with 6-8 times its weight of sodium carbonate and a gram of sodium nitrite. Digest the cold mass with water. Filter off the insoluble residue which contains ferric oxide, sodium titanate, etc. 5 Dry and ignite the residue to burn off the filter paper. Solution of Titanium Oxide. Fuse the residue with 1 5-20 times its weight of potassium pyrosulphate. When cold, add 2-3 c.c. of concentrated sulphuric acid, and heat again until all is melted. Leave a piece of thick platinum wire in the fused mass. When cold, heat the crucible, to just soften the cake in contact with the crucible, and transfer the cake, by means of the piece of platinum wire, to a 600-c.c. beaker. Wash the crucible and lid with 5 per cent, sulphuric acid, and make the washings, etc., in the beaker to 150-200 c.c. with the 5 per cent, sulphuric acid. Precipitation of Titanic Oxide. Add 50 c.c. sulphurous acid. 6 Warm the solution, but not hotter than can be held comfortably by the hand. This 1 A pink or green coloration is due to manganese. 2 Iron oxide, if present, will also remain insoluble. 3 W. F. Hillebrand, Bull. U.S. Geol. Sur., 422. 136, 1910. 4 A. A. Blair, The Chemical Analysis of Iron, Philadelphia, 184, 1908 ; F. J. Pope, Trans. Amer. Inst. Min. Eng., 29. 372, 1899 ; J. W. Bain, Journ. Amer. Chem. Soc., 2$. 1073, 1903 ; J. J. Morgan, Chem. News, 75. 134, 1897; G. B. Waterhouse, ib., 85. 198, 1902; E. Roer, Chem. Ztg., 33. 1225, 1909 (ilmenite) ; K. Borneman and H. Schirmeister, Met., 7. 71, 1911. 5 Nearly all the phosphorus and aluminium pass into solution as sodium phosphate and aluminate. The nitrate may be used for the determination of other constituents chlorine, fluorine, sulphur, etc. See pages 652, 637, 618, etc. 6 Or 5 e.c. of a saturated solution of ammonium bisulphite. 2 1C A TREATISE ON CHEMICAL ANALYSIS. accelerates the solution of the pyrosulphate cake, and the acid prevents the pre- cipitation of the titanium at this temperature. If necessary, filter the solution from any silica which might separate. Add ammonia 1 until the precipitate which forms redissolves with difficulty. 2 Treat the solution with 10 c.c. of sulphurous acid, 20 grms. of sodium acetate dissolved in a little water, and one- sixth the total volume of the solution of acetic acid (sp. gr. 1 -04, that is, about 49 per cent.). Heat the solution to boiling, and boil two or three minutes. Flocculent titanic oxide is precipitated. Digest the solution on a steam bath for half an hour. 3 Let the precipitate settle. Filter and wash first with hot water containing 5 per cent, of acetic acid, and finally with hot water. Dry the precipitate. If the precipitate is dark-coloured, instead of white, the titanic oxide may be contaminated with vanadic acid, etc., 4 in addition to phosphoric acid, alumina, sulphates, etc. The precipitate is purified as indicated for Gooch's process, page 208. 96. The Computation of the Results for "Alumina." The analytical results may now be treated as follows : Crucible and ammonia precipitate ...... 20 7003 grms. Crucible alone 20 "4520 . Ammonia precipitate and silica residue ..... 0*2483 grm. Ash 0-0003 Alumina ferric oxide, titanic oxide, etc '2480 grm. Ferric oxide (p. 198) . . . . . . . . . 0'0156 ,, Titanic oxide (p. 204) 0'0120 ,, Phosphoric oxide (p. 595) O'OOOO ,, Manganese oxide (p. 372) O'OOOO ,, Extra silica (p. 185) 0'0009 ,, Sum 0'0285 grm. Alumina, ferric oxide, etc. . . . . . . . . 0-2480 ,, Sum of ferric oxide, etc . . 0*0285 ,, Alumina . 0*2195 grm. The eight alumina determinations in the clay mentioned above thus furnished : 0-2184; 0-2188; 0-2191; 0-2193; 0*2187; 0-2199; 0-2185; 0-2195 grm. A1 2 3 . The arithmetical mean is 21*90 per cent., and the deviations range between the limits 0'10. If the complete analysis proves faulty when submitted to the test mentioned on page 246, it will be generally found that the fault lies with the alumina determination. In addition to the disturbing effects of phosphates and fluorides (discussed elsewhere) on the ammonia precipitate, the presence of borates and of oxalates may lead to the precipitation of the barium, strontium, calcium, and magnesium salts by ammonia; while the presence of citric and tartaric acids and sugars hinders or prevents the precipitation of iron, aluminium, and chromium. 5 1 Or add a slight excess of ammonia, and then a few drops of sulphuric acid until the precipitate redissolves. ' 2 If platinum is to be removed, the solution should be now treated with hydrogen sulphide. 3 If filtered immediately, some titanium may be found in the filtrate. 4 Iron, if present in the solution, would be carried down in the ferrous condition or as basic acetate. L. J. Cartman and H. Dabin, Journ. Amer. Chem. Soc., 34. 1493, 1912. CHAPTER XVI. THE DETERMINATION OF CALCIUM AND MAGNESIUM. 97. The Properties of Calcium Oxalate. CALCIUM is precipitated from alkaline solutions as hydrated calcium oxalate CaC. 2 4 .H 2 which is converted by calcination into calcium oxide CaO. The calcium oxalate obtained by the first precipitation is so contaminated with sodium and magnesium salts that a second precipitation is advisable. If more magnesium than lime be present, the second precipitation is imperative ; if not, one precipitation may be sufficiently exact for the purpose. In illustration, Fresenius 1 gives the following results : One precipitation Two precipitations Calcium oxide. Magnesium oxide. Exp. 1. Exp. 2. Exp. 1. Exp. 2. 0-2059 0-2051 0-2063 0-2049 0-4912 0-4927 0-4904 0-4928 Hence, with one precipitation the lime was O'OOll grm. (mean) too high, and the magnesia Q'0020 grm. (mean) too low. Calcium oxalate is precipitated from feebly ammoniacal solutions, and also from solutions acidified with acetic acid, 2 oxalic acid, salicylic acid, or citric acid, by means of ammonium oxalate. 3 Ammonium chloride or sulphate, 4 or an excess of ammonium oxalate, 5 do not interfere very much, but in presence of ammonium salts the precipitate is said to be more difficult to filter. If, however, the pre- cipitation be made in hot solutions, or in a solution acidified with acetic acid, the precipitate will be crystalline or granular, and filter easily. Solubility. The solubility of calcium oxalate in cold water is, for analytical purposes, negligibly small, but in hot water the solubility becomes appreci- 1 R. Fresenius, Zeit. anal. Ohem., 7. 310, 1868 ; W. Gibbs, Amer. J. Science (2), 44. 213. - Chem. News, go. 248, 1904 ; W. Herz and G. Muhs, Ber., 36. 3715, 1903. 3 Chem. News, go. 248. 1904 ; F. Utz, Oesler. Chem. Ztg., 7. 510, 1904. 4 For the solubility of calcium oxalate in ammonium sulphate solutions, see R. Fresenius. Zeit. anal. Chem., 30. 594, 1891. 5 Zinc, barium, and lead, and also cobalt and nickel, if appreciable amounts be present, should be removed before adding the ammonium oxalate. Small quantities of nickel do not interfere when two precipitations are made. Uranium and arsenic do not interfere with the result. Manganese, copper, aluminium, molybdenum, and phosphoric acid, if present, will contaminate the precipitate more or less. Chromic acid does not interfere unless it has been reduced to chromic oxide by standing some time Chem. News, 90. 248, 1904. 212 A TREATISE ON CHEMICAL ANALYSIS. able, 1 particularly if the precipitate be subjected to a prolonged washing. The graph, fig. 108, shows the effect of temperature on the solubility of this salt in water. 100 c.c. of water at 95 will dissolve 0*0015 grm. of calcium oxalate, and at 18* 0*0007 grin. When first precipitated, calcium oxalate is more soluble than after standing some time, presumably owing to the fact that the fine particles grow into larger grains. 2 Influence of Magnesium. A small propor- tion of magnesium salt accompanies the calcium oxalate precipitate even when a very large amount of ammonium chloride is present, and this particularly when the amount of mag- nesium is relatively large, or the amount of calcium oxalate relatively small. On the other hand, a certain amount of calcium escapes precipitation. Thus, if care be not taken in ' separating calcium and magnesium, the result fUQ / / - / i i ft. j 50 c i E , .. f f V > > X lit r - f r \ - 0-L FIG. 108. Solubility of calcium oxalate in water. can only represent the truth by a fortuitous balancing of errors, the magnesium precipitated with the calcium compensating the calcium retained in the solution. When the calcium is reprecipitated two or three times, the result must be sensibly low." 3 This difficulty arises from the fact that calcium oxalate is soluble in magnesium chloride solutions, while magnesium oxalate is but slightly soluble in water. Hence, sufficient ammonium oxalate must be added to transform all the magnesium chloride into oxalate, and the solution must be so dilute that the magnesium oxalate remains in solution. 4 Free ammonia and warm solutions favour, and the presence of ammonium chloride retards, the precipitation of magnesium oxalate along with the calcium oxalate. 5 Hence, as recommended by Scheerer in 1859, two or three precipitations in the presence of a large excess of ammonium chloride are needed to get rid of the magnesium. Accordingly, many prefer to precipitate the " lime " as calcium sulphate by adding sodium sulphate and an excess of 90 per cent, alcohol, in which calcium sulphate is almost insoluble, while magnesium sulphate is fairly soluble in the same menstruum see page 524. Action of Heat. Calcium oxalate, dried at 100, has the composition CaC 2 4 .H 2 0. 6 At 200 it loses water, forming CaC 2 4 ; and at 500 it begins to decompose into calcium carbonate, with the separation of carbon. At the same time the carbon imparts a greyish colour to the mass. At a still higher temperature the carbon burns off, and the carbonate decomposes completely into calcium oxide CaO which remains as a white hygroscopic powder. 1 A. F. Holleman, Zeit. phys. Chem., 12. 125, 1893 ; F. Kohlrausch and F. Rose, ib., 12. 234, 1893 ; T. W. Richards, C. T. M'Caffrey, and H. Bislee, Proc. Amer. Acad., 36. 375, 1901 ; Zeit. anorg. Chem., 28. 71, 1901. 2 W. H. Wollaston, Phil. Trans., 103. 51, 1813 ; W. Ostwald, Zeit. phys. Chem., 34. 495, 1900 ; G. Hulett, ib., 37. 385, 1901 ; 47. 357, 1904. 3 T. Scheerer, Journ. prakt. Chem. (1), 76. 424, 1859 ; G. C. Wittstein, Zeit. anal. Chem., 2. 318, 1863 ; E. Sonstadt, Chem. News, II. 97, 1865 ; 29. 209, 1874 ; E. Divers, ib., n. 144, 1865 ; M. Longchamp, Ann. Chim. Phys. (1), 12. 255, 1819 ; E. Lenssen and A. Souchay, Liebig's Ann., 99. 31, 1856 ; 100. 308, 1856 ; T. Scheerer, ib., no. 236, 1859 ; H. Oeffinger, Schweiz. Woch. Pharm., 6. 265, 1868 ; F. Hundeshagen, Zeit. bffent. Chem., 15. 85, 1909. 4 H. Hager, Pharm. Centr. (2), 6. 226, 1865 ; 10. 241, 1869. 5 M. von Paguireff, Journ. Russ. phys. Chem. Soc., 34. 195, 1906. E. Murmann (Zeit. anal. Chem., 49. 688, 1910) says that "a good separation from magnesium is only possible by adding to the solution of the salts in 90 per cent, alcohol just sufficient sulphuric acid to form calcium sulphate, and washing with 90 per cent, alcohol." "The error is then within 0*15 per cent, of the true value; while the error in the oxalate method is from 0'5 to 2*0 per cent." A. Chizynski, ib., 4. 348, 1865. 6 A. Souchay and E. Lenssen, Liebiq's Ann., 100. 322, 1856. THE DETERMINATION OF CALCIUM AND MAGNESIUM. 213 98. The Gravimetric Determination of Calcium. If manganese has been determined, the nitrate from the manganese sulphide is boiled for an hour, and filtered from the precipitated sulphur; 1 and if manganese is not to be separated, the filtrate from the aluminium and ferric hydroxides is evaporated to about 300 c.c. The solution already contains con- siderable amounts of ammonium chloride ; it should contain about 10 grms. per 100 c.c. of the solution per 0-0015 grm. of magnesia. First Precipitation? Heat the solution to boiling, and add 2 c.c. of acetic acid per 100 c.c of the neutral solution. Meanwhile, say, 1 grm. of oxalic acid is dissolved in a little hot water, and added to the solution. 3 In about 5 minutes add a slight excess of ammonia. Let the solution cool for 2 or 3 hours. Pour the cold, clear solution through a 7-cm. filter paper. Wash three times by decantation with dilute ammonia (1 : 10), or a 1 per cent, solution of ammonium oxalate. Second Precipitation. The precipitate is contaminated with sodium and magnesium salts. 4 Dissolve the precipitate in dilute nitric acid (1 : 5), and collect the runnings in the same beaker in which the lime was first precipitated. About 50 c.c. of acid are needed. Tilt the beaker so that the acid wets the sides all round \ 1 the beaker, in order to dissolve any adhering precipitate. Add a slight excess of ammonia and a few drops of oxalic acid solution. Boil. Allow to stand a couple of hours. When cold, filter 5 and wash as before. Transfer all the precipitate to the filter paper. The sides of the beaker will want carefully rubbing with the " policeman," since the precipitate adheres very tenaciously, and is sometimes difficult to see when wet. The Ignition of the Precipitate. Place the moist filter paper in an ignited and weighed crucible. Heat gently so as to char the paper fig. 96 or fig. 112. Incinerate the precipi- tate with the crucible inclined at an angle of, say, 45, so as to burn the carbon mixed with the oxalate. Blast about 10 minutes, or heat over a Meker's or Teclu's burner with a Winkler's chimney (fig. 109), 6 so as to raise the temperature above that possible with an unprotected flame. Cool the crucible in a desiccator, 7 and weigh as quickly as possible to avoid the FIG. 109. Winkler's chimney. 1 Note the possible formation of sulphuric acid, and precipitation of, say, barium as sulphate C. de la Harpe, Bull. Soc. Ind. Muhlhouse, 245, 1885. 2 E. Murmann, Monats. Chem., 32. 105, 1911 ; Oester. Chem. Ztg., 12. 305, 1909 ; Zeit. anal. Chem., 49. 688, 1910. 3 At least four times as much ammonium oxalate is required beyond that needed to form calcium oxalate with the lime, and magnesium oxalate with the magnesia. Note that com- mercial ammonium oxalate sometimes contains calcium salts. 4 T. Scheerer, Journ. prakt. Chem. (1), 79. 424, 1859 ; A. Cossn, Zeit. anal. Chem., 8. 141, 1869; T. W. Richards, Zeit. anorg. Chem., 23. 383, 1900; W. C. Blasdale, Journ. Amer. Chem. Soc, 31. 917, 1909 ; C. Stolberg, Zeit. angew. Chem., 17. 741, 769, 1903 ; R Hefelmann, Zeit. offent. Chem., 3. 193, 1897; N. Knight, Chem. News, 89. 146, 1904 ; C. Liesse, Bull. Assoc.'Chim. Sucr. Dist., 28. 559, 1910 ; F. H. M'Crudden, Journ. Biol. Chem , 10. 187, 1911. 5 A Gooch's crucible may be used. 6 0. Brunck, Zeit. anal. Chem., 45. 80, 1906. They are made by the Kgl. Sachs. Muldenhiitte bei Freiberg i. S., and have proved very useful for general work in the laboratory. 7 Containing concentrated sulphuric acid. According to 0. Brunck (Zeit. angew. Chem., 17. 953, 1904), the oxide should be dried over sulphuric acid, because the carbon dioxide of the 214 A TREATISE ON CHEMICAL ANALYSIS. absorption of carbon dioxide and moisture 1 from, the atmosphere. If the pre- cipitate be large, ignit*e once again and reweigh in order to make sure that the conversion of the oxalate to oxide was complete. 2 Other Methods, of Weighing the Precipitate. Instead of weighing the calcium in the form of oxide, 3 some prefer to convert the oxalate into carbonate, or sulphate, or fluoride. The objection is made that in converting the oxalate into oxide the crucible loses weight during the 10 minutes' blasting, 4 and the oxide is difficult to weigh on account of its hygroscopicity. If small quantities are in question, it might be well to convert the oxalate into sulphate or fluoride. Care must then be taken to avoid loss by the spurting which will occur if the crucible be heated too rapidly. 5 Transformation of Calcium Oxalate to Calcium Sulphate. First ignite the oxalate at a comparatively low temperature, sufficient to burn the paper and convert most of the oxalate into oxide. Add one or two c.c. of water, gradually, drop by drop, to avoid spurting. This treatment will transform the calcium oxide into hydroxide. Add a slight excess of dilute sulphuric acid ; drive off the excess with a small flame; and finally ignite the crucible at a dull red heat. 6 Weigh the resulting calcium sulphate CaS0 4 and multiply the weight by 0*41195 in order to get the equivalent amount of CaO. Souchay 7 examined the results obtained by weighing the lime as oxalate, as carbonate, as sulphate, and as caustic lime, with the following results : Oxalate. Carbonate. Sulphate. Oxide. CaO ... 38-12 38-09 38 '06 38 "12 percent. Hence, it follows that, with care, all the methods furnish satisfactory results. In cases like this, where different methods furnish equally reliable results, the choice of any particular process is determined by convenience, risk of error, and time. Errors. The following numbers represent the results of eight independent determinations by the " oxalate to oxide " process on one sample of clay : 0-0137; 0-0141; 0-0147; 0-0142; 0-0152; 0-0152; O'OHl ; 0-0142; with a mean value of 0*0145 grm. CaO corresponding with 1*45 per cent. CaO. The deviations from the mean are 0'08. air may lead to the evolution of chlorine by the calcium chloride. It might be questioned whether this action could affect the result appreciably. 1 To illustrate the hygroscopicity of the powder, R. Fresenius (Quantitative Chemical Analysis, London, 2. 633, 1900) states that 0'5599 grm. of calcium oxide weighed 0'5605 grm. after standing 2 minutes on the pan of the balance ; 0*5609 grm. after standing 6 minutes ; and 0'5625 grm. after standing 17 minutes. 2 If the clay contains strontia, this will be precipitated with the lime. To remove the strontia, convert the precipitates to nitrates and digest the mixture with a mixture of absolute alcohol and ether (page 514). The calcium nitrate is washed away, and the insoluble strontium nitrate is ignited and weighed as SrO. The CaO is obtained by subtraction from the weight of mixed CaO + SrO. 3 A. Fritzsche, Zeit. anal. Chem., 3. 177, 1864. 4 0. Bnmck, Zeit. anal. Chem., 45. 77, 1906. 5 E. Kettler, Zeit. angew. Chem., 17. 685 ; 1904; 0. Brunck, ib., 17. 953, 1904; E. Murmann, Zeit. anal. Chem., 49. 688, 1910 ; A. N. Clark. Journ. Amer. Chem., Soc., 26. 110, 1904. F. B. Guthiie and C. R. Barker (Journ. Roij. Soc. N.S. W. , 36. 132, 1902) ignite the oxalate with ammonium nitrate equal to the bulk of the lime in the crucible ; and W. H. Hess (Journ. Amer. Chem. Soc., 22. 477, 1POO) then adds twice as much ammonium sulphate ; ignites at a dull red heat ; and weighs as calcium sulphate. A. Schrotter (Die Che)tne,Wien, 2, 1849) uses ammonium sulphate alone. For the decomposition of calcium and magnesium sulphates by heat, A. Mitscherlich, Journ. prakt. Chem. (1), 83. 485, 1861; J. Boussingault, Covn.pt, Rend., 64. 1159, 1867 ; W. Schiitz, Metallurgie, 8. 228, 1910. 6 The use of the ring burner, fig. 96, page 170, reduces to a minimum the risk of loss by spurting. 7 A. Souchay, Zeit. anal. Chem., 10. 323, 1871 ; R. Fresenius, ib., 10. 326, 1871. THE DETERMINATION OF CALCIUM AND MAGNESIUM. 215 99. The Volumetric Determination of Lime Kraut's Process. For routine work with calcareous clays, it is quickest to determine the lime volumetrically. 1 Spread the paper carrying the precipitated and washed calcium oxalate on the side of the beaker \ and wash the calcium oxalate from the paper by means of a jet of hot water, and then with dilute sulphuric acid (1 : 4). Remove the paper, add sufficient water to make the solution up to about 50 c.c. Add 10 c.c. of concentrated sulphuric acid, and titrate the hot solution (60-70) with standard permanganate, as indicated for sodium oxalate (page 194), until the solution is tinged with a permanent pink colour. The reaction which takes place during the titration is represented by the equation : 5CaC 2 4 + 2KMn0 4 + 8H. 2 S0 4 = K 2 S0 4 + 2MnS0 4 + CaS0 4 + 8H 2 + 10C0 2 . This shows that 1 grm. of KMn0 4 corresponds with 0*887 grm. of CaO. EXAMPLE. One gram of the sample was taken, The permanganate solution used in a titration contained 2'4 grins, of KMn0 4 per litre, and 20'5 c.c. were used. Here 1 c.c. of the permanganate has 0'0024 grm. ; hence, 20'5 c.c. have 0'0492 grm. of KMn0 4 ; and the sample contains 0'0492 x 0'887 = 0-0436 grm. CaO. The sample used for the analysis had 4 '36 per cent, of calcium oxide CaO. The method is not recommended when but a few determinations are made, and it is only used when a large number of analyses have to be conducted concurrently. ioo. The Properties of Ammonium Magnesium Phosphate. A precipitate of ammonium magnesium phosphate is obtained in the deter- mination of magnesia by adding a soluble phosphate to the ammoniacal solution containing the magnesium compound. Composition of the Precipitate. According to Neubauer, 2 the precipitation is practically complete, even in the presence of comparatively large quantities of ammonium salts, including the oxalate, but the composition of the precipitate is largely determined by the nature of the solution. The ammonium-magnesium phosphate may exist in three different forms, according to the composition of the mother liquid at the time of precipitation : 1. In neutral or ammoniacal solutions, the precipitate is Mg(NH 4 ) 4 (P0 4 ) 2 . This contains less magnesium than the normal MgNH 4 P0 4 . The former com- pound, on calcination, decomposes, forming magnesium metaphosphate, water, and ammonia : Mg(NH 4 ) 4 (P0 4 ) 2 -* Mg(P0 3 ) 2 + 4NH 3 + 2H 2 0. 1 W. Hempel, Menwires sur Vemploi de Vacide oxalique dans les dosages a liqueurs titrdes, Lausanne, 1853 ; K. Kraut, Henneberg's Landioirthsch. (1), 4. 112, 1856 ; Zeit, anal. Chem., 26. 629, 1887 ; L. T. Bowser, Journ. Ind. Eug. Chem., 3. 82, 1911 ; G. P. Baxter and J. C. Zanette, Amer. Chem. Journ., 33. 500, 1905; H. Walland, Chem. Ztg., 27. 922, 1906; C. H. Schultze, ib., 29. 508, 1905; B. Enright, Journ. Amer. Chem. Soc., 26. 1003, 1904 ; T. Ulke, Monit. Scient. (4), 14. 775, 1908 ; M. Kruger, Zeit. physiol. Chem., 16. 445, 1892 ; G. Lunge, Zeit. angew. Chem., 17. 265, 1904; J. Volhard, Liebig's Ann. , 198. 333, 1879; E. Rupp and A. Bergdolt, Archiv Pharm. , 242. 450, 1890 ; C. A. Peters, Zeit. anorg. Chem., 29. 145 r 1902 ; H. M. Davy, Chem. News, 29. 250, 1874 ; Compt. Rend., 78. 978, 1874 ; A. Heifer, Tonind. Ztg., 18. 535, 1894 ; K. von Radlowski, ib., 18. 592, 3894. 2 H. Neubauer, Ueber die Zuverlassigkeit der Phosphorsaurebcstimmung als Magnesiumpyro- phosphat, Rostock, 1893 ; Journ. Amer. Chem. Soc., 16. 289, 1894; W. Heintz, Zeit. Chem., (2), 6. 479, 1870; 0. Popp, ib. (2), 6. 395, 1870; K. Bube, Ueber Magnesiumammonium- phosphat, Weisbaden, 1910 ; R. Reidenbach, Ueber die quantitative Bestimmung des Magnesiums als Magnesiumpyrophosphat, Kusel, 1910. 2l6 A TREATISE ON CHEMICAL ANALYSIS. On further calcination, the metaphosphate decomposes into magnesium pyro- phosphate and phosphoric anhydride : and some phosphoric anhydride is volatilised. 1 2. When an excess of magnesium salt is present, and no excess of ammonia, the precipitate has the normal composition MgNH 4 P0 4 and the results are correct. 3. If an excess of magnesium salt and an excess of ammonia be present, the precipitate contains more magnesium than the normal phosphate, and the calculated phosphoric acid will be too high. Nerl has verified the first, and Reidenbach the second observation of Neubauer ; but Reidenbach considers that the third observation is not correct. He found the precipitate contained less, not more, magnesium than Neubauer's statement represents. In any case, it is necessary to precipitate the ammonium- magnesium phosphate in a solution containing definite amounts of ammonia, ammonium salts, magnesia, and phosphoric acid in order to obtain concordant results. The precipitant should also be added to the acid solution, and the ammonia in slight excess added afterwards. Solubility of Ammonium- Magnesium Phosphate. The precipitate is readily soluble in dilute acids, and 100 c.c. of water at 10 dissolve 0'0065 grm. of the salt. It is much less soluble in aqueous ammonia. 2 Thus, according to Stiinkel, Wetzke, and Wagner, the amounts of ammonia indicated in the first line of the following scheme dissolve per litre : Ammonia ..... 1 2 3 per cent. MgO ...... 0-00050 0-00023 '00008 grm. P 2 5 . . . . ' . . 0-00088 00038 '00015 ,, The solubility of the ammonium magnesium phosphate in ammonia is also illustrated by graph, fig. 110, which shows that the solubility of normal ammonium magnesium phosphate decreases very rapidly with increasing con- centration of ammonia. 3 According to Jorgensen, the solubility in 2'5 per cent. ammonia is approximately 0-00006 grm. of MgO per 100 c.c. This is negligibly small. 4 The solubility is increased in the presence of ammonium chloride, so that O'OOIS grm. is dissolved per 100 c.c. in the presence of 2 -5 per cent. ammonia containing a gram of ammonium chloride. The effect of ammonium chloride on the solubility of the magnesium ammonium phosphate is illustrated by the graph, fig. 111. 1 D. Campbell, Phil. Mag. (4), 24. 380, 1862. 2 G. Jorgensen, Mem. Acad. Roy. Soc. Dane mark (7), 2. 141, 1905 ; C. Stiinkel, T. Wetzke, and F. Wairner, Zeit. anal. Chcm,, 21. 353, 1882 ; T. R. Ogilvie, Chem. News, 31. 274, 1875 ; 32. 5, 12, 70, 1875 ; E. W. Parnell, ib., 32. 222, 1875 ; 23. 145, 1871. 3 According to A. Bolis (Chem. Ztg., 27. 1151, 1903), 2 grms. of MgNH 4 P0 4 . 6H 2 in contact with 100 c.c. of a solution of ammonium citrate (400 grms. of citric acid per litre) loses, by solu- tion, an average 0'457 per cent, in weight at ordinary temperatures, and 0'587 per cent, at 50. 4 There is therefore no need to correct for the solvent action of the wash liquids as recom- mended by C. R. Fresenius, Anleitung zur quantitative chemischtn Analyse, Braunschweig, I. 333 1863; Liebig's Ann., 55. Ill, 1845 ; H. Warington, Journ. Chem. Soc., 18. 27, 186f> ; W. Kiibel, Zeit. anal. Chem., 8, 125, 1869 ; O. Abesser, W. Jani, Marcker, ib., 12. 239, 1873 ; F. A. Gooch, Amer. Chem Journ., I. 391, 1879 ; T. S. Gladding, Chem. News, 46. 213, 1882 ; 47. 71, 1883. To get some idea of the effect of the solubility of the precipitate, three solutions containing the same amount of magnesia were treated in exactly the same manner, but the precipitates were respectively washed with a litre of 1, 2, and 3 per cent, ammonia. The corresponding precipitates gave 0'1956, 0-1967, 0'1968grm. of ammonium magnesium phosphate. Theory required 01971 grm. Nothing like a litre of washing liquid is required in practice, so that the errors from this cause must be negligibly small. THE DETERMINATION' OF CALCIUM AND MAGNESIUM. 217 The presence of ammonium chloride, curiously enough in view of fig. Ill, furnishes precipitates with positive, not negative errors. Otherwise expressed, the results were too high. For example, in solutions containing NH 4 C1 . Mg 2 P 2 7 . Error 0-5567 2-67 0-5603 + 0-0036 5-35 0-5612 + 0-0045 10-70 0-5619 + 0-0052 grms. per litre grm. The solution from which the magnesium phosphate is first precipitated usually contains not only ammonium chloride and ammonia, but also ammonium oxalate, and considerable amounts of sodium chloride. The effect of ammonium oxalate, 10 - .Co. 1-0 2-0 FIG. 110. Effect of ammonia on the solubility of ammonium magnesium phosphate. FIG. 111. Effect of ammonium chloride on the solubility of ammonium magnesium phosphate. like ammonium chloride, is to raise the weight of the precipitate above the normal. Thus, in solutions containing (NH 4 ) 2 C 2 4 M g2 ? 2 7 Error 0-5567 3-55 5900 + 0-0333 4-26 0-5864 + 0-0297 4-97 0-5863 + 0-0296 grms. per litre grm. The presence of sodium chloride in the mother liquid also considerably augments the weight of the resulting precipitate. Thus, in solutions containing Nad Mg 2 P 2 7 Error 0-585 5-85 23 '4 grms. per litre 0-5567 0-5585 0'5689 0'5770 grm. + 0-0018 +0-0122 +0-0203 The effect with potassium chloride is much greater than with sodium chloride. The practical lesson to be learned from these observations is that the magnesium ammonium phosphate first precipitated must be dissolved in dilute acid, and reprecipitated, otherwise the result will be of little value. The precipitate is more soluble in hot solutions. With hot solutions there is also a loss of ammonia. Hence, the precipitation should be made in colci solutions, with an excess of the precipitating agent, and in the presence of at least 2'5 per cent, of ammonia. The precipitate made in solutions containing ammonium salts chloride and oxalate and sodium chloride is very impure, and it must in consequence be dissolved in dilute acids, and reprecipitated as described below. Action of Heat. Ammonium magnesium phosphate Mg^H 4 P0 4 .6H 2 loses 5H 2 at about 100, and the remaining H 2 with the ammonia, at a red heat, forming magnesium pyrophosphate. 2MgNH 4 P0 4 . 6H 2 -> Mg 2 P 2 O, + 2NH 3 + 14H 2 0. 2l8 A TREATISE ON CHEMICAL ANALYSIS. If the temperature be raised still further, the pyrophosphate becomes incan- descent, owing to an intermolecular change according to Popp, 1 a passage from the crystalline to an amorphous condition. The magnesium pyrophosphate fuses at 1220. If magnesium pyrophosphate be exposed to a reducing atmosphere at a high temperature, phosphorus, phosphorus hydride, and phosphorus oxide are said to be volatilised. 2 These vapours attack the platinum crucibles. The pyrophosphate is soluble in dilute nitric and hydrochloric acid, and but sparingly soluble in water. Precipitation of Ammonium Magnesium Phosphate for the Determination of Phosphorus. An excess of magnesium chloride (magnesia mixture, page 597) reduces the solubility 3 of magnesium ammonium phosphate even more than an excess of the phosphate solution. No phosphoric acid, for example, could be detected in a nitrate which would, without the excess magnesium chloride, have contained the equivalent of O0025 grm. P 2 5 . An excess of a soluble phosphate also reduces the solubility of the ammonium magnesium phosphate in a similar manner. ioi. The Gravimetric Determination of Magnesia. The magnesia is determined in the filtrate from the calcium oxalate by the addition of a soluble phosphate ; a precipitate of ammonium magnesium phosphate separates. Unfortunately, the composition of the precipitate is considerably modified by the composition of the solution in which the precipita- tion takes place. 4 The precipitate is therefore redissolved, and reprecipitated under definite conditions whereby the precipitate MgNH 4 P0 4 is obtained. This is ignited and weighed as Mg 2 P 2 7 . First precipitation. Gradually add a solution containing, say, 2 grms. of sodium ammouiufu phosphate 5 (dissolved in 15 c.c. of water) to the solution under investigation, 6 with constant stirring. While still stirring the 1 0. Popp, Zeit. anal. Chem., 13. 305, 1874. 2 H. Struve, Journ. prakt. Chem. (1), 79. 349, 1860 ; R. Weber, Pogg. Ann., 73. 146, 1848. 3 W. Heintz, Zeit. anal. Chem., 9. 16, 1870 ; E. Kessel, ib., 8. 173, 1869 ; W. Kiibel, ib., 8. 125, 1869 ; R. Weber, Pogg. Ann., 73. 139, 1848. Basic magnesium phosphate is precipitated if the magnesia mixture be in large exce.ss, and magnesium sulphate is used in place of magnesium chloride for compounding the magnesia mixture (page 597). " 4 H. Struve, Zeit. anal. Chem., 36. 289, 1897 ; 37. 485, 1898 ; K. K. Jarvinen, ib., 43. 279, 1904 ; 44. 333, 1905 ; H. Fresenius, H. Naubauer, and E. Luck, ib., IO. 133, 1870 ; C. Schumann, ib., ii. 382, 1872 ; H. Schmidt, ib., 45. 512, 1906 ; H. Neubauer, Zeit. angew. Chem., 9. 435, 1896 ; F. Raschig, ib. t 18. 374, 1905 ; F. A. Gooch and M. Austin, Amer. Journ. Science (4), 7. 187, 1899 ; W. Gibbs, ib. (3), 5. 114, 1873 ; Chem. News, 28. 51, 1873 ; R. W. C. Maclvor, ib., 28. 69, 1873 ; T. R. Ogilvie, ib. t 21. 205, 1870 ; F. A. Gooch, Amer. Chem. Journ., I., 391, 1879; A. K. Christomanos, Zeit. anorg. Chem., 41. 305, 1904; T. S. Gladding, Journ. Amer. Chem. Soc., 4. 135, 1882 ; Chem. News, 46. 213, 1882 ; E. Raffa, Gas. Chim. Ital, 38. ii., 556, 1908. 5 J. J. Berzelius (Lehrbuch der Chemie, Dresden, 2. 650, 1826) used disodium phosphate. C. Mohr prefers sodium ammonium phosphate microcosmic salt (Zeit. anal. Chem., 12. 36, 1873 ; W. Gibbs, I.e.) because it precipitates more rapidly and completely. By precipitating more and more dilute solutions of magnesia, he arrived at a point where sodium phosphate no longer gave a precipitate under conditions where sodium ammonium phosphate did. L. Blum (Zeit. anal. Chem., 28. 452, 1889 ; W. Heintz, ib.,g. 16, 1867) prefers sodium phosphate, because the precipitate settles more rapidly than when sodium ammonium phosphate is the precipitating agent. 6 In order to get rid of the ammonium salts which have accumulated in the solution which, it will be remembered, has been treated for both alumina and lime, some prefer to evaporate the filtrate nearly to dryness in a large dish. Add concentrated nitric acid ; evaporate the solution to dryness, and heat the residue until all the ammonium salts have volatilised. Dissolve the mass in water, filter off any insoluble matter, and treat the solution as described in the text. The insoluble residue may be examined for alumina. This, if present, is precipitated, washed, and weighed as " extra alumina." J. Jambor, Zeit. anal. Chem., 49. 733, 1910. See page 224. THE DETERMINATION OF CALCIUM AND MAGNESIUM. 219 solution, 1 gradually add, drop by drop, about one-third its volume of aqueous ammonia; cover the solution with a watch-glass, 2 and let it stand 12-24 hours. 3 Filter. Wash the precipitate with dilute ammonia (1 : 10, that is, about 2'5 per cent.). 4 Reject the nitrate, which should give a precipitate with "magnesia mixture" (page 597). When the washing is completed, the runnings will give no precipitate with silver nitrate in acid (HN0 3 ) solution. Reject the nitrate and washings. Second Precipitation. Redissolve the precipitate in warm dilute nitric acid (1 : 5), and collect the washings in the beaker in which the first precipitation was made. About 50 c.c. of acid are needed. While the acid is running through the funnel, turn the beaker round so that the acid runs all round the sides of the beaker. Add an aqueous solution containing a little sodium ammonium phosphate, and then add ammonia gradually with constant stirring, as in the first precipitation ; wash as before ; reject the filtrate and washings. Ignition of the Precipitated The precipitate may be collected on a Gooch's asbestos 6 crucible prepared in the usual manner. After the crucible has been dried and heated slowly over a Bunsen burner (in a Gooch's crucible saucer, page 106), it is heated over the Teclu's or Maker's burner for about 10 minutes. Cool in a desiccator and weigh. Ignite 4 or 5 minutes, cool, and weigh again. Repeat the ignition, if necessary, until the weight is constant. If the precipitate be collected on a 7-cm. filter paper, the moist paper is placed in the crucible, dried and carbonised very slowly in the (platinum or porcelain) crucible, 7 placed obliquely on a triangle over, say, an argand burner, 1 This point requires careful attention. The precipitate will be more contaminated with impurities if quickly made. In illustration, a mean of six experiments in which the solutions were mixed suddenly gave 0'2028 grm. of Mg 2 P 2 7 ; another six experiments with the ammonia gradually mixed with constant stirring gave 0'1972 grm. Mg 2 P 2 7 . Theory required 0'197l grm. Mg 9 P 2 7 . The stirring rod should be kept from the sides and bottom of the beaker. See H. Lasne, Bull. Soc. Chim. (3), 17. 823, 1897 ; Chem. News, 76. 270, 1897. 2 Some keep the beaker under a bell jar resting on a greased ground-glass plate in order to prevent the evaporation of ammonia. 3 Precipitates which take a long time to separate on standing frequently come down quickly if the solution be vigorously ngitated by, say, bubbling a current of air through the solution, or mechanical agitation L. Briant, Chem. News, 53. 99, 1886 ; 0. Texter, Journ. Anal. App. Chem., 7. 279, 1893 ; V. Markoonikotf, Liebig's Ann., 289. 254, 1895; H. B. Yardley, New Remedies, g. 333, 1880. 4 There is no particular need to here clean the precipitate from the sides and bottom of the beaker. 5 K. Brookman (Zeit. anal. Chem., 21. 551, 1882) dissolves the precipitate in nitric acid, and evaporates the solution to dry ness in a weighed dish, ignites, etc. The object is to avoid loss by (1) the sticking of the precipitate to the walls of the beaker ; (2) the " crawling" of the finely divided precipitate above the edge of the filter -paper on to the funnel during washing ; (3) as "dust" during the transfer of the dry filter paper to the crucible; and (4) loss as ' ' dust " during the collapse of the ash of the filter paper during the ignition in the crucible. L. L. de Koninck (Zeit. angew. Chem., 2. 187, 1888; Zeit. anal. Chem., 29. 165, 1890; R. Fresenius, ib., 15. 224, 1876 ; 16. 63, 1897) recomrnends a similar procedure for precipitates liable to reduction from the combustion of the filter paper, e.g., ammonium and potassium chloroplatinates, ammonium magnesium arsenates and phosphates ; cadmium and zinc carbonates, etc. H. N. Warren (Chem. News, 61. 63, 1890) recommends a plug of gun-cotton (pyroxylin) in place of asbestos. This burns away on ignition. . 7 It is not generally advisable to rest the triangle on the chimney of the argand burner, because the draught may be choked. The crucible is best supported on a triangle as shown in the diagram. If the precipitate be large, it should be dried and separated from the filter paper, so that the filter paper can be ignited alone. If a platinum crucible be employed for the ignition of the phosphate, and reducing agents (like carbon from the filter paper, reducing gases, hydrogen liberated from the decomposition of ammonia at high temperatures, etc. ) be present, phosphides maybe formed. These attack the platinum (W. C. Heraeus, Zeit. angew. Chem., 15. 917, 1902; Chem. News, 97. 102, 1903 ; W. P. Headden, Proc. Colorado Scientific Soc., 8. 45, 1905). Under ordinary conditions, however, there is little danger of this occurrence 22O A TREATISE ON CHEMICAL ANALYSIS. fig. 112. The temperature is gradually raised to redness, and there maintained until the precipitate appears white. 1 Finish the ignition over a weak blast. It is a bad practice to remove the last trace of carbon by blasting. The precipitate should be white before the blast is applied. If the contents of the crucible have a dirty white appearance, Fresenius 2 recommends moistening the FIG. 112. Ignition over Argand burner. precipitate with a few drops of nitric acid. Dry, and ignite as before. There is always a slight loss after the nitric acid treatment, 3 probably owing to the fact during the ignition of the magnesium phosphates if the temperature of the blast be not high enough to fuse the pyrophosphate. The action, however, may easily occur if Gooch's platinum crucibles containing old precipitates be employed. The action is most marked if the ignition be conducted quickly in covered crucibles. 1 When the precipitate is ignited too rapidly, there is a rapid shrinking and sintering. Particles of carbon are liable to be enclosed with the precipitate. It is then almost impossible to burn off the carbon by ignition, even in a blast (R. Bunsen ; F. Muck, Zeit. anal. C/tetn., 19. 131, 1880). L. L. de Koninck (ib., 29. 165, 1890) considers the frequent blackening of magnesium pyrophosphate is not due to particles of carbon derived from the filter paper, but to the presence of organic bases (e.g. pyridine) in commercial ammonia and its salts. - R. Fresenius, Quantitative Chemical Analysis, London, 2. 190, 1876. 3 According to D. Campbell (Chem. News, 6. 206, 1862), the treatment of magnesium pyrophosphate with nitric acid is objectionable, because the liberated phosphoric acid may volatilise at the high temperatures required for expelling the last traces of the nitric acid. J. 0. Handy, Journ. Amer. Chem. Soc., 22. 31, 1900; E. Luck, Zeit. anal. Chem., 13. 255, 1874. THE DETERMINATION OF CALCIUM AND MAGNESIUM. 221 that some H 4 Mg(P0 4 ) 2 is formed. This on ignition forms magnesium meta- phosphate Mg(P0 3 ) 2 which decomposes as indicated on page 216. After ignition to constant weight, cool in a desiccator, and weigh as Mg 2 P 2 7 . This weight multiplied by 0'3620 l gives the equivalent weight of MgO. 2 Instead of multiplying, the result may be read at once from Table XC. Corrections. Magnesia determinations are usually a little high owing to the presence of some impurities lime, manganese, silica, etc. (a) Lime. There is not usually sufficient magnesia present in clays to render it advisable to apply Hillebrand's correction for lime. 3 If, however, over, say, 2 per cent, of magnesia be present, the correction may be made if the work is intended to be exact. Digest the pyrophosphate in a little dilute sulphuric acid, and add 9 to 9J times its volume of absolute alcohol. After standing overnight, filter off the scarcely visible precipitate of calcium sulphate, and wash it free from phosphates by means of alcohol. Dry the precipitate ; dissolve in hot water slightly, acidulated with hydrochloric acid, and precipitate the lime in ammoniacal solution by means of ammonium oxalate (page 213). Filter, wash, ignite, and weigh as CaO. Add the result to the lime, and subtract it from the magnesium pyrophosphate. (b) Manganese. If this element 4 be present, it can be determined colori- metrically in the nitric acid solution of the precipitate. If present in the filtrate from the magnesia, evaporate to dryness; drive off the ammoniacal salts by ignition; and again evaporate the residue to dryness two or three times with nitric acid (or once with sulphuric acid) to drive off' the chlorine. Determine the manganese colorimetrically, as indicated on page 382 ; and make the necessary allowance. (c) Silica. If silica be present, the magnesium pyrophosphate will leave a residue 5 when treated with the dilute sulphuric acid. This is filtered off, ignited, and weighed as indicated on page 185.- The amount of silica is sub- tracted from the magnesium pyrophosphate. (d) Barium. See page 517 for barium in silica. 6 Errors. Some idea of the results may be gathered from the eight inde- pendent determinations on one sample of clay : 0-0106; 0-0114; 0*0109; 0-0110; 0-0107 ; 0-0103; 0-0117; 0-0118; with a mean of O'OllO grm. MgO, representing I'lO per cent, of MgO. The deviations from this value range between 0'08. 1 Or by 0-6378 for the equivalent weight of P 2 5 page 597. 2 M. Schmorger, Zeit. anal. Chem., 37. 308, 1895 ; H. Mastbaum, ib., 37. 581, 1898. 3 W. F. Hillebrand, Bull. U.S. Geol. Sur., 422. 127, 1910. 4 The magnesium pyrophosphate has generally a pink colour if appreciable amounts' of manganese be present. Some platinum sulphide may also be present (see page 186). B. Tollens (Journ. Landw. , 3 30. 48, 1883) says that if basic lime or magnesia phosphates are precipitated with the ammonium magnesium phosphate, and a little silver nitrate be added, a yellow colour will be developed when the precipitate is warmed. CHAPTER XVII. THE DETERMINATION OF THE ALKALIES. 102. Meretricious Methods for Estimating the Alkalies. 1. Difference Method. We shall see very shortly that the sum of the different constituents in a clay analysis may vary between 99'5 and 100*5. Assuming that the tolerated errors in the determination of the different constituents correspond with a total lying within 99 '5 and 100*5, it follows that if the alkalies be determined by summing the different constituents actually determined, and subtracting the result from 100, the alkalies may be 0'5 per cent, greater or less than the number obtained by the process of subtraction. This means that a clay with 1 per cent, of alkalies might be reported with or 1 per cent., and the numbers mean that the clay might have anything between 0*5 and 1*5 percent., on the assumption that the analysis has been conducted accurately. The number representing the alkalies would therefore be quite misleading unless interpreted as I have just indicated. It would certainly be better and more honest to report, " Alkalies not determined." It will also be noticed that when a constituent say the alkalies is determined by difference, and the whole analysis totals 100, there is no check on the accuracy of the analysis, and the numbers are accordingly under more or less suspicion. 2. Calculation from the Supposed Quantities of Potash and Soda. There is another unsatisfactory method employed for estimating the alkalies. 1 The mixed chlorides are isolated and weighed. It is then assumed that the potassium and sodium chlorides are present in equal or some other proportions. The correspond- ing potash and soda are then calculated by arithmetic and reported as " alkalies." This method is possibly less objectionable than the preceding one, but it too must be condemned. It pretends to represent by number a magnitude which has not really been determined. 2 There is no method known for determining alkalies without separating them into "potash and soda." 103. The Separation of the Alkalies as Mixed Chlorides- Smith's Process. The alkalies in silicates are usually determined by first decomposing the clay by hydrofluoric acid, or by fusion with bismuth, lead, or boric oxides, etc. 3 The constituents other than alkalies are then removed. The mixed alkali salts remain behind as a residue. 4 One of the two following methods is generally employed. The 1 At present rather too common. 2 In illustration of both these vicious methods, see the two clay analyses, Brit. Clayworker, ig. 155, 1910. 3 E. Makinen (Zeit. anorg. Chem., 74. 74, 1912) fuses the silicate with calcium chloride, and otherwise proceeds as described in the text. 4 E. Bonjean, Chem. News, 80. 240, 1899 ; A. Verweij, Zeit. anal. Chem., 48. 760, 1909 ; E. W. Dbrfurt, ib. t 51. 755, 1912. 222 THE DETERMINATION OF THE ALKALIES. 223 most important, and one of the best for general work, is that devised by J. Lawrence Smith. 1 The following are the directions : The Crucible and Furnace. An ordinary covered platinum crucible may be employed, but great care is then necessary not to heat the crucible at too high a temperature, nor should the crucible be heated to redness more than three-fifths of its height. The crucible recommended by Smith is best. It is a platinum crucible about 8 cm. long, 1'8 cm. wide at the mouth, and 1*5 cm. wide at the bottom. This crucible is placed at an angle of about 45 through the side of a fireclay cylinder, and heated to the full temperature of a Bunsen's burner as described below. Wrap a thin strip of asbestos paper round the crucible. This strip comes between the crucible and the fireclay jacket. It prevents the crucible jamming tightly in the aperture of the fireclay cylinder. The disposition of the FIG. 113. Furnace for L. Smith's process. fireclay cylinder and crucible is indicated in fig. 113. The source of heat is an ordinary Bunsen's burner with a flat-flame nozzle (fig. 113). Preparation of the Sample for the Decomposition. Spread about half a gram of calcium carbonate 2 as a layer on the bottom of the crucible. Intimately mix 1 J. L. Smith, Amer. J. Science (2), 50. 269, 1871 ; Chem. News, 23. 222, 234, 1871 ; P. Holland, ib., 54. 242, 1886 ; F. Field, ib., I. 193, 217, 1860 ; T. Boring, Zeit. anal. Chem., 49. 158, 1910. R. L. Steinlen (Chem. Ztg., 29. 264, 1905 ; B. M. Margosches, ib., 29. 385, 1905) has devised a special cooler for the lid of the crucible, to prevent loss of alkalies by volatilisation. 2 CALCIUM CARBONATE. This is best made by dissolving calcite in hydrochloric acid. Heat the solution of calcium chloride to boiling, and add a hot concentrated solution of ammonium carbonate. Wash the precipitate thoroughly with hot water on a Biichner's funnel (page 103), and dry. This procedure gives a dense, coarse granular powder very suitable for this work. The calcium carbonate is generally contaminated with sodium chloride, but free from potassium chloride. This circumstance, and the fact that alkalies may be dissolved by the action of hot water on the glass vessels used in the analysis, render it necessary to find what correction must be made for the alkalies derived from the reagents, glass vessels, etc. This is done by treating the ^ grm. of ammonium chloride, and the 4^ grms. of calcium carbonate without the clay, 224 A TREATISE ON CHEMICAL ANALYSIS. half a gram of ammonium chloride l with half a gram of the finely powdered and dried clay, and 3 grms. 2 of calcium carbonate. Transfer the mixture from glazed paper into the platinum crucible. Rinse the mortar and pestle and paper with another gram of calcium carbonate, and transfer the rinsings to the crucible. Ignition. Heat the crucible, inclined as indicated above, very gently over a small flame for about 15 minutes, in order to volatilise the ammonium chloride. Then raise the temperature until the lower three-fifths of the crucible is at a dull red heat, and maintain this temperature for about an hour. Let the crucible cool. Beginners generally err by heating the crucible at too high 3 a temperature, when the mass vitrifies and then disintegrates with difficulty when treated with water. Dissolution of the Cake in the Crucible. Remove the cake from the crucible by adding 2 to 3 c.c. of water. After five or ten minutes, add more water. The sintered cake, not fused, usually comes away from the crucible quite readily. Transfer the contents of the crucible to a large porcelain or platinum dish. Heat the mass for about half an hour on the water bath with about 50 c.c. of water, 4 and restore that lost by evaporation from time to time. Triturate any large particles to powder with a small pestle. Decant the clear liquid through a filter paper. Wash about four times by decantation, and transfer the residue to the filter paper. Wash until the washings only give a faint turbidity with silver nitrate. 5 If the clay has been perfectly decomposed, the residue on the filter paper should leave no undecomposed residue when treated with hydrochloric acid. 6 Removal of Lime. Add ammonia and ammonium carbonate to the filtrate. Heat to boiling ; filter ; and again digest the precipitate with ammonia and ammonium carbonate. Filter, and allow the filtrate to collect with that from the previous filtration. Evaporate 7 the filtrate to dryness in a platinum or porcelain dish. Remove the ammonium salts by gentle ignition in a moving flame. 8 Treat the cold mass with water. 9 Remove the last traces of lime by the exactly as if the clay were present. A blank test made in this way gave '00022 grm. Na 2 and no K 2 for the materials quoted in the text. See M. Grager, Neues Jahrb. Pharm., 29. 158, 1868. For barium carbonate in place of calcium carbonate and ammonium chloride, see G. Werther, Journ. prakt. Chem. (1), 91. 321, 1864. 1 AMMONIUM CHLORIDE. The ammonium chloride is made by neutralising pure ammonia with pure hydrochloric acid, or by resublimin^ commercially pure ammonium chloride. J S. Stas, (Euvres Completes, Bruxelles, I. 468, 1894 ; Chem. New*, 15. 194, 217, 231, 1867. 2 If the silicate contains much iron, or is liable to sinter with the lime, increase the amount of calcium carbonate. 3 A. Verweij (Zeit. anal. Chem., 48. 760, 1909) uses a platinum crucible 4*5 cm. high and 3 '5 cm. in width. He covers the mixture with a layer of 3 grms. of calcium carbonate, and after the expulsion of ammonia, heats the mixture for an hour over a Teclu's or powerful Bunsen's burner. The cold mass is boiled 15 minutes with water, etc. 4 According to T. Doring (Zeit. anal. Chem., 49. 158, 1910), at least 500 c.c. of filtrate should be obtained per gram of sample, because the alkaline chlorides are retained tenaciously by the insoluble mass. 5 The faint turbidity is possibly due to the presence of calcium oxychlorides which are slowly dissolved from the residue. Borates will be found in the insoluble residue. 6 Some flecks of silicic acid may separate. 7 G. H. Bailey (Journ. Chem. Soc., 65. 445, 1894) noticed a loss of alkaline chlorides during the evaporation of f-yN- solutions of alkaline chloride. For instance, lithium chloride lost 0'25 mgrm. per litre; sodium chloride, 0'81 mgrm. per litre; potassium chloride, 1'22 mgrms. ; rubidium chloride, 2*95 mgrms. ; and caesium chloride, 3*35 mgrms. per litre. 8 In specially exact work, the removal of the ammonium salts is best effected by heating the dish placed high above the flame (fig. 100). This requires about 45 minutes. If the ignition be conducted too rapidly, or the dish be too strongly heated, several milligrams of alkali may be lost; otherwise, A. Mitscherlich (Journ. prakt. Chem. (1), 83. 459, 1861; J. Boussingault, Compt. Rend., 64. 1159, 1867) has shown that there is no loss of alkali during the expulsion of the ammonium salts. See page 218. 9 If sulphur be present, add a drop of barium chloride solution, and remove the excess of barium by means of ammonium carbonate. THE DETERMINATION OF THE ALKALIES. 225 addition of ammonium oxalate to the boiling solution, and let the mixture stand overnight about 12 hours. 1 Determination of the Mixed Alkaline Chlorides. Filter the liquid into a weighed platinum dish. Evaporate to dryness. Ignite. Cool. Moisten the residue with concentrated hydrochloric acid. Again evaporate to dryness. Ignite gently. 2 Weigh. The increase in weight represents the weight of the mixed sodium and potassium chlorides. 3 The Accuracy of the Results. The following numbers represent the results obtained with eight independent determinations on half-gram samples of the same clay : 0-0226; 0-0229; 0'0233 ; 0'0229 ; 0'0228; 0'0221 ; 0'0225; 0-0222. The mean value is 0-0227 grm. per half gram of clay ; the deviations range approximately 0*0006 per grm. of clay. There is also a constant error due to the loss of alkali. L. Smith says : " Usually an amount of alkali remains behind amounting to 0'2 to I'O percent, of the materials used." Smith recovers this by reheating the residue after the first sintering with the ammonium chloride mixture, and mixing the aqueous extract with the extract of the sintered mass obtained during the first heating. By proceeding as described above, the errors from this source will rarely exceed 0'002 grm. Holland has investigated the magnitude of the loss due to the retention of alkalies by the precipitate during the first extraction. He obtained : Table XXXIX. Loss of Alkalies in Smiths Process. Nature of silicate. Mixed chlorides, in grms. One fusion. Re-fusion of residue. Basalt (Wales) 0-1305 0-1118 0-2472 01434 0-1634 0-0040 0-0032 0-0034 0-0038 0-0029 Basalt (Westmoreland) .... Leucite (Rieden, Germany) . . . Red spongy lava (Pompeii) .... Syenite (North Wales) Hence it is very probable that practically all but about 2 per cent, of the total alkali is extracted during the first operation. Under these conditions, about 2 per cent, of the total alkali is lost. This agrees with Dittrich's observation cited Table XLIIL, page 247, and the result may be taken to represent the constant incidental to the process. The other chief errors arise from (1) too high a tem- perature for the fusion ; (2) imperfect leaching of the fused mass ; (3) driving off the ammonium chloride too rapidly ; (4) spitting by too rapidly heating to drive off the ammonium salts; (5) igniting the mixed chlorides at too high a 1 Schaffgotsch's or Gooch and Eddy's solutions can also be used for the removal of lime and magnesia, as described in the next section. 2 A. Mitscherlich (Journ. prakt. Chem. (1), 83. 485, 1861) has shown that a six minutes' heating to the melting point of potassium chloride resulted in a loss of 0'0040 per cent., and 0-0042 per cent, of sodium chloride H. Rose, Pogg. Ann., 31. 133, 1833; G. H. Mulder, Archiv Pharm. (2), 129. 231, 1867 ; Reichmann, Journ. Gasbeleuchtung, 7. 9, 1864 ; P. M. Delacharlonny, Compt. Rend., 103. 1128, 1886; H. B. von Adlerskron, Zeit. anal. Chem., 12. 390, 1873. 3 If the residue dissolves in water, all is well. If not, filter off the insoluble matter, ignite and weigh. Deduct the weight of the insoluble residue from the weight of the total chlorides. 15 226 A TREATISE ON CHEMICAL ANALYSIS. temperature; (6) imperfect separation of magnesium and calcium salts; and (7) the presence of sulphates. If the silicate contained appreciable quantities of sulphur, this element will be found as alkali sulphates along with the chlorides. The sulphates can be converted into chlorides by adding a little barium chloride before the final precipitation of the calcium is made. The excess of barium is removed by the ammonium carbonate and oxalate treatments. 104. The Separation of the Alkalies as Mixed Chlorides Berzelius' Process. It is sometimes convenient to decompose the silicate by a mixture of sulphuric and hydrofluoric acids, as recommended by Berzelius. 1 The alumina, lime, magnesia, etc., may be precipitated by mercuric oxide, 2 barium oxide, 3 or ammonium carbonate. 4 The alkali salts remain in solution. The following plan gives good results. It is rather quicker than Smith's process, but more expensive materials are needed. The "yield" of mixed chlorides is also about 1 per cent, (on the "total alkali") higher than in Smith's process. Berzelius' process is not recommended when the silicate contains boron compounds. In Smith's process the boron remains behind as insoluble calcium borate. Removal of Silica, Alumina, and Iron. Mix 1 grm. of the clay in a platinum crucible with 50 c.c. of concentrated sulphuric acid, and add carefully, in small quantities at a time, about 50 c.c. of hydrofluoric acid. 5 Heat gently on a sand bath ; when the hydrofluoric acid has evaporated, add another 50 c.c. of hydrofluoric acid, and heat as before. When the hydrofluoric acid has nearly gone, evaporate the solution to dryness. Cool. Add concentrated aqueous ammonia, and heat gently. When the mass is quite disintegrated, filter and wash with hot water. Transformation of the Sulphates into Chlorides. It is now advisable to trans- form the sulphates in the filtrate into chlorides 6 by precipitating the sulphates 1 J. J. Berzelius, Pogg. Ann., I. 169, 1824; E. A. Wiilfing, Ber., 32. 2214, 1899; C. Reinhardt, Stahl Eisen, 16. 448, 1896 ; W. Knopp, Zeit. anal. Chem., 22. 421, 558, 1883 ; 18. 462, 1879 ; Chem. News, 48. 110, 1883 ; A. H. Low, ib., 67. 185, 1893 ; Journ. Anal. App. Chem.,t. 666, 1892. 2 E. F. Smith and P. Heyl, Zeit. anorg. Chem., 7. 82, 1894 ; Chem. News, 70. 193, 204, 1894 ; G. Starck, Zeit. anal. Chem., 48. 415, 1909. 3 J. N. von Fuchs, Schweigger's Journ., 62. 184, 1831 ; H. Rose, Pogg. Ann., 83. 137, 1851 ; C. Zimmermann ; J. E. Thomsen, Journ. Amer. Chem. Soc., 30. 420, 1908. 4 F. G. Schaffgotsch, Pogg. Ann., 104. 482, 1858; H. Weber, Vierteljahr. prakt. Pharm., 8. 161, 1859. Schaffgotsch's solution is made by dissolving 230 grms. of ammonium carbonate in 180 c.c. of aqueous ammonia (sp. gr. 0'92) and making the solution up to a litre. F. A. Gooch and E. A. Eddy, Chem. News, 97. 280, 1908 ; Amer. J. Science (4), 25. 444, 1908 ; F. A. Gooch and M. A. Phelps, ib. (4), 22. 488, 1906 ; F. A. Gooch, ib. (3), 48. 216, 1893 ; P. E, Browning and W. A. Drushel, Zeit. anorg. Chem., 54. 151, 1907 ; E. Divers, Journ. Chem. Soc., 61. 196, 1892 ; E. Bonjean, Bull. Soc. Chim. (3), 21, 1899 ; Chem. News, 80. 248, 1899. 5 R. Kayser (Zeit. offent. Chem., 5. 107, 1900) found 2 '4 percent, of potash in one sample, and 1*4 percent, of potash and - 6 per cent, soda in another sample of so-called " chemische reine Flusssaure." Hence the necessity for testing this acid for alkalies. See pages 163 and 446. 6 According to H. Rose (Pogg. Ann., 74. 568, 1848), when alkaline sulphates are mixed with ammonium chloride and exposed to a red heat, the alkaline sulphate is partly, and on repeated application of the process wholly, transformed into the chloride. E. Nicholson (Chem. Neivs, 26. 147, 1872) claims that this reaction is of no use as an analytical process. According to F. C. Phillips (Zeit. aiml. Chem., 13. 149, 1874 ; M. Chikashige, Chem. News, 71. 17, 1895), for complete conversion, the temperature of ignition should be just short of the melting point of sodium and potassium chlorides, and the best way of converting the sulphates to chlorides is to evaporate the solution to dryness in a platinum basin with ammonium chloride. Mix the residue with dry powdered ammonium chloride and calcine to a constant weight in a covered crucible to avoid loss by volatilisation. THE DETERMINATION OF THE ALKALIES. 227 as barium sulphate, and removing the excess of barium chloride as barium carbonate l in the following manner : Acidify the nitrate with hydrochloric acid, 2 and heat the solution to boiling. Add an excess of a hot solution of barium chloride 3 to the boiling solution. Heat the mixture to boiling, and let it stand overnight. Filter ; wash with hot water ; evaporate the nitrate to dryness in a platinum dish ; ignite to drive off the ammonium salts ; and cool. The barium is removed during the next operation. Removal of Magnesia and Lime. The main difficulty in this process is the removal of the magnesia. The ammonium magnesium carbonate is appreciably soluble in ammoniacal ammonium carbonate (Schaffgotsch's solution), and an exact separation is not possible by this reagent. Gooch and Eddy have shown that an alcoholic " Schaffgotsch's " solution is quite effective, since the precipita- tion of the magnesia is then complete. Evaporate the solution to about 50 c.c. ; add 50 c.c. of absolute alcohol, and 50 c.c. of an alcoholic solution of ammonium carbonate Gooch and Eddy's solution. 4 Stir the mixture thoroughly, and let it stand about 20 minutes. Filter, and wash with Gooch and Eddy's reagent. Dissolve the precipitate in dilute hydrochloric acid, add an excess of ammonia, and repeat the treatment. Collect the filtrate in a weighed platinum dish, and evaporate the liquid to dryness. Drive off the ammonium carbonate by gentle heating. Add a drop of hydrochloric acid. Ignite, cool, and weigh the mixed chlorides. 105. The Indirect Determination of Potash and Soda. In some cases it is possible to determine, quite accurately, the potash and soda in the mixed chlorides 5 by the so-called " indirect process." The weight of the mixed chlorides is first determined, and afterwards the total chlorine by, say, Mohr's process (page 79). From the data so obtained, it is possible to deduce the respective amounts of soda and potash by arithmetic. Let KC1 denote the weight of potassium chloride and Nad the weight of sodium chloride in the mixture. Let w> "denote the weight of the mixed chlorides, and u the weight of chlorine in the mixture. Then, obviously, the weight of the mixed chlorides is t0 . (1) One part by weight of sodium chloride contains 0*6047 part of chlorine, and one part by weight of potassium chloride contains 0'4756 part of chlorine. Hence, the total weight of chlorine may be written : 0-6047NaCl + 0'4756KC1 = u . (2) 1 Better results are obtained at a later stage by working with chlorides, although some get the alkalies at the last stage as mixed sulphates instead of mixed chlorides, and determine the potassium as potassium chloroplatinate in the mixed sulphates. The small amount of sulphuric acid liberated during the double decomposition: K 2 S0 4 + H 2 PtCl 6 = K 2 PtCl 6 + H 2 S0 4 , does not appreciably affect the results. 2 There is a slight loss of alkaline chlorides by adsorption. The barium sulphate carries down less potassium salts when the precipitation is made in solutions but feebly acid B. West, Zeit. anal. Chem., 2O. 357, 1881. J. F. de Vries (Chein. PTeekblad, 5. 261, 1908) found 26'6 instead of 27 '0 per cent. K 2 0. The loss was due to the potassium salt retained by the barium sulphate. 3 Commercial "guaranteed pure" barium chloride not infrequently contains potassium salts. Hence attention must be paid to this matter, otherwise high results may be obtained. 4 GOOCH AND EDDY'S SOLUTION. Mix 180 c.c. of ammonia, 800 c.c. of water, 900 c.c. as absolute alcohol, and saturate the mixture with ammonium carbonate. Filter off any insoluble salt. 5 The method for the mixed sulphates is quite similar in principle. 228 A TREATISE ON CHEMICAL ANALYSIS. These two relations suffice for the computation. Solve the first equation for NaCl, and substitute the result in the second. We thus obtain, on reduction, KC1 = 4-609^ - 7-622^ ; and NaCl = w - KC1 by substituting the value of KC1 so obtained, in the first equation. The con- version of the amounts of the two chlorides so obtained into the equivalent oxides is effected in the usual manner, namely: Amount of KC1 x O6317 = Amount of K 2 ; Amount of NaCl x O5303 = Amount of Na 2 0. This method does not give good results with the small amounts of alkalies found in clays, unless very special care be taken to ensure the absence of magnesium and other metallic chlorides. Collier 1 obtained quite satisfactory results with mixtures of pure salts : for instance : Table XL. Test Analyses Indirect Separation of Potassium and Sodium Chlorides. Mixed together. Found. Chlorine. KC1. NaCl. KC1. NaCl. Found. Calculated. 0-0582 0-0573 0-0009 0-0278 0-0277 0-1284 0-0067 0-1284 0-0067 0-0651 0-0651 0-0967 0-0102 0-0967 0-0102 0-0522 0-0522 0-0782 0-0317 0-0783 0-0316 0-0564 0-0564 0-0305 0-0379 0-0304 0-0380 0-0375 0-0375 00101 0-1029 0-0104 0-1026 0-0672 0-0672 0-0065 0-1100 0-0058 0-1107 0-0699 0-0698 0-0590 ... 0-0605 0-0360 0-0358 According to Virgili, the relative errors involved in the indirect processes for sodium and potassium salts are those indicated in Table XLL, on the assumption that pure salts are being treated. Table XL I. Relative Errors in Indirect Processes for the Determination of Soda and Potash. Determination. Mixture contains Relative error. Sodium salt. Potassium salt. Total chlorine in the mixed ( chlorides \ Conversion chlorides into/ sulphates \ Total S0 4 in the mixed ( sulphates \ NaCl : KC1 NaCl : 10KC1 NaCl : KC1 NaCl : 10KC1 Na 2 S0 4 : K 2 S0 4 Na 2 S0 4 : 10K 2 S0 4 NaCl 0-018 NaCl 0-102 NaCl +0-021 NaCl 0-118 Na 2 S0 4 +0-021 Na 2 S0 4 +0-118 KC1 +0-014 KC1 0-008 KC1 +0-017 KC1 0-010 K 2 S0 4 +0-017 K 2 S0 4 0-010 1 P. Collier, Amer. J. Science (2), 37. 344, 1865 ; Cheni. News, 10. 182, 1864 ; F. M. Lyte, ib., 29. 159, 1874 ; E. Fleischer, ib., 19. 265, 300, 1869 ; F. Mohr, Lehrbuchder Titricrmethode, Braunschweig, 364, 1859; Zeit. anal. Ohem., 7. 173, 1868; M. Kretschy, ib., 15. 37, 1876; H. Schiff, Liebig's Ann., 105. 219, 1858 ; T. Thomson, ib., 20. 205, 1836 ; K. List, ib., 81. 117, 1852 ; S. Panpushka, Journ. Phys. Chem. Ges. St. Petersburg, 19. 106, 1888 ; Zeit. anal. Chem., 27. 160, 1888 ; L. W. Winkler, Chem. Ztg., 24. 816, 1900 ; G. Errera, Gazz. Chim. Ital., 18. 244, 1889 ; E. K. Landris, Journ. Amer. Chem. Soc., 17. 466, 1895 ; 18. 132, 1896 ; G. Werther, Journ. prakt. Chem. (1), 91. 324, 1864 ; J. J. Berzelius, De V Analyse des Corps Inorganiques, Paris, 69, 1827; J. P. Wuite, Chem. Weekblad, 4. 19, 1907. A. J. Sofianopoulos (Bull. Soc. Chim. (4), 5. 632, 1909) converts the mixed chlorides into fluorides. The process has not been thoroughly tested in general work. THE DETERMINATION OF THE ALKALIES. 229 To illustrate the effect of impurities on the determination, suppose a mixture of potassium chloride, O200 grm., and sodium chloride, 0*020 grm., has O002 grm. of magnesium chloride as impurity, the calculated potassium and sodium chlorides (supposing that there are no analytical errors) will be respectively 0*190 grm. potassium chloride and 0*027 grm. sodium chloride, in place of 0*200 grm. and 0020 grm. respectively. Suppose further that a similar mixture of potassium and sodium chlorides has 0*002 grm. of an inert impurity, say, 0*002 grm. of magnesia. Then, under the above conditions, 0*206 grm. of potassium chloride and 0*016 grm. of sodium chloride would be obtained, in place of 0*200 grm. and 0*020 grm. respectively. This all means that the indirect process ivill give exact results ivith large or small amounts of the mixed chlorides when no impurity is present. Consequently, when there are no means of establishing the purity of the mixture, and when the alkalies appear at the end of a long series of separations, the method will not be very reliable, since the method of calculation multiplies a small trace of impurity into a relatively large error. Some consider that the errors by the indirect process are less than the experimental errors by the direct process (e.g., List, Rose, etc.). This is probably the truth about the method with the limitations just stated. 1 106. A General Study of Indirect Separations. This seems rather an attractive and speedy method of resolving the bases in a mixture, and the student might well inquire why the principle is not more extensively applied for determining, say, potash and soda ; cobalt and nickel * lime and magnesia; lime and strontia * etc. It is therefore necessary to study the principle in more detail. The problem may be stated in general terms : Given a mixture of two salts ivith different bases and one acid, or with different acids and one base, to find the amount of each constituent. Let the salts have the general formulae MA and NA ; 2 and further, let the total weight of the mixture be w, and the weight of the common constituent u. It is required to find the weights of M and N in the mixture. For convenience, let x denote the weight of M, and y the weight of N. Let a denote the known weight of A in the salt MA, and b the weight of A in the salt NA. Hence, it follows : (x + a) -f (y + b) = w ; and, a + b = u. Further, it follows from the laws of chemical combination that a : x = A : M, and b\y = A\N, where A, M, and N now denote the equivalent weights of the respective constituents represented by the letters. Then, a__A L _b__A_ ~x~M' 7~T' These four equations suffice for the algebraic solution of the four unknowns x, y, a, and b from the given data. Solving these equations in the usual manner, we get A A /o\ 1 R. Fresenius, Anleitung zur quantitativen chemischen Analyse, Braunschweig, 441, 1863; H. Rose, Handbuch der analytischen Chemie, Braunschweig, 2. 17, 1871. R. Bunsen (Zeit. anal. Chem., ip. 400, 1871) considers the indirect process a useful means of controlling the results by the direct process. J. Dennant (Rep. Australasian Assoc., 2. 385, 1890) objects. 2 If A be bivalent, and M or N univalent, say, M 2 A and N 2 A, then we write M' = 2M in place of M ; and conversely, if A be univalent, and N or M bivalent so that the salts are MA 2 and N A 2 , we write A' = 2A in place of A. In the first case, M', N', and A are chemically equivalent, and in the second case M, N, and A' are chemically equivalent. 230 A TREATISE ON CHEMICAL ANALYSIS. which corresponds with equation (1) above. From the first pair of equations, we get x + y = w - u . . . . . . (4) Solving this equation for y, and substituting the result in the first equation, we get M u(N+A)-wA N u(M+A)-wA /KX X =A' ' N-M -> and y= ^' M-N The last equation is not needed in the actual arithmetic, because it is easier to calculate y from the value of x 9 w, and u in equation (4). These equations (5) generalise the method employed on page 227 in solving a particular problem. Conditions for Success. Let us interpret the equations further. It is easy to see that if u(M + A) - wA = 0, x must be zero, and u w N+A This means that if, in the analysis of a mixture of, say, the alkaline chlorides, we get a relation such that u_ 35-5 ^35-5 w "23 + 35-5" 58-5' no potassium chloride is present. If ./Vbe less than J/in magnitude, we have u(N+A) greater than wA, since, if wA were the greater, x would be negative, and negative values of x are impossible. Consequently, we have u w "N+A Every addition of KC1 to NaCl makes the ratio u : w greater than A : N+A ; or greater than 35-5 : 58-5. When M = N, as is nearly the case with a mixture of cobalt arid nickel salts, the indirect analysis is impossible because M-N will then be zero, and equations (5) cannot be solved. By rearranging the first of equations (5), and expressing the percentage amount of x in a mixture weighing w grams, we have (u(N+A)-wA\ I per cent, of M . A(N-M)\ w We see at a glance that the factor on the left is constant for a particular pair of salts, and that the greater the value of this factor, the greater the influence of errors in the determination of u and w on the final result. With a mixture of potassium and sodium chlorides, the factor is nearly 7 ; with magnesium and calcium sulphates, 9; and with nickel and cobalt sulphates, 555. It is thus easy to see, other things being equal, that a mixture of magnesium and calcium sulphates can be estimated nearly, but not quite, as accurately as a mixture of potassium and sodium chlorides, while the indirect separation of nickel and cobalt sulphates is hopeless, because the errors in the determination are multiplied enormously. Hence, other things being equal, the calculated values of x and y will be the more accurate (1) The smaller the numerical values of the equivalent weights of the constituents M and N ; (2) The greater the difference between the equivalent weights of the con- stituents M and N '; THE DETERMINATION OF THE ALKALIES. 231 (3) The greater the equivalent weight of the single constituent A and (4) The nearer the ratio of the weights u : w to the ratio of their equivalent weights M:N. 1 107. The Properties of Sodium and Potassium Chloroplatinates. In place of the preceding process it is usual to determine the potassium chloride separately, and subtract this from the weight of the mixed chlorides. The remaining weight of sodium chloride is multiplied by 0'5303 to get the equivalent amount of sodium oxide. The process of separation depends upon lOOr ~\ v 90 - **. s v "^ \ ftp * \ , 1 \ \ \ S V 70 - v Qi_ ^s ^ U 60 - ^ 11 r ' S p f i '/ jty-- SO - 1 [ ( | r 6 1C FIG. 114. Potassium chloroplatinate (Archibald, Watson, and Buckley). FIG. 115. Sodium chloroplatinate (Peligot). the facts: (1) that a mixture of sodium and potassium chlorides in contact with a solution of hydrochloroplatinic acid H 2 PtCl 6 forms both sodium and potassium chloroplatinates ; 2 and (2) there is a marked difference in the solu- bility of the two salts in alcoholic solution. Solubilities of Sodium and Potassium Chlot^oplatinates in Alcoholic Solutions. Both sodium and potassium chloroplatinates are appreciably soluble in water. The solubilities of sodium and potassium chloroplatinates in different proportions of ethyl and methyl alcohol and water are indicated by the graphs, figs. 114-115. 3 These results show that the greater the concentration of the alcohol, the less the solubility of the chloroplatinate, and that the potassium salt is less soluble in ethyl than in methyl alcohol. Different investigators have different ideas as to the best strength of the alcoholic solution, and we find numbers ranging from 70 to 100 per cent alcohol 4 recommended for the washing, etc. If we were guided only by the 1 When the errors in the determination of u and w are the same. For a more exact study of the relations between the calculated values of x and y and the experimental errors in the determination of u and w, see J. W. Mellor, Higher Mathematics, London, 539, 1909; and for a discussion on the indirect methods of analysis, see J. Pages y Virgili, Die indirekten Methoden der analytischen Chemie, Stuttgart, 1911. 2 Also called platinichlorides, not platinochlorides. 3 M. Peligot, Monit. Scient.' (4), 6. 872, 1892; H. Precht, Chem. Ztg., 20. 209, 1896; R. Bunsen and G. Kirchhoff, Pogg. Ann., 113. 337, 1861; E. H. Archibald, W. G. Watson, and B. G. Buckley, Journ. Amer. Chem. Soc. t 30. 747, 1908. 4 H. Precht (Zeit. anal. Chem., 18. 509, 1879; H. Precht, H. Vogel, and H. Haefcke, Lands. Ver. Stat., 47. 97, 1896) recommends absolute alcohol. R. Finkener (Pogg. Ann., 29. 232 A TREATISE ON CHEMICAL ANALYSIS. solubility, we should recommend the use of absolute alcohol. But the Versuchstationen at Halle and Darmstadt l have shown that alcohol of greater concentration than 96 per cent, gives too high results. Decomposition of Sodium Chloroplatinate by Concentrated Alcoholic Solutions. Morozewicz 2 claims to have traced the high results with concentrated alcohol to the decomposition of part of the sodium chloroplatinate, by alcohol of greater concentration than 90 per cent., into sodium chloride 3 and soluble platinum chloride ; and three or four times more of the hydrochloroplatinic acid is needed to convert all the sodium into sodium chloroplatinate. The presence of the insoluble cubic crystals of sodium chloride intermixed with the potassium chloroplatinate is supposed to explain the high results obtained when the alcohol employed for the washing, etc., is too concentrated. There is no danger of the formation of the sodium chloride with 80 per cent, alcohol, but another danger arises, owing to the solubility of the potassium chloroplatinate in the alcohol. 4 This means that O'OOOl grm. of potassium chloride, or 0'000064 grm. of potassium oxide, must be added to the final result for every 10 c.c. of 80 per cent, alcohol which comes in contact with the precipitate. 5 As a rule, less than 50 c.c. of the alcoholic solution are used in clay analyses, and in consequence the solubility correction may usually be neglected. 6 Effect of the Concentration of the Solution on Precipitated Potassium Chloro- platinate. We now inquire : should the hydrochloroplatinic acid be added to a concentrated or to a dilute solution of the two chlorides 1 Fresenius 7 recommends a concentrated solution. In this case, a fine pulverulent precipitate is formed, which appears under the microscope to be a mixture of stellate groups of 637, 1866) and D. Lindo (Ohem. News, 44. 77, 86, 97, 129, 1881), 98 per cent, alcohol ; the Stassfurter Kaliindustrie, 96 to 100 per cent. ; B. Sjollema (Ch&m. Ztg., 21. 739, 1897), 90 per cent, alcohol ; and A. Atterberg (Zeit. anal. Chem., 36. 314, 1897 ; Chem. Ztg., 22. 523, 538, 1898), 80 per cent, alcohol. H. Fresenius (Chem. Ztg., 34. 1032, 1910) states that 70 per cent, alcohol gives too low results, 85 per cent, too high ; and 80 percent, alcohol gives correct results. 1 Lands. Ver. Slat., 45. 374, 1894 ; 46. 181, 1895. 2 J. Morozewicz, Ber.Acad. Sciences Cracovie, 796, 1906. F. P. Treadwell (Kurzes Lehrbuch der analytischen Chemie, Leipzig, 2. 38, 1911) does not think that the sodium chloroplatinate, is decomposed in the manner stated by Morozewicz. W. A. Davis (Chem. World, i. 219, 1912) agrees with Morozewicz. 8 F. Rottger and H. Precht (Ber., 18. 2076, 1885) show that, at 15, 100 grms. of alcohol dissolve the following amounts of sodium and potassium chlorides : Alcohol 90 92 '5 95 percent. Sodium chloride .... 0'345 0'223 0*146 grm. Potassium chloride . . . . 0'073 0'043 0'028 grm. 4 According to Peligot, the solubility of the potassium chloroplatinate is rather less in methyl alcohol (but see fig. 114), and, in consequence, some recommend the use of methyl alcohol in place of ordinary ethyl alcohol. P. Rohland (Zeit. anorg. Chem., 15. 412, 1897 ; 16. 306, 1898 ; Zeit. anal. Chem., 49. 358, 1910) points out that methyl alcohol has additional advantages. If the precipitation be made in the presence of barium chloride, the barium chloroplatinate is liable to decompose into barium chloride and platinum chloride. The former is fairly soluble in methyl alcohol (78 parts of absolute methyl alcohol, at 15, dissolve 1 part of barium chloride), and practically insoluble in ethyl alcohol even at 80 per cent, concentration. After washing in methyl alcohol Rohland finishes off by washing with ether. The potassium chloroplatinate is less soluble in a mixture of alcohol and ether than in alcohol alone, and R. Finkener (I.e.) recommends a mixture of absolute alcohol and ether in the proportions 2:1, while B. C. Corenwinder and G. Contarnine (Compt. Rend., 89. 907, 1879) recommend a mixture of 95 per cent, alcohol with ether in the proportions 9:1; and H. N. Warren (Chem. News, 75. 256, 1897) recommends a mixture of amyl alcohol and ether. 5 The volume is easily measured if the filtration be conducted in a Gooch's crucible in a Witt's filtration jar (page 99), containing a measuring cylinder. The measurement, however, need be only approximate. 6 H. Fresenius and P. H. M. P. Brinton, Zeit. anal. Chem., 50. 21, 1911. 7 R. Fresenius, Zeit. anal. Chem., 16. 63, 1877 ; 21. 234, 1882. THE DETERMINATION OF THE ALKALIES. 233 crystals with a few octahedral crystals. If the hyc>rochloroplatinic acid be added to a dilute solution, and subsequently evaporated down, the crystals are mainly octahedral. In both cases the crystals belong to the cubic system. Those formed in concentrated solution appear to be octahedra distorted into rod-like crystals. The crystals formed in concentrated solution also contain a relatively large number of cavities, which enclose minute globules of the mother liquid, while the octahedral crystals contain, as a rule, comparatively few inclusions. These facts are rather important. The crystals with the liquid inclusions must give high results, since the resulting precipitate contains potassium chloroplatinate plus mother liquid. The crystals can be more or less perfectly dried, but the result will then still be high, presumably owing to the residue left on evaporation of the liquid to dryness. Difficulties are also encountered in drying the potassium chloroplatinate formed in concentrated solutions. This is exemplified by the following experiment, due to Winton. 1 In one case, a mixture of 1'018 grms. of potassium chloride and 0*541 grm. of sodium chloride was made up to 100 c.c. ; and in another case, a mixture of 13 063 grms. of potassium chloride and 0*541 grm. of sodium chloride was also made up to 100 c.c. Corresponding volumes were treated with hydrochloro- platinic acid. The resulting precipitates were treated in the same way, and their weights, after drying for the periods stated in Table XLII., and at the stated temperatures, were determined. The results are shown in Table XLII. The Drying of Potassium Chloroplatinate. Obviously, a more protracted drying at a relatively high temperature is needed to dehydrate crystals formed in a concentrated solution, than when the crystals are formed in a dilute solution. This result also explains how different investigators have made different recommendations at this stage of the analysis. Fresenius, for example, says that 30 hours at 130 are needed, while Eggertz and Nilson say that from 10 minutes to 4 hours suffice. 2 Table XLII. Effect of drying Potassium Chloroplatinate formed in Dilute and in Concentrated Solutions. Temperature of oven. C. Hours drying. Precipitated in dilute solution. Precipitated in concentrated solution. 100 2 99-57 100-22 100 8 99-57 100*03 100 14 99-57 99-95 100 20 99*56 99-93 130 26 99-56 99-80 130 38 99-55 99-70 130 66 99-55 99-67 160 84 99-52 99-56 160 102 99-51 99-50 The "Atomic Weight" of the Platinum in Potassium Chloroplatinate. The older analysts used the number 197 for the atomic weight of platinum, whereas recent work shows that the atomic weight of this element is nearer 195. The 1 A. L. Winton, Journ. Amer. Chem. Soc., 17. 453, 1895 ; Ruer, Chem. Ztg., 20. 270, 1896 ; F. T. B. Dupre, ib., 20. 305, 1896. 2 R. Fresenius, I.e.; C. G. Eggertz and L. F. Nilson, Konigl. Land. Akad. Hand. Tids., 35. 326, 1898 ; G. Krause, Archiv Pharm. (3), 2. 407, 1874. 234 A TREATISE ON CHEMICAL ANALYSIS. former number gives the factor 0'3056 for converting the weight of the potassium chloroplatinate into potassium chloride, while the later number gives the factor 0'30688. The "wrong" number apparently gives a result nearer the truth, unless certain modifications be made in the conduct of the analysis (page 250). The Transformation of Potassium Chloroplatinate into Metal before Weighing. Owing to possible errors arising from the contamination x of the potassium chloroplatinate when precipitated in contact with barium chloride, as is necessary in some special cases, some reduce 2 the chloroplatinate to metal, wash out the soluble salts, and multiply the weight of the metal by the necessary factor to get the equivalent potassium chloride or oxide. The chlorine may also be determined in the soluble salts by titration according to Volhard's or Mohr's processes (pages 76 and 79), and the result calculated to potassium chloride or oxide. 3 These methods are not usually employed in silicate analyses. 108. The Separation of Potassium as Potassium Chloroplatinate. Precipitation of Potassium Chloroplatinate. Treat a dilute solution of the mixed chlorides in a small porcelain basin with 0'3 c.c. more than the amount of hydrochloroplatinic acid calculated 4 on the assumption that the mixed chlorides 1 A. Atterberg, Chem. Ztg., 21. 261, 1897 ; Zeit. anal. Chem., 36. 214, 1897 ; L. Tietjens and Apel, ib., 36. 315, 1897 ; A. H. Allen, B.A. Rep., 24, 1876 ; Chem. Neivs, 35. 259, 268, 1877 ; 36. 17, 38, 47, 1877. 2 With hydrogen gas (R. Finkener, Pogg. Ann., 129. 637, 1866); zinc dust (J. Diamant, Chem. Ztg., 22. 99, 1898); mercury (E. Sonstadt, Journ. Chem. Soc., 67. 984, 1895); mag- nesium ribbon (L. L. de Koninck, Zeit. anal. Chem., 12. 137, 1873 ; 21. 406, 1882 ; C. Favre, Compt. Rend., 122. 1331, 1896 ; A. Villiers and F. Borg, Bull. Soc. Chim. (3), 9. 602, 1893 ; A. Fiechter, Zeit. anal. Chem., 50. 629, 1911 ; R. Trnka, ib., 51. 103, 1912 ; A. Atterberg, ib., 51. 483, 1912); mercurous chloride (A. Mercier, Bull. Assoc. Belg. Chem., 10. 403, 1897); formic acid (H. N. Warren, Chem. News, 75. 256, 1897) ; sodium formate (R. Bottger, Zeit. anal. Chem., 13. 176, 1874 ; F. Jean and J. A. Trillat, Bull. Soc. Chim. (3), 7. 228, 1892 ; B. C. Corenwinder and G. Contamine, Compt. Rend., 89. 907, 1879 ; Woussen, Ann. Agronom., 13. 431, 1888) ; calcium formate (L. L. de Koninck, Chem. Ztg., 19. 901, 1895) ; sodium oxalate (F. Mohr, Zeit. anal. Chem., 12. 137, 1873 ; 21. 216, 1883) ; thioacetic acid (A. Atterberg, Zeit. anal. Chem., 36. 314, 1897 ; Chem. Ztg., 22. 522, 538, 1898). 3 J. Diamant, Chem. Ztg., 22. 99, 1898; A. Atterberg, ib., 22. 522, 538, 1898; L. L. de Koninck, Zeit. anal. Chem., 35. 72, 1896 ; F. Mohr, ib., 12. 137, 1873. 4 Sufficient hydrochloroplatinic acid should be added to convert both the sodium and the potassium chlorides into the corresponding chloroplatinates. The solution of hydrochloro- platinic acid should contain O'l grm. of platinum per c.c. Assume that the mixed chlorides are all sodium chloride, multiply their weight by 17, and add 0'3 to the product. The result represents the number of cubic centimetres of the hydrochloroplatinic acid solution to be used. In the above example (page 225), '037 x 17 = '629 c.c.; and '629 + '3 = 1 c.c. nearly. If sufficient hydrochloroplatinic acid has been added, a drop of the solution, under the microscope, will show golden yellow octahedral crystals of potassium chloroplatinate, and orange-coloured needles of sodium chloroplatinate. If colourless cubes of sodium or potassium chloride be present, insufficient hydrochloroplatinic acid has been added. The hydrochloroplatinic acid should Be free from potassium and ammonium chloroplatinates. Potassium chloroplatinate is appreciably soluble in ammonium chloride (0*0015 grm. of potassium chloroplatinate dissolved in 10 c.c. of ammonium chloride solution). The import- ance of using pure hydrochloroplatinic acid has been emphasised by H. Vogel and H. Haefcke (Lands. Ver. Stat., 47. 97, 1896), H. Precht (Zeit. anal. Chem., 18. 509, 1879), A. F. Holleman (Ckem. Ztg., 16. 35, 1892), and C. R. Fresenius (Zeit. anal. Chem., 33. 358, 1892). S. Zuck- schwerdt and B. West (Zeit. anal. Chem., 20. 185, 1881) state that potassium chloroplatinate is soluble in hydrochloroplatinic acid, and consequently, if too great an excess of this acid be present, low results will be obtained. An excess, however, must be added to ensure complete precipitation. The comparatively small excess usually added cannot have an appreciable influence, even accepting Zuckschwerdt and West's data (100 c.c. of a solution of hydro- chloroplatinic acid containing 7 grms. of platinum per 100 c.c. dissolve 0'3250 grm. of potassium chloroplatinate in 30 hours). W. Dittmar and J. M c Arthur (Trans. Roy. Soc. THE DETERMINATION OF THE ALKALIES. 235 are all sodium chloride. Evaporate the solution 1 to a syrupy consistency 2 a water bath. 3 Cool. The mass should form a solid cake. on FIG. 116. Air bath and thermostat. Filtration and Washing. Treat the residue with a few cubic centimetres of an 80 per cent, solution of alcohol. 4 Stir with a glass rod. Decant the liquid Edin., 33. ii. 561, 1887 ; Journ. Soc. Chem. Ind., 6. 799, 1887 ; R. R. Tatlock, Chem. News, 30. 71, 1874; 43. 273, 1881) state that "the precipitate of potassium chloroplatinate has a remarkable tendency to carry down platinum chiefly as hydroxide, if produced in the absence of a large excess of hydrochloroplatinic acid." There appears to be a kind of hydrolysis in the dilute solution. For objections to sulphates with the " platinum chloride," see A. F. Holleman, Chem. Ztg., 16. 35, 1892. 1 Sufficient water should be present to form a clear solution when first heated on the water bath. 2 H. Precht, Zeit. anal. Chem., 18. 514, 1879. P. Rohland (Zeit. anorg. Chem., 15. 412, 1897 ; 16. 306, 1898) does not recommend evaporation to dryness, since the dehydrated sodium chloroplatinate is less soluble in alcohol than the crystalline salt. G. Ulex (Zeit. anal. Chem., 17. 175, 1878) recommends adding 1 to 5 c.c. of a 20 per cent, solution of glycerol to prevent the sodium chloroplatinate becoming too dry, since otherwise it might not be dissolved. R. Fresenius, Anleitung zur quantitativen chemischen Analyse, Braunschweig, 2. 290, 1905 ; Eng. edit., 2. 220, 1900. 3 The final result is independent of the kind of dish (platinum or porcelain) used in the evaporation ; the presence of a little free hydrochloric or sulphuric acid ; and the temperature of the water bath (A. L. Winton, Journ. Amer. Chem. Soc., 17. 453, 1895). Cases are on record where an insoluble platinous compound is formed by a reaction between the platinum of the dish and the salt. W. F. Hillebrand, Bull. U.S. Qeol. Sur., 422, 173, 1910. 4 A. Mitscherlich (Journ. prakt. Chem. (1), 83. 460, 1861) recommends adding the alcohol after the platinum salt, as indicated in the text. 236 A TREATISE ON CHEMICAL ANALYSIS. through a weighed Gooch's crucible. Treat the mass again with 80 per cent, alcohol; and again decant. Repeat the decantations with 80 per cent.' alcohol until the alcohol running through the Gooch's apparatus appears colourless ; and the precipitate appears golden yellow, not orange yellow. Transfer the precipitate to the Gooch's crucible 1 by the aid of a "policeman." Wash with 80 per cent, alcohol by half filling the crucible about six times. Drying the Precipitate. Drain off the alcohol, 2 and dry the precipitate at 130 to a constant weight in an air bath whose temperature is maintained constant by means of a thermostat, A, fig. 116 ; or in an amyl alcohol bath, fig. 90. ^represents the thermometer. The copper flange CC on the front and sides of the bath is to deflect the products of combustion from the flame up the back of the bath, away from the door. Cool the crucible in a desiccator, and weigh. The increase in weight represents the potassium chloroplatinate. 3 Calculation. Multiply the weight of the potassium chloroplatinate by 0*3068 (or see Table LXXXVII.) to get the corresponding amount of potassium chloride, and by 0'1941 (or see Table LXXXVIII.) to get the corresponding amount of potassium oxide. Subtract the amount of potassium chloride from the weight of the mixed chlorides in order to get the amount of sodium chloride in the given sample, and multiply the result by 0-5303 (or see Table LXXXIX.) to get the corresponding amount of sodium oxide. The following are the weighings obtained with half a gram of clay from which 0'0227 grm. of the mixed chlorides was obtained : Dish and mixed chlorides ..... Empty dish . 26'2867grms. . 26'2640grms. Mixed chlorides ..... > . . 0'0227 grm. Gooch's crucible and potassium chloroplatinate . Gooch's crucible alone ...... . 27'6898 grms. . 27 '6351 grms. Potassium chloroplatinate ...... '0547 grm. Since 0'0547 x 0-3068 = 0-0168 grm. of potassium chloride, Mixed chlorides 0'0227 grm. Potassium chloride . . . . . . . . '01 68 grm. Sodium chloride '0059 grm. Hence, Weight of potassium chloroplatinate x 0*1941 = 0106 grm. K 2 0. Weight of sodium chloride . . x '5303 = '0031 grm. Na 2 0. If the amount of ammonium chloride and calcium chloride used in Smith's process for the mixed chlorides contained a grm. of K 2 0, and b grm. of Na 2 0, a must be subtracted from the potash, and b from the soda. A blank determination showed that the materials used in the above determination con- tained no potassium oxide (a = 0), but did contain (b = ) - 00022 grm. of Na 0. Hence, the amount of sodium oxide is 0'0031 less 0-0002 = 0-0029 grm. per half 1 The results are generally too high with tared filter papers F. H. van Leent, Zeit. anal. Chem., 40. 569, 1901 ; R. Caspari, Zeit. angew. Chem., 6. 68, 1893 ; F. Bolm, Zeit. anal. Chem., 38. 348, 1899 ; C. R. Fresenius, Anleitung zur quantitativen chemischen Analyse, Braunschweig, 2. 290, 1905; Eng. edit, 2. 221, 1900; H. J. F. de Vries, Chem. Weekblad, 4. 231, 455, 1907 ; 5. 176, 1908. 2 C. G. Eggertz and L. F. Nilson (Konigl. Land. Akad. Hand. Tids., 35. 326, 1898) here recommend drenching the precipitate with ether. 3 According to H. Fresenius and P. H. M. P. Brinton (Zeit. anal. Chem., 50. 21, 1911), if the weighed potassium chloroplatinate be caked into lumps, and not pulverulent, it should be redissolved in boiling water, and evaporated in a platinum dish on a water bath in which the water is not quite boiling, again dried at 130, and weighed. The reasons Avill appear from page 233. THE DETERMINATION OF THE ALKALIES. 237 gram of the clay. The clay in question thus contained the equivalent of 2'12 per cent, of K 2 0, and 0'58 per cent. Na 2 0. Errors. Some idea of the magnitude of the accidental errors in the deter- mination of the potash by this and Smith's process can be obtained from the following eight independent determinations made on one sample : 0-0106; 0-0103; 00108; O'OlOl ; 0'0109; 0'0108; 0-0104; O'OIOS. The mean is 0*0105 grin, per 0'5 grm. of clay, that is, 2'10 per cent. The deviations range approximately between 04 grm. per 100 grms. of clay. It is necessary to mention that if the clay or silicate contained sulphur, and sulphur be not removed as indicated on page 226, some sodium sulphate is almost certain to contaminate the potassium chloroplatinate. In fact, the sulphates must be con- verted into chlorides by treatment with barium chloride before applying the "platinum" process. Care must be taken that there is no excess of either sulphate or barium chloride, or the results will be high. 109. The Separation of Potassium as Potassium Perchlorate. Potassium perchlorate is practically insoluble in concentrated alcoholic solutions, while sodium, barium, and magnesium perchlorates are soluble under the same conditions. In 1831, Serullas 1 proposed a method for the determina- tion of potassium based on this property, but owing to the difficulty in getting pure perchloric acid at that time, and to some mistaken ideas on the properties of perchloric acid, Serullas' proposal did not get the attention it deserved. Quite satisfactory perchloric acid can now be easily obtained in commerce, and some of the principal sources of error have been investigated, 2 and methods devised by which results rivalling the hydrochloroplatinic acid process can be obtained, and this in less time and at a less cost hydrochloroplatinic is un- comfortably expensive for commercial work. The greater simplicity of the process also renders it less liable to error when once the manipulation is mastered. The details of the method to be used in the separation are as follows : The Process. Dissolve the mixed chlorides 3 in from 10 to 15 c.c. of hot water, and then add two or three times as much perchloric acid as is theoretically required to precipitate the mixed perchlorates. Evaporate the mixture on a 1 M. Serullas, Ann. Chim. Phys. (2), 46. 294, 1831. 2 T. Schloesing, Compt. Rend., 73. 1269, 1871 ; K. Kraut, Zeit. anal. Chem., 14. 152, 1875 ; A. Bertrand, Monit. Scient. (3), 2. 961, 1881 ; L. Grandeau, Tmite & analyse des matieres agricoles, Paris, I. 419, 1897 ; D. A. Kreider, Chem. News, 73. 8, 17, 1896 : 72. 241, 251, 261, 1895 ; Zeit. anal. Chem., 9. 343, 1870 ; D. A. Kreider and J. E. Breckenbridge, ib., 13. 161, 1897 ; Chem. News, 74. 227, 1896 ; F. S. Shiver, ib., 79. 269, 281, 1899 ; Journ. Amer. Chem, Soc., 21. 33, 1899 ; R. Caspari, Zeit. angeiv. Chem., 6. 68, 1893 ; Zeit. anal. Chem., 36. 709, 1897 ; V. Schenke, Landw. Ver. Stat., 47. 36, 1895 ; C. Aumann, ib., 60. 231, 1901 ; A. Strigel and J. Dodt, ib., 78. 179, 1912 ; V. Schenke and P. Kriiger, ib., 67. 145, 1907 ; W. AVense, Zeit. angeiv. Chem., 4. 691, 1891; 5. 233, 1892; H. Kolbe, Journ. prakt. Chem. (2), 5. 93, 1872 ; F. Hamel, Chem. News, 26. 27, 1892 ; H. Precht, International Cong. App. ' Chem., 7. i, 146, 1909 ; G. SUIT, Min. Eng. World, 36. 605, 1912 ; W. A. Davis, Chem. World, I. 219, 1912 ; A. AVityn, Buss. Journ. Exp. Agric., 2. 192, 1912. 3 The amount is calculated as follows : Assume that the mixed chlorides are all sodium chloride, and that the perchloric acid used is 30 per cent. (sp. gr. 1'20) solution. Multiply the weight of the mixed chlorides by 6, and the product represents the number of cubic centimetres of acid required for the work. Perchloric acid of approximately 60 per cent, strength has a specific gravity of T54 ; 30 per cent, 1'20 ; and 20 per cent., 1'12. There is no difficulty about keeping the acid. It is not decomposed by hydrochloric or by sulphurous acid. The acid slowly volatilises at 138 without decomposition. It is not affected by exposure to light. The strength of the acid is easily determined by titration of a known amount with standard sodium hydroxide with phenol phthalein as indicator. 238 A TREATISE ON CHEMICAL ANALYSIS. water bath to a syrupy liquid until the fumes of perchloric begin to appear ; cool a little. Take up the mass with hot water, and add 5 to 6 c.c. of perchloric acid. Re-evaporate until the fumes of perchloric acid again begin to appear. 1 The object of this treatment is to remove the hydrochloric acid. This is important. Stir the cold mass -with about 20 c.c. of 96 or 97 per cent, alcohol containing 0"2 per cent, by weight of perchloric acid. Keep the potassium perchlorate as coarsely granular as possible. Let settle. 2 Decant through a dried and weighed Gooch's crucible. Wash the residue by decantation through the Gooch's crucible three times. About 20 c.c. of the alcohol will be needed for the washing. Transfer the precipitate to the Gooch's crucible by means of alcohol. 3 Some prefer to wash the residue at this stage with 20 c.c. of a mixture of equal parts of 97 per cent, alcohol and ether. Dry the precipitate at 120 to 130 for about half an hour, 4 and then weigh as KC10 4 . Calculations. Assume that 0*5 grm. of clay has been treated by Smith's process, and that 0-0243 grm. of the mixed chlorides have been obtained. Crucible and perchlorate 67223grms. Empty crucible 6 '6951 grms. Potassium perchlorate 0'0272 grm. Multiply the weight of the potassium perchlorate by 0*5381 in order to find the equivalent amount of potassium chloride or use Table LXXXVIL Hence, 0-0272 X 0-5381 =0'0146 grm. of potassium chloride. Multiply the weight of the potassium perchlorate by 0*33992 in order to find the equivalent amount of potassium oxide K 2 or use Table LXXXVI. Hence, 0-0272 x 0-34 = 0-0092, or 1-84 percent, of K 2 0. Subtract the weight of the potassium chloride from the weight of the mixed chlorides, and the result represents the weight of the sodium chloride. Mixed chlorides '0243 grm. Potassium chloride '0146 grm. Sodium chloride '0097 grm. Multiply the weight of the sodium chloride by 0-5303 in order to find the equivalent amount of sodium oxide Na 2 or use Table LXXXIX. Thus, 0-0097 x 0-530 = 0-0052 ; or 1'04 per cent Na 2 0. The Results. The perchlorate process does its work fairly well in the presence of sulphates. Consequently it is not always necessary to remove the sulphates by means of barium chloride (page 226). Thus, with a mixture of potassium chloride and sulphate containing the equivalent of 0*0307 K 2 0, Davis found K 2 . . . 0-0305; 0'0312; 0'0305 ; 0-0306; 0-0307; and with 0-1 grm. K 2 S0 4 containing 0'0541 grm. K 2 0, Davis obtained 0-0536 and 0-0536 grm. K 2 0. Quite correct results were also obtained with potassium chloride mixed with twice its own weight of sodium chloride, sodium phosphate, 1 Some prefer to evaporate to dryness at this stage. '* The addition of a drop of ether aids filtration E. Murmann, Oester. Chem. Ztg., 13. 227, 1910. 3 A few drops of the filtrate, evaporated to dryness on platinum foil, should show no residue. 4 Potassium perchlorate does not decompose below 400. THE DETERMINATION OF THE ALKALIES. 239 calcium chloride, or barium chloride. If phosphoric acid be present, a larger excess of perchloric acid is needed for the precipitation. Small amounts of magnesia do no particular harm, provided it is not present as sulphate when the precipitation is made ; large amounts of magnesium salts give high results, and they must accordingly be first removed. Ammonium salts, if present, should be removed by boiling with sodium hydroxide, owing to the sparing solubility of ammonium perchlorate. 1 10. The Determination of Sodium Oxide. It will be observed that the soda has been determined by difference, and, in consequence, the resulting error is the joint effect of the error in the determina- tion of the mixed chlorides and in the determination of the potash. The two errors may exactly or partially neutralise one another, or they may act both in one direction. Hence, a comparison of the actual values obtained for the soda Na 2 in eight determinations of one sample of clay, indicated below, is particularly interesting : 0-0029; 0-0030; 0-0034; 0*0035; 0*0030; 0-0029 ; 0-0031; 0-0030; with a mean value of 0'0031 per half gram, and an error ranging between 0*0003. The theory of errors furnishes a value for the error 0-00035 which is very near the value actually obtained when the soda Na 2 Q is determined by difference. The accidental errors per 100 grms. of clay do not differ much from 0'03 grm. Hence, this method of finding the amount of sodium oxide in a clay is quite reliable, and there is no need to determine the soda directly. Indeed, the direct processes at present available are somewhat risky and offer no advantage over the difference method. The soda can be determined directly in the filtrate from the potassium chloroplatinate by evaporating the filtrate to a small volume, and reducing the platinum to metal by means of formic acid. Filter off the platinum ; evaporate to dryness in a weighed dish ; and finally weigh as sodium chloride, 1 or as sodium sulphate. 2 Determination as Sodium Antimoniate. Beilstein and Blaese 3 recommend the following direct process for sodium : Add a solution of potassium antimoniate to a solution of the mixed chlorides or nitrates, not carbonates, and let the mixture stand 24 hours in the cold. Decant the clear liquid through a Gooch's crucible, and wash the precipitate with a solution of potassium acetate (7 grms. per litre), and finally with 50 per cent, alcohol. Calcine the precipitate, and weigh as NaSb0 3 . In order to correct the result for solubility of the sodium antimoniate, add 0'0233 grm. to the weight of the precipitate for every 100 c.c. 1 Some of the lithium in the silicate will be found with the sodium chloroplatinate in the nitrate (page 236) ; most will be precipitated with the carbonate in removing lime and magnesia by Gooch and Eddy's process, but not by Smith's process. Most of the caesium and rubidium, if present, will be found with the potassium, since both csesium and rubidium chloroplatinates are even less soluble than the corresponding potassium salt W. Crookes, Chem. News, g. 37, 205, 1864 ; R. Bunsen and G. Kirchhoff, Pogg. Ann., 113. 337, 1861 ; Zeit. anal. Chem., I. 62, 1862 ; R. Bunsen, ib., 2. 161, 1863. For the separation of the -three chloroplatinates, see O. D. Allen, Amer. J. Science (2), 34. 367, 1862; Zeit. anal. Chem., 2. 68, 1863; Chem. News, 6.. 265, 1862. F. Stolba's process (Zeit. anal Chem., 12. 440, 1873) with stannous chloride is not good ; while F. Godeffroy's process (Her., 7. 375, 1876) with antimony chloride is very fair (A. Cossa, Acad. dei Lincei (3), 2.). H. C. Wells, Amer. J. Science (3), 43. 17, 1892 ; (3), 46. 186, 1893 ; H. H. Johnson and 0. D. Allen, ib. (2), 35. 94, 1863. 2 A. Mitscherlich, Zeit. anal. Chem., I. 59, 1862 ; G. Werther, Journ. praJcf. Chem. (1), 91. 321, 1864. 3 F. Beilstein and 0. von Blaese, Bull. Acad. Sciences St. Petersburg, 33. 209, 1895. 240 A TREATISE ON CHEMICAL ANALYSIS. of the mother liquid, not washing fluid, which is in contact with the precipitate. I have had very little experience with the method. I am told by an analyst who has tested the process that it is unsatisfactory for general work. I have had no experience with Ball's process of separation based on the insolubility of sodium bismuthi-nitrate. 1 in. The Preparation of Hydrochloroplatinic Acid from Platinum Residues and Scraps. Although usually called platinum chloride PtCl 4 the reagent employed for separating potassium in the form of potassium chloroplatinate is really hydrochloroplatinic acid H 2 PtCl 6 . In order to keep down expenses, it is very necessary to collect all platinum residues, alcoholic washings, platinum scraps, etc., in bottles provided for the purpose, and then to recover the platinum at convenient intervals. The recovered platinum is usually converted into hydro- chloroplatinic acid. 2 1. Alcoholic Washings. Distil off the alcohol 3 and take up the residue with water; 4 add the potassium chloroplatinate residues to the solution. Pour the solution into a solution of sodium hydroxide 5 mixed with 8 per cent, of glycerol. Heat the solution to boiling, and the platinum separates as a heavy black powder. Wash with water, then with hydrochloric acid, and then with water again. Dry. Ignite the powder in order to destroy organic compounds, and weigh. 2. Platinum Scraps. These may contain iridium, which is itself insoluble in aqua regia, but soluble in this menstruum when alloyed with platinum. Dissolve the platinum in aqua regia (3HC1 -f HN0 3 ) in a capacious flask on a water bath. When solution is complete, evaporate to a syrupy consistency in a large porcelain basin. Take the residue up with water, and gradually add sodium carbonate and sodium formate until the solution is alkaline. 6 Heat to boiling. Both platinum and iridium are precipitated as black powders. Decant off the supernatant liquid. Wash the residue with dilute hydrochloric acid to remove sodium salts, and then with water to remove the acid. Dry the powder, 1 W. C. Ball, Journ. Chem. Soc., 97. 1408, 1910. ' 2 L. Opificus, Zeit. anal. Chem., 23. 207, 1884; H. Precht, ib., 18. 509, 1879 ; G. Krause, ib., I4.J84, 1875 ; W. Dittmar and J. M'Arthur, Trans. Roy. Soc. Edin., 33. ii., 561, 1887 ; H. C. Weber, Journ. Amer. Chem. Soc., 30. 29, 1908 ; W. C. Zeise, Pogg. Ann., 21. 498, 1830 ; 40. 234, 1837; E. Duvillier, Compt. Rend., 84. 444, 1877 ; T. Knosel, Ber., 6. 1159, 1873; H. W. Wiley, Journ. Amer. Chem. Soc., 19. 258, 1897 ; Chem. News, 75. 214, 1897. 3 PURIFICATION OF ALCOHOL. The alcohol may be recovered by re-distilling from quicklime. Put 1| litre of alcohol in a 2-litre Winchester, add 180 grms. freshly burnt quicklime in coarse powder. Agitate every now and again for about 8 days. Distil the alcohol into a large flask. Add about 120 grms. freshly burnt quicklime per litre. The alcohol which is distilled from this will be very nearly absolute J. L. Smith, Chem. News, 30. 234, 1874. The alcohol may then be mixed with powdered potassium permanganate until it is distinctly coloured ; allow to stand some days until the permanganate is decomposed, and manganese oxide is deposited. Add a little quicklime, and distil slowly. When 10 c.c. of the distillate gives no perceptible yellow coloration when boiled with 1 c.c. of a concentrated solution of caustic soda or potash, the subsequent distillate is collected for use. The alcohol so obtained is neutral, and gives no coloration on boiling with silver nitrate or caustic alkalies E. Waller, Journ. Amer. Chem. Soc., II. 124, 1889. 4 On evaporating the solution of hydrochloroplatinic acid to dryness in the presence of alcohol, hydrochloroplatinous acid H 2 PtCl 4 and ethylene are formed. These react to produce ethylene platinous chloride, which dissolves in alcohol, forming an explosive powder when evaporated to dryness. The dry powder is insoluble in acids. 5 21 grms. of sodium hydroxide made up to 100 c.c. with water. Add 8 to 10 c.c. of glycerol. 6 Take care that there is no loss by spurting during effervescence. THE DETERMINATION OF THE ALKALIES. 24! and ignite in a weighed porcelain crucible over a blast lamp, whereby the indium becomes insoluble in aqua regia. Weigh. Conversion of the Metal into Hydrochloroplatinic Acid. Dissolve the grey powder at as low a temperature as possible in hydrochloric acid, and add nitric acid in small quantities at a time. In this way some nitrosoplatinic chloride PtCl 4 (NO) 9 is formed. This must be destroyed, since it leads to low results in the determination of potassium. Evaporate the solution with water, whereby the nitrosoplatinic acid is decomposed into hydrochloroplatinic acid with evolu- tion of nitrogen oxides. Some of the latter still remain in solution. Hence, add more water and hydrochloric acid, and again evaporate. These operations are repeated until no more nitrous fumes are evolved. 1 During these operations, some hydrochloroplatinous acid is formed H 2 PtCl 4 . This is particularly objectionable, since it leads to high results when the chloroplatinic acid is used for the determination of potassium. In order to convert hydrochloroplatinous into hydrochloroplatinic acid, saturate the warm solution with chlorine gas, and evaporate the solution at as low a temperature as possible to a syrupy con- sistency. Dissolve the yellowish-brown mass in water (cold), and if any insoluble indium be present, filter, wash, ignite, and weigh the residue. Subtract the result from the weight of metal taken. The difference gives the amount of platinum which has passed into solution. Dilute the solution until it contains the equivalent of 1 grm. of platinum (metal) per 10 c.c. 1 W. Dittraar and J. M' Arthur (I.e.) doubt if it is possible to destroy all the nitrosoplatinic chloride in this manner. They recommend acting on the metal with hydrochloric acid and chlorine from a "chlorine Kipp." H. Precht (Zeit. anal. Chem., 18. 509, 1879) and W. A. Noyes and H. C. P. Weber (Journ. Amer. Chem. Soc., 30. 13, 1908) have also emphasised the necessity for removing nitric acid from the hydrochloroplatinous acid. J. S. Stas (Chem. News, 73. 5, 1896) removes the chlorine by dissolving the salt in water; raises the temperature of the solution to its boiling point ; saturates the solution with chlorine ; and keeps the temperature of the solution at 100 until it ceases to smell of chlorine. 16 CHAPTER XVIII. ABBREyiATED ANALYSES AND ANALYTICAL ERRORS. 112. Exhaustive u. Works Analyses. IT will now be well to review our results. The hydrochloric acid solution of the sodium carbonate was evaporated twice, with an intervening filtration, and the silica altered off, ignited, weighed, and treated with hydrofluoric acid. The filtrate was precipitated twice with ammonia and ammonium chloride. The precipitate was ignited with the residue from the silica, and weighed as a mixture of aluminium, ferric, titanic, and phosphoric oxides. This was fused with potassium pyrosulphate, the insoluble silica filtered off, and the solution made up to 200 c.c. Of this, 50 c.c. were used for the determination of iron colorimetrically ; and 100 c.c. for the titanic oxide. 1 The filtrate from the alumina was evaporated down to a small bulk, and alumina separated in the usual way. 2 The filtrate was treated with ammonium oxalate. The precipitated calcium oxalate was filtered off; and the filtrate was treated with ammonium sodium phosphate, when a precipitate of magnesium phosphate was obtained. The next filtrate was rejected. The alkalies were determined on a separate sample. Summarised (solids to left, solutions to right) : (Fluorine absent) HC1 sol. of Na 2 C0 3 fusion (two evaporations) SiO 2 ; HF treatment I Add aq. NH 3 and NH 4 C1 Res Ext F | idue SiF 4 volatilises Free J ipitate Evapc >rate KTH 3 sol. oxalate hosphate ect Extra A1 2 O 3 H 2 S in 1 A1 2 Ignite ; weigh ; fus( { , etc. 5 K 2 S 2 7 , digest with water to 200 c.c. 2 1 MnO Amm. ra SiO 2 Filtrate CaO Amm. p I BaOa P 2 5 Ti MgO Re It is important in devising analytical schemes to keep the object of the analysis clearly in view. If extreme accuracy be desired for research and other 1 And, if desired, 100 c.c. for the phosphorus (page 595). 2 The solution can then be treated with ammonia and hydrogen sulphide to precipitate the manganese sulphide (q.v.}. 242 ABBREVIATED ANALYSES AND ANALYTICAL ERRORS. 243 purposes, no precautions must be neglected which will ensure exact results. The analysis may have to be criticised while the analyst is in the witness box in the County Court. In such cases he must be prepared to furnish clear, concise, complete, and conclusive proofs of the accuracy of his statements. The purity of the reagents should have been established by blank or other tests ; and the degree of accuracy of the analytical process should be known. Analyses for reports on new clays, etc., usually call for more exhaustive details than are needed for general practice. If the analysis is to be made for industrial work, accuracy and speed are of prime importance. Such precautions must be adopted as will ensure the required degree of accuracy. Ultra-refined processes waste time. Superfluitas, said Bacon, impedit multum et reddit opus abominabile. A scheme of analysis might serve a given purpose admirably, and yet appear grotesque if applied with another object in view. The determination of the 0'03 per cent, of lithia usually present in Cornish stone, for instance, would be useless for ordinary technical requirements. We should not know how to apply the information if we had it. A certain amount of care is imperative in applying the principle, "near enough for our purpose," because unsuspected sources of faults, etc., might easily be overlooked. The prime object of chemical analysis in industrial practice is to prevent errors of commercial importance. These errors may arise from (1) a variation in the composition of the raw materials; (2) a wrong proportioning of the clays in a body mixture, etc. ; (3) the need for checking the efficiency of processes of purification, grinding, etc., at different stages in the manufacture ; (4) the introduction of deleterious impurities with the raw materials; (5) payment for raw materials invoiced, possibly, higher than their market value ; etc. 113. Abbreviated Schemes of Analysis. In purchasing raw materials, the analysis must frequently be conducted in a much shorter time than is possible by the scheme indicated in the preceding pages. The necessary time is not available, and an exact analyses may be no more useful than a close approximation. The material may have been sold before the analyst has completed his work. The analysis, when completed, is accordingly useless for his own firm a rival has bought the material. The methods of analysis taught in the schools are not those which have developed under the stress of competitive practice. The analytical chemist must therefore exercise his analytical faculties, not only in manipulative skill, but also in distinguishing between necessary precautions and unnecessary exercises in chemical gymnastics. The two faculties are not always located in the same man. The one is mere mechanical dexterity ; the other is the quality which makes a man valuable. Simplified Scheme of Analysis. A simpler scheme suitable for certain analyses in routine work may be employed for many purposes. This will be understood from the following representation (solids to left, solutions to right) : HOI sol. of Na 2 C0 3 fusion (one evap. ) SiO 2 Make sol. to 200 c.c. 100 c.c. Add ammonia Fe 2 O 3 (Reinhardt's process) A1 2 O 3 + Fe 2 O 3 CaO, MgO 244 A TREATISE ON CHEMICAL ANALYSIS. For simply checking the correctness of a " mixing," the determination of the silica may suffice. In some cases a mere determination of the loss on ignition of the dried (110) sample will show whether it is necessary to proceed further with the analysis. The idea is to pick out one or two components which admit of easy determination lead and lime, for instance, in a glaze. If these be quite normal, it is sometimes sound reasoning to infer that the different con- stituents have been properly proportioned. Rapid Clay Analyses. It is possible to determine the silica, alumina, and ferric oxide in a clay without using a platinum crucible, and in a comparatively short time, by the following process : * Mix O5 grm. of the clay with six times its -weight of " peroxide fusion mixture" 2 in a 30-c.c. nickel crucible with a nickel spatula. Fuse the mass for about 5 minutes at a dull red heat. 3 Cool. Place the crucible in an evaporating basin 12 cm. diameter, or in a covered beaker. Add water slowly. The action may be somewhat vigorous, and care must be taken to avoid loss by spurting. The heat generated during the action will lead to a rapid dissolution of the cake. Add an excess of hydrochloric acid. Evaporate the solution to dryness. Grind the residue to powder. Dehydrate the mass at about 110 for an hour. Digest the mass with hydrochloric acid ; filter ; wash and ignite the residue in a porcelain crucible for total silica. Make the nitrate up to 200 c.c. Precipitate the aluminium and ferric hydroxides in 100 c.c. of the solution by adding an excess of ammonia in the usual manner, and weigh as mixed Al 2 3 + Fe 2 3 . Determine the ferric oxide in the other 100 c.c. by Reinhardt's process. All this involves 3 or 4 hours' work, excluding the 2 to 4 hours required for the evaporation of the silica. 4 Hence two such partial analyses can be made in a day. 114. The Indirect Determination of Lime and Magnesia. The lime and magnesia can be determined much more quickly by the indirect process than by precipitation respectively as oxalate and as ammonium magnesium phosphate. Schaffgotsch 5 showed that calcium and magnesium carbonates separate from a solution containing a great excess of ammonium carbonate and ammonia, probably in the form of double ammonium carbonates. A certain amount of magnesium carbonate is, however, hydrolysed under these conditions, and the precipitation of the magnesia is not therefore quite complete. Gooch and Eddy 6 have shown that the separation is completed in a relatively short time if the solution contains an excess of alcohol. 1 See E. P. Fleming, Western Chem. Met., 5. 396, 1909; J. H. Walton, Journ. Amer. Chem. Soc., 29. 481, 1907. See E. Ladd, page 460. 2 PEROXIDE FUSION MIXTURE. Mix 4 parts by weight of the purest sodium peroxide with 1 part of the purest sodium hydroxide. Sodium peroxide usually contains traces of silica. The amount can be determined by a blank experiment and an allowance made for the 3 grins, of peroxide fusion mixture used. 3 For a still more rapid method of decomposition, see page 267. 4 For systems of accelerated evaporation, see H. J. S. Sand, Journ. Soc. Chem. Ind., 26. 1225, 1907 ; W. Hempel, Ber., 21. 900, 1888 ; C. Jones, Journ. Anal. App. Chem., 3. 121, 1889; R. Fessenden, Chem. News, 61. 4, 1890; E. Donath, Chem. Ztg. t 32. 1107, 1908; T. Brugnatelli, Gaz. Chim. Ital., 16. 1878 ; J. W. Gunning, Zeit. anal. Chem., 26. 725, 1887 ; A. Gawalovski, ib., 12. 181, 1873 ; C. Zengelis, ib., 45. 758, 1906. 5 F. C. Schaffgotsch, Pogg. Ann., 104. 482, 1858 ; 106. 294, 1859; A. K. Christomanos, Zeit. anal. Chem., 42. 606, 1903. See page 227. 6 F. A. Gooch and E. A. Eddy, Amer. J. Science (4), 25. 444, 1908; Chem. News, 97. 280, 1908; J. M. Stillman and A. J. Cox, Journ. Amer. Chem. Soc., 25. 732, 1903; E. Dreschel, Journ. prakt. Chem. (2), 16. 169, 1878 ; 0. Foote, Gaz. Chim. Ital., 24. i, 207, 1894. According to 0. Bertrand (Monit. Scient. (3), 10. 477, 1880), at 10, 1 part of calcium carbonate ABBREVIATED ANALYSES AND ANALYTICAL ERRORS. 245 Determination of the Mixed Oxides. After the sodium carbonate fusion the silica and alumina have been removed in the usual manner, and the nitrate from the alumina is boiled down to a small volume, 1 and an equal volume of alcohol and 50 c.c. of Gooch and Eddy's solution (page 227) added. In about half an hour the precipitate is washed with the same solution, and dissolved in a small volume of dilute hydrochloric acid; neutralised with ammonia; and an equal volume of Gooch and Eddy's solution and alcohol added. The precipitate is washed as before, ignited in a weighed platinum crucible, and weighed as a mixture of CaO + MgO. 2 Transformation of the Mixed Oxides to Sulphates. Add, very carefully, suffi- cient dilute sulphuric acid 3 to combine with all the lime and magnesia in the crucible. Evaporate the solution to dryness ; gradually raise the temperature of the crucible to the full heat of a Bunsen's burner 4 for about a quarter of an hour Weigh 5 the contents of the crucible as mixed CaS0 4 + MgS0 4 . Calculations. Suppose that W represents the weight of the mixed oxides, w the weight of the, mixed sulphates ; then, if u represents the weight of the S0 3 present, w- W=u. By the method of page 229, therefore, we have the two equations : MgS0 4 + CaS0 4 = w ; and 0'665 MgS0 4 + 0-588 CaS0 4 = u. Hence, w - CaS0 4 . In illustration, suppose the mixed oxides weighed 0'0265 grm., and the mixed sulphates 0'0703 grm., it follows that the mixture contains 0'0383 grm. of CaS0 4 and 0'0320 grm. of MgS0 4 . But CaS0 4 x 0-412 = CaO ; and MgS0 4 x 0'335 = MgO. Hence, the given mixture has 0*0160 grm. CaO, and 0'0107 grm. MgO. 115. Permitted Errors. If all the constituents in any given silicate have been determined, the numbers, in the ideal case, should add up to 100 per cent. The proximity of the actual sum to the ideal 100 is a valuable check on the accuracy of the work. Absolute identity would not represent perfect work. 6 The errors due to incomplete washing ; dust ; inevitable impurities in the best of reagents ; action of the reagents and solutions on the glass and porcelain, etc., all tend to make the is soluble in 13,980 parts of water containing ammonium chloride ; in 8380 parts of water con- taining ammonium sulphate ; and in 14,438 parts of water containing ammonium nitrate. 1 If alumina separates, it must be of course filtered off. 2 If baryta and strontia be present, they will be included with these bases. To evaluate the mixture by a titration process, see A. Trabert, Gompt. Rend., 119. 1009, 1894 ; Chem. News, 71. 26, 1895. 3 The addition of sulphuric acid is not a very safe operation for an analytical process. It is best to add an ammoniacal solution of ammonium sulphate containing a little ammonium chloride W. L. Scott, Chem. News, i. 144, 1860. 4 If no fumes of sulphuric acid came from the crucible when it was being ignited at the higher temperature, insufficient sulphuric acid was probably added. In that case, more sulphuric acid must be added when the crucible has cooled. A large excess of sulphuric acid should be avoided. 5 The mixture is a little hygroscopic, and must not be needlessly exposed to the air before weighing. 6 F. Jordis, Zeit. anorg. Chem., 45. 362, 1905. If the weighings be not all reduced to " weight in vacuo" it is easy to prove that the sum of the several constituents with ordinary analytical weights, and with perfect work, cannot add up to 100. The proof is easy to see from 10, page 23. The clay is weighed with, say, a 1-grm. brass weight, the other constituents of the clay with platinum weights. The difference is illustrated in Table V. , page 23. 246 A TREATISE ON CHEMICAL ANALYSIS. total greater than 100. On the other hand, mechanical losses through imperfect cleaning of the vessels in transferring liquids and solids from one vessel to another ; accidental spilling of drops; the slight solubility of the precipitates, etc., all tend to reduce the total below 100 per cent. There is but a remote probability perhaps exceeding one in a hundred that the two sets of errors will exactly balance one another, and the ideal 100 be obtained. The coincidence would be mere chance, and when it does occur it may be somewhat embarrassing. It is necessary to decide on limiting deviations above and below 100 for satisfactory work. If these limits be exceeded, the analysis is to be condemned. Hillebrand l places these limits at 99'75 and 100-5 ; Washington 2 prefers 99'5 and 100-75. If the total falls below 99-5, there is strong presumptive evidence that some con- stituent has been either overlooked or ignored. In the analysis of clays for technical purposes no pretence is made to exhaust the possible constituents. Only those constituents of technical importance are determined. Hence, it is advisable to extend the lower limit. In reports on clays and related materials I make 0-5 the permitted limiting deviations from 100. If the total falls below 99-5, it is advisable to find the missing constituent. If the analysis pretends to be exhaustive, the lower limit should be raised. The practice of adjusting the results of an analysis to an exact 100 is utterly bad although it rather appeals to a business man who is ignorant of what is implied, and likes to see the data "properly balanced," as he calls it. Apart altogether from the ethics of the computation, and the temptation to "cook" a defective analysis, the adjustment, as Fresenius 3 has pointed out, "prevents others from judging the accuracy of the results," and, in consequence, makes chemists reason- ably sceptical as to the value of the work. The figures must always be given as they are obtained, and it is just here that the integrity of the analyst meets its first test. 4 There is an impression that a satisfactory summation is a sufficient criterion of accurate work. As a matter of fact, a satisfactory summation is no proof that the separations have been correctly performed. This is well demonstrated by the ratio Si0 2 : A1 2 3 in different analyses of the same clay by different men. 5 Another meretricious system of reporting commercial analyses may be illustrated by quoting the last three lines of an analysis of ball clay from a " clay expert's " report : Alkalies 1'83 Undetermined constituents . . . . .0*18 Total 100-00 The " undetermined constituents " will not deceive a chemist who has grasped the significance of the errors incidental to all methods of analysis, but it may mislead 6 those who have not devoted special attention to the subject. 1 W. F. Hillebrand, Bull. U.S. Qeol. Sur., 305. 26, 1907. 2 H. S. Washington, Manual of the Chemical Analysis of Hocks, New York, 24, 1904. For the limits of accuracy in the analysis of alloys, see E. A. Lewis, Metal Ind., 2. 304, 1910 ; Journ. Soc. Chem. Ind., 31. 96, 1912. See also J. Grossmann, ib., 18. 977, 1896. 3 R. Fresenius, Quantitative Chemical Analysis, London, 2. 101, 1900. 4 Clay analyses have been published where the total number of constituents makes over 104 per cent., and that not a misprint. We cannot but admire the honesty of the analyst a pro- fessional, by the way. The work should have been repeated. Its publication was ill-advised. 5 Of course, something is then wrong. If baryta were overlooked in a glaze analysis, the "total" might be satisfactory, and yet the alumina, tin, etc., might be high, especially if sulphates be present. 6 " 100 per cent, analyses" are frequently "crocks." As a matter of fact, there was a dis- ABBREVIATED ANALYSES AND ANALYTICAL ERRORS. 247 Dittrich l has analysed artificial mixtures containing known quantities of the principal constituents which occur in silicate rocks, and found the limits of error 2 to be as follows : Table XLIII. Accidental and Constant Errors in Silicate Analyses. Constituent. Limits (per cent.). Alumina ... -0-15 and -0-25 Ferric oxide + 0-2 and -0-3 Lime .... -o-i Magnesia ... -o-i Potash (Bunsen's process) + 0-1 Potash (Smith's process) -O'l and -0-2 Soda (Bunsen's process) Soda (Smith's process) . + 0*2 and -0-1 and -0-3 -0-2 Each limit will naturally vary with the skill of the analyst, with the process of analysis, and be dependent upon the number of separations involved in the analysis. In general, the greater the number of separations, the greater the errors of the analysis. In order to find what accidental errors might be expected in clay analyses, eight independent analyses of one homogeneous sample of clay were made in my laboratory by an analyst accustomed to work with the processes recommended in this book. Details have been indicated in the preceding text. The results are here collected in Table XLIV. Table XLIV. Accidental Errors in Eight Analyses of a Clay. Constituent. Mean values per cent. Maximum and minimum deviations. Silica 60-47 + 0-07 Titanic oxide . 1-20 + 0-07 Alumina . 21-90 +0-10 Ferric oxide . 1-50 + 0-06 Lime 1-45 + 0-08 Magnesia 1-10 + 0-08 Potash . 2-11 + 0-04 Soda 0-62 + 0-04 Loss on ignition 9-30 + 0-05 The above numbers take no account of constant errors. Dittrich's Table XLIII. indicates both constant and accidental errors. The errors would be different 1 M. Dittrich, Neues Jahrb. Min., 2. 69, 1903. 2 By " accidental error " is understood the irregular deviations from the arithmetical mean which are just as likely to have a positive as a negative value. A "constant error " is an error due to well-defined causes which make the error incline more in one direction than in another e.g., errors due to a defect in the pipette, burette, measuring flask, solubility of precipitate, etc. The correction tables pages 29 and 49 are intended to neutralise constant errors due to the causes named. For a discussion on accidental and constant errors, see J. W. Mellor, Higher Mathematics for Students of Chemistry and Physics, London, 510, 537, 1909. The accuracy of the results can be expressed in several different ways, e.g. : "the results are within, say, O'l of each other"; "the results are accurate to 0'1 per 100 parts of the sample"; "the results are accurate to 0'1 per 100 parts of the given constituent in the sample " ; etc. 248 A TREATISE ON CHEMICAL ANALYSIS. if different amounts of each constituent were present, and different methods of analysis employed. The relative error involved in separating a relatively small amount of a constituent is greater than in separating a large quantity, because, in the former case, a small quantity of impurity and the slight solvent action of the mother liquid have a greater influence on the final result. The sulphur and chlorine have not here been determined ; without these, and certain other constituents, the total is 99*59 per cent. This series of analyses, with some hundreds of other analyses, have led me to place the check for commercial analyses at 100 0*5, as indicated above. It will be obvious that if each constituent had its maximum deviation, or if each constituent had its minimum deviation, the total might fall outside the assigned limit, and the analysis be condemned although the error with no con- stituent exceeded the tolerated limits. The chance of this event happening is over 1 in 100,000,000 analyses. The magnitude of the errors might be reduced by working in a clearer atmosphere than sometimes prevails in the testing laboratory of a works, 1 and using platinum utensils. The numbers given above represent analyses made under routine conditions by the method described in what precedes. The Committee of the American Chemical Society " On Uniformity in Technical Analysis " 2 reported the results of 35 analyses of one sample of an argillaceous limestone made by different laboratories. These are somewhat startling in their want of agreement. I have given the mean of two concordant analyses by the two referees in the second column of the table, and the maxima and minima results sent to the committee by the different analysts in the last two columns. Table XL V. Comparative Analyses of an Argillaceous Limestone. Constituent. Standard. Minimum. Maximum. Silica . 18-14 16-58 18-92 Titanic oxide , 0-22 0-11 0-82 Alumina . 570 4-42 7-35 Ferric oxide . 171 1-06 2-83 Manganese oxide 0-04 none 170 Lime 37-65 35-26 41-98 Magnesia 1-93 0-92 3-05 Potash . 1-14 0-46 2-68 Soda 0-33 0-11 2-00 Loss on ignition Phosphoric oxide 32-27 0-18 31-94 0-12 32-88 0-65 Sulphur . ; 0-27 none 0-71 Sulphur trioxide 0-012 none 0-69 Carbon . 0-64 0-41 2-03 Carbon dioxide 30-68 28-65 31-65 1 I am here reminded of a quaint explanation, given by the Brit. Clayworker, 19. 155, 1910, of the fact that two analyses of a clay differed in the amounts of silica and iron. The one with the higher silica was made by a clay analyst, the one with the higher iron by an iron and steel works analyst. The claim was made that the iron in the latter case was high because the utensils of the iron and steel analyst were probably contaminated with iron. The tu quoque retort might have been made with reference to the silica of the clay analyst. 2 Journ. Amer. Chem. Soc., 26. 1652, 1904 ; 28. 223, 1906. See also " Report of the Sub- committee on the Uniformity in Analysis of Materials for the Portland Cement Industry," Journ. Soc. Chem. Ind. t 21. 12, 1223, 1902 ; H. W. Stanger, 21. 1216, 1902. ABBREVIATED ANALYSES AND ANALYTICAL ERRORS. 249 116. The Chief Sources of Error in Analyses. Every careless step in an analysis shows itself in material mistakes. The student must reason closely to keep his solutions correct. He cannot go on long with mere enthusiasm and boasting. His own results bring him the greatest reproaches. His experiments silently humble him, and he is laughed at by the forces which he cannot avenge. W. CROOKES. 1 It may be here instructive to follow Jiiptner's plan and summarise the more important sources of error. 2 Most of these have already been discussed in detail. (1) Imperfections in sampling. This, as already indicated, is a prolific source of discrepancies in analytical results. (2) Errors due to mistakes and lack of skill. Mixing the samples ; arithmeti- cal errors in calculations ; wrong reading of burettes, weights, tables ; sticking of Erdinann's float ; dirty vessels ; etc. The susceptibility of a worker to errors of this kind is greater in badly equipped laboratories, imperfect illumination ; over- work, with consequent fatigue and lapses of attention, and with a laboratory near a dusty road. Experiments with clays favour the development of dust, and clay dust settling in beakers, funnels, etc., conduces to high results. (3) Impure reagents. Phosphates in ammonium nitrate used for washing the alumina and the ammonium phosphomolybdate precipitates; iron and carbon in the zinc used for reducing ferric salts for the permanganate titration ; fluorine in the hydrogen peroxide used for the titanium determinations ; silica in the sodium carbonate flux ; alkalies in the calcium carbonate used in Smith's process ; vana- dium in the potash ; chlorides or sulphates in the sodium carbonate or nitrate ; etc. (4) Faulty measuring apparatus. The weights may be in error ; the volu- metric apparatus may not be consistent ; the volumetric measurements may not be corrected for variations of temperature ; etc. 3 (5) Errors in manipulation. Action of water on the glass vessels ; intro- duction of sulphur from rubber stoppers used in sulphur determinations ; moisture in the apparatus used in fluorine determinations ; losses by spitting when an alkaline carbonate is neutralised by an acid ; etc. (6) Errors due to faults in the process of analysis. Examples occur during the precipitation of barium sulphate (page 618) ; incomplete oxidation of carbon in the wet. combustion process (page 546) ; solubility of silica (page 186) ; adsorp- tion of salts by precipitates e.g., the "alumina" precipitate, and by the calcium carbonate in the Smith's process for alkalies ; the solvent action of hydrochloric acid on sulphides e.g., lead sulphide. (7) Personal errors. Defects in the perception of colour in colorimetric processes ; methods of reading burette ; method of adding reagents to solutions (page 178) ; etc. The result is that one chemist gets consistently better results with one process than with another, whereas the converse may be true for another chemist. 4 The differences of opinion as to the relative merits of the cyanide, iodine, and electrolytic processes for copper might be cited in illustra- tion. By studying the properties of precipitates, the number of filtrations, and 1 Chem. News, 19. 1, 1869. 2 H. von Jliptner, Journ. I.S. Inst., 49. 80, 1896; Chem. News, 74. 81, 1896; C. B. Dudley, Journ. Amer. Chem. Soc., 15. 501, 1893 ; A. B. Prescot, Chem. News, 53. 78, 88, 1886 ; S. H. Collins, Univ. Durham Phil. Soc., 1, 1909 ; E. A. Lewis, Journ. Soc. Chem. Ind. t 31. 96, 1912. 3 H. Lunden, Svensk. Kern. Tids., 24. 96, 1912. 4 This explains some puzzling statements which confront the student who " consults the originals." Given three processes A, B, and C for a particular determination, the author of the process A may quote analyses to prove that A is superior to B and C ; the author of B may try to prove that his process is superior to A or C ; and similarly, the author of process C may set out to prove that C is superior to A and B. 250 A TREATISE ON CHEMICAL ANALYSIS. of transfers from vessel to vessel, common sense will sometimes decide in favour of one of a number of rival processes. Other things being equal, the " margin of safety" that is, the risk of error, is less the smaller the number of separations involved in isolating a given constituent. The more complex the method of separa- tion, the greater the influence of the personal factor. Other things being equal, that method of analysis is safest which depends least on the skill of the operator. In order to reduce the personal factor to a minimum in the determination of phosphorus and manganese in iron and steel, where accuracy and speed are of vital importance, and where the general composition and range of variation of the substances to be analysed are known, methods have been devised, chiefly by C. H. and N. D. Risdale, 1 for the rapid determination of the constituents just mentioned. These methods styled mechanicalised processes and andloid processes are followed "mechanically" with prescribed quantities of solvents, measured reagents, and ready-made tablets of solid reagents introduced at definite assigned temperatures, etc. In this way the trouble, expense, and uncertainty involved when differences occur in analyses made by different chemists are reduced to a minimum. Proposals to standardise processes of analysis crop up from time to time. It is suggested that the methods of analysis in which manufacture and sale depend upon the results shall be standardised by an authoritative committee 2 and periodically revised. In this way, a greater uniformity in results might be expected. Progress would not necessarily be arrested, because improved methods would be examined by the committee periodically, and the less satisfactory methods cancelled. (8) Errors due to calculations based on different atomic weights. In this work we assume that the table of atomic weights indicated on page 684 is used. It may seem heresy to inquire if this is the best we can do. It will be obvious that the atomic weight question is of the greatest importance. It is not uncommon to find a difference of a few per cent, between the maximum and minimum values for the atomic weights found by different or by the same experimenters. The maximum value for the reliable determinations of the atomic weight of magnesium given by Clarke is 24*706 ; the minimum value 24'01 6. 3 In calculating the ratio MgO : Mg 2 P 2 7 , page 221, we assumed the atomic weight of magnesium to be 24*32. It seems that under the ordinary conditions of chemical analysis, the weight may deviate at least (possibly more than) + % 3 from this value. This means that instead of taking the factor 0*362 for con- verting a given weight of magnesium pyrophosphate into magnesia, we require a factor somewhere between 0*3643 and 0*3603. With small amounts of magnesia, it does not matter which be used 0*36 will do quite well. With larger amounts of magnesia, the result may be affected by over one-fourth per cent. It might therefore be personally interesting for an analyst to find if he works under the 0*3643 or the 0*3603 conditions. Many chemists recommend the use of an old discarded value for the atomic weight of platinum 197 in potash determinations for calculating the amount of KC1 or K 2 equivalent to the K 2 PtCl 6 precipitate. 4 There can be little doubt 1 0. H. andN. D. Risdale, Journ. I. S. Inst., I. 332, 1911 ; W. H. Herdsmann, Journ. West, Scot. I. S. Inst., 4, 1912. 2 For discussions on the standardisation of analytical processes, see Journ. Soc. Chern. Ind. , 3 2 210, 345, 356, 1884; J. Pattinson, ib., 3. 17, 1884; R. R. Tatlock, ib., 3. 307, 1884; J. C. Bell, ib., 2. 109, 1883 ; A. R. Ling, ib., 22. 677, 1903; C. Richardson, ib., 20. 334, 1901 ; J. White, ib. t 24. 390, 1905 ; H. D. Richmond, ib., 22. 676, 1903 ; G. Lunge, ib., Chcm. News, 47. 210, 1883 ; B. Blount, ib., 86. 177, 1900 ; H. von Jiiptner, ib. t 74. 81, 1896 ; Journ. I. S. Inst., 49. 80, 1896 ; W. D. Richardson, Journ. Ind. Eng. Chem., 2. 99, 1910. Page 248. 3 F. W. Clarke, A Recalculation of the Atomic Weights, Washington, 222, 1910. 4 F. Wolfbauer, Chem. Ztg., 14. 1246, 1890. ABBREVIATED ANALYSES AND ANALYTICAL ERRORS. 251 that the best chemists working with the best available instruments, under the best conditions, with the purest available materials, and with a few selected reactions, get a number nearer 195 than 197. Why then recommend 197? If, by a given analytical process, 197 gives a result nearer the correct value than 195, it is obvious that the former number should be employed in that process. It is, indeed, possible that if the atomic weight had been determined by the same reaction used in the analytical process, a number different from 195 would have been obtained. The conclusion is obvious. Each analytical process requires its own set of atomic factors, which may, or may not, coincide with the atomic weights given in the tables. Consequently, to get the best results in technical analyses, each chemist should find for himself the set of factors (or atomic weights) peculiar to that process. 1 If the ratio K 2 : K 2 PtCl 6 under one set of conditions be O19308, and under another set of conditions 0*19384, it would be absurd to employ the former number for conditions where the latter number would have been obtained. 117. The Statement of the Results. It would be an advantage if different analysts adopted one uniform practice in stating their results, since analyses would then be much easier to read. The Fifth International Congress of Applied Chemistry at Berlin agreed that 1. The name of the constituent is to be followed by the formula ; 2. By the name of an acid, the acid itself and neither its anhydride nor its ion is to be understood ; and 3. When the results are calculated in the form of metal oxide and acid anhydride, the latter is to be written either as " sulphuric anhydride, S0 3 ," or as " sulphuric acid (calculated as anhydride) S0 3 ." 2 Washington and Hillebrand 3 recommend stating the results somewhat in the following order : Si0 2 , A1 2 3 , Fe 2 3 , FeO, MgO, CaO, Na 2 0, K 2 0, H 2 (ignition), H 2 (below 110"), C0 2 , Ti0 2 , Zr0 2 , P 2 5 , S0 3 , Cl, F, S, (FeS 2 ), Cr 2 3 , V 2 3 , MnO, NiO, CoO, CuO, ZnO, BaO, SrO, Li 2 0, C, NH 3 , and organic matter. The idea is to keep the important oxides at the head of the list, so that the general character of the silicate can be seen at a glance. This is no doubt excellent. For clays, I prefer to keep the titanium among the important constituents, since, in the higher grades of clay, it is often as important a constituent as iron, and more important than magnesia. The P 2 5 , S0 3 , Cl, F, C0 2 , and carbon are best kept as a separate group. For glazes, frits, etc., I also prefer to keep certain bases in a separate group. The following lists represent the order I use for commercial analyses of clays and glazes and similar materials, where the purpose of the analyses is different from that of Washington and Hillebrand : CLAY DRIED AT 109 C. GLAZE DRIED AT 100 C. Hygroscopic moisture lost at 109 C. Hygroscopic moisture lost at 100 C. Silica (Si0 2 ). Silica (Si0 2 ). Titanic oxide (Ti0 2 ). Alumina (Al a 3 ) and Ferric oxide (Fe 2 3 ). 4 Alumina (A1 2 3 ). Lime (CaO). Ferric oxide (Fe 2 3 ). Magnesia (MgO). Manganese oxide (MnO). Potash (K 2 0). 1 Or to the analyst himself. ' 2 W. Fresenius. Zeit. anal. Chem., 44. 32, 1905. See also Chem. News, 53. 186, 1886 ; E. C. C. Stanford, 16. 29, 190, 1874 ; C. G. Hopkins, Journ. Amer. Chem. Soc., 29. 1312, 1907. 3 H. S. Washington, Amer. J. Science (4), 10. 59, 1900 ; Prof. Paper U.S. Geol. Sur., 14. 24, 1903; 28. 7, 1904; W. F. Hillebrand, Bull. U.S. Geol. Sur., 305. 27, 1907. 4 There is usually no need to separate the constituents of the ammonia precipitate unless phosphates be present. 252 A TREATISE ON CHEMICAL ANALYSIS. CLAY DRIED AT 109 C. GLAZE DEIED AT 100 C. Magnesia (MgO). Soda (Na 2 0). Lime (OaO). Loss when calcined over 100 C. 1 Potash (K a O). Soda (Na 2 0). Lead oxide ( PbO). Loss when calcined over 109 C. Tin oxide (Sn0 2 ). Zinc oxide (ZnO). Chlorine (Cl). Barium oxide (BaO). Fluorine (F). Phosphoric oxide (P 2 5 ). Fluorine (F). Sulphur (S). Phosphoric oxide (P 2 5 ). Sulphur trioxide (S0 3 ). Sulphur trioxide (S0 3 ). Carbon (C). Carbon dioxide (C0 2 j. Carbon dioxide (C0 2 ). Boric oxide (B 2 O 3 ). If there is any evidence showing the way the acids and bases are combined, the amounts may be stated separately e.g., calcium carbonate (whiting), white lead, calcium fluoride, etc. When the way the acids and bases are combined is open to doubt, it is best to leave the adjustment to those who intend to transform the analysis of, say, a glaze into a recipe, etc. 2 It is sometimes difficult to decide how best to report the separate con- stituents in an analysis, since curious practices prevail in buying and selling which are not always those most convenient for the consumer. Manganese ores, for instance, may be sold on their per cent, of metallic manganese, or manganese peroxide ; cerium is weighed as cerium dioxide, but the earths are sold on their per cent. Ce 2 8 ; chromium ores are valued on their O 2 3 contents ; tungsten ores on their W0 3 contents ; tantalum and vanadium ores on the amount of the respective elements equivalent to Ta 2 5 and V 2 5 they contain ; while uranium ores are valued on their equivalent U 3 8 . 3 Analysts must then bow to the inexorable fetish custom. 1 The interpretation of " loss on ignition " is somewhat indefinite with glazes, and is rather a qualitative indication of the character of the glaze in question. 2 For an example of the discrepancies which may arise when the analyst attempts to go beyond his facts, and show the mode in which the " bases " and "acids " are combined, compare H. M. Noad, Journ. Ghem. Soc., 14. 43, 1862; and A. Voelcker, ib., 14. 46, 1862; Chem. News, 3. 77, 285, 315, 1861. 3 G. T. Holloway, Trans. Inst. Min. Met., 21. 569, 1912. CHAPTER XIX ELECTRO-ANALYSIS 118. Some Definitions. IF two platinum plates be dipped in an aqueous solution of, say, copper sulphate, and the plates be connected by wire with the poles of a battery or accumulator, metallic copper will be precipitated on one plate, and if the condi- tions be suitable, practically all the copper will separate in the form of a compact coherent metallic film. If the plate with the film of copper be washed, dried, and weighed, the increase in the weight of the plate before and after the experiment will represent the amount of copper which was present in the solution under investigation. The process of decomposition is called electrolysis ; the solution undergoing decomposition is called an electrolyte ; the two plates by which the current enters and leaves the electrolyte are called electrodes. The electrode by which the current is conventionally supposed to enter the electrolyte is called the anode ; and the other electrode is called the cathode. 1 What has just been stated about the solution of copper sulphate applies to many other metallic salts. When such solutions are electrolysed, the metal is deposited on the cathode. But with solutions of lead and manganese salts, the peroxides are sometimes deposited on the anode. In applying these facts to practical analysis, it is convenient to make the dish holding the electrolyte one of the electrodes generally the cathode on which the deposit is to be collected. The quantity of electricity which passes through the solution in unit time, or the speed of the current, is measured by an ammeter. The unit is called an ampere? The unit of the electric pressure 3 driving the electric current along the circuit is called the voltf and the voltage of the current is measured by the voltmeter. The current is regulated by' coils of wire, incandescent lamps, etc., which obstruct the steady flow of the current of electricity and fritter the energy of the electric current down into heat. The unit of resistance is called the ohm. 5 1 It is always easy to determine which wire from the battery or accumulator belongs to the cathode, and which to the anode, by wetting a piece of blotting paper with an aqueous solution of potassium iodide and starch, and allowing the tips of the two wires to touch the wet paper about a quarter of an inch apart. The paper in the vicinity of the wire to be attached to the anode will be coloured blue. If the blotting paper be soaked in a solution of sodium chloride a solution of silver nitrate in water, deposits metallic silver at the rate of '001118 grm. per second." 3 Also called the electromotive force, or potential. 4 The volt ' ' is represented by the electrical pressure that if steadily applied to a conductor whose resistance is one ohm will produce a current of one ampere." 5 The ohm is the unit of resistance and "is represented by the resistance offered to an unvarying electric current by a column of mercury at the temperature of melting ice 14 '4521 grms. in mass, of a constant sectional area and a length of 106 '3 cm." 253 254 A TREATISE ON CHEMICAL ANALYSIS. There is an interesting relation Ohm's law between these magnitudes : Quantity of electricity (amps. ) ^ / ressure ^ vo ^ . Resistance (ohms) Consequently, decreasing the resistance opposed to, or increasing the voltage of a current will increase the quantity of electricity passing through a circuit per second (amperage) ; and, conversely, increasing the resistance or decreasing the voltage will decrease the quantity of electricity passing through the circuit per second. 1 119. Some Factors which determine Success in Electro-analysis. 1. Pressure of the Current Decomposition Voltage. For the decomposition of every metallic solution, the pressure of the electric current must exceed a certain minimum value. This minimum voltage is characteristic of the particular solution under investigation. For instance Sulphates. Chlorides. volts. volts. Nickel sulphate . . . 2'09 Nickel chloride . . .1-85 Cadmium sulphate . . . 2 '03 Cadmium chloride . . . 1 "88 Cobalt sulphate . . . T92 Cobalt chloride . . .178 A current of many amperes might be passed through a solution of cobalt sulphate without any visible decomposition; only when the voltage exceeds 1'92 volts will cobalt be deposited. If a solution contains a mixture of salts of different metals, the electrodes appear as if they exerted a kind of selective action which really depends upon the minimum voltage required to effect the decomposition of the different metallic salts. This peculiarity of the salts can be employed to effect the separation of the different metals. In a solution containing zinc and cobalt sulphates, for example, cobalt will be deposited with a voltage less than 2*02 volts, and greater than 1*92 volts. When all the cobalt is deposited, increasing the voltage above 2 '08 volts will lead to the deposition of zinc. 2. Strength of the Current Current Density. The weight of a given metal deposited by a current of electricity is proportional to the quantity of electricity (amperes) passing through the solution ; and the weights of different metals deposited by the same quantity of electricity (amperes) is directly proportional to the chemical equivalents of the metals in the solutions (Faraday's laws). Thus, a given quantity of electricity will deposit the relative quantities of the different metals indicated in the last line of the following scheme : Al Ni Co Sn(ic) Cu ' Cd Atomic weight .... 27'1 58'68 58'97 119'0 63'57 112'4 Chemical equivalent . . . 903 29'34 29'48 2975 3178 56'2 The quantity of metal deposited in a given time, that is, the rate of deposition, is dependent upon the strength of the current in amperes. The deposited metal redissolves in the electrolyte, and, consequently, the rate at which the metal is deposited must exceed the rate at which it redissolves. The metal is only deposited from the solution in the immediate vicinity of the cathode. Hence, other things being equal, the greater the area of the cathode the more metal available for deposition, and, mutatis mutandis, the greater the rate at which the metal is redissolved. Hence, the amount of current necessary to deposit the metal will depend upon the area of the cathode. The amount of current per unit area (per second) is called the current density 100 sq. cm., that is, 1 sq. decimetre, is generally taken as unit area. Hence, " a current density of 2 amperes " means that 2 amperes should be used for each 100 sq. cm. of cathode area. 1 J. W. Richards, Chem. News, 94. 5, 20, 31, 1906 " Electrochemical Calculations." ELECTRO-ANALYSIS. 255 While the tendency of a metal to redissolve in the solution fixes a lower limit to the amount of current which may be used for electro-analysis, the tendency of the metal to form spongy, non-adherent films when deposited too rapidly prevents the use of currents exceeding a certain maximum strength (amperage per unit area of cathode). It is highly important, for accurate work, to precipitate the metal in a coherent film which is easily washed and weighed. The condition of the deposited metal is not only dependent upon the voltage and the current density, but is also determined by the nature of the solution from which the metal is deposited, the amount of free acid, the temperature, etc. These "optimum" conditions can only be determined by trial and failure. 3. Nature and Concentration of Electrolytes. Enough salt should be present in the solution to carry the current, otherwise hydrogen, or some other element, may cause the film of metal to become spongy and impure. The more con- centrated the solution, the greater the amount of metal in solution in the neighbourhood of the cathode. The nature of the salt undergoing decomposition is of importance. Copper nitrate, for example, gives better results than copper sulphate ; and nickel sulphate is far more satisfactory than nickel nitrate. Better deposits are frequently obtained with complex salts like the metallo-cyanides, double oxalates, etc., than with the simple sulphates, nitrates, etc. 4. Nature and Concentration of Foreign Salts in the Electrolyte. A current density which gives a perfectly pure coherent deposit of, say, copper when no other metal is present may give a very impure deposit when, say, arsenic is present, particularly if the decomposition voltage of the foreign salt be near that of the copper salt. The greater the concentration of the foreign salt, the greater the danger of its simultaneous deposition with the copper. 5. Temperature. The time required for the complete electrolysis of certain solutions, e.g., copper or nickel sulphate, lead nitrate, etc., is frequently ab- breviated by working at a higher temperature, but the deposits are not always so good as those formed at lower temperatures. When working at an elevated temperature, care must be taken not to heat the solution to its boiling point, or the deposit may be loosened from the cathode, and an accurate determination is then impracticable. 6. Condition of the Electrolyte. Only those metallic salts in the vicinity of the electrodes can be decomposed by the current, and the diffusion of more salt from the body of the solution to the vicinity of the electrodes is comparatively slow. Hence, adequate agitation of the solution will greatly accelerate the speed of precipitation by supplying fresh metal to take the place of the metal withdrawn from the solution faster than it can be supplied by simple diffusion. When artificial means are used to agitate the solution during electrolysis, a greater maximum current density can be used. Only one of the electrodes is usually rotated, and the consequence is that the current may be considerably increased without injuring the deposit; and deposits which require several hours under ordinary conditions may be completed in as many minutes. This subject will be discussed later. In giving directions for electrolytic separations, or in stating the results of experiments, the following factors, apart from those connected with the apparatus, are to be considered essential for success : 1. Pressure of electric current in volts. 2. Current density in amperes per square decimetre. 3. Nature and concentration of the electrolyte. 4. Nature and concentration of foreign salts dissolved in the electrolyte. 5. Temperature of the electrolyte. 6. Condition of the electrolyte at rest, or agitated. 2 5 6 A TREATISE ON CHEMICAL ANALYSIS. If attention be paid to these details necessary for successful work, many electrolytic methods of analysis rival in accuracy, neatness, and rapidity the best of the gravimetric and volumetric processes available for the same metal. In consequence, a number of electrolytic methods have won a permanent place in analytical chemistry. True enough, the time required for the electrolysis is prolonged, but very little attention is needed during the actual electrolysis. Indeed, many electrolyses can be safely left 12 hours say overnight without attention. Those methods involving the use of rotating electrodes enable a de- termination to be made in 10 or 15 minutes. An example is given on page 337. 120. The Apparatus for Electro-Analysis. When a relatively small number of determinations is made, the outfit will be different from that employed when determinations are frequently practised. The former will alone be considered here. For the latter, special text-books 1 must be consulted. The determination of copper is conveniently taken as a standard process for reference. This is also particularly appropriate not only because of the typical character of the process, but also because the Mansfield Oberberg und Hiittendirection, in 1867, offered a prize for a rapid and accurate method for the determination of copper in ores. The prize was won by C. Luckow 2 for an electrolytic process. This gave an impetus to electro- analysis generally. A plan of the outfit for occasional work is shown in fig. 117. The current from an accumulator B passes through the resistance R to the decomposition cell C, then through the ammeter J., and back to the accumulator. A voltmeter G is placed in metallic contact with the cathode and anode. Usually a couple of 2- volt E.P.S. accumulator cells B will suffice for the current. The key K is for making or breaking the passage of the current as desired. A photograph of the apparatus is FIG. 117. Plan of fig. 118. shown in fig. 118. The electrolytic cell shown at C is the type recommended by Classen. It consists of a platinum basin, about 200 c.c., with or without an inner matted surface. 3 This serves as the electrode for collecting the deposit. A perforated platinum disc fixed to a stout platinum wire may be used for the 1 A. Classen, Quantitative Analyse durch Electrolyse, Berlin, 1908 ; B. B. Boltwood's trans., New York, 1908; A. Fischer, Electroanalytische Schnellmethoden, Stuttgart, 1908; F. M. Perkin, Practical Methods of Electro-chemistry, London, 1905 ; J. Riban, TraiU tf analyse Chimique Quantitative par Electrolyse, Paris, 1899 ; B. Hollard and L. Bertiaux, Analyse des Metaux par Electrolyse, Paris, 1909 ; E. F. Smith, Electro-analysis, Philadelphia, 1907 ; B. Neumann, Theorie und Praxis der analytischen Elektrolyse der Metallc, Halle a. S., 1897 J. B. C. Kershaw's trans., London, 1898. 2 C. Luckow, Zeit. anal. Chem., 8. 23, 1869 ; 19. 1, 1880. 3 The nature of the surface of the electrodes is of importance. Some metals give less satis- factory deposits on hammered surfaces than on spun and polished surfaces. In the determination of lead as peroxide, the precipitate only adheres firmly to the electrode when the surface has been roughened by means of, say, a sand-blast. A woven platinum gauze electrode flag electrode is also recommended in special cases. The electrode must be quite clean or a poor deposit will be obtained. ELECTRO-ANALYSIS. 257 other electrode. 1 The dish rests on a brass "retort ring" to which three platinum points have been fixed to ensure electrical contact between the retort ring and the platinum dish. The ring is clamped to a glass rod fixed to a heavy The platinum wire supporting the disc is also fixed by a suitable iron base. FIG. 118. The apparatus for an electro-analysis (electrodes stationary). clamp to the same glass rod. The glass rod serves to insulate the disc from the dish. If otherwise, the current would "short circuit" and not pass through the decomposition cell. Binding screws are fixed to the clamps, so that the dish and disc can be put into electrical contact with the battery. 2 drea of cathode surface^ sq cm. Depth in Dish in cm. Capacity of Dish in c.c FIG. 119. Classen's dish. Classen's dish, shown in fig. 119, is 9 cm. in diameter, and 4'2 cm. deep. It holds about 250 c.c., and weighs between 35 and 37 grms. The dish shown in the diagram held 231 c.c. when filled to within 1 cm. of the upper edge; and 156 c.c. when filled to within 2 cm. of the upper edge. With a disc anode, the best distribution of the current is obtained by adjusting the disc so that it is exactly in the centre of the dish and about 2 or 3 cm. below the upper edge. 1 Several other types of anode and cathode are in use dishes, cones, cylinders, gauzes, spiral wires, etc. The important points about the electrodes are : (1) they must not be attacked by the electrolyte, nor absorb gas ; (2) they must be of such a shape that the density of the current on the electrode which receives the deposit is as homogeneous as possible ; and (3) the shape must favour rapid diffusion of the electrolyte from one electrode to the other. s It is easy to make less expensive supports than the one here described. A lot depends upon the "taste" of the worker. The construction of apparatus for electrolytic work lends itself to what Ostwald calls "Basteln" pottering. The voltmeter and ammeter stand in the background of fig. 118 has ammeters and voltmeters for running two independent determina- tions at the same time. 17 258 A TREATISE ON CHEMICAL ANALYSIS. If the electrolytic cell is to be heated, a thin asbestos board or quartz plate is placed under the dish, and the vessel is heated by a small burner (fig. 118) so as to ensure a uniform temperature. The resistance or rheostat R, shown in the photograph, is made from platinoid wire wound on an asbestos-covered brass tube, or on a slate frame. By moving the sliding contact along a bar, any desired resistance can be obtained within the capacity of the instrument. 1 The ammeter A is an instrument with a small internal resistance so small that it can, for most purposes, be neglected. In consequence, the ammeter is kept in the main circuit all the time an experiment is in progress. A 20-amp. meter reading to - amp. will suffice for the work described in this book. 2 The voltmeter G has a relatively large internal resistance, and if placed in the main circuit would oppose so much resistance to the passage of the current that very little current would pass at all, or the coils of the voltmeter would be burnt out. If, for instance, the voltmeter has a resistance of 1000 ohms, and the electrolytic cell 2 ohms, the ratio of the current passing through the decom- position cell to that passing through the voltmeter will be as 500 : 1. The volt- meter may be left in the "shunted " circuit all the time an electrolysis is in progress. A voltmeter reading to 20 volts graduated in half- volts is sufficient. 3 121. The Electrolytic Determination of Copper. Copper can be satisfactorily deposited from solutions acidified with nitric or sulphuric acids, but not hydrochloric acid, except under special conditions. Nitric acid gives better results than sulphuric acid. The determination will be here described in some detail to serve as a type for later references. 4 1. Cleaning the Electrodes. The electrodes must be perfectly clean. Never touch the depositing surfaces of the electrodes with the fingers, to avoid danger of contaminating the surfaces with grease. Clean electrodes are absolutely necessary for good deposits. Silver soap or round-grained sea sand applied with a soft cloth or small sponge is commonly used for scouring and polishing. Grease is removed either by heating the electrodes to redness, 5 or by immersion in a saturated solution of chromic acid (or potassium dichromate) in concentrated sulphuric acid. Wash with distilled water, and dry by warming. The cathode is cooled in a desiccator, and weighed. 2. Preparation of the Electrolyte. Dissolve a gram of copper sulphate in 140 c.c. of distilled water; add 5 to 10 c.c. of nitric acid (sp. gr. 1'42), so as to make the solution between 8' and 10 per cent, nitric acid. The solution is placed in the weighed platinum dish. If the electrolysis is to be conducted at an elevated temperature, say 60, warm the solution in the dish, and adjust the gas so as to maintain the necessary temperature. 6 1 Numerous other types are used. Some of these will be found described in the text-books cited above. a Ammeters with a range 0-2 amps, reading in hundredths, provided with a shunt for reading 0-20 amps, in tenths, will be ample for other Avork than that described here. 3 In purchasing a new instrument, one reading 0-3 volts in tenths, provided with a shunt for reading 0-30 in half-volts, will be found useful for work with rotative electrodes. 4 W. C. Blasdale and W. Cruess, Journ. Amer. Chem. Soc., 32. 1231, 1910 ; D. J. Demorest, Journ. Ind. Eng. Chem., 5. 216, 1913. 5 Never heat platinum electrodes unless every particle of the previous deposit has been removed, otherwise an alloy may be formed which will spoil the apparatus. See page 114. 6 If the solution contained much free nitric acid, it should be evaporated to dryness, and the residue dissolved in the prescribed amount of nitric acid and water. If the solution contained much free sulphuric acid, the solution should be neutralised with ammonia, and the necessary amount of nitric acid added. ELECTRO- ANALYSIS. -259 3. Adjustment of the Apparatus. The apparatus is fitted up as indicated in fig. 118. See that all the metallic contacts battery connections, connections with the voltmeter, ammeter, resistance, and electrolytic stand are clean and rigid. If otherwise, the contacts may, later on, offer so much resistance that the current is either weakened or interrupted. The anode should be adjusted in the centre of the dish, and at a sufficient distance (O5 to 1 cm.) from the cathode to present sufficient resistance to prevent the necessity of using a very high voltage when the current has started. It is best to arrange the resistance, etc., so that when the circuit is closed a current no greater than the maximum required passes through the circuit. Cover the platinum dish with a watch- glass cut in two pieces, and provided with notches for the anode, and also for a thermometer if the electrolysis is to be conducted at an elevated temperature. The watch-glass is intended to prevent loss by the spray carried off with the gases liberated at the anode. 4. The Electrolysis. Complete the circuit and adjust the resistance so that a current of from 2'0 to 2*5 volts 1 passes through the circuit, and a current density of 0'5 to I'O 2 amp. 3 A bright red film will flash over the cathode surface as soon as the circuit is closed. The electrolysis will be finished in about 4 hours. If the electrolysis is to run overnight, a current density of about O'l amp. will suffice. A little more nitric acid 2 c.c. should also be added, since some of the nitric acid is converted into ammonia by the hydrogen" liberated at the cathode, and this is inclined to cause a spongy deposit. To find if all the copper has been deposited, raise the level of the solution by the addition of a few c.c. of distilled water. If no copper is deposited on the newly exposed cathode surface after a run of about 15 minutes, it may be assumed that the electrolysis is finished. 4 Or, a few drops of the electrolyte may be transferred by means of a pipette to a test-tube, and test for copper in the ordinary manner, say, make alkaline with ammonia, then acidify with acetic acid, and add a few drops of potassium ferrocyanide. A brownish-red p IG i 2 0. precipitate shows that all the copper has not been deposited, and that it is necessary to continue the electrolysis. When no coloration appears on repeating the test, the electrolysis is completed. 5. Washing and Drying the Deposited Metal. Since the precipitation has been made in an acid solution, some consider that the current should not be stopped until the acid liquid has been removed, otherwise the dissolution of the deposited copper will start before the plates can" be washed, and thus lead to low results. In most cases, however, sufficiently accurate results will be obtained by breaking the current, and immediately ^ pouring the solution into 1 The voltage and current density depend upon the resistance of the electrolytic cell, which in turn depends upon the conductivity of the solution, the size and shape of the electrodes, and the distance apart of the electrodes. It is frequently inconvenient to regulate these two factors to correspond exactly with prescribed directions. In that case, bring the voltage to the desired value, and let the current adjust itself to the required amperage. This is particularly the case with separations where the decomposition voltage is of prime importance. 2 If other metals are present in the solution, keep to the lower amperage. 3 From fig. 119 we see that the dish with 150 c.c. of solution offers a cathode surface of approximately 120 sq. cm. Hence, if 0'8 amp. is needed per 100 sq. cm., a cathode surface of 120 sq. cm. will require a current of 1 amp. read on the ammeter. 4 Classen (I.e.) places a small strip of bright platinum foil in contact with the cathode (F, fig. 120), but not touching the anode. If after half an hour no deposit is formed on the foil immersed in the liquid, it is safe to assume that the electrolysis is complete. 5 F. Kudorff (Zeit. anyew. Chem., 6. 5, 1892) adds 10 drops of a saturated solution of sodium acetate just before breaking the circuit. The acetic acid which is set free does not attack the 26o A TREATISE ON CHEMICAL ANALYSIS. an empty beaker, and rapidly rinsing out the basin with hot distilled water. Finally, wash three times with about 5 c.c. of alcohol, 1 and once with ether. 2 Dry the precipitated copper in an air-bath at about 80 ; cool in a desiccator, and weigh. When it is desired to wash the deposit free from acid before stopping the current the acid liquid is syphoned off (by suction) into a nitration flask, and at the same time fresh distilled water is poured into the basin until the washings are free from acid before stopping the current. A method of doing FIG. 121. Washing the cathode dish. this is illustrated in fig. 121. Then wash the deposit with hot distilled water, etc., as indicated above. 6. Errors. If the current density be too high, the deposit may be "burnt." In that case, the copper, instead of appearing as a bright red coherent film, is coloured more or less brown, and appears more or less pulverulent and non- adherent. If the current be acting for too long a time, there is a possibility of some of the sulphuric acid being reduced by the hydrogen liberated at the cathode to form hydrogen sulphide. In that case, some copper sulphide will be formed, and this leads to the development of dark brown spots on the copper. If arsenic or antimony be present, these elements may be deposited with the copper, giving it a dull and maybe a grey appearance. In that case, the copper immediately, and in consequence there is sufficient time to wash the plates before appreciable action has occurred. 1 The alcohol should give no residue on evaporation to dryness. The alcohol should also have been distilled over lime. Instead of absolute alcohol, methylated spirit free from mineral oil can be used. The methylated spirit is purified by standing in contact with caustic soda for a few days, and distilling. Freshly burnt lime is now added to the distillate, and after standing 24 hours, the spirit is decanted into a distilling flask containing some freshly burnt lime, and redistilled. 2 The ether is supposed to have been distilled over caustic potash. Some omit the ether treatment. ELECTRO-ANALYSIS. 26 1 copper deposit may be heated to dull redness to volatilise the oxides of arsenic and antimony. 1 Dissolve the resulting copper oxide in nitric acid, and repeat the electrolysis. Hollard and Bertiaux say that the addition of a little ferric sulphate lessens the danger of precipitating arsenic ; and the addition of a little lead nitrate, the deposition of antimony. If much arsenic, antimony, or bismuth be present, they should be separated chemically before the electrolysis. Silver, if present, will be deposited with the copper, and should therefore be removed from the solution before the electrolysis, or the metal deposited on the cathode can be weighed as " silver + copper." The mixed metals are then dissolved in nitric acid, and the silver determined by the addition of a little hydrochloric acid in the usual manner. Tin and mercury, if present, may also be deposited with the copper. In illustration of the results which may be obtained with the process just described : Used 25-08 25 '08 25'08 25 '08 percent. Found 25-08 25 '05 25 '07 25 '04 ,, 7. Recording the Results. The results entered in the note-book will include the following data : 1. Current pressure : 2 '2-2 '5 volts. 2. Current density : 5 amps. 3. Electrolyte : 1 grro. of copper sulphate with 8 to 10 per cent, of nitric acid, made up to 150c.c. 4. Foreign salts : no foreign metallic salts were present. 5. Temperature: 15-20. 6. Electrodes : stationary. Time, 4J hours. One gram of the sample was dissolved as indicated on page 258, and used as the electrolyte. The weighings were : Cathode and deposit ......... 36 '3542 grins. Cathode 36 '0734 Copper deposit . . . , . . . 0'2808 grm. Hence, the sample contained 28*08 per cent, metallic copper. 8. Cleaning the Deposit from the Electrode. In the case of copper, the deposit can be readily removed from the platinum dish by the action of dilute nitric acid, and subsequent washing with distilled water. Electrolysis of Sulphuric Acid Solutions. It may be desirable to precipitate the copper from a sulphuric acid solution free from nitric acid. The deposits of copper are not then quite so red in colour as when nitric acid is present. Between 7 and 10 per cent. H 2 S0 4 is added, that is, about 4 c.c. per 100 c.c. of solution. According to C. Engels, the deposits are better if from 1 to 1J grms. of hydroxylamine sulphate be present. 2 About 2 -5 volts and a current density of 0'5 amp. are employed. The best temperature is from 70 to 80. Between 1 and 1J hours are needed for the electrolysis. The process is used when copper is to be separated from zinc, cadmium, nickel, and less than O'l grm. of iron. If over this amount of iron say, up to 0*6 grm. be present, the electrolysis must be conducted at atmospheric temperatures. 3 The time required for the electrolysis is then between 2 and 2J hours. For an over- ' Remember that a platinum dish will not stand much of this sort of treatment. 2 A. Classen recommends urea for the same purpose. In that case, some carbon is deposited with the copper, and the subsequent correction is troublesome. 8 0. Foerster, Zeit. angew. Chem., 19. 1890, 1906. 262 A TREATISE ON CHEMICAL ANALYSIS. night electrolysis, use rather less sulphuric acid, and O5 grm. of hydroxylamine sulphate, also a current density of about Ol amp. Electrolysis of Ammoniacal Solutions. Add ammonia to a solution of copper chloride containing less than 2 grms. of copper. When the precipitate has re- dissolved, add 20-25 c.c. of ammonia (sp. gr. 0'96) for quantities of copper up to 0*5 grm., and 30-35 c.c. for quantities up to 1 grm. Add 3-4 grms. of ammonium nitrate. If up to 0'5 grm. of copper be present, use a current density of 0'5 amp., and so proportionally up to 2 amps, for 2 grms. of copper. Time : about 2 hours. If insufficient ammonia be present, a brown non-adherent deposit collects on the anode, which falls off and contaminates the cathode deposit. 1 1 F. Oettel, Chem. Ztg., 18. 47, 879, 1894 ; F. Riidorff, er. t 21. 3050, 1888. PART III. ANALYSIS OF GLASSES, GLAZES, COLOURS, AND COMPLEX SILICATES. CHAPTER XX. THE ANALYSIS OF GLAZES, GLASSES, ENAMELS, AND COLOURS. 122. The Selection of the Sample. THE glaze to be analysed may be partly or wholly raw or fritted. 1 In the latter case, the method of analysis is the same as for glasses, enamels, and frits. Again, if a fired glaze be in question, it is usually necessary to chip the glaze from the body with a chisel and hammer. In that case, it is impossible to prevent part of the body being analysed with the glaze, because (1) the glaze while being fired dissolves some of the constituents of the body ; and (2) the mechanical separation of chippings of the glaze from the body, even under a good lens, is laborious and unsatisfactory. 2 In the case of iron enamels, the analysis will include part of the intermediate flux between the enamel and the metal. 3 To remove a glaze or enamel from a body, de Luynes 4 recommends roughening the surface of the glaze with emery or carborundum paper or a file, and coating the roughened surface with wet glue. The glue is dried and baked in an air-bath. During the drying, the glue sometimes drags part of the glaze or enamel from the body. The glue can be removed by means of boiling water, and washing the fragments on a filter paper. As a rule glazes contain silica, boric oxide, alumina, ferric oxide, lime, magnesia, lead oxide, and alkalies. 5 There may also be present tin oxide, baryta, phosphoric and sulphuric oxides, and fluorine. .In special cases, zirconia, antimony, and arsenic may be present. In the case of colours, zinc, chromium, cobalt, nickel, manganese, copper, bismuth, cadmium, titanium, uranium, molybdenum, and gold may be found. The last six elements named are rare. 6 Selenium is occasionally found in clear glasses and frits. The general scheme for the separation of the more common elements is as follows : I. Hydrogen sulphide group. (1) Copper group. Insoluble in sodium sulphide e.g., copper, lead, bismuth, cadmium. (2) Arsenic group. Soluble in sodium sulphide e.g., arsenic, anti- 1 An examination of the constituents removed by treatment -with dilute acids, and the determination of the carbon dioxide and water, will often furnish valuable data for reconstructing such glazes from the analysis. a The interpretation of the analysis is then more or less obscure and vague. 3 Fragments of iron may be removed with a magnet. 4 V. de Luynes, Comp. Rend., 134. 480, 1902. 5 Along with carbon dioxide and combined water with raw and partially fritted glazes. 6 Molybdenum, gold, silver, iridium, rhodium, and platinum are extremely rare. Tungsten is rare, although traces are common enough in tin glazes. It is introduced as an impurity with the tin oxide, and its presence is not always objectionable. 265 266 A TREATISE ON CHEMICAL ANALYSIS. II. Ammonia group, or basic acetate group e.g., aluminium, iron, titanium, chromium. III. Ammonium sulphide group e.g., zinc, manganese, cobalt, nickel. The treatment when the rarer elements are present is discussed later. The boric oxide is determined on a separate sample by the methods described later. We first discuss the metals precipitated by hydrogen sulphide in acid solutions ; and follow by a discussion of the processes required when zirconium, manganese, uranium, cobalt, nickel, chromium, zinc, etc., are present. Tn any case, the given sample is ground to a fine powder (page 121). 123. "Opening" the Sample. The method to be employed for the analysis of a glaze is determined by the result of the qualitative analysis. The colour of the glaze, etc., is often a good indication of the colouring oxides present. If the glaze contains unfritted white or red lead, a preliminary digestion with dilute acetic, hot hydrochloric, or nitric acid will remove the lead. The insoluble residue is fused 1 with sodium carbonate (as indicated page 164), and the fused mass digested with, say, dilute nitric acid. The solution may be added to that obtained by the preliminary digestion with nitric acid. If all or part of the lead be fritted, and the frit is not completely decomposed by digestion with these acids, 2 another procedure must be followed. The lead frit can be fused in a platinum crucible provided there be no possibility of reducing conditions developing in the interior of the crucible. If tin oxide be present, it will dissolve in the molten sodium carbonate slowly. Fusion with Sodium Hydroxide or Peroxide. An enamel, glaze, or glass containing antimony, 3 arsenic, or tin oxide can sometimes be conveniently fused with eight to ten times its weight of sodium hydroxide 4 (or sodium peroxide) in a silver or nickel crucible at a dull red heat until the mass is fused. Dissolve the fused mass, when cold, in a little water or dilute hydrochloric acid. The objection to the use of caustic alkalies is their tendency to froth over, and the time necessary for the solution of the powder. 5 The following is the best way of conducting the operation : The crucible 6 is placed in a circular hole cut in a sheet of -" asbestos millboard, so that the crucible when pressed tightly into the aperture projects on the upper side about a quarter of an inch. The lid of the crucible is tapped on an agate mortar with a round-faced hammer, so that the lid fits the crucible with its convex side downwards. Any portions 1 Care must be taken in using a platinum crucible when metallic oxides and salts are present, or the crucible may be attacked during the sodium carbonate fusion, etc. 2 For the joint effect of acids and a metal like zinc, see T. Moore, Chem. News, 67. 267, 1893. 3 For the volatilisation of antimony during the fusion with sodium carbonate, see H. N. Warren, Chem. News, 67. 16, 1893. 4 A "pinch " of powdered wood charcoal quarter gram accelerates the decomposition of tin oxides, cassiterite, etc. The solution of the cassiterite will be complete in three or four minutes, but the heating is continued a little longer in order to burn off the carbon C. A. Burghardt, Chem. News, 61. 260, 1890; Proc. Manchester Lit. Phil. Soc. (4), 3. 171, 1890; A. Gilbert, Zeit. offent. Chem., 16. 441, 1910. This process also works well with chrome iron ore, wolframite, etc. H. T. Loram (Proc. Vhem. Soc., 27. 60, 1910) recommends fusing the "tin ore" in a silver crucible, with six or seven times its weight of potassium hydroxide, and its own weight of potassium cyanide. Extract the cold mass with water. Dissolve all in dilute hydrochloric acid. Boil to expel cyanogen compounds, etc. 5 Cassiterite may take 45 to 60 minutes E. S. Simpson, Chem. News, 99. 243, 1909 ; W. B. Giles, ib., 99. 1, 25, 1909 ; J. Gray, Journ. Chem. Met. Soc. S. Africa, 10. 312, 1910 ; H. Milou and R. Fouret, Int. Cong. App. Chem., 8. 373, 1912. 6 Nickel crucibles are recommended for the determination of metals precipitated by hydrogen sulphide in acid solutions tantalum, niobium, tin, etc. ; silver crucibles for the determination of silica, alumina, iron, manganese, cerium, etc. THE ANALYSIS OF GLAZES, GLASSES, ENAMELS, AND COLOURS. 267 projected from the melting mass on the under side of the lid travel to the centre of the lid, and drop back into the crucible. A single Bunsen's burner will heat a small charge to a dull red heat. This usually suffices for the decomposition. The mouth of the crucible is kept cool, and the fused salt does not creep over the edges of the crucible. The results are usually excellent. A similar remark applies to the fusion with sodium peroxide as described for chromite, 1 page 474. The heat required for the decomposition may be furnished by a reaction between the sodium peroxide and the organic matter of the substance 2 under investigation ; or organic matter purposely added. 3 When starch and organic matter are used the fusion does not usually proceed quietly, but is attended by slight explosions which project the fused mass on to the sides and lid of the crucible. Walton and Scholz 4 observed that the fusion is much quieter if zinc sulphide be substituted for the starch, and Parr 5 found that the addition of a small amount of potassium persulphate led to better fusion and more complete decomposition. Suppose a lead frit be under investigation. The operation is conducted as follows : A round-bottomed nickel crucible about 30 c.c. capacity and 4 cm. in diameter is charged with 8 grms. of powdered sodium peroxide; 0*5 grm. of the finely powdered (200's lawn) frit; (say) 1*2 grm. of finely powdered zinc sulphide ; and 0*3 grm. of potassium persulphate. 6 Each constituent must be quite dry, 7 and all thoroughly mixed by stirring with a glass rod. Partial mixing means partial decomposition. The rod is then brushed clean. The crucible is placed in a dish of cold water, taking care, of course, to keep the inside of the crucible dry. The cover is placed on the crucible a little to one side, so that a piece of lighted magnesium ribbon, 8 just over 1 cm. long, may be dropped into the crucible, and the cover placed in proper position immediately the mixture ignites. When the mass has cooled a little, about a minute after ignition, place the crucible in a clean evaporating basin, and add about 100 c.c. of cold water. When the violence of the reaction is over, the solution can be acidified with hydrochloric acid, and the silica determined by two evaporations in the usual manner. For comparison purposes, two samples of a frit treated by the standard process gave 30'06 and 30'04 per cent, of silica ; and two samples of the same frit, treated by the process just described, gave 29'98 and 30O6 per cent, of silica. The principal advantages of the method now under discussion are : (1) the vessel in which the decomposition occurs is much less attacked than 1 J. Darroch and C. A. Meiklejohn, Eng. Min. Journ., 82. 818, 1906; H. Angenot, Zeit. angew. Chem., 17. 1274, 1904. For opening copper pyrites by fusion with six times its weight of potassium persulphate, see L. Majewski, Kosmos, 35. 597, 1910. 2 S. W. Parr, Journ. Amer. Chem; Soc., 22. 646, 1900 ; 30. 764, 1908 (coal in a steel bomb) ; H. H. Pringsheim, Amer. Chem. Journ,, 31. 386, 1903 ; Ber., 37. 2155, 1904 ; 38. 2436, 1906 (halogens, arsenic, and phosphorus in organic compounds with a steel bomb). 3 F. vcn Konek, Zeit. angew. Chem., 17. 771, 1904 ; F. von Konek and A. Zohls, ib., 17. 1093, 1904 ; H. H. Pringsheim, ib., 17. 1454, 1904. 4 J. H. Walton and H. A. Scholz (Amer. Chem. Journ., 29. 771, 1908 ; Chem. News, 98. 61, 76, 1908) used zinc sulphide either alone or mixed with iron pyrites W. B. Pollard, ib., 98. 211, 1908. 5 S. W. Parr, Journ. Amer. Chem. Soc., 24. 167, 1902. 6 If lead be present, it will be found associated with the silica as lead sulphate. 7 The mixture is so easily "ignited" that if any moisture be present, the heat evolved by the reaction with sodium peroxide suffices for the decomposition of the silicate. 8 A piece of twine about 2 cm. long, soaked in alcohol, may be employed when the introduction of magnesia is not desired. There are many cases where the presence of zinc and sulphur is not particularly objectionable ; in other cases such additions would present insuper- able objections. 268 A TREATISE ON CHEMICAL ANALYSIS. usual thus, an 18-gram crucible lost nearly O34 grm. in 16 fusions; 1 and (2) the whole process occupies but a few minutes. The mixing, ignition, cooling, and dissolution of the melted mass can be done in about 5 minutes. The method can be advantageously used for clays, galena, 2 lead glazes, lead slags, monazite sand, etc. The decomposition with basic substances like chromite, franklinite, and bauxite is not satisfactory. If the presence of iron be not objectionable, the chromite and franklinite can be completely decomposed by substituting 2'0 grms. of iron pyrites and 0'3 grms. of magnesium powder in place of the 1-2 grms. of powdered zinc sulphide. If the iron be objectionable, some other method of decomposition must be used. A similar remark applies to the addition of zinc sulphide. 3 Reduction Process. In the case of glazes containing tin oxide, it is sometimes advisable to subject the finely divided and dried material to a preliminary heating in a reducing atmosphere, as recommended by Wells, 4 in order to con- vert the oxide to metallic tin soluble in hydrochloric acid. The powdered material may or may not be first digested in hydrochloric acid and dried. A thin layer of the dried material is spread on the bottom of a porcelain boat about 7 cm. long and 1 cm. broad. The boat is weighed, as usual, before and after the addition of the powder. 5 The boat is placed in a hard glass tube about 30 cm. long and 2 cm. wide drawn out at one end, as indicated at A, fig. 122. The wide end of the tube is connected by means of a perforated stopper with a wash- bottle and a Kipp's hydrogen apparatus ; the opposite end of the tube is allowed to dip in a beaker containing dilute hydrochloric acid (1 : 10). The hydrogen is passed through a wash-bottle 6 about two bubbles per second. When the air has been expelled from the tube, the gas jet 7 is lighted, and arranged so that about 15 cm. of the tube, in the neighbourhood of the boat, is heated to dull redness 8 for three or four hours. The boat is allowed to cool while the current of gas is still passing. 9 When cold, transfer the contents of the boat 10 to a 400-c.c. beaker, and treat the mass with 100 c.c. of hydrochloric acid with a few drops of nitric acid in order to convert the stannous into stannic chloride. Let the vessel 1 For the losses with direct fusions, see page 163. 2 The galena is itself oxidisable (combustible), and the proportion of zinc sulphide may be accordingly reduced. Thus the charge for galena may be : O'f> grm. powdered (200's lawn) ore ; 8'0 grms. of sodium peroxide ; 0'8 grm. zinc sulphide ; 0'3 grm. potassium persulphate. Sulphuric acid was used for neutralising the alkaline solution, and sufficient acid to make about 2 per cent, excess H 2 S0 4 was added. Sodium bisulphite was added to reduce the lead peroxide, and the solution boiled 5 minutes to get rid of the sulphur dioxide. The lead sulphate can then be determined gravimetrically, or volume trically (molybdate process). The trial experiments were quite satisfactory. 3 Unless some other sulphide be available. 4 J. S. Wells, School Mines Quart., 12. 295, 1891 ; Journ. Amer. Chem. Soc., 20. 687, 1898 ; Chem. News, 64. 294, 1891 ; M. W. lies, ib., 50. 194, 1884 ; 85. 179, 1902 ; A. Hilger and H.Haas, ib., 63. 195, 1891; Per., 23. 458, 1890; J. A. Miiller, Bull. Soc. Chim. (3), 25. 1004, 1901 ; Chem. News, 85. 147, 1902 ; W. Hampe, Chem. Ztg., u. 19, 1887 ; H. W. Rennie and W. H. Derrick, Journ. Soc. Chem. Ind., II. 662, 1892 ; G. L. Mackenzie, Trans. Inst. Min. Met., 13. 87, 1903. 5 The boat funnel of Stoltzenberg is convenient for filling boats with powder. 6 That shown in the diagram is J. Habermann's (Zeit. anal. Chem., 24. 79, 1883). 7 A Weston's cap on an ordinary Bunsen's burner is very convenient for tbjs purpose, as shown in the diagram. 8 The reduction of the tin oxide commences about 170 ; lead oxide about 310 W. Miiller, Pogg. Ann., 107. 136, 1869. H. Haas ( Ueber die quantitative Trennung dcs Zinn'sund Titan's, Erlangen, 1890) separates tin from titanium by calcination of the mixture in a current of hydrogen, whereby tin alone is reduced to the metal and is subsequently dissolved by treatment with hydrochloric acid. 9 Coal-gas may be used, but the objectionable sulphur, present in coal-gas, may form volatile sulphides with some of the constituents of the powder. 10 The boat should not be heated high enough to vitrify the powder. THE ANALYSIS OP GLAZES, GLASSES, ENAMELS, AND COLOURS. 269 stand in a warm place until the action has subsided. Boil 3 minutes. Dilute with an equal volume of hot water, and filter through a hot-water funnel (page 324). Ignite the filter paper and contents, and fuse 1 the residue with four or five times its weight of sodium carbonate and a gram of sodium nitrite in a platinum crucible. Take up the cold mass with hydrochloric acid and a drop of nitric acid. Mix the solution with that derived from the digestion of the "reduced" glaze with hydrochloric acid. Evaporate the solution to dry ness for the separation of the silica (page 167). Nitric acid is sometimes preferred to hydrochloric acid to avoid dealing with the sparingly soluble lead chloride. If nitric acid be used with tin and antimony glazes, metastannic and aritimdnic acids will contaminate the silica. 2 If tin and lead be associated with the regular silica, FIG. 122. Opening lead and tin silicates by a preliminary reduction. alumina, magnesia, etc., precipitate the tin and lead as sulphides, as described below. Fusion with Potassium Cyanide. If tin and phosphorus be present, the silica may be contaminated with a metastannic phosphate (possibly 2Sn0 2 .P 2 5 ). 3 In that case, the residue left after the ignition of the silica is treated with hydrofluoric acid to drive off the silica. After weighing, the residue is fused with at least three times its weight of pure potassium cyanide, free from sulphides, in a porcelain crucible. If the crucible be rotated and tapped while hot, the separate beads of tin will unite to form a larger bead. When cold, extract with water, filter, and wash the bead of metallic tin. The bead can be weighed as metal, or 1 It is not safe to assume that all the reducible oxides have been reduced to the metal and dissolved in the acids. A second treatment with the gas, or fusion of the residue as described in the text, is sometimes advisable. 2 In that case, drive off the silica with hydrofluoric acid, and take up the residue by fusion with sodium carbonate, etc. Add the ac,id solution to the main solution. 3 If arsenic be present, an insoluble metastanniQ arsenate soluble in hydrochloric acid may be formed. 270 A TREATISE ON CHEMICAL ANALYSIS. dissolved in hydrochloric acid, and added to the main solution. The difference between the weight of the tin calculated to Sn0 2 and the weight of the "silica residue " may be taken as phosphoric oxide P 2 5 . r ^ ne filtrate from the tin can be boiled under a good hood with hydrochloric acid until the fumes of the highly poisonous cyanogen compounds have been driven off. Evaporate the solution to a small volume, and precipitate the phosphorus in the usual manner (page 595). 1 The objection to this process turns on the fact that some tin may be lost owing to the formation of soluble alkali stannates. According to Bloxarn, 2 if the potassium cyanide contains sulphide, an insoluble tin sulphide or a soluble thiostannate may be formed. If tungsten be present, it will be found in the solution as alkaline tungstate. Fusion with Sodium Carbonate and Sulphur. According to Miller, 3 the hydrogen reduction for the determination of tin in, say, cassiterite, gives low results, and he prefers Rose's process by fusion with sulphur and sodium carbonate. This Hofman conducts in the following manner : Intimately mix 0*5 grm. of the sample with 3 grms. of a mixture of equal parts of sodium carbonate and sulphur 4 in a porcelain crucible (No. 1 Berlin), which, in turn, is placed in a larger porcelain crucible, and this again in a graphite crucible which has a layer of calcined fireclay spread on the bottom, so that the top of the crucible is nearly on a level with the top of the graphite crucible. 5 The whole is then heated in a crucible furnace for about 1-1 \ hours at a red heat. The cold mass is treated with hot water. The tin dissolves as sodium thiostannate along with some copper and iron. The latter will be precipitated by the addition of sodium sulphite to the deep brown liquid. 7 Filter, wash with water containing sodium sulphite in solution, and finally with water containing hydrogen sulphide in solution. The precipitate contains iron, copper, and lead. The solution contains arsenic, antimony, and tin. These can be separated as described below. The objection to this process turns on the fact that it is tedious and dirty, particularly if a re-fusion be necessary. Hydrofluoric Acid Process. The problem with enamels and coloured glazes is often very difficult, since many metals of both the second and third groups may be present. If tin, bismuth, antimony, or arsenic be present, the hydro- 1 H. Rose, Pogg. Ann., no. 425, 1870 ; F. Oettel, Chem. Ztg., 20. 19, 1896 ; J. A. Miiller, Bull. Soc. Chim. (3), 25. 1004, 1901. The potassium cyanide fusion can also be employed for separating the metals reduced by this agent W. H. Rennie and W. H. Derrick, Journ. Soc. Chem. Ind., II. 662, 1892 ; T. Moore, Chem News, 67. 267, 1893 ; H. Y. Lorara, Pro.\ Chem. Soc., 27. 60, 1912. 2 C. L. Bloxam, Journ. Chem. Soc., 18. 97, 1865. 3 E. H. Miller, Journ. Anal. App. Chem.., 6. 441, 1892 ; H. Rose, Ausfilhrliches Hand- buch der analytischen Chemie, Braunschweig, 2. 286, 1851 ; H. 0. Hofman, Berg. Hiltt. Ztg., 49. 342, 350, 357, 1890; Chem. Neivs, 62. 57, 1890; Tech. Quart., 3. 112, 261, 1890; F. Becker, Zeit. anal. Chem., 17. 185, 1878 ; J. Mitchell, Manual of Assaying, London, 481, 1881 ; J. F. C. Abelspies, Trans. Inst. Min. Met., 13. 99, 1903 ; F. W. Rennie and W. H. Derrick, Journ. Soc. Chem. Ind., n. 662, 1892 ; E. D Campbell and E. C. Champion, Ind. and Iron, 267, 1898; 0. Beck and H. Fischer, School Mines Quart., 20. 372, 1899; Chem. News, 80. 259, 1899 (comparison of methods); L. Medri and C. Gastaldi, Soil. chim. farm., 48. 893, 1910. 4 H. Rose's flux is a mixture of equal parts of sulphur and sodium carbonate. Chauvenet substituted potassium carbonate for the sodium carbonate. A. Froehde (Pogg. Ann., 119. 317, 1875) and E. Donath (Zeit. anal. Chem., 19. 23, 1880) prefer powdered sodium thiosulphate, which has previously been fused in order to remove the water. 5 The object is to cut off the supply of air, otherwise decomposition will be incomplete. 6 Gritty particles insoluble in water -show that the action was not complete. In that case, filter, wash, dry, and repeat the treatment with the residue. 7 Containing sodium polysulphide, which dissolves some iron and copper sulphides. Sodium sulphite changes sodium polysulphide to sodium monosulphide, in which iron and copper are practically insoluble. THE ANALYSIS OP GLAZES, GLASSES, ENAMELS, AND COLOURS. 271 chloric acid evaporation for silica may cause an appreciable loss by volatilisation of the chlorides. 1 It may then be advisable to make a special decomposition for metals precipitated by hydrogen sulphide in an acid solution. This is often done with hydrofluoric acid. 2 Two grains of the impalpable powder are treated with about 20 c.c. of concentrated hydrofluoric acid (40 per cent.) and an equal volume of hydrochloric acid or nitric acid in a platinum dish, and the solution taken nearly to dryness. 3 If the residue 4 does not dissolve easily in hydro- chloric acid, filter off the insoluble residue, and reduce the latter in coal-gas, as indicated above ; or fuse it with sodium carbonate. Take up the cold cake with hydrochloric acid, and add the solution to the main nitrate. Separate the metals of the hydrogen sulphide group as indicated below. 124. The Behaviour of Metals of the Hydrogen Sulphide Group in the Silica Determination. Volatilisation of the Chlorides. Care must be taken in boiling solutions evaporation for silica, etc. con- taining chlorides of arsenic, antimony, tin, bismuth, and mercury, since serious losses may occur by the volatilisation of the chlorides. Arsenious chloride, that is, arsenic trichloride, volatilises at 134 ; antimony trichloride at 223 ; and stannic chloride at 114. But these compounds volatilise at a much lower temperature in steam which arises when the aqueous solutions are boiled. (1) Tin. Concentrated solutions of stannic chloride in the presence of hydrochloric acid (20 per cent.) lost, after 20 minutes' boiling at 107, nearly OO014 grm. of SnCl 4 . A solution of stannous chloride SnCl 2 in hydrochloric acid can be evaporated to dryness without appreciable loss. Hence, the evapora- tion of stanniferous glazes, without arsenic and antimony, is best made in the presence of a reducing agent. (2) Antimony. A solution of antimony trichloride SbCl 3 in hydrochloric acid (20 per cent.) may be heated to 110 without serious loss, but an appreciable quantity is volatilised at higher temperatures. Antimony trichloride is not so volatile under these conditions as stannic chloride. Antimony peritachloride, in hydrochloric acid (20 per cent.) solution, may be evaporated to dryness with a negligibly small loss. If, however, a mixture of stannous chloride and antimony pentachloride be added, the latter will be " reduced " and the former " oxidised " to the more volatile chlorides. According to Hoffmann, the addition of, say, 12 grms. of potassium chloride 5 retards from the volatilisation of these chlorides and permits these solutions to be evaporated without appreciable loss. (3) Arsenic. Arsenious chloride AsCl 3 and arsenious salts in solutions containing hydrochloric acid are volatilised, during evaporation and boiling, in comparatively large quantities ; but the arsenic salts can be evaporated to dry- ness with no appreciable loss. Hence the evaporation of antimonical and arsenical solutions is best made in the presence of an oxidising agent. 6 (4) Mercury. When 30 c.c. of aqueous O'l to 1-0 per cent, solutions of mercuric chloride are distilled, Minozzi 7 found that when half the liquid had lV T. M. Drown and G. F. Eldri.lge, Tech. Quart., 5. 136, 1893. 2 H. N. Warren, Chem. News, 67. 16, 1983. ' 3 Watch the hydrofluoric acid for "lead " impurity. 4 A bluish residue indicates tungsten. 5 M. Hoffmann, Beitrdge zur Kenntnis der analytischen Chemie des Zinns, Antimons, und Arsens, Berlin, 10, 1911. 6 J. I. D. Hinds, Inter. Cong. App. Chem., 8. 227, 1912. 7 A. Minozzi, Boll. Chim. Farm., 43. 745, 1904 ; E. Esteve, Chem. Ztg., 35. 1152, 1911 ; P. Bohrisch and F. Kiirschner, Pharm. Ccntralkalle, 52. 1367, 1911. 272 A TREATISE ON CHEMICAL ANALYSIS. distilled over, the distillates contained from -00025 to 0'0022 grm. of mercuric chloride. The greater the concentration of the solution, the greater the loss by volatilisation. The error introduced into an analysis, under these conditions, may amount to 0'2 per cent. Similar results were obtained in the presence of hydrochloric and phosphoric acids, and sodium chloride. Separation of Insoluble Constituents with the Silica. The silica residue may be very complex. Insoluble tin phosphates, tin arsenates, antimony and bismuth compounds, 1 etc., may separate with the silica. The determination of the silica is therefore liable to error. In view of these difficulties, it is better to determine the members of the hydrogen sulphide group on a separate sample by, say, the hydrofluoric acid process of decomposition, page 270. With the portion in which the members of the hydrogen sulphide group are not being determined, losses during the evaporation for silica and after the silica has separated may be neglected. In the analysis of stanniferous slags, containing tin, in addition to the usual constituents of clays, Bailey 2 evaporates the solution from the sodium carbonate fusion to dryness with 20 c.c. of concen- trated nitric acid ; and boils the residue with 20 c.c. of concentrated hydro- chloric acid. 3 The insoluble residue of stannic oxide and silica is mixed with an equal volume of water ; two sticks of zinc about 2 J cm. long are placed in the mixture, and the whole is allowed to stand in a warm place for some time. Carefully scrape off any adherent spongy tin from the sticks of zinc ; filter and wash the tin and silica, and transfer the insoluble mixture back from the filter paper into the dish. Add 10 c.c. of concentrated hydrochloric acid and a few drops of nitric acid, 4 and warm the mixture until the silica appears white. Dilute the solution, filter, and wash. The silica and tungsten oxide will be found on the filter paper ; 5 the tin, etc., in the filtrate. Dott 6 separates stannic oxide from the silica by heating the mixture with three or four times its weight of hypophosphorous acid over a Bunsen's flame for about 30 minutes. The tin is converted into a stannous phosphate soluble in warm hydrochloric acid. The silica is not affected by this treatment, and can be filtered off. 125. The Theory of Precipitation by Hydrogen Sulphide. The terms "soluble" and "insoluble" are purely relative. With increasing refinements in the methods of measurement, the list of substances insoluble in a given solvent becomes shorter and shorter. Many substances once said to be " insoluble " in a given solvent are known to be appreciably soluble. It is all a question of delicacy of measurement. Hence, some leap beyond the domain of demonstrated fact, and say, " No substance is perfectly insoluble in water." The method of classifying certain elements into two groups those which form soluble and those which form insoluble sulphides in hydrochloric acid frequently conveys wrong ideas of the properties of the sulphides. The solubility of the sulphides depends upon the concentration of the acid. For instance, if hydrogen sulphide be passed into 5 c.c. of a solution of 2 grms. of tartar emetic 1 Also tungsten, niobium, and tantalum compounds, if these elements be present. 2 H. Bailey, Chem. News, 73. 88, 1896. 3 There is no danger of losing tin, because it is here insoluble. 4 Just sufficient to oxidise the tungsten, if any be present. 5 Separate by the ammonia process, page 409. 6 D. B. Dott, Pharm. Journ., 81. 585, 1908. THE ANALYSIS OF GLAZES, GLASSES, ENAMELS, AND COLOURS. 273 in 15 c.c. of hydrochloric acid (sp. gr. 1'175) and 85 c.c. of water, 1 antimony sulphide will be precipitated, but not if 12 c.c. of hydrochloric acid with no further addition of water had been employed. In the one case, the reaction may be represented 2SbCl 3 + 3H 2 S = Sb 2 S 3 + 6HC1 ; and, in the second case, Sb 2 S 3 + 6HC1 = 3H 2 S + 2SbCl 3 . In other \vords, the antimony sulphide, in the second case, is decomposed as fast as it is formed. Similarly, no lead will be precipitated by hydrogen sulphide from a solution containing over 3 per cent. of hydrochloric acid (HC1), and if the solution has 2*5 per cent, of acid, the lead will be imperfectly precipitated : part will be precipitated, and a certain proportion will be decomposed as fast as it is formed, and thus remain in solution. Similarly a 5 per cent, boiling solution of hydrochloric acid will prevent the precipitation of cadmium. 2 If, then, a metallic sulphide be treated with hydrochloric acid, hydrogen sulphide and a metallic chloride will be formed : RS + 2HC1^^RC1 2 + H 2 S. Conversely, when a metallic chloride in aqueous solution is treated with hydrogen sulphide, the metallic sulphide and hydrochloric acid are produced : RC1 + HS^rRS + 2HC1. Hydrochloric acid thus accumulates in the solution as the action goes on. After the hydrochloric acid has attained a certain concentration, if any more sulphide is produced, the excess of sulphide will be decomposed by the acid. There are thus two simultaneous reactions: (1) formation of sulphide and hydrochloric acid ; and (2) formation of chloride and hydrogen sulphide. In further illustra- tion, if a current of hydrogen sulphide be passed through a saturated solution of zinc chloride, part of the metal is precipitated ; but when the hydrochloric acid has attained a certain concentration, the action apparently ceases because the reverse change sets in. Hence, the precipitation of the zinc as sulphide will be incomplete. 3 Similar remarks apply for the other metals. We can get more precise ideas than this. Take the case of lead chloride : PbCl 2 + H 2 S^=PbS + 2HC1. When equilibrium is established, 4 the solution contains lead chloride, hydrogen sulphide, and hydrochloric acid. If bracketed chemical symbols be employed to represent the concentration (gram-molecules per litre) of the respective com- 1 The solution of tartar emetic will keep a couple of hours. 2 M. Martin, Journ. praU. Chem. (1), 67. 374, 1856 ; C. C. Hutchinson, Phil. Mag. (5), 8. 433, 1879 ; Chem. News, 41. 28, 1880. In separating cadmium from zinc by hydrogen sulphide in an acid solution, some prefer to precipitate most of the zinc with the cadmium, and then digest the precipitate in a solution containing about 5 '5 per cent, of hydrochloric acid without heating, but with vigorous agitation. It is claimed that the zinc passes into solution, and the excess of hydrogen sulphide in the solution prevents the dissolution of the cadmium. 3 M. Baubigny, Compt. Rend., 107. 1148, 1888 ; G. Ohesneau, ib., III. 269, 1890. In the case of zinc, the accumulation of hydrochloric acid can be prevented by the use of certain organic salts ammonium or sodium acetate, sodium formate, etc. These substances react with the hydrochloric acid, producing sodium chloride and an acetate (or formate, etc.). This is a very convenient way of substituting a weak acid say, acetic acid or formic acid for a strong acid hydrochloric acid. The solubility of the zinc sulphide in, say, acetic acid is so small that in- appreciable amounts of zinc remain in solution, although iron, nickel, cobalt, and manganese sulphides are dissolved by the acid. * 4 At an early stage in the reaction between hydrogen sulphide and lead chloride, lead thio- chloride probably PbS.PbCl 2 appears to be formed, since a brick-red precipitate of this com- pound sometimes separates when hydrogen sulphide is passed into a solution of lead chloride in hydrochloric acid. E. H. E. Reinsch, Journ. prakt. Chem. (2), 13. 130, 1876 ; V. Lehner, Journ. Amer. Chem. Soc., 23. 680, 1901 ; F. Parmentier, Com.pt. Rend., 114. 298, 1892. 18 274 A TREATISE ON CHEMICAL ANALYSIS. pounds in solution, the law of mass action * requires that the product of the con- centration of the given chloride and concentration of the hydrogen sulphide, divided by the square of the concentration of the hydrogen chloride, shall always have the same value. 2 In symbols, for equilibrium, we have [PbCU x [H 2 S] m 2 = Constant . . (1) This agrees with facts, and when the phenomenon is described in this way, it is easy to see that if the concentration of the acid be increased, and the concentra- tion of the hydrogen sulphide be constant, the amount of lead chloride which remains in a given solution (i.e. escapes precipitation) must increase in order to keep the numerical value of the ratio constant. Conversely, if it be desired to keep the amount of lead chloride in the solution as low as possible, it is necessary to keep the concentration of the acid down to a minimum value. 3 The concen- tration of the hydrogen sulphide in solution is practically constant (0'0073 gram- molecule per litre at 20) when the gas is passing through the solution. If the concentration of the hydrogen sulphide in solution were large, and the concentra- tion of the metallic chloride in- solution small, a very small excess of acid would not suffice to keep the metals in solution. It will be observed, however, that the concentration of the hydrogen sulphide under ordinary conditions is small. In consequence, a comparatively small amount of acid is sufficient to prevent the separation of sulphides of zinc, iron, and manganese. This may be expressed another way : if the solubility of the hydrogen sulphide had been greater than it is, some of the metals zinc, iron, nickel, etc. would have been included in the " hydrogen sulphide " group ; and, conversely, had the solubility of the hydrogen sulphide been less than it is, some of the present members of the " hydrogen sulphide " group would not have been there for instance, stannous tin, lead, cadmium. These deductions have been experimentally realised by Bruni and Padoa. 4 By causing the hydrogen sulphide to react under pressure, the solu- bility of the hydrogen sulphide in the liquid was augmented, and iron, nickel, and cobalt were precipitated ; by working under diminished pressure, the solubility of the hydrogen sulphide in the acid liquid was reduced, and cadmium was not precipitated under conditions where otherwise the sulphide would have separated. Under ordinary conditions, the solubility 5 of the precipitated sulphides in dilute hydrochloric acid, starting with the least soluble, 6 is approximately as follows reading from above downwards : Molybdenum Platinum Gold Arsenic Silver Copper Antimony Bismuth Stannic tin Mercury Cadmium Lead Stannous tin Zinc Titanium Iron Nickel Cobalt Manganese 1 J. W. Mellor, Chemical Statics and Dynamics, London, 156, 1904. 2 Neglecting disturbances due to the presence of foreign substances in the solution. 3 This, of course, is limited by the necessity for keeping the zinc in solution, when separat- ing lead and zinc by this method. 4 G. Bruni and M. Padoa, Atli Accad. Lincei (5), 14. ii. 525, 1905. 5 The term "solubility" is here understood to refer to the amount of the element which remains in solution, or escapes precipitation, when the hydrogen sulphide is passed through an acid solution of a given concentration. The order of solubility will be the same as if the numerical values of the constants of the series indicated in equation (1) above were arranged in ascending order. 6 The order is only approximate, and varies with the strength of the acid. For the solubility of the sulphides in water, see 0. Weigel, Zeit. phys. Chem., 58. 293, 1907 ; W. Bottger, ib. t THE ANALYSIS OF GLAZES, GLASSES, ENAMELS, AND COLOURS. 275 Elements wide apart in the list can be easily separated by hydrogen sulphide in acid solution, but elements close together require a very careful adjustment of the amount of acid in solution before satisfactory separations can be made. For instance, the separation of cadmium or lead from zinc by means of hydrogen sulphide is only satisfactory when the concentration of the acid is very carefully adjusted. If too much acid be present, cadmium or lead will be imperfectly precipitated; while if too little acid be present, zinc will be precipitated with the cadmium or lead. 1 Hence, no sharp line of demarcation can be drawn between metals precipitated and metals not precipitated by hydrogen sulphide in acid solution. All depends upon the concentration of the acid. 2 This is con- veniently adjusted so that the lead and .stannous sulphides are precipitated, while zinc sulphide remains in solution. A solution containing 4 c.c. of hydro- chloric acid (sp. gr. 1*12) per 100 c.c. will serve this purpose. 3 If more acid be present, there will be a danger of incomplete precipitation of stannous tin and lead ; if less than this amount of acid be present, some zinc, iron, nickel, cobalt, or manganese may be precipitated. The adjustment of the acid cannot be perfect, but it can generally be made so that inappreci- able quantities of the sulphides to be precipitated remain in solution. This is illustrated by fig. 123, which shows the relation between the amount of antimony chloride which remains in solution in the presence of hydrochloric acid and the concentration of this acid when the solution is saturated with hydrogen sulphide. The proximity of the curve to the lower horizontal axis 2-0, Grms HCl per 100 grm.Sol. FIG. 123. Effect of hydrochloric acid on the solubility of antimony sulphide. shows that for 20 grms. and less hydrochloric acid per 100 grms. of solution, very little antimony will escape precipitation. Filtration. In washing sulphide precipitates, say, copper sulphide in acid solution, with hydrogen sulphide water, a colourless filtrate is usually obtained. As the concentration of the acid in the mother liquid diminishes by washing, the filtrate sometimes becomes turbid, and, in the case of copper, the filtrate may be tinted green. 4 There are several distinct actions going on. First, the 46. 531, 1903 ; M. Hoffmann, Beitrdge zur Kenntnis der analytischen Chemie des Zinns, Antimons, und Arsens, Berlin, 49, 1911. 1 A. W. Hofmann (Liebig's Ann., 115. 286, 1860 ; Journ. Chem. Soe. 13. 78, 1860) separates copper and cadmium sulphides with sulphuric acid (1 : 5) the latter alone dissolves. 2 L. Loviton (Journ. Pharm. Chem. (5), 17. 361, 1888 ; Zeit. anal. Chem., 29. 345, 1890) has devised a method for separating antimony and tin based upon the solubility of the sulphides in hydrochloric acid of different concentrations; and E. Neher's process (Zeit. anal. Chem., 32. 45, 1893) for the separation of arsenic, antimony, and tin is based on the same property. A. A. Abel and F. Field, Journ. Chem. Soc., 14. 290, 1862 ; F. Field, Chen,. News, 3. 114, 1861 ; A. H. Low, Journ. Amer. Chem. Soc., 28. 1715, 1906 ; 0. Kohler, Arch. Pharm. (3), 27. 406, 1889 ; G. Panajotow, Ber., 42. 1296, 1909 ; E. Lesser, Ueber einige TrennungsundBestimmungs- Methoden des Arsens, des Antimons, unddesZinns, Berlin, 1886 ; W. R. Lang and C. M. Carson, Journ. Soc. Chem. Ind., 21. 1018, 1902 ; with J. C. Mackintosh, ib., 2l r 748, 1902 ; J. and H. S. Pattinson, ib., II. 211, 1898 ; F. Kietreiber, Osterr. Chem. Ztg. (2), 13. 146, 1910. 3 A. A. Noyes and W. C. Bray, Journ. Amer. Chem. Soc., 29. 137, 1907 ; Tech. Quart., 19. 191, 1906. 4 The filtrate becomes darker in colour, and finally flakes of copper sulphide separate from the filtrate. 276 A TREATISE ON CHEMICAL ANALYSIS. sulphide is oxidised to sulphate x ; second, the colloidal sulphides may be defloc- culated (page 96) ; and, third, if water alone be used in the washing, the dilution of the mother liquid may lead to the precipitation of some of the elements belonging to the next group, say, zinc sulphide, and so contaminate the precipitate being washed. Hence, the washing liquid should be kept acidified. 2 126. The Separation of the Metals precipitated by Hydrogen Sulphide in Acid Solutions. In the extreme case, suppose that a qualitative examination shows that the filtrate from the silica contains antimony, arsenic, 3 lead, bismuth, cadmium, tin, and copper, as well as alumina, etc. Adjust the solution so that it contains 10 c.c. of hydrochloric acid (sp. gr. 1*12) per 100 c.c. Heat the solution in an Erlenmeyer's flask at 70 to 80, and pass hydrogen sulphide through the hot solution 4 for about an hour while the temperature of the solution is main- tained. 5 Let the mixture cool. When cold, add one and a half times its own bulk of water (100 c.c. becomes 250 c.c.), 6 and saturate in the cold with hydrogen sulphide passing in a slow stream for about 15 minutes. Cork the flask and let the solution stand two or three hours. Filter and wash the precipitate with a dilute acid solution containing 20 c.c. hydrochloric acid (sp. gr. 1'12) per litre and saturated with hydrogen sulphide. The precipitate may contain lead, bismuth, copper, cadmium, arsenic, anti- mony, and tin sulphides. The precipitate will generally be practically free from 1 Hence, the filtration should be conducted as rapidly as possible, and the filter paper with the precipitate should be kept filled with the hydrogen sulphide wash-water. 2 G. Larsen, Zeit. anal. Chem., 17. 312, 1878; E. Berglund, ib., 22. 184, 1883; W. Dederichs, Pharm. Ztg., 44. 178, 1899. Mineral acids are objectionable if the filter paper is to be afterwards dried, because the acid is concentrated on the paper during the drying, and the paper is attacked. The paper then readily disintegrates. A dilute solution of acetic acid, saturated with hydrogen sulphide, gives good results. 3 According to F. Wohler, more or less zinc sulphide is precipitated in the presence of arsenic acid in comparatively strongly acid solutions, but not if arsenious acid be present. For this reason, and the reason stated in the next footnote, it is well to reduce the arsenic to arseni- ous salts, if arsenic salts be present. 4 Elements at the upper end of the series, page 274, are precipitated more or less imperfectly in the cold. Arsenic sulphide, for instance, continues separating a long time after the solu- tion is saturated, hence the current of gas is continued an hour longer. The sulphides which separate from a hot solution can be filtered and washed more easily than precipitates formed in cold solutions. Arsenious acid H 3 As0 3 reacts at once with hydrogen sulphide, forming As 2 S 3 ; with arsenic acid H 2 As0 4 the formation of As 2 S 3 is the joint effect of three consecutive re- actions : (1) the formation of thioarsenic acid H 3 AsS0 3 ; this (2) slowly decomposes into sulphur and arsenious acid H 3 As0 3 ; this latter (3) reacts with the hydrogen sulphide, as indicated above. Hence, the initial and end products are represented by the equation : 2H 3 As0 4 H-5H 2 S = As 2 S 3 + S 2 + 8H 2 0. One object of the boiling is to accelerate the very slow decomposition of the thioarsenic acid. Hence, before passing the hydrogen sulphide, some prefer to reduce the arsenic acid to arsenious acid with sulphur dioxide, or by warming the solution with a mixture of hydro- chloric acid and hydriodic acid (or potassium or ammonium iodide) sulphur dioxide is not recommended as a reducing agent L. L. de Koninck, Bull. Soc. Chim. Belg., 23. 88, 1909. There is also a very slow side reaction 2H 3 AsO 4 + 5H 2 S = As 2 S 5 + 8H 2 particularly in cold, feebly acid solutions (J. P. Bouquet and S. Cloiz, Ann. Chim. Phys. (3), 13. 44, 1845 ; B. Brauner and F. Tomiczek, Monats. Chem., 8. 607, 1887 ; L. R. W. McCay, Journ. Amer. Chem. Soc., 24. 661, 1902; Zeit. anal. Chem., 27. 632, 1888; F. Neher, ib., 32. 45, 1893; R. Bunsen, Liebig's Ann., 192. 305, 1878 ; H. Rose, Pogg. Ann., 107. 186, 1859). 5 The current of gas passes at the rate of about two bubbles per second. 6 If sulphuric acid be used in place of hydrochloric acid, use 375 c.c. of sulphuric acid (sp. gr. 1'84) with the same dilutions. The greater acidity of the boiling solution prevents the separation of titanium hydroxide during the boiling. The acid is also useful in preventing the separation of bismuth and antimony oxychlorides. The great dilution required for tiie separa- tion of lead and stannous tin would lead to the precipitation of the oxychlorides in question. THE ANALYSIS OF GLAZES, GLASSES, ENAMELS, AND COLOURS. 277 aluminium, iron, zinc, nickel, cobalt, and manganese. The filtrate is used for the determination of the alumina, etc. (page 177). 1 Precipitation in Pressure Flasks. When molybdenum, platinum, gold, or selenium are present, the solution to be treated with hydrogen sulphide is placed in a thick-walled mineral water bottle (or " pressure flask "), 2 heated nearly to boil- ing, and saturated with hydrogen sulphide. The solution is cooled and saturated, in the cold, with hydrogen sulphide. Close the bottle with a cork and wire, or screw the stopper down as at D, fig. 124. Heat the bottle in a vessel of water for about an hour. The pressure bottle should not rest directly on the bottom of the water bath, but rather be suspended by a wire (7, fig. 124, so that it is immersed up to the neck in the cold water. The water bath is gradually heated until the water boils. Boiling water is added to the bath, if necessary, as the water evaporates. The new water should not be poured directly on to the pressure bottle. At the end of the experiment, the water bath is allowed to cool ; the bottle is then removed from the bath, and opened. Fig. 124 shows a Linter's "pressure flask " A ready for lowering into the bath B. The bottle can be surrounded with a stout wire gauze so that no damage may be FIG. 124. Linter's "pressure flask." done if it bursts. Molybdenum, platinum, gold, selenium (also tellurium), if present, will be found with the precipitated sulphides of tin, arsenic, antimony, lead, cadmium, etc. 127. The Separation of Tin, Arsenic, and Antimony from the remaining Metals. The precipitate is now treated with a concentrated aqueous solution of sodium monosulphide 3 in a covered beaker heated to about 70. The lead, 1 The hydrogen sulphide must be boiled off from the filtrate, and the free sulphur removed by filtration. The iron is oxidised with a few drops of nitric acid, or hydrogen peroxide, or bromine water. - Pressure flasks Linter's, Salamon's, etc. are made specially for the purpose. A. Gawalovski, Zeit. anal. Chem., 22. 526, 1883 ; F. Allihn, ib., 23. 406, 1884 ; H. Rempel, .tfer. , 18. 621, 1885 ; A. Eiloart, Chem. News, 55. 148, 1887. See also page 412. 3 SODIUM MONOSULPHIDE SOLUTION. This reagent is prepared as follows : Dissolve 333 grms. of pure sodium hydroxide (made from the metal) in a litre of air-free water. Pour the solution into a flask, and pass a rapid stream of washed hydrogen sulphide into the solution through a wide glass tube (1 cm. bore) fitted into a double-bored stopper, so as to protect the contents of the flask from air as much as possible. The object of the wide delivery tube is to avoid choking the tube with the separated sulphide. There is an increase in the volume of the solution such that 1000 c.c. becomes 1218 c.c. When the solution is saturated, the pale yellow liquid may be poured into small glass-stoppered bottles and sealed with paraffin ; or the solution may be evaporated in a platinum or porcelain dish until a film of crystals begins to 278 A TREATISE ON CHEMICAL ANALYSIS. copper, 1 cadmium, 2 and bismuth 3 present remain undissolved, while the arsenic, antimony, and tin pass into solution. 4 The precipitate is washed with the sodium sulphide solution. The arsenic, antimony, and tin in the solution are separated by the methods indicated below ; the precipitate is dissolved in hydrochloric acid, and the lead, bismuth, copper, and cadmium are also separated one by one, as indicated later on. If one or more of the elements in question be absent, the corresponding steps are omitted. Short cuts are also advisable in special cases. It must also be added that in many technical analyses, it is not usual to work through a complex series of separations with one portion of a given sample, but several separate portions are taken and one group of constituents in each portion is determined by special methods of isolation. In a long series of separations, some of the stages may be more suited for one element than for another, and a process of analysis is selected which gives the best average for all the constituents to be determined, and not the one most suited for any particular constituent. When, however, a separate portion of the original sample is taken for the determination of each constituent, a method of separation specially favourable for that con- stituent can be employed. Theoretical. Freshly precipitated arsenic, antimonic, and stannic sulphides are quickly dissolved by sodium sulphide ; 5 stannous sulphide is but slowly dissolved. 6 Soluble alkaline thioarsenate AsS(SNa) 3 ; thioantimoniate SbS(SNa) 3 ; and thiostannate SnS(SNa) 3 appear to be formed ; while arsenious and antimonious sulphides form the corresponding As(SNa) 3 and Sb(SNa) 3 respectively. 7 form on the surface, and the hot liquid bottled and sealed. The liquids may deposit crystals of sodium monosulphide Na 2 S. 9H 2 on standing or cooling. The bottled solutions keep indefinitely when protected from the atmospheric air. A small quantity of black iron, nickel or silver sulphide, may settle on the bottoms of the bottles on standing. For use, the solution is diluted to a specific gravity 1'14. Sodium hydroxide pure by alcohol does not give so satisfactory a solution as that prepared from the metal, since a solution prepared from the former will be coloured with colloidal sulphides, which only separate after long standing. According to E. Prothiere and A. Revaud (Journ. Pharm. Chim., 16. 484, 1902), a layer of almond or olive oil on the surface of a solution of the sulphide cuts off the air without forming deposits or an emulsion. The solution then keeps indefinitely. For the action of polysulphides see H. SchifF, Liebig's Ann., 115. 68, 1860. 1 Freshly precipitated copper sulphide is appreciably soluble in colourless ammonium sulphide, and still more soluble in yellow ammonium sulphide C. L. Bloxam, Journ. Chem. Soc., 18. 94, 1865. 2 A. Ditte (Compt. Rend., 85. 402, 1887 ; Chem. News, 36. 109, 1877) says that cadmium sulphide is appreciably soluble in ammonium sulphide, but E. Donath and J. Mayerhofer (Zeit. anal. Chem., 2O. 384, 1881) state that this is not the case. H. Fresenius (Zeit. anal. Chem., 20. 26, 1681) agrees with Ditte. G. Vortmann, Monats. Chem., I. 952, 1880; E. Zettnow, Pogg. Ann., 130. 328, 1867. 3 T. B. Stillmann ( Journ. Amer. Chem. Soc., 18. 683, 1896) showed that if a solution con- taining bismuth be made alkaline with sodium hydroxide, and then heated with an excess of sodium sulphide, a considerable amount of bismuth remains in solution; but G. C. Stone (ib., 18. 1091, 1896) pointed out that if the bismuth sulphide be first precipitated from an acid solution, it is not dissolved by subsequent treatment with an alkaline sulphide. J. Knox, Journ. Chem Soc., 95. 1760, 1909. 4 Molybdenum, gold, platinum, and selenium, if present, may be partly dissolved with the arsenic, etc., and partly retained by the insoluble sulphides. 5 R. Bunsen, Liebig's Ann., 192. 320, 1878 ; H. Thiele, ib., 265. 65, 1891 ; B. Brauner and F. Tomicxek, Monats. Chem., 8. 607, 1887 ; F. Neher, Zeit. anal. Chem., 32. 45, 1893; L. R. W. M'Cay, ib., 27. 632, 1888; 34. 725, 1895; R. Fresenius, ib., I. 192, 1862 ; J. J. Berzelius, Pogg. Ann., 7. 1, 1826 ; C. Rammelsberg, ib., 52. 191, 1841 ; C. F. Nilson, Journ. prakt. Chem. (2), 14. 149, 1877 ; (2), 19. 170, 1879. 6 Hence, some use a little nitric acid with the hydrochloric acid in order to oxidise the stannous sulphide and accelerate the rate of solution. 7 Gold and platinum sulphides, in the same group as arsenic sulphide, dissolve with difficulty in the alkaline sulphide, and thus form an intermediate link between sulphides soluble and THE ANALYSTS OF GLAZES, GLASSES, ENAMELS, AND COLOURS. 279 Sodium monosulphide, in aqueous solution, is partly hydrolysed or decom- posed by water, forming sodium hydroxide and hydrosulphide : Na 2 S + H 2 O^NaOH + NaHS. The amount of hydrolysis depends upon the concentration of the solution. With dilute solutions, say y^N-solution, 86'4 per cent of the Na 2 S is hydrolysed, 1 but with increasing concentrations the amount of hydrolysis is diminished. Solutions of sulphur in the alkaline sulphides are much less hydrolysed than the monosulphide. This question of hydrolysis is important, because the products of hydrolysis may attack sulphides 2 which would otherwise be insoluble. Aqueous solutions of ammonium monosulphide are more readily hydrolysed than sodium monosulphide ; and, in consequence, sodium monosulphide is preferable to the ammonium salt, particularly in the presence of copper sulphide, which is slightly soluble in ammonium monosulphide. On the other hand, the ammonium salt is nearly always used if mercury salts be present, because mercuric sulphide is insoluble in ammonium monosulphide, 3 but readily soluble in sodium or potassium monosulphides owing to the formation of a tnio-salt : S = Hg(SNa) 2 . The mercuric sulphide is soluble in ammonium sulphide if a little potassium or sodium hydroxide be present; 4 and it is also more readily soluble in potassium or sodium sulphide if a little of the alkaline hydroxide be also present. 5 The latter reaction enables mercury sulphide to be separated from the sulphides of lead, silver, bismuth, and copper. sulphides insoluble in the alkaline sulphide. According to J. Riban (Compt. Rend., 85. 283, 1877 ; Bull. Soc. Chim. (2), 28. 241, 1877), platinum sulphide PtS 2 alone is practically insoluble in ammonium and sodium mono- and polysulphides. In the presence of arsenic, antimony, and tin sulphides, appreciable quantities of platinum and gold sulphides pass into solution, and this the more the greater the quantity of those elements present. The separation of gold and platinum, as well as molybdenum and selenium, is not therefore satisfactory by this process, since part will be found in the solution, and part with the precipitate. 1 F. W. Kiister and E. Heberlein, Zeit. anorg. Chem., 43. 53, 1905 ; J. Walker, Zeit. phys. Chem., 32. 137, 1900 ; H. Rose, Pogg. Ann., 55. 533, 1842; C. F. Sammet, An Investigation on the Production, Precipitation, and Migration of Colloids, Boston, Mass., 1903. 2 This, for instance, is the case with copper sulphide in ammonium monosulphide, and particularly with ammonium polysulphide. Copper and lead are also attacked by sodium polysulphides, apparently not altogether because of the hydrolysis. The mode of action of the polysulphides has not yet been made quite clear. According to V. Hassreidter (Zeit. angew. Chem., 18. 292, 1905), the copper can be recovered from a solution of the sulphide in, say, sodium polysulphide by boiling the solution with the cautious addition of sodium sulphite until colourless ; sodium monosulphide and thiosulphate are formed, which have no solvent action on copper sulphide. 3 But slightly soluble in ammonium polysulphide A. Glaus, Liebig's Ann., 129. 209, 1863; Chem. News, 9. 145, 1864; C. Barfoed, Journ. prakt. Chem. (1), 93. 230, 1865; Zeit. anal. Chem., 3. 139, 1864 ; 4. 435, 1865. 4 J. Volhard, Liebig's Ann., 255. 252, 1891. 5 A. Ditte, Compt, Rend., 98. 1271, 1884 ; K. Polstorff and C. Billow, Arch. Pharm., 229. 292, 1891. CHAPTER XXI. THE DETERMINATION OF ARSENIC. 128. The Separation of Arsenic from Antimony and Tin by Distillation. THE distillation process, proposed by Schneider and by Fyfe, depends upon the volatility of arsenious chloride. 1 If arsenic compounds be reduced to arsenious compounds in hydrochloric solution by ferrous sulphate, Fischer 2 showed that the arsenic may be quantitatively separated as arsenious chloride from solutions containing antimony and tin. 3 By this means arsenic may also be separated from the other metals of the hydrogen sulphide group. 4 Details of the modified Fischer's process are as follows : 1 See page 271. 2 E. Fischer, Ber., 13. 1778, 1880; Liebig's Ann., 208. 182, 1881 ; A. Classen and R. Ludwig, Ber., 18. ilO, 1885 ; F. Hufschmidt, ib , 17. 2245, 1884 ; Chem. News, 50. 269, 1884 ; W. Odling, ib. t 8. 27, 1863 ; T. Gibb, ib., 45. 218, 1882 ; F. C. Schneider, Pogg. Ann., 85. 433, 1852 ; A. Fyfe, Phil. Mag. (4), 2. 487, 1851 ; Journ. prakt. Chem. (1), 55. 103, 1852 ; P. Jannasch and E. Heimann, ib. (2), 74. 473, 488, 1906 ; 0. Ducru, Compt. Rend., 127. 227, 1898; Chem. News, 78. 73, 1898; A. Hollard and L. Bertiaux, ib., Si. 242, 1900; Bull. Soc. Chim. (3), 23. 300, 1900 ; A. Hollard, ib. (3), 23. 292, 1900 ; Chem. News, 81. 258, 1900 ; E. Azzarello, Qazz. Chim. ItaL, 39. ii, 450, 1910 ; J. Clark, Journ. Soc. Chem. Ind., 6. 353, 1887 ; A. Gibb, ib., 20. 184, 1901 ; J. E. Stead, ib., 14. 444, 1895; Rieckher, Zeit. anal. Chem., 9. 516, 1870; J. A. Kaiser, ib., 14. 250, 1875; H. Beckurts, Arch. Pharm., 222. 653, 1884 ; A. Kleine, Stahl Eisen. 24. 248, 1894 ; J. Clark, Journ. Anal. App. Chem., 6. 277, 1892 ; E. Ruff and F. Lehniann, Arch. Pharm., 250. 382, 1912 ; H. Hagen (Pharm. Centr. (3), 22. 169, 1882) claims that the use of ferrous sulphate is due to himself and not to Fischer. 3 Ferrous chloride, ferrous sulphate, ferrous ammonium sulphate (Classen and Ludwig), or cuprous oxide may be used. The presence of large quantities of the ferrous salt seems to prevent the well-known volatilisation of mercuric chloride with water vapour. Oxidising agents nitric acid should be absent, since they will oxidise the ferrous salt and make it inert. If present, nitric acid should be removed by evaporation with sulphuric acid. A little sulphuric acid does no harm. 0. Piloty and A. Stock (Ber., 30. 1649, 1897; G. T. Morgan, Journ. Chem. Soc., 85. 1001, 1904) reduce by passing a stream of hydrogen sulphide through the boiling solution. If the arsenic be all present as arsenious salt, the ferrous salt is not needed. F, A. Gooch and E. W. Banner (Amer. J. Science (3), 42. 308, 1891 ; F. A. Gooch and B. Hodge, Zeit. anorg. Chem., 6. 268, 1894 ; Chem. News, 70. 23, 1894) reduce by means of potassium iodide and hydrochloric acid: F. A. Gooch and M. A. Phelps (Zeit. anorg. Chem., 7. 123, 1894) used potassium bromide and hydrochloric acid ; M. Rohmer (Ber., 34. 33, 1565, 1901) reduced with sulphur dioxide, and H. B. Bishop (Journ. Amer. Chem. Soc., 28. 178, 1906) passed sulphur dioxide through the solution during the distillation and omitted the ferrous salt. This process gives good results, and can be used when the presence of iron is objectionable. C. Friedheim and P. Michaelis (Ber., 28. 1414, 1895 ; H. Cautoniand J. Chautems, Arehiv Sci. Phys. Nat. Geneve (4), 19. 364, 1905 ; L. Moser and F. Perjatel, Monats. Chem., 33. 779, 1912 ; S. W. Collins, Analyst, 37. 229, 1912) reduce with methyl alcohol ; P. Jannasch and T. Seidel (Ber., 43. 1218, 1910 ; P. Jannasch and E. Heimann, Journ. prakt. Chem. (2), 74. 437, 1906) reduce with 3 grins, of hydrazine sulphate, or hydrochloiide, with excellent results. The time required is about an hour. This process can be recommended when the presence of iron is objectionable, and the hydrazine salt is available. M. Hoffmann (Beitrdge zur Kenntnis der analytischen Chemie des Zinns, Antimons,-und Arsens, Berlin, 15, 1911) used potassium iodide with a current of hydrogen chloride gas. 4 F. Wohler, Die Mineral- Analyse in Beispeilen, Gottingen, 223, 1861. 280 THE DETERMINATION OP ARSENIC. 28l Oxidation of Arsenious to Arsenic Salts. The mixed sulphides obtained in the preceding operation are transferred to a beaker. Add, say, 2 to 3 grms. of potassium chlorate, and pour in cautiously 40 c.c. of dilute hydrochloric acid (30 c.c. of concentrated acid, 10 c.c. water). Cover the beaker with a clock-glass, and heat the mixture on a water bath until decomposition is complete, and all the chlorine oxides have been expelled. The object of the oxidation is to transform the arsenious salts to arsenic salts. The former are volatile with steam, while arsenic acid is not volatile. Preparation for the Distillation. Transfer the contents of the beaker to a 500-600 c.c. distillation flask. Wash the beaker with concentrated hydrochloric acid. Add, say, 25 grms. of ferrous ammonium sulphate and make the solution up to about 200 c.c. with concentrated hydrochloric acid. 1 Incline the flask at an angle of about 45-60, as illustrated in the diagram, fig. 125. Connect this FIG. 125. Apparatus for the separation of arsenic by distillation. flask with an Erlenmeyer's flask (3 or 4 litres) containing about a litre of concentrated fuming hydrochloric acid, saturated with sodium or ammonium chloride. 2 This flask is fitted with a delivery tube and a dropping funnel containing concentrated sulphuric acid. When the latter is dropped into the hydrochloric acid, hydrogen chloride is given off. The rate of evolution of the gas is determined by the rate at which sulphuric acid is dropped from the funnel. The neck of the distillation flask is connected with a condenser which has a tube dipping into water in a receiver (200 c.c.). 3 The receiver is 1 If antimony is to be afterwards determined by the distillation process (page 282), add 50 c.c. of a solution of zinc chloride made by saturating concentrated hydrochloric acid with metallic zinc, and evaporating till the solution boils at 108. 2 L. L. de Koninck (Zeit. anal. Chem., 19. 467, 1880) uses concentrated sulphuric acid and solid ammonium chloride in a small gas generator say Kipp's apparatus. 3 L. Brandt (Chem. Ztg., 33. 1114, 1909) condemns'the use of a condensing apparatus, but advocates a delivery tube with a bulb, and drawn to a narrow point, and dipping into water in a beaker. The beaker is kept cool by being placed in a larger beaker containing cold water. 282 A TREATISE ON CHEMICAL ANALYSIS. fitted l with a bulb tube to act as a water seal in case any chloride escapes condensation. If the condenser be adequately cooled, this will not be needed. The exit of the water seal must be directed outside the room. The Distillation. A slow current of hydrogen chloride 2 is passed through the apparatus for about an hour, so as to saturate the cold solution in the flask. The flask is then heated to boiling over an asbestos pad, or in an oil bath with a thermometer. The distillation is allowed to proceed while a rapid current of hydrogen chloride is passed through the system. A temperature of 108-110 ie best. 3 When about 80 c.c. have collected in the receiver, the distillation is stopped. A current of cold water is passed through the condenser all the time the distillation is in progress. Determination of Arsenic in the Distillate. The arsenic is best determined in the distillate by the volumetric process of Mohr 4 or Pearce. The arsenic may also be precipated as sulphide by diluting the solution with hot water and saturating the boiling solution with hydrogen sulphide, 5 and it can be weighed in the form of arsenic trisulphide, silver arseniate, or magnesium pyrophosphate. The magnesium pyrophosphate process is described in the next section. If the arsenic is weighed as trisulphide, the precipitate from the hydrogen sulphide treatment is prepared as described for antimony trisulphide on page 296. The antimony and tin may be determined in the residue in the flask by precipitating the mixed sulphides of tin and antimony with hydrogen sulphide, and separating these two elements as indicated on page 305. 6 This method gives excellent results. It occupies between two and three hours. The main objections are : (1) the need for a special apparatus ; and (2) the 1 All rubber tubing and stoppers should have been previously boiled with dilute potash to get rid of the sulphur. 2 It is important to saturate the solution in the flask with hydrogen chloride before commencing the distillation. Fischer simply distilled from concentrated hydrochloric acid without keeping the solution saturated with a current of hydrogen chloride. This, however, renders repeated distillation necessary. One addition of hydrochloric acid, by Fischer's process, suffices for O'Ol grin, arsenic, but four distillations are needed for 1 grm., and the arsenic is not completely distilled off after the tenth distillation. With the current of hydrogen chloride, how- ever, one distillation suffices for 0'5 grm. arsenious oxide. After 50 c.c. have passed over, no more arsenic can usually be detected in the distillate. The addition of calcium chloride to the solution in the flask gives better results in distilling from concentrated hydrochloric acid with- out passing hydrogen chloride through the system. This method may be used for small amounts of arsenic. For the action of hydrochloric acid on arsenic, see R. Fresenius, Zeit. anal. Chem., i. 448, 1862. 3 If the temperature of distillation exceeds 125, some antimony, if present, will be found in the distillate F. Flatten, Journ. Soc. Chem. Ind., 13. 324, 1894. 4 If the solution has to stand any length of time, and Mohr's process is intended, the solution should be nearly neutralised with sodium hydroxide and an excess of sodium bi- carbonate added. This retards the rate of oxidation of the solution of arsenious chloride. 5 H. Rose, Handbuchderanalytischen Chemie, Berlin, 343, 1829 ; 390, 1871 ; R. E. 0. Puller, Zeit. anal. Chem., 10. 45, 1871 ; C. Friedheim and P. Michaelis, ib., 34. 505, 1895 ; R. Bunsen, Liebig's Ann., 192. 305, 1878. 6 Antimony can be separated by distillation, after the arsenic, by raising the temperature of the distillation flask up to ]80. The distillation of the antimony chloride commences about 125. Under ordinary conditions, the contents of the distillation flask would become dry, and the antimony chloride would sublime on to the neck of the flask. It is therefore necessary to use a mixture in the retort which is not liable to evaporate to dry ness, and which does not decompose under the conditions of the experiment. A. Gibb uses a solution of zinc chloride (page 280). After the distillation of the arsenic chloride, let the flask cool, change the receiver, and add 20 c.c. of concentrated hydrochloric acid. Fit up the apparatus as before, and gradually raise the temperature up to 200, still maintaining the rapid current of hydrogen chloride. The antimony chloride will collect in the receiver. The antimony may be determined volumetrically by titration exactly as described for arsenic. The solution in the retort may be dissolved in hot water, and the tin determined as described below. There may be a loss of stannous chloride by this process if tin be present (L. A. Youtz, Zeit. anorg. Chem., 35. 55, 1903 ; W. Plato, ib., 68. 26, 1910). THE DETERMINATION OF ARSENIC. 283 consumption of time. If this process be inconvenient, the arsenic may be separated from the antimony and tin as ammonium magnesium arsenate. 129. The Separation of Arsenic from Antimony and Tin as Magnesium Ammonium Arsenate. Theoretical. In 1846 Levol 1 showed that a precipitate of magnesium ammonium arsenate, resembling magnesium ammonium phosphate (page 215), is obtained when a magnesium salt is added to an ammoniacal solution of an arsenate. The precipitate so obtained is slightly soluble in water 100 grms. of water dissolve 0*036 grms. of MgNH 4 As0 4 . The magnesium ammonium arsenate is hydrolysed by hot water forming "free arsenic acid, ammonium arsenate, and magnesia " : MgNH 4 As0 4 + 2H 2 = Mg(OH) 2 + NH 4 H 2 As0 4 . If an aqueous solution be evaporated to dryness and the residue calcined, some arsenic is volatilised. " In the cold, magnesium ammonium arsenate is not hydrolysed by water." The precipitate is much less soluble in ammoniacal solutions. A 25 per cent, solution of ammonia dissolves 0*0063 grm. of the salt per 100 grms. of the solvent. In consequence, it is usual to make the precipitation in ammoniacal solution. The presence of ammonium chloride augments the solubility of the precipitate. Thus, 100 grms. of a 1*41 per cent, solution of ammonium chloride dissolve 0*0735 grm. of the anhydrous MgNH 4 As0 4 , while 100 grms. of a 12-5 per cent, solution dissolve 0'113 grm. of the salt. The presence of am- monium tartrate increases the solubility of the salt. Thus, 100 grms. of a 1*5 per cent, solution of tartaric acid, feebly ammoniacal, dissolves 0*07 grm. of MgNH 4 As0 4 . The ammonium tartrate is required when the precipitation is made in the presence of tin and antimony salts. The solubility of the pre- cipitate is also much reduced in alcoholic solution, and consequently many prefer to make the precipitation in a solution containing both alcohol and ammonia. The precipitate is fairly soluble in acids. To find what amount of arsenic escaped precipitation, Hoffmann 2 determined the arsenic in the filtrate after application of the method described below. Starting with solutions containing the equivalent of 0*3 grm. of tin, 0*3 grm. of antimony, and As (used). . . 0-05 0*05 0'15 0-15 0'3 0*3 grm. As in filtrate . . O'OOOG 0'0005 0*0008 0'0012 0*0006 0*0008 grm. Traces only of tin and antimony were found in the precipitated magnesium ammonium arsenate. Fresenius proposed a correction for the unprecipitated magnesium arsenate 1 mgrm. arsenic per 30 c.c. of fluid. Dacru considers Fresenius' correction too small, and Virgili considers it too high. The latter recommended an allowance of 0*0012 grm. of arsenic per 100 c.c. of fluid, and this agrees best with Hoffmann's work. 1 A. Levol, Ann. Chim. Phys. (3), 17. 501, 1846 ; Compt. Rend., 23. 57, 1846 ; W. Hampe, Chem. Ztg., 18. 1900, 1894 ; J. F. Virgili, Zeit. anal. Chcm., 44. 492, 1905 ; R. Fresenius, 16., 3. 206, 1864 : G. C. Wittstein, ib., 2. 19, 1863 ; R. E. 0. Puller, ib., 10. 63, 1871 ; H. Lesser, ib., 27. 218, 1888 ; R. Brauner, ib., 16. 57, 1877 ; L. F. Wood, ib., 14. 356, 1875 ; Amer. J. Science (3), 6. 368, 1893 ; F. A. Gooch and M. A. Phelps, ib. (4), 22. 488, 1906 ; K. Reicliel, Zeit. anal. Chcm., 20. 89, 1881 ; C. Friedheim and P. Michaelis, ib., 34. 505, 1895 ; Her., 28. 1414, 1895; R. Bunsen, Liebig's Ann., 192 305, 1878 ; C. Ullgren, ib. t 69. 364, 1849; M. A. von Reis, Stahl Eisen, 9. 270, 1885 ; H. Rose, Fogg. Ann., 76. 534, 1849 ; E. Raffa, Oaz. Chim. Ital., 39. i, 154, 1909 ; J. C. Briinnich and F. Smith, Zeit. anorg. Chem., 68. 292, 1910. 2 M. Hoffmann, Beitrage zur Kenntnis der analytischen Chemie des Zintis, Antimons, und Arsens, Berlin, 1911. 284 A TREATISE ON CHEMICAL ANALYSIS. Magnesium ammonium arsenate is a white crystalline solid. Its composition corresponds with MgNH 4 As0 4 . 6H 2 when dried over sulphuric acid in a desiccatoK and MgNH 4 As0 4 . JH 2 when dried at 100 on a water bath. When dried afnigher temperatures, more water and possibly some ammonia are lost. On account of the indefinite character of the hydrated salt, it is not advisable to attempt to use this substance for the final weighing, as recommended by Koehler. 1 On ignition, the salt loses ammonia and water, forming magnesium pyroarsenate Mg 2 As 2 7 . If the ignition be conducted at too high a tempera- ture, the magnesium pyroarsenate appears to decompose, and some arsenic oxide is volatilised. For instance, Friedheim and Michaelis 2 found that a precipitate which theoretically should have been 0'3621 grm. Mg 2 As 2 7 changed in weight on ignition as indicated in the following table : Table XL VI. Influence of Calcination on Magnesium Pyroarsenate. Burner. Time of heating. Min. Weight. Grms. Bunsen's 30 0-3612 60 0-3611 Blast 30 0-3602 ii 55 0-3596 ii 85 0-3587 J25 0-3576 ii 155 0-3572 There is thus a loss of 4 milligrams on blasting. The numbers also show that 30 minutes' calcination on a good Bunsen's burner is sufficient. The Determination. The solution of the freshly precipitated sulphides in sodium sulphide is evaporated nearly to dryness, and the residue digested with hydrochloric acid and potassium chlorate in a flask with a reflux condenser (page 303), in order to prevent loss by the volatilisation of arsenic chloride. 3 The solution should occupy about 100 c.c. Add (say, 12 grms.) tartaric acid, 4 ammonium chloride (about 2 grms. per 50 c.c.), and an excess of concentrated ammonia (one-third the total volume of the solution). 5 The arsenic is then precipitated from the clear ammoniacal solution by adding magnesia mixture (page 283) drop by drop with constant stirring. The volume of the magnesia 1 0. Koehler, Arckiv Pharm. (3), 27. 406, 1889. 2 E. W. Parnel, Ghem. News, 21. 133, 213, 1870 ; R. W. E. Macivor, ib., 32. 283, 1875 ; F. Field, ib., 21. 193, 1870; Journ. Chem. Soc., 26. 6, 1873 ; C. Rammelsberg, Ber., 7. 544, 1874 ; L. Chevron and A. Droixhe, Bull. Acad. Belg. (3), 16. 475, 1888 ; F. Reichel, Zeit. anal. Chem., 20. 89, 1881 ; R. Brauner, ib., 16. 57, 1877 ; C. Friedheim and P. Michaelis, ib., 34. 505, 1895 ; M. Austin, Zeit. anorg. Chem., 23. 146, 1900 ; C. Lefevre, Ann. Chim. Phys. (6), 27. 55, 1892. 3 0. Piloty and A. Stock, Ber., 30. 1649, 1897. 4 The object of the tartaric acid is to keep the tin and antimony in solution. If much tin be present, the addition of ammonia may produce a turbidity, showing that insufficient tartaric acid is present. In that case, decant off the clear, dissolve the precipitate in tartaric acid on a water bath, and then mix the solutions (page 295). 5 Some recommend the addition of one-third the volume of 95 per cent, alcohol at this stage. F. Field, Journ. Chem. Soc., 26. 6, 1873 ; R. E. 0. Puller, Zeit. anal. Chem., 10. 57, 1871 ; C. R. Fresenius, ib., 3. 206, 1864 ; F. Reichel, ib., 20. 89, 1881 ; C. Friedheim and P. Michaelis, ib., 34. 505, 1895 ; H. Rose, ib., i. 417, 1862 ; Pogg. Ann., 76. 534, 1849 ; L. F. Wood, Amer. J. Science (3), 6. 368, 1873 ; M. Austin, ib. (4), 9. 55, 1900 ; 0. C. Beck and H. Fisher, Chem. News, So. 259, 1899 ; School Mines Quart., 20. 372, 1899. THE DETERMINATION OF ARSENIC. 285 mixture to be added is approximately one-third the total volume of the solution. 1 Let the mixture stand in a cool place for about 1 2 hours. Decant the clear liquid through a Gooch's crucible packed with asbestos. Wash with dilute ammonia water (2-5 per cent, of ammonia), first by decantation, and finally transfer the precipitate to the crucible. 2 Note the volume of nitrate and washings. Dry the precipitate at about 109. Place a crystal of ammonium nitrate in the crucible, 3 and gradually raise the temperature until the saucer containing the Gooch's crucible (page 106) is bright red. Too high a temperature is bad, since some arsenic may be lost. Cool in a desiccator and weigh as Mg 2 As 2 7 . 4 The weight of the magnesium pyroarsenate multiplied by 0*6374 represents the corresponding amount of As 2 3 . Except for the disadvantage which attends the slow precipitation, especially when two precipitations are made, the method is quite satisfactory. A correction of 0*0016 grm. of As 2 3 per 100 c.c. of filtrate and washings may be allowed. EXAMPLE. Suppose the precipitate has been in contact with approximately 150 c.c. of liquid, and Crucible and precipitate . ... . . 10*31428 grm. Crucible . . 10'OOU grm. Mg. 2 As 2 7 0*3128 grm. Hence 0-3128x0-6373 = 0-1993 grm. As 2 3 . Since 100 c.c. of filtrate involves a correction of +0 0016 grm., we have 0-1993 + 0-0024 = 0*201 7 grm. of As 2 3 in the given sample, which \veighed 5 grms. Hence the sample has the equivalent of 4*03 per cent. As 2 3 . Comparison Determinations. In a separation of arsenic from mixed tin and antimony sulphides by the distillation process, and the arsenic subsequently determined by this and the two processes described below, the percentage arsenic obtained was : Level's process. Mohr's process. Pearce's process. 8*25 8-25 8*15 8*30 8-29 8*19 ^-_ ^^ - Distillation process of separation. The filtrate is just acidified with hydrochloric acid and saturated with hydrogen sulphide, whereby the sulphides of tin and antimony are precipitated as indicated on page 276 ; or the solution may be divided into two parts, and the antimony and tin determined separately, as indicated on page 305. 5 1 The precipitation is facilitated by the addition of a little alcohol E. Murmann, Oester. Chem. Ztg., 13. 227, 1910. 2 Some filter here through filter paper, and dissolve the precipitate in hydrochloric acid. The magnesium arsenate is reprecipitated in order to eliminate a possible contamination of the first precipitate with magnesia, more particularly if magnesium sulphate is one of the components of the magnesia mixture. 3 Or ignite in a current of oxygen in a Rose's crucible. 4 If filter paper be employed, some arsenic may be lost by volatilisation. Destruction of the filter paper by nitric acid gives good results L. L. de Koninck (Zeit. anal. Chem., 29. 165, 1890). W. Hampe prefers to dissolve the precipitate in hydrochloric acid, remove the arsenic by passing hydrogen sulphide, and determine the magnesia in the filtrate by precipitating as magnesium phosphate (page 218). A better plan is to determine the arsenic volumetricaliy by dissolving the precipitated magnesium ammonium arsenate in acid and proceeding as indicated for Mohr's process, or Pearce's process. 5 It will be remembered that many kinds of glass, as well as caustic alkalies, contain arsenic (page 184), but not in quantities likely to affect quantitative results by the processes here described R. Fresenius, Zeit. anal. Chem., 6. 201, 1867 ; W. Fresenius, ib., 22. 397, 1883; J. Marshall and C. Pott, Amer. Chem. Journ., 10. 425, 1888 ; S. R. Scholes, Journ. Ind. Eng. Chem., 4. 16, 1912. 286 A TREATISE ON CHEMICAL ANALYSIS. 130. Notes on Iodine, Potassium Iodide, Starch, and Sodium Thiosulphate. In 1853, Bunsen 1 demonstrated the general applicability of reactions in which iodine was liberated from potassium iodide for volumetric analysis. Bunsen used sulphurous acid for measuring the iodine set free. Schwarz 2 used sodium thiosulphate for titrating iodine in neutral or acid solutions, and Mohr extended the process to solutions of arsenious and antimonious salts in alkali carbonate solutions. F. Stromeyer first used starch for developing the tint of free iodine. 3 Starch, iodine, potassium iodide, and sodium thiosulphate are now used so fre- quently in certain types of volumetric work, that some notes on the use of these substances may now be made. Starch Solutions. Starch is used as an indicator in titrations with iodine solutions. . Free iodine colours an aqueous solution of starch blue only in the presence of a soluble iodide. The sensitiveness of the reaction is determined by the composition of the solution. Sodium sulphate makes the reaction very sensitive, e.g., a 4N- solution of sodium sulphate (with starch) gives a colour with a 0'0000017N-iodine solution. 4 Some believe the blue substance is a solid solution, 5 while others con* sider it to be a compound of hydriodic acid with an iodine addition product of starch 6 C 24 H 40 20 I 4 . HI. The addition product C 24 H 40 20 I 4 is supposed to be colourless in aqueous solution. It is further supposed that the blue compound C 24 H 40 20 I 4 . KI is dissociated in dilute aqueous solutions into colourless C 24 H 40 20 I 4 and KI. When the concentration of the latter is increased, the concentration of the blue C 24 H 40 20 I 4 . KI is increased, and the solution appears blue. This latter hypothesis agrees best with the facts. One gram of potato (arrowroot, corn, or rice) starch is triturated with 10 c.c. of cold water until a smooth paste is obtained. Add sufficient boiling water with constant stirring to make 200 c.c. of a thin translucent fluid. If the fluid be not transparent, it must be boiled two or three minutes, but prolonged boiling must be avoided, since it converts some of the starch into dextrine. Cool the solution. Let it settle overnight, and decant off the dear. This solution will not keep more than a couple of days. A larger quantity can be made and the clear liquid poured into small 50-c.c. phials up to the neck. The phials are placed in a water bath, heated for a couple of hours, and then closed by cork stoppers. The solution so preserved will keep indefinitely. When a phial is opened, the contents deteriorate in a few days, hence the use of small bottles, which permits the solution to be used before it spoils. In order to prevent deterioration, preservatives are frequently added, e.g., a few drops of chloro- form, oil of cassia, salicylic acid, zinc chloride, zinc iodide, or mercuric chloride (2 grms. per litre of paste). 7 J R. Bunsen, Liebig's Ann., 86. 265, 1853. 2 C. L. H. Schwarz, PraTctische Anleitung zu Maassanalysen, Braunschweig, 1853. 3 F. Stromeyer, Schweigger's Journ. , 12. 349, 1814. 4 J. Pinnow, Zeit. anal. Chem., 41. 485, 1902; C. Meineke, Chem. Ztg., 18. 157, 1894; A. Eckstadt, Zeit. anorg. Chem., 29. 51, 1901. 5 F. W. Kiister, Liebig's Ann., 283. 360, 1895. 6 F. Mylius, Bar., 20. 688, 1881 ; C. Lonner, Zeit. anal Chem., 33. 409, 1894; L. W. Andrews and H. M. Goettsch, Journ. Amer. Chem. Soc., 24. 865, 1902 ; H. B. Stocks, Chem. News, 56. 212, 1887. 7 G. Gastine, Bull. Soc. Chim. (2), 50. 172, 1888; M. Musculas, Compt. Rend., 78. 1483, 1874 ; M. Miitnianski, Zzit. anal. Chem., 36. 220, 1897 ; C. Reinhardt, ib., 25. 37, 1886 ; F. Mohr, ib., 14. 79, 1875 ; A. Mliller, ib., 22. 76, 1883 ; A. Wroblewski, Journ. Pharm. Chim. (6), 8. 314, 1907; 0. Forster, Chem. Ztg., 21. 41, 1897; L. Mathieu, Bull. Assoc. Chem. Sucr. List., 27. 1166, 1910. THE DETERMINATION OF ARSENIC. 287 Zulkowsky's 1 soluble starch is very convenient for titrations. It is sold in the form of a thick paste. A small portion is taken from the stoppered bottle on the end of a clean glass rod, and then mixed with water in a test-tube for use. The so-called "soluble starch" is made by grinding, say, 100 grms. of rice starch with a 2 per cent, potash solution until a homogeneous solution is obtained. Stir the solution with more potash, so that the volume of the mixture is from 600 to 800 c.c. Heat the -mixture on a water bath until all is quite liquefied, and then 30 to 40 minutes over a free flame. Filter. Add an excess of acetic acid to the filtrate. Finally precipitate the starch by adding an equal volume of 95 per cent, alcohol. Redissolve the precipitate, again precipitate the starch, and dissolve the precipitate in the least possible quantity of water. Pour the solution in a thin stream into a large quantity of absolute alcohol. Filter the precipitate and wash the precipitate with alcohol, and finally with ether. Dry in vacuo. The yield is about 50 or 60 per cent. " Soluble starch " is a commercial article. It is sold as a white powder, which dissolves readily in boiling water, forming a clear solution. The reaction of iodine with these purified starches is sensitive and sharp. If impure starch be used for the work, a loss of iodine may occur owing to the presence of erythrodextrine, which gives a reddish colour with iodine. 2 If a preservative has been used with the starch, it may be necessary to be prepared for certain disturbances, e.g., the presence of zinc chloride would interfere in the titration of sulphides and carbonates. In using starch as an indicator with iodine titrations, if nitrogen fumes are in the atmosphere of the laboratory, a rapid "after-blueing" of the solution may lead to high results. 3 Potassium Iodide. The potassium iodide used in volumetric work must be free from iodates. To test if the potassium iodide is suitable for the work, dissolve a gram of the salt in 11 c.c. of water, and add a drop of pure hydrochloric acid. If iodates be present, free iodine will be formed : KI0 3 + 5KI + 6HC1 = 6KC1 + 3I 2 + 3H 2 0. Shake the mixture with a little chloroform. If the chloroform shows no colora- tion, or, at the worst, a faint tint, the potassium iodide is suitable for the work, 4 1 K. Zulkowsky, er., 13. 1395, 1874. 2 F. E. Hale, Amer. J. Science (4), 13. 379, 1902; C. Lonnes, Zeit. anal. Chem., 33. 409, 1894; C. Meineke, Chem. Ztg., 18. 157, 1894; 19. 5, 1895 ; G. Rivat, Chem. Ztg., 34. 1141, 1910 ; L. Mathieu, Ann. Chim. Anal., 16. 51, 1911. 3 F. Sinnatt (Analyst, 35. 309, 1910 ; 37. 252, 1912) recommends methylene blue in place of starch in titrating with standard iodine solutions. '05 grm. of the dye is dissolved in water, and the solution made up to 50 c.c. One c.c. of this solution per 50 c.c. of the solution to be titrated is used as indicator. The end point is indicated by a change from blue to yellowish green. It might here be added that B. Schwezoff (Zeit. anal. Chem., 44. 85, 1905) and A. Bobierre (Monit. Sclent. (2), 5. 951, 1868 ; Chem. News, 18. 265, 1868) substitute for the starch reaction with iodine the intense red coloration furnished by the solution of iodine in colourless benzene, a reaction suggested by M. Moride in 1852 (Compt. Rend., 35. 789, 1852 ; B. M. Margosches, Zeit. anal. Chem., 44. 392, 1905). M. Bertin (Journ. de Medecine de V Quest, 31. 201, 1868) also used the same reaction. A. Dupre (Ann. Chim. Pharm., 94. 365, 1855) used chloroform, which is less convenient than benzene. It is claimed that the benzene reaction is 2^ times as sensitive as starch. E. Borgmann (Nederl. Tijdschr. Pharm., 8. 140, 1896) used au alcoholic solution of kakotelin. 4 F. Pollaci, Oaz. Chim. Ital., 3. 474, 1873. This test is often prescribed for iodates, but traces of iron or cuprous oxides, in the presence of dissolved oxygen, also give the pink colour. L. W. Andrews (Journ. Amer. Chem. Soc., 31. 1035, 1909) detects iodates by using potassium hydrogen tartrate in place of hydrochloric acid. 288 A TREATISE ON CHEMICAL ANALYSIS. and sufficiently free from contamination with substances which decompose potassium iodide. Iodine. Iodine I. Atomic and equivalent weights : 126 '92. The most commonly employed solutions are : Grras. I per c.c. AN-Iodine '01 2692 T ijjN-Iodine . 001269 Since iodine is but sparingly soluble in water, it is necessary to augment the solubility by using solutions of potassium iodide as a solvent for the iodine. Hence, dissolve 12*69 grms. of iodine and 18 grms. of potassium iodide in about 300 c.c. of water in a litre flask, and make the solution up to a litre. The solution is standardised by decinormal sodium thiosulphate, or by means of arsenious oxide. 1 Commercial iodine may contain chlorine, bromine, water, cyanogen, etc. To purify iodine 2 on a large scale, dissolve 4 kilograms of iodine in a solution con- taining 2 kilograms of potassium iodide, dissolved in 2 litres of water. Pour the solution into water, and let it stand until the iodine is precipitated. The chlorine or bromine present react with the potassium iodide, liberating iodine and forming potassium chloride or bromide. Pour off the supernatant liquid ; wash the iodine with water 3 until it is free from potassium iodide ; filter through a layer of sand assisted by suction. Transfer the iodine to a shallow dish, and dry the mass over concentrated sulphuric acid (12 days). Place the iodine in a com- bustion tube slightly inclined, with a plug of asbestos placed so as to prevent the melted iodine running down the tube. The lower end of the combustion tube is connected with a couple of drying tubes, one containing calcium chloride, and the other phosphorus pentoxide, in order to dry the air which is to pass through. The upper end of the combustion tube is covered by a bottle fitted with a rubber stopper to catch any iodine vapours not previously condensed. The stopper is fitted with a glass tube connected with a cylinder, which acts as an aspirator. The suction is sufficient to enable the iodine to condense far enough from the heated portion to prevent liquefaction of the sublimed crystals. To prevent an accident from the sudden heating of the tube, place a piece of iron pipe over the part of the tube containing the asbestos and iodine, and extending about 10 cm. beyond the tube. A layer of asbestos is placed between the glass and iron tubes. This process is not convenient if but small quantities of iodine have to be treated. To purify small quantities of iodine, grind, say, 6 grms. of commercial iodine with 2 grms. of potassium iodide. Put the dry mixture in a small dry beaker (fig. 126) fitted with a Gockel's condenser. 4 The beaker is surrounded with a cylindrical asbestos jacket (not shown in the diagram). Place the beaker on a wire gauze, or a hot plate, and heat the apparatus by means of a small flame. 1 See next section. 2 L. L. de Koninck, Bull. Assoc. Chim. Belg., 17. 15, 1904 ; J. S. Stas, (Euvres Completes, Bruxelles, I. 563, 1894 ; G. C. Wittstein, Dingler's Journ., 20O. 310, 1871 ; C. F. Mohr, Lehrbuch der chemisch-analytischen Titrirmethode, Braunschweig, 269, 1874 ; A. Gross, Journ. Amer. Chem. Soc., 25. 987, 1903 ; Glum. News, 88. 274, 1903 ; L. W. Andrews, ib., go. 27, 1904 ; Amer. Chem. Journ., 30. 428, 1903 ; B. Lean and W. H. Whatmough, Proc. Chem. Soc., 14. 5, 1898; C. Meineke, Chem. News, 68. 272, 1893; Chem. Ztg., 16. 1219, 1230, 1892; Z. Musset, Zeit. anal. Chem., 30. 45, 1891; G. Lunge, Zeit. angew. Chem., 7. 234, 1894; A. Ladenburg, er., 35. 1256, 1902. 3 The washings, particularly at first, are retained and the iodine recovered. 4 H. Gockel, Zeit. angew. Chem., 12. 494, 1899. THE DETERMINATION OF ARSENIC. 289 The condenser is full of cold water, at the temperature of the room. When violet vapours have ceased to come from the bottom of the beaker, let the apparatus cool. A crust of iodine will be found on the condenser. Pass a current of cold water through the con- denser. The glass contracts, and the crust of iodine can be easily removed by pushing it with a glass rod into a similar beaker. The sublimation is repeated without the potassium iodide at as low a temperature as possible. Grind the iodine in an agate mo*tar, and dry in a desiccator over calcium chloride, not sul- phuric acid, or the iodine may be con- taminated. If the cover of the desiccator is greased, the iodine may attack the grease, forming hydriodic acid, which might contaminate the iodine. Preservation of Iodine Solutions. Iodine solutions should only be kept in glass-stoppered bottles preferably made of brown glass, to cut off the light. Iodine solutions diminish in strength with time, particularly if the bottle with the stock solution is frequently opened, owing to the volatilisation of the iodine. Accord- ing to Schmatolla, 1 a y^N-solution of iodine scarcely changes if kept for a year, particularly if the space between the stopper and the neck of the bottle be kept dry. The presence of an excess of potassium iodide lowers the vapour pressure of the iodine vapour and makes the solution keep better. Sodium Thiosulphate. Anhydrous sodium thiosulphate Na 2 S. 2 3 ; crystalline sodium thiosulphate Na 2 S 2 3 . 5 FJ 2 0. Molecular weight of the crystalline salt: 248*22; equivalent weight: 248 '22. Mole- cular weight of anhydrous salt : 158'14 ; equivalent weight : 158*14. The most commonly employed solution has the strength : FIG. 126. Purification of Iodine. thiosulphate Grins, per c.c. ( 0*024822 (crystalline salt) 10*015814 (anhydrous salt) the Sodium Thiosulphate. This salt is much used in iodine titrations thiosulphate, in acid solutions, 2 is transformed by iodine into tetrathionate : 2Na 2 S 2 3 + 1 2 = 2NaI + Na 2 S 4 6 . The anhydrous sodium thiosulphate is made by recrystallising the "pure " com- mercial salt from warm solutions, saturated at 30 35, by cooling and constant stirring. The fine-grained crystals so obtained are dried on filter paper at the 1 0. Schmatolla, Apoth. Ztg., 17. 248, 1902. 2 In alkali or sodium hydrogen carbonate solutions, sodium thiosulphate and tetrathionate are partly oxidised to sulphate. G. Topf, Zeit. anal. Chem., 26. 137, 277, 1887 ; J. P. Batey, Analyst, 36. 132, 1911 ; E. Abel, Zeit. anorg. Chem., 74. 395, 1912 ; C. Friedheim, Zeit. angcw. Chem., 4. 415, 1891. 19 290 A TREATISE ON CHEMICAL ANALYSIS. temperature of the room. The salt is then dehydrated over concentrated sulphuric acid until it has fallen to powder, and a portiop heated in a test tube shows no signs of fusion when heated to 50. The dehydration is completed at 80 in an air bath with repeated stirring of the powder. For relatively small quantities, two hours will suffice. Preserve the salt in well- stoppered bottles. Young 1 finds that the salt so prepared keeps well and may be itself used for standardising iodine solutions. The molecular weight of the crystalline salt Na 2 S 2 3 . 5H 2 is 248'22 ; and of the anhydrous salt Na 2 S 2 3 158 '14. For a litre of the normal solution, one molecular weight of the salt expressed in grams is employed. The solutions should be made with water free from carbon dioxide, since the latter reacts with the thiosulphate, forming sodium sulphite and free sulphur : Na 2 S 2 3 + 2H 2 C0 3 - Na 2 S0 3 + H 2 + 2C0 2 + S. The sodium sulphite reacts with more iodine than the corresponding thiosulphate, and hence the solution appears to become more concentrated on standing. After all the carbon dioxide in the water has reacted with the thiosulphate, the solution may be kept for months without appreciable change. 2 Tread well found no change in the titre of a solution against iodine after it had been kept for eight months. The addition of ammonium carbonate, sometimes recommended to preserve the solution, really acts in the opposite direction, and makes the solution less stable. 3 containing arsenious oxide, the latter is oxidised 4 to arsenic oxide, and a colour- 131. Mohr's Iodine Volumetric Process for Arsenic. Theoretical. When a solution of iodine is gradually added to a solution taining arsenious oxide, the latter is less solution of hydriodic acid is formed : As 2 3 + 2H 2 + 2I 2 = 4HI + As 2 5 . When all the arsenious oxide has been oxidised, any further addition of iodine produces a yellow colour, or, if a solution of starch be present, blue starch iodide is formed. The starch solution enables one part of iodine in over three million parts of the solution to be detected. It is best to work with cold solutions, since the starch iodide dissociates on heating and loses its blue colour.- The blue colour re-forms on cooling. The solution should be alkaline in order to neutralise the hydriodic acid formed in the reaction. Alkaline hydroxides will not do, because they react with starch iodide, and with free iodine (GNaOH + 3I 2 = 5NaI + NaI0 3 + 3H 2 0) ; sodium carbonate is partly hydrolysed 5 in aqueous solutions (Na 2 C0 3 + H 2 = NaOH + NaHC0 3 ). It therefore reacts with iodine, but to a less extent than 1 S. W. Young, Journ. Amer. Chem. Soc., 26. 1028, 1904; G. Topf, Zeit. anal. Chem., 26. 140, 1887. 2 C. Meineke, Chem. Ztg., 18. 33, 1894 ; F. P. Treadwell, Kurzes Lehrbuch der analytischen Chemie, Leipzig, 2. 530, 534, 1911. 3 See S. U. Pickering (Chem. Neivs, 44. 277, 1881) for the deterioration of sodium thiosulphate with keeping. The solution should be kept in amber-coloured small glass-stoppered bottles, carefully protected from dust, air, and light. It should be re- standardised frequently. 4 E. Waitz, Zeit. anal. Chem., 10. 158, 1871 ; S. Avery and H. T. Beans, Journ. Amer. Chem. Soc., 23. 485, 1901 ; J. K. Haywood, ib., 25. 963, 1903 ; S. Avery, ib., 25. 1096, 1903 ; E. W. Washburn, ib., 30. 31, 1908 ; F. Mohr, Lehrbuch der chemisch-analytischen Titrirmethode, Braunschweig, 1859 ; J. P. Batey, Analyst, 36. 132, 1911 ; B. E. Curry and T. 0. Smith, Journ. Ind. Eng. Chem., 4. 198, 1912 ; G. S. Jamieson, ib., 3. 250, 1911. 6 H. N. M'Coy, Amer. Chem. Journ., 24. 437, 1900. THE DETERMINATION OF ARSENIC. 2 9 I sodium hydroxide. For example, working with a solution of iodine (1 c.c = 0-005 grm. iodine) and a saturated solution of sodium carbonate, Waitz found : Water ... 20 20 50 20 100 c c Sod. carb. solution .5 10 10 20 20 c!c! Iodine sol. required . 0'25 0'3 0*35 0*8 1*1 cV It is generally supposed that sodium bicarbonate does not react with iodine, and Beringer l quotes the following experiment to show that " a large variation in the quantity of bicarbonate has no effect " on the consumption of iodine : Sodium bicarbonate 1 2 5 10 grms. Iodine solution required . 20 '1 20 '0 20 '1 20 '0 c.c. The truth is that the sodium bicarbonate is slightly hydrolysed in aqueous solu- tions, forming sodium carbonate and carbonic acid (2NaHC0 3 = Na t CO +H CO ). The sodium carbonate is further hydrolysed, as indicated above. The amount hydrolysed is increased at elevated temperatures and in concentrated solutions. 2 Hence, in using sodium bicarbonate to neutralise the hydriodic acid, the solution should be as nearly neutral as possible at the end of the titration to get the best results. A deficiency of the sodium bicarbonate is very objectionable. A slight excess does no serious harm, since the results will be sufficiently exact for commercial work. 3 The solution to be titrated should be dilute and cold. The titration should also be performed as quickly as possible to avoid variations in the concentration of the acid in solution by the escape of carbon dioxide. The Determination. The acid solution of the arsenious compound 4 is nearly neutralized with ammonia or sodium hydroxide. 5 Add 20 c.c. of a saturated solution of sodium bicarbonate, 6 and 2 or 3 c.c. of a solution of starch, and titrate with standard iodine solution 7 until a permanent blue tinge remains 1 C. and J. J. Beringer, A Textbook of Assaying, London, 387, 1908. 2 W. A. Puckner (Proc. Amer. Phil. Assoc., 43. 408, 1904) considers the amount of hydrolysis is also slightly greater in large flasks than in small flasks owing to the escape of more carbon dioxide. The carbon dioxide is evolved during the titration by the decomposition of the carbonic acid (H 2 CO 3 = H.,0 -f C0 2 ). Hence, the titration should be made in Erlenmeyer's flasks, not in beakers 3 The labour involved in the exact adjustment of the bicarbonate for " unknown " solutions might be considered "finicking" by a commercial analyst. Every 100 c.c. of ^N-iodine solution needs about 5 grms. of sodium bicarbonate, on the assumption that the solution is neutral before adding this amount of sodium bicarbonate. E. W. Washburn (Journ. Amer. Chem. Soc., 30. 31, 1908) avoids the disturbances with sodium bicarbonate by using 11 grms.' sodium phosphate Na 2 HP0 4 . 12H 2 for every 100 c c. of the ^N-iodine solution. The sodium phosphate is supposed to be added to the neutral solution. For errors with an excess of sodium bicarbonate, see J. P. Batey, Analyst, 36. 132, 1911 ; M. Bialobrzcsky (Zeit. anal. Chem., 37. 444, 1898) recommends ammonium acetate in place of sodium bicarbonate. 4 If arsenic salts be present, the solution must be reduced by, say, a crystal of potassium iodide and an excess of sulphur dioxide. The sulphur dioxide destroys the free iodine. If all the sulphur dioxide be boiled olf before the reduction is complete, the first drop of iodine from the burette in the titration will probably colour the solution yellow. If the reduction be complete, no permanent coloration will be produced when a crystal of potassium iodide is dropped into the solution. The solution must also be boiled free from sulphur dioxide, (test for sulphur dioxide as indicated page 192). Avoid prolonged boiling (page 30u). 5 The ammonia should be free from pyrrol. This is readily obtained by tinging ammonia with potassium permanganate ancl afterwards decanting from the deposit formed. 6 Free from nitrites, chlorates, etc. 7 STANDARD IODINE SOLUTION. Shake, say, 1*05 grms. of resublimed iodine along with 4 grms. of potassium iodide (free from iodates) with about 2 c.c. of water in a litre flask until the iodine is dissolved. This occupies but two or three minutes. Make the solution up to a litre. Standardise the solution of iodine by dissolving 0*4092 grm. of pure arsenious oxide in a hot solution containing about 20 grms. of sodium hydroxide (free from sulphur) in 100 c.c. of water (B. Penot, Dingler's Journ,, 127. 134, 1853 ; L. Miiller, ib., 129. 286, 1853 ; D. Hancock, Journ. Amer. Chem. Soc., 16. 431, 1894) ; acidify the solution with hydrochloric acid, and then make 292 A TREATISE ON CHEMICAL ANALYSIS. suffused through the solution. The end point is very sharp, and consequently the iodine must be added very gradually towards the end of the titration. EXAMPLE. Suppose 1-5 grms. of material be under investigation and the solution requires 26*2 c.c. of iodine containing the equivalent of 0*0004092 grm. of As 2 O 3 per cubic centimetre, then the material contains the equivalent of 0-0004092x26-2x100 t = 0*715 percent, of As 2 3 . 1*O Correction for the Volume of the Liquid being titrated. With the same amount of arsenic, and different volumes of liquid, different volumes of the standard iodine solution may be needed to develop the starch blue. Hence the solution to be titrated should have nearly the same concentration as the solution used for standardising the iodine. The error is not very marked with T VN- and stronger iodine solutions, but with more dilute solutions, say T J^N-, the error is quite appreciable. The more dilute the solution, the greater the amount of iodine needed to produce the coloration. When the solution is less than 150 c.c. and potassium iodide is present, the same amount of the iodine solution is needed to produce the blue colour; but if the volume of the solution to be titrated be greater than 150 c.c., more iodine is necessary the greater the volume of the solution. Thus, Treadwell 1 found, with no other potassium iodide than that present in the standard solution : Water . . . . .50 100 150 200 c.c. yfo N-Iodine .... 0*15 0'30 0*47 0*64 c.c. In the presence of 1 grm. potassium iodide, Treadwell found : Water . . . 50 100 150 200 500 c.c. . 0'04 0'04 0'04 0'14 0'32 c.c. If the solutions to be titrated vary in volume, it is therefore necessary to apply a correction for the amount of the standard solution required to produce a coloration after the reaction between iodine and arsenious oxide is completed (page 200). 132. Pearce's Volumetric Process for Arsenic. In Bennett's modification of Pearce's process, 2 the compound containing the arsenic is oxidised. The arsenic oxide is precipitated as silver arsenate. The silver in the silver arsenate is determined by Volhard's volumetric process, and the corresponding As 2 3 is computed by rule of three. In the older process of E. Reich (1864), the silver arsenate is determined gravimetrically, and this may be advisable when standard solutions are not available, and only an occasional arsenic determination has to be made. 3 the solution alkaline with sodium bicarbonate ; cool ; make the solution up to a litre. Pipette 50 c.c. of this solution into a flask ; add a couple of drops of starch solution, and titrate with the iodine solution until the permanent blue tinge is obtained. If 50 c.c. of the iodine solution are required, 1 c.c. of iodine solution represents 0*0004092 grm. of As 2 3 . The solution of iodine keeps very well if kept in. a cool place in the dark. 1 K. P. Treadwell, Kurzes Lehrbuch der analytischen Chemie, Leipzig, 2. 536, 1911 ; Eng. trans., New York, 2. 513, 1904. 2 R. Pearce, Proc. Colorado Scientific Soc., I. 14, 1883 ; Chem. News, 48. 85, 1883 ; F. Reich, Zeit. anal. Chem., 25. 411, 1886 : A. H. Low, Journ. Amer. Chem. Soc., 28. 1715, 1906 ; J. F. Bennett, ib. } 21. 431, 1899 ; L.' R. W. McCay, Amer. Chem. Journ., 8. 77, 1886 ; 12. 547, 1890; Chem. News, 48. 7, 168, 1883 ; 0. J. Frost, ib., 48. 85, 1883 ; 53.. 221, 232, 243, 1886 ; T. Brown, ib., 8z. 178, 184, 1900 ; Journ. Amer. Chem. Soc., 21. 1899 ; R. C. Canby, Trans. Amer. Inst. Min. Eng., 17. 77, 1888 ; G. W. Lehmann and W. Mager, Amer. Chem. Journ., 7. 112, 1885 ; H. E. Hooper, Eng. Min. Journ., 94. 706, 1912. 3 In that case, the silver arsenate is precipitated as described in the text, filtered and washed with a dilute solution of ammonium nitrate. The precipitate is dissolved in ammonia, THE DETERMINATION OF ARSENIC. 293 Precipitation of Silver Arsenate. The arsenic compound is dissolved in nitric acid warming the solution if necessary. This oxidises the arsenious to arsenic oxide. 1 Add a few drops of phenolphthalein to the cold solution, and add sodium hydroxide until the indicator turns pink. Then add acetic acid, drop by drop, with constant stirring until the colour is discharged, 2 and add one drop of dilute acetic acid 3 in excess. 4 The solution now occupies about 100 c.c. Add an excess, say, 10 c.c., of a neutral solution of silver nitrate 5 with vigorous stirring. Let the precipitate settle a few minutes ; decant the clear ; wash twice by decantation; and, finally, filter off the brick-red precipitate of silver arsenate Ag 3 As0 4 . Wash the precipitate with cold water until it is free from silver acetate and nitrate when tested with a drop of sodium chloride solution. Titration for the Silver. Dissolve the precipitate in 5-10 c.c. of dilute nitric acid (1 : 1), and collect the filtrate, washings, etc., in a beaker. 6 Dilute the solu- tion, if necessary, to 100 c.c. with water. Add 5 c.c. of a saturated solution of ferric ammonium alum, and titrate with a standard solution of ammonium thio- cyanate 7 until a permanent red tint appears, according to Volhard's well-known process (page 76). Shake the flask well during the titration to break up any clots of silver thiocyanate, which might enclose some of the solution to be titrated. The process works well with small amounts of arsenic ; with large quantities, the bulky precipitate is inconvenient, and silver is expensive. Then use Mohr's process, page 290. Phosphates and molybdates should, of course, be absent. 133. The Evaluation of Arsenious Oxide. The amount of arsenic in commercial arsenious oxide can be estimated by dissolving, say, 0'5 grm. of the sample in 20 grms. of sodium hydroxide and 100 c.c. of water. Acidify the solution with hydrochloric acid, and add about 100 c.c. of a saturated solution of sodium bicarbonate. Titrate with standard iodine (a more concentrated standard solution than that indicated above may be used, say that indicated on page 312). 8 and the solution evaporated and dried in a weighed platinum dish. From the weight of Ag 3 As0 4 so obtained, the corresponding amount of arsenic or arsenious oxide can be readily computed. 1 If necessary, boil to expel carbon dioxide. 2 The precipitate formed later is soluble in nitric acid, ammonia, and ammonium nitrate solutions (J. W. Mallet and J. R. Santos, Chem. News, 38. 94, 1878). Nitric acid may be liberated during the separation of silver arsenate: Na 2 HAs0 4 + 3 AgN0 3 = Ag 3 As0 4 + 2NaN0 3 + HN0 3 . Hence the solution must be carefully neutralised. These two reagents nitric acid and ammonia presented a difficulty in Pearce's original method. Canby tried to eliminate the trouble with zinc oxide. As stated in the text, Bennett used a slightly acidified solution of sodium acetate in which "silver arsenate is practically insoluble." 3 Sodium acetate facilitates the subsequent separation of silver arsenate C. E. Avery, Amer. J. Science (2), 47. 25, 1867. 4 If the solution be alkaline, silver oxide may separate. 5 SILVER NITRATE SOLUTION. An ordinary nitric acid solution of silver nitrate will notdo. Dissolve 17 grms. of silver nitrate crystals in 500 c.c. of distilled water. One c.c. will pre- cipitate 0'005 grm. of arsenic, that is, 1 percent, if 0'5 grm. of sample were originally taken. Hence, 10 c.c. is usually an excess. 6 A white precipitate of silver chloride may remain on the filter paper undissolved. This can be neglected. 7 AMMONIUM THIOCYANATE SOLUTION. Dissolve 7 '61 7 grms. of ammonium thiocyanate in a litre of water. Standardise the solution by titration against known weights of silver nitrate (0'8 grm. per 100 c.c. acidified with nitric acid), and 5 c.c. of a saturated solution of iron alum as indicator. 8 If copper is present, add 2-3 grms. sodium potassium tartrate after the sodium bicarbonate, to prevent precipitation of the copper during the titration S. Avery and H. T. Beans, Journ. Amer. Chem. Soc., 23. 485, 1901. 294 A TREATISE ON CHEMICAL ANALYSIS. A mixture of arsenic and arsenious oxides may be treated in a similar manner. The titration gives the amount of arsenious oxide. A separate portion is dissolved in a similar manner, and the arsenic oxide (As 2 5 ) is reduced 1 by boiling with an excess of hydrochloric acid and potassium iodide. Boil the solution until all the free iodine has been driven off, and titrate the solution by Mohr's process. This gives the total arsenious oxide. The difference between the amount obtained in the two titrations is expressed in terms of As.->0 5 by multiplying the difference by 1 '1616. 1 For reduction by potassium iodide, see L. Rosen thaler, Zeit. anal. C'hem., 45. 319, 1906 ; for reduction with hydriodic acid, F. A. Gooch and P. E. Browning, Amer. J. Science (3), 39. 188, 1890; (3), 40. 66, 1890; for reduction with sulphur dioxide L. R. W. McCay, Amer. Chem. Journ., 7. 373, 1885. CHAPTER XXII. THE DETERMINATION OF ANTIMONY. 134. Antimony Sulphide. ANTIMONY sulphide is fairly soluble in hydrochloric acid * containing over 20 per cent. HC1, as indicated on page 275. Hence, it is necessary to work with dilute solutions if all the antimony is to be separated as sulphide. If the solution be too dilute, antimony oxychloride will separate. It is therefore best to work with a solution acid enough, say 20 per cent. HC1, in order to prevent the precipita- tion of the oxychloride. The acid solution is saturated with hydrogen sulphide, diluted with, say, an equal volume of water, and again saturated with the gas, 2 as indicated on page 276. At low temperatures, and particularly with rapid streams of hydrogen sulphide, Sb 2 S 5 is precipitated. 3 The higher the temperature, and the slower the stream of gas, the greater the amount of Sb 2 S 3 mixed with the Sb 2 S 5 . 4 In general work, therefore, we may assume that the precipitate will be a mixture of the two sulphides along with free sulphur. The Ignition of Antimony Sulphide. When the antimony is to be weighed as trisulphide, the excess of sulphur is removed from the precipitate by washing with, say, carbon disulphide, and the precipitate is thoroughly dried in a neutral atmosphere, say, carbon dioxide. In illustration, a precipitate of trisulphide, after drying 4 hours at 110, weighed 0'5861 grm. ; after washing with carbon disulphide and alcohol, and drying at 150 for two hours, 0*5552 grm. ; again washing with carbon disulphide, and drying one hour at 180, 0'5547 grm. Further drying an hour at Temperature . . 200 220 250 280' 300 Weight . . .0-5540 0*5538 0'5526 0'5523 0'5516 grm. when the weight remained practically constant. Influence of Oxalic Acid. The addition of oxalic acid to a solution of antimony or tin, feebly acid with hydrochloric acid, produces a white precipitate of antimony or tin oxalate ; if the solution be alkaline, fairly soluble double oxalates are formed, and the solution will remain clear without turbidity, 1 J. Lang, Ber., 18. 2714, 1885; J. Theile, Liebigfs Ann., 263. 361, 1891; Zeit. anal. Chem., 30. 473, 1891 ; T. Wilm, ib., 30. 428, 1891 ; B. Brauner and F. Tomiczek, ib., Monats. Chem., 8. 607, 1887 ; 0. Bosek, Chem. Neivs, 71. 195, 1895 ; B. Brauner, ib., 71. 196, 1895 ; M. Berthelot, Compt. Rend., 102. 22, 84, 86, 1886 ; A. Ditte, ib., 102. 168, 212, 1886 ; G. C. Wittstein, Vierteljahr. prakt. Pharm., 18. 531, 1869. 2 If the solution be too acid, the clear filtrate will become turbid as soon as it comes in contact with water say a moist beaker. In that case the solution must be diluted with water, and refiltered. 3 When the amount of hydrochloric acid in solution exceeds 10 per cent., the greater the amount of hydrochloric acid, the greater the proportion of Sb 2 S 5 precipitated. 4 Dark brown Sb. 2 S 5 is precipitated from cold solutions of antimonic salts, but not from antimonious salts, by an excess of hydrogen sulphide in aqueous solution. 295 296 A TREATISE ON CHEMICAL ANALYSIS. provided sufficient alkali be present. Hydrogen sulphide produces an orange- coloured precipitate of antimony trisulphide in solutions of an antimony salt ; and in solutions of a stannous salt, a black precipitate of stannous sulphide. 1 In the presence of oxalic acid, antimonic sulphide is precipitated while stannic salts are imperfectly precipitated by hydrogen sulphide. If the solution be hot, and a sufficient excess of oxalic acid be present, the stannic sulphide is not precipitated at all. Influence of Tartaric Acid. When this reaction is employed for the separation of antimony and tin, it is necessary to oxidise the stannous salts to stannic. This is conveniently done by means of hydrogen peroxide. Tin, however, may be carried down with the antimony sulphide. Hence, the precipitate must be dissolved in sodium sulphide, more oxalic acid added, and the antimony sulphide again precipitated. The addition of an excess of tartaric acid prevents the precipitation of tin along with antimony sulphide, and renders a second pre- cipitation unnecessary. This is the principle upon which Clarke and Henz's process 2 for the separation of tin and antimony is founded. The presence of tartaric acid also retards the separation of chloride possibly SbOCl or SbSCl with the precipitate. The chloride, if present, volatilises when the precipitate is heated to 300 in a stream of carbon dioxide, and thus leads to low results. 3 The greater the amount of tartaric acid in the solution, the less the danger of loss from this cause. Thus, with the precipitate from a gram of antimony the loss on ignition was as follows : Tartaric acid used 2 '5 5 '5 grins. Precipitate lost on ignition .... 0'37 0*18 0'05 percent. 135. The Gravimetric Separation of Antimony and Tin- Clarke and Henz's Process. Preparation of the Solution for the Precipitation. Dissolve, say, the mixed sulphides containing less than the equivalent of O3 grm. of the mixed metals in a solution of sodium or potassium sulphide (page 277) in a 500-c.c. beaker. Add 6 grms. of the purest potassium hydroxide ; 4 3 grms. of tartaric acid ; 5 and as much hydrogen peroxide (30 per cent.) as is necessary to decolorise the solution. Then add a volume of hydrogen peroxide equal to that already added. Boil the solution for a few minutes until the evolution of oxygen is over, and the thio-salt 1 Stannons sulphide is black ; stannic sulphide yellow ; a mixture of the two, maroon colour. C. F. Barfoed, Zeit. anal. Chem., 7. 260, 1868 ; T. Scheerer, Journ. prakt. Ghem. (2), 3. 472, 1871 ; A. Carnot, Compt. Rend., 103. 258, 1886. 2 F. W. Clarke, Amur. J. Science (2), 49. 48, 1870 ; Chem. News, 21. 124, 1870 ; Zeit. anal. Chem. 9. 487, 1870 ; 21. 114, 1882 ; A. Rossing, ib , 41: 1, 1902 ; G. Vortmaim and A. Metzl. ib., 44. 525, 1905; E. Lesser, ib., 27. 218, 1888; J. A. Miiller, ib., 34. 171, 1895 ; A. Czerwek, ^.,45. 505,1906; F. Henz, Zeit. anorg. Chem., 37. 1, 1894; A. Fischer, ib., 42. 372, 1903 ; A. Gutbier and C. Brunner, Zeit. angew. Chem., 17. 1137, 1904; C. Ratner, Chem. Ztg., 26. 873, 1902; J. Clark, Journ. Soc. Chem. Ind.,1^. 255, 1896; R. Buusen, Liebig's Ann., 192. 317, 1878 ; Zeit. anal. Chem., 18. 261, 1879 ; A. Carnot, Compt. Rend., 103. 258, 1886 ; Chem. News, 54. 89, 1886 ; H. N. Warren, ib., 62. 216, 1890; G. C. Wittstein and A. B. Clark, Vierleljdlir. prakt. Pharm., 19. 551, 1870 ; F. P. Dewey, Amer. Ghem. Journ., I. 244, 1879 ; Chem. News, 40. 257, 1879 ; 0. Klenker, Journ. prakt. Chem. (2), 59. 353, 1899 ; T. Brown. Journ.. Amer. Chem. Soc., 21. 780, 1899; Chem. News, 8l. 178, 184, 1900; C. Hallmann, Vergleichende Untersuchmig iiber Methoden der quantitatiren Antimon lestim- mung, Aachen, 1911 ; A. Inhelder, Beitrag zur Trennung des Antimons und Ziwiis und zur Analyse von Lagermetallen, Zurich, 1911. 3 H. Rose, Pogg. Ann., 98. 455, 1856; 0. Petriciolli and M. Heuter, Zeit. angew. Chem., 14. 1179, 1901 ; L. A. Youtz, Journ. Amer. Chem. Soc., 30. 975, 1908. 4 That is, one-third the sum of the weights of the mixed tartaric and oxalic acids. Use potassium hydroxide "pure by alcohol." 5 That is, 10 times the maximum weight of the mixed metals tin and antimony in the solution, THE DETERMINATION OF ANTIMONY. 297 is oxidised. 1 Cool. Cover the beaker with a clock-glass; add 15 grms. of pure oxalic acid. 2 Carbon dioxide comes off vigorously. Boil the solution briskly for 10 minutes to decompose the hydrogen peroxide completely. The solution, now occupying about 80 or 100 c.c., is allowed to cool. Precipitation of Antimony Sulphide. Pass a rapid current of hydrogen sulphide through the cold solution and then heat it to the boiling point, all the time maintaining the current of hydrogen sulphide. Keep the flask or beaker in a boiling water bath. 3 In about 15 minutes dilute the solution to about 250 c.c. Continue the current of gas another 15 minutes. Remove the source of heat, and continue the current of gas another 10 minutes longer. Let the precipitate settle until the solution is cold. Washing the Precipitate. Decant the dense precipitate of antimony sulphide through a Gooch's crucible 4 which has been heated to about 300 in a stream of carbon dioxide for about an hour before cooling in the desiccator and weighing. Wash the precipitate twice by decantation with a 1 per cent, solution of oxalic acid, and twice by decantation with very dilute acetic acid. Both wash- ing liquids should be boiling and kept saturated with hydrogen sulphide. 5 The filtering and washing should be done as quickly as possible, since the antimony sulphide is liable to decompose or deflocculate and pass into solution. 6 The pre- cipitate will be contaminated with some sulphur. It is best to remove 7 most of the sulphur by washing three times with alcohol, then with a mixture of equal parts of alcohol and carbon disulphide, then with alcohol, and finally with ether. The Ignition of the Precipitate. Dry the precipitate and heat it gradually to between 250 and 300 in a stream of carbon dioxide for between 30 and 60 minutes. The heating is conveniently done in Paul's drying oven, 8 fig. 127. The sulphur in the precipitate is volatilised, and the antimony pentasulphide is transformed into the grey trisulphide. The sulphur will all have volatilised in about an hour. 9 The crucible is cooled and weighed. When two successive weighings do not differ by more than 0'0005 grm., the operation is complete. The weight of the precipitate Sb 2 S 3 multiplied by 0*8568 represents the corresponding amount of Sb 2 3 . 1 All the peroxide cannot be decomposed at this stage. 2 That is, 50 times the weight of the mixed metals in the solution. The hot solution should be saturated with oxalic acid. 3 If the solution be kept cold, the antimony sulphide will be difficult to filter so as to form a clear solution ; if the gas be passed at once into the boiling solution, the precipitate is liable to stick tenaciously to the sides of the beaker ; and if the gas be passed into the cold solution and the solution gradually heated to boiling during the passage of the gas, the precipitate is granular, and washes easily S. P. Sharpies, Amer. J. Science (2), 50. 248, 1870 ; Chem. News, 22. 259, 1870. 4 If arsenic be .present, it too will be precipitated as sulphide. Filter tubes, figs. 53 B and C, can be used. W/Gibbs and E. R. Taylor, Amer. J. Science (2), 44. 215, 1867 ; R. Fresenius, Zeit. anal. Chem., 8. 155, 1869. 5 At this stage the antimony can usually be determined volumetrically, with a great saving in time, if a standard solution is ready. 6 C. Friedheim and P. Michaelis, Zeit. anal. Chem., 34. 505, 1895 ; J. Theile, ib., 30. 479, 1891. The treatment described in the text is better than washing with carbon disulphide alone, since the latter may be entangled with the precipitate, and not pass through the filter paper. Antimonic sulphide Sb 2 S 5 is not decomposed into antimonious sulphide SbgSg and sulphur by the treatment with carbon disulphide. For an extraction apparatus, see fig. 134, page 343. 7 See page 343 for a discussion on removing sulphur from sulphides. 8 T. Paul, Zeit. anal. Chem., 31. 537, 1892. In fig. 127, T represents the thermometer, B a Habermann's wash-bottle containing sulphuric acid for cleaning the carbon dioxide. The carbon dioxide must be free from air, or some of the antimony will be oxidised to Sb 2 4 . F. Henz's gas generator (Chem. Ztg. , 26. 386, 1902) is one of the best " C0 2 generators " for this experiment. 9 The heating should not extend much more than an hour, or some antimony sulphide may be lost by volatilisation. See L. A. Youtz, Journ. Amer. Chem. Soc., 30. 975, 1908. 298 A TREATISE ON CHEMICAL ANALYSIS. EXAMPLE. The sulphide from 1 grm. of material, after half an hour's ignition in Paul's oven, as indicated above, gave, on first weighing, 0*02423 grm. Sb 2 S 3 . After ignition for another 15 minutes, the sulphide weighed 0*02418 grm. Hence, 0'0242 x 0-8568 = 0-0207 grm., that is, 2'07 per cent of Sb 2 3 . FIG. 127. Ignition of antimony sulphide. The Accuracy of the Process. One objection to this process is the great amount of oxalic and tartaric acid required to keep the tin in solution. 1 The chief errors arise from (1) the presence of free mineral acids ; (2) the use of too concentrated solutions ; and (3) error in assuming the ignited precipitate is normal trisulphide, j Sb 2 S 3 . In illustration of the kind of separations which can be obtained with this method, the following numbers are cited for mixtures of known amounts of antimony and tin : Table XLVI1. Test Analyses of Mixtures of Antimony and Tin. Antimony. Tin. Used. Found. Used. Found. 0-0463 0-0463 0-0924 0-1855 0462 0-0461 0-0923 0-1853 0-2555 0-1017 0-0103 0-1017 0-2532 1011 0-0113 0-0999 1 W. Dancer (Journ. Soc. Chem. IncL, 16. 403, 1897) and J. Marburg (Zeit. anal. Chem., 39. 47, 1900) first precipitate the tin with an excess of lirne water. THE DETERMINATION OF ANTIMONY. 299 If the antimony is to be separated from a complex enamel, the values for the antimony will not, of course, be so concordant as this. The tin was here determined by the electrolytic process, page 312. The method thus gives good results, although it is somewhat laborious. Consequently, volumetric processes are employed wherever practicable. The volumetric process, however, is not satisfactory when less than 1 per cent, of antimony is present. For many purposes the " iron precipitation " of antimony is the most convenient process for the separation of antimony from tin. 1 Effect of Molybdenum. The separation of arsenic and tin is even sharper than the separation of antimony and tin by Clarke's process. To separate molybdenum from "stannic" tin requires a modification of- the process. Antimony and molybdenum are precipitated together. 2 "By adding an alkaline sulphide in excess to a solution containing a molybdate," says Clarke, "and then decomposing the thio-salt with a considerable quantity of dilute hydrochloric acid, and allowing the whole to stand overnight in a warm place, the molybdenum is precipitated. The sulphide thus obtained can be easily washed with a mixture of dilute hydrochloric acid and ammonium chloride. If now, by this process, we throw down tin and molybdenum together, every trace of the former metal may be dissolved out by boiling the mixed sulphides for about three-quarters of an hour with oxalic acid (20 grms. of oxalic acid per gram of tin). It is best to have present in the solution, while boiling, a little dilute hydrochloric acid. If antimony be present, it is necessary, just before ceasing to boil, to add to the solution an equal volume of a saturated solution of hydrogen sulphide to reprecipitate any antimony which may have gone into solution." 3 136. Waller's Volumetric Iodide Process for Antimony. The solution containing the antimony and tin may be divided into two parts : the antimony determined in one part by Weller's or by Gyory's process, and the tin in the other part by one of the processes indicated on page 310 et seq. In Weller's process 4 the antimony is all oxidised to antimony penta- chloride ; mixed with potassium iodide ; and the amount of iodine liberated by the reduction of the antimonic to antimonious chloride determined by titration with sodium thiosulphate, as indicated on page 352, or by stannous chloride the alternative process here described. Which process thiosulphate or stannous chloride is used will often be determined by the most convenient standard solutions. The Oxidation of Antimonious to Antimonic Chloride. The sulphide is trans- ferred to a 600-c.c. Erlenmeyer's flask, and dissolved by boiling with a mixture of concentrated hydrochloric acid and potassium chlorate. The latter is added in small portions at a time. The object is to convert the antimony trichloride 1 For the " rapid " electro-deposition of both antimony and tin from the ammonium sulphide solution of the two sulphides ; digestion of the mixed metals witli aqua regia ; and titration of the antimony by Weller's process, see D. J. Demorest, Journ. Ind. Eng. Chem., 2. 80, 1910 ; Chem. News, 101. 260, 1910. 2 For the separation of antimony and molybdenum, see page 415. 3 Tungsten gives discordant results. Sometimes the tungsten sulphide seems to dissolve like tin ; at others, the solution is only partial F. W. Clarke (I.e.). 4 A. Weller, Liebig's Ann., 213. 364, 1882 ; L. A. Youtz, School Mines Quart., 24. 135, 407, 1903; Zeit anorg. Chem., 37. 337, 1903 ; A. Kolb and R. Formhals, ib., 58. 189, 202, 1908 ; E. Schmidt, ib., 34. 453, 1910 ; H. Causse, Compt. Rend., 125. 1100, 1897 ; H. Giraud, Bull. Soc. Chim. (2), 46. 504, 1886 ; G. Rollin, Ann. Chim. Anal. App., 6. 114, 1902 ; G. von Knorre, Zeit. angew. Chem., I. 155, 1888; M. Rohmer, Ber., 34. 1565, 1901; J. Darroch, Chem. Eng., 4. 162, 1906 ; G. S. Jamieson, Journ. Ind. Eng. Chem., 3. 250, 1911. 300 A TREATISE ON CHEMICAL ANALYSIS. SbCl 3 into the pentachloride SbCl^ 1 When the free chlorine and chlorine oxides have been driven off by boiling, and the volume of the solution is about 50 c.c., let the solution cool. When cold, add 20 c.c. of concentrated hydro- chloric acid, and dilute the solution to about 200 c.c. with recently boiled distilled water (cold). The Liberation of Iodine. Add, say, 3 grms. of potassium iodide dissolved in 10 c.c. of water. Stir the solution thoroughly. The antimony pentachloride is reduced to the trichloride. The reaction is represented : SbCl 5 + 2KI^:SbCl 3 + 2KC1 + I 2 . The solution now has a brown colour due to the liberated iodine. The strong acidity of the solution prevents the precipitation of antimony oxychloride. Dilute the solution to about 400 c.c. Titration with Stannous Chloride. Titrate the solution, at once, 2 with a standard solution of sodium thiosulphate or of stannous chloride. 3 The solution should be well shaken during the titration, particularly towards the end. When the end point is near, the solution becomes pale yellow. Then add a couple of drops of starch solution (page 286). Shake the solution vigorously, and titrate until the blue colour of the starch iodide has disappeared. The end point is very sharp, and, consequently, the standard solution must be added cautiously drop by drop. If any after-bluing occurs, it will be due to secondary reactions and may be ignored. The reaction which takes place during the titration may be represented : SnCl 2 + 1 2 + 2HC1 = SnCl 4 + 2HL Influence of Foreign Metals. Arsenic reacts in a similar manner to antimony, and cupric salts also liberate iodine from potassium iodide. 4 Bismuth iodide resembles the colour of free iodine in solution and consequently obscures the end point, but does not otherwise interfere. Tin, as stannic chloride, does no harm. 1 Any stannous salts which may be present are also oxidised to stannic salts. 2 Iodine titrations are usually made in Erlenmeyer's flasks, not in beakers, on account of the volatility of the iodine during the titration. A large excess of potassium iodide lessens the danger of losing iodine in this way. For a similar reason, iodine titrations should be per- formed immediately the iodine has separated ; and the solution should be cold. J. Wagner (Zeit. anal. Chem., 27. 137, 1888 ; C. R. A. Wright, Chem. News, 21. 163, 1870) found that with the same amount of solution of sodium thiosulphate : Temperature . . 16 30 38 50 75 83 91 Iodine sol. . . . 39'6 39'8 40'0 401 407 40'9 41'4c.c. These numbers show that the higher the temperature, the less the amount of thiosulphate required in titrating a given amount of iodine ; and conversely. E. Sherer, Chem. News, 21. 141, 1870. 3 STANDARD STANNOUS CHLORIDE. Dissolve, say, 10 grms. of stannous chloride in 60 grms. of hydrochloric acid (sp. gr. 1'12). Dilute the solution to a litre. Dissolve 0'2 grm. of metallic antimony in concentrated hydrochloric acid and potassium chlorate (or bromine) in the cold. Metallic antimony is but slowly attacked by hot hydrochloric acid and potassium chlorate. Boil the solution to get rid of the chlorine oxides (or bromine). Cool. Add potassium iodide as indicated in the text. The titration is made as indicated in the text. The antimony equivalent of the standard solution of stannous chloride is thus simple arithmetic. Weller titrates with standard sodium thiosulphate, but I prefer the stannous chloride, if the preparation of the solution is not inconvenient. The objection to the stannous chloride titration is the instability of the solution due to oxidation. The objection to the thiosulphate titration rests on tlie known reaction between sodium thiosulphate and strong acids. For the sodium thio- sulphate titration, see page 352. 4 Arsenic can be removed by boiling the solution with concentrated hydrochloric acid ; copper and bismuth can be removed by washing the precipitated sulphides on a filter paper with sodium sulphide. THE DETERMINATION OF ANTIMONY. 301 Influence of Hydrochloric Acid. Too much hydrochloric acid gives high results owing to the action of the acid on the potassium iodide. The solution should not contain much more than about one-fifth of its volume of concentrated hydrochloric acid (sp. gr. 1-16). Too little hydrochloric acid leads to the separation of basic chlorides or iodides of antimony. By following the above directions boiling down to 50 c.c. an acid of constant strength is obtained approximately 20 per cent. HC1 and diluting the solution as indicated, satisfactory results will be obtained. If the acidity of the solution used for standardising the stannous chloride be the same as that of the sample under investigation, and the general treatment be the same, 1 determinations can be made more quickly and as accurately as the gravimetric process when over 1 per cent, of antimony is present. If less than 1 per cent, be present, use the gravimetric process of Clarke. Relation of Weller's to Mohr's Process. Mohr's process for arsenic (page 290) is quite satisfactory when applied to antimony ; and Weller's process for antimony is quite satisfactory when applied to arsenic. 2 A comparison of the two processes is interesting. The same general equation applies to both : R 2 3 + 2I 2 + 2H 2 0^rR 2 5 + 4HI <-Weller In Mohr's process, the reverse action indicated in the equation is prevented by keeping down the concentration of the hydriodic acid by means of an excess of alkali (sodium bicarbonate) in the solution. In Weller's process, the reverse action is prevented by the relatively strong acidity of the solution. Comparison Determinations. To illustrate the results obtained in a separa- tion of antimony and tin, the following duplicate determinations of the percentage amount of antimony with each of the processes here recommended, might be quoted : Henz and Clarke's process. Weller's process. Gyory's process. 26-96 26'33 26'52 2610 26-23 26-59 137. Gyory's Volumetric Bromate Process for Antimony. This process 3 may be employed as an alternative to that which precedes. Chemists who use this process regularly are enthusiastic about its merits. The sample under investigation is dissolved in a 250-c.c. Erlenmeyer's flask with 20 c.c. of concentrated hydrochloric acid and a little potassium chlorate or a few drops of bromine. Boil the solution. Reduction of the Antimonic and Arsenic Salts. Add, say, 0'75 grm. of crystalline sodium sulphite in order to reduce the antimonic to the antimonious chloride, and the arsenic to arsenious chloride. 4 Boil vigorously until the solution is reduced to about half its former volume. This drives off the sulphur 1 Some prefer to make a blank test and find the amount of stannous chloride or sodium thiosulphate needed when no antimony is present. For the influence of organic substances on iodine titrations, see J. Klaudie, Listy ChemicJce", 12. 91, 1888. Fatty acids and sugar have no action ; aldehyde, phenol, tannin, and the aromatic alcohols give high results. 2 T. Smith, Journ. Amer. Chem. Soc., 21. 769, 1899; F. A. Gooch and P. E. Browning, Amer. J. Science (3), 40. 66, 1891. 3 S. Gyory, Zeit. anal. Cliem., 32. 415, 1893 ; H. Nissenson and P. Siedler, Chem. Ztg.,2fj. 749, 1903; E. Schmidt, ib., 34. 453, 1910; J. B. Duncan, Chem. News, 95. 49, 1907; H. W. Rowell, Journ. Soc. Chem. lnd. t 25. 1181, 1906 ; F. Foerster, Zeit. Elcktrochem. , 15. 232, 1909 ; A. Christensen, Pharm. Ztg., 41. 326, 1896. 4 The stannic chloride is not reduced. 3O2 A TREATISE ON CHEMICAL ANALYSIS. dioxide, 1 and volatilises the arsenic. Rinse the sides of the flask with hot water, and add 10 c.c. of hydrochloric acid to the solution. The Titration. Heat the solution to about 80 or 90*, and titrate with a standard solution of potassium bromate 2 until nearly all the antimony is oxidised to antimonic chloride. 3 Add 3 drops of methyl orange, and continue the titration until the tint of the methyl orange is destroyed. 4 The solution should be thoroughly agitated during the titration, so that local excesses of bromate are not formed. 5 If very little bromate be needed, the indicator may be added before the litration. This, however, is an extreme case. If over 10 c.c. of bromate are needed, it is best to run in the greater part of the bromate before adding the indicator. The indicator has a tendency to fade, and if it be used from the start on an " unknown," it is necessary to add more from time to time as the titration progresses. The first titration may be used for finding the approximate amount of bromate needed. The solution is not run from the burette much faster than 30 c.c. per minute. The reaction is represented by the equation : KBr0 3 + HC1 + 3SbCl 3 = KC1 + HBr + 3Sb + 3H 2 0. Any further addition of bromate destroys the colour of the methyl orange, probably owing to the liberation of bromine by the action of free acid on the bromate, for if bromate be added after the methyl orange has lost its colour, the yellow colour of the bromine appears in a short time. Influence of Foreign Substances. The process works well in the presence of lead, zinc, tin,- silver, chromium, and sulphuric acid, since these substances have no appreciable effect on the result. The presence of large amounts of calcium, magnesium, and ammonium salts furnishes high results. Iron and copper are partly reduced by the sodium sulphite in the solution, and hence react with the bromate, giving high results. .Rod well states that 1 per cent. of iron raises the amount of antimony by 0*02 per cent., but 5 per cent, of iron raises the amount of antimony but little more. Rodwell also states that, for every O'l per cent, of copper in the sample up to 1 per cent., 0'012 per cent, of antimony should be subtracted. 6 Accuracy of the Process. The most important sources of error are: (1) incomplete expulsion of sulphur dioxide ; (2) the imperfect volatilisation of arsenic, if present ; and (3) over-titration, when too little hydrochloric acid is present, owing to the slowness of the reaction. In illustration of the accuracy of the process, the following determinations, by Rowell, may be quoted : 1 The amount of antimony "lost" by volatilisation or reoxidation is inappreciable. Most of the arsenic will be volatilised, since arsenious chloride is fairly volatile from boiling solutions. See page 271. 2 POTASSIUM BROMATE SOLUTION. Dissolve, say, 27852 grms. of the pure salt in a litre of water. To standardise the solution, dissolve 0'2 grm. of metallic antimony in hydrochloric acid and potassium chlorate as indicated, page 300. The reduction and titration of the resulting solution are conducted as indicated in the text. 3 If lead be present, and the lead chloride separates as the solution cools, boil the solution again in order to keep the lead chloride in solution while the titration is in progress. 4 Some prefer to use, as indicator, a solution of indigo made by dissolving powdered indigo in fuming sulphuric acid. Neutralise the solution with calcium carbonate ; dilute with 10 times its volume of water ; and filter the blue liquid. Add 3 drops of this solution at the start, and more when the reaction is nearly completed. The blue passes through various shades of greenish yellow into pale green. The colour is discharged with a drop of bromate in excess. The indigo solution is not so sensitive an indicator as the methyl orange, and the results with the indigo are a little higher than with methyl orange. 5 Otherwise the results will be high owing to the loss of bromate before it attacks the antimony. 6 For the reduction of the bromate by stannous salts, see F. Fichter and E. Miiller, Chem. Ztg., 37. 309, 1913. THE DETERMINATION OF ANTIMONY. Table XL VIII. Test Analyses of Antimony Ores. 33 Nature of mixture. Antimony (per cent.). Volumetric bromate process. Gravimetric (as Sb 2 S 3 ). Antimony sulphide .... Stibnite (As, Fe) . Lead : antimony . . . Lead : arsenic : antimony Lead : antimony : tin . 71-40 71-20 9-42 8 65 2-13 71-35 71-35 9-37 870 2-20 The analysis can thus be conducted with an accuracy approaching 0*1 per cent, in samples containing 10 per cent, of antimony. It may be a little FIG. 128. Reflux condenser (see page 582). difficult at first to get concordant results, but once the manipulation is mastered, the process gives \ little trouble. J. B. Duncan informs me that "duplicates on a straight assay running 99'6 per cent, metallic antimony should agree exactly." The Influence of Arsenic. The process indicated above can be employed for the determination of arsenic, but the boiling of the solution in open vessels to expel the sulphur dioxide will not do. The boiling must be conducted in a 304 A TREATISE ON CHEMICAL ANALYSIS. flask fitted with a reflux condenser (fig. 128). If the antimony is to be determined, and arsenic is present in the sample, the latter must be removed, as indicated above, by boiling the solution down to half its volume in presence of sodium sulphite. If more than 2 or 3 per cent, of arsenic be present, add 20 c.c. more concentrated hydrochloric acid, and 5 c.c. of a saturated aqueous solution of sulphur dioxide. Boil the liquid down again. There is no danger of an appreciable loss of antimony if the temperature be kept below 120. An artificial mixture of antimony and arsenic so treated furnished 91*09 per cent, of antimony, when 90 '91 per cent, was actually present. 138. The Volumetric Determination of Antimony and Arsenic in presence of Tin. The arsenic and antimony in a solution containing antimony, arsenic, and tin can be separately determined by the bromate process. The mixture is dissolved and oxidised as indicated for Gyory's process. Make the solution up to, say, 200 c.c. Boil off the arsenic in an aliquot portion, say, 100 c.c., as indicated above, and titrate the solution for antimony with potassium bromate. The arsenic and antimony can be determined together in the other aliquot portion by proceeding as indicated above, but boiling down the solution in a reflux condenser (fig. 128). The difference in the two titrations represents the potassium bromate which is equivalent to the arsenic. 1 EXAMPLE. A solution was made up to 200 c.c. as indicated above. The antimony in 100 c.c. required 24 '7 c.c. of the potassium bromate solution. This corresponds with 24'7 x 0-0006432 = 0'01 59 grm. of Sb 2 3 in half the sample. The other portion, boiled with a reflux condenser, required 52'1 c.c. of the bromate solution. Hence, 52*1 24 7 = 27*4 c.c. of the bromate were required for the arsenic titratioii 27'4x 0-000441 7 = 0-0121 grm. of As 2 3 in half the sample. 139. Metallic Precipitation. In the ordinary zinc : carbon cell, if the metallic zinc be replaced successively by manganese, aluminium, and magnesium, the voltage of the cell is increased ; and conversely, if the zinc be similarly replaced by cadmium, iron, and cobalt, rttage is diminished. The order in which the metals can be thus arranged is the same as the order in which the metals displace one another in their salts. The order is approximately: K, Na, Mg, Al, Mn, Zn, Cd, Fe, Co, Ni, Sn, Pb, Sb, Bi, As, Cu, Hg, Ag, Ft, Au, and the series is called the electrochemical series of the metals. The order varies a little with different solutions. A metal on the left in the series will generally displace another metal on the right from its salt solution. Secondary reactions may prevent the actual precipitation of the metal. In many cases the displacement is so complete that the reaction can be used in quantitative analysis. What metals are precipitated often depends upon the acidity of the solution as well as on the metal used as precipitating agent. In illustration of the more important metals used as precipitating agents, magnesium precipi- tates zinc, cadmium, thallium, iron, cobalt, nickel, tin, lead, 2 antimony, 3 bismuth, 1 Rather more hydrochloric acid is required in the case of antimony than arsenic in order to prevent the precipitation of antimony during the titration, as a result of the increasing dilution of the solution. 2 W. Schulte, Metallurgie, 6. 214, 1909. 3 S. Kern, Chem. News, 32. 309, 1875 ; E. G. Bryant, ib., 79. 75, 1899. THE DETERMINATION OF ANTIMONY. 305 copper, 1 mercury, silver, 2 platinum, and gold. Some part of the antimony may be evolved as a gas when solutions containing antimony are treated with magnesium or zinc ; and arsenic is also lost in this way when its solutions are treated with zinc, magnesium, and iron. 3 Aluminium precipitates lead, 4 antimony, 5 tin, copper, and silver, 6 etc. Zinc precipitates cadmium, cobalt', 7 nickel, 7 tin, lead, 8 antimony, bismuth, arsenic, copper, 9 mercury, platinum, gold, etc. ; but not iron. 10 Cadmium precipitates copper, lead, silver, 11 antimony, tin, etc. Iron precipitates bismuth, 12 antimony, 18 copper, mercury, silver, lead, gold, etc. In the presence of stannic chloride, arsenic is also precipitated. Metallic iron also reduces ferric to ferrous salts, stannic to stannous salts, etc. Tin precipitates antimony, arsenic, copper, mercury, silver, gold, etc. Copper pre- cipitates mercury, silver, gold, etc., but not tin ; lead precipitates bismuth, 14 etc. Precipitation of Antimony in the Presence of Tin Tookey's Process. The separation of antimony from tin by precipitation of the antimony, as metal, from a hot hydrochloric acid solution by means of metallic iron 15 was proposed by Tookey in 1862. 16 The process can be conducted in the following manner : Dissolve the mixed sulphides in a warm mixture of hydrochloric acid and potassium chlorate, and boil off the chlorine. The solution should contain about 12 per cent., and not less than 2 per cent, of hydrochloric acid, or some tin may be precipitated. Place some coarse granules of pure 17 metallic iron in the 1 A. Villiers and F. Borg, Bull. Sac. Chim. (3), 9. 602, 1893 ; Chem News, 68. 263, 1893. 2 E. G. Bryant, Chem. News, 79. 75, 1899. 3 A. Roussin, Journ. Pharm. Chim. (4), 3. 413, 1866 ; Chem. News, 14. 27, 1866. W. N. Hartley, ib., 14. 73, 1866 (failed to precipitate iron with magnesium) ; S. Kern, ib., 33. 236, 1876 ; T. L. Phipson, ib., 9. 219, 1864 ; A. Commaille, Compt. Rend., 63. 556, 1866 ; F. Clowes and R. M. Caven, Chem. News, 76. 297, 1897 ; E. G. Bryant, ib., 79. 75, 1899. J. G. Hicks and E. F. Smith (Journ. Amer. Chem. Soc., 16. 822,' 1894) have studied the action of magnesium on manganous salts. 4 A. H. Low, Journ. Anal. App. Chem , 4. 12, 1891 ; 6. 664, 1892 ; J. E. Williams, Eny. Min. Journ., 53. 641, 1892. 5 W. Schulte, Metallurgie, 6. 214, 1909. 6 N. Tarugi, Qaz. Chim. Red., 33. ii. 223, 1903. 7 J. L. Davies, Journ. Chem. Soc., 28. 311, 187:,. 8 F. Stolba (Journ. prakt. Chem. (1), 101. 150, 1867 ; Chem. News, 17. 2, 1868) precipi- tates lead from lead salts quantitatively by metallic zinc in the presence of hydrochloric acid on a water bath. L T. Merrill, Eng. Min. Journ., 91. 56, 1911 ; A. Eckenroth, Pharm. Ztg. t 40. 528, 1895 ; F. Mohr, Zeit. anal. Chem., 12. 142, 1873; C. Rossler, ib., 24. 1, 1885. a J. C. Shengel and E. F. Smith, Journ. Amer. Chem. Soc., 21. 932, 1899 ; Chem. Neivs, 8l. 134, 1900. 10 M. Demargay, Bull. Soc. Chim (2), 32. 610, 1879. 11 A. W. Clasen, Journ. prakt. Chem. (1), 97. 217, 1866 ; Chem. Neivs, 13. 232, 1866. 12 J. Clark (Journ. Soc. Chem. Ind., 19. 26, 1900) for the separation of bismuth and lead. J. G. Gallety and G. C. Henderson (Analysis, 34. 389, 1909) recommend the process using a hot solution containing 2J per cent, of free nitric acid 13 A. Thiel and K. Keller, Zeit. anorg. Chem., 68. 42, 1910. 14 C. Ullgren, Berzelius > Jahrb., 21. 148, 1842 ; A. Patera, Zeit. anal, Chem., 5. 226, 1866, separate bismuth from lead in this way ; but according to 0. Steen (Zeit. angew. Chem., 8. 531, 1895) the error lies between 20 and 30 per cent. 15 J. L. Gay Lussac precipitated antimony in the presence of tin by means of metallic tin in hydrochloric acid solution J. L. Gay Lussac, Ann. Chim. Phys (2), 46. 222, 1831 If zinc is used instead of iron, some antimony may be lost as hydride. No antimony is lost as hydride 1870 ; A. W. Clasen, Journ. prakt. Chem. (1), 92. 477, 1864 ; Chem. News, 13. 232, 1866 ; 0. Low, Vierteljahr. prakt. Chem., 14. 406, 1864; A. Carnot, Compt. Rend., 114. 587, 1892; J. H. Mengin, ib., 117. 224, 1894; M. Hoffmann, Beitrdge zur Kenntnis der analytischen Chemie des Zinns, Antimons, und Arsens, Berlin, 27, 1911. - 7 C. Rammelsberg recommends iron reduced in hydrogen gas. Bright piano wire will do quite well. 20 306 A TREATISE ON CHEMICAL ANALYSIS. solution very nearly at the boiling temperature. Add more hydrochloric acid as the iron dissolves. When all the antimony is precipitated, 1 filter the solution as quickly as possible, since some antimony may redissolve when the solution is exposed to the air. An excess of iron is necessary, and the precipitated antimony must be mixed with the iron during the filtration. 2 The mixture of metallic antimony and iron can then be dissolved in a mixture of hydrochloric acid and potassium chlorate, and the antimony determined by precipitation as sulphide, etc. Working with known mixtures of stannic chloride and antimony trichloride, Hoffmann found the following results : Antimony O'l 0'2 0'4 O'l 0'2 0'4 grm. Tin used . . . 0'4343 0'4343 0'4343 0'8686 0-8686 0'8686 grm. Tin found. . . 0'4336 0'4337 0'4338 0'8684 0'8687 0'8685 grm. Error. . . . -O'OOO? -0'0006 -0'0005 -0'0002 +0'0001 -O'OOOl grm. The results are therefore quite satisfactory. 3 I have had no experience with the electrolytic processes for antimony. 140. The Evaluation of Antimony Compounds. Antimony oxide may be evaluated by a method similar to that used for the arsenic compounds (page 293). Add, say, O'l grm. of antimonious oxide to 20 c.c. of water ; heat the solution to boiling ; add tartaric acid in small quantities at a time until the oxide is completely dissolved. 4 Neutralise the solution with sodium bicarbonate ; and add 10 c.c. more of the sodium bicarbonate solution. Titrate with iodine using starch as indicator as described for Mohr's process for arsenic, page 290. 5 Since Sb 2 3 + 2H 2 + 2I 2 = 4HI + Sb 2 5 , every gram of iodine represents 0*564 grm. of Sb 2 3 . Antimonic compounds are reduced as indicated under Gyory's process, and estimated by Mohr's or by Gyory's process. 6 1 If the volume of the solution is small this will take about 20 minutes. 2 A. W. Clasen (I.e.) has shown that the precipitated antimony is perceptibly soluble in hot or cold hydrochloric acid of various strengths, and hence antimony may be "lost." By following the plan described in the text, this loss is made negligibly small. 3 The tin was determined in the nitrate, after the separation of the antimony, by precipita- tion as sulphide. G. Panajotow (Ber., 42. 12fe6, 1909) separates antimony from tin by pre- cipitating the antimony as sulphide from a solution containing 15 per cent, of HC1, at ordinary temperatures. The tin remains in solution. The results seem to be good A. Inhelder, Beitrag zur Trennung des Antimons und Zinns und zur Analyse von Lagermetallen, Zurich, 1911. 4 The tartaric acid keeps the antimony oxide in solution F. H. Alcock, Pharm, Journ., 362, 1900. 5 Titrate at once, or antimonious hydrate may be precipitated. 6 R. Rickmann (Zeit. angew. Chem., 2$. 1518, 1912) detects antimony in enamels, by cleaning the enamel free from adhering iron, and then boiling the powdered enamel with a 4 per cent, solution of acetic acid (or 2 per cent, tartaric acid). Divide the solution into two parts. Test one part with hydrogen sulphide, and titrate the other part with standard permanganate. If antimony is present, and no permanganate is consumed, it is inferred that the enamel contains a metantimoniate which Rickmann says is harmless. CHAPTER XXIII. THE DETERMINATION OF TIN. 141. The Metallic Precipitation of Tin. TIN may be conveniently precipitated in the metallic state by means of metallic zinc, aluminium, or cadmium. This reaction also offers a convenient method of separating tin in the nitrate from the antimony sulphide obtained in Clarke and Henz's process. 1 Boil off the hydrogen sulphide. Precipitation of Metallic Tin. Place a piece of zinc 2 (foil, ribbon, or stick) in the solution, and heat the solution nearly to the boiling point over a small flame. In about 20 minutes, test a few drops of the clear liquid for tin by means of "H 2 S" water. If necessary, heat the solution a little longer until all the tin is precipitated. Decant the liquid through a small filter paper, and transfer the metallic tin and zinc to the filter paper. Wash with water. 3 Conversion of the Metallic Tin into Stannic Oxide. Rinse the metallic tin and zinc into a 250-c.c. beaker. The total volume of the liquid will be about 10 c.c. Add 10 c.c. of nitric acid, and cover the beaker with a clock-glass. 4 When all the zinc is dissolved, and the tin oxidised, dilute the solution to 4050 c.c. Carefully heat the solution to boiling, and stir vigorously. Let settle. Add 5 grms. of ammonium nitrate ; filter and wash with a 5 per cent, solution of ammonium nitrate. Burn the filter paper in a crucible. Moisten the ash with nitric acid ; dry in a water bath ; ignite the precipitate, gently at first, and finally over a blast. Weigh the precipitate as Sn0 9 . Tin and antimony can be together precipitated by cadmium, and weighed. By digesting the mixed metals in hydrochloric acid, the tin dissolves, and antimony remains behind. The antimony can be washed, dried, and weighed. The amount of tin is obtained by difference. 5 Stannic salts are reduced to stannous salts by the metallic iron, but no tin is precipitated. 142. The Precipitation of Tin as Hydroxide Lowenthal's Process. There are two stannic oxides corresponding with Sn0 2 . The one is called a-stannic acid, or crmetastannic acid ; the other, /^-stannic acid, or /3-meta- 1 C. Ratner, Chem. Ztg., 26. 873, 1902; G. Buchner, ib., 18. 1904, 1894; L. Vignon, Compt. Rend., 107. 734, 1888. 2 Free from tin, For lead in place of zinc, see A. Pleischl, Dingler's Journ., 164. 200, 1862. 3 If a stick of cadmium is used for the precipitation, the cadmium can be easily separated from the tin. The latter is washed first with water ; then with alcohol ; melted to a button under stearic acid in a porcelain crucible ; and cleaned by washing with benzene L. Moissenet, Compt. Rend., 51. 205, 1860. 4 If the reaction be too violent, dilute the solution with cold water. If the action ceases, warm the solution. 5 J. H. Mengin, Compt. Rend., 119. 224, 1894; J. L. Gay Lussac, Ann. Chim. Phys. (2), 46. 222, 1831. 307 308 A TREATISE ON CHEMICAL ANALYSIS. stannic acid. 1 a-stannic acid dissolves quickly in cold dilute mineral acids nitric, hydrochloric, and sulphuric acids. The solution, on prolonged boiling, deposits the /3-acid, which is practically insoluble in dilute acids. 2 Salts corre spending with the two stannic acids are known. In Lowenthal's process 3 for the determination of tin, ammonia is added tci the solution containing stannic chloride 4 until a permanent precipitate just begins to form. Add dilute hydrochloric acid, drop by drop, with constant stirring, until the precipitate is all just redissolved. Add a cold saturated solution of j ammonium nitrate 5 to the solution and boil for some time. 6 When all the tin is precipitated, let the precipitate settle, and wash by decantation with a 5 per cent, solution of ammonium nitrate until the precipitate is free from chlorides. 7 After burning the filter paper, moisten the ash with nitric acid, evaporate to dryness on a water bath, and ignite on a gradually rising temperature. Finish the ignition on a blast ; 8 weigh as stannic oxide Sn0 2 . Tin is sometimes separated from its solutions by evaporating the solution to dryness with nitric acid. The tin is thus converted into the insoluble metastannic acid. Errors. There are several objections to this process The great difficulty is involved in the washing of the precipitate. If iron be present in the solution, the precipitate, after washing, is almost sure to be contaminated with appreciable quantities of iron oxide. 9 Other oxides may also contaminate the precipitate. Antimony hydroxide, for example, may be nearly all precipitated with the tin. Hoffmann 10 thinks that an insoluble tin antimoniate 3Sn0 2 .2Sb0 2 is formed. Arsenic also is precipitated, probably as tin arsenate n 2Sn0 2 . As 2 5 a reaction which Hoffmann has suggested for the gravimetric determination of arsenic. Lepez and Storch (I.e.) have shown that bismuth ,too, if present in considerable quantities, will contaminate the precipitated stannic oxide. Hence, in a general way, avoid the determination of tin by separating the tin 1 J. J. Berzelius, Ann. Ghim. Phys. (1), 87. 50, 1813 ; R. Engel, Compt. Rend., 124. 766, 1897 ; 125. 464, 651, 709, 1897 ; C. F. Barfoed, Zeit. anal. Chem., 7. 260, 1868 ; J. W. Mellor, Modern Inorganic Chemistry, London, 791, 1912. 2 D. B. Dott, Pharm. Journ. (4), 27. 486, 1908. 3 J. Lowenthal, Journ. prakt. Chem. (1), 56. 366, 1852. 4 If stannous chloride be present, oxidise with bromine or chlorine water. 5 Or sodium sulphate. H. Rose (Fogg. Ann., 112. 164, 1861 ; Chem. News, 5. 87, 1862) used sulphuric acid for the precipitation. 6 To make sure all the tin is precipitated, add a few drops of the clear solution to a hot solution of ammonium nitrate or sodium sulphate. If all the tin is precipitated, no further precipitation will occur. 7 If washed with boiling water, a turbid nitrate may be produced. R. Bunsen's plan (Liebig's Ann., 106. 13, 1858) of washing with ammonium acetate or ammonium nitrate removes this difficulty. For the colloidal stannic acids, see J. M. van Bemmelen, Zeit. anorg. Chem., 23. 124, 1900. M. Liebschutz (Chem. News, 102. 213, 1910) adds a dilute solution of albumen to the solution from which the colloidal metastannic acid is to be precipitated, and heats the solution for a short time. The stannic acid is entangled with the albumen as the latter curdles. The curd is filtered at once. Any copper, etc., entangled with the tin is removed by boiling the solution with dilute nitric acid. The method is also recommended for the sulphides of zinc, lead, etc. Paper pulp acts well with this and similar precipitates. See also pages 96 and 179. 8 Mere heating to redness does not suffice to expel all the water J. B. Dumas. Liebig's Ann., 105. 104, 1858. 9 According to H. Rose (Pogg. Ann., 112. 169, 1835) the iron becomes soluble when the solution is evaporated to dryness. C. Lepez (Monats. Chem., 10. 283, 1889) shows that iron, chromium, and cerium lead to the incomplete precipitation of metastannic acid in dilute nitric acid solutions ; while aluminium, uranium, cobalt, nickel, and copper do not. F. H. van Leent (Monit. Scient. (4), 12. 866, 1899 ; Chem. News, 78. 320, 1898) states that chromium and aluminium, as well as iron, retard the precipitation of metastannic acid. 10 M. Hoffmann, Beitrdge zur Kenntnis der analytischen Chemie des Zinns, Antimons. und Arsens, Berlin, 45, 1910. 11 E. Haeffely, Phil. Mag. (4), 10. 220, 1855. THE DETERMINATION OF TIN. 309 as stannic acid by the action of nitric acid, etc., since the stannic oxide may be contaminated with silica, phosphates, arsenates, antimoniates, lead, bismuth, ferric oxides, etc. 143. The Precipitation of Tin as Sulphide. The precipitation of tin by metallic zinc is much easier than the usual process of separation by precipitation as sulphide or hydroxide and weighing as stannic oxide. Whenever convenient, analysts avoid the precipitation of tin both as sulphide and hydroxide (stannic acid). Precipitation of Tin Sulphide. The combined nitrates containing the tin are neutralised with ammonia, acidified with acetic acid, and saturated with hydrogen sulphide. The stannic sulphide is but very slightly soluble in acetic acid solution containing oxalates. 1 Let the precipitate stand in a warm place over-night. 2 Decant the clear liquid through a Gooch's crucible, and wash the precipitate four or five times by decantation with a solution of ammonium nitrate or acetate. 3 Finally transfer the precipitate to the crucible and wash free from chlorides. Conversion of Sulphide into Oxide. 4 Dry the Gooch's crucible, and heat the crucible very gently in its saucer with free access of air until the smell of sulphur dioxide is no longer perceptible. 5 If the temperature be raised too rapidly, at this stage fumes of stannic sulphide would be evolved and burnt to stannic oxide, consequently giving low results. The crucible should be covered with a lid to prevent loss by decrepitation during the earlier stages of the ignition. Remove the lid, and raise the temperature gradually with free access of air. Finish off by blasting the crucible for half an hour. Cool the crucible in a desiccator. Removal of Sulphates. The stannic oxide Sn0 2 so formed holds a little sulphuric acid 6 very tenaciously. Hence, weigh the crucible. Place a piece of ammonium carbonate about half a centimetre diameter in the crucible, and blast it again. This treatment with the ammonium carbonate is repeated until two successive weighings do not differ by more than 0*0005 grm. 7 The result is generally a little too high because of the adsorption of alkalies mentioned above, and the difficulty which attends the complete conversion of tin sulphide to tin oxide. A little sulphate also is nearly always retained by the oxide. As a consequence of these facts, the gravimetric determination of tin as sulphide is employed as little as possible. 8 Electrolytic or volumetric processes are far more satisfactory. 1 If Clarke and Henz's process has been used, the solution contains ammonium oxalate and tartrate. These exert an appreciable solvent action on the tin sulphide. Another reason for condemning the gravimetric process. Clarke (page 208) destroyed the organic acids by means of potassium permanganate before precipitating the tin as sulphide. G. W Wdowiszewski (Stahl Eisen, 27. 781, 1907) destroys organic matter (tartaric acid) by digesting the solution in a covered beaker with 50 c.c. nitric acid (sp. gr. 1 '4) and 10 c.c. of sulphuric acid (sp. gr. 1'65). The solution is evaporated on a sand-bath down to about 30 c.c. The black or brown mass becomes almost colourless when mixed with 20 c.c. nitric acid. In the case of tin, of course, stannic oxide would be precipitated by this treatment. a The tin sulphide is a difficult precipitate to deal with, because it separates in a slimy condition, especially if the solution be boiled during the passage of the gas. The slimy precipitate also retains alkaline salts very tenaciously, and is difficult to wash, particularly in the absence of ammonium salts. See page 96. 3 R. Bunsen, Liebig's Ann., 106. 13, 1858. 4 J. Lowenthal, Journ. praU. Chem. (1), 56. 366, 1852. 5 For the volatilisation of stannic sulphide during the roasting, see C. J. Brooks, Chem. News, 73. 218, 1896. 6 That is, sulphates derived from the oxidation of the sulphide. F. Henz, Zeit. ano-rg. Chem., 37. 39, 1903. 7 H. Rose, Ausfiihrliches Handbuch der analytischen Chemie, Braunschweig, 2. 284, 1851. 8 It is generally used when but minute traces have to be determined. 3 io A TREATISE ON CHEMICAL ANALYSIS. 144. The Volumetric Determination of Tin Mene's Ferric Chloride Process. This method is founded upon the fact that tin sulphide can be dissolved in hydrochloric acid, and the resulting stannic chloride reduced to stannous chloride by means of an excess of metallic iron. 1 The solution of stannous chloride is then titrated with a standard ferric chloride solution, 2 which oxidises the stannous chloride back to stannic chloride : SnCl 2 + 2FeCl 3 = SnCl 4 + 2FeCl 2 . 3 Dissolution of the Tin.li other metals are present, the strongly acid solution is reduced by warming it at 80 or 90" with a piece of iron wire, when lead, arsenic, antimony, copper, and bismuth, if present, are precipitated. The solution is filtered, and neutralised with thin strips of zinc, whereby tin and lead are pre- cipitated. When the action is over and a drop of the clear filtrate gives no reaction for tin with "hydrogen sulphide water," pour off the clear liquid and wash the residue twice by decantation so as to keep the precipitated metals in the flask. 4 The Titration. Add 150 c.c. of hydrochloric acid (sp. gr. 1'16), and protect the contents of the flask from oxidation by a trap (page 188). Heat the solution to boiling, 5 and when everything is dissolved, titrate with the standard solution of ferric chloride. The titration should be conducted rapidly, since the stannous chloride is very sensitive to oxidising agents, oxygen dissolved in the water, etc. 6 The end of the reaction is indicated when a drop of ferric chloride imparts a permanent pale yellow tinge to the solution. The End Point. F. Mohr, one of the pioneers in volumetric analysis, 7 says that, " in Mene's process, the coloration of the solution by the ferric chloride is not sufficiently marked to enable the operator to recognise with certainty if a drop be added in excess." With moderately concentrated solutions, a person with normal colour, vision will soon recognise the effect when one drop of the ferric chloride has been added in excess, but with dilute solutions of ferric chloride Mohr's objection is quite valid. However, Morgan 8 has shown that if a blue Bunsen's flame be examined by looking through the solution being titrated, it will appear to have a greenish colour as soon as a trace of ferric chloride is in the solution. Morgan claims that a solution containing the equivalent of 0*00005 grin, of ferric oxide in 25 c.c. of water can be so recognised. The blue flame of a small Bunsen's burner is placed about 1 3 mm. below the bottom of the 1 C. Mene, Compt. Rend., 31. 82, 1850; Dingler's Journ., 117. 230, 1850; K. Pellet and A. Allart, Bull. Soc. Chim. (2), 27. 43, 438, 1877 ; J. A. Sanchez, ib. (4), 7. 890, 1910 ; H. Nelsmann, Zeit. anal. Chem., 16. 50, 1877 ; H. J. B. Rawlins, Chem. News, 107. 53, 1913. ' 2 STANDARD FERRIC CHLORIDE. Dissolve, say, 37 grms. of piano wire in hydrochloric acid and dilute the solution to a litre. Or evaporate 90 grms. of pure commercial ferric chloride to dryness with hydrochloric acid ; dissolve the residue in 150 c.c. of hydrochloric acid, and make the solution up to a litre. Standardise by dissolving a gram of pure tin in hydrochloric acid as indicated for tin in the text. Oxidise with nitric acid or hydrogen peroxide ; evaporate twice to dryness, and dissolve fh 200 c.c. of hydrochloric acid. Make the solution up to a litre. 3 This is the converse of J. Lowenthal and A. Stromeyer's process for iron (R. Fresenius, Anleitung zur quantitativen chemischen Analyse, Braunschweig, i. 365, 1875). 4 If no zinc is present, add a small piece to fill the flask with an atmosphere of hydrogen. 5 Do not boil too vigorously, or the hydrochloric acid will be weakened in strength before all the tin is dissolved. 6 A. C. Campbell, Journ. Anal. App. Chem., 2. 287, 1888. 7 F. Mohr, Lehrbuch der chemisch-aiialytischen Titrirmethode, Braunschweig, 264, 1874. 8 F. H. Morgan, Journ. Anal. App. Chem., 2. 169, 1888; C. L. H. Schwarz (Praktische Anleitung zu Maassanalysen, Braunschweig, 132, 1853) added a drop of potassium thiocyanate to the solution and titrated until a persistent red tint appeared ; C. Victor (Chem. Ztg,, 29. 179, 1905) titrates until starch and potassium iodide indicator is blued. THE DETERMINATION OF TIN. 3 I I flask in which the solution is being titrated. A good clear glass flask must be used, and the titration conducted in a darkened room, or in a dark corner. Errors. Mohr also points out that " Mene has stated, but not proved, that the decomposition is complete. A reverse action between the stannic chloride and the ferrous chloride proceeds very slowly in the cold, and is more marked on boiling. The reaction is only complete if an appreciable excess of ferric chloride be present. On the other hand, ferric chloride is wholly reduced by an equivalent quantity of stannous chloride ; but stannous chloride cannot be oxidised by an equivalent, but only by an excess of ferric chloride." As a matter of fact, the results are very good when the precautions are taken to prevent oxidation of the stannous chloride either during or before the titration. Molyb- denum, antimony, titanium, and tungsten interfere with the ferric chloride titra- tion ; uranium does not interfere. The titration in the presence of uranium wants careful watching, since the uranous salts are peroxidised after the tin and form a yellow solution. The ferric chloride tint gradually fades as the uranium is peroxidised. Rawlins obtained a maximum error of 0'15 per cent, by the process. Lowenthal and Stromeyer^s Process. Owing to the difficulty in determining the end point when the more dilute ferric chloride solutions are employed, many prefer the Lowenthal modification, 1 where an excess of the standard ferric chloride is added to the solution. The stannous chloride reduces part of the ferric to ferrous chloride. The amount of ferrous chloride so reduced is deter- mined by the permanganate titration Reinhardt's process, page 452. The chemical equation 2FeCl 3 + SnCl 2 = 2FeCl 2 + SnCl 4 corresponds with the fact that 1 grm. of potassium permanganate represents 2 '3888 grms. of stannic oxide Sn0 2 in the given solution. 145. Lenssen's Volumetric Iodine Process for Tin. In Lenssen's iodine process, 2 the stannous chloride is titrated in an alkaline 3 solution with a standard solution of iodine. The results are usually very slightly low. This is due to the action of air on the stannous chloride, and it is generally considered best to work with acid solutions, which do not oxidise so readily as alkaline solutions. The reaction is generally represented by the equation : SnCl 2 + 1 2 + 2HC1 = SnCl 4 + 2HI. This is supposed by Young to take place in two stages: (1) 2SnCl 2 + 4I = SnCl 4 + SnI 4 , when an excess of stannous chloride is present ; and (2), the reduc- tion of stannic iodide to stannous iodide by the stannous chloride SnI 4 + 2SnCl 2 = 2SnI 2 + SnCl 4 . If too much hydrochloric acid be present, the results will be high owing to the action of the acid on the potassium iodide. The solution should 1 J. Lowenthal, Journ. praTct. Chem. (1), 76. 484. 1859 ; A. Stromeyer, Liebig's Ann., 117. 261, 1861 ; C. Zengelis (Ber., 34. 2046, 1901) titrates the excess of ferric chloride with stannous chloride until a drop of the solution colours a drop of ammonium molybdate blue ; and T. Moore (Chem. News, 67. 267, 1893) titrates the excess of ferric chloride with cuprous chloride. 2 G. de Claubry, Grnn.pt. Rend., 23. 101, 1846 ; J. A. Miiller, Bull. Soc. Chim.(3), 25. 1002, 1901 ; Chem. News, 85. 114, 1902 ; A. Fraenkel and J. Fasal, Mitt. K. K. Tech. Gew. Wien, 7. 227, 1897 ; A. Scheurer-Kestner, Chem. News, 4. 101, 192, 1861 ; F. Ibbotson and H. Brearley, ib., 84. 167, 1901 ; E. Lenssen, Journ. praU. Chem. (1), 78. 200, 1859 ; Liebig's Ann., 114. 114, 1860 ; G. Topf, Zeit. anal. Chem., 26. 163, 1887 ; C. Friedheim, Zeit. anorg. Chem., 4. 145, 1893; W. H. Low, Journ. Amer. Chem. Soc., 29. 66, 1907 ; S. W. Young and M. Adams, ib., 19. 515, 1897 ; S. W. Young, ib., 19. 845, 851, 1897 ; 23. 21, 119, 453, 1901 ; B. Job, Journ. Soc. Chem. Ind., 17. 325, 1898 ; T. Benas, Zur massanalytischen Bestimmung des Zinns und uber einige Zinnoxydulsahe, Berlin, 1884 ; J. A. Sanchez, Butt. Soc. Chim. (4), 7. 890, 1910 ; H. J. B. Rawlins, Chem. News, 107. 53, 1913 ; A. Frankel and J. Fasal, ib., 78. 100, 1898 ; A. Jolles, Chem. Ztg., 12. 597, 1888. 3 In the presence of potassium sodium tartrate and sodium bicarbonate. 112 A TREATISE ON CHEMICAL ANALYSIS. \J contain about one-fifth its volume of concentrated hydrochloric acid (sp. gr. 1*16). Sulphates, phosphates, iodides, bromides, iron, nickel, cobalt, zinc, man- ganese, molybdenum, tungsten, titanium, uranium, aluminium, lead, chromium, and bismuth do not interfere with the results unless their colour is sufficient to mask the colour of the indicator starch blue. Arsenic spoils the result. According to Sanchez, the presence of antimonious chloride, SbCl 3 , does no harm. If copper be present, the iodine must be added slowly, and the solution briskly agitated, otherwise the results will be too high. When copper is present, one- third the volume of hydrochloric acid should be present, instead of one-fifth. deduction of the Stannic to Stannous Chloride. The hydrochloric acid solu- tion containing the stannic chloride is evaporated to about 50 c.c. in a wide 400-c.c beaker. A piece of clean iron wire l is placed in the solution against the side of the beaker, and the whole covered by a watch-glass. The solution is warmed to 80 or 90 on a sand bath or quartz plate. In about half an hour the arsenic, antimony, and copper will be precipitated, and the stannic chloride reduced to stannous chloride. 2 Cool the beaker in a stream of water; wash the clock-glass, etc., rapidly in a stream of cold, recently boiled distilled water. 3 The Titration. Add a few drops of starch paste (page 26) to the solution, and titrate the mixture rapidly with a standard solution of iodine 4 until the blue colour of the starch appears. For details of the "iodine" titration, see page 353. Ibbotson and Brearley's test analyses with the powdered "antimony reduction " show : Tin used. . . ./. 01750 O'lOOO 0'0480 0'0200 O'OOSO grm. Tin found . . . 01752 0'0997 0'0490 0'0201 0'0083 grm. Rawlins obtained a maximum error of 0*10 per cent, by the process. In all volumetric processes for tin, there is a tendency to low results owing to the rapidity with which the reduced tin salt is oxidised. 5 146. Henz and Classen's Electrolytic Process for Tin. Tin is readily deposited from solutions of ammonium oxalate in the presence of an excess of oxalic acid. 6 If ammonium oxalate be used alone, it is converted during the electrolysis into ammonium carbonate. The electrolyte then smells 1 C. and J. J. Beringer (A Textbook of Assaying, London, 288, 1906) recommend inserting a piece of nickel foil 20 cm. long and 5 cm. wide coiled on itself to form a cylinder into boiling solu- tion for the reduction. F. Ibbotson and H. Brearley (Chem. News, 84. 167, 1901) prefer reducinfl the stannic chloride to the stannous condition by boiling the solution with powdered antimony. Cool m an atmosphere of C0 2 (fig. 127, page 298), and titrate without removing the excess of antimony. J. A. Sanchez (Bull. Soc. Chim. (4), 7. 890, 1910) reduces with aluminium wire added in small portions until all is dissolved. D. B. Dott, Pharm. Journ. (4), 27 486 1908 The assumption is justified in practice, although there is no convenient test 3 The black precipitate antimony, arsenic, and copper if present does not interfere with the result. Antimony precipitated by iron does very slowly reduce stannic chloride, and there- fore the ' starch blue obtained in the titration is not quite permanent. Antimony which has been fused cooled, and ground m an agate mortar has a scarcely perceptible reducing action in the cold, although it very quickly reduces a boiling solution. Hence Ibbotson and Brearlev's process of reduction indicated in a preceding footnote. When the solution has cooled in an atmosphere of C0 2 , there is no need to remove the excess of antimony for the titration 4 Page 288. About 21 '32 grms. of iodine and 45 grms. of potassium iodide are made up to a litre One c c. of this solution will represent nearly O'Ol grm. of tin. The solution should however, be standardised against a known weight of tin. 5 E. A. Lewis, London Min. Journ., 606, 1911. 6 F. Henz, Zeit. anorg. Chem., 37. 31, 1903; A. Fischer, ib., 42. 363 1904- J M M Dormaar ib '"* 0/tn 1 r >' 7 - * __ j /s -n ^ ' ', " and M. A. von 2060, 1894 THE DETERMINATION OF TIN. 313 of ammonia, and stannic acid may separate, particularly if much tin be present. The stannic acid is dissolved by an excess of oxalic acid. Hence the electro- lyte must always be kept acid by the addition of, say, oxalic acid. The Electrolyte. In gravimetric analysis the tin is obtained as sulphide, and it is usually dissolved in sodium sulphide. To convert this solution into acid oxalate, acidify the solution with dilute acetic acid. Dissolve equal parts of ammonium oxalate and oxalic acid in hot water, so that the total amount of oxalate and oxalic acid per Ol grm. of tin amounts to 3*5 grms. Heat the solutions to boiling and pour the oxalate solution into the tin solution. The resulting solution may be slightly turbid owing to the separation of some sulphur, but it forms an excellent electrolyte for the deposition of tin. The Electrolysis. Use a current density of about 0*2 to 0'3 amp., and an E.M.F. of 2 to 3 volts. In about six hours most of the tin will have deposited. Add 8 c.c. of sulphuric acid (1:1) (or add more oxalic acid), and continue the electrolysis for another twenty-four hours. All the tin will then probably be precipitated. 1 En gels says that the electrolysis can be shortened by the addition of hydroxylamine sulphate to the electrolyte (see page 261). Results. The tin is precipitated as a compact shining silver- white metal. It is washed, dried, and weighed in the usual manner. In illustration of the results, Henz gives the subjoined numbers : Table XLIX.Test Analyses with the Electrolytic Process for Tin. Used Current Ammonium Sulphuric acid added. Total duration Found Error. Sn. density. oxalate. of the Sn. electrolysis. Grm. Amp. Grms. After hours. c.c. Hrs. Grm. Grm. Per cent. 0-1017 0-75-1-00 15 H 10 28 0-1036 + 0-0019 1-7 0-1017 0-48-0-05 15 3 2 10 27 0-1021 + 0-0004 0-4 0-2555 0-35-0-24 15 7 10 24 0-2555 o-oooo o-o 0-2555 0-38-0-20 15 8 10 24 0-2548 -0-0007 0-3 0-2555 O'30-O-IO 15 71 10 23 0-2550 -0-0005 0-2 0-2555 0-38-0'20 30 3 5 22 0-2551 -0-0004 0-2 0-1017 0-15-0-10 15 6 2 21 0-1016 -o-oooi o-i This table shows that the proportions indicated above are likely to give satisfactory results. The time factor is a rather serious objection. Removal of Tin from the Electrode. There is a difficulty in removing tin from the cathode, since the metal does not readily dissolve in acids even boiling hydrochloric acid dissolves the metal very slowly. Nitric acid forms a coating of stannic oxide, which must be frequently removed in order to expose a fresh surface of the metal to attack. Classen recommends warming the deposit with a mixture of 2 grms. of tartaric acid, 8 c.c. of water, and 2 c.c. of concentrated A. Scheen, ib., 14. 257, 1908 ; F. Forster and J. Wolf, ib., 13. 205, 1907 ; H. J. S. Sand, ib., 13. 327, 1907 ; H. Ost, Zeit. angew. Chem., 10. 325, 1897 ; 14. 817, 1901 ; A. Hollard, Bull. Soc. Chim. (3), 29. 262, 1903 ; A. Inhelder, Beitrag zur Trennung des Antimons und Zinns und zur Analyse von Lagermetallen, Ziirich, 1911 ; A. H. Cushrnan and E. B. Wettengel, Journ. Ind. Eng. Chem., 5. 217, 1913. 1 To make sure, withdraw about 1 c.c. by means of a pipette; acidify the solution with sulphuric acid, add CC H 3 S water," and warm gently. If yellow or brown stannous or stannic sulphides separate, the "electrolysis is not completed. Do not confuse free sulphur with tin sulphide. 314 A TREATISE ON CHEMICAL ANALYSIS. nitric acid ; or fusion with potassium bisulphate. Some recommend coating the cathode with a film of copper of silver before precipitating the tin in order to prevent the platinum being seriously attacked during the subsequent removal of the deposited tin. Another plan is to make the dish the anode and electrolyse a solution of dilute sulphuric acid with a piece of copper wire as cathode. 147. The Evaluation of Commercial Compounds containing Tin. Tin Oxide. The amount of tin oxide can be determined by reducing the sample to metal in a current of hydrogen or coal gas (page 268), and dissolving the metal in hydrochloric acid. The solution may be titrated for tin as indicated above. The metal may also be obtained by reduction with potassium cyanide (page 269). The button of metal should be specially analysed for tin, since it may be contaminated with other metals. Tin Ash (Calcine). Here, tin and lead are to be separated. The hydro- chloric acid solution is treated with an excess of sodium hydroxide, and saturated with hydrogen sulphide. Lead is precipitated ; tin remains in solution. The tin solution may be treated as indicated, pages 307-309 ; the lead sulphide as on page 320. 1 Sodium Stannate. Commercial sodium stannate may contain relatively large quantities of sodium arsenate. 2 The analyst may thus have to determine both the tin and the arsenic. The arsenic may be precipitated by placing best piano wire in the solution acidified with hydrochloric acid, as in the case of antimony by Tookey's process. The tin can be determined in the filtrate. The sodium can be determined as chloride by applying part of the process (page 224). 1 For the solubility of lead sulphate in stannous chloride, see M. de Jong, Zeit. anal. Chem., 41. 596, 1901 ; A. van Raalte, ib., 43. 36, 1903. 2 E. Haeffely (Phil. Mag. (4), 10. 220, 1855) adds a known excess of arsenic, boils with an excess of nitric acid, weighs the tin as tin arsenate 2Sn0 2 . As 2 5 and determines the excess of arsenic in the nitrate. See T. Goldschmidt, Dingier 's Journ., 162. 77, 1861 ; A. Scheurer- Kestner, Rep. Chim. App., 4. 221, 1862 ; P. T. Austen, Amer. Chem. Journ., 5. 210, 1883. CHAPTER XXIV. THE DETERMINATION OF LEAD. 148. The Properties of Lead Sulphate. LEAD is perhaps most frequently separated and weighed as sulphate. Lead sulphate is a heavy white powder sparingly soluble in water : 100 c.c. of water dissolve nearly 0-0038 grm. lead sulphate. 1 The solubility is decreased in the presence of small quantities of sulphuric acid, and increased in more concentrated solutions, as well as in solutions of hydrochloric and nitric acids. This is illustrated by the following table : 2 Table L. Solubility of Lead Sulphate in Mineral Acids. Sulphuric acid. Hydrochloric acid. Nitric acid. Grm. PbS0 4 per 100 grms. sol. Approximate normality of acid. Grm. PbS0 4 per 100 grms. sol. Approximate normality of acid. Grm. PbS0 4 per 100 grms. sol. Approximate normality of acid. 0-003 O'Oll 0-039 27 31 36 0-33 0-59 078 3 5 10 0-14 0-35 0-95 2 3 5 The effect of sulphuric acid in reducing the solubility of lead sulphate is illustrated by the graph, fig. 129. It is generally supposed that lead sulphate is best precipitated from a liquid approximately N-H 2 S0 4 , and the precipitate is best washed with an acid of the same strength. The solu- bility of lead sulphate is also reduced in the presence of alcohol. Hence, many prefer to dilute the liquid with two or three times its volume of alcohol or methylated spirit, and wash with a dilute alcoholic solution of sulphuric acid containing 10 per cent, of alcohol and 1 per cent, of sulphuric acid, in order to reduce the loss of lead sulphate to a minimum. There is little difference in the actual results with the alcohol and the dilute sulphuric acid treatments, 1 H. C. Dibbits, Zeit. anal. Chem., 13. 139, 1874; F. Kohlrausch, Zeit. phys. Chem., 50. 356, 1905; W. Bbttger, ib., 46.602, 1903; R. Fresenius, LieUg's Ann., 59. 125, 1876; G. F. Rodwell, Chem. News, II. 50, 1865 ; M. Pleissner, Ueber die Loslichkeit einiger Bleiver- bindungen in Wasser, Berlin, 1903. 2 G. F. Rodwell, Journ. Chem. Soc., 15. 59, 1862 ; C. S. Sellack, Pogg. Ann., 133. 137, 1868 ; Zeit. anal. Chem., g. 464, 1870; H. Struve, ib., g. 34, 1870; W. Stadel, ib., 2. 180, 1863; C. Stammer, Ding ter's Journ., 165. 209, 1862 ; Chem. Ztg. t 6. 63, 1884 ; J. Sehnal, Compt. Rend., 148. 1394, 1909 ; K. Beck and P. Stegmiiller, Arb. Kais. Gesund. Amt., 34. 446, 1910. 3 i6 A TREATISE ON CHEMICAL ANALYSIS. but the use of alcohol is apt to give high results, because very few sulphates are readily soluble in alcohol. According to Sehnal, lead sulphate is " absolutely insoluble in water containing one part of H. 2 S0 4 per 1000 c.c." If "absolutely " here means that " the amount of soluble lead sulphate cannot be determined by ordinary analytical methods," I quite agree. The solubility of lead sulphate is increased in the presence of alkaline chlorides and many other salts ; and it is fairly soluble in ammonium acetate, 1 tartrate, citrate, chloride, nitrate, etc., and in sodium thiosulphate, sodium acetate, caustic alkalies, etc. 2 In illustration, the following table represents the 0-005 0-005 0-01 FIG. 129. Effect of sulphuric acid on the solubility of lead sulphate. solubility of lead sulphate in ammonium acetate (Noyes and Whitcomb) and in sodium acetate (Dunnington and Long). The solubility is increased by raising the temperature. Table LI. Solubility of Lead Sulphate in Ammonium and Sodium Acetates. Ammonium acetate. Sodium acetate. 25 100 Grm. PbS0 4 Grm. acetate Grm. PbS0 4 Grm. acetate Grm. PbS0 4 Grm. acetate per 100 per 100 per 100 per 100 per 100 per 100 c.c. c.c. c.c. c.c. c.c. c.c. 0-0041 O'OOO 7-12 28 2-05 0-054 0-0636 0-798 9-88 32 8-20 0-9 0-138 1-596 10-58 37 41 11-2 0-302 3-192 11-10 45 ... ... 1 H. C. Dibbits, Bull. Soc. Chim. (2), 2O. 258, 1873 ; F. P. Dunnington and J. C. Long, Amer. Chem. Journ., 22. 217, 1899; E. Lenssen, Journ. prakt. Chem. (1), 85. 89, 1862; J. Lowenthal, ib. (1), 60. 267, 1853. 2 L. Kahlenberg, Zeit. phys. Chem., 17. 577, 1895 ; Chem. News, I. 152, 1860 ; J. Lowe, Journ. prakt. Chem. (1), 74. 348, 1858 ; A. A. Noyes and W. H. Whitcomb, Journ. Amer. Chem. Soc., 27. 747, 1905. THE DETERMINATION OF LEAD. 317 It is therefore necessary to precipitate lead sulphate from solutions as free as possible from ammonium salts, alkaline chlorides, nitric and hydrochloric acids, etc. If potassium salts be present, Levol l has pointed out that there is a danger of precipitating potassium sulphate with the lead sulphate in the form of a double lead and potassium sulphate K 2 Pb(S0 4 ) 2 . There is not the same risk with sodium sulphate. The solubility of lead sulphate in sodium thiosulphate and ammonium acetate solutions is frequently employed in separating lead sulphate from barium sulphate. The latter is supposed to be insoluble in ammonium acetate and sodium thiosulphate. When the latter is used, the temperature must be kept below 68, or an insoluble lead sulphite may be formed. If basic iron sulphate 2 or barium sulphate 3 be precipitated with the lead sulphate, there is a difficulty in dissolving the last traces of the lead sulphate in ammonium acetate, thus leading to low results. With ammonium acetate, there is a small error owing to the slight solubility of barium sulphate in this menstruum. Kernot, 4 for example, finds that at 25, 0*133 grm. of barium sulphate dissolves in a litre of a solution of ammonium acetate containing 300 grins, of the salt. Solutions of carbonates, but not bicarbonates, 5 convert lead sulphate at ordinary temperatures into lead carbonate, which is soluble in dilute nitric acid, and this reaction is also sometimes used for separating lead sulphate from barium sulphate (see page 515), which does not react with carbonates in the same way. Lead sulphate is practically unaltered when exposed to the air, and also when ignited at a low red heat. It fuses at 848 without decomposition, provided reducing agents be absent. 6 If heated for a long time, there may be an appre- ciable loss by volatilisation. 7 Thus, 1'4082 grms., when heated in a porcelain tube in a current of carbon dioxide, lost in weight, according to Williams, 0*134 per cent, either by volatilisation or partial decomposition. 149. The Determination of Lead as Sulphate. The solution, if dilute, is concentrated by evaporation, and if hydrochloric or nitric acids be present, the solution is evaporated with about 1 c.c. of sul- phuric acid 8 (1:1) until copious fumes of sulphuric acid are evolved. This drives off the nitric and hydrochloric acids, which interfere with the subsequent precipitation of the lead sulphate. Cool the solution. Add water until the volume of the liquid is about 50 c.c. and the solution contains between 1 and 2 per cent, of H 2 S0 4 about JN-H 2 S0 4 . Filter the solution through filter paper, or through a Gooch's crucible packed with asbestos, or, better still, through a Munroe's crucible, 9 since there is a risk of the acid attacking 1 A. Levol, Rep. Chim. App., 4. 21, 1862; Chem. News, 5. 144, 1862; J. J. Fox, Proc. Chem. Soc., 23. 199, 1907 ; Journ. Chem. Soc., 95. 878, 1909: F. G. Belton, Chem. News, 91. 191, 1905. 2 J H. Walton and H. A. Scholz, Amer. Chem. Journ., 39. 771, 1908; Chem. Neu-s, 98. 61, 76, 1908. 3 J. C. Bull, School Mines Quart., 23. 348, 1903 ; Chem. News. 87. 40, 52, 66, 1903. 4 G. Kernot, Rend. Acad. Sci. Napoli (3), 15. 155, 1909. 5 H. Rose, Fogg. Ann., 95. 426, 1855. 6 0. L. Erdmann, Journ. prakt. Chem. (1), 62. 381, 1854. 7 C. P. Williams, Chem. News, 23. 236, 1871; A. Mitscherlich, Journ. prakt. Chem. (1), 83. 485, 1861 ; J. Boussingault, Compt. Rend., 64. 1159, 1867. 8 Be careful in adding concentrated acid to dilute solutions, or some liquid may be lost by spurting. 9 Filter paper can be used to collect the lead sulphate. In that case, dry the washed pre- cipitate ; separate the precipitate from the paper on a watch-glass ; burn the paper in a weighed 318 A TREATISE ON CHEMICAL ANALYSIS. asbestos. Wash the precipitate with dilute sulphuric acid (JN- to N- and finally with alcohol, until the washings are free from sulphuric acid. 2 Dry the precipitate at 100. Ignite the precipitate over the full flame of a Meker's burner, 3 and, when cold, weigh the ignited precipitate as lead sulphate PbS0 4 . The weight of the lead sulphate multiplied by 0*7359 represents the correspond- ing amount of lead oxide PbO or use Table XCIIL, or fig. 23. The results are usually excellent. 4 150. The Separation of Lead from Bismuth, Copper, and Cadmium. The process just indicated gives excellent separations of lead from copper, cadmium, zinc, etc. We therefore take up the problem left on page 276 : the separation of the constituents of the residue left on treatment of the hydrogen sulphide precipitate with sodium monosulphide. The residue is dissolved in concentrated hydrochloric acid. There are two cases : (1) Bismuth absent. This is nearly always the case in the analysis of pottery colours arid glazes. The solution is evaporated on a water bath along with a little sulphuric acid. The evaporation is completed on a hot plate, or over a ring burner, until the sulphuric acid begins to fume. The object is to drive off the hydrochloric acid, which leads to an imperfect precipitation of the lead sul- phate. The lead sulphate is determined by diluting the solution as indicated for lead sulphate (page 317). (2) Bismuth present. The solution should contain just enough hydrochloric acid to prevent the precipitation of basic bismuth salts when the solution is diluted. The right amount of hydrochloric acid is determined 5 as follows : Evaporate the solution on a water bath, and put a drop of the evaporating solution on a watch-glass with a drop of water. If no precipitate forms, too much hydrochloric acid may be present. Continue the evaporation. If a white precipitate forms, the evaporation has gone too far and hydrochloric acid is added, a few drops at a time, until no precipitate of basic bismuth salt forms, when a drop of the solution is diluted with water. The lead is then precipitated as lead sulphate, and the bismuth separated from the copper and the cadmium by Jannasch's process (page 348). If no bismuth be present, proceed to copper (page 350). The process just indicated for the separation of lead from bismuth, copper, and cadmium requires great care and some practice before good results can be ensured. Some might prefer to employ Lowe's process for the separation of bismuth from copper, cadmium, and lead, and afterwards separate the lead from the copper and cadmium in the form of lead sulphate. The separa- tion of bismuth is discussed on page 347 et seq. crucible ; moisten the ash with nitric acid ; add a drop of sulphuric acid ; evaporate to dryness on a low flame. Transfer the precipitate from the watch-glass to the crucible, and ignite over the Meker's burner to a constant weight. 1 L. L. de Koninck (Bull. Soc. Chim. Belg., 21. 141, 1907) washes the precipitate with water containing 0*7 per cent, of ammonium sulphate. 2 If the sulphuric acid be not washed out, particularly from filter paper, the paper may be charred as the acid becomes more and more concentrated during the drying of the precipitate. With the Munroe's crucible there is not so much need for the alcohol washing. 3 Some merely dry the precipitate at 110 before weighing. 4 According to A. G. Blakeley and E. M. Chance (Journ. Soc. Chem. Ind., 30. 518, 1911), in the ordinary method of precipitating lead as sulphate, the presence of tin and antimony interfere because basic salts are precipitated with the lead sulphate. P. H. Walker and H. A. Whitman, Journ. Ind. Eng. Chem., I. 519, 1909. 5 0. Steen, Zeit. angew. Chem., 8. 530, 1895; H. Rose, Pogg. Ann., no. 432, 1860; C. Friedheim, Leitfaden fur quantitative chemische Analyse, Berlin, 300, 1905 ; H. Herzog, Journ. Anal. App. Chem., I. 245, 1887; Chem. News, 58. 129, 1888. THE DETERMINATION OF LEAD. 319 The scheme for separating the constituents of the precipitate produced by hydrogen sulphide may now be conveniently summarised (mercury absent) solids to left, solutions to right : Digestion with sodium sulphide | Dissolve in HOI ; add H 2 S0 4 Add NH 3 and magnesia mixture | ! r i r "~i Lead Ammonia ; H 2 2 Arsenic H 2 S ; oxalic acid I ! I Bismuth Ammon. thiocyanate Antimony Metallic zinc Copper Hydrogen sulphide Tin Reject I. I Cadmium Reject Any step can be omitted if the qualitative process shows the corresponding element is absent. The scheme for dealing with the next group is given on page 386. 151. The Analysis of White Lead. There are several alternative schemes for the analysis of white lead. 1 The following process brings out the information required from the chemical analysis for most technical operations : Hygroscopic Moisture. Dry a weighing bottle in a steam oven ; cool in a desiccator and weigh. Transfer about 10 grms of the sample to the weighing bottle, and weigh again. Dry the sample in the steam or air bath at from 100 to 105 until no further loss in weight occurs overnight generally suffices when further heated. The loss in weight represents hygroscopic moisture. 2 Dissolution of the White Lead. Transfer a gram of the dried sample to a 100-c.c. flask and boil for 5 minutes with an excess of dilute acetic acid. 3 Filter through a 6'5-cm. filter paper, and wash once with hot dilute acetic acid and then with hot water. The residue may contain insoluble sand (silica), silicates (clays), lead sulphate, barium carbonate, calcium sulphate, and possibly alumina, magnesia, and free lead. The Soluble Portion. Heat the solution ; add 1 c.c. of a saturated solution of mercuric chloride. 4 Pass hydrogen sulphide through' the boiling liquid for 1 P. Drawe, Zeit. angew. Chem., 13. 174, 1902; A. Coppalle, Ann. Chim. Anal. App., 12. 62, 1907; G. W. Thompson, Journ. Soc. Chem. Ind., 24. 487, 1905; G. Tissandier, Chem. Neivs, 23. 268, 1871 ; A. Neujean, ib., 22. 251, 1870 ; A. Adriani, ib., 4. 43, 1861. 2 W. A. Da vis and C. A. Klein (Journ. Soc. Chem. Ind., 26. 848, 1907) say heat has no action below 110, and decomposition begins at 120. L. Joulin (Chem. News, 27. 211, 1873) considers that at temperatures over 50 there is no security against partial decomposition and loss of carbon dioxide. At 150 the pressure of cerusite was below 30 mm. ; at 250, 75 mm. ; and at 300 decomposition was complete. I have met with only one sample of white lead which showed an appreciable decomposition when dried in the steam oven, but usually, drying in the steam oven is quite safe. 3 DILUTE ACETIC ACID. 10 c.c. of glacial acetic acid with 25 c.c. water. Some prefer dilute nitric acid (1 : 5). If too much nitric acid be present, the lead sulphate is not completely precipitated. 4 When very small amounts of lead are to be precipitated as sulphide, E. Murmann (Monats. Chem., 19. 404, 1899 ; M. Antony and T. Benelli, Guz. Chim. Ital., 26. i, 218, 1906) adds mercuric chloride to the solution before passing the gas. This renders the precipitate easy to filter. The mercuric sulphide is volatilised on roasting the precipitate. 320 A TREATISE ON CHEMICAL ANALYSIS. about half an hour. 1 Let the precipitate settle. Filter. 2 Wash with hot water. 3 The precipitate might contain zinc and lead sulphides; the filtrate, calcium, barium, and possibly magnesium salts. Boil the filtrate to expel the hydrogen sulphide, and precipitate the barium with sulphuric acid (page 618) and filter. Calculate the corresponding amount of barium carbonate by multiplying the weight of barium sulphate by 0*84555. Precipitate the lime as oxalate (page 213), and calculate the corresponding amount of calcium carbonate by multiplying the weight of calcium oxide by 1*7844. Precipitate the magnesia as phosphate (page 218), and calculate the corresponding amount of magnesium carbonate by multiplying the weight of the magnesium pyrophosphate with the factor 0*7576. The precipitated mixture of zinc and lead sulphides is ignited in a porcelain crucible at a low temperature, 4 in order to burn off the filter paper. Brush the residue into a small 400-c.c. beaker and wash out the crucible by means of dilute nitric acid (1 : 5). Add about 1 c.c. of sulphuric acid (1:1); evaporate until copious white fumes of sulphuric anhydride come off; and separate the lead sulphate as described above. Weigh the ignited precipitate of lead sulphate as PbS0 4 . The weight of the lead sulphate, multiplied by 0*8526, represents the weight of normal white lead 2PbC0 3 . Pb(OH) 2 in the given sample. The zinc is precipitated from the cold filtrate as zinc ammonium phosphate (page 366), and the corresponding amount of ZnO or ZriC0 3 computed in the usual manner. The Insoluble Portion.- Boil the residue with 20 c.c. dilute hydrochloric acid (1 : 1) and 5 grms. of ammonium chloride for 5 minutes. Dilute to 400 c.c. Boil 5 minutes. The lead arid calcium sulphates are dissolved ; barium sulphate, 5 silica, and silicates remain undissolved. Filter and wash with hot water. The lead is precipitated from the filtrate by hydrogen sulphide. 6 The lead sulphide is treated as described above and reported as lead sulphate. The filtrate is treated with ammonia 7 and ammonium oxalate to precipitate the lime as calcium oxalate (page 213). Report as calcium sulphate. The residue (insoluble in ammonium chloride and hydrochloric acid), containing the barium sulphate, sand, and clay, is ignited in a platinum crucible and 1 Calcium chloride in a dilute solution of lead chloride in presence of hydrochloric acid or nitric acid may give no precipitate with hydrogen sulphide owing to the solubility of lead sulphide in presence of calcium chloride and acid. K. H. Mertens, Pharm. Centr., 34. 273, 1895 ; H. Hagerand E. Geissler, ib., 34. 273, 1885. 2 A. Gawalovski (Zeit. anal. Chem., 26. 51, 1887) suggests saturating the margin of the filter paper with paraffin or other fat (free from ash) in order to prevent the creeping of the precipitate over the edge of the paper. 3 F. Moldenhauer (Chem. Ztg., 22. 256, 1898; Chem. Neivs, 79. 182, 1899) washes first with water and then with warm ammonium sulphide. 4 For the alleged volatilisation of lead sulphide, see H. Rose, Pogg. Ann., no. 134, 1860; G. F. Rodwell, Journ. Chem. Soc., 15. 43, 1863 ; A. Souchay, Zeit. anal. Chem., 5. 63, 1886 ; A. Classen, Journ. prakt. Chem. (1), 96. 257, 1865. 5 It must be remembered that heavy spar is slightly soluble in nitric acid. Hence, the "insoluble" heavy spar can only be determined approximately, unless the solution be evaporated to dryness, and the residue taken up with water, filtered, washed, and calcined at a low red heat. Note that the "total heavy spar" does not necessarily represent the amount of heavy spar added to the white lead, because the heavy spar used is not usually pure barium sul- phate, but contains, say : 90 per cent. BaS0 4 ; 2 per cent. CaO ; 7 per cent. Si0 2 ; 1 per cent. A1 2 O 3 and Fe 2 3 . Since different solvents may dissolve different amounts of the heavy spar, it is easy to see how different "total insolubles " may be reported by different analysts. In exact work it is as well to fuse the " insoluble " with sodium carbonate as indicated in the text. 6 For the precipitation of lead sulphide from hydrochloric acid solutions, the solution should not contain more than 3 or 4 c.c. of hydrochloric acid (sp. gr. 1*12) per 100 c.c. of solution in the cold. 7 If alumina be precipitated when ammonia is added, filter, wash, ignite, weigh, and report as alumina. THE DETERMINATION OF LEAD. 32! weighed. Fuse the residue with ten times its weight of sodium carbonate. Extract with water ; filter ; and wash. Dissolve the precipitate in 5 c.c. of hydrochloric acid;, boil; add sulphuric acid to precipitate barium sulphate (page 618); weigh the precipitate; and report as barium sulphate. Deduct this weight from the "insoluble," and report the difference as "clay, sand, and insoluble silicates." Collecting these results together, it will be seen that we have followed the subjoined scheme : I. Soluble in dilute acetic acid. Treat with H 2 S. A. Soluble. Add sulphuric acid. (a) Soluble. Add ammonium oxalate. 1. Soluble. Add microcosmic salt. (i.) Soluble. Reject. (ii. ) Insoluble Mg salt . . . MgC0 3 . 2. Insoluble Ca salt . . . . . . CaC0 3 . (b) Insoluble barium sulphate . . . . . . BaCO s . B. Insoluble. Take up with HN0 3 , add H 2 S0 4 . (a) Soluble. Add microcosmic salt. 1. Soluble. Reject. 2. Insoluble Zn Salt - . . . . . ZnO. (6) Insoluble Pb salt 2PbCOo.Pb(OH) 2 . II. Insoluble. Digest with HC1 and NH 4 C1. A. Soluble. PassH 2 S. (a) Soluble. Add ammonium oxalate. 1. Soluble. Reject. 2. Insoluble Ca salt. . . . . . . CaS0 4 . (6) Insoluble Pb salt . . . . . " . . . PbS0 4 . B. Insoluble. Weigh, fuse sod. carb., etc. (a) Soluble. Add HoS0 4 . 1. Soluble. Reject. 2. Insoluble Ba salt BaS0 4 . (b) Insoluble. Subtract weight of II. B., (a) 2, from II. B., Insolubles. Carbon dioxide can be determined as indicated on page 553. Water is determined as indicated on page 561 or 574. Iron is determined by dissolving 5 grms. of the original sample in dilute nitric acid. Precipitate the lead as sulphate ; filter ; evaporate to dryness ; dissolve the residue in 1 c.c. of hydrochloric acid and a little water ; and make the solution up to 200 c.c. Determine the iron colorimetrically 1 (page 200) in one 100 c.c. Copper is determined colorimetrically by the ammonia process (page 355) in the 100 c.c. remaining after the iron has been determined. Acetic acid is determined by alternate distillation and distillation in steam from a mixture of the sample with phosphoric acid. 2 The distillate is titrated with y^N-NaOH and phenolphthalein as indicator until 10 c.c. of the distillate require but a drop of the alkaline solution to redden the indicator. Soluble salts are isolated by boiling 510 grms. of the sample with 100 c.c. of water. Filter and wash. The sulphates are often determined by the barium chloride process (page 618). The barium sulphate may be expressed either in terms of S0 3 , or, as is sometimes done, as ZnS0 4 . Sulphur dioxide or sulphites are sometimes found in the quick process white leads. Mix 2 grms. of the dried sample with 100 c.c. of recently boiled and cold water in a 300-c.c. Erlenmeyer's flask ; add 5 c.c. of concentrated sulphuric acid. Thoroughly agitate the mass. In about 15 minutes, titrate the mixture with j-^N-iodine solution, using starch as indicator. Continue the titration until the blue colour develops. One gram of iodine represents 0*252 grm. of S0 2 . 1 A. Lecrenier, Bull. Soc. Chim. Belg., 18. 404, 1904; J. A. Schoeffer, Journ. Ind. Eng. Chem., 4. 659, 1912. 2 -G. W. Thompson, Journ. Soc. Chem. Ind., 24. 587, 1905. 21 322 A TREATISE ON CHEMICAL ANALYSIS. 152. The Analysis of Red Lead. Red lead is usually made by roasting " pig lead " until oxidation ceases. The red lead may contain as impurities or adulterants heavy spar, lead sulphate, silica, ferric oxide, copper oxide, etc. It also contains metallic lead, and various lead oxides PbO, Pb 2 3 , Pb 3 4 . Silica and clay may be derived from the floor of the muffle in which" the red lead was roasted. Owing to the difference in the specific gravity of heavy spar and red lead, good mixing is difficult, and an intimate mixture is liable to segregate when travelling in casks. Hence samples drawn from different parts of the same cask may give different results. Care must therefore be exercised in the sampling. Dissolve, say, 10 grms. of the given sample in dilute nitric acid (sp. gr. T40), or in acetic acid. The dissolution is greatly facilitated by the addition of about 2 c.c. of hydrogen peroxide (5 per cent.) 1 just after the addition of the nitric acid. Heat the solution for about an hour on the water bath. Filter and wash with hot water. The insoluble residue may contain clay, silica, and heavy spar barium sulphate. 2 This is treated as indicated for white lead (page 320). The clear solution is evaporated to dryness with hydrochloric acid in order to make the silica insoluble, and the latter filtered off. The filtrate from the silica is treated with sulphuric acid in the usual manner, in order to precipitate the lead as sulphate. Keep the amount of acid as low as possible 3 c.c. of concen- trated sulphuric acid diluted to 100 c.c. per 10 grms. of red lead will suffice. The filtrate from the lead sulphate is treated with hydrogen sulphide to precipitate the arsenic, antimony, copper, etc., as sulphides. Separate the arsenic and antimony from the copper by means of sodium mono-sulphide (page 277); determine the copper colorimetrically (page 355). If no arsenic or antimony be present, determine the copper in the filtrate from the lead. Iron, zinc, and lime can be determined, if present, in the filtrate from the hydrogen sulphide precipitate (page 276). Metallic Lead. Red lead and litharge frequently contain free metallic lead diffused throughout the mass in fine granules maybe up to 3 per cent. 3 To determine the amount, dissolve a known weight of the red lead in dilute acetic acid at 40 in the presence of hydrogen peroxide. Wash the residue with an acetic acid solution of ammonium acetate, and then with water. Dry and weigh the residue. If other impurities be present, the metallic lead can generally be separated by treatment with dilute nitric acid, and either the washed residue dissolved, dried, and weighed, or the lead determined in the nitric acid solution in the usual manner. Lead Monoxide. The presence of lead monoxide, due to imperfect oxidation, 1 Hydrogen peroxide of commerce sometimes contains sulphuric acid, and in consequence may precipitate lead sulphate. Merck's "perhydrol" 30 per cent, hydrogen peroxide is satisfactory. Cane sugar, oxalic acid (R. Fresenius, Anleitung zur qaantitativen Analyse, Braunschweig, 2. 484, 1903), lactic acid {A. Partheil, Chem. Ztg., 31. 941, 1907), methyl alcohol, glycerol, formaldehyde, phenylhydrazine, and hydroxylamine hydrochloride have also been suggested (E. Pieszczek, Pharm. Ztg., 52. 922, 1908 ; F. P. Dunnington, Zeit. anal. Chem., 28. 338, 1889; L. Opificus, ib., 28. 345, 1889; H. Schlossberg, ib., 41. 740, 1902), but the hydrogen peroxide will be difficult to beat. 2 Barium sulphate is not a common impurity of normal red lead, since it makes the lead incline to an orange tint. The heavy spar in "tinted red leads" may vary from 10 to 70 per cent. Highly adulterated red lead has been stained with aniline dyes to a good standard colour (M. Frehse, Ann. Chim. Anal. App., II. 176, 1906). The dye can be detected by the tint acquired by digesting the red lead in a suitable solvent. Shake 20 c.c. of 95 per cent, alcohol with 2 grms. of the sample, heat to boiling, and let settle. Pour off the alcohol and boil with 20 c.c. of water ; let settle, and proceed in a similar way with ammonium hydroxide. If any of these three solvents have been coloured, the red lead is probably coloured with an organic dye. 3 G. C. Wittstein, Dingier 's Journ., 194. 84, 1869 ; Chem. News, 20. 249, 1869. THE DETERMINATION OF LEAD. 323 is common. Digest the dried sample with a neutral saturated solution of lead acetate in which lead monoxide dissolves. Filter off the insoluble red lead on a weighed Gooch's crucible. Wash, and dry at 100 to constant weight. 1 Lead Peroxide. Digest, say, one gram of dry red lead in a porcelain dish with 20-30 c.c. of dilute nitric acid (1 : 5) on a water bath for a few minutes, 2 when insoluble lead dioxide 3 and soluble lead nitrate are formed. The former may be filtered off, washed, dried at 110, and weighed. For volumetric determination, add 50 c.c. of iN-oxalic acid and heat to boiling. 4 Titrate the excess of oxalic acid in the hot solution without filtering by means of a iN-potassium permanganate solution. If 31 c.c. of permanganate are used, the 50-39 = 11 c.c. of iN-oxalic acid are used for the reduction of the gram of jred lead. From the equation Pb0 2 + H 2 C 2 4 = PbO + H 2 + 2C0 2 it follows that 1000 c.c. of the iN-oxalic acid represents 23 -9 grms. of Pb0 2 , or 1 c.c. represents 0"0239 grm. of Pb0 2 . Hence 11 c.c. of oxalic acid correspond with 1 1 x 0-0239 = 0'2629 grm. Pb0 2 . Hence, one gram red lead has 0-2629 grm. Pb0 2 , or 26 '29 per cent. Pb0 2 . The lead monoxide may be determined by calculating the Pb0 2 to Pb 3 4 and subtracting this from the total lead. 5 In the gravimetric determination of lead peroxide, the sample is digested with warm dilute nitric acid (1 : 5) and allowed to stand in a cool place for 24 hours. Filter through a Gooch's crucible, wash with hot water, dry 6 hours, between 105 and 110, and weigh as Pb0 2 . Calculate the corresponding amount of Pb 3 4 by multiplication with 2 -866. The difference between this weight and the weight of the original sample represents litharge. This method is quite unreliable, because lead sesquioxide Pb 2 3 dissolves in the nitric acid and is in consequence reported as litharge. The Pb0 2 of red lead containing sodium nitrite is attacked by nitric acid more than if the nitrite were absent. 1 Lead sesquioxide Pb 2 3 is scarcely attacked by the lead acetate E. E. Dunlap, Journ. Amer. Chem. Soc., 30. 611, 1908; L. Qpificius, C/iem. Ztg., 12. 477, 1898. D. Woodman (Journ. Amer. Chem. Soc., 19. 339, 1897), using this process, found a variation of 41-92 per cent, in the amount of red lead, and from 8-59 per cent, of lead monoxide, in commercial red leads on the American market. J. Lowe (Dingler's Journ., 271. 472, 1889) recommends lead nitrate. 2 F. Lux, Zeit. anal. Chem., 19. 153, 1880 ; E. Rupp, ib., 42. 732, 1903 ; Schlossberg, ib., 41. 735, 1902 ; P. Beck, ib., 47. 465, 1908 ; H. Fleck, Ber., 20. 855, 1881 ; W. P. Joshua, Analytische Beltrage zur Brstimmung von Bleisuporoxyd neben Blei uns Bleioxyd, Ziirich, 1906; A. Chevala and H. Colle, Gazz. Chim. Ital, 41. ii. 551, 1911 ; Zeit. anal. Chem., 50. 209, 1911 ; J. F. Sacher, Chem. Ztg., 35. 731, 1911. 3 It is simply afacon de parler to say "the Pb0 2 content of red lead," since this statement does not imply the existence of Pb0 2 in red lead. Red lead is probably not a mixture of Pb0. 2 and 2PbO. See J. W. Mellor, Modern Inorganic Chemistry, London, 795, 1912. 4 Impurities like sand, barium sulphate, lead sulphate, remain undissolved. 5 G. Topf (Zeit. anal. Chem., 26. 296, 1887; W. Diehl, Dingler's Journ., 246. 196,1882) heats the red lead with hydrochloric acid and marble, and passes the chlorine evolved into a standard solution of potassium iodide (5 grms. neutral KI, free from iodates) in 100 c.c. of water contained in each of a couple of Peligot's tubes, and estimates the liberated iodine in the regular way, p. 300. V. Farsoe, Zeit. anal. Chem., 46. 308, 1907. After the action, the iodine solution is titrated with j^N-sodium arsenate. In a sample weighing 0*3624 grm., 10 c.c. of the T VN-sodium arsenate solution were needed. Hence, the red lead contained 32 '9 per cent, of PbO ? . C. Marchese (Gazz. Chim. Ital., 37. ii. 289, 1907) has examined several methods of estimating lead peroxide in red lead H. Forestier's method (Zeit. angew. Chem., II. 176, 1898 ; Atmal. Lab. Chim. R. Gabelle, 5. ii. 486, 1906), in which the red lead is heated for half an hour in a water bath with 10 c.c. of 10 per cent, acetic acid solution, and 20 c.c. of water. The insoluble residue after washing represents "lead peroxide and other insoluble matter." E. Szterkher's method (Ann. Chim. Anal. App., 7. 214, 1902) depends on the insolubility of lead peroxide in dilute nitric acid 2 c.c. of acid (sp. gr. 1'18) and 30 c.c. of water free from nitrous acid. Forestier's method requires a longer heating or stronger acid than is prescribed by its author. Topf's and Szterkher's methods are considered best. A. Partheil ( Ver. Ges. deut. Naturforsch. Aertze, 159, 1907) prefers Topf's process. 324 A TREATISE ON CHEMICAL ANALYSIS. 153. The Analysis of Lead Chromates. The chrome yellows and chrome reds are lead chromates or basic lead chromates with more or less lead sulphate, white lead, lead oxide, calcium carbonate, calcium sulphate, zinc oxide, ferric oxide, etc., along with soluble salts due to imperfect washing potassium sulphate, potassium bichromate, potassium perman- ganate, etc. A chrome yellow may be con- sidered adulterated if it contains anything besides insoluble lead chromates and lead compounds. The scheme of analysis depends upon the information desired. 1 Insoluble Matter. Digest 1 grm. of the dry powder (100's lawn) in boiling concen- trated hydrochloric acid, adding half a dozen drops of alcohol to the boiling liquid one at a time. This accelerates the dissolution of the lead chromates. Lead chloride PbCl 2 and chromic chloride CrCl 8 are formed. Evaporate the solution to dryness. Add a few drops of hydrochloric acid and 100 c.c. of hot water. Filter the hot solu- tion (a hot- water funnel, 2 fig. 130, is here useful). Wash the residue with hot water, 3 ignite, and weigh. The result is sometimes reported as "insoluble matter"; it contains barium sulphate, silica, etc. The silica can be volatilised by treatment with hydrofluoric acid and sulphuric acid (page 169). The residue is then reported as "barium sulphate," and the loss in weight by the hydrofluoric acid treatment is reported as "silica." 4 FIG. 130. Hot funnel. 1 E. F. Scherubeland E. S. Wood, Journ. Ind. Eng. Chem., 2. 482, 1910 ; P. H. Walker, Bull. U.S. Dept. Agric. Chem., 109. 29, 1910 ; M. Lachaud and C. Lapierre, Bull. Soc. Chim., (3), 6. 335, 1891 ; Compt. Rend., no. 1035, 1890. 2 HOT FUNNEL. There are numerous types of hot funnel very useful for filtering hot solutions which have a tendency to crystallise on cooling. In filtering saturated salt solutions liable to crystallise, it is best to use funnels with the stem cut off, e.g., Gattermann's funnel. A good temporary hot-water jacket can be made by wrapping a piece of flexible copper, lead, or "compo" gas piping spirally two or three times round the funnel and blowing steam through the spiral by connecting the pipe with a tin can containing boiling water, fig. 130 (A. Horvath, Liebig's Ann., 171. 135, 1874 ; 0. von Liebreich, Chem. Ztg. Repert., u. 153, 1887 ; Zeit. anal. Chem., 27. 387, 1888; 24. 582, 1885; V. Brudny, Zeit. Mikros., 26. 418, 1910). Another simple form is a plain sheet zinc or copper bath with a sheet zinc or copper funnel soldered into the bath, so that the glass funnel fits into the metal funnel, with the stem of the glass funnel below the bath. If necessary, the bath can be fitted with a condenser or with a constant level attachment. In Paul's steam-heated funnel (T. Paul, Ber., 25. 2208, 1886) the steam is blown through a copper tube bent in a spiral form to fit the funnel. The condensed steam returns to the boiler. A simple method of heating by an electric current is also available. 3 If lead sulphate be present, it must now be washed out by digesting the precipitate with sodium thiosulphate, or ammonium acetate, or ammonium chloride, before ignition, and precipitated from the solution by treatment with hydrogen sulphide (page 276). If the residue be coloured, repeat the digestion with a few drops of hydrochloric acid. Some basic lead chromates may be present, which dissolve very slowly in acids ; in that case, the addition of alcohol or hydrogen peroxide is advisable to hasten the process of dissolution. 4 There are objections to this practice. The alumina of clay, if clay be present, would thus be reported as barium sulphate. For a more exact method of treating the insoluble, see white lead, page 320. THE DETERMINATION OF LEAD. 325 Calcium Sulphate. Add a hot solution of barium chloride to the boiling nitrate, and determine the sulphates in the soluble portion as barium sulphate (page 322). Calculate the corresponding amount of calcium sulphate by multiplying the weight of the barium sulphate by 0*5837. Zinc Oxide and Calcium Carbonate. Digest a gram of the sample with concentrated acetic acid, boil, dilute to 100 c.c., filter, and wash with dilute acetic acid. The zinc oxide, the calcium salts, and some lead oxide pass into solution. Precipitate the lead and zinc by means of hydrogen sulphide, filter, and wash, as described on page 319. Add ammonia to the filtrate until the solution is alkaline ; precipitate the lime as calcium oxalate, and proceed as described on page 213. The weight of the calcium oxide so obtained, multiplied by 1 '7844, represents the corresponding amount of calcium carbonate. The weight of the calcium sulphate obtained in the preceding determination, multiplied by 0-7351, gives the equivalent amount of calcium carbonate. The difference in the two results represents the amount of calcium carbonate in the given sample. The mixed lead and zinc sulphides are treated as described on page 320, but the lead sulphate is rejected. The zinc is weighed as phosphate, page 366, and reported as zinc oxide. Ferric Oxide and Lead Chromate. Boil 1 grm. of the powdered sample with about 10 grms. of potassium hydroxide and 100 c.c. of water, filter, and wash with hot water. The alkaline nitrate containing the lead chromate in solution is acidulated with acetic acid, and insoluble lead chromate will be precipitated. This is collected on a weighed Gooch's crucible, washed with water, dried at 110, and weighed as lead chromate PbCr0 4 . The filter paper contains the ferric oxide, barium sulphate, silica, and lead oxide. Ignite the mixture slowly in a porcelain crucible so as to prevent the mass sticking to the sides of the crucible. Digest the residue with 5 c.c. of hot concentrated hydrochloric acid, and transfer the mixture to a beaker with 50 c.c. of water. Filter off the insoluble silica and barium sulphate. Nearly neutralise the filtrate with ammonia and precipitate the lead as lead sulphide. Wash and reject the precipitate. The filtrate is boiled to expel the hydrogen sulphide, and treated with a few drops of nitric acid to oxidise the iron. The iron is precipi- tated with ammonia, washed, ignited, and weighed as ferric oxide (page 182). Total Lead Oxide. One gram of the sample is digested with concentrated hydrochloric acid, evaporated to dryness, and the mixture treated with a few drops of hydrochloric acid arid about 100 c.c. of hot water. Boil the solution until it has a clear green colour. Add 5 to 10 grms. of sodium acetate to replace the free hydrochloric acid by acetic acid. 1 Add 5 c.c. of acetic acid and a, gram of potassium chromate. Stir the solution vigorously. The lead chromate is precipitated. Filter through a weighed Gooch's crucible, wash the precipitate with water, dry at 110, and weigh as lead chromate PbCr0 4 . 2 The difference in the weight of the lead chromate so determined and that previously obtained is multiplied by 0-6903, and the result reported as lead mon- oxide PbO. White Lead. If white lead be present, determine the amount of carbon dioxide in the given sample (page 552 or 553), and deduct the carbon dioxide corre- sponding with the calcium carbonate previouslydetermined. The result, multiplied 1 If iron be present, the solution turns reddish- brown. 2 The separation of lead as chromate is sometimes advantageous under conditions where the sulphate, sulphite, or molybdate separations would -be less convenient. This, for instance, is the case in separating small quantities of lead from copper and zinc in acetic acid solutions, and in separating lead from silver in ammoniacal solutions. If bismuth be present, some bismuth chromate will be precipitated with the lead chromate, and the two must be separated by other methods W. Diehl, Chem. Ind., 6. 157, 1883. 326 A TREATISE ON CHEMICAL ANALYSIS. by 17-62, gives the corresponding amount of white lead Pb(OH) 2 . 2PbC0 3 . The amount of white lead is multiplied by 0*2877 to get the corresponding amount of lead oxide PbO. This is deducted from the lead oxide previously obtained. Soluble Salts. The soluble salts acetates, bichromates, sulphates, nitrates, etc. present are generally derived from imperfect washing in the manufacture of the chromate. The soluble salts are determined by weighing, say, 5 grms. of the dry (109) impalpable powder on to an asbestos Gooch's crucible (dried at 109), and washing six times with 25 c.c. of cold water. Dry the contents of the crucible at 109 and weigh. The loss in weight represents the soluble salts removed by the water. The washings may be examined, if desired, and the salts just named specially determined qualitatively, or quantitatively. 1 Results. Scherubel and Wood's test analysis on a known mixture gave the following percentage results : Si0 2 . BaS0 4 . Fe 2 3 . CaS0 4 . ZuO. CaC0 3 . PbCr0 4 . PbO. Used . . 476 476 476 476 9'52 476 57'14 9'54 Found . .4-52 470 4 '88 4 -90 4 '54 4 '64 57 '04 9 '68 These results must be considered satisfactory for technical work. Naples Yellow. In evaluating this substance, lead and antimony are to be separated. Determine the antimony volumetrically in an aliquot portion of the solution : the lead may be determined by precipitation as sulphide by hydrogen sulphide in an alkaline solution, hot, if necessary, to keep the lead chloride in solution. 2 154. The Determination of Silver in Lead Compounds by the Turbidity Method. Silver in small quantities is best determined by cupellation. The amount present in red lead, white lead, etc., is usually too small to be determined satisfactorily by other gravimetric 3 or volumetric processes. Blunt 4 determines the silver by the following turbidity process : Dissolve 10 grms. of, say, red lead in 50 c.c. of dilute nitric acid (sp. gr. 1*42, 1 vol.; water 4 vols.) quite free from chlorides. Dilute to 120 c.c.; filter; wash the residue with a little distilled water; make the solution up to 200 c.c. ; pipette 100 c.c. into the left test glass of the colorimeter. The other 100 c.c. is treated with a drop of concentrated hydrochloric acid and stirred. The lead chloride passes into solution and the silver chloride produces a white turbidity. Let all stand an hour ; and filter the solution into the right test glass of the colorimeter through a small 5*5 -cm. filter paper. Wash with distilled water. Add a standard solution of silver nitrate 5 drop by drop with constant stirring until the turbidity in both vessels appears the same. The volume of the silver nitrate solution used represents the amount of silver in the given solution. See page 654. 155. The Determination of Silver and Gold by Cupellation. This process is very old, and an enormous number of determinations of gold and silver by this process are made daily in assay offices in different parts of the 1 In this case probably more powder will be needed. 2 The lead may also be precipitated as oxalate, and the lead oxalate determined by titration with potassium permanganate, as indicated for lime (page 215). 3 Precipitation as silver chloride or iodide. 4 T. P. Blunt, Chem. Neivs, 32. 3, 1875 ; J. Krutwig, Ber., 15. 307, 1264, 1882. 5 SILVER NITRATE SOLUTION. Dissolve 1'575 grms. silver nitrate in water and make the solution up to a litre. 1 c.c. represents O'OOl grm. of silver Ag. THE DETERMINATION OF LEAD. 327 world. A known weight of the lead compound is reduced to the metal. The metal is then heated in a vessel, called a cupel, made of bone ash, or some other suitable material, whereby the lead is oxidised and absorbed by the cupel ; the silver remains behind as a small bead. The silver will be alloyed with gold, if gold be present. Roasting and Fusion. The object of the fusion is to collect the silver in a button of metallic lead silver alloy. Intimately mix 100-120 grms. of dry litharge, red lead or white lead, with the flux sodium bicarbonate 60 grms., argol 2 grms. on a sheet of glazed paper. The mixture is transferred to a fire- clay crucible so as to fill the crucible not more than three-quarters full, 1 and the whole contents covered with a layer of finely powdered dry borax glass. The lid is placed on the crucible and the whole heated gradually to prevent breaking the crucible or prevent the charge from blowing. When the contents of the crucible have ceased bubbling and all is in a state of quiet fusion, the crucible is removed from the fire and gently tapped, sides and bottom, in order to assist the molten metal inside to collect on the bottom of the crucible. The molten contents of the crucible are usually poured into conical iron moulds; or the crucible may be allowed to cool, and then broken. The slag is separated from the lead button by hammering the button on an anvil into a cube with its corners flattened. 2 The button should weigh about 20 grms. " If it weighs more, use less argol ; if less, more argol. Cupellation. The object of the cupellation is to remove the "base" metals by oxidation and absorption in the cupel. The " noble" metals gold and silver remain on the cupel bottom after the operation in the form of a small bead. Place three or four empty cupels 3 in a red-hot muffle. 4 In about 10 minutes the button of lead is placed by means of the "cupel tongs" in the centre of a hot cupel free from cracks. Close the door of the muffle. The lead melts (326) ; a black scum develops on its surface. This disappears in a few minutes (about 675), and the molten lead has a bright silvery appearance. The door of the muffle should be opened, but not wide enough to allow cold air to impinge directly on the muffle. Oxidation now sets in vigorously. The flakes of oxide which form on the surface of the button "slide" down the convex surface of the button and are absorbed by the cupel. The temperature of the cupel should not exceed 750 a temperature below the melting point of litharge (about 900). The most common mistake at this stage is too high a temperature. If the temperature of the muffle is right, "feathers" of litharge appear on the side of the cupel near the doors and on the upper rim of the cupel. If the temperature is too low, feathers form low down in the cupel. If the muffle be not hot enough, put a lump of charcoal in front of the muffle. This will raise the temperature a little. Sometimes the action stops during cupellation and the button solidifies. This is called " freezing." This arises when the lead has been oxidising more rapidly than the lead oxide has been absorbed by the cupel. In that case add more lead to the cupel. If " freezing " be due to the low temperature of the 1 Say, a Battersea round, approximately size E. 2 If portions of lead or pasty-looking masses adhere to the sides of the crucible, the fusion is defective, and another fusion must be made. 3 Shallow cups of bone ash or other suitable material. The success of the work is largely dependent on the quality of the cupel. If many cupels are required, they can be made with good bone ash in "cupel moulds " ; if but few determinations are to be made, old and dry cupels of good quality can be purchased from the dealers. A good cupel will absorb its own weight of lead oxide. The cupel should be nearly twice as heavy as the lead button. 4 That is, varying from " red "to " bright red." Not "dull red," nor "white," 328 A TREATISE ON CHEMICAL ANALYSIS. muffle, the temperature of the latter must be raised a little. The result of a cupellation with a button which has frozen is to be regarded with suspicion. Traces of foreign metals copper, antimony, iron, zinc, etc. are partly absorbed by the cupel, and partly volatilised as oxides. As the absorption of lead continues, the button becomes more and more spherical. At this stage the temperature of the muffle should be raised, or the cupel pushed into a hotter part of the muffle. When the last of the lead has gone, the button appears to revolve axially in the cupel, and it also seems to be covered with an iridescent film. The colours disappear ; the bead becomes dull and acquires a silvery tinge. If the temperature of the muffle be below the melting point of the button (962 for silver), or if the cupel be withdrawn from the muffle, the bead suddenly becomes very bright "flashing." Close the doors of the muffle for one or two minutes so as to remove the last traces of lead. The cupel may then be covered by placing a hot empty cupel over the one containing the bead, and the cupels removed from the muffle. The object, of course, is to ensure slow cooling in order to prevent " sprouting " caused by the evolution of occluded oxygen by the silver bead. If the bead should sprout, start again. When cold, the bead is removed with a pair of forceps. . If the bead be excessively small, merely touching it with a wet pin and lifting it on to a watch-glass will suffice. Dry the bead by warming, and weigh. 1 The bead should be examined with a lens before weighing to make sure that no particles of bone ash adhere to the bead. If the bead is so contaminated, hold it in the "bead forceps," and brush it clean with a hard brush "button brush." The bead contains both gold and silver. If the bead is white, it contains more than half its weight of silver ; if yellow or reddish-yellow, it has about three-quarters of its weight gold. Intermediate tints represent different pro- portions of gold and silver. For the determination of the amount of gold in the bead, see page 432. In the case of red and white leads, if the silver be deter- mined at all, it will usually suffice to report the weight of the bead as " silver." 156. The Analysis of Galena. Native galena PbS may be accompanied by products of the decomposi- tion of lead sulphide anglesite and cerusite quartz, silicates, blende, calcite, fluorspar, heavy spar, and pyrites. Certain varieties also contain arsenic and antimony ; and argentiferous galena may contain up to 1 *0 per cent, of silver. Decomposition 'of the Galena. Moisten, say, half a gram of the finely divided sample in a 250-c.c. basin with dilute nitric acid ; and, after the cold mixture has stood for a few minutes, digest on the water bath with 15 c.c. of concentrated nitric acid. The basin is covered by a clock-glass during the reaction to prevent loss by spurting. When the violence of the action is over, evaporate the mixture to dryness on the water bath. 2 Add 10 c.c. more of concentrated nitric acid, 1 If the bead of silver be too small to weigh, its volume can be measured with the microscope. Suppose the bead has a diameter 0'18 mm. For the method of measuring under the microscope, see the next volume of this work. The measurement is supposed to be the average of measure- ments in two or three different directions. The volume of the bead is 0*5236 x (0'18) 3 cub. mm. If the specific gravity of the silver be 10'5, the bead of silver will weigh 10'5 x 0'5236 x (0'18) 3 =: 0-032 mgrm. or '000032 grm. V. Goldschmidt, Zeit. anal. Chem., 17. 142, 1878; 16. 434, 1877; C. F. Fohr, ib., 22. 195, 1883; Chem. News, 50. 114, 1884; G. Tate, ib., 61. 43, 54, 1890 ; G. A. Gozdorf, ib., 54. 231, 1886 ; D. Forbes, ib., 15. 231, 1867 ; G. A. Goyder, ib., 70. 194, 203, 1894 ; J. W. Richards, Journ. Amer. Chem. Soc., 23. 203, 1901 ; J. S. Curtis, Berg. Hiitt.Ztg., 47. 3, 1888; M. Guerreau, Bull. Soc. CMm. (3), 27. 1902; Chem. News, 86. 194, 2 A little hydrochloric acid generally aids the dissolution of the galena by dissolving the film of lead nitrate, which is almost insoluble in the strong acid, and thus protects the galena from further attack, THE DETERMINATION OF LEAD. 329 some water, and 10-15 drops of bromine. Heat the mixture with frequent stirring on the water bath, so as to complete the oxidation of the sulphide. To decompose the bromates, add more nitric acid, and evaporate to dryness. Repeat this operation in all three times. Boil the dry residue with 60 c.c. of water and 20 c.c. of concentrated hydrochloric acid so as to dissolve the lead sulphate. Filter off 1 the residual gangue ; wash well with boiling water. If desired, the residue maybe ignited, weighed, and reported as "insoluble gangue." This may contain the quartz, insoluble silicates, some silver chloride, and the greater part of the barium sulphate. 2 Transformation of the Decomposition Products into Sulphates. Add 10 c.c. of dilute sulphuric acid (1 : 1) to the nitrate (or an aliquot portion of the nitrate) from the insoluble gangue, and evaporate the mixture until copious white fumes of sulphuric acid appear. The evaporation should be carried nearly to dryness in order to ensure complete conversion of the lead salts into sulphates ; to prevent too large an excess of acid ; and to drive off the nitric and hydrochloric acids. Let the mixture cool. Separation of the Lead Sulphate. Gradually add 100 c.c. of cold water, and boil the mixture so as to dissolve the ferric sulphate, etc. The lead sulphate remains behind, insoluble. Decant the clear solution through a filter paper, and wash by decantation with dilute sulphuric acid, as indicated on page 317. If the lead sulphate is to be determined gravimetrically, transfer it 3 to a Gooch's crucible or filter paper, as indicated on page 318; if the lead is to be determined volumetrically, 4 keep as much solid in the beaker as possible. Wash the paper with alcohol to remove sulphuric acid. Dissolve the precipitate on the paper by pouring a hot concentrated solution of ammonium acetate, slightly acidulated with acetic acid, through the filter paper into the beaker containing the remainder of the lead sulphate. The amount of lead in the ammonium acetate solution can then be determined by the molybdate volumetric process (page 333). Bad results by this process can often be traced to an imperfect dissolution of the lead sulphate on the filter paper, etc. Determination of Other Constituents. Copper, antimony, and arsenic can be precipitated as sulphides from the filtrate, and the copper separated from the antimony and arsenic, as indicated on page 277. The copper can be determined as indicated, page 351 ; the antimony and arsenic determined as indicated, page 304. The filtrate is boiled to drive off the hydrogen sulphide, oxidised with a few drops of hydrogen peroxide, and heated to boiling with an excess of ammonium chloride and a slight excess of ammonia. The precipitate is treated as indicated on page 182. Zinc, if present, will be found in the filtrate, and determined as indicated on page 366. The sulphur can be determined by oxidis- ing the powdered sample with nitric acid and bromine, and finally weighed as barium sulphate (page 618). Other Methods of " Opening " Galena. Other methods of decomposition are sometimes useful alternatives. Galena, for example, can be reduced to "metal " 5 by fusing a mixture of, say, 25 grms. with 75 grms. of potassium cyanide, and 1 If lead only is to be determined by a volumetric process, this filtration may be omitted. 2 The difference in weight after treatment with hydrofluoric acid will give the silica. Fusion with sodium carbonate and extraction with water will give an insoluble residue of the carbonates. These can be dissolved in nitric acid, the silver precipitated with hydrochloric acid, and the barium with sulphuric acid (pages 652 and 618). The total silver may be determined by cupellation. 3 Of course, previously freed from "insoluble gangue." 4 If the amount of lead is alone to be determined volumetrically, there is no need to filter off the " insoluble gangue." 5 A. W. Warwick, Chem. News, 63. 30, 145, 1891 ; F. Jean, Bull. Soc. Chim. (3), 9. 253, 1893. 330 A TREATISE ON CHEMICAL ANALYSIS. covered with a layer of 12 grms. of powdered anhydrous borax. Cover the crucible with a lid, and heat the mixture to redness for about 5 minutes (page 269). The metallic button is washed free from cyanide, dissolved in dilute nitric acid (1 : 5), and treated for lead by the sulphuric acid process (page 317). The sodium peroxide fusion (page 266) is often most convenient for the determination of sulphur ; and the process indicated by footnote, page 326, enables a determination of the lead to be made very quickly. 1 157. The "Government Test" for the Solubility of Lead Frits. The directions given in the Home Office Circular* for this test are as follows : "No glaze into the composition of which the fritted lead enters shall be regarded as satisfying the requirement as to insolubility which yields to a dilute solution of hydrochloric acid more than 2 per cent, of its dry weight of a soluble lead compound calculated as lead monoxide, when determined in the following manner : A weighed quantity of dried material is to be continuously shaken for one hour, at the common room temperature, with 1000 times its weight of an aqueous solution of hydrochloric acid containing 0'25 per cent, of HC1. This solution is thereafter to be allowed to stand for one hour and to be passed through a filter. The lead salt contained in an aliquot portion of the filtrate is then to be precipitated as lead sulphide and weighed as lead sulphate." The directions here are incomplete. It is difficult, indeed impossible, to filter some glazes satisfactorily through filter paper ; and if some silica gets into solution, it may, later on, be precipitated with the lead sulphide. Hence, dry the glaze at 100. 3 Transfer, say, 0'5 grm. of the powder into a 500-c.c. Stohmann's shaking bottle closed with a rubber stopper, and fill the bottle up to the mark on the neck 4 with 0'25 per cent, hydrochloric acid. Shake the mixture for an hour in, say, Wagner's shaker, fig. 131. Let the bottles stand for one hour. Filter off a convenient quantity, say, 425 c.c. of the clear solution. 5 Evaporate to dryness on a water bath. Take up the residue with dilute hydrochloric acid, so that the solution is but slightly acid (page 276). Filter and wash the silica. Add 1 c.c. of a saturated solution of mercuric chloride (page 319). The mercury may be ignored, because it is volatilised later on. Precipitate the lead 1 R. Benedict (Chem. Ztg., 16. 43, 1896) decomposes galena by digestion with hydriodic acid (sp. gr. 1 7). Subsequent digestion with nitric acid transforms the lead iodide into nitrate, and the lead is then precipitated as sulphate. According to F. H. Storer (Chem. Neivs, 21. 137, 1870; A. Mascazzini, Zeit. anal. Chem., 10. 491, 1871 ; F. Stolba, Journ. prakt. Chem. (1), 101. 150, 1867 ; F. Mohr, ib., 12. 142, 1873), galena is decomposed rapidly and completely by dilute hydrochloric acid (1 : 4) in contact with metallic zinc. Besides galena, metallic lead may be precipitated quickly and completely from the sulphate, nitrate, chromate, oxide, carbonate, and chloride. The clear liquid decanted from the residue shows no trace of lead with hydrogen sulphide. C. Boucher (Ball. Soc. Chim. (3), 29. 933, 1904 ; Chem. News, 89. 56, 1904) attacks pyrites and galena by heating the powdered mineral intimately mixed with 4 to 5 times its weight of a mixture of 3 parts of sodium persulphate with 1 of ammonium nitrate in a dish or flask on a sand bath. For other methods of analysing galena, see J. A. Muller, Bull. Soc. Chim. (3), 31. 1303, 1904 ; Chem. News, 92. 15, 1905; P. Jannasch and H. Kammerer, ib., 72. 78, 1895; Ber., 28. 1409, 1895; J. K. Meade, Journ. Amer. Chem. Soc., 19. 374, 1897 ; Chem. Eng., u. 49, 1910 ; Chem. News, 101. 137, 1910 ; W. Stahl, Berg. Hutt. Ztg., 48. 237, 1889. a K. E. Digby, Home Office Circular, Dec. 14, 1889. 3 If a slop glaze be supplied, it must be thoroughly agitated and a portion evaporated to dryness, and the dry powder thoroughly mixed. Special attention is here needed, because the glaze is not usually homogeneous after it has been dried in a basin on the water bath. 4 500 grms., or 499'4 c.c. of 0'25 per cent, acid at 15. The method in the text is sufficiently exact. 5 Reject the first 10 c.c. which pass through the paper. THE DETERMINATION OF LEAD. 331 as sulphide by means of hydrogen sulphide. Let the gas bubble slowly through the cold solution for about half an hour. Let the precipitate settle about an hour. Filter and wash with cold " H 2 S " water. The precipitate is supposed to be lead sulphide. 1 Incinerate the precipitate in a porcelain crucible at a low temperature 2 in order to burn off the filter paper. Brush the residue into a 250-c.c. beaker. Wash the crucible with dilute nitric acid and proceed as indicated for lead sulphate (page 317). The weight of the lead sulphate, multiplied by O7359, gives the corresponding amount of lead monoxide PbO in the 425 c.c. of solution. Hence, by proportion, calculate the amount in the 500 c.c., that is, in 0*5 grm. of the glaze. The result multiplied by 200 gives the so-called " percentage solubility of the lead frit," or "per cent, of soluble lead" in terms of lead monoxide. The differences obtained in duplicate determinations with quantities of " soluble FIG. 131. Shaking apparatus. lead" up to 10 per cent, run to about 0'1 per cent. ; for larger quantities 20 to 40 per cent, "soluble lead" differences in the duplicates run about 0'2 per cent. 158. The Gravimetric Determination of Lead as Molybdate. Lead can also be advantageously determined as molybdate and as chromate. The advantage of the molybdate precipitation is that the precipitate may be 1 Many glazes contain a little zinc oxide, and some zinc sulphide might be precipitated from the dilute acid with the lead sulphide. Note, other members of the hydrogen sulphide group might be present, and unless a separation were made, the "PbO solubility" would be too high. 2 If much lead be present, the filter paper had better be ignited alone (page 317). Instead of following the method described in the text, the lead sulphide and filter paper can be incinerated until the paper is charred to carbon ; when nearly cold, carefully add a little fuming nitric acid, a drop at a time, from a pipette. The carbon will be oxidised, and the lead sulphide transformed into sulphate. This method requires a little practice, but it saves a great deal of time and is more exact than the roundabout method described in the text, once the manipulation is mastered. 332 A TREATISE ON CHEMICAL ANALYSIS. ignited along with the paper, and no particular harm results from a prolonged ignition at a higher temperature than is required to destroy the paper. The errors in manipulation are also reduced more in calculating the weight of the molybdate to PbO than with the corresponding sulphate or chromate. The reaction has been studied in particular by Chatard l and Brearley, and Chatard's description of the disturbing effect of an excess of the precipitating agent has led to the process being viewed with a suspicion which the method may not deserve. Suppose that the lead salt be in solution, add a slight excess of ammonia and then acidify the solution with acetic acid. Add about 5 grms. of ammonium chloride, and heat the solution to boiling. 2 An excess of an aqueous solution of ammonium molybdate 3 is then gradually added with constant stirring. Boil the solution two minutes. Let the cream-coloured precipitate gradually settle until cold. Filter through a weighed Gooch's crucible. Wash with hot water contain- ing about 2 per cent, of ammonium acetate in solution. Ignite and weigh as lead molybdate PbMo0 4 . Multiply the weight of the lead molybdate so obtained by O6077 to get the corresponding weight of lead monoxide PbO. The result is not more than 1 per cent, out with the elements which form insoluble molybdates nickel, zinc, cobalt, magnesium, aluminium and the effects of these elements, together with calcium, 4 strontium, and barium, if present, can be eliminated by reprecipitation. Arsenates and phosphates should not be present. The slight excess of free molybdic acid which may con- taminate the first precipitate is eliminated by reprecipitation. For reprecipitation, dissolve the precipitate of lead molybdate in hydro- chloric acid ; heat the solution to boiling ; add a sufficient excess of ammonium acetate to neutralise the free hydrochloric acid, and proceed as described above. The lead may be first precipitated as sulphate. Dissolve the sulphate in a concentrated solution of ammonium acetate, and add an excess of ammonium molybdate, as indicated above. Excellent results can be obtained by this method. 159. Conversion Factors. In gravimetric analyses it is comparatively rare to find that the constituent to be determined can be weighed directly, as were silica, alumina, ferric oxide, and lime. In most cases the constituent to be determined is separated as an "insoluble" salt barium sulphate, lead sulphate, silver chloride, magnesium pyrophosphate, etc., and the corresponding amount of baryta, lead oxide, silver, magnesia, etc., determined by multiplying the observed weight by a factor, the so-called conversion factor. The errors of experiment are obviously multiplied at the same time. Any error in the preparation and weighing of the precipitate of, say, lead sulphate PbS0 4 in the determination of the lead oxide will be distributed between the lead oxide and the sulphuric anhydride. In lead sulphate: PbO : S0 3 = 3 : 1, very nearly; or, more exactly stated, the conversion factor for transforming the lead sulphate to lead oxide is O7359. Hence, a total error of O8 per cent, in the determination of the lead sulphate corresponds with an error of O6 in the determination of the lead oxide. If the lead be determined as lead molybdate PbMo0 4 we have very nearly : PbO : Mo0 3 = 3:2; or, more exactly, the conversion factor for transforming the weight of the lead 1 T. M. Chatard, Chem. News, 24. 175, 1871 ; Amer. J. Science (3), I. 416, 1871 ; J. F. Sacher, Chem. Ztg., 33. 1257, 1909 ; see page 414. 2 The precipitate is difficult to filter in the absence of this or some other coagulating salt. 3 AMMONIUM MOLYBDATE SOLUTION. Dissolve 34 '34 grms. of ammonium heptamolybdate in water ; make the solution faintly acid with acetic acid ; and make the solution up to litre 4 Calcium molybdate is very prone to precipitation with the lead molybdate. THE DETERMINATION OF LEAD. 333 molybdate to lead oxide is 0*6077. Hence, an error of 0*8 per cent, in the determination of the lead molybdate means that there will be an error of 0-5 in the determination of the lead oxide. Consequently, other things being equal, the smaller the factor the less the influence of errors in the determination upon the final result. Hence, given two rival analytical processes which are liable to errors of experiment of the same magnitude, and which involve the same amount of manipulation, that process will be chosen which has the least conversion factor. For this reason, among others, silver is better precipitated as bromide than as chloride. The conversion factor in the former case is 0*5744, and in the latter 0*7526. As a matter of fact, in spite of the theoretical advantages of the bromide process, the chloride process is preferred by many because the sources of error and corrections have been well explored for the chloride process, but not for the bromide process. 160. The Volumetric Determination of Lead Molybdate Process. Schindler, in 1888, proposed to determine lead as lead molybdate by means of a standard solution of ammonium molybdate. The titration is made in a hot ammonium acetate solution of lead sulphate slightly acidified with acetic acid. 1 The reaction is symbolised : Pb(C 2 H 3 2 ) 2 + (NH 4 ) 2 Mo0 4 -> PbMo0 4 + 2NH 4 C 2 H 3 2 . According to Kroupa, the presence of arsenic, antimony, bismuth, phosphorus, and zinc does not interfere with the process. Barium and strontium salts lead to low results. These substances appear to act by retarding the dissolution of the lead sulphate. The difficulty is overcome by repeatedly boiling the sulphate with hot ammonium acetate to ensure complete solution. Bannister and M'Namara found that the results are high if calcium salts be present. E.g., with test solutions containing the equivalent of 0*2005 grm. of lead and to 0*2415 grm. of calcium sulphate, the results ranged from 100 to 109 per cent, of lead in place of 100. About a gram of the lead say, white lead is dissolved in acetic acid or in dilute nitric acid ; for the dissolution of red lead, see page 322. The insoluble sul- phates are filtered off. The lead is precipitated as sulphate, filtered, and washed, as indicated, page 319. If calcium salts be present, special care must be taken to wash out the calcium sulphate before proceeding further. The precipitate is dissolved in a concentrated solution of ammonium acetate. Acidify the solution with 2-3 c.c. of acetic acid, and heat the solution to boiling. Titrate the hot solution with standard ammonium molybdate 2 from a burette until a drop gives a slight but distinct yellow coloration with a drop of freshly prepared 3 aqueous 1 C. Schindler, Zeit. anal. Chem., 27. 137, 1888 ; H. Weber, ib., 42. 628, 1903 ; I. C. Bull, ib. t 41. 653, 1902; School Min. Quart., 23. 348, 1903; Chem. News, 87. 40, 52, 66, 1903 ; H. H. Alexander, Eng. Min. Journ., 55. 298, 1893 ; J. F. Sacher, Chem. Ztg., 33. 1257, 1909 ; J. A. Miiller, Bull. Soe. Chim. (3), 31. 1303, 1904 ; R. Kroupa, Berg. Hutt. Ztg., 53. 411, 1894 ; J. F. Sacher, Chem. Ztg., 33. 1257, 1909 ; C. 0. Bannister and W. M'Namara, Analyst, 37. 242, 1912. See Chem. Eng., 16. 36, 1912. 2 STANDARD AMMONIUM MOLYBDATE SOLUTION. Dissolve 8*022 grms. of ammonium molybdate in water and dilute the solution to a litre. If the solution be turbid, add some ammonia before the dilution is completed 1 c.c. will be nearly equivalent to O'Ol grm. PbO. To standardise the solution, dissolve, say, five known amounts of lead sulphate in hot ammonium acetate ; add 2-3 c.c. of acetic acid ; and titrate as indicated in the text. Schindler also uses the converse reaction, titration with standard lead acetate, for the volumetric determination of molybdenum. 3 TANNIN SOLUTION. About 0*5 grm. of tannin dissolved in 100 c.c. of water. Ammonium molybdate gives with tannin a coloration varying from blood-red to a pale yellow, according to 334 A TREATISE ON CHEMICAL ANALYSIS. solution of tannin spot test. If the solution cools during the titration, it should be heated again to finish the titration. Titrating with "Spot-test" Indicators. It is easy to over-titrate the solution when using "spot-test" indicators, and three methods of titrating to rectify or avoid over-titration may be indicated : (1) Some prefer to use a standard solution of lead nitrate 1 for checking the first titration. Suppose the titration is finished, add 10 c.c. of the standard lead nitrate solution, and continue the titration. 0*1 grin, of lead must be subtracted from the result of the titration. (2) For the titration it is well to place about half the solution to be titrated in a separate beaker and titrate the other half by adding a cubic centimetre of the ammonium molybdate solution at a time, in order to get a rough idea of the amount needed. Suppose the one portion requires between 15 and 16 c.c. for the titration (no coloration with 15 c.c. ; yellowish-brown coloration with 16 c.c.). Add the untitrated portion to the portion just titrated, and 14 c.c. can be safely run from the burette, and the titration finished by small additions from the burette. (3) In titrating with a spot test, particularly with slow reactions, the testing should be done systematically. Suppose the first stop be made at 29 c.c., the second at 29 '2 c.c., the third at 29 '4 c.c., and so on. If the brown coloration develops on the sixth spot, it would mean that 29 + (n 1)0*2 = 30*0 c.c. has been added (n = 6). But the coloration may develop on the fourth spot after it has been passed. It is then easy to determine the burette reading for the fourth spot namely, 29 + (n - 1)0*2 = 29*6 c.c., where n = 4. Lead carbonate may also be precipitated by the addition of ammonium car- bonate ; the precipitate dissolved in acetic acid ; and the solution titrated with potassium ferrocyanide, 2 as indicated under zinc. The results are good. 161. The Electrolytic Process for the Determination of Lead. Metallic lead can be readily deposited from alkaline electrolytes containing phosphates, alkaline plumbate, double oxalates, double cyanides, etc. But it is so difficult to dry the lead for weighing without oxidation, that it is considered better to take advantage of the fact that lead is deposited from acid solutions .in the form of lead dioxide Pb0 2 on the anode, not the cathode. Hence, the dish, with its inner surface matt, is fitted as described on page 256, fig. 118; but the direction of the current is reversed so that the dish is the anode, and the platinum disc the cathode. The details indicated on page 258 are to be followed, with the modifications indicated in the following paragraphs. The Electrolyte. Suppose that 1 grm. of lead nitrate be dissolved in 30 c.c. of distilled water. Add 25-35 c.c. of concentrated nitric acid. Dilute the solution to 150-160 c.c. The Electrolysis. Electrolyse the solution with a current density of 0*5 to 1*8 amps., 3 and a voltage of 2*0 to 2'5 volts, at atmospheric temperatures. 4 As the concentration of the solutions. The colour is visible with a dilution 1 : 400,000 (Schindler). Lead molybdate gives no coloration. Concentrated solutions of lead acetate give a faint greenish- yellow colour, which cannot be confused with the coloration due to ammonium molybdate. J. L. Danziger uses a solution formed by saturating acetic acid with crystalline stannic chloride and saturating the solution with ammonium thiocyanate. When this is used instead of tannin as indicator, the end of the reaction is indicated by a pink coloration spot test. 1 LEAD NITRATE SOLUTION. 16 grms. of lead nitrate per litre corresponds with O'Ol grm. lead per c.c. 2 M. Yvon, Journ. Pharm. Chem. (5), 19. 18, 1889 ; A. H. Low, Journ. Amer. Chem. Soc., 15. 548, 1893. 3 For an overnight electrolysis (10-12 hours), use a current density of 0'5 amp. 4 If the operation be conducted at 50-55, the time required for the electrolysis is shortened to between 1 and 1| hours, using a current density of 1*3 to 1*6 amps., and 2'2 to 2*6 volts. THE DETERMINATION OF LEAD. 335 soon as electrical circuit is closed a yellow deposit appears on the anode (dish), which becomes orange, red, and finally dark brown or black. The electrolysis occupies between 2 and 3 hours. 1 If a deposit of metallic lead should appear on the platinum disc (cathode) during the electrolysis, add a little more concen- trated nitric acid ; or interrupt the current for about half a minute during the middle and towards the end of the electrolysis. When the electrolysis is complete, the deposit is washed with water, absolute alcohol, and absolute ether in the usual way. If the electrolyte be syphoned off before the current is stopped (fig. 119, page 260), keep the deposit quite covered with liquid all the time the syphon is in action. In other words, pour distilled water into the dish as fast as it is syphoned off. Drying and Weighing the Lead Dioxide. The lead dioxide retains water with great tenacity, 2 and the results will be too high if the regular factor 0*8662 for converting the lead dioxide into the equivalent amount of metallic lead be employed. Classen recommended drying the lead dioxide at 180-190, and Hollard at 200. Smith has shown that deposits of lead dioxide approximately 0'5 grm. in weight do not really have a constant composition until they have been dried at 230. For instance, with a deposit containing the equivalent of 0-4992 grm. of Pb : Time heated ... 30 30 30 30 30 min. Temperature . . . 200 200' 230 240 270 Pb0 2 0-5788 0-5790 0'5780 0'5780 '5780 grm. Similarly, with the equivalent of approximately 0*25 grm. Pb, the lead dioxide attains a constant weight at 200. In the former case, the factor of conversion corresponds with the empirical value 0-8634 ; and in the latter case, 0*8643. With still less quantities of the dioxide, say, 0*05 to O'l grm., a special factor is not needed. If the theoretical factor be used the results will be high. High results can be prevented by the use of empirical factors. Fischer recommends : Amount of Pb . Below O'l O'l Bet. 1 and '3 0'5 1*0 grm. Factor . . . 8660 0'8658 0*8652 0'8629 0*8610 Instead of following this procedure, May 3 recommends gentle ignition of the dioxide at a low temperature to convert the dioxide into lead monoxide PbO before weighing. The results are then excellent. Thus, Treadwell gives for four experiments : Mean. Pb0 2 found . . . 0*2202 0*2200 0'2203 0'2202 0*2202 grm. PbO found . . . 0*2042 0'2046 0'2043 0'2044 0-2044 grm. According to J. G. Fairchild (Journ. Ind. Eng. Chem., 3. 902, 1911), the essential conditions for good deposits are : (1) hot solutions (50-60) ; and (2) an initial low amperage. 1 To recognise the end of the electrolysis, withdraw a few drops of the solution from the dish ; make the solution alkaline with ammonia ; add a few drops of H 2 S Avater, or ammonium sulphide, when a black or dark brown precipitation represents lead and shows that the electro- lysis is not completed. Of course other recognised qualitative tests for lead are available. If the level of the electrolyte be raised a little by the addition of distilled Water, the appearance of a yellow or orange film on the newly immersed surface of the anode also shows that the electrolysis is not ended. 2 It is very doubtful if the excessive weight of the lead dioxide is due to the formation of a higher oxide. A. Hollard, Bull. Soc. Chim. (3), 29. 151, 1903 ; (3), 31. 239, 1904 ; Compt. Rend., 136. 229, 1904 ; Chem. News, 79. 122, 1899 ; R. 0. Smith, Journ. Amer. Chem. Soc., 27. 1287, 1905 ; R. C. Benner, Journ. Ind. Eng. Chem., 2. 348, 1910 ; J. G. Fairchild, ib., 3. 902, 1911 ; F. Utz, Chem. Zeit., ii, 1788, 1912. 3 W. C. May, Amer. J. Science (3), 6. 255, 1873 ; F. P. Treadwell, Kurzes Lehrbuch der anahjtischen Chemie, Leipzig, 2. 148, 1911 ; H. J. S. Sand, Chem. News, IOO. 269, 1909. 336 A TREATISE ON CHEMICAL ANALYSIS. In another series of trials the following average results were obtained : Lead used in each trial . . '1898 grm. Lead found by weighing as Pb0 2 . '1907 grm. Lead found by converting Pb0 2 ->PbO .... '1898 grm. Removing the Deposited Lead Peroxide from the Electrode. Digest the dioxide with nitric acid to which a little hydrogen peroxide has been added (say, 1 c.c.). The lead dissolves rapidly and quickly. Smith recommends removing the deposit by the action of dilute nitric acid along with a rod of copper or zinc. Henz 1 recommends a solution of sodium or potassium nitrite acidified with nitric acid. Preparation of Lead Sulphate for Electrolysis. Lead is commonly separated as sulphate in gravimetric work. According to Classen, lead sulphate is brought into a condition for electrolysis by warming the precipitated lead sulphate with an excess of ammonia, whereby lead hydroxide is formed. Pour the mixture with constant stirring into a platinum dish containing about 20 c.c. of warm concen- trated nitric acid. If any lead sulphate appears as a precipitate, it will soon dissolve in the solution. The solution is then ready for electrolysis. Lead sulphate is also brought into a condition for electrolysis by dissolving it in ammonium acetate, adding 20 c.c. of concentrated nitric acid, and electrolysing the solution at 60 with a current density of 1-5 to 1-7 amps. 2 Marie 3 places the lead sulphate in the dish to be used for the electrolysis ; the dish is heated on the water bath along with some dilute nitric acid. Grad- ually add crystals of ammonium nitrate until all is dissolved. Every 0*3 grm. of lead sulphate requires 5 grms. of ammonium nitrate. Dilute the solution with warm water so that the liquid contains about 10 per cent, of free acid, and electrolyse the solution at 60-70. For lead silicate, decompose the fine powder with sulphuric and hydrofluoric acids. Too great an excess of sulphuric acid prevents the solution of lead sulphate by the ammonium nitrate. 4 Effect of Foreign Electrolytes. The process indicated above will separate lead from the alkalies, alkaline earths, chromium, beryllium, zirconium, iron, uranium, zinc, nickel, cobalt, and cadmium. Chlorides should be absent from the electro- lyte. 5 The determination of lead by this process in the presence of appreciable amounts of arsenic will be unreliable. 6 The results will be low. Classen says that "when enough (0*5 grm.) is present, no lead will be deposited as dioxide on the anode, but metallic lead mixed with arsenic will be deposited on the cathode. If the electrolysis be continued for some time, the arsenic will be gradually driven from the cathode as arsenic hydride, and the precipitated lead will pass into solution. Finally, if the electrolysis be sufficiently prolonged, all the lead will be deposited as dioxide on the anode." The action of selenium is similar. According to Vortmann 7 the results are high in the presence of sulphuric, selenic, and chromic acids. Hence, he recommends reprecipitation by electro- lysis. In the presence of arsenic and phosphorus the results will be low. In that case Vortmann recommends depositing the metal as lead on the cathode, dissolving the metal in nitric acid, and reprecipitating the lead as dioxide. 1 F. Henz, Zeit. anorg. Chem., 37. 2, 1903. 2 H. Nissensen and B. Neumann, Chem. Ztg., 19. 1142, 1895. 3 C. Marie, Compt. Rend., 130. 1032, 1900 ; Chem. Neivs, 82. 51, 1900. 4 For lead chromate, Marie proceeds in a similar manner, but less ammonium nitrate is needed 2 grms. of ammonium nitrate suffice for 0*5 grm. of lead chromate. 5 When chlorides are present, evaporation with sulphuric acid and treatment of the sulphate as described in the text will furnish a solution ready for electrolysis. 6 B. Neumann, Chem. Ztg., 20. 382, 1896. 7 G. Vortmann, LieUg's Ann., 351. 283, 1907. For the effect of gums, and colloids generally, see H. Freundlich and J. Fischer, Zeit. Elektrochem., 18. 885, 1912. THE DETERMINATION OF LEAD. 337 162. The Rapid Deposition of Lead Dioxide by a Rotating Electrode Exner's Process. The enormous gain in the speed of deposition which attends the use of a rotat- ing electrode removes one of the most serious objections to electro-analysis. In the case of copper, for example, a rotating anode enables a satisfactory determination to be made in 5 minutes, using a current of 10 volts and 13 amps. ; nickel, in 15 minutes; cobalt, in half an hour; cadmium, in 5 minutes; etc. The following diagram (fig. 132), after Fischer, emphasises very forcibly the great gain in time obtained by the use of rotating electrodes. The ordinates show the amounts of copper precipitated by rotating and stationary electrodes after the elapse of different intervals of time (abscissae). Knobukow 1 introduced rotating electrodes in electro-analysis in 1886, but the subject did not attract much attention until the American chemists Gooch and Smith took up the subject about 1903. These chemists and their co- workers extended Knobukow's idea, and showed that a still further gain in time 100 200 300 400 FIG. 182. Time of deposition, stationary and rotating electrodes. 500 attends the use of currents of greater density and voltage. For further details see the works cited on page 256. The Apparatus. The determination of lead as dioxide by the use of rotating electrode may now be described, on the assumption that pages 253 to 262 have been mastered. Fit up the apparatus illustrated in fig. 133 the lettering is similar to that in figs. 117 and 118, from which it differs in that the platinum dish is here used as the anode, not the cathode. The current accordingly passes through the electrolytic cell in the reverse direction to what was the case with copper. The disc electrode is replaced by a spiral of heavy platinum wire with the spirals fixed in position by twisted platinum binding wires. The centre of the spiral is depressed to give it the form of a shallow bowl about 5 cm. in diameter. This electrode is clamped to the axis of a wheel which can be rotated by a motor or turbine between 450 and 600 revolutions per minute, and yet be in electrical contact with the battery. 2 The Electrolysis. The solution, prepared as indicated in the preceding section 1 N. von Knobukow, Journ. prakt. Chem. (2), 33- 473, 1886; F. A. Gooch and H E. Medway, Amer. J. Science (4), 15. 320, 1903 ; F. F. Exner, Journ. Amer. Chem. 8oe. % 2$. yb, 1903 ; R. 0. Smith, ib., 27. 1287, 1905. , . 2 Numerous types of cells for rotating electrodes have been devised- some are described m the text-books indicated on page 256. That described in the text may be the simplest but not necessarily the best when many determinations have to be made. 33* A TREATISE ON CHEMICAL ANALYSIS. for stationary electrodes, is placed in a platinum basin, matted on its inner surface Add 20 c.c. of concentrated nitric acid and dilute the solution to make a total volume of 110-125 c.c. Heat the solution to about 70, and start the electrode rotating at about 500 revolutions per minute. 1 A current of elec- tricity at 4-5 volts, and current density 10-11 amps., is passed through the solution. The lead will soon be deposited as a black adherent velvety him of dioxide on the dish. The maximum period required for 0'25 grm. of the metal was found by Smith to be 15 minutes; and for 0'5 grm., 25 minutes. 1 Smith FIG. 133. Electrolysis by rotating electrode. found that the rate of precipitation from a solution containing the equivalent of 0'5787 grm. of Pb0 2 was as follows : Time PbO, 5 0-4940 10 0-5708 15 0-5747 20 25 30 0-5770 0-5787 0-5789 grm. In about 25 minutes, when the decomposition is finished, stop the rotator, and reduce the current by the introduction of resistance. Add water to cover the lead peroxide ; syphon off water from the dish while the dish is kept full of water. Wash the deposit with alcohol and ether in the usual manner. To illustrate the results which might be expected : Pb0 2 found. . , . 0-0566 0-1137 0'2887 0'5781 Q'5788 grm. Pb (by factor 0-8662) . . 0'0490 0'0985 0'2501 0-5008 0'5015 grm. Pb(used) .... 0-0491 0'0982 0-2496 0'4992 0'4996 grm. The calculation of the amount of lead corresponding with the deposited lead dioxide here presents the same difficulty as was encountered in dealing with stationary electrodes. The factor 0-8662 furnishes too high results. 1 There is no need to heat the solution during the electrolysis, because the high current used keeps the liquid hot. With higher speeds the electrolyte may sweep round the edge of the dish and be thrown against the cover glass. This will do no particular harm if the amount of liquid in the basin does not exceed 125 c.c. 2 F. F. Exner, Journ. Amer. Chem. Soc., 2$. 896, 1903 ; R. 0. Smith, ib., 27. 1287, 1908. THE DETERMINATION OF LEAD. 339 163. The Colorimetric Determination of Lead. Pelouze l first proposed to estimate small amounts of lead in a given solution from the intensity of the brown coloration * produced when the lead is converted into sulphide. 3 The method has been frequently used for estimating the small amounts of lead in water, citric acid, etc. 4 The method has also been recom- mended for the determination of the amount of lead in factory dusts, glazes, etc. 5 The results are usually too low in presence of traces of free acid acetic or hydro- chloric acid since some lead is not then converted into the sulphide. This will be obvious from our study of the action of hydrogen sulphide on lead salts in acid solutions (page 273). The results are very much better in alkaline solutions using sodium or ammonium sulphide as precipitant. Influence of Iron Salts. Dark brown iron sulphide is precipitated in alkaline solutions, and since traces of iron are nearly always associated with lead, it is necessary to eliminate the disturbing effects produced by this agent. Warington used the colorimetric process for estimating small quantities of lead in citric and tartaric acids, and, in order to avoid the effects produced by iron, he re- commends making the solution " alkaline with ammonia, treating the solution with a few drops of potassium cyanide, and heating it to near the boiling point." The iron is thus converted into a complex cyanide probably ferro- cyanide and it is not then affected by the subsequent addition of alkaline sulphide. The method works quite satisfactorily provided the iron is all present in the ferrous condition, and the liquid is strongly alkaline. Ferric salts are not converted by the treatment with potassium cyanide into colourless substances unaffected by alkaline sulphides. It is therefore necessary to reduce ferric salts to the ferrous condition before applying the test. Wilkie recommends sodium thiosulphate in acid solution as a reducing agent. Its action is probably re- presented by the equation : 2Na 2 S 2 3 + 2FeCl 3 = 2FeCl 2 + Na 2 S 4 6 + 2NaCl. The Determination. Wilkie made an artificial mixture of 12 grams of citric acid, 0-004 grm. of ferric iron, 0-00005 grm. of lead. This mixture was placed in a " Jena flask and made up to about 35 c.c. with water. 2 c.c. of y^N-sodium thiosulphate 6 1 T. J. Pelouze, Ann. Chim. Phys. (3), 79. 108, 1841. 2 It is possible to detect 1 part of lead per 16,000 parts of solution, according to F. Jackson (Journ.Amer. Chem. Soc., 25. 992, 1903); 1: 26,900, according to T.G.Wormley;andl: 1,000,000, according to A. B. Prescottand E. C. Sullivan, and R. Warington (Jou rn. Soc. Chem. Ind., 12. 97, 1893); C. H. Pfaft', Handbuch der analytischen Chemie, Altona, 1824 ; J. L. Lassaigne, Jo-urn. Chim. Mid., 8. 581, 1832 ; P. Halting, Journ. prakt. Chem. (1), 22. 45, 1841 * :i A. Trillat (Compt. Rend., 136. 1205, 1903) showed that lead peroxide produces a blue coloration in contact with tetramethyl-diamidodiphenylmethane in a solution acidified with acetic acid. By converting the lead into peroxide, and estimating the amount of the latter from the intensity of the blue coloration, the results are far from satisfactory. Similar attempts to convert the lead into peroxide and subsequently bringing the peroxide in contact with a solution of potassium iodide and estimating the amount of lead from the intensity of the colour of the liberated iodine in the presence and in the absence of starch likewise failed to give constant results without an abnormal expenditure of time in isolating the lead peroxide. 4 J. M. Wilkie, Journ. Soc. Chem. Ind., 28. 636, 1909 ; R. Warington, ib., 12. 97, 1893 ; C. A. Hill, Chemist Druggist, 66. 388, 1905 ; H. W. Woudstra, Zeit. anorg. Chem., 58. 168, 1908 ; F. L. Teed, Analyst, 17. 142, 1892 ; E. R. Budden and H. Hardy, ib., 19. 169, 1894 ; B. Kiihn, Arbeit. Kais. Gesand., 23. 389, 1906 ; A. G. V. Harcourt, Journ. Chem. Soc., 97. 841, 1910 ; M. Lucas, Bull. Soc. Chim. (3), 15. 39, 1896; A. Liebrich, Chem. Ztg., 22. 225, 1898; G. Bischof, Zeit. anal. Chem., 18. 73, 1879; V. Antony and T. Benelli, Gazz. Chim. ItaL, i. 218, 1871; 2. 194, 1872; P. Carles, Journ. Pharm. Chim. (6), 12. 517, 1900; L. Libermann, Pharm. Centralhalle, 29. 10, 1889. L. W. Winkler, Zeit. angew. Chem., 26. 38, 1912, recommends addition of ammonium chloride to get more reliable results. 5 H. R. Rogers, Report Departmental Committee appointed to inquire into the Dangers incident on the Use of Lead, 2. 118, 1910 ; A. G. V. Harcourt, ib., 2. 120, 1910. Rogers's method cannot be regarded seriously as a quantitative process. 6 It is very necessary to test all reagents to ensure their freedom from traces of lead ammonia, sodium hydroxide, etc. 340 A TREATISE ON CHEMICAL ANALYSIS. were then added, and the whole heated to incipient boiling, and the flame removed. After about 5 minutes, the solution became perfectly water-white, and to it was immediately added 1 c.c. of a 10 per cent, solution of potassium cyanide, 1 and then 2 an excess (13 c.c.) of 0'880 ammonia, and the whole gently boiled until colourless." Assuming that the solution under investigation is treated as the mixture just described, a similar solution is made with a known amount of lead. This solution is systematically diluted with water in Nessler's glasses, and one drop of ammonium sulphide 3 added to each. The colours so obtained are compared with that of the solution under investigation in the usual way see Colorimetry. 4 Disturbing Factors. Woudstra found that with the colorimetric process : Lead present. . . . O'OOl O'OOl 0'0009 0'0009 grm. Estimated lead in error . . - 18 - 8 + 1 1 '1 - 3 '3 per cent. It must be borne in mind that the lead sulphide precipitated in alkaline solution is colloidal, 5 and that the tint is to some extent determined by the size of the colloidal particles, which, in turn, is dependent upon the nature of the salts in the solution under investigation, and on slight variations in the way the solution has been prepared. Vigorous agitation, for instance, may coagulate the suspended colloid and cause a precipitation of the lead sulphide. Salts of the alkalies and alkaline earths rapidly coagulate the colloidal sulphide. According to Kiihn, barium chloride is 100 times more active than sodium nitrate in coagulating the suspended colloid. Sometimes the colloidal sulphide imparts a smoky opalescence to the solution. This is not favourable for satisfactory comparisons. Warington recommended the addition of (approximately) an equal volume of glycerol to the solution under investigation ; and Harcourt recommended mixing the solution with about one- fifth its volume of a clear solution of cane sugar (half sugar, half water), in order to eliminate this difficulty. Each of these additions makes the tint of the lead sulphide slightly paler, and less opaque, and thus facilitates the work of comparison. The effect of iron has been already discussed. In the case of lead glazes, cobalt oxide may have been added to the glaze, or dissolved from the body by the glaze. Cobalt salts in alkaline sulphide solutions give a dark brown coloration (page 386), which would make the reported amount of lead too high. 6 To summarise : In order to get reliable results it is necessary that the solutions under comparison have the same general character otherwise the result may be over 50 per cent, in error. With unknown solutions, this can only be assured by elaborate preparations which occupy so much time that the main advantage of the colorimetric process rapidity of execution is nullified. 1 It is best to use an excess of potassium cyanide, say 25 of potassium cyanide to 1 of iron. 2 Wilkie considers that the results are better if the potassium cyanide is added to the acid solution before the ammonia, since the formation of the ferrocyanide proceeds faster in the acid solution. 3 According to Warington, ammonium sulphide is a more delicate reagent than hydrogen sulphide. 4 A. G. V. Harcourt (Journ. Chem. Soc., 97. 841, 1900) makes a series of permanent colour standards matching the tints of solutions containing variable amounts of lead. The colour standards are made from mixtures of ferric, cobalt, and copper sulphates in suitable proportions. These are labelled and preserved in hermetically sealed glass cylinders. J. W. Lovibond (Measurement of Light and Colour Sensations, London, 124, 1910) made combinations of his standard glass slips to match the tints of solutions containing known amounts of lead. 5 W. Spring, Bull. Acad. Roy. Belg., 483, 1909. 6 G. D. Elsdon (Pharm. Journ. (4), 89. 143, 176, 1912) points out that in filtering very dilute solutions of lead salts part of the lead may be retained by the filter paper. If lead solution be acidified with 0*6 per cent, acetic acid the absorption does not occur, and the lead may be washed from the paper by '6 per cent, acetic acid. CHAPTER XXV. THE DETERMINATION OF BISMUTH AND MERCURY. 164. The Separation of Mercury from Lead, Bismuth, Copper, and Cadmium Rath's Process. As indicated on page 279, mercuric sulphide is almost insoluble in ammonium sulphide, but readily soluble in sodium or potassium sulphides. If, therefore, the sulphides precipitated by hydrogen sulphide be digested with an alkaline sulphide, the mercury will be found with the arsenic and antimony group. Indeed, mercury, tin, arsenic, and antimony sulphides can be sharply and con- veniently separated from lead, silver, bismuth, and copper sulphides by digesting the mixed sulphides in a mixture of potassium sulphide and hydroxide. 1 If cadmium (and zinc) should be present, the mercury sulphide is but imperfectly dissolved. 2 The net result of this operation is a solution containing tin, mercury, arsenic, and antimony sulphides, and a precipitate containing copper, lead, and bismuth sulphides assuming that cadmium is absent. If tin also be absent, this method of analysis offers some advantages, because the mercury can be readily separated from the arsenic and antimony by treating the solution with ammonium chloride. If tin be present, much tin will be precipitated with the mercury, and the separation by this process is inconvenient. If the sulphides precipitated by hydrogen sulphide be digested in ammonium sulphide, 3 a good separation can be made in the absence of tin and copper. Mercury, bismuth, lead, copper, and cadmium will remain with the insoluble residue, while arsenic and antimony, with some copper sulphide, will be dissolved. If tin be present, a somewhat soluble compound of tin and mercury sulphide appears to be formed, because part of the mercury sulphide will pass into solution, and part of the tin will remain with the precipitate. 4 The plan of analysis must, therefore, be modified to suit these different conditions. Fortunately, mercury is comparatively rare in silicate analysis, and one of the two distillation processes indicated later, or simply " loss on ignition " of the original sample, will suffice. Rath's process 5 is one of the most convenient for separating mercury from the sulphides insoluble in ammonium monosulphide. 1 Prepared by saturating half the prepared volume of a 15 per cent, solution of potassium hydroxide with hydrogen sulphide. Mix the two parts, and filter after the solution has stood for some days (cf. page 277). 2 K. Billow, Zeit. anal. Chem., 31. 697, 1892 ; Chem. News, 67. 174, 1893. 3 AMMONIUM MONOSULPHIDE. Saturate three volumes of aqueous ammonia (sp. gr. 0'88) with hydrogen sulphide, and add two volumes of fresh ammonia (sp. gr. 0'88) to the saturated solution. (Ammonium polysulphide is made by dissolving 25 grms. of " flowers of sulphur" in a mixture of 500 c.c. of ammonium monosulphide and 500 c.c. of water.) When mercury is present, the freshly prepared colourless ammonium sulphide is used in place of sodium mono- sulphide, for the reasons stated on page 279 E. Donath, Chem. Ztg. Rep., 15. 68, 1895. 4 T. Wilm, Ber., 20. 232, 1887. 3 G. von Rath, Pogg. Ann., 96. 322, 1855. 342 A TREATISE ON CHEMICAL ANALYSIS. It is based on the fact that the mercuric sulphide is insoluble in boiling dilute nitric acid (sp. gr. 1'2-1'S), 1 while the remaining sulphides silver, 2 bismuth, copper, cadmium, and lead pass into solution. A little insoluble lead sulphate may be formed by the action of the nitric acid on the sulphide. The mercuric sulphide can then be separated from the lead sulphate by digesting the mixture with a little aqua regia ; dilute the solution with water, filter off the precipitated sulphur and lead sulphate, and wash with water. A trace of lead sulphate may be dissolved by the filtrate. This is recovered later. The mercury is separated as sulphide by Volhard's process. 165. The Gravimetric Determination of Mercury as Sulphide Volhard's Process. In this process 3 the mercury is precipitated as sulphide by the addition of ammonium sulphide to a nearly neutral solution ; the mercuric sulphide is dissolved in caustic alkali, from which it is reprecipitated as mercuric sulphide, by the addition of ammonium nitrate. Precipitation of the Mercuric Sulphide. Almost neutralise the acid filtrate from the preceding operation with sodium carbonate ; add a slight excess of colourless ammonium sulphide (freshly made, page 341); and, finally, with constant agitation, a solution of pure sodium hydroxide. When the dark colour begins to lighten, heat the solution to boiling, and add more sodium hydroxide until the liquid is perfectly clear. 4 The mercury dissolves forming a thio-salt Hg(SNa) 2 . Add an excess of ammonium nitrate, and boil the solution as long as ammonia is given off. The thio-salt is decomposed by the ammonium nitrate : Hg(SNa) 2 '+ 2NH 4 N0 3 = 2NaN0 3 + (NH 4 ) 2 S + HgS. Wash the precipitated mercuric sulphide 5 two or three times with hydrogen sulphide water, then with hot water by decantation through a Gooch's crucible until the water no longer reacts with silver nitrate. Transfer the precipitate to the crucible, dry between 110 and 112 , 6 and weigh as mercuric sulphide HgS. Every gram of the mercuric sulphide represents 0'8618 gram of mercury, or 0'9305 gram mercuric oxide HgO. The results are generally a little high owing to the presence of some sulphur with the precipitate. 7 Removal of Sulphur from Metallic Sulphides. The sulphur may be removed either by boiling the precipitate with a little sodium sulphite before filtering so as to convert the sulphur into soluble sodium thiosulphate (S + Na 2 S0 3 = Na 2 S 2 3 ), 1 J. Torrey, Amer. Chem. Journ., 7. 355, 1886 ; J. L. Howe, ib., 8. 75, 1886. Mercuric sulphide is not appreciably attacked by boiling dilute nitric acid, but a single drop of dilute hydrochloric acid (1 : 3) will convert black mercuric sulphide into a yellow compound, and if a few drops of hydrochloric acid be present, some of the mercury will pass into solution. Hence, the nitric acid must be perfectly free from chlorides and hydrochloric acid. 2 If silver be present, it will be precipitated as chloride by the hydrochloric acid. 3 J. Volhard, Liebig's Ann., 255. 255, 1889. 4 A precipitate of insoluble lead sulphate maybe present owing to the dissolution of some lead sulphate in the aqua regia. If present, it must be filtered off and the precipitate washed with dilute sodium hydroxide. The lead sulphate is mixed with that previously obtained, and treated by the method of page 317. 5 The precipitate so obtained is much more compact and easily filtered than mercuric sulphide precipitated by hydrogen sulphide. 6 The time of drying is materially lessened by a final washing with alcohol R. S. M'Bride, Journ. Phys. Chem., 14. 189, 1910. 7 F. P. Treadwell, Kurzes Lehrbuch der analytischen Chemie, Leipzig, 2. 140, 1911 ; L. Vignon, Cum.pt. Rend., 116. 584, 1893. THE DETERMINATION OF BISMUTH AND MERCURY. 343 or by extraction with carbon disulphide. 1 The following are convenient methods of conducting the process of extraction : The Gooch's crucible containing the mixture is placed on a glass tripod whose feet rest on the bottom of a beaker containing some carbon disulphide. The beaker is covered with a round- bottomed flask containing cold water. The beaker, etc., is placed on a hot-water bath. 2 The carbon disulphide boils at 46, and, after condensing against the bottom of the flask, drops into the crucible and passes back to the bottom of the beaker. Wiley's or Drehschmidt's extraction apparatus 3 is convenient for wash- ing precipitates free from sulphur by extraction with carbon disulphide. Both have a receptacle, A, for the Gooch's crucible below a condenser, C ; Drehschmidt's FIG. 134. Extraction apparatus. is shown in fig. 134. A current of cold water is passed through the condenser, and the flask, , containing the solvent, etc., is placed in a hot- water bath, or supported over an incandescent electric lamp. The carbon disulphide boils, and the condensed liquid runs through the crucible back to the bottom of the flask. In about half an hour the apparatus is allowed to cool, and the carbon disulphide is washed from the precipitate in the crucible by one treatment with alcohol, and 1 C. Friedheim and P. Michaelis, Zeit. anal. Chem., 34. 526, 1895 ; H. W. Wiley, ?&., 23. 586, 1884 ; G. Vortmann, Uebungsbeispie/e aus der quantitativen chemischen Analyse, Leipzig, oq 1 Q1 A ' 2 Note, no naked flames must be near enough to risk ignition of the carbon disulphide. 3 H. W. Wiley, Journ. Anal. App. Chem., 7. 65, 1893; W. D. Richardson and E. F. Scherubel, Journ. Ind. Eng. Chem., 4. 220, 1912 ; H. J. C. Curr, ib., 4. 535, 1912. 344 A TREATISE ON CHEMICAL ANALYSIS. one treatment with ether. Dry the sulphide at 110, and weigh as indicated above. The results are very fair, generally less than O'l per cent. low. 166. The Distillation Process for Mercury. Erdmann and Marchand's Process. All mercury compounds, 1 with the exception of the iodide, are quantitatively decomposed when heated with quicklime. The reaction, in the case of the chloride, is represented : HgCl 2 + CaO = CaCl 2 + OHg. Hence, it is often most convenient to determine the mercury on a separate sample by a distillation process, and ignore the mercury when the other con- stituents are determined, because the mercury compound which might contaminate a precipitate is volatilised when the precipitate is ignited before weighing. FIG. 135. Erdrnann and Marchand's distillation process for mercury. Mercury can be determined in colours containing, say, mercuric chromate, and in gold amalgams, "best" gold, etc., by placing a plug of asbestos at the end of a combustion tube 45 to 50 cm. long, and 1 '5 cm. wide then an 8-cm. layer of freshly burned calcium oxide ; a 10-cm. layer of an intimate mixture of a weighed quantity of the given substance with an excess of calcium oxide ; a 20-cm. layer of calcium oxide ; and, finally, a loose plug of asbestos fibre. 2 Bend the tube as shown in the diagram, (7, fig. 1 35. Fit the bent end of the combustion tube, by means of a piece of rubber tubing, with one arm of a small Peligot's tube, J3, as shown in the diagram. The other arm of the Peligot's tube, D, is loosely packed with pure gold leaf. The other end of the combustion tube is connected with a wash- bottle, A, containing sulphuric acid, and with a tube delivering coal gas. The coal gas is allowed to bubble through the apparatus at the rate of about three bubbles per second for about half an hour. Gradually raise the temperature of the combustion tube, starting with the 20-cm. layer, and gradually carry the 1 A simple determination of the "loss on ignition" is sometimes sufficient for the mercury in a dry sample of mercury chromate. The last trace of mercury is difficult to expel from gold M. J. Personne, Compt. Rend., 56. 63, 1862. 2 The tube is often closed at one end, and the closed end is packed with magnesite or sodium bicarbonate, which, on heating towards the end of the operation, evolves a stream of carbon dioxide ; this drives out the mercury vapour. The mercury is collected in water. The results are satisfactory with rich (7-8 per cent, mercury) ores. THE DETERMINATION OP BISMUTH AND MERCURY. 345 flame backwards until finally the whole tube is being heated at the same time. The part C of the combustion tube can be heated with a naked Bunsen's flame, and any mercury here condensed is thus driven forward into the Peligot's tube, but the rubber connection must not be scorched. The current of coal gas is continued all the time the combustion is in progress, and while the apparatus is cooling. Most of the mercury collects in the lower bulb of the Peligot's tube ; a small part is arrested by amalgamation with the gold leaf ; l and some of the mercury may condense in the narrow portion of the combustion tube near C. The Peligot's tube is disconnected 2 and a current of air, dried by passing through a drying tower, fig. 164, is passed through the apparatus for about half an hour. Weigh the Peligot's tube ; the increase in weight represents mercury derived from the given sample. Now cut the combustion tube where the mercury is condensed. Weigh. Heat this portion of the com- bustion tube while a current of air is passed through. This volatilises the mercury. Cool the tube in a desiccator and weigh again. The apparent loss in weight represents the mercury which was condensed in the tube. Add this to the preceding result to get the total mercury. 3 Holloway's Modification of EschkcCs Process. This method is based on the fact that when mercuric sulphide is heated with iron filings, iron sulphide and volatile mercury are formed. 4 The mercury is condensed on a gold or silver plate, and an amalgam is formed. The in- crease in the weight of the plate from this cause represents the amount of mercury in the sample. The process is used for sulphides and amalgams. The process can also be used with "best gold" and other amalgams. 5 The Apparatus. The upper edge of a deep, glazed porcelain crucible about 4 cm. diameter, and 4'5 cm. FIG. 136. Holloway's apparatus, high is ground flat. The crucible, A, figs. 136 and 137, is supported in a hole in an asbestos or quartz plate, H. A 5-cm. disc of silver 6 plate (weighing about G'4-0'5 grm. per sq. cm.) is annealed by holding it in the Bunsen's flame for a minute or two. The disc is rubbed between two flat surfaces until it lies perfectly flat on top of the crucible Weigh this disc, B. Place a rather larger metal disc, (7, on top of the weighed disc, B, in order to keep the latter clean. A metal condenser, D, through which a current of cold water is flowing, is placed on the discs to keep them cool. 1 Instead of using gold leaf, etc., water is sometimes used in the bottom of the Peligot's tube ; the combustion tube has one end closed, and a layer of magnesite is placed at the closed end. 2 Watch that no mercury falls from the combustion tube after the Peligot's tube has been disconnected. 8 . H. Rose, Fogg. Ann., HO. 542, 1860 ; 0. L. Erdmann and R. F. Marchand, Journ. prakt. Chem. (1), 31. 385, 1844 ; C. R. Konig, ib. (1), 70. 64, 1857 ; A. C. Gumming and J. Macleod, Jour. Chem. Soc., 103. 513, 1913. 4 As in Jordan's test for mercury W. J. Jordan, Schweigger's Journ. , 57. 339, 1829. 5 A. Eschka, Chem. News, 26. 22, 1872 ; Zeit. anal. Chem., II. 344, 1872 ; Dingier 1 s Journ. , 204. 47, 1872 ; C. T. Holloway, Analyst, 31. 66, 1906 ; Chem. Eng., 4. 169, 1906 ; R. E. Chism, Eng. Min. Journ., 66. 480, 1898. 6 Gold has a greater "collecting" power for the mercury than silver; but silver has the greater "collecting" power weight for weight, and is also a better heat conductor. Hence Holloway recommended silver, Eschka and Chism used gold. 146 A TREATISE ON CHEMICAL ANALYSIS. is weighed 3 The condenser is held in place by a 500-grm. weight, TF, on top of the condenser. A gutter, G, runs round the bottom of the condenser to collect any condensed moisture trickling down the condenser. 1 The moisture is removed from time to time by means of a piece of blotting paper. Chare/ing the Crucible. The finely powdered (120's lawn) and dry 2 sample tidied 3 into the crucible and mixed with 10 grms. of fine iron filings. 4 The mixture is covered with 5 grms. of the coarse iron filings. Everything is placed in position, as shown in the diagrams, figs. 136 and 137. Volatilisation of the Mercury. The crucible is heated with a small flame sufficient to raise the bottom of the crucible to redness without the flame coming in contact with the sides of the crucible. 5 After heating from 20 to 30 minutes, let the system cool for 15 minutes with the condenser at work. Wash the gold or silver disc with alcohol; 6 dry in a desiccator; and weigh. The apparent increase in weight of the disc represents the mer- cury. It requires a little practice to adjust the size of the flame and the time of heating to be sure all the mer- cury is volatilised under the conditions just indicated. 7 To make quite sure, reheat the crucible with a fresh gold or silver disc in position. If all the mercury was volatilised during the first heating, there will be no increase in weight of the disc. If any increase in weight is obtained, add the result to the mercury obtained with the first plate. The plates can be freed from mercury, and prepared for another determination, by simply heating them to drive off the mercury. 8 A determination occupies 40 to 50 minutes. FIG. 137. Holloway's apparatus. 1 The whole outfit can be bought for about 5s. 2 If metallic mercury be present, the drying must be conducted with care on account of the tendency of mercury to volatilise. 3 If the amount of mercury in the sample is less than 1 per cent., take 2 grms. of the sample ; if between 1 and 2 per cent., take 1 grm. ; if between 2 and 5 per cent., take 0'5 grm. ; and if over 5 per cent, is present, grind the sample with 10 grms. of dry sand (120's lawn), and take an aliquot portion so as to keep approximately within the indicated limits. 4 Clean iron filings free from oils and fats are prepared by heating the filings to redness for an hour in a covered crucible. Sift the filings through an 80's lawn, and also sift some through a 30's lawn. Keep each in a separate bottle. 5 The tip of a small Meker's flame is very suitable. The plate does not absorb mercury well if it is hot. 6 To remove tarry and organic matters, etc., which might have collected on the gold or silver plate. The gold amalgams used in pottery, if already mixed with "fat oil," should be washed with ether to remove the oil before treatment by this process. 7 If the plate does not show a clear circular stain, either too much ore has been used, or the lid did not fit properly. If the stain extends beyond the edge of the crucible, there will be an element of uncertainty owing to the probable loss of mercury. The plate and crucible should fit close enough to prevent this. 8 The plate seems to improve with use, since it gets more porous and absorbs mercury better. THE DETERMINATION OF BISMUTH AND MERCURY. 347 Special Precautions. (1) Avoid the excessive heating of the crucible; the edge of the foil should never feel warm to the fingers. (2) The silver or gold foil should be thoroughly cleaned and fit on the crucible quite flat. (3) The foil and cooler should be carefully adjusted to ensure contact all round the top of the crucible. (4) The foil should be quite dry. (5) Draughts are objectionable. (6) Allow 15 minutes for cooling before removing the foil to prevent danger of losing uncondensed mercury vapour. Duplicate determinations on a 1 per cent, ore should agree to within O05 per cent. (Holloway), a result which cannot be equalled by the wet method. 167. The Separation of Bismuth from Lead, Cadmium, and Copper Lowe's Process. Lowe J has based a process on the fact that water converts bismuth nitrate into an insoluble basic salt under conditions where lead, copper, and cadmium salts undergo no such transformation : Bi(N0 3 ) 3 + 2H 2 0^=^Bi(OH) 2 N0 3 + 2HN0 3 . The basic nitrate is washed with a solution of ammonium nitrate ; this allows the precipitate to be washed without decomposition. If water alone be used, the precipitate becomes more and more basic, the filtrate consequently acquires an acid reaction, and some bismuth passes into solution. The process can, how- ever, be used for the separation of bismuth from the elements just named. 2 Precipitation. The nitric acid solution of the sulphide is evaporated on a water bath to a syrupy consistency, and mixed with hot water and thoroughly stirred with a glass rod. Take care to loosen any crusts which may have formed on the sides of the basin. The solution is again evaporated, and the addition of water and the evaporation are repeated until further addition of water pro- duces no turbidity three or four evaporations usually suffice. This shows that the reaction indicated above is complete. Evaporate the solution to dryness, and when the dry mass has ceased to smell of nitric acid, cool. Washing. Add a cold solution of ammonium nitrate (2E), and after standing some time with frequent agitation to make sure that all the lead nitrate has passed into solution, filter. Wash the precipitate with the solution of ammonium nitrate, and dry in an air bath. Ignition. Remove the dry precipitate from the filter paper, and preserve it in a watch-glass. Ignite the filter paper in a porcelain crucible at a low temperature. Moisten the ash with nitric acid. Evaporate to dryness very cautiously to prevent spurting. Transfer the precipitate to the crucible and ignite filter-paper ash and precipitate together. Try to keep the temperature below the fusing point of the oxide, since, if the temperature be too high, the oxide melts and attacks the glaze. 3 When the weight has become constant, weigh the precipitate as bismuth oxide Bi 2 3 . Purification of the Bismuth Oxide. The precipitate may be contaminated with a little iron, mercury, and copper if these elements be present. Hence, some prefer to redissolve the precipitate before ignition and repeat the separation. The combined filtrates are evaporated to dryness and calcined at a low tempera- 1 J. Lowe, Journ. prakt. Chem. (1), 74. 344, 1858 ; C. H. Pfaff, Handbuch der analytischen Chemie, Altona, 1821. 2 This process can be used for evaluating bismuth nitrate and bismuth oxide. The latter is soluble in nitric acid. 3 The error from this cause can be neglected in most cases. If reducing gases be present inside the crucible, some of the oxide will be partially reduced. 348 A TREATISE ON CHEMICAL ANALYSIS. ture to destroy the ammonium salts, which interfere with the subsequent pre- cipitation of the lead as sulphate (page 317). If sulphuric or hydrochloric acid be present, a basic sulphate or chloride may be formed, which is not converted to oxide on ignition. The results will ac- cordingly be high. In that case, Rose 1 recommends fusing the mass for about 1 5 minutes with four or five times its weight of, say, 98 per cent, potassium cyanide in a covered crucible. When all is fused the crucible is gently tapped to collect the little beads of metal into one button. The crucible should not be heated above low redness. Wash the cold mass with water to remove the cyanides and cyanates. The button is then washed with alcohol, dried at 100, and weighed as metallic bismuth. This weight multiplied by 1'1154 gives the corresponding amount of bismuth oxide Bi 2 3 . 2 168. The Separation of Bismuth from Copper and Cadmium Jannasch's Process. The bismuth can be separated 3 from copper and cadmium, if cadmium be present, by evaporating a solution containing salts of these elements to dryness. Dissolve the residue in, say, 5 c.c. of nitric acid (sp. gr. 1-4) and 25 c.c. of water, and pour the solution into a beaker containing 25 c.c. of concentrated ammonia, and 50 c.c. of a 4 per cent, solution of hydrogen peroxide, with constant stirring. A dull yellow precipitate of basic bismuth hydroxide separates. 4 Let the pre- cipitate settle ; decant the clear ; add more of the ammoniacal hydrogen peroxide ; decant ; transfer to a filter paper ; wash with hot dilute ammonia (1 : 8), and finally with hot water, until a drop of the wash-water gives no precipitate when evaporated on a piece of platinum foil. To remove any copper or cadmium which might be precipitated with the bismuth, dissolve the precipitate on the filter paper in hot dilute nitric acid ; evaporate to dryness, and repeat the precipitation as described above. After complete washing, dry the precipitate at 90-95 ; ignite it in a porcelain crucible until its weight is constant ; and weigh as Bi 2 3 . The ignition of the precipitate, etc., is described in detail, page 347. 5 169. The Determination of Bismuth Colorimetrically. Small amounts of bismuth are conveniently determined Colorimetrically. Bismuth iodide 6 forms an intense yellow, orange, or red coloration, 7 which, unlike the somewhat similar colour by iodine, is not destroyed by sulphur dioxide. The process is not often used, and in consequence it has not been 1 H. Rose, Pogg. Ann., 91. 104, 1854 ; no. 136, 426, 1860. 2 Sometimes the bismuth forms a kind of metallic lustrous film inside the crucible. 3 P. Jannasch, Zeit. anorg. Chem., 8. 302, 1895 ; Leitfaden der Gewichtsanalyse, Leipzig, 107, 110, 1904; P. Jannasch and E. von Cloedt, Zeit. anorg. Chem., 10. 398, 1895 ; Chem. Neivs, 72. 64, 1895. 4 It' lead were present, it too would be precipitated with the bismuth as a peroxide. 5 A. L. BenkertandE. F. Smith's process (Journ. Amer. Chem. Soc., 18. 1055, 1896 ; A. F. V. Little and E. Cahen, Analyst, 35. 301, 1910), by precipitation as bismuth formate, is an excellent process for the separation of bismuth from lead, cadmium, etc. 6 F. Field, Chem. News, 36. 260, 1877 ; F. A. Abel and F. Field, Journ. Chem. Soc., 14. 290, 1862; M. Planes, Chem. News, 89. 10, 1904; Journ. Pharm. Chim. (6), 385, 1903; L. L. de Koninck, Bull. Soc. Chim. Belg., 19. 91, 1905 ; F. B. Stone, Journ. Soc. Chem. Ind., 6. 416, 1887 ; T. C. Cloud, ib., 23. 523, 1904. 7 One part of bismuth per 10,000 parts of water gives an orange colour ; 1 part of bismuth in 40,000 parts of water, a light orange ; and 1 part of bismuth in 100,000, a perceptible yellow coloration T. C. Thresh, Pharm. Journ., 641, 1880. THE DETERMINATION OF BISMUTH AND MERCURY. 349 subjected to that critical examination which would have been the case had it been in common use. The use of comparatively large amounts of glycerol is an objectionable feature in the process. Test Solution. The sample under investigation is dissolved in a flask with just sufficient nitric acid and water. Add 10 c.c. of glycerol and 10 c.c. of potassium iodide J solution, and make the solution up to 50 c.c. with glycerol. Place this solution in one test glass of the colorimeter. Standard Solution. Mix 10 c.c. of a standard solution of bismuth 2 with 10 c.c. of the potassium iodide solution, and make the solution up to 50 c.c. with glycerol and water. Comparison. The solutions may then be compared in the colorimeter. The solutions to be compared would have approximately the same concentration. Hence, it may be necessary to alter the amounts of standard and test solution indicated in the text. The calculations are made in the usual manner (pages 200 and 206). 1 POTASSIUM IODIDE SOLUTION. Dissolve 5 grms. of potassium iodide in 5 c.c. of water, and make the solution up to 100 c.c. with glycerol. 2 STANDARD SOLUTION OF BISMUTH NITRATE. Dissolve 1 grm. of bismuth in 3 c.c. of nitric acid (sp. gr. 1'39) and 2 '8 c.c. of water. Make the solution up to 100,c.c. with glycerol. The object of the glycerol is to keep the bismuth iodide in solution. The glycerol need not be used when dealing with small amounts of bismuth say, 0001 to 0'0075 grm. per c.c. For larger amounts of bismuth use the glycerol. CHAPTER XXVI. THE DETERMINATION OF COPPER AND CADMIUM. 170. Rivot's Thiocyanate Process for Copper. THE very valuable electrolytic process for copper has been described on page 258. This will not do in the presence of cadmium. If cadmium be present, Rivot's process 1 of separation may be used. This depends upon the fact that an alkaline thiocyanate produces a precipitate of cuprous thiocyanate CuSCN in neutral or feebly acid solutions of a copper salt in the presence of a reducing agent. The solution is best slightly acidified with sulphuric or hydrochloric acid. An excess of acid is injurious. The solution should be free from oxidising agents nitric acid, free chlorine oxides, etc. The process enables copper to be quantitatively separated from cadmium, and, indeed, many other metals. 2 The Precipitation. The solution 3 is neutralised with ammonia, if necessary ; acidified with a couple of drops of sulphuric acid ; and treated with an excess of sulphur dioxide, or ammonium bisulphite. 4 Then add, drop by drop, with constant stirring, an aqueous solution of ammonium thiocyanate. The greenish precipitate of mixed cuprous and cupric thiocyanates soon becomes white. Let the mixture gtand overnight. Washing and Drying the Precipitate. Filter the solution, with the bulky precipitate, through a Gooch's crucible, previously dried and weighed ; and wash with cold water 5 until the washings give but a faint red coloration with a solution of ferric chloride. 6 Then wash six times with 20 per cent, alcohol. 1 L. E. Rivot, Compt. Rend., 38. 868, 1854 ; R. G. von Name, Zeit. anorg. Chem., 31. 92, 1902 ; Chem. News, 83. 258, 1901 ; E. Busse, Zeit. anal. Chem., 17. 53, 1878 ; 30. 122, 1891. ; H. Tamm, Chem. News, 24. 91, 1874 ; E. Fleischer, ib., 19. 206, 1869 ; G. Fernekes and A. A. Koch, Journ. Amer. Chem. Soc., 27. 1224, 1905 ; W. Harape, Chem. Ztg., 17. 1691, 1893 ; Journ. Soc. Chem. Ind., 13. 421, 1894 ; B. Blount, Analyst, 19. 92, 1894. 2 Copper maybe also separated from cadmium (Pb, Mg, Mn, Hg, Zn, etc.) by the nitroso-|8- naphthol process, as indicated on page 394 G. von Knorre, Zeit. anal. Chem., 28. 234, 1889 ; and from cadmium, nickel, cobalt, aluminium, chromium, etc., by ammonium nitroso-phenyl- hydroxylamine ammonium " cupferron " as indicated on page 455. 3 Remaining after the separation of the bismuth. * Made by saturating aqueous ammonia with sulphur dioxide. Some use for the precipitation a mixture of 120 grms. of potassium thiocyanate and 120 grms. of sodium hydrogen sulphite dissolved in 2 litres of water (H. Tamm, I.e.}. 5 The precipitate is practically insoluble in cold, but appreciably soluble in hot water. 6 The copper in the thiocyanate may be determined volumetrically with some advantage if the standard solutions are ready made. E. Fleischer (Chem. News, 19. 206, 1869) digests the precipitate in a solution of caustic alkali and washes the red precipitate of cuprous oxide with hot water until the washings give no red coloration with ferric chloride, and determines the copper by Haen's process (page 351). S. W. Parr (Journ. Amer. C/iem. Soc., 22. 685, 1900; 24. 580, 1902; H. A. Guess, ib., 24. 708. 1902; W. E. Garrigues, ib., 19. 940, 1897 ; R. K. Meade, ib., 20. 610, 1898 ; Chem. News, So. 67, 1899 ; J. Volhard, Liebig's Ann., 190. 251, 1877) dissolves the precipitated thiocyanate in 10 c.c. of a solution of potash (10 per cent.), adds 10 c.c. of ammonia (sp. gr. 0'96), and titrates with potassium permanganate until the 3 so THE DETERMINATION OF COPPER AND CADMIUM. 351 Dry the precipitate between 110 and 120; and weigh as CuSCN. The drying must be repeated until the weight is constant. 1 The weight of the precipitate multiplied by 0'6541 gives the corresponding amount of cupric oxide CuO. - Accuracy of the Results. The results are excellent. For example, with known amounts of copper, Fernekes and Koch found : Copper (taken) . . . 0-0939 0'0939 0'0939 0'0188 0'0188 grm. Copper (found) . . . 0'0941 OD939 0'0939 0'0189 0'0188 grm. The main objections to the method are the slight solubility of the precipitate in an excess of the precipitating reagent, and in water ; and the tardy separation of the precipitate. The method can be used to separate copper from zinc, cadmium, iron, cobalt, nickel, bismuth, tin, arsenic, and antimony, because these elements are not precipitated under the conditions of the experiment. If desired, the precipitate can be roasted to drive off the cyanogen compounds ; the residue dissolved in acid, and the copper determined volumetrically ; or the thiocyanate can be dissolved in about 2 c.c. of concentrated nitric acid, boiled for a few minutes, treated with an excess of ammonia and the excess boiled off, treated with 2 to 3 c.c. of acetic acid, and the solution titrated with iodine as described below. Not more than 0*0001 grm. of copper will be found in the nitrate with cadmium if that element be present. 171. De Haen's Volumetric Iodine Process for Copper. E. de Haen's process 2 is based on the fact that when an excess of potassium iodide is added to a concentrated solution of a copper salt, acidulated with acetic acid, cuprous iodide is formed, and an equivalent amount of iodine is liberated. The amount of free iodine is determined by titration with sodium thiosulphate. If the solution be too dilute, the cuprous iodide produced in the reaction just indicated re-formg the original salt. The "back reaction" proceeds as the green colour persists after warming the solution at 45-55 for a short time. Then add an excess of the permanganate solution say one-third or one-fourth of the amount already added. Let the mixture stand five minutes. Acidify the solution with 25 c.c. of dilute sulphuric acid (1 : 2), and titrate with permanganate at 60-70 until the pink coloration appears. One gram of potassium permanganate corresponds with 07193 grams of cupric oxide CuO. 1 Cuprous thiocyanate begins to decompose if heated above 170. A. Glaus (Journ. prakt. Chem. (1), 15. 401, 1838) found 3'0 percent, of water in a sample dried at 115 ; and M. Meitzen- dorfi(Pogg. Ann., 56. 63, 1842) found T54 per cent, when dried at 100. Practically all the water is removed by working as described in the text. 2 E. de Haen, Liebig's Ann., 91. 237, 1854; A. Riimpler, Journ. prakt. Chem. (1), 105. 193, 1868; M. Flajolet, ib. (2), u. 105, 1894 ; D. Vitali, Zeit. anal. Chem., 36. 549, 1897 ; L. Moser, ib., 43. 597, 1904 ; 44. 196, 1904 ; E. V. Videgren, ib., 48. 539, 1909 ; F. M. Litter- scheid, ib., 41. 219, 1902 ; Chem. Ztg., 33. 263, 1909 ; E. Victor, ib., 29, 179, 1905 ; F. M. Litterscheid, ib. t 33. 263, 1909; G. Vortmann and J. von Orlowsky, Zeit. anal. Chem., 20. 416, 1881 ; Monats. Chem., 7. 418, 1886 ; M. Willenz, Rev. Chim. Anal. App., 5. 355, 1896 ; Chem. News, 76. 243, 1897 ; L. Gamier, Journ. Pharm. Chim. (6), 9. 326, 1899 ; F. Pisani, Conipt. Rend, 47. 294, 1858 ; M. Haupt, Pharm. Centr., 10. 509, 1899 ; Chem. News, 70. 206, 1894 ; L. de Bruyn, Rec. Chim. Pays-Bas, 10. 119, 1891 ; H. Cantoni and M. Rosenstein, Bull. Soc. Chim. (3), 35. 1069, 1906; P. Gerlinger, Zeit. angew. Chem., 19. 520, 1906; L. Moser, Zeit. anorg. Chem., 56. 143, 1907; Chem. Ztg., 31. 77, 1907; G. Fernekes and A. A. Koch, Journ. Amer. Chem. Hoc., 27. 1224, 1905 ; R. S. Dulin, ib., 17. 346, 1895 ; A. H. Low, ib., 18. 457, 1896 ; 24. 1082, 1902 ; Chem. Neivs, 74. 52, 1896 ; A. M. Fairlie, Eng. Min. Journ., 78. 787, 1905 ; A. H. Low, ib. t 59. 124, 1896 ; E. H. Miller, ib., Si. 519, 1906 ; P. E. Browning, Amer. J. Science (3), 46. 280, 1893 ; F. H. Heath, ib. (4), 25. 513, 1908 ; F. A. Gooch andF. H. Heath, ib. (4), 24. 65, 1907 ; Chem. News, 97, 174, 187, 1908; E. 0. Brown, Journ. Chem. Soc., 10. 65, 1857 ; J. W. Walker and M. V. Dover, ib., 87. 1584, 1905 ; R. Williams, Chem. News, 58. 273, 1888 ; J. W. Westmoreland, ib., 58. 78, 1888 ; Journ. Soc. Chem. Ind., 5. 48, 1886; U. Tsukakoski, Eng. Min. Journ., go. 969, 1910; F. E. Lathe, ib., 93. 1071, 1912 ; A. W. Peters, Journ. Amer. Chem. Soc., 34. 422, 1912 ; W. C. Bray and G. M. J. MacKay, ib., 32. 1193, 1910. 352 A TREATISE ON CHEMICAL ANALYSIS. separated iodine is removed by the sodium thiosulphate to produce the pheno- menon of "after-blueing"; 1 and it is retarded by increasing the concentration of the potassium iodide. This shortens the time required for the titration. Free mineral acids react with the potassium iodide, forming hydriodic acid. A little free acid does no particular harm, but if much acid be present, some cuprous iodide is dissolved. This acts as a catalytic agent, accelerating the oxidation of the hydriodic acid by atmospheric oxygen. This leads to " after-blueing " and to high results. The Preparation of the Solution. If the copper has been precipitated as thiocyanate (or sulphide), the precipitate is digested with, say, 5 c.c. of nitric acid until the solution of the copper is complete. Evaporate the solution on a water bath with hydrochloric acid in order to expel the red fumes of nitrogen oxides. 2 Let the solution cool. Redissolve the residue in 25 c.c. of water, and neutralise the acid by the addition of a few drops of ammonia. 3 Boil the solution in order to expel the ammonia. Add 2 c.c. of concentrated acetic acid in order that the acid may be in slight excess. 4 If necessary, boil to ensure the com- plete dissolution of the copper. The cold solution occupies about 50 c.c. If not, make the solution up to this volume with water. The Titration. Dissolve 3 grms. of potassium iodide in the solution. 5 Nearly white cuprous iodide separates and iodine is liberated. Shake the solution vigorously. The reaction is represented by the equation : " 2Cu(C 2 H 3 2 ) 2 + 4KI^^Cu 2 L 2 + 1 2 + 4KC 2 H 3 2 . The free iodine which separates colours the solution brown. The solution should be cold, and kept cold, in order to prevent loss of iodine. To avoid oxidation, etc., titrate the solution at once with a standard solution of sodium thiosulphate 6 until the brown colour of the iodine has changed to a faint straw yellow. 7 Then 1 When the blue colour of the starch iodide has been removed by the sodium thiosulphate, and more iodine passes into solution, the blue starch iodide colour will again appear. If this blue reappears after the colour has once been discharged by the titration, the phenomenon is called "after-blueing." J. H. Davies and E. P. Perman, Chem. News, 93. 225, 1908; K. Sugiura and P. A. Kober, Journ. Amer. Chem. Soc., 34. 818, 1912. 2 This is very important. A. H. Low (I.e.) recommends the addition of 5 c.c. of bromine water and boiling the solution until the bromine is expelled in order to ensure the removal of the nitric acid ; E. C. Kendall (Journ. Amer. Chem. Soc., 33. 1947, 1911) recommends boiling with sodium hypochlorite and afterwards taking up the free chlorine with phenol. 3 Or sodium bicarbonate, not the carbonate (page 290). Note that iodine may be carried off if much carbon dioxide escapes from the solution after the addition of the potassium iodide. 4 Avoid a large excess of acetic acid. The solution should not contain more than 3 c.c. of nitric, hydrochloric, or sulphuric acid, or 25 c.c. of 50 per cent, acetic acid, per 100 c.c. of solu- tion. For the action of hydrochloric acid on sodium thiosulphate titrations, see J. T. Norton, Amer. J. Science (4), 7. 287, 1899 ; Chem. News, 80. 27, 1899. S. U. Pickering (Journ. Chem. Soc., 37. 135, 1880) showed that more iodine is required to oxidise the thiosulphate as the pro- portion of hydrochloric acid increases. 5 Or add about 6 c.e. of a solution of potassium iodide containing 50 grms. of the solid per 100 c.c., that is, 1 c.c. contains about half a gram of the solid. An excess of potassium iodide does no harm ; too little will make the subsequent titration tedious owing to the gradual solu- tion of precipitated iodine. An excess is therefore necessary to keep the precipitated iodine in solution. A " supra-excess," of course, means waste, and potassium iodide is expensive. 6 STANDARD SOLUTION OF SODIUM THIOSULPHATE. Dissolve 12 grms. of the pure anhydrous salt in a litre of pure, recently boiled distilled water. If the crystalline salt is used, take 19 grms. The latter is first reduced to powder and dried between sheets of blotting paper. After the solution has stood about a fortnight, it is standardised by weighing 0*10, 0'15, and 0'20 grm. of pure electrolytic copper foil separately in three 250 -c.c. Erlenmeyer's flasks. Warm the copper with 5 to 10 c.c. of nitric acid (sp. gr. 1 '20). Make the volume of the solution in each flask up to about 25 c.c., and evaporate on a steam bath in order to expel red fumes of nitrogen oxides. Then treat each solution as indicated in the text. 7 If the end point does not appear before 25 c.c. of the thiosulphate have been added, add 2 grms. more potassium iodide. THE DETERMINATION OF COPPER AND CADMIUM. 353 add a couple of drops of a cold solution of starch (page 286) to develop the " iodine blue," and continue the titration cautiously, drop by drop, until the iodine blue is discharged. A drop in excess will suffice. The action of the thiosulphate on the iodine is represented by the equation : 2Na 2 S 2 3 + I 2 -> 2NaI + Na 2 S 4 6 . The blue colour of the starch iodide returns on exposure to the atmosphere. If the solution remains colourless two minutes, the titration may be considered finished. If 1 c.c. of the standard solution of sodium thiosulphate represents 0*0159 grm. of copper oxide, and 25 c.c. of the sodium thiosulphate solution be required for the titration, then the solution under investigation contained the equivalent of 25 x 0-0159 = 0-3975 grm. of copper oxide. Influence of Foreign Substances. If the blue colour of fche starch iodide, after a titration, returns almost immediately, and this again after the addition of two more drops of thiosulphate, the titration will probably be unreliable owing to the presence of oxidising agents probably arising from the incomplete expulsion of the red fumes of nitrogen oxides at an earlier stage of the operation. Excessive dilution, excessive amounts of acetic acid, and the presence of sodium or ammonium acetates retards the reaction between the copper acetate and the potassium iodide. This leads to low results. Hence the care t recommended above to remove the ammonia. In the presence of these retarding agents, the titration is somewhat tedious, since the " after-blueing " has to be followed up for a long time. The presence of a large amount of alkaline salts particularly sulphates and nitrates leads to wrong results. Ferric acetate liberates iodine from potassium iodide ; ferric phosphate does not. Hence, if a little iron be present, some recommend the addition of sodium phosphate. 1 If much iron be present, the copper should first be separated as sulphide, metal, or thiocyanate. Bismuth does not interfere beyond obscuring the end point and making it difficult to recognise when the reaction is complete. Too much thiosulphate may accordingly be run in before the starch is added. Bismuth forms a brown-coloured iodide which is very like the colour of the iodine in solution. Bismuth also imparts a dirty green colour to the starch indicator, so that the change is not from blue to colourless, but from dirty green to yellowish white. Lead, arsenic, and antimony interfere 2 with the determination, and should be removed before titrating. 3 Lead may be removed as sulphate, bismuth as phosphate, or both as peroxides by ammonium persulphate in alkaline solution (page 348). Errors. The chief points requiring attention are: (1) Errors due to the use of insufficient potassium iodide ; (2) too much acid liberates iodine from potassium iodide in presence of air, but sufficient acid must be present to give a prompt liberation of iodine and a sharp end point during the titration sulphuric and acetic acids can be used ; nitric and hydrochloric acids are not so good ; (3) loss of iodine by volatilisation (page 300) ; (4) the solution to be estimated should occupy as small a volume as possible and be titrated with the most dilute thiosulphate solution which will give a good end point; and (5) the end point is modified in the presence of a large excess of cuprous iodide. 1 L. Moser (Ze.it. anal. Chem., 43. 597, 1904) adds an excess of sodium pyrophosphate Na 4 P 2 7 to determine copper in the presence of arsenic and iron by the iodine process. Complex phosphates are formed ; of these, the copper salt is alone decomposed by acetic acid. 2 According to Fernekes and Koch, cadmium, zinc, aluminium, arsenic, antimonic and stannic salts do not interfere. 3 C. and J. J. Beringer, A Textbook of Assaying, London, 201, 1908. 354 A TKEATISE ON CHEMICAL ANALYSIS. The mixture of starch and cuprous iodide, just before the end point is reached, assumes a chocolate-brown coloration, and this changes to a pale buff colour with the last necessary drop of thiosulphate. The end point has occurred when another drop of thiosulphate does not diminish the prevailing light tint of the mixture. So long as a drop of thiosulphate falling on the quiet surface of the liquid being titrated produces a perceptible white area, the end point has not been reached. 172. The Evaluation of Copper Oxide and Carbonate. Dissolution of the Copper Oxide. Digest, say, 0'5 grm. of the oxide or carbonate on a hot plate with sulphuric acid ; or with 6 to 10 c.c. of nitric acid and 7 c.c. of sulphuric acid, until the volatile acids are expelled. Heat the solution until sulphuric acid fumes begin to come off. Calcined copper oxide dissolves rather slowly, while copper oxide which has not been calcined at a high temperature dissolves quickly. When cold, add 25 c.c. of water ; * and, if necessary, filter. Precipitation of the Copper. Place two strips of metallic aluminium 2 say, 15 cm. by 2 '5 crn. by 1 mm. in the solution so that one end of each strip rests against the side of the beaker. Metallic copper is precipitated. Heat the solution to boiling. Cover the beaker with a clock-glass to prevent loss by spurting. 3 When all the copper is precipitated about a quarter of an hour transfer the liquid to a second beaker. Decant through a 9-cm. filter paper. Wash the copper 4 on the aluminium plates with a weak solution of hydrogen sulphide in air-free water to prevent oxidation of the copper. Wash the copper in the second beaker in the same manner. Reject the nitrate and washings. 5 Dissolution of the Copper. Place the second beaker containing the copper below the funnel. Put 5 to 6 c.c. of nitric acid (sp. gr. 1'3) in the first beaker, to dissolve the copper from the aluminium plates, and pour the acid solution through the filter paper ; collect the solution in the beaker containing the copper below the funnel. Warm the solution until the copper is all dissolved. Wash the beaker and the filter paper. Expel the excess of acid, and treat the solution as described under de Haen's process (page 351). Errors. The Committee of the American Chemical Society, " On Uniformity in Technical Analysis," 6 reported in 1904 that the following numbers represent the extremes sent in by nineteen analysts for a sample of cupriferous slag : 1 Copper is slightly soluble in concentrated sulphuric and hydrochloric acids ; hence, the acidity of the solution should be feeble. 2 D. Tommasi, Bull. Chim. Soc. (2), 37. 443, 1882 ; Chem. News, 46. 62, 1882 ; A. H. Low, Journ. Amer. Chem. Soc., 18. 458, 1896 ; G. E. Perkins, ib.,2/[. 478, 1902 ; Chem. Neivs, 86. 86, 1902; E. V. Videgren, Zeit. anal. Chem., 48. 539, 1909. For magnesium: A. Villiers and F. Horg, Compt. Rend., 116. 1524, 1893. For zinc : G. H. Pfatf, Handbuch cler analytischen Chemie, Altona, 269, 1822 ; F. Mohr, Zeit. anal. Chem., I. 143, 1862 ; C. Ullgren, ib., 7. 442, 1868 ; F. Field, Chem. News, i. 62, 73, 1860. For cadmium : A. Classen, Journ. prakt. Chem. (1), 96. 259, 1865. For iron: A. A. Julien, Chem. News, 26. 9, 1871 (page 188). The iron should dissolve uniformly without the separation of black particles and the formation of ridges on the surface. 3 Cadmium used in place of aluminium does not spurt so much. For magnesium, see E. G. Bryant, Chem. News, 76. 30, 1897 ; S. A. Sworn, ib., 76. 59, 1897. 4 The copper film has its own characteristic red colour. If the solution contained any arsenic or antimony, the copper will be contaminated and appear dirty brown. 5 At this stage Field washed, dried, and weighed the metallic copper. He found that the copper precipitated by either zinc or iron always contained traces of these metals ; at the same time, a small amount of copper in the primary solution always escapes precipitation. In commercial analyses, Field considers that the small amount of foreign metal precipitated with the copper compensates for that left in the primary solution. 6 Journ. Amer. Chem. Soc., 26. 1644, 1904. THE DETERMINATION OF COPPER AND. CADMIUM. 355 Si0 2 . Fe. A1 2 3 . CaO. MgO. Zn. Mn. Cu. S. Highest . . 35-15 32'20 7 '16 13'53 3'21 4'25 1'53 0'46 1'98 Lowest . . 31-27 30 '33 3 '24 1073 I'OO 1'87 O'll 0'20 T45 The remarks made on the analyses of the argillaceous limestone, on page 248, are applicable here. 173. The Colorimetric Determination of Copper Carnelly's Process. In cases where but small quantities of copper are likely to be present, in red lead, for example, the copper is usually determined colorimetrically, 1 from the intensity of the colour of ammoniacal solutions ; 2 or from the intensity of the colour of ferrocyanide solutions. 3 Potassium ferrocyanide, in acid solutions of copper, produces an earthy brown coloration which can be detected when 1 part of copper is present in 1,000,000 parts of solution. In neutral solutions, 1 part of copper can be detected in 1,500,000 parts of solution; and in neutral solutions containing ammonium nitrate, 1 part of copper can be detected in 2,500,000 parts of solution. 4 Hence, the ferrocyanide test is made with neutral solutions in the presence of ammonium nitrate. A similar result occurs if ammonium chloride be substituted for the nitrate. The Standard Solution. Pipette 1 c.c. of an aqueous solution of potassium ferrocyanide 5 into a 100-c.c. flask; add 5 c.c. of an aqueous solution of ammonium nitrate, and make the solution up to 100 c.c. with distilled water. Transfer this solution, or an aliquot portion, to the right test glass of the colorimeter. The Test Solution. The solution under investigation must be neutralised. 7 If free potash be present, it is first neutralised and a slight excess of acid added, then a slight excess of ammonia. Boil off the excess of ammonia until the solution is neutral. If the solution is acid, add a slight excess of ammonia, and boil off the excess until the solution is neutral. Make the cold solution up to a definite volume, say, 100 c.c. Pipette 1 c.c. of the solution of potassium ferrocyanide into a 100-c.c. flask ; add 5 c.c. of the ammonium nitrate solution, and then add an aliquot portion, say, 50 c.c., of the solution under investigation, and 1 A. E. von Hubert, Berg. Hiltt. Ztg., 8. 667, 1849 ; IO. 804, 1851 ; V. Eggertz, ib., 21. 218, 1862; Zeit. anal. Chem., 2. 434, 1863 ; F. Delims, ib., 3. 218, 1864 ; Dingler's Journ., 172. 160, 1864 ; G. Panten, ib., 170. 391, 1863 ; C. Stammer, ib., 159. 641, 1861 ; A. Payen, ib., 27. 372, 1828 ; V. A. Jacquelain, Journ. prakt. Chem. (1), 46. 174, 1849 ; A. Miiller, ib. (1), 60. 474, 1853; Zeit. anal. Chem., 2. 434, 1863 ; M. Bergeron and L. 1'Hote, Compt. Rani., 80. 268, 1875 ; Wagmeister, (Jester. Zeit. Berg. Hutt., 13. 270, 1865 ; J. Parry and J. J. Morgan, Trans. Amer. Inst. Min. Eng., 30. 851, 1901 ; L. VV. Winkler, Zeit. angew. Chem., 26. 38, 1913. 2 G. Bischof, Dingler's Journ., 184. 433, 1867: J. Milbauer and V. Stanek, Zeit. anal. Chem., 46. 644, 1907 ; A. Austin, Min. World, 33. 753, 1910. 3 T. Carnelly, Chem. News, 32. 308, 1875. 4 This is nearly the same delicacy as hydrogen sulphide, which gives a brown coloration reaction is not recommended for the colorimetric test because of the disturbing effects of lead, etc. 5 POTASSIUM FEKROCYANIDE SOLUTION. Dissolve 1 grm. of potassium ferrocyanide in 25 c.c. of water. 6 AMMONIUM NITRATE SOLUTION. Dissolve 100 grms. of the salt in a litre of water. 7 Free acids make the tint paler ; free ammonia dissolves the precipitate produced by the ferrocyanide ; and free potash decomposes it. 356 A TREATISE ON CHEMICAL ANALYSIS. make the solution up to the mark with water. Transfer the whole solution, or an aliquot portion, to the left test glass of the colorimeter. The Comparison. Fill a burette with a standard solution of copper sulphate. 1 The burette may read to, say, -^th c.c. Run the copper sulphate solution gradually from the burette into the right test glass of the colorimeter, with constant stirring, until the tints of the test solution and the standard solution are the same. Calculations. A gram of red lead furnished a solution which was made up to 100 c.c. 100 c.c. of the standard solution required 0*78 c.c. of the standard copper solution to produce uniformity of tint. The 0*78 c.c. of the standard copper solution contained 0*000078 grm. of copper oxide. Hence, 100*78 c.c. of the standard contain the same amount of copper oxide as the 100 c.c. of the given sample. Hence, the sample contains 0*0078 per cent, of copper oxide. Accuracy of the Results. The following results illustrate the accuracy of the process with solutions containing known quantities of copper oxide : Used . . 100 10 1*15 0*9 0*8 07 0*5 mgrm. Found . . 102-01 11*2 1-19 0*91 0*82 0*71 0'52 mgrm. The amount of copper oxide so determined is rather too high with the more concentrated solutions, but the results are quite good with small quantities. Influence of Foreign Substances. With moderate proportions, say, 0*25 to 2 c.c., an excess of the ferrocyanide does not affect the accuracy of the method. Similar remarks apply to the ammonium nitrate. For example, no difference could be detected by Carnelly in the results obtained with solutions containing 5 and 15 c.c. of the respective salt solutions. Ammonium chloride, sodium chloride, calcium chloride, calcium sulphate, magnesium sulphate, and sugar did not appear to affect the results. It is, however, best to destroy the organic matter, if present, by evaporation with nitric acid. Influence of Lead. Lead salts form a white precipitate with potassium ferro- cyanide which does not interfere with a comparison of the colours. Carnelly made up a solution containing 2 grms. of lead nitrate (1*25 grins. Pb) with 0*255 grm. of copper nitrate in a litre of water. Varying proportions of this solution were taken, with the following results : Cuused .... 077 070 0'49 0*51 0*35 mgrm. Cu found . . ' . 0*80 0*75 0*51 0*49 0*38 mgrm. Hence, small amounts of lead have no appreciable effect on the accuracy of the comparison. Influence of Iron. If iron be present, the solution is oxidised with a few drops of nitric acid and evaporated to a small bulk. Precipitate the iron with ammonia, filter, and wash. Redissolve the precipitate in nitric acid and re- precipitate with ammonia. Filter and wash. Mix the nitrates, and boil the solution to drive oif the ammonia. The following represent the results obtained with varying amounts of iron, 2 which were separated before the copper was determined : Fe used ..... 0'20 2 '40 3*00 mgrm. Cuused 0*66 0'51 0*61 076 mgrm. Cu found 0*66 0'53 0*69 079 mgrm. The results are therefore quite satisfactory. 1 STANDARD SOLUTION OF COPPER SULPHATE. Dissolve 0*3138 grm. of pure copper sulphate CuS0 4 . 5H 2 in a litre of water. One c.c. of the solution contains 0*0001 grm. of CuO. 2 Small amounts of iron are determined by the thiocyanate process (page 200). THE DETERMINATION OF COPPER AND CADMIUM. 357 174. The Gravimetric Determination of Cadmium as Sulphate. Precipitation as Cadmium Sulphide. If any cadmium should be present, it will be found in the slightly acid nitrate from the copper thiocyanate. Treat the nitrate with hydrogen sulphide, 1 and cadmium sulphide, 2 more or less contaminated with basic salts, 3 will be precipitated. Conversion of the Sulphide to Sulphate. The sulphide is filtered ; washed with water containing about 5 per cent, of ammonium nitrate; dissolved in hot hydrochloric acid (1 : 3) ; and heated on a water bath with a slight excess of sulphuric acid. The solution is then evaporated to dryness in the weighed platinum crucible supported over the ring burner (fig. 96), so that there may be no loss by spurting. The crucible is then placed in a larger crucible, and heated to redness 4 until no more white fumes of sulphuric acid are evolved. Cool in a desiccator. If the sulphate is tinged with a yellow colour, some oxide is present. In that case, add a drop of sulphuric acid to " moisten " the mass, and again ignite. Cool in a desiccator, and weigh as cadmium sulphate CdS0 4 . The cadmium sulphate so obtained should be white in colour, and dissolve to a perfectly clear solution in water. 5 Every gram of cadmium sulphate CdS0 4 corresponds with O6159 gram of cadmium oxide CdO. The results are excellent. 6 For instance, Follenius found with solutions containing 0'4036 grm of cadmium sulphate : CdS0 4 found . . . 0'4036 0'4036 0-4033 0'4038 grm. Error . '. . -0'0003 +0'0002 grm. 1 The adjustment of the acid wants attention. The solution may contain between 2 and 7 c.c. of concentrated sulphuric acid per 100 c.c. If 100 c.c. of the solution contains more than the equivalent of 14 c.c. of hydrochloric acid (sp. gr. I'll), the separation of cadmium sulphide will not be complete in the cold ; and at 70, more than 5 c.c. of this acid will lead to incomplete precipitation of the cadmium sulphide. No precipitate at all will be produced if over 22 c.c. of hydrochloric acid be present in 100 c.c. of the cold solution, and 19 c.c. in hot solutions (70). The separation of cadmium sulphide is complete in dilute sulphuric acid (sp. gr. 1*19) in the cold. If but 5-40 c.c. of this acid be present in 100 c.c. of solution, the precipitate is finely divided and difficult to filter; with 40-70 c.c., the precipitate is more compact and easier to filter; and with 70-100 c.c. of this acid in 100 c.c. of solution the precipitation is complete, but the gas must be conducted through the solution a long time. In hot solutions (70), if over 30 c.c. of this acid be present, traces of cadmium will remain in solution 0. Follenius, Zeit. anal. Chem., 13. 411, 1874. 2 The colour of the precipitate is determined by the nature of the mother liquid, temperature, etc. The colour of the cadmium sulphide may vary from light yellow to an orange brown. N. von Knobukow, Journ. prakt. Chem. (2), 39. 412, 1889 ; G. Biichner, Chem. Ztg., II. 1087, 1107, 1887. 3 E.g., CdS.CdCl 2 ; OdS.CdS0 4 0. Follenius, Zeit. anal. Chem., 13. 411, 1874. 4 On evaporating a solution of cadmium sulphate with sulphuric acid, crystals of CdS0 4 . H 2 are formed (B. Kiihn, Arch. Pharm. (2), 50. 286, 1847 ; Schweigger's Journ., 60. 344, 1830 ; K. von Hatier, Journ. prakt. Chem. (1), 64, 477, 1855 ; (1), 72. 372, 1857 ; G. Wyrouboff, Bull. Soc. Mim., u. 275, 1888 ; 12. 366, 1889 ; F. Mylius and R Funk, Ber., 30. 832, 1897 ; Zeit. anorg. Chem., 13. 157, 1896; M. de Schulten, Compt. Rend., 107. 405, 1888).^ At 100 the water of crystallisation is given oft', and anhydrous cadmium sulphate is obtained. Cadmium sulphate can be heated for a long time at a red heat, under the conditions described in the text, without decomposition. At a still higher temperature, the sulphate begins to decompose, first acquiring a yellow tint, and finally passing to a dark brown-coloured oxide. If ammonium chloride be present, ammonium sulphate and cadmium chloride may be formed. 5 One part of water dissolves 0'59 part of the anhydrous sulphate at 23 (von Hauer). 6 H. Rose, Ausfiihrliches Hnndbuch der analytischen Chemie, Braunschweig, 2. 149, 1871 ; 0. Follenius, Zeit. anal. Chem., 13. 272, 1874; A. Carnot, Compt. Rend., 102. 621, 1886; Bull. Soc. Chim. (2), 46. 812, 1886 ; H. Baubigny, Compt. Rend., 142. 577, 792, 959, 1906. 358 A TREATISE ON CHEMICAL ANALYSIS. 175. The Volumetric Determination of Cadmium Berg's Process. The cadmium may be determined volumetrically, instead of gravimetrically as sulphate, when a standard solution of iodine is available. 1 Collect the precipitated cadmium sulphide on a filter paper or in a Gooch's crucible packed with asbestos. Wash the sulphide with air-free water, and transfer the precipitate by washing with 250 c.c. of air-free water to a 500-c.c. Erlenmeyer's flask. An excess of iodine solution 2 is added along with 10 c.c. of hydrochloric acid (sp. gr. 1*19). The reaction which occurs is represented by the equation : CdS + 2HC1 + 1 2 = CdCl 2 + 2HI + S. Dilute the solution to about 300 c.c., and titrate the excess of iodine with a standard solution of sodium thiosulphate, as indicated for copper (Haen's process, page 352). 176. The Electrolytic Determination of Cadmium Beilstein and Jawein's Process. The electrolytic process for cadmium (page 258) generally gives more accurate results than gravimetric or volumetric processes. The following method is due to Beilstein and Jawein. 3 The Electrolyte. Add a drop of phenolphthalein to a solution of a cadmium salt containing, say, 0'5 grm. of the sulphate or acetate. Then add potassium hydroxide until a permanent red colour is obtained. Cadmium hydroxide is precipitated. Add a solution of 98 per cent, potassium cyanide, slowly, with constant stirring, until the precipitate is all redissolved. An excess of the potassium cyanide disturbs the action. 4 Make the solution up to about 150 c.c. The Electrolysis. The solution so prepared is electrolysed in the cold 5 with a current density 0*5 to 0'7 amps., 6 and an electrical pressure of 4*6 to 5 volts. 7 In about 6 hours increase the current density to 1*0-1 '2 amps., and continue the electrolysis for about an hour. 8 If the electrolysis is finished, stop the current, wash, dry, and weigh the deposited metal in the usual manner. There is not so much danger of loss by re-solution of the metal when the current is stopped as is the case with copper (page 259). Preparation of Cadmium Sulphide for Electrolysis. Cadmium sulphide is usually separated in gravimetric analysis. Dissolve this in nitric acid. Expel the excess of acid by evaporation. Dissolve the dry residue in water, and add potassium hydroxide, etc., as indicated above. Removal of the Cadmium Deposit from the Electrode. The removal of cadmium from the cathode presents no particular difficulty. The metal readily dissolves in nitric acid, although it dissolves but slowly in hot hydrochloric and sulphuric acids. 1 P. von Berg, Zeit. anal. Chem., 26. 23, 1887. 2 Say 7 '94 grins, of iodine, and 16 grins, of potassium iodide per litre. One grrn. of iodine corresponds with 0*5058 grm. CdO. Hence, 1 c.c. of iodine represents O'OOl grm. of sulphur or 0-00401 6 grm. CdO. 3 F. Beilstein and L. Jawein, JBer., 12. 759, 1879 ; A. L. Davison, Journ. Amcr. Chem. Soc., 27. 1275, 1905 ; L. G. Kollock and E. F. Smith, ib., 27. 1527, 1905. 4 E. H. Miller and R. W. Page, School Mines Quart.. 22. 391, 1901 ; Zeit. anorq. Chem., 28. 233, 1901. 5 For hot solutions, 50-60, a current density of '1-0 '3 amp. and 3 3-4'5 volts will require between 5 and 6 hours. 6 The metal is not all precipitated after 12 hours with a current density of 0'5 amp. If 1 amp. be employed at the start, the deposit is non-coherent, and liable to drop off. 7 If left overnight, use a current density of 0'06 amp. and 3 '2 volts. 8 Test the solution for cadmium by acidifying a few drops with hydrochloric acid ; boil to expel the hydrocyanic acid (note, the gas is very poisonous) ; add a little H 2 S water : a yellow precipitate shows that cadmium is still present. CHAPTER XXVII. THE DETERMINATION OF ZINC. 177. The Analysis of Silicates containing Zinc Compounds. IF a silicate containing zinc be treated as indicated for clays, 1 several points require special attention. Silica Evaporation. It is highly important to remove all the silica during the " evaporation for silica," because, when ammonia is added to a solution con- taining zinc and silica, in order to remove the iron and alumina, 2 a zinc com- pound, said to be zinc silicate, will be precipitated. Again, although zinc chloride alone does not volatilise below a red heat, if an acid solution containing zinc and ammonium chlorides be evaporated to dryness, and gradually heated, the chlorides begin to sublime about 145. Hence, the evaporation of solutions containing silica, zinc, and ammonium chlorides should not be finished at temperatures much higher than 100. Ammonia Precipitation. The addition of ammonia to precipitate iron and aluminium hydroxides leads to the simultaneous precipitation of some zinc hydroxide, 3 and two, better three, reprecipitations are needed to eliminate the zinc from the ammonia precipitate, 4 even when zinc alone is to be determined. For instance, Waring obtained the following results with artificial mixtures of zinc with iron, etc.: Table LII. Contamination of the Ammonia Precipitate ~by Zinc. Used. Zinc found. Zinc. Iron. NH 4 01. Water. Ammonia solution. First nitrate. Second filtrate. Total zinc. Per cent. 60 60 60 60 60 Per cent. o 2 2 10 10 Grm. 2 2 2 4 6 c.c. 33 66 132 66 66 c.c. 5 5 5 10 10 59-10 59-18 59-02 58-87 58-90 0'94 0-82 1-03 1-23 1-00 60-04 60-00 60-05 60-10 59-90 1 Zinc ores are usually decomposed by digestion with acids, say, equal volumes of hydro- chloric and nitric acids, and heated on a hot plate until red fumes cease to be evolved. Add 3"5grms. of ammonium chloride, and heat the solution until it becomes thick and "pasty." When the mass is nearly dry, add 30 c.c. of hot water. If silica be present, iilter and wash thoroughly ; if an insoluble residue remains, fuse the ignited mass with a little sodium carbonate, take up the cold residue with water, and separate the silica by evaporation to dryness, etc. Add the acid liltrate to the first filtrate. 2 E. Prost and V. Hassreidter, Zeit. angew. Chem., 5. 166, 1892 ; W. G. Waring, Journ. Amer. Chem. Soc., 26. 4, 1904. 3 According to G. Bertrand and M. Savillier (Bull. Soc. Chim. (4), I. 63, 1907), crystals of calcium zincate CaH 2 Zn 2 4 . 4 H 2 are deposited on boiling an ammoniacal solution of a zinc salt and an excessive proportion of a calcium salt. 4 W. Funk, Zeit. angew. Chem., 18. 1687, 1905. 360 A TREATISE ON CHEMICAL ANALYSIS. In the analysis of three different samples of zinc blende, Waring also obtained ] : Table LHI. Contamination of the Ammonia Precipitate by Zinc. Per cent, of zinc. Total iron per cent. First filtrate. Second filtrate. Third filtrate. 8-62 47-20 3-00 0-58 8-75 46-05 4 75 0'40 14-32 16-50 2-05 0-55 It may also be regarded as a general rule that hydroxides precipitated by ammonia contain silica and phosphates, unless these substances have been pre- viously removed. Manganese Precipitation. If manganese be precipitated, before the zinc, by one of the usual oxidising agents bromine, hydrogen peroxide, ammonium persulphate, etc. some zinc hydroxide will be precipitated with the manganese hydroxide. 2 Stone found that manganese oxide precipitated in the presence of bromine from a solution (left after the basic acetate separation) containing 0'2069 grm. of zinc, and 0'2260 grm. of manganese, really contained Manganese 0'3045 0'2895 0'2840 0'2860 grm. Zinc 0-1860 0'1780 0'1858 0'1795 grm. These experiments show that the error is quite serious, particularly if much manganese and little zinc are present. According to Waring, in order to prevent the precipitation of zinc with the manganese, it is necessary to separate the latter as hydrated peroxide, and if the separation be effected from ammoniacal solutions, " the solution should be concentrated and contain a large amount of ammonium chloride." Precipitation of Zinc as Sulphide. The zinc can be determined either by the volumetric ferrocyanide process, or precipitated as sulphide. White zinc hydrosulphide is precipitated from neutral solutions of zinc salts by means of hydrogen sulphide, 3 but not if much over 4 c.c. hydrochloric acid (sp. gr. 1*12) be present per 100 c.c. In that case, however, the precipitation can be made if sufficient alkaline acetate, formate, oxalate, etc., be present to form alkaline chloride and set free the organic acid. Many organic acids may be present in considerable quantities without interfering with the precipitation of the sulphide. This is the case, for instance, with formic, acetic, citric, succinic, and chloracetic acids. The theory of the precipitation of zinc sulphide in hydro- chloric acid solutions is discussed on page 272 ; 4 the practice, on page 364. 1 According to L. L. de Koninck and E. von Winiwarter (Bull. Soc. Chim. Belg., 26. 238, 1912) the retention of zinc by ferric oxide is due to the formation of a double hydroxide, and not to adsorption. The retention is prevented by adding 5 per cent, of magnesium (as chloride) and sufficient ammonium chloride to prevent the precipitation of the magnesium hydroxide. E. Beyne, ib., 26. 355, 1912. 2 M. Dittrich and C. Hassel, Ber., 35. 15, 3266, 1902; H. Baubigny, Compt. Rend., 135. 22, 965, 1902 ; 136. 7, 449, 1903 ; W. G. Waring, Journ. Amer. Chem. Soc., 26. 4, 1904. 3 H. Baubigny, Compt. Rend., 107. 1148, 1888. 4 According to S. Glixelli (Zeit. anorg. Chem., 55. 297, 1907), the theory of the action of hydrogen sulphide on zinc salts does not depend upon equilibrium conditions similar to those indicated on page 273. He says that the reaction ZnS0 4 + H 2 S = ZnS + H 2 S0 4 is not reversible, but that a kind of " false equilibrium " occurs in acid solutions which may be very persistent. For the theory of false equilibrium, see J. W. Mellor, Chemical Statics and Dynamics, London, 417, 1904. Glixelli's view doe.s not include all the facts. THE DETERMINATION OF ZINC. 361 178. The Theory of the Basic Acetate Separation. It may now be well to recapitulate our treatment of a complex mixture of colouring oxides with a silicate. The acidified filtrate from the silica was treated with hydrogen sulphide. Lead, bismuth, copper, cadmium, mercury, arsenic, antimony, and tin sulphides were precipitated. The filtrate was boiled to expel the hydrogen sulphide, and treated with an excess of ammonium chloride and ammonia, when aluminium, titanium, and ferric hydroxides were precipitated. 1 Instead of using ammonia as precipitating agent, many prefer the so-called basic acetate separation of iron and aluminium from zinc, manganese, cobalt, arid nickel. This separation depends upon the fact that solutions of acetates of iron, aluminium, titanium, chromium, zirconium, and vanadium are decomposed (page 181) when heated, and insoluble sub- or basic acetates are deposited. On the contrary, the acetates of magnesium, manganese, zinc, nickel, and cobalt are stable enough to remain undecomposed when their solutions are boiled for a few minutes. 2 When sodium carbonate is added to a solution of ferric chloride, the pre- cipitate first formed dissolves in the ferric chloride. One part of ferric chloride will dissolve approximately ten parts of ferric hydroxide. 3 Any further addition of sodium carbonate produces a permanent precipitate, since the ferric chloride is already saturated with the hydroxide. If a solution of ferric chloride, just saturated with the hydroxide, be heated, the mixture decomposes, and ferric hydroxide is precipitated. If sodium or ammonium acetate be added to a solu- tion of ferric hydroxide in ferric chloride, ferric acetate is produced. This is hydrolysed on heating, and a basic ferric acetate is precipitated, while acetic acid passes into solution. The hydrolysis (page 181) which occurs on boiling is only completed in dilute solutions at least 500 c.c. of water should be present per gram of ferric hydroxide. For complete precipitation, therefore, as much sodium carbonate should be added as is possible without causing a permanent precipitate ; and only sufficient sodium acetate should be added to replace the combined chlorine of the ferric chloride in solution. If a mixture of aluminium, ferric, nickel, manganese, zinc, and cobalt salts be so treated, the acetates of manganese, zinc, cobalt, and nickel are hydrolysed at a much higher temperature (175) than the aluminium and ferric acetates. If a large excess of alkaline acetate be present, the acetates of manganese, nickel, cobalt, and zinc are decomposed at a much lower temperature, and they may, in consequence, be precipitated along with the iron and aluminium on boiling the solution. Manganese acetate does not appear to be hydrolysed under these conditions when less than twenty times the required amount of alkaline acetate is present ; nickel acetate is more liable to hydrolysis than cobalt ; while zinc is next to manganese in its tendency to hydrolyse and form the hydroxide or basic acetate and free acetic acid. 4 1 We ignore pro tempore the possible occurrence of certain constituents in both the hydrogen sulphide and in the ammonia groups. 2 Formates, succinates, and benzoates have been Suggested in place of the acetates, but it is generally said that the precipitations are not then so complete (F. Schulze, Chem. Centr. (2), 6. 3, 1861 ; W. Hisinger, Afhandlingar i Fysik, 3. 153, 1810). 3 F. Kessler, Zeit. anal. Chem., u. 258, 1872 ; 18. 8. 1879 ; Chem. Neivs, 27. 14, 1873. 4 A. Mittasch, Zeit. anal. Chem., 42. 492, 1903 ; W. Funk, ib., 45. 181, 1906 ; C. Stockmann, ib., 16. 172, 1877 ; C. Kramer, ib., 16. 334, 1877 ; G. Matzurke, ib., 17. 78, 1878 ; 0. Brunck, Chem. Ztg., 28. oil, 1904 ; G. Rosenthal, Dingler's Journ., 225. 154, 1877 ; C. Meineke, Zeit. angew. Chem., I. 224, 1888; J. Jewett, Amer. Chem. Journ., I. 251, 1879; Chem. News, 40. 273, 1879 ; H. Brearley, ib., 75. 13, 1897 ; 76. 165, 222, 1897 ; 79. 193, 1899 ; A. Jeannel, ib., 362 A TREATISE ON CHEMICAL ANALYSIS. Acetic acid is usually added to the solution before the alkaline acetate in order to lessen the danger of the joint precipitation of manganese, zinc, cobalt, and nickel along with the aluminium and iron. But since aluminium hydroxide (and phos- phate) is soluble in dilute acetic acid, and since alumina usually predominates in silicate analyses, while zinc, nickel, and cobalt, if present at all, only occur in minute quantities, it is best to work with as little free acetic acid as possible when dealing with clays and related substances. The presence of 11 per cent, of acetic acid will prevent the precipitation of the iron, and 5 per cent, will prevent the precipitation of aluminium. In illustration, two solutions containing the equivalent of 0'2 grm. of iron, and 0*2 grm. each of manganese, zinc, nickel, and cobalt, were mixed with 2 grms. of sodium acetate, and each made up to 300 c.c. ; 12 c.c. and 3 c.c. of acetic acid (sp. gr. 1*044) that is, 4 per cent, and 1 per cent, respectively of acetic acid were added. After treatment by the basic acetate process, the precipitate, according to Brearley, contained the equivalent of Acetic acid. Manganese. Zinc. Nickel. Cobalt. 1 per cent. . . 0'0585 0'2046 0'1770 0'0590 4 per cent. . . . trace nil OT219 0'0315 The precipitate by the basic acetate process contains titanium, most of the zirconium, 1 vanadium, uranium, 2 as well as phosphorus and arsenic, if present. Chromium (q.v.) is precipitated with the iron and aluminium, but the separa- tion is not satisfactory with chromium. This process can also be used for separating iron and aluminium from the rare earths. 3 The basic acetate process requires careful manipulation when much manganese, nickel, and cobalt are present. A second application of the process is then advisable, by redissolving the precipitate in hydrochloric acid, etc. Moore says four precipitations are necessary, Mackintosh says six ! If the process be con- ducted with due respect for the idiosyncrasies indicated in the text, two acetate precipitations, or one acetate precipitation followed by an ammonia precipitation, will generally suffice. Practical directions for the process now follow. 179. The Basic Acetate Separation. Neutralisation of the Solution. Gradually add a concentrated aqueous solu- tion of sodium carbonate from a burette, with constant stirring, until the precipitate begins to disappear slowly ; 4 then add a dilute solution in a similar manner until the small precipitate does not disappear with one or two minutes' stirring. If much iron be present, the solution will now have a reddish tint ; but if little iron be present, the solution will be almost colourless. Add 3 c.c. of acetic acid (sp. gr. 1 *044) to dissolve the precipitate, and allow the solution to stand a couple of minutes. 5 If the solution be not clear, add another drop of dilute acetic acid, and proceed as before. If necessary, repeat the addition of 17. 286, 1868 ; T. Moore, &>., 56. 75, 1887 ; R. L. Leffler, ib. t 77. 265, 1898 ; J. B. Mackintosh, ib., 56. 64, 1887 ; W. Gibbs, ib., II. 102, 1865 ; Amer. J. Science (2), 39. 58, 1865 ; W. Herz, Zeit. anorg. Chem., 20. 16, 1899 ; B. Reinitzer, Monats. Chem., 3. 256, 1882 ; G. Rudorf, Zeit. phys. Chem., 43. 262, 1903; V. von Eggertz, Berg. Hiitt. Ztg.. 26. 187, 1867 ; W. Hampe, Zeit. Berg. Hiitt. Sal., 25. 253, 1877 ; F. Mayer, Ber., 22. 2627, 1889 ; R. B. Riggs, Journ. Anal. App. Chem., 6. 94, 1892. 1 E. Linnemann (Monats. Chem., 6. 335, 1885) says the separation is not complete. Some zirconium remains with the iron precipitate. 2 H. Rheineck, Chem. News, 24. 233, 1871. 3 A. Beringer, Liebig's Ann., 42. 134, 1842 ; T. Scheerer, Pogg. Ann., 51. 467, 1840. 4 According to H. Tamm (Chem. News, 26. 37, 1872), oxidation with nitric acid leads to the precipitation of manganese with the iron and alumina. It is claimed that the oxidation of the iron is best effected with potassium chlorate. 5 Some add 3 c.c. of acetic acid (sp. gr. 1*044 ; that is, 33 per cent.), in excess, for every 100 c.c. of solution in the beaker previous to the dilution. As indicated in the text, an excess of THP] DETERMINATION OF ZINC. 363 dilute acetic acid, drop by drop, until the solution does become clear under the conditions stated. Precipitation. Dilute the solution to about 400 c.c. and heat to boiling; add about 6 c.c. of sodium acetate solution. 1 Do not boil the mixture more than a minute, or the precipitate will become slimy, and some may dissolve in the acetic acid. The precipitate is bulky and difficult to filter 2 and wash. Filter and wash with hot water containing 50 grms. of sodium acetate per litre. 3 Suck the precipitate dry at the pump. The precipitate is seldom or never ignited and weighed, because it is always contaminated with alkaline salts. Hence, the precipitate must be dissolved in nitric or hydrochloric acid, and reprecipitated with ammonia, as described on page 182. The process just outlined gives satisfactory results. The presence of sulphates leads to imperfect precipitations. 4 Tartaric, citric, and organic acids should be absent. Indeed, it might here be emphasised that analysts are very chary about introducing organic acids into their solutions, because, if later separations have to be made, there is some uncertainty as to the effect of the organic acids' on subsequent precipitations. Tartaric, citric, oxalic, malic, and other organic acids, dextrin, and the sugars, for instance, prevent the precipitation of aluminium, chromium, and ferric hydroxides by ammonia ; 5 and of manganese sulphide by ammonium sulphide. 6 Evaporate the combined filtrates to about 200 c.c. If a precipitate should separate, filter it off, dissolve the precipitate in hydrochloric acid, and precipitate the iron and alumina as basic acetates ; add the filtrates to the main filtrate. Suppose that zinc, cobalt, and nickel be absent, the manganese may now be precipitated by a number of different processes by bromine, by ammonium sul- phide, etc. If cobalt, nickel, or zinc be present, remove the zinc by the formic acid process, and treat the filtrate from the zinc sulphide by the method described below. acid, however, leads to an incomplete precipitation of the iron and aluminium ; but too little acid will lead to the precipitation of manganese, zinc, nickel, and cobalt, if these elements be present. 1 SODIUM ACETATE SOLUTION. Dissolve 167 grms. of sodium acetate in water and make the solution up to 500 c.c. One c.c. of this solution will contain nearly one-third of a gram of sodium acetate. Ammonium acetate is not so suited for the work, although it is sometimes recommended, chiefly because it is removed from the precipitate on ignition. Ammonium acetate has the disadvantage of being more readily hydrolysed (page 181) than sodium acetate in aqueous solution. H. C. Dibbits (Bull. Soc. Chim. (2), 18. 490, 1872 ; (2), 2O. 258, 1873) says that 7'6 per cent, of ammonium acetate is hydrolysed against 0'14 per cent, of sodium acetate. Hence, on boiling solutions of ammonium acetate ammonia is evolved, and the solution becomes more acid. Hence, the acidity of solutions of ammonium acetate is less easy to control than solutions of sodium acetate. W. Gibbs (Chem. News, n. 102, 1865) adds the acetate to the cold solution and then heats the solution to boiling. H. Brearley (ib., 79. 193, 1899) says a large excess of acetate always leads to an imperfect separation of, say, nickel and iron, because (1) the excess of acetate transforms the nickel as well as iron into acetate ; and (2) the nickel acetate is partially decomposed along with the ferric acetate. " When the acetate is added to the cold solution, both these factors exert their maximum influence. On adding acetate to the boiling solution, the first portion of it precipitated the iron ; when the remaining portion is added, the conditions are changed, for the iron is present in suspension only, and its influence is greatly lessened. " 2 H. N. Warren (Chem. News, 6l. 63, 1890) agitates finely powdered glass with the pre- cipitate. He claims that if this is properly done, the glass carries down the precipitate, and enables the washing to be done more quickly. The glass does no harm, since the precipitate is afterwards dissolved in acid. a The filtrate may be tested for iron by potassium ferrocyanide, not thiocyanate. 4 Sulphates must not be determined in the filtrate from the basic acetate process on account of the great probability of contamination with sulphates by the large amounts of reagents used. s M. Juette, Compt. Mend., 66. 417, 1868 ; Chem. News, 18. 63, 1868. 6 J. Spiller, Journ. Chem. Soc., 10. 110, 1858 ; Chem. News, 8. 280, 1863 ; 19. 166, 1869 ; H. How, ib., 19. 137, 1869; H. N Draper, ib., 8. 260, 1863; F. Field, ib., 3. 65, 1861 ; F. Pisani, ib., 3. 257, 1861 ; L. J. Curtmau and H. DuBin, Journ. Amer. Chem. Soc., 34. 1485, 1912; E. Salkowski, Zeit. physiol. Chem., 83. 159, 1913. 364 A TREATISE ON CHEMICAL ANALYSIS. 180. The Separation of Zinc from Manganese, Cobalt, and Nickel. Assume that the filtrate from the ammonia or basic acetate separation contains zinc, 1 manganese, 2 cobalt, and nickel, as well as magnesia, lime, and the alkalies. Precipitation of Zinc Sulphide. The ammoniacal solution is neutralised with formic acid, and about 5 c.c. additional formic acid (sp. gr. 1-2) is added per 150 c.c. of solution. Not more than about 3 per cent, excess of formic acid should be present, or some of the zinc may escape precipitation. Enough sodium formate should be present to react with the inorganic acids set free by the action of the hydrogen sulphide, and so prevent free hydrochloric acid accumulating in the solution. 3 Heat the solution to 50 to 60. Pass hydrogen sulphide (two bubbles a second) through the warm solution for an hour in order that the zinc sulphide may coagulate, and filter quickly. 4 Zinc sulphide is practi- cally insoluble in formic acid of the strength just indicated ; and, like many other sulphides, changes from the colloidal form after precipitation. Hence, it is better to let the mixture stand overnight in a warm place with a slow current of hydrogen sulphide passing through the solution, or else let it stand some time in a corked flask. The larger the excess of formic acid, the slower the separation of the zinc sulphide, and the larger the excess of alkaline acetates or formates present, the more slimy the precipitated sulphide and difficult to filter and wash. Filtration, and Washing the Zinc Sulphide. Filter off the precipitated 1 Zinc may be found in reagents kept in Jena and "nonsol" glass bottles page .145. Acids transported in carboys with glass covers secured by zinc rings are liable to contamination owing to the splashing of the acid into the neck of the carboy. This explains how zinc has been reported when no zinc was present in the original sample. A. Vita, Stahl Msen, 32. 1532, 1912 ; Sprech., 42. 787, 1912. 2 The manganese rnay or may not have been removed prior to the zinc. 3 W. Hampe, Glum. Ztg., 9. 543, 1885 ; Chem. News, 52. 313, 1885 ; M. Bragard, Chem. Ztg., 10. 729, 1886; E. Dohler, ib., 23. 399, 1899 ; P. von Berg, Zeit. anal. Chem., 2$. 512, 1886 ; G. Naumann, ib., 28. 57, 1889 ; W. Funk, ib. t 46. 93, 1907 ; H. Kinder, Stahl Eisen, 16. 675, 1896; H. Delffs, Zeit. Chim. Pharm., 4. 1860; Chem. News, 41. 279, 1880; T. Cockburn, A. D. Gardiner, and J. W. Black, Analyst, 37. 443, 1912. According to Bragard, the solution should have 5 c.c. of formic acid (sp. gr. 1'1136) per 0'03 grm. of nickel in solution, and this amount should not be exceeded if the zinc is to be precipitated completely. If much nickel be present, the solution should be diluted to 500-600 c.c. If the solution be heated, more acid is needed to prevent the precipitation than if the solution be " gassed " cold. A very slight excess of hydrochloric acid added to a neutral solution is sometimes used instead of formic acid, but there is a difficulty in regulating the correct amount of acid, and the results are satisfactory when only small amounts of zinc are present. C. Brunner, Dingler's Journ., 150. 369, 1858 ; A. Klaye and A. Deus, Zeit. anal. Chem., IO. 190, 1871 ; F. Oettel, ib., 27. 16, 1888 ; J. B. Kirkland, Rep. Australasian Assoc., 2. 397, 1890. For acetic acid, see A. Rosenheim and E. Huldschinsky, Zeit. anorg. Chem., 32. 84 ; 1902 ; A. Thiel, ib., 33. 1, 1902 ; H. Nissenson and W. Kettembeil, Chem. Ztg., 29. 950, 1905 ; J. Flath, ib., 25. 564, 1901 ; W. G. Waring, Journ. Amer. Chem. Soc , 26. 4, 1904 ; W. Funk, Zeit. anal. Chem., 46. 104, 1907 ; M. Bragard, ib., 26. 209, 1887 ; H. Kinder, Stahl Eisen, 16. 675, 1896 ; J. Platz, ib., 9. 494, 1889 ; H. Baubigny, Compt. Rend., 108. 236, 450, 1888. For citric acid, see F. Beilstein, er., n. 1715, 1878. For oxalic acid, A. Carnot, Compt. Rend., 102. 678, 1886. For succinic acid, H. Alt and J. Schultze, er., 22. 3259, 1889 ; H. Delffs, Chem. News, 41. 279, 1880. For monochloracetic acid, P. von Berg, Zeit. anal. Chem., 25. 512, 1886. For trichloracetic acid, J. J. Fox, Journ. Chem. Soc., 91. 964, 1907. It is claimed that this acid is better for the separation of zinc and cadmium, since the acidity limits are not so narrow as with hydrochloric acid. For benzolmonosul phonic acid, see H. Schilling, Chem. Ztg., 36. 1352, 1912. 4 Zinc sulphide is a difficult precipitate to filter and wash. It may be gelatinous or in a very finely divided condition suspended in the fluid ; in the former case washing is very slow ; in the latter case the precipitate passes through the paper (pages 178 and 275). P. Pipereaut and A. A. Vila (Internat. Cong. App Chem., 7. i. 141, 1909) say that zinc sulphide is pre- cipitated in a very dense form, readily washed, by the addition of finely divided sulphur to the boiling alkaline solution. The precipitate is white, and when the precipitation is complete, a pink coloration appears. E. Murmaim (Monats. Chem., 19. 404, 1898) adds a little mercuric chloride to the solution before "gassing." The results are good. See page 319. THE DETERMINATION OF ZINC. 365 sulphide and test the filtrate with hydrogen sulphide to make sure that all the zinc is precipitated. Wash the precipitate by decantation two or three times with water saturated with hydrogen sulphide, and containing a little formic acid. Collect the washings in separate beakers, so that if the filtrate should commence to run through turbid, 1 as sometimes occurs when the ammonium salts are nearly all washed out, it will not be necessary to re-filter a large quantity of liquid. The precipitate is usually free from manganese, cobalt, and nickel. In case a relatively large amount of these elements be present, the precipitate may be contaminated, and it is advisable to dissolve the precipitate in dilute hydrochloric acid, neutralise with ammonia, and reprecipitate with hydrogen sulphide. Add the second filtrate to the first. The joint filtrates are reserved for the separation of manganese, cobalt, and nickel, as indicated on page 388. The precipitate of zinc sulphide is then treated as indicated in the next section. Influence of Formic Acid. To illustrate the relation between the amount of acid in the solution and the efficiency of the formic acid separation, the following numbers are quoted from Bragard's work : Table LIV. Effect of Formic Acid on the Separation of Zinc and Nickel as Sulphides. Zinc, grm. Nickel. grin. Volume solution, c.c. Formic acid, c.c. Zinc ppd. Nickel ppd. with zinc per cent. 0-0325 0-0048 300 5 0-0324 54-2 0-0325 0-0048 300 6 0--0325 37-5 0-0325 0-0048 300 7 0-0322 22-9 0-0325 0-0048 300 10 0-0325 o-o 0-0325 0-0096 500 10 0-0325 o-o 0-3252 0-0960 400 10 0.3252 28-1 0-3252 0-0960 500 20 0-3247 18-4 0-3252 0-0960 600 30 0-3252 8-0 0-3252 0-0960 600 40 0-3248 o-o Manganese, aluminium, cobalt, and iron are not usually precipitated if the solution has 15-20 c.c. of formic acid (sp. gr. 1*2) per 250-500 c.c. of solution ; but if much manganese, etc., are present it is well to make sure that the precipitate is free from contamination, and if necessary a reprecipitation must be made. The 1 904 report of the Committee on Uniformity in Technical Analyses 2 shows analyses of three samples of zinc ore by forty-two chemists. As a result : SAMPLE A. SAMPLE B. SAMPLE C. Zn. Fe. Zn. Fe. Zn, Fe. Highest results .... Lowest results , 39-22 12-20 21-92 18-04 38-86 28-90 15-00 8-40 59-79 56-03 3-26 2-10 1 0. Miihlhauser, Zeit. angew. Chem., 15. 731, 1902. 2 Journ. Amer. Chem. Soc., 26. 1644, 1904. 366 A TREATISE ON CHEMICAL ANALYSIS. The results for samples A and B are surprisingly discordant, and seriously undermine confidence in many published analyses. Each chemist used the particular process he thought best. 181. The Determination of Zinc as Phosphate. Zinc sulphide is not an easy precipitate to prepare for the balance, and, in consequence, many adopt Tamm's l process, in which the zinc is precipitated as phosphate. This is an easy gravimetric process. Although, in practice, with the standard solutions ready made, it is far easier to get good results with the unwashed sulphide by volumetric processes, than by gravimetric methods which involve the washing of the precipitated zinc sulphide. Dissolve the precipitated sulphide in hot dilute hydrochloric or nitric acid. Boil the solution to expel the hydrogen sulphide. Filter off any sulphur which separates. The acid solution, occupying about 150 or 200 c.c., 2 is neutralised with ammonia and then made faintly acid with hydrochloric acid. Heat the solution to boiling, add 15 to 20 times as much ammonium phosphate 3 as there is zinc in the solution. If a precipitate does not form, carefully add ammonia until it does. 4 The solution will now be alkaline. Keep the vessel and contents warm on a water bath 10-15 minutes and the amorphous ammonium zinc phosphate ZnNH 4 P0 4 first precipitated will crystallise. If the solution be boiled it bumps badly. Let the precipitate settle. 5 Filter through asbestos 6 in a Gooch's crucible. Clean 7 the beaker with the mother liquid, and wash with hot water containing about 1 per cent, of ammonium phosphate until the precipitate is free from chlorides (silver nitrate with nitric acid is the test). Finish the washing with alcohol. Dry at 100, and weigh the precipitate as ZnNH 4 P0 4 . This weight, when multiplied by 0*4561, represents the corresponding amount of zinc oxide ZnO. 8 The precipitate, if desired, can be ignited to low redness, and weighed as zinc pyrophosphate Zn 2 P 2 7 . If the ignition be conducted at too high a temperature, the pyrophosphate may melt, and be absorbed by the asbestos, 1 H. Tamra, Chem News, 24. 148, 1871 ; G. Lbsekann and T. Mayer, Chem. Ztg., 10. 729, 1886; L. Jawein, ib., II. 347, 1887; M. Bragard, ib., 10. 1605, 1886; G. C. Stone, Journ. Amer. Chem. Soc., 4. 26, 1882 ; R. W. Langley, ib., 31. 1051, 1909 ; A. C. Langmuir, ib.,2i. 125, 1899; Chem. News, 79. 183, 1899; H. D. Dakin, Zeit. anal. Chem., 39. 273, 1900 ; Chem. News, 82. 101, 1900 ; M. Austin, Amer. J. Science (4), 8. 206, 1899 ; R. C. Boyd, School Mines Quart., II. 355, 1890; J. Clark, Journ. Soc. Chem. Ind., 15. 866, 1896; P er 3 If microcosmic salt be employed, there is more difficulty in washing the precipitate free from alkalies than when the ammonium salt is used. 4 The precipitate is soluble in acids and ammonia, and slightly soluble in large quantities of ammonium salts. For chromium phosphate, see H. Kammerer, Zeit. anal. Chem., 12. 375, 1873. 5 The important points to observe in precipitation are a large excess of reagent, and sufficient ammonium salts to flocculate the precipitate. If materials other than alkaline salts be present the precipitate may be contaminated accordingly. See the preceding footnote ; if too great an excess of ammonium salts be present, a little zinc phosphate may pass into solution. 6 A slight loss of zinc by reduction to metal, and subsequent volatilisation may occur if filter paper be used and the precipitate be afterwards ignited. If the precipitate be ignited, the asbestos used in the Gooch's crucible should have been previously calcined. Otherwise, drying at 100 will suffice. 7 The precipitate is inclined to stick tenaciously to the walls of the beaker in which the precipitation is made. It may then be necessary to dissolve the precipitate from the walls with a little acid, evaporate to dryness in a weighed crucible, etc. 8 Tests made under various conditions show that here less than 0'0003 grm. of zinc phosphate escapes with the nitrate and in the washings. THE DETERMINATION OF ZINC. 367 with but a slight loss in weight possibly O'l per cent. The weight of the pyrophosphate multiplied by '53 402 gives the corresponding amount of zinc oxide ZnO. 182. The Volumetric Ferrocyanide Process for Zinc. Zinc ferrocyanide is precipitated from a hydrochloric acid solution of a zinc salt by potassium ferrocyanide. When all the zinc has been converted into zinc ferrocyanide, any further addition of the potassium ferrocyanide will cause the solution to give a brown coloration with a uranium salt. 1 Theory of Process. If the ferrocyanide be gradually added to the solution, there is first a slow formation of normal zinc ferrocyanide : 4ZnCl 2 + 2K 4 FeCy 6 ->8KCl + 2Zn 2 FeCy 6 . If the solution be cold (20 to 25), the solution will colour a drop of uranium acetate a fading yellowish brown. When about three-quarters of the required amount of potassium ferrocyanide has been added, the solution no longer affects the indicator uranium salt. The precipitate becomes flocculent. The slow reaction just indicated is followed by a faster reaction : 6Zn 2 FeCy 6 + 2K 4 FeCy 6 -> 4K 2 Zn 3 (FeCy 6 ) 2 . The indicator is coloured permanently when the second reaction is completed. The whole reaction is therefore written : 3ZnCl 2 + 2K 4 FeCy 6 . 3H 2 -> Zn 3 K 2 (FeCy 6 ) 2 + 6KC1 + 3H 2 0. After a little practice, the transient colour first produced will not be mistaken for the final permanent coloration. The first coloration gradually fades, and the permanent tint becomes more intense. The solution during the former stage is bluish white, and during the latter stage, pale cream. The first transient coloration of the indicator does not appear if the solutions be hot, because the first reaction proceeds quickly. This reaction is the basis of Galetti's process 2 for the volumetric determination of zinc. If the process be carried out in a regular, uniform manner, comparable and satisfactory results can be obtained. " In my opinion," says Seaman, " the results for zinc by the ferrocyanide process more nearly approach the absolute amount of zinc in an ore than the results obtained by working slowly with the gravimetric process." 1 E. H. Miller, Journ. Amer. Chem. Soc., 18. 1100, 1896 ; 22. 541, 1900; 24. 226, 1902 ; E. H. Miller and J. A. Matthews, ib. s 19. 547, 1897 ; E. H. Miller and M. J. Falk, ib., 26. 952, 1904; E. H. Miller and J. L. Danziger, ib., 24. 827, 1902; G. C. Stone and D. A. van Ingen, ib., 19. 542, 1897 ; E. H. Miller and E. J. Hall, School Mines Quart., 21. 267, 1900; Chem. News, 82. 177, 1900 ; F. Reindel, Dingler's Journ., 190. 395, 1868 ; K. Zulkowsky, ib., 249. 175, 1893; M. Bragard, Beitrage zur Kenntnis der quantitativen Bestimmung des Zinks, Berlin, 1887 ; G. Wyrouboff, Ann. Chim. Phys. (5), 8. 444, 1876 ; L. L. de Koninck and E. Frost, Zeit. angew. Chem., 9. 460, 564, 1896 ; Chem. News, 76. 6, 15, 29, 38, 51, 1897. 2 M. Galetti, Zeit. anal. Chem., 4. 213, 1865; 8. 135, 1869; 14. 190, 1875; Butt. Soc. Chim. (2), 2. 83, 1864 ; J. Muller, ib. (4), I. 13, 61, 1907 ; C. Fahlberg, Zeit. anal. Chem., 13. 379, 1874 ; L. Blum, ib., 29. 271, 1890; 31. 60, 1892 ; E. Murmann, ib., 45. 174, 1906 ; Monats. Chem., 19. 404, 1898 ; L. L. de Koninck and E. Frost (I.e.} : A. Renard, Compt. Rend., 67. 450, 1868 ; M. Pouget, ib., 129. 45, 1899 ; W. G. Waring, Journ. Amer. Chem. Soc., 26. 4, ]904 ; 29. 265, 1907 ; A. H. Low, ib., 15. 550, 1893 ; 22. 198, 1900 ; W. H. Seaman, ib., 29. 205, 1907 ; W. H. Keen, ib., 30. 225, 1908 ; G. C. Stone, ib., 17. 475, 1895 ; 30. 904, 1908 ; Reports on this subject, ib., 29. 262, 1907 ; Chem. News, 67. 517, 1893 ; F. M. Lyte, ib., 31. 222, 1875; W. H. Keen, ib., 98. 201, 1908; Journ. Amer. Chem. Soc., 30. 225, 1908; R. W. Mahon, Amer. Chem. Journ., 4. 53, 1882 ; H. S. Pattinson and G. C. Redpath, Journ. Soc. Chem. Ind., 24. 228, 1905 ; E. Donath and G. Hattensauer, Chem. Ztg., 14. 323, 1890; Report in Chem. News, 67. 5, 17, 1893 ; E. Rupp, Archiv Pharm., 241. 331, 1903 ; E. Rupp, Chem. Ztg., 33. 3, 1909 ; C. Kirschnick, ib., 31. 960, 1908. 3 68 A TREATISE ON CHEMICAL ANALYSIS. Disturbing Agents. Anything which oxidises or decomposes the ferrocyanide solution should be absent strong acids, chlorine, etc. Metals which give insoluble or sparingly soluble ferrocyanides aluminium, cobalt, cadmium, 1 copper, iron, 2 manganese, 2 lead, nickel, arsenic, antimony, and magnesium should also be absent, Some of these substances, however, produce no appreciable effect if only present in small quantities e.g., aluminium and lead. 3 The composition of the precipitate varies according as the solution is alkaline, neutral, or acid ; according as the solution is acid with acetic or hydrochloric acids ; if the hydrochloric acid be in excess, the potassium ferrocyanide 4 will be decomposed, forming a blue solution which spoils the work. This all shows that uniform conditions are indispensable for accurate /cork. The end point is not very sensitive in the presence of hydrochloric acid, but it is sharpened a little in presence of ammonium chloride. In consequence, it is necessary to deduct from the burette reading the amount of the standard ferrocyanide solution in excess of that actually required for the above reaction needed to give the brown coloration with 'the uranium nitrate used as indicator. Since uranium ferrocyanide is soluble in hydrochloric acid, this excess is dependent upon the amount of hydrochloric acid present. Hence, it is necessary to work under certain definite conditions which experience has shown to be the best, and to determine the allowance to be made for the indicator. /Standardising the Potassium Ferrocyanide Solution. Dissolve 34*64 grms. of crystalline potassium ferrocyanide K 4 FeCy 6 . 3H 2 in water, 3 and make the solu- tion up to a litre if not clear, filter. 5 1 c.c. corresponds with 0*01 grm. of zinc. In order to standardise this solution, ignite pure zinc oxide, 6 and cool in a desic- cator to ensure freedom from moisture and carbonates. Dissolve 12 '45 grms. in concentrated hydrochloric acid, add ammonia until a slight permanent precipitate is formed ; dissolve this in a drop or two of dilute hydrochloric acid ; add 30 c.c. of concentrated hydrochloric acid, and 50 grms. of ammonium chloride, and make the solution up to a litre. 1 c.c. corresponds with 0'01245 grm. of ZnO. 1 Cadmium salts are fatal to the success of the determination of zinc by the ferrocyanide process. Cadmium ferrocyanides Cd 2 FeCy 6 and K 2 CdFeCy 6 , or a mixture of the two are formed (E. H. Miller, Journ. Amer. Cliem. Soc., 22. 541, 1900 ; 24. 226, 1902). If cadmium, copper, and antimony be present, a difficult separation by hydrogen sulphide may be necessary. It is generally considered best to precipitate the copper, lead, cadmium, and antimony by boiling the dilute hydrochloric acid solution with a piece of metallic aluminium (page 304). If arsenic be present, some iron may escape precipitation with ammonia. In that case, the solution may be evaporated to dry ness. Boiling of the residue with concentrated hydro- chloric acid and bromine will expel the arsenic. Lead will precipitate copper in an acidified solution. Members of the iron group should be absent. If manganese be present, it can be removed by means of bromine (page 177). A. Kenard (Bull. Soc. Chim. (2), n. 473, 1869) removed manganese by sodium phosphate in ammoniacal solution, and obtained for the analysis of an untreated sample 1*036 grm. zinc; and for the same sample treated previously with sodium phosphate to remove manganese, 0'998 grm. zinc. If copper is to be precipitated with aluminium, ammonium salts should be absent. If much aluminium be present, derived from the solution of metallic aluminium, the results of the ferrocyanide titration will be irregular (E. H. Miller and E. J. Hall, School Mines Quart., 21. 270, 1900: Chem. News 82. 177, 1900). 2 C. Fahlberg (Zeit. anal. Chem., 4. 213, 1865) removes iron and manganese by shaking the solution with lead dioxide and filtering before titrating. 3 Y. Lehner and C. C. Meloche (Journ. Amer. Chem. Soc., 35. 134, 1913) show that lead does no harm in the ordinary ferrocyanide titration for zinc. 4 L. Blum (Zeit. anal. Chem., 24. 285, 1895) deals with the impurities in commercial potassium ferrocyanide. 5 F. Moldenhauer (Chem. Ztg., 15. 223, 1891 ; Chem. News, 64. 150, 1891) proposes to prevent the decomposition of the solution by adding 1 to 2 grms. of potassium hydroxide per litre. The standard solution should be preserved in darkness. 6 Zinc oxide is preferable since it is more easily procured free from iron, lead, etc. , than metallic zinc. J THE DETERMINATION OF ZINC. 369 The Titration. Pipette 25 c.c. of the standard solution of zinc oxide into a 600 c.c. flask or beaker, and add sufficient water to make about 180-200 c.c. Warm the solution to about 70-80, and pour about half into another beaker. Add the standard ferrocyanide solution from a burette, 1 c.c. at a time, until the solution has a greyish colour, and a drop in contact with a drop of uranium nitrate solution 1 on a porcelain plate (page 454) gives a distinct brown coloration. Mix the two solutions. Suppose that 14 c.c. of the ferrocyanide gave no coloration with the first half, and 15 c.c. gave a distinct brown. Then at least 28 c.c. of ferrocyanide solution will be required for the titration. Hence, a total 28 c.c. of ferrocyanide may be added without fear of exceeding the limit. Then add the ferrocyanide drop by drop, and test for the end by means of the spot test. Then pour the solution backwards and forwards from . one beaker to the other, and finish the titration. This procedure is less tiresome than titrating the whole solution directly, and there is less danger of exceeding the limit. A little time should be allowed for the spot test to change colour, or the end point may be exceeded. This may be corrected in the following manner : Let the test drops be added in regular order, and keep a memorandum of the corresponding burette readings. The first drop which shows the brown coloration after standing a short time is the proper reading, ^rhis gives the volume of the ferrocyanide solution corresponding with the zinc in the given solution. Hence the amount of zinc oxide is readily computed. 2 Indicator Allowance. An allowance should be made for the amount of ferrocyanide in excess of that required to convert all the zinc to ferrocyanide in order to produce the brown coloration with the indicator. Pour 180 to 200 c.c. of water into a beaker, and add 10 grms. of ammonium chloride, 6 c.c. of concen- trated hydrochloric acid. Warm the mixture to 70-80. Put a number of drops of uranium nitrate on a white tile, and add ferrocyanide from the burette with constant stirring until the solution gives a distinct brown with the indicator. The volume needed should not be greater than 0*5 c.c., and, in subsequent titrations, it should be subtracted, as a correction, from each burette reading. 183. The Evaluation of Zinc Oxide. When the zinc oxide is to be determined in a commercial sample of the oxide, proceed as indicated in the preceding section for the preparation of standard zinc oxide solution, and keep the conditions as nearly the same as possible : the 1 URANIUM NITRATE SOLUTION. 15 grms. of uranium nitrate in 100 c.c. of water (C. Fahlberg, Zeit. anal. Chem., 4. 213, 1865). H. Nissenson and W. Kettembeil (Chem. Ztg. t 20. 591, 1905) and W. G. Waring (Journ. Amer. Chem. Soc., 26. 4, 1904) use a 1 per cent, solution of ammonium heptamolybdate instead of uranium nitrate or acetate, provided hydrogen sulphide is absent. A trace of hydrogen sulphide can be destroyed by a small crystal of sodium sulphite. F. Moldenhauer (Chem. Ztg., 13. 1220, 1889 ; 15. 223, 1891 ; Chem. News, 64. 150, 1891) draws a narrow streak of a 4 per cent, solution of copper sulphate along strips of white filter paper by means of a camel-hair brush. The strips are dried quickly, and preserved in stoppered bottles. When a drop of liquid containing potassium ferrocyanide is placed on the white portion of a strip, if free potassium ferrocyanide soaks into the portion containing copper sulphate a reddish mark is produced. G. C. Stone (Journ Amer. Chem. Soc., 17. 473, 1895) recommends a dilute solution of cobalt nitrate as indicator, and claims that it gives^better results than uranium, copper, and iron salts. Note that commercial "uranium acetate" may be either uranyl acetate or sodium uranyl acetate. 2 There is a danger of under-titrating when the solution is cold, owing to the slow reaction between the ferrocyanide and the zinc. F. Schulz (Chem. Ztg., 33. 1187, 1909) places an open tube (12-15 mrn. diameter) in the liquid to be titrated, and then titrates as usual ; while rotating the beaker, the liquid inside the tube has not been acted upon. Hence, by using the tube as a stirrer, the two liquids mix, and any over-titration is neutralised. The last drops of the standard solution are now added very cautiously. 370 A TREATISE ON CHEMICAL ANALYSIS. volume of solution to be titrated, 180-200 c.c. ; the solution contains 6 c.c. concentrated hydrochloric acid ; 10 grms. ammonium chloride ; and the tempera- ture should be between 70 and 80. A little lead in the solution will not affect the result appreciably. 1 Zinc oxide is soluble, even after ignition, in a mixture of equal volumes of ammonia (sp. gr. 0'924), solution of ammonium carbonate (20 per cent.), and ammonium chloride (20 per cent.). Hence, Tambon 2 determines the zinc oxide in "zinc white," "grey zinc," etc., by digesting 10 grms. of the sample with 300 c.c. of the above solution and shaking the mixture a few minutes. After standing 10 minutes, 3 filter, wash, and dry the insoluble. The difference between this weight and the original sample represents the zinc oxide. Instead of estimating the soluble zinc by difference, it can be determined volumetrically. If soluble zinc salts are present, they must be first removed by washing with warm water before digesting with Tambon's solution. 1 If more hydrochloric acid be employed, the indicator allowance will be larger than 0'5 c.c. and the result will be uncertain. 2 J. Tambon, Bull. Soc. Chim. (4), i. 823, 1907. 3 "Zinc grey" requires 30 minutes' digestion. CHAPTER XXVIII. THE DETERMINATION OF MANGANESE. 184. The Effect of Manganese on Silicate Analyses. FIRECLAYS often contain up to 0*2 per cent, of manganese. This element is usually ignored in clay analyses, and in that case, Hillebrand l has shown that manganese, if present, will be found distributed between the ammonia precipitate, the lime, and the magnesia, even when a double ammonia precipita- tion is made. Much remains in the ammonia precipitate, presumably because the manganese is peroxidised, when it is precipitated in ammoniacal solutions. For instance, Hillebrand found : Table L V. Distribution of Manganese among the Different Constituents of a Clay Analysis. Composition of rock. Manganese found. A1A Fe 2 3 . CaO. MgO. Total. * by diff. CaO. MgO. 9-35 11-84 2-81 0-311 0-036 0-023 0-252 1271 11-98 4-30 0-442 0-088 016 0-338 4-80 50-51 1-04 0700 0-301 0-087 0-312 3-49 3-99 0-92 0-016 0-016 nil nil 1-00 28-04 19-11 0-574 0-032 0-101 0-441 If the manganese is to be determined in a silicate or clay, one of two methods may be adopted. 2 The manganese is either precipitated with the aluminium hydroxide, etc., by peroxidising the manganese as indicated on page 177 ; or the iron and aluminium hydroxides are precipitated by the basic acetate process, redissolved, and reprecipitated by ammonia. The manganese will be found in the combined filtrates. 3 In the former case, the manganese can be determined by the colorimetric process in an aliquot portion of the pyrosulphate fusion. In the latter case, the manganese can be precipitated by ammonium sulphide, or by bromine, and subsequently determined colorimetrically, or gravimetrically as 1 W. F. Hillebrand, Bull. U.S. Geol. Sur., 305. 96, 1907 ; P. de Sorray, Bull Assoc. Chim. Slier. Dist., 27. 671, 1910 ; H. Rose, Pogg. Ann., HO. 292, 1860 ; Chem. News, 2. 266, 1860. 2 Of course, a great number of methods are available. 3 It may be well to bear in mind that laboratory glass generally contains a little manganese, which is dissolved out by alkali K. A. Gortner, Amer. Chem. Journ., 39. 157, 1908. This remark is only applicable when minute amounts of manganese are in question. 372 A TREATISE ON CHEMICAL ANALYSIS. phosphate. The basic acetate process is generally employed for the separation of alumina and iron from manganese, zinc, cobalt, and nickel. 185. The Precipitation of Manganese by the Bromine Process. Add 15 c.c. of sodium acetate to the filtrates from, say, the basic acetate separation, and then add 2-3 c.c. of liquid bromine. 1 Heat the solution to boiling. If the solution, on standing, has a yellow colour, sufficient bromine is present; if not, add more bromine. Filter the precipitate through a close- grained filter paper. Add more bromine to the nitrate, and boil again. If more manganese is thrown down, filter, and repeat the process until no manganese is precipitated. Wash the precipitate. 2 If the later washings carry any manganese, re-filter. The mixed precipitates are now ignited and weighed as Mn 3 4 . 3 This weight multiplied by 0'93007 gives the corresponding amount of MnO. If the crucible be surrounded by oxidising gases during the ignition, the composition of the precipitate will not deviate appreciably from Mn 3 4 ; but if reducing gases be present, this formula will not represent the composition of the ignited precipi- tate (page 184). MnO is formed if the manganese compound be calcined at a red heat in a Rose's crucible (fig. 138) in a vigorous current of hydrogen; and the higher oxides form Mn 3 4 if ignited in a current of carbon dioxide ; and into the sesquioxide, Mn 2 3 , by ignition in a current of oxygen. 4 The MnO formed as just described is said to be a "convenient and accurate form in which to weigh manganese." The chief disadvantages of the bromine precipitation are : (1) the contamina- tion of the precipitate with alkalies ; and (2) uncertainty in the composition of the ignited oxide. A reprecipitation by bromine and ammonia will generally free the precipitate from foreign contaminations. Instead of weighing the precipitate as Mn 3 4 , many prefer to dissolve the moist precipitate in dilute acid and determine the manganese volumetrically ; colorimetrically ; or gravi- metrically by the phosphate process. Other oxidising agents ammonium, potassium, or sodium 5 persulphate, 1 Liquid bromine is recommended because it keeps down the volume of the solution. Some add bromine water, that is, water saturated with bromine. The latter is made by keeping water in a bottle with an excess of liquid bromine. Water at dissolves 4 "2 grms. of bromine per 100 c.c. ; at 5, 37 grms. ; at 10, 3 '4 grms. ; at 15, 3 '25 grms. ; and at 20, 2 '2 grms. L. L. de Koninck(^e^. anal. Chem., 18. 468, 1880 ; Chem. News, 43. 34, 1881) recommends a saturated solution of bromine in a 10 per cent, aqueous solution of potassium bromide. Others recommend a saturated solution of bromine in concentrated hydrochloric acid. The latter solution is strongly acid ; Koninck's solution is neutral. For the presence of bromoform in commercial bromine, see S. Reymann, Ber., 8. 792, 1877. 2 It is very difficult to wash the precipitate free from the alkalies carried down by the manganese oxide. Ammonium acetate in place of sodium acetate in the basic acetate separation helps a little, but the separation of manganese is not so good. A. G. M'Kenna, Tech. Quart., 3. 333, 1890 ; Chem. News, 63. 184, 1891 ; V. Eggertz, ib. t 18. 232, 1868 ; ib., 43. 226, 1881 ; 0. Reinhardt, Chem. Ztg., 10. 323, 357, 372, 1896; C. Holthof, Zeit. anal. Chem., 23. 491, 1884 ; F. Kessler, ib., 18. 1, 1879 ; N. Wolff, ib., 22. 520, 1883 ; P. Waage, ib., JO. 206, 1871 ; H. Kammerer, Ber., 4. 218, 1865. 3 S. U. Pickering, Chem. News, 43. 225, 1881 ; C. R. Wright and A. P. Luff, Journ. Chem. Soc., 33. 525, 1878 ; W. Dittmar, ib., 17. 294, 1864 ; J. and H. S. Pattinson, Journ. Soc Chem. Ind., 10. 333, 1891 ; E. H. Saniter, ib., 13. 112, 1894 ; H. D. Richmond, Analyst, 19. 99, 1894 ; F. A. Gooch and M. Austin, Zeit. anorg. Chem., 17. 268, 1898 ; R. Schneider, Pogg. Ann., 107. 605, 1869 ; C. Meineke, Zeit. angew. Chem.. i. 3, 1888 ; W. C. Heraeus and W. Geibel, ib., 20. 1892, 1907 ; A. Gorgeu, Compt. Rend., 106. 743, 1888 ; W. W. Randall, Amer. Chem. Journ., 19. 682, 1897 ; St C. Deville, Compt. Rend., 56. 977, 1863 ; Chem. News, 7. 294, 1863. 4 P. N. Raikowand P. Tischkoff (Chem. Ztg., 35. 1013, 1911). 5 L Dede, Chem. Ztg., 35. 1077, 1911. THE DETERMINATION OF MANGANESE. 373 hydrogen peroxide, etc. are often used instead of bromine for precipitating manganese as hydrated peroxide. 1 186. The Precipitation of Manganese by Ammonium Sulphide. Ammonium sulphide alone precipitates manganese, zinc, nickel, and cobalt sulphides very imperfectly ; but if ammonium chloride be present, precipitation is practically complete. Fresenius 2 says that these elements can be precipitated from solution containing the equivalent of 400 * 000 of manganese oxide, and 80U 1 OUO of nickel or cobalt oxides in presence of an excess of ammonium chloride. 3 If free ammonia be present, the precipitation of manganese and nickel sulphides is retarded, and a certain amount of these elements remains in solution. There are at least two varieties of manganese sulphide. The one is pink or flesh-coloured and colloidal, and readily passes through the filter paper. Nor does the pink variety settle readily from the solution. The other sulphide is green in colour, coarse grained, crystalline, and settles quickly. The latter can be easily filtered and washed. The pink variety, in presence of an excess of ammonium sulphide, 4 soon passes into the green variety when heated, but the change is retarded by the presence of ammonium chloride. Hence, ammonium chloride facilitates the separation of manganese sulphide, but hinders the trans- formation of the pink into the green sulphide. The pink modification is the first product of the reaction, and this is transformed into the green modification later on. The analyst must therefore employ methods which ensure a complete conversion of the manganese into sulphide, and the transformation of the pink into the green sulphide. The best conditions are: (1) a large excess of ammonium chloride ; and (2) a large excess of ammonium sulphide in hot solutions. The ammonium sulphide must be free from yellow polysulphide, since manganese sulphide is slightly soluble in ammonium polysulphide. 5 Suppose, then, that the manganese is to be precipitated by ammonium sulphide. Evaporate the filtrate, as indicated above, to about 200 c.c. Pour the solution into an Erlenmeyer's flask; add 10 c.c. of ammonium chloride solution, and ammonia until the solution is alkaline. Pass hydrogen sulphide through the boiling solution 6 for about 10 minutes. Cork the flask and let it stand for about 24 hours in a warm place. 7 This procedure gives a granular 1 M. Dittrich and C. Hassel, Ber., 36. 284, 1423, 1903 ; Zeit. anal. Chem., 43. 382, 1904 ; M. Dittrich, Ber., 35. 4072, 1902 ; M. E. Pozzi-Escot, Ann. Chim.AnaL, 7. 376, 1902 ; H. Baubigny, Compt. Rend., 135. 965, 1110, 1902 ; 136. 449, 1325, 1903 ; J. von Knorre, Zeit. angew. Chem., 14. 1149, 1901 ; 16. 905, 1903 ; Zeit. anal. Chem., 43. 1, 1904 : 44. 88. 1905 ; E. Donath, ib., 44. 698, 1905; H. Liidert, Zeit. angew. Chem., 17. 422, 1904 ; H. P. Smith, Chem. News, 90. 237, 1904 ; H. Rubricus, Stahl Eisen, 25. 890, 1905. 2 R. Fresenius, Journ. praht. Chem (1), 82. 265, 1861 ; Zeit. anal. Chem., II. 419, 1872; H. Raab and L. Wessely, ib.,^2. 433, 1903; A. Classen, ib., 16. 319, 1877; 8. 370, 1869; C. Meirieke, Zeit. angew. Chem., I. 3, 1888 ; A. Volker, Liebig's Ann., 59. 38, 1846 ; P. de Cleremont and H. Guiot, Bull, Soc. Chim. (2), 27. 353, 1877 ; J. C. Olsen and W. S. Rapalje, Journ. Amer. Chem. Soc., 26. 1615, 1904 ; J. C Olsen, E. S. Clowes, and W. 0. Weidmann, ib., 26. 1622, 1904 ; F. Muck, Zeit. Chem. (2), 5. 580, 1869 ; (2), 6. 6, 1870 ; H. How, Chem. News, 19. 137, 1870 ; W. Bottger, Ber., 33. 1019, 1900 ; Chem. News, 82. 247, 1900. a And 80U) 1 000 of zinc oxide, if zinc be present. 4 Not with sodium or potassium sulphides (Muck, I.e.}. 5 For the retarding action of salts of organic acids on the precipitation of manganese sulphide, see page 363"; and for the retarding action of ammonium salts, see H. Rose, Chem. Neivs, 2. 302, 1860. 6 E. Murmann (Monats. Chem., 19. 404, 1898) adds a little mercuric chloride before passing the hydrogen sulphide through the hot solution. The precipitate of the green sulphide so gbtained is easily filtered and washed. The mercury, of course, volatilises as soon as the precipitate is ignited. 7 If the solution happens to contain much lime, as sometimes occurs in the analysis of blast- furnace slags, the prolonged standing leads to the formation of crystals, probably calcium thiosulphate (L. Blum, Zeit. anal. Chem., 28. 454,1889). In that case, it is better to follow 374 A TREATISE ON CHEMICAL ANALYSIS. precipitate. 1 Collect the precipitate 2 on a small filter paper say 7 '5 cm. diameter. Wash with water containing a little colourless ammonium sulphide. 3 Dissolve the precipitate in a little dilute sulphuric acid if the manganese is to be determined colorimetrically, or in hydrochloric acid if the manganese be determined gravimetrically as phosphate. 187. The Gravimetric Determination of Manganese Gibb's Phosphate Process. Manganese is separated from the alkaline earths by precipitation with bromine, or ammonium sulphide, etc. The precipitate is dissolved in hydro- chloric acid (1:1), and alkaline phosphate is added to the solution. A pre- cipitate of ammonium manganese phosphate is obtained similar to the precipitate obtained with magnesium. 4 This is converted into the pyrophosphate by ignition and then weighed. First Precipitation. The cold solution of the manganese oxide in hydro- chloric acid (1 : 1) is supposed to contain no more manganese than is represented by O'l grm. MnO per 100 c.c. of solution. Add an excess, say 5 c.c., of a cold saturated solution of microcosmic salt 5 with constant stirring. Then add a slight excess of dilute ammonia. Heat the mixture until the precipitate becomes crystalline, and let the whole stand for about 1J hours, till cold. Filter and wash with cold, slightly ammoniacal water, dry the precipitate at a gentle heat, and ignite as described below. If the manganese salt is associated with other salts in solution, a second precipitation may be made. A large excess of microcosmic salt is necessary in order to render the precipitate insoluble, especially in the presence of ammonium salts. 6 H. Rose (Ausfuhrliches Hand/buck der analytischen Chemie, Braunschweig, I. 167, 1864) and boil the solution while adding ammonium sulphide again and again. Filter at once. No calcium thiosulphate is then formed. 1 A black residue indicate* that manganese, zinc, nickel, cobalt, copper, or platinum may be present. With clays, however, there is little, chance of this, but platinum may be derived from the platinum crucible during the pyrosulphate fusions. For the separation when zinc, etc. , is present, see page 364. 2 If some sulphide sticks to the walls of the flask or beaker, wash the vessel with dilute nitric acid. The resulting solution is either added to that obtained by dissolving the precipitate, or it is evaporated to dryness in a weighed crucible, calcined to Mn 3 4 , and the result added to the weight of the main precipitate. It is important to test the filtrate for manganese, and to test the precipitate for silica, barium, etc., if only one precipitation followed by roasting to Mn ? 4 be made J. and H. S. Pattinson, Chem. News, Si. 193, 1900. 3 L. Blum (Zeit. anal. Chem., 44. 7, 1905) says there is frequently a slight oxidation of sulphide to sulphate, and in consequence traces of barium and strontium sulphates, if present, may be precipitated. 4 W. Gibbs, Chem. News, 17. 195, 1868 ; Amer. J. Science (2), 44. 216, 1867 ; Zeit. anal. Chem., 7. 101, 1868; R. Fresenius, ib., II. 415, 1872; F. Kessler, ib., 18. 8, 1879; H. D. Dakin, ib., 39. 784, 1890 ; Chem. News, 83. 37, 1900 ; W. Bottger, ib., 82. 101, 1900 ; er., 33. 1019, 1900 ; A. G. M'Kenna, Tech. Quart., 3. 333, 1890 ; Chem. News, 63. 184, 1891 ; T. Moore, ib., 63. 66, 1891 ; G. L. Norris, Journ. Soc. Chem. 2nd., 20. 551, 1901 ; E. H. Saniter, ib., 13. 112, 1894 ; R. C. Boyd, School Mines Quart., n. 355, 1890 ; L. Riirup, Chem. Ztg. t 2O. 285, 337, 1896 ; A. Ledebur, ib., 8. 910, 927, 963, 1884 ; F. A. Gooch and M. Austin, Zeit. anorg. Chem., 18. 339, 1898 ; Amer. J. Science (4), 6. 233, 1898 ; Chem. News, 78. 239, 246, 1898 ; C. E. Munroe, Amer. Chem., 7. 287, 1877. 5 AMMONIUM SODIUM PHOSPHATE SOLUTION. A saturated solution has nearly 170 grms. per litre. 1'5 grms. suffice for O'l grm. of MnO. Hence 9 c.c. will be required per O'l grm. MnO. 6 The proportion of ammonium chloride to the ammonium manganese phosphate should be as 50 : 1. A very large excess of ammonium chloride may be added without any perceptible solvent action. R. Fresenius (I.e.) says that 1 part of the precipitate dissolves in 32,092 parts of cold water, in 20,122 parts of boiling water, and in 1775 parts of a solution of ammonium chloride (1 : 70). This latter statement, however, does not hold good when an excess of the phosphate used for precipitating is present. THE DETERMINATION OF MANGANESE. 375 Second Precipitation. Redissolve the precipitate in an excess of sulphuric or hydrochloric acid. Heat the solution to boiling and add a slight excess of ammonia and microcosmic salt solution. The flocculent white gelatinous pre- cipitate of manganese ammonium phosphate so produced passes into a flesh- coloured crystalline precipitate NH 4 MnP0 4 when the solution is boiled, or allowed to stand for some time. 1 The precipitation is best performed in a platinum vessel; but a glass vessel is quite satisfactory. When cold, filter either through an asbestos-packed Gooch's crucible or through filter paper, and wash with cold water or, better, very dilute ammonia until the wash- water acidified with nitric acid gives no turbidity with a drop of silver nitrate solution. This is important. 2 Add more microcosmic salt to the filtrate. If a precipitate settles after standing several hours, filter it through a small filter paper. Dry the precipitates. Ignition of the Precipitate. If the filtration has been done through filter paper, transfer the dry precipitate to a watch-glass. Ignite the paper separately in a porcelain crucible. Then transfer the precipitate from the watch-glass, and ignite at a red heat. The temperature should be raised very gradually, in order to prevent any solid from being carried away with the ammoniacal vapours. The ignited precipitate should be white or pale pink. Calculation. The weight of the calcined manganese pyrophosphate Mn. 2 P 2 7 multiplied by O5 (or, more exactly, by 0'4998), and divided by the weight of the sample, represents the amount of manganese oxide MnO in the given sample. Errors. If the precipitate be coloured brown, the manganese was not all converted into the phosphate. In that case, redissolve the precipitate in hydrochloric acid (1:1), and repeat the precipitation with more microcosmic salt. If the precipitate is not all soluble in the hydrochloric acid, some silica was probably precipitated with the manganese. In that case, filter the solution, wash, ignite, and weigh the insoluble silica. Deduct the weight of the silica from the weight of the manganese phosphate. The process does not give good results in the presence of zinc, nickel, copper, and other metals which give precipitates of sparingly soluble phosphates in ammoniacal solutions. It is an excellent method for converting precipitates by bromine, etc., into a weighable form, and for separating manganese from the elements which are not liable to precipitation in ammoniacal phosphate solutions. Gooch and Austin found, in using 50 c.c. of a saturated solution of micro- cosmic salt, 20 grms. of ammonium chloride, and a solution diluted to 200 c.c. containing : MnO used . . . 0'0942 0'0942 0'0942 0'1885 0'1885 0'1885 grin. MnO found . . . 0'0951 0'0955 0'0956 G'1888 0'1886 0'1889 grm. Error .... 0'0009 0*0013 0014 Q'0003 O'OOOl Q'0004 grin. The positive error here observed appears to be due to the slight adsorption of the microcosmic salt by the precipitated phosphate. The average error in duplicate determinations did not, therefore, exceed O'OOl grm. when expressed in terms of MnO. 1 A large excess of ammonium chloride favours a rapid transformation; ammonium nitrate is not so good. 2 Ammonium chloride would, of course, be volatilised during the ignition, but a trace of manganese chloride might be formed. There is no danger from this if the washing be con- ducted as described in the text. 376 A TREATISE ON CHEMICAL ANALYSIS. 188. The Volumetric Determination of Manganese Pattinson's Process. A large number of rapid methods have been suggested, by precipitating the manganese as hydrated manganese dioxide possibly MnO(OH) 2 by potassium chlorate ; l zinc oxide and bromine ; 2 bleaching powder and ferric chloride ; 3 etc., and subsequently determining the manganese dioxide by one of the many available volumetric processes. These processes can be made to give accurate results under special conditions, but there is some uncertainty as to the com- position of the precipitate, and in routine work an allowance is frequently made for the deviation in the composition of the precipitate from the assumed MnO(OH) 2 . When ores are only occasionally analysed, the corrections are troublesome. Ledebur, Saniter, Riirup, and others have examined Volhard's, Pattinson's, and other methods. Pattinson's gives excellent results. This process is based on the fact that the whole of the manganese in a solution of manganese chloride can be precipitated as manganese dioxide in the presence of ferric or zinc chloride by an excess of an aqueous solution of calcium hypochlorite, or bromine water. Dissolution of the Mineral. Digest 10 grms. of the finely powdered and dried (110) mineral in 100 c.c. of concentrated hydrochloric acid. Add 5 c.c. of nitric acid, and evaporate down to a small volume. 4 Transfer the solution to a 100-c.c. flask and make up to the mark- with water. Conversion of the Manganese Chloride into Manganese Peroxide. Agitate the contents of the flask, and pipette 20 c.c. into a litre beaker or Erlenmeyer's flask. 5 Add granular precipitated calcium carbonate, in small quantities at a time, until the free acid is neutralised, and the solution, though clear, has a reddish-brown tinge. Then add 30 c.c. of a solution of zinc chloride, 6 and 50 c.c. of a solution 7 of " chloride of lime " (or 25 c.c. of a saturated aqueous solution of bromine). Add more calcium carbonate, with constant stirring, until the latter remains undissolved. Now stir in 700 c.c. of hot water at about 70. The supernatant liquid should be colourless. If it be pink (calcium permanganate), add 2 c.c. methyl alcohol and boil ; if it still be pink, repeat 1 W. Hampe and M. Ukena, Zeit. anal. Chem., 24. 431, 1885 ; 32. 369, 1893 ; A. P. Ford, Trans. Amer. Inst. Min. Eng., 9. 397, 1880; F. Williams, ib., 10. 100, 1881; R. Boiling, Journ. Amer. Chem. Soc. t 23. 493, 1901. 2 A. H. Low, Journ. Anal. App. Chem., 6. 663, 1892 ; Chem. Neivs, 67. 162, 1893. 3 J. Pattinson, Journ. Chem. Sac., 35. 365, 1879; Journ. Soc. Chem. Ind., 5. 422, 1886 ; J. and H. S. Pattinson, ib., 10. 333, 1891 ; R. W. Atkinson, ib., 5. 365, 1886 ; E H. Saniter, ib., 13. 112, 1894; J. Pattinson, Chem. News, 21. 267, 1870 ; 41. 179, 1880 ; G. Lunge, ib., 41. 78, 120, 141, 179, 181, 1880 ; W. Weldon, 41. 207, 1880 ; C. R. Wright, Journ. Chem. Soc., 37. 22, 49, 1880; A. Ledebur, Chem. Ztg., 8. 910, 927, 963, 1884; F. Jean, Bull. Soc. Chim., (3), 9. 248, 1895. 4 The insoluble residue may be filtered off, ignited, and fused with sodium carbonate, and if it be coloured greenish blue, take up the melt with water and dilute hydrochloric acid and add it to the main solution. 5 Add sufficient ferric chloride to make the amount of ferric and manganese chlorides present in the solution approximately equal. This ensures the precipitation of manganese dioxide instead of some lower oxide. 6 Containing the equivalent of half a gram of metallic zinc per 30 c.c. If ferric chloride be used in place of zinc chloride, an error amounting to 0'4-0'5 per cent, of manganese may be introduced into the determination owing to the presence of manganese in this salt. To correct, the amount of manganese in the ferric chloride can be determined by boiling the solution with ammonia and a little hydrogen peroxide, washing the precipitate in nitric acid and a little sulphurous acid, and determining the manganese by the colorimetric process. 7 Mix 15 grms. of fresh bleaching powder with 100 c.c. of water, and, after standing some time to settle, decant the clear. THE DETERMINATION OF MANGANESE. 377 the treatment with methyl alcohol. Let the precipitate settle. Decant the clear supernatant liquid through an asbestos-packed Gooch's crucible. 1 Wash four times by decantation with 300 c.c. of hot water (70). Transfer the precipitate to the Gooch's crucible without attempting to remove the last portions of the precipitate from the sides of the beaker. Wash the precipitate with hot (70) water until the nitrate gives no blue coloration with starch paper. The Titration. The precipitate and the asbestos are placed in the original beaker. Dissolve the precipitate in dilute sulphuric acid (1:1). When the brown colour has disappeared, 2 add 50 c.c. of a freshly standardised solution of ferrous sulphate 3 acidified with sulphuric acid. Mix the solution thoroughly, and then titrate the excess of ferrous sulphate 4 with y^N-potassium dichromate solution. 5 Calculations. Ten grms. of the ore were dissolved in 100 c.c. of solution, and 20 c.c., equivalent to 2 grms. of the ore, were taken. 50 c.c. or 3-5 grms. of ferrous ammonium sulphate solution, representing 89' 24 c.c. of the y^N-dichromate, were added ; but only 38 '4 c.c. of the dichromate solution were required in the titration. Hence, 2 grms. of the ore represented 50 '84 c.c. of the dichromate solution. But 1 c.c. of y^N-dichromate solution represents 0'004347 grm. of Mn0 (or 0'003547 grm. MnO) ; hence, 50'84 c.c. of the dichromate solution represents 0*221 grm. Mn0 2 (or 0*180 grm. MnO). That is, 2 grms. of the ore has the equivalent of 0-221 grm. Mn0 2 (or 0'180 grm. MnO); that is, O'll per cent, of Mn0 2 (or 0*09 per cent. MnO). Errors. This method will give results within 0*1 per cent, of the true percentage of manganese in a given sample. Lead, copper, nickel, cobalt, and chromium lead to high results. The interference of up to 1 per cent, of lead, copper, and nickel is not serious, but cobalt and chromium spoil the results. Higher oxides of these elements are probably precipitated with the manganese dioxide, and, later on, oxidise the ferrous sulphate. For instance, 100 parts of Mn 3 4 , with 1 part of the elements indicated, gave the following results : Lead. Copper. Nickel. Cobalt. Chromium. 100-23 100-23 100*23 10064 100'60 percent. In the absence of these disturbing elements, 1000'1 per cent, would have been obtained. 189. The Volumetric Determination of Manganese Volhard's Process. Guyard has shown that if a feebly acid or neutral solution of manganese sulphate or nitrate be treated with a solution of potassium permanganate, a 1 The asbestos should be tested by a blank experiment with ferrous sulphate to make sure that it contains nothing which will reduce the manganese dioxide. G. W. Sargent and J. K. Faust (Journ. Amer. Chem. Soc., 21. 287, 1898) use a filter tube packed first with glass-wool, then with sand, and finally with asbestos. 2 If the ore contains organic matter, this must be filtered off before attempting to oxidise the ferrous salt, since organic matter will reduce the ferric salts and give a high result. Or the solution may be filtered through a Gooch's crucible and the subsequent titrations made with permanganate instead of the dichromate, as indicated in the text. 3 Containing the equivalent of about 10 grms. of metallic iron per litre ; i.e. 70 grms. of ferrous ammonium sulphate per litre. 4 A. Terrell, Bull. Soc. Chim. (2), 35. 551, 1881. 5 In any case the dichromate or permanganate used for the titration should be standardised with zinc or ferric chloride equal to the amount used in the determination. 6 A. Guyard, Chem. News, 8. 292, 1863; Zeit. anal. C/iem., 3. 373, 1864; Bull. Soc. Chim. (2), 6. 88, 1863 ; F. Jean, ib. (3), 9. 248, 1893. 378 A TREATISE ON CHEMICAL ANALYSIS. compact dark brown precipitate of a manganic acid is formed, which is sometimes said to be H 2 Mn0 2 . The permanganate is decolorised as long as any manganese salt remains in solution ; any further addition of the permanganate produces a pink coloration. In reality, the manganic acid MnO(OH) 2 is not precipitated, but rather a manganese manganite whose composition varies with the conditions of the experiment. It is accordingly difficult to get uniformly good results with Guyard's process. The reaction has been the subject of many investiga- tions. 1 Volhard showed that the process is more under control, for analytical purposes, if a strongly basic oxide be present in the solution. Mercury, calcium, magnesium, barium, and zinc salts may be used, but the latter appears to be most suitable. Meineke considered that a considerable amount of zinc sulphate 25-30 grms. is needed for the purpose. The titration can be conducted with a greater degree of precision in the presence of some zinc sulphate, and the consumption of permanganate then corresponds with : 3MnS0 4 + 2KMn0 4 + 2H 2 = 5Mn0 2 + K 2 S0 4 + 2H 2 S0 4 . Ferrous salts should be absent, since they are transformed by the permanganate into ferric salts. Bromine, hydrogen peroxide, etc., can be used to oxidise the iron, but the excess of the oxidising agent must be removed by boiling. If ferric salts be present, they can be precipitated by the addition of zinc oxide, or by sodium bicarbonate, as in Sarnstrom's method. It is somewhat difficult to see when the reaction is complete, because the manganic oxide suspended in the liquid masks the rose colour of the permanganate. The precipitate, however, coagulates on warming, and then settles quickly. Volhard's process, more or less modified, is as follows : Dissolution of the Manganese. Dissolve a gram of the sample in a porcelain basin with a suitable acid, say, 10 c.c. of concentrated hydrochloric acid, 2 assisted, if necessary, by nitric acid, 3 particularly if sulphides be present, and ferrous iron is to be oxidised. Add 10 c.c. of concentrated sulphuric acid, and heat the mixture until the sulphuric acid fumes copiously. 4 When cold, add 25 c.c. of 1 J. Volhard, Liebig's Ann., 198. 218, 1879; Chetn. News, 40. 207, 1879; F. W. Daw, ib., 79. 25, 1899 ; T. Morawski and J. Stingl, ib., 38. 297, 1878 ; Journ. praU. Chem. (2), 18. 96, 1878 ; A. Ghilian, Rev. Univ. Mines (3), 3. 270, 1888 ; Chem. News, 59. 121, 1889 ; E. Donath, ib., 43. 253, 1881 ; F. W. Daw, ib., 79. 25, 58, 1899 ; H. Brearley, ib., 79. 47, 83, 1899 ; L. Riirup, Chem. Ztg., 15. 149, 1891 ; A. Ledebur, ib. t 12. 927, 1888 ; 8. 829, 1884 ; W. Hampe, ib., 7. 1104, 1883 ; E. Dliss, ib., 34. 237, 1910 ; E Donath, ib., 34. 437, 1910 ; A. Kaysser, ib., 34. 1225, 1910 ; R. Schoffel and E. Donath, Oester. Zeit. Berg. Hutt., 31. 229 ; 1883; C. G. Sarnstrom, Berg. Hiitt. Ztg., 40. 425, 1881; Zeit. anal. Chem., 22. 84, 1883; L. Blum, ib., 30. 210, 1891; M. Orthey, ib.. 47. 547, 1908; L. Karaoglanoff, ib., 49. 419 1910; C. Winkler. ib., 3. 423, 1864; R. Habich, ib.,3. 474, 1864; W. M. Fischer, ib., 48. 751, 1909; N. Wolff, ib., 43. 564, 1904; Stahl Eisen, 4. 702, 1884; C. Reinhardt, ib., 5. 782, 1885 ; 6. 150, 1886 ; C. Meineke, Zeit. anal. Chem., 24. 423. 1885 ; Rep. anal. Chem., 3. 337, 1883; 5. 1, 1885; Zeit. angew. Chem., i. 228, 1888; E.' W. Mayer, ib., 2O. 1980, 1907 ; A. Longiand S. Camilla, Gaz. Chim. Ital., 27. 87, 1897 ; G. C. Stone, Journ. Amer. Chem. Soc., 18. 228, 1896 ; W. S. Thomas, ib., 17. 341, 1895; W. A. Noyes, ib., 24. 243, 1902 ; G. Auchy, ib., 17. 943, 1895 ; 18. 498, 1896 ; C. T. Mixer and H. W. du Bois, ib iS 385, 1896 ; E. Cahen and H. F. V. Little, Analyst, 36. 52, 1911 ; W. Heike, Stahl Eisen. 29. 1921, 1909 ; E. Miillerand P. Koppe, Zeit. anorg. Chem., 68. 160, 1910 ; P. Slawik, Chem. Ztg., 36. 106, 1912 ; L. Karaoglanoff, Jahrb. Univ. Sofia, 33, 1911. 2 Chlorides, over 0'5 grm. per litre, should be absent, or the results will be high. If chlorides be present, the precipitate may have a reddish colour; if absent, dark brown. Evaporation with sulphuric acid until the acid fumes copiously will drive off the combined chlorine. 3 If the manganese is to be determined in the insoluble residue, fuse the insoluble matter with sodium carbonate, dissolve the resulting mass in acid, and add the result to the main solution. 4 Organic matter should be absent. It can be destroyed by calcination, or evaporation of the solution to dryness with sulphuric acid, or with nitric acid followed by calcination. THE DETERMINATION OF MANGANESE. 379 water, and boil the solution a short time to dissolve the ferric sulphate. Transfer the mixture to a 500-c.c. flask. Precipitation of Iron. Nearly neutralise the solution with sodium carbonate, and, if iron be present, add gradually an emulsion of zinc oxide l in slight excess. Shake the mixture well after each addition, and avoid a large excess of the zinc oxide. Fill about three-fourths of the flask with water. Agitate the contents of the flask, 2 and let the mixture stand to allow the ferric oxide to settle. If the solution be not colourless, more zinc oxide is probably needed. Make the solution up to the mark with distilled water. Agitate the solution, and let the matter in suspension settle. The Titration. Pipette 100 c.c. of the clear supernatant liquid into a 250-c.c. flask. 3 Heat the solution, and titrate, while hot, with a standard solution of potassium permanganate. The permanganate produces a precipitate which discolours the liquid, and it is necessary to titrate cautiously by agitating the flask after each addition, and allowing the precipitate to settle sufficiently after each addition to show whether or not the liquid is coloured pink. 4 The colour is best observed by holding the flask against a white background, and observing whether or not the upper edges of the liquid are coloured pink. Warm the solution to be titrated, but the liquid must not be boiled. 5 If the pink colour fades, add more permanganate. When the pink colour is permanent, take a final reading of the burette. Calculations. The permanganate is prepared by the method employed on page 193. On comparing the equation on page 198 for ferrous salts with that indicated above, page 378, 6 it will be observed that, in the case of iron, 2KMn0 4 represent 5Fe 2 3 ; and here, 2KMn0 4 represent 3MnO. Consequently, 3MnO correspond with 5Fe 3 . Hence, 1 grm. of Fe. 7 3 corresponds with 0*2665 grm. of MnO. 7 Errors. A comparison of the results with the three processes here indicated on a commercial sample of " manganese oxide " gave : 1 ZINC OXIDE EMULSION. A mixture of pure zinc oxide and water will generally do the work "generally," because some samples of commercial zinc oxide are not effective in separating iron from manganese, possibly owing to the crystalline structure of the powder. A better emulsion is made by dissolving zinc chloride in water ; or by dissolving zinc oxide in hydrochloric acid, heating the mixture with a little bromine, and filtering off the excess of zinc oxide. Precipitate zinc hydroxide from the solution by the addition of ammonia. Do not add an excess of ammonia, or the zinc hydroxide will dissolve. Wash the precipitate several times by decantation with hot water, and wash the oxide into a bottle, which is stoppered and preserved. Shake the mixture well before use. F. A. Emmerton, Trans. Amer. Inst. Min. Eng., 10. 201, 1881. A. Guyard (Compt. Rend., 97. 673, 1883 ; Chem. News, 48. 193, 1883) reports the presence of manganese in zinc oxide. L. L. de Koninck tests the suitability of the zinc oxide for the determination by triturating 3 gnus, with 30 c.c. of water containing 1 grm. of iron alum in solution. The mixture is agitated with sufficient 6N-sulphuric acid to dissolve all the zinc oxide. Avoid a large excess of acid. One drop of the permanganate solution should give a permanent pink coloration. If not, metallic zinc or zinc sulphide may be present. 2 Many here recommend the addition of a couple of drops of nitric acid. 3 The precipitate in the flask may appear bulky, but, as a matter of fact, it occupies very little volume. See page 73 for a discussion on the volume of suspended precipitates. The error is here negligible. 4 It sometimes saves time to take two aliquot portions. Titrate one by adding 1 c.c. of the permanganate at a time, and in the other the titration can be carried to a greater precision without an inordinate expenditure of time. 5 Owing to the well-known instability of the permanganate in the presence of the solid manganese oxide (page 196). 6 See J. W. Mellor, Modern Inorganic Chemistry, London, 480, 1912. 7 Some workers deduct 0'2 c.c. from the volume of permanganate employed in the titration before the calculation is made, in order to allow for the presence of the two drops of nitric acid which is supposed to facilitate settling and to counteract the effect of traces of organic matter. 380 A TREATISE ON CHEMICAL ANALYSIS. Gibb's phosphate process . . . 74 '84 74*85 74 '80 per cent. MnO. Volhard's process . . . . 74 '68 74 '66 74 '56 per cent. MnO. Pattinson's process .... 74 '71 7470 7474 per cent. MnO. The values by the phosphate process are probably O'l per cent. high. Volhard's method is inclined to give too low values when the permanganate is standardised against iron or sodium oxalate. The permanganate should be standardised against manganese sulphate of known strength. 1 Gorgeu and Carnot 2 appear to think that the low values arise from the formation of a manganous manganite Mn0.5Mn0 2 by the manganese in the solution, and this retards the further action of the permanganate on the manganous oxide under investigation. When zinc sulphate is present, zinc manganite ZnO. 5Mn0 2 is formed, and this leaves only part of the manganese in solution to react with the permanganate. Bemmeleri 3 considers that the precipitated manganic acid adsorbs part of the manganese salts in the solution, and so removes a little manganese from the "sphere of action" of the permanganate. He further assumes that when a zinc salt is present the zinc salt is adsorbed instead of the manganese salt. The chief errors arise from the presence of organic matter ; the addition of too much zinc oxide ; and standardising the permanganate against iron instead of against manganese reduced from permanganate. Cobalt and chromium interfere with the process. The method gives good results with compounds rich in manganese, although some object to the process, saying it gives erratic results. This criticism is too severe ; when peculiarities of the method are understood, it is all right, and some analysts have said that they consider it to be "the simplest, quickest, and most accurate process for the volumetric determination of manganese." Fischer's Modification of Volhard's Process. This process, as recommended by Cahen and Little, is as follows : The solution of the manganese salt in hydro- chloric or sulphuric acid is neutralised with caustic soda until a slight precipitate persists on shaking. The precipitate is just redissolved by the addition of a drop or two of dilute sulphuric acid. Add 10 grms. of zinc sulphate and heat the mixture to boiling. Add 1 grm. of freshly ignited zinc oxide, and titrate with permanganate, with frequent heating nearly to boiling, until the permangan- ate is no longer decolorised. Cool the mixture under the tap for a minute or two, add 1-2 c.c. of glacial acetic acid, and thoroughly agitate the solution. The hot (not boiling) liquid is then titrated with permanganate, added a few drops at a time, with vigorous shaking after each addition, until the supernatant liquid retains its pink colour after being well shaken several times. The end point is easily observed, because the precipitate settles very quickly in the presence of acetic acid. A difficulty arises during the titration unless the volume of the per- manganate is known to within 3 or 4 c.c., owing to the slowness with which the finely divided oxide settles in the presence of zinc oxide. The time required for a titration is very long, and the end point is difficult to detect. If the acetic acid be added before any permanganate is added the result is too low, but the titration is rapidly effected, and the result serves as a guide for the titration proper. 1 SeeG. Auchy, Journ. Amer. Chem. Soc., 18. 498, 1896 : Chem. News, 74. 214, 248, 262 1896. 2 A. Gorgeu, Bull. Soc. Chim. (3), 9. 490, 1893; A. Carnot, Compt. Rend., 116. 1375, 1893. 3 J. M. van Bemmelen, Joarn. prakt. Chem. (2), 23. 387, 1888. THE DETERMINATION OF MANGANESE. 381 The results leave little to be desired as far as accuracy is concerned. With two samples of pyrolusite, the following comparative results were obtained : No. 1. No. 2. Volhard-Fischer's process . . . . .4875 50'35 Pattinson's process . . . . .48*72 50 '50 190. The Evaluation of Manganese Dioxide Mohr's Process. By the simultaneous action of sulphuric acid and an excess of ferrous sulphate or oxalic acid, the manganic dioxide is reduced to manganous sulphate, and the ferrous sulphate or oxalic acid is simultaneously oxidised. If known quantities of ferrous sulphate or oxalic acid be employed, the excess, not oxidised, can be determined by titration with standard permanganate. 1 The reaction with oxalic acid is represented : Mn0 2 + H 2 C 2 4 + H 2 S0 4 = MnS0 4 + 2C0 2 + 2H 2 0. The following operations furnish sufficient data to calculate the amount of the manganese dioxide in the given sample. The Determination. Digest 0*4 grm. of finely divided, dry (110) 2 sample in an Erlenmeyer's flask with 75 c.c. of a N-oxalic acid 3 or sodium oxalate solution, and 20 c.c. of sulphuric acid (1 : 4), until the black particles have passed into solu- tion. Add about 200 c.c. of hot water (70), and titrate the warm solution with approximately N-potassium permanganate until a permanent pink blush is suffused throughout the liquid. The titration is repeated as a blank experiment on 50 c.c. of the oxalic acid solution, and the volume of the permanganate corre- sponding with 75 c.c. calculated from the result. Calculation. From the equation representing the action of oxalic acid, or rather sodium oxalate, on potassium permanganate (page 194) and on manganese dioxide, it follows that f of 158*03 grms. of KMn0 4 correspond with 86*93 grms. of Mn0 2 . Hence, 1 grm. of KMn0 4 represents 1*3752 grms. of Mn0 2 . Suppose that 1 c.c. of the permanganate solution has 0*0031606 grm. KMn0 4 per c.c., it follows that 1 c.c. of the permanganate solution will represent 0*0043464 grm. of Mn0 2 . Suppose, in an experiment, KMn0 4 sol. 75 c.c. oxalic solution alone ..... 75 c.c. 75 c.c. oxalic solution with sample .... 9 c.c. 0*4 grm. sample requires . . . . .66 c.c. Hence, 0*4 grm. of sample has the equivalent of 66 x 0*00435 = 0*287 grm. Mn0 2 . Hence, the sample contains J x 0*287 x 1000 = 71*7 per cent, of Mn0 2 . Errors. The process indicated above represents the amount of oxygen "available oxygen" which is given off when the sample is decomposed by sulphuric acid, and since other manganese oxides Mn 2 3 , Mn 3 4 , etc. react in a similar way with the oxalic and sulphuric acids : 1 W. Hempel, Memoire sur I'emploi de Vacide oxalique dans Us dosages A liqueurs titrees, Lausanne, 1853 ; F. Mohr, Zeit. anal. Chem., 8. 314, 1869. R. Fresenius and H. Will (Neue Verfahrungsiveise zur Priifung der Potasche, etc., sowie des Braunstein, Heidelberg, 1843) deter- mined the amount of manganese dioxide from the loss in weight due to the evolution of carbon dioxide; H. Kolbe (Liebig's Ann., 119. 130, 1861) absorbed the carbon dioxide in weighed potash bulbs ; G. Bodlander (Zeit. angew. Chem., 8. 430, 1894) measured the volume of the gas evolved during the action of the acid. 2 The sampling in bulk for moisture requires special attention (page 127). J. E. de Vry (Liebig's Ann., 6l. 248, 1847); R. Fresenius, Dingier s Journ. , 135. '277, 1855. 3 Or 4*7 to 4*8 grms. of crystals of oxalic acid. The normal oxalic acid contains 63 '024 grms. of the crystalline salt H 2 C 2 4 . 2H 2 per litre. 382 A TREATISE ON CHEMICAL ANALYSIS. 2Mn 2 3 + 2H 2 S0 4 + H 9 C 2 4 = 2MnS0 4 + 3H 2 + 2C0 2 , 2Mn 3 4 + 3H 2 S0 4 + H 2 C 2 4 - 3MnS0 4 + 4H 2 + 2C0 2 , it follows that, if these oxides be present, the calculation might indicate more Mn0 2 and 1 less MnO than is really present. Some of the calculated Mn0 2 might be present as Mn 2 8 , Mn 3 4 , etc. If the total manganese be determined as described on pages 376 or 377 Pattinson's orVolhard's process the amount of Mn0 2 multiplied by 0'816, and the product subtracted from the total MnO, will represent the approximate amount of manganese oxide MnO in the sample. If the pyrolusite contains reducing agents ferrous oxide, carbonaceous matter, etc., low results will be obtained, because reactions are set up which reverse that produced by the peroxide. Any ferrous iron which might be present will react with the permanganate and give low results by apparently diminishing the amount of oxalic acid broken down by the manganese dioxide. Carbonates do not interfere. 191. The Colorimetric Determination of Manganese- Walter's Process. The colorimetric determination of manganese is based upon the ease with which manganese solutions are oxidised to pink or violet permanganate. Brunner l fused the sample with alkali while exposed to an Oxidising atmosphere, and deduced the amount of manganese from the intensity of the colour of the solution of the fused cake. The results by this method are not satisfactory. Lead peroxide, 2 sodium bismuthate, 3 and ammonium persulphate 4 are usually employed as oxidising agents. The manganese in the given solution is thus oxidised to pink permanganic acid. The intensity of the coloration depends upon the amount of manganese present. The tint of a test solution so prepared can be compared with the tint of a similar solution containing a known amount of manganese. 5 Something of the order O'OOOOl grm. of manganese in 100 c.c. of solution can be determined by means of this process. Preparation of the Standard Solution. A stock solution of manganese sulphate containing O'l grm. of MnO per litre is prepared. Dissolve 0'2225 grm. of potassium permanganate in water ; acidify with sulphuric acid ; reduce 1 A. Brunner, Dingier' s Journ. , 2IO. 278, 1873. a T. M. Chatard, Amer J. Science (3), i. 416, 1871 ; Chem. News, 24. 196, 1871 ; S. Peters, ib., 33. 35, 1876 ; Dingier 's Journ., 221. 486, 1876 ; P. Picard, Compt. Rend., 75. 1821, 1872 ; A. Lecterc, ib., 75. 1209, 1872; Chem. News, 26. 296, 1872; L. L. de Koninek, Rev. Univ. Mines (3), 5. 308, 1889 ; T. E. Thorpe and F. J. Hambly, Journ. Chem. Soc., 53. 182, 1888 ; A. Ledebur, Berg. Riitt. Ztg., 41. 417, 1882 ; F. C. G. Miiller, Stahl Eisen, 6. 98, 1886 ; F. Osmond, Bull. Soc. Chim. (2), 43. 56, 1885 ; Deshays, Chem. News, 38. 70, 1878. 3 L. Schneider, Dingler's Journ., 269. 224, 1893; L. Dufty, Chem. News, 84. 248, 1901 ; J. Reddrop and H. Ramage, Journ. Chem. Soc., 67. 268, 1895 ; F. Ibbotson and H. Brearley, Chem. News, 82. 269, 1900 ; 84. 247, 302, 1901 ; 85. 58, 1902; H. Ramage, ib., 84. 209, 269, 1901 ; 85. 24, 95, 1902 ; A. A. Blair, Journ. Amer. Chem. Soc , 26. 793, 1904 ; R. S. Weston, ib., 29. 1074, 1907; F. J. Metzger and R. F. M'Cracken. ib., 32. 1250, 1910; W. Blum, ib., 34. 1379, 1912; P. H. M. P. Brinton, Journ. Ind. Eng. Chem., 3. 237, 376, 1911 ; W. F. Hillebrand and W. Blum, ib., 3. 374, 1911 ; D. J. Demorest, ib., 4. 19, 1912 ; J. R. Cain, ib., 3. 360, 1911 ; G. Bertrand, Bull. Soc. Chim. (4), 9. 361, 1911 ; H. Rubricus, Stahl Eisen, 30. 957, 1911 ; R. A. Gortner and C. 0. Rost, Journ. Ind. Eng. Chem., 4. 522, 1912 ; F. J. Metzger and L. E. Marrs, ib., 3. 333, 1911 ; 5. 125, 1913 ; H. F. U. Little, Analyst, 37. 554, 1912. 4 H. E. Walters, Proc. Eng. Soc. West Pa., 17. 257, 1901; Journ. Amer. Chem. Soc., 25. 392, 1903 ; 27. 1550, 1905 ; Chem. News, 84. 239, 1901 ; H. Marshall, ib., 83. 73, 1901 ; M. R. Schmidt, Journ. Amer. Chem. Soc., 32. 965, 1910 ; H. Rubricus, Stahl Eisen, 30. 957, 1911 ; H. Kunze, ib., 32. 1914, 1912 ; P. Holland, Chem. News, 96. 2, 1907 ; J. J. Boyle Journ. Ind. Eng. Chem., 4. 202, 1912 ; M. Stanichitch, Rev. Met., 8. 891, 1911. 5 The amount of permanganate in the test solution can also be verified volumetrically by titration with a standard solution of ferrous sulphate. THE DETERMINATION OF MANGANESE. 383 with sulphurous acid ; and make the solution up to a litre. Pipette sufficient of this solution say 2 c.c. into a 100-c.c. flask. Add 10 c.c. of silver nitrate solution (containing 2 grms. per litre), add 1 grm. ammonium persulphate, and warm on a water bath until a pink colour is developed. By the time the flask has cooled the colour will have acquired its maximum intensity. Make up to the 100-c.c. mark. Pipette from 2 to 10 c.c. into the test glass of the colorimeter. If insufficient silver nitrate has been added, a brown precipitate will be pro- duced in the solution after the addition of the ammonium persulphate. In that case, add more sulphurous acid and more silver nitrate. Reoxidise with ammonium persulphate as before. Preparation of the Test Solution. The solution obtained by dissolving the cake from the pyrosulphate fusion ; or the solution remaining after the colori- rnetric titanium determination ; or the manganese sulphide can be dissolved in dilute sulphuric acid. The products from the pyrosulphate fusion generally contain chlorides, which must be removed before applying the test. Hence, it is usually quickest to assume that chlorides are present, and add a little silver nitrate solution to the boiling solution under investigation. Filter and wash the precipitate. Collect the filtrate and washings in a 200-c.c. flask ; acidify the solution with sulphuric acid ; add 10 c.c. of the silver nitrate solution as indicated above ; and also add ammonium persulphate and warm as indicated for the preparation of the standard solution. When the solution is cold, make it up to the mark with water, and pour a quantity into the test glass of the colorimeter. The Comparison. The standard solution is diluted with water from a burette until an aliquot portion has the same tint as the test solution. The amount of water required for the purpose is measured. See iron, page 200, for further details. Calculations. Suppose that 2 c.c. of the standard solution of manganese sulphate was made up to 100 c.c., and that 5 c.c. of this solution required the addition of 83 c.c. of water to bring it to the same tint as the test solution. 1 c.c. standard has OO001 grm. MnO ; .'. 2 c fc c. has 0-0002 grm.; this diluted to 100 c.c. has 0-000002 grm. MnO per c.c. ; 5 c.c. of this required 83 c.c. water, so that 5 + 83 = 88 c.c. of the solution has O'OOOOl grm. MnO, and 200 c.c. has 200 x 0-00001 OO But the clay was made up to 250 c.c., and 100 c.c. was taken ; hence, v 0* x = 0-00004 grm. ; or 0-004 per cent. MnO. 1 ou With practice, and normal colour vision, differences of tint corresponding with O'OOOOl grm. of MnO can be detected. If the quantity of manganese in the portion of the sample under investigation is less than O'OOl grm., it is well to work with a larger quantity of the sample. The results with quantities of manganese over about 2 per cent, are not so good as by gravimetric or volumetric processes. The presence of a little iron seems to favour the oxidation of the manganous oxide, MnO, to the permanganate. Jervis l shows that the colour is more intense with increasing amounts of manganese up to a certain maximum, and after that 1 H. Jervis, Chem. News, 8z. 171, 1900. 384 A TREATISE ON CHEMICAL ANALYSIS. a decrease in the intensity of the colour occurs with increasing amounts of manganese. Thus : MnS0 4 added .... 2 10 20 30 40 50 c.c. KMn0 4 found with iron . . 3'45 14'8 247 21 "0 1375 8'0 c.c. KMn0 4 found, no iron . . 3'35 16'45 25'85 22'9 18'0 12-15 c.c. Aluminum and molybdenum do not interfere ; copper and nickel exercise no further influence than that due to the colour of their salts. The Effect of Chromium Salts. If chromium be present, the yellow colour of the chromate produced by the oxidising action of the persulphate spoils the tint of the permanganate, and the chromium must be removed before the comparison can be made. Dittrich 1 does this by boiling the discoloured solution with ammonia for a short time, iron and manganese hydroxides are precipitated, and the chromium remains in solution as chromate. Filter and wash. Remove the silver from the nitrate by the addition of common salt ; again filter and wash. Evaporate the nitrate down to between 50 and 100 c.c. and estimate the chromium as indicated on page 473. The manganese precipitate on the filter paper is dis- solved in dilute sulphuric acid mixed with some sulphurous acid or hydrogen peroxide. The manganese is then determined colorimetrically as indicated above. 192. The Analysis of Wads, and Manganese Earths. Manganese peroxide is mainly used in the manufacture of chemicals, in glass- making, and for the manufacture of dry batteries. There are many manganese ores which serve excellently for the manufacture of pottery colours and yet contain little or no peroxide. Mohr's process is not then applicable. These ores are usually sold on a basis of 50 units of manganese 2 one unit means 1 per cent. at so much per unit, with a bonus or penalty per unit respectively above or below 50. A maximum of, say, 8 per cent, silica and 0*2 per cent, phosphorus may be allowed, with a deduction of, say, Jd. per unit of silica above the maximum, and an agreed deduction for each 0'02 per cent, of phosphorus above the agreed maximum. 3 The Caucasian, Indian, and Brazilian ores are fairly constant in composition, and run from about 50 to 55 per cent, manganese, 0'03 per cent, phosphorus, and 10 per cent, silica; the Turkish and Japanese ores run from 43 to 56 per cent, manganese, 0*5 per cent, phosphorus, and 7 to 10 per cent, silica. Japanese "brown stone " may run as much as 87 per cent. Mn0 2 , and such an ore sells for twice as much as 70 per cent, ore. 4 Dissolution of the Solid. Digest 1 grm. of the sample in a 250-c.c. Erlenmeyer's flask with concentrated hydrochloric acid. The attack generally begins in the cold. When the first action is over, gradually warm the flask up to the boiling point. It may be necessary to add a little more acid to com- plete the action. Add water and filter. The insoluble matter may be light- coloured silica, or some dark-coloured mineral not decomposed by the acid. The filter paper and contents should be ignited in a weighed platinum crucible and 1 M. Dittrich, Zeit. anorg. Chem., So. 171, 1913 2 Pure manganous oxide, MnO, runs 77 '5 per cent, manganese (Mn) ; the dioxide, Mn0 2 , 63 '2 per cent. ; the sesquioxide, Mn. 2 3 , 69 '6 per cent. ; and the manganomanganic oxide, MiigO^, 72*0 per cent. 3 E.g., one consumer recently purchased its ore on the following basis (1911): Ores with 50 per cent. Mn to be purchased (delivered) at 13d. per unit per ton ; 46-50 per cent. Mn, 12d. ; 43-46 per cent. Mn, 12d. ; 40-43 per cent., ll^d. For each per cent of silica above the maximum 8 per cent., deduct 7Jd. per ton, and for each - 02 per cent, phosphorus above 0*2 per cent., deduct Id. per unit of manganese per ton. Sample for analysis to be dried at 100 ; the percentage of moisture in the sample to be deducted from the weight. 4 G. T. Holloway, Trans. Inst. Min. Met , 21. 569, 1912. THE DETERMINATION OF MANGANESE. 385 weighed. This may be reported as "sand and insoluble matter," or fused with alkaline carbonate in an oxidising atmosphere. If the mass on cooling is white or pale greenish blue in tinge, it may be rejected. If the fused mass be green, dissolve in hydrochloric acid, evaporate to dryness, and take up with water and hydrochloric acid as indicated on page 164. Add the filtrate from the silica, to the main solution. The silica may be determined in the usual manner. Removal of Alumina and Iron Oxide. Precipitate the joint alumina, ferric oxide, and phosphoric oxide by the sodium acetate process (page 362) from a solution acidified with acetic acid. If much manganese be present, two, three, or four precipitations may be necessary. If the combined filtrates exceed 300 c.c., evaporate to about 200 c.c. If a white precipitate separates during the evapora- tion, it may be neglected ; but if a reddish precipitate separates, it must be filtered off, dissolved in hydrochloric acid, the iron precipitated as basic acetate, and the filtrate added to the main filtrate. The alumina, iron, and phosphoric oxide may be treated by the methods of page 177. Precipitation of Manganese. Add approximately 5 grms. of sodium acetate 15 c.c. of solution and 2 or 3 c.c. of liquid bromine. If a saturated aqueous solution of bromine be used, the solution may become rather bulky. The solu- tion should have a yellow tint, showing that an excess of bromine is present. Filter. Add more bromine to the filtrate and boil again. This ensures the complete precipitation of the manganese. If any precipitate be formed, filter. The precipitated manganese is dissolved in a hot dilute solution of nitric acid to which is added either sulphurous acid or a little sodium bisulphite. The reducing agent facilitates the solution of the precipitate. The manganese may now be determined in the solution by volumetric, colorimetric, or gravimetric processes. If needed, the lime and magnesia can be determined in the combined filtrates. Determination of Carbon. Dissolve 10 to 20 grms. of the ore in concentrated hydrochloric acid as indicated above. Dilute with water. Filter the residue through a Gooch's crucible charged with ignited asbestos, wash, dry at 110, and weigh the crucible and contents. Burn off the carbon, and re-weigh. The loss in weight represents the organic matter. To eliminate the obvious errors, the asbestos may be ignited and the resulting carbon dioxide determined as indicated on page 563. The wet combustion process may also be employed as indicated on page 546. CHAPTER XXIX. THE DETERMINATION OF COBALT AND NICKEL. 193. The Detection of Cobalt and Nickel. THE usual scheme for the qualitative analysis of mixtures leaves finally a precipitate containing the mixed sulphides of cobalt and nickel. There are several distinguishing tests. To get the sulphides into solution, boil the filter paper and contents in a small flask with 10 c.c. of hydrochloric acid (1:4) and 1 c.c. of nitric acid (1 : 3). Filter off the precipitated sulphur and the filter paper. Collect the filtrate in a basin, and evaporate to dryness to expel the excess of acid. Dissolve the residue in two or three drops of hydrochloric acid and 10 c.c. of water. Tests for Cobalt. (1) Ilinsky and Knorre's Test. 1 Add a slight excess of a saturated solution of nitroso-/3-naphthol (page 394) in acetic acid. Agitate the solution. A brick-red precipitate represents cobalt. Confirm as usual. (2) Skey's Test. 2 Add a saturated solution of potassium thiocyanate and shake up the mixture with a mixture of amyl alcohol and ether. The ethereal layer will be blue if 1 c.c. of a 1 : 50,000 aqueous solution be used. (3) Danziger's Test. 3 To about 5 c.c. of the colourless solution, acidified with hydrochloric acid, add a little solid ammonium thio-acetate CH 3 . COSNH 4 a few drops of stannous chloride, 4 and an equal volume of amyl alcohol, or a mixture of acetone and ether, or alcohol and ether. Shake. Let settle. If cobalt be present, the upper layer will be coloured blue. This test will indicate 1 part of cobalt in 500,000 parts of water, and is more delicate than Skey's test. Tests for Nickel. (1) Tschugajeffis Test. 5 If an excess of dimethylglyoxime be added to a strongly ammoniacal solution containing a mixture of cobalt and nickel, and the solution be boiled a short time, a rose coloration or a scarlet-red precipitate will be obtained, according to the amount of nickel present (page 394). This reagent is reported to detect nickel in 1 c.c. of a 1 : 200,000 solution, or in 4 c.c. of a 1 : 500,000 solution. (2) Parr's Test. 6 Freshly precipitated nickel hydroxide liberates iodine from potassium iodide, but the corresponding cobalt compound does not. Hence, 1 M. Ilinsky and G. von Knorre, Ber., 18. 699, 1885. 2 W. Skey, Chem. News, 16. 201, 1867. This is generally, but incorrectly, called " Vogel's test." H. W. Vogel, er., 8. 1533, 1875 ; 12. 2314, 1879 ; F. P. Treadwell, Zeit. anorg. Chem., 26. 108, 1901. See page 397. 3 J. L. Danziger, Journ. Amer. Chem. Soc., 24. 578, 1902. 4 To reduce the ferric salts which give a red colour. The solid ammonium thio-acetate ensures a concentrated solution. s L. Tschugajeff, Ber. , 38. 2520, 1905. For the sensitiveness of different tests for cobalt and nickel, see A. del Campo y Cerdan and J. Ferrer, Aiiales Soc. Espan. Fis. Quim., g. 201, 272, 1911 ; S. R. Benedict, Journ. Amer. Chem. Soc., 27. 1360, 1905. 6 S. W. Parr, Journ. Amer. Chem. Soc., 19. 341, 1897; S. R. Benedict, ib. t 26. 695, 1904. 386 THE DETERMINATION OF COBALT AND NICKEL. 387 add bromine water to the solution containing either cobalt or nickel warm the solution ; add an excess of sodium hydroxide, and boil. Filter. ' Wash the precipitated hydroxide on a filter paper. Pour a hot solution of potassium iodide through the paper. Free iodine in the filtrate te evidence of nickel. The iodine is best detected by shaking up the filtrate with a little benzene. The separation of cobalt and nickel is not usually of any particular importance in silicate analyses, but the methods here indicated represent the type of what is wanted in colour and glaze analyses. The analysis of cobalt oxides is more frequently wanted than nickel. The scheme for the sepcaration of iron, aluminium, titanium, zinc, manganese, nickel, cobalt, magnesia, and lime may be summarised (solids to left, solutions to right) :-. Basic acetate separation (page 362) Fe, Al, Ti (page 177) Zn, Mn, Ni, Co, Mg, Ca (Formic acid and H 2 S, page 364) Zn (page 366) Mn, Ni, Co, Mg, Ca (Ammonium sulphide, page 373) n, Ni, Co, Mg, Ni, Co, Mn Mg, Ca (page 211) (Acetic acid and H 2 S, page 388) I | Mn (page 374) Ni, Co, Potassium nitrite (page 390) Co Ni. 194. The Properties of Cobalt and Nickel Sulphides. Simultaneous Precipitation of Traces of Cobalt and Nickel with the Alumina and Iron. The addition of ammonium chloride prevents, to a great extent, the precipitation of cobalt and nickel by ammonia ; but if insufficient ammonium chloride be present, some nickel and cobalt may be carried down with the aluminium and iron hydroxides. Baumhauer * determined the amount of cobalt and nickel which were precipitated along with the iron when known mixtures of iron, cobalt, and nickel were treated with the regular precipitating agents. He found : Per cent, nickel. Per cent, cobalt. Ammonia (page 177) .... 27 48 Sodium acetate process (page 362) 18 9 Barium carbonate in the cold (page 470) . 8 15 But even when plenty of ammonium chloride is present, appreciable amounts may be precipitated with the iron and alumina as well as with the zinc. If these leakages are not guarded against by double precipitations, 2 the loss of cobalt and 1 E. H. von Baumhauer, Zeit. anal. Chem., IO. 217, 1871. 2 T. Moore, Chem. News, 65. 75, 1892 ; A. Thomas, ib. t 35. 187, 1877 ; T. H. Laby, ib., 89. 280, 1904 ; H. von Jiiptner, Oester. Zeit. Berg. Hiitt., 41. 616, 1894. For clean separations with a large excess of ammonium chloride, see J. T. Dougherty, Chem. News, 95. 261, 268, 1907 ; V. Hassreidter, Zeit. angew. Chem., 22. 1492, 1909. According to F. Ibbotson and H. Brearley (Chem. News, 81. 193, 1900), the nickel is absorbed by the precipitated ferric hydroxide in the same way that filter paper absorbs salts from a solution ; the nickel is absorbed from ammoniacal solutions only because nickel is removed from the precipitate by making the solu- ^88 A TREATISE ON CHEMICAL ANALYSIS. nickel, when precipitated at the end of a complex series of separations, may bf quite serious, In the present case, the iron, aluminium, and titanium are separated by the ammonia or the basic acetate process ; the zinc as zinc sulphide ; and finally the cobalt and nickel are precipitated as sulphides from a solution acidified with acetic acid. Manganese is not precipitated under these conditions. There is need for a few special remarks on the sulphides of nickel and cobalt. Action of Ammonium Poly sulphide on Nickel and Cobalt Salts. These elements are not precipitated by hydrogen sulphide from solutions acidified with the mineral acids, but the sulphides are precipitated in the presence of acetic acid, particularly if ammonium or sodium acetate be present and the solution is warm. Ammonium sulphide precipitates the black or dark brown-coloured sulphides from neutral or ammoniacal solutions of nickel and cobalt. The precipitate is but sparingly soluble in acetic acid, and in very dilute hydrochloric acid. If yellow ammonium sulphide be employed, that is, ammonium poly- sulphide, more or less nickel sulphide passes into solution and the filtrate will be coloured brown. It is generally supposed that either a soluble complex ammonium thionickelate, (NH 4 ) 2 NiS 2+n , or a colloidal nickel persulphide is formed. The greater the excess of ammonium sulphide and the longer the solution is exposed to the air, the greater the tendency of the nickel to pass into solution in this manner. 1 Action of Ammonium Monosulphide on Nickel Salts. In the presence of ammonium hydroxide, colourless ammonium sulphide forms a violet-coloured solution with nickel salts. This soon becomes red, dark brown, and finally a black precipitate separates. On filtration, the filtrate appears colourless ; but if the colourless solution be heated, a black precipitate of nickelous sulphide separates. Free ammonia is not necessary for this action. Colourless ammonium sulphide, free from the polysulphide, precipitates nickel completely as sulphide, and the filtrate, in the absence of air, will be free from nickel and colourless. Ammonium monosulphide rapidly oxidises to polysulphide when exposed to the air, and it is almost impossible to prevent some nickel passing into the filtrate, particularly when working with large quantities of this metal. It is useless to try and get the brown solution clear by filtration. The best method of dealing with the coloured filtrate is to coagulate the nickel sulphide by 5 or 10 minutes' boiling ; re-filtration is then generally successful. These troubles are alleviated by passing hydrogen sulphide into the warm ammoniacal solution, If the solution of nickel be free from foreign salts ammonium chloride, etc. colourless ammonium sulphide forms a brown colloidal solution of the sulphide which passes through the filter paper. The sulphide is then precipitated by adding, say, ammonium chloride. Action of Acids upon Nickel and Cobalt Sulphides. These sulphides, once tion so slightly acid that the ferric hydroxide is scarcely dissolved ; and less nickel is absorbed by increasing the amount of ammonium salt chloride, nitrate, or sulphate in the solution, and decreasing the amount of ammonia. If, therefore, the precipitation of ferric hydroxide be made in presence of ammonium chloride, and so little ammonia that the filtrate is perceptibly acid, the separation of iron and nickel is a "good one." Schwarzberg's method of separation (P. Schwarzberg, Liebig's Ann., 97. 216, 1856 ; J. F. W. Herschel, Ann. Chim. Phys. (3), 49. 306, 1837) depends on the nice adjustment of the "neutralisation" so that the fluid loses its transparency without showing the least trace of a distinct precipitate, and fails to recover its clearness after standing some time ; the solution is boiled, and the iron hydroxide is said to be precipitated comparatively free from nickel (R. Fresenius, Quantitative Chemical Analysis, London, I. 437, 1876). 1 A. Lecrenier, Chem. Ztg., 13. 436, 449, 1889 ; R. Fresenius, Journ. prakt. Chem. (1), 82. 257, 1861 ; A. Villiers, Compt. Rend., 119. 1208, 1263, 1894 ; U. Anthony and G. Magri, Oazz. Chim. Ital., 31. ii., 265, 1901. THE DETERMINATION OF COBALT AND NICKEL. 389 'ormed, are generally said to be insoluble in acetic and hydrochloric acids. The alleged "insolubility" is a misnomer, because relatively large amounts of the two sulphides do dissolve in dilute hydrochloric acid 1 volume acid, sp. gr. 1'12, with 5 volumes of water. 1 The reverse action, the precipitation of the sulphides in acid solution by hydrogen sulphide, is also exceedingly slow. If over a trace of acetic or hydrochloric acid be present, neither sulphide will be precipitated in the cold ; but if the solution be hot, both nickel and cobalt sulphides will be pre- cipitated in the presence of ammonium or sodium acetate. Cobalt and nickel sulphides become " insoluble " in these acids on standing for some time, or on heating. They then require digesting with aqua regia for their solution. 2 195. The Separation of Manganese from Cobalt and Nickel. Precipitation of Nickel, Cobalt, and Manganese Sulphides? The filtrate from the zinc sulphide is neutralised with ammonia, and an excess of ammonia (free from carbonate) is added to the solution. Hydrogen sulphide is passed through the warm (70-SO) solution, whereby the mixed ' sulphides of manganese, cobalt, and nickel 4 are precipitated. The alkalies and alkaline earths remain in solution. Filter 5 at once, 6 and wash with water containing a little ammonium sulphide and chloride in solution. The nitrate may come through brown. In any case clear or brown evaporate the nitrate to about 50 c.c. ; add freshly prepared ammonium sulphide ; acidify with acetic acid ; and boil for some time. This will curdle the sulphides not retained on the filter paper, and allow them to be readily removed by filtra- tion. The danger of losing nickel and cobalt sulphides is here so serious that, if any metal escaped with the first filtrate, it is advisable to test the filtrate again to make sure that all the cobalt and nickel have been precipitated. Precipitation of Nickel and Cobalt Sulphides. The washed precipitate is dissolved in aqua regia, and boiled to expel the excess of acid. Add an excess of sodium carbonate to the solution, and then acetic acid 7 until the solution is faintly acid. Add 3 to 5 grms. of sodium or ammonium acetate ; 8 dilute the solution to, say, 200 c.c. ; pass hydrogen sulphide through the warm solution (70-80). The sulphides of cobalt and nickel 9 are precipitated, the manganese remains in solution. Filter at once through a close-packed filter paper, and wash with hydrochloric acid (sp. gr. 1'025) saturated with hydrogen sulphide in order to remove any manganese sulphide precipitated with the cobalt and nickel sulphides. The risk of losing cobalt, and particularly nickel, is here very great. Hence, make sure that the precipitation is complete by evaporating the clear filtrate to 50 c.c. ; add an excess of ammonium sulphide ; acidify with an excess of acetic 1 H. Baubigny, Compt. Rend., 94. 963, 1183, 1251, 1417, 1473, 1715, 1882; 95. 35, 1882; 105. 751, 806, 1887 ; 106. 132, 1888 ; Chem. News, 57. 55, 1888. 2 W. Herz, Zeit. anorg. Chem , 27. 390, 1901 ; 28. 342, 1901. 3 R. Fresenius, Anleitung zur quantitativen Analyse, Braunschweig, I. 579, 1875 ; London, I. 429, 1876. 4 Also copper, if present. 5 A close-packed filter paper must be used, say, C. Schleicher and SchiiU's No. 589, or M. DreverhofTs No. 331. 6 All sulphides precipitated from boiling solutions should be filtered and washed at once, so as to prevent oxidation. S. P. Sharpies, Amer. J. Science (2), 50. 248, 1870 ; Chem. News, 22. 259, 1870. 7 W. Funk (Zeit. anal. Chem., 45. 562, 1906) uses formic acid. 8 Say 5 grms. of ammonium acetate per gram of cobalt or nickel. 9 Also copper, zinc, and uranium, if present. W. Gibbs, Amer. J. Science (2), 39. 62, 1865 ; Chem. News, n. 147, 1865. 390 A TREATISE ON CHEMICAL ANALYSIS. acid, and warm the solution. If nickel or cobalt be present, a precipitate will be formed. Filter. Test the nitrate as before. When all the cobalt and nickel has been precipitated, wash as before. 1 Dry the filter paper and contents. Incinerate to burn off the paper. Dissolve the precipitate in hydrochloric acid mixed with a little nitric acid, and determine the nickel and cobalt separately as indicated below. 196. The Separation of Cobalt and Nickel Fischer's Nitrite Process. Fischer's process 2 depends on the formation of an "insoluble" potassium cobaltinitrite Fischer's salt under conditions where the corresponding nickel salt is soluble. If necessary, evaporate the solution containing the mixed cobalt and nickel salts to dryness. Take up the residue with one or two drops of concentrated hydrochloric acid, and as little water as possible. It is said that the presence of two parts of cobalt per million can be detected in this way. Precipitation of Potassium Cobaltinitrite. Add caustic potash 3 to the solution under investigation until the precipitate is no longer formed on adding another drop of alkali. The solution will then have an alkaline reaction. Acidify the solution with an excess of acetic acid. 4 Suppose that the total volume is between 5 and 10 c.c. Add half this volume of a 50 per cent, solution of potassium nitrite, 5 and stir the solution vigorously. Let the solution stand for 24 hours in a warm place. Test if the precipitation is complete by pipetting off a little clear solution and adding a little more potassium nitrite. If precipitation be not complete, transfer the small portion back to the main solution, and repeat the treatment with the potassium nitrite solution, etc. Filter the solution containing the yellow crystalline precipitate. 6 Use the' clear filtrate for washing the precipitate on to the filter paper. Wash the precipitate with a barely acid 5 per cent, solution of potassium nitrite 7 until 1 The manganese in the filtrate from the nickel and cobalt sulphides is determined as indicated on page 374. 2 A. Duflosand N. W. Fischer, Pogg. Ann., 72. 475, 1847 ; 74. 115, 1849 ; H. Rose, ib., HO. 411, 1860 ; A. Stromeyer, Liebig's Ann., 96. 218, 1855 ; W. Gibbs and F. A. Genth, ib., 104. 309, 1857 ; Chem. News, 28. 51, 1873 ; H. Baubigny, Ann. Chim. Phys. (6), 17. 103, 1889 ; F. Gauhe, Zeit. anal. Chem., 4. 56, 1865 ; 5. 74, 1866 ; A. Brauner, ib., 16. 195, 1877 ; 0. Brunck, Zeit. angew. Chem., 20. 834, 1847, 1907 ; H. Herrenschmidt and E. Capilli, Le Cobalt et le Nickel, Rouen, 1888 ; Chem. Neivs, 69. 112, 128, 142, 1894 ; Zeit. anal. Chem., 32. 607, 1893 ; W. Funk, ib., 46. 1, 1907. 3 Some of the very best grades of caustic alkali contain nickel derived from the dishes in which the alkalies were made. Iron, alumina, and silica also appear in all but the very best grades of caustic alkali. 4 L. L. de Koninck (Bull. Soc. Chim. Belg., 23. 11, 200, 1909) considers that the precipita- tion is more complete if the solution contains a little free nitric acid and more potassium nitrite is added than suffices to neutralise the nitric acid. 5 POTASSIUM NITRITE SOLUTION. 1 grm. of the salt per 2 c.c. of water, and just neutralise the solution with acetic acid. The solution is prepared for use as required. If the solution has any flecks of insoluble alumina or silica, filter. Always test each batch of potassium nitrite for silica, alumina, and lead before it is used. 6 If alkaline earths, copper, or lead be present, some nickel may be precipitated with the cobalt owing to the formation of triple nitrites of lead, potassium, nickel, and the third element. It is also important to use freshly prepared ammonium sulphide, or ammonia free from carbonates, in separating the alkaline earths from the ammonium sulphide group. H. Baubigny, Compt. Rend., 107. 1148, 1888. Zinc and cadmium do not give precipitates when treated by the nitrite process. 7 Some recommend washing with a 10 per cent, solution of potassium acetate containing a little potassium nitrite, because the potassium acetate can be removed by washing with alcohol, in which it is fairly soluble, while potassium nitrite is but sparingly soluble in this solvent. B. Brauner, Zeit. anal. Chem., 16. 195, 1877. There is no need for the alcohol washing here, and potassium nitrite alone gives better results. THE DETERMINATION OF COBALT AND NICKEL. 391 1 c.c. of the wash-water boiled with hydrochloric acid and treated with caustic potash and bromine gives no black precipitate of nickel hydroxide, or until a portion of the nitrate neutralised with ammonia is not coloured brown with ammonium sulphide. Precipitation of Cobalt Hydroxide. Transfer the precipitated potassium cobaltinitrite * to a porcelain dish, cover with a clock-glass, and add hydrochloric acid until no more nitric oxide is evolved, showing that the nitrite is decomposed. Treat the solution with an excess of caustic potash and bromine ; 2 take care that the solution is kept alkaline with potash. The cobalt is precipitated as black cobalt hydroxide. Filter through a close-packed filter paper and wash by decantation with hot water. Dry and ignite the filter paper and precipitate in a Rose's crucible. Cool and weigh. Ignite the contents of the crucible in a stream of hydrogen, and weigh as metallic cobalt. 3 Purification from Silica. Owing to the large amount of alkali used in these determinations, the precipitates are particularly liable to contamination with silica and alumina from the glass, etc. To remove the silica, treat the oxide with hydrochloric acid in a porcelain crucible ; evaporate the mass to dryness ; mix with concentrated hydrochloric and nitric acids ; add hot water ; filter ; wash with hot water ; ignite the paper and contents ; and weigh as silica Si0 2 . Dede 4 claims that if potassium or sodium persulphate is used in place of bromine, so slight an excess of alkali hydroxide is needed that the precipitated cobalt (or nickel) oxides can be easily washed. Ignition in a Rose's Crucible. Rose's crucible 5 has an opening in the lid for the introduction of an earthenware pipe A for leading gas into the red-hot crucible. 6 The arrangement for this particular experiment is indicated in fig. 138. The hydrogen 7 is generated in a Kipp's apparatus, B. The hydrogen should be freed from arsenic, antimony, phosphorus, and carbon compounds by washing in a solution, (7, of potassium permanganate in concentrated sulphuric acid ; from sulphur compounds by washing in a concentrated solution of caustic soda, D ; and dried by passing through a tower of calcium chloride, E. s 1 0. L. Erdmann, Journ. prakt. Chem. (1), 97. 397, 1866; M. St Evre, ib. (1), 54. 84, 1851 ; Compt. Rend., 33. 166, 1851 ; A. Remele, Zeit. anal. Chem., 3. 313, 1864 ; W. Braun, ib., 6. 72, 1867 ; 7. 313, 1868 ; A. Rosenheim and I. Koppel, Zeit. anorg. Chem., 17. 35, 1898 ; R. Wegscheider, ib., 49. 441, 1906 ; T. Rosenbladt, Ber., 19. 2535, 1886 ; S. P. Sadtler, Amer. J. Science (2), 49. 189, 1870; Chem. News, 22. 8, 15, 26, 1870; W. Blomstrand, Chemie der Jetztzeit, Heidelberg, 414, 1869. 2 L. Dede (Chem. Ztg., 35. 1077, 1911) recommends sodium or potassium persulphate in place of bromine, since only enough alkali is then needed to ensure alkalinity after the addition of the persulphate. The precipitation occurs in cold after standing one or two hours with frequent shaking. 3 T. Bayley, Chem. Neivs, 34. 81, 1876 ; A. Carnot, ib., 59. 183,' 1889 ; Compt. Rend., 108. 741, 1889 ; B. Brauner, Zeit. anal. Chem., 16. 195, 1877. 4 L. Dede, Chem. Ztg., 35. 1077, 1911. 5 H. Rose, Pogg. Ann., no. 128, 1860. In the absence of a Rose's crucible, an ordinary crucible and a common clay pipe of such a size that the mouth of the inverted bowl will just pass into the crucible may be used. E. Murmann (Monats. Chem., 19. 403, 1898) has a Gooch's crucible with a tube extension below the perforated base whereby the sulphide can be filtered as in Gooch's crucible, and subsequently heated in a current of any desired gas. The tubes are easily broken, and they are therefore expensive. 6 W. Gibbs (Chem. News, 28. 30, 1873 ; R. H. Lee, ib., 24. 234, 1871 ; Amer. J. Science (3), 2. 44, 1871) recommends a circular disc of porous earthenware above the substance to be heated, and below the gas inlet pipe. The gases pass through the disc to the substance to be reduced by diffusion. Mechanical loss is thus prevented. The "soft" porous capsule can be easily filed to tit the crucible perfectly. 7 Not coal gas, because of the formation of cobalt carbides. 8 E. Schobig, Journ. prakt. Chem. (2), 14. 289, 1877 ; E. Varenne and E. Hebre, Bull. Koc. Chim. (3), 28. 523, 1902. 392 A TREATISE ON CHEMICAL ANALYSIS. Explosions may be prevented by placing a glass tube containing cotton-wool between discs of wire gauze immediately after the drying tower. 1 See page 150. The Hose's crucible is placed in position, and when the air has been expelled by the hydrogen generated in the Kipp's apparatus, light the burner and gradually raise the temperature of the crucible and contents to bright redness. 2 The current of gas 3 should be so regulated that from about two to four bubbles per second pass through the wash-bottle during the earlier stages of the reduction, when comparatively large volumes of steam are being evolved. The velocity of the stream may then be increased to about eight bubbles per second. In about 10 or 15 minutes the oxide will no doubt be all reduced to metal ; FIG. 138. Reduction of cobalt oxide. remove the flame ; and let the crucible cool in the current of gas. Place the crucible in a desiccator, and, when cold, weigh as metallic cobalt. For example : Rose's crucible plus metal 15 '3572 grms. Rose's crucible alone . . . . . . . 14-8731 grms. Metallic cobalt . , 0*4841 grm. The reduction is necessary because oxides of varying degrees of oxidation are formed by the ignition in air. Some prefer to convert the oxides into sulphates before weighing, and thus avoid the reduction. 4 1 R. Fresenius, Zeit. anal. Chem., 12. 73, 1873 ; C. G. Hopkins, Journ. Aiwr. Chem. Soc., 21. 645, 1899 ; Chem. News, Si. 134, 1900. 2 The reduction of cobalt oxide commences about 132 W. Miiller, Pogg. Ann., 136. 51, a J. Habermann (Zeit. anal. Chem., 28. 88, 1889) recommends an alloy of tin with 83-84 per cent, of zinc in preference to zinc alone. The form of Habermann's alloy remains the same alter the zinc has dissolved, and in consequence none falls into the lower bulb of Kipp's generator. w 4 T E o J ' A 5f umen ^ Compt. Rend., 79. 179, 1874 ; F. Gauhe, Zeit. anal. Chem., 4. 53, 1865; W. J. Russell, Journ. Chem. Soc., 16. 51, 1863. THE DETERMINATION OF COBALT AND NICKEL. 393 Determination of Nickel. The nickel is determined l in the nitrate from the cobalt by acidifying the solution with hydrochloric acid in order to decompose the nitrite. Precipitate the black nickel hydroxide by adding caustic potash and bromine as just indicated for cobalt.- The dark brown precipitate of Ni(OH) 3 is reduced to the metallic state in a Rose's crucible as described for cobalt. The precipitate may also be ignited in an ordinary crucible, and weighed as nickel oxide NiO. 3 The potassium nitrite process is by no means perfect, since nickel can be afterwards detected with the cobalt, and cobalt with the nickel. The error in the determination of the cobalt is, however, almost balanced by the slight solubility of the potassium cobaltinitrite precipitate in the mother liquid. The precipitate should not be allowed to stand much over the 24 hours without attention. This process, as indicated above (and also Liebig's cyanide process), breaks down if alkaline earths be present. 197. The Separation of Nickel Liebig's Cyanide Process. This process is based on the fact that nickel hydroxide alone is precipitated by bromine from an alkaline solution containing an excess of potassium cyanide. The nickel probably occurs in solution as a double cyanide of nickel arid potassium, whereas the cobalt occurs in solution as potassium cobalticyanide. Bromine reacts with the former, not with the latter. 4 The solution of the mixed sulphides is neutralised with potassium hydroxide, and treated with a solution of "pure" potassium cyanide until the precipitate first formed redissolves ; add more potassium cyanide in all, 3 ' or 4 grms. usually suffice. Then about 5 grms. of, potassium hydroxide are added, and about 5 c.c. of bromine, with constant stirring, until the nickel is all precipitated. If necessary, add more potash in order to keep the solution alkaline throughout the whole process, or the precipitation will not be complete. The nickel should all be precipitated in about an hour. Dilute with about 800 c.c. of cold water, and determine the nickel as indicated above. 5 The cobalt remaining in the filtrate as potassium cobalticyanide is evaporated with dilute sulphuric acid in a platinum dish on a water bath ; then add con- centrated sulphuric acid, and evaporate on a sand bath until dense white fumes are evolved and effervescence has ceased. This shows that the colourless cobalticyanide is all changed to rose-red cobalt sulphate. Cool. Dissolve the residue in water, and precipitate the cobalt with bromine in alkaline solution as indicated above. In special cases, certain other methods for the determination of cobalt and nickel are useful. If suitable apparatus be available, the electrolytic methods of separation are splendid. The following are also useful in special cases. 1 The electrolytic process, page 394, may be used, and in fact is strongly recommended, if convenient. 2 If the bromine be omitted, the apple-green precipitate of nickel hydroxide Ni(OH) 2 is difficult to filter and to wash from alkali. 3 It is very difficult to wash the precipitate free from alkali. The oxide can be washed with water, dried and weighed Silica can be determined as indicated in the text for cobalt, or the solution can be treated with ammonium sulphide and the precipitate ignited and weighed as NiO or metal. 4 J. von Liebig, Liebig's Ann., 65. 244, 1848 ; 87. 128, 1853 ; F. Wohler, ib., 70. 256, 1849 ; F Gauhe, Zeit. anal. Chem., 5. 75, 1866; C. Krauss, ib., 30. 227, 1891 ; Chem. News, 63. 254, 264, 280, 293, 1891 ; W. Gibbs, ib., u. 125, 1865 ; Amer. J. Science (2), 39. 58, 1865. 5 Or use the electrolytic process, page 394, which gives better results than the process indi- cated in the text. 394 A TREATISE ON CHEMICAL ANALYSIS. 198. The Separation of Small Amounts of Cobalt from Large Amounts of Nickel Ilinsky and Knorre's Nitroso-/3-naphthol Process. This process 1 depends on the fact that a solution of nitroso-/3-naphthol precipitates cobalt from a mixed solution of cobalt and nickel. The reagent is so sensitive that a visible turbidity is produced in solutions which give no sign of cobalt by the nitrite process. The precipitate is rather bulky, so that the process is most convenient for separating small quantities of cobalt from com- paratively large amounts of nickel. Add a little sulphuric acid to the solution of the mixed sulphides indicated above, and evaporate the solution on a sand bath until white fumes of sulphuric acid are evolved. Cool. Dilute. Add 5 c.c. of concentrated hydrochloric acid, and then a hot acetic acid solution of freshly prepared nitroso-^-naphthol 2 as long as a precipitate of cobalti-nitroso-/?-naphthol is produced. Let the volumin- ous precipitate settle. Test the clear solution with more reagent to find if the precipitation is complete. Let the mixture stand overnight. Filter. Wash with cold water, then with a hot 12 per cent, solution of hydrochloric acid so as to remove the nickel salt, which is soluble in hydrochloric acid. Wash with hot water until the filtrate is free from acid. Dry the precipitate and place it in a Rose's crucible. Add a little pure oxalic acid, 3 and raise the temperature gradually in order to avoid loss by spurting. Finish the ignition over a Teclu's burner. When the carbon of the filter paper is all consumed, reduce the^ cobalt to metal by heating in a current of hydrogen. Cool, and weigh the metallic cobalt. 4 The greatest difficulty in this process is the elimination of carbon. The cobalt seems to form a compound with carbon which is not destroyed even by ignition over a blast. Ilinski and Knorre's reagent also precipitates copper, chromium, and iron, but it is not affected by the presence of magnesia and lime. It does not precipitate aluminium, lead, cadmium, manganese, nickel, mercury, and zinc, although these constituents may be carried down mechanically with the cobalt precipitate. The latter must then be purified by solution and re- precipitation. 199. The Separation of Small Amounts of Nickel from Large Amounts of Cobalt Brunck's a-Dimethylglyoxime Process. If an excess of a-dimethylglyoxime be added to a strongly ammoniacal solution, containing both nickel and cobalt salts, and if the solution be boiled a short time, 1 part of nickel in the presence of 500 parts of cobalt will pro- duce a rose coloration. When small quantities of nickel are in question, the 1 M. Ilinsky and G. von Knorre, Ber., 18. 699, 2728, 1885 ; 2O. 283, 1886 ; Chem. News. 52. 301, 1885 ; C. Krauss, Zed. anal. Chem., 30. 227, 1891 ; L. L. de Koninck, Rev. Univ. Mines, 9. 243, 1890; Chem. News, 62. 19, 1890; C. Meineke, Zeit. angew. Chem., i. 3, 1888; M. Ilinsky, Chem. Ztg., 19. 1421, 1885; H. Copaux, Bull. Soc. Chim. (3), 29. 301, 1903; Che-m. News, 87. 291, 1903. 2 NiTR.oso-0-NAPHTHOL SOLUTION.- -Dissolve 8 grms. of the solid in 300 c.c. of cold glacial acetic acid. Dilute with 300 c.c. of water. Filter. The solution does not keep very well. Make up a fresh solution about once a month. Nitroso--uaphthol costs 8s. per 100 grms. 3 If the oxalic acid be omitted, the results are usually a little high. 4 The nickel can be determined in the filtrate by evaporating to a small bulk with sulphuric acid, and expelling most of the acid on a sand bath. The nickel may then be precipitated with caustic potash and bromine water as indicated above. THE DETERMINATION OF COBALT AND NICKEL. 395 i solution should be evaporated almost to dryness. 1 This coloration is best seen when the solution is filtered and the precipitate shaken with water. Larger amounts of nickel give a scarlet red precipitate. If the precipitate be dissolved in a mixture of chloroform and alcohol, and evaporated to dryness, red needles of nickel-a-dimethylglyoxime are formed. 2 The reaction is applied quantitatively as follows : Evaporate the solution of mixed sulphides to dryness on a water bath. Dilute- the residue to about 300 or 500 c.c., such that the cobalt is less concen- trated than O'l grm. per 100 c.c. Warm the solution to about 50 ; add O2 grm. of dimethylglyoxime in alcoholic solution, 3 and then add about 2 grms. of sodium acetate. 4 Stir the solution thoroughly, and let it stand about half an hour. A voluminous precipitate of the nickel oxime separates. 5 Filter the solution through an asbestos-packed Gooch's crucible. Wash with warm water (50), and dry for about an hour at 110-120, after which there should be no further loss on drying. 6 When cold, weigh the mass as C 8 H 14 N 4 4 Ni, and multiply the weight so obtained by O2033 to get the corresponding amount of nickel oxide, NiO. There is no danger of loss of nickel if the precipitate be heated below 250 . 7 At this temperature the salt begins to sublime undecomposed. The process can be employed for separating nickel from zinc, manganese, iron, aluminium, and chromium in ammoniacal solutions in the presence of sodium acetate. With zinc-nickel mixtures, for example, the dimethylglyoxime is destroyed by boiling the filtrate from the nickel with hydrochloric acid, and the zinc precipitated by Gibb's process. 200. The Electrolytic Process for Cobalt and Nickel Fresenius and Bergmann's Process. Nickel and cobalt cannot be precipitated satisfactorily from solutions con- taining free acids, but these metals are readily precipitated from solutions of the double cyanides, double oxalates, and double sulphates, or in the presence of 1 L. Tschugajeff, Ber., 38. 2520, 1905; K. Kraut, Zeit. angew. Chem., 19. 1793, 1906; 0. Brunck, ib., 2O. 834, 1844, 1907 ; Chem. News, 99. 275, 1909; A. Ivanicki, Stahl Eisen, 2fj. 358, 1908 ; H. Wdowiszewski, ib., 27. 960, 1908; 29. 358, 1910 ; P. Bogoluboff, ib., 30. 458, 1911 ; L. V. W. Spring, Journ. Ind. Eng. Chem., 3. 255, 1911; F. Ibbotson, Chem. News, 104. 224, 1911 ; H. Pederson, Met., 8. 335, 1911 ; S. W. Parr and J. M. Lindgren, Trans. Amer, Brassfounders' Assoc.,$. 120, 1912. For the separation of nickel and palladium, see M. Wunder and V. Thuringer, Ann. Chim. Anal., 17. 201, 1912. Platinum is qualitatively, not quantitatively, precipitated M. Wunder and V. Thuringer, ib., 17. 328, 1912. 2 H. Grossmann and B. Schuck (Chem. Ztg., 31. 535, 643, 911, 1907) detect nickel under similar circumstances by means of dicyanodiamidine sulphate. The process is also used quantitatively H. Grossmann, B. Schuck, and W. Heilborn (Bull. Soc. Ind. Rouen, 38. 116, 125, 1910). Dicyanodiamidine costs about 5d. per 10 grms. y a- DIMETHYLGLYOXIME SOLUTION. Dissolve 0'2 grm. of a - dimethylglyoxime CH 3 .C : NOH C : NOH.CH 3 in 20 c.c. of 98 per cent, alcohol. Filter the solution if necessary. The solution will not keep very long. In 1908 dimethylglyoxime cost 12s. per 10 grms. ; and in October 1911 the price had fallen to Is. 6d. per 10 grms. Hence the cost factor of this useful process is no longer so serious as it was. 4 The precipitate is soluble in free mineral acids. Ammonium and sodium acetates render the mineral acids inert. 5 The solution should not have more than half its volume of the alcoholic dimethylglyoxime, or appreciable amounts of the nickeloxime may be dissolved. Otherwise an excess of the solu- tion does no harm. Theoretically, four times as much dimethylglyoxime by weight is needed as nickel to be precipitated. A small excess suffices say five times ; but if much cobalt be present, a greater excess is needed, owing to some of the glyoxime forming a complex salt with the cobalt. 6 P. Bogoluboff, Stahl Eisen, 29. 458, 1910. 7 The cobalt can be determined in the filtrate by evaporating it to a small bulk and pro- ceeding as indicated above. Brunck recommends the joint determination of cobalt and nickel, direct determination of nickel, and cobalt by difference. 396 A TREATISE ON CHEMICAL ANALYSIS. alkaline acetates, tartrates, citrates, etc. 1 In Fresenius and Bergmann's process, 2 which is most generally used, the double sulphate of potassium and nickel, or potassium and cobalt in ammoniacal solutions, is employed. The presence of sodium phosphate 3 or ammonium sulphate is favourable to the deposition of these metals. Cobalt deposits rather more slowly than nickel, and the results with cobalt are usually a little too high. The Electrolyte. Mix the given solution with 5-10 grms. of ammonium sulphate and 30-40 c.c. of concentrated ammonia for every 0-25-0-30 grm. of the nickel or cobalt sulphate or chloride 4 in the solution under investigation. Dilute the solution with distilled water until it occupies a volume of about 150 c.c. The Electrolysis. The solution is electrolysed at the room temperature 5 with a current density of 1-0 to 1'5 amps., and 2-8 to 3*5 volts. The electrolysis is completed in the case of nickel in about 2| to 3 hours; and in the case of cobalt, in 5 to 6 hours. 6 Break the current ; pour off the exhausted electrolyte (page 359); wash, dry, and weigh the deposit. The deposited nickel adheres firmly to the cathode, and it has a bright silver- grey colour, sometimes closely resembling the appearance of the platinum itself. The cobalt deposit is generally brown or black in colour. It is rare to see a brilliant cobalt deposit. The results with the dark-coloured cobalt deposits are quite as satisfactory as with the bright deposits. If too little ammonia has been used in the electrolyte, some nickel may be deposited on the anode, giving low results. Too much ammonia retards the deposition of the metals. The presence of ammonium chloride or ammonium nitrate also retards the deposition of the metals. 7 A couple of test experiments quoted from Fresenius and Bergmann's paper show that good results can be obtained. This agrees with general experience. Nickel. Cobalt. Used " . , ', . . 0-1233 0'1280 grm. Found . . v ,. . 0-1233 0'1286 grm. Treatment of Mixed Cobalt and Nickel Sulphides. In actual analyses, mixed cobalt and nickel sulphides are often obtained. To prepare these for analysis, ignite the filter paper in the usual manner (page 390). Dissolve the sulphides and ash of the paper in hydrochloric acid with a little nitric acid. Evaporate the solution to dryness on a water bath. Dissolve the residue in a little dilute sulphuric acid, and transfer the solution to the platinum basin for electrolysis ; add 5 grms. of ammonium sulphate, 40-60 c.c. of ammonia ; dilute the solution to about 150 c.c. with distilled water; and proceed with the electrolysis as indicated in the text. Weigh the mixed deposit as metallic " cobalt + nickel." Dissolve the metals in hot nitric acid, and determine either the nickel or the 1 W. Gibbs, Zeit. anal. Chem., 3. 336, 1864 ; u. 10, 1872 ; 22. 558, 1883 ; F. Wrightson, ib., 15. 300, 1876 ; T. Schweder, ib., 16. 344, 1877 ; W. Ohl, ib., 18. 523, 1899 ; C. Luckow, Dingier' 's Journ., 177. 235, 1850. 2 H. Fresenius and F. Bergmann, Zeit. anal. Chem., 19. 314, 1880. 3 M. S. Cheney and E. S. Richards, Amer. J. Science (3), 14. 178, 1877. 4 Nitrates should be absent. If nitrates be present in the solution, evaporate to dryness with sulphuric acid in order to expel the nitric acid. 5 If the solution is electrolysed at, say, 45-50, 1 J to 2 hours are needed for the electrolysis. 6 To test for the end of the electrolysis in a colourless solution, transfer about 1 c.c. by means of a pipette to a test tube. Add H 2 S water. A brown coloration will show whether the metal is all deposited. Yellow ammonium sulphide is not so good for making the test, because its yellow colour may mask the colour produced by small quantities of nickel or cobalt. 7 If the electrolysis of ammoniacal solutions be too protracted, the cathode increases in weight as soon as all the nickel is deposited, possibly owing to the dissolution of the anode F. P. Tread well, Kurzes Lehrbuch der analytischen Chemie, Leipzig, 2. 115, 1911. THE DETERMINATION OF COBALT AND NICKEL. 397 cobalt by one of the processes described in this work. The other metal can be obtained by difference. Removal of the Deposit from the Electrode. Nickel is difficult to remove from the cathode, because it is inclined to become "passive." Warm nitric or sul- phuric acid may be used for dissolving the metal. Owing to the close resem- blance between the deposited nickel and the platinum, special care must be taken to dissolve off all the nickel before the cathode is heated ready for the next determination. It is very difficult to remove the stain left on platinum when it has been heated in contact with nickel. The removal of cobalt presents no particular difficulty, since it does not assume the passive condition, and it is readily soluble in concentrated nitric acid. 201. The Colorimetric Determination of Small Quantities of Cobalt. Lampadius' attempt l to determine small quantities of cobalt from the reddish- brown colour produced when ammonia is added to the solution of a cobalt salt was not very successful. The colour changes on exposure to the air. 2 Miiller used a colorimeter for estimating the amount of cobalt. This measured the cobalt in terms of the intensity of the colour of solutions of its salts; 3 and Winkler, recognising that the colour of cobalt solutions is complementary to that of nickel solutions, 4 and that the one solution therefore neutralises the tint of the other, found that the maximum decolorisation occurred when the proportion of nickel was to cobalt as 3:1. Hence Winkler proposed adding nickel or cobalt to a given solution until the maximum decolorisation occurred. This furnished data sufficient to compute an approximation to the amount of nickel and of cobalt in a given solution. The process is not very reliable. Skey, 5 as indicated on page 385, found that, when solutions of cobalt thio- cyanate are shaken up with ether or alcohol, a blue superincumbent layer is obtained, and this reaction is sometimes inaptly called " Vogel's test " for cobalt. Skey suggested shaking out the solution of cobalt thiocyanate with ether as a means of separating cobalt and nickel ; and this idea was later employed by Rosenheim and Huldschinsky, 6 and recommended for separating cobalt arid nickel by a process similar to that employed by Rothe (page 456) for separation of iron. Skey further showed that the blue colour is destroyed by sodium thiosulphate, sodium acetate, and mercuric chloride. The red colour of ferric thiocyanate, if 1 W. A. Lampadius, Journ. prakt. Chem. (1), 13. 385, 1838. 2 T. Bodemann, Anleitung zur Berg- und Huttenmdnnischen Probierkunst, Clausthal, 456, 1857. 3 A. Miiller, Das complementar Colorimeter, Chemnitz, 1854 ; Journ. prakt. Chem. (1), 60. 474, 1853 ; F. Dehms, Zeit. anal. Chem., 3. 218, 494, 1864. 4 R. Wagener, Journ. prakt. Chem. (1), 61. 129, 1854; C. Winkler, ib. (1), 97. 414, 1866; R. W. Challinor, Journ. Roy. Soc. N.S.W., 38. 406, 1905; E. J. Maumene, Compt. Rend., 30. 209, 1850; J. H. Gladstone, Phil. Mag. (4), 9. 535, 1855 ; T. Bayley, ib. (5), 6. 15, 1878 ; Journ. Chem. Soc., 37. 828, 1880; J. Bottomley, Proc. Mancliester Lit. Phil. Soc., 19. 164, 1880; Chem. News, 42. 56, 1880; M. Knieder, Berg. Hiltt. Ztg., 53. 243, 1894; F. Dehms, Dingier 's Journ., 172. 440, 1864 ; 173. 436, 1864. 5 W. Skey, Chem. News, 16. 201, 1867; H. W. Vogel, Ber., 8. 1533, 1875; 12. 2314, 1879; Zeit. anal. Chem., 21. 563, 1882; T. T. Morrell, ib., 16. 251, 1877; Pharm. Centr., 17. 394, 1877 ; W. H. Bettink, Nederl. Tijdschr. Pharm., u. 43, 1899. H. Vitz (Chem. Ztg., 25. 109, 1901) studied the changes of colour of cobalt salts by the action of glycerine, ether, alcohol, acetone, etc. Skey isolated blue acicular crystals from the ethereal solution of cobalt thiocyanate. F. P. Treadwell (Zeit. anorg. Chem., 26. 108, 1901) later represented the composi- tion of the crystals by the formula (NH 4 ) 2 Co(SCN) 4 . 6 A. Rosenheim and E. Huldschinsky, Zeit. anal. Chem., 40. 809, 1901; Ber., 34. 2050, 1901 ; A. Rosenheim and A. Cohn, ib., 33. 1111, 1900 ; F. W. Dootsou, Proc. Cambridge Phil. JSoc., 12. 125, 1903. 398 A TREATISE ON CHEMICAL ANALYSIS. present, masks the blue colour of the cobalt thiocyanate, and, in consequence, the iron must be removed. Wolff 1 proposed removing the iron by first adding an excess of ammonium thiocyanate, and then neutral sodium carbonate until the blood-red colour of the ferric thiocyanate disappeared. The precipitated ferric oxide was removed by filtration, and the filtrate shaken with ether to get the cobalt-blue coloration. This is quite satisfactory. Bettink preferred reducing the ferric iron to the ferrous condition by means of sodium thiosulphate until the red colour of the ferric thiocyanate disappeared. The filtered solution was then treated with ether as before. There are several objections to the use of sodium thiosulphate. These reactions can be applied to the colorimetric deter- mination of small quantities of cobalt 2 in the following manner : Standard /Solution. Pipette 1 c.c. of standard solution of cobalt sulphate 3 into a graduated Nessler's tube, 4 25 to 30 c.c. capacity. Add 1 c.c. of dilute sulphuric acid; 5 c.c. of water; 0'2 grm. of solid ammonium thiocyanate; 5 and water up to the 10-c.c. mark. Test Solution. Add 1 grm. of ammonium thiocyanate per 50 c.c. of the solu- tion under investigation, and then add concentrated sodium carbonate Na 2 C0 3 until the red colour of the ferric thiocyanate disappears. 7 Filter and wash. Neutralise the filtrate with dilute sulphuric acid, and make the solution up to 100 c.c. Pipette, say, 5 c.c. into a Nessler's tube; add 1 c.c. of dilute sulphuric acid ; and make the solution up to the 10-c.c. mark with water. The Comparison. Fill both tubes up to the 20-c.c. mark with a mixture of ether and alcohol. 8 Shake the mixture. As soon as the ethereal- and aqueous layers have separated, compare the colour of the ethereal solutions, and note if more or less standard solution is needed to make the tints of the two solutions the same. Add a diluted solution (say 1 in 10 c.c.) of the standard cobalt solution to the paler- tinted solution from a burette reading to -g 1 ^ c.c., and an equivalent amount of water to the other solution, until the tints of the two solutions are the same. Or make up other standard solutions, if necessary, containing more or less standard cobalt sulphate, until the tints of the standard and test solutions are the same. The calculations are then made in the usual manner. EXAMPLE. Suppose 5 c.c. of the "filter press water "of a pottery be evaporated to dry ness and the residue washed, and the solution, about 5 c.c., be treated as described in the text. After a comparison, etc., the test solution was the paler. It required 2-58 c.c. of diluted cobalt sulphate (10 c.c. of standard was made up to 100 c.c. with water) to bring the tints the same. Hence the standard solution had 0*000002 grm. of cobalt, and the test solution x + 0*00000258 grm. of cobalt. Hence = 0-00000742 grm. of cobalt per 5 c.c. of filter press water. But 0'00000742 x 2'204 = 0'00001635 grm. cobalt chloride, CoCl 2 , per 5 c.c. ; 0'00033 grm. per litre ; or the equivalent of 0'00054 oz. of anhydrous cobalt chloride was escaping per gallon of the filter press water. The aqueous layer retains about one-tenth of its volume of ether, and also a trace of cobalt ; but since the standard and test solutions are similar, the error is 1 C. H. Wolff, Zeit. anal. Chem., 18. 38, 1879; C. Zimmermann, ib., 20. 414, 1881; Liebig's Ann. , 199. 1, 1879 ; W. H. Bettink, Nederl. Tijdschr. Pharm,, II. 64, 1899. 2 J. W. Mellor, Trans. Eng. Cer. Soc., 8. 132, 1908. 3 STANDARD SOLUTION OF COBALT SULPHATE. Dissolve O'Ol grm. of cobalt in dilute sulphuric acid and make the solution up to a litre with water. This furnishes a solution con taining O'Ol grm. cobalt per 1000 c.c., or 1 c.c. is equivalent to O'OOOOl grm. of cobalt. 4 The Nessler's cylinders are about 17 cm. high and 13mm. internal diameter, and made from colourless glass. The cylinders should be as nearly alike as possible, and stoppered. See page 85. stained body or glaze." carbonate is unnecessary. ETHER- ALCOHOL MIXTURE. Methylated ether, 5 vols. ; isoamyl alcohol, 5 vols. No naked flame should be near the bench when ether is being used. THE DETERMINATION OF COBALT AND NICKEL. 399 negligibly small. Otherwise the ethereal solutions would have to be removed and the aqueous layer washed out with ether, and the two solutions compared in any convenient way. This method is useful for estimating the small amounts of " cobalt " employed for "bleaching" china clays, pottery bodies, etc. 1 202. The Evaluation of Cobalt and Nickel Oxides. The standard for commercial " black oxide " of cobalt is 70 per cent, of metallic cobalt, corresponding with 89 per cent, of cobalt monoxide CoO but it generally contains the equivalent of 71 to 72 per cent, metallic cobalt. The standard for " prepared oxide " is 7 4 '5 per cent, of metal, corresponding with 9 4 '6 per cent of CoO ; but commercial samples sometimes run as high as 76 to 77 per cent, of metallic cobalt. This corresponds with 96 to 98 per cent, of cobalt oxide CoO. Hence, in evaluating these oxides, a determination of the percentage of cobalt or nickel as metal generally suffices. 2 A qualitative test should always be made. If any of the metals indicated below be absent, the process can be abbreviated accordingly. Assume, in the extreme case, that the oxide contains silica, some metals in the hydrogen sulphide group, iron, aluminium, manganese, zinc, nickel, and cobalt. In evaluating the oxide, therefore, everything is got into solution and the im- purities are removed, step by step. For a full analysis, the method indicated in the schemes, pages 319 and 387, may be used ; for the cobalt and nickel oxides alone, proceed as described below. 1. Dissolution of the Oxide. Digest half a gram of the oxide in a 200-c.c. flask with 10 c.c. of concentrated hydrochloric acid. Filter and wash. Fuse the dried and ignited insoluble residue with sodium carbonate; take up the mass with hydrochloric acid ; evaporate the mixed solutions to dryness. 2. Removal of Silica, Iron, and Aluminium. Add 35 c.c. of concentrated ammonia and ammonium chloride, and warm the mixture until the residue is disintegrated ; filter ; wash with hot water ; rinse the residue on the filter paper back into the flask ; pour hot dilute hydrochloric acid (1:2) through the filter paper to dissolve the precipitate, collecting the runnings in the flask under the funnel ; wash ; boil until solution is complete ; add ammonium chloride and excess of ammonia to the boiling solution ; filter ; wash thoroughly with hot water ; and unite the two filtrates. 3 3. Precipitation of Copper by Hydrogen Sulphide. Boil off any excess of ammonia ; just acidify the solution with hydrochloric acid ; and then add 5 c.c. of the acid in excess ; boil a few minutes ; dilute to about 250 c.c. with water ; pass hydrogen sulphide to precipitate copper, lead, etc. Filter, and wash with water saturated with hydrogen sulphide; and boil the filtrate to expel the hydrogen sulphide. 4. Removal of Manganese. Add a slight excess of ammonia ; acidify the solution strongly with acetic acid ; add 1 to 2 grms. of ammonium acetate ; heat the solution to 70 or 80 ; saturate the solution with hydrogen sulphide ; filter, and wash with hot water. The manganese remains in solution ; the cobalt, zinc, and nickel are precipitated. The filtrate probably still contains 1 For the colorimetric determination of nickel by the use of potassium thiocarbonate, see M. Lucas, Bull. Soc. Chim. (3), 21. 432, 1899; Chem. Nexus, 80. 39, 1899. 2 For the analysis of cobalt ores, see H. Copaux, Bull. Soc. Chim. (3), 29. 301, 1903 ; and for the analysis of nickel ores, A. Hollard, Ann. Chim. Anal., 8. 401, 1903. 3 The amount of iron and alumina is usually too small for the result to be affected by the adsorption of salts of nickel and cobalt by the precipitated iron ; if otherwise, use the basic acetate separation. 4<3O A TREATISE ON CHEMICAL ANALYSIS. traces of cobalt and nickel. To precipitate these, concentrate the solution by evaporation ; add colourless ammonium sulphide ; acidify the solution with acetic acid; warm; and filter, if necessary, into a separate beaker. Test the filtrate in the same way ; and, if necessary, repeat the operation. 5. Removal of Zinc. Wash the precipitated sulphides from the paper as completely as possible ; dry, and burn the paper ; dissolve everything in hydrochloric acid with a little nitric acid. The solution contains cobalt, nickel, and zinc. To remove the latter, add 2 or 3 grms. of finely crystalline ammonium chloride ; x evaporate the solution to dryness on a water bath ; heat the solid mass until all the ammonium chloride is expelled. Zinc is volatilised at the same time. When cold, dissolve the mass in aqua regia and evaporate to dryness. The solution is now ready for the separation of cobalt and nickel as described on pages 389 et seq. Smalt or zaffre may be fused with sodium carbonate and sodium nitrite (page 461), or broken down by treatment with acids. The silica can be separated by evaporation of the hydrochloric acid solution (page 167), and the filtrate treated as described above for the separation of iron, or by the basic acetate process (page 362). Nickel, cobalt, bismuth, iron, lead, and arsenic, as well as silica, alumina, alkalies, and alkaline earths, may be present. The matt blues rich in alumina are best fused with potassium bisulphate, or pyrosulphate. Zinc, phosphoric and arsenic oxides, as well as cobalt, nickel, silica, etc., may here be present. 203. The Volumetric Determination of Cobalt and Nickel. The methods which have been suggested up to the present for the determination of cobalt volumetrically are not satisfactory except under special conditions, and hence no volumetric method for cobalt 2 has won a place in general practice. Volumetric methods for nickel also fail in the presence of cobalt, although the cyanide process for nickel 3 is quite good when cobalt is absent, or only present in minute quantities, and when copper, silver, gold, and metals of the platinum group are absent. Theory of the Cyanide Process. The cyanide process, suggested by Moore and by Campbell and Andrews, depends on the fact that, if potassium cyanide be added to a feebly ammoniacal solution of a nickel salt, a double cyanide is formed, say, NiCl 2 + 4KCN = 2KC1 + Ni(CN) 2 . 2KCN ; 1 R. Fresenius (Zeit. anal. Chem., 21. 229, 1881) says that 5 grms. of ammonium chloride suffice for '2 grin, of zinc oxide. 2 C. Winkler (Zeit. anal. Chem., 3. 420, 1864) titrates with potassium permanganate in the presence of mercuric oxide. H. B. Harris (Journ. Amer. Chem. Soc., 20. 173, 1898) obtained fair results with hot dilute solutions, not with concentrated solutions. N. M'Culloch (Chem. News, 59. 51, 1889) titrates cobalt cyanide with potassium bichromate after the addition of Mohr'ssalt. E. Fleischer (Journ. prakt. Chem. (2), no. 48, 1870) precipitates the cobalt with alkali hypochlorite, adds an excess of Mohr's salt, and titrates back with potassium per- manganate. E. Donath (Ber., 12. 1868, 1879; Chem. News, 41. 15, 1880) oxidises the with zinc emulsion ; adds an excess of potassium permanganate, and titrates back with a solution of Mohr's salt. C. Rossler, Liebig's Ann., 2OO. 323, 1880 ; E. Rupp and F. Pfennig, /^Zi^/wj 5V /v f% A QOO 1 Q1 f\ * n G To-rviT^cn-vf* 7", , / /i./i-i //,>*. A"Y7, -, Cf** ** *7K*7 iniA Chem. Ztg., 34. 322, 1910 ; G. S. Jamieson, Journ. Amer. Chem. Soc., 32. 757, 1910. 3 T. Moore, Chem. News, 59. 160, 292, 1889 ; 72. 92, 1895 ; H. Brearley and H. Jervis, ib., 78. 177, 196, 1898; E. D. Campbell and W. H. Andrews, Journ. Amer. Chem. Soc., 17. 125, 1895 ; E. D. Campbell and W. Arthur, ib., 30. 1116, 1908 ; Chem. News, 98. 38, 1908 ; G. W. Sargent, ib., 21. 854, 1899 ; C. M. Johnson, ib., 29. 1201, 1907 ; H. Grossmann, Chem. Ztg., 32. 1223, 1908; H. Grossmann and B. Schiick, Zeit. anal. Chem., 47. 169, 1908 ; G. Raulin, Monit. Sclent., 74. 84, 1911. THE DETERMINATION OF COBALT AND NICKEL. 401 and if the solution contains a little silver iodide (or a mixture of silver nitrate and potassium iodide) the solution will remain turbid until enough potassium cyanide has been added to transform all the nickel to the double cyanide. Further additions of potassium cyanide dissolve the silver iodide to a clear solution. Thus, Agl + 2KCN = AgCN.KCN + KI. When the silver iodide is all dissolved, it will be obvious that the trans- formation of the nickel to the double cyanide is completed. Hence, the potassium cyanide used in the titration represents the amount of potassium cyanide used for the formation of the double cyanide, and for the dissolution of the silver iodide. If the amount of the potassium cyanide required for the dissolution of the silver iodide be known, the amount required for the transformation of the nickel can be found by subtraction. The Determination. Add 2 c.c. of concentrated sulphuric acid to the solution under investigation. Neutralise the solution with ammonia. 1 Add 20 c.c. of 2N-NH 3 to the neutral solution, and then 2 c.c. of a 2 per cent, solution of potassium iodide, and 5 c.c. of a solution of silver nitrate (5*85 grms. per litre). 2 The solution will now be turbid owing to the presence of silver iodide. Titrate 3 the cold solution with a standard solution of potassium cyanide 4 until the solution is clear, showing that all the silver iodide is dissolved. A black background is best for the titration. Add more silver iodide, so as to make the solution turbid again. Again titrate with the potassium cyanide drop by drop until the last drop clears up the opalescence. The temperature of the solution should not exceed 20, or disturbing side reactions may set in. EXAMPLE. Suppose it be found that 31 c.c. of the potassium cyanide solution are needed for the whole titration ; that 1 c.c. of the potassium cyanide solution represents 0'0037 grm. NiO ; suppose 3 c.c. of silver nitrate are needed to reproduce the turbidity the second time, and that 1 c.c. of the silver nitrate solution represents ^ c.c. of the potassium cyanide solution. In all, 5 + 3 = 8 c.c. of silver nitrate has been added, and this corresponds with ^x8 = 2'7 c.c. of the potassium cyanide. Hence the nickel corresponds with 31 less 2'7 = 28-3 c.c. of potassium cyanide, or 28'3 x 0'0037, i.e. 0105 grm. of NiO in the given solution. Influence of Foreign Substances? Free alkalies, alkaline earths, alkaline car- bonates, chlorides, bromides, phosphates, have no marked effect. If not present in excessive amounts, chromates, manganese, cobalt, arsenic, bismuth, molybdenum, tin, lead, and uranium have no marked effect. With manganese the indication is 1 The ammonium sulphate so formed makes the indicator more sensitive. If barium, or elements which form insoluble sulphates, be present, use nitrates or chlorides, not sulphates. 2 Since 1 grm. of silver nitrate corresponds with 076 grm. of potassium cyanide, 1 c.c. of the silver nitrate, containing the equivalent of O'OOS grm. of silver iodide, will require 0'00445 grin, of potassium cyanide, or 0*345 c.c. of the solution of potassium cyanide described below. The silver nitrate should be standardised by a blank experiment without nickel. Find how many cubic centimetres of the standard solution of potassium cyanide are required to dissolve the silver iodide corresponding with 1 c.c. of silver nitrate solution. Do not trust to the calculation. 3 In some cases, e.g. in the presence of manganese, if the titration be left partly completed, a precipitate may separate. This interferes with the result later on. Hence, the titration should be completed without delay. The ammonium sulphate retards the formation of the precipitate. 4 STANDARD POTASSIUM CYANIDE. Make 12'9 grms. of potassium cyanide and 5 grms. of potassium hydroxide up to a litre. The potassium hydroxide makes the solution more stable. The cyanide can be standardised by dissolving 0'08 grm. of nickel in sulphuric acid, or '539 grm. of nickel ammonium sulphate in water, and proceeding with the solution as described above. The cyanide should be freed from sulphur. This is sometimes effected by agitating the solution with lead monoxide, or bismuth oxide. 5 H. Brearley and F. Ibbotson, The Analysis of Steel Works Materials, London, 139, 183, 1902 ; C. M. Johnson, Journ. Amer. Chem. Soc., 29. 1201, 1907. 26 402 A TREATISE ON CHEMICAL ANALYSIS. not quite so sharp or delicate. The manganese may be prevented from precipi- tating during the titration by the addition of ammonium chloride. For lead, use ammonium nitrate instead of sulphate, as indicated, page 336. If arsenic and antimony be present, use a little tartaric acid. Arsenic and tin should be oxidised with a little nitric acid and bromine water, and the excess boiled off. The results with iron, aluminium, chromium, and zinc are low. This is due to the precipitate 1 carrying down some nickel. If, therefore, these elements can be kept in solution, the titration is satisfactory. The chromium should be oxidised to chromate. Moore recommends the use of sodium pyrophosphate for keeping the zinc, iron, and aluminium in solution ; while Brearley and Ibbotson recommend tartaric or citric acids to minimise the chances of error. Brearley and Ibbotson accordingly recommend the following procedure when iron and manganese have not been separated : Add 3 grms. of citric acid and 2 grms. of ammonium sulphate per gram of sample. Add enough ammonia to make the solution feebly but distinctly alkaline. Add 2 c.c. of a 2 per cent, solu- tion of potassium iodide, and silver nitrate until the turbidity is apparent. Titrate with potassium cyanide until the turbidity dis- appears. The alkaline citrate solution, con- taining iron in solution, is dark-coloured, and the changes are not easy to distinguish. Hence, a strong beam of light should be sent across the solution. Fig. 139 shows Lupp's arrangement, useful for titrations which re- quire strong illumination to distinguish the end point. The light is focussed in the body of the liquid from a reflecting mirror. 2 If chromium be present, the turbidity disappears before the formation of the double nickel potassium cyanide is completed, but it soon returns. The phenomena of appearance and disappearance of the turbidity recur a number of times, and it is difficult to decide just when the reaction is at an end. Hence, Johnson 3 prefers to use 12 grms. of citric acid instead of 3 grms. Johnson's plan then enables him to say that the titration can be accomplished " at almost the full speed of the burette." If zinc be present, proceed as for iron and aluminium, but make the solution alkaline with sodium carbonate instead of ammonia. Copper should be removed by the thiocyanate process (page 350) : acidify the solution with nitric acid and proceed as described above. If the amount of the other interfering elements cobalt, silver, gold, etc. be known, an allowance can be made ; otherwise, the result of the titration gives nickel plus cobalt, silver, gold, and platinum, if these be present. Cobalt darkens the solution, owing to the oxidation of the cobaltocyanide to cobalticyanide. If the amount of cobalt be less than 10 per cent, of the total nickel, the sum Ni + Co can be determined by the titration, and the cobalt determined separately. Or the nickel can be precipitated with dicyari- diamide or dimethylglyoxime ; dissolved in warm dilute hydrochloric acid ; and the ainmoniacal solution titrated as described above. FIG. 139. Illumination of turbid solutions during titration. 1 H. Brearley, Chem. News, 74. 16, 1896 ; G. T. Dougherty, Iron Age. 70. 1274, 1907. 2 A. Lupp, Zeit. anal. Chem., 34. 182, 1895. 3 C. M. Johnson, Rapid MetJwdsfor the Chemical Analysis of Special Steels, New York, 105, 1909. PART IV. SPECIAL METHODS-BASES. CHAPTER XXX. THE DETERMINATION OF MOLYBDENUM, TUNGSTEN, NIOBIUM, AND TANTALUM. 204. Molybdenum, Tungsten, Niobium, and Tantalum in Silicate Analyses. THE "opening up" of minerals containing tungsten, molybdenum, niobium, and tantalum 1 follows the lines indicated in different parts of this work (pages 164 and 266). First the acids alone or mixed 2 are tried. If these be impotent, the fluxes sodium peroxide, 3 potassium hydroxide, potassium bisulphate or pyrosulphate, 4 etc. are tried. Potassium hydroxide (or carbonate) is recom- mended as a general flux for the minerals containing the elements discussed in this chapter. The reason will appear later. The oxides of these elements form soluble salts when fused with an excess of potassium carbonate or hydroxide. When the aqueous solution is treated with a slight excess of hydrochloric acid, a certain amount of the oxides separates, and if more acid be present, the soluble salts are decomposed, and the separation of the oxides is often practically complete. In silicate analyses, therefore, the oxides of these elements may be found with the silica when the acid solution is evaporated to dryness. 5 If tungsten be the only member of this series present, the problem resolves into the separation of silica from tungsten oxide. Some molybdenum, if present, will separate with the silica as silicomolybdate. Niobium, tantalum, antimony, 6 tin, 7 arsenic, 8 and phosphorus, 9 if present, may also be associated with the silica. Niobium, by the way, is often called columbium and symbolised Cb. In dealing with complex mixtures of this type, it is generally best to extract the sodium or potassium carbonate fusion with water. The aqueous solutions may contain one or more of the following : alkaline tungstate, molybdate, stannate, sulphate, aluminate, chloride, fluoride, arsenate, phosphate, antimoniate, chromate, vanadate, niobate, tantalate, silicate, etc., provided certain combina- tions which form insoluble precipitates are absent. The residue may contain alkaline zirconate, ferric oxide, calcium, barium, and strontium carbonates, etc. 1 For tantalum in clays, see A. Terrell, Compt. Rend., 51. 94, 1860. 2 For hydrochloric acid and aqua regia, see H. Bartonec, Oestr. Chem. Ztg., 12. 114, 1909. H. W. Hutchin and F. J. Fouks (Trans. Inst. Min. Met., 18. 425, 1909) recommend boiling alkaline hydroxide (25 per cent, solution). 3 E.g. W. Hempel (Zeit. anorg. Chem., 3. 193, 1895) decomposed wolframite .in a few minutes by heating with four times its weight of sodium peroxide. 4 E.g., H. Cremer, Eng. Min. Journ., 59. 345, 1895. 5 Molybdic oxide is soluble in an excess of acid. 6 Antimony oxychloride. 7 Stannic chloride gives a yellow precipitate with alkaline tungstates. 8 Tin or tungsten arseniate. 9 Tin or tungsten phosphate ; and phosphomolybdates. 4OC 406 A TREATISE ON CHEMICAL ANALYSIS. The treatment for the separation of the different combinations revealed by a qualitative analysis may be based on the processes indicated in this work. 205. The Detection of Tungsten, Tantalum, Molybdenum, and Niobium. Fuse, say, half a gram l of the finely powdered mineral with six times its weight of potassium hydroxide in a silver or nickel crucible. Extract the cold mass with hot water. Treat the filtered solution with 25 c.c. of hydrochloric acid (sp. gr. 1*16). Boil the solution, and filter. The residue may contain niobium, tantalum, tin, tungsten, silica, molybdenum, and antimony. Digest the moist precipitate with yellow ammonium sulphide. Molybdenum, tin, tungsten, and antimony dissolve ; niobium and tantalum remain. The ammoniacal sulphide solution containing molybdenum, tin, tungsten, and antimony is acidified with hydrochloric acid and boiled ; digest the washed pre- cipitate with a little hydrochloric acid and a little nitric acid ; filter and wash. The precipitate contains sulphur and tungsten, if present. Treat the solution with a few pieces of metallic zinc. Antimony and tin 2 are precipitated, molybdenum and zinc remain in solution. If the solution is blue, molybdenum is probably present. Molybdenum. Evaporate the solution along with a little nitric acid to dryness. Dissolve the residue in ammonia. (1) Pour a portion of the solution into moderately concentrated hydrochloric acid, add a few drops of a solution of potassium thiocyanate; if no red colour develops, iron is absent. Place a little metallic zinc in the acid solution ; the development of a carmine-red colour indicates the presence of molybdenum. 3 (2) Another portion of the solution is heated with a few drops of sulphuric acid on the lid of a platinum crucible until it begins to fume. When cold, an ultramarine blue colour will develop, when the acid is brought in contact with a little alcohol, if molybdenum be present. 4 Tungsten. (1) The residue of tungsten and sulphur indicated above, or tungstic oxide itself, gives a blue solution when digested with a little zinc in the presence of hydrochloric or sulphuric acid. (2) Stannous chloride gives a yellow pre- cipitate with a soluble tungstate ; 5 if the solution be acidified with hydrochloric acid and boiled, a blue colour is formed. (3) Defapqz's test. ( > Heat a mixture of the oxide with four or five times its weight of potassium bisulphate and a few drops of sulphuric acid in a platinum capsule. Add enough sulphuric acid to prevent the mixture solidifying when it cools. On cooling, add a drop of this 1 More than this is sometimes required, say, 5 grms., but only if small quantities of these elements are present. P. Nicolardot, Compt. Rend., 144. 859, 1907. 2 Note, if arsenic and antimony be present, poisonous hydrides may be evolved. For the recognition of tin and antimony, see any of the Text-books on Qualitative Analysis. 3 The blood-red tint with ferric salts is destroyed by the addition of phosphoric acid. This is not the case with molybdenum R. Fresenius, Qualitative Chemical Analysis, London, 185, 1897. 4 In the event of antimony or tin being present with the molybdenum, the mixture should be evaporated to dryness with phosphoric acid before it is treated with the sulphuric acid R. Fresenius, Qualitative Chemical Analysis, London, 185, 1897. For the sulphuric acid test, see H. Schonn, Zeit. anal. Chem., 8. 379, 1869 ; 0. Maschke, ib., 12. 383, 1873 ; Arch. Pharm. (3), 6. 125, 1874; F. von Kobell, Zeit. anal. Chem., 14. 317, 1874; Chardkteristik der Mineralien, Nuremberg, 109, 1831 ; Grundzuge der Mineraloyie, Nuremberg, 284, 1338. 5 Fuse the residual oxide with an excess of sodium carbonate, and take the mass up with water to get a soluble sodium tungstate. 6 E. Defayqz (Compt. Rend., 123. 308, 1896 ; Chem. News, 74. 88, 1896) claims that this reaction will indicate the presence of between 0*0000025 and 0*000002 grm. of tungsten, whereas the blue colour developed by zinc and hydrochloric acid will only detect O'OOl grm. MOLYBDENUM, TUNGSTEN. NIOBIUM, AND TANTALUM. 407 solution to a drop of phenol ; if tungsten be present, an intense red colour is developed; if hydroquinone be used in place of phenol, an amethyst-violet coloration is formed with concentrated solutions, and a rose colour with dilute solutions. 1 Titanium and vanadium interfere with the test. Niobium and Tantalum. Digest the insoluble residue, 2 in- dicated above, in a platinum capsule with a little hydrofluoric acid in slight excess; add a saturated solution of potassium fluoride. Evaporate the solution to a very small volume. Let the solution cool slowly. If acicular rhombic crystals of potassium fluotantalate (fig. 140) separate, tantalum is present; if plates of potassium-niobium oxyfluoride separate (fig. 141), niobium is also present. A good hand lens is a convenient means of recognising the crystals if they are small. It must be added that isomorphous potassium fluotantalate, K 2 TaF r , and potas- sium fluoniobate, K 2 NbF 7 , separate from concentrated solutions of hydrofluoric acid ; 3 and if an excess of hydrofluoric acid be not present, the solution containing niobium alone fur- nishes the plates of oxyniobate, K 2 NbOF 5 .H 2 0. By boiling late K 2 TaF 7 . dilute aqueous solutions of both salts hydrolysis occurs, and tantalum and niobium oxyfluorides are formed. The latter is much more soluble than the former, so that the appear- ance of a turbidity when dilute aqueous solutions of the two salts are boiled enables small traces of tantalum to be detected in the presence of niobium. Pick out a few of the acicular crystals and heat them on FIG. 141. Potassium- ^ ne lid of a platinum crucible with concentrated sulphuric niobium oxyfluoride acid (sp. gr. 1'29) until the acid fumes strongly, so as to K 2 NbOF 5 . H 2 0. drive off the hydrofluoric acid. Treat the cold mass with an excess of water in a small glass capsule, and boil the solution so as to precipitate the tantalic oxide. The precipitate dissolves in an excess of hydrochloric acid, giving an opalescent solution. Metallic zinc along with hydrochloric acid does not colour the solution ; tannic acid with a drop of the solution gives a light brown precipitate if tantalum be present. 4 If a few of the crystalline plates of potassium fluoniobate or potassium niobium oxyfluoride be treated in a similar manner, and the precipitate dissolved in sulphuric acid, metallic zinc will give a blue coloration ; 5 and tannic acid, an orange-red coloration. 6 fluotanta- 206. The Determination of Tungsten as Tungsten Trioxide. If the substance under investigation can be decomposed by an acid, 7 tungsten trioxide W0 3 will remain as an insoluble yellow powder. Repeated 1 For quinine and strychnine, see F. Scheibler, Journ. prakt. Chem. (1), 80. 204, 1860. 2 Antimony can be detected in the residue by fusion with potassium cyanide. Wash. Digest the insoluble matter with hydrochloric acid (and a crystal of potassium chlorate). The antimony passes into solution. The washed residue is treated as described in the text. 3 The form of the crystals is also modified by the nature of the mother liquid, and the .temperature of crystallisation C. Marignac, Bibl. Univ. Arch. Geneve, 23. 249, 1865 ; ib., 26. 89, 1866 ; (Euvres Completes, Geneve, 2. 258, 384, 1894. 4 Tungstic oxide under similar conditions gives a brown precipitate with tannic acid. 5 See above for a blue coloration with tungstic acid under similar conditions. 6 Molybdic acid gives an orange-red coloration with tannic acid M. E. Pozzi-Escot, Compt. Rend., 138. 200, 1904 ; vanadic acid, a blue coloration C. Matignon, ib., 138. 82, 1904. 7 K. W. Scheele, 1781 Opuscula Chemica et Physica, Lipsise, 2. 119, 1789 used nitric or hydrochloric acids ; F. Margueritte (Ann. Chim. Phys. (3), 17. 475, 1846), dilute sulphuric 408 A TREATISE ON CHEMICAL ANALYSIS. evaporation of a soluble tungstate (say, three times) with nitric acid or hydro- chloric acid, washing, and baking at 110 to 120 will render the tungstic oxide insoluble in acids. The dry mass is moistened with concentrated nitric acid, and after a 15 minutes' digestion add, say, 20 c.c. of a hot 5 per cent, solution of ammonium nitrate. Filter. 1 Wash with the 5 per cent, ammonium nitrate solution acidified with a few drops of concentrated nitric acid until all the alkali is removed. Dry. The filter paper must be ignited separately to avoid reduction ; hence, transfer the powder to a watch-glass, and ignite the filter . paper in a weighed porcelain crucible, then transfer the powder carefully from the watch-glass to the crucible, and ignite until the greenish tinge passes to a clear yellow. If the green colour persists, add a couple of drops of concentrated nitric acid, and repeat the ignition. 2 Weigh the ignited mass as tungstic oxide -W0 3 . Unlike the corresponding molybdenum oxide, this oxide can be calcined at the highest temperature of a Bunsen's burner without fear of appreciable loss by volatilisation. If the ignition be conducted over a blast, a 20 minutes' ignition is said to have caused a loss of about 7 per cent. Thus, tungstic oxide weighing 0'3007 grm. was reduced after 20 minutes' blasting to 0*2872 grm. 3 207. The Gravimetric Determination of Tungsten Berzelius' Process. A concentrated alkaline solution of the tungstate is neutralised with nitric acid, and a few drops of -nitric acid in excess are added. The solution is there- fore feebly acid. Add an excess of mercurous nitrate 4 solution. Agitate the solution. Then add ammonia, drop by drop, with constant stirring, 5 until a brown precipitate separates. Heat the solution to boiling. Let the precipitate settle. Filter and wash with water containing 2 per cent, of mercurous nitrate. Dry, ignite, and weigh the residual tungstic oxide as W0 3 . 6 acid. R. Hermann, Zeit. anal. Chem., 51. 736, 1912 ; C. Scheibler, Jo-urn, prakt. Chem. (1), 83. 273, 1861 ; Chem. News, 6. 182, 1862. See J. W. Mallet, ib., 31. 276, 1877, for the solubility of tungstic acid when an excess of concentrated hydrochloric acid is added to an alkali tungstate. 1 A Gooch's crucible can be used if the precipitate is not to be subjected to further treatment. ' 2 If the alkalies have not all been removed in the washing, the green tinge cannot be always removed by this treatment. H. L. Wells and F. J. Metzger, Journ. Amer. Chem. Soc., 23. 356, 1901. 4 MERCUROUS NITRATE SOLUTION. Digest 60 grins, of pure mercury with 25 c.c. of nitric acid (sp. gr. 1'4) and 75 c.c. of water on a water bath for about 1| hours. Let the mixture stand overnight. Dilute the solution to 400 c.c. (|E). The addition of about 20 c.c. usually suffices for the precipitation. The solution should give no residue when evaporated to dryness and the mercurous nitrate volatilised. 5 Some prefer to add the mercurous nitrate solution directly to the alkaline solution. 0. F, von Pfordten (Liebig's Ann., 222. 152, 1883 ; Chem. News, 50. 18, 1884 ; Zeit. anal. Chem., 24. 92, 1885) says the mercurous nitrate may be added to the neutral solution if it is free from carbon dioxide. 6 J. J. Berzelius, Schweigger's Journ., 16. 476, 1816; W. W. Hutchin, Analyst, 36. 398, 1911 ; E. Bagley and H, Brearley, Chem. News, 82. 270, 1900; L. Desvergnes, Ann. Chim. Anal., 9. 321, 1904 ; W. Gibbs, Amer. Chem. Journ., I. 219, 1879 ; H. W. Hutchin and F. J. Fouks, Inst. Min. Met., 18. 425, 1909. Instead of mercurous nitrate, lead acetate is sometimes used e.g. T. M. Chatard, Amer. J. Science (3), i. 416, 1871 ; Chem. News, 24. 175, 1871 ; F. A. Bernoulli, Poqg. Ann., in. 573, 1860 ; Chem. News, 5. 116, 1862 ; H. H. Brearley, ib., 79. 64, 1899 ; F. Ibbotaon and H. Brearley, ib., 80. 293, 1899 ; E. Zettnor, ib., 16. 12, 1867 ; J. S. de Benneville, Journ. Amer. Chem. Soc., 19. 377, 1897. Other organic precipitating agents have been recommended, e.g. quinine acetate or sulphate (J. Lefort, Journ. Pharm. Chim. (3), 4. 221, 326, 1881; Chem. News, 45. 57, 1882); cinchonine (H. Cremer, Eng. Min. Journ., 59. 345, 1895); benzidine (G. von Knorre, Ber., 38. 783, 1905 ; Zeit. anal. Chem., 47. 37, 1908 ; Stahl Eiscn, 28. 984, 1908 ; H. Wdowiszewski, Chem. Ztg., 34. 1365, MOLYBDENUM, TUNGSTEN, NIOBIUM, AND TANTALUM. 409 208. The Separation of Tungsten and Silica. Hydrofluoric Acid Process. This is best effected by the volatilisation of the silica in the usual manner by treatment of the mixture with sulphuric and hydrofluoric acids (page 169). Prolonged ignition is not always sufficient to drive off the last traces of sulphuric acid. In that case, a piece of pure solid ammonium carbonate in the crucible will remove the acid in question. 1 Wells and Metzger 2 have shown that Herting's assertion 3 that some tungstic oxide is volatilised by this treatment is unfounded. The mixture, however, should not be blasted, for the reason indicated above. Ammonia Process. The method 4 of separating tungstic oxide from silica based on the solubility of moist tungstic oxide in ammonia is not to be used if it can be avoided, since the silicic acid is slightly soluble in the same menstruum (page 172) ; but the error is relatively small. If the mixture be baked, so as to reduce the solubility of the silica, some of the tungstic oxide also becomes insoluble. Potassium Bisulphate Process. The methods 5 of separation depending upon the fusion of tungstic oxide and silica with, say, 5 parts of potassium bisulphate, and subsequent digestion of the residue with ammonium carbonate or ammonium sulphide, (NH 4 ) 2 S, 6 to dissolve the tungstic oxide, take no account of the slight solubility of silica in potassium bisulphate (page 186), and of the failure of these menstrua to dissolve all the tungstic oxide. 7 The errors are small, and the experimental results, at present, meet commercial requirements. 209. The Separation of Tungsten and Tin. Rose and Rammelsberg's Process* Rose 8 has shown that stannic oxide may be removed from tungstic oxide by repeated ignition with ammonium chloride. The ammonium chloride reacts with stannic oxide, forming a volatile stannic chloride, while the tungstic oxide remains behind. Rammelsberg 9 recommends the following method of conducting the operation : Mix the weighed residue from the hydrofluoric acid treatment with six to eight times its weight of ammonium chloride. 10 Place the covered crucible containing the mixed stannic and tungstic 1910; F. W. Hinrichsen, Mitt. Konig. Materialpruf. Gross. Lichter. West., 25. 308, 1907) also used for the separation of tungsten from phosphoric acid; and a-naphthylamine (M. Tschilikin, Ber., 42. 1302, 1909). B. Mdivani (Bull. Soc. Chim. (4), 9. 122, 1911), precipi- tates W 2 5 by adding a solution of stannous chloride to a soluble tungstate. 1 H. Rose, Ausfilhrliches Handbuch der atialytischen Chemie, Braunschweig, 2. 343, 1851. 2 H. L. Wells and F. J. Metzger, Journ. Amer. Chem. Soc., 23. 356, 1901 ; Chem. News, 83. 3, 1901 ; A. G. McKenna, Proc. Eng. Soc. Pennsylvania, 16. 119, 1900 ; Eng. Min. Journ., 66. 607, 1898. 3 0. Herting, Zeit. angew. Chem,, 14. 165, 1901 ; Chem. News, 84. 75, 1901. 4 H. Borntrager, Zeit. anal. Chem., 39. 361, 1900; J. Preusser, ib., 28. 173, 1889; Chem. News, 60. 37, 1889 ; S. Kern, ib., 35. 67, 1877 ; J. Parry and J. J. Morgan, ib., 67. 260, 1893 ; A. Cobenzl, Monats. Chem., 2. 259, 1881 ; J. S. de Benneville, Journ. Amer. Chem. Soc., 19. 377, 1891 ; H. Cremer, Eng. Min. Journ., 59. 345, 1895 ; H. F. Watts, Chem. News, 95. 19, 1907; L. and G. Ca'mpredon, Ann. Chim. Anal., g. 41, 1904; L. Wolter, Chem. Ztg., 34. 2, 1910; R. Namias, Stahl Eisen, n. 757, 1891. 5 L. Schneider and F. Lipp, Zeit. anal. Chem., 24. 292, 1885 ; Chem. News, 51. 297, 1885 ; R. Schoffel, ib., 41. 31, 1880 ; Ber., 12. 1866, 1879 ; 0. Herting, Zeit. angew. Chem., 14. 165, 1901 ; C. Marignac, Ann. Chim. Phys. (4), 3. 6, 1843. 6 H. Rose, Pogg. Ann , 100. 146, 1857. 7 R. D. Hall, Journ. Amer. Chem. Soc., 26. 1235, 1904 ; E. F. Smith, Proc. Amer. Chem. Soc., 44. 151, 1905. 8 H. Rose, Ausfuhrlich.es Handbuch der analytischen Chemie, Braunschweig, 2. 352, 1871 ; W. P. Dexter, Pogg. Ann., 92. 335, 1854. 9 0. Rammelsberg, Pogg. Ann., I2O. 66, 1864 ; Chem. Neivs, g. 25, 1864. 10 The ammonium chloride must be tested to make sure that it yields no non-volatile con- stituents when heated in a platinum dish. 410 A TREATISE ON CHEMICAL ANALYSIS. oxides in a larger crucible. The latter is covered with a lid. Heat the crucibles as long as vapours of ammonium chloride issue from the outer crucible. Repeat the treatment three times. The object of using two crucibles is to prevent the formation of a coat of stannic oxide on the outside of the smaller crucible. The stannic oxide is formed by a reaction between the vapours of stannic chloride which issue from the crucible and the moisture of the air. The contents of the inner crucible become green and finally almost black in colour. The yellow colour is restored when the small crucible is ignited while exposed to the air. If a fourth treatment with ammonium chloride gives the same weight as the third, it is assumed that all the stannic oxide has been driven off. The inner crucible is then ignited alone and finally weighed as indicated in 206, page 408. The following numbers, by Rammelsberg himself, illustrate the accuracy of the process : Stannic oxide taken .'"'.. . . . 0'6977 0'554 grm. Tungstic oxide taken ; . V . . 07335 1'332 grm. Tungstic oxide found . . ... . 07255 V337 grm. Donath and Mutter's Process. According to Rammelsberg, 1 Rose's method of reduction by heating a mixture of tin and tungsten oxides in a current of hydrogen, and dissolving out the reduced tin with hydrochloric acid, is not very exact. 2 Donath and Miiller 3 obtained better results by mixing the two oxides with twice their weight of zinc dust, and, after 15 minutes' ignition in a covered crucible, dissolving out the tin by boiling with dilute hydrochloric acid (1 : 2). Add sufficient potassium chlorate to the cold solution to oxidise the blue tungsten oxide to the yellow oxide. Dilute the solution with 1-5 times its volume of water, and after 24 hours, filter, ignite, and weigh the precipitate as W0 3 . The tin can be determined in the filtrate by precipitation with hydrogen sulphide, etc. 210. The Separation of Tungsten from Tin and Antimony Talbot's Process. The two processes which precede will sometimes remove any antimony which may be present, but Hallopeau 4 has shown that when a mixture of soluble anti- mony and tungsten salts is treated with mercurous nitrate, a mercurous anti- monio-tungstate 3HgSb0 3 . 2W0 3 is precipitated, and this, on ignition, furnishes antimony tetroxide and tungsten trioxide. The separation, according to Hallopeau, is best effected by Talbot's process for the separation of tungsten and tin. 5 In Talbot's process 6 the mixed oxides 7 are fused with potassium cyanide as indicated on page 269. If too little potassium cyanide be employed, a black residue containing tungsten may be formed. One part of the mixed oxides with 1 C. Rammelsberg, Pogg. Ann., 120. 66, 1864 ; Chem. News, 9. 25, 1864. 2 For reduction by hydrogen between 600 and 900, forming either a lower oxide or metal, see C. Marignac, Ann. Chim. Phys. (4), 3. 9, 1864 ; E. Defagqz, Compt. Rend., 146. 1319, 1908 ; 144. 848, 1907 ; C. Friedheim, W. H. Henderson, and A. Pinagel, Zeit. anorg. Chem., 45. 396, 1905. The process is recommended for the separation of tungstic oxide and silica by L. E. l-fc . . J m T\l - /rt\ --. TOO *Or/\ A TT'l 1 TT T1 f- j"si-ts\^:r. Rivot, Ann. Chim. Phys. (3), 30. 188, 1850 ; A. Hilger and H. Haas, Eer., 23. 458, 1890. 3 E. Donath and F. Miiller, Monats. Chem., 8. 647, 1887; Chem. News, 59. 73, 1889; E. Donath, Zeit. anqew. Chem., 19. 473, 1906; H. Angenot, ib., 19. 956, 1906; J. Preusser, Zeit. anal. Chem., 28. 173, 1889. 4 L. A. Hallopeau, Bull. Soc. Chim. (3), 17. 170, 1897. 5 Hallopeau found that the fusion of the mixture of antimony and tungsten oxides with sodium hydroxide and extraction of the sodium tungstate with alcohol is not satisfactory. Some sodium antimoniate passed into solution. 6 J. H. Talbot, Amer. J. Science (2), 50. 244, 1870 ; Chem. News, 22. 229, 1870 ; R. Helmhacker, Eng. Min. Journ., 60. 153, 1896 ; B. Setlik, Chem. Ztg., 13. 1474, 1889 ; Chem. News, 61. 54, 1890 ; E. D. Desi, Journ. Amer. Chem. Soc., 19. 239, 1897. 7 Tin and tungsten ; antimony and tungsten ; or tin, antimony, and tungsten. MOLYBDENUM, TUNGSTEN, NIOBIUM, AND TANTALUM. 411 1 2 parts by weight of potassium cyanide usually suffices. The tin and antimony remain behind as metals ; the alkaline tungstate formed at the same time passes into solution when the mass is leached with water. Filter and wash. The aqueous solution is boiled with an excess of nitric acid l to drive off the volatile cyanogen compounds. The tungstate is then precipitated by the usual process. If phosphorus be present, it will be found in the solution with the tungsten. The separation is described below. The metallic bead is dissolved in acid, and analysed in the ordinary manner mixture of tin and antimony. 2 211. The Separation of Tungsten from Arsenic and Phosphorus Kehrmann's Process. The separation of arsenic and tungsten is exceedingly difficult, because part of the arsenic is retained very tenaciously by the tungsten as a complex salt. The distillation process for separating arsenic and tungsten does not give satisfactory results. 3 The following process is due to Kehrmann. 4 The same process can be applied, mutatis mutandis, to the separation of tungsten and phosphorus. Fuse between 1 and 2 grms. of the sample with twice as much sodium hydroxide as is required to combine with the arsenic oxide. Add enough water to dissolve the resulting cake. Boil the solution in an Erlenmeyer's flask for about half an hour. Cool. Add three times as much ammonium chloride as is needed to form chlorides with the alkalies present. Add ammonia about one fourth the volume of the solution under investigation and cold magnesia mixture. In about 12 hours the precipitate of ammonium magnesium arsenate is filtered and washed with a solution of ammonia and ammonium nitrate (page 285). It is advisable to dissolve the precipitate in dilute acid, repeat the precipitation a number of times, 5 and treat the precipitate as indicated on page 285. The tungstic acid is separated from the combined filtrates by evaporation with nitric acid (page 408). The precipitate is washed with a mixture of nitric acid and ammonium nitrate, and weighed as indicated on page 408. It is difficult to get the tungsten quite free from magnesia. Gooch 6 first determines the total weight of tungsten and arsenic, then separates the tungsten by mercurous nitrate process, and estimates the arsenic by difference. 212. The Precipitation of Molybdenum as Sulphide. Molybdenum is precipitated from an acid solution as sulphide^ MoS 2 in the hydrogen sulphide group along with copper, etc. It is almost impossible to 1 Note the poisonous fumes of the cyanogen compounds which are, evolved render it necessary to perform the operation under a hood. 2 If any antimony adheres tenaciously to the walls of the crucible, remove the metal by treatment with acids ; or, if antimony or tin alone be present, wash the crucible, dry and weigh. Add some ammonium chloride, and heat the crucible to redness ; the antimony volatilises. The loss in weight represents the amount of metal which adhered to the crucible. 3 C. Friedheim and P. Michaelis, Ber., 28.. 56, 1888 ; but see S. Hilpert and T. Dieckmann, Ber., 46. 152, 1913. 4 F. Kehrmann, Liebig's Ann., 245. 56, 1888; Ber., 2O. 1813, 1887. F. W. Hinrichsen (MM. Konigl. Malerialprufungsamt., 28. 229, 1910), for the separation of both phosphorus (P 2 5 ) and tungsten (W0 3 ). recommends the mercurous nitrate process (page 408) ; the ignited precipitate is fused with alkali carbonate. The phosphorus can be precipitated from the aqueous solution of the cold cake by first precipitating the phosphates with magnesia mixture, and after taking up the precipitate with nitric acid, re precipitate (page 598). 5 The arsenic and phosphorus can be separated in the usual manner if they are present together. 6 F. A. Gooch, Proc. Amer. Acad., 16. 134, 1881 ; Amer. Chem. Journ., I. 412, 1879 ; W. Gibbs, ib., 7. 337, 1885; H. Bullnheimer, Chem. Ztg., 24. 870, 1900; Chem. News, 85. 184, 1902. 412 A TREATISE ON CHEMICAL ANALYSIS. effect complete separation in a hydrochloric acid solution. The separation is more complete in sulphuric acid solutions, but even then the separation is not satisfactory in a reasonable time. For instance, a solution of ammonium molybdate acidified with five volumes of concentrated sulphuric acid was saturated with hydrogen sulphide. The solution was filtered after it had been allowed to settle for an hour, and the filtrate was again treated with hydrogen sulphide. This sequence of operations was repeated in all five times. A little molybdenum sulphide was precipitated each time. Hence, molybdenum, in acid solutions, is but imperfectly precipitated by hydrogen sulphide at atmospheric pressures. 1 If a pressure bottle be employed, the results are satisfactory. The pressure flask (page 277) of 300, 500, or 1000 c.c. capacity is closed by a ground-glass plate. 2 About 250 c.c. of the solution are placed in the flask, and saturated with hydrogen sulphide by passing a rapid stream of gas through the solution. 3 Dilute the solution to about 500 c.c. with water saturated with hydrogen sul- phide. The flask is then placed in a cold water-bath and gradually heated to boiling. The bath is kept at that temperature for about an hour. Let the bottle cool ; empty the contents into a beaker ; and wash the bottle with dilute acid (sulphuric acid 1, water 50) saturated with hydrogen sulphide. Let the precipitate settle. Filter, and wash the precipitate with jbhe dilute acid saturated with hydrogen sulphide. If several members of the hydrogen sulphide group be present, the precipi- tated sulphides are digested with ammonium monosulphide for a couple of hours, when molybdenum, arsenic, antimony, and tin, 4 if present, will pass into solu- tion. Vanadium and uranium, if present, are not precipitated by the above treat- ment. Tungsten is partly precipitated, but not if 3 to 5 grms. of tartaric acid be added to the solution before passing the hydrogen sulphide. 5 If much iron be present, an appreciable quantity may be carried down with the sulphides. For instance, a solution containing 0*1076 grm. of iron with different amounts of molybdenum in the same solution gave the following numbers : Molybdenum . . . 0*00492 0'00984 0*02460 0'04921 grm. Ferric oxide . . . 0*0001 0*0002 0*0003 0*0005 grm. when the iron was separated from the precipitated molybdenum sulphide and weighed as ferric oxide. To recover iron from the molybdenum sulphide, dissolve the washed precipitate in a mixture of 10 c.c. of hydrochloric acid, 5 c.c. of nitric acid, and 10 c.c. of concentrated sulphuric acid. Evaporate the solution until copious fumes of sulphur oxides are evolved. Cool. Add 50 c.c. of water and an excess of ammonia. The ferric hydroxide which separates may be filtered off. If no other member of the ammonium sulphide sub-group be present, the warm solution is acidified with hydrochloric or sulphuric acid, and the reddish- yellow solution is boiled to drive off the hydrogen sulphide. 6 Filter and wash 1 E. Dohler, Chem. Zeit., 24. 537, 1.900; Cfam. News, 82. 294, 1900; F. P. Treadwell, Kurzes Lehrbuch der analytischen Chem.ie, Leipzig, 2. 183, 1911 ; F. van Dyke Cruser and E. H. Miller, Journ. Amer. Chem. Soc., 26. 675, 1904 ; Chem. News, go. 204, 218, 1904. 2 An empty "citrate of magnesia " bottle makes a good pressure flask for the operation. 3 A fast current of gas causes less trouble by the sticking of the sulphide to the walls of the tube than a slow current of the gas. B. Herstein (Bull. U.S. Dept. Agric., Chem., 150. 44, 1912) uses 0*75 c.c. thioacetic acid per 0*1 grm. Mo0 3 in place of hydrogen sulphide. 4 Selenium, tellurium, and germanium, if present, will be found mainly in the ammonium sulphide solution. 5 H. Rose, Handbuch der analytischen Chemie, Braunschweig, 2. 358, 1871. If tungstic oxide be present, most of it will be found in the residue with the silica. The silica is driven off by means of hydrofluoric acid (page 169), and the residue is treated as indicated on page 408 for W0 3 . . 6 F. E. Zenker, Journ, prakt. Chem. (1), 58. 257, 1853. MOLYBDENUM, TUNGSTEN, NIOBIUM, AND TANTALUM. 413 the brown precipitate with hot water acidified with hydrochloric (or sulphuric) acid saturated with hydrogen sulphide. The precipitated molybdenum sulphide may now be treated in several different ways gravimetric or volumetric. Arsenic can be separated from molybdenum by the magnesia-mixture process described on page 283. Two or three precipitations are 'needed to obtain a precipitate of ammonium magnesium arsenate free from molybdenum. 1 Tin sulphide can be separated by treatment with a solution of oxalic acid slightly acidified with hydrochloric acid, as indicated in Clarke's process (page 296). 213. The Gravimetric Determination of Molybdenum as Oxide, and as Sulphide. Determination of Molybdenum as Oxide. If molybdenum alone be present, the sulphide can be conveniently transformed into the oxide by washing it first with dilute sulphuric acid (1 : 20), and finally with alcohol until all the acid is removed. The moist filter paper is placed in a large porcelain crucible and dried in an air bath. Carbonise the filter paper over a small flame while the cover is on the crucible. Remove the cover. Burn the carbon from the sides of the crucible with as small a flame as possible, and raise the temperature of the crucible very gradually. When the evolution of sulphur dioxide has ceased, cool. Add a little mercuric oxide suspended in water. Stir up the mixture in the crucible, and evaporate to dryness on the water bath. Drive off the mercuric oxide by gentle ignition. 2 Weigh as molybdenum trioxide Mo0 3 . The oxide should be white, not tinted blue. JSrrors. The chief difficulties in determining molybdenum as trioxide arise from the tendency of the sulphide to oxidise so violently as to project particles from the crucible during calcination ; and the tendency of the oxide to volatilise at comparatively low temperatures. 3 When the sulphide is calcined, a blue oxide is first formed. This contains less oxygen than the trioxide, and the cal- cination must therefore be continued until the mass in the crucible is white, not blue, when cold. This requires so prolonged a calcination that the oxide begins to volatilise before all the blue oxide is transformed into trioxide and the weight is constant. For instance, Collett and Eckardt 4 roasted the sulphide 5 at a low temperature and weighed the crucible and contents after the lapse of different intervals of time. The results are illustrated in fig. 142. The curve shows that the weight decreases rapidly as the sulphide is oxidised. Immediately after all has been converted into the trioxide (A, fig. 142), the oxide continues to lose weight, and after some time, about a couple of hours, the weight remains nearly 1 Arsenic can also be separated by the distillation process indicated on page 280. In order to avoid the introduction of large amounts of iron, C. Friedheim and P. Michaelis (Ber., 28. 1414, 1895) recommend distillation from methyl alcohol saturated with hydrochloric acid. 2 The mercuric oxide helps to burn off the carbon, and also assists in the removal of some sulphur oxide. 3 E. Collett and M. Eckardt, Chem. Ztg., 33. 968, 1909. M. Seligsohn (Journ. prakt. Chem. (1), 67. 472, 1856) fuses the sulphide with lead oxide and ammonium nitrate. The excess in weight over the amount of PbO used represents the molybdic oxide Mo0 3 ; E. Dohler (Chem. Ztg., 24. 537, 1900) prefers gentle ignition of the sulphide in a Rose's crucible in a current of hydrogen and weighing as MoS 2 ; or ignition in a current of hydrogen at a higher temperature and weighing as metal, as recommended by 0. F. von Pfordten (Ber., 17. 734, 1884) ; W. T. Taggart and E. F. Smith (Journ. Amer. Chem. Soc., 18. 1053, 1896) recommend igniting the sulphide with oxalic acid and weighing as Mo0 3 ; and C. Friedheim and H. Enler (Ber., 28. 2061, 1895) ignite the sulphide at a high temperature in air and weigh as Mo0 3 . 4 E. Collett and M. Eckardt, Chem. Ztg., 33. 968, 1909. 5 Similar results were obtained by calcining ammonium molybdate, and fig. 145 applies to both this salt and to the sulphide. A TREATISE ON CHEMICAL ANALYSIS. constant, although the weight of the substance is really 1-2 per cent, less than it should be. This is due to the volatilisation of the oxide, 1 although the oxide in the crucible still has a blue tint. If the temperature of calcination be raised, the loss by volatilisation is greater. This is illustrated by the quick descent of 60 50 40 20 JO Mo% __Cpristant_w too I2S 150 /75 200 225 rqinutes FIG. 142. Loss in weight during the ignition of molybdenum sulphide. the curve CD (tig. 142). With some practice, it is possible to conduct the operation quite satisfactorily. Collett and Eckardt, however, consider that the process should be abandoned, and the molybdenum weighed as sulphide. Determination of Molybdenum as Sulphide. This determination is less liable to error, and the result is rather more certain than the preceding method, but the apparatus required is a little more complex. The washed sulphide pre- cipitate is dried ; the precipitate separated from the paper, and preserved between two watch-glasses while the filter paper is burned to ash in a crucible. The sulphide is transferred to the same crucible, which is then fitted with a perforated lid (Rose's crucible, fig. 138, page 392), and the whole calcined to a constant weight in a current of hydrogen gas. The contents of the crucible are finally weighed as molybdenum sulphide MoS 2 . The equivalent weight of molybdenum trioxide is obtained by multiplying the weight of the sulphide by 0-8992. For instance, working with a known amount of molybdenum trioxide, it was found that : Mo0 3 used .... 01413 0'1422 0'1500 grm. MoS 2 found .... 0-1573 0'1578 01672 grm. Mo0 3 calculated from sulphide . 0-1414 0'1419 0'1503 grm. The results are therefore quite satisfactory. 214. The Gravimetric Determination of Molybdenum as Lead Molybdate. If a gravimetric process be required for the molybdenum in soluble salts like ammonium molybdate, dissolve, say, 0'5 grm. of the salt in 200 c.c. of water and add a few drops of ammonia, followed by 2 or 3 c.c. of (33 per cent.) acetic acid, 2 1 The sublimation of the oxide is shown by the presence of small glistening crystals above the mass in the crucible. 2 If more acetic acid be present, some lead molybdate may be dissolved. MOLYBDENUM, TUNGSTEN, NIOBIUM, AND TANTALUM. 415 and 4 or 5 grms. of ammonium chloride. Heat the solution to boiling, and while boiling add gradually, with constant stirring, 45-50 c.c. of a solution of lead acetate. 1 Boil two or three minutes more with vigorous agitation. The cream- coloured granular precipitate settles and filters rapidly. When the precipitate has settled, filter through an asbestos-packed Gooch's crucible, and wash by decantation with boiling water containing 5 grms. of ammonium chloride or ammonium acetate and five drops of acetic acid per 200 c.c. Transfer all the precipitate to the Gooch's crucible, and when the washings are free from lead, wash the precipitate twice with boiling water. Heat the crucible on an asbestos plate until dry, and then over a naked flame. Cool in a desiccator, and weigh as lead molybdate PbMo0 4 . Multiply the weight of the lead molybdate so obtained by 0*39236 to get the corresponding amount of Mo0 3 . Errors. If sulphates be present, as would be the case if the precipitated sulphide were dissolved in hot nitric acid, lead sulphate will be precipitated with the lead molybdate. Hence, it is necessary to roast molybdenum sulphide to oxide before it is dissolved in acid, preparatory to precipitating as lead molybdate. Cruser and Miller give the following trials with known amounts of molybdenum : Used . . . 0-14765 0'14793 0-17302 G'17211 (H9657 grin. Found . . . 0-14745 0-14793 0'17301 0-17210 0'19656 grm. Vanadium, if present, interferes with the test. According to Brearley, the presence of acetic acid, lead acetate, alkaline nitrates, chlorides, and acetates, salts of manganese, copper, cobalt, nickel, zinc, magnesium, mercury, barium, strontium, calcium, arsenic, cadmium, phosphorus, aluminium, and uranium do not interfere : silicates give slightly high results ; iron and chromium should be removed, by the addition of sodium hydroxide to the boiling solution. 2 215. The Volumetric Determination of Molybdenum by Potassium Permanganate. Dissolve the sulphide in a beaker or dish by digestion with a mixture of 10 c.c. of concentrated sulphuric acid, 10 c.c. of hydrochloric acid, and 5 c.c. of nitric acid. Evaporate 3 the solution until sulphurous fumes are evolved. Let the solution cool, and then neutralise it with ammonia, add 10 c.c. of concentrated sulphuric acid, and make the solution up to 200 c.c. When the solution has a temperature of 70-75, pour it through a reductor (page 191) with a column of granulated zinc 4 (20-30 mesh) about 37 or 38 cm. long. Arrange the suction so that it takes 6 minutes for the solution to pass through the reductor tube, 5 1 LEAD ACETATE SOLUTION. Dissolve 4 grms. of the salt in 100 c.c. of water. 2 For some properties of lead molybdate, see page 331. T. M. Chatard, Amer. J. Science (3), I. 416, 1871 ; Chem. News, 24. 175, 1871 ; P. Guichard, Compt. Rend., 131. 389, 419, 1900 ;" H. Brearley, Chem. News, 78. 203, 1898 ; 79. 2, 14, 1899; F. Ibbotson and H. Brearley, ib., 79. 3, 1899 ; 81. 269, 1900 ; Analysis of Steel Works Materials, London, 85, 273, 1902 ; L. Schindler, Zeit. anal. Chem., 27. 137, 1888 ; F. van Dyke Cruser and E. H. Miller, Journ. Amer. Chem. Soc., 26. 675, 1904. 3 The solution is inclined to bump badly. Blowing air through the solution during the evaporation prevents this. 4 F. Pisani, Compt. Rend., 59. 301, 1864 ; C. Rammelsberg, Pogg. Ann., 127. 281, 1866 ; A. Werneke, Zeit. anal. Chem., 14. 1, 1875 ; H. Borntrager, Zeit. anal. Chem., 37. 438, 1898 ; E. Knecht and F. W. Atack, Analyst, 36. 98, 1911. 5 There has been some discussion on the product obtained by reducing molybdic salts. Some say that Mo 2 3 is formed (W. A. Noyes and E. D. Frohman, Journ. Amer. Chem. Soc., 16. 553, 1894) ; others consider that Mo^O^ is obtained (A. A. Blair and J. E. Whitfield, Journ. 416 A TREATISE ON CHEMICAL ANALYSIS. Titrate the solution immediately with standard potassium permanganate. 1 The test experiments are excellent. 216. The Separation of Tungsten and Molybdenum Hommel's Process. Tungsten and molybdenum can be separated by the action of warm sulphuric acid (sp. gr. 1-378) upon the moist freshly precipitated 2 oxides Mo0 3 and W0 3 . 3 The former passes into solution, the latter remains insoluble. Hommel says that this process only gives satisfactory results when the moist oxides are first digested with concentrated sulphuric acid, and a few drops of dilute nitric acid, say, in a porcelain dish over a naked flame for about half an hour. 4 When the solution is cold, add about three times its volume of water. Filter, wash the precipitate with dilute sulphuric acid (1 : 20), and finally wash it two or three times with alcohol, ignite as indicated on page 408, and weigh as W0 8 . The molybdenum in the filtrate may be precipitated by hydrogen sulphide (page 277), or, if only small quantities are present, the solution may be evaporated to dryness and weighed in the platinum dish. Hommel's process is quite satis- factory. For example, Hommel obtained the following results with artificial mixtures of tungsten and molybdenum trioxides : Table LVI. Test Analyses, Tungsten and Molybdenum Mixtures. Tungstic trioxide. Molybdenum trioxide. Used. Found. Used. Found. 0-9616 0-8732 07029 0-5737 0'9614 0-8738 0-7035 0-5740 0-1077 0-0501 0-1030 0-1184 0-1080 0-0503 0-1028 0-1180 Amer. Chem. 8oc., 17. 747, 1895) ; others again say that the reduction does not go so far as this. E. H. Miller and H. Frank (ib., 25. 919, 1903) did not get Mo 2 3 , in the reductor, but they could find no definite stopping place at Mo 24 37 , although reduction under the usual conditions proceeds very close to this. With the reductor arranged as described in the text, the iron standard of the permanganate multiplied by 0'88 gives the equivalent Mo0 3 standard, and by 0*01579, the phosphorus standard. With amalgamated zinc, the factors were 0'8842 and 0-01586 respectively. If the reduction had been to Mo 24 O 37 , the factors would have been 0-8832 and '01584 respectively. It is best to standardise the potassium permanganate by the method of reduction with known solutions of molybdate. See page 599. 1 STANDARD SOLUTION OF AMMONIUM MOLYBDATE. Dissolve 6'132 grms. of ammonium molybdate (NH 4 ) 6 Mo 7 24 . 4H 2 in a litre of water. The solution contains the equivalent of 0*005 grm. Mo0 3 per c.c. The strength of the solution may be verified by precipitation of the molybdenum as lead molybdate page 414. The potassium permanganate solution is approxi- mately T VN, and it is standardised by running the standard ammonium molybdate solution through the reductor and titrating as described in the text. 2 If the oxides have been ignited, it is best to fuse them with from four to six times their weight of sodium carbonate, dissolve the fused mass in water, evaporate the solution to dryness, and add the concentrated sulphuric and nitric acid as described in the text. 3 E. D. Desi, Journ. Amer. Chem. Soc , 19. 213, 1897 ; M. Rueginberg and E. F. Smith, ib., 22. 772, 1900 ; Chem. News, 83. 5, 1901 ; W. Hommel, Ueber die quantitative Trennung von Wolfram und Molybddn, Zurich, 1902. 4 If the solution is greenish-coloured, add a drop or two of dilute nitric acid to oxidise the tungstic oxide to yellow tungstic acid. MOLYBDENUM, TUNGSTEN, NIOBIUM, AND TANTALUM. 417 Ease's Tartaric Acid Process. In Rose's method, 1 the alkaline salts under investigation are treated with an excess of sulphuric acid, and the tungsten is kept in solution by the addition of a considerable amount of tartaric acid 2 while the molybdenum is precipitated by hydrogen sulphide in a pressure flask. The tartaric acid is afterwards destroyed by repeated evaporation with nitric acid (page 208), and the tungsten finally separated in the usual manner as W0 ? . Friedheim and Meyer 3 have raised the question : Is all the molybdenum precipi- tated in the presence of the tartaric acid ? The trial results with known mixtures of molybdenum and tungsten are, however, satisfactory. The objection to the process rests on the time consumed in the destruction of the large quantities of tartaric acid. 217. The Separation of Tungsten and Molybdenum Pochard's Process. Pechard's process 4 is based on the fact that if a weighed mixture of tungsten and molybdenum (oxides or alkaline salts) be placed in a boat, and heated FIG. 143. Separation of Molybdenum from Tungsten. between 250 and 270 in a current of dry hydrogen chloride, the molybdenum is volatilised as Mo0 3 .2HCl and deposited in the form of long acicular white crystals in the cooler parts of the tube, while tungsten oxide remains behind in the boat. The Apparatus. The arrangement indicated in fig. 143 may be employed for the purpose. A piece of wide glass tubing D holds a boat containing a weighed quantity of the mixed oxides. The tube is connected at one end with an apparatus for generating hydrogen chloride (fig. 125, page 281). hydrogen chloride gas is passed through a wash-bottle A containing concentrated hydrochloric acid, and dried by passing through a wash-bottle B containing 1 H. Rose, Handbuch der analytischen Chemie, Braunschweig, 2. 358, 1871. 2 J. Lefort, Ann. Chim. Phys. (5), 9. 93, 1877. :i C. Friedheim and R. Meyer, Zeit. anorg. Chem., I. 76, 1892. 4 E. Pechard, Compt. Rend., 114. 173, 1891 ; Zeit. anorg. Chem., I. 262, 1892 ; Chem. News, 65. 89, 1892. 41 8 A TREATISE ON CHEMICAL ANALYSIS. concentrated sulphuric acid. The opposite end of the tube is connected with an absorption tube E containing a little water. The Sublimation. The tube containing the boat is heated to 270 1 in the tube oven (7, fitted with thermometer and thermostat. The sublimate of Mo0 3 . 2HC1 is driven forward into the absorption tube, every now and again, by heating the part of the combustion tube D with a naked Bunsen flame. The Tungstic Oxide. When sublimation has ceased (about 1| hour), the boat may contain tungsten oxide with or without sodium chloride, according to the nature of the substance under investigation. The sodium chloride, if present, is removed by washing the residue with water ; the tungsten oxide is filtered into a weighed Gooch's crucible, ignited as indicated on pago 106, and weighed as tungsten trioxide W0 3 . The Molybdic Oxide. The molybdic hydrochloride decomposes in contact with the water, forming a brick-red chloride Mo 3 6 Cl 8 insoluble in hydrochloric acid, but readily soluble in nitric acid. The sublimate is washed from the com- bustion tube and from the absorption tube into an evaporating basin with water containing a little nitric acid. The solution is evaporated to dryness on a water bath ; the residue is dissolved in a little ammonia ; evaporated to dryness ; ignited as indicated on page 413 ; and weighed as molybdenum trioxide Mo0 3 . The Results. The separations by this process are excellent. For instance, with artificial mixtures : Tungstic oxide used . . 0*2834 grm. Molybdic oxide used . . 0'0386grm. Tungstic oxide found . . 0'2838 grm. Molybdic oxide found . . 0'0380grm. The .comparatively elaborate apparatus is considered a disadvantage in the laboratory of a works. This objection would not be serious if a large number of determinations had to be made. The time factor is then significant. 2 218. The Determination of Niobium and Tantalum- Simpson's Process. The elements niobium and tantalum are often associated with tungsten, and, if present, they will be precipitated with the tungsten trioxide. Many tantalum minerals are free from tungsten, and since tantalum is now one of the "industrial elements," the analysis of compounds containing tantalum is sometimes required. Opening Tantaliferous Minerals. Several methods of opening the minerals have been suggested, and possibly some have special advantages in particular cases. Fusion with potassium bisulphate or pyrosulphate is generally recom- mended, 3 but the fusion is then usually very protracted 5 hours is not always 1 If the temperature be much greater than 270, tungsten may sublime; if much less, the separation may be incomplete. 2 For the separation of tungsten and vanadium, see C. Friedheim, Ber., 23. 352, 1890 ; Chem. News, 65. 27, 1892; A. Rosenheim and C. Friedheim, Zeit. anorg. Chem., I. 313, 1892 ; C. Friedheim, W. H. Henderson, and A. Pinagel, ib., 45. 396, 1905 ; C. Friedheim and C. Castendyck, Ber., 33. 611, 1900; W. Gibbs, Zeit. anal. Chem., 23. 543, 1884; 25. 544, 1886; Amer. Chem. Journ., 4-377, 1883; 5. 378, 1883; I. 219, 1879; F. v Mohr, Lehrbuch der chemischen-analytischen Titrirmethode, Braunschweig, 314, 1877 ; A. Safarik, Liebig's Ann., 109. 84, 1859; C. R. von Hauer, Ber. Wien. Akad., 39. 448, 1860; A. Rosenheim, Ber. t 23. 3208, 1890 ; P. E. Browning and R. J. Goodmann, Zeit. anorg. Chem., 13. 427, 1897 ; C. Reichard, Chem. Ztg., 27. 4, 1903 ; F. Rothenbach, Ber., 23. 3050, 1890 ; A. Carnot, Compt. Rend., 104. 1803, 1850, 1887 ; 105. 119, 1887. 3 W. Gibbs, Amer. J. Science (2), 37. 355, 1864 ; W. P. Headden, ib. (3), 41. 91, 1891 ; T. B. Osborne, ib. (3), 30. 229, 1885 ; Chem. News, 53. 43, 1886 ; W. B. Giles, ib., 95. 1 37, 1907 ; 99. 1, 1909 ; L. Smith, ib. t 48. 13, 29, 1883 ; Amer. Chem. Journ., 5. 44, 73, 1885 ; M. E. MOLYBDENUM, TUNGSTEN, NIOBIUM, AND TANTALUM. .419 sufficient for the work. The platinum crucible is much attacked by the opera- tion, and the subsequent analysis is beset with difficulties arising from the partial solubility of the oxides under in-vestigation in the aqueous extract of the fused cake. Fusion with potassium hydroxide 1 gives better results than potassium bisulphate. Many finely powdered minerals are decomposed by a half-hour's fusion at a dull red heat. The following process is due to Simpson. 2 Three grams 3 of pure potassium hydroxide are fused in a 4-cm. silver or nickel crucible resting on a wire gauze. When the potassium hydroxide in the crucible is in a state of tranquil fusion, remove the lid, and drop into the crucible 0'5 grm. of the finely powdered mineral. 4 Mix the contents by quickly rotating the crucible. Heat the mass to a dull red heat for 10 minutes longer ; remove the lid ; and again agitate the mass. Place the crucible in a hole in an asbestos pad, as described on page 266. Heat the crucible over a naked flame at a bright red heat for about half an hour. Remove the cover carefully to avoid losing the blob of flux on the under side, and let the lid cool upside down. Tilt the crucible on one side, so that the molten contents solidify on one side of the crucible. When cool, pour 10 c.c. of dilute hydrochloric acid (sp. gr. 1'08) into a 300-c.c. beaker. Place the crucible in a beaker, and fill the crucible two-thirds with warm water. Immediately cover the beaker with a clock-glass. The action is somewhat violent, and if any spurts from the crucible it will be caught by the basin. The action will subside in a couple of minutes. Then transfer the contents of the crucible to the beaker. Wash crucible, basin, and lid with water, assisted by a "policeman," then with dilute .acid, and finally with water. About 20 c.c. of acid are required for the washing, and the solution occupies from 80 to 100 c.c. Add a drop or two of alcohol to destroy the potassium manganate. Isolation of the Tantalum and Niobium Oxides. Boil the solution with 5 to 10 c.c. of hydrochloric acid (sp. gr. 1*1611) on a hot plate. 5 The tantalum and niobium hydroxides are precipitated. To make sure that the precipitation is complete, dilute the solution to 200 c.c. and boil 15 minutes more. Let the precipitate settle; decant the clear liquid through a 12'5-cm. close-packed filter paper. Filter and wash the residue with dilute hydrochloric acid (sp. gr. 1'0843), until the washings give no indication of iron. The filtrate contains the tin, iron, magnesium, calcium, manganese, copper, nickel, and titanium. The residue contains tantalum, niobium, tungsten, silica, antimony, and some of the tin. Pennington, Amer. Chem. Soc. Journ., 18. 38, 1896; Chem. News, 75. 8, 18, 31, 38, 1897; C. Marignac, Ann. Chim. Phys. (4), 13. 5, 1857; T. Prior, Min. Mag., 15. 83, 1910; W. Blomstrand, Jo urn. prakt. Chem. (1), 99. 40, 1866 ; R. Hermann, ib. (2), 5. 66, 1872 ; L. Weiss and M. Landecker, Zeit. anorg. Chem., 64. 65, 1909 ; Chem. Neivs, 101. 2, 13, 26, 1910 ; E. F. Smith, ib., 92. 209, 1905 ; Proc. Amer. Phil. Soc., 44. 151, 1905 ; R. D. Hall and E. F. Smith, ib., 44. 177, 1905 ; Chem. News, 92. 220, 232, 242, 252, 262, 276, 1905 ; G. Chesneau, Compt. Rend., 149. 1132, 1910 ; 0. Ruff and E. Schiller, Zeit. anorg. Chem., 72. 329, 1911 ; R. J. Meyer and 0. Hauser, Die Analyse der seltenen Erden mil der Erdsduren, Stuttgart, 296, 1912. 1 Potassium in preference to sodium hydroxide because of the greater solubility of the potassium salts of the oxides under investigation. 2 E. S. Simpson, Bull. West Australia Geol. Sur., 23. 72, 1906; Chem. News, 99. 243, 1909. 3 If but small quantities of these elements are in question, it is advisable to take as much as 30 grms. of potassium hydroxide, and 5 grms. of the mineral. 4 Note that the proportion of mineral to flux is as 1 : 6. 5 Less acid can be used if no titanium is present ; the titanium is kept in solution by the hydrochloric acid. 6 If the nitrate becomes turbid, it may be necessary to dilute the solution and boil for a longer period. 42O A TREATISE ON CHEMICAL ANALYSIS, A small amount of the niobium may not be precipitated if much titanium be present. It seems as if "a soluble double chloride of titanium and niobium is formed, part of the niobium only being precipitated." Hence, the results for the niobium may be a little low, and the tantalum correspondingly high. In the absence of appreciable amounts of titanium, 1 it is claimed that the per- centage of tantalum and niobium obtained by this method can be relied upon to be accurate to O'l per cent. Removal of Tin, Antimony, Tungsten, and Silica from the Residue. If niobium and tantalum alone be present, dry the precipitate, burn the filter paper (page 408); ignite at a bright red heat for 15 minutes, and weigh as Nb 2 5 + Ta 2 5 . If tungstic acid be present, digest the moist precipitate 2 with ammonia, 3 in which tungstic oxide is soluble. Filter and wash. 4 Ignite the precipitate and weigh. The loss in weight represents the tungstic oxide. Treat the mass with sulphuric acid and hydrofluoric acid (page 169), ignite, and weigh again. 5 The loss in weight represents the silica. Place the weighed residue in a weighed porcelain boat, and heat the mixture in a current of hydrogen (page 269). The tin is reduced. Digest the mixture in hydrochloric acid. 7 Ignite the washed residue, and weigh as Nb 2 5 + Ta 2 5 . The loss in weight represents the tin oxide. Determination of Tin. The filtrate contains chloride of tin, etc. The solu- tion is treated with hydrogen sulphide in the usual manner (page 309). The tin so obtained is added to that recovered from the first precipitate. A nickel crucible is used for the determination of the tantalum, niobium, and tin; but for the determination of the iron, aluminium, manganese, and rare earths, etc., it is best to fuse a separate portion in a silver crucible, and reject the precipitate of tantalum, niobium, etc., and the hydrogen sulphide precipitate. The analysis is continued with the filtrate from the hydrogen sulphide precipitate. 1 According to L. Weiss and M. Landecker (Zeit. anorg. Chem., 64. 65, 1909 ; Chem. News, 101. 2, 13, 26, 1910), the formation of this compound is hindered by the addition of an oxidising agent sodium nitrate to the alkali. After dissolving in water and filtering, very little titanium remains in solution, and this is completely precipitated by hydrogen sulphide, with- out carrying down the other earth acids. The titanium can also be determined colorimetrically in the mixed oxides, since the colour is not affected by niobium and tantalum G. Chesneau, , 149. 11 94. 298, 1906 ; R. D. Hall and E. F. Smith, Journ. Amer. Chem. Soc., 27. 1369, 1905. Compt. Rend., 149. 1132, 1909 ; C. H. Warren, Amer. J. Science(^), 22. 520, 1906 ; Chem. News, by mot . J. Scii 2 Washed free from hydrochloric acid test with dilute solution of silver nitrate. 3 Some recommend digesting the moist residue with yellow ammonium sulphide to remove tin, tungsten, etc. H. Rose, AusfiihrlicJies Handbuch der analytischen Chemie, Braunschweig, 2. 349, 1851 ; F. Wohler, Die Mineral- Analyse in Beispielen, HO, 1861 ; W. P. Headden, Amer. J. Science (3), 41. 89, 1898; M. E. Pennington, Journ. Amer. Chem. Soc., 18. 38, 1896. In that case, there is a slight loss of the niobic acid if sulphates are present (W. B. Giles, Chem. News, 99. 1, 1909). The reduction process offers many advantages in analysis of tin ores where small amounts of titanium, niobium, and tantalum have to be separated from large amounts of tin A. Hilger and H. Haas, Ber., 23. 458, 1890. See also R. D. Hall, Journ. Amer. Chem. Soc., 26. 1235, 1904 ; E. F. Smith, Proc. Amer. Phil. Soc., 44. 151, 1905. 4 Some silica may then be found contaminating the tungstic oxide. The ammoniacal solution is evaporated to dryness, ignited, and weighed as W0 3 + Si0 2 . The latter is determined by the hydrofluoric acid treatment and added to the main silica. 5 F. D. Metzger and C. E. Taylor, Zeit. anorg. Chem., 62. 383, 1909. 6 Simpson recommends the removal of antimony and tin by digestion with ammonium sulphide ; re-fuse the residual oxides and repeat the operations. 7 The hydrogen may be passed through a beaker of water (fig. 122) and the tin recovered. If the potassium cyanide fusion be adopted (page 269), insoluble niobic and tantalic acids may envelop the metallic tin, and protect the tin from the action of the acid later on ; and some may dissolve in the potassium cyanide. The insoluble residue is digested with acids to remove tin (and antimony if present). Niobic and tantalic acids remain insoluble. The niobic and tantalic oxides dissolved in the potassium cyanide can be recovered by boiling the solution with nitric acid. The tungstic oxide is separated from the tantalic and niobic oxides by ammonia as indicated above. MOLYBDENUM, TUNGSTEN, NIOBIUM, AND TANTALUM. 42 1 219. The Separation of Tantalum and Niobium Marignac's Process. Marignac's process 1 for the separation of tantalum and niobium is probably the best. It is based upon the different solubilities of potassium fluotantalate K 2 TaF 7 and of potassium fluoniobate K 2 NbOF 5 .H 2 0. 2 One part of the latter salt is soluble in 12-13 parts of cold water, while 1 part of potassium fluotantalate dissolves in 150-157 parts of cold water. The separation is tedious, and sometimes " not worth while " 3 see the next section. It is convenient to use four platinum dishes " A " dish, 7'5 cm. in diameter ; "B" dish, 9 cm. ; "C" dish, 6 cm. ; and "D" dish, 5 cm. in diameter for the separation. Fuse the (weighed) mixed oxides of niobium and tantalum in a platinum crucible with eight times their weight of potassium carbonate. When solution is complete, cool the mass. Digest the cake in water. 4 Boil the mixture with dilute hydrochloric acid to precipitate the niobic and tantalic acids. Collect the precipitate on a filter paper ; wash into, say, dish " C." The filter paper, folded inside out, is placed in dish "D," covered with hot water, and a few drops of hydrofluoric acid are added. Warm the mixture on a hot plate for a few minutes. Pour the solution into dish " C." Repeat the washing of the filter paper with very dilute hydrofluoric acid, and finish by washing the paper four times with hot water. First Crop of Crystals. The dish " C " now contains all the niobium and tantalum oxides. Put the dish on a hot plate. If solution be not complete in a few minutes, add another drop of hydrofluoric acid, but avoid an excess of this acid. Add slowly, with constant stirring, a boiling aqueous solution of 0*7 grm. of potassium fluoride to the boiling solution in basin "C." Evaporate the contents to about 10 c.c. 5 Wash down the sides of the basin with a few drops of hot water. Cool the vessel slowly to about 15. Decant the clear solution con- taining the niobium and part of the tantalum through a 7-cm. filter paper 6 into dish "B." Wash the felted mass of crystals of potassium fluotantalate four times with cold water. Second Crop of Crystals. Evaporate the mixed filtrate and washings down to about 5 c.c. Cool the solution slowly as before. Decant the solution through a 5-5-cm. filter paper into dish "A." Wash the crystals four times with cold water. If flat plates of potassium-niobium oxyfluoride (fig. 141) are present, wash the mass until they are removed. Third Crop of Crystals. Evaporate the solution to dry ness on a water bath. 1 C. Marignac, Ann. Chim. Phys. (4), 8. 5, 49, 1865 ; E. S. Simpson, Bull. West Australia Geol. Sur., 23. 71, 1906 ; A. Tighe, Journ. Soc. Chem. Ind., 25. 681, 1907 ; 0. Ruff and E. Schiller, Zeit. anorg. Chem , 72. 329, 1911 ; E. Meimberg, Zeit. angew. Chem., 26. 83, 1913. 2 Or 2KF. NbOF 5 . H 2 0. G. Kriiss and L. F. Nilson, er., 20. 1676, 1887. 3 Commercial ores of"tantalum are valued at, say, 1 per unit percentage see page 384. The value of the ore is lessened in proportion to the amount of niobium associated with the tantalum. 4 If any remains insoluble, filter, ignite the residue, and fuse with a little potassium car- bonate. Add the aqueous extract to the main solution. 5 It is sometimes stated that solutions of tantalum fluorides when evaporated lose tantalum fluoride with the vapour of hydrofluoric acid and steam. This seems to be an error O. Ruff' and E. Schiller, Zeit. anorg. Chem., 72. 329, 1911. No loss was detected by evaporating solutions of tantalum fluoride in sulphuric acid to dry ness. Similar remarks apply to the calcination of tantalic oxide with ammonium fluoride. Potassium fluotantalate alone may be slightly volatile when heated, but Ruff and Schiller could detect no loss when potassium fluotantalate was heated up to its melting point. (i A rubber funnel, or a glass funnel coated with cerasine or the wax mixture of page 634, is used. 422 A TREATISE ON CHEMICAL ANALYSIS. Cool, add one drop of hydrofluoric acid. Run 1 to 5 c.c. 1 water from a burette into the solution. Heat the vessel quickly to dissolve the residue. Add O'l grm. of potassium fluoride dissolved in 1 c.c. of water. Note the volume of the solution for "correction a." Cool the solution for about an hour at 15 or less ; filter it into a small platinum dish ; and wash three or four times with a few drops of water at 15 or less. Note the volume of the liquid used for the washing "correction 6." Determination of the Niobium. 12 Add 8 c.c. of sulphuric acid (sp. gr. 1*29) to the filtrate and washings. Evaporate the solution on a sand bath until fumes of sulphur oxides appear. Continue the heating for about 20 minutes longer in order to drive off all the fluorine compounds. See that no fluorides remain unattacked on the side of the dish out of reach of the acid. Cool. Pour the mixture into 150 c.c. of water in a 300-c.c. beaker. Wash the dish with cold water assisted by a "policeman." Boil the solution 20 minutes in order to precipitate all the niobium and tantalum. Filter, wash with boiling water, and dry. Ignite until the filter paper is burnt, add a gram of solid ammonium carbonate, cover the crucible, and re-ignite to constant weight. Weigh the dish, which contains practically all the niobium and a small amount of tantalum. Correction for the Tantalum. The amount of the tantalum admixed with the columbium is determined from the observation that 1 c.c. of the water, slightly acidifed with hydrofluoric acid, used in the final crystallisation dissolves the equivalent of 0-002 grm. of Ta 2 5 (correction a), and 1 c.c. of water used in the final washing (correction b) dissolves the equivalent of 0*00091 grm. of Ta 2 5 . In illustration, suppose that 0'5 grm. of tantalite furnished 0-3819 grm. of the mixed tantalum and niobium oxides. 4 c.c. of the acidified solution (correc- tion a) were in contact with the crystals of potassium fluotantalate, and 4 c.c. of water (correction b). Platinum crucible plus precipitate . . . . . . 14 '3550 grms. Platinum crucible alone . . '.. . ..'"*. . . 14'3211 ,, Precipitate. , , . ... .. . . '0339 grm. Correction a (0'002x 4). . ."'.. . . '0080 grm. Correction b (0 -00091x4) . . . . . . . 0'0036 ,, Total correction . . . . * . . . 0'0116 ,, Precipitate . . . . -. '0339 grm. Correction ...... ,V . . . . . 0'0116 ,, Niobium oxide . . . t .. . . . 0*0223 ,, Mixed oxides 0'3819 , Tantalum oxide . , V . . . . . 0'3406 Hence, the sample had 78-12 per cent, of tantalum oxide, and 4*26 per cent. of niobium oxide. It is claimed that this tedious method of separation is accurate to about 1 per cent. 3 1 The amount depends upon the amount of niobium expected to be present. 1 c.c. per 0'035 grm. Nb 2 O 5 . Usually about 3 c.c. suffices. 2 For a colorimetric process for estimating niobium, see E. Meimberg, Zeit. angew. Chem., 3 L. Weiss and M. Landecker (Zeit. anorg. Chem., 64. 65, 1909 ; Chem. Neivs, 101. 2, 13, 26, 1910) conduct the analysis by fusion with potassium bisulphate ; extract with hot water acidulated with sulphuric acid r add sulphurous acid to the boiling solution until the precipitate becomes flocculent ; in 20-30 minutes, filter and wash with sulphurous acid. The residue, con- taining tantalum, niobium, titanium, and tin, is neutralised with ammonia; extracted with ammonium sulphide ; heated in a platinum crucible ; and fused with a mixture of sodium carbonate and nitrate. Extract with water, filter, treat with hydrogen sulphide. Filter, and precipitate the niobic and tantalic acids with sulphurous acid as before. Fuse the ignited MOLYBDENUM, TUNGSTEN, NIOBIUM, AND TANTALUM. 423 220. The Estimation of Niobium and Tantalum Oxides- Specific Gravity Method. The great difference between the specific gravities of tantalic oxide (8-710) and niobic oxide (4'552) has suggested methods for evaluating ores l and also the oxides 2 of these elements from the specific gravity of the mixture. The specific gravities of the oxides appear to become constant after blasting for an hour. The specific gravity of the mixture is not quite a linear function ; an empirical table is therefore constructed, from which the specific gravity of all possible mixtures can be obtained by interpolation. Determination of the Specific Gravity. Foote and Langley use the ordinary specific gravity bottle with the stopper extra carefully ground. Place the unweighed powder in a small beaker half full of water. Boil the water for half an hour by passing an electric current through a fine spiral of platinum wire suspended in the water. 3 Agitate the powder frequently. Cool. Decant most Nb 2 5 10 20 30 4-0 50 60 70 80 90 Ta 2 5 FIG. 144. Specific gravities of mixtures of niobic and tantalic oxides (Foote and Langley). of the water from the beaker. Wash the residue into the specific gravity bottle through a small funnel by means of boiled water. The bottle is filled to over- flowing, and any powder or air-bubbles floating on the surface are swept off with a glass rod. Remove any air-bubbles from the sides of the bottle by means of a platinum wire. Place the bottle in a dish of water at 20 for 20 minutes. Insert the stopper. Wipe the top. Dry and weigh. Transfer the contents to a platinum dish ; 4 evaporate to dryness ; ignite over the blast for 5 minutes ; precipitate of mixed niobic and tantalic acids with sodium carbonate and nitrate ; digest with warm water ; cool ; and pass carbon dioxide through the solution. Tantalic acid alone is precipitated. E. Wedekind and W. Maass, Zeit. angeiv. C'hem., 23. 2314, 1910 ; 0. Hauser and A. Lewite, ib. t 25. 100, 1912. 1 S. L. Pentield and W. E. Ford, Amur. J. Science (4), 22. 61, 1906 (stibiotantahte) ; E. S. Simpson, Bull. West Australia GeoL Soc., 23. 71, 1906 (iron- and mangano-tantalites). The estimations are usually accurate to less than 5 per cent. 2 H. W. Foote and R. W. Langley, Amer. J. Science (4), 30. 393, 1910 ; Chem. News, 103. ' 3 Boiling in the ordinary manner is impracticable owing to "bumping." The powder is heavy and settles rapidly. 4 In transferring the powder to the dish, loss can be avoided by inverting the bottle over the dish (fig. 146, page 434) and shaking the bottle slightly until the powder runs out. A trace of powder sticks to the sides of the bottle. Hence, dry the bottle at 120, and weigh. The increase in weight represents the powder in the bottle, usually not more than 0*002 grm. 424 A TREATISE ON CHEMICAL ANALYSIS. and weigh. Hence calculate the specific gravity. Duplicate determinations agree to within 2*2 per cent. EXAMPLE. Suppose the empty specific gravity bottle weighed 26'9 grins., and, after due correction, the mixed oxides weighed 6'2 grms. ; the bottle, mixed oxides, and water weighed 69'8 grms. ; and the bottle and water only 64'8 grms. Since weight of mixed oxides Spec fie gravity = weight of an equal volume of water ' it follows that the weight of a volume of water equal to the volume of the mixed oxides is: 64-8 + 6'2-69'8= 1-2. Hence the required specific gravity is 6-2/l'2 = 5-17. The corresponding percentage amounts of tantalic and niobic oxides can be read from fig. 144. Given a mixture of two constituents A and B, of known specific gravity 8j and s 2 respectively, to find what proportion of each is present when the specific gravity of the mixture, S, is known. Let W denote the weight of the mixture, S its specific gravity; let x of constituent A be present, then W-x of B is present. Since the weight of a substance divided by its specific gravity represents its volume, we have W x W-x Ws^s.-S) -=-+ 01> x = - ' Hence, the mixture contains -#) v ' percent, of A; inn 100ai(* 2 - S) 100 - >-ii ~\ P er EXAMPLES. (1) A mixture of clay (sp. gr. 2'6) and lignite (sp. gr. 12) has a specific gravity 2'0 : what proportions of clay and lignite are present ? Here =2'4 ; s 2 = 2'6 ; fc = l-2; hence, the mixture contains 7'14 per cent, of lignite and 92'86 per cent. of clay. (2) A mixture containing tantalic and niobic oxides has a specific gravity 5 - 85, niobic oxide alone has a specific gravity 4'552, and tantalic oxide, 8*71 : what is the percentage composition of the mixtures ? Answer Nearly 54 per cent, of tantalic oxide and 46 per cent, of niobic oxide. Preparation of the Mixed Oxides. The method of preparing the oxides should be as fixed and as definite as possible. The specific gravity of the mixed oxides becomes constant after igniting for an hour over the blast in a platinum crucible. Artificial mixtures of tantalic and niobic oxides gave specific gravities : Tantalic oxide . 10 20 30 40 50 60 70 80 90 100 Specific gravity . 4 '552 4746 4 '929 5 '200 5 "474 5 '850 6 '434 7'083 7 '654 8 "098 8716 These numbers are plotted in fig. 144. Hence, given the specific gravity of the mixed oxides, the proportion of niobic and tantalic oxides can be determined by interpolation. The results are probably more exact than with Marignac's gravimetric process. In illustration, two determinations of tantalic and niobic oxides on a sample of niobite or columbite were made by the specific gravity and by Marignac's process : Specific gravity process. Marignac's process. Tantalic oxide . . . . 16 '5 4 15 '81^ ' 14'43 13'60^ Niobic oxide .... 62 "23 63 -23 63 '68 6575 CHAPTER XXXI. DETERMINATION OF GOLD AND SELENIUM. 221. The Precipitation of Gold and Platinum. GOLD is precipitated from cold acid solutions by hydrogen sulphide as a dark brown sulphide ; ] if the solution be hot, some metallic gold is also precipitated. The sulphide dissolves in alkali monosulphide extremely slowly, but a little more quickly in the polysulphide, and forms a brownish-red solution. If, therefore, the attempt be made to separate gold from the precipitate of mixed sulphides by ammonium or sodium monosulphide, part of the gold will be found with the tin, antimony, and arsenic; and part with the metals insoluble in alkaline sulphide. What has here been stated with respect to gold sulphide might be repeated for platinum sulphide. 2 The solubility of platinum sulphide in ammonium or sodium monosulphide is facilitated by the presence of arsenic and other sulphides. On account of the difficulty involved in separating gold and the platinum metals from the other members of the hydrogen sulphide group, it is best to separate these elements before precipitating the hydrogen sulphide group ; or to redissolve the hydrogen sulphide precipitate in acids, and separate the gold and platinum metals from the solution of the sulphides. To separate gold, advantage is taken of the ease with which gold compounds are reduced to the metal. Several reducing agents are available : sulphurous acid, 3 oxalic acid, 4 ferrous chloride (or sulphate), 4 hydrazine hydrochloride, 5 hydroxylamine hydrochlbride, 6 alkaline formaldehyde solutions, 7 alkaline hydrogen peroxide solutions, 8 chloral hydrate, 9 magnesium, 10 nickel, 11 hypophosphorous acid, 12 cane sugar, 13 etc. Sulphurous acid, oxalic acid, and hydrazine hydrochloride are most con- venient. The choice of the right precipitating agent is determined by the 1 Probably AuS. A. Levol, Ann. Chim. Phys. (3), 30. 356, 1850 ; U. Antony and A. Lucchesi, Gazz. Chim. ItaL, 19. 545, 1889 ; 2O. 601, 1890 ; L. Hoffmann and G. Kriiss, Ber., 20. 2369, 2704, 1887 ; W. Bettel, 56. 133, 1887 ; J. Riban, Bull. Soc. Chim. (2), 28. 241, 1877. 2 To facilitate subsequent nitration, R. Gaze (Apoth. Ztg., 27. 959, 1912) adds a little mercuric chloride. 3 P. Berthier, Ann. Chim. Phys. (3), 7. 82, 1843. 4 L. Hoffmann and G. Kriiss, Liebig's Ann., 238. 66, 1887 ; Chem. News, 56. 83, 1887. 5 P. Jannasch and 0. von Meyer, Ber., 38. 2129, 1905. (i C. Winkler, Ber., 22. 890, 1889 ; A. Lainer, Monats. Chem., 12. 639, 1891 ; P. Jannasch and 0. von Meyer, Ber. 38. 2129, 1905. 7 L. Vanino, Ber., 31. 1763, 1899 ; Zeit. Chem. Ind. Roll, I. 272, 1906. For formic acid, W. Bettel, Chem. News, 56. 133, 1887. 8 L. Vanino and L. Seemann, Ber., 32. 1968, 1899 ; Chem. News, 82. 70, 1900 ; L. Rbssler, Zeit. anal. Chem., 49. 739, 1910. 9 P. J. Dirvell, Bull. Soc. Chim. (2), 46. 806, 1886. 10 C. Scheibler, Ber., 2. 295, 1869. 11 V. Goldschmidt, Zeit. anal. Chem., 45. 87, 1906. 12 F. Treubert, Dissertation, Miinchen, 1909. 13 P. Leidler, Zeit. Chem. Ind. Koll, 2. 103, 1907. 426 A TREATISE ON CHEMICAL ANALYSIS. nature of the accompanying elements. The precipitation of gold by these reducing agents generally leads to low results. For instance, Hoffmann l found : Table LVII. Comparison of Different Precipitating Agents for Gold. Precipitating agent. Used. Found. Per cent, loss. Sulphurous acid . Sulphurous acid . 3-12360 2-14739 3-11916 2-14688 0-14 0-03 Oxalic acid . . . . . 215292 2-15040 0-12 Ferrous chloride . 1-65596 1-65535 0-04 This loss arises from the fact that the gold is precipitated in a very fine state of subdivision, and minute particles of gold pass through the filter paper. In consequence, it is best to work with solutions as concentrated as possible, and to let the mixture stand in contact with the acid mother liquid for some time in a warm place, and agitate the solution thoroughly from time to time. This treatment favours the coagulation of the particles of gold, and thus facilitates " clean " filtrations. 2 Precipitation by Sulphurous Acid. The acid solution under investigation is concentrated on a w^ater bath until crystals begin to form as a drop of the solution is cooled on the end of a glass rod. Add sufficient water to dissolve all the crystals, 3 and digest the solution for half an hour on a water bath with an excess of sulphurous acid. At the end of that time the liquid should be clear, and smell of sulphur dioxide. The addition of sulphur dioxide from time to time may be necessary. The solution is then filtered, washed with dilute hydro- chloric acid, and the metallic gold ignited in a weighed porcelain crucible. This precipitating agent is superior to both oxalic acid and ferrous chloride. There is no danger of loss by spurting, as occurs with oxalic acid, which decom- poses with the formation of carbon dioxide ; and the precipitated gold is more easily washed free from salts than is the case when ferrous salts are employed. The time required for complete precipitation half an hour is shorter than that required for complete precipitation with oxalic acid and ferrous chloride. If selenium be present, it will be precipitated with the gold. If platinum be present, and the solution be not sufficiently acid, the precipitated gold will be contaminated with platinum, and the results for gold will then be a little high. Oxalic acid, however, gives the best separations of gold from platinum, although, in separating gold from platinum, it is usual to first precipitate the latter metal as ammonium or potassium chloroplatinate. Sulphurous acid gives better separations of gold from palladium than ferrous chloride or oxalic acid, and there is little to choose between these three agents for the separation of gold from iridium, rhodium, and ruthenium. A large excess of hydrochloric acid generally retards the precipitation of gold. Nitric acid is, of course, supposed to be absent; this is removed by 1 1 1. Hoffmann, Untersuchung uber das Gold, Erlangen, 11, 1887 ; E. Pfiwoznik, Oester. Zeit. Berg. Hiitt., 59. 639, 1911 ; W. Bettel, Chem. News, 56. 133, 1887. 2 J. Yolhard (Lielrig's Ann., 198. 331, 1879) recommends warming the gold solution with mercuric oxide on the water bath. The particles of gold are said to aggregate with the mercuric oxide, and are, in consequence, easily filtered and washed. The mercuric oxide vola- tilises on ignition. 3 The gold chloride may be partially decomposed into metallic gold, etc., during the evapora- tion. This does not matter. DETERMINATION OF GOLD AND SELENIUM. 427 repeated evaporation with hydrochloric acid. 1 No notice need be taken of any separation of gold during the evaporation. Precipitation by Oxalic Acid The solution is concentrated as described above, and the acid solution is neutralised and warmed with oxalic acid, until further additions of oxalic acid give no further separation of gold. 2 The reduction is best made in a covered vessel, on account of the risk of losses by the spurting which attends the oxidation of the oxalic acid to carbon dioxide. The vessel is allowed to stand in a warm place for about 24 hours, and the metallic gold is filtered, washed with water, 3 ignited in a porcelain crucible, and weighed as metallic gold. The oxalic acid process is not satisfactory in the separation of gold from pal- ladium, since some of the palladium is precipitated with the gold about \ per cent, is precipitated per gram of gold. When lead is present, it will be precipitated as oxalate with the gold. Hence, the lead must be removed or another pre- cipitating agent substituted. If the lead be removed as sulphate, there is a danger of the latter carrying down some of the gold. If much lead and little gold be present, probably all the gold will be found with the precipitate of lead sulphate. 4 Selenium and molybdenum are not precipitated ; but if insufficient hydrochloric acid be present, tellurium, if present, may give a precipitate which is removed from the gold by washing with hydrochloric acid. If much copper be present, the gold will be contaminated with copper oxalate, even if a great excess of hydrochloric acid be employed. To separate the gold from the copper oxalate, Purgotti 5 proposes the following process : Evaporate the aqua regia solution on a water bath to dryness, dissolve the residue in water, add oxalic acid, and digest the mixture in a warm place for 48 hours. Neutralise the boiling solution with potassium hydroxide. If a great excess of oxalic acid has not been added, more will be needed so as to form a soluble double copper and potassium oxalate, which imparts an ultramarine blue tint to the solution. The precipitated gold can then be filtered from the solution and washed in the usual manner. When the above-mentioned disturbing elements are present, it is best to use the sulphurous acid or the hydrazine process. Precipitation by Hydrazine Hydrochloride. Hydrazine hydrochloride may 1 It is interesting to observe in this connection that the evaporation of alkali nitrates to dryness a number of times with hydrochloric acid is not sufficient to effect the complete conversion of the nitrates to chlorides. For instance, 0*1893 grm. of potassium nitrate furnished after a number of evaporations : Evaporation . . . First. Second. Third. . . . Ninth. . . . Twelfth. Potassium nitrate . . 0*1136 0*0581 0*0149 ... 0*0013 ... 0*0013 grm. B. Lucanus (Zeit. anal. Chem., 3. 403, 1864) prefers to ignite the alkali nitrate with four to six times its weight of grape sugar, so as to form the carbonate ; and when this is treated with hydrochloric acid, the chloride will be obtained free from the nitrate. 2 Note that most of the oxalates (excepting those of the alkalies and magnesium) are but sparingly soluble in neutral solutions and feebly acid solutions. Some dissolve in alkalies, and all dissolve in concentrated mineral acids, while some are almost insoluble in dilute acids e.g., copper oxalate in dilute nitric acid (C. Luckow, Zeit. anal. Chem., 26. 9, 1887 ; Chem. News. 55. 73, 1887 ; G. Bornemann, Chem. Zty., 23. 565, 1899 ; H. L. Ward, Zeit. anorg. Chem., 77. 257, 269, 1912) hence, when other metals are present the solution should be sufficiently acid to prevent the separation of insoluble oxalates. 3 If some of the less soluble oxalates be present, they can be removed by washing the pre- cipitated gold with dilute hydrochloric acid. 4 C. Whitehead, Chem. News, 66. 19, 1892. 5 E. Purgotti, Zeit. anal. Chem., 9. 128, 1870. y R. Fresenius (Quantitative Chemical Analysis, London, I. 477, 1876) says that the oxalic process is not satisfactory in the presence of lead, mercurous, and silver salts. Since the determinations are usually made in hydrochloric acid solutions, if the solution be dilute, practically all the silver will be precipitated as chloride. Mercurous salts cannot remain in contact with gold trichloride without reducing the latter, so that if mercury and gold are present together in solution, the mercury must be in the higher stute of oxidation. 428 A TREATISE ON CHEMICAL ANALYSIS. be employed to precipitate gold from neutral, acid, or alkaline solutions con- taining potassium, sodium, barium, strontium, calcium, magnesium, aluminium, chromium, zinc, manganese, iron, uranium, nickel, cobalt, cadmium, mercury, lead, copper, but not tin. When platinum metals are associated with the gold in a solution, the gold may be first precipitated by adding hydroxylamine hydro- chloride to the solution acidified with hydrochloric acid, and digesting the mixture on a water bath for some time. The filtrate must be tested to make sure that all the gold is precipitated, because hydroxylamine hydrochloride is not so vigorous an agent as the hydrazine salt. The palladium, iridium, platinum, rhodium, and osmium remain in solution. The first four elements, if present, are precipitated by hydrazine hydrochloride in alkaline solution, but not in acid solutions. 1 222. The Separation of Gold and Platinum from Tin, Arsenic, and Antimony. The tin group of elements cannot be quantitatively separated from the gold-platinum group by reducing agents ; nor can tin be dissolved from the platinum metals by treatment with acids ; and the sulphides cannot be separated by concentrated hydrochloric acid nor by sodium hydroxide. The volatilisation of tin sulphide in a stream of hydrogen sulphide is tedious and slow. The chlorides of tin, antimony, and arsenic can be volatilised in a stream of chlorine FIG. 145. The [Separation of the Tin from the Gold-Platinum group. or hydrogen chloride. The volatilisation of the bromides in a stream of bromine gives better results. Volatilisation Process. Transfer the washed sulphides to a porcelain boat along with the filter-paper ash. 2 Place the boat B in a hard glass combustion tube, and pass a stream of bromine or hydrogen chloride through the tube while the boat is heated as indicated in fig. 145. The exit tube leads into an absorption flask A containing dilute hydrochloric acid. The sulphides of arsenic, antimony, and tin are volatilised and retained by the absorption flask ; any sublimate in the tube is driven away from the boat by means of the naked flame 1 Osmium and ruthenium are only partially precipitated in alkaline solutions by the hydrazine salt, and not at all in acid solutions. - The filter paper is separated from the sulphides, and burned alone. The sulphides may also be intimately mixed with 3-5 times their weight of ammonium chloride and their own weight of ammonium nitrate, and calcined. The gold and platinum remain in the crucible. DETERMINATION OF GOLD AND SELENIUM. 429 from a Bunsen's burner. The gold and platinum remain in the boat. 1 The combustion tube is washed with dilute hydrochloric acid, and the washings run into the absorption flask. The tin, arsenic, and antimony in the latter can be de- termined in the usual manner. The process is accurate, but it occupies a long time. DirvelVs Process. Dissolve the precipitated sulphides in aqua regia. Add a small quantity of a saturated neutral solution of sodium oxalate and of oxalic acid. Then add a considerable excess of sodium hydroxide pure by alcohol. No notice need be taken of any separation of sodium oxalate. Heat the solution to boiling, and while boiling, add, drop by drop, a solution of chloral hydrate. Boil the solution under a hood. 2 When an excess of chloral hydrate has been added, and the gold and platinum have precipitated, filter the boiling solution, wash, and weigh the mixed metals in the usual manner. The filtrate is diluted with water and boiled to drive off the chloroform. 3 The antimony, arsenic, and tin are determined in the filtrate in the usual mariner. The washed gold and platinum are dissolved in aqua regia and separated either by precipitating the gold with oxalic acid, or the platinum with ammonium chloride (page 436). 223. The Colorimetric Determination of Gold. Gold can be estimated, somewhat approximately, by the tint obtained when gold solution is brought in contact with a solution containing a mixture of stannous and stannic chlorides. 4 The gold sulphide is dissolved in 10 c.c. of aqua regia (three volumes of concentrated hydrochloric acid, and one volume of concentrated nitric acid), and treated with a saturated solution of stannous chloride until the yellow colour is bleached. The purple tint which develops in less than a minute is compared with a set of artificial standards made by mixing solutions of copper and cobalt salts in the required proportion. A coloration is obtained when but one part of gold per million parts of solution is present. The usual precautions for colorimetric determinations are here of exceptionally great importance in order to get reliable results. The difficulty in controlling the exact tint of the precipitated "purple of Cassius " considerably limits the scope of this process. 224. The Analysis of Colours containing Purple of Cassius. In the analysis of gold colours which usually contain lead oxide, silver carbonate, stannic oxide, gold, boric oxide, silica, and soda the sodium peroxide fusion (pages 226 and 475) is an excellent method of opening up the compound. One portion of the sample is fused in a porcelain crucible, and the resulting cake is employed for the determination of gold, silver, tin, and lead. The boric oxide is determined in another portion of the sample fused in a silver crucible. The alkalies are determined by the method described on page 226 ; the ignition is conducted in a silver crucible. The silica, alumina, etc., are determined on 1 R. Fresenius, Zeit. anal. Chem., 2$. 200, 1886 ; L. Wbhler and A. Spengel, ib., 50. 165, 1911 ; L. L. de Koninck and A. Lecrenier, Rev. Univ. Mines (3), 2. 98, 1888 ; L. Eisner, Journ. prakt. Chem. (1), 35. 310, 1845; A. Bechamp and C. St Pierre, Compt. Rend., 52. 757, 1861; P. de Clermont, ib., 88. 972, 1879; G. Campari, Ann. di Chim., 74. 1, 1882; T. Bailey, Journ. Chem Soc., 49. 735, 1886; V. Antony and L. Niccoli, Gazz. Chim. Ital., 22. ii, 408, 1892. 2 Vapours of chloroform are evolved. :i P. J. Dirvell, Bull. Soc. Chim. (2), 46. 806, 1886 4 E. Sonstadt, Chem. News, 26. 159, 1872 ; T. K. Rose, ib., 66. 271, 1892 ; R. N. Maxou, Amer. J. Science (4), 21. 270, 1906; Chem. News, 94. 257, 1906; A. Carnot, Compt. Rend., 97. 105, 1883 ; J. Moir, Journ. Chem. Met. Soc. S. Africa, 4. 125, 1903 ; A. Prister, ib., 4. '235, 1903 ; H. R. Cassel, Eng. Min. Journ., 76. 661, 1903 ; E. Rupp, Ber., 36. 3961, 1903 ; M. E. Pozzi-Escot, Ann. Chim. Anal., 12. 90, 1907 ; W. Bettel, Min. Eng. World, 35. 987, 1912. For the detection of traces of gold, see J. E. Saul, Analyst, 38. 54, 1913. 430 A TREATISE ON CHEMICAL ANALYSIS. another portion of the sample fused in a silver crucible. The separation of the gold, silver, lead, and tin alone requires special mention. Separation of Silver. The fused cake is taken up with water and the solu- tion warmed with an excess of hydrochloric acid. The gold passes into solution as gold chloride, 1 the silver is precipitated as silver chloride along with a certain amount of silica and lead. The solution is diluted, so that it contains no more than the equivalent of 0'25 to 0'30N-HC1 after the alkali has been neutralised. The hot solution is filtered through a hot-jacketed funnel, and washed with JN- hydrochloric acid. The precipitated silver chloride is freed from silica by treat- ment with hydrofluoric acid, and then dissolved in ammonia. 2 ' The properties of silver chloride should here be studied page 650. Purification of Silver Chloride from Lead Chloride. The lead chloride can be removed from the ammoniacal solution of silver chloride by adding 5 c.c. of a concentrated solution of ammonium nitrate, 40 c.c. of a 2 per cent, solution of hydrogen peroxide per 25 c.c. of the cold ammoniacal solution. 3 In 3 or 4 hours the yellowish-brown flocculent precipitate of lead peroxide may be filtered and washed, first with dilute ammonia, and lastly with cold water. The lead peroxide is dissolved in nitric acid, assisted by the addition of a little hydrogen peroxide. The lead is transformed into sulphate and weighed in the usual manner. The silver is determined by acidifying the solution with nitric acid and boiling it for a short time. The precipitated silver chloride is filtered, etc., as indicated on page 652. Separation of Tin. The nitrate remaining after the removal of the silver chloride contains the gold, lead, and tin. These metals are precipitated as sulphides by means of hydrogen sulphide. The sulphides are washed as in- dicated on page 309, and dried. The paper is ignited after the sulphides have been transferred to a porcelain boat. The ash of the filter paper is also brushed into the boat, and the tin volatilised by heating the boat, etc., in a stream of bromine (fig. 145). 4 Separation of Gold and Lead. The boat containing the gold and lead bromides is again introduced into a combustion tube and heated in a current of hydrogen gas with the usual precautions against explosion. Metallic gold and lead remain in the boat. The latter is alone soluble in dilute nitric acid, hence the separation is easy. ,Add the weight of lead sulphate finally obtained to the lead sulphate obtained from the silver, and compute the corresponding amount of lead oxide in the usual way. The gold is ignited, weighed, and reported as metallic gold. The amount of gold in these colours is relatively small, and there is a slight loss of gold with each precipitation. The silver and gold are best determined by the fusion, cupellation, and parting processes on a separate sample, and the tin and lead by usual processes. In illustration, the analysis of a purple of Cassius colour furnished : Si0 2 . Al. 2 0.<. Sn0 2 . Na 2 0. Ag. PbO. Au. B 2 3 . C0 2 . 15-10 1'21 30-19 4-49 7'00 29'23 1 '44 10'50 1-20 percent. The carbon dioxide is determined by the method of page 553. 1 Although hydrogen peroxide in alkaline solution precipitates metallic gold, the hydrogen peroxide in the presence of hydrochloric acid forms chlorine, which dissolves the gold as chloride. The tin is in the form of stannic chloride, and stannic chloride gives no precipitate with gold chloride. 2 It is sometimes difficult to dissolve all the silver chloride in ammonia. In that case, suc- cessive treatment of the precipitate on the filter paper with ammonia, boiling water, and dilute nitric acid will bring about the solution of the silver chloride. P. Jannasch, Eer., 26. 1496, 1893. 4 If bismuth be present, it will be volatilised with the tin, and the separation of bismuth and tin can be effected by the usual methods. DETERMINATION OF GOLD AND SELENIUM. 431 Mylius' Ether Process. 1 When acidified aqueous solutions of gold chloride in presence of many other chlorides are treated with ether, the ethereal solution which floats on the aqueous layer carries most of the gold chloride, and the chlorides of the other metals remain below in the aqueous layer. When 100 c.c. of ether were shaken with 100 c.c. of an aqueous solution containing the equivalent of 1 grm. of metal, Mylius found that the following percentage amounts passed into the ethereal solution : HgCl 2 . AuUl 3 . FeCl 3 . SbCl 3 . SnCl 4 . AsCl 3 . 10 percent. HC1 . . . 0'4 98'2 8 22 23 7 '3 1 per cent. HC1 . . . 130 85'0 tr. 0'3 0'8 0'2 And 3-0 per cent, of FeCl 3 ; 0'05 CuCl 2 ; 0'03 ZnCl 3 ; O'Ol NiCl 2 ; O'Ol Pt01 4 ; O'Ol PdCl 2 ; 0-02 H 2 IrCl 6 from the 10 per cent, acid solutions, and only traces from the 1 per cent, acid solutions. No lead or silver chloride was dissolved by the ether. Hence, by treating a solution of gold chloride containing 5-10 per cent, of gold and 5-10 per cent, of free hydrochloric and (HC1) four or five times with ether in a separating funnel, or in the apparatus, figs. 149 or 151, very good separations aro said to be effected. The ether is distilled off, and the gold reduced with sulphurous acid, etc. The method promises to be useful for the separation of gold from certain metals. 225, The Determination of Gold and Silver by Cupellation and Parting. The fusion and cupellation are performed very much as indicated on page 326. but some modifications must be introduced. 2 The powdered sample should be thoroughly mixed with the flux, so that as soon as the lead is reduced it may come into contact with the gold and silver. The lead dissolves these metals, and the solution finally collects at the bottom of the crucible. The amount of reducing agent added to the fusion mixture should be proportioned to give a button weighing about 30 grms. The sample under investigation may be reducing, oxidising, or neutral, and too much or insufficient lead may be obtained. Preliminary Assay. It is therefore necessary to make a preliminary assay by fusing an intimate mixture of 5 grms. of the powdered sample with : Litharge, 50 grrns. ; sodium bicarbonate, 18 grms. ; calcined borax, 5 grms. The mixture is placed in a crucible of such a capacity that the crucible is less than three-fourths filled, and covered with a layer of common salt. 3 The fusion is made as indicated on page 327. Break the cold crucible, if the molten mixture has not been poured into the mould. Hammer off the slag, and weigh the button. The button from 30 grms. of the sample should weigh between 16 and 20 grms. Hence, the button from the 5 grms. of the sample should weigh nearly 3 grms. Weigh the button. Suppose : (1) The button weighs nearly 3 grms. In that case the above-mentioned mixture is the right one. Hence, fuse 30 grms. of the sample with : Litharge, 80 grms. ; sodium bicarbonate, 18 grms. ; calcined borax, 5 grms. 1 F. Mylius, Zeit. anorg. Chem., 70. 203, 1911 ; F. Mylius and C. Huttner, Her., 44. 1315, 1911. 2 "Gold lustres," "liquid gold," etc., can be prepared for cupellation by evaporating a weighed quantity in a porcelain dish, and incinerating the carbonaceous residue to drive off the carbon. The residue may be dissolved in aqua regia and analysed by the wet process, or wrapped in sheet lead, cupelled, etc., for gold and silver. 3 Previously fused and ground to powder. 432 A TREATISE ON CHEMICAL ANALYSIS. in a suitable crucible. The mixture is covered with a layer of common salt as indicated on page 327. (2) The button weighs over 3 grms. The sample is therefore a strong- reducing agent, and this must be corrected by the addition of an oxidising agent, say, nitre. Too large a button gives low results for silver, since some silver is then lost owing to absorption by the cupel and by volatilisation, and this the more the greater the amount of lead present. EXAMPLE. Suppose that the button weighs 7 grms. Then 30 grms. of sample would give a button weighing 42 grms., that is, 24 grms. too much lead. Since 1 grm. of nitre oxidises about 4 grms. of lead, it will be obvious that of 24 = 6 grms. of nitre must be used with the mixture of litharge, etc., indicated in (1). (3) -If the weight of the button be less than 3 grms , the sample has but a slight reducing action, but not sufficient to give a button of normal weight. There is then a danger of some gold and silver being left in the slag. A little reducing agent charcoal, argol, etc. must then be added to the mixture indicated in (1). EXAMPLE. Suppose the button weighed 2 grms., the button from 30 grms. of the sample would weigh 2x6 = 12 grms. Hence, in order to get a button weighing 18 grms., enough reducing agent to give 6 more grms. of lead is needed. One grm. of argol is found to reduce 10 grms. of lead from litharge. Hence, it is necessary to add r ^ of 6 = 6 grm. of argol to the mixture indicated in (1). (4) If no button is obtained, the sample is either neutral or oxidising. In the former case, the fusion can be made with 2 grms. of argol, 30 grms. of the sample, and the normal fusion mixture. If the sample be oxidising, the button may still be too small. In that case, the weight of the button obtained will furnish data for calculating how much more argol will be required to give a button of normal weight. Example of Assay. The preliminary fusion of a sample of gold colour showed no button of lead, and the fusion of 30 grms. of the sample with 50 grms. of litharge, 1'5 grms. of argol, 18 grms. of sodium bicarbonate, and 5 grms. of calcined borax, gave a button weighing just less than 20 grms. If the charge in the crucible contains much nitre, and uncalcined borax, it is liable to froth over. To prevent this, start the fusion at a low temperature and raise the temperature very slowly. If the bubbling be still troublesome, it may be advisable to roast the sample in a shallow clay " roasting dish " in the muffle, and take portions of the roasted sample for the preliminary and final assay. The final results must of course be expressed in terms of the roasted sample. We may now assume that the fusion and the cupellation have been performed as indicated on page 326. The bead of silver and gold is weighed. The separate determination of the gold and silver involves the two operations of inquartation and parting. 1 Inquartation and Parting. The latter term is applied to the process of separating gold and silver by the action of acid on the bead or "prill" obtained during cupellation. Nitric or sulphuric acids may be used, but the former is almost universally employed. Silver is soluble in nitric acid, gold is not. If an alloy of silver and gold be digested with nitric acid, silver, not gold, is dissolved. It is necessary that at least twice as much silver as gold be present to ensure the dissolution of all the silver by nitric acid of specific gravity not less than 1'26, and boiling for half an hour. But, as a matter of fact, in parting all bullion assays, five times as much silver as gold is considered necessary. For successful parting, therefore, special attention must be paid to : 1 The "gold" obtained by burning "gold rags" from the "decorating shop" of a pottery is evaluated in this way. DETERMINATION OF GOLD AND SELENIUM. 433 (1) The concentration of the nitric acid. For general work, an acid of the strength recommended by Keller 1 may be used, namely, 1 part of concentrated nitric acid (sp. gr. 1'42) with 9 parts of distilled water. With this acid, Keller says the gold remains as a coherent mass, even if 500 times as much silver as gold be present. The beads are supposed to be boiled in the acid about 15 minutes. There are, however, certain advantages in employing acids of two different strengths, as indicated below. (2) The temperature of the acid. The acid should be boiling when the bead to be parted is dropped into the flask. If the bead be dropped into cold acid, and then heated up to the boiling point, the gold is liable to disintegrate into a finely divided condition especially if the proportion of silver is large, and this may lead to loss in subsequent operations. (3) The ratio -of silver to gold in the bead. Although the ratio 5:1 is usually recommended, this is not quite under control. If less than this amount of silver be present, more silver can be added to the crucible when the fusion is made for the cupellation ; or to the lead button during cupellation when gold alone, and not silver, is to be determined. If silver and gold are to be deter- mined, the bead, after cupellation, is weighed. This weight represents silver + gold. A small piece of pure sheet lead weighing about 2 grms. is folded in the shape of a hollow cone. The necessary amount of pure silver is placed in the cone, which is then closed and folded into a little packet. This is then cupelled. Instead of this recupellation, some prefer to alloy the silver with the bead by fusion on charcoal before the blowpipe. The latter operation requires some practice for successful work. The process of alloying silver with the cupelled bead in order to prepare it for parting is called " inquartation," because formerly at least three parts of silver to one of gold were considered necessary for success- ful parting. With practice, the amount of silver needed can be estimated from the colour of the bead. If the bead is white, it contains more than three parts of silver to one of gold, and inquartation may be unnecessary, or the bead may be inquarted with its own weight of silver. If the bead is greenish yellow, it probably con- tains less than three parts of silver to one of gold, and the bead is inquarted with about twice its weight of silver. If the bead is yellow or reddish yellow, the gold predominates, and it is inquarted with two or three times its weight of silver. A set of standard prills can be easily made for comparison. The "quarted" alloy is removed from the cup with the "bead forceps," cleaned with the "button brush," and hammered on an anvil to a flat disc about 1 mm. thick. The metal becomes hard and brittle during the hammering, and the disc is accordingly annealed by heating it to dull redness while supported on a clean cupel, and cooling it rapidly on a piece of brass foil. The bead is then rolled between two steel rollers so as to form a long strip. If the disc had been rolled without annealing, the edges of the strip would probably be rough, and little pieces would drop off during the action of the acid. The strip is again annealed and rolled. The roll is placed in a small parting flask 2 con- taining about 30 c.c. of nitric acid 3 (sp. gr. 1'2), previously heated to about 90. The acid is boiled 4 for about 20 minutes ; cooled ; and decanted off. The roll is washed twice with distilled water. 30 c.c. of boiling concentrated nitric acid 1 E. Keller, Trans. Amer. Inst. Min. Eny., 36. 3, 1905 2 There are several different forms. The flasks are better without a lip. Some prefer to conduct the parting in small glazed porcelain crucibles. The " cupping " described in the text is then unnecessary. 3 Free from chlorine, sulphuric and sulphurous acids, and sulphides. 4 If the bead turns black and the action stops, insufficient silver is probably present. The bead is then flattened and fused with two or three times its weight of silver and re-parted. 434 A TREATISE ON CHEMICAL ANALYSIS. (sp. gr. 1-3) are then added to the flask containing the roll. After 20 minutes' boiling, 1 decant off the acid, and wash three times by decantation with distilled water. 2 Fill the flask to the very top with distilled water, and place a close- fitting porcelain capsule "parting cup" over the mouth of the flask. Invert the flask (fig. 146). The roll settles in the parting cup. The flask is removed by raising its mouth to the level of the water in the crucible, and moving the flask at right angles away from the parting cup to allow the water to run away. The water is decanted from the parting cup. The cup and contents are dried by holding the cup with a pair of tongs in the flame of a Bunsen's burner until the cup is red-hot. 3 The gold is annealed, and changes from a soft dark brown condition to hard yellow gold. 4 When cold, weigh. Subtract the weight of the gold so obtained from the weight of the original prill to get the approximate weight of silver in the given sample. Errors. The errors in the cupellation process involve : (1) The retention of gold by the slag during fusion. If the amount of collect- ing lead is ample, the error is not serious. 5 Buttons less than about 20.grms. should not be used. The loss is largely depend- ent on the nature of the impurities which pass into the slag. (2) The loss of gold by volatilisation. Gold begins to volatilise about 1070, and silver at about 680 ; hence gold and silver may be volatilised during fusion and cupellation. The loss is greater the higher the temperature of cupellation. 6 (3) The absorption of gold by the FIG. 146. "Cupping." cupel. The amount varies with the nature of the bone ash used in making the cupel. The loss of gold is greater the higher the temperature of cupellation." (4) The retention of lead by the cupelled beads. A little lead is almost invariably retained by the bead cupelled at the regular temperatures. Hillebrand and Allen 8 found about 0*3 per cent, loss with a 0'09 grm. bead of gold. (5) The solution of gold by the parting acid. During parting a trace of gold 1 By first boiling with a dilute acid, and afterwards with a more concentrated acid, the removal of the silver is more complete, the gold becomes more compact, and there is less risk of disintegrating the residual gold. 2 If the gold disintegrates, there will probably be losses in transferring. 3 The drying must be carefully done, or particles of gold may be spurted from the cup with the steam. The drying, etc., may be done by placing the cup in the red-hot muffle for a short time. 4 If the dish shows a black stain after it has been heated, the washing was imperfectly done, and the analysis or assay should be repeated. If, after parting, the amount of gold is more than one-third the weight of the bead, it must be fused with about three times its weight of silver and again parted. 5 E. H. Miller and C. H. Fulton, School Mines Quart., 17. 160, 1896; W. F. Hillebrand and E. T. Allen, Bull. U.S. Geol. Sur., 253. 24, 1905. 6 G. H. Makins, Journ. Chem. Soc., 13. 77, 1860 ; T. K. Rose, ib., 63. 707, 1893; J. Napier, ib., 10. 229, 1858 ; W. Witter, Chem. Ztg., 23. 522, 1889 ; H. Rossler, Dingler's Journ., 206. 189, 1884; W. F. Hillebrand and E. T. Allen, Butt. U.S. Geol. Sur., 253. 19, 1905 ; K. Friedrich, Zeit. angew. Chem., 16. 269, 1903 ; J. W. Richards, Chem. News, 74. 2, 1896. 7 A. F. Crosse, Journ. Chem. Met. Soc. S.A., 2. 325, 1902 ; T. L. Carter, Eng. Min. Journ., 78. 728, 1902 ; W. F. Hillebrand, Bull. U.S. Geol. Sur., 253. 19, 1905. 8 W. F. Hillebrand and E. T. Allen, Bull. U.S. Geol. Sur., 253. 23, 1905. DETERMINATION OF GOLD AND SELENIUM. 435 appears to be dissolved by nitric acid, even if it be free from the deleterious impurities hydrochloric acid or chlorine. This amounts to about O'Ol to 0'03 per cent, on a half-gram bead. 1 (6) The retention of silver by the parted gold. The amount of silver which resists attack by the acid varies with the percentage of silver alloyed with the gold to be parted. 2 It amounts to from 0'05 to (HO per cent, under ordinary condi- tions, and it is generally considered to be least when the alloy contains between I'O and 2 - 5 per cent, of silver. If more or less silver be present, the gold will be contaminated with more silver after parting. Other foreign elements copper, tellurium, etc. may be retained by the bead. (7) The occlusion of gases by the gold. According to Graham, 3 about two parts by weight per 10,000 parts of gold are retained by the annealed gold. The amount depends upon the temperature of annealing. Some of these errors obviously lead to low results, others to high results. Experience shows that under normal conditions both sets of errors are small, and tend to neutralise one another. The algebraic sum of the losses and gains is called the surcharge, and the surcharge can be approximately determined by control experiments with the purest available gold and silver. Since the purity, i.e. the fineness, of gold is usually reported in parts per 1000, a surcharge of + 0'2 means that the gold actually reported weighed 0'2 part per 1000 more than was present ; and conversely with a surcharge of - 0*2, It must be added that the greatest care must be exercised in weighing the gold. A balance even more sensitive than -^^ mgrm. is a distinct advantage. The precautions indicated in the first chapter require serious attention. Under favourable conditions, and weighing correctly to 0*01 per 1000, Rose 4 considers that the operations can be conducted with an error not exceeding 0'02 part per 1000. 226. The Determination of Platinum and the "Platinum Metals." The Cupellation Process for Gold, Silver, and Platinum. When gold is determined by the process described in the preceding section, the bead will have a greyish colour if small amounts of platinum be present ; 5 and if larger amounts be present, the bead will have a rough " frosted " appearance, because it " freezes " before all the lead is oxidised. The method used for separating platinum, under these conditions, depends on the solubility of platinum in nitric acid when the metal is alloyed with at least 12 times its weight of silver. 6 Platinum can be separated from gold by cupelling the cornet with about 12 times its weight of silver, and parting with hot nitric acid (sp. gr. 1'42). 7 The residual gold is 1 A. H. Allen, Chem. News, 25. 85, 1872 ; V. Lenher, Journ. Amer. Chem. Soc., 26. 552, 1904. 2 W. F. Lowe, Journ. Soc. Ohem. 2nd., 8. 687, 1889 ; W. F. Hillebrand and E. T. Allen, Bull. U.S. Geol. Sur., 523. 25, 1905; T. K. Rose, The Metallurgy of Gold, London, 485, 1906. 3 T. Graham, Phil. Trans., 156. 433, 1866. 4 T. K. Rose, Journ. Chem. Soc., 63. 700, 1893. 5 In cupelling alloys of platinum, gold, silver, and lead, the greater the proportion of platinum the higher the temperature required for the cupellation. Alloys with over 50 per cent, of platinum cannot be freed from lead except at the temperature of the oxyhydrogen flame. 6 C. Winkler, Zeit. anal. Chem., 13. 369, 1874: P. Oehmichen, Berg. Hult. Ztg., 60. 137, 1901 ; M. Trenkner, Met., 9. 103, 1912; Min. Eng. World, 37. 342, 1912. 7 J. Spiller (Proc. Chem. Soc., 18 118, 1897) states that nitric acid (sp. gr. 1'42) will dissolve 0'75 to 1'25 per cent, of platinum along with silver; while a weaker acid (sp. gr. 1'2) will dissolve only 0'25 per cent, of platinum ; and a stronger acid will lead to the separation of platinum black. E. Pfiwoznik, Berg. Hutt. Ztg., 44. 325, 1895 ; H. Carmichael, Journ. Soc. Chem. Ind., 22. 1324, 1903. 436 A TREATISE ON CHEMICAL ANALYSIS. washed, dried, and weighed. 1 The operations are repeated until the residual gold has a constant weight, showing that all the platinum has been removed. 2 The silver solution is largely diluted with water, or, better, evaporated to drive off the excess of acid and then treated with a dilute solution of hydrogen sulphide. The silver sulphide which separates carries down the platinum as well. 3 Let the mixture stand overnight. Filter and wash in a porcelain dish. Dry the precipitate and ignite it with the filter paper at a low temperature. Wrap the residue in a small piece of "assay lead-foil," and cupel. The resulting bead is parted with concentrated sulphuric acid, when the platinum remains behind as a dark spongy mass. The sponge is again boiled with fresh acid, washed by decantation, dried, and weighed as platinum. 4 The first cupellation enables the gold, silver, and platinum to be estimated ; the result of the nitric acid parting gives the gold; and the sulphuric acid parting, the platinum ; the silver is obtained by difference. The Separation of Platinum by Potassium or Ammonium Chloride. Platinum 5 - can be separated from solutions containing the copper and aluminium groups 6 by almost neutralising the acid solution with ammonia, evaporating the solution to the crystallisation point, and adding water in sufficient quantity to just dissolve the crystals. Then add an excess of a saturated solution of potassium chloride to the feebly acid solution, and, after well mixing, add an excess of alcohol. Cover the vessel with a clock-glass, and let the mixture stand in a warm place for about 24 hours. Collect the precipitate on an asbestos-packed Gooch's crucible, and wash the precipitated K 2 PtCl 6 with 80 per cent, alcohol, as indicated for the determination of potassium (page 234). The potassium chloroplatinate can be dried and weighed ; or reduced to metal and the metal weighed. Ammonium chloride is preferable to potassium chloride as precipitant, because the resulting ammonium chloroplatinate, (NH 4 ) 2 PtCl 6 , can be washed with a saturated solution of ammonium chloride, and ignited very gently at first to avoid the risk of loss by spurting. The metallic platinum which remains behind can be weighed directly. If iridium be present in the solution, it would also be precipitated with the ammonium chloroplatinate, to which it imparts a reddish tinge. If the ignited mass be digested with dilute aqua regia (1 : 5) 7 at 40, the platinum dissolves and metallic iridium remains as an insoluble black powder. This is washed, dried, and weighed. 1 If the gold be very finely divided, it is best to filter and wash, so that the fine particles of gold will not contaminate the silver solution. 2 If palladium be present, it will dissolve witli the silver ; while if iridium be present, it will remain with the gold. E. H. Miller, School Mines Quart., 17. 26, 1896. 3 F. P. Dewey, Journ. Ind. Eng. Chem., 4. 257, 1912; Min. Eng. World, 36. 503, 1912. Traces of gold in solution of silver could also be gathered with the silver in a similar way. 4 To make sure the residue is platinum, dissolve it in a few drops of aqua regia, evaporate the solution to drive off the excess of acid, and test it qualitatively with potassium iodide, or ammonium chloride. If palladium was present in the original sample it will be found in the acid solution after parting. According to A. C. Dart (Met. Chem. Eng., 9. 75, 1911 ; 10. 219, 1912), the silver can be precipitated as chloride ; and the palladium precipitated as metal, by boiling the filtrate, made ammoniacal with ammonia and acidified with formic acid. The silver chloride, it may be added, is very liable to carry down traces of other metals. If iridium be associated with the gold, the latter can be removed by digesting the mixture at 40, with a mixture of nitric acid (sp. gr. 1'34) with three times its volume of hydrochloric acid, and all diluted with five times its volume of wacer. The washed and dried residue is iridium. The difference in the two weighings represents the gold. 5 The general properties of platinum were discussed in dealing with gold. 6 Metals precipitated by soluble chlorides silver, lead, and mercurous salts are supposed to be absent. 7 Iridium is soluble when heated with concentrated aqua regia. "Iridium grey" and " platinum grey" colours can be analysed by these processes. DETERMINATION OF GOLD AND SELENIUM. 437 General Analysis of the Platinum Metals. 1 It is not often that an analysis of this kind is needed. 2 The sample is treated with concentrated aqua regia at 70, and evaporated to a small volume. 3 More hydrochloric acid is added, and the solution evaporated to dryness. The mass is digested with warm water, and the insoluble residue probably osmiridium filtered off. Fuse one part by weight of sodium hydroxide in a copious nickel crucible, and add gradually an intimate mixture of one part of the insoluble residue with four parts of sodium peroxide. Keep the mass in a semi-fluid condition for about half an hour, and stir frequently with a nickel spatula. Dissolve the cold mass in dilute hydrochloric acid, and add the liquid to the main solution. The acid solution may contain salts of gold, osmium, rhodium, platinum, palladium, ruthenium, iridium, along with chromium, manganese, iron, etc., and of nickel from the crucible. 4 1. Removal of Osmium and Ruthenium. 1 * If the solution contains osmium and ruthen- ium, these elements are best removed at this stage by connecting a distilling flask (fig. 147) FIG. 147. Separation of osmium and ruthenium from the other "platinum metals." by means of ground glass joints with a pair of Wolbling's absorption flasks 6 contain- ing dilute hydrochloric acid (1 : 2). The Wolbling's flasks are placed in a dish con- 1 L. E. Rivot, Dodmasie, Paris, 4. 1103, 1866 ; R. Jagnaux, Analyse Chimique des Substances, Commercials, Minerales et Organiques, Liege, 1888 ; W. Crookes, Select Methods in Chemical Analysis, London, 437-477, 1905 ; H. St C. Deville and H. Debray, Ann. Chim. Phys. (5), 56. 439, 1859 ; R. W. Bunsen, Liebig's Ann., 146. 265, 1868; C. Glaus, Beitrage zur Chemie der Platinmetalle, Dorpat, 1854 ; M. C. Lea, Chem. News, IO. 279, 301, 1864 ; u. 3, 13, 1865 : Amer. J. Science (2), 38. 81, 248, 1864 ; W. Gibbs, ib. (2), 31. 63, 1861 ; (2), 34. 353, 1862 ; E. Leidie, Compt. Rend., 131. 888, 1901 ; Bull. Soc. Chim. (3), 27. 179, 1901 ; F. Mylius and F. Forster, Eer., 25. 665, 1892 ; H. Arnold, Zeit. anal. Chem., 51. 550, 1912; E. V. Koukline, Rev. Met., 9. 815, 1912 ; T. Wilm, Ber., 18. 2536, 1885. 2 Some difficulties in the manufacture of "platinum lustres" have been attributed to the presence of elements of this group other than platinum. 3 If osmium and ruthenium be present in the soluble portion, which is not likely, the operation must be conducted in a flask fitted with condensing tubes as illustrated in fig. 151. 4 E. Leidie and M. Quennssen, Bull. Soc. Chim. (2), 25. 840, 1901 ; (2), 27. 179, 1902. 5 A. Joly, Encyc. Chim. de Fremy, Paris, 3. 236, 1892. 6 H. Wolbling, Chem. Zty., 33. 499, 1909. 438 A TREATISE ON CHEMICAL ANALYSIS. taming ice-cold water. The liquid under investigation is poured into the distilling flask with an excess of sodium hydroxide, and a slow current of chlorine l is bubbled through the liquid. When the liquid in the flask is saturated with chlorine, raise its temperature to about 70. The osmium and ruthenium form tetroxides Ru0 4 and Os0 4 which collect in the condensing flasks. The liquid in the distilling flask must be kept alkaline 2 to prevent the action of hydrochloric acid on the iridium tetroxide, and the subsequent volatilisation of iridium chloride. When all the ruthenium has volatilised, 3 the current of chlorine is stopped, and the apparatus cooled. The ruthenium tetroxide in the con- densing flask forms a stable trichloride, RuCl 3 , while the osmium tetroxide undergoes no change. 2. The Separation of Ruthenium and Osmium. The liquid from the condensing flasks is treated with hydrogen sulphide so as to precipitate both sulphides. Wash and dry the precipitate, and heat it in a weighed platinum boat in a current of oxygen fig. 122. The ruthenium remains behind as oxide, while the osmium and sulphur pass on to the condensing flasks containing a 12 per cent, solution of sodium hydroxide mixed with 2 per cent, of alcohol. The solution of sodium osmiate in the condensing flasks has probably a violet colour. If not, warm the flask gently until a violet coloration does app'ear if osmium is present. Add strips of aluminium in small quantities at a time. The aluminium dissolves in the sodium hydroxide and reduces the osmium to metal. When the solution is decolorised, and all the aluminium has dissolved, wash the osmium by decantation first with water, and then with 5 per cent, sulphuric acid. Collect the precipitate in a weighed, stoppered filter tube (fig. 204) packed with asbestos. Wash with 5 per cent, sulphuric acid, and then with water. Dry at a dull red heat in a current of hydrogen, and when cold displace the hydrogen with carbon dioxide, and weigh. To check the result, expel the osmium from the tube by heating it at a red heat in a current of oxygen, and weigh again. 3. The Separation of Iridium. Boil the solution to expel free chlorine, and ignore the separation of a precipitate from the faintly acid solution. Add ammonium chloride to the solution, and then two-thirds its volume of alcohol. Let the mixture stand about 24 hours. The precipitate contains most of the platinum and iridium, and traces of palladium and rhodium. Filter and wash the precipitate with dilute alcohol. The nitrate is evaporated almost to the point of crystallisation, and if a solid separates, filter and wash it with a solution of ammonium chloride. Call the filtrate A. Calcine the two mixed precipitates in a Rose's crucible (fig. 138) in an atmosphere of hydrogen. The reduced metals are digested with dilute aqua regia (1 : 5). The platinum and palladium pass into solution, the iridium and rhodium remain undissolved. Filter and wash. Call the filtrate B. The insoluble residue is carefully fused with potassium bisulphate so as to avoid loss by spurting (page 185). A rhodium salt is formed which passes into solution ; the iridium is oxidised, but it does not dissolve. The cold cake is digested in water acidified with sulphuric acid, and the insoluble mass is reduced by calcination in an atmosphere of hydrogen, washed, dried, and weighed as iridium^ The filtrate tinged red, pink, or yellow is boiled with an excess of sodium carbonate ; acidified with hydrochloric acid ; and the precipitate is calcined, and set aside, labelled No. I. 4. The Separation of Platinum. The dilute aqua regia solution, filtrate B, contains platinum and a trace of palladium ; it is evaporated to dryness with sodium carbonate ; calcined in a Rose's crucible in an atmosphere of hydrogen ; washed with boiling water ; and the insoluble mass weighed as platinum. 5 The filtrate and washings are rejected. 1 Say, from a chlorine bomb and a wash-bottle (fig. 147). - If necessary, add more sodium hydroxide via the stoppered funnel. :f A drop of the distillate will give no blackening with hydrogen sulphide if the action is over. 4 This may contain a trace of rhodium. For a further purification, see Glaus (I.e.), and Deville and Debray (I.e.). 5 The precipitate may contain traces of iridium and palladium. For a further purification, see Deville and Debray (I.e.). DETERMINATION OF GOLD AND SELENIUM. 439 5. The Separation of Palladium. The alcoholic filtrate A is diluted with water, and the acid nearly but not quite neutralised with ammonia. The liquid is saturated with hydrogen sulphide. Filter and wash. Call the filtrate C. The precipitate, containing insoluble sulphides of gold, copper, palladium, and rhodium, is roasted and then digested in hydrochloric acid. The insoluble mass is labelled No. II.; it contains gold and rhodium; the filtrate containing palladium and copper is treated with potassium chloride and alcohol to precipitate brownish- red potassium chloropalladite, K 2 PdCl 4 . This is decomposed by ignition; washed ; dried ; and the palladium weighed as metal. The filtrate containing copper is treated in the usual manner (page 350). 6. The Separation of Rhodium and Gold. The filtrate C from the precipitated sulphides of gold, etc., contains possibly a trace of gold and of rhodium together with metals of the iron-aluminium group. When this liquid is boiled, the gold, rhodium, and some sulphur are precipitated. The precipitate is washed, calcined, and digested with hydrochloric acid, and labelled No. III. The filtrate is treated with nitric acid and evaporated to dryness ; the residue is calcined ; washed with hydrochloric acid ; dried ; and labelled No. IV. The washings contain the iron, nickel (from the crucible), etc. These elements are separated as usual. The four residues, labelled Nos. I. to IV., are mixed together and digested with aqua regia. The gold dissolves, and the rhodium remains insoluble. Filter and wash. Precipitate the gold from the filtrate in the usual manner ; wash, dry, and weigh the insoluble powder as metallic rhodium. 1 The above method gives a very fair approximation. The results check very fairly against more elaborate schemes. The operations may now be summarised : Digest in aqua regia. ; digest HC1. ,k Distil in Chlorine. IA mium; Ruthenium. Add NH 4 C1 ; alcohol. use KHS0 4 ; water. Evaporate Na 2 C0 3 . Reduce H ; digest aqua regia. Dilute ; pass H 2 S. IB 1 Reduce H. Boil Na 2 CO 3 . Iridium. Platinum. Add KC1 ; alcohol. Boil. Palladium. Copper. II. III. Boil HN0 3 , etc IV. I Iron, etc. Digest aqua regia. Rhodium. Gold. 227. The Detection of Selenium. Selenium and tellurium determinations 2 are very rarely needed in technical analyses for the silicate industries. Selenium is used in the preparation of 1 It will be observed that traces of the platinum metals which escape precipitation accumulate with the gold and rhodium. 2 Tellurium determinations are chiefly required in the analysis of cupriferous pyrites and auriferous ores for metallurgical purposes. 440 A TREATISE ON CHEMICAL ANALYSIS. certain reddish colours, and in bleaching and decolorising glasses and special enamel frits. 1 In the latter case, the amount in question may vary from 0'002 to 0'004 per cent., and a qualitative test may then reveal whether or not selenium is present ; but quantitative determinations are only satisfactory if a relatively large amount of the material is taken for the analysis. In the regular course of qualitative analysis, selenium (and tellurium) appear as members of the sub-group of elements whose sulphides are preci- pitated by hydrogen sulphide in acid solutions, and whose sulphides are also soluble in sodium monosulphide. If the selenium sulphide be precipitated from a cold solution, the lemon -yellow selenium sulphide SeS 2 is readily dissolved by alkaline sulphides. If the solution be hot, the precipitate is orange-yellow, and is much less soluble. Hydrogen sulphide separates selenium from the elements which do not give precipitates with hydrogen sulphide in acid solutions, but selenium sulphide 2 cannot be quantitatively separated from the mixed sulphides of the hydrogen sulphide group by the action of sodium or ammonium sulphides. Part of the selenium passes into solution, and part remains associated with the insoluble sulphides. In fact, from two-thirds to seven-eighths of the total selenium may remain with the insoluble sulphides of copper, bismuth, and lead. For example, a solution of copper and selenium in dilute hydrochloric acid, just acid enough to prevent the precipitation of copper selenite, was treated 3 with hydrogen sulphide. The precipitate was digested with concentrated sodium monosulphide, washed with water containing a little sodium sulphide, the water containing a little hydrogen sulphide, and finally with strong alcohol. The insoluble part, in four experiments, contained : Sulphur .... 0-3787 0-1649 0'1730 0'0416 grm. Selenium . . . . 0'2828 0'1726 0-1565 0'0239 grm. Copper ._ .. . . 0-9880 0'4880 0'4909 0'0979 grm. Hence, the alkali sulphide effects a very imperfect separation when selenium and copper are together precipitated in an acid solution. The process, however, is used in qualitative analysis. Selenium is precipitated from the ammonium or sodium sulphide solution by treatment with acids. 4 The washed and dried precipitate is mixed with twice its weight of a flux made by mixing equal parts of sodium nitrite and sodium carbonate. The mixture is added to the crucible containing two parts by weight of previously fused sodium nitrite. The fluid mass is poured on to a porcelain slab, and the cold masses (in the crucible and on the slab) are extracted with water sodium antimoniate, stannic oxide (gold, platinum, and iridium) remain insoluble ; selenic (telluric), molybdic, and arsenic acids are dissolved. Add an excess of hydrochloric acid, and boil the solution to reduce the sodium selenate, Na 2 Se0 4 , to selenious acid, H 2 Se0 3 : Na 2 Se0 4 + 4HC1 = 2NaCl + H 2 Se0 3 + H 2 + C1 2 . Then boil the solution with ammonium sulphite, sulphurous acid, hydrazine 1 F. Kraze, Sprechsaal, 45. 214, 227, 1912 ; P. Fenaroli, Chem. Ztg., 36. 1149, 1912. - According to B. Rathke (Liebig's Ann., 152. 181, 1869; er., 18. 1534, 1885 ; 36. 594, 1903), the precipitate is a mixture of SeS 2 , Se 2 S, and S. Some hold that the precipitate from selenious acid is a mixture of selenium and sulphur. H. Rose, Pogg. Ann., 107. 186, 1859 ; 113. 473, 1861 ; E. Divers and M. Shiraose, Chem. News, 51. 199, 1885. 3 E. Keller, Journ. Amer. Chem, Soc., 19. 771, 778, 1897. 4 The ammonium or sodium sulphide solution contains selenium, molybdenum, tin, anti- mony, arsenic, tellurium, gold, platinum, etc. DETERMINATION OF GOLD AND SELENIUM. 44! sulphate, hydroxylamine, 1 or some other suitable reducing agent. 2 Brown or red-coloured selenium is precipitated. 3 For opening glasses and frits, treat 5 grms. of the finely powdered material with sulphuric and hydrofluoric acid in the usual manner. Evaporate the solution to dryness. Digest the residue with water, and filter. Selenium can be precipitated from the residue by boiling with a reducing agent. The aqueous extract of the sodium carbonate fusion can also be acidified and treated in a similar manner. 228. The Gravimetric Determination of Selenium Sulphurous Acid Process. Mix the powdered substance with 6 parts by weight of sodium carbonate and 1 part of sodium nitrite in a nickel crucible. 4 Raise the temperature very gradually until the mass is fused. Extract the cold mass with water. Sodium selenate passes into solution. 5 Acidify the solution with hydrochloric acid, and boil it until chlorine 15 is no longer evolved starch paper held in the steam will show this. 7 The boiling solution is saturated with sulphur dioxide. 8 Selenium is precipitated. The boiling is continued for about 15 minutes. 9 The 1 P. Jannasch and M. Miiller, Ber., 31. 2388, 2393, 1898; A. Gutbier, Ber., 34. 2724, 1901; R. Rosenheim and M. Weinheber, Zeit. anorg. Chem., 69. 266, 1911. For tellurium it is recommended that the reduction be effected in a weakly alkaline solution with a 10 per cent, solution of hydrazine sulphate contained in a beaker placed in an autoclave heated to 130 under a pressure of 3 or 4 atmospheres. 2 Phosphorous acid (H. Rose, Handbuch der analytischen Chemie, Braunschweig, 2. 441, 1871 ; A. Gutbier, Zeit. anorg. Chem., 41. 448, 1904) ; potassium iodide and hydrochloric acid (A. W. Peirce, Amer. J. Science (4), I. 416, 1896 ; Zeit. anorg. Chem., 12. 409, 1896; F. A. Gooch and W. G. Reynolds, ib. (3), 50. 254, 1895 ; W. Mathmann and J. Schafer, Ber., 26. 1008, 1893) ; glucose (F. Stolba, Zeit. anal. Chem., u. 437, 1872); magnesium or aluminium (L. Kastner, ib., 14. 142, 1875) ; hypophosphorous acid (A. Gutbier and E. Rohn, Zeit. anorg. Chem., 34. 448, 1903; A. Gutbier, ib., 32. 295, 1902); stannous chloride (A. Grak and J. Petren, Svensk. Kem. Tidschrift, 24. 128, 1912). For a general criticism of reducing agents, see A. Gutbier, G. Metzner, and J. Lohmann, Zeit. anorg. Chem.. 41. 291, 1904. Hydrazine salts and sulphurous acid are recommended. 3 H. "Rose (Zeit. anal. Chem., i. 73, 1862; Pogg. Ann., 113. 472, 624, 1861 ; Chem. News, 5. 185, 1862) has shown that the presence of hydrochloric acid is an essential factor in the complete reduction of selenium salts. 4 Platinum crucibles are attacked by certain selenides. The crucible must be selected according to the nature of the material under investigation and the constituents to be sought. For selenium only, a porcelain crucible can be used. A nickel crucible with sodium peroxide fusions may also be used. 5 If a soluble alkaline selenate be in question, the fusion, of course, is not needed. 6 If the solution contains nitric acid, this must be removed by repeated evaporation with concentrated hydrochloric acid (page 427). 7 No appreciable loss of selenium will occur if the solution contains sodium or potassium chlorides (B. Rathke, Liebig's Ann., 152. 194, 206, 1869 ; Journ. prakt. Chem. (1), 108. 249, 1869 ; Zeit. anal. Chem. , g. 484, 1870). One of these salts is also necessary, if the nitrate be evaporated and boiled with hydrochloric acid, in order to make sure that all the selenium is precipitated. 8 H. Rose, Pogg. Ann., 113. 472, 1861 ; Zeit. anal. Chem., I. 73, 1862. Sodium or ammonium bisulphite may also be used. A. W. Peirce (Amer. J. Science (4), I. 416, 1896) reduces selenium with a mixture of potassium iodide and hydrochloric acid. The reduction is more rapid than with sulphur dioxide, and a second reduction of the filtrate is not necessary. For a volumetric process based on this reaction, see J. F. Norris and H. Fay, Amer. Chem. Journ., 18. 703, 1896 ; 23. 119, 1901 ; P. Klason and H. Mellquist, Arkiv Kem. Min. Geol., 4. xviii. 1, 1911 ; ib., 4. xxix. 1, 1912. The thiosulphate titration fails if arsenic be present because sodium thiosulphate is oxidised by arsenic acid. 9 According to Rose (I.e.), solutions of selenious acid cannot be heated above 100 without appreciable loss of selenium by volatilisation. In the case of nitric acid solutions, the removal of nitric acid by evaporation on a water bath as recommended by Rose (I.e. ; E. Divers and M. Schimose, Journ. Chem. Soc., 57. 439, 1890 ; P. Michaelis, er.,3p. 2827, 1897 ; R. E. Lyons and F. L. Shinn, Journ. Amer. Chem. Soc., 24. 1087, 1902) leads to low results. Boiling in a reflux condenser with concentrated hydrochloric acid, in order to remove the nitric acid, is not successful even after boiling for 6 hours. Lyons and Shinn add a quarter as much silver or zinc nitrate as is sufficient to combine with the selenium. The selenite so formed is stable at 442 A TREATISE ON CHEMICAL ANALYSIS. precipitate becomes more compact and easier to filter after boiling. The colour also changes from red to dark brown. Test the filtrate, or a little of the clear liquid, with more reducing agent say ammonium bisulphite to make sure that precipitation is complete. If not, add more reducing agent and repeat the sequence of operations. 1 Collect the red precipitate on a Gooch's crucible ; wash with hot water, then with alcohol, and finally with ether. Dry the crucible between 100 and 105, and weigh as selenium Se. 2 The following numbers represent the type of results which can be obtained with this process in the presence of a large excess of copper sulphate : Sensed . O'OIOO O'OIOO 00500 0-0500 0*0100 O'OIOO grra. Se found . 0'0094 0'0096 0-0507 0'0496 0-0103 0-0094 grm. Effect of the Acidity of the Solution on the Precipitation of Selenium and Tel- lurium by Sulphur Dioxide. This method of precipitating selenium boiling with sulphur dioxide in the presence of hydrochloric acid separates selenium from all metals not precitated by sulphur dioxide, and also from arsenic, antimony, and tin. 3 If much antimony be present, however, tartaric acid must be added to prevent the separation of antimony oxychloride, unless the solution be very strongly acid. The concentration of the hydrochloric acid needs special attention. There is a maximum and a minimum concentration of the acid between which " best " results can be obtained. Keller's important experiments 4 on this sub- ject show that with hydrochloric acid of specific gravity 1*175, and with solutions containing O'lOOO grm. of tellurium, arid also solutions containing O'lOOO grm. selenium, and saturated in the cold with sulphur dioxide, the amounts of selenium and tellurium which separated from solutions of different acidity were as follows : Table LVIII. Effect of the Acidity of the Solution on the Precipitation of Selenium and Tellurium by Sulphur Dioxide. Acidity per cent. Selenium. Tellurium. Acidity per cent. Selenium. Tellurium. 0-5 total 40 i total total 1 6 0-0761 50 total total 2 0-0525 60 total total 3 trace 0-0653 65 total 0-0965 5 0-0124 0-0745 70 total 0-0882 8 ... 0-0931 75 total 0-0411 10 0-0349 total 80 total nil 20 0-0935 total 90 total nil 30 total total 100 total nil 1 i 100, and insoluble enough in cold water to enable the precipitate to be washed free from nitric acid and nitrates. Evaporate the solution to dryness. Wash down the sides of the vessel with a little water ; evaporate again. Treat the residue with 50 c.c. of dilute ammonia and evaporate the solution to dryness ; again add ammonia, and again evaporate to dryness. The double Ag 2 Se0 3 .NH 3 so formed (B. Boutzoureauo, Ann. Chim. Phys. (6), 18. 289, 1889) is insoluble in water and stable at 100. Add water twice, and evaporate after each addition to drive off the ammonia. When cool, take up the residue with hydrochloric acid, filter, wash, reduce the selenium by sulphur dioxide, etc. 1 G. Pellini (Gazz. Chim. ItaL, 33. i. 515, 1903 ; G. Pellini and E. Spelta, ib., 33. ii. 89, 1903) adds 50 to 100 c.c. of a saturated solution of ammonium tartrate and warms the solution for a couple of hours at 50* to 60 with hydrazine sulphate. Selenium is said to be precipitated, tellurium not precipitated. Collect the precipitate in a Gooch's crucible in the usual manner. The tellurium is precipitated from the nitrate by hydrogen sulphide, etc. 2 For converting selenium into the dioxide by burning it in a current of oxygen, see P. Klason and H. Mellquist, ArUv Kern. Min. GeoL, 4. xviii. 1, 1911. 3 Some of the heavy metals may be partially precipitated by sulphur dioxide. 4 E. Keller, I.e. ; C. Alexi, Ueber die Bestimmung von Selen und Tellur und die Unter- suchung von selen- und tellurhaltigen Handelskupfer, Berlin, 1905. DETERMINATION OF GOLD AND SELENIUM. 443 These numbers are plotted in fig. 148. The left-to-right downwards cross-hatch- ing shows where selenium is wholly precipitated ; the right-to-left downwards cross-hatching shows where tellurium is wholly precipitated ; and left-to-right and right-to-left cross-hatching shows where both tellurium and selenium were precipitated. If selenium alone be present, total precipitation occurs if the acidity of the solution exceeds 30 per cent. If both selenium and tellurium be present, tellurium and selenium will be precipitated if the acidity of the solution be kept 20 4Q 60 80 100 FIG. 148. Influence of the acidity of the solution on the precipitation of selenium and tellurium. between 30 and 50 per cent. But outside these limits either selenium or tellurium may be wholly or partially precipitated, according to the strength of the acid. Selenium will alone be precipitated from a mixed solution if the acidity of the solution exceeds 80 per cent. 1 Hence, if selenium be first pre- cipitated from a solution over 80 per cent, acidity, and the solution made up to twice its former volume with water, the tellurium will be precipitated by a repetition of the sulphur dioxide treatment. The tellurium may be filtered off, dried, and weighed in a similar manner to the process described for selenium. 2 1 The precipitates are more voluminous from feebly acid solutions than from strongly acid solutions. 2 For the reduction of tellurium solutions with hydrazine salts, see A. Rosenheim and M. Weinheber, Zeit. anorg. Chem., 69. 266, 1910; P. Jannasch and M. Miiller, Ber., 31. 2377, 1898; A. Gutbier, ib. , 34. 2724, 1901. For the determination of tellurium in minerals, etc., see C. H. Fulton, Journ. Amer. Chem. Soc., 20. 586, 1898; E. Keller, ib., 19. 771, 1897; V. Lehner, -ib., 21. 347, 1899 ; E. Donath, Zeit. angew. Chem., 3. 216, 1890 ; R. W. E. Maclvor, Chem. News, 86. 308, 1902 ; T. Egleston, ib., 47. 51, 1883. CHAPTER XXXII. THE DETERMINATION OF ALUMINIUM AND BERYLLIUM COMPOUNDS. 229. The Gravimetric Determination of Alumina Hess and Campbell's Process. THE errors in the determination of the titanium and iron accumulate on the alumina. So also do the errors arising from an imperfect ignition of the mixed oxides (Fe 3 ->Fe 3 4 ), imperfect washings, and imperfect separations. It is therefore interesting to find the magnitude of the error affecting the alumina under the conditions of the analysis by the process just indicated. We have seen that with the clay in question there was an error of 0*07 per cent, in the determination of the titanic oxide, an error of 0*06 per cent, in the determination of the ferric oxide, and an error of 0'13 per cent, in the determination of the mixed oxides. Consequently the error in the determination of the alumina by difference l amounts to about 0'09 per cent. This is probably not far different from what would be obtained by a direct determina- tion. The error does not multiply up as it did when the alkalies were determined by difference (page 222). The above estimate of the error becomes less favour- able, the less the proportion of alumina in the mixed precipitate, and the greater the proportion of ferric oxide and titanic oxide. A small positive error in the determination of the two oxides iron and titanium might altogether mask small amounts of alumina. It is therefore desirable to use a direct method for the de- termination of small amounts of alumina in the presence of large amounts of iron. Several processes have been suggested for the direct determination of alumina, but most of them are not very satisfactory in general work. The sodium thio- sulphate process will be discussed later, when dealing with zirconium (pages 495, 501). Hess and Campbell 2 give a process which promises well. After a critical examination of the method, Allen said that "minute quantities of alumina may be accurately separated from large amounts of iron by this process." It is based on the fact that aluminium hydroxide is quantitatively precipitated from solutions by the addition of phenylhydrazine. Any phosphorus which might be present will be precipitated with the aluminium hydroxide as aluminium phosphate. Chromium, 3 titanium, zirconium, 4 and thorium, 5 if present, will be 1 For the methods of calculation, see J. W. Mellor, Higher Mathematics, London, 528, 1909. 2 W. H. Hess and E. D. Campbell, Chem. News, 8l. 158, 1900 ; Journ. Amer. Chem. Soc., 21. 776, 1899; E. T. Allen, ib., 25. 421, 1903 ; Chem. News, 89. 43, 63, 76, 88, 103, 1904 ; M. Wunderand N. Cheladze, Ann. Chim. anal., 16. 205, 1911. 3 The precipitation of chromium is slowest in chloride solutions, and quickest in nitrate and sulphate solutions. 4 A. M. Jefferson (Journ. Amer. Chem. Soc., 24. 543, 1902) says that zirconium is not pre- cipitated by phenylhydrazine, but Allen shows that this is a mistake. 5 Thorium is precipitated completely from nitrate solutions, cerium only partially. Titanium and zirconium can be separated from beryllium in chloride solutions by this precipitant. DETERMINATION OF ALUMINIUM AND BERYLLIUM COMPOUNDS. 445 precipitated with the aluminium hydroxide ; but neither ferrous iron, 1 beryllium, 2 manganese, zinc, cobalt, nickel, cadmium, mercury, magnesia, barium, strontium, nor calcium will be precipitated. 3 First Precipitation. The solution is diluted to 100 or 200 c.c., according to the amount of alumina to be precipitated, and heated to about 90. Ammonia is added until the precipitate formed just redissolves. Add 5 to 20 drops, according to the amount of iron in solution, of a neutral saturated solution of ammonium bisulphite, 4 with constant stirring, until the solution smells strongly of sulphur dioxide. 5 If the solution turns deep red, it is not acid enough. The red colour shows that ferric sulphite is present. In that case, add a few drops of hydrochloric acid, then add ammonia quickly 6 to make the solution neutral as before. Add a couple of drops of dilute hydrochloric acid. Add phenyl- hydrazine 7 (1 to 3 c.c.), with constant stirring, to the liquid smelling strongly of sulphur dioxide, until all the alumina is precipitated. When the precipitate has become flaky and settles quickly, filter in the usual manner. The filtra- tion should be done quickly, or a brownish scum may form on the surface of the solution and on the sides of the dish. The solution should not stand more than an hour before filtration. The precipitate is generally coloured brown. The colouring agent is not necessarily ferric hydroxide, but rather organic impurities in the precipitating agent. Wash the precipitate with hot water containing dilute phenylhydrazine bisulphite 8 in solution. Continue the washing until the wash-water gives no indication of iron when mixed with a few drops of ammonium sulphide, or no indication of chlorides. Test the filtrate with a couple of drops of phenylhydrazine to make sure that precipitation is complete. Second Precipitation. It is best to redissolve the precipitate on the filter paper in hot dilute hydrochloric acid ; wash with hot water ; neutralise the Beryllium is not precipitated from chloride solutions, but a large proportion is precipitated from sulphate solutions, and a trace from nitrate solutions. The method for the separation of aluminium and chromium from beryllium by this process, which naturally here suggests itself, does not work. 1 Phenylhydrazine is a powerful reducing agent, and it thus plays a double role: (1) main- taining the iron in the ferrous condition ; and (2) precipitating the alumina. Ferric oxide, it may be added, is partially precipitated by phenylhydrazine. It is necessary to convert the ferric into ferrous salts before adding the phenylhydrazine, because the latter does not reduce ferric salts very rapidly. 2 In chloride solutions, as indicated in a previous footnote. 3 If zinc, cobalt, nickel, mercury, or cadmium be present in concentrated solutions, the corresponding hydroxides may be precipitated. Molybdenum salts give a bright red coloration L. Spiegel and T. A. Maass, Ber., 36. 512, 1903. 4 AMMONIUM BISULPHITE SOLUTION. Pass sulphur dioxide into a cooled solution of ammonia (1:1) until the solution becomes yellow. The object of the sulphur dioxide is not only to reduce the iron, but to keep it in the reduced condition so as to prevent its precipitation with the alumina. 5 Sulphur dioxide is best purchased in "syphons" of the liquefied gas. The "syphons" are a convenient source of sulphur dioxide for analytical work. 6 If the operation be done slowly, some ferric hydroxide may be formed which does not dissolve readily in the dilute hydrochloric acid. 7 Phenylhydrazine costs about 5s. 6d. per Ib. To prevent a waste, it is best to add 1 or 2 c.c. of phenylhydrazine, and, if no precipitate is formed, add dilute ammonia drop by drop until a precipitate is just visible. Clear this up by adding one drop of acid, and add more phenylhydrazine. If too little phenylhydrazine has been added, a few drops of the filtrate will show the mistake. The phenylhydrazine should be free from foreign organic matters. Tin has been detected in commercial samples. For preservation, see G. Deniges, Bull. Trav. Soc. Pharm. Bordeaux, 52. 513, 1912. 8 PHENYLHYDRAZINE BISULPHITE SOLUTION. Add a cold saturated solution of sulphurous acid gradually to a few cubic centimetres of phenylhydrazine until the precipitate of phenyl- hydrazine sulphite, first formed, redissolves to a yellow solution. If the odour of sulphur dioxide appears after standing a few minutes, add a few drops of phenylhydrazine to neutralise the sulphur dioxide. This solution keeps indefinitely in a well-stoppered bottle. For use, mix from 5 to 10 c.c. of the solution so prepared with 100 c.c. of hot water. 446 A TREATISE ON CHEMICAL ANALYSIS. solution with ammonia ; acidulate with a couple of drops of hydrochloric acid (1:1); and precipitate the aluminium hydroxide from the small volume of liquid by the addition of, say, 0'5 c.c. of phenylhydrazine. Wash as before. The Ignition. The precipitate is dried in a platinum crucible, charred at a low temperature, and ignited at a bright red heat until its weight is constant, A second ignition is always advisable to make sure the weight is really constant. Weigh as alumina A1 2 3 . Errors. Allen found that when aluminium hydroxide is precipitated from pure solutions the results are inclined to be rather low. For example : Used, 0-2500 grm. A1 2 3 ; found, 0-2487 grm. A1 2 3 ; error, - 0*0013 grm. But the separation of aluminium, titanium, and zirconium from iron, from a solution con- taining a mixture of all four salts, gave excellent results. Thus : Used . . . 0-1817 0'1648 0'1059 0'0726 0'0595 grm. Found . . 0-1816 0'1654 0'1063 0'0717 0'0603 grm. Error . . . -O'OOOl + 0*0006 +0*0004 -0*0009 4-0 '0008 grm. If phosphorus be present, it must either be determined on a separate sample, or the ignited alumina fused with sodium carbonate or potassium bisulphate (page 185), and the phosphorus determined as indicated on page 595. If the phosphorus be in excess of that required to form aluminium phosphate A1P0 4 Hess and Campbell recommend the addition of an excess of a solution of aluminium chloride of known strength, and proceeding as indicated above, making due allowance for the extra alumina. If desired, the alumina and iron may be precipitated with ammonia in the usual manner (page 182), or with the basic acetate process (page 361), and the alumina precipitated as indicated above from the pyrosulphate fusion. The iron can be determined by difference, or precipitated with ammonium sulphide. 1 230. The Analysis of Bauxite. An accurate and complete analysis of this mineral can be made by a process similar to that employed for clays. Time may be saved in analyses for com- mercial work by using, say, Handy's process, 2 which may be regarded as a type of many others. In Handy's process, 1*5 grms. of the finely powdered bauxite, dried at 100, is mixed with 50 c.c. of acid, 3 in a 300-350 c.c. evaporating basin. Boil the solution until white fumes are evolved, and continue heating the mixture for another 15 minutes. Cool. Add 100 c.c. of water; stir; boil for 10 minutes. The silica is thus dehydrated and precipitated. Filter the solution into a 300-c.c. beaker. Wash the precipitate with water. The filtrate and washings will occupy between 150 and 200 c.c. The insoluble matter 4 is ignited in a platinum crucible and weighed. Add 3-4 drops of sulphuric acid and about 5 c.c. of hydrofluoric acid. Evaporate slowly to dryness. Ignite and weigh. The loss in weight represents silica. Fuse the residue with a little potassium bisulphate. When all is dissolved, cool. Take up the mass with water, and add the solution to the main solution. Any 1 Along with other metals precipitated by that reagent, if such elements be present. 2 J. 0. Handy, Journ. Amer. Chem. Soc.. 18. 766, 1896 ; H. Lienau, Chem. Ztg., 27. 422, 1903 ; 29. 1280, 1905 ; M. Taurel, Ann. CUm. Anal, 9. 323, 1904. For a rough method of evaluating aluminous minerals, see J. C. y Leon, Annies Soc. Espan. Fis. Quim.,S. 281, 1911. 3 Acid mixture: 100 c.c. of nitric acid (sp. gr. 1'42), 300 c.c. of hydrochloric acid (sp. gr. 1'20), and 600 c.c. of sulphuric acid (sp. gr. 1*18). 4 According to E. Baud (Rev. Chim. pure appl., 6. 368, 1903),' the residue will contain silica, and in addition titanic oxide, corundum, and a little alumina. These are subsequently brought into solution by the potassium bisulphate fusion, as described in the text. DETERMINATION OF ALUMINIUM AND BERYLLIUM COMPOUNDS. 447 insoluble residue will be silica, which is filtered off, ignited, and weighed. The weight is added to the previous result for silica. Make the mixed solution up to 300 c.c. Divide the 300 c.c. into three portions of 100 c.c. In one, precipitate a mixture of alumina, titanic oxide, and ferric oxide with ammonium chloride and ammonia. 1 The filtrate from this solution is used for the determination of lime (page 2J3) and magnesia (page 218).' 2 The precipitate is washed, ignited, and weighed (page 183). The titanium can be determined, if necessary, by the gravimetric process (page 207). In the second 100 c.c., determine the iron by reduction with zinc and the permanganate titration (page 187). 3 The alkalies are determined by Smith's process (page 222) ; the loss on ignition (organic matter and water), as on page 157. 231. The Analysis of Alumina Hydrated and Calcined. The alumina "hydrate" on the market may have from about 40 to 65 per cent, of A1. 2 3 . It is sometimes sold with a guarantee of 60 per cent. A1 2 3 . The loss on ignition (page 157) includes water and carbon dioxide. For the silica, heat 5 grms. of the hydrate with 25 c.c. of sulphuric acid (sp. gr. 1 '4). The alumina dissolves. Cool. Add 100 c.c. of water and boil. The insoluble silica separates. Filter., wash, and ignite the insoluble matter. Fuse the residue with a little potassium bisulphate. Take up the cold mass with water. Filter, wash, ignite in a platinum crucible, and weigh. Treat the mass with sulphuric and hydro- fluoric acids as described above. Weigh again. The loss in weight represents silica. Fuse the mass with potassium bisulphate, take up with water, and add it to the main solution. The iron can be determined colorimetrically in this solution (page 200). The sodium can be determined (1) in the hydrochloric acid solution of the hydrate by the method indicated page 239 ; or (2) use Smith's process (page 222) ; or, better still, (3) the sodium can also be determined by calcining 5 grms. of the "hydrate" at a red heat, 4 and digesting the calcined mass on a water bath with an excess of N-H 2 S0 4 , and titrating back with N-NaOH (page 70). This gives the "total sodium." The "soluble sodium" is deter- mined by boiling 5 grms. of the "hydrate" in water and titrating the filtered solution with N-H 2 S0 4 , using phenolphthalein as indicator. The difference between the "total" and the "soluble" sodium represents the "combined" sodium. Calculate the soluble sodium to per cent. C0 2 , and subtract the result from the percentage loss on ignition. The difference represents water. The alumina is usually determined by difference, although it can be determined by precipitation from an aliquot portion of the "main solution" by the method indicated under bauxite. Calcined alumina 5 may contain 99 per cent, of alumina. Fuse a gram of the dried alumina with, say, 10 grms. of potassium bisulphate. If the molten mass be not clear, add 2 more grams of the flux to the cold mass and fuse again. Dissolve the cold mass in water. Filter off the insoluble residue and fuse it with sodium 1 This method will give low results if a great excess of ammonium chloride be not employed (page 181). The sample can be fused with sodium carbonate, or sodium peroxide, and the cake taken up with dilute hydrochloric acid. 2 K. W. Jurich, Die Fabrication von schwefelsaurer Tonerde, Berlin, 45, 1894. 3 There is the difficulty with permanganate titrations mentioned on page 189, when much titanium is present. In that case special precautions must be adopted (see page 190). 4 This is to make the alumina insoluble in acid. 5 Alundum is made by fusing bauxite in the electric furnace. It contains about 1 per cent, of silica, ferric and titanic oxides. Alundum is scarcely attacked by aqueous acids and alkalies, and it is but superficially attacked by fused alkaline carbonates. 44^ A TREATISE ON CHEMICAL ANALYSIS. carbonate. Take up the cold mass with water in a 300-c.c. evaporating basin. Add 25 c.c. of sulphuric acid, and, when all action has subsided, evaporate the solution until the sulphuric acid begins to fume. Cool; add 100 c.c. of water. Boil the solution. Filter, wash, ignite, and weigh the residue. Treat the residue with sulphuric acid and hydrofluoric acid as described under bauxite. Ignite, and weigh again. The loss represents silica. For sodium, use Smith's process. The alumina is generally determined by difference. These methods will suffice to. demonstrate adulterations with china clay or ground flint. 232. The Analysis of Cryolite. Silica. Although cryolite contains a large amount of fluorine, the rela- tively small amount of silica, which is also present, does not all react with the fluorine, to form volatile silicon fluoride, when the powdered mineral is digested with sulphuric acid. Part remains with the residue undecomposed. This fact makes the determination a little laborious. The silica must be sought both in the residue and in the escaping gases. The solutions cannot be acidified with hydrochloric acid and evaporated in the usual manner, because some silica would be then lost as silicon fluoride. Fresenius and Hin1 z J recommend a process which certainly has not the merit of simplicity. The method indicated on page 637 may be here employed. Fuse the cryolite with an excess of sodium carbonate, and extract the melt with water. 2 Precipitate silica, alumina, lime, etc., with Seemann's solution, and determine the silica, etc., as indicated on page 640. 3 For the sodium, use Smith's process (page 222). The processes indicated on pages 642 and 646 are available for fluorine. 233. The Detection of Beryllium. If beryllia be present, it will invariably be found associated with the alumina and ferric oxide precipitated by ammonia. The properties of beryllia are closely related to those of alumina. In testing for beryllium, 4 the silica is separated by evaporation, and the hydrogen sulphide group precipitated in acid solution in the usual manner. The nitrate is evaporated down to about 25 c.c., and, when cold, add a couple of grams of sodium peroxide (solid). Boil and filter the alkaline solution ; acidify the filtrate with nitric acid ; and add an excess of ammonia. If no precipitate be formed, beryllium is absent; if a precipitate be formed, transfer it to a small beaker containing 2 or 3 grms. of solid sodium bicarbonate per 20 c.c. of liquid (10 per cent, solution). Raise the temperature rapidly to boiling, and boil half a minute. Alumina is precipitated. Filter. Dilute the filtrate with 10 volumes of water (so as to make approximately a 1 per cent, solution) and boil. If beryllium be present, a precipitate will be formed. To distinguish beryllia from alumina, dissolve the precipitate in acid 5 ; nearly neutralise the solution with ammonia ; add ammonium carbonate. The white precipitate may be either aluminium or beryllium hydroxide. The latter alone 1 C. R. Fresenius and E. Hintz, Zeit. anal. Chem., 28. 324, 1889. - E. C. Uhlig (Chemical Analysis for (ttassmakers, Pittsburg, 80, 1903) recommends fusing the cryolite with sodium carbonate as if it were a typical silicate. He ignores the action of the fluorides on the porcelain basin, on the precipitation of alumina (page 180), and the volatilisa- tion of silica as silicon fluoride. 3 The nitrate is used for the fluorine determination (page 637). 4 C. L. Parsons, The Chemistry and Literature of Beryllium, Easton, Pa., 6, 1908. 5 Freshly precipitated beryllium hydroxide is readily soluble in potassium or ammonium carbonate, sodium hydroxide, or dilute acids. If the hydroxide be allowed to stand for some time at ordinary temperatures, or if it be heated with water, dilute ammonia, alkaline carbonates, or alkaline hydroxides, the hydroxide appears to become very sparingly soluble in these menstrua. F. Haber and G. van Oordt, Zeit. anorg. Chem., 38. 377, 1904. DETERMINATION OF ALUMINIUM AND BERYLLIUM COMPOUNDS. 449 is soluble in an excess of the reagent, and a white precipitate of basic carbonate is obtained on boiling the clear solution. 1 Again, both aluminium and beryllium salts give white precipitates with caustic soda ; with both, the precipitates are soluble in an excess of the reagent ; but the solution with beryllia alone gives a white precipitate on boiling the solution. Beryllium, by the way, is sometimes called " glucinum " and symbolised Gl. 234. The Gravimetric Determination of Beryllium- Parsons and Barnes' Process. When beryllium is present, two. points require special attention in the treat- ment of the ammonia precipitate : (1) When the precipitate is washed with water, and the ammonium chloride has nearly all gone, the beryllium hydroxide begins to pass through the paper, and this more rapidly than is the case with alumina under similar conditions. The addition of ammonium salts say ammonium acetate to the water used for washing effectively prevents this action, and the washing then presents no difficulty of this nature. (2) Small amounts of beryllium hydroxide may adhere very tenaciously to the walls of the vessel in which the precipitation is made. 2 Hence, after as much beryllia has been removed as is convenient, it is advisable to wash the walls with a little dilute nitric acid, and re-precipitate the beryllia which has been dissolved. The beryllia found in the ammonia precipitate along with, say, iron, chromium, and aluminium oxides can be separated by fusing the precipitate for two or three hours with sodium carbonate ; the beryllia remains with the iron on washing with water, while the aluminium and chromium pass into solution as aluminates and chromates. The beryllia and iron can be separated by fusion with sodium pyrosulphate, leaching with water, and precipitating the iron with potassium hydroxide. The precipitate is filtered and washed ; the filtrate is acidified with hydrochloric acid, and the beryllia precipitated by ammonia. 3 Parsons and Barnes' 4 process for the quantitative separation of beryllium is based on the insolubility of aluminium and ferric hydroxides in a 10 per cent, solution of sodium bicarbonate, and the ready solubility of beryllium hydroxide 5 in the same reagent. First Precipitation of Aluminium and Iron. After the first precipitation by ammonia, 6 the hydroxides are dissolved in hydrochloric acid, and oxidised, if necessary, with a drop of hydrogen peroxide or nitric acid. 7 Neutralise the 1 M. Wunder and N. Cheladze, Ann. Chim. Anal., 16. 205, 1911. 2 B. Bleyer and K. Boshart (Zeit. anal. Chem., 151. 748, 1912) study the precipitation of beryllium by ammonia and by ammonium sulphide in ordinary glass, in Jena glass, in porcelain, and in platinum vessels. The two last-named gave the best results. 3 M. Wunder and P. Wenger, Zeit. anal. Chem., 51. 470, 1912. 4 C. L. Parsons and S. K. Barnes, Journ. Amer. Chem. Soc., 28. 1589, 1906 ; S. L. Penfield and D. 1ST. Harper, Amer. J. Science (3), 32. 112, 1886 ; Chem. Neivs, 54. 90, 102, 1886 ; F. A. Gooch and F. S. Havens, Zeit. anorg. Chem., 13. 435, 1896 ; F. S. Havens, ib., 16. 15, 1897 ; L. A. Aars, Uebcr die analytische Bestimmung von Beryllium und den sogenannten selten Erden, Kristiania, 1905. 5 Uranium, if present, also dissolves. 6 In the case of the basic acetate separation, C. L. Parsons (Journ. Amer. Chem. Soc., 26. 738, 1904 ; C. L. Parsons and W. 0. Robinson, ib., 28. 555, 1906) proposes to separate beryllium from iron and aluminium by digesting the mixed precipitate in hot glacial acetic acid, which dissolves basic beryllium acetate from the dried precipitate, and, after filtering the solution through a hot-water funnel, basic beryllium acetate separates from the solution on cooling. The small amounts of ferric and aluminium acetates which dissolve, do not separate on cooling. F. Haber and G. van Oordt (Zeit. anorg. Chem., 40. 465, 1904) propose to separate beryllium from the mixed basic acetate by treatment with chloroform, in which the basic beryllium acetate dissolves. 7 Although hydrogen peroxide is usually a better oxidising agent than nitric acid, it is far more likely to introduce impurities into the solution, owing to the greater purity of " commercial nitric acid." Merck's ' ' perhydrol " is the best form of hydrogen peroxide ; it is quite pure enough. 20 450 A TREATISE ON CHEMICAL ANALYSIS. solution with ammonia, and to the cold solution add 10 grms. of solid sodium bicarbonate 1 per 100 c.c. of solution. Cover the beaker 2 with a clock-glass, and heat the solution to boiling as quickly as possible. 3 Boil the solution one minute. 4 Place the beaker in cold water, and, when cold, filter. Collect the filtrate in a litre beaker, and wash the residue three times with a hot (70-80) 10 per cent, solution of sodium bicarbonate. Second Precipitation of Aluminium and Iron. The precipitated ferric and aluminium hydroxides generally retain a little beryllium. 5 Hence, the pre- cipitate is re-dissolved in as little dilute hydrochloric acid (1 : 1) as possible. Collect the filtrate in the same beaker in which the precipitation was first made. Make the solution up to about 100 c.c., neutralise with ammonia, and treat the mixture as the first one was treated. 6 Collect the filtrate and washings in the same litre beaker. Third Precipitation of Aluminium and Iron. It is impracticable to wash the sodium bicarbonate from the precipitated ferric and aluminium hydroxides ; hence, it is best to again dissolve the precipitate in dilute hydrochloric acid and re-precipitate the iron and aluminium hydroxides with ammonia in the ordinary manner (page 182). Precipitation of Beryllium. The joint filtrates from the two precipitations are carefully acidified with concentrated hydrochloric acid with the beaker covered so as to avoid " spurting losses " by the escaping gas. 7 Boil the solution to remove carbon dioxide. 8 Add a slight excess of ammonia, boil, let the precipitate settle, filter, wash with ammonium acetate (2 per cent, solution) until the washings are free from chlorides. Ignite, and weigh as beryllium oxide BeO. The Accuracy of the Separation. In illustration of the accuracy of the method of separation, Parsons and Barnes quote the following separations with mixtures of varying proportions of ferric, aluminium, and beryllium chlorides when the two former were several times the amount of the latter : Used . . - . . 0-2152 O'OQll 0'0825 0'1020 grin. BeO. Found . ... . 0-2146 0'0906 O'OSOG 0'0995 grm. BeO. Hence, the " separation is almost complete. The resulting beryllia was found to be analytically pure." Consequently, the method of separation just described " leaves little to be desired." 1 Free from all but traces of sodium carbonate as shown by the phenolphthalein test- page 62. 2 When the solution is warmed, loss by spitting may occur, if the beaker be not covered. 3 If too much carbon dioxide be lost on boiling, aluminium may pass into solution. 4 The brisk evolution of carbon dioxide must not be mistaken for boiling. 5 And uranium, if present. 6 A cloudiness in the filtrate is probably due to the dilution of the concentrated solution of sodium bicarbonate with water. 7 Just before neutralisation, some beryllium may be precipitated, but this redissolves on adding more acid. 8 Otherwise ammonium carbonate may be formed later on. Beryllium hydroxide is soluble in ammonium carbonate solutions. CHAPTER XXXIII. SPECIAL METHODS FOR IRON COMPOUNDS. "When Nature lakes up the brush, iron is almost always on the palette." R. J. HAUY. 235. The Volumetric Determination of Iron in Hydrochloric Acid Solutions Reinhardt's Process. IRON compounds are so common as components of colours, and the effects of iron are so important in the silicate industries, that it is necessary to indicate some alternative methods of analysis other than those described on pages 187 et seq. It lias been shown that the permanganate titration is useless in the presence of organic matter, 1 and of much hydrochloric acid, unless special precautions be taken. In the latter case, because of the disturbing effects of a side reaction, probably indicated by the equation : 16HC1 + 2KMn0 4 = 4C1 2 + 2MnCl 3 + 2KC1 + 8H 2 0. The chlorine acts as an oxidising agent and furnishes high results (page 194). This effect was pointed out by Lowenthal and Lenssen 2 in 1863. Kessler 3 showed that the presence of manganese sulphate inhibits the side reaction ; and Zimmermann 4 further proved that satisfactory results can be obtained in the presence of hydrochloric acid, provided sufficient manganese sulphate be present. Action of Manganous Sulphate. According to Manchot, 3 the first action of the permanganate in the titration of a ferrous salt is to form what he calls a " primary oxide " probably Fe 2 5 . In symbols : Mn 2 7 + 2FeO = Fe 2 5 + 2Mn0 2 . 1 Ferric salts in the presence of organic matter may be treated by adding a little copper sulphate as catalytic agent, and titrating the solution with standard thiosulphate until a drop gives no red coloration with ammonium thiocyanate. J. T. Hewitt and G. R. Mann, Analyst. 37. 179, 1912. 2 J. Lowenthal and E. Lenssen, Pogg. Ann., 118. 41, 1863; 119. 225, 1863; Zeit. anal. Chem., i. 329, 1862; E. Lenssen, ib., 2. 169, 1863; R. Fresenius, ib., I. 361, 1862; H. Kinder, Stahl Eisen, 27. 344, 1907 ; 28. 508, 1908 ; P. Lehnering, ib., 27. 202, 601, 1907 ; A. Miiller, ib., 26. 147, 1906 ; 27. 204, 1907; H. Wdowiszewski, Zeit. anal. Chem., 42. 183, 1903; H. Weber, ib., 46. 788, 1907; 47. 249, 1908; J. Brown, Chem. News, 93. 59, 1906; Amer. J. Science (i), 19. 31, 1905. 3 F. Kessler, Pogg. Ann., 118. 17, 1863 ; 119. 218, 1863 ; Zeit. anal. Chem., 2. 280, 1863 ; 21. 219, 1882. 4 C. Zimmermann, Btr., 14. 779, 1881; Liebig's Ann., 213. 305, 1882; H. P. Cady and A. P. Ruediger, Journ. Amer. Chem. Soc., 19. 575, 1897 (HgS0 4 ) ; C. T. Hood, Chem. News, 50. 278, 1884 (MgSOJ ; J. Krutwig and A. Cocheteux, Ber., 16. 1534, 1883 ; Chem. News, 48. 102, 1884 ; N. W. Thomas, Amer. Chem. Journ., 4. 359, 1882. H. Rose (Handbuch der analytischen Chemie, Braunschweig, 2. 926, 1871) says that potassium fluoride and potassium sulphate inhibit the reaction ; 0. Follenius (Zeit. anal. Chem., u. 177, 1872) denies this. 5 W. Manchot, Liebig's Ann , 325. 105, 1902 ; J. Volhard, ib. t 198. 337, 1879 ; A. Skrabal, Zeit. anal. Chem., 42. 359, 1903. 452 A TREATISE ON CHEMICAL ANALYSIS. The manganese peroxide so formed oxidises the ferrous salt to a ferric salt. In symbols : Mn0 2 + 2FeO = Fe 2 3 + MnO. In the presence of a sufficient amount of manganous salt, the ferric "primary oxide " Fe 2 5 oxidises the manganous oxide MnO back to Mn0 2 : Fe 2 5 + 2MnO = Fe 2 3 + 2Mn0 2 . On the other hand, if but little manganous oxide be present in the solution, and hydrochloric acid is also present, the latter is oxidised by the "primary oxide." In symbols : Fe 2 5 + 10HC1 = 2FeCl 3 + 5H 2 + 2C1 2 . "According to this theory," says Manchot, 1 "the manganese salt acts in two ways. In the first place, it modifies the velocity of the reaction between the ferrous oxide and the permanganate ; since, according to Volhard, the perman- ganate first reacts with the manganese salt to form manganese peroxide, which, in turn, reacts with the ferrous salt. In the second place, the manganous salt 'carries' oxygen from the ferric 'primary oxide' Fe 2 5 and gives it up to the unoxidised ferrous salt. In both cases, it is essential that the manganese peroxide shall not react rapidly with the hydrochloric acid, and that the quantity of manganous salt shall not exceed the amount of iron present in the solution." Whatever be the right explanation of the action, the ill effects of the hydro- chloric acid are counteracted when a manganous salt is present in suitable proportions, and the amount of hydrochloric acid is not too great. The Action of Phosphoric Acid. The presence of ferric chloride in the solu- tion interferes with the end point, on account of the yellow colour of this salt which is formed during the titration. This difficulty was overcome by Reinhardt, 2 who showed that in the presence of phosphoric acid the solution remains colourless until the pink colour of the permanganate appears. This is probably due to the formation of a colourless iron phosphate. Some contradictory statements have been published as to the efficiency of the subjoined Reinhardt's method 3 of conducting the permanganate titration in the presence of hydrochloric acid, but it can be made to furnish results as con- sistent as the best volumetric methods, provided a uniform and favourable pro- cedure be adopted. In rapid routine work, where the titanic oxide is neglected, 4 the solution obtained after filtering off the silica may be divided into two equal portions say 100 c.c. each. In one, the iron and alumina are precipitated in the usual manner; in the other portion, the iron can be determined by the permanganate titration in the following manner : The Reduction of Ferric Salts by Stannous Chloride? The amount of hydro- chloric acid in the 100 c.c. under investigation should not exceed 4N-HC1. 1 For the theory of induced reactions, see J. W. Mellor, Chemical Statics and Dynamics, London, 315, 333, 1904. 2 C. Reinhardt, Stahl Eisen, 4. 704, 1889; Chem. Ztg.. 13. 323, 1889; Zeit. anal. Chem., 36. 794, 1897. 3 L. Brandt, Chem. Ztg., 32. 812, 830, 840, 851, 1908 ; C. T. Mixter and H. W. Dubois, Journ. Amer. Chem. Soc., 17. 405, 1895 ; G. P. Baxter and J. E. Zanetti, Amer. Chem. Journ., 33. 500, 1904 ; G. P. Baxter and H. L. Frevert, ib., 34. 109, 1905 ; F. W. Harrison and F. M. Perkin, Analyst, 33. 47, 1908; G. C. Jones and J. H. Jeffery, ib., 34. 304, 1909; J. A. N. Friend, Journ. Chem. Soc., 95. 1228, 1909 ; W. C. Birch. Chem. News, 99. 61, 73, 1909 ; F. A. Gooch and C. A. Peters, ib. t Bo. 78, 91, 1899 ; Amer. J. Science (4), 7. 461, 1899 ; M. Wunder and A. Stoicoff, Ann. Chim. Anal., 17. 361, 1912. 4 If titanium is to be determined, reserve an aliquot portion of the solution for that purpose. 5 Ferric chloride solutions can be reduced by amalgamated zinc D. L. Randall, Zeit. anorg. Chem. t 48. 389, 1906. SPECIAL METHODS FOR IRON COMPOUNDS. 453 Heat the solution to boiling, add a solution of stannous chloride l gradually, drop by drop, with constant stirring, until all the ferric chloride is reduced to ferrous chloride. 2 It is important to add as little stannous chloride as possible in excess of that required to reduce the iron. 3 The difficulties incidental to this method of reduction are : (1) the danger of adding insufficient stannous chloride ; and (2) the tendency to add too much. If much iron be present, the colour is a fair guide when the reduction is nearly complete ; but when the colour is very faint, the stannous chloride is added until a drop of the solution brought quickly into contact with a drop of ammonium thiocyanate solution on a white tile gives no red coloration. When the solution is cold, add rapidly 5 c.c. of a solution of mercuric chloride 4 in order to destroy the excess of stannous chloride. If no precipitate forms, sufficient stannous chloride was not added, and some ferric chloride has not been reduced ; if a grey precipitate forms, too much s^annous chloride was added. This reacts with the permanganate, spoil- ing the titration. If a white precipitate of mercurous chloride is formed, it will do no harm, 5 and the solution, after standing 10 minutes, is ready for titration. The Titration ivith Potassium Permanganate. Wash the ferrous chloride solu- tion into a large 600-c.c. beaker, containing a mixture of 25 c.c. of Reinhardt's solution with about 300 c.c. of distilled water and one drop of the solution of permanganate. 7 Titrate the solution at once with permanganate, which is added gradually, drop by drop, with constant stirring. The calculations, etc., are made as indicated on page 198. The whole determination occupies about 20 or 30 minutes. Amounts of copper less than 0*01 per cent, may be neglected in ordinary technical analysis. If more than this amount be present, the copper should be removed. 8 236. The Volumetric Determination of Iron Penny's Dichromate Process. Instead of titrating for iron, in hydrochloric acid solutions, with potassium permanganate according to Reinhardt's process, Penny's dichromate process 9 is commonly used. The ferrous salt is here oxidised with potassium dichromate instead of potassium permanganate. The presence of hydrochloric acid 10 does no particular harm, although the end point is a little sharper with sulphuric than 1 STANNOUS CHLORIDE SOLUTION. Dissolve 29 grms. of crystalline SnCl 2 in 330 c.c. of concentrated hydrochloric acid, and make the solution up to -a litre with water. Add granulated tin and boil the solution until it is clear and colourless. Preserve the solution in a well- stoppered bottle. Some keep a stick of metallic tin in the stoppered bottle with the stock solu- tion. 1 c.c. of the stannous chloride solution will reduce about 0'015 grm. of ferric chloride. 2 Titanium chloride is not reduced by stannous chloride in hot or cold solutions, and hence the iron determination is not affected by the presence of titanium. :! W. F. Stock and W. E. .Jack, Chem. News, 31. 63, 1875 ; 30. 221, 1874. 4 MERCURIC CHLORIDE SOLUTION. A saturated solution containing 50 to 60 grms. per litre. 1 c.c. of the mercuric chloride will oxidise 1 '2 c.c. of the stannous chloride solution. 5 The mercurous chloride is not affected by potassium dichromate (Kessler), nor by potassium permanganate (Reinhardt). For the action of mercurous chloride on ferric salts (HgCl + FeCl 3 = HgClo + FeCl 2 ) see C. Meineke, Zeit. bffent. Chem., 4. 433, 1898. Hence the titration should be quickly done, to prevent the mercurous salt standing in contact with the ferric salt longer than is necessary A. E. Haswell, Rep. anal. Chem., 2. 841, 1882. 6 REINHARDT'S SOLUTION. Dissolve 200 grms. of crystalline manganese sulphate in 1000 c.c. of water ; add to this a cold mixture of 400 c.c. of concentrated sulphuric acid (sp. gr. 1*8), 600 c.c. of water, and 1000 c.c. of phosphoric acid (sp. gr. 1'3). 7 To oxidise traces of organic matter in the water, etc. 8 K. Schroder, Zeit. offent. Chem., 14. 477, 1908. 9 F. Penny, Chem. Gaz., 8. 330, 1850 ; L'Institut, 18. 27, 1850; Schabur, Sitzber. Akad. Wiss. Wien, 6. 396, 1851 ; F. Kessler, Zeit. anal. Chem., II. 249, 1872. 10 Organic matter decomposes hot or cold solutions as in the case of permanganate, but potassium dichromate is not so much affected as the permanganate. See page 195. 454 A TREATISE ON CHEMICAL ANALYSIS. with hydrochloric acid. The end point is not so easily determined in the dichro- mate process, because a "spot" indicator must be used. 1 Hence, it is easy to "over- titrate." The process is not quite so exact as the permanganate process ; but, with practice, there is little to choose between the two so far as accuracy is concerned. Dissolve 2 grms. of fused potassium dichromate 2 in a litre of water. Standardise this solution by dissolving 15*9716 grms. of Mohr's salt in a litre flask with water and concentrated sulphuric acid. Pipette 25 c.c. of this solu- tion into a beaker, and charge the burette with the solution of potassium dichromate. Place 10 or 12 drops of a freshly prepared solution of potassium ferricyanide 3 (1 : 100) separately as spots on a white tile. 4 At intervals during the titration, remove a drop of the solution undergoing titration on a glass rod and mix it with one of the drops on the tile "spot test." A blue coloration shows that the reaction is not complete and that the solution still contains ferrous iron. When a drop of the solution gives no sign of a blue coloration, the titration is complete. If a drop of the solution being titrated be touched against the side of a drop of the potassium ferricyanide, a little practice will enable one to judge when the end of the reaction is near. 5 The reaction is represented by the equation : 6FeCl 2 + K 2 Cr 2 7 + 14HC1 = 6FeCl 3 + 2CrCl 3 + 2KC1 + 7H 2 0. Zinc is objectionable as a reducing agent for the dichromate titrations, because the zinc ferricyanide formed in the "spot .test" interferes with the in- dicator. Reductions for the dichromate titration are usually made with stannous chloride (page 452). The solution reduced with stannous chloride is titrated with the standard dichromate as indicated above. The calculation is obvious. 1 A solution of s-diphenylcarbazide CO(NH.NH.C 6 H 5 ) 2 in the presence of an excess of hydrochloric acid has been suggested as an internal indicator. This turns violet when the reaction is over. P. Cazeneuve, Chem. Ztg. t 24. 684, 1900 ; Bull. Soc. Chim. (4), 23. 592, 701, 769, 1900; L. Brandt, Zeit. anal. Chem., 45. 95, 1906; 0. L. Barnebey and S. R. Wilson, Journ. Amer. Chem. Soc., 35. 156, 1913. The solution of s-diphenylcarbazide is prepared by dissolving O'l grm. of the carbazide in 35 c.c. of acetic acid and diluting the solution to 100 c.c Use 5 c.c. as indicator. The solution decomposes if kept very long. 10 grms. will suffice for 2000 titrations. The indicator reduces the standard dichromate, but when a suitable correction is made the results are good. 2 POTASSIUM DICHROMATE SOLUTION. Fuse about 5 grms of the pure crystallised salt in a platinum crucible below a red heat. The salt fuses in the neighbourhood of 400. Cool in a desiccator. The fusion must not be made at too high a temperature. R. W. Atkinson (Chem. News, 49. 117, 1884) states that, "however carefully the heat be regulated, on dissolving in water, and allowing to settle, a green deposit of chromic oxide will be found at the bottom of the flask or beaker. This shows that a small amount of decomposition occurs during the fusion, and that the value of the standard must be lower than is theoretically required for the amount of dichromate weighed out. " 3 The ferricyanide should be free from ferrocyanide, which gives a blue coloration Avith ferric salts. For the disturbing effects of copper, see J. S. Parker, Chem. Neivs, 22. 313, 1870. Potassium ferricyanide will show 1 part of iron in 500,000 parts of water A. Wagner, Zeit. anal. Chem., 20. 349, 1881. 4 Some prefer a slab of paraffin wax, or a white tile covered with a layer of paraffin, owing to the form the drops take on the wax, and the ease in cleaning the waxed slab. R. Seligmann and F. J. Willott, Journ. Soc. Chem. Ind., 24. 1278, 1905. 5 The development of the ferricyanide "blue" is slow with dilute solutions of ferrous salts, and the end point may thus be judged prematurely. To get the true end of the reaction, let the mixed drop stand 5 minutes covered with a crucible lid to shut out the light. The reaction is disturbed by elements which form insoluble ferricyanides the brown colour with manganese, nickel, or copper, for instance, masks the reaction with dilute ferrous solutions. Hence, the ferricyanide is generally made slightly "acid" to prevent the formation of insoluble ferri- cyanides. H. Vogel (Chem. News, 23. 142, 1871 ; R. W. Atkinson, ib., 49. 217, 1884 ; L. van Itallie, Pharm. Weekblad, 40. 490, 1903) says that 30 seconds' exposure to sunlight is sufficient to make potassium ferricyanide solutions give a blue colour with ferric salts. 6 For titrating ferric chloride solutions directly with stannous chloride, see page 310, where the converse process is described. For cobalt chloride or cobalt nitrate indicators, see F. P. SPECIAL METHODS FOR IRON COMPOUNDS. 455 237. The Gravimetric Determination of Iron Ilinsky and Knorre's Process. Several methods are available for the quantitative precipitation of iron in the presence of aluminium salts. 1 Ilinsky and Knorre's process 2 gives satisfactory results in separating ferric or ferrous iron from aluminium, chromium, man- ganese, nickel, zirconium, zinc, lead, cadmium, antimony, arsenic, magnesium, calcium, beryllium, but not vanadium, tungsten, copper, and cobalt. Silver, tin, bismuth, as well as molybdenum, interfere with the precipitation of iron, cobalt, vanadium, tungsten, and copper by the nitroso-/^-naphthol process. Some phosphorus, if this element be present, will be precipitated with the iron. The following is the method : The solution of the chloride or sulphate under investigation is neutralised with sodium carbonate, and acidified with a few drops of hydrochloric acid. The cold solution is stirred with a concentrated acetic solution of nitroso-j8-naphthol. 3 A voluminous precipitate of Fe(C 10 H 6 . N0) 3 separates on standing, say, overnight. Try if any more reagent causes any further precipitation in the clear solution. Wash the precipitate with 50 per cent, acetic acid, and finally with cold water. 4 The precipitate is dried and ignited in a capacious crucible, 5 and heated so that the temperature gradually rises. When all the carbon has burned off, the crucible is cooled, and the weight of Fe 2 3 determined. Trial results on known mixtures of alumina and iron are quite good. If necessary, the organic matter in the filtrate is destroyed by boiling with hydrochloric acid and potassium chlorate. 238. The Gravimetric Determination of Iron Baudisch's Cupferron Process. Baudisch 6 has shown that ammonium nitrosophenylhydroxylamine called for brevity " cupferron " or Baudisch's reagent promises to be a valuable agent for separating ferric and cupric salts from most other metals mercury, tin, lead, bismuth, and silver, however, may be partly precipitated with the iron and copper if these elements be present. The precipitated iron and copper salts Venable, Journ. Anal. App. Chem., I. 312, 1887; A. C. Campbell., ib., 2. 4, 1888; F. H. Morgan, ib., 2. 172, 1888. 1 Trimethylaraine precipitates iron in the presence of aluminium and chromium compounds L. Vignon, Journ. Pharm. Chim. (5), 12. 677, 1885 ; Zeit. anal. Chem., 26. 631, 1887. A solution of naphthenic acid in benzene precipitates ferrous oxide in presence of aluminium, etc. K. Charitschkoff, Chem. Ztg., 35. 463, 671, 1911 ; E. Pyhala, ib., 36. 869, 1912 ; 1. 1. Lutshinnskii, ib., 35. 1204, 1911; E. Pyhala, ib., 36. 869, 1913. Pyridine also precipitates iron in the presence of manganese, nickel, and cobalt R. B. Moore and I. Miller, Journ. Amer. Chem. Soc., 30. 593, 1908 ; Chem. News, 98. 105, 1908 and J. A. Sanchez (Bull. Soc. Chim. (4), 9. 880, 1911) recommends the pyridine precipitation for separating iron and manganese. Zinc should be absent, for it is precipitated by pyridine. Hydrazine hydrate also precipitates ferric iron quantitatively E. Schirm, Chem. Ztg., 35. 897, 1911. 2 M. Ilinsky and G. von Knorre, Her., 18. 2728, 1885 ; G. von Knorre, Zeit. anal. Chem., 28. 234, 1889; Zeit. angew. Chem., 6. 264, 1893; 17. 641, 1904; R. Burgass, ib., 9. 596, 1896 ; L. L. de Koninck, Rev. Univ. Mines (2), 9. 243, 1890 ; M. Schleier, Chem. Ztg., 16. 420, 1892 ; M. Beard, Ann. Chim. Anal., IO. 41, 1905. . 3 NITROSO--NAPHTHOL SOLUTION. 8 grms. of nitroso-/3-naphthol are dissolved in 300 c.c. of glacial acetic acid. The solution is diluted with. 300 c.c. of water, and filtered. The solution will keep about a month in a dark place. Nitroso-j8-naphthol costs about 8s. per 100 grms. 4 There is no particular difficulty in washing the precipitate, except its comparatively large volume. r> Capacious, because the precipitate swells very much during the earlier stages of the ignition. ti 0. Baudisch, Chem. Ztg., 33. 1298, 1909; 35. 913, 1911 ; 0. Baudisch and V. L. King, Journ. Ind. Eug. Chem., 3. 629, 1911 ; H. Biltz and 0. Hodtke, Zeit. anorg. Chem., 66. 426, 1910 ; H. Biltz and A. Soukup, ib., 68. 52, 1910 ; R. Fresenius, Zeit. anal. Chem., 50. 35, 1911 ; H. Nissenson, Zeit. angew. Chem., 23. 969, 1911. 456 A TREATISE ON CHEMICAL ANALYSIS. are but slowly attacked by 2N-HC1 in the cold ; but they are decomposed by hot acid. Cold dilute sodium carbonate has no appreciable action, but the alkaline hydroxides decompose the precipitate rapidly. Ammonia does not affect the iron salt, but it quickly dissolves the copper salt, and it is therefore possible to separate the iron and copper salts by digesting the mixed salts with ammonia, cold. The precipitation is best made in strongly acid (hydrochloric, sulphuric, or acetic x ) solutions, thus : Add 20 c.c. of concentrated hydrochloric acid to 100 c.c. of solution at the room temperature. Add a cold solution of Baudisch's reagent 2 slowly, with constant stirring. A reddish-brown (iron) or greyish-white (copper) flocculent precipitate separates. The end of the precipitation is easy to recognise, since small white crystals of nitrosophenyl-hydroxylamine C 6 H 5 (NO)OH separate as soon as all the copper or iron has been precipitated. It is then necessary to add a further excess of the reagent, say, one-fifth of the volume of that already added. In about 15 minutes, filter on a paper filter by suction. Wash first with 2N-HC1, then with water, and then with ammonia in the case of iron, and sodium carbonate in the case of copper, so as to remove the excess of the reagent. Finish the washing with water. The precipitate can be ignited very slowly in the ordinary way and weighed as oxide Fe 2 3 (or CuO). 3 Trial separations of iron from aluminium, nickel, chromium, and of copper from zinc and cadmium, are excellent. The chief value of the process lies in the fact that it furnishes a useful means of precipitating iron in the presence of cobalt, zinc, manganese, nickel, chromium, aluminium, alkaline earths, alkalies, phos- phates, and sulphates. 4 The separation of copper from cadmium and zinc by this process offers, as yet, no particular advantages over the older process (page 350). It has also been recommended for separating titanium from aluminium. 5 239. The Separation of Iron Ether Process. When a solution of ferric chloride with arsenic, antimony, aluminium, titanium, zirconium, chromium, vanadium, manganese, uranium, cobalt, nickel, copper, calcium, and magnesium chlorides is acidified with hydrochloric acid, and shaken with ether, the mixture separates into two layers an ethereal solution of ferric chloride above, and an aqueous solution of the other chlorides below. The two layers can be separated in a suitable funnel. Some gold, tungsten, and molybdenum, 6 if present, are more or less completely removed with the iron in the ethereal layer. The same remark applies to thallium, stannous and mercuric chlorides. 7 W. Skey first proposed to separate iron from 1 Acetic acid is best for copper. 2 BAUDISCH'S CUPFERRON REAGENT. Dissolve 6 grms. of the white crystalline solid ammonium nitrosophenyl-hydroxylamine in water and make the solution up to 100 c.c. The solution will keep in the dark for about a week. If exposed to the light, it becomes turbid owing to the formation of nitrobenzene. Old turbid solutions should be filtered before use. The crystalline solid costs about 8s. 9d. per 100 grms. 3 Finish the ignition with a Winkler's chimney (page 213). 4 For instance, say, 5 grms. of a brown iron ore are digested with 60 c.c. of concentrated hydro- chloric acid, and boiled with potassium chlorate. When cold, make the solution up to 500 c.c. Pipette 25 c.c. into a beaker, add 20 c.c. of hydrochloric acid and 100 c.c. of cold distilled water. Stir in the equivalent of 3 grms. of the cupferron reagent, etc. 5 I. Bellucci and L. Grassi, Atti R. Accad, del Lincei, Rome, 22. i. 30, 1913. 6 Molybdenum appears to be but imperfectly separated in the absence of ferric chloride W. Skey, Chem. News, 36. 48, 1880; 16. 201, 324, 1867; A. A. Blair, Journ. Amer. Chem. Soc., 30. 1229, 1908 ; R. de Jong, Zeit. anal. Chem., 41. 596, 1902 ; 0. D. Braun, ib., 2. 36, 1863 ; 6. 86, 1867. 7 The chlorides of the alkalies, calcium, nickel, zinc, and cadmium ; and the thiocyanates of nickel, copper, and zinc, are soluble in anhydrous but not in aqueous ether W. Skey, Chem. News, 36. 48. 1880. Stannous chloride is soluble in ether M. de Jong, Zeit. anal. Chem., 41. 596, 1902. SPECIAL METHODS FOR IRON COMPOUNDS. 457 many associated elements in this manner ; and he also suggested the correspond- ing separations : cobalt from nickel ; and gold from platinum. 1 Rothe, 2 however, has applied Skey's idea to the technical analysis of iron compounds, and the method has now won a place in analytical practice. Two or three extractions with ether are sufficient to remove practically all the iron. Effect of Acid. The acidity of the solution requires attention. 3 If the solution has much above or below " 20 per cent. HC1," the efficiency of the ex- traction will be impaired. For instance, Speller found that when 0'8 grm. of iron (as ferric chloride) was dissolved in 100 c.c. of hydrochloric acid, of the strength stated, shaken with twice its volume of ether, and allowed to stand 30 minutes at 17-18, the iron was divided between the ethereal and aqueous layers as indicated in Table LVIII." Table LIX. Relation between the Strength of the Acid and the Partition of Iron (as Ferric Chloride) betiveen Ether and Dilute Acid. Strength of hydrochloric acid. (Sp. gr.) Per cent, of iron. Aqueous solution. Ethereal solution. 1-193 99-0 1-0 1-164 97'5 2-5 1-158 92-84 7-16 1-151 74 26-0 1-123 7-2 92-8 1-115 2-4 97-6 1-111 2-0 98 1-105 1-95 98-05 1-103 1-95 98-05 T091 3-3 967 1-0825 10-0 90-0 1-069 47-5 52-5 1-06 87-0 13'0 1-0525 98'4 1-6 1-0424 99-6 0-4 1 K. Willstiitter (Ber., 36. 1830, 1903) separates gold chloride from platinum chloride in aqueous solutions by extracting with ether F. Mylius, Zeit. anorg. Chem., 70. 203, 1911. See page 431. 2 J. W. Rothe, Mitt, konig. tech. Ver., 10. 132, 1892; Chem. 'News, 66. 182, 1892; Stahl Risen, 12. 1052, 1892 ; A. Ledebur, ib. t 13. 222, 1893 ; E. Hanriot, Bull. Soc. Chim. (3), 7. 161, 1892 ; F. N. Speller, Chem. News, 83. 124, 1901 ; E. Pinerua, ib., 75. 193, 1897 ; A. A. Noyes, W. C. Bray, and E. B. Spear, Tech. Quart., 21. 14, 1908; Journ. Amer. Chem. Soc., 30. 481, 1908 ; G. W. Sargent, ib., 21. 854, 1899 ; A. C. Langmuir, ib., 22. 102, 1900 ; E. F. Kern, ib., 23. 685, 1901 (separation of iron from uranium) ; J. M. Matthews, Journ. Amer. Chem. Soc. , 20. 846, 1898 ; Cliem. Neivs, 79. 97, 112, 1899 (separation of iron from zirconium, thorium, cerium, and titanium) ; F. A. Gooch and F. S. Havens, ib., 74. 296, 1896 ; Amer. J. Science, (4), 2. 416, 1896 (separation iron and aluminium; ; A. Carnot, Methodes d* Analyse des Fontes des Fers et des Aciers, Paris, 123, 1895. For the separation of iron and vanadium, E. Deiss and H. Leysaht, Chem. Ztg., 35. 878, 1911 ; for iron and nickel, J. P. Thompson, Journ. Ind. Eng. Chem., 3. 950, 1911. G. L. N orris (Journ. Soc. Chem. Ind., 20. 551, 1901) prefers a mixture of acetone and ether rather than ether alone. 3 It must be pointed out that, since the ether initially contains no hydrochloric acid, the amount of hydrochloric acid in the aqueous layer is decreased by each extraction if ordinary ether be used. Hence, some use ether slightly acidified with HC1. Ether dissolves 3 per cent, of its volume of water. 10 c.c. of a solution of ether which had been shaken with an equal volume of hydrochloric acid (sp. gr. T1033) contained but 0'0019 grm. of HC1. Speller found that when 150 c.c. of anhydrous ether were shaken with 100 c.c. of hydrochloric acid, and allowed to stand 30 minutes between 17-18, 100 c.c. of the acid contained : 14-4 c.c. . 140 89 64 30 19 15'5 1-177 1-140 1-123 1-103 1-075 1'063 1-055 Ether Sp. gr. acid The presence of ferric, copper, cobalt, and nickel chlorides in the hydrochloric acid slightly modified the solubility of ether in the acid. 458 A TREATISE ON CHEMICAL ANALYSIS. Hence, hydrochloric acid of specific gravity 1 -100-1 '11 5 (25 '5 grms. of HCl per 100 c.c. of solution) is most favourable to the extraction of ferric chloride by ether. Effect of Phosphorus. If phosphorus be present, Wysor * has shown that there is an increasing amount of phosphorus lost in the ether layer with an increase in the proportion of iron and phosphorus. But "a fairly constant percentage of the phosphorus present in the solution remains with the iron in the ether separation. The approximate error may therefore be calculated and the proper correction applied." The first line in the following table represents the amount of soluble phosphorus pentoxide in the given sample : and the second line, the weight of phosphorus pentoxide to be added to the weight of the A1 2 3 residue found per gram of iron : PA .... Correction per 1 grm. Fe Correction per 2 grms. Fe Correction per 3 grms. Fe 0-0023 0-0046 0-0005 0-0009 0-0015 0-0017 0-0092 0-0182 0-0366 grm. 0-0016 0-0031 0-0060 0-0029 0-0055 0*0108 0-0032 0-0064 0'0130 It is assumed that the correction for the intermediate values of phosphorus and iron can be obtained by interpolation. Extraction of the Iron by Rothe's Pipette? Rothe's pipette, mounted ready for use, is illus- trated in fig. 149. Two mixing cylinders, each about 200 c.c., are connected with a 3-way cock. The other ends of the cylinder are fitted with 2-way stopcocks. The boring of the 3-way cock will be obvious from fig. 150, which shows five different positions, numbered 1 to 5. The 3-way cock is place in the position No. 1 ; the solution under investigation, 3 con- taining about 20 per cent, of hydrochloric acid, and occupying 50-60 c.c., is poured into the left cylinder by means of the long funnel as shown in the diagram, fig. 149. The beaker and funnel are washed with 20-23 per cent, hydrochloric acid. Close the left stopcock. Place the funnel in the other cylinder, and introduce 100 c.c. of ether. 4 Remove the funnel, and close the stopcock. Connect the ether cylinder with the rubber bellows, as shown in the diagram. Open the right stopcock and blow a small quantity of air into the ether cylinder. Turn the 3-way stopcock into the position No. 2, and ether will bubble through the solution into the left cylinder. The ether becomes warm as it mixes with the solution. 5 When about nine-tenths of the ether has bubbled through the solution, restore the 3-way cock to No. 1 position. Close the other two 1 R. J. Wysor, Journ. Ind. Eng. Chem., 2. 45, 1910. 2 Many other forms of " separating funnels " or " separating pipettes " have been devised, e.g. A. Carnot, Methodes d' 'Analyse des Fontes des Ferset des Aciers, Paris, 124, 1895. 3 The solution should be quite clear ; if not, filter. Nitric acid and chlorine should be absent. If present, they must be removed by evaporation. 4 No naked flames must be near while the ether is being used. Some ferric chloride is reduced to ferrous chloride. But this does not matter. i FIG. 149. Rothe's ether pipette. SPECIAL METHODS FOR IRON COMPOUNDS. 459 cocks. Cool the cylinder under a stream of cold water, and thoroughly agitate the solution under the water tap. Open the right cock. Force more air FIG. 150. Bore of stopcock in Rothe's pipette. into the right cylinder, and turn the 3-way cock into position No. 2. The remaining ether passes into the left cylinder. 1 Turn the 3-way stopcock into position No. 1. Close the two remaining cocks and shake the apparatus vigorously. Let the apparatus remain at rest two or three minutes. The liquid in the left cylinder separates into two layers. Open the two 2-way stopcocks, and turn the 3-way cock into position No. 2. When the lower aqueous layer has passed into the right c}4inder, 2 so that but a little of the aqueous layer remains in the capillary tube on the left, turn the 3-way cock into position No. 1. Close the two 2-way cocks. Shake the cylinder, whereby any drops of aqueous liquid adhering to the walls of the left cylinder collect on the bottom. The ethereal layer will be clear in 5-10 minutes. Then blow the remaining aqueous liquid into the right cylinder until a little of the ethereal layer appears in the capillary at the bottom of the right cylinder. Place a beaker below the 3-way stopcock, and turn this cock into the position No. 3. Open the left 2-way cock, and the ethereal solution containing the FeCl 3 flows into the beaker. 3 Turn the 3-way cock into the position No. 1. Pour a little ether into the left cylinder. Close the stopcocks ; shake ; and if any aqueous solution separates, blow it into the right ... . , , . , ,, T T a* ,1 j-i, u FIG. 151. Extraction of liquids with etner. cylinder and run off the excess ether by turning the 3-way cock into the position No. 3 as before. Turn the 3-way cock into position No. 1. Pour about 50 c.c. of ether into the left cylinder and repeat the preceding operations, simply reversing the cylinders. 4 This leaves the aqueous solution in the left cylinder. 1 This procedure prevents error arising from the non- extraction of the liquid in the capillary tubes connecting the two cylinders. 2 Assisted, if necessary, by connecting the blower with the left cylinder. 3 If much copper and cobalt be present, appreciable quantities of copper and cobalt chloride will be dissolved by the ether. In that case, add 10 c.c. of hydrochloric acid (sp. gr. 1'104) to the ethereal solution, shake, and draw off the aqueous layer containing the copper and cobalt into a second beaker. 4 Note the position of the 3-way cock Nos. 4, 2, 5 for working the right cylinder, compared with Nos. 1, 2, 3 while working the left cylinder. 460 A TREATISE ON CHEMICAL ANALYSIS. Repeat the operation a third time with another 50 c.c. of ether. This leaves the aqueous solution in the right cylinder. Treatment of Ethereal Layer. The three ether extractions have removed practically all the iron from the aqueous layer. The ethereal solution is trans- ferred to a porcelain basin, and evaporated on a water bath, not too hot, in order to drive off the ether, and evaporate to dryness. The residue may then be dissolved in hydrochloric acid and the iron determined in the usual manner (page 198). Treatment of Aqueous Layer. The aqueous solution freed from iron is also run off into another porcelain dish. The funnels are washed with 20 per cent, hydrochloric acid, and the solution is evaporated to dryness to remove the ether. It is then ready for any other separations which may be required. Extraction of the Iron by Soxhlet's Extractor. Instead of " shaking out " with ether by means of the separating funnel, it may be more convenient to extract the iron from the solution by a modification of Soxhlet's extractor, 1 say Taylor's extractor, illustrated in fig. 151. The solution 25-30 c.c. is placed in the receptacle A ; ether is placed in the flask B. The latter is heated by a water bath, or an electric lamp. The ether boils, the vapour is condensed in C. The condensed liquid runs into A, bubbles through the liquid, and is finally syphoned into the flask B. This circulation of the ether will extract all the iron from the solution in a short time. At the end of, say, an hour, when the solution has cooled, the apparatus is disconnected. The ethereal solution of iron in the flask j5, and the extracted solution in the receptacle A, are treated as described above. 240. The Analysis of Iron Oxides, Red Earths, and Iron Ores. The iron in iron ores, iron oxides, and red marls can sometimes be extracted by boiling the powdered sample with concentrated hydrochloric acid with or without a little nitric acid. Nitric acid should always be added if sulphides be present, in order to oxidise the sulphides to sulphates. 2 The silicate ores are sometimes more easily decomposed by digestion in acids if they be first calcined at a low temperature ; 3 and even if the ore dissolves readily in acids, the silica obtained in the subsequent evaporation (page 167) is cleaner and more pure than if the uricalcined ore had been digested directly in acids. 4 If the sample be not decomposed by digestion in acids, the sodium peroxide fusion may be employed, as described on pages 266 and 475. If iron alone is to be determined, the peroxide fusion is usually quite satisfactory for red clays, cal- cined iron oxides, etc. ; but several alternative methods have been suggested. 5 1 H. P. Smith, Chem. News, 83. 152, 1901 ; H. Gockel, Zeit. angew. Chcm., 10. 683, 1897. 2 Otherwise, the results of, say, the bichromate titration will be high. W. F. K. Stock and W. E. Jack, Chem. News, 30. 221, 1874. A. Esilman (Chem. News, 30. 243, 1874) re- commends filtering off the insoluble matters before titrating for the iron, in order to eliminate as far as possible the secondary effects of carbonaceous matters, pyrites, etc. 3 E. L&dd(Min. Eng. World, 36. 1350, 1912) proposes to avoid the fusion in the analysis of china clays by utilising this principle, but the method cannot be recommended. W. H. Worthington (Min. Science, 63. 521, 1911) recommends opening the sample of ore (in the absence of lime and baryta) with potassium bisulphate, because the silica can be filtered immediately the aqueous extract of the cold cake has been boiled in water. Fusion, filtration, and washing of the silica occupy about 30 minutes. 4 Of course the iron oxides resist attack by the acids better after calcination, but this is not necessarily the case with silicates calcined at, say, dull redness. H. Rocholl, Zeit. anal. Chem., 20. 289, 1881 ; G. W. Dean, Journ. Amer. Chem. Soc., 28. 882, 1906; H. E. Ashley, Chem. News, 90. 274, 1904. 5 R. W. E. Maclvor (Chem. News, 29. 246, 1874 ; B. Krieger, Chem. Ztg.,3$. 1054, 1911) states that the iron in haematites, etc., readily dissolves as ferrous iron when the finely powdered SPECIAL METHODS FOR IRON COMPOUNDS. 461 The iron, when dissolved, is reduced and determined by titration with per- manganate or dichromate. For a complete analysis, the sodium carbonate fusion (page 164) may be employed. A little sodium nitrite (1 to 1*5 grms.) l should be added in order to oxidise the sulphides to sulphates. Care must be taken to heat the mixture slowly, or loss by spurting, etc., may occur. The hygroscopic moisture, com- bined water, silica, 2 elements precipitated by hydrogen sulphide in acid solution, alumina, iron oxide, phosphoric and titanic oxides are determined as indicated on page 210. The sulphur can be determined in an aliquot portion of the filtrate from the silica ; chromium can be determined as indicated on page 469. The filtrate from the basic acetate separation can be used for the determination of manganese, zinc, nickel, cobalt, lime, and magnesia. The zinc is first pre- cipitated as sulphide by hydrogen sulphide in dilute (formic, hydrochloric, etc.) acid solution (page 364). The mixed sulphides of manganese, 3 nickel, and cobalt are then precipitated by hydrogen sulphide in ammoniacal solution as indicated on page 389. The magnesia and lime are determined in the filtrate. The residual mixed sulphides are taken up with aqua regia (page 389) ; the solution neutralised with sodium carbonate ; acidified with acetic acid and treated with hydrogen sulphide, whereby nickel and cobalt sulphides are precipitated, as described on page 389. 241. The Determination of Ferrous Oxide. Iron existing in the ferrous condition is rarely reported in clay analyses, and when the relative proportions of ferrous and ferric oxides are indicated the results are usually of little or no value. One reason is that no satisfactory general method is known for the determination of the ferrous iron in clays containing organic matter. A number of different methods have been sug- gested for decomposing silicates without interfering with the state of oxidation of the iron. For example : (1) Fusion with sodium carbonate. The objection to this process is the risk of absorbing air during the fusion, and the consequent oxidation of the manganese. material is digested in a long-necked flask with sulphuric acid and metallic zinc until all the iron is dissolved. The iron is in the ferrous condition ready for titration with potassium per- manganate. H. Borntrager (Zeit. anal. Chem., 35. 170, 1896 ; 38. 774, 1899) says that ignited ferric oxide dissolves at once in acid provided a little iron-free manganese dioxide be added. This is probably due to the formation of chlorine. A. Classen (Zeit. anal. Chem., 17. 182, 1878) says that ignited ferric oxide dissolves at once in concentrated hydrochloric acid if the powdered material be previously boiled to a flocculent condition with caustic potash solution. The alkaline liquid is decanted off and the acid added. T. M. Brown (Iron, 361, 1878 ; Dingier 's Journ., 228. 378, 1878 ; E. Donath and R. Jeller, Zeit. anal. Chem., 25. 361, 1886) ignites the powdered material intimately mixed with from half to its own volume of powdered zinc. The mixture is ignited in a porcelain crucible for 5-8 minutes. The mass is transferred from the crucible to a flask, and dissolved in dilute sulphuric acid (1 : 2). Portions adhering to the crucible can be removed by washing with dilute acid. The zinc must be free from iron. The ferric oxide is reduced to metal. Chromium oxide and alumina are not reduced to metals by the ignition with zinc powder. H. von Jiiptner (Oester. Zeit. Berg. Hutt., 42. 469, 1894) recommends magnesium powder in place of zinc. E. Hart (Chem. News, 34. 65, 1876 ; J. S. Maclaurin and W. Donovan, Journ. Soc. Chem. Ind., 28. 827, 1909) reduces the ore in a stream of hydrogen gas for 10-30 minutes (page 269). The sample is then usually easily decomposed by hot dilute acid. F. Michel, Chem. Ztg., 36. 345, 1912. 1 Potassium nitrate is generally recommended for this purpose. The nitrite neither froths nor attacks the crucible so much as the nitrate. 2 G. W. Dean, Journ. Amer. Chem. Soc., 28. 882, 1906 ; 29. 1208, 1907 ; T. G. Pimby, ib. t 30. 614, 1908 ; H. E. Ashley, Chem. News, 90. 274, 1904. 3 V. Maori, Monit. Scient. (4), 20. 18, 1906 ; A. Kaysser, Chem. Ztg., 35. 94, 1911 (bog iron ores) ; F. Michel, Chem. Ztg., 36. 345, 1912. 462 A TREATISE ON CHEMICAL ANALYSIS. The maiiganates, later on, oxidise the ferrous salts, leading to low results when the solution of the fused mass is titrated with permanganate. 1 (2) Fusion with borax Hermann's process. Here high results are obtained owing to the reduction of ferric salts during the fusion. 2 (3) Digestion in a sealed glass tube with sulphuric acid, or a mixture of hydrofluoric 3 and sulphuric acids, at 150-200. This method furnishes very fair results in the absence of sulphides. 4 (4) Digestion in a platinum crucible with a mixture of hydrofluoric and sulphuric acids, in an atmosphere of carbon dioxide. 5 The silicate, when decomposed by one of the methods (4) for preference indicated above, is titrated with a standard solution of potassium permanganate for the ferrous iron. 6 Dissolution of the Ferrous Iron. From 0'3 to 1 grm. of the sample, powdered as coarsely as will permit of its decomposition by the acids in a reasonable time, 7 is placed in a platinum crucible or dish approximately 40-50 c.c. and stirred up with 15 c.c. of dilute sulphuric acid (1 : 3). s The stirring must be thorough, since the powder may " cake " at the bottom of the dish and escape attack by the acids later on. Place the dish A on a support, fig. 152, in Treadwell's decomposition vessel, 9 which consists of a leaden box B supported in a paraffin 10 1 W. Early, Chem. News, 30. 169, 1874. 2 R. Hermann, Journ. prakt. Chem. (1), 15. 105, 1338; W. Suida, TschermaVa Mitt. (1), 5. 176, 1876; C. Bodewig, Pogg. Ann., 158. 222, 1876; C. Rammelsberg, Zeit. deut. Oeol. Ges., 24. 69, 1872 ; H. Rose, Handbuch der analytischen Chemie, Braunschweig, 2. 699, 1871. 3 Presumably the hydrofluoric acid is not exhausted by attacking the glass before it has had time to affect the substance under investigation. According to W. F. Hillebrand (Bull. U.S. Geol. Sur., 422. 158, 1910), the addition of hydrofluoric acid is "of doubtful utility." 4 A. Mitscherlich, Journ. prakt. Chem. (1), 8l. 116, 1860 ; (1), 83. 445, 1861 ; Zeit. anal. Chem., I. 56, 1862; C. Doelter, ib., 18. 50, 1879 ; Tschermak's Mitt. (1), 7. 281, 1877; (2), 3. 100, 1880 ; A. Remele, Notizblatt, 3. 160, 1867 : 4. 173, 1868. W. Michaelis (ib., 5. 204, 1869) says all the clay is not decomposed. For hydrochloric acid in sealed tubes A. H. Allen, Chem. Neivs, 22. 57, 1870 ; S. Pina de Rubies, Anales Fis. Quim., 10. 78, 1912. For operating with sealed tubes, consult page 493. 5 G. Werther, Journ. prakt. Chem. (1), 91. 321, 1864 ; J. P. Cooke, Amer. J. Science (2), 44. 347, 1867 ; J. H. Pratt, ib. (3), 48. 149, 1894 ; F. P. Treadwell, Kurzes Lehrbuch der analytischen Chemie, Leipzig, 2. 415, 1911 ; F. Mohr, Zeit. anal. Chem., 7. 450, 1867 ; A. R. Leeds, ib., 16. 323, 1877 ; C. E. Avery, Chem. News, 19. 270, 1869. C. A. Wilbur and W. Whittlesey (ib., 22. 2, 1870) use a mixture of calcium fluoride and hydrochloric acid ; and A. H. Chester and F. I. Cairns (Amer. J. Science (3), 34. 113, 1883) use a mixture of ammonium fluoride and sulphuric acid. 6 J. M. Eder (Monats. Chem., i. 137, 140, 1880) suggests the following gravimetric process for the determination of ferrous oxide in the presence of organic matter and ferric oxide. The solution must not be too strongly acid: Add an excess of neutral potassium oxalate and silver nitrate. In a few minutes add sufficient tartaric acid to prevent the precipitation of ferric hydroxide on addition of ammonia. Add an excess of ammonia. The addition of some ammonium chloride here favours the filtration and washing of the precipitate. The precipitated silver may be dissolved in nitric acid, precipitated, and weighed as silver chloride. From this, the ferrous oxide can be determined by computation from the equation : 2FeO + Ag 2 = Fe 2 3 + 2Ag. Hence, 1 grm. of silver represents 0'666 grm. FeO ; or 1 grm. of silver chloride represents 0-5017 grm. FeO. 7 Some silicates particularly those poor in silica and rich in magnesia, are liable to "cake" on the bottom of the crucible and escape attack by the acid. M. Dittrich (Ber., 44. 990, 1911) recommends that the powdered mineral be intimately mixed with coarsely ground quartz, which resists attack by the hydrofluoric acid long enough for the mixed acids to attack the mineral. The quartz also exposes a larger surface of the mineral to the attack. 8 If carbonates be present the mixture may effervesce. Hence, the acid must be added slowly to the dish ami covered with a watch-glass to prevent loss by spurting. The cover must be after- wards rinsed into the dish. There is little danger of oxidation of the ferrous iron at this stage. 9 F. P. Treadwell, Kurzes Lehrbuch der analytischen Chemie, Leipzig, 2. 416, 1911. 10 Instead of paraffin, L. W. Bosart (Journ. Amer. Chem. Soc. t 31. 724, 1909 ; Chem. Neius, 100. 238, 1909) recommends a mixture of 10 parts refined cotton seed oil and 1 part of bees- wax. It fumes but little below 250, and its flash-point in an open cup is above 300 ; hard SPECIAL METHODS FOR IRON COMPOUNDS. 463 bath C. The leaden cover has two holes D and E. A rapid stream of carbon dioxide gas 1 is passed through D for about 5 minutes. The lead cover is removed, and about 10 c.c. of concen- trated hydrofluoric acid 2 (40 per cent.) is added to the crucible. The dish is placed directly under the opening E. This permits the contents of the crucible to be stirred with a stout platinum wire. The bath is now heated to about 100, and maintained at that temperature for about an hour. The aperture T allows a thermometer to be placed in the paraffin bath C. Some of the excess of hydrofluoric acid may be driven off FIG. 152. -Treadwell's apparatus (section), by raising the temperature to 120 for another hour. It is important not to work too long at the higher temperature, for the reason indicated above. There should now be no grit representing undecomposed mineral in the crucible. The use of silicic acid later on renders a prolonged heating to drive off hydrofluoric acid unnecessary. Keep the stream of carbon dioxide flowing through the apparatus all the time. Let the system cool, while still maintaining the current of gas. An alternative and simpler method of decomposing the mineral not quite as effective is as follows : Fit a 50-c.c. platinum crucible tightly into* a hole in the centre of a piece of asbestos millboard so that the crucible passes about half way through, and the joint between the crucible and asbestos is nearly air-tight. Moisten half a gram of the coarsely powdered sample in the crucible with water. Put in a couple of coils of platinum wire to prevent bumping. Add a cold mixture of 10 c.c. of hydrofluoric acid and 15 c.c. of dilute sulphuric acid (1 : 3). Cover the crucible with a tight-fitting lid. Connect the top of a 15-cm. funnel 3 with a carbon dioxide generator. The funnel is placed over the crucible, and it should fit close to the asbestos millboard. When the air has been expelled by the carbon dioxide, place a Bunsen's burner below the crucible (fig. 153) so that the tip of the flame is about 3 inches below the bottom of the crucible, and the flame is raised oj lowered so that the contents of the crucible boil gently. The motion prevents the particles of silicate from caking on the bottom. In 5 or 10 minutes the mineral will be decomposed. The burner is removed. When the steam in the crucible has condensed, the crucible is lifted with a pair of tongs (fig. 66), 4 without removing the lid, and plunged beneath the surface of 400-500 c.c. of cold, recently boiled distilled water contained in an 800-c.c. beaker. The lid of the crucible is removed and the contents are ready for titration. 5 Titration for the Ferrous Oxide. Thoroughly stir about 10 grins, of pure paraffin under same conditions flashed at 215. Like paraffin, the mixture solidifies on cooling, but it melts more quickly than paraffin. Special oils, with a flash-point over 300, are now sold for charging oil baths H. Giesen, Duisburg-Wanheimerort. 1 Purified as indicated on page 169. 2 Hydrofluoric acid may contain reducing agents : arsenious acid, hydrogen sulphide, etc. C. Jehn, Zeit. anal Chem., 13. 176, 1874 ; Archiv Pharm. (3), I. 481, 1873. If the acid be made in leaden vessels it frequently contains sulphurous acid. If present, it must be destroyed by the addition of potassium permanganate until the colour just ceases to be discharged W. Hampe, Chem. Ztg., 15. 1777, 1891. See pages 169 and 226. 3 Coated inside with paraffin wax. 4 A loop of platinum wire may be fixed round the crucible before it is placed on the asbestos millboard, for convenience in lifting and transferring the hot crucible to the beaker of water. 5 A reddish-brown sediment which dissolves when stirred up with water is not to be mistaken for undecomposed mineral. 464 A TREATISE ON CHEMICAL ANALYSIS. silicic acid l with about 100 c.c. of water in a 600-c.c. beaker, and add 20-25 grms. of potassium sulphate. Wash the hot contents of the platinum crucible into the beaker. Titrate as rapidly as possible with standard permanganate (page 198). The first permanent pink blush is the end point. This fades in a short time. It is well to confirm the method by a blank test, or with ferrous sulphate solution which has been standardised with hydrofluoric acid. One grm. of KMn0 4 r FIG. 153. Opening silicates for the determination of ferrous iron. represents 2*27412 grms. FeO. The amount of ferric oxide corresponding with the ferrous oxide so determined is calculated by multiplying the amount of ferrous oxide by I'll 1343. Subtract the result from the total amount of ferric oxide determined in the sodium carbonate fusion as indicated on page 182. EXAMPLE. The sodium carbonate fusion, etc. (page 182), furnished 3'65 per cent, of Fe 2 O 3 . In the above determination, 1 grm. of the sample required 6*2 c.c. of per- manganate containing 0'00098 grrn. KMn0 4 per c.c. That is, 0'006076 grm. KMn0 4 , i.e. 0-006076 x 2-274 grms. FeO. Hence, the substance has 0'0138 per cent, of FeO. But 0'0138jrm. FeO corresponds with 0;0138x 1-1113 = 0'0152 grm. Fe 2 O 3 , that is, T52 per cent. -Fe 2 3 . It is probable that the results for ferrous oxide are usually a little too low. 242. Disturbing Factors in the Determination of Ferrous Oxide. This subject has been carefully studied by Hillebrand and Stokes, and after a comparison of the different methods in use for the determination of ferrous oxide in silicates insoluble in the mineral acids, it has been stated that, " despite the utmost care in practical manipulation, the exact determination of ferrous iron in rocks is fraught with extraordinary difficulties and uncertainties. Only in the absence of decomposable sulphides and carbonaceous matter, and when Fe 2 3 . Hence, the sample has 3 '65 less 1 '52 = 2 '13 per cent, of ferric oxide Or enough to neutralise all the hydrofluoric acid which is present. SPECIAL METHODS FOR IRON COMPOUNDS. 465 the amount and condition of the vanadium are known and relatively coarse powders can be used, is it permissible to regard the result as fairly above suspicion." 1. The Influence of Fine Grinding. The unavoidable oxidation of ferrous compounds during grinding was discussed on page 124. "Nearly all mineral analyses which have been made in the past are affected by more or less serious errors due to the oxidation of the iron during grinding. The error is greater, the more finely the sample is powdered " (Hillebrand). 2. The Influence of Sulphides. When sulphides decomposable by the acids are present, some hydrogen sulphide, and possibly also some sulphur dioxide, are formed. 1 These gases will reduce some of the ferric oxide to ferrous oxide, and consequently the decomposed silicate will contain more ferrous iron than the undecomposed silicate. This is more particularly noticed when the decomposi- tion of the silicate is effected in sealed tubes. The difference is not very marked when but small quantities 2 per cent. of ferrous iron are present, but the error may amount to nearly 20 per cent, when about 10 per cent, of ferrous iron is present. O'Ol per cent, of sulphur may cause the ferrous iron to appear (H35 per cent, too high. The oxidation of the sulphides, and the consequent reduction of the ferric oxide, are greatly accelerated by the presence of ferric salts. The acid mixture will scarcely attack pyrite, but in the presence of ferric salts the action is relatively quick. The iron of the sulphide say, pyrrhotite may also appear in the final titration as ferrous iron. 2 3. The Influence of Organic Matter. Carbonaceous matters will reduce the sulphuric acid in the sealed tube ; and, in general, organic matter will reduce the ferric oxide and render the determination of the ferrous iron in a mixture nugatory. 3 4. The Influence of Vanadium. If appreciable quantities of vanadium be present as, say, V 2 3 , 4 a correction must be made for the permanganate used in converting V 2 3 ->V 2 5 during the titration (page 200). Every gram of V 2 3 found in the sample will be equivalent to 1*5757 grms. of FeO in the consump- tion of permanganate. If the vanadium occurs in the form of V 2 5 , no correction will be required. EXAMPLE. If a silicate contains O'll per cent, of V 2 3 , and. the FeO, calculated from the permanganate titration, is 2'11 per cent., it follows that O'll x 1 -5767 = 0-17 per cent, is not FeO. The FeO corrected for the vanadium will then be 2'11 less 0'17 = 1'94 per cent. FeO. 5. Influence of Hydrofluoric Acid on the Permanganate Titration. A solution of ferrous sulphate in fairly concentrated sulphuric acid oxidises somewhat slowly; but if a little hydrofluoric acid be added to the solution, the rate of 1 E. A. Wulfing, Ber., 32. 2217, 1899; J. H. L. Vogt, Zeit. prakt. Geol., 7. 250, 1899; L. L. de Koninck, Ann. Soc. Geol. Belgique, 10. 101, 1883 ; Zeit. anorg. Chem., 26. 123, 1901 ; W. F. Hillebrand and H. N. Stokes, Journ. Amer. Chem. Soc., 22. 625, 1900 ; H. N. Stokes, Atner. J. Science (4), 12. 414, 1901. T. Scheerer (Pogg. Ann., 124. 98, 1850) compared the results obtained by the different methods. See also A. Leonhard, Ueber die Bestimmung des Eisenoxyduls in Gesteinen, Heidelberg, 1912. 2 If the sulphides have been determined, this can sometimes be corrected accordingly. 3 J. T. Hewitt and G. R. Mann (Analyst, 37. 179, 1912) titrate solutions of ferric iron in presence of organic matter with standard thiosulphate solution, using ammonium thiocyanate as indicator and a small quantity of copper sulphate as catalyst. The titration is continued until the red colour of the ferric thiocyanate disappears. The end point is not sharp, and better results are obtained by adding a slight excess of the standard thiosulphate, and titrating back the excess with an iodine solution, using starch as indicator. 4 C. Czudnowicz, Pogg. Ann., 120. 20, 1863 ; 0. Lindemann, Zeit. anal. Chem., 18. 99, 1879 ; W. F. Hillebrand and F. L. Ransom e, Amer. J. Science (4), 10. 120, 1900. 466 A TREATISE ON CHEMICAL ANALYSIS. oxidation is considerably increased. Hence it is very important to expose the solutions containing ferrous iron as little as possible to the air before titration. This discussion all shows that the presence of the hydrofluoric acid interferes with the accuracy of the permanganate titration. In the presence of hydro- fluoric acid, the permanganate is decolorised very quickly, so that the permanent pink blush normally obtained with permanganate titrations is very fugitive in the presence of hydrofluoric acid. It is therefore difficult to get a distinct end point. At the best, the pink blush lasts but a few seconds, and this is even more fugitive with increasing amounts of ferrous oxide, and of hydrofluoric acid. If the solution be boiled some time with the idea of driving off part of the hydro- fluoric acid, part of the ferrous iron may -be transformed to ferric iron by the oxidising action of the sulphuric acid as it becomes more and more concentrated. Gage l states that hydrofluoric acid does not interfere with the titration in the presence of calcium phosphate, but, according to Leonhard, the action is then so slow that the end-point is uncertain. Dittrich and Leonhard recommend the addi- tion of 1 to 2 grms. of potassium or sodium sulphate (not ammonium sulphate) to the mixture to facilitate the recognition of the end point of the permanganate titration. 2 That being the case, it is advisable to spend less time in driving off the excess of hydrofluoric acid, and avoid getting low results arising from the oxidation of ferrous iron during the concentration of the solution. Fromme also avoids the danger of reoxidation during the removal of hydrofluoric acid by neutralising the hydrofluoric acid with pure silica. The resulting hydrofluosilicic acid enables the titration to be conducted more accurately than in the presence of hydrofluoric acid. 6. Influence of Manganous Salts. The manganous salt formed by the reaction between the permanganate and the ferrous iron is itself readily oxidised by permanganate in the presence of hydrofluoric acid, so that a sharp end reaction is difficult to obtain if much hydrofluoric acid be present, or if much manganous sulphate be present. The latter will occur when much ferrous iron is under investigation, so that the pink tint of the permanganate persists longer when but little ferrous iron is in question than with larger amounts. The more hydrofluoric acid used, the greater the apparent amount of iron found. 3 7. Influence of Titanium. According to Dittrich and Leonhard, titanium sesquioxide, Ti 2 3 , in the presence of ferric salts produces disturbing effects on the permanganate titration, and titanium is nearly always present in the analysis of silicate rocks. These effects can be obviated by pouring the contents of the crucible containing the hydrofluoric acid, etc., into a 600-c.c. beaker containing 100 c.c. of water mixed with about 10 grms. of precipitated silica, and 20-25 grms. of potassium sulphate. The mixture is quickly titrated with 1 H. Fromme, Tschermak's Mitt., 28. 329, 1909 ; R. B. Gage, Journ. Amer. Chem. Soc.. 31. 381, 1909 ; M. Dittrich and A. Leonhard, Ber. Vers. Oberrhein. geol. Ver., 2. 92, 1910 ; Zeit. anorg. Chem. ,74. 21, 1912 ; 0. Follenius, Zeit. anal. Chem., u. 177, 1872 ; E. Rupp, Ber.,$>. 164, 1905 ; A. Leonhard, Ucber die Bestimmung des Eisenoxyduls in Gesteinen, Heidelberg, 1912. 2 E. Deussen (Zeit. anorg. Chem., 44. 425, 1905 ; Monats. Chem., 28. 163, 1907) recommends manganese sulphate ; but this is apparently a mistake W. F. Hillebrand, Bull. U.S. Geol. Sur., 422. 162, 1910. 3 F. J. Metzgerand L. E. Marrs, Journ. Ind. Eng. Chem., 3. 333, 1911. CHAPTER XXXIV. THE DETERMINATION OF CHROMIUM, VANADIUM, AND URANIUM. 243. The Errors due to the Presence of Chromium, Vanadium, and Uranium. VANADIUM is a fairly common constituent of fireclays. 1 If chromium, vanadium, and uranium be present in a silicate, they will be precipitated with the aluminium, iron, and titanium. 2 Consequently, these elements will be found in the products of the pyrosulphate fusion, and, if ignored, they may introduce two errors in the analysis : 1. The iron determination will be too high. Vanadic acid V 2 5 is reduced by hydrogen sulphide to V 2 4 ; by magnesium in sulphuric acid solution, and by ammonium bisulphite, to V 2 3 ; and by metallic zinc, more or less completely to V 2 2 . 3 The colour of the solution undergoing reduction passes from yellowish (V 2 5 ) to blue (V 2 4 ), to green (V 2 8 ), to lavender (V 2 2 ). The reduced oxides V 2 4 , V 2 3 , and V 2 2 are converted to V 2 5 by titration with permanganate. Consequently, the action of, say, one molecule of V 2 3 on the potassium per- manganate will be equivalent to the effect of two molecules of reduced ferric oxide ; and the action of a molecule of V 2 4 , the same as one molecule of reduced ferric oxide. Hence, with the ammonium bisulphite reduction, every gram of vanadium pentoxide, V 2 5 , corresponds with 0'875 grm. of ferric oxide Fe 2 3 and with the magnesium reduction, to 1751 grms. of ferric oxide. 4 EXAMPLE. A fireclay containing 0'09 per cent, of vanadium pentoxide showed 2'15 per cent, of ferric oxide by the ammonium bisulphite reduction and permanganate titration. Hence, the effect of the reduced vanadic oxide was the same as 0-09 x 0-8755 = 0'08 grm. Fe 2 O 3 . Hence, the clay had 215 less 0'08 = 2'07 per cent, of ferric oxide. 1 A. Terrell, Compt. Rend., 51. 94, 1860 ; see Vol. II. 2 Note, if certain members of the hydrogen sulphide group be removed before the ammonia precipitation is made, some uranium may be precipitated in that group. The addition of an excess of hydrogen peroxide before ammonia will prevent vanadium precipitating with aluminium, etc. W. VV. Clark, Met. Chem. Eng., u. 91, 1913. 3 The statement by B. Glasmann (Ber., 38. 600, 604, 1905) that metallic zinc reduces V 2 5 to V 2 0. 2 , while metallic magnesium reduces V 2 5 to V 2 3 , while both agents reduce molybdic acid to molybdenum sesquioxide, was also" proposed as a volumetric process, with the permanganate titration, for the simultaneous determination of vanadium and molybdenum. F. A. Gooch and G. Edgar (Amer. J. Science (4), 25. 233, 1908 ; Chem. News, 98. 2, 1908), however, have shown that the action of magnesium is somewhat irregular. For a study of the conditions under which the sulphur dioxide reduction will reduce vanadium pentoxide to the tetroxide, and not attack molybdenum trioxide, see G. Edgar, Amer. J. Science (4), 25. 332 1908 ; Chem. News, 97. 245, 1908. A solution containing vanadium dioxide cannot be exposec he air momentarily without undergoing oxidation. 1 Gooch and Newton's plan (page 189) for "the oxidation of TLj0 3 without the oxidation aes not work with V 2 4 , although it converts V 2 3 to V 2 4 if the solution be iecthas not been urooerlv investigated. " W. F. Hillebrand, Bull. U.S. Geol. to the air momentarily without undergoing oxidation. 1 Gooch and N of ferric oxide does warm. This subject has not been properly investigated Sur., 422. 109, 1910. 468 A TREATISE ON CHEMICAL ANALYSIS. Chromic acid and the chromates are reduced by hydrogen sulphide, metallic zinc, magnesium, and aluminium in acid solution, sulphites, and stannous chloride. Uranium salts are also reduced by similar agents. The reduced salts are reoxidised by potassium permanganate. Hence, both chromium and uranium salts, in the pyrosulphate fusion, will introduce errors in the determination of iron by the titration process. If, therefore, appreciable quantities of these elements are present, other methods of separation must be employed, and an allowance made. 2. The titanium determination will be too high. Vanadium compounds give a reddish-brown "brick-red" coloration with hydrogen peroxide. This is a more intense coloration than the yellowish-orange tint due to titanium, 1 and, in consequence, vanadium may intensify the effect of titanium in Weller's colori- metric process, and thus lead to high results. With practice it is possible to tell if much vanadium is present from the tint obtained with hydrogen peroxide in Weller's process for titanium. 2 244. The Detection of Chromium and Vanadium. Fuse, say, 5 grms. of the finely powdered silicate with 20 grms. of sodium carbonate 3 and 3-4 grms. of sodium nitrite NaN0 2 . Extract the cold cake with hot water, and add a little alcohol to reduce the sodium manganate. Filter the solution. The filtrate contains some alumina, silica, arsenic, phos- phorus, fluorine, chlorine, sulphur, molybdenum, chromium, vanadium, tungsten, etc. ; the residue may contain titanium, iron, uranium, 4 barium, zirconium, rare earths, etc. Nearly neutralise the filtrate with nitric acid, but keep the sodium carbonate in slight excess. Evaporate the solution to dryness, take up the mass with water, and filter. Add mercurous nitrate to the alkaline solution, when mercurous arsenate chromate, molybdate, and tungstate may be precipitated. Boil the solution and filter. Dry and ignite the precipitate in a platinum crucible, fuse with sodium carbonate, and extract the fused mass with water. Chromium'. A yellow solution probably indicates chromium. 5 The tint of the solution deepens on the addition of, say, sulphuric acid. Pour a little of the 1 An experienced eye can frequently recognise the presence of vanadium from the tint of the solution prepared for the colorimetric determination of titanium. 2 According to H. J. H. Fenton (Journ. Chem. Soc., 93. 1064, 1908), a cold aqueous solution of dihydroxymaleic acid gives a straw-yellow coloration in the presence of 1 part of titanium (as titanic salt, say TiCl 4 ) in 1,000,000 parts of solution ; a lemon-yellow with 1 : 150,000 ; and an intense orange-red with 1:15,000. The coloration is approximately "fifteen to twenty times as delicate as the hydrogen peroxide test," and " it is not given by vanadium." Molyb- dates and uranyl salts give red or brown colours which are destroyed by acids the former by heating, the latter in the cold. Tungstic acid gives a brown coloration which immediately becomes blue. Negative results are obtained with salts of thorium, cerium, zirconium, silica, and tin. Ferric salts interfere by destroying the reagent. Ferrous salts do not interfere. Hence the reaction can be used for detecting titanium in presence of vanadium. 3 H. L. Robinson (Chem. News, 70. 199, 1894) reports the presence of vanadium in commercial sodium hydroxide. 4 Treat the residue with nitric acid, and test by boiling the solution with an excess of sodium carbonate. It is perhaps advisable to indicate here a few of the important reactions of uranium salts. Potassium hydroxide with uranic salts precipitates a yellow potassium uranate, U0 2 .0(OK) 2 ; ammonium hydroxide, a yellow ammonium uranate, U0 2 .0(ONH 4 ) 2 ; sodium carbonate, a yellow uranyl sodium carbonate, U0 2 Na 4 (C0 3 ) 3 . Ammonium sulphide precipitates a brown uranyl sulphide, U0 2 S, which is soluble in ammonium carbonate and in dilute acids. Uranyl salts accordingly give no precipitate with ammonium sulphide in the presence of ammonium carbonate. Potassium ferrocyanide gives a brown precipitate of uranyl ferrocyanide, either J0 2 . K 2 FeCy 6 or (U0 2 ) 2 FeCy 6 . Sodium phosphate gives a yellowish-white uranyl phosphate, U0 2 HP0 4 , which passes into ammonium uranyl phosphate in the presence of ammonium salts. 5 Note, zinc and chromates may form insoluble zinc chromate in alkaline solutions G. Chancel, Compt. Rend., 43. 927, 1856. DETERMINATION OF CHROMIUM, VANADIUM, AND URANIUM. 469 solution into a test tube, add an excess of hydrogen peroxide, and shake up the mixture with 3 c.c. of ether. A blue coloration in the ethereal solution floating on the aqueous solution represents chromium. 1 See page 484. Vanadium. Acidify a portion of the solution with sulphuric acid, and precipitate molybdenum, if present, by hydrogen sulphide in a pressure flask (page 277). If tin be present, oxalic acid will keep it in solution. Molybdenum, arsenic, traces of platinum, etc., will be precipitated, if present. The filtrate from the precipitated sulphides is boiled while a current of carbon dioxide is passing through the solution in order to drive off the hydrogen sulphide. Evaporate the solution to dryness, and remove the excess of sulphuric acid in an air bath. Dissolve the residue in water, and add a few drops of hydrogen peroxide. A characteristic "brick-red" coloration develops if vanadium be present. 2 245. The Separation of Chromium, Iron, Titanium, Aluminium Vanadium, and Uranium from Manganese, Cobalt, Nickel, and Zinc. Chromium in the Basic Acetate Separation. In applying the basic acetate process, it may be remembered that both chromium acetate and chromium hydroxide are easily soluble in acetic acid. It might therefore be supposed that some of the chromium will be dissolved when the basic precipitation is made in the presence of free acetic acid. Some claim that practically all the chromium can be precipitated with the iron and aluminium, even in the presence of some free acetic acid, because the chromium acetate seems to assume a kind of "passive state" towards acetic acid, particularly when the chromium is in very large excess ; but it is necessary to make sure the chromium has all been precipitated by adding hydrogen peroxide to the filtrate. A blue coloration shows that chromium is present in the filtrate. 3 Some recommend trans- forming the chromium salts into chromates before applying the basic acetate process. Most of the chromium then passes into the filtrate along with some of the aluminium as aluminium chromate, but a part remains with the precipitate in the form of basic chromates of iron and aluminium. 4 The error due to the formation of basic ferric chromates is least when the precipitation of the iron from the chromate solution is made with sodium hydroxide in sufficient excess to decompose the basic ferric chromate. If but 1 A. Terrell (Bull. Soc. Chim. (2), 3. 30, 1865 ; Chem. News, u. 136, 1865) precipitates the chromium as lead chromate. This is all right if molybdenum, etc., are absent. W. Gibbs (Amer. J. Science (2), 39. 59, 1865 ; (3), 5. 110, 1873 ; Chem. News, 28. 63, 1873) separates uranium from chromium quantitatively or qualitatively by precipitating the uranium with a slight excess of sodium hydroxide. The uranium will be found in the precipitate, the chromium in the nitrate. If a little chromium be present in the precipitate, it can be removed by dissolving the precipitate in dilute nitric acid, boiling to expel nitrous fumes, etc., as described below, and finally precipitating the chromium as mercurous or barium chromate. 2 C. M. Johnson (Rapid Methods for the Chemical Analysis of Special Steels, New York, 5, 1909) states that ferrous ammonium sulphate will discharge the brick-red coloration of vanadium more quickly than the yellowish-orange tint of titanium compounds. Hence, if ferrous ammonium sulphate be added to a solution containing both vanadium and titanium in the presence of hydrogen peroxide, the brick-red colour will fade to a bright yellow tint before the colour is discharged ; if titanium be absent, the brick-red coloration will fade to a colourless solution without showing the yellow tint. Fenton's reaction page 468 gives better results in detecting titanium in the presence of vanadium. * B. Reinitzer, Monats. Chem., 3. 249, 1882 ; Chem. News, 48. 114, 1883 ; F. Mayer, Ber., 22. 2627, 1889. One writer says : "The method of separating ferric and aluminium salts in the form of basic acetates entirely loses its applicability in the presence of chromium salts. " 4 H. Brearley, Chem. News, 76. 175, 1897 ; 77- 49, 179, 1898. 470 A TREATISE ON CHEMICAL ANALYSIS. a small amount of iron be present, the excess need not be very large. 1 Sodium car- bonate is quite as good provided also a sufficient excess be used. Thus Brearley 2 found that in a litre of solution containing the equivalent of 1 grm. of iron (and O5 grm. of aluminium), and 0*375 grm. of potassium chromate, the addition of a varying excess of alkali led to the results shown on the left of Table LX. ; and with solutions containing half a gram per litre of aluminium in place of iron, a varying excess of alkali furnished the results shown on the right of the same table. Table LX. Effect of an Excess of Alkali on the Separation of Iron, Aluminium, and Chromium. Percentage separation. Percentage separation. Excess of Excess of alkali alkali 2N-sol. Sodium carbonate. Sodium hydroxide. Ammonia. 2N-sol. Sodium bi- carbonate. Ammonium carbonate. Ammonia. 54-6 69-3 24-8 10 94-0 81-3 89-0 10 90-4 100-0 '/0'4 20 95-0 93-6 90-5 20 97-5 99-8 77-0 30 99-0 94-4 93-1 30 987 100-0 83-6 50 1001 ... 87-8 The results show that sodium carbonates or sodium hydroxide give good separa- tions if a sufficient excess be used. But when chromium is to be separated from aluminium, a large excess of alkali inhibits the precipitation of aluminium hydroxide, and favours the precipitation of ferric hydroxide. Consequently, as indicated on page 362, avoid the basic acetate process for the separation of chromium from manganese, cobalt, nickel, zinc, and ferrous salts. In many cases, the chromium will be precipitated with the aluminium, iron, and titanium both in the "basic acetate" and in the "ammonia" pro- cesses. 3 It then remains to separate the chromium from the iron, aluminium, and titanium by some other process, say Knorre's method, or the mercurous nitrate process. The Barium Carbonate Method. This method originated with Fuchs and Rose, 4 and it is more satisfactory than the basic acetate process when chromium is present. The barium carbonate process depends upon the fact that ferric, aluminium, chromium, vanadium, titanium, and uranium hydroxides are precipitated from cold dilute solutions in a few hours by barium carbonate, while manganese, nickel, cobalt, zinc, and ferrous salts are not precipitated. The barium is subsequently removed from both the precipitate and nitrate. Ferric salts, etc., are readily hydrolysed (pages 181 and 362) in aqueous solutions : FeCl 3 + 3H 2 0==Fe(OH) 3 + 3HC1. 1 Similar remarks apply to the separation of molybdenum by this process F. Ibbotson and H. Brearley, Chem. News, 8l. 269, 1900 ; 79. 3, 1899. 2 H. Brearley, Chem. News, 76. 175, 1897 ; tt. 49, 131, 179, 216, 1898 ; W. Galbraith, #., 77. 187, 1898. 3 It must be borne in mind that aluminium hydroxide and chromium hydroxide are slightly soluble in ammonia, so that the nitrate from the ammonia precipitate must be boiled to recover the aluminium and chromium, as indicated on page 479. 4 J. N. von Fuchs (Schweigger' s Journ. , 62. 184, 1831) also used calcium carbonate. H. Rose, Pogg. Ann., 83. 137, 1851 ; J. F. W. Herschel, Ann. Chim. Phys. (3), 49. 306, 1837. DETERMINATION OF CHROMIUM, VANADIUM, AND URANIUM. 471 In the presence of barium carbonate, the free acid will be removed as fast as it is formed : BaC0 3 + 2HC1 = BaCl 2 + H 2 + C0 2 ; equilibrium is disturbed ; and the hydrolysis is completed. The net result of the action is represented by the equation : 2FeCl 3 + 6H 2 + 3BaC0 3 = 3BaCl 2 + 2Fe(OH) 3 + 3H 2 + 3C0 2 . If the solution be heated, the salts of zinc, nickel, etc., may also be hydrolysed, and,, in consequence, precipitated by the barium carbonate treatment. If much nickel and cobalt be present, small amounts of these elements will be precipitated by the barium carbonate even in the cold. The presence of ammonium chloride, 1 however about 5 grms. per 100 c.c. of solution prevents the precipitation of nickel and cobalt, and the results are then quite satisfactory. The method has also been recommended for separating vanadium and chromium from ferrous iron 2 and manganese. With these elements, the separation is effected after a few minutes' boiling with barium carbonate. 3 Zinc hydroxide 4 and cadmium carbonate 5 have also been proposed in place of barium carbonate. The cadmium can be removed by subsequent treatment with hydrogen sulphide. The modus operandi is as follows : The slightly acid solution containing nitrates or chlorides of the elements indicated above, but free from sulphates, 6 is treated in an Erlenmeyer's flask with a solution of sodium carbonate, as in the basic acetate process, until a slight permanent turbidity is produced. This is cleared by a drop or two of dilute hydrochloric acid. The dilute solution is thoroughly agitated with an emulsion of barium carbonate (that is, barium carbonate suspended in water). 7 The flask is closed and allowed to stand several, hours, with occasional shaking. Decant the clear solution through a filter paper, and wash the residue three times with cold water ; wash the precipitate on to the filter paper by means of cold water. (a) The precipitate is dissolved in hydrochloric acid ; heated to boiling to remove the carbon dioxide; and the iron, vanadium, chromium, aluminium, titanium, and uranium are either precipitated by ammonium sulphide, or the barium is removed by means of sulphuric acid or sodium sulphate. The latter process is not so good as the former, because the barium sulphate is liable to carry down some of the metals from the solution. (b) The filtrate from the barium carbonate is heated to boiling and treated with sulphuric acid or sodium sulphate to precipitate the barium. The precipitate is washed. The filtrate contains manganese, cobalt, nickel, zinc, and ferrous salts, 8 if present. 1 P. Schwarzenberg, LieUg's Ann., 97. 216, 1856. 2 It is almost impossible to prevent the precipitation of a little iron from " ferrous solutions," but the error from this effect is so small that C. R. Fresenius (Quantitative Chemical Analysis, London, i. 434, 1876) recommended the process in special cases for the separation of ferrous iron from ferric iron, etc. 3 A. Steffan, Ueber die Bestimmung von kleinen Mengen an Chrom und Vanadin in Gestein und Stahlarten, Zurich, 1902. 4 P. Slawik, Chem. Ztg., 34. 648, 1910. 5 J. R. Cain, Bull. Bur. Stw , _ Standards, 7. 377, 1911. 6 The solution must be quite free from sulphates because the other elements may be pre- cipitated, e.g. : ZnS0 4 + BaC0 3 ^--ZnC0 3 + BaS0 4 . 7 BARIUM CARBONATE. the barium carbonate must be. tested by dissolving a portion of it in hydrochloric acid and precipitating the barium by the addition of sulphuric acid ; the nitrate must leave no residue when evaporated to dryness in a platinum dish, showing that the barium carbonate is free from alkaline carbonates, etc. 8 As a rule, the ferrous iron is oxidised to the ferric state, so that no iron remains in the filtrate. 472 A TREATISE ON CHEMICAL ANALYSIS. The main objection to the cold process particularly with chromium is the time required for the precipitation, and the subsequent removal of barium. The precipitation can, of course, be allowed to proceed overnight when the results are not urgently wanted. As indicated above, in the special case where chromium and vanadium have to be separated from manganese and ferrous salts, the solution can be heated to boiling, and small additions of barium carbonate made every two or three minutes until an excess of about 1 to 2 grms. 1 has been added. After 10 to 15 minutes' boiling, let the mixture stand a few minutes to allow the precipitate to settle, filter and wash the residue with hot water. Place the filter paper and precipitate in a large platinum crucible ; burn off the filter paper ; and fuse the residue with about 2 grms. of sodium carbonate and \ grm. of sodium nitrite. The fused mass is extracted with water ; ferric oxide remains undissolved, while sodium aluminate, chromate, and vanadate pass into solution. The chromate is treated with acetic acid and lead acetate, or with dilute nitric acid and lead nitrate, for the precipitation of lead chromate and lead vanadate, as indicated below. 246. The Separation of Chromium and Vanadium in the Analysis of Silicates. Decomposition of the Silicate? Fuse 5 grms. of the powdered sample with about 20 grms. of sodium carbonate and 3 grms. of sodium nitrite. Extract with water ; reduce the sodium manganate with a few drops of alcohol ; 3 nearly neutralise the solution with nitric acid about 150 c.c. of acid (sp. gr. 1'5) will be needed. The solution must not be acid or nitrous acid will be liberated, and this, in turn, will reduce some of the chromium and some of the vanadium. 4 Evaporate the solution to dryness ; take up the mass with water ; and filter. Ignite the precipitate; treat the ignited mass with sulphuric and hydrofluoric acids ; evaporate to dryness ; fuse with sodium carbonate ; take up the mass with water ; neutralise with nitric acid ; boil ; filter ; and add the filtrate to the main solution. The object of this treatment is to recover chromium which is retained by the precipitated silica. Precipitation of Chromium and Vanadium. Add an almost neutral solution of mercurous nitrate 5 to the cold, barely alkaline solution until no further 1 Too much barium carbonate makes the subsequent extraction of the chromium difficult. 2 W. F. Hillebrand, Amer. J. Science (3), 6. 210, 1893 ; Chem. Neivs, 78. 216, 1893. The higher oxidation products chromates are not precipitated by ammonia. Hence, the silicate can be decomposed with the sodium peroxide fusion in a nickel crucible and the cake taken up with hydrochloric acid. The silica is removed as usual. The alumina is precipitated by boiling the solution made alkaline with ammonia and hydrogen peroxide. The precipitate is washed with a solution of ammonium nitrate, redissolved, and reprecipitated with ammonia and hydrogen peroxide as before. The alumina is then practically free from vanadium. The vanadium and chromium remain in the nitrate W. Trautmann, Stalil Eisen, 30. 1802, 1910. 3 E. Classen (Amer. Chem. Journ., 8. 437, 1896) shows that the undecomposed residue with some minerals requires re-fusion, since, if considerable vanadium be present and the vanadium is to be determined, the insoluble residue will always contain vanadium. 4 If an excess of acid be added inadvertently, add a slight excess of sodium carbonate. The amount of acid required to neutralise 20 grms. of sodium carbonate can be determined by a blank test. 5 Page 408. For the precipitation of vanadium as mercury vanadate, see C. R. von Hauer, Journ. prakt. Chem., 69. 385, 1856 ; C. Rammelsberg, Ber., I. 158, 1868 ; E. Classen. Amer. Chem. Journ., 7. 349, 1885 ; C. Radau, Liebig's Ann., 251. 114, 1889 ; R. Holverscheit Ueberdie quantitative Bestimmung des Vanadins und die Trennung der Vanadinsdur e von Phosphor sdure, Berlin, 10, 1890. It was formerly the custom to add a slight excess of mercurous nitrate and then mercuric oxide to neutralise the excess of nitric acid which is present in the solution of mercurous nitrate; but W. F. Hillebrand (I.e.} has shown that the mercuric oxide is not necessary, since the basic mercurous carbonate is sufficient to remove the small quantity of free nitric acid introduced along with the mercurous nitrate. Chlorides should be absent, or vanadium may be lost during the subsequent ignition of the precipitate. E. B. Auerbach and K. Lange, Zeit. angew. Chem., 25. 2522, 1912. DETERMINATION OF CHROMIUM, VANADIUM, AND URANIUM. 473 precipitation takes place. A bulky precipitate 1 containing mercurous carbonate, chromate, and vanadate is obtained. 2 Boil ; filter ; wash with water containing a little mercurous nitrate or ammonium nitrate in solution ; dry the precipitate ; and transfer as much as possible to a platinum crucible. Burn the filter paper separately, and add the ash to the main precipitate, which is ignited to remove the mercury. Fuse the residue with sodium carbonate, leach with water, and filter the yellow solution of sodium chromate into a small flask 25-50 c.c. If chromium alone be present, and the amount small, it is determined colori- metrically ; if the amount be relatively large, it is determined volumetrically or gravimetrically. If both chromium and vanadium be present, they can be determined as described below, 247 and 252 ; or 253. Knorre's Process 3 for separating Chromium. Dissolve the " ammonia " precipi- tate in dilute sulphuric acid, and add an excess of ammonium persulphate, with sufficient sulphuric acid to prevent the precipitation of basic ferric sulphate. On boiling, the dilute solution is converted into chromic acid. The iron and aluminium are then precipitated by ammonia in the usual manner. The precipitate is dis- solved in dilute sulphuric acid, and the operation is repeated so as to eliminate the trace of chromium precipitated with the aluminium and iron. 4 The chromium is determined in the joint filtrate in the usual manner by gravimetric or volu- metric processes 249, 250, and 251, pages 476 to 479 ; and the precipitate, containing the aluminium, titanium, and iron, is treated as indicated on page 210. 247. The Colorimetric Determination of Chromium. A solution of an alkaline chromate is yellow. The intensity of the colour is proportional to the amount of chromate in solution. 5 If a solution (standard solution) containing a known amount of chromate has the same tint as another solution of equal depth of liquid (test solution), it is assumed that both solutions have the same amount of the alkaline chromate in solution. Standard Solution of Potassium Chromate. This is prepared by dissolving 0'5105 grni. of potassium chromate in a litre of water made alkaline with sodium carbonate. One cubic centimetre of this solution represents 0*002 grm. of Cr 2 3 . Pipette, say, 10 c.c. into the right test glass of the colorimeter. Test Solution. This is prepared by concentrating the solution under investigation and making it up to a definite volume 25, 50, or 100 c.c. such that its colour is weaker than the standard solution. Pour this solution into the left test glass of the colorimeter. Comparison. Dilute the standard solution with water from a burette until its tint is the same as the test solution. If the tint of the test solution be faint, 1 If the precipitate is inconveniently large, cautiously add a little nitric acid and then a drop of mercurous nitrate to make sure that precipitation is complete. 2 Tungsten, molybdenum, phosphorus, and arsenic, if present, will be precipitated. 3 G. von Knorre, Zeit. angew. Chem., 16. 1097, 1903. 4 R. B. Riggs (Amer. J. Science (3), 48. 409, 1894; Chem. News, 70. 311, 1894 ; W. J. Sell, ib., 54. 299, 1886 ; Journ. Chem. Soc., 35. 292, 1879) digests the precipitate in 100 c.c. of water, 10 c.c. of hydrogen peroxide, and 1 grm. of sodium or potassium hydroxide. When effervescence has ceased, separate the ferric hydroxide by filtration. The filtrate contains aluminium, and chromium (as chromate). Acidify with acetic acid, precipitate the aluminium with ammonia ; and the chromium is determined in the filtrate either by the volumetric or gravimetric process. J. Clarke (Journ. Chem. Soc., 63. 1082, 1893) added sodium peroxide until the solution was alkaline, and instead of using hydrogen peroxide and sodium hydroxide T. Poleck, Chem. News, 69. 285, 1894 ; 0. Kassner, Archiv Pharm., 232. 226, 1895 ; M. E. Pozzi-Escot, Bull. Soc. Chim. (4), 5. 558, 1909. 5 L. L. de Koninck, P/iarm. Ztg., 78. 594, 1889 ; F. W. Richardson, W. Mann, and W. Hanson, Journ. Soc. Chem, Ind., 22. 614, 1903 ; W. F. Hillebrand, Bull. U.S. Geol. Sur., 176. 80, 1900 ; F. L. Langmuir, Ueber die quantitative Bestimmung des Chroms auf geivichts- analytischen auf kolorimetrischen Wege, Freiburg i. Br., 1906. 474 A TREATISE ON CHEMICAL ANALYSIS. Nessler's tubes may be used and the solution examined through a vertical column instead of horizontally as in the titanium determination. EXAMPLE. 250 c.c. of a solution (from, say, 1 grm. sample) were treated as indicated above. 48 c.c. of water were required to make 10 c.c. of a solution containing, say, 10 c.c. of the standard per 100 c.c. of solution (that is, 0'0002 grin, of Cr 2 3 per c.c.) the same tint. Hence, 48 + 10 = 58 c.c. of the standard solution were equivalent to the solution from the given sample. Consequently, the 250 c.c. (1 grm. of clay) had : = . 0008 irf DO that is, the sample has the equivalent of 0'086 per cent, of chromic oxide. The amount of chromic oxide can be deducted from the ammonia precipitate. According to Horn, 1 the tint is most sensitive when about 0;00002 grm. of potassium chromate is present per 100 c.c. of solution. 2 The smallest amount of chromium which can be detected in distilled water is 0*000013 grm. per 100 c.c., but if the solution contains much less than about 0-004 grm. of chromic oxide, Cr 2 3 , per 100 c.c. the comparison is not satisfactory, and a greater quantity of the original sample must be taken. Errors. The following numbers represent the results which can be obtained with known mixtures : Cr 2 3 used .' 8'88 8'88 8'88 8*88 1T67 6*34 mgrms. Cr 2 3 found . . . 8'45 9'01 8'91 9'29 11*54 6'27 mgrms. Error ... . . -0'43 +0*13 +0*03 + 0-41 - 0*13 -0*07 ragrm. If vanadium be present, the solution can be reserved for the determination of vanadium by the volumetric process ; if manganese be present, see page 384. 248. The Analysis of Chromites and Chromic Oxides. In the analysis of chromite (chrome iron ore), and also of the ignited chromic oxides used as colouring agents, we have a special difficulty arising from the insolubility, or rather the extremely slow rate of solution, of the materials in acids. We therefore depend upon a preliminary fusion of the material with a suitable flux. Here again the action of the flux is so slow that it is important to grind the materials under investigation to a very fine powder, and to heat the mixture for a comparatively long time. There is a wide choice in the selec- tion of the flux. 3 Sodium peroxide is now in almost general use. Here, the 1 D. W. Horn, Amer. Chem. Journ., 35. 253, 1906; M. Dittrich, Zeit. anorg. Chem., So. 171, 1913. ' 2 For a colorimetric reaction based upon the colour produced by mixing diphenylcarbazide and chromic acid, see P. Cazeneuve, Zeit. angew. Chem., 13. 958, 1900; Bull. Soc. Chim., (3), 25. 758, 1901 ; A. Moulin, ib. (3), 31. 295, 1904 ; Chem. Neivs, 89. 268, 1904 ; F. L. Langmuir, I.e. See page 454. 3 FLUXES FOR OPENING CHROMITE. Numerous fluxes have been used for the chromite fusion (S. Rideal and S. Roseblum, Chem. News, 73. 1, 1896 ; A. Miiller, Bull. Soc. Chim. (4), 5. 1133, 1909). For example: SODIUM HYDROXIDE ALONK (L. Duparc, Ann. Chim. Anal., 9. 201, 1904 ; L. Duparc and A. Leuba, Ann. Chim. Anal. Chim., 9. 201, 1908 ; H. N. Morse and W. C. Day, Amer. Chem. Journ., 3. 163, 1881 ; Chem. News, 44. 43, 1881) ; POTASSIUM HYDROXIDE AND CHLORATE(H. Schwarz, LiebigsAnn., 69. 212, 1849 ; Zeit. anal. Chem., 22. 83, 530, 1883 ; H. Pellet, Berg. Hutt. Ztg., 40. 224, 1881) ; MAGNESIA OR LIME WITH CAUSTIC SODA (J. Clark, Journ. Soc. Chem. Ind., u. 501, 1893; Chem. News, 24. 286, 304, 1871 ; A. Christomanos, er., 10. 16, 364, 1877 ; Zeit. anal. Chem., 17. 244, 1878 ; P. Veksin, Berg. Journ., 4. 437, 1908 ; J. Clouet, Dingler's Journ., 193. 33, 1869 ; Zeit. anal. Chem., 17. 244, 1878). Clark's flux is 1 part of the sample with 8 parts of a mixture of 5 parts of sodium hydroxide and 3 of calcined magnesia. Heat in a Bun sen's flame in a platinum crucible for about 40 minutes ; WITH CALCIUM CHLORIDE (J. Massignon, Journ. Anal. App. Chem., 5. 465, 1891 ; J. Massignon and E. Watel, Bull. Soc. Chim. (3), 5. 371, 1891) ; WITH SODIUM CARBONATE, (R. Kayser, Zeit. anal. Chem., 15. 187, 1876); WITH A MIXTURE OF SODIUM AND POTASSIUM CARBONATES (J. E. Stead, Journ. I. S. Inst., i. 160, 1893). Stead's mixture is sometimes called DETERMINATION OF CHROMIUM, VANADIUM, AND URANIUM. 475 action is quick, and the extreme care in grinding is not so very important as with less active fluxes. The fusion with sodium peroxide, or with a mixture of potassium hydroxide and sodium peroxide, can be made in a nickel, iron, 1 copper, or silver 2 crucible. Platinum crucibles should not be used with caustic alkalies. If iron and chromium are alone to be determined, a porcelain crucible may be used. The procedure is as follows : Sodium Peroxide Fusion. Grind the substance to a fine impalpable powder the tribasic flux : 4 parts lime, 1 sodium carbonate, 1 potassium carbonate. It resembles Kayser's flux 2 parts of sodium carbonate, and 3 parts of lime; WITH BORAX (R. Fieber, Chem. Ztg., 24. 333, 1900; R. W. E. Maclvor, Chem. News, 82. 97, 1900). SODA LIME AND SODIUM NITRATE (F. Calvert, Dingier' s Journ., 125. 466, 1852; J. Fels, ib. , 224. 86, 1877) ; WITH POTASSIUM CHLORATE (E. Reinhardt, Zeit. anal. Chem., 13. 430, 1889; J. B. Britton, Chem. News, 21. 266, 1870 ; Zeit. anal. Chem., g. 487, 1870 ; J. Fels, ib., 18. 498, 1879); AND BORAX GLASS (L. Perl and V. Stefko, Stahl Eisen, 24. 1373, 1904). SODIUM PEROXIDE (W. Hempel, Zeit. anorg. Chem., 3. 193, 1893 ; H. K. Tompkins, Chem. News, 68. 136, 1893 ; J. Clark, Journ. Chem. Soc., 63. 1079, 1893 ; E. H. Saniter, Journ. I. S. Inst.,^. 153, 1895 ; Journ. Soc. Chem. Ind., 15. 156, 1896 ; S. Rideal and S. Roseblum, ib., 14. 1017, 1895 ; C. Glaser, Journ. Amer. Chem. Soc., 20. 130, 1898; Chem. News, 77. 123, 1898 ; G. Tate, ib., 80. 235, 1899 ; H. Freseniusand H. Bayerlein, Zeit. anal. Chem., 37. 31, 1898 silver crucible) ; WITH SODIUM CARBONATE (L. Luechese, Ann. Chim. Anal. App., g. 450, 1908; M. Hbhnel, Archiv Pharm., 232. 222, 1886; C. Glaser, Chem. Ztg. t 18. 1448, 1894); WITH SODIUM NITRATE AND SODIUM CARBONATE (J. A. Muller, Bull. Soc. Chim. (4), 5. 1133, 1909) ; WITH SODIUM HYDROXIDE (J. Spuller and S. Kalman, Chem. Ztg., 17. 880, 1207, 1412, 1893). POTASSIUM PEROXIDE AND SODIUM CARBONATE (J. A. Miiller, Bull. Soc. Chim. (4), 5. 1133, 1909), BARIUM PEROXIDE (E. Donath, Dingier' s Journ., 263. 245, 1887); WITH SODIUM CARBONATE (L. P. Kennicut and G. W. Patterson, Journ. anal. Chem., 3. 131, 1889). POTASSIUM CARBONATE WITH SODIUM HYDROXIDE (C. Hausermann, Chem. Ztg., 15. 1601, 1891) ; WITH SODIUM CARBONATE AND BORAX (W. Dittmar, Exercises in Quantitative Analysis, Glasgow, 128, 1887 ; Dingler's Journ., 221. 450, 1878 ; P. Hart, Journ. prakt. Chem. (1), 67. 320, 1856 ; E. Waller and H. T. Vulte, Chem. News, 66. 17, 1892 ; 0. Nydegger, Zeit. angew. Chem., 24. 1163, 1911). SODIUM CARBONATE AND POTASSIUM CHLORATE (R. Fresenius and E. Hintz, Zeit. anal. Chem.,2g. 29, 1890). AMMONIUM NITRATE AND CAUSTIC SODA (C. A. Burghardt, Chem. News, 6l. 260, 1890). POTASSIUM BISULPHATE (E. Clark, Journ. Amer. Chem. Soc., 17. 327, 1895 ; H. Tamm, Chem. News, 24. 307, 1871 ; C. L. Oudesluys, ib. , 5. 255, 1862 ; C. O'Neill, ib., 5. 199, 1862 ; F. A. Genth, ib., 6. 30, 1862 ; Zeit. anal. Chem., I. 498, 1862 ; R. Namais, Stahl Eisen, IO. 977, 1890 ; T. S. Hunt, C. O'Neill, and F. A. Genth, Amer. J. Science (3), 5. 418, 1873), 1 part sample with 15 parts of potassium bisulphate ; WITH SODIUM FLUORIDE OR CRYOLITE (H. Hager, Untersuchungen, Leipzig, I. 263, 1888 ; P. C. Dubois, Zeit. anal. Chem., 3. 401, 1864 ; S. Kern, Chem. News, 35. 107, 1877 ; F. W. Clarke, ib., 17. 232, 1868; Amer. J. Science (2), 45. 173, 1868); AND SODIUM NITRATE (H. Rose, Ausfuhrliches Handbuch der analytischen Chemie, Braunschweig, 2. 376, 1851). POTASSIUM HYDROXIDE AND THE ELECTRIC CURRENT (E. F. Smith, Ber., 24. 2182, 1891). HYDROCHLORIC ACID UNDER PRESSURE (P. Jannasch and H. Vogtherr, Ber., 24. 3206, 1891). HEATING IN CURRENT OF CHLORINE, etc. (R, Fresenius and E. Hintz, Zeit. anal. Chem., 29. 28, 1890). HEATING WITH BROMINE WATER IN SEALED TUBES some days (E. F. Smith, Amer. J. Science (3), 15. 198, 1877); WITH SULPHURIC ACID (A. Mitscherlich, Zeit. anal. Chem., i. 54, 1861 ; F. C. Philips, ib., 12. 189, 1873); J. Jones, Chem. News, 65. 8, 1892). TREATMENT WITH A MIXTURE OF POTASSIUM CHLORATE AND NITRIC OR SULPHURIC ACID. This IS not particularly suited for chromites, but it is satisfactory for some chromic oxides (F. H. Storer, Proc. Amer. Acad., 4. 352, 1869 ; Chem. News, 21. 195, 1870 ; C. Brunner, ib., 4. 57, 1861). 1 H. N. Morse and W. C. Day, Amer. Chem. Journ., 3. 163, 1881. 2 All these crucibles are attacked. Porcelain may last from two to four times before it is cut through by the flux. Silver crucibles lose about half a gram per fusion. There is a danger in the use of silver crucibles from the fact that the crucible can be melted if heated at too high a temperature on the Bunsen's flame. Silver, copper, and lead may be found in the fusion after using a silver crucible W. Dittmar, Chem. Ztg., 15. 1521, 1580, 1891. W. Bettel (Chem. News, 43. 94, 1881) recommended a platinum crucible gilded inside for fusions with alkalies. See also J. L.[Smith, Chem. News, 31. 55, 1875. Gold crucibles would be cheaper, in view of the higher price of platinum. Nickel crucibles are generally used. They lose about O'l grm. per fusion under the conditions described in the text. One crucible is said to last 15-25 fusions after a little practice in their use but this estimate is rather high. Gold is not attacked chemically by fused potassium hydroxide, but nickel forms nickel hydroxide M. le Blanc and O. Weyl, Ber., 45. 2300, 1912. Some use copper crucibles. The contamination from copper crucibles is said to give least trouble in many complex analyses. 476 A TREATISE ON CHEMICAL ANALYSIS. in an agate mortar, and, by means of a platinum spatula, intimately mix 0'5 grm. of the dry (110) powder with about 8 grms. of sodium peroxide l in, say, a nickel crucible of not less than 50 c.c. capacity, and fitted as described on page 266. Heat the crucible gently over the tip of a Bunsen flame until the contents of the crucible are fluid this takes about 10 minutes. Raise the temperature to low redness say another 10 minutes. Cool. Place the crucible in a beaker (400 c.c.), cover the crucible with a watch-glass, and gradually add about 50 c.c. of cold water in such a way that the first addition of water flows slowly between the watch-glass and the crucible. The mixture bubbles up and dissolves in a few minutes. 2 Remove the crucible and rinse it carefully with water. Boil the solution about 10 minutes to decompose the sodium peroxide. 3 The oxides of iron, manganese, cobalt, and nickel (mainly from the crucible) remain as insoluble precipitates. The chromium, aluminium, and most of the silica, with a little magnesia and lime, remain in solution. The solution is now so strongly alkaline that it would destroy the filter paper, if an attempt at filtration be made. Hence, add about 8 grms. of ammonium carbonate, so as to neutralise the greater part of the sodium hydroxide but still leave the latter in slight excess. Filter the solution into a litre flask, and wash the residue on the filter paper with water. See note, page 494. 249. The Volumetric Determination of Chromium. Make the solution up to the litre mark with water and determine the chromium in an aliquot portion. 4 Hence, pipette a portion from the litre flask ; acidify the solution with a large excess of dilute sulphuric acid (1 : 4) ; and dilute to about 500 c.c. with cold water. Place some solid sulphate in a weighing bottle. Weigh. Add the solid salt in small portions at a time until the yellow colour of the chromate has disappeared, and a blue coloration is obtained when a drop is touched on a spot of potassium ferricyanide solution. The chromate is reduced and the ferrous sulphate is oxidised. Weigh the bottle again. The loss represents the amount of ferrous ammonium sulphate added to the solution. The action is usually represented : 2K 2 Cr0 4 + 6FeS0 4 . (NH 4 ) 2 S0 4 . 6H 2 + 8H 2 S0 4 = Cr 2 (S0 4 ) 8 + etc. Titrate the solution with standard potassium permanganate solution. Then dissolve, say, 1-5 grms. of the ferrous ammonium sulphate in 500 c.c. of water acidulated with sulphuric acid (10 c.c. concentrated acid). Also make a blank experiment with acidulated water 500 c.c. We have now sufficient data to calculate the amount of chromium as chromic acid in the aliquot portion taken 1 Reject the crust, if any, formed at the top of the peroxide bottle. This is mainly carbonate and oxide. L. Archbutt (Analyst, 20. 3, 1895) reports a sample of sodium peroxide with 0'5 per cent, of alumina and ferric oxide. 2 Sometimes it requires warming a little. If the residue be insoluble in dilute sulphuric acid (1 : 4), this means that some of the substance has not been decomposed. In that case, the portion not dissolved must be re-fused with the flux. If nitric acid be used to take up the cake, some chromate may be reduced by the acid. A. Leuba, Ann. Chim. Anal. App., 9. 303, 1908. Note the formation of volatile chromyl chloride when sulphuric acid is heated with a chloride and a chromate. 3 To ensure the decomposition of the peroxide, some here recommend an evaporation to dry- ness". Peroxides must be absent in order to prevent reducing the chromates to chromic oxide Cr 2 3 when the solution is acidified. If the solution is purple-coloured, add a gram of sodium peroxide, and boil another 10 minutes. 4 H. Schwarz, Zeit. anal. Chem., 22. 530, 1883 ; H. Bollenbach, Chem. Ztg., 31. 760, 1907 ; H. Vignal, Bull Soc. Chim. (2), 45. 171, 1886; Chem. News, 33. 195, 1886; T. W. Hogg, Journ. Soc. Chem. Ind., 10. 340, 1891 ; R. L Leffler, Chem. News, 77. 156, 1898 ; W. Galbraith, ib., 35. 151, 1877; 77. 187, 1898. DETERMINATION OF CHROMIUM, VANADIUM, AND URANIUM. 477 from the litre flask. Every gram of ferrous ammonium sulphate oxidised represents 0-06447 grrn. of chromic oxide Cr 2 3 ; or 0'0853 grin, of chromic trioxide CrOg. 1 Error due to the Filtration of Chromate Solutions through Filter Paper. 2 If chromic acid solutions be filtered through paper, low results may be obtained owing to a partial reduction of the chromate. This error can be avoided by filtering the solution through asbestos, or by oxidising the solution after the filtration. The effect is produced by both acid and alkaline solutions it is more marked with hot than with cold solutions ; and also more marked with the lower- grade papers. For instance, 200 c.c. of a solution containing the equivalent of 25 c.c. of potassium bichromate (1 c.c. = 0'01 grm. Fe) were digested for half an hour with a 12'5-cm. filter paper. In one case 50 c.c. of N-NaOH were added to the solution, and in another case 50 c.c. of N-H 2 S0 4 . The results are shown in Table LXI. Table LXI. Effect of Filter Paper on Chromate Solutions. Filter paper. Ash of filter paper. Equivalent in c.c. K 2 Cr 2 7 remaining. Hot (100). Cold. Acid. Alkaline. Acid. Alkaline. Munktell's Swedish Schleicher and Schiills' 589 . Ordinary English . French grey .... 0-0005 0-0003 0-0028 0-0139 24-6 23-8 23-6 22-5 24-9 24-2 24-6 24-0 250 25-0 24-0 25-0 25-0 247 The error with the ordinary filtration of chromate solutions may be equivalent to just over O'l per cent, of chromium. The residue on the filter paper can be dissolved in dilute sulphuric acid (1 : 4) 3 and the iron determined by titration (page 198). The result is usually calculated to FeO. The silica can be determined by evaporating the sodium peroxide fusion with hydrochloric acid, as described on page 167. An aliquot part of the filtrate is treated with ammonia for alumina as indicated under clay analysis. An aliquot portion of the solution from the silica filtrate is used for the determination of sulphur. Magnesia, manganese, and lime can be determined by the methods of pages 213, 218, and 374. 250. The Gravimetric Determination of Chromium as Barium Chromate. The Solubility of Barium Chromate. This process 4 is based on the sparing solubility of barium chromate in water. One part of the salt dissolves in 87,000 parts of cold, and in 23,000 parts of boiling water. 5 The presence of acetic acid increases the solubility of the chromate ; thus, 1 part of the salt dissolves in 3670 parts of water containing 1 per cent, of acetic acid. The presence of ammonium salts also makes the salt more soluble. For instance, 1 part of barium chromate dissolves in about 23,000 parts of water containing 0*5 per cent, of ammonium chloride ; in 45,000 parts of water containing 0'5 per cent, of ammonium nitrate ; in 50,000 parts of water containing 0'75 per cent, of ammonium acetate ; and in 1 This process can also be used for evaluating alkaline chromates and bichromates. 2 H. Jervis, Chem. News, 77. 133, 1898 ; A. Allison, ib., 96. 1, 1907. 3 If any remain undissolved, the original fusion was not complete. 4 W. Gibbs, Zeit. anal. Chem., 12. 309, 1893 ; G. Chancel, Compt. Rend. 43. 927, 1856. 5 R. Fresenius, Zeit. anal. Chem., 29. 414, 1890. 478 A TREATISE ON CHEMICAL ANALYSIS. 24,000 parts of a 1*5 per cent, solution of ammonium acetate. Hence the solution in which the precipitation is made must be but faintly acid with acetic acid. Mineral acids should be absent, or, if present, approximately neutralised by the addition of sodium carbonate, followed by sodium acetate. 1 Oxidation to ^Chromic Acid. If the chromium be present in the lower state of oxidation, it must be converted into chromic acid by the addition of an oxidising agent. For example, warm the alkaline solution with potassium persulphate, or sodium peroxide, or hydrogen peroxide,' 2 or bromine, 3 or chlorine. 4 Nitric acid is rather slow in action. 5 Potassium chlorate acts fairly well in the presence of hydrochloric acid, but there is a difficulty in removing the last traces of chlorine oxides, especially in dilute solutions. 6 If the chromium be already present in the form of chromate, there is no need for the oxidation. Precipitation of Barium Chromate. Neutralise the solution with acetic acid. 7 Add a hot solution of barium acetate, 8 slowly, drop by drop, 9 to the hot neutral or weakly acidified solution. Let the mixture stand until cold ; and, when the precipitate has settled, wash five times by decantation with hot water. 10 Filter n through asbestos felt in a Gooch's crucible. 12 Wash the precipitate with dilute alcohol (alcohol 1 vol., water 10 vols.), and dry it in an air bath at 110 . 13 The Ignition. Heat the covered crucible in a saucer, gently at first, and finally over the full flame of a Teclu's burner. In about 5 minutes, remove the lid, and heat the crucible until the precipitate has a uniform yellow colour. 14 Cool the crucible in a desiccator, and weigh as barium chromate BaCr0 4 . Multiply the weight by 0*3 or by 0'2999 to get the equivalent amount of chromic oxide Cr 2 3 . Results. Working with solutions containing the equivalent of 0'6886 grm. of BaCr0 4 , the following numbers were obtained : BaCr0 4 . . . 0'6922 0'6918 0'6876 0'6861 grm. Error .... +0-0036 +0*0032 -O'OOIO -0'0025 grm. The precipitate may be dissolved in hydrochloric acid, mixed with a solution of potassium iodide, and the liberated iodine titrated with sodium thiosulphate as indicated on page 352 for copper, etc. 15 Instead of barium acetate, lead nitrate 1 L. Schulemd, Journ. prakt. Chem. (2), 19. 36, 1890. 2 W. J. Sell, Chem. News, 54. 299, 1886 ; A. Carnot, Compt. Rend., 107. 997, 1888. 3 Bromine is a favourite oxidising agent. It does not attack platinum. R. von Wagner, Deut. Ind. Ztg., 19. 114, 1878 ; Dingler's Journ., 218. 332, 1875 ; 219. 544, 1876 ; P. Waage, Zeit. anal. Chem., 10. 206, 1871 ; Chem. News, 25. 282, 1872; H. Kammerer, Ber., 4. 218, 1871 ; E. Reichardt, Arch. Pharm. (3), 5. 1, 1876 ; G. Vulpius, ib. (3), 5. 422, 1876. 4 W. Gibbs, Amer. J. Science (2), 39. 58, 1865. 5 If the solution also contains chlorides, nitric acid must not be used in a platinum dish. 6 F. H. Storer (Proc. Amer. Acad., 4. 342, 1869 ; Amer. J. Science (2), 45. 190, 1868) used a mixture of potassium chlorate and nitric acid for oxidising chromium salts B. Pawolleck, Ber., 16. 3008, 1883. 7 For the reduction of chromic by acetic acid, see H. Basset, Chem. News, 79. 157, 1899. " Small quantities of acetone, aldehyde, etc., are present in the inferior (grades of) acid." 8 Barium nitrate or chloride furnish rather high results, and are n()t so satisfactory as the acetate F. L. Langmuir, Ueber die quantitative Bestimmung des Chroms auf gewichtsanalytischen und Jcolorimetrischen Wege, Freiburg i. Br., 1906. 9 If the barium acetate be added too quickly, barium acetate, etc. , will be carried down with the chromate R. H. Richards, Chem. News, 21. 198, 1870. 10 A. H. Pearson (Chem. News, 21. 198, 1870) recommends a solution of ammonium acetate for the washing. 11 The precipitate is very awkward to filter through paper. 12 Use very gentle suction. If strong suction be applied, the felt gets choked, and subsequent filtration and washing will be very slow. 13 It will be remembered that vanadium will be precipitated as barium vanadate, if this metal be present A. Carnot, Compt. Rend., 104. 1803, 1887. 14 The edges near the crucible sometimes appear green. This is due to a slight reduction by dust and alcohol. When heated in the open crucible, the green colour gradually disappears. 15 K. Zulkowsky, Journ. prakt. Chem. (2), 103. 351, 1868 ; L. Crismer, Ber., 17. 642, 1884. DETERMINATION OF CHROMIUM, VANADIUM, AND URANIUM. 479 or mercurous nitrate may be used as precipitating agents. The precipitate is washed with a solution of the salt used for the precipitation. In the case of mercurous chromate, ignition converts the precipitate into chromic oxide Cr 2 O 3 . 251. The Gravimetric Determination of Chromium as Chromic Oxide Chromium is most easily and quickly determined by volumetric methods. The gravimetric process is only employed in special cases. If the chromium be in solu- tion in the lower state of oxidation, it can be precipitated by the addition of a slight excess of ammonia in the presence of ammonium salts ; 1 of freshly prepared ammonium sulphide to the boiling solution ; or of hydrazine sulphate. 2 Filter, wash, calcine, and weigh as chromic oxide Cr 2 3 . Details of the procedure are as follows : Reduction of Chromates. If chromates be present, they must be reduced to a form suitable for treatment by this process. Boil the solution with 10 c.c. of hydrochloric acid, and 10 c.c. of alcohol; and continue the boiling until all the alcohol has been expelled. Sulphurous acid and sulphites are also excellent reducing agents. Jannasch and Mai 3 reduce the solution by the addition of 5 c.c. of hydrochloric acid and a couple of grams of hydroxylamme hydrochloride. This, later on, facilitates the precipitation and washing. Slightly better results are obtained with sulphuric acid 4 and hydroxylamine hydrochloride. Suppose that 25 c.c. of a solution of potassium chromate be under investigation, acidify the cold solution with 1 to 5 c.c. of dilute sulphuric acid, and add 1 gram of hydroxylamine hydrochloride. 5 The Precipitation. Add a slight excess of ammonia, 6 and boil the solution to drive off the excess of ammonia. Filter quickly while still hot, and wash with a dilute solution of ammonium nitrate as indicated for alumina (page 183). The precipitate is dissolved in hydrochloric acid and again precipitated with ammonia, etc., as just described. 7 Ignition. The wet precipitate is placed in a platinum crucible, and the temperature gradually raised so as to prevent small particles of the precipitate being projected from the crucible. After heating over the Bunsen's flame, finish the ignition over the blast. Weigh as Cr 2 3 . The following numbers were obtained with one precipitation, using the equivalent of 0'1034 grm. of Cr 2 3 , 2 c.c. of the tannin solution, 1 c.c. of sulphuric acid, and 1 grm. of hydroxylamine hydrochloride : Cr 2 3 found . . . 0'1033 0'1031 0'1034 grm. Error . . . . -O'lO -0'29 -O'OO percent. Errors. As a matter of fact, the results are generally high, possibly owing to the formation of a little alkaline chromate. The alkalies come from the 1 M. Z. Jowitschitsch, Monats. CKem., 34. 225, 1913. 2 T. Hanusand T. Lucas, Chem. Ztg., 36. 1134, 1912. 3 P. Jannasch and J. Mai, Ber., 26. 1786, 1893; P. Jannasch and F. Riihl, Journ. prakt. Chem. (2), 72. 10, 1905 ; C. Friedheim and P. Hasenclever, Zeit. anal. Chem., 44. 594, 1905. 4 It is not advisable to use less than 1 c.c. of dilute sulphuric acid. If too little acid be employed, the reduction is disturbed by a side reaction attended by the development of nitrous fumes, and the formation of a brown instead of a violet-coloured solution. 5 The addition of 2 c.c. of a 2 '5 per cent, solution of tannin before the addition of the ammonia gives a precipitate much easier to filter and wash R. E. Divine, Journ. Soc. Chem. Ind.,24. 11, 1905. 6 Chromium hydroxide is soluble in an excess of ammonia, forming a red solution. This can easily be demonstrated by adding ammonia to a dilute solution of a chromium salt. No precipitate can be observed until the solution has been boiled a long time. 7 Take care in the washing, or the results may be 2 per cent, high when only one precipitation is made. 480 A TREATISE ON CHEMICAL ANALYSIS. reagents, imperfect washing, etc. 1 Genth 2 recommends boiling the precipitate with sulphurous acid before it is weighed, in order to make sure that the chromates are reduced ; Clouet 3 recommends boiling the precipitate with alcohol ; and Clark 4 suggests removing the chromates by washing with dilute acetic acid, and estimating them by titration with Mohr's salt ; the chromates may be also estimated colorimetrically. With the method described in the text there is no need for any of these methods of purification. Hence, the two important errors in this determination arise from the solubility of the hydroxide in ammonia, and the presence of alkalies in the calcined oxide. If the excess of ammonia be expelled by boiling, the precipitate becomes slimy, and difficult to wash, and so aggravates the second source of error. 5 The effect of a slight excess of ammonia need not be feared when hydro- xylamine is used, because hydroxylamine, not free ammonia, will be present, and chromium hydroxide is not appreciably soluble in this menstruum. Again, the addition of tannin enables the precipitate to be more readily washed, and thus facilitates the removal of alkalies. It is interesting to note that the precipitated chromium hydroxide may be coloured either violet, or green, or intermediate tints. The green precipitate 6 is commonest with the alcohol reduction, the violet with the hydroxylamine reduction. The violet precipitate is nearly always obtained by the method described in the text. If tannin is present, the precipitate appears grey. The green precipitate seems to have a greater avidity for alkalies than the violet- coloured one, and thus gives rather higher results. Effect of Phosphoric Acid. When phosphoric acid is present, chromium phosphate will be precipitated. 7 In that case, fuse the dried precipitate with sodium carbonate and a little sodium nitrite. Dissolve the melt in water acidified with nitric acid, and precipitate the phosphoric acid with ammonia and magnesia mixture. The chromium may be determined by difference, or precipitated from the filtrate as barium chromate, or determined volurnetrically. 252. The Volumetric Determination of Vanadium. One method for the determination of vanadium is based on the reduction of vanadium pentoxide V 2 5 by sulphur dioxide to vanadium tetra-oxide V 2 4 , and the re-oxidation of the latter by a standard solution of potassium per- manganate. 8 1 T. Wilm, Per., 12. 2223, 1887 ; Chem. News, 41. 222, 1880 ; F. P. Treadwell, Ber., 15. 1392, 1880. A. Souchay (Zeit. anal. Chem., 4. 66, 1865) thinks that the high results are due to the action of the ammonia on the glass vessels. 2 F. A. Genth, Amer. J. Science (2), 5. 418, 1873. 3 P. Clouet, Ann. Chim. Phys. (4), 16. 90, 1849. 4 J. Clark, Chem. News, 24. 304, 1871. 5 J. HanusandJ. Lukas (Inter. Cong. App. Chem., 8. 209, 1912) quantitatively precipitate chromium from neutral or alkaline solutions containing chromates, or chromic salts by means of hydrazine hydrate or some of the derivatives of hydrazine hydrazine sulphate, phenyl- hydrazine, thiosemicarbazide. The precipitation is more rapid and complete in the presence of ammonium chloride, and thiosemicarbazide gives best results. 6 The violet hydroxide is possibly Cr(OH) 3 ; the green, Cr. 2 0(OH) 4 . 7 H. Baubigny, Bull. Soc. Chim. (2), 41. 291, 1884 ; Chem. News, 50. 18, 1884. 8 W. F. Hillebrand, Journ. Amer. Chem. Soc., 20. 461, 1898 ; A. Bettendorff, Pogg. Ann., 160. 126, 1877; C. Czudnowicz, ib., 120. 17, 1863; C. Rammelsberg, Ber., I. 158, 1868; B. W. Gerland, ib., 10. 1513, 1516, 1877 ; 0. Manasse, LieUg's Ann., 240. 23, 1887 ; Zeit. anal. Chem., 32. 225, 1893 ; 0. Lindemann, ib., 18. 99, 1879 ; F. A. Genth and G. von Rath, Chem. News, 53. 218, 1886 ; 34. 78, 1896 ; J. R. Cain, Bull. Bur. Standards, 7. 377, 1911 ; E. de M. Campbell and C. E. Griffin, Journ. Ind. Eng. Chem., I. 661, 1909 ; G. Auchy, ib., I. 455, 1910 ; G. Edgar, Amer. J. Science (4), 25. 332, 1908 ; F. A. Gooch arid L B Stookay, ib. (4), 14. 369, 1902 ; F. A. Gooch and R. D. Gilbert, ib. (4), 15. 389, 1903 ; F. A. Genth, ib. (3), 12. 32, 1876 ; F. A. Gooch and G. Edgar, ib. (4), 25. 233, 1908 ; Ber., 38. 600, 1905 ; DETERMINATION OF CHROMIUM, VANADIUM, AND URANIUM. 481 Suppose we have to deal with the solution remaining after the determination of chromium (page 473). 1 The boiling solution is reduced with sulphur dioxide, and the excess of sulphur dioxide is expelled by passing a current of carbon dioxide through the solution, which is then titrated, while hot, with a dilute solution of potassium permanganate approximately Tlyo~~N, that is, O3163 grm. of KMn0 4 per litre 2 until a permanent pink colour is obtained. If but a small quantity of permanganate be employed, verify the presence of vanadium as follows : Evaporate the solution to dryness in order to expel the sulphuric acid ; take up the residue with 3-4 c.c. of water ; acidify the solution with a few drops of nitric acid ; and add two or three drops of hydrogen per- oxide. A brownish-red colour indicates vanadium. If the solution be titrated cold, the end point is not so sharp. 3 After the titration, reduce the boiling solution with a current of sulphur dioxide, when ^2^2(^^4)3 passes to V 2 2 (S0 4 ) 2 ; or, generalised : Boil the solution while a rapid current of carbon dioxide is passing through until the escaping gas no longer decolorises a solution of potassium permanganate. Repeat the titration on the hot solution. The reduction and titration can be repeated once again. The last two results will probably be lower than the first. The mean of these titrations represents the vanadium in the given solution. Hillebrand's Correction for Excess of Chromium. If over 0*005 grm. of chromium be present, a correction must be made, since an appreciable quantity of the permanganate is needed for oxidising the chromium. Make a solution of potassium chromate containing the same amount of chromium as the solution under investigation. Reduce with sulphur dioxide as indicated for vanadium, and titrate with the standard permanganate. The amount of permanganate used must be subtracted from that consumed in titrating for vanadium, and the difference represents that used in converting V 2 4 ->V 2 5 . It is perhaps easier to determine the chromic and vanadic acids volumetrically by the following process than to apply these corrections. Volumetric Determination of Chromic and Vanadic Acids. It might be added that Edgar 4 has devised a process for the simultaneous determination of these two compounds. The process is based on the fact that hydrobromic acid will reduce vanadic acid to vanadium tetroxide V 2 5 -W 2 4 and chromic acid to chromic oxide 2Cr0 3 ->Cr 3 ; and hydriodic acid will reduce vanadium tetroxide to the trioxide V 2 4 ->V 2 () 3 , and leave chromic oxide unaffected. The mixture of E. P. Alvarez, Chem. Ztg., 33. 149, 1909 ; P. Slawik, ib., 36. 171, 1912 ; E. Campagne, Bull. Soc. Chim, (3), 31. 962, 1906 ; Compt. Rend., 137. 570, 1898 ; A. Ditte, ib., Mi. 698, 1885 ; C. Matignon and P. Monnet, ib., 134. 542, 1902 ; W. Heike, Stahl Eisen, 2$. 1357, 1905 ; C. Hensen, Kritische Untersuchung der Vanadin Bestimmungsmethoden, Aachen, 36, 1909. H. E. Roscoe (Journ. Chem. Soc. , II. 928, 1871) considered the process by zinc reduction not reliable. D. J. Demorest, Journ. Ind. Eng. Chem., 4. 249, 895, 1912 ; J. R. Cain and J. C. Hostetter, #.,4. 250, 1912 ; J. R. Cain and D. J. Demorest, ib., 4. 256, 1912 ; F. Garratt, ib., 4. 256, 1912. 1 If molybdenum arid arsenic be present, the solution, after the determination of chromium (page 473), is acidified with sulphuric acid, and treated with hydrogen sulphide (page 277), when both arsenic and molybdenum together with traces of platinum from the crucible are pre- cipitated. Boil the filtrate to expel hydrogen sulphide. 2 1 grm. KMn0 4 corresponds with 2 '883 grms. V 2 5 for the oxidation V 2 4 ->V 2 5 . 3 The temperature of the solution being titrated is of importance ; the best temperature is between 70 and 80. If the titration be conducted at too low a temperature, the end point is not sharp enough, and if at too high a temperature, the permanganate may be reduced by the hot sulphuric acid solution of manganous sulphate, as shown by A. C. Sarkar and J. M. Dutta (Zeit. anorg. Chem., 67. 225, 1910). 4 G. Edgar, Amer. J. Science (4), 26. 333, 1908. 31 482 A TREATISE ON CHEMICAL ANALYSIS. chromic and vanadic oxides (about 0'2 grm.) is boiled in a flask A, fig. 154, with 15 c.c. of hydrochloric acid, and 2 grms. of potassium bromide. The flask is arranged so that a slow current of hydrogen from a Kipp's apparatus and wash- bottle, B> passes while the solution is being boiled. The bromine vapour evolved owing to the reaction: V 2 5 + 2Cr0 3 + 8HBr = V 2 4 + Cr 2 3 + 4H 2 + 8Br, is led through a pair of Wolbling's flasks, (7, containing an alkaline solution of potassium iodide. When the reaction is over, the contents of the absorption bulbs are washed into an Erlenmeyer's flask ; the solution is acidified with hydrochloric acid \ and the free iodine, liberated by the action of the bromine on the potassium iodide Br 2 + 2KI = 2KBr + I 2 is titrated with sodium thio*- sulphate in the usual manner. Two grams of potassium iodide, 15 c.c. of hydrochloric acid, and 3 c.c. of syrupy phosphoric acid are now added to the solution in the boiling flask, and FIG. 154. Simultaneous determination of chromic and vanadic acids. the distillation repeated as before in a current of hydrogen. The absorption flasks are charged as before, and the iodine liberated by the reaction V 2 4 + 2HI = V 2 3 + H 2 + 1 2 is absorbed by the alkaline potassium iodide, and titrated as before. The second titration furnishes data for calculating the amount of vanadic acid in the original solution ; and the first and second titra- tions give data for calculating the amount of chromic acid. Volumetric Determination of Vanadium and Molybdenum. 1 li a solution of molybdic and vanadic acids be treated with sulphur dioxide, the vanadic acid, V 2 5 , is alone reduced to vanadium tetroxide, V 2 4 , provided the solution contains no more than O2 grm. of molybdic acid per 50 c.c., and approximately 1 c.c. of free sulphuric acid. When the reduction is complete, pass a current of carbon Iioxide through the boiling liquid to drive off the sulphur dioxide. The vanadium tetroxide can be titrated in the usual manner with T VN-potassium permanganate. ihen pass the solution through a reductor containing amalgamated zinc, and 1 G. Edgar, Amer. J. Science (4), 25. 332, 1908. DETERMINATION OF CHROMIUM, VANADIUM, AND URANIUM. 483 received into 50-60 c.c. of a 10 per cent, solution of ferric alum ; add 8-10 c.c. of syrupy phosphoric acid to decolorise the iron salt, and titrate with- y^-N potassium permanganate as before. The permanganate required for the latter titration includes that required for converting the V 2 2 to V 2 5 , and for the conversion Mo 2 3 to 2Mo0 3 ; the former is three times the amount consumed for the V 2 4 ->V 2 5 titration. Volumetric Determination of Vanadium and Iron. 1 If a mixed solution of ferric and vanadic oxides, acidified with a little sulphuric acid, be treated in a similar manner, the first titration (after the sulphur dioxide reduction) corresponds with the conversion of V 2 4 to V 2 5 , and of 2FeO->Fe 2 3 ; and the second titration (after the zinc reduction) with V 2 2 ->V 2 5 , and of 2 FeO->Fe 2 3 . Hence, with y^N-permanganate, the difference between the two titrations multiplied by 0*00456 gives the amount of vanadic acid originally present, and the amount of iron can then be calculated from either titration. 253. The Gravimetric Separation of Chromium and Vanadium. Instead of applying the colorimetric process for chromium, followed by the volumetric process for vanadium, these two constituents can be determined gravimetrically 2 in the following manner : Hydrochloric acid, if present, is expelled from the solution by repeated evaporation with nitric acid. 3 The solution is just neutralised with sodium hydroxide, and acidified with acetic acid, or lead acetate is added to the boiling solution. Boil three or four minutes. The voluminous precipitate of lead chromate and lead vanadate 4 soon settles, provided too great an excess of lead acetate has not been added. 5 1 G. Edgar, Amer. J. Science (4), 26. 79, 1908. . . 2 The chromium may be separated, if desired, by reducing the solution with sulphurous acid. Pour the boiling solution slowly, with constant stirring, into a boiling solution of sodium hydroxide containing 100 grms. of solid per litre. Boil, filter, and wash. Neutralise with nitric acid (blue litmus just red) and add a few drops of sodium hydroxide ; boil and filter. The precipitate contains the chromium, nickel, iron, manganese and copper if present. A trace of chromium may also pass into the filtrate, since the manganese peroxide formed during the boiling of the solution oxidises the chromium to soluble chromate (J. R. Cain, Journ. Ind. Eng. Chem., 3. 476, 1911 ; 4. 17, 1912 ; A. A. Noyes and W. C. Bray, Tech. Quart., 21. 14, 1908). Most of the vanadium passes into the filtrate, though a little may be retained by the precipitate, and two or three precipitations may be needed to get all the vanadium in the filtrate (W. Trautmann, Stahl Eisen, 30. 1802, 1910). This is attended by the passage of a relatively large amount of chromium into the filtrate. Vanadium, if present, is generally precipitated by ammonia along with aluminium, hydroxide, etc. Warm the precipitate with ammonium phosphate on a water bath. Ammonium vanadate passes into solution, and aluminium and ferric phosphates remain insoluble. Wash the precipitate with ammonium chloride solution 0. Bettendorff, Fogg. Ann., 160. 126, 1877. 3 If the fused cake obtained in, say, the barium carbonate precipitation is under treatment, dissolve the mass in hot water; add 2-3 c.c. of hydrogen peroxide to the filtrate in order to destroy any nitrates present and prevent their reducing chromium and vana'dium. Boil 10-15 minutes to destroy the hydrogen peroxide and prevent its reducing chromic acid when the solution is acidified (perchromic acid, of course, would be first formed, but this rapidly decom- poses). In case the cake is under treatment, it is well to make sure the insoluble matter does not contain chromium by re-fusing, it with the mixture (sodium carbonate 4 ; sodium nitrite 1). If the fused cake is white when cold, chromium is absent ; if yellow, determine the chromium coloiimetrically. 4 T. Fischer, Ueber die Bestimmung von Vanadinsdure, Berlin, 1894; G. Sefstrom, Fogg. Ann., 21. 43, 1831 ; J. J. Berzelius, ib., 22. 1, 1831 ; H. E. Roscoe, Liebig's Ann. Suppl, 8. 95, 1872 ; E. Claassen, Amer. Chem. Journ., 7. 349, 1886 ; H. Cormimbceuff, Ann. Chim. Anal. App., 7. 258, 1902. H. F. Watts ( Western Chem. Met., 5. 408, 1909 ; Chem. News, 101. 34, 1910) precipitates lead vanadate in a solution just acid with nitric acid and containing 1 to 2 grms. of sodium acetate. The precipitate is dissolved in dilute nitric acid, and the lead and nitric acid removed by repeated evaporation with sulphuric acid. 5 If the sodium carbonate fusion be treated with a slight excess of dilute nitric acid (1 : 1) and boiled vigorously a few minutes to expel the carbon dioxide, just neutralised with sodium 484 A TREATISE ON CHEMICAL ANALYSIS. Separation of the Chromium. Evaporate the solution to about 30 c.c. on a, water bath. Add 2 grms. of solid potassium carbonate ; evaporate the solution to dryness; take up the residue with a little hot water; let the precipitate settle, and decant the clear liquid through a filter paper. Wash the filter paper with water, and then wash any precipitate which may be on the paper back into the vessel containing the bulk of the precipitate. Evaporate the filtrate down to about 15 c.c. and repeat the operations just described. If the precipitate is white, 1 wash the precipitate ; if yellowish, repeat the treatment with potassium carbonate. The filtrate contains the chromic acid. Acidify the solution with acetic acid carbon dioxide is evolved, and lead chromate is precipitated. The solution is heated to boiling, and more lead acetate is added until all the chromium is precipitated, and the precipitate is treated gravimetrically as indicated on page 325 ; or the lead chromate precipitate is collected on the asbestos mat in a Gooch's crucible, and then dissolved in hot dilute hydrochloric acid (1 : 4). The solution is cooled'; treated with an excess of ferrous sulphate ; and titrated with standard bichromate or permanganate as directed on pages 198 and 453. Precipitation of the Vanadium. The precipitate of lead carbonate and lead vanadate 2 is dissolved in nitric acid ; evaporated to dryness ; and the solution taken up with dilute hydrochloric acid. The lead is precipitated as sulphide. Filter. Evaporate the solution down to a small volume, and treat it with concentrated nitric acid; evaporate to dryness; filter into a platinum crucible; and after calcination (fusion) weigh as V 2 5 . (For the effect of chlorides, see page 472.) The vanadium may here be determined volume trically (page 480) or colorimetrically. 254. The Rapid Determination of Vanadium Cain and Hostetter's Process. In the absence of titanium, the hydrogen peroxide colorimetric process described for titanium can be applied, mutatis mutandis, to vanadium. 3 Gregory's colorimetric process 4 can be used for the colorimetric determination of vanadium in the presence of titanium. It is based on the orange coloration which is developed when an acid solution of vanadic sulphate is brought in contact with strychnine. The solution is at first coloured deep violet ; this gradually changes to an intense orange. The intensity of the orange coloration depends upon the amount of vanadium in the solution. The process gives fairly satisfactory hydroxide, and 2 c.c. of nitric acid (1 : 1) per 100 c.c. of solution, followed by, say, 20 c.c. of a 20 per cent, solution of lead nitrate, lead chromate alone is precipitated from the cold solution containing chromium, vanadium, and uranium, according to A. A. Noyes, W. C. Bray, and E. B. Spear (Tech. Quart., 21. 14, 1908; Journ. Amer. Chem. Soc., 30. 481, 1908; R. Cain, Journ, Ind. Eng. Chem., 4. 17, 1912), only O'QOOl-0'0003 grm. of chromium remains in solution, and no precipitate was obtained with solutions containing O'l grm. of vanadium. Lead vanadate is sparingly soluble and can be precipitated quantitatively in presence of a weaker acid acetic acid and ammonium acetate. Lead uranate is precipitated from neutral or slightly alkaline solutions. If the nitrate is afterwards required, the lead is removed with hydrogen sulphide. 1 The ether and hydrogen peroxide test for chromium can be made. G. Werther, Journ. prakt. Chem. (1), 83. 195, 1861. According to F. H. Storer, this reaction detects 1 part of potassium chromate in 40,000 parts of water. W. J. Karslake, Journ. Amer. Chem. Soc., 31. 250, 1909 ; C. Reichard, Zeit. anal. Chem., 40. 577, 1901 ; M. Martinon, Bull. Soc. Chim. (2), 45. 862, 1886. 2 R. Holverscheit, Ueber die quantitative Bestimmung des Vanadins und die Trennung der Vanadinsdure von Phosphorsdure, Berlin, 21, 1890. 3 V. von Klecki, Zeit. anorg. Chem., 5. 374, 1894 ; L. Maillard, Bull. Soc. Chim. (3), 23. 422, 559, 1900; Chem. News, 82. 19, 1900; L. C. Barreswil, Compt. Rend., 16. 1085, 1843 ; G. Worthier, Journ. prakt. Chem. (1), 83. 195, 1861 ; C. Reichard, Zeit. anal. Chem., 40. 577, 1901 ; 42. 95, 293, 1903. 4 A. W. Gregory, Chem. News, 100. 221, 1909 ; J. R. Cain and J. C. Hostetter, Journ. Ind. Eng. Chem., 4. 250, 1912. The lower oxides of vanadium do not give the colour test. For other colour reactions see L. Levy, Compt. Rend., 103. 1195, 1886 ; C. Matignon, ib., 138. 82, 1904 ; C. Reichard, I.e. ; V. von Klecki, I.e. ; C. Laar, Ber., 15. 2086, 1882 ; 0. N. Witz, Dingier 's Journ., 250. 271, 1883 ; Bull. Soc. Chim. (2), 45. 309, 1886. DETERMINATION OP CHROMIUM, VANADIUM, AND URANIUM. 485 results in the presence of titanium, molybdenum, and tungsten, but it is not satisfactory in the presence of iron. Of course the iron can be removed by fusion with sodium carbonate, and extraction with water. The use of concen- trated sulphuric acid is a very unpleasant feature in this process. The following volumetric method, due to Cain and Hostetter, enables a vanadium determina- tion to be made in about half an hour. The process is based on the fact that vanadic acid can be quantitatively precipitated by ammonium phosphomolybdate, and the vanadium reduced to the quadrivalent condition by hydrogen peroxide without interference from the associated iron and molybdenum. The solution can then be titrated with permanganate in the usual manner. The solution under investigation in the Erlenmeyer's flask containing possibly iron, titanium, chromium, vanadium, manganese, etc. is heated to boiling and treated with ammonia ; the precipitate is washed once or twice by decantation, and dissolved in nitric acid (sp. gr. 1'135). Boil the solution until it is free from fumes, and oxidise it with potassium permanganate. Dissolve the pre- cipitate by treatment with sodium sulphite, and boil the mixture until it is free from nitrous fumes. Nearly neutralise the solution with ammonia, and add an amount of a solution of sodium phosphate l equivalent to ten times as much phosphorus as there is vanadium present. Heat the solution to boiling, and add sufficient ammonium molybdate (page 595) to precipitate all the phosphorus added as sodium phosphate. Agitate the solution for about a minute. The precipitate, which settles rapidly, is washed three times by decantation with hot (80) acid ammonium sulphate solution. 2 The washing liquid is filtered by suction through an asbestos-packed Gooch's crucible. The last washing should be decanted as completely as possible from the precipitate in the flask, and the filter sucked dry. Change the filtration flask, and draw hot concentrated sulphuric acid through the Gooch's crucible by suction. This dissolves the precipitate. Transfer the sulphuric acid solution to the flask in which the precipitation was made. Wash the filtration flask with concentrated sulphuric acid. Every 0*01 grm. of phosphorus requires a final volume of 5 to 8 c.c. of the sulphuric acid. Heat the contents of the flask until the precipitate is all dissolved, add a few drops of nitric acid (1 : 25), and in two or three minutes, when the fumes are coming off vigorously, remove the flask from the hot plate. When cold, add 3 per cent, hydrogen peroxide solution, in small quantities at a time, with vigorous shaking after each addition, until the solution assumes a deep brown colour owing to the action of the peroxide on the molybdate. The brown colora- tion gives way to a clear green or blue. 3 Put the flask on the hot plate, and when the solution has been fuming four or five minutes, cool, add sufficient water to give an acidity 1 to 5 by volume, 4 and titrate at a temperature between 70 and 80, as indicated in 252, page 481. It will be remembered that, if appreciable amounts of vanadium are present, the colorimetric process for titanium will be erroneous. The volu- metric process is not sensitive enough for the traces of vanadium usually found in British fireclays, unless relatively large quantities are taken for analysis. ' ti 1 SODIUM PHOSPHATE SOLUTION. Dissolve 124 grams of Na 2 HP0 4 .12H 2 in a litre of water. One c.c. will precipitate O'OOl grm. of vanadium, and it is equivalent to nearly '01 3 grm. of phosphorus. G. P. Baxter, Amer. Chem. Journ., 28. 301, 1902. 2 ACID AMMONIUM SULPHATE SOLUTION. Mix 15 c.c. of aqueous ammonia (sp. gr. 0'9)^j \ with 1000 c.c. of water and 25 c.c. of sulphuric acid (sp. gr. 1'84). 3 If the vanadium be not reduced, traces of nitric acid are probably present in the solution,-' ' since this acid readily oxidises quadrivalent vanadium. The nitric acid must be removed by f Q the fuming process. 4 If the solution has a greater acidity than 1 :2, the end point will be uncertain because of the deep yellow colour which quinquevalent vanadium imparts to concentrated sulphuric acid. A dilution of 1 : 5 gives a sharp end point. 486 A TREATISE ON CHEMICAL ANALYSIS. 255. The Simultaneous Determination of Small Quantities of Titanium and Vanadium colorimetrically. Principle of the Method. 1 When dilute acidified solutions of titanium and of vanadium sulphates, coloured by the addition of hydrogen peroxide, are measured in Lovibond's tintometer, and the results with the red rays plotted on squared paper, curves resembling two of the three shown in fig. 155 are obtained. If titanium solutions are coloured with Fenton's acid (p. 468) 2 di- hydroxymaleic acid, C 4 H 4 6 . 2H 2 the curve, also shown in fig. 155, is produced. 3 The curves for the red rays in duplicate experiments are quite concordant, and the readings for the curves fig. 155 do not deviate more than 0'02 on Lovibond's scale. With practice, too, the variations can be reduced to 0-01. The yellow rays are more difficult to manage, because the eye requires FIG. 155. Intensity of red rays in solutions of vanadium and titanium coloured with hydrogen peroxide and Fenton's acid (1" trough). more practice to detect differences of the yellow ray. Hence it is best to confine the attention to the intensity of the red ray. The intensities of the red and yellow tints of mixed solutions of titanium are directly proportional to the amounts of each of these elements present in the solution. The coloration of solutions of titanium sulphate with Fenton's acid is not affected by the presence of vanadium ; consequently, the intensity of the colour of a given solution of titanium with this acid enables the amount of titanium to be computed graphically from the charts drawn on a larger scale than fig. 155 or from the equation of the curve : y=151-2(k (1) 1 H. J. H. Fenton, Journ. Chem. Soc., 9. 1064, 1908 ; J. W. Mellor, Trans. Eng. Cer. Soc.. 12. 18, 1912. 2 For the coloration of titanium solutions with phenols, naphthols, phenolcarboxylic acids, see 0. Hauser and A. Lewite, er., 45. 2480, 1912 ; V. Lehner and W. G. Crawford, Inter. Cong. App. Chem., 8. 285, 1912 ; Journ. Amer. Chem. Soc., 35. 138, 1913. 3 With more concentrated solutions the curves take on the exponential form y ae~ bx and the error is greater. DETERMINATION OF CHROMIUM, VANADIUM, AND URANIUM. 487 where y denotes the intensity of the red ray on Lovibond's scale (with the "inch " trough), and x the amount of titanium oxide, Ti0 2 , expressed in grams, per litre of solution. It is very important to keep the conditions constant, because the tint is liable to vary with small modifications. This is the weakest stage of the process. When the amount of titanium in the mixture has been so determined, the corresponding value of the red ray on Lovibond's scale can be computed for the hydrogen peroxide coloration either graphically (fig. 155) or from the equation of the curve : ^ = 39-03^ . . . . . (2) where x 1 denotes the amount of titanic oxide in grams per litre, and y l the corresponding effect on the red ray on Lovibond's scale. Since the intensity of the red ray for the mixed solution of titanium and vanadium sulphates can be determined directly, the effect due to the titanium can be deducted, and the remainder represents the intensity of the red tint produced by the vanadium. The corresponding amount of vanadium can be computed graphically (fig. 155) or from the equation of the curve : , . (3) where y 9 denotes the reading on Lovibond's scale for a solution containing x 2 grams of vanadium per litre coloured by hydrogen peroxide. 1 Conduct of the Analysis, During the analysis of a silicate say a fireclay the vanadium and titanium oxides will be found with the iron and aluminium oxides in the ammonia precipitate. This is washed, calcined, and weighed in the usual way page 183. The mass is then fused with eight times its weight of sodium pyrosulphate, and the resulting cake, when cold, is taken up with water (page 186) and treated by Gooch's process (page 208) or by the ether process (page 456) to remove the iron. 2 The solution is acidified with sulphuric acid, 3 and an aliquot portion say, one-half is treated with hydrogen peroxide as indicated on pages 204-6. The colour of the solution is measured in Lovibond's tintometer (page 84). The remaining aliquot portion of the solution is treated in a similar way, but about 0'2 gram of Fenton's acid 4 is added instead of hydrogen peroxide. Let the solution stand about an hour, 5 and then read its colour in Lovibond's tintometer. EXAMPLE. The "ammonia" precipitate furnished by a gram of fireclay was treated as described above, and the solution made up to 200 c.c. Half of this solution was treated with Fenton's acid and made up to 200 c.c. This solution gave 1'25 red units with Lovibond's tintometer. From equation (1), when y = 1 '25, x = -00826, hence the solution contained the equivalent of 0'00826 gram of titanic oxide per litre ; or 0'0033 gram per 400 c.c., or 0'0033 gram of titanic oxide per gram of clay. Again, from equation (2), when x = 0^ = '00826, ^ = 0-32 : hence, since the combined titanium and vanadium 1 The constants in fhe equation naturally change if a different observation trough is used, and they possibly also vary a little with different observers. The method of computing the constants for the equations is described in J. W. Mellor, Higher Mathematics for Students of Chemistry and Physics, London, 322, 1909. There is no need to use the equations it caret graphs are drawn, of course on a larger scale than fig. 155. 2 The iron oxide can be determined by dissolving it in acid, etc. , as usual. 3 If the solution be not distinctly acid, the vanadium does not develop its proper colour ; m fact, the strength of the acid can be so reduced that the colour shows with titanium but r with vanadium. T , 4 The acid is not yet on the market, but it will no doubt soon find its way in commerce. is best to add the solid as described in the text ; if much more acid be used in one experimer than in another, the tints will be slightly different. 5 The colour develops gradually unless the solution is warmed a little. The rate of change of the colour after standing an hour is slow enough not to interfere with the accuracy of the work. 488 A TREATISE ON CHEMICAL ANALYSIS. give 0'36 red units, 0-36- ?/! = 0'36 -0-32 = 1/2 = 0-04. When ?/ 2 = 0'04, ic 2 =0'00077 from (3). Consequently, if 1000 c.c. of the solution contain 0-00077 gram of vanadic oxide, 400 c.c., corresponding with 1 gram of clay, contain 0*0003 gram of vanadic oxide. Hence, the clay contained : Vanadic oxide (V 2 6) .... 0'03 per cent. Titanic oxide (Ti0 2 ) .... 0'33 per cent. 256. The Separation of Uranium. 1 The filtrate from the silica ; the pyrosulphate fusion ; or from the hydrogen sulphide precipitation 2 is evaporated to about 100 c.c. and treated with an excess of ammonia, 3 ammonium carbonate, and afterwards with ammonium sulphide. 4 The solution is allowed to stand bvernight in a corked flask, filtered, and washed with water containing ammonium carbonate and sulphide in solution. The precipitate contains the beryllium, titanium, zirconium, aluminium hydroxides, 5 and the hydroxides of the rare earths, also the cobalt, zinc, ferric, and manganese sulphides, if present. If much alkalies, alkaline earths, iron, or manganese be present, dissolve the precipitate in acid and repeat the precipitation. Collect the filtrates together. The filtrate contains all the uranium (possibly as uranyl ammonium carbonate), along with alkalies, alkaline earths, etc. Separation of Rare Earths and Calcium from Uranium.* If the rare earths be present, some may be found with the ammonium uranate. In that case, the precipitate from the ammonium sulphide is dissolved in nitric acid, and the solution evaporated to dryness. Add water ; heat the solution to boiling ; and, while boiling, add oxalic acid and a few drops of ammonium oxalate. Wash the precipitated oxalates with a weak solution of oxalic acid. The precipitate is dissolved in nitric acid ; and reserved for further examination of the rare earths. 7 The filtrate is evaporated to dryness, and ignited to destroy the oxalates, re- 1 F. Pisani, Compt. Rend., 52. 106, 1861; C. Friedel and E. Cumenge, ib., 128. 532, 1899 ; Amer. J. Science (4), 10. 135, 1900 ; G. Edgar, ib. (4), 26. 79, 1908 ; (4), 25. 332, 1908 ; H. P. Foullon, Jahrb. K. K. Geol. Reichanst., 33. 23, 1887 ; C. Rammelsberg, Chem. Centr. (3), 15. 806, 1884 ; C. Zimmermann, Liebig's Ann., 213. 285, 1882 ; G Alibigoff, ib., 233. 117, 143, 1886 ; Zeit. anal. Chem., 26. 632, 1887 ; H. Weber, ib., 44. 420, 1905 ; A. Remele, ib., 4. 379, 1865 ; 26. 631, 1888 ; R. Fresenius and E. Hintz, ib. t 34. 437, 1895 ; A. Borntrager, ib., 37. 436, 1898 ; C. Winkler, ib., 8. 357, 1869 ; Chem. News, 43. 153, 1881 : A. Guyard, ib., 10, 13, 1864 ; A. C. Langmuir, ib., 84. 224, 1901 ; Journ. Amer. Chem. Soc., 22. 102, 1900 ; A. N. Finn, ib., 28. 1443, 1906 ; Chem. News, 95. 17, 1907 ; P. Fritchle, ib., 82. 258, 1900 ; Eng. Min. Journ., 70. 548, 1900 ; W. Gibbs, Zeit. anal. Chem., 12. 310, 1873 ; Amer. J. Science, (2), 39. 62, 1865 ; A. Patera, ib. (4), 16. 229, 1903; Dingier s Journ., 180. 242, 1866 ; Zeit. anal. Chem., 5. 228, 1866; W. Trautmann, Zeit. angew. Chem., 24. 61, 1911. 2 Boiled to expel hydrogen sulphide, filtered to get rid of sulphur, and oxidised with nitric acid. For the hydrogen sulphide precipitation, the acidity should correspond with about 1 c.c. concentrated hydrochloric acid (sp. gr. 1'20), or 1 c.c. of concentrated nitric acid (sp. gr. 1'42), per 50 c.c. of the solution to get a clean separation from lead, cadmium, copper, etc. If much more acid be present, some lead will remain in solution, as indicated on page 274. 3 According to W. F. Hillebrand (Bull. U.S. Geol. Sur.,jS. 43, 1890), ammonia is necessary, otherwise the earths thrown down on neutralisation of the originally acid solution will be only partially redissolved. 4 H. Rose, Pogg. Ann., 96. 352, 1865 ; Chem. News, 7. 159, 1863 ; 8. 99, 1863 ; Zeit. anal. Chem., I. 410, 1862; W. Trautmann, Zeit. angew. Chem., 24. 61, 1911; 25. 19, 1912. H. Boetticher (ib., 43. 99, 1903) deals with the separation of metals precipitated with ammonium sulphide. For the basic acetate separation, see H. Rheineck, CJiem. News 24. 233, 1871 ; A. Remele, Chem. Neivs, 10. 123, 158, 1864. 5 G. Lbsekann (Ber., 12. 56, 1881) deals with the precipitation of aluminium and chromium by alkaline sulphides. See L. Storch (ib , 16. 2014, 1885) for the precipitation of iron by alkaline sulphides, and J. A. Norblad (Bull. Soc. Chim. (2), 23. 64, 1875) for the separation of vanadium. ; W. F. Hillebrand, Bull. U.S. Geol. Bur., 78. 47, 1891. 7 The rare earths are here separated from lime by precipitating with ammonia (page 504). DETERMINATION OF CHROMIUM, VANADIUM, AND URANIUM. 489 dissolved in nitric acid ; after removal of all the nitric acid by repeated evapora- tion with sulphuric acid, the residual sulphates are dissolved in water, and two to three times the bulk of alcohol added. In the course of 1 2 hours, the precipitate will contain the calcium not thrown down by the oxalic treatment for the rare earths. Collect the precipitate on a filter ; wash with alcohol ; dry ; dissolve in dilute nitric acid ; and separate the rare earths from the lime by means of ammonia. 1 The alcoholic filtrate is evaporated to dryness, and ignited. The residue is dissolved in nitric acid and the uranium precipitated by ammonia in the form of ammonium uranate. 2 257. The Gravimetric Determination of Uranium as Uranium Oxide. Evaporate the filtrate containing the uranium to dryness in order to get rid of the greater part of the ammonium carbonate. Acidify the solution with hydrochloric acid and a few drops of nitric acid ; boil to expel all traces of carbon dioxide ; filter off the sulphur which separates ; add a slight excess of ammonia to the boiling solution, 3 when a lemon-yellow voluminous precipitate of ammonium uranate (NH 4 ) 2 U 2 7 .7iH 2 separates. The precipitate rapidly becomes less bulky, darker in colour, and more easy to filter and wash ; but at best, the precipitate filters with difficulty. Wash the precipitate with a 2 per cent, solution of ammonium nitrate, 4 i.e. 2 grms. of salt in 100, c.c. of water. When the precipitation is made with ammonia in the presence of alkalies or alkaline earths, the ammonium uranate is sure to be contaminated with these bases. 5 Hillebrand found that, when precipitated from hot solutions in the presence of ammonium chloride, the ammonium uranate was practically free from alkalies when the precipitation was repeated twice ; and after three pre- cipitations the uranate was practically pure. The precipitate is ignited in a porcelain or platinum crucible with the filter paper, slowly at first so as to burn the paper, then 15 minutes in the blast, and finally, let the precipitate cool slowly in a gradually decreasing flame. The crucible is kept in a slanting position to ensure a free circulation of air. The 1 The calcium oxide so obtained is added to that obtained in the previous precipitation. 2 The uranium oxide finally obtained is not perfectly free from the rare earths, and, in the analysis of some uranium minerals, Hillebrand (I.e.) recovered the uranium oxide amounting to a quarter or a third per cent, of the whole material by treating the dried nitrates with ether ; filtering ; washing with ether ; dissolving the nitrates of the earths containing some uranium in water acidulated with a drop of nitric acid ; adding oxalic acid to the hot solution ; after evaporating to dryness. The earths were added to the main portion. A second precipitation by ether will ensure the removal of the last traces of the earths from the uranium. 3 The ammonia should be free from carbonates. If carbonates be present, the precipitation will be incomplete. Ammonium sulphide may also be employed in place of ammonia. Ammonium uranate is soluble in alkaline carbonates, and also slightly soluble in water ; but it is not soluble in water containing ammonia, ammonium nitrate, or ammonium chloride. The presence of tartaric acid, oxalic acid, and many other non-volatile organic acids prevents the pre- cipitation R. Fresenius. If iron be separated from uranium by ammonium carbonate, the iron is free from uranium, but the uranium contains iron the iron can, however, be removed by fusing the precipitate with potassium bisulphate, dissolving in water, filtering, neutralising the filtrate, and precipitating the iron with ammonium sulphide F. Glaser, Chem. Ztg., 36. 1166, 1912. 4 E. F. Kern (Journ. Amer. Chem. Soc., 23. 685, 1901) obtained the same results by igniting with or without the filter paper. 5 R. Fresenius, Quantitative Chemical Analysis, London, i. 533, 1876 ; W. F. Hillebrand, Amer. J. Science (4), 10. 136, 1900. 4QO A TREATISE ON CHEMICAL ANALYSIS. crucible is finally cooled in the desiccator, and weighed as the dark green or velvety black powder of the oxide UgOg. 1 The nitrates contain lime, magnesia alkalies, and possibly some of the rare earths. 258. The Gravimetric Determination of Uranium as Uranium Phosphate. Owing to the difficulty in obtaining an oxide of constant composition when ammonium uranate is ignited, many prefer to precipitate the uranium as phosphate. 2 The old objection that the precipitate is too slimy to filter and wash with comfort, does not apply if the precipitation be made as described below. Dissolve the oxide in dilute nitric acid. 3 Add ammonia until a pre- cipitate just begins to form. Add just the right amount of nitric acid to clear the solution, and then 2-3 grms. of microcosmic salt, 4 dissolved in a little water, together with 10 grms. of crystalline sodium thiosulphate. 5 The solution becomes yellow, and deposits a voluminous precipitate. Boil about 15 minutes. The precipitate then coagulates and settles rapidly. The precipitate is readily washed by decantation with water containing a little ammonium nitrate. 6 Transfer the precipitate to a Gooch's crucible containing ignited asbestos, wash, and dry. Ignite the precipitate at a red heat for 10 to 20 minutes. If the pre- cipitate be greenish coloured, 7 add a few drops of nitric acid (say, sp. gr. 1'42), dry over the flame, and re-ignite at low redness over a Bunsen's burner. The lemon-yellow compound (U0 2 ) 2 P 2 7 is formed on ignition. Multiply its weight by 1*29 to get the equivalent amount of U 3 8 . The precipitate is somewhat hygroscopic, and should not therefore be needlessly exposed to the air before weighing. In illustration of the results which can be obtained by precipitating uranium nitrate solutions, containing the equivalent of 0*1925 grm. of uranium, with microcosmic salt, Kern gives : 0-1921; 0-1927; 0'1925; 0-1905; 0'1903; 0*1921; 0-1918. 259. The Separation of Uranium as Uranium Ferrocyanide Fresenius and Hintz's Process. It is difficult to separate the members of the hydrogen sulphide group from uranium, since the precipitate arsenic and copper sulphides must be dissolved 1 This usually contains some patches of a yellowish-brown colour. C. Zimmermann (Liebig's Ann., 232. 273, 1886) says that the theoretical U 3 8 is only formed if the calcination be done in a current of oxygen, and W. F. Hillebrand (Bull. U.S. Geol. Sur., 78. 53, 1891) confirms this. However, E. F. Kern (I.e.) considers that the oxide is sufficiently near U 3 8 for all practical requirements, when the ignition is performed as described in the text. Kern also tried igniting the ammonium uranate in a current of hydrogen and weighing as U0 2 , but the results were not so satisfactory as when the precipitate was weighed as U 3 8 . 2 C. Leconte, Pharm Journ., 13. 80, 1854 ; F. Pisani, Ghem. News, 3. 211, 1862 ; Compt. Rend., 52. 106, 1861. 3 Sulphuric acid is used if the precipitate is to be evaluated volumetrically as described later. 4 That is, about ten times the amount actually required to precipitate the uranium. 5 In the event of lime and magnesia being present, acidify the solution with acetic acid ; add 5 c. c. of concentrated ammonium acetate to prevent the precipitation of calcium phosphate. Lime is determined in the filtrate by neutralising most, but not all, of the free acetic acid, and pre- cipitation as oxalate and estimation as sulphate (H. Brearley, The Analytical Chemistry of Uranium, London, 24, 1903). The magnesia is determined in the filtrate as usual. 6 Test the filtrate for uranium with potassium ferrocyanide. 7 This is always the case if the ignition temperature be rather high. The green precipitate is generally supposed to be U 2 3 . P 2 7 . DETERMINATION OF CHROMIUM, VANADIUM, AND URANIUM. 491 and reprecipitated a number of times to eliminate the uranium. Phosphoric acid too prevents clean separations of iron and uranium. In this special case, Fresenius and Hintz l add an excess of potassium ferro- cyanide to the slightly acid filtrate from the silica, and saturate the liquid with sodium chloride to facilitate the flocculation of the precipitated uranium, copper, and iron ferrocyanides. Otherwise the precipitate is exceedingly difficult to filter and wash. Wash the precipitate by decantation, and afterwards on a filter paper with a solution of sodium chloride. Digest the precipitate in the cold with dilute potassium hydroxide, and when the transformation of the ferro- cyanides to hydroxides is complete, decant the liquid, wash the precipitate with water containing ammonium chloride and ammonia until ferrocyanide can no longer be detected in the acidulated nitrate. Dissolve the precipitate in hydro- chloric acid 2 and, if necessary, concentrate the solution by evaporation. Neutralise the greater part of the acid with ammonia ; add ammonium carbonate to the clear liquid ; filter and wash the ferric hydroxide with water containing a little of the filtrate. Heat the filtrate to expel most of the ammonium carbonate, acidify with hydrochloric acid, and boil the solution so as to dissolve the yellow flocculent precipitate. Remove copper from the solution by passing hydrogen sulphide through the hot liquid. The uranium can be determined in the filtrate as usual page 490 or 491. 260. Belohoubek's Volumetric Process for Uranium. In 1867, Belohoubek 3 devised a method for the volumetric determination of uranium based on the reduction of uranium salts to uranous U0 2 salts by means of zinc in acid solutions, precisely in the same way .that ferric salts are reduced, and, as with iron, the subsequent titration with permanganate to re-oxidise the uranous salts. The reducing action is represented : U0 2 S0 4 + Zn + 2H 2 S0 4 = ZnS0 4 + U(S0 4 ) 2 + 2H 2 0. There is not a general agreement as to the accuracy of the results. Some consider that zinc reduces further than the U0 2 stage, and that the reduced solution must be exposed to the air to re-oxidise it to U0 2 before the perman- ganate titration is made. 4 Kern, however, has shown that the reduction does not proceed beyond the U0 2 stage when sulphuric acid is used, even upon five hours' boiling. 5 The washed precipitate of, say, uranous ammonium phosphate is dissolved in sulphuric acid, and the uranyl sulphate U0 2 S0 4 reduced with zinc, magnesium, or aluminium 6 to uranous sulphate in an excess of sulphuric acid. According to 1 R. Fresenius and E. Hintz, Zeit. anal. Chem., 34. 437, 1895 ; Chem. News, 72. 206, 1895. 2 If any ferrocyanide remains undissolved, repeat the treatment. 3 A. Belohoubek, Journ. prakt. Chem. (1), 99. 231, 1867 ; Zeit. anal. Cliem., 6. 120, 1867 ; 0. Follenius, ib., II. 179, 1872 ; C. Zimmermann, Liebig's Ann., 213. 285, 1882 ; F. Ibbotson andS. G. Clarke, Chem. News, 103. 146, 1911 ; E. deM. Campbell and C. E. Griffin, Journ. Ind. Eng. (Jliem., i. 455, 1910. ' 4 0. S. Pullman, Amer. J. Science (4), 16. 229, 1903 ; H. M. Goettsch, Journ. Amer. Chem. Soc., 28. 1541, 1906 ; H. N. M'Coy and H. H. Bunzel, ib., 31. 367, 1909. 5 E. F. Kern, Journ. Amer. Chem. Soc., 23. 685, 1901 ; Chem. News, 84. 224, 236, 250, 260, 271, 283, 1901 ; E. deM. Campbell and C. E. Griffin, ib., 101. 7, 1910 ; Journ. Ind. Eng. Chem., 1. 661, 1909. tt According to Kern (I.e.), uranyl salts are not reduced by hydrogen sulphide. Hence, there is a probability that iron and uranium could be determined in one solution by consecutive reductions with hydrogen sulphide and zinc. The results with zinc, magnesium, and aluminium are the same. Magnesium reduces fastest, zinc slowest. The reductor also gives satisfactory results. Kern did not get satisfactory results with stannous salts. Hydrochloric acid is also objectionable. Nitric acid must, of course, be absent. 492 A TREATISE ON CHEMICAL ANALYSIS. Kern, the ratio of the volume of free sulphuric acid to the total volume of the solution should not be less than 1 : 5, nor more than 1 : 6. During the reduction, the colour of the solution passes from yellow to light green, and finally to a green tinged with blue. This colour is retained even when the reduction has been in progress for four hours. Owing to the fact that there is no satisfactory test to determine whether all the uranyl sulphate has been reduced, it is best to let the reduction proceed for about an hour. 1 Then dilute the solution with recently boiled distilled water until it is almost colourless, and then titrate with N- potassium permanganate 2 until a permanent pink blush remains suffused through the solution. The oxidation by the permanganate is represented by the equation : 5U(S0 4 ) 2 + 2KMn0 4 + 2H 2 S0 4 + 2H 2 0->5U0 2 S0 4 + 2MnS0 4 + 2KHS0 4 + 3H 2 S0 4 . In illustration of the results obtained with solutions of known strength, with zinc reduction, the following numbers might be quoted : 3 Used . . . 015406 0-15464 0'09643 0'19286 0-23050 grm. Found . . . 0-15400 0'15400 "09625 Q'19250 0'23100 grm. Volumetric Determination of Uranium and Vanadium. Sulphur dioxide does not reduce uranyl solutions. Hence, vanadium can be determined volumetrically in presence of uranyl salts. Griffin 4 tried to determine one in presence of the other by first reducing with sulphur dioxide, and titration with permanganate ; and then reducing both by zinc, and titrating. By subtraction, the amount of permanganate required for the oxidation of each can be readily found. Difficulties were encountered owing to the fact that, while zinc reduces the vanadic salts, V 2 5 , to vanadous salts, V 2 2 , the latter is practically oxidised by air before one can titrate, no matter how rapidly the work is done. Reduction in the reductor, fig. 104, page 191, gave good results, bat even then an atmosphere of carbon dioxide above the liquid in the flask is needed for good results. The method is otherwise similar in principle to those indicated on page 482 for vanadium and iron, and for vanadium and molybdenum. 5 Evaluation of Commercial Salts. Sodium uranate and uranium oxides, soluble in sulphuric acid, can be treated in a similar manner. There is a difficulty in dissolving the oxide U 3 8 . It dissolves very slowly in sulphuric acid. In that case, a little nitric acid, say 20 c.c. of concentrated acid, facilitates the dis- solution. The solution must then be boiled to drive off the nitric acid, before the permanganate titration, etc. 6 Hillebrand 7 effects solution by heating, say, 0'2 grm. of the oxide in 10 to 15 c.c. of sulphuric acid (free from nitric acid), at 150 to 175, in a sealed glass tube. The dry powder is placed in a thick glass tube, sealed at one end, and 1 Kern says that 1 hr. is sufficient for O'l grm. of uranium, 1| hr. for 0'2 grm. uranium. 2 According to C. E. Griffin (Min. Eng. World, 37. 247, 1912), permanganate solutions not exceeding ^N in strength give a sharper end point than more concentrated solutions. 3 For interference with iron and titanium salts, see V. Auger, Compt. Rend., 155. 647, 1912. In this case, add a large excess of sodium tartrate and reduce with titanous chloride, using nitro- induline as indicator ; then titrate the uranium with a ferric salt in acid solution, using potassium thiocyanate as indicator. 4 C. E. Griffin, Min. Eng. World, 37. 247, 1912. 5 If uranium and vanadium occur together, A. N. Finn (Journ. Amer. Chem. Soc., 28. 1443, 1906) proceeds as follows: Take up the sodium carbonate fusion with water, and treat the acidified solution with sodium carbonate. The filtrate contains uranium, vanadium, etc. Make the solution alkaline with ammonia, and precipitate the uranium as phosphate (page 490). The filtrate contains the vanadium, the precipitate the uranium. Dissolve the latter in sulphuric acid and determine the uranium by the volumetric process (page 491) ; the vanadium in the filtrate is acidified with sulphuric acid, reduced with sulphurous acid, etc., as described page 480. 6 For the rate of oxidation of uranous solutions, see H. N. M'Coy and H. N. Bunzel, Journ. Amer. Chem. Soc., 31. 367, 1909. 7 W. F. Hillebrand, Bull. U.S. Geol. Sur., 78. 90, 1889. DETERMINATION OF CHROMIUM, VANADIUM, AND URANIUM. 493 the sulphuric acid (water 6, acid 1) added through a long-necked funnel in such a way as to keep the open end of the tube dry. When the carbon dioxide has had sufficient time to escape, draw out the open end in the blast blowpipe to a capillary point, and seal oft' the end. After the digestion in a suitable furnace so that there is no risk if the tube bursts, let the tube cool, open by scratching the capillary end with a file and applying a hot glass rod. Wash the contents of the tube into a beaker, reduce with zinc, and titrate with permanganate as described above. 1 261. Operations with Sealed Tubes. Closing One End. A piece of glass tubing between 35 and 40 cm. long and 16-17 mm. internal diameter, with walls over 3 mm. thick is carefully sealed at one end so that the glass is not thickened into a blob. If a blob does form, it can generally be removed by alternately heating and blowing. Tubes ready sealed at one end can be purchased from the dealers. Charging the Tube. The tube is washed (page 36) and dried. The dry powder under investigation is introduced into the tube so that the powder does not touch the side of the tube near the open end. The necessary acid is added by means of a long-stemmed funnel without wetting the sides of the tube near the open end. Sealing the Tube. About 5 cm. of the glass at the open end is very gradually heated by revolving it for several minutes in the smoky flame of a gas blowpipe. The tube is held inclined at an angle of about 45. The blast is gradually turned on and the tube revolved until the glass begins to soften. One end of a glass rod, about 13 cm. long, is pressed against the edges of the glass tube a, fig. 156. The blowpipe flame is reduced to about 8 or 10 cm. length, and directed at a point about 2 or 3 cm. away from the end to which the glass rod is attached. All the time the tube is being heated, it is slowly revolved so that the glass is heated as FIG. 156. Three stages in sealing the tube, uniformly as possible, but not drawn out. 2 The glass begins to thicken, and the inside diameter of the tube contracts (see the photograph, fig. 157). When the inside diameter of this part of the tube has contracted to about 3 mm., the tube is withdrawn from the flame, and the thickened portion is drawn out to form a capillary end 6, fig. 156. In a few seconds, when the capillary has cooled until it is rigid, seal the tip as shown at c, fig. 156. The tube is allowed to cool, capillary end upwards. Heating the Sealed Tube. The sealed tube is then placed in a metal cylinder with a screw cap. The metal cylinder is placed in a cold tube furnace Meyer's, Gattermann's, Volhard's, Habermann's, etc. 3 The tube furnace is gradually heated to the desired temperature. The furnace is placed so that no damage will be done if the tube bursts. The gas is extinguished when the tube has been heated long enough, and the tube allowed to cool. 1 With much stronger acid, Hillebrand obtained green crystals of a complex uranium salt in the sealed tube. 2 Hard glass tubes are not usually thickened, but drawn out at once into a wide capillary about 1J cm. long. The flame is directed at the base of this capillary, and the glass gradually drawn out so as to form a capillary 2 or 3 cm. long. This is sealed otf at the tip. An oxy-coal gas flame gives the best results with hard glass. 3 L. Meyer, Ser., 14. 1087, 1883 ; J. Volhard, Liebig's Ann., 284. 233, 1895 ; J. Haber- niann, Zeit. anal. Chem., 13. 165, 1874. 494 A TREATISE ON CHEMICAL ANALYSIS. Opening the Sealed Tube. When cold, withdraw the tube from the iron casing so that the capillary end, and only the capillary end, projects. If the reaction in the tube is such as to develop gas, place the tip of the capillary in ' the Bunsen's flarne so as to drive out the liquid which condenses there. Then heat the tip until the glass softens. The pressure inside the tube blows a vent- FIG. 157. Sealing the glass tube. hole in the glass. When the pressure is relieved, and not before, withdraw the tube from its metal casing, file a mark on the glass tube about 2 cm. below the base of the capillary, and touch the file mark with the red-hot tip of a drawn- out piece of glass tubing. The tube will crack, and the crack is carried round the tube by means of the hot glass tip. The end of the tube is then removed. The tube is kept nearly horizontal daring this operation, so as to prevent frag- ments of glass dropping into the tube. The contents of the tube are then washed into a beaker, and treated as required. NOTE FOR PAGE 476. Valuation of Chrome Iron Ores. Chrome iron ores are valued on their Cr 2 3 contents. The market value fluctuates. The unit for high-grade, hand-picked, cobbed ores is 50 per cent. Cr 2 3 , with a premium of, say, 2s. per unit per ton over 50 per cent., and a deduction of, say, 2s. per unit below 48 per cent. Cr 2 3 . Second-grade ores are referred to a 45 per cent, unit, with a premium of, say, 2s. per unit up to 48 per cent. Cr 2 3 . Ores carrying between 40 and 45 per cent. Cr 2 3 inclusive are third-grade ores. The standard is 40 per cent. Cr. 2 3 , and a premium or deduction is allowed for Qvery unit above or below this standard. The so-called " concentrates" are divided into first and second grades ; the former are referred to the 50 per cent, unit, and the latter to the 45 per cent, unit, with premiums and deductions as before. Much crude ore is sold for furnace linings for steel furnaces. A minimum of 5 per cent, silica was formerly allowed, but ores with 8 per cent, of silica now find a market in the chromitc brick industry. CHAPTER XXXV. ZIRCONIUM, THORIUM, AND THE RARE EARTHS. 262. The Separation and Detection of Zirconium. MINUTE quantities of zirconium occur in many if not most silicates of the type of Cornish stone. Zirconia is also used as a constituent of certain white enamels and opaque glazes, and in special cases as a high temperature refractory material. In the regular course of a silicate analysis, zirconia appears with the ammonia or ammonium sulphide precipitate, and, if ignored, would be reported as alumina. If the zirconia is to be determined, its weight must be deducted from the weight of the alumina as determined by the method of pages 182 and 210. The Separation of Zirconium and Aluminium from Iron. The oxides are fused with potassium pyrosulphate, and the resulting cake, when cold, is dissolved in water. The solution is nearly neutralised with sodium carbonate and made up to 300-400 c.c. Then an excess (say 5 grms.) of sodium thio- sulphate Na 2 S 2 3 is added and the solution is boiled until the smell of sulphur dioxide is no longer perceptible (one to two hours suffice). The ferric iron, it will be observed, is reduced to the ferrous condition. The precipitation seems to be a result of hydrolysis (pages 181 and 207), and the thiosulphate serves to keep the solution neutral, but does not directly effect the precipitation. The reaction is usually symbolised : 2A1C1 3 + 3Na 2 S 2 3 + 3H 2 = 2A1(OH) 3 + GNaCl + 3S + 3S0 2 . The precipitate may contain sulphur mixed with the hydroxides of aluminium, 1 zirconium, titanium, 2 and thorium, if present. 3 The nitrate contains salts of iron, beryllium, cerium, lanthanium, didymium, etc. The precipitate is washed, dissolved in dilute sulphuric acid, and filtered from the sulph'ur, if necessary. This method gives good separations of zirconium and aluminium from iron. The Separation of Zirconium from Aluminium. Zirconium (with thorium, if present) can be separated from aluminium as a sparingly soluble basic zirconium 1 Chancel's process for the separation of aluminium from iron (M. F. Chancel, Compt. Rend., 46. 987, 1858 ; F. Stromeyer, Liebig's Ann., 113. 127, 1860) ; P. T. Cleve, Zeit. Kryst., 16. 362, 1890 ; R. Hermann, Journ. prakt. Chem. (1), 97. 330, 1866 ; E. W. Parnell, Chem. News, 21. 54, 1870 ; J. E. Stead, ib., 63. 11, 172, 1891 ; Journ. Soc. Chem. Ind., 8. 965, 1889 ; A. Carnot, Chem. News, 64. 73, 1891 ; Compt. Rend., in. 914, 1890; H. Lasne : ib., 121. 63, 1896 ; Bull. Soc. Chim. (3), 15. 118, 1896 ; E. Donath and R. Jeller, Rep. anal. Chem., 7. 35, 1887). W. Gibbs (Zeit. anal. Chem., 3. 389, 1864) and A. Zimmermann (Ueber die Trennung dcr Thonerde und der Beryllerde, Berlin, 1887) have shown that the separation is not complete unless the boiling be very prolonged. 2 F: Stromeyer (Liebig's Ann., 113. 127, 1860) separates zirconium and titanium from iron by this process. M. Dittrich and S. Freund, Zeit. anorg. Chem., 56. 337, 1907 ; R. Hermann, Journ. prakt. Chem. (1), 97. 330, 1866. J. T. Norton (Amer. J. Science (4), 12. 115, 1901 ; Chem. News, 84. 254, 261, 1901) discusses the influence of pressure on precipitations by sodium thiosulphate. 3 If thorium should be present, it must be separated by the oxalic process described later. 495 496 A TREATISE ON CHEMICAL ANALYSIS. iodate in neutral or feebly acid solutions. 1 Suppose the solution has the equivalent of O'l grm. of zirconium per 100 c.c. in a solution acidified with hydrochloric acid. Add sodium carbonate to the solution until a faint turbidity appears when the solution is thoroughly agitated. Add the smallest amount of dilute hydrochloric acid necessary to just give a clear solution, and then an excess of sodium iodate. Heat the mixture for about 15 minutes, and then let it stand for about 12 hours. Filter off the precipitated iodate, and wash with boiling water. Dissolve it in hydrochloric acid, and precipitate the zirconium hydroxide by the addition of ammonia. 2 The precipitate is washed, etc., as if it were aluminium hydroxide (page 182), and finally weighed as zirconium oxide, Zr0 2 . The aluminium in the filtrate from the basic zirconium iodate is determined by the ammonia precipitation in the usual manner. The Separation of Zirconium from Titanium, Aluminium, and Iron. 3 There is no particular difficulty in separating zirconium in the absence of titanium, nor in separating titanium in the absence of zirconium, but together the individuality of the elements, as Crookes 4 expresses it, is destroyed. The reactions these elements undergo when they occur alone do not necessarily occur when the two elements are mixed together. For instance, titanium hydroxide is completely precipitated when a dilute solution of titanium sulphate is boiled for some time ; but if zirconium be present, there may be either an incomplete precipitation or none at all. The phenomenon is by no means uncommon : aluminium hydroxide, for example, is dissolved by ammonium carbonate solutions in appreciable quantities when beryllium hydroxide is present, but alone, the alumina is scarcely affected. 5 A similar observation was made respecting niobium and titanium on page 420. A fairly concentrated solution of hydrogen peroxide (say, 30 per cent., or a " 100 vol." solution) gives no precipitate with a solution of zirconium nitrate or oxy chloride neutral or acidified with sulphuric acid. On the other hand, a white, flocculent precipitate is obtained with solutions of zirconium sulphate. The precipitate is said to be a hydrated peroxide Zr 2 5 . 4H 2 0. If the solution be agitated and warmed, and the hydrogen peroxide is present in great excess, the reaction is quantitative j but if the hydrogen peroxide be in dilute solution, the separation is very slow. The separation is best performed in a feebly acid solution, for if too much acid be present, no precipitation occurs. The exact proportions have not been worked out, but the precipitate is practically insoluble in 1 per cent, solutions of sulphuric acid. 6 Bailey considers that this reaction furnishes a convenient means of separating zirconium from titanium, iron, aluminium, niobium, tantalum, thorium, yttrium, cerium, etc. The method gives good results only under very special conditions, and it is accordingly rarely used. When sodium phosphate Na 2 HP0 4 is also added to the solution just after the addition of the hydrogen peroxide, Hillebrand has shown that a white, slimy, voluminous precipitate of zirconium 1 J. T. Davis, Journ. Amer. Chem. Soc., II. 26, 1889 ; R. J. Meyer and M. Speter, Chem. Ztg., 34 306, 1910 ; R. J. Meyer, Zeit. anorg. Chem., 71. 65, 1911. Cerium and yttrium iodatesare not very soluble in water, but they are readily soluble in nitric acid. Zirconium and thorium iodates are not soluble in nitric acid provided an excess of alkaline iodate be present. Accordingly, thorium, the rare earths, titanium, and iron may give precipitates in feebly acid solutions. 2 There is too much risk of loss by decrepitation to render it safe to calcine the iodate. 3 M. Dittrich and R. Pohl (Zeit. anorg. Chem., 43. 236, 1905) precipitate the titanium and zirconium hydroxides ; wash, etc., and weigh. The titanium is determined colorimefrically, and the zirconium by difference. 4 W. Crookes, Select Methods in Chemical Analysis, London, 139, 1905. 5 C. A. Joy, Amer. J. ^Science (2), 36. 83, 1863 ; Chem. Neivs, 8. 183, 197, 1863. 6 Hydrochloric acid dissolves the precipitate with the liberation of chlorine. See W. Biltz and W. Mecklenburg, Zeit. angew. Chem., 25. 2110, 1912. ZIRCONIUM, THORIUM, AND THE RARE EARTHS. 497 phosphate possibly Zr 5 H 4 (P0 4 ) 8 . 6H 2 separates. The precipitate is not soluble in water, but it is readily soluble in dilute acids. The precipitate can be purified from titanium and rare earths by pouring dilute hydrofluoric acid (1:5) several times through the filter paper supported in a celluloid or rubber funnel, and collecting the runnings in a platinum dish. The rare earth phos- phates are not soluble in hydrofluoric acid. Evaporate the solution with a few drops of sulphuric acid, and, when the hydrofluoric acid is expelled, repeat the operation with hydrogen peroxide and sodium phosphate. The Turmeric Test for Zirconium. Dissolve the purified zirconium phosphate in a little dilute hydrofluoric acid (1 : 5), evaporate the solution almost to dryness, and dip a piece of turmeric paper into it. Dry the paper in a test tube in a steam oven (about 100). A pink coloration represents zirconium. The presence of O'OOOl grrn. of zirconium is nearly the limit of the test. 1 Titanium gives a similar coloration, and it is even more sensitive than zirconium. Greater quantities of titanium give a brown coloration, 2 but not if the titanium be first reduced to titanium oxide 3 by means of zinc in an acid solution. In the latter case, the titanium may be oxidised as the paper dries, and so develop the titanic acid coloration. Blank tests with hydrofluoric acid are not always satisfactory, and the turmeric test is not therefore very reliable except under favourable conditions. 263. The Gravimetric Determination of Zirconium as Phosphate. Decomposition of the Silicate. Two grams of the dry powder are fused with from 12 to 15 grms. of sodium carbonate. 4 Digest the cold mass with water. When the fused cake is quite disintegrated, add a drop or two of alcohol or hydrogen peroxide to reduce the sodium manganate. Wash the disintegrated residue with a 2 per cent, solution of sodium carbonate. 5 The filtrate contains chlorine, fluorine, sulphur, silica, alumina, etc. 6 The residue on the filter paper is washed into a beaker by projecting a strong jet of warm (70) dilute sulphuric acid (1 : 20) 7 against all parts of the filter paper. About 50 c.c. of the acidulated water should suffice. Add two or three drops of sulphurous acid to dissolve any brown manganese hydroxide which may be present. Warm the solution to about 70. When all effervescence has ceased, add a few drops 1 B. E. Schlesinger, A Systematic Procedure for the Qualitative Detection of the Rare Metals of Certain Groups, Boston, Mass., 1902; C. R. Bohm, Die Darstelhtng der seltenen Erden, Leipzig, 1905; P. E. Browning, Introduction to the Rarer Elements, New York, 1908; J. Only, Analysis, Detection, and Commercial Value of the Rare Metals, Denver, 1903 ; P. Truchot, Les Terres Rares, Paris, 1898 ; J. Herzfeld and 0. Korn, Chemie der selten Erden, Berlin, 1901 ; J. von Panayeff, Verhalten der Wichtigsten seltenen Erden zu Reagentien, Halle a. S., 1909; R. J. Meyer and 0. Hauser, Die Analyse der seltenen Erden und Erdsduren t Stuttgart, 1912 ; W. Crookes, Select Methods in Chemical Analysis, London, 1905. 2 Boric, niobic, tantalic, and molybdic oxides also give the brown coloration with turmeric paper. The paper reddened with boric acid becomes blue when touched with a dilute solution of potash, while the reddening with zirconia is only affected in the ordinary manner. K. Kraut, Zeit. anal. Chem., 4. 168, 1865 ; A. M tiller, Joum. prakt. Chem. (1), 80. 118, 1860. 3 F. Pisani, Compt. Rend., 59. 301, 1864 ; Chem. News, 10. 218, 1864. 4 If sulphur is to be determined, add between 1 and 2 grms. of sodium nitrite and note the precautions stated on page 617. 5 If water be employed, the later washings sometimes run through turbid. The sodium carbonate should be free from bicarbonate. 6 If desired, the filtrate can be made up to, say, 150 c.c., and equal portions taken for the determination of fluorine, sulphur, and chlorine. But 50 c.c., that is, one-third of 2 grms., may not contain enough for a determination of each of these materials. 7 M. Wunder and B. Jeanneret (Zeit. angew. Chem., 50. 733, 1911) remove the iron from the residue by extraction with hot hydrochloric acid (1:1); but L. Weiss and W. Trautmann (Zeit. anal. Chem., 51. 303, 1912) say that the sodium zirconate, Na 2 Zr0 3 , is appreciably soluble in this liquid. 32 498 A TREATISE ON CHEMICAL ANALYSIS. more of the acidulated water to make sure that all action is over. 1 The liquid should be distinctly acid. 2 Filter the liquid into an Erlenmeyer's flask (150 c.c.), 3 and wash with warm (70) acidulated water. The nitrate contains the iron, aluminium, and most of the zirconium. Some zirconium remains in the residue along with the silica and any barium sulphate which might be present. To recover this zirconium, ignite the filter paper and contents in a platinum crucible, and treat the mass with hydrofluoric and sulphuric acids in order to expel the silica, which holds back some of the zirconium very tenaciously. Take up the residue with hot dilute sulphuric acid (1 : 20). Barium sulphate remains insoluble (page 618). The filtrate contains titanium, aluminium, iron, zirconium, rare earths, etc. Combine the two filtrates. Precipitate the zirconium by the hydrogen peroxide and sodium phosphate process described below. The sequence of the operations employed for separating the zirconium from an insoluble silicate is therefore represented by the scheme : c A , C Soluble in water . Chlorine, sulphur, etc. bod. carb J , Soluble Na H P0 4 . . Titanium, etc. fusion. | Inso i uble ( SolubleH 2 S 4{ i ns oluble Na 2 HP0 4 . . Zirconium. ( Insoluble in H 2 S0 4 Barium, etc. Precipitation of Zirconium Phosphate* Add sufficient (say, 5 c.c.) of a concen- trated solution of hydrogen peroxide to the filtrate to maintain a permanent yellow coloration. The object of the hydrogen peroxide is to keep the titanium 5 in solution ; and the acidity prevents the precipitation of the aluminium, iron, etc. Add about 1 c.c. of a saturated solution of sodium phosphate Na 2 HP0 4 . Fill the flask nearly up to the neck with water, and set it aside in a cool place for a couple of days. If the yellow colour at any time fades away, add more hydrogen per- oxide. Gelatinous zirconium phosphate is precipitated. Filter through, say, a 7-cm. filter paper, and wash with water containing a few drops of dilute sulphuric acid. A little titanium phosphate and rare earth phosphates may be carried down with the precipitate. In the former case, the precipitate will probably have a yellow tint. The filtrate is reserved for the determination of the rare earths (page 502). Purification of the Zirconium Phosphate. Ignite the precipitate and filter paper in a platinum crucible ; fuse with a little sodium carbonate ; extract the cold mass with water ; filter, wash as before, ignite the filter paper and precipitate ; fuse with a little potassium pyrosulphate ; dissolve the cold mass in hot water ; acidify the solution with a couple of drops of sulphuric acid ; and repeat the treatment with hydrogen peroxide and sodium phosphate. Weigh the ignited precipitate as Na. 2 0.4Zr0 2 .3P 2 5 . Hence, multiply the weight of the pre- cipitate by 0*5, or more strictly by 0'5014, to get the corresponding amount of zirconia Zr0 2 . 6 The weight of the zirconia, if appreciable, is to be deducted from the collected ammonia precipitate (page 210). 1 If the digestion be protracted, gelatinous silica may separate and give trouble in the subsequent nitration. 2 An excess of sulphuric acid hinders the precipitation of the zirconium phosphate later on, and retards nitration by closing the pores of the filter paper. Not more than 1 per cent, of free H 2 S0 4 should be present in the solution. 3 Filter through the paper used in the previous filtration. 4 G. H. Bailey, Journ. Chem. Soc., 49. 149, 481, 1886 ; Chem. News, 53. 55, 260, 287, 1886; W. F. Hillebrand, Bull. U.S. Geol. Sur., 176. 75, 1900; M. Dittrich and R. Pohl, Zeit. anorg. Chem., 43. 236, 1905; E. Wedekind, ib., 33. 83, 1903: H. Geisow and P. Horkheimer, ib., 3% 374, 1892. 5 Otherwise titanium phosphate Ti( OH )P0 4 . H 2 would be precipitated V. Merz, Journ. prdkt. Chem. (1), 99. 157, 1856. 6 It is always well to make certain that the precipitate is really zirconia by fusing the mass with a little sodium carbonate. Take up the cold cake with hydrochloric acid ; evaporate the solution until it is nearly dry, and apply the turmeric paper test (page 497). A little thorium, if present in the sample, may even yet be found with the zirconia. ZIRCONIUM, THORIUM, AND THE RARE EARTHS. 499 264. The Analysis of Zircon. Opening up the Mineral. Fuse a gram of the finely powdered and dried (110) mineral with 4-5 grms. of sodium bisulphate. 1 The mineral dissolves with great difficulty, and it is necessary to keep it in a molten condition for some time. 2 Determination of Silica. When cold, the mass is digested with water. The white insoluble silica is filtered off, ignited, and weighed. Drive off the silica from the ignited mass by treatment with hydrofluoric and sulphuric acid in the usual manner (page 169). This gives the "first silica." Fuse the residue with a little sodium bisulphate, and add the aqueous solution of the fused mass to the main solution. Hydrogen sulphide is passed through the combined solutions to remove any platinum derived from the crucible. The filtrate is evaporated twice to dryness, with an intervening filtration for silica (page 167). The silica is removed by treatment with hydrofluoric acid. This gives the "second silica." The weight of silica so obtained is added to the previous result. The residue is fused with a little sodium bisulphate and added to the main solution. Determination of Iron and Alumina. The solution is approximately neutralised with ammonia, and mixed with sufficient ammonium oxalate and tartrate to prevent the precipitation of aluminium, etc., with ammonia. Add ammonium sulphide to the warm solution (page 208). Keep the solution warm for about an hour. 3 Practically all the iron and a little alumina are precipitated. These may be determined in the usual manner. 4 Determination of Zirconia. Evaporate the filtrate from the iron sulphide to dryness in a platinum dish, and ignite the residue until it is almost white. The combustion of the carbon can be facilitated by repeatedly moistening the residue with ammonium nitrate and repeating the ignition. Fuse the white residue with sodium bisulphate. Dissolve the cold mass in water. Add a concentrated solution of hydrogen peroxide (page 496), and follow on with an aqueous solution of sodium hydroxide, added drop by drop, with constant 1 J. M. Matthews, Journ. Amer. Chem. Soc., 20. 815, 1898. The opening up of zircon is very difficult. Alkaline carbonates, recommended by F. Wohler (Licbig's Ann., 31. 122, 1839), P. Berthier (Ann. Chim. Phys. (1), 50. 302, 1832), W. Henneberg (Journ. prakt. Chem. (1), 38. 508, 1846), H. L. Wells and H. W. Foote (Zeit. anorg. Chem,, 10. 434, 1895), L. M. Dennis and A. E. Spencer (Journ. Amer. Chem. Soc., 18. 673, 1896), and C. F. Chandler (Pogg. Ann., 102. 446, 1857), act very slowly (T. Scheerer, ib., 59. 481, 1843). Alkaline hydroxides with or without alkaline fluorides are better E. Linnemann, Monats. Chem., 6. 335, 1885 : Cliem. News, 52. 233, 240, 1885; G. H. Bailey, ib., 53. 55, 260, 287, 1886; 60. 6, 17, 32, 1889; F. P. Venable, Journ. Amer. Chem. Soc., 16. 469, 1894 ; F. P. Venable and W. Belden, ib., 20. 273, 1898; F. P. Venable and T. Clarke, ib., 18. 434, 1896; Chem. Neivs, 74. 44, 1896; C. Marignac, Ann. Chim. Phys. (3), 60. 260, 1860 ; F. Dubois and A. A. da Silveira, ib. (1), 14. Ill, 1820; J. J. Berzelius, Pogg. Ann., 4. 117, 1825; M. Melliss, Bull. Soc. Chim. (2), 14. 204, 1870 ; M. H. Klaproth, Beitrage zur chemischen Kenntniss dcr Mineralkorper, Berlin, I. 203, 227, 1795. For the method indicated in the text see L. Weiss and R. Lehmann, Zeit. anorg. Chem., 65. 178, 1910 ; M. Tcharviani and M. Wunder, Ann. Chim. Anal., 16. 1, 1911 ; M. Wunder and B. Jeanneret, Zeit. anal. Chem., 50. 733, 1911. 2 For the method of conducting bisulphate fusions, see page 185. 3 The iron sulphide which first separates is very finely divided, and inclined to pass through the filter paper. The precipitate coagulates and filters easily after standing for an hour. 4 When the amount of iron is great in comparison with the alumina, a relatively large proportion of the total alumina, according to L. Weiss and R. Lehmann (Zeit. anorg. Chem. , 65. 178, 1909), is precipitated with the iron. If much alumina and little iron be present, most of the alumina will remain in solution. The zircon may have so little alumina that the determination of the alumina with the iron will be sufficiently exact for technical purposes. In that event, the alumina in the analysis will be a little low, and the zirconia a little high, in consequence of the alumina which escapes precipitation with the iron. 500 A TREATISE ON CHEMICAL ANALYSIS. agitation. Zirconium hydroxide separates ; the pertitanic acid remains in solu- tion. 1 Dissolve the precipitate in hydrochloric acid ; add ammonia ; boil off the excess of ammonia ; filter ; wash ; ignite ; and weigh as Zr0 2 . The zirconia so obtained is generally contaminated with both silica and alumina. The former is removed by digesting the mass with hydrofluoric and sulphuric acids 2 and repeated calcination with ammonium carbonate in order to remove the fluorine. 3 The zirconia is now contaminated with alumina. For the separation of zirconium and aluminium oxides, see page 495. Determination of Titanium. The excess of hydrogen peroxide is decomposed by boiling the filtrate. Take care to avoid loss by spurting owing to the de- composition of the hydrogen peroxide. Precipitate the titanium by the addition of ammonia, or apply the colorimetric process. 4 265. The Rare Earths. The term " rare earth " is applied to certain trivalent metallic oxides yttria and ceria earths which were formerly regarded as elementary bodies. They fall into the same analytical group as aluminium, iron, beryllium, etc., from which they can be separated as oxalates insoluble in dilute acids, and in cold ammonium oxalate solutions. The solubilities of the oxalates 5 of some of these earths in water, in solutions of ammonium oxalate, and in sulphuric acid solutions are shown in the following table : Table LXII. Solubilities of the Rare Earth Oxalates. Oxalate. Water. Grm. per litre. Ammonium oxalate. One gram of oxalate in 38 grms. of water dissolves grm. Normal sulphuric acid. Grms. of anhydrous oxalate per litre. Lanthanum . .. Praseodidymium Neodidymium . Cerium . . Yttrium . Ytterbium . . Thorium ... . 0-00062 0-00074 0-00049 0-00041 o-ooioo 0-00334 0-00023 0-00026 0-00034 0-00042 0-00256 0-02437 0-62000 2-56 1-23 I'OO 1-64 0-190 There are some doubts about the identity of many of the rare earths, and the individuality of some has not yet been satisfactorily demonstrated. The following are the more important types : 1. YTTRIUM EARTHS e.g., yttria, erbia, terbia, holmia, thulia, dysprosia, ytterbia (neoytterbia), lutecia, europia, etc. 1 A. Classen, Ber., 21. 370, 1888. ' 2 Zirconium may be volatilised when the silica is removed by treatment with hydrofluoric acid, unless, say, the mass is mixed with 20 times its bulk of sulphuric acid and 45 times its volume of hydrofluoric acid. The zirconia is finally weighed as oxide, Zr0 2 . E. Wedekind, Ber., 44. 1753, 1911. 3 This gives the "third silica," which, added to the " first" and "second" silica, gives the "total silica." 4 The lime and magnesia are determined alkalies by J. L. Smith's process (page 222). earth elements, see the index. 5 E. Rimbach and A. Schubert, Zeit. Chem. Soc., 73. 951, 1898. as indicated on pages 213 and 218, and the For traces of niobium, vanadium, and the rare Chem., 67. 183, 1909 ; B. Brauner, Journ. ZIRCONIUM, THORIUM, AND THE RARE EARTHS. 501 2. CERIUM EARTHS e.g., ceria, lanthana, neodidymia, praseodidymia, samaria, scandia, decipia, gadolinia, etc. Thorium and zirconium earths are not included in the above definition of the rare earths, but in general analytical work it is often convenient to include both thoria and zirconia with the rare earths, because they all have many analogous properties. The Separation of Thoria, Ceria, and the Yttria Earths from the Ammonia Precipitate. In determining the rare earths, it is often convenient to dissolve the ammonia precipitate containing the aluminium, iron, yttrium, cerium, thorium zirconium, uranium, etc., hydroxides or phosphates in the smallest possible quantity of hydrochloric acid; and treat the solution with oxalic acid. The yttrium, cerium, and thorium oxalates are precipitated, while zirconium, aluminium, iron, and uranium 1 remain in solution. 2 The latter can be separated by processes previously described. The chief source of difficulty is the fact that the acidity of the solution requires careful adjustment, or appreciable quantities of the cerium, yttrium, and thorium oxalates will pass into solution, or else the precipitate may be contaminated with zirconium oxalate. The conditions for successful work with ceria 3 have been investigated, but not for yttria. 4 The best results are obtained when the ammonia precipitate is dissolved in JN- to JN- hydrochloric acid about 50 c.c. of acid are required per gram of the earths. Add 40-50 c.c. of a 10 per cent, solution of oxalic acid, and keep the solution at about 60 for about 12 hours. The precipitate is washed with warm water. The precipitate can be calcined to convert the oxalates into oxides, and weighed. If it be desired to separate the thoria, yttria, and ceria, the washed pre- cipitate is digested on a water bath with concentrated nitric acid (sp. gr. 1 -4) and one drop of T ^N-potassium permanganate, which accelerates the rate of dissolu- tion by the nitric acid. The dish should be covered with a clock-glass to prevent loss by spurting. The oxalates will be decomposed in a short time. The excess of acid is removed by evaporation, and the thorium precipitated by the hydrogen peroxide process (page 505), or by the sodium thiosulphate process. Separation of Thorium by Sodium Thiosulphate. The nitrate solution is repeatedly evaporated with hydrochloric acid to convert the nitrates into chlorides (page 427). Dilute the solution and heat it to boiling. Add an excess of a concentrated solution of sodium thiosulphate. Basic thorium thio- sulphate is precipitated in a flocculent mass which, after standing about 12 hours, is easily filtered and washed. The precipitate with its filter paper is digested in concentrated hydrochloric acid, and again treated with the thiosulphate so as to get rid of all but a trace of ceria which contaminated the first precipitate. The precipitate is dissolved in hydrochloric acid, and treated with oxalic acid as described above. The first filtrate contains the yttrium and cerium earths. The hydroxides are precipitated by ammonia, and separated by the double potassium 1 According to 0. Hauser (Zeit. anal. Chem., 47. 677, 1908), the presence of uranyl salts makes the cerium oxalates very soluble, unless a great excess of oxalic acid is added. 2 The precipitated oxalates will be contaminated with phosphates, if phosphorus compounds be present. A reprecipitation may then be necessary ; or the precipitate may be washed into a dish and digested with fuming hydrochloric acid, then warmed with oxalic acid and the whole diluted with water. In this way most of the phosphoric acid passes into solution. 3 E. Hintz and H. Weber, Zeit. anal. Chem., 36. 213, 1897 ; P. Drossbach, Zeit. angew. Chem., 14 655, 1901 ; E. Benz, ib., 15. 297, 1902; 0. Hauser and F. Wirth, ib., 22. 484, 1909 ; C. Glazer, Zeit. anal. Chem., 36. 213, 1897 ; R. J. Meyer and R. Jacoby, ib., 27. 364, 1901; R. Finkener, ib., 3. 369, 1864; N. Engstrom, Zeit. Kryst., 3. 191, 1879; W. Blomstrand, ib., 15. 99, 1889 ; H. Backstrom , i b. , 16. 83, 1890 ; C. Jones, Amer. Chem. Journ., 20. 345, 1898 ; H. Gorceix, Compt. Rend., 100. 357, 1885 ; M. Holzmann, Journ. prakt. CJiem. (1), 75. 321, 1858 ; R. Ruer, Zeit. anorg. Chem., 42. 87, 1904. 4 L. F. Nilson, Ber., 13. 1437, 1880 ; P. T. Cleve and M. Hoglund, ib., 6. 1468, 1873; N. J. Berlin, Pogg. Ann., 43. Ill, 1838. 502 A TREATISE ON CHEMICAL ANALYSIS. sulphate process described on page 508 for separating zirconium and yttrium. The cerium salt, Ce 2 (S0 4 ) 3 . 3K 2 S0 4 , is but sparingly soluble, while the yttrium salt, Y 2 (S0 4 ) 3 .3K 2 S0 4 , is readily soluble. No exact quantitative methods are known for separating the individual members of these groups, and all known methods give more or less approxi- mate analytical separations. This is because the reactions of the different members of the group of rare earths, with the possible exception of cerium, differ among themselves by minute differences ; so much so, that it is difficult to draw a real line of demarcation between the yttrium and cerium earths. In all attempts at separation it is necessary to dissolve and reprecipitate a number of times, and in some cases an elaborate process of fractional crystallisation must be employed. These methods fall outside the range of analytical work, and details must be sought in the original memoirs. Opening the Minerals. The minerals containing the rare earths are opened by the usual methods. 1 When sulphuric and hydrofluoric acid are employed, the fluorine is driven off by heating. If hydrochloric acid be used, the solution is evaporated to dryness to make the silica insoluble. The residue is treated with a little hydrochloric acid, diluted with water, and filtered. If sulphuric acid, or potassium pyrosulphate or bisulphate, be employed for opening the mineral, the mass must be stirred with cold water to get the required solution ; otherwise, titanium hydroxide may be precipitated. If hydrofluoric acid alone be employed in the cold, a gelatinous precipitate of the silico-fluorides and fluorides of the earths is obtained. The precipitate is collected and washed with dilute hydro- fluoric acid. Niobium, tantalum, zirconium, etc., pass into solution. 2 The insoluble fluorides are decomposed by sulphuric acid. Fusion of the mineral with sodium or potassium hydroxides or carbonates, and leaching with water, gives a residue of insoluble hydroxides. 266. The Determination of "Rare Earths" in Silicates. The determination of zirconium in silicates has been discussed on page 498. If rare earths be also present, they will be found in the filtrate from the zirconium. Several methods of separating the rare earths are available. The following process is adapted for silicates containing but small quantities of the rare earths. It is based upon the very sparing solubility of the fluorides of the rare earths and thorium, and the ready solubility of the fluorides of the remaining elements (including zirconium, if present) in dilute hydrofluoric acid. 3 Start with the filtrate from the zirconium phosphate (page 498), which we can suppose contains salts of the rare earths, beryllium, titanium, aluminium, iron, and uranium. Treat the solution with an excess of sodium hydroxide, in which beryllium, uranium, and aluminium hydroxides are soluble, while titanium, iron, and the rare earth hydroxides are almost insoluble. Dilute the solution. Decant the clear through a filter paper ; wash the precipitate twice by decanta- tion ; and then wash the precipitate into a platinum dish. Treat the precipitate with hydrofluoric acid ; evaporate the mixture nearly to dryness ; and add a little water with a few drops of hydrofluoric acid. Collect the insoluble precipitated fluorides on a filter paper, using a rubber funnel and 1 H. B. Hicks (Journ. Amer. Chem. Soc.,33. 1492, 1911) recommends sulphur monochloride vapour for opening the rare earth minerals. Volatile chlorides are formed. This treatment decomposes rutile, wolframite, scheelite, tantalite, chromite, etc. See also E. F. Smith, ib., 22. 289, 1898 ; R. D. Hall, ib., 26. 1243, 1904. 2 M. Delafontaine, Chem. News, 75. 229, 1897. 3 W. F. Hillebrand, Bull. U.S. Geol. Sur., 176. 77, 1900 ; J. J. Chydenius, Pogg. Ann., 119. 49, 1861 ; M. Delafontaine, Chem. News, 75. 229, 1897 ; J. J. Berzelius, Pogg. Ann., 16. 385, 1829 ; A. Rosenheim, V. Samter, and I. Davidsohn, Zeit. anorg. Chem., 35. 424, 1903. ZIRCONIUM, THORIUM, AND THE RARE EARTHS. 503 platinum cone. Wash the precipitate with water acidulated with hydrofluoric acid, and finally transfer the precipitate to a platinum dish. Evaporate the solution to dryness with sulphuric acid, to decompose the fluorides. Ignite the filter paper separately, and add the ash to the contents of the dish. Dissolve the residual sulphates in dilute hydrochloric acid ; precipitate the rare earths with ammonia ; redissolve the precipitate in hydro- chloric acid ; and evaporate to dryness. Treat the residue with a few drops of an aqueous solution of oxalic acid ; filter, and wash the insoluble oxalates. Dry the precipitate ; ignite the oxalates; weigh the resulting oxides ; and report as " rare earths." 267. The Analysis of a Mixture of Rare Earths. The analysis of a " rare earth mineral," and the separation of some of the more important members of the rare earth group, is well illustrated by the following method for the analysis of monazite sands, etc. 1 The analysis occupies about seven days. 1. The Fusion. About a gram 2 of the finely powdered mineral is fused in a platinum crucible with five to six times its weight of sodium pyro- or bisulphate. 3 The temperature is raised very gradually to a red heat so as to avoid loss by spurting. The fusion temperature is kept as low as possible to avoid the forma- tion of sparingly soluble basic sulphates. When all is dissolved, the crucible and contents are allowed to cool. The cold cake is taken up with cold water, and evaporated to dryness on a sand bath. Digest the residue in 10 c.c. of concen- trated sulphuric acid, pour the solution into 200 c.c. of cold water, and boil the mixture in a flask with a reflux condenser for two or three hours. Most of the zirconia and ferric oxide dissolve. Silica, titanic, niobic, and tantalic oxides, 4 etc, 5 are precipitated. Filter. Let the solution stand 24 hours, and re-filter if necessary. Wash the precipitate, which may contain the tungstic, stannic, titanic, tantalic, and niobic oxides and part of the zirconia and ferric oxide. The filtrate contains the rare earth oxides, uranium and thorium oxides, alumina, part of the zirconia and ferric oxides, the alkaline earths, etc. 2. The Separation of Tin, Tungstic, Titanic, Tantalic, and Niobic Acids. Digest the residue with ammonium sulphide so as to remove the stannic and tungstic oxides (page 406). The insoluble mass is treated with a mixture of equal volumes of 10 per cent, sulphuric acid and 3 per cent, hydrogen peroxide. 6 Dilute the filtrate to 200 c.c., add sulphurous acid, and boil in a large flask fitted with a reflux condenser (fig. 128) until a little of the clear gives no titanium coloration when tested with hydrogen peroxide. About two or three hours' boiling suffices to precipitate all the titanium, tantalum, and niobium oxides. The ferric oxide and zirconia remain in solution. 7 Filter off the insoluble residue, and wash. Add the filtrate and washings to the main solution, or determine the iron and zirconium they contain as indicated below. 1 C. Glaser, Journ. Amer. Chem. Soc., 18. 782, 1896 ; Chem. Ztg. t 20. 619, 1896 ; Zeit. anal. Chem., 36. 213, 1897 ; Chem. Neivs, 75. 145, 157, 1897 ; G. Chesneau, Compt. Rend., 153. 429, 1911. 2 More frequently between 10 and 30 grnis. are taken if some of the rarer of the "rare earths " are to be determined by the process here described. 3 The sodium salt is preferred to the potassium salt because the latter forms sparingly soluble double sulphates with some of the oxides under investigation. 4 A trace of tantalic oxide and silica may dissolve in the sodium bisulphate solution. 5 Lead and barium sulphates, if present, remain undissolved ; so also do the stannic and tungstic oxides. 6 L. Weiss and M. Landecker, Zeit. anorg. Chem., 64. 65, 1909. 7 Or, if necessary, fuse with potassium bisulphate. Cold water must be used in dissolving the cake of the bisulphate fusion. 504 A TREATISE ON CHEMICAL ANALYSIS. 3. Separation of Titanic Oxide from Niobic and Tantalic Oxides. The moist residue is boiled three or four hours with a large excess of salicylic acid. The titanic oxide dissolves, while the niobic and tantalic oxides remain insoluble. 1 The nitrate and washings are treated with ammonia, and the titanic hydroxide is filtered off, washed, ignited, and weighed as Ti0 2 . The niobic and tantalic oxides are calcined and weighed. If desired, the two last-named oxides can be separated by Marignac's process (page 421). 4. Separation of the Hydrogen Sulphide Group. Saturate the hot acidified solution with hydrogen sulphide, and, when cold, again saturate the solution with the same gas. The members of the hydrogen sulphide group are precipitated. Filter, and boil the filtrate to expel the hydrogen sulphide. The residue on the filter paper contains the metals of the hydrogen sulphide group copper, bismuth, etc. These, if present, are separated as described on page 319. 5. Separation of the Thorium and Cerium Group from the Zirconium and Yttrium Groups. Thorium oxalate is somewhat soluble in nearly neutral solutions of ammonium oxalate, 2 and if a large excess of ammonium oxalate be used, thorium may not be precipitated at all. But an excess is necessary to keep the zirconium oxalate in solution. Hence, some thorium may escape pre- cipitation, and, if not separated, will, later on, appear with the zirconium precipitate. Glaser therefore recommended adding concentrated hydrochloric acid to the boiling filtrate. On cooling, insoluble thorium oxalate is precipitated. The zirconium oxalate is soluble in the oxalic acid liberated by the action of the hydrochloric acid. 3 The method here employed is based on the solubility of yttrium and zirconium oxalates and the " insolubility " of thorium and cerium oxalates in nearly neutral solutions of ammonium oxalate in the presence of hydrochloric acid. An alter- native process is indicated on page 501. 4 Either evaporate the filtrate and wash- ings from the preceding operation to about 100 c.c. ; or treat the solution with ammonia so as to precipitate the phosphates and hydroxides of the rare earths, thorium, zirconium, aluminium, iron, and manganese. Filter and wash (page 183). Dissolve the precipitate in nitric or hydrochloric acid. The filtrate contains the alkaline earths. Nearly neutralise the solution (sulphates, nitrates, or chlorides) with ammonia, so that the solution has the least possible excess of acid. Boil. Add a large excess of a boiling solution of ammonium oxalate. Dilute the solution to nearly twice its volume with water, and add concentrated hydrochloric acid, drop by drop, until the addition of more acid to the clear solution produces no further precipitation. Let the mixture stand overnight. Filter off the precipitated oxalates, and wash with a solution of ammonium oxalate. 5 The precipitate contains the cerium earths, and the thorium as oxalates. 6 The filtrate 1 M. Dittrich and S. Freund, Zeit. anorg. Chem., 56. 344, 346 1908; 0. Hauser and H. Herzfeld, Zenlr. Min., 759, 1910. 2 J. F. Bahr, Liebig's Ann., 132. 231, 1864. 3 But even with Glaser's improvement the operation is not altogether satisfactory. The thorium and zirconium may be first removed by Metzger's process, and the thorium and zirconium separated by the oxalic process. The method indicated above will usually satisfy commercial requirements. 4 The presence of uranium salts augments the solubility of cerium and lanthanum oxalates, and therefore the precipitation will be imperfect in the presence of uranium salts, unless a large excess of oxalic acid is added 0. Hauser, Zeit. anal. Chem., 47. 665, 1908. 5 Dissolve, say, 2 grms. of solid ammonium oxalate in a little water, and make the solution up to 100 c.c. 6 E. Hintz and H. Weber, Zeit. anal. Chem., 36. 27, 1897 ; P. Drossbach, Zeit. angew. Chem 14. 655, 1901 ; E. Benz, ib., 15. 297, 1902 ; T. Scheerer, Pogg. Ann., 56. 498, 1842 ; 51. 470, 1840; C. Rammelsberg, ib., 108. 48, 1859; H. Rose,#., 118/502, 1863; T. Thomson, Trans. Roy. Soc. Edin.,6. 371, 1811; A. Connell, Edin. Phil. Journ., 20. 300, 1842; R. Hermann, Journ. praJct. Chem. (1), 82. 387, 1861 ; M. Holzmann, ib. (1), 84. 78, 1861 ; ZIRCONIUM, THORIUM, AND THE RARE EARTHS. 505 is reserved for the determination of phosphorus, aluminium, iron, manganese, yttrium, zirconium, calcium, etc. Ignite the mixed precipitates to convert the thorium, cerium, lanthanum, and didymium oxalates into oxides. 1 I. The Ammonium Oxalate Precipitate. 6. Separation of Thorium from Cerium, Lanthanum, and Didymium. This separation is based on the solubility of thorium oxalate, and the " insolubility " of cerium, lanthanum, and didymium oxalates, in a solution of ammonium oxalate containing a little ammonium acetate. A large excess of ammonium acetate should not be used, because cerium oxalate is slightly soluble in that reagent. When an excess of ammonium oxalate is present, a few c.c. of a solution of ammonium acetate will suffice to keep the thorium oxalate in solution. Dissolve the mixed oxides in sulphuric acid, and nearly neutralise the solution with ammonia. Add to the boiling solution an excess of a boiling solution of ammonium oxalate. In a short time, when the oxalates of the cerium earths have formed, and before the liquid has cooled, add a solution of ammonium acetate.' 2 Cerium oxalate will be precipitated, while thorium oxalate remains in solution. Let the solution stand overnight. Filter and wash. Dissolve the precipitate in sulphuric acid, and repeat the treatment, since otherwise some thorium will be precipitated, and, later, contaminate the cerium. 3 7. Determination of Thorium. Add an excess of ammonia to the filtrate, and thorium hydroxide will be precipitated. It is the general rule, in separating the rare earths in the presence of alkaline solutions, to dissolve the hydroxide, and reprecipitate ; since the hydroxides are particularly liable to absorb alkaline salts from the mother liquid. The thorium hydroxide may be purified by dis- solving the precipitate, well washed in boiling water, in nitric acid. Neutralise the solution with ammonia; add 10 per cent of ammonium nitrate 4 per 100 c.c. of solution ; heat the solution between 60 and 80 ; add 10 c.c. (per 50 c.c. of solution) of a 10 per cent, solution of hydrogen peroxide; 5 and boil the solution for a few minutes. A flocculent precipitate of thorium perhydroxide Th(OOH)(OH) 2 separates. This can be easily washed. 6 Test some of the filtered solution by another treatment with an equal volume of hydrogen peroxide and boil. If no precipitate appears, filter the main solution while hot as rapidly as possible; and wash with a hot solution of 5 to 10 per cent, of ammonium T. Lange, ib. (1), 82. 135, 1861 ; M. Delafontaine, Liebig's Ann., 131. 105, 1864 ; J. Brush and S. L. Penfield, Amer. J. Science (3), 25. 459, 1883 ; G. Bodmann, Zeit. anal. Chem., 27. 251, 1901 ; H. du Bois and 0. Liebknecht, Ber., 32. 3346, 1899. 1 If much yttrium earths be present, the precipitate may be contaminated with them. The precipitate is then redissolved and the operations repeated. A similar remark applies if phosphorus be present. 2 If the ammonium acetate be added before the oxalate precipitate has formed, the mixture is inclined to give a turbid filtrate. 3 E. Hintz and H. Weber, Zeit. anal. Chem., 36. 27, 1897 ; E. Benz, Zeit. angew. Chem., 14. 297, 1902. For the solubility of thorium oxalate, see R. Bunsen, Pogg. Ann., 155. 380, 1875. E. Benz says that "under no circumstances is it possible to get a satisfactory separation of thorium from cerium by ammonium oxalate in the presence of ammonium acetate." If there is any difficulty, use Metzger's process, page 510. 4 If the ammonium nitrate be omitted, it is so difficult to calcine the voluminous gelatinous precipitate without loss of fine dust, that it is advisable to dissolve the precipitate in hydro- chloric acid, and reprecipitate with ammonia. 5 Traces of phosphoric oxide, or traces of cerium, if present, will be precipitated as cerium phosphate, etc., with the thorium. If an appreciable time elapses between the precipitation and the filtration of the thorium hydroxide, the greater the risk of contamination with ceria. 6 This behaviour with hydrogen peroxide distinguishes thorium and zirconium from cerium and yttrium compounds. c;o6 A TREATISE ON CHEMICAL ANALYSIS. nitrate. Add the filtrates to the solution of cerium earths employed in the next operation. Wash the precipitate, ignite in a platinum crucible, and weigh as thorium oxide Th0 2 . The precipitate should be white. 1 If the precipitate has a yellow tinge, a little ceria may be present. A reprecipitation is necessary to get white thoria free from ceria. 2 8. Separation of Cerium from Lanthanum and Didymium. The precipitated cerium earths may be ignited and weighed as " cerium earths." The precipitate would include lanthana and didymia. To separate the cerium, 3 dissolve the mixed oxides in hot hydrochloric acid. Add a solution of potassium or sodium hydroxides, so as to precipitate the gelatinous hydroxides of cerium, lanthanum, and didymium. Wash the hydroxides three or four times by decantation ; add a concentrated solution of potassium hydroxide so as to make the mixture occupy about 200 c.c Pass a slow current of chlorine gas through the solution, and agitate the mixture from time to time. When the liquid no longer has an alkaline reaction, and is saturated with chlorine, the cerium hydroxide will have been oxidised to lemon-yellow cerium dioxide, which remains as an insoluble pre- cipitate, whereas the lanthanum and didymium hydroxides will have dissolved. 4 Let the mixture stand in a corked flask for 24 hours. Filter, and wash the cerium dioxide. While the precipitate is still moist, dissolve it in hydrochloric acid. Add ammonium oxalate to the solution. Wash and ignite the precipitated cerium oxalate, and weigh as cerium oxide Ce0 2 . Cerium oxide has a light rose colour, and it dissolves forming a yellow solution in sulphuric acid. The colour is bleached by the addition of sulphurous acid, and restored by the addition of hydrogen peroxide or sodium peroxide. 9. Determination of Lanthanum and Didymium. Owing to the laborious nature of the process for the separation of these two bases, they are often pre- cipitated together as oxalates in the filtrate from the cerium dioxide, calcined, and weighed as a mixture of "La 2 3 + Di 2 3 ." 5 Damour and Deville's process 6 of separation is based on the fact that, when a mixture of didymium and lanthanum nitrates is heated, the didymium nitrate decomposes before the lanthanum nitrate and forms a sparingly soluble didymium subnitrate. The filtrate from the cerium dioxide is boiled to eliminate the chlorine, and treated with an excess of ammonia. The precipitated didymium and lanthanum hydroxides are washed, and dissolved in nitric acid. The solution is evaporated 1 P. T. Cleve, Bull. Soc. Chim. (2), 43. 53, 1885 ; G. Wyrouboff and A. Verneuil, ib. (3), 19. 219, 1898 ; Chem. News, 77. 245, 1898 ; Compt. Rend., 126. 340, 1898 ; 127. 412, 1898 ; 128. 1331, 1899 ; L. de Boisbaudran, ib., 100. 605, 1864 ; E. Benz, ZeU. angew. Chem., 15. 297, 1902. For another method of purifying the thorium oxide, see page 510 Metzger's process. 2 White thorium hydroxide, Th(OH) 4 , readily dissolves in ordinary mineral acids ; the oxide Th0 2 is practically insoluble in these acids, but it dissolves in hot fuming sulphuric acid, and it can be converted into a soluble sulphate by fusion with sodium pyrosulphate. 3 G. Mosander, Journ. prakt, Chem. (1), 30. 267, 1843 ; H. St C. Deville, Compt. Rend., 59. 272, 1864 ; P. Schiitzenberger, Compt. Rend., 120. 663, 962, 1143, 1895 ; 124. 481, 1897. 4 0. Popp (Liebig's Ann., 131. 359, 1864) used sodium hypochlorite in place of chlorine gas for the separation ; W. Gibbs (ZeU. anal. Chem., 3. 396, 1864) oxidised the solution with lead 19. 219, 1898; Chem. News, 77. 245, 1898), hydrogen peroxide; G. Bricout (Compt. Rend., 118. 145, 1894) suggested a separation based on the solubility of cerium carbonate in chromic acid ; M. M. Pattison and J. Clarke (Chem. News, 16. 259, 1867) decomposed the chromate by heat, etc. ; and H. Robinson (ib., 54. 229, 1886) based a process on the different solubilities of the nitrates. 5 The term " didymium" is here used for a mixture of praseodidymium and neodidymium. 6 E. Damour and H. St C. Deville, Bull. Soc. Chim.. (2), 2. 339, 1864 ; P. Schiitzenberger and 0. Boudouard, Compt. Rend., 122. 697, 1896; 123. 782, 1896- 126. 1648 1898 L F. Nilson, ib. t 88. 642, 647, 1879. ZIRCONIUM, THORIUM, AND THE RARE EARTHS. 507 to dryness in a small weighed platinum basin. The dried mass usually has a pale rose colour. Heat the basin to a temperature of about 450 in a small muffle. 1 Care must be taken not to overheat the mixed salts near the bottom of the basin. The salts fuse and give off nitrous fumes. It is difficult to ensure uniform heating if a large quantity of the mixed nitrates is being treated. The result is better when small quantities are treated. In about half a minute remove the basin from the source of heat. When nearly cold, add hot water. The lanthanum nitrate dissolves, while greenish-grey flakes of didymium sub- nitrate remain undissolved. After the mixture has stood a couple of hours, boil and filter. If the nitrate has a pink colour, evaporate again with nitric acid, and repeat the operation until a colourless liquid is obtained. Two or FIG. 158. Didymium sulphate. FIG. 159. Lanthanum sulphate. three repetitions of the process may suffice. According to Cleve, a repetition of the process fourteen times will give a solution in which the spectroscope can detect no didymium. The didymium subnitrate is calcined until its weight is constant, and weighed as didymium oxide Di 2 3 . The didymia so obtained varies in tint from pure white to pale cinnamon brown usually a pale yellowish- brown oxide is obtained which is readily soluble in dilute nitric acid. The lanthanum solution is also evaporated to dryness in a weighed dish, ignited, and weighed as lanthanum oxide La 2 3 . A slight trace of lanthanum nitrate may be decomposed when the nitrates are calcined ; as a result, the amount of didymia Di 2 3 reported above is usually a little high, and the lanthania La 2 3 a little low. The lanthanum oxide generally has a pale yellowish-brown colour (if pure it would be white), and it is soluble in acids e.g. dilute nitric acid. When didymium oxide is treated with sulphuric acid, reddish-violet crystals 1 A small electric furnace \vith a pyrometer is excellent for the purpose. 508 A TREATISE ON CHEMICAL ANALYSIS. (oblique rhomboidal prisms) of didymium sulphate (fig. 158) are obtained; while lanthanum sulphate furnishes colourless needles (right rhomboidal prisms) (fig. 159). Ideally perfect crystals are also shown in outline. Some needles of lanthanum sulphate can usually be detected among the plates of didymium sulphate, and conversely. II. The Filtrate from the Ammonium Oxalate Precipitate. 10. Determination of Aluminium and Phosphorus. Add an excess of ammonia to the solution l remaining after the separation of the oxalates of the thorium and cerium groups. Filter and wash the precipitated hydroxides and phosphates. The filtrate may contain traces of cerium and thorium, which are usually neglected. The washed precipitate is dried and fused with sodium carbonate. The cold cake is extracted with water. Filter and wash the insoluble matter. The solution contains sodium phosphate and sodium aluminate. Bring the solu- tion to a definite volume, and determine the phosphorus in one portion (page 595), and the combined aluminium and phosphoric oxides in another (page 182). 11. Determination of Calcium. Dissolve the insoluble residue left, after leaching the sodium carbonate fusion with water, in dilute hydrochloric acid. 2 Add ammonia, filter and wash. Precipitate the lime as calcium oxalate from the filtrate in the usual manner (page 213). 12. Determination of Iron and Manganese. Dissolve the precipitate in hot dilute hydrochloric acid, and neutralise the solution with dilute ammonia. Pour the solution slowly, with constant stirring, into a cold mixture of ammonium carbonate and sulphide. 3 Iron and manganese sulphides are precipitated, while zirconium, beryllium, and yttrium remain in solution. The precipitated iron and manganese sulphides can be separated in the usual manner. 13. Separation of Beryllium from Zirconium and Yttrium. Boil the filtrate for an hour; beryllium, zirconium, and yttrium hydroxides are precipitated. Filter and wash. Dissolve the precipitate in dilute hydrochloric acid. 4 Treat the solution with an excess of sodium hydroxide, when zirconium and yttrium hydroxides are precipitated. The beryllium remains in solution. Boil the diluted filtrate one hour, when beryllium hydroxide is precipitated. Wash the precipitate, and treat as described on page 449. 14. Separation of Zirconium and Yttrium. This separation 5 is based on the fact that potassium yttrium sulphate, Y 2 (S0 4 ) 3 .3K 2 S0 4 , is soluble in 7 to 8 parts of a feebly aqid solution of potassium sulphate, while the corresponding potassium zirconium sulphate is practically insoluble in the same menstruum. 6 1 This solution may contain aluminium, phosphorus, iron, manganese, calcium, magnesium, beryllium, zirconium, and yttrium. 2 To make sure that all the zirconium is dissolved, incinerate the filter paper, and fuse with a little sodium carbonate. Add the solution to the main solution. 3 The amount of ammonium carbonate should more than suffice to retain the oxides of yttrium, beryllium, and zirconium in solution ; and the ammonium sulphide should suffice to precipitate all the iron and manganese. 4 Or add hydrochloric acid to the solution and boil to expel the carbon dioxide. Cool, and treat the solution with sodium hydroxide, etc. Incinerate the filter paper as described in a preceding footnote. 5 J. J. Berzelius, Pogg. Ann., 4. 135, 1825; C. M. Warren, ib., 102. 449, 1857; E. Linnemann, Monats. Chem., 6. 335, 1885; C. Marignac, Ann. Chim. Phys. (5), 20. 535, 1880 ; Compt. Rend., 90. 899, 1880; Chem. News, 41. 250, 1880; M. Delafontaine, ib., II. 241, 253, 1865 ; G. Kriiss, ib., 64. 65, 75, 100, 120, 1891 ; Liebig's Ann., 265. 1, 1891 ; G. Urbain, Ann. Chim. Phys. (6), 19. 184, 1900. 6 When zirconia compounds are melted with potassium bisulphate, the insoluble double sulphate remains behind when the melted mass is extracted with water. Neither sodium nor ammonium sulphate can be used in place of the potassium salt, because they do not give a sparingly soluble double salt. ZIRCONIUM, THORIUM, AND THE RARE EARTHS. 509 The operation is conducted in the following manner : Dissolve the precipitate l containing the mixed zirconium and yttrium hydroxides in a little concentrated sulphuric acid. 2 Stir the ice-cold solution with a saturated solution of potassium sulphate. 3 This precipitates a double potassium zirconium sulphate 4 possibly Zr(S0 4 ) 2 .2K 2 S0 4 .2H 2 which is practically insoluble in ice-cold water and in solutions of potassium sulphate. Hence, filter 5 and wash with a cold solution of potassium sulphate. The double sulphate is decomposed by boiling with a concentrated solution of sodium hydroxide. The compact zirconium hydroxide is washed, first by decantation, and then on the filter paper. The precipitate is dissolved in hydrochloric acid, 6 and reprecipitated with ammonia, filtered, washed, ignited, and weighed as zirconia Zr0 2 (page 498). The filtrate from the potassium zirconium sulphate is treated with an excess of ammonia. 7 The precipitate is filtered, washed, dissolved in acid, re- precipitated with ammonia, washed, ignited, and weighed as yttria Y 2 3 . The yttrium so obtained is usually pale straw-yellow (if pure, it would be white), and soluble in warm acids. The scheme of analysis may now be summarised (precipitates and solids on left, filtrates and solutions on right) : Boil result of NaHS0 4 fusion with much water and H 2 S0 4 . Digest with (NH^S. Digest with H 2 S0 4 and H 2 2 . Tin ; Tungsten. I I Lead ; Silica. Boil with S0 2 . Digest with salicylic acid. Treat H^. Titanium. Niobium; Tantalum. H 2 S group. Neutral sol. ;' amm. oxalate. I H 2 S0 4 ; amm. oxalate and acetate. NH 3 ; fuse Na^COs ; digest H 2 0. lite ; HC1 ; KOH and Cl. Thorium. Dissolve in HC1 ; add NH 3 . Phosphorus ; Aluminiut Cerium. Nitrates ; calcine ; treat H 2 0. HC1 ; amm. carb. and sulphide. Calcium ; Magnesium. Didymium. Lanthanum. Iron ; Manganese. Boil ; HC1 ; NaOH. ; HC1 ; H 2 S0 4 ; treat K 2 S0 4 . Beryllium. | I Zirconium. Yttrium. 1 Incinerate the filter paper, etc., as indicated in a preceding footnote. 2 Sulphuric acid is preferable to nitric and hydrochloric acids, since, if chlorides or nitrates be present, some yttrium may be precipitated. 3 POTASSIUM SULPHATE. The finely powdered salt is ground in a mortar with water at about 30, but not at a higher temperature. 4 The corresponding sodium and ammonium salts are fairly soluble in water. 5 The precipitate sometimes sticks tenaciously to the sides of the beaker, 6 The filter paper is incinerated, etc. , as usual. 7 C. Kersten, Pogg. Ann., 47. 392, 1839 ; H. Rose, ib., 118. 508, 1863 ; R. Fresenius and E. Hintz, Zeit. anal. Chem., 35. 532, 1896 ; 0. Boudouard, Bull. Soc. Chim. (3), 19. 11, 1898 ; 510 A TREATISE ON CHEMICAL ANALYSIS. 268. The Gravimetric Determination of Thorium Metzger's Process. Metzger's process 1 is based on the fact that a saturated solution of fumaric acid in 40 per cent, alcohol quantitatively precipitates white flocculent thorium oxide from neutral solutions containing 40 per cent, alcohol by volume. The only other metals precipitated are zirconium (completely), erbium (partially), silver, and mercury. The trial results are excellent. Cerium, 2 lanthanum, didymium, yttrium, samarium, gadolinium give no precipitates in hot or cold solutions under the same conditions. Similarly with salts of copper, gold, man- ganese, calcium, strontium, barium, zinc, cadmium, boron, aluminium, tin, lead, phosphorus, arsenic, antimony, bismuth, uranium, vanadium, tungsten, iron, cobalt, nickel, platinum. First Precipitation. The precipitated oxalates (page 504) are washed and rinsed into a beaker, mixed with 20-25 c.c. of concentrated caustic potash, and boiled. Dilute the solution with water, filter and wash. Dissolve the pre- cipitated hydroxides in dilute nitric acid (1:1), and evaporate the solution to dryness on a water bath. Dissolve the residue in 50 c.c. of water, and add sufficient alcohol and water to make a solution containing 40 per cent, of alcohol, by volume, and occupying 200 c.c. Add 20 to 25 c.c. of fumaric acid, 3 and heat the solution to boiling. Filter while hot, and wash the precipitated thorium several times with hot 40 per cent, alcohol. Second Precipitation. Return the filter paper and the precipitate to a beaker. Add 25-30 c.c. of dilute hydrochloric acid (1:1). Heat the solution to boiling, and filter off the paper. Wash with dilute acid, and evaporate the filtrate and washings to dryness on a water bath. Agitate the solution, and wash down the sides of the basin every now and again, to prevent the residue clinging to the sides. While still on the water bath, loosen the residue from the bottom of the basin by stirring with a " policeman." The carbonaceous matter does not interfere. Add sufficient alcohol and water to make the solution occupy about 150 c.c. j add 10 c.c. of fumaric acid ; and heat the solution to boiling. Filter, and wash with 40 per cent, alcohol. Ignite, and weigh as thorium oxide Th0 2 . Evaluation of Thorium Minerals* The thorium minerals are evaluated on G. Kriiss and L. F. Nilson, Ber., 20. 1677, 1887. According to C. F. Whittemore and C. James (Journ. Amer. Chem. Soc , 34. 772, 1912) the precipitation of yttrium hydroxide by sodium or ammonium hydroxide gives high results in the presence of potassium or sodium salts, and they recommend one precipitation with ammonium sebacate in the presence of sodium salts, and two precipitations in presence of potassium salts. 1 F. J. Metzger, Journ. Amer. Chem. Soc., 24. 901, 1902. A. C. Neish (Journ. Amer. Chem. Soc., 26. 780, 1904 ; Chem. News, go. 196, 201, 1904) separated thorium by metanitro- benzoic acid ; L. M. Dennis and'F. L. Kortright (Amer. Chem Journ., 16. 79, 1894; Journ. Amer. Chem. Soc., 18. 947, 1896), by potassium trinitride KN 3 : J. J. Chydenius (Pogg. Ann., 119. 45, 1861 ; R. Fresenius and E. Hintz, Zeit. anal. Chem., 35. 530, 1896 ; J. P. Drossbach, Zeit. angew. Chem., 14. 656, 1901 ; E. Benz, ib., 15. 302, 1902), by boiling with sodium thiosulphate ; T. 0. Smith and C. James (Journ. Amer. Chem. Soc., 34. 281, 1912) by sebacic acid; and M. Koss (Chem. Ztg., 36. 686, 1912; A. Rosenheim, ib., 36. 821,1912; F. Wirth, Zeit. angew. Chem., 25. 1678, 1912) by sodium hypophosphate. See also B. L. Hart- well, Journ. Amer. Chem. Soc., 25. 1128, 1903; Chem. News, 89. 15, 27, 1904. Thorium is precipitated quantitatively by ammonium molybdate F. J. Metzger and F. W. Zons, Journ. Ind. Eng. Chem., 4. 493, 1912. 2 If much cerium be present, a little maybe precipitated during the first precipitation of the thorium oxide. 3 FUMARIC ACID SOLUTION. Dissolve 1 grm. of fumaric acid in 100 c.c. of water. Fumaric acid costs about 2s. for 10 grms. 4 E. White, Thorium and its Compounds, London, 1912 ; 0. R. Bohm, Die Fabrication der Gluhkorper fur Gasgluhlicht, Halle a. S., 1910. ZIRCONIUM, THORIUM, AND THE RARE EARTHS. 5 1 I their thorium contents. Most of the commercial thoria comes from monazite sand, which has about 6 per cent, of Th0 2 ; a little comes from thorianite, which ' has about 80 per cent, of Th0 2 . The following process of evaluation has been tacitly recognised by buyer and seller as a kind of standard : Heat 12*5 grms. of the sand to 180 9 -200 with 50 c.c. of concentrated sulphuric acid. In 2 or 3 hours the grains will be all broken up. Dilute the cold mixture with 300-400 c.c. of water, and filter. Make the solution up to 500 c.c. Agitate 200 c.c. (5 grms. of sample) with 180 c.c. of a cold saturated solution of oxalic acid. Let stand overnight (about 12 hours), and niter. Wash the precipitate with dilute, hydrochloric acid or water until the runnings are free from phosphates. Reject the filtrates. Ignite the dried precipitate and dissolve it in hydrochloric acid (sp. gr. 1*16). Evaporate to dryness, add a few c.c. of water, and again evaporate to dryness. Dissolve the acid-free chlorides in 200 c.c. of water, and add 9 grms. sodium thiosulphate in 30 c.c. of water. After standing 12 hours, boil for 10 minutes, filter, and wash the precipitate until ammonia gives no turbidity. Boil the filtrate for an hour, and collect the precipitate A as before. Reject the filtrate. The first precipitate is dissolved in 5 per cent, hydro- chloric acid. The acid solution is evaporated to dryness, taken up with 150 c.c. of water, and treated with 3 grms. of sodium thiosulphate as before. The filtrate is treated with ammonia and boiled. The precipitate is washed and mixed with the precipitate A. The thiosulphate precipitate is dissolved in acid, evaporated, etc., and treated with thiosulphate as before. The precipitate is washed until it gives no turbidity with ammonia. The filtrate is precipitated with ammonia, while the thiosulphate precipitate is again subjected to the preceding treatment until the filtrate gives no precipitate with ammonia. Three thiosulphate precipitations generally suffice. The last thiosulphate precipitate is dissolved in 5 per cent, hydrochloric acid, and the solution made up to 150 c.c. ; 10 c.c. of hydrochloric acid are added, and then 30 c.c. of oxalic acid. The liquid is kept between 30 and 40 for 2 to 3 hours, and, after standing overnight, filtered, washed, dried, ignited, and weighed as Th0 2 . The ammonia precipitates and the precipitate A are mixed and re- worked for thoria. The yield is about 0*005 grm., or O'l per cent, on the 5 grms. sample. 269. The Volumetric Determination of Cerium Knorre's Process. Cerium can be determined volumetrically in the presence of thorium, lan- thanum, and didymium, by Knorre's process. 1 This is based on the fact that cerous salts are oxidised to yellow eerie salts by the action of ammonium per- sulphate in sulphuric acid solution, and the eerie salts are reduced to colourless cerous salts by the action of hydrogen peroxide : 2Ce0 2 + H 2 2 = Ce 2 3 + 2 + H 2 0. If, therefore, an excess of a solution of hydrogen peroxide of known strength be added to the yellow eerie salt, the excess of hydrogen peroxide can be determined by back titration with potassium permanganate. With freshly prepared solu- tions of the eerie salts, the reduction with hydrogen peroxide is instantaneous and the permanganate is also decolorised at once. With solutions which have been exposed to the air for some time, the reduction may take a quarter of an hour. Hence, if an old solution be titrated with permanganate within a quarter of an hour after adding the hydrogen peroxide, the results will be low. When the 1 G. von Knorre, Zeit. angew. Chem., 10. 685, 717, 1897 ; -Her., 33. 1924, 1900; A. Job, Compt. Rend., 128. 101, 180, 1899 ; E. Hintz and H. Weber, Zeit. anal. Chem., 37. 94, 504, 1908; Chem. News, 79. 25, 41, 1909; G. P. Drossbach, Ber., 29. 2452, 1896; A. Waegner and A. Miiller, ib., 36. 282, 1903 ; F. J. Metzger, Journ. Am>er. Chem. Soc., 31. 523, 1909 ; L. Schneider, Dingier 's Journ., 269. 224, 1888 ; W. Muthmann and L. Weiss, Liebig's Ann., 331. 1, 1904 ; R. J. Meyer and A. Schweitzer, Zeit. anorg. Chem., 54. 104, 1907. 512 A TREATISE ON CHEMICAL ANALYSIS. reduction is completed, the total permanganate consumed with both old and new solutions is the same. If an old solution be boiled for a few minutes with dilute sulphuric acid, and cooled before adding the hydrogen peroxide, the rate of the reduction is accelerated. Oxidation of Cerous to Ceric Salts. Acidify the solution in an Erlenmeyer's flask with dilute sulphuric acid, and add 2 grms. of ammonium persulphate to the cold solution. Heat the solution to boiling for a couple of minutes. Cool . to between 40 and 60 by dipping the vessel containing the solution in cold water. Add half a gram of ammonium persulphate, 1 heat the solution to boiling for about 5 minutes; cool, add another half gram of the persulphate, and boil 15 minutes, adding more dilute sulphuric acid towards the end of the boiling in order to decompose all the ammonium persulphate. Cool. 2 The Titration. Acidify the solution with sulphuric acid, and then add an excess of hydrogen peroxide. The hydrogen peroxide is added from a burette until the solution is decolorised, and then a small excess of the peroxide is run into the solution. When the reaction between the yellow eerie salt and the hydrogen peroxide is over, titrate the residual hydrogen peroxide with potassium permanganate until the solution acquires a rose colour which persists for at least half a minute. 3 Calculation. Suppose that 12*2 c.c. of the permanganate solution are re- quired, and suppose that 25 c.c. of hydrogen peroxide were added to the solution, and that 25 c.c. of the peroxide require 2 9 '8 c.c. of permanganate. The eerie salts in the solution correspond with 29'8 less 12'2 = 17'6 c.c. of the perman- ganate. The reaction between the permanganate and the hydrogen peroxide is represented by the equation : 2KMn0 4 + 5H 2 2 + 3H 2 S0 4 = K 2 S0 4 + 2MnS0 4 + 8H 2 + 50 2 . On comparing this equation with that indicated on page 198, and with 2Ce(S0 4 ) 2 + H 2 2 = Ce 2 (S0 4 ) 3 + H 2 S0 4 + 2 , it follows that 55-84 grms. of Fe correspond with 140*25 grms. of Ce, or 172*25 grms. of Ce0 2 . It was found that 1 c.c. of the permanganate solution used in the above titration 4 corresponded with 0'00365 grm. Fe; hence, 1 c.c. will correspond with O00365 x 3'0714 = 0-01121 grm. Ce0 2 . Consequently 17'6 c.c. o the permanganate solution used in the back titration corresponded with 17-6 x 0-01121 =0-1973 grm. Ce0 2 in the given solution. Phosphoric and titanic acids disturb the action. The results are very fair if the above directions be carefully followed. 1 3 grms. of ammonium persulphate suffice for the oxidation of 0*2 to 0'3 grm. cerium. 2 Of course, if all the cerium is present as eerie sulphate, this preliminary oxidation is not needed. 3 The slow decomposition of potassium permanganate by cerous sulphate does not interfere with the recognition of the end point in the permanganate titration. 4 The permanganate solution should not be more concentrated than is represented by 2 grms. per litre. CHAPTER XXXVI. SPECIAL METHODS FOR THE DETERMINATION OF BARIUM, STRONTIUM, CALCIUM, AND MAGNESIUM. 270. The Influence of Barium and Strontium on the Calcium and Magnesium Precipitates. THE spectroscopic test for barium and strontium is generally applied to the ignited calcium oxalate precipitate when these elements are sought in a silicate analysis. 1 These elements have characteristic spectra (Plate I.). It is assumed that sufficient barium and strontium will be precipitated with the lime to reveal their presence, in spite of the fact that barium and strontium oxalates are more 1 A small direct- vision spectroscope with a scale is a very convenient adjunct for deciding if certain precipitates are properly washed. Fig. 160 illustrates the method of using the FIG. 160. Spectroscopic Test. spectroscope. The loop of platinum wire in the Bunsen's flame (porcelain burner) has been dipped in a drop of the solution under investigation. F. von Kobell (Journ. prakL Chem. (2), 3. 176, 1871) applies the spectroscope test to silicates by placing the powdered mineral, moistened with hydrochloric acid, on a piece of platinum foil, perforated with small holes and bent in the form of a trough. The arrangement is held in the flame by a pair of platinum- tipped tongs. The spectra of barium, strontium, and calcium are illustrated in Plate I. The spectroscopic method will detect y^Vff mgrm. of baryta ; ^^o mgrm. of strontia ; and looSfob m g rm - of lime (Bunsen). 5,3 33 514 A TREATISE ON CHEMICAL ANALYSIS. soluble than the corresponding calcium oxalate. Hillebrand ] says that " this assumption, in the case of strontium, is well founded, but it may be entirely fallacious in the case of barium." For instance, a mineral containing 0*76 per cent, of barium gave no indication of barium under the conditions of the test. This is due to the solvent action of ammonium chloride upon the barium oxalate. In a series of experiments with artificial mixtures of calcium with barium and with strontium, it was found that strontium and barium are but very incom- pletely precipitated by the addition of a slight excess of ammonium oxalate, but a greater proportion is precipitated when a large excess of ammonium oxalate is employed. With traces of strontium in the presence of a large excess of calcium, most of the strontium is precipitated with the calcium ; but this action was not noticed with barium, since a considerable proportion of the barium escapes precipitation with the calcium, and is therefore to be sought in the nitrate. With a double precipitation of the lime, much of the strontium, and practically no barium, will be associated with the lime, provided the amount of barium does not exceed 0'2 or 0*3 per cent. If more than this amount of barium be present, the precipitation by ammonium oxalate can be repeated a third and fourth time, or the strontium and barium can be separated from the calcium ; and the barium and strontium, which are recovered from the lime precipitate, can be separated by the process described below. If barium be present, it may also contaminate the ammonium magnesium phosphate precipitate as barium phosphate. In that case, the barium must be removed 2 before precipitating the magnesium ammonium phosphate. Add 3 drops of sulphuric acid to the filtrate from the calcium oxalate ; evaporate the solution to dryriess; ignite the residue in a porcelain dish to drive off the ammonium salts ; take the residue up with water acidulated with hydrochloric acid ; filter off the carbonaceous matters ; add a drop of sulphuric acid ; let the mixture stand about 12 hours; and if a precipitate forms, filter, and treat the precipitated barium sulphate as described below. Determine the magnesia in the filtrate as indicated on page 218. 271. The Separation of Calcium from Strontium and Barium Stromeyer and Rose's Process. If strontium be present, most of it will be found associated with the lime. The separation of barium and strontium is conveniently done by Stromeyer's process improved by Rose. 3 The mixed oxalates are ignited as usual and weighed as oxides. Calcium. The oxides are dissolved in dilute nitric acid (1 : 5) in a small 25-c.c. stoppered flask, 4 and the solution evaporated to dryness at 150 to 160 in a stream of dry air in an attachment resembling that employed for lithium 1 W. F. Hillebrand, Journ. Amer. Gkem. Soc., 16. 81, 83, 1894; Bull. U.S. Geol. Sur., 422. 120, 1910 ; Chem. News, 69. 142, 147, 1894. 2 R. Langley (Amer. J. Science (4), 26. 123, 1908) removes the barium by precipitation with sulphuric acid just after the separation of silica ; redissolves the precipitate in concentrated sulphuric acid, and reprecipitates it with water, in order to remove ferric and other sulphates. P. Stromeyer, Gilbert's Ann., 54. 245, 1816 ; H. Rose, Pogg. Ann., no. 292, 1860; R. Fresenms, Zeit anal. Chem., 29. 20, 143, 413, 1890 ; 30. 18, 452, 583, 1891 ; 32. 189, 1893 ; Chem. News, 68. 213, 1893 ; J. L. M. van der Horn van der Bos, Chem. Weekblad, 8. 5, 1911. Ton? 1 " nm g ( Amer - J - Science (3), 43. 50, 1892 ; (3), 44. 462, 1892 ; Chem. News, 67. 45, 53, 1893 ; 65. 2/1, 282, 1892 ; 66. 3, 1892) recommends amyl alcohol in place of the mixture of alcohol and ether employed in Stromeyer's and Rose's process ; L. Moser and L. Machiedo (Chem. Ztg. 35. 337, 1911) say that amyl alcohol offers no advantages over the ether-alcohol mixture. 4 I use a weighing bottle like fig. 3, b, for this purpose. BARIUM, STRONTIUM, CALCIUM, AND MAGNESIUM. 515 (fig. 163, page 536). Treat the dry mass in the stoppered flask with about 10 times its weight of a mixture of equal volumes of ether and alcohol, in which the barium and strontium nitrates are but sparingly soluble, 1 while calcium nitrate is readily soluble. Let the flask stand about 12 hours, with occasional shaking. Filter through a small 5'5-cm. filter paper moistened with a couple of drops of the ether-alcohol mixture. Wash with the mixture of ether and alcohol until a drop of the filtrate gives no residue when evaporated to dryness on a clean piece of platinum foil 2 about six washings usually suffice. Evaporate the filtrate to dryness. Dissolve the calcium nitrate in water, and precipitate the calcium as oxalate in the usual manner (page 213). Strontium. The strontium can be determined by difference, or directly Dissolve the residue on the filter paper in dilute nitric acid ; collect the washings in a 50-c.c. beaker and add sulphuric acid (1:2) almost equal in volume to the liquid in the beaker. In about 12 hours, filter off the precipitate, ignite, 3 and weigh as strontium sulphate SrS0 4 . The weight of strontium sulphate so ob- tained, multiplied by 0'5641, gives the corresponding amount of strontia SrO. Barium. If barium be present in the calcium oxalate precipitate, it will also be found with the strontium sulphate. The barium and strontium can be separated by the chromate process described below. 4 If barium and strontium are to be determined, it is advisable to combine the filtrates from the calcium oxalate with the strontium and barium separated from the oxalate itself, and then apply the chromate process to the combined filtrates. But there are so many leakages of barium, and other sources of error in the analysis, that it is best to determine the barium on a separate sample, either by the hydrofluoric decomposition process, or by the process of decomposition indicated on page 498. 272. The Separation of Barium from Strontium and Calcium Chromate Process. The separation of barium, strontium, and calcium can be conveniently effected by (1) removing the calcium nitrate by Stromeyer and Rose's process, and then separating the barium and strontium by the chromate process ; or (2) separating barium from the neutral solution of the mixed chlorides by the 1 About 0'0017 per cent, of strontium nitrate. 2 Take care that no flame is near the ether-alcohol. The mixture is very inflammable. 3 If the ignition temperature be too high, some of the sulphate may decompose M. Darmstadt, Zeit. anal. Chem., 6. 376, 1867. For the loss on heating strontium sulphate, see M. Darmstadt, ib., 6. 376, 1867 ; A. Mitscherlich, Journ. prakt. Chem. (1), 83. 485, 1861 ; J. Boussingault, Compt. Rend., 64. 1159, 1867. 4 A rough separation of barium from strontium can be effected by digesting the mixed sulphates in a cold concentrated solution of ammonium carbonate. Plug the narrow end of the funnel with a piece of cork, or use a funnel with a stopcock, and fill the funnel with the solution of ammonium carbonate Let the whole stand 12 hours. The strontium sulphate is converted into strontium carbonate, while the barium sulphate is but slightly affected. The precipitate is then washed with hot water, dilute hydrochloric acid, and finally with water. Ignite, and weigh as barium sulphate. The strontium is determined by neutralising the filtrate with ammonia, precipitation with ammonium carbonate, converting the ignited strontia into sulphate, and weighing as strontium sulphate. If the barium predominates, an appreciable quantity of strontium sulphate escapes decomposition ; and when the strontium is in excess, an appreciable quantity of barium sulphate will be decomposed. R. Fresenius, Zeit. anal. Chem., 29. 20, 1890; S. G. Rawson, Journ. Soc. Chem. Ind., 16. 113, 1897; H. Rose, Fogy. Ann., 95. 286, 299, 427, 1855. See P. Schweitzer (Contrib. Lab. State Univ., 1, 1876 ; Proc. Amer. Assoc., 187, 1877) for the theory of the reaction. L. Moser and L. Machiedo (Chem. Ztg., 35. 337, 1911) consider the sulphate process gives an inexact separation of strontium, and believe that the best way of separating strontium from calcium and barium is to wash the anhydrous nitrates with a mixture of ether and alcohol. The strontium nitrate dissolves, while calcium and barium nitrates remain undissolved. 516 A TREATISE ON CHEMICAL ANALYSIS. chromate process, precipitating the calcium and strontium as carbonates by means of ammonium or sodium carbonate, and after transforming the carbonates into nitrates, using Stromeyer and Rose's process for the calcium and strontium. The transformation of the mixed carbonates or oxides into nitrates offers no particular difficulty ; but in the case of the chlorides, repeated evaporation with concentrated nitric acid will not do, because the transformation is not complete. 1 It is then best to precipitate the carbonates with ammonium or sodium carbonate, and digest the carbonates with as little nitric acid as possible in a small flask, taking care, of course, to avoid loss by spurting. First Precipitation of Barium Chromate? Evaporate the mixed nitrates to dryness, and dissolve the mixture in water. Neutralise any great excess of acid with ammonia, and add an excess, say 10 c.c., of a solution of ammonium acetate ; 3 heat the solution to boiling, and gradually add, with constant agitation, 5 c.c. of ammonium bichromate solution. 4 Let the precipitate settle, and when the solution is cold, decant the clear liquid through a filter paper, and wash the precipitate by decantation with a dilute solution of ammonium acetate 5 until the filtrate is no longer perceptibly coloured. About 100 c.c. of liquid will be required for the washing. Some strontium, if present, may be carried down with the barium chromate ; hence the precipitate is dissolved in dilute acid and reprecipitated. Second Precipitation of Barium Chromate. Place a beaker under the funnel and dissolve the precipitate on the filter paper by running warm dilute nitric acid (1:5) through the filter paper, and collect the "runnings" in the vessel in which the precipitation was first made. Wash the paper. Add ammonia to the solution until the precipitate which forms no longer redissolves when the solution is agitated. Add, with constant agitation, 10 c.c. of the concentrated solution of ammonium acetate ; heat the solution to boiling. Let the solution settle till cold, and then filter through a weighed Gooch's crucible packed with asbestos felt. Wash the precipitate as before. Dry, ignite, and weigh as described on page 478. 6 The weight of the barium chromate multiplied by 0'60507 gives the corresponding amount of barium oxide BaO. Determination of Strontium. The strontium may be precipitated from the combined filtrates by the addition of ammonia and ammonium carbonate after the solution has been concentrated in the presence of an excess of nitric acid. The precipitated strontium chromate may contain a little chromic acid. This is removed either by dissolving the precipitate in hydrochloric acid and pre- cipitating the strontium as sulphate (page 515) in the presence of alcohol, if strontium alone and no calcium be present ; or, by the application of the ether- alcohol process (page 5 14) in the event of both strontium and calcium being present. 7 1 If magnesium be present, it too will be precipitated, and the chromium must later on be separated from the nitrate before the magnesium can be precipitated as phosphate. The chromium is separated by reducing the chromate as described on page 479, and precipitating the chromium as hydroxide by the addition of ammonia. 2 The properties of barium chromate were discussed on page 477. It might be added that barium chromate is not soluble in water containing acetic acid when so much ammonium chromate is present that the solution contains only alkaline acetate and bichromate (H. N. Morse, Amer. Chem. Journ., 2. 176, 1880). 3 AMMONIUM ACETATE. Neutralise an aqueous solution of 300 grms. of " pure " commercial ammonium acetate with ammonia, and make the solution up to a litre. See R. Reik (Monats. Chem., 23. 1033, 1902), A. Mittasch (Zeit. anal. Chem., 42. 492, 1903), and L. H. Duschak (Journ. Amer. Chem. Soc., 30. 1827, 1908) for the impurities in the commercial salt. 4 AMMONIUM BICHROMATE SOLUTION. Dissolve 200 grms. of the salt in a litre of water. 5 Made by diluting 20 c.c. of the above concentrated solution of ammonium acetate to a litre. 6 Simply drying at 110 is not sufficient ; the precipitate must be gently ignited (see page 478). 7 For indirect method, claimed to be the most accurate, see R. L. y Gamboa, Anal. Fis. Quim., 10. 389, 1912. BARIUM, STRONTIUM, CALCIUM, AND MAGNESIUM. 517 Kammerer l first used potassium chromate for the qualitative precipitation of barium in the presence of calcium and strontium from solutions containing acetic acid and ammonium acetate ; Frerichs applied the reaction quantitatively ; and Russmann showed that very fair results could be obtained with the process. Fresenius investigated the conditions which favoured success and failure; and Skrabal and Neustadl have developed the process in its present form, whereby a satisfactory separation can be made. 273. The Determination of Barium in Insoluble Silicates. Decomposition by Fusion ivith Sodium Carbonate. It is not advisable generally to separate barium quantitatively from a solution which has previously been employed for the determination of the members of the hydrogen sulphide, ammonia, and ammonium sulphide groups, because of the leakage or loss of barium entailed during the separations. In the hydrogen sulphide group the main loss is due to the reduction of ferric chloride by hydrogen sulphide and oxidation of the sulphide to sulphate, 2 which means that barium sulphate will be precipitated. Curtman and Frankel found a loss of 14*7 ingrms. of barium in a solution containing 100 mgrms. of iron as ferric chloride per 100 c.c. when the ferric hydroxide was precipitated by ammonia, probably owing to the absorption of carbon dioxide by the ammonia. In the precipitation by ammonium sulphide, between 2 and 3 mgrms. of barium were also lost. Losses of barium may also occur owing to the presence of traces of sulphates in the reagents, and also to the retarding influence of ammonium salts on the precipitation of barium sul- phate or carbonate. Hence it is best to determine barium on a special sample, and not on the sample used for the determination of silica, iron, etc. On page 497, it may be remembered, the silicate was fused with sodium carbonate, the resulting "cake" was digested with water, and the residue treated with sulphuric acid for the determination of zirconium. The insoluble portion contained the barium, strontium, calcium, silica, etc. To determine the total barium, ignite the filter paper containing the insoluble residue. Fuse with a gram of sodium carbonate for 10 to 15 minutes. Dissolve the mass in warm water, filter and wash. Place a 250-c.c. beaker below the funnel, and dissolve the precipitate in dilute hydrochloric acid. Wash the filter paper well. Neutralise the filtrate with sodium carbonate and make the solution up to 150 c.c., so that the solution occupies at least 30 c.c. per gram of calcium, barium, and strontium oxides present. This prevents the precipitation of calcium sulphate later on. Precipitate the barium by the addition of a hot solution of dilute sulphuric acid (1 : 300) to the boiling solution. The addition is made gradually, with constant stirring. The reasons will appear from the discussion on page 618. Wash by decantation ; 3 collect the 1 H. Kammerer, Zett. anal. Chem., 12. 375, 1873; J. Meschezerski, ib., 21. 399, 1882; E. Fleischer, ib., 9. 97, 1870 ; R. Fresenius, ib., 29. 20, 143, 413, 1890 ; 30. 18, 452, 583, 1891 ; 32. 183, 312, 1893; A. Skrabal and L. Neustadl, ib., 44. 742, 1906; W. Fresenius and F. Ruppert, ib., 30. 672, 1891; A. Russmann, ib., 29. 447, 1890; Chem. News, 63. 13, 44, 1891 ; P. Schweitzer, Proc. Amer. Assoc., 187, 1877 ; H. Baubigny, Bull. Soc. Chim. (3), 13. 326, 1895 ; H. Baubigny, ib. (4), i. 55, 1907 ; H. Caron andM. Raquet, ib. (3), 35. 1061, 1906 ; F. Frerichs, Her., 7. 800, 956, 1874 ; H. Robin, Ann. Chim. Anal., 18. 445, 1903 ; B. Kahan, Analyst, 33. 12, 1908 ; J. L. M. van der Horn van der Bos, Chem. WeekUad, 8. 5, 1911 ; 9. 1002, 1912 ; E. Beyne, Chimiste, 3. 256, 1912 Sr in zinc blendes. 2 L. J. Curtman and E. Frankel (Journ. Amer. Chem. Soc., 33. 724, 1911) found a loss of 1'4 mgrms. from this cause; A. A. Noyes and W. C. Bray (ib. t 29. 137, 1907) state that the presence of 500 mgrms. of FeCl 3 in solution leads to the precipitation of as much as 20 mgrms. of barium as sulphate in the hydrogen sulphide group. 3 The nitrate contains some silica, titanium, iron, aluminium, niobium, tantalum, and tin, if these elements be present. 518 A TREATISE ON CHEMICAL ANALYSIS. precipitate in a small filter paper ; and ignite tha barium sulphate, with the precautions indicated on page 616. The weight of the barium sulphate multi- plied by 0*6570 gives the corresponding amount of barium oxide BaO. Purification from Calcium. If much calcium be present, the ignition should be made in a platinum crucible, and the ignited precipitate fused with sodium carbonate. The resulting " cake " is treated with water, and acidified with acetic acid, which is added drop by drop. The barium is precipitated by the addition of sulphuric acid as before. 1 Hydrofluoric Acid Decomposition. The total barium is conveniently determined in silicates 2 by treating 2 grms. of the finely powdered, dry (110) silicate with 10 c.c. of sulphuric acid (1 : 4) and 5 c.c. of hydrofluoric acid in a large platinum crucible. Evaporate the solution on a water bath ; add more hydrofluoric acid ; and repeat the evaporation. If no sandy grains can be detected with the platinum spatula, further treatment with the hydrofluoric acid is not necessary. Heat the mixture until most of the sulphuric acid has been driven off. Let the crucible cool, and pour its contents into, say, 25 c.c. of water. The precipitate of barium sulphate will probably be free from calcium, but a little strontium may be present. Filter off the barium sulphate. Ignite in a platinum crucible ; cool. Purification from Calcium and Strontium.-* The strontium and calcium can be removed by dissolving the barium sulphate in concentrated sulphuric acid, and again pouring the solution into water. If the amount of barium sulphate be less than about 0'002 grm., there is no need to purify the salt further. 3 Ignite the precipitated barium sulphate in a platinum crucible, and weigh as barium sulphate. This weight, multiplied by 0'6570, gives the corresponding amount of BaO. 274. The Complete Analysis of Limestones, Gault Clays, etc. Limestones, gault clays, marls, dolomite, magnesite, and similar carbonate rocks may be analysed from several different points of view ; and, in consequence, a more or less incomplete analysis may serve all requirements. Some abbrevi- ated methods will be described later. Dissolution of the Sample. A gram of the powdered and dried material is digested in a 100-c.c. beaker with 20 c.c. water, and 2 c.c. of concentrated hydrochloric acid and 2 drops of nitric acid. The acids are added slowly, and the beaker is kept covered by a watch-glass with its convex side downwards, so as to avoid loss during the effervescence. When effervescence has ceased, heat the solution to its boiling point on a hot plate, so as to drive off the carbon dioxide. Rinse the watch-glass with water. Filter the solution into an evaporating basin, and wash with water. Ignite the insoluble residue in a platinum crucible, 4 and fuse with a little sodium carbonate. Remove the fused mass with water and dilute hydrochloric acid : keep the crucible covered during the action to avoid loss by spurting. Add the solution to the main solution. If the limestone does not contain much more than about 5 per cent, alumina 1 Silica is not precipitated with barium sulphate from dilute solutions of sodium silicate. 2 With glazes, lead sulphate may be present. This can be removed by digestion with ammonium acetate as decribed on page 316. 3 If zirconium be present, it will be associated with the barium sulphate, and it must be removed as indicated on page 498. 4 The insoluble residue is sometimes reported as such, or as "siliceous minerals," "clay and sand," etc. As a matter of fact, the term "insoluble residue" is more or less ambiguous, since the attack on the siliceous minerals is dependent on the strength of the acid, and on the state of subdivision of the powder. The finer the sample is ground, the less the "insoluble residue." BARIUM, STRONTIUM, CALCIUM, AND MAGNESIUM. 519 and ferric oxide, and 15 per cent, of silica, strong ignition over a blast 1 in a covered platinum crucible will frequently give a powder wholly soluble in hydrochloric acid (1:1), except, possibly, a little flocculent silica which does not matter. If too much siliceous matter be present to render this treatment successful, as is sometimes the case with siliceous limestones, magnesian lime- stones, cement rocks, and highly calcareous marls, Meade 2 recommends igniting the limestone with just less than its own weight of sodium carbonate. The sintered, not fused, mass is then easily broken down by hydrochloric acid (1:1). Determination of Silica, Alumina, Titanium, etc. Evaporate 3 the combined nitrates to dryness for silica (page 167) ; determine the alumina, titanium, iron, phosphorus, 4 lime, 5 magnesia, and manganese 6 as described for clays (pages 177 seq.). The carbon dioxide (page 553), water (page 571), and the alkalies are determined on separate samples. 7 Chlorine and Fluorine. For the chlorine, digest, say, 10 grms. with water and nitric acid at a gentle heat. Filter, and proceed by the method of page 652. For the fluorine, digest 10 grms. of the dry powdered sample in acetic acid 8 without boiling or filtering until the magnesium and calcium carbonates are decomposed. Evaporate the solution to dryness to expel the excess of acetic acid, add a slight excess of sodium carbonate, extract the residue with water ; and the precipitate of calcium fluoride, carbonate, and insoluble residue is treated by the process of page 639. Determination of Sulphur. Sulphides as well as sulphates are often present in limestones and calcareous clays. To determine the "sulphate or soluble sulphur," digest, say, 2 grms. of the powdered sample in a large porcelain basin with 40 c.c. of water and 4 c.c of concentrated hydrochloric acid. Evaporate the solution to dryness. Boil the residue with water, filter and 1 T. Engelbach, Liebig's Ann., 123. 260, 1862. * R. K. Meade, Portland Cement, Easton, Pa., 254, 1906. :! The evaporation is more rapid than with cl;iys because of the absence of large amounts of alkaline chlorides, and a smaller volume of liquid is used for the dissolution of the fused cake. 4 Phosphatic limestones may have less alumina than is needed to combine with the phosphorus, and in that case some calcium phosphate will be precipitated with the alumina and iron hydroxide. Some, therefore, add a ferric salt to the solution before adding the ammonia, and afterwards deduct the amount of iron added from the ferric oxide obtained later see page 606. 5 In the case of magnesites where but little calcium and much magnesium is present, the oxalate process for lime is not suitable (F. Hundeshagen, Zeit. o/ent. Chem., 15. 85, 1907). It is best to treat the filtrate occupying, say, 30 c.c. from the silica with 4 grms. of sodium sulphate, and 40 c.c. of 90 per cent, alcohol. After standing about five hours at 17-20 in a covered vessel, filter the solution, and wash the precipitated calcium sulphate with 50 per cent. Jilcohol. In order to free the precipitate completely from magnesia, redissolve the sulphate in hot dilute hydrochloric acid, and precipitate the lime as usual by the oxalate process. The combined filtrates from the calcium sulphate and oxalate are evaporated to drive off the alcohol, the iron is oxidised with hydrogen peroxide, and aluminium and ferric hydroxides precipitated by add- ing ammonia. Magnesia is determined in the filtrate in the usual manner. 6 The manganese will be found in the filtrate from the ammonia precipitate ; or it may be precipitated with the alumina, etc., by adding bromine as indicated on page 177. Copper, lead, zinc, nickel (sulphides or carbonates), rare earths, chromium, vanadium, etc., are to be determined on large quantities of the sample say 50-500 grms. . 7 As a rule, in Smith's process, about half as much calcium carbonate is needed for lime stones and the more calcareous clays as is indicated in the standard directions (page 222). The precipitated calcium carbonate is much more effective in the work of decomposition than the native crystalline carbonate. If no precipitated carbonate be used, the results, with Smith's process, are usually rather low. T. Engelbach (Liebig's Ann., 123. 260, 1862) uses a modifica- tion of Berzelius' process for the alkalies in the "soluble" portion of limestones. Ignite the powdered mineral strongly over a blast, boil with a little water, filter, neutralise the solution with hydrochloric acid. Treat with ammonia and ammonium carbonate, etc. (page 226). 8 G. Jenzsch, Fogg. Ann., 96. 145, 1855. 520 A TREATISE ON CHEMICAL ANALYSIS. wash. 1 Precipitate the sulphuric acid in the filtrate as barium sulphate by the method of page 618. The weight of barium sulphate so obtained, multiplied by 0*3430, represents the corresponding amount of sulphur trioxide S0 3 ; and when multiplied by 0-5833, the corresponding amount of calcium sulphate. To get the "total sulphur," place 2 grms. of the powdered sample in a porcelain evaporating basin, and cover the mass with bromine water. Decompose the carbonates by adding 25-30 c.c. of hydrochloric acid, in small quantities at a time, to the cold solution. The " sulphide sulphur " is liberated as hydrogen sulphide, and immediately oxidised by the bromine water. If the acid be added gradually, and the bromine water be in excess, there will be no appreciable loss of sulphide. The solution is evaporated to dryness, filtered, and treated with barium chloride, etc., for barium sulphate, as indicated on page 618. The difference between the weights of barium sulphate obtained with and without the bromine water, multiplied by 0-1374, represents the corresponding amount of "sulphide sulphur." 2 EXAMPLE. Suppose the following results have been obtained : Barium sulphate (with bromine) . . '. 0-0341 grm. Barium sulph-xte (without bromine) . * . . . .' . . '0251 grm. Barium sulphate 0'0090 grm. Sulphide sulphur (0-0090x0 -1374) ..... . . . 0*0012 grm. Sulphur trioxide (0-0251x0-3430) . . . . ,_ . . '0086 grm. Calcium sulphate (0-0251x0 -5833) .' -,.'.-', . . . 0-0146 grm. Sometimes the total sulphur is determined by one of the methods described on pages 621 et seq. Or the powdered sample can be fused with sodium carbonate and sodium nitrite (page 46 1). 3 Extract the mass with water. Evaporate the solution to dryness, acidify with hydrochloric acid to dryness. Boil the residue with dilute hydrochloric acid, filter, and determine the sulphates in the filtrate by the method of page 618. Organic Matter is found in nearly all limestone rocks. Schaffgotsch's process for the determination of organic matter is to fuse, say, 5 grms. of borax glass in a platinum crucible over a Bunsen's burner 4 until the weight of the crucible and contents is constant. 5 Place 1 grm. of the sample on the borax, put the lid of the crucible in position, and fuse over the Bunsen's burner. Do not remove the lid until the contents are fused, otherwise loss by decrepitation 1 If phosphorus is to be determined, it is sometimes advisable to take double the amount of the sample for the operation just described, and now make the nitrate up to 100 c.c. Take 50 c.c. for the sulpbur determination, and to the other 50 c.c. add 5 c.c. of nitric acid, and determine the phosphorus by Woy's process (page 595). 2 Pyrite FeS 2 is the commonest sulphide in limestone rocks. Some of the sulphur may come from the organic matter. a A. Petzboldt, Journ. prakt. Chem. (1), 63. 194, 1854; J. Roth, ib. (1), 58. 84, 1853; J. .1. Ebelmen, Compt. Mend., 33. 881, 1851 ; H. St C. Deville, ib., 37. 1001, 1853. 4 Note, if a blast be used, the crucible and contents must be afterwards brought to constant weight over the Bunsen's burner. Borax glass can be kept in a state of fusion some 15-30 minutes without appreciable volatilisation, but there is a decided loss after a few minutes' blasting (page 578). R. Fresenius, Zeit. anal. Chem., I. 65, 1862; A. Mitscherlich, Journ. prakt. Chem. (1), 83. 485, 1861. 5 According to 0. Lutz and A. Tschischikow (Journ. Russ. Phys. Chem. Ges., 36. 1274, 1904), microcosmic salt may be used in place of borax. The salt is heated until ammonia and water have been expelled and the weight is constant. The advantage of microcosmic salt is : (1) it fuses more quickly ; (2) is not so liable to loss by volatilisation. H. Rose (Pogg. Ann., 116. 131, 686, 1862 ; Zeit. anal. Chem., I. 183, 1 862) recommends potassium dichromate; W. Bottger (Zeit. anal. Chem., 49. 487, 1910) recommends sodium metaphosphate (see page 661); F. A. Gooch and S. B Kuzirian (Amer. J. Science (4), 31. 497, 191 1) recommend sodium paratungstate ; and H. Rose (Pogg. Ann., 116. 635, 1862 ; T. W. Richards and E. H. Archibald, Proc. Amer. Aead., 38. 443, 1903) recommends silica. BARIUM, STRONTIUM, CALCIUM, AND MAGNESIUM. 521 may occur. When the evolution of gas has ceased, cool in a desiccator, and weigh. Again ignite, cool, and weigh. -When the weight is constant, the differ- ence between the weights of the crucible and contents with and without the powdered sample represents the carbon dioxide, water, and organic matter. If the carbon dioxide and water have been previously determined, the amount of organic matter follows by difference. Petzholdt l recommends the following process : Dissolve 20 grms. of the sample in dilute hydrochloric acid, and boil the solution carefully to expel carbon dioxide. Filter any undissolved residue through ignited asbestos, and wash well with water, dry, and transfer the asbestos to a porcelain boat, which is then placed in a combustion tube charged with copper oxide (page 563) and ignited in a current of oxygen, so that the carbon dioxide can be collected in weighed potash bulb ; or the carbon may be determined by the wet process (page 546). 275. The Partial Analysis of Limestones, Dolomites, Magnesites, Marls, Cements, etc. A process which occupies about three hours may now be described. The abbreviated processes are not usually quite so exact as the preceding process, but they are quite accurate enough for many purposes, particularly when a great number of analyses of one type of carbonate rock have to be made rapidly. 2 See the remarks on pages 243 and 248. 1. Silica. Cautiously add 5 c.c. of concentrated hydrochloric acid to a gram of the powdered sample, keeping the basin covered during the attack by the acid so as to avoid loss by spurting. When the effervescence has ceased, evaporate the mixture to dryness. 3 Cool. Digest the residue with a little hot water and a few drops of hydrochloric acid. Heat the mixture to boiling and filter. Wash the residue on the filter paper with hot water. Ignite, and weigh as Si0 2 . 2. Alumina and Ferric Oxides. Precipitate the mixed ferric and aluminium hydroxides with ammonia and ammonium chloride in the usual manner (page!82), and finally weigh as A1 2 3 + Fe 2 3 . 3. Calcium Carbonate. Acidify the filtrate from the ammonia precipitate with hydrochloric acid. Heat the solution to boiling, and, while still boiling, gradually add, crystal by crystal, 4 approximately 3 grms. of solid oxalic acid per gram of mixed calcium and magnesium oxides. 5 When all the oxalic acid has been added, add a slight excess of ammonia. Let the solution stand until the precipitate has settled -2 or 3 minutes. 6 According to Schoch, 7 a double pre- 1 A. Petzholdt (Journ. pralcl. Chem. (1), 63. 194, 1854) "considers that 58 parts of carbon correspond with 100 parts of humus. If carbon and hydrogen are both determined by com- bustion, 4 '5 parts of hydrogen correspond with every 58 parts of carbon. The remaining hydrogen is supposed to be derived from the water. Note that hydrogen may come from the asbestos of the filter tube if the drying be imperfect, and also from the hydrated minerals. It is generally best to represent the result as "carbon derived from organic matter," without making any assumptions as to the composition of the organic matter. See page 573. 2 K. J. Sundstrom, Journ. Soc. Chem. Ind., 16. 520, 1897 ; F. Hundeshagen, Zeit. offent. Chem., 15. 85, 1909; N. Knight, Chem. News, 92. 61, 108, 1905; 96. 126, 1906; F. Clowes and J. B. Coleman, ib., 92. 259, 1905 ; W. H. Stanger and B. Blount, Journ. Soc. Chem. Ind., 21. 1216, 1902 ; Report, ib.; 21. 12, 1223, 1902 ; G. Hentschel, Chem. Ztg., 36. 821, 1912. :i When the solution is evaporating, determine the mixed calcium and magnesium carbonates as described below. 4 Otherwise loss might occur from spurting. Stir all the time the acid is being added. 5 The magnesium must all be converted to the oxalate to prevent the solution of calcium oxalate in the magnesium chloride (page 212). The approximate amount is determined from " 4 " below. (J For more exact results, let stand overnight, or at least 3-4 hours. 7 C. Schoch, Die moderne Aufbereitung und Wertung der Mortel-materialien, Berlin, 36, 1896; W. 0. Blasdale, Journ. Amer. Chem. Soc., 31. 917, 1909; E. H. Schultze, Chem. Ztg , 522 A TREATISE ON CHEMICAL ANALYSIS. cipitation is unnecessary ; but if much magnesium be present, it is better to dissolve the washed calcium oxalate in hot hydrochloric acid, and reprecipitate by the addition of ammonia and a little ammonium oxalate to the boiling solution. Filter the calcium oxalate, and wash with cold water until the filtrate is free fronPcnlorides and oxalates. 1 Igrrfte the precipitate in a platinum crucible, and finish the ignition with half aTThour's blasting. Cool in a desiccator, and weigh as CaO. 2 Multiply the result by 1'7844 to get the corresponding amount of calcium carbonate 4. Mixed Calcium and Magnesium Carbonates. Weigh 1 grm. of the finely powdered sample into a small porcelain basin. Add 25 c.c. of N-HC1. Cover with a watch-glass, and, when effervescence has ceased, heat the solution to boiling. When cold, titrate the free acid with NJIaOH and with a drop of methyl orange as indicator (page 70). EXAMPLE. One gram of limestone required 5'15 c.c. of N-sodium hydroxide for the titration after adding 25 c.c. of N-hydrochloric acid. Hence, the mixed calcium and magnesium carbonates corresponded with 19'85 c.c. of N-HC1. Again, 0538 grm. of CaO was obtained. This corresponds with 1 "7844 x 0-538 = 09603 grm. CaC0 3 . But 1000 c.c. of N-HC1 correspond with 28 grms. of CaO ; hence, 1 grm. CaO corresponds with 35-7 c,c. of N-HC1 ; or 0'538 grm. CaO corresponds with 0'538x35'7 = 19-21 c.c. of N-HC1. Hence, 19'85 - 19'21 =0'64 c.c. of N-HC1 was needed for the magnesium carbonate. But 1 c.c. of N-HC1 represents - 042 grm. MgC0 3 , and 0'64 c.c. corresponds with 0-042 x 0-64 = 0-0269 grm. of MgCO 3 . Hence, the analysis reads (per cents.) : Calcium carbonate V .96*03 Magnesium carbonate . . . . . . . . 2 '69 Silica 1-06 Alumina and ferric oxide . . . . . . . .0*00 The magnesium carbonate determined by this indirect process should not differ more than \ per cent, from that obtained by more elaborate processes. Passon' s Process for the Determination of Lime in the Presence of Alumina, etc. If the amount of lime only is needed, Passon 3 says that calcium oxalate can be precipitated in the presence of iron, aluminium, magnesium, and phos- phorus by adding a drop of phenolphthalein to the solution, and neutralising with dilute ammonia (1:5) until the precipitate is no longer redissolved on shaking. Add 25 c.c. of Wagner's solution. 4 If the precipitate does not all dissolve, repeat the neutralisation with ammonia. Add 12 c.c. more of Wagner's solution ; dilute the solution to 200 c.c. with water. Heat the solution to boiling, and add solid ammonium oxalate until no more precipitate is produced. After standing some time (overnight), filter, wash, and ignite the calcium oxalate. 5 Newberry's Process for the Simultaneous Determination of Magnesium and Calcium Carbonates. Newberry 6 based a process for the volumetric determination v : 29. 508, 1905 ; M. J. van Kruijs, Chem. Weekblad, 4. 29, 1907 ; C. Liesse, Ann. Chim. Anal., 16. 7, 1911 ; B. Enright, Journ. Amer. Chem. 8oc., 26. 1003, 1904; C. Stolberg, Zeit. angew. Chem., 17. 741, 769, 1904 ; R. F. Young and B. F. Baker, Chem. Neivs, 86. 148, 1902. 1 Magnesium can be determined in the combined nitrates, if desired, by the method of page 218. W. F. Koppeschaar, Zeit. anal. Chem., 44. 184, 1905. The calcium oxalate can ) also be determined by the volumetric permanganate process (page 215). H. B. Kinnear, Chem. Eng., 13. 247 : 1911. 2 The CaO can be dissolved in 25 c.c. of N-HC1 and titrated with N-NaOH in the usual manner (page 70), as a check on the weighings. 3 M. Passon, Zeit. angew. Chem., n. 776, 1898 ; 12. 48, 1899 ; 14. 285, 1901. 4 WAGNER'S SOLUTION. Dissolve 25 grms. of citric acid and 1 grm. of salicylic acid in water, and make the solution up to a litre. 5 R. K. Meade (Chem. Eng., I. 21, 1905; Portland Cement, Easton, Pa., 189, 1906) and K. Balthaser (Chem. Ztg., 33. 646, 1909) describe volumetric processes for the determination of lime without the separation of silica, etc. 6 S. B. Newberry, Cement Eng. News, 15. 35, 1903 ; Tonind. Ztg., 27. 833, 1903. BARIUM, STRONTIUM, CALCIUM, AND MAGNESIUM. 523 of calcium and magnesium carbonates on the fact that magnesium hydroxide is sufficiently soluble in water to colour phenolphthalein ; and whe'n boiled with a dilute solution of sodium hydroxide, the magnesium is completely precipitated and separated from the calcium. Dissolve half a gram of limestone in an Erlenmeyer's flask fitted with a long "condenser tube" (tig. 185, page 582) to serve as condenser. Add 60 c.c. of iN-hydrochloric acid ("first acid"), 1 and boil the solution for 2 minutes with the condenser tube in position. Wash the interior of the condenser tube into the flask by means of a wash-bottle. Remove the tube and cool the solution thoroughly in the cooling-box (page 72). Add 5 drops of phenolphthalein solution and titrate with i N-sodium hydroxide ("first alkali ") until a faint pink colour appears in the solution. This may fade in a few seconds.' 2 Transfer the neutral solution to a large test tube (12 inches long and 1 inch internal diameter) provided with a mark 3 corresponding with 100 c.c Heat the solution to boiling, add iN-NaOH, J c.c. at a time, and boil after each addition. When a deep red colour is obtained which does not become paler on boiling, read the burette ("second alkali"). Dilute the solution to 100 c.c.; boil for a moment ; and let the precipitate settle. When the precipitate has settled, pipette 50 c.c. of the clear solution into a flask and titrate with 4N-HC1 until the pink colour has gone (" second acid"). 4 EXAMPLE OF CALCULATION. Suppose 11 '6 c.c. alkali (" first alkali ") \\ere needed for the first titration, cold ; and 3'55 c.c. of alkali ("second alkali") were needed for the second titration, hot. Suppose, further, that 50 c.c. of the solution required 0'45 c.c. of acid for decolorisation (" second acid "). Then, (Second alkali - second acid)2 x 0'84 = 1-68(3-55 - 0'90) = 4-45 per cent, of magnesium carbonate. {First acid - (First alkali + second alkali - second acid)}2 x 1*00 = 2{60-(1 1-60 + 3-55 -0-90)} = 91'50 per cent, of calcium carbonate. If the results are to be expressed in terms of MgO and CaO, substitute 0'40 for 0'84, and 0'56 for TOO, in the preceding formulae. The results for magnesium carbonate are usually a little low, and for lime a little too high. This is probably due to the formation of calcium carbonate by the absorption of carbon dioxide of the air during the precipitation of magnesia, and to the insolubility of the lime compounds in the sample. If much soluble alumina and iron be present, the end point is obscured and the readings are more difficult. With the high-grade limestones the results are suitable for technical work. With marble containing 54'62 per cent. CaO and 0'84 per cent. MgO by the gravimetric process, 54'99 and 0'88 respectively were obtained by Newberry's process. Witli a magnesian marl containing 5*43 per cent, of alumina and ferric oxide, 52'08 per cent, of silica, 19-52 per cent, of CaO, and 2'24 per cent, of magnesia, Newberry's process gave : 16-83 ^av^ MgO ... A sample of dolomitic marl w CaO MgO ... 2-08 2-30 2-00 percent, th 26-50 per cent. CaO and 18'76 MgO : . 26-38 25-42 25'53 percent. . 17-80 18-40 18-96 percent. 1 Standardised against Iceland spar. 2 If the titration be carried until A permanent red coloration appears, the lime will be too low. If the limestone under investigation contains inappreciable amounts of magnesia, this completes the determination : (first acid - first alkali) x 2 x 0'56 = the per cent, of CaO. n A band of paper or etched ring. 4 If more than 1 c.c. of acid be required, too much alkali was added, and the magnesium hydroxide will have carried down some Ca(OH) 2 . 524. A TREATISE ON CHEMICAL ANALYSIS. were obtained. Experiments with artificial mixtures of Iceland spar and magnesium carbonate gave excellent results. The results with the less pure lime- stones and calcareous marls are not so good. Magnesite. In the case of magnesites, 1 the joint magnesia and lime 2 may be determined by boiling 0'5 grin, of the magnesite with a known excess of standard sulphuric acid and titrating back the excess with standard sodium hydroxide. The lime can be precipitated from the solution by adding an excess of 90 per cent. alcohol 250 c.c. of alcohol per 100 c.c. of the solution. Let the mixture stand in a covered vessel overnight. Filter and wash with 50 per cent, alcohol, and proceed as indicated in the footnote, page 212. After subtracting the calcium oxide from the joint magnesium and calcium oxides obtained by the titration, the amount of magnesia follows at once. The magnesia may exist in the form of oxide, hydroxide, or carbonate. 3 The hydroxide can be determined from the loss in weight which occurs when the sample is baked to a constant weight at 300-350 8 , and the carbonate from the loss in weight which occurs when the same sample is calcined to constant weight at a red heat. The difference between the amounts so determined and the total magnesia represents the magnesium oxide. Of course, the water and carbon dioxide can be determined by the more laborious processes, pages 559 et seq.* The value of burnt magnesite depends upon the proportion of "active magnesia.'' The "active magnesia" is the difference between the total magnesia, MgO, and that combined with carbon dioxide and water. 5 276. The Mineralogical Analysis of Limestones and Marls. Calcium carbonate is undoubtedly the prevailing constituent of limestone rocks, and magnesium carbonate is probably next in importance. With increasing proportions of the latter, the carbonate rock passes into dolomite, magnesian limestone, and finally magnesite. Ferrous and manganese carbonates are probably present in small quantities, and when these compounds predominate, siderite and other ores of iron and manganese result. It is usual, in technical analyses, to report the lime and magnesia as carbonates, and the iron and 1 For volumetric process, Chem. Ztg., 33. 545, 1909. - If the amount of calcium relatively with the magnesium is very small, 0. Kallauner and I. Preller (Chem. Ztg., 36. 449, 462, 1912) show that the oxalate process (pages 212-3) is trust- worthy only if dilute solutions are used, and if a large excess of ammonium salts and ammonium oxalate are present, and if the solution is filtered immediately after precipitation ; or if the pre- cipitation is repeated : otherwise the lime is always 0*15 to 0'2 per cent, low, owing to resolution during washing. Most methods involving the precipitation of calcium sulphate in alcoholic solutions are unreliable if much magnesia is present. Good results are obtained as follows : (1) Precipitate the lime as oxalate, filter and wash, convert the oxalate into sulphate, dissolve the sulphate in hydrochloric acid, reprecipitate as oxalate, and weigh as CaO. Or (2), evaporate a solution of the mixed chlorides to dryness with lithium sulphate, treat the residue with a mixture of 10 volumes of ethyl alcohol with 90 volumes of methyl alcohol saturated with lithium sulphate. The precipitated calcium sulphate is dissolved in hydrochloric acid and re precipitated by the oxalate process E. C. Carron, Ann. Chim. Anal., 17. 127, 1912. 3 For the total magnesia in magnesites, J. Mayrhoffer (Zeit. angeiv. Chem., 21. 592, 1908) decomposes 5 grms. of the finely powdered mineral with aqua regia on a water bath. Evaporate the solution to dryness to render the silica insoluble. Digest with acidulated water, filter, arid make the solution up to a litre ; mix 40 c.c. (20 c.c. if burnt magnesite is in question) with 5 c.c. of sulphuric acid and 100 c.c. of citrate solution (100 grms. citric acid, 333 c.c. of ammonia sp. gr. 0-91 made up to a litre with water), 20 c.c. of a 10 per cent, solution of disodium hydrogen phosphate, and 15 c.c. of ammonia. Stir the mixture 5 minutes without touching the walls of the beaker. Filter, etc. , after standing 5 minutes. The calcium is not precipitated in presence of the ammonium citrate. 4 For the determination of magnesia in the presence of magnesium carbonate from the heat of the reaction with hydrochloric acid, see V. Fortini, Chem. Ztg., 36. 270, 1912. 5 0. Kallauner, Chem. Ztg., 36. 711, 1912 ; L. Dede. BARIUM, STRONTIUM, CALCIUM, AND MAGNESIUM. 525 manganese as oxides. In all probability, the two latter more generally exist as carbonates. A determination of the carbon dioxide will sometimes show that more carbon dioxide is present than is required for the lime and magnesia. Some of the lime and magnesia may also be combined with the silica and alumina, and, in consequence, even if the amount of carbon dioxide does satisfy the lime and magnesia, some may be combined with the iron and manganese. Phosphorus may be combined with the iron and aluminium, or be present as apatite, etc. The silicate minerals associated with the calcareous clays and limestones are prob- ably as varied as the silicate minerals found in clays (q.v.). 1 Many attempts have been made to isolate the different minerals in limestones and marls by treatment with different solvents. Struckmann 2 analysed various marls by digesting the samples first with water, then with acetic acid, and finally with hydrochloric acid. Each solution and the residue was examined separately. Bolton 3 studied the action of citric and tartaric acids, and Jannettaz, 4 the action of potassium bisulphate. Deville 5 tried to dissolve calcium carbonate from mortars by boiling the mixture with a solution of ammonium nitrate, which he supposed did not attack the silicates ; but Gunning showed that Deville's assumption is faulty. Browne and Harrison's 6 method of dealing with the problem is as follows : Clay and Quartz. Digest 5 grms. of the dried (air) and powdered sample in dilute hydrochloric acid (2'5 per cent.) 7 in the cold (30). When effervescence has ceased, decant the solution through a filter paper. Heat the residue on a water bath several hours with a concentrated solution of sodium carbonate containing some sodium hydroxide, 8 and wash the residue at the filter pump (1) with the same solution ; (2) with water ; and (3) with dilute hydrochloric acid. Ignite, and weigh the residue as " clay and siliceous minerals." Reduce the residue to an impalpable powder in an agate mortar, and boil a weighed portion repeatedly with concentrated sulphuric acid until the clay is quite decomposed. Drive off the excess of acid, and heat the residue with dilute hydrochloric acid to dissolve basic aluminium and ferric sulphates. Heat the residue with the sodium carbonate solution to dissolve the silica set free during the decomposition of the clay. The final residue represents the crystalline 1 W. F. Hurne, Chemical and Mineralogical Researches on the Upper Cretaceous Zones of the South of England, London, 1893; S. Pfaff, Ueber die unloslichen Bestandtheile der Kalk und Dolomite, Halle, 1878 ; L. Cayeux, Ann. Soc. Geol. du Nord, 16. 342, 1890 ; A. Vesterberg, Bull. Geol.Inst. Upsala, 5. 97, 1900 ; 6. 254, 1904 ; K. Haiishofer, Sitzber. K.K. Akad. Wiss. Miinchen, u. 220, 1881. 2 C. Struckmann, Liebig's Ann., 74. 170, 1858 ; C. J. B. Karsten, Archiv Min., 22. 592, 607, 613, 1848 ; E. Damour, Bull. Soc. Geol. (2), 6. 313, 1849 ; G. Forchhammer, Journ. prakt. Chem. (1), 49, 55, 1850 ; J. Roth, Zeit. deut. geol. Ges., 4. 565, 1852 ; T. Liebe, ib., 7. 406, 1855 ; F. Hoppe-Seyler, ib., 27. 499, 1875 ; F. Pfaff r Fogg. Ann., 82. 488, 1851 ; T. S. Hunt, Amer. J. Science (2), 28. 181, 371, 1859; M. Schafhautl, Neues Jahrb. Min., 812, 1864 ; C. Knausz, Chem. Centr. (1), 26. 244, 1855 ; G. Doelter and R. Homes, Jahrb. yeol. Reichsanst., 69, 1875 ; C. Schmidt, N. Petersb. Acad. Bull., 16. 205, 1871. 3 H. C. Bolton, Ann. N. Y. Acad. Science, I. 1, 153, 1877-80 ; Proc. Amer. Assoc. Adv. Science, 31. 271, 1883 ; Chem. News, 36. 249, 1877 ; 37. 14, 24, 65, 86, 98, 1878 ; 38. 168, 1878 ; 43. 31, 39, 1881 ; 47. 251, 1883 ; J. W. Richards and N. S. Powell, Journ. Amer. Chem. Soc., 22. 117, 1900. 4 E. Jannettaz, Campt. Rend., 77. 838, 1873 ; 78. 852, 1874. . 5 H. St C. Deville, Compt. Rend., 37. 1001, 1853; J. W. Gunning, Journ prakt. Chem., (1), 62. 318, 1854. 6 A. J. J. Browne and J. B. Harrison, "On the Geology of Barbados, " Journ. Geol. Soc., 48. 170, 1892. 7 Dilute acetic acid was tried, but it only partially attacked the zeolites, and the silica set free from these minerals was estimated as colloidal silica. 8 LUNGE'S SOLUTION. Dissolve 100 grms. of crystalline sodium carbonate and 10 grms. of sodium hydroxide in water, and make the solution up to a litre with water. Remember that the solution will take silica from the glass of the bottle if it be stored in glass bottles. 526 A TREATISE ON CHEMICAL ANALYSIS. siliceous minerals, mostly quartz. The difference between this result and the preceding, multiplied by M64, represents the amount of clay, provided it be assumed that the clay can be represented by Al 2 03.2Si0 2 .2H 2 0, and that ignited clay is all dissolved by the treatment in question. For the limitations involved in the assumption see pages 661 et seq. Colloidal Silica. Treat 5 grms. of the (air) dried and powdered sample in dilute hydrochloric acid (2 '5 per cent.) in the cold (30). Evaporate the solution to dryiiess in a porcelain basin, with the usual precautions against loss by spurting (page 167), in order to render the silica insoluble Digest the residue in warm hydrochloric acid, add water, filter and wash. Ignite the residue and weigh. The insoluble matter consists of the residue insoluble in dilute hydrochloric acid, together with the colloidal silica originally present in the sample, and that formed by the decomposition of certain hydrated silicates and aluminium silicate, which might be present in small quantities. The difference between the weight of this mixture and the " clay and quartz " resulting from the preceding operation represents the colloidal silica. l EXAMPLE. Suppose that 5 grms. of gault clay furnished : Clay and quartz residue . .. . . ".". 0'25 14 grm. Quartz residue . . .... . . '1255 grm. Calcined clay 1259 grra. Siliceous residue . . . . ' ... 0'3819 grm. Clay and quartz residue ........ 2514 grm. Colloidal silica '1305 grm. The calcined clay corresponds with 0'1259x 1'164 = 0'1465 grm. of clay. Hence, the sample would be reported to contain : Quartz residue . . . . . '. . . 2 '51 percent. Clay residue . . 2 '93 per cent. Colloidal silica . ... . . - . . . 2'61 per cent. Alumina, ferric oxide, etc., can be determined in the filtrate from the silica in the usual manner. For the determination of sand, see " Elutriation " in the second volume of this work. 2 Clay in Limestones. A number is obtained by the following process which is generally supposed to represent the amount of " clay " in the given limestone : Digest 2 grms. of the powdered sample with 40 c.c. of water and 4 c.c. of concentrated hydrochloric acid as indicated on page 525. When effervescence has ceased, boil the solution for about 10 minutes to drive off" the carbon dioxide. Precipitate any iron and aluminium which may have gone into solution by the addition of ammonia. Filter through a Gooch's crucible, wash, and ignite at a low red heat. Between one and two hours are required for the determination. Archetti 3 considers that, if much magnesia be present, the results will be high, because some magnesium hydroxide will be precipitated by the ammonia. In that case, consequently, Archetti prefers to treat 2 grms. of the sample with 10 per cent, hydrochloric acid mixed with y^th its volume of 10 per cent, nitric acid. When solution is complete, add an excess of ammonium chloride and a slight excess of ammonia. Wash the precipitate, etc., as before. 1 The term "colloidal silica" is here synonymous with "silica soluble in dilute hydro- chloric acid." 2 Note that burnt limestones and burnt magnesites may contain more " soluble silica " than the raw material, owing to the formation, during calcination, of silicates, which can be decom- posed by treatment with acids. 3 A. Archetti, Bull. Chim. Pharm., 48. 409, 1909. BARIUM, STRONTIUM, CALCIUM, AND MAGNESIUM. 527 277. The Determination of Free Lime in Quicklime, Mortars, etc. Free lime may be detected in whiting by boiling a couple of grams for about 5 minutes with 100 c.c. of water. An alkaline reaction development of a red colour with phenolphthalein is generally taken to represent free lime, although, of course, the coloration may possibly be due to the partial hydrolysis of the carbonate, and to the presence of alkaline hydroxides, or carbonates in the original sample. Winkler' s Process. To determine the free lime in quicklime, 1 carefully slake 10 grms. of the sample with water, and transfer the resulting cream to a 250- c.c. flask and make the contents up to the mark with distilled water. Pipette 25 c.c, ( = 1 grm. of the sample) of the thoroughly mixed contents into an Erlenmeyer's flask; add a few drops of phenolphthalein 2 as indicator, and titrate slowly with N-hydrochloric acid until the red colour disappears. The colour change takes place before the calcium carbonate is attacked by the acid, and it is possible to strike the point where all the calcium hydroxide is just neutralised. The small amount of carbon dioxide liberated by the next drop or two of the standard acid decolorises the phenolphthalein, and indicates the end of the titration. The acid must be added very slowly with constant stirring to get reliable results. Stone and Schenck's Process. 3 This is based on the fact that calcium oxide forms soluble compounds with a solution of sugar under conditions where calcium carbonate, alumina, and ferric oxides remain unaffected. Hence, shake, say, 1 grm. of the powdered sample with 500 c.c. of a 10 per cent, solution of cane sugar for 20 minutes. Filter, and wash with the solution of sugar. The lime in the nitrate can be determined either by precipitation as calcium oxalate, or volumetrically by titraiion with standard hydrochloric acid. The authors say that " magnesia is not soluble to any appreciable extent in the solution of cane sugar containing lime, under the stated conditions." Fruhling's Method* 5 grms. of the powdered sample are introduced into a stoppered Erlenmeyer's flask. Add 100-150 c.c. of water, and 30 drops of phenolphthalein solution. Shake the contents thoroughly. Wash the sides of the flask and the stopper. Add, say, 30 c.c. of standard hydrochloric acid 5 from a burette. The red colour of the solution may disappear, but it will re- appear if the flask be agitated, because "quicklime" generally contains more than 60 per cent, of CaO. Add a gram of ammonium chloride. The remaining calcium hydroxide decomposes the ammonium salt, liberating an equivalent amount of ammonia. This facilitates subsequent titration. Add the hydro- chloric acid 1 c.c. at a time until the red colour no longer reappears when the flask is agitated. As soon as the red colour takes an appreciable time to return, 1 Also, of course, in slaked lime. A. Winkler, Journ. prakt. Chem. (1), 67. 444, 1856 : W. Richter, Ton-bid. Ztg., 27. 1862, 1943, 1903. 2 A. Gawalovski, Zeit. anal. Chem., 22. 397, 1883 ; F. W. Kiister, Zeit. anorg. Chem., 13. 141, 1897 ; G. Lunge, Zeit. angew. Chem., 10. 41, 1897 ; C. Winkler, Praktische Uebungenin der Maassanalyse, Leipzig, 1888. For phenacetolin as indicator, see page 63. 3 W. E. Stone and F. C. Schenck, Journ. Amer. Chem. Soc., 16. 721, 1894; Chem. News, 70, 278, 1894 ; J. Hendriclc, Analyst, 32. 320, 1907. For the solubility of lime in saccharine solutions, see J. Weisberg, Bull. Soc. Chim. (3), 21. 773, 1899; 23. 740, 1900; Chem. News 82. 284, 1900. See also W. Heldt, Journ. prakt. Chem. (1), 94. 129, 1865; L. C. Levoir, Rec. Travs. Pays-Bas, 5. 59, 1886; C. Parsons, Dent. Top. Zeig. Ztg., 2. 585, 1888; W. P. Mason, Journ. Amer. Chem. Soc., 16. 733, 1894; 0. Rebbufat, Tonind. Ztg., 23. 782, 823, 883, 900, 1899 ; Gcasz. Chim. Ital., 28. ii, 209, 1899. 4 R. Friihling, Tonind. Ztg., 8. 393, 1884 ; H. Seger and E. Cramer, ib., 31. 619, 1907. 5 1000 c.c. of the hydrochloric acid have 130'3 grms. of HC1, and this is equivalent to 100 grms. of CaO. Hence, 1 c.c. of the standard solution corresponds with 01 grm. of CaO. 528 A TREATISE ON CHEMICAL ANALYSIS. the acid is added O'l c.c. at a time. If the red colour does not appear after the agitated solution has stood 5 minutes, read the volume of acid used, multiply the result by 2 to get the corresponding amount of CaO in the given sample. For instance, if 38 c.c. of hydrochloric acid had been employed, the sample contained 76 per cent, of CaO. Friihling's process is also applied to the determina- tion of lime in mortar. 1 The mortar is well mixed by passing it through a sieve with a stiff brush. Weigh 100 grms. 2 into a stoppered flask. Add 25 grms. of ammonium chloride, 100 c.c. of water, and 20 drops of phenolphthalein. Add 50 c.c. of the standard hydrochloric acid. The red colour disappears temporarily, but reappears on shaking. This shows that free lime is still present. Continue adding acid 1 c.c. at a time, until the red colour reappears but slowly. Then add the acid in smaller portions at a time, say 0'5 c.c. Finally, when the red colour does not appear until the solution has stood for 5 minutes, the titration is complete, and the calculation is made as before. 3 Estimates of the free lime in cements are now seldom published in technical papers, because no method is yet known which will give satisfactory results. Some of the calcium silicates and aluminates are decomposed, by hydrolysis, at the same time as the "free lime" is dissolved. Solutions of glycerol 4 in water ; iodine solutions ; 5 dilute hydrochloric acid ; 6 ammonium salts (page 525) ; 7 etc., 8 have been proposed and found unsatisfactory. The same might 1 M. Holmblad, Tonind. Ztg., 13. 143, 1889 ; H. Seger and E. Cramer, ib., 26. 1719, 1902. 2 A small brass cylinder, closed at one end, and fitted with a movable piston, is sometimes used when the mortar is evaluated by volume, not by weight. The cylinder holds 50 c.c. of mortar, and it is filled by placing the mouth of the cylinder with the piston depressed into the mass of mortar. By elevating the piston, and pressing the cylinder into the mortar, the latter is easily filled. By depressing the piston, 50 c.c. of mortar can be delivered into the glass cylinder for analysis. The final result is then represented as kilograms (or Ibs.) per cubic metre (or cubic foot). 3 If the sand mixed with the lime contains calcium carbonate, the lime content of the mortar may come out too high, because some of the acid may have attacked the carbonate. The calcium hydroxide is, however, attacked before the carbonate, as indicated under ' ' Winkler's process," above. See H. E. Kiefer, Journ. Ind. Eng. Chem., 4. 358, 1912 ; A. H. White, ib., I. 6, 1909. 4 F. Hart, ib., 24. 1674, 1900; M. Mayard, Bull. Soc. L'him. (3), 27. 851, 1902; Chem. News, 87. 109, 1903. 5 F. Hart, Dingier' 's Journ., 175. 208, 1865. 6 C. Zulkowsky, Tonind. Ztg., 22. 285, 1898; H. Hauenschild, ib., 19. 239, 1895; E. Fremy, Compt. Rend , 67. 1205, 1868 ; E. Laudrin, ib.,g6. 156, 379, 841, 1229, 1883 ; F. Schott, Dingier' s Journ., 202. 434, 1871; G. Feichtinger, ib., 174. 437, 1864 ; A. Schulatschenko, ib., 194. 355, 1869. 7 G. Berjuand W. Kosinenko (Landw. Ver. Stat., 60. 419, 1904; A. BodenbenderandE. Ihlee, Zeit. Ver. Rubenzuckerind. , 29. 714, 1907) use ammonium nitrite; M. Heyer ((/hem. Ztg., 33. 102, 1157, 1909 ; P. Philossophoff, ib., 33. 67, 1909; M. Popel, Zeit. angew. Chem., 21. 2080, 1908; F. Knapp, Dingier' s Journ., 265. 184, 1887 ; M. Tomei, Tonind. Ztg., 19. 177, 1895 ; H. Hauenschild, ib., 19. 239, 1895 ; S. Wormser, ib., 24. 1636, 2072, 1900 ; E. Michel, Journ. prakt. Chem. (1), 33. 548, 1844) uses ammonium chloride ; R. Brandenberg (Chem. Ztg., 33. 880, 1909) distils with an alcoholic solution of ammonium bromide. The ammonia in the distillate corresponds with the calcium oxide. J. Hendrich (Journ. Soc. Chem. Ind., 28. 775, 1909 ; 30. 520, 1911) determined the free lime in basic slags by distilling them with solutions of ammonium sulphate and chloride, and estimated the basicity of the slag from the amount of ammonia given off: (NH 4 ) 2 S0 4 + CaO = CaS0 4 + H 2 + 2NH 3 . The ammonia evolved owing to the hydrolysis of the ammonium salts (W. Smith, Journ. Soc. Chem. Ind., 15. 3, 1896 ; 30. 253, 1911 ; V. H. Veley, Journ. Chem. Soc., 87. 26, 1905) is negligibly small, and he claims the results are satisfactory. G. Feichtinger (Bayer. Kunst. Geiverb., 69, 1858) used ammonium carbonate ; M. Tomei (Tonind. Ztg., 19. 177, 1895) used ammonium hydroxide, and also ammonium acetate; S. Wormser (Tonind. Ztg., 24. 28, 1900) ammonium oxalate ; C. Winkler (Dingier' s Journ. , 175. 208, 1865). 8 M. Rischoft (Chem. Ztg., 29. 82, 1903) recommends water for extracting lime from slags, since he thinks sugar solutions dissolve calcium hydrogen carbonate. E. H. Reiser and S. W. Forder (Amer. Chem. Journ., 31. 153, 1904) propose to determine lime in commercial quicklime, slags, and cements by digesting with water free lime is attacked at once, calcium silicates but BARIUM, STRONTIUM, CALCIUM, AND MAGNESIUM. 529 be stated concerning the attempt to determine this constant from the amount of "heat" developed during hydration. 1 278. The Analysis of Calcium Sulphate, Plaster of Paris, and Gypsum. Gypsum, terra alba, plaster of Paris are more or less pure varieties of calcium sulphate. The analysis of calcium sulphate generally involves the determination of lime (CaO), sulphur trioxide (S0 ? ), water, silica, iron, aluminium oxides, magnesia, and carbon dioxide. 2 Calcining and physical tests show value of gypsum and plaster of Paris better than chemical analysis. For the loss of moisture when grinding the sample, see page 124. Moisture. Heat about a gram of the sample one hour at 105, and note the loss in weight. 3 Heat a weighed amount of the sample (say, 1 grm.) in a crucible on a hot plate, and finally at a dull red heat, over a Bunsen's burner for about 10 minutes, and at a rather lower temperature for about 40 minutes. Cool and weigh. Repeat the ignition until a constant weight is obtained. This gives the total water. EXAMPLE. Suppose that : Crucible and 1 grm. of raw gypsum . . . ~. . . 20'2682 grins. Crucible and gypsum at 105 (1 hr.) . 20'1114 ,, Crucible and gypsum at 105 ( hr. more) . . ... 20-1113 ,, Water lost at 105 . 0'1569 grm. Crucible and raw gypsum . . . .... . . 20'2682 grms. Crucible and gypsum at dull redness . . . * . . 20 '0592 ,, Total water . . . .. . . - '. ' v -,. . 0^090 grm. Water lost at 105 . . .....". 15'69 percent. Total combined water . . . .' ... 20 '90 ,, Frey' s Process for the Different Forms of Plaster. In view of the different forms of plaster on the market, it is required to determine the different modi- fications 4 of plaster in a given sample. The problem, at present, has not been solved completely, but Frey 5 has made the best attempt. The hemihydrate CaS0 4 .JH 2 is estimated by finding the amount of water taken up at ordinary temperatures, and retained at 60. 5 grms. of the sample are spread in slowly (C. Winkler, Ckem. Centr. (2), 3. 482, 1858 ; H. le Chatelier, Bull. Soc. Chim. (2), 41. 377, 1884 ; L. C. Levoir, Sec. Travs. Pays-Bas, 5. 59, 1886 ; C. Zulkowsky, Chem. Ind., 24. 290, 1901); water and carbon dioxide (G. Feichtinger, Dingier 's Journ., 174. 437, 1864 ; C. Winkler, I.e. ; L. C. Levoir, I.e.); limewater (E. Laudrin, Lc. ; H. le Chatelier, I.e.} ; sodium carbonate (G. Feichtinger, Bayer. Kunst. Gewerb., 69, 1858 ; F. Schott, Dingier' s Journ. , 202. 434, 1871); potassium carbonate (G. Oddo and E. Manselle, Gazz. Chim. Ital., 25. ii, 101, 1895 ; G. Feichtinger, I.e.) ; sodium hydroxide (F. Hart, Tonind. Ztg., 24. 1674, 1900) ; magnesium chloride (F. Knapp, Dingier' s Journ., 265. 184, 1887); magnesium nitrate (C. Zulkowsky, Zeit. Niederosterreich Ing. Ver., 1863); calcium chloride (L. C. Levoir, Eec. Travs. Pays-Bas, 3. 55, 1885) ; aluminium chloride (S. Wormser and 0. Spanjer, Tonind. Ztg., 23. 1785, 1899) ; water glass (W. Heldt, Journ. prakt. Chem. (1), 94. 129, 1865) ; hydrogen sulphide (B. Steuer, Tonind. Ztg., 23. 1604, 1899). 1 W. Ostwaldand R. Blank, Riga. Ind. Ztg., 9., 208, 1883. 2 E. H. S. Bailey, Univ. Geol. Sur. Kansas, 5. 166, 1899. 3 Gypsum CaS0 4 .2H 2 commences to lose water when heated to 80, and all is driven off at about 300. Three-fourths of the water is driven off at 160, and the remainder at 400. Gypsum heated to about 400 forms ' ' dead burnt " and flooring plaster. V. Zunino, Gaz. Chim. Ital., 30. i. 333, 1910. G. Surr (Eng. Min. World, 36. 467, 1912) recommends report- ing as "moisture" the loss of weight gypsum suffers on heating to 80, and "combined water " at 200. 4 For the five different modifications, see J. H. van't Hoff, Zeit. phys. Chem., 45. 257, 1903 ; P. Rohland, Zeit. anorg. Chem., 35. 201, 1903 ; 36. 332, 1903. 5 0. Frey, Tonind. Ztg., 33. 1229, 1909. 34 530 A TREATISE ON CHEMICAL ANALYSIS. a thin layer over a porcelain basin, and just covered with water from a wash- bottle. In 30 minutes, the basin is placed in a drying oven at 60, and kept at that temperature until the weight is constant. If a grms. of water are taken up, the weight y of hemihydrate in the sample is y = 5-37a ...... (1) unless soluble anhydrite be also present. In that case, the weight of the anhydrite z must be taken into account from (3) below : The soluble anhydrite is estimated by exposing 5 grms. of the sample in a thin layer as before to water vapour at the ordinary temperature for a period of 7 days under a bell-jar. The whole of the soluble anhydrite is thus converted into hemihydrate, and if b represents the increase in weight after drying at 60-70, To detect soluble anhydrite, a portion of the plaster is lawned into a small cylinder containing 200 c.c. of water in which a sensitive thermometer is immersed. If soluble anhydrite % be present, the temperature rises almost at once, and continues to rise for about 2 minutes. The rise of temperature which occurs when the hemihydrate is alone present begins some 5 minutes after the mixing, and continues rising for about 20 minutes. The proportion of flooring plaster l found in plaster of Paris which has been burnt at too high a temperature is estimated by wetting 5 grms. as in the estimating of the hemihydrate, allowing it to remain 7 days in a moist atmosphere, and finally drying at 60. If c denotes the increase in weight, the amount e of the flooring plaster is e = 3'78(c-a) ...... (4) The constituents indicated above are all capable of hydration, but, in addition to these active components, the sample may contain unburnt gypsum, dead- burnt plaster, natural anhydrite, and impurities clay and sand. The unburnt plaster is estimated by finding the loss in weight on ignition. If the loss in weight be w, the weight of unburnt gypsum r is r = 4-78(w-0-062y) . . . . . (5) The total calcium sulphate as determined from the weight of the sulphate present is S. The weight t of natural anhydrite and of dead-burnt plaster is then * = S-(0-93y + z + e + 0-79r) . . . . (6) The sand and clay are determined by difference. Silica. Digest 1 grm. of the powdered sample in an evaporating basin with 20 c.c. of concentrated hydrochloric acid and 20 c.c. of water. Evaporate the solution to dryness. Boil the mass several times 2 with 100 c.c. of dilute * Flooring or hydraulic plaster Estrichgips is gypsum which has been calcined at about 400 , and which has accordingly lost all its water, but, unlike "dead-burnt plaster," is able to combine with water to form a cement. Flooring plaster and dead- burnt plaster are both dehydrated, but the former is active and admits of rehydration, the latter is inactive. J. H. van't Hoff and J. Just, Sitzber. K. Preuss. Akad. Wiss., i. 249, 1903. 2 A single boiling may not suffice to dissolve all the calcium sulphate. There is here a danger of error owing to some calcium sulphate escaping solution. BARIUM, STRONTIUM, CALCIUM, AND MAGNESIUM. 531 hydrochloric acid (1 : 3). Filter, and collect the insoluble matter on a filter paper, wash with hot water, ignite, weigh, and report as insoluble siliceous minerals, or treat the residue as in clay analysis with hydrofluoric acid. 1 Alumina and Ferric Oxide. Make the filtrate to 500 c.c. in a measuring flask. Transfer 300 c.c. to a 600-c.c. Erlenmeyer's flask. Precipitate the ferric and aluminium hydroxides with ammonia, ignite the washed precipitate with the residue left after treating the siliceous residue with hydrofluoric acid. Lime and Magnesia. The filtrate from the ammonia precipitate is concen- trated by evaporation, and the lime and magnesia determined in the usual manner (pages 213 and 218). Sulphuric Acid. The 200 c.c. remaining in the measuring flask are treated with barium chloride as described on page 618. The amount of barium sulphate, multiplied by 0'5833, represents the corresponding amount of calcium sulphate ; and by 0'2402 gives the corresponding amount of CaO. If the amount of CaO so determined is less than the amount obtained by the oxalate process, the remaining CaO will probably have been combined with carbon dioxide ; and the amount of calcium carbonate is determined by multiplying the difference between the two results by 1'7844. Carbon dioxide is not usually present. If it is, the amount can be determined by the process described on page 552. EXAMPLE. Suppose that the analysis of a sample of gypsum furnished : Barium sulphate . . . . . . . . 1 '3156 grm. Calcium oxide . . . . -. . '3295 grm. Hence, 1'3156 xO'2402 = 0'3160 grm. CaO ; and Q'3295 less 0'3160 = 00135 grm. CaO. Consequently, the sample has 0'0135 x 1 '7844 = 0-0240 grm. CaC0 3 ; and I'3156x0'5833 = 0-7674 grm. CaS0 4 . 279. The Analysis of Barytes and Witherite. Barytes, heavy spar, or barite is a variety of native barium sulphate ; and witherite a native barium carbonate. These minerals are so cheap that they are very seldom adulterated, and the analysis is therefore directed to the deter- mination of the natural impurities.' 2 The minerals are often found in commerce ground to impalpable powders. The analysis may involve the determination of barium and calcium sulphates, barium and calcium carbonates, magnesia, silica, alumina, ferric oxide, and moisture. This latter is determined by drying about 2 grms. at 110 for a couple of hours. Traces of lead, and copper sulphides, zinc, are sometimes present. Fluorides (fluorspar) are not at all uncommon. A high- grade sample may run : BaS0 4 . Si0 2 . A1 2 3 . Fe 2 3 . MgC0 3 . CaC0 3 . H 2 0. 96-5 0-5 0-5 0-2 0'2 1'2 0'5 Barium Sulphate and Silica. Boil a gram of the sample in dilute hydro- chloric acid (1 : 3), and evaporate the solution to dryness. Moisten the mass with dilute acid, digest with water, boil, filter, wash. The filtrate contains free iron oxide, alumina, magnesia, etc. ; 3 the insoluble residue contains barium sulphate and "silica." Ignite and weigh in a platinum crucible 4 as "total insoluble." 5 This is sometimes reckoned as barium sulphate. To correct the 1 Or fuse the residue with sodium carbonate ; digest with hydrochloric acid ; evaporate to dryness for silica ; digest the dry mass with dilute hydrochloric acid ; filter ; wash ; and add the nitrate to the main solution. 2 J. Aron, NotizUatt, 8. 293, 1372. 3 Barium carbonate, if present, is also dissolved. The barium chloride can be removed from the filtrate by the addition of dilute sulphuric acid (page 517), and reported as barium carbonate. 4 Assuming that the qualitative analysis has shown the absence of lead. 5 Some prefer to fuse insoluble matter with sodium carbonate. Take up the mass with water, etc., as described below. 532 A TREATISE ON CHEMICAL ANALYSIS. " total insoluble " for silica, add sulphuric and hydrofluoric acids. Evaporate to dryness, ignite, and weigh. The loss in weight represents silica. The alumina, ferric oxide, lime, and magnesia are determined in the filtrate in the usual manner. Soluble Sulphate. Boil 1 grm. of the powdered sample with 20 c.c. oi concentrated hydrochloric acid; add 200 c.c. of hot water: boil; filter; and wash. Determine the sulphates in the filtrate by the addition of bariurr chloride (page 618). Multiply the weight of the barium sulphate obtained b) 0'5833 to get the corresponding amount of calcium sulphate ; and by 0'2402 tc get the corresponding amount of CaO. If the amount of CaO so determined b less than that obtained by the oxalic process, above, and if carbonates are present, report the equivalent of the remaining calcium oxide as calciuir carbonate ; and if carbonates are absent, report the remaining CaO as lime. Loss on Ignition. On ignition of a given sample in a platinum crucible, the loss in weight represents water (free, or combined with gypsum or clay) ; carbor dioxide from the whiting ; and organic matter. Complete Analysis. A more exact analysis can be made by fusing a coupl( of grams of the sample with sodium carbonate in a platinum crucible. Digesl the fused mass with hot water J until the alkaline sulphates and carbonates an all dissolved. Filter the hot solution, and wash the insoluble residue with water A small excess of hydrochloric acid is added to the solution, which is ther evaporated to dryness. The silica is removed by evaporation in the usual manner and sulphuric acid is determined in the filtrate by the method of page 618 The insoluble residue is dissolved in dilute hydrochloric acid, evaporated tc dryness, and the silica filtered off in the usual manner. The two filtrates car then be combined aluminium and ferric hydroxides are precipitated by th( addition of ammonia (page 181); the ammoniacal filtrate is treated for barium strontium, and lime, etc., by the method of 271, or 272, pages 514 and 515. 1 Note, if hydrochloric acid be used for the extraction, insoluble barium sulphate is re-formed During the fusion: BaS0 4 + Na 2 C0 3 =BaC0 3 + Na 2 S0 4 . The latter is soluble in water; th< barium carbonate is insoluble. If hydrochloric acid be used, barium carbonate dissolves, anc barium sulphate is reprecipitated by the soluble sulphate. CHAPTER XXXVII. SPECIAL METHODS FOR THE DETERMINATION OF ALKALIES AND THEIR SALTS. 280. The Gravimetric Determination of Lithium Kahlenberg's Process. LITHIUM can be detected in many clays and in most natural silicates by means of the spectroscope, but it is usually present in such small quantities in clays that a quantitative separation is impracticable. Some samples of Cornish stone, how- ever, may have appreciable amounts of lithium. 1 It is generally stated that the lithium will be found in the ordinary course of analysis, mixed with the chlorides of potassium and sodium. 2 As a matter of fact, lithium carbonate is precipitated when a solution of lithium chloride is treated with an alkali carbonate, for 100 c.c. of water, at 10, dissolve but 1*4 grms. ; and 100 c.c. of alcohol, 0*06 grm. of the carbonate. 3 The solubility is, however, much increased in the presence of ammonium salts, so that no precipitate will be obtained if a sufficient excess of these salts be present. Again, when calcium is precipitated as oxalate, lithium is retained so tenaciously that it is exceedingly difficult to wash the lithium from the precipitate. This uncertainty about the location of the lithium salts makes it desirable to proceed a little differently from the method indicated on page 223, when lithium is to be determined. The Isolation of Lithium, Salts. After treating the calcium oxide, etc., formed during the ignition, in Smith's process (page 222), with water, the aqueous extract is evaporated to dryness in a platinum dish with the occasional addition of con- centrated hydrochloric acid. Dry at 110 for about half an hour. Extract the residue with 25 c.c. of 95 per cent, alcohol, and wash with alcohol until the lithium is all removed. 4 Evaporate the washings, etc., to dryness 5 along with a little hydro- chloric acid. The residue is now extracted with absolute alcohol until the runnings are free from lithium. Evaporate to dryness as before, and treat the residue with a little dilute hydrochloric acid. Make the solution alkaline with calcium 1 The separation of alumina in the analysis of lithium silicates according to K. and E. Sponholz (Zeit. anal. Chem., 31. 521, 1892) requires four or five precipitations of the alumina with ammonia to get the alumina free from lithium. The basic acetate separation, however, is said to present no particular difficulty. 2 Along with rubidium and caesium chlorides. A. A. Anderson, Min. Eng. World, 36. 1055, 1912. For the contamination of potassium chloroplatinate with lithium, see G. Jenzsch, Pogg. Ann., 104. 102, 1858. :5 P. A. Fliickiger, Arch. Pharm. (3), 25. 542, 1887 ; C. N. Draper, Chem. News, 55. 169, 1887 ; G. Geffeken, Zeit. anorg. Chem., 43. 197, 1905. 4 Lithium may be detected spectroscopically in the residue when it is present in un- weighably small amounts. A small direct-vision spectroscope with a slit will show a crimson lithium band between the yellow sodium line and the red potassium line (Plate II.). The spectra of the alkalies are shown on Plate II. The spectroscopic method will detect 3^ mgrm. of soda ; 115 ^ mgrm. of potash ; 1 ^oo mg rm - lithia ; v $ v mgrm. rubidia (Bunsen). 5 Recover alcohol in the usual way. 533 534 A TREATISE ON CHEMICAL ANALYSIS. hydroxide, 1 and filter off the precipitated magnesium hydroxide. Precipitate the lime with ammonia and ammonium oxalate in the usual way. Filter and thoroughly wash with water 100 c.c. usually suffice to free the precipitate from lithium. Eva- porate the filtrate to dryness, drive off the ammonium salts, and repeat the treatment with dilute acid, ammonia, and ammonium oxalate. After evaporation to dryness as before, the residue containing the lithium chloride, along with traces of potassium and sodium chlorides, 2 is treated by the pyridine or the amyl alcohol process. The Solubility of Lithium Chloride in Pyridine. The separation of lithium from a mixture of alkaline chlorides is based on the fact that the chlorides of sodium and potassium 3 are practically insoluble in solutions of pyridine con- taining less than 5 per cent, of water, while pyridine with less than 3 per cent, of water has no appreciable solvent action on the chlorides of sodium and potassium. On the other hand, lithium chloride is fairly soluble in anhydrous as well as in aqueous solutions of pyridine. 4 The solubility of lithium chloride in anhydrous and in 97 per cent, pyridine solutions at different temperatures is shown in the diagram, fig. 161. The curious break in the curve with anhydrous pyridine is supposed to be due to the fact that below 28 the solution contains LiC1.2C 5 H 5 N, and above 28, LiCl.C 5 H 5 N. The table shows that at 20, 100 grms. of anhy- drous pyridine dissolve 13 '39 grms. of lithium chloride, and 100 grms. of the 97 per cent, pyridine (97 vols. anhy- drous pyridine, 3 vols. water) at 22 dissolve 14-31 grms. of lithium chloride. Kahlenberg and Krauskoff have based upon these facts a process for the separation of lithium chloride from the other alkaline chlorides. 1. First Extraction. Evaporate the aqueous solution of the mixed chlorides 5 to dryness. Add a couple of drops of hydrochloric acid. 6 Evaporate the mass again to dryness. Digest the residue 1 L. R. Milford, Journ. Ind. Eng. Chem., 4. 595, 1912. Some prefer lime water to baryta water for precipitating magnesium, and thus avoid introducing a foreign substance. Calcium is already present, and calcium oxalate is less soluble and more easily washed than barium oxalate. 2 If iodides or bromides were present, these salts would be found with the chlorides, since the bromides and iodides are more soluble than chlorides. In that case, an evaporation to dryness with hydrochloric acid and chlorine water, followed by a gentle ignition, will ensure a residue consisting of chlorides only L. R. Milford (I.e.). 3 Also rubidium and caesium. According to M. Serullas (Ann. Chim. Phys. (2), 46. 294, 1831), perchloric acid does not precipitate lithium perchlorate from concentrated solutions oi lithium chloride. 4 A. Neumann and J. Schroeder, Ber., 37. 4609, 1904 ; L. Kahlenberg, Journ. Amer. Chem. Soc., 24. 401, 1902; L. Kahlenberg and F, C. Krauskoff, ib., 30. 1104, 1908; E. Murrnann, Zeit. anal. Chem., 50. 171, 273, 1911. 5 Weighing less than 2 grms. 6 In order to convert any litliium hydroxide into chloride. The hydroxide is formed by th hydrolysis of the chloride, and the hydroxide is not so soluble in the pyridine. 2/i | s 3 JJ 2 * 22 ~? j > / n c * on Q J V^ J- \ <; / > f X ^ 10 g / b 3 -5 ^ V ) ^ t tfi V, ' ^ t s. ^ '^ ^ p y/ a if *^ " 10 '0 20 4-0 60 100 FIG. 161. Solubility of lithium chloride in pyridine. DETERMINATION OF ALKALIES AND THEIR SALTS. 535 with about 25 c.c. of boiling anhydrous pyridine 1 for 5 or 10 minutes. Break up any large masses with a stirring-rod with a rounded end. Let the insoluble residue settle. Decant the clear through a small filter paper 6-5 cm. Collect the filtrate in a small Erlenmeyer's flask, and wash the residue with about 5 c.c. of hot pyridine. 2. Second Extraction. Lithium chloride is retained with great tenacity by the insoluble alkali chlorides, and a second extraction is there- fore necessary. Dissolve the residue in a small amount of water with a drop of hydrochloric acid. Evaporate the solution to dryness, and repeat the extraction with hot pyridine. 2 Combine the fil- trates from the different extractions and wash- ings. Distil off the pyridine by connecting the Erlenmeyer's flask with a small condenser (fig. 162). When the residue in the flask is nearly dry, let the mass cool. 3. Transformation of Lithium, Chloride to Lithium Sulphate. Dissolve the mass in water acidulated with sulphuric acid in order to trans- form the lithium chloride into lithium sulphate. Filter off any organic residue which may be present, and collect the filtrate in a weighed platinum dish. Evaporate to dryness. Drive off the excess of sulphuric acid at a gentle heat. Fuse and weigh. Multiply the weight of the lithium sulphate so obtained by 0*37467 to get the corresponding amount of lithia Li 2 0. The most important objection to this process is the unpleasant odour of the pyridine. Hence, the work should be done under a well-ventilated hood. If the odour of the pyridine be considered an insuperable objection, the following is the next best procedure. In illustration of the excellent results which may be obtained by this process, the following experiments with artificial mixtures of lithium chloride with potassium, sodium, and barium chlorides are reported: Table L XIII. Test Analyses for Lithium. FIG. 162. The recovery of pyridine. Lithium chloride. Residue. Used. Found. Used. Found. 0-0958 0-0907 0-0956 0-0903 0-6007 0-5867 0-6001 0-5863 Besides lithium chloride, the mixture contained : 0-2020 KC1 ; 0-3082 NaCl ; 0-0905 BaCl 2 0-1933 KC1 ; 0'2913 NaCl ; 0'1021 BaCl. 2 The Determination of the Potassium and Sodium Chlorides. If the mixture of alkaline chlorides of L. Smith's process has been under treatment, the residues 1 Kahlbaum's pyridine may be used for this purpose. The cheaper material (5s. per Ib.) should be puritied by distillation from fused potash, and the fraction boiling between 114 and 116 be used for the work. 2 If a relatively large amount of alkaline chloride and a relatively small amount of lithium chloride be present, a third extraction may be advisable. 536 A TREATISE ON CHEMICAL ANALYSIS. on the filter papers which have been washed by the 95 per cent, and the absolute alcohol as well as all the other succeeding residues, are collected together, dissolved in water, and treated as indicated on page 227 et seq., for potassium and sodium. 281. The Gravimetric Determination of Lithium Gooch's Process. Gooch's process l is based upon the fact that lithium chloride is soluble in amyl alcohol 10 c.c. dissolve 0'66 grm. of lithium chloride whereas sodium and potassium chlorides are but sparingly soluble in the same menstruum. Gooch has shown that 10 c.c. of the alcohol dissolve 0-0041 grm. of sodium m \Jsl FIG. 163. Removal of water from amyl alcohol. chloride, and 0*0051 grm. of potassium chloride. Hence, if a mixture of lithium, potassium, and sodium chlorides be washed with amyl alcohol, lithium chloride will be removed, while sodium and potassium chlorides remain behind as sparingly soluble residues. The Separation. The mixed chlorides say, 0'2 grm. 3 or less are dissolved in water and filtered into a 50-80 c.c. Erlenmeyer's flask. Evaporate the solution slowly on an asbestos pad over a small flame until the salts show signs of crystallising (1-2 c.c.). Add a couple of drops of water, and a couple of drops of concentrated hydrochloric acid. Treat the concentrated solution of the 1 F. A. Gooch, Proc. Amer. Acad., 22. 177, 1886; Amer. Chem. Journ., g. 33, 1887; Ohem. News, 55. 18, 29, 40, 56, 78, 1887 ; E. Waller, ib., 62. 173, 181, 1890 ; Journ. Amer. Chem. Soc., 12. 214, 1890; W. J. Schieffelin and W. R. Lamar, ib., 24. 392, 1902; B. Feigenberg, Eine neue Trennungsmethode des Lithiums von anderen Allcalimetallen, Berlin, 1905 ; A. A. Anderson, Eng. Min. World, 36. 1055, 1912 ; W. W. Skinner and W. D. Collins, Bull. U.S. Dept. Agric. Chem., 153. 1, 1912. 2 Rubidium and caesium chlorides are also practically insoluble in the same solvent less than 0-009 grm. is dissolved by 10 c.c. of hot amyl alcohol. 3 If more than this amount of the mixed chlorides be present, concentrate the lithia by digesting the mixed chlorides with 25 c.c. of a mixture of equal volumes of anhydrous alcohol and ether for about an hour in a corked flask. The ether should have been distilled from quick- lime ; otherwise, the concentration of the lithium is not effected. Some lithium hydroxide remains in the residue. Hydrochloric acid is usually added to retard the hydrolysis of the chloride: LiCl + H 2 0^LiOH + HC1. DETERMINATION OF ALKALIES AND THEIR SALTS. 537 mixed chlorides so obtained with 10-15 c.c. amyl alcohol (boiling point, 129- 132) in the Erlenmeyer's flask (fig. 163) fitted with a cork through which passes a capillary tube dipping in the liquid, and also with another tube connected with an aspirator. The flask is heated on an asbestos plate to the boiling point while a current of air is slowly aspirated through the liquid. This prevents any violent bumping of the alcohol, and facilitates the escape of water vapour through the upper layer of amyl alcohol. When the water and about half the amyl alcohol have boiled off, the chlorides of sodium and potassium and a little lithium hydroxide l are deposited, while the lithium chloride remains in solution in the alcohol. Let the liquid cool ; add two or three drops of concentrated hydrochloric acid in order to transform the lithium hydroxide into chloride. Heat the contents of the flask to the boiling point, and let the whole stand half a minute while aspirating a current of air through the fluid. Filter the hot liquid through a Gooch's crucible packed with asbestos. Wash the precipitate and flask with hot amyl alcohol which has been boiled to remove moisture, until the filtrate shows no trace of lithia. When but small quantities of lithia are in question, redissolve the precipitate in a little water, and repeat the treatment. The volume of the hot alcohol which has been in contact with the undissolved chlorides is noted. 2 Evaporate the united filtrates to dryness in an air bath at a temperature not exceeding 125 . 3 Dissolve the residue in a little water acidulated with sulphuric acid. Filter off the organic residue. Collect the filtrate in a weighed platinum dish, evaporate on a water bath and drive off the excess of sulphuric acid at a gentle heat. Calcine the lithium sulphate in the platinum dish to make sure all the carbon left by the amyl alcohol is burnt. The sulphate is fused one minute. When the dish is partly cooled, cover it with a watch-glass to return any fragments of lithium sulphate which may spit off as the cooling mass contracts. Cool in a desiccator and weigh ; again heat the dish up to the fusion point of the sulphate, cool as before, and weigh. The weight is usually constant. Multiply the weight of lithium sulphate so obtained by O2 7 26 to get the corresponding amount of Li 2 0. Corrections. A deduction must be made for the potassium and sodium chlorides dissolved by the boiling amyl alcohol in contact with the mixed chlorides. The potassium and sodium chlorides dissolved by the cold amyl alcohol used in the washings are neglected. Subtract '00050 grm. for every 10 c.c. from the lithium sulphate of the first filtrate if the lithium is being separated from sodium chloride ; arid '00059 if the lithium is being separated from potassium chloride ; and '00 109 if the lithium is being separated from both sodium and potassium chlorides. Determination of the Sodium and Potassium Chlorides. Dissolve the mixed chlorides in the Gooch's crucible in water ; evaporate the solution to dryness in a weighed porcelain dish, and weigh. Separate the two chlorides as indicated on page 234. If lithium and sodium are alone being treated, add 0*00041 grm. to the weight of the sodium chloride for every 10 c.c. of amyl alcohol in the filtrate, and 0-00051 grm. to the potassium chloride, if potassium and lithium chlorides are being treated ; and 0'00092 grm. if the three chlorides sodium, potassium, and lithium are in question. Errors. Duplicates with this process agree very well. The chief errors arise from those incidental to Smith's process (page 222), as well as (1) bumping when 1 Formed by the hydrolysis of the chloride. Lithium hydroxide is but sparingly soluble in amyl alcohol. 2 Estimate the amount of amyl alcohol in the flask by pouring about the same amount of water into a flask the same size, and measuring the amount of water. 3 The amyl alcohol rapidly evaporates at this temperature without loss of lithium by spattering. 538 A TREATISE ON CHEMICAL ANALYSIS. driving off the amyl alcohol ; (2) filtering the amyl alcohol solutions hot ; (3) neglect to evaporate the amyl alcohol in the dish at a temperature below the boil- ing point of amyl alcohol ; (4) foaming or spurting when driving off the sulphuric acid from the lithium sulphate ; and (5) neglect to protect dish from dust, etc., when the lithium sulphate is cooling. A lithium determination by this process, starting from the dry silicate, occupies about two days. 282. The Evaluation of Potassium and Sodium Salts. Sodium Nitrate Chili Saltpetre. Sodium nitrate or Chili saltpetre generally contains between 95 and 98 per cent, sodium nitrate. 1 The general impurities are potassium nitrate (up to 9 per cent, in extreme cases), 2 sodium chloride, sodium sulphate, sodium iodate, a sodium perchlorate, 4 sodium carbonate (rare), magnesium sulphate (rare), and insoluble matter. A common method of evaluating the salt is to determine the percentage moisture, sodium chloride, sodium sulphate, and insoluble matter. These are grouped together as the "refraction," and the remainder is supposed to be sodium nitrate. There are many objections to this process, more particularly if the -salt is to be used in the manufacture of nitric acid. Thus potassium nitrate may be present, and this would interfere with the calculated yield of the nitrate. Determination of Moisture and Insoluble Matter. For the determination of the moisture, heat 10 grms. of the nitrate in an air bath at 120-130 until the weight is constant. This takes about two hours. The insoluble matter is de- termined by digesting 50 grins, of the salt in water and filtering the solution through a tared filter paper, or Gooch's crucible, in the usual manner. Determination of Calcium, Magnesium, and Iron ; Chlorides and Sulphates. Place 20 grms. of the sample in a filter paper resting in a funnel in the neck of a litre flask. Pour boiling distilled water through the paper. When the solution is cold, make it up to the mark on the neck of the flask with cold distilled water. Pipette 50 c.c. for the determination of chlorides gravimetrically (page 652) or volumetrically (pages 76, 79). Another 50 c.c. are treated with barium chloride, and the barium sulphate determined in the usual manner (page 618). The lime is determined in 200 c.c. by precipitation as calcium oxalate, etc. (page 213). The magnesia is determined in another 200 c.c. by precipitation with ammonium phosphate (page 218). Iron can be determined colorimetrically (page 200). Determination of Sodium Carbonate. For the sodium carbonate, evaporate 100 c.c. of the solution to dryness with sulphuric acid, and ignite the residue until the weight is constant. The residue contains sodium, calcium, magnesium, and potassium sulphates Subtract the amount of calcium, magnesium, and potassium sulphates already determined, and the difference represents sodium sulphate. Calculate the equivalent amount of sodium carbonate by multiplying the weight by 0'746. Determination of Potassium. The potassium is determined by dissolving half a gram of the salt in an Erlenmeyer's flask in 25 c.c. of dilute hydrochloric acid. Heat the solution to boiling, and add just sufficient barium chloride, drop by 1 R. Albert! and W. Hempel, Zeit. angew. Chem., 5. 101, 1892. 2 G. Lunge, Chem. Ind., 9. 269, 1886; H. Hagen, Chem. Ztg., 15. 1528, 1891 ; Zeit. angew. Chem., 6. 495, 698, 1893. 3 H. Beckurts, Pharm. Centralhalle, 233, 1886. 4 H. Beckurts, Archiv Pharm., 224. 333, 1886 ; H. Fresenius and H. Bayerlein Zeit anal Chem., 37. 501, 1898. DETERMINATION OF ALKALIES AND THEIR SALTS. 539 drop, to precipitate the sulphates as barium sulphate. Wait after each addition until the liquid has a clear layer, and add the barium chloride until a drop ceases to produce a cloudiness. If too much barium chloride has been added, a drop or two of dilute sulphuric acid will probably make things right by removing the excess of barium chloride. When the solution is cold, treat the clear filtered solution with hydrochloroplatinic acid, etc., as indicated on page 234. See also page 540. Potassium Nitrate, Saltpetre, Nitre. The moisture is determined as in the case of sodium nitrate. Less than J per cent, of moisture is usually present. For the chlorine, dissolve 100 grms. of the salt in water, filter, and wash with hot water. The solution is treated with silver nitrate and the chlorine determined by the turbidity method, or gravi- metrically. Less than O'Ol per cent, is usually present. The insoluble matter, sulphates, lime, and magnesia are determined as in the case of sodium nitrate. The sodium is determined as indicated on page 239. Potassium Carbonate, Pearlash. The moisture and insoluble matter are determined as just indicated for sodium nitrate. The available alkali is determined by titration as on page 69. The chlorides, sulphates, insoluble matters, silicates, and phosphates are determined in the usual manner. Iron is determined colorimetrically. The total potassium is determined as indicated under sodium nitrate. The results are expressed by calculating the amount of potassium chloride from the silver chloride, and the amount of potassium sulphate from the amount of barium sulphate. Then calculate the amount of potassium carbonate from the difference between the total potash and that corresponding with the potassium chloride and sulphate. Calculate the amount of sodium carbonate from the difference between the available alkali and the potassium carbonate calculated as indicated above. For a further discussion on sodium and potassium carbonates see page 72. 283. The Determination of Potassium Colorimetrically Cameron and Failyer's Process. When potassium chloroplatinate is dissolved in water and mixed with an excess of a solution of potassium iodide, the intensity of the red colour so obtained is proportional to the amount of potassium chloroplatinate 1 in the solution. This reaction may therefore be employed quantitatively for the determination of very small quantities of potassium. Ammonium chloroplatinate produces a similar coloration, and therefore all the reagents, and the atmosphere of the laboratory in which the determination is made, must be free from ammonia. Standard Solution. Dilute *10 c.c. of standard potassium chloroplatinate solution 2 to about 30 c.c. ; add a drop of hydrochloric acid (1 : 1) and J c.c. of potassium iodide solution. 3 Let the mixture stand for about an hour to permit the colour to fully develop, and then dilute to 100 c.c. The colour of the test solution is developed at the same time. 1 F. K. Cameron and G. H. Failyer, Journ. Amer. Ghem. Soc., 25. 1063, 1903 ; L. A. Hall, ib., 25. 990, 1903 ; T. T. Morrell, ib., 2. 145, 1880 ; 0. Schreinerand G. H. Failyer, Bull. U.S. Dept. Agric. (Soils), 31. 31, 1906. 2 STANDAKD POTASSIUM CHLOROPLATINATE SOLUTION. Dissolve 0'0516 grm. of carefully recrystallised potassium chloroplatinate-K 2 PtCl 6 in water, and dilute the solution to a litre. Each c.c. of this solution is equivalent to '00001 grm. of K 2 0. 3 POTASSIUM IODIDE SOLUTION. Dissolve 25 grms. of potassium iodide in water, and make the solution up to 100 c.c. 540 A TREATISE ON CHEMICAL ANALYSIS. Test Solution. The potassium chloroplatinate is isolated in the usual manner, as indicated on page 234. 1 The salt is collected on the asbestos felt 2 of a Gooch's crucible rather than on filter paper, because of the difficulty in getting filter paper quite free from ammonia. The dried precipitate of potassium chloroplatinate is dissolved in hot water by washing the crucible until about 25 c.c. of filtrate have been obtained. Cool the filtrate ; add a drop of hydrochloric acid, and J c.c. of potassium iodide solution. After the solution has stood for an hour, dilute to 100 c.c. The comparison is made in the usual manner. 3 g 284. The Determination of Potassium Cobaltinitrite Process. A solution of sodium cobaltinitrite, [Co(N0 2 ) 6 ]Na 3 also called Koninck's reagent precipitates canary-yellow potassium cobaltinitrite when added to a solution of a potassium salt. 4 The corresponding salts of lithium, magnesium, barium, strontium, calcium, iron, aluminium, zinc, etc., are soluble, and hence potassium salts can be readily detected in the presence of these compounds. Caesium, rubidium, and ammonium 5 salts, like potassium salts, give sparingly soluble cobaltinitrites. The reaction proceeds in presence of chlorides, sulphates, nitrates, phosphates, and acetates. The reaction is quantitative so far as the separation of potassium from sodium is concerned, 6 but the precipitate contains variable quantities of sodium unless the conditions under which the precipitation is made be fairly constant. 7 The precipitate can then be made to approximate closely to dipotassium sodium cobaltinitrite, [Co(N0 2 ) 6 ]K 2 Na.H 2 0. It is maintained by some that the precipitation of potassium as cobaltinitrite is not advisable because of the variable nature of the precipitate just indicated, and because the precipitate is extremely difficult to wash clean. When first formed, the precipitate is granular and settles easily, but it has a tendency to become more or less colloidal as the washing liquids become poor in dissolved salts. The precipitate then filters slowly, and is liable to run through the filtering medium. It has been found possible, however, to cope with these difficulties by evaporating the solution containing the precipitate to a " pasty " mass before filtration and washing. The evaporation also eliminates an error due to the variable character of precipitates formed in solutions of different con- centration, and variations which occur when the precipitate is allowed to stand 1 Special care must be taken to make sure all the ammonia salts are driven off during the ignition of the mixed chlorides, etc. 2 The asbestos is washed (page 104), and ignited in a platinum dish to remove all traces of ammonia. Preserve the asbestos in alcohol in a well-stoppered bottle. 3 If the standard or test solution is the stronger, dilute the stronger solution still more, or take an aliquot portion and dilute it to 100 c.c. 4 0. L. Erdmann, Journ. prakt. Chem. (1), 97. 385, 1866 ; G. P. Sadtler, Amer. J. Science (2), 49. 189, 1870 ; L. L. de Koninck, Zeit. anal. Chem., 20. 390, 1881 ; E. Bjilmann, ib., 39. 284, 1900 ; C. 0. Curtmann, Ber., 14. 1951, 2121, 1881 ; G. Sarr, Min. Eng. World, 37. 288, 1912. Potassium can be detected if but 60 parts are present per 1,000,000 parts of solution ; if the test be made in the presence of T f N-silver nitrate, the sensitiveness of the test is in- creased to one part per million (L. L. Burgess and 0. Kamm, Journ. Amer. Chem. Son., 34. 652. 1912). 8 Ammoniacal fumes should be absent when the work tests, etc. is in progress. 6 F. H. van Leent, Zeit. anal. Chem., 40. 567, 1901 ; H. Weber, ib., 38. 171, 1899; W. Autenrieth and R. Bernheim, Zeit. physiol. Chem., 37. 29, 1902 ; W. Autenrieth, Centr. Min., 513, 1908 ; K. Gilbert, Die Bestimmung des Kaliums nach quantitativer Abscheidung desselben als Kaliumnatriumkobaltinitrit, Tubingen, 1898. 7 R. H. Adie and T. B. Wood, Journ. Chem. Soc., 77. 1076, 1900 ; M. Cunningham and F. M. Perkin, ib., 95. 1562, 1909 ; K. A. Hofmann and 0. Burger, Ber., 40. 3298, 1907 ; E. A. Mitscherlich, K. Celichowski, and H. Fischer, Landw. Ver. Stat., 76. 139, 1911 ; E. A. Mitscherlich and H. Fischer, ib., 78. 75, 1912. DETERMINATION OF ALKALIES AND THEIR SALTS. 541 in contact with the mother liquid for different lengths of time. Under these conditions fairly satisfactory results have been obtained. The precipitate need not be ignited and weighed, for it reacts with potassium permanganate as symbolised by the complicated equation : 10[Co(N0 2 ) fi ]K 2 Na H.,0 + 22KMn0 4 + 58H 2 S0 4 = 21K 2 S0 4 + 5Na 2 S0 4 + 10CoS0 4 + 22MnS0 4 + 60HN0 3 + 38H 2 0. Hence, the precipitate, and accordingly also potassium, can be de- termined by a volumetric process. 1 It follows from the equation that 22KMn0 4 are equivalent to 10K 2 0, and 2KMn0 4 are equivalent to -j-f K 2 ; or 1 c c of T?rN- KMn0 4 = 0-0008573 grm. K 2 0. Preparation of the Sample. One to five grams of the powdered mineral is decomposed by evaporation to dryness with a mixture of sulphuric and hydro- fluoric acids (page 226). The dry mass is extracted with about 50 cc. of hot water, and boiled with a saturated solution of sodium acetate 2 for 10-15 minutes in a beaker of Jena glass. The precipitated basic acetates of iron, aluminium, titanium, etc., are filtered and washed a few times with hot water. The filtrate is boiled with sodium carbonate. The precipitated magnesium and calcium salts are washed with hot water, and the filtrate evaporated down to about 15 or 20 c.c. Precipitation of Potassium Cobaltinitrite. The solution prepared as just described, or by any other convenient process, is mixed in a porcelain basin with, say, 30 c.c. 3 of Koninck's reagent, 4 and 1 c.c. of concentrated acetic acid; 5 evaporated on a water bath down to a pasty consistence ; 6 and, after cooling, stirred with 25-50 c.c. of water. The precipitate is filtered through an asbestos- packed Gooch's crucible, and washed with sufficient water to remove the excess of the reagent. . Volumetric Determination of the Precipitated Cobaltinitrite. Transfer the precipitate and felt 7 to a 300-c.c. beaker, and carefully remove any precipitate adhering to the sides of the crucible by means of a rubber- tipped rod and jet of water. An excess of T 1 pN-KMn0 4 is added, and the contents of the beaker are 1 R. H. Adie and T. B. Wood (I.e.) ; W. A. Drushel, Amer. J. Science (4), 24. 433, 1907 ; (4), 26. 329, 555, 1908 ; Bull U.S. Agric. Dept. (Ohem.), 152. 42, 1912 ; E. L. Baker, ib. t 152. 28, 1912 ; P. F. Trowbridge, ib., 152. 184, 191:2 ; L. T. Bowser, ib. t 152. 45, 1912 ; Journ. 2nd. Eng. Chem., I. 791, 1909 ; 0. M. Shedd, ib., I. 302, 1909 ; 2. 379, 1910 ; E. A. Mitscheilich, K. Celichowski and H. Fischer, Landw. Vers. Stat., 76. 139, 1912. 2 The solution of 5 grms. of the crystalline salt in water should give no yellow precipitate with sodium Cobaltinitrite (Koninck's reagent) after standing 24 hours. 3 A large excess of Koninck's reagent is needed to ensure the coagulation of the precipitate and prevent its running through the asbestos. 4 KONINCK'S REAGENT. Dissolve 30 grms. of cobalt nitrate, Co(N0 3 ) 2 . 6H 2 0, in 60 c.c. of water, and add 100 c.c. of a saturated solution (corresponding to about 50 grms.) of sodium nitrite, and 10 c.c. of glacial acetic acid. In a few seconds, a brisk effervescence occurs owing to the escape of nitric oxide, the colour of the solution simultaneously changes to a dark yellowish brown. Since commercial sodium nitrite nearly always contains some potassium salt, some potassium Cobaltinitrite usually separates when the mixture has stood a couple of days. Filter the clear solution. The reagent undergoes very little change in three to four weeks if it is kept in dark-coloured glass bottles. The sensitiveness of the reagent which has been kept some time can be tested by adding a few drops to 5 c.c. of distilled water containing one drop of a 10 per cent, solution of potassium chloride. The reagent can be used if it at once produces a yellow precipitate. The reagent which has decomposed and is accordingly useless for the purpose has lost its yellowish-brown colour and become rose-red W. Autenrieth (I.e.). Some prefer to keep the solutions of cobalt nitrate and sodium nitrite in separate bottles and mix the proper proportions for use as required. 5 The acetic acid prevents subsequent difficulties arising from the sticking of the precipitate to the porcelain dish. If the difficulty still persists, add more acetic acid next time the analysis is made. 6 If the heat be here too vigorously applied, the viscid liquid is liable to spurt. 7 It is best to remove the plate to avoid the danger of its becoming coated with manganese hydroxide later on. 542 A TREATISE ON CHEMICAL ANALYSIS. boiled until the colour begins to darken. Add 10 c.c. of sulphuric acid (1 : 1) ; l thoroughly stir the mixture ; and titrate with ^N-oxalic acid solution until the pink colour of the permanganate just disappears. 2 The difference between the number of c.c. of the j-^N-permanganate added, and the number of c.c. of oxalic acid used in the titration, multiplied by 0*0008573 represents the number of grams of K 2 in the given sample. Thus, with a sample of felspar, 0-5 grm. 50 c.c. of y^N-permanganate were added, and 8 c.c. of y^N-oxalic acid were used in the subsequent titration of the undecomposed permanganate. Hence, (50 - 8) x 0-0008573 = 0'036 grm. of K 2 per 0'5 grm. of sample ; or the sample contained the equivalent of 7 '2 per cent, of potash (K 2 0). 3 Determination of the Potassium in the Cobaltinitrite as Potassium Chloro- platinate or Perchlorate. According to Autenrieth, the method is most usefully employed as a rapid control process, and for separating the potassium from the other constituents prior to its exact determination as perchlorate or chloro- platinate. In this case, the precipitated cobaltinitrate is collected on a filter paper ; the paper is ashed in a platinum crucible, and the precipitate added ; the whole is then calcined to dull redness for a few minutes, and the alkali nitrites extracted with hot water. Filter off the cobalt oxide ; evaporate to dry- ness with concentrated hydrochloric acid (sp. gr. 1-124); and determine the potassium as indicated on page 234 or page 237. 1 F. H. MacDougall (Journ. Amer. Chem. Soc., 34. 1684, 1912). " If the sulphuric acid be added to the yellow salt before the permanganate, the amount of permanganate required will be greater than in the regular method . . . owing to the failure of cobaltic cobalt to oxidise nitrites in sulphuric acid solutions under the conditions of the described method." 2 If manganese hydroxide sticks to the sides of the beaker, add a slight excess of oxalic acid and loosen the precipitate by means of a rubber-tipped rod. Remove the rod and rinse it, since rubber must not be left in contact with potassium permanganate. 3 After trying the effect of various salts on the results, Bowser (I.e. ) said: "The greatest danger seems to exist in the case of MgCl 2 , the iron and the soluble calcium salts. It is unsafe to trust a determination when any of these metals are present. Adie and Wood advised pre- cipitating out all these interfering metals by boiling with sodium carbonate, and this would seem to be the best procedure." PART V. SPECIAL METHODS-ACIDS AND NON-METALS. CHAPTER XXXVIII. THE DETERMINATION OF CARBON-FREE AND COMBINED. 285. The Direct Determination of Carbon. A ROUGH idea of the amount of carbon in clays, graphites, etc., can often be obtained by digesting the clay with sulphuric and hydrofluoric acids. The temperature must not be high enough to cause a reaction between the carbon and the sulphuric acid. A slight action can scarcely be avoided. When the acid has done its work, dilute the solution with water filter through a Munroe's crucible or a tared filter paper ; wash ; l and weigh the dried residue. In view of the many theoretical objections which can be urged against this process, it is surprising how good an approximation can be obtained with a little practice. 2 Graphitic carbon in the presence of other varieties of organic matter (vegetation, etc.) which cannot be removed mechanically is determined by Mackintosh 3 by the following process, which is based on Schoffel's method for purifying graphite : Melt some potash in a silver crucible. When the fused mass has stopped spitting, and is in a state of quiet fusion, add the dry and powdered sample under investigation. Keep the temperature of the mass just above the melting point of the potash, and stir occasionally with a silver wire. Raise the temperature a little towards the end of the fusion. When the decomposition is complete, take up the mass with water, filter through a weighed Gooch's crucible, and wash with dilute hydrochloric acid, and finally with a little ammonia in order to dissolve any silver chloride formed by the reaction between the potash and the silver crucible. The graphite remains behind, insoluble ; it is scarcely attacked by the treatment. Dry and weigh. Mackintosh says the whole opera- tion occupies about two hours. Lowe 4 fuses the sample with sodium carbonate, digests the resulting cake with water, and then boils the residue with a solution of sodium hydroxide. The residue is then boiled with hydrochloric acid, washed with water, and dried in a weighed filter, and its weight determined. Silica, etc., can be determined in the aqueous solution from the fused cake. There is always some doubt in both Lowe's and in Mackintosh's processes if the residue is all carbon, and in conse- quence processes have been devised to burn the carbonaceous residue, and estimate the amount of carbon it contains either by " loss on ignition " or as carbon dioxide. 1 Not through asbestos, owing to the attack by the acids. Of course the acids may be neutral- ised before the filtration. 2 For filtration with carbon tubes, fig. 53, page 102 ; and subsequent combustion of the residue, see page 563. 0. Johanuseu, Stahl Eisen, 30. 458, 1910. 3 J. B. Mackintosh, Chem. News, 51. 147, 1885; R. Schoffel, Jahrb. geol. Rcichsanttalt, 16. 270, 1866. 4 J. Lowe, Dinner's Journ., 137. 445, 1885 ; E. Weinschenk, Zeit. Kryst., 28. 300, 1897. 545 35 546 A TREATISE ON CHEMICAL ANALYSIS. 286. The Wet Combustion Process for the Determination of Carbon as Carbon Dioxide. In the so-called wet combustion process, the finely powdered sample is digested with a mixture of sulphuric acid and chromic acid, or potassium dichromate, or potassium permanganate. 1 The oxygen evolved during the decomposition of the chromates oxidises the carbon to carbon dioxide. The gases are dried, and the carbon dioxide is absorbed as indicated below. These methods appear very simple ; so they are. But it is not easy to get good results.. It is therefore well to practise the method with a known weight of cane sugar mixed with, say, china clay known to contain no carbon. One gram of cane sugar furnishes the equivalent of 1'544 grms. of carbon dioxide. In this way the FIG. 164. Wet combustion of carbon. beginner may acquire confidence in his work. Details of the procedure are indicated below. The results are usually a little low. 1. The Apparatus. A Corleis' flask 2 A (fig. 164) contains the powdered clay under investigation. It is closed by a ground-glass condenser B. A side tube, reaching nearly to the bottom of the flask, is connected witli a 1 R. Warington and W. A. Peake, Journ. Chem. Soc., 38. 6, 17, 1880; 0. F. Cross and E. J. Bevan, ib. t 53. 889, 1888; A. H. Elliott, ib., 7. 182. 1869 ; R. Phelps, Zeit. anorg. Chem., 16. 85, 1898 ; R. Finkener, Mitt, konigl. Ver. Anstalt, Berlin, 156, 1889 ; Zeit. anal. Chem., 29. 666, 1890; R. E. and W. M. Rogers, Amer. J. Science (2), 5. 352, 1848; (2), 6. Amer. Chem., 2. 140, 1872. Errors C. F. Cross and E. J. Bevan, Chem. Ztg., 36. 1226, 1912. 2 A. Corleis, Stdhl Eisen, 14. 381, 624, 1894 ; A. Kleine, Chem. Ztg., 33. 376, 1909. Over a score of different forms of flask have been devised for this purpose. A substitute can readily be improvised from an ordinary flask and condenser. THE DETERMINATION OF CARBON FREE AND COMBINED. 547 plug stopcock c, and a bulb b for the reception of the acid to be subse- quently brought in contact with the clay under investigation. The side tube is connected with a Fresenius' tower C containing an aqueous solution of caustic potash in the lower part, and soda lime in the upper part. The object of the tower C is to remove carbon dioxide from any air subsequently aspirated through the system. 1 The purpose of the condenser is to return steam, etc., arising from the flask daring an experiment, and so prevent the entry of an excess of moisture into the absorption tubes. The gases travelling past the condenser traverse a White's 2 absorption tube D containing an acid solution of silver sulphate 3 to absorb chlorine and vapours of sulphur compounds. The gases then pass through a capillary tube E of quartz 4 or platinum, heated to a dull redness by means of the Bunsen's burner F. The object of this tube is to oxidise any hydrocarbons, carbon monoxide, etc., which might be imperfectly oxidised by the chromic acid in Corleis' flask. 5 The gases are then dried by passing through a Miiller's absorption tube G. This tube contains concentrated sulphuric acid. The tube G must follow the hot capillary tube E> since sulphuric acid is liable to absorb hydrocarbon gases. Any carbon dioxide in the gas is then absorbed by the weighed potash tube 6 //, which has a calcium chloride tube / attached. Drechsel's wash-bottle, 7 containing potash, protects the weighed potash tube from atmospheric carbon dioxide. The open end of the train J is connected later on with an aspirator or suction apparatus in order that a current of air may be drawn through the system when required. The whole may be mounted on a wooden stand or on a Miiller's retort stand as shown in the diagram. The condenser is connected with the tubes in the upper part of the stand by means of rubber tubing. The lower tubes of the stand are respectively connected with the water supply and the sink. 1 For the difficulties involved in removing the last traces of carbon dioxide from air, see C. W. Eliot and F. H. Storer, Proc. Amer. Acad., 5. 52, 1861 ; Chem. News, 3. 178, 1861 ; R. E. and W. M. Rogers, I.e. ; C. Brunner, Pogg. Ann., 24. 571, 1832 ; H. Hlasiwetz, Wiener Akad. Ber., 20. 189, 1856. 2 J. T. White, Chem. News, 58. 166, 1888. Mitscherlich's tube (A. Mitscherlich, ZeiL anal. Chem., 15. 389, 1876) or various other forms of absorption tube may be employed, and also other types of flask. Thus, an Erlenmeyer's flask with a three-hole rubber stopper fitted with a separating funnel, a reflux condenser, and a gas leading tube dipping to the bottom of the flask, and connected, when required, with a suitable absorption apparatus for removing carbon dioxide. L. L. de Koninck (ZeiL anal. Chem., 27. 463, 1888) suggests adding the silver sulphate to the contents of the flask A ; but although dilute sulphuric acid (2 acid, 1 water, does not appear to act on silver chloride (J. Volhard, ib., 18. 281, 1879), concentrated sulphuric acid decomposes silver chloride with the evolution of hydrogen chloride (A. Sauer, ib., 12. 176) 1873 ; L. Blum, ib., 28. 450, 1889). Hence, de Koninck's proposal will not do. 3 SILVER SULPHATE SOLUTION. Dissolve 0*624 grm. of silver sulphate in water, add 5 c.c. of normal sulphuric acid, and make the solution up to 100 c.c. The solution is -feE. A tube containing warm lead dioxide is sometimes used to absorb sulphurous gases and haloids. If the lead dioxide be free from lead monoxide, as is the case with " Dennstedt's lead dioxide," there is no danger of the absorption of carbon dioxide. M. Dennstedt and F. Hassler, Zeit. anal. Chem., 42. 417, 1903 ; M. Dennstedt, ib., 41. 525, 1902; F. Kopfer, ib., 17. 28, 1878 ; C. M. Warren, Amer. J. Science (2), 41. 40, 1866. Sometimes a heated tube containing silver gauze or wire is used F. Kopfer, Zeit. anal. Chem., 17. 28, 1878. The silver is restored by heating in a current of hydrogen K. Kraut, ib., 2. 242, 1863. Antimony is sometimes recommended for removing chlorine. A copper spiral was used by C. Glaser, LieHg's Ann. Suppl., 7. 213, 1870.; G. Stadler, Liebig's Ann., 69. 334, 1849. For fluorine, see H. Moissan, Compt. Rend., 107. 992, 1888 ; A. Volcker, Chem. Gaz., 7. 245, 1849. 4 M. Widemann, Chem. Ztg., 33. 1186, 1909. 5 C. F. Mabery, Journ. Amer. Chem. Soc., 20. 510, 1898 ; G. Auchy, ib., 20. 243, 1898 ; J. Widmar, Zeit. anal. Chem., 29. 160, 1890 ; Chem. News, 62. 274, 1890 ; J. W. Langley, ib. , 62. 218, 1890; Trans. Amer. Inst. Min. Eng., 19. 614, 1890 ; A. von Reis, Stahl Eisen, 14. 581, 1893 ; J. J. Messinger, Ber., 21. 2910, 1888 ; A. Miiller, Chem. Zeit., 28. 795, 1904. 6 I prefer Berl's or Landsiedl's form of potash tube for the reasons indicated on page 551. 7 E. Drechsel, Zeit. anal. Chem. , 15. 446, 1876. 548 A TREATISE ON CHEMICAL ANALYSIS. 2. Charging the Absorption Tubes. The BeiTs or Landsiedl's potash bulb (figs. 165 and 166) is kept under a desiccator over sulphuric acid or calcium chloride, and it should stand in the balance case about an hour before each weighing. This apparatus is filled with an aqueous solution of caustic potash. 1 Place the potash solution in an evaporating basin ; remove the calcium chloride tube B ; attach a piece of rubber tubing, or a suction tube, to the end which was attached to the calcium chloride tube ; dip the opposite end in the potash solution, and suck at the rubber tube until the three bulbs are nearly three- fourths full. Do not suck too vigorously, or potash may be drawn into the mouth. If the tube A is wide enough, the potash solution may be introduced FIG. 165. BerPs potash bulbs. FIG. 166. Landsiedl's potash bulbs. by means of a pipette. Clean that part of the apparatus C which dipped in the solution by means of filter paper. Now fill the small tube B by placing a plug of glass-wool at the end of the tube ; then a layer of fragments of granular (not fused) calcium chloride less than half a centimetre in diameter and sifted free from dust ; then a similar layer of fragments of soda lime ; 2 and finally another plug of glass-wool. 3 The joint A is slightly greased. The potash apparatus will absorb about 0'2 grm. of carbon dioxide without re-filling. The contents of the potash apparatus are protected, when not in use, by glass 1 POTASSIUM HYDROXIDE SOLUTION. A solution of specific gravity 1 '4 is made by dissolving 4 parts of the solid in 6 parts of water by weight. Caustic potash is better than caustic soda, since sodium bicarbonate is less soluble than the corresponding potassium salt. The latter is therefore less likely to block the tubes by the crystallisation of the bicarbonate. If the potash should contain any nitrites high results may be obtained, owing to the formation of nitrates S. C. Jutsum, Chem. News, 41. 17, 1880. For the relative efficiency of the different absorption media, see R. Fresenius, Zeit. anal. Chem., 5. 87, 1866 ; J. Lowe, ib., 9. 220, 1870. 2 Soda lime loses moisture in air dried by calcium chloride or concentrated sulphuric acid. Hence, some prefer bits of potash in place of the soda lime. 3 If the soda lime tubes, etc., be packed too tightly, the expansion which occurs as the solid becomes moist may burst the tubes. It is best to interpose layers of glass-wool between the fragments of soda lime. Calcium chloride may contain basic chlorides or free lime, which absorb carbon dioxide and lead to incorrect results if such be placed in tubes before the potash apparatus. The faulty calcium chloride will react alkaline to litmus. To render the contamination inert, it is usual to pass a stream of dry carbon dioxide through the calcium chloride for a couple of hours, and remove the carbon dioxide by passing a current of dry air over the calcium chloride for half an hour. C. Winkler (Zeit. anal. C/iem., 21. 545, 1882) denies the efficacy of the treatment. He maintains that the lime in the chlorides cannot be completely saturated with carbon dioxide, since the inside of the granules remains caustic after the treatment. Under any circumstances the error mentioned by Winkler is very small, and in the present case this difficulty does not enter into the problem. THE DETERMINATION OF CARBON FREE AND COMBINED. 549 caps ground on to the free ends, or short pieces of rubber tubing 2 cm. long with one end closed by a plug of glass rod about 1-5 cm. long, and rounded at both ends. The apparatus is weighed before and after an experiment with the caps in position. The caps are only removed when the apparatus is to be connected with the train of absorption tubes. The lower part of the Fresenius'. tower C is filled to a depth of about 2 cm. with potash lye (sp. gr. 1-4). A plug of glass-wool is placed in the constricted part, and rather coarser fragments of soda lime than those used for the tube B, and free from dust, are placed in the upper part. When not in use, the ends of the absorption tubes should be plugged with rubber tubing and glass rod to shut off the air. The stoppers of C and G render plugs unnecessary. 3. Charging the Combustion Flask. Transfer, say, 1 grm. of the clay, 1 finely powdered and dry (110), to the Corleis' flask, and add from 5-10 grms. of coarsely granular dry potassium bichromate. If the clay be rich in organic matter less clay may be taken, and the clay is often advantageously mixed with about 10 grms. of dry calcined sand say Calais sand. Add about 30 c.c. of water, and shake the mixture up, taking care that none of the powder is left stranded above the level of the fluid in the flask. Connect the apparatus as shown in the diagram (fig. 164), and aspirate air throughout 'the flask and train of absorption tubes for about 10 minutes. Weigh the potash apparatus, and place it in the position shown in fig. 164. 4. The Oxidation. Close the stopcock c, light the Bunsen burner F. Pour concentrated sulphuric acid into the bulb b, and allow the acid to run very slowly into the flask. If no vigorous action occurs, raise the temperature of the flask very slowly. 2 The bubbles of gas should not at any time pass so rapidly that they cannot be easily counted. If the velocity of the reaction be too rapid, the gases may pass through the potash apparatus without being washed free from carbon dioxide, and the calcium chloride tube may not absorb the moisture driven from the potash apparatus. The bulb b should be emptied before the flask is warmed, so that the stopcock can be quickly opened in case a sudden pressure be generated in the flask. If the flask should burst and hot concentrated sulphuric acid be sprayed about the operator, the consequences would be serious. With the Corleis' flask the condenser tube or the plug stopcock will be lifted if the pressure becomes very great. 3 A wire gauze between the gas and the Corleis' flask, A, is preferable to a sand bath, or to an asbestos or quartz plate, because the temperature of the flask is then under better control. When all the carbon is oxidised, a current of air is aspirated through the system for about 10 minutes. 5. The Weighings. 1 grrn. of clay was treated as just described. The effect on the potash apparatus was as follows : Weight of potash bulbs (after) 787221 grms. Weight of potash bulbs (before) , . . 78 '5637 grms. Carbon dioxide absorbed . 01584 grm. 1 The amount of clay to be taken depends on the amount of carbon it contains. The present sample was a black Devonshire ball clay. The potash apparatus should not be made to take more than 0'2 grm. of C0 2 . ' 2 The bichromate may be dissolved in the sulphuric acid before it is mixed with the clay, as recommended by F. K. Cameron and F. Breazeale (Journ. Amer. Ohem. Soc., 26. 29, 1903). 3 An explosion might occur if boiling sulphuric acid struck the condenser and caused it to crack. At any rate, a stream of cold water from the cracked condenser descending into the boiling sulphuric acid would make things unpleasant. Hence the acid should boil without bumping. Corleis' flasks can now be obtained with the side tube arranged to intercept the drops from the condenser so that they run quietly into the hot acid. 550 A TREATISE ON CHEMICAL ANALYSIS. Every 44 grms. of carbon dioxide are equivalent to 12 grma. of carbon ; hence, 0-1584 grm. of carbon dioxide is equivalent to O0432 grm. of carbon per 1 grm. of clay ; or the sample has 4'32 per cent, of carbon. It will be observed that an error of 1 per cent, in weighing, etc., only makes about 0'3 per cent, difference in the determination of the carbon. The agreement in a number of duplicates does not prove that the results are accurate. As indicated on page 171, such an agreement shows nothing more than that the method of analysis used will furnish consistent results when conducted in a uniform way. Suppose a constant percentage of carbon escaped decomposition, the duplicates would be consistently too low. In order to make reasonably sure that a result is accurate, it is necessary to show that a similar result is obtained by several different methods. The time involved is, of course, out of question in commercial work ; but it is necessary in establishing the accuracy of a method of measurement when it is not possible to operate with mixtures of known composition. 1 As stated above, there is no special reason why the particular form of apparatus here recommended should be adopted. Once the principle is mastered, there will be no difficulty in devising innumerable modifications. Fresenius, for instance, used nothing but U-tubes and flasks connected by rubber and glass tubing, but he required nine U-tubes and two flasks ! By means of more powerful absorbing apparatus, it is now possible to make the apparatus more compact, and to lessen the labour fitting up the apparatus. The time-factor here again presses upon us. A neat and compact apparatus frequently costs more than a " home-made " apparatus, but in routine work the neat and compact apparatus frequently saves time and renders it more difficult to make mistakes. It is generally advisable to let students use "home-made" apparatus in order to acquire skill, ingenuity, and adaptability in an unfavourable environment. Students reared among neat time-saving apparatus may be found wanting when it is necessary to improvise apparatus. 287. The Errors in Analyses involving the Weighing of Absorption Tubes. It requires some practice to get constant results. In addition to the precautions already mentioned, it may be useful to point out a few more sources of danger. 2 1. The air leaving the weighed potash bulbs should have the same state of humidity as the air which enters ; otherwise, a loss or gain of moisture might ensue. A second weighed absorption tube containing, say, concentrated sulphuric acid can be placed after the potash bulb to prove that no moisture escaped absorption in the absorption tubes during an experiment. A temoin (witness) tube in the form of a duplicate potash bulb may follow the one used for the absorption. The object of this vessel is to make sure that all the carbon dioxide passing through the system is absorbed in the first potash bulbs. If the gas from the clay has a tendency to come off in sudden rushes, this second tube is a necessary adjunct to the train (fig. 164). 2. The volume of the weighed tubes is often great enough to render necessary the corrections mentioned on page 25 for differences in the buoyancy of air due to variations of temperature and pressure during an experiment. 3. There is a difficulty in weighing the potash apparatus on account of the difference in the amounts of moisture and gases condensed on the surface of the 1 See the cane sugar and china clay experiment, page 546. Dittrich's experiments, page 247. B. Blount and A. J. Levy, Analyst, 34. 94, 1909 ; A. J. Levy, ib., 37. 3., 98. 38, 1908. 4 Landsiedl's (fig. 166) and Berl's (fig. 165) bulbs are filled by removing B (A is a slightly greased joint) and introducing the absorbent l>y means of a pipette, etc. Apply suction at C. B is plugged with glass-wool, and filled with calcium chloride as indicated page 548. Total weight about 60 gnus. 5 J. Lowe, Zeit. anal. Chem., 7. 224, 1868. " U. Kreusler, Zeit. anal. Chem., 5. 216, 1866; P. Claesson, Ber., 9. 174, 1876; M. Pettenhofer, Journ. Chem.' Soc., 10. 292, 1857 ; Journ. prakt. Chem., 82. 32, 1861 ; 85. 179, 1862 ; Liebig's Ann. Suppl^ 2. 23, 1863 ; W. Spring and L Roland, Mem. Acad. Roy. Belg , 37. 1, 1887 ; J. A. Aupperle, Journ. Amer. Chem. Soc. , 28. 858, 1906 ; A. Gregoire, J. Hendrick, E. Carpiaux, and E. Germain, Ann. Chim. Anal , 18. 1, 1913. 7 J. Wiborg, Berg. Hiltt. Ztg., 46. 233, 1887 ; A. A. Blair, Journ. Amer. Chem. Soc., 18. 223, 1896 ; G. Lunge and L. P. Marchlewski, Zeit. angeiv. Chem., 4. 229, 412, 1891 ; E. Donath and W. Ehrenhofer, Oester. Zeit. Berg. Hutt., 45. 285, 1897 ; W. Hempel, Verhand< Ver. Beford. Geiverb., 640, 1893. 552 A TREATISE ON CHEMICAL ANALYSIS. 288. The Detection of Carbon Dioxide. The presence of carbonates is indicated by a sudden effervescence when the substance is treated with dilute acids sulphuric, hydrochloric, or phosphoric acid. An excess of acid should be used, since some bicarbonates are soluble in water. The experiment is made by placing a little of the substance in a test tube ; covering it with dilute sulphuric acid ; and warming the contents of the tube. The carbon dioxide which is formed is heavier than air, and may be poured into a second test tube containing a little lime or baryta water. 1 If the second test tube be shaken, a turbidity will be produced if carbonates be present in the sample under investigation. 2 A drop of lime water hanging from the end of a glass rod held just inside the test tube, which contains the sub- stance under investigation, will show a film of calcium carbonate if carbonates be present. The test may also be conducted in the following manner. 3 Draw out the end of a narrow test tube as shown at A, fig. 167, and fit a capillary funnel or conical tube as shown at B. Put the FIG. 167. substance under investigation in the tube A, and a drop of baryta water in B. Dip the test tube at A in hydrochloric acid. 4 The presence of carbonates is shown by a white cloud forming in the baryta water as carbon dioxide passes through. 289. The Rapid Determination of Carbon Dioxide in Carbonates. As Fresenius remarks, there are few methods of analysis equal in accuracy to a process somewhat similar to that described on page 546 ; but it requires care and time. There are several short cuts, some of which are quite satis- factory ; others, not so accurate, are only justified when time must be abbreviated. There is a remarkable variety 5 of remarkably ingenious and compact instruments 1 The lime water should be kept in contact with calcium carbonate so that the solution is saturated with calcium carbonate, otherwise minute amounts might escape detection owing to the slight solubility of calcium carbonate in lime water C. L. Berthollet, Ann. Chim., 3. 68, 1789 ; H. A. von Vogel, Schiveigger's Journ., 33. 207, 1821 ; F. H. Storer, Amer. J. Science (2), 25. 42, 1858. 2 Owing to the fact that air contains carbon dioxide, baryta water will nearly always give a turbidity with a blank test. Hence, the turbidity produced by the carbonate should be more intense than the turbidity produced by the blank test. The two should show a marked difference. 3 0. Rossler, Ber., 20. 2629, 1887. In special cases, the test must be conducted in air free from carbon dioxide see F. P. Treadwell, Analytical Chemistry, New York, I. 307, 1903. Consult any work on qualitative analysis for confirmation tests. 4 Dilute sulphuric acid does not always produce effervescence with magnesite, siderite, and dolomite. The ' ' best acid " to use depends upon the nature of the carbonate. Sulphuric acid is suitable for alkaline carbonates ; hydrochloric acid for magnesium and calcium carbonates ; and nitric acid for lead carbonates white lead. With nitric and hydrochloric acids there is always a danger of losing a little acid, since some acid may be carried out of the system with the carbon dioxide, especially if the latter comes off rapidly. A. Mayer (Lands. Ver. Stat., 51, 339) says that if ferrous carbonate be present, dilute acetic acid is bes"t (1 acid, 2 water), because ferrous carbonate is not then attacked, while magnesium and calcium carbonates are attacked, H. Borntrager (Zeit. anal. Chem., 29. 141, 1890) prefers nitric to hydrochloric acid for small quantities, because hydrogen chloride and chlorine gases are not absorbed by the drying agent sulphuric acid whereas nitric and nitrous gases are. 5 See the dealers' catalogues for a few of the multitudinous forms which have been invented. For a process based on the fusion of sample with sodium paratungstate, see F. A. Gooch and S. B. Kuzirian, Amer. J. Science (4), 31. 497, 1911. The loss in weight of the mixture THE DETERMINATION OF' CARBON FREE AND COMBINED. 553 in which the acid, the carbonate, and the desiccating agent are placed in separate compartments of one instrument. All are weighed together. The acid is then brought in contact with the carbonate, and when the action is over, the loss in weight represents the carbon dioxide which has left the system. It is not easy to get a degree of accuracy less than about half per cent, with these instruments, except by taking precautions which involve the expenditure of nearly as much time as the gravimetric process next described. If an accuracy of about one per cent, be desired, 1 we can select an instrument from the dealers' catalogues. We natur- ally reject those which are cumbrous, fragile, difficult to clean, and expensive. Rohrbeck's and Schrbtter's are favourite forms. 2 Fig. 168 represents an apparatus which resembles an ordinary weighing bottle in external appearance. 3 It is easy to clean and fill. Place about a gram 4 of the carbonate under investigation in B, and some concentrated sulphuric acid in the outer vessel A. The pipette F is filled with the required acid (about 5 c.c.) either by suction or by resting the tip B on a rubber A stopper and introducing the acid by means of a pipette at E!> The air in the capillary prevents acid running through, and the stopper E is inserted before F is lifted from the rubber. Everything is then fitted as indicated in the diagram. Wipe the outer tube, and FIG. 168. C0 2 apparatus, when the vessel has stood in the balance case a short time, it is weighed. The caps D and E are then removed, acid runs on to the carbonate, the gas bubbles through the concentrated sulphuric acid and escapes via D. When all action is over, a slow current of air may be aspirated through the system, by applying suction at D, and connecting E with a soda lime and calcium chloride tube. Replace the caps, and weigh as before. The loss in weight represents the carbon dioxide which has left the system. The results are improved a little if the weighing bottle be placed in a little warm water while the air is being aspirated through the system. The folio wing, numbers represent the results which might be expected : Used(CaC0 3 ) . . 0-2153 0-2011 0-2004 0'2006 0-2020 . 0'2018 grm. Found (CaC0 3 ) . . 0'2149 0'2010 0'1999 0'2003 ' 0'2017 0'2017 grm. Error . . . -0-0004 -O'OOOl -0'0005 -0'0003 -0'0003 -O'OOOl grm. 290. The Gravimetric Determination of Carbon Dioxide. With clays containing carbonates it is advisable to determine the carbon dioxide independently of the carbon, and subtract the carbon dioxide from before and after fusion represents the carbon dioxide expelled. Processes like this have a limited application. See page 520. 1 The results are better with substances rich in carbonates. 2 A. R. von Schrbtter's (Ber. Wien. Acad., 63. 471, 1871), J. Davies' (Brit. Pat. No. 533209, 1908), J. L. Kreider's (Amer. J. Science (4), 19. 188, 1905; Zeit. anorg. Chem., 44. 154, 1905 ; Chem. Neivs, 93. 62, 1906), C. H. Cribb's (Analyst, 21. 62, 1896), etc., are useful forms. The apparatus fig. 168 was made for me by Gallenkamps. 3 The apparatus, slightly modified, can be easily made from regular laboratory apparatus old pipettes, a weighing bottle, etc. 4 The amount is determined by the percentage of carbonates in the given sample. We generally have a rough idea. Two or three grams may be used when the substance is poor in carbonates. 5 The acid should run out slowly in small quantities at a time. This is best arranged by making the capillary tubes C and B narrow. 554 A TREATISE ON CHEMICAL ANALYSIS. the decomposition of the carbonate from the carbon dioxide derived from the combustion of the carbonaceous matters. The carbon dioxide can be accurately determined in an apparatus (fig. 164) modified as indicated in fig. 169. 1 The quartz capillary is to be replaced by a double U-tube E, one half filled with pumice saturated with anhydrous copper sulphate, 2 and the other half with calcium chloride. The White's tube D contains sulphuric acid. The other tubes, etc., are filled as indicated on pages 348 and 349 in the wet combustion process for carbon. The amount of clay to be taken for an experiment depends upon the FIG. 169. Determination of carbon dioxide in carbonates. amount of carbonates present. From 5 to 50 grms. of the dried clay are transferred to the flask, and this is made into slip with boiled distilled water free from carbon dioxide. Hydrochloric or sulphuric acid is added slowly by opening the plug cock. When effervescence 3 has ceased and enough acid has been added to decompose all the carbonates, the flask is gently heated for about 15 minutes while a current of air is slowly aspirated through the system. 1 R. Fresenius, Quantitative Chemical Analysis, London, 2, 340, 1876 ; Zeit. anal. Chem., 14. 174, 1875; A. Classen, ib., 15. 288, 1876; H. Kolbe, Liebig's Ann., 119. 130, 1861; J. Hessert, ib., 176. 136, 1875; J. Volhard, ib., 176. 142, 1875; H. Rose, Pogg. Ann., Il6. 131, 1862 ; J. Persoz, Compt. Rend., 53. 239, 1861; L. T. Bowser, Journ. Ind. Eng. Chem., 4. 203, 1912; H. W. Bnibaker, ib., 4. 599, 1912 ; E. W. Gaither, ib., 4. 611, 1912. ' 2 F. Stolba, Dingier 's Journ. . 164. 128, 1862 ; Zeit. anal. Chem., 10. 76, 1871 ; R. Fresenius ib., 14. 174, 1875. The pumice is prepared by sifting about 60 grms. of fragments of pumice, about the size of peas, free from dust. The fragments are placed in an evaporating basin with a concentrated solution of copper sulphate (30-35 grms.). Evaporate the solution ,to dry- ness with constant stirring. Heat 4 or 5 hours in an air bath at 150-160, not more, or sulphur dioxide will be formed. The pumice so prepared will remove hydrogen sulphide and hydrogen chloride from the gases. 3 If the substance froths and foams, it may be necessary to use a comparatively large flask. THE DETERMINATION OF CARBON FREE AND COMBINED. 555 In illustration of the weighings, the following numbers may be cited : 2 grms. of clay were treated with dilute hydrochloric acid, and : Weight of potash bulbs (after) * . 78'1863grms. Potash bulbs (before) ......... 78 "1326 grms. Carbon dioxide '0537 grm. 0'0537 grm. of carbon dioxide per 2 grms. of clay represents 2*68 per cent, of carbon dioxide, or 6 '09 per cent, of calcium carbonate, since weight of carbon dioxide multiplied by 2 27432 represents the corresponding amount of calcium carbonate, CaC0 3 ; and if the weight of carbon dioxide be multiplied by 1'2748 the corresponding amount of calcium oxide, CaO. Of course, the carbon dioxide may be wholly or in part combined as magnesium carbonate ; as dolomite ; as ferrous carbonate (siderite) ; or a basic iron carbonate. If w denotes the weight of carbon dioxide, we have 2 - 2748w? = Per cent, calcium carbonate, CaC0 3 = Per cent, magnesium carbonate, MgC0 3 = Per cent, dolomite, M^C0 3 . CaC0 3 2 -6330^ = Per cent, ferrous carbonate, FeC0 3 Success in this determination largely depends upon the speed at which the carbon dioxide is liberated. If the gas comes off too rapidly, some carbon dioxide may pass through the system without giving up its moisture to the desiccating agent. Concentrated hydrochloric or nitric acid, or dilute sulphuric acid, may be used as indicated on page 552 Morgan 1 recommends orthophosphoric acid (sp. gr, 1'75) for decomposing the carbonates, and then the addition of a moderate excess of chromic acid (3-4 grms.), whereby the carbon is oxidised. Thus carbon dioxide and carbon may be determined separately on the same sample. I have had no experience with this phosphoric acid process. 291. The Volumetric Determination of Carbon Dioxide in Carbonates Scheibler and Dietrich's Process. The volume of the carbon dioxide obtained by the action of acids on car- bonates can be measured quite accurately. A large number of instruments have been devised and modified for this purpose. 2 Scheibler's instrument is one of the oldest, and, with some modifications, one of the best. Scheibler and Dietrich's instrument, illustrated in fig. 170, is fairly common. The Apparatus. In this apparatus, a measuring tube A is connected with a levelling tube B by means of a piece of thick-walled rubber tubing. A is graduated from to 200 c.c. ; B slides on the upright of the stand. The 1 G. T. Morgan, Journ. Chem. Soc., 85. 1001, 1904. 2 C. Scheibler, Anleitung zum Gebrauch des Apparates zur Bestimmunq des Koklensauren KalJcerde in der Knockenkohle, Berlin, 1862; Chem. News, 22. 75, 1870; R. Warington, ib., 31. 253, 1875; E. Nicholson, ib., 2g. 245, 1875; H. Fresenius, Zeit. anal. Chem., 19. 206, 1880; E. Dietrich, ib., 3. 162, 1864; 4. 141, 1865; 5. 49, 1866; D. Sidersky, ib., 25. 93, 1886; 26. 336, 1887; E. Jager and G. Kriiss, ib., 27. 721, 1888; R. Baur, ib., 23. 371, 1884 ; A. Classen, ib., 15. 288, 1875; F. Schulze, ib., 2. 289, 1863; A. Gawalovski, ib., 18. 244, 560, 1879; G. Burkhardt, Neue Zeit. Rub. Lid., 16. 115, 1886; D. Sidersky, Zeit. Ver. Rub. Ind., 2O. 919, 1885; W. Borchers, Journ. prakt. Chem. (2), 17. 353; G. Lunge, Chein. Ind., 8. 166, 1885; F. Fuchs, Chem. Ztg., 13. 873, 1889; R. Finkener, Zeit. angew. Chem., 3. 273, 1890; W. Thorner, ib.,2. 041, 1889; E. Cramer, Tonind. Ztg.. 18. 577, 1894. 556 A TREATISE ON CHEMICAL ANALYSIS. FIG. 170. Scheibler and Dietrich's apparatus. upper end of A is connected with a three- way cock d which connects A either with the air, or with a third tube C to cool the gases which come from F, and flask F. A is tilled with a 1 per cent, solution of boric acid. If all the connections are good, when the liquid in A is at zero, it will remain at zero if the levelling tube be depressed for some time, and then raised again. Adjustment of the Apparatus. When the instrument is connected up as in the diagram, read the barometer and thermo- meter. A table is supplied with the instru- ment. This indicates half the amount of substance to be weighed out for all variations of temperature and pressure between 10 and 25, and 720 mm. and 770 mm. The right amount of the powdered substance l is weighed into the flask F, and a small tube containing 5 c.c. of hydrochloric acid (sp. gr. 1*124) is placed in the same flask. The acid tube is reared against the side of the flask, so that the acid does not come in contact with the powder. The flask is closed by means of a rubber stopper. The liquid in A is set so that the lower level of the menis- cus is at the zero of the scale when the liquid in the levelling tube is at the same level. Evolution of Gas. The generating flask is now tilted so that the acid comes in con- tact with the carbonate. When all action has ceased, and the apparatus has cooled two or three minutes, the level of the liquid in A and B is adjusted, and the lower level of the meniscus in A is read. 2 Add to this a correction for the volume of the carbon dioxide dissolved by the liquid in the gener- ating flask F. z * Half the sum represents the percentage amount of carbon dioxide in the given sample. 292. The Volumetric Determination of Carbon Dioxide in Carbonates Lunge and Marchlewski's Process. The apparatus just mentioned gives good results, but it has some weak points. Lunge and Marchlewski 4 have introduced several improvements, but 1 As indicated in the table, multiplied by 2. The reason for taking the amount of sample different with different temperatures and pressures is to avoid calculating the weight of gas corresponding with the volume measured in A. 2 If after the first action, and the volume of the gas has been read, gas is still slowly evolved from the generating flask, magnesium carbonate is present. 3 In 0. Petterson's method (Ber., 23. 1402, 1890) the carbon dioxide is driven from the gener- ating flask by the simultaneous action of the acid on iron or aluminium wire. Carbon dioxide and hydrogen collect in the measuring flask. The former is determined in an Orsat's apparatus. 4 G. Lunge and L, Marchlewski, Zeii. angew. Chem., 4. 229, 412, 1891 ; 6. 395, 1893 ; THE DETERMINATION OF CARBON FREE AND COMBINED. 557 their apparatus, shown in fig. 171, is somewhat expensive (3, 5s.) Still, it is one of the best instruments for the purpose on the market. The Apparatus. In the diagram (fig. 171) A is a flask (about 30 or 40 c.c. capacity) closed by a soft rubber stopper. The stopper is fitted with a dropping funnel a and a capillary tube bent at right angles, as shown in the diagram. B is the gas measuring tube, D is the levelling tube, and G is a compensating tube whose function is described below. These tubes are connected by means of thick-walled rubber tubing, and mounted as shown in the diagram. E is an Orsat's absorption tube con- taining soda solution or potash. 1 The soda lime tube F protects the contents of the Orsat's bulb from atmospheric carbon dioxide. G and H are three-way stopcocks. The apparatus is used in the following manner : Adjustment of the Apparatus. The decomposition flask A is cleaned perfectly free from acid, and an amount of substance which will give not more than 50 c.c. of carbon dioxide is weighed into the flask along with 0*8 grm. of thin aluminium wire. 2 Raise the mer- cury in B until it reaches the stopcock 6r. 3 Connect A and B, but not E and B. Depress the mercury in B by lowering D as far as possible. Turn the stopcock G so that when the levelling tube is raised the air in B will be driven out of the system. When the mercury reaches G, connect A with B, and repeat the operation three or four times so as to partially evacuate the vessel A. Evolution of the Gas. With the mercury in the levelling tube C lower than in B, and C and B in communication, add about 10 c.c. of hydrochloric acid (3 vols. water, 1 vol. acid) via the funnel a. Close the stopcock of the funnel just before the last drop of acid runs into the flask. Carbon dioxide is evolved at W. Thorner, ib., 2. 644, 1889 ; R. Lorenz, ib., 6. 395, 411, 1893 ; M. A. von Reis, Stahl Eisen, 8. 257, 1887 ; 0. Vogel, ib., II. 486, 1890 ; C. Reinhardt, ib., 12. 648, 1040, 1891 ; G Lunge, Chem. Ztg., 12. 821. 1 SODIUM HYDROXIDE SOLUTION. 104 grms. of ordinary caustic soda dissolved in 130 c.c. of water will make about 200 c.c. of a solution approximately 13N. See page 548. 2 Weigh, say, 10 metres of the thin wire, and calculate the length required to furnish 0-8 grm. Lengths cut this size can be kept in a bottle ready for use. 3 This requires care, or mercury will be driven where it is not wanted. FIG. 171. Lunge and Marchlewski's apparatus. 558 A. TREATISE ON CHEMICAL ANALYSIS. once. The mercury in B is depressed. Lower the levelling tube C, so that, all the time, the mercury in B is lower than in C. The flask A is gently warmed with a spirit lamp for about 2 minutes. Add more acid, and repeat the operation. 1 In this manner nearly all the carbon dioxide is driven from the liquid in A by the escaping hydrogen. When all has dissolved, add dilute acid through the funnel to the flask until the flask A, as well as the connecting tube, are filled almost as far as the stopcock G. Let all stand for about 10 minutes so as to acquire the temperature of the room. Meanwhile the flask A and its connecting tube can be removed and cleaned. Measuring the Volume of the Gas. Adjust the levels of the mercury in C and D until the mercury in C is at the 100-c.c. mark, and at the same level in B. Now read the volume of the gas confined in B from the top of the mercury meniscus. Suppose it be 147 '5 c.c. The level of the liquid in the Orsat's tube E is adjusted by means of the three-way cock ff, and by blowing through the tube F until the soda solution stands at the mark just below the stopcock. Close the stopcock H. Let the mercury in the levelling tube stand at a higher level than in B. Connect the Orsat's bulb E with B. Raise D gradually until the mercury in B is near the stopcock G. Then depress D until the soda solution stands at its former level. Repeat this operation three times. The soda solution absorbs the carbon dioxide. The motions require care, because it is easy to get the soda solution into the measuring tube B, and mercury into the Orsat's tube E. A little practice and attention will soon give control of the movements of the fluids. When the level of the soda solution is at its mark below the stopcock H, close the cocks H and Na 2 Si03 + 2C. Hence, if a dark-coloured residue remains, it does not necessarily mean undecomposed powder. It may be unburnt carbon. The suspended undecomposed matter is filtered off, washed, calcined in a platinum crucible, and re-fused with sodium carbonate. The cooled mass is taken up with water and acid and mixed with the main portion. THE DETERMINATION OF CARBON FREE AND COMBINED. 565 Very few investigations have been made pn the analysis of carborundum and silicon carbides. We sometimes have to deal with silicon carbide present in old graphite crucibles, 1 and with carborundum admixed with clay in com- pounding certain bodies. In that case, Miihlhaeuser's method of grinding must be rejected. The 1-min. powder has a different composition from the 5-min. powder. The grinding need not be so protracted if the alkalies have not to be determined. 2 The substance may then be first broken in a hard steel mortar, and finished with an agate mortar and pestle. 0'5 grm. may be calcined for loss on ignition 3 in a platinum crucible. A crucible with 1 grm. of powdered car- borundum weighed 27*9834 grms. After 20 minutes' calcination over a Bunsen's burner, the whole weighed 27-7413; after another 12 minutes' calcination, 27-7390 grms. ; after 25 minutes' further calcination on the blast, 27*7398 ; after 20 minutes' more blasting the weight was 27*7415. No marked change in weight occurred with a more prolonged blasting. The increase is probably due to oxidation of ferrous oxide (page 157). The calcination for loss on ignition is to be conducted over a Bunsen's burner until no further loss in weight occurs. Carborundum gave a loss of 0'88 per cent., followed by a slight gain in weight on further calcination. A known weight is mixed with 5 to 6 grms. of red lead, 4 or lead carbonate, and a gram of sodium nitrite. The mixture is fused over a small Meker's burner to ensure oxidising condi- tions. When the melt is clear, cool the hot crucible with its lid in position by plunging the hot crucible in cold distilled water. The cake is usually easy to loosen with dilute nitric acid (not hydrochloric acid). Transfer the mass to an evaporating basin and evaporate with concentrated nitric acid to dryness. The fragments gradually soften, and they can be rubbed with an agate pestle from time to time. The dry mass is treated for silica, as described for clays, but nitric acid is used in place of hydrochloric acid. The filtrate is treated in the cold with hydrochloric acid in order to precipi- tate most of the lead as chloride. Filter off the lead chloride, and wash with hydrochloric acid (1:1). Evaporate the filtrate to dryness and take up the residue with hydrochloric acid (1:4). Saturate the solution with hydrogen sulphide. Filter off the lead sulphide. Boil the filtrate to expel the hydrogen sulphide. The alumina may then be determined in the solution in the usual manner (page 182). The Estimation of Silicon Carbide SiC. There is a difficulty in calculating the amount of silicon carbide from the analytical data. For instance, suppose that the analysis furnishes : 1 Formed in the body of the crucible itself while in use. - Which is not usually the case. Wdowiszewski (I.e.) puts for his final analysis : ' ' KjO + Na 2 4- analytical errors 1 percent." 3 Note the difficulty with " loss on ignition " owing to the fact that silicon carbides do not burn even in a current of oxygen. The object of the ignition is to burn off as much of the carbon as possible in order to prevent the reduction of the lead oxide to metallic lead, which would ruin the platinum crucible. If alkalies are to be determined, the sodium nitrite must, of course, be omitted. 4 The red lead must be specially purified, or a blank analysis made so as to be able to correct the final product for silica, alumina, etc. If the red lead contains particles of metallic lead, it will spoil the platinum crucible. The red lead is conveniently made by treating an aqueous solution of pure commercial lead nitrate with a saturated solution of oxalic acid containing 3 per cent, of nitric acid. The precipitate is separated, washed, dried, and ignited at a dull red heat for some hours. Lead carbonate is made by precipitating a solution of lead acetate with the calculated amount of ammonium carbonate, etc. If the sample under investigation contains organic matter, the carbon should be destroyed by gentle ignition. A. F. Crosse (Journ. Chem. Met. Soc. S. Africa, 2. 182, 1897 ; Chem. News, 76. 253, 1897), in "assaying graphite crucibles, " mixes graphite crucibles with enough manganese dioxide to burn all the graphite. 566 A TREATISE ON CHEMICAL ANALYSIS. Silica . Carbon Alumina anc Magnesia Lime . Potash . Soda . ferri c oxic * e Per cent. 48-34 50-12 16-07 0-53 1-02 0'21 0-09 Total. . . .:, . *:.. . 116-38 This result assumes that all the silicon is present as silica. Some is present as silicon carbide, presumably SiC. It is assumed that if the amount of SiC, free silica and combined silicon, and free and combined carbon had been determined, the total would have been 100 instead of 116'38. Part of the oxygen which we have supposed to be combined with the silica was not present in the given sample. Since 32 parts of oxygen correspond with 60 parts of silica, 16 -38 parts of oxygen correspond with 30 '71 parts of silica. 1 Hence, 48 -34 less 30 '71 = 17 '63 per cent, of free silica. Again, 60 parts of silica correspond with 40 parts of silicon carbide SiC; hence, 30'71 parts of silica correspond with 20'47 parts of SiC. Again, 40 parts of silicon carbide correspond with 12 parts of carbon; hence, 20'47 parts of silicon carbide correspond with 6 '14 parts of carbon as silicon carbide. Hence, 50'12 less 6-14 represents 43'98 per cent, of free carbon. The first two terms of the preceding analysis may now be revised. They read : Silica 17-6 Silicon carbide 20 '5 Carbon 44'0 The collected errors in the analysis are here distributed between the silica, silicon carbide, and the carbon, and the method can only be regarded as a method of approximation. According to Parr, 2 the free silica is removed by evaporation with hydrofluoric acid, which does not attack silicon carbide SiC. Hence, determine the total silicon by the fusion process, and the free silica in a separate sample by means of hydrofluoric acid. He expresses the analysis : silica (volatilised with HF), 8-27 ; metallic iron (removed by a magnet), 4-37 ; silicon as Si (by fusion process), 63*58 ; carbon (volumetric from fusion mixture, page 562), 23'67 per cent. 295. The Analysis of Siloxicon. Siloxicon is a greenish-grey refractory substance which varies in composition between Si 2 C 2 and Si 7 C. 2 0. It is vigorously decomposed, but not completely, by a fused mixture of sodium carbonate and potassium nitrate. Hydrofluoric and sulphuric acids have but a slight action. It is vigorously attacked by heated lead peroxide. 3 Spielmann 4 analysed a sample of the substance in the following manner : Silicon, Carbon, and Moisture. The amount of moisture is determined, in the usual way, by drying the sample to a constant weight at 110. The substance is " opened " by fusing the sample with sodium peroxide in a nickel 1 It will be noted that we use rounded atomic weights. The resulting error is negligibly small in comparison with the approximate nature of the method of computation. 2 S. W. Parr, Journ. Amer. Chem. Soc., 30. 764, 1908. 3 P. Jannasch and H. J. Locke, Zeit. anorg. Chem., 6. 168, 321, 1894 ; A. Leclerc, Compt. Rend., 125. 893, 1897 ; P. Jannasch, Zeit. anorg. Chem., 8. 364, 1895 ; Chem. News, 72. 51, 1895. 4 P. E. Spielmann, Journ. Soc. Chem. Ind., 24. 654, 1905. THE DETERMINATION OF CARBON FREE AND COMBINED. 567 crucible (page 266). The siloxicon is vigorously attacked, and the resulting cake, when cold, is digested in water, and then with dilute hydrochloric acid. The resulting solution is treated in the usual manner, and the amount of silica obtained is multiplied by 0'46932. This represents the amount of silicon Si in the given sample. The total carbon is determined by fusing the sample with sodium peroxide in an iron boat in a combustion tube in a gentle stream of oxygen. In one experiment a potash bulb was fitted to the tube to collect any carbon dioxide which might escape ; but none did escape, and hence the carbon was afterwards estimated from the amount of carbonate in the product of the fusion of the siloxicon with the sodium peroxide in a crucible. The carbon dioxide in the fused mass was determined by the action of dilute sulphuric acid l in the usual manner (page 552). Alumina, Iron, and Si 2 C 2 0. Another sample was heated in a stream of chlorine in a porcelain boat in a hard combustion tube, whereby silicon tetrachloride, ferric chloride, and a trace of aluminium chloride were volatilised. These were absorbed by passage (1) over the surface of water in a flask; and (2) through a U-tube loosely packed with cotton-wool. The silicon tetrachloride was hydrolysed to silicic acid. When all the ferric chloride had volatilised, the combustion tube was heated by the blast* to as high a temperature as the tube would stand without squatting. The reaction is at an end when the contents of the boat cease to glow. The cotton-wool was burnt in a crucible, and the oxides of silicon and iron remained. The aluminium, iron, and silicon collected in the flask and by the cotton-wool were determined as usual. The residue in the boat was heated in a current of oxygen, and the carbon dioxide determined by absorption, etc., in potash bulbs. The residue in the boat was found to be almost pure Si 2 C 2 0. 2 Calculations. Spielmann collects his results in the following manner : Per cent. Total silicon (Na 2 2 fusion) .'. . . - ' . . 50 '31 Total carbon (Na 2 2 fusion) . . ,' 31 ; 39 Iron (heating in chlorine) . . , . ... . . 1 '07 Aluminium (heating in chlorine) . . ' . . . ... trace Moisture - ' 19 82-96 Oxygen (difference from 100) . . . ... 17'04 Residue in boat (Si 2 C 2 0) . . . ' !, ?? Carbon on heating chlorine residue in oxygen . . . .LI '79 The latter represents the carbon which occurred in siloxicon partly as graphite and partly as carborundum (decomposed by chlorine). This quantity, with the amount calculated for Si 2 C 2 and subtracted from the total carbon, leaves 1-47 per cent, of carbon. The oxygen in Si 2 C 2 subtracted from the tote, oxygen leaves 5*24 per cent, of oxygen. This, with the 1'47 per cent, carbon, is that required for SiC0 3 . This amount of silicon subtracted from the silicoi volatilised by the chlorine (7 -54 per cent.) gives the amount of silicon present as carborundum, i.e. 4'08 per cent. The amount of carbon thus combined 1-73 per cent. This subtracted from the weight of carbon burnt on oxygen leaves the amount of carbon present as graphite. Collecting thes< together, the analysis reads : 1 If hydrochloric acid be used, some chlorine will be formed. 2 A portion of the sample was boiled with hydrofluoric acid. The filtrate contained trace of ammonium silicofluoride. This was probably due to the presence of a trace o nitride in the original sample. 568 A TREATISE ON CHEMICAL ANALYSIS. Si 9 C 2 . . . . . ... . . . 71-39 SifcO, ........... 10-18 SiC (carborundum) ......... 5'81 Graphite .......... 10'06 Fe P ............ 1-07 VolatUes ........... 0'19 Ai 2 o 3 ; si 3 N 4 (by difi.) -67 The method of calculation can only be regarded as the roughest of approximations. 296. The Analysis of Graphite, and Graphite Crucibles. If the sample has no chemically combined water, as is the case with some of the artificial graphites, the carbon can be simply determined by heating the dried sample in a capacious Rose's crucible fitted so that a current of dry oxygen can be passed into the crucible while the calcination is in progress. The loss in weight represents the graphite burnt. 1 With 'natural graphites, however, this method is risky, because some silicates may be present which lose water only at elevated temperatures. Again, if' pyrites, FeS 2 , be present, the sulphur is burnt to the dioxide, and ferric oxide is formed such that one part of pyrites furnishes two-thirds its weight of ferric oxide. If appreciable quantities of occluded oxygen, hydrogen, nitrogen, and sulphur be present, this method will give erroneous results. 2 Berthier's old process 3 is not infrequently employed when rapid work is needed. 0*5 grm. of the finely powdered sample is intimately mixed with 12 grms. of powdered lead monoxide, and placed in an unglazed porcelain crucible. The surface of the mixture is covered with 1 2 more grms. of lead monoxide, and heated slowly. The lead monoxide is reduced by the carbon so that 1 grm. of reduced lead represents 0*03 grm. of graphite. Quite good results can be obtained by this process if care be exercised in selecting the lead monoxide, and if the sample be free from sulphides. The sample can be readily powdered in a hardened steel mortar with a ball pestle (fig. 69), and finished in an agate mortar. The wet combustion process gives rather low results. 4 Hence, pack a hard glass combustion tube with an asbestos plug as indicated on page 563. Then fill about 45 cm. of the tube with granulated or wire copper oxide 5 instead of granulated lead oxide. Then insert another plug of the asbestos. Mix, say, O2 to 0*3 grm. of the powder with powdered copper oxide, 6 and transfer the mass to a porcelain boat. Place the boat in the combustion tube. Follow this with a roll of copper gauze, 13 cm. long, bound 1 E. Donath, Der Graphite, Wien, 163, 1904. See also page 545, 285. 2 E. Donath, Der Graphite, Wien, 168, 1904 ; G. Auchy, Journ. Amer. Chem. Soc., 22. 47, 1900. 3 P. Berthier, Dingler's Journ., 58. 391, 1835 ; Traite des Essais par la Voie Seche, Paris, i. 222, 1847 ; W. F. Gintl, Zeit. anal. Chem., 7. 423, 1868 ; G. C. Wittstein, Dingler's Journ. , 2l6. 45, 1875. G. Forchhammer (Berg. Hutt. Ztg., 5. 465, 1846) recommended a mixture of three parts of lead monoxide and one part of lead chloride in place of lead monoxide alone, because the former fuses at a lower temperature and does not corrode the crucible so much. A. Schrotter (Dingler's Journ., 116. 115, 1850) used lead oxychloride. 4 J. Widmer, Zeit. anal. Chem., 29. 160, 1890 ; Chem. News, 62. 274, 1890. J. Goldstein, (Chem. Ztg., 35. 1134, 1911) considers that the combustion with sulphuric and chromic acids in Corleis' flask, etc., page 546, gives best results. 5 C. Reischauer, Viert. prakt. Pharm., u. 38, 1862; E. C. C. Stanford, Chem. News, 7. 81, 1863 ; E. Erlenmeyer, Zeit. Pharm., 6. 156, 1854. 6 If the ash is liable to sinter, calcined magnesia or zinc oxide is mixed with it S. S. Sadtler, Journ. Franklin Inst., 144. 201, 1907 otherwise particles of unburnt carbon may be sealed in the slag and thus protected from oxidation. THE DETERMINATION OF CARBON FREE AND COMBINED. 569 with copper wire arranged with a hook at one end. The roll should slide easily into the combustion tube, after the boat 1 see fig. 177. The combustion Spiral Boat Asbestos- Plug - Copper Oxtde >l Asbestos FIG. 177. Diagrammatic sketch of combustion tube. is conducted as indicated on page 563. 2 When silicon carbides are absent, the loss on ignition, 3 silica, etc., can be determined as for clays. 4 1 The copper gauze should be heated in a Bunsen's flame so as to burn off oil and combustible matters. E. Calberia (Journ. prakt. Chem. (1), 104. 232, 1869) prefers silver gauze. 2 The tube should be heated in a current of oxygen before the boat is introduced in order to burn out every particle of dust and moisture. 3 F. Stolba, Dingler's Journ., 198. 213, 1870. Due allowance is of course made for the carbon. 4 F. Mayer (Chem. Ztg., 35. 1024, 1911) estimates the amount of carbon in graphite from its calorific power. CHAPTER XXXIX. THE DETERMINATION OF WATER. 297. Brush and Penfield's Method. IT is sometimes desirable to determine directly the amount of water evolved when clays and related materials are ignited, although the determination is seldom asked for in industrial work. Brush and Penfield l conduct the process in the following manner : One or two bulbs are blown on a piece of hard glass combustion tube about 25 cm. long, with an internal diameter of about 0*6 mm., FIG. 178. Determination of water Penfield's process. as shown in the diagram, fig. 178. Air is blown through the hot tube by means of a piece of glass tubing reaching nearly to the Jbottom of the combustion tube so that the tube may be thoroughly dried. Weigh the tube with its support. 2 Introduce about 0*5 grm. of the powdered sample by means of a kind of thistle funnel, a, fig. 179, 3 without soiling the tube away from the closed end. The 1 G. J. Brush, Amer. J. Science (1), 46. 240, 1868 ; S. L. Penfield, ib. (3), 48. 31, 1894 ; Zeit. anorg. Chem., 7. 22, 1894; F. A. Gooch, Amer. Chem. Journ., 2. 247, 1880; W. F. Hillebrand, Bull. U.S. Geol. Sur., 422. 72, 1910. For a general study of the subject, see G. N. Huntley and J. H. Coste, Journ. Soc. Chem. Ind., 32. 62, 1913. 2 Fora brass " tube support" for steadying the tube on the pan of the balance, see fig. 3a. 3 Can be made from a small pipette. 570 THE DETERMINATION OF WATER. 571 powder should occupy 2 or 3 cm. of the tube. Weigh the tube, support, and sample together. Take care not to roll the powder in the tube towards the bulbs. 1 The open end of the tube may be fitted with a piece of rubber tube holding a glass tube drawn out to a capillary end, b, fig. 180. Tap the tube so as to form a free passage for steam, etc., above the powder. Support the tube with a very slight slope downwards from the closed end ; wrap a strip of filter paper or cloth about the bulb and tube near the open end fig. 178. Keep the filter paper moist and cold so as to ensure condensation of the moisture expelled from the FIG. 179. powder. The powder is now heated gradually up to the full heat of the burner. If the heated end tends to sink, turn the tube around from time to time. 2 In about 15 minutes, drive the water in the tube a safe distance from the closed end, and seal off the part containing the water by means of a blowpipe flame. Let this portion of the tube cool in a horizontal position. Wipe the outside clean, and weigh. Blow the steam and moisture from the other tube by placing a small tube inside the larger one and reaching to the bottom. The loss in weight represents the total moisture. The weighings, etc., are made as indicated in the following example : Weight of tube plus powder Weight of empty tube ...... 69-9214 grms. 69-4321 grms. 0'4893 grm. Weight of portion containing water . . . . Weight after expulsion of water . . 20 '8 142 grms. 20*7997 grms. Weight of water . . . ' .>" . . 0-0245 grm. Hence the sample contained 5'01 per cent, of water. 298. Jannasch's Process for Water. Jannasch 3 determines the moisture in the presence of sulphur, fluorine, etc., by fusing the substance with about six times its weight of lead oxide and collect- ing the water evolved in calcium chloride tubes. The other gases are said to be retained by the lead oxide. A hard glass tube, 26 cm. long, with an internal diameter of 1 cm., has a long bulb, A, blown about 11 cm. from one end. The 1 If the sample has not been dried, it may now be dried at, say, 109, and the system weighed again for " water lost at 109." ? For minerals like talc which do not give up all their water even when blasted, Penfae.d uses a cylinder of platinum foil (about 0'07 mm. thick) sprung tightly inside the part ot the glass tube which is to be heated intensely, so as to prevent the glass squatting. The outside of the tube is surrounded with a piece of asbestos board. The end of the tube to be neatec partly surrounded with blocks of charcoal and fireclay so arranged that the tube is in a small furnace, which, later on, can be heated by a blast gas blowpipe. 3 P. Jannasch, Praktischer Leitfaden der Qewicktsanalyse, Leipzig, 357, 1904 ; P Jannasch and P. Weingarten, Zeit. anorg. Chem., 8. 352, 1895 ; P. Jannasch and J. Locke ib , 6. 163, 1894; M. Dittrich and W. Eitel. ib., 75- 373, 1912; Sitzber. Heidelberger Akad. Wiss. 19 1911 ; W. Eitel, Die Bestimmung des Wassers in Silikat- Miner alien und Gestemcn, Frankturt a. M., 1912. 572 A TREATISE ON CHEMICAL ANALYSIS. bulb must have thick walls, and the whole tube thoroughly dried. In fig. 180, b b are plugs of glass-wool ; c, a loosely packed layer of Pb0 2 and PbO (page 547) about 5 cm. long. The object of the lead oxides is to retain the chlorine, fluorine, sulphur, etc. This part of the tube must be kept hot while an experiment is in progress. One end of this tube is fitted with a calcium chloride tube B, and potash bottles H H; the opposite end is fitted with a calcium chloride tube (page 561) C ; a potash bulb D with its calcium chloride tube E ; guard tube F \ and aspirator G. The object of the potash bottles H ff, and calcium chloride tube B, is to remove moisture and carbon dioxide from the aspirated air (page 567). From 0'5 to 1 grm. of the powder is mixed with six times its weight of lead monoxide 1 on glazed paper, and the mixture is transferred to the middle of the bulb by means of a long strip of glazed paper and a feather. The operation requires care. No powder must be lost, and none left sticking to the sides of the tube. 2 Now raise the temperature of the bulb and contents gradually, and finally finish the ignition at a full red heat. The calcium chloride absorbs the water ; 3 the FIG. 180. Direct determination of water. potash, the carbon dioxide. After about 10 minutes' ignition the reaction should be completed. Let the tube cool in a current of dry air freed from carbon dioxide. The tube C and the set of bulbs DE are then weighed. The increases in weight in the respective tubes represent the absorbed water and carbon dioxide. The simultaneous determination of carbon or carbon dioxide and moisture has been discussed on pages 562-4. Assuming that the hygroscopic moisture has been expelled by drying the sample at 110, a certain proportion of the remaining water in carbonaceous clays will be derived from the clay ; and the remainder, from the organic matter. In that case, digest a weighed portion of the sample with a mixture of two volumes of faming hydrofluoric acid and one volume of hydrochloric acid (sp. gr. 1'18); evaporate to dryness. Repeat the operation three times. Digest the residue with hot water, filter through an asbestos-packed Gooch's crucible or carbon tube, wash, and dry at 110. After weighing, the carbon may be detached from the filter and the ratio of carbon to hydrogen determined by combustion of the organic matter to carbon dioxide 1 Freed from carbon dioxide by heating in a porcelain tube. 2 For the amount of moisture which calcium chloride, sulphuric acid, and phosphoric oxide leave in a gas, see page 156. 3 E. Ludwig, Tschermatfs Mitt. (1), 2. 214, 1875; Zeit. anal. Chem., 17. 206, 1878; L. Sipocz, ib., 17. 207, 1878; Sitzber. K.K. Akad. Wiss. Wien, 86. 51, 1877. The former used a platinum tube ; the latter, a platinum boat. THE DETERMINATION OF WATER. 573 and water. 1 Jannasch's process or one of the methods indicated in the pre- ceding chapter may be used for the combustion. EXAMPLE. A carbonaceous clay gave the following percentage results : Jannasch's process . . Total water . . . . . . 14-21 Lissners process . . Total organic matter . . t . . . 4'32 Combustion process . . Carbon (calculated from C0 2 ) . . . . 4*16 Hydrogen (calculated from H 2 0) . . . O'lO Organic matter 4 '26 Hence, 4*26 2 grms. of organic matter have 4'16 grms. of carbon, and O10 grm. of hydrogen. But 010 grm. of hydrogen corresponds with 0'90 grm. of water. Hence, the sample of clay contained : Water . . . .' . . . . 13'31 per cent. Organic matter . ..... . 4 '3 2 per cent. These numbers can only be regarded as an approximate representation of what they are supposed to designate. According to Danne, 3 the amount of moisture can be determined in many substances by heating them with calcium carbide, and measuring the volume of acetylene obtained. The method has been extended to the determination of the " water of crystallisation " in many salts. The method has also been modified by allowing the acetylene to escape, and estimating the water from the loss in weight of the apparatus owing to the escape of acetylene. 299. Fractional Dehydration : Water lost at Different Temperatures. Some zeolites, and hydrated alumino-silicates, give off water 4 below 110, and in that case the clay can be heated at, say, 100, 110, 120, 130, ... to find if the substance loses appreciable amounts of water at progressively higher temperatures. The results are then represented as "water below 100," "water between 100 and 110," etc. When the water is to be determined at different temperatures in this manner, a copper cylinder is fitted up as shown in fig. 181 with a thermostat, A, and thermometer, B. The glass tube, say 30 cm. long and 3'5 or 4 cm. wide, has a hollow stopper ground at one end C ; the stopper is fitted with a tube and stopcock, G. The other end of the glass tube is drawn out and fitted with a drying tower, E. The copper cylinder has an outer jacket of asbestos. The glass tube is plugged in the copper tube with carded asbestos. The thermostat is regulated to keep the furnace at the desired temperature any length of time. Hence, any material placed in a small squat uncovered weighing bottle resting in the glass tube can be heated to any desired temperature while a current of air is drawn through the apparatus. The air is dried by passing it through a drying tower at the end E. This tube is packed with soda lime and calcium chloride, the former to remove carbon dioxide, the latter, moisture. A weighed calcium chloride tube, F, is placed at the end C to absorb any moisture driven from the substance in the weighing bottle. The increase in weight of this tube during an experiment represents the moisture driven from the substance. The result may 1 A. Lissner, Chem. Ztg., 34. 37, 1910. See page 521. 2 The difference between 4 '32 and 4 '26 represents analytical errors, or oxygen, nitrogen, etc. 3 H. A. Danne, Proc. Soc. Chem. Ind. Victoria, 1900; P. V. Dupre, Analyst, 30. 266, 1905 ; 31. 213, 1906 ; R. A. Cripps and J. A. Brown, ib., 34. 519, 1909 ; R. W. Roberts and A. Frazer, Journ. Soc. Chem. Ind., 29. 851, 1910 ; F. H. Campbell, ib., 32. 67, 1913 ; I. Masson, Journ. Chem. Soc., 97. 851, 1910 ; Chem. News, 103. 37, 1911 ; A. C. D. Rivett, ib., 104. 261, 1911 ; H. C. McNeil, Chem. Eng., 16. 38, 1912. 4 For the gases in silicates, see A. P. Lidoff, Zeit.anal. Chem., 46. 357, 1907 ; R.T. Chamberlain, The Gases in Rocks, Washington, 1908. 574 A TREATISE ON CHEMICAL ANALYSIS. be checked by also weighing the loss of weight of the weighing bottle. The two may not necessarily coincide, since volatile substances other than water may be driven off. The small bulbs g contain sulphuric acid. This tube shows the FIG. 181. Fractional dehydration. rate at which air is passing through the apparatus, by counting the number of bubbles per second. If the cock C be closed, and the other end connected with a suction pump, the substance can be heated under reduced pressure. In that case, the tube F is fitted to the opposite end of the glass tube. 1 1 J. W. Mellor (Trans. Eng. Cer. Soc., 7. 114, 1908 ; J. W. Mellor and A. D. Holdcroft, ib., II. 1, 1911) describes a method for heating the tube electrically at any desired temperature. If the joints of the furnace in the preceding diagram be brazed, and the thermostat, A, replaced by a reflux condenser, liquids whose boiling points are known can be employed. For instance : Table LXIV. Liquids for Vo,pour-Baths. Liquid. Temperature available. Price per Ib. Benzene . . . 80-81 7d. Water . . ' -, Toluene . . . . 99-100 109-112 9d. Ethyl butyrate .... Amyl alcohol ..... 120-121 128-132 Is. 6d. Is 7d. Xylene . . . Cumene . ." . . . Pseudocumene . . , Salol 139-141 152-lf>3 165-168 172-173 8d. Is. 5d. Is. 2d. 3s Od Aniline ...... Dimethylaniline .... Nitrobenzene ..... 180-182 191-193 207-209 lid. 3s. 5d. lid. The temperature of the inner chamber is generally 2 to 5 lower than the boiling point of the liquid in the outer chamber. THE DETERMINATION OF WATER. 575 Guttmann's x weighing bottle, fig. 182, is useful for heating substances to a constant weight in a current of gas, and hence determining the loss in weight on heating to different temperatures. The gas entry and exit tubes have ground caps for protecting the contents from air during the weighing. Meyer's 2 vapour bath or oil bath is useful when a small amount of a substance is to be heated to a given temperature below about 200. The bath is FIG. 182. Guttmann's weighing bottle. FIG. 183. Meyer's vapour bath. shown in section in fig. 183. A porcelain cylinder is placed inside a double- walled copper cylinder. The crucible containing the substance under investiga- tion is placed in the inner cylinder. A suitable liquid is placed in the copper vessel. The latter is fitted with a condenser, C. A small flame and a small amount of liquid are needed for the work. For temperatures higher than 200, oils may be used. A current of air circulates outside the inner cylinder, entering at a and leaving at b. It seems fairly certain that some clays lose more moisture when dried under reduced pressure over sulphuric acid 3 than when they are heated to constant weight in the hot oven, fig. 90, page 156. Skertchly 4 has shown that many organic substances do the same thing. 1 L. F. Guttmann, Journ. Amer. Chem. oc., 28. 1667, 1906. 2 V. Meyer, er., 18. 2999, 1885 ; A. Fock, ib., 18. 1124, 1885. 3 E. Lowenstein, Zeit. anorg. Chem., 63. 69, 1909. 4 W. P. Skertchly, Journ. Soc. Chem. Ind., 32. 70, 1913. CHAPTER XL. THE DETERMINATION OF BORON. 300. The Detection of Boric Oxide. THE following remarks on this subject supplement those given in the regular text-books on qualitative analysis. It is generally necessary to get the borate in a soluble condition. This is done by fusing the borosilicate with five to .ten times its weight of sodium carbonate, and extracting the cold mass with water. The Alcohol Flame Test. Boric acid dissolves in alcohol, and the alcohol burns with a flame tinged with green. Similarly, if a borate be decomposed by treatment with an acid, 1 and the mass be mixed with alcohol, the alcohol, when ignited, burns with a greenish flame. The test is carried out by igniting the mixture on a watch-glass. 2 It is possible to detect O'OOl per cent, of boric oxide in a solution by the flame test, and a glaze containing O'l per cent of boric oxide will give a distinct reaction. If copper be present, the test is not satisfactory. The copper must be first removed by hydrogen sulphide. Barium salts also colour the flame green. If sulphuric acid be used to decompose the borate, non-volatile barium sulphate is formed, and barium does not then interfere with the test. When copper or barium is present, the flame test is best made by placing the mixture in a six- inch test tube fitted with a cork and gas-jet as illustrated in fig. 184. The test tube is heated, and when the alcohol begins to boil, light the flame with a second burner. The copper and barium do not then interfere. If metallic chlorides be present, ethyl chloride may be formed, and this colours the flame green, thus spoiling the test. This difficulty is easily avoided by using sulphuric or nitric acid, not hydrochloric acid. The Glycerol Flame Test? The powdered and calcined borate is moistened with sulphuric acid and heated on a platinum foil until the acid is expelled. Moisten the mass with glycerol. The glycerol burns with a green flame if boric oxide be present. The glycerol test will indicate O'OOl per cent, of boric oxide. For the disturbing agents, see the alcohol flame test. Turmeric Test. If a borate be just acidified with dilute hydrochloric acid, and a strip of turmeric paper 4 be half immersed in the solution, no apparent change occurs; but if the paper be dried on a watch-glass at 100, the half which has been dipped in the boric acid shows a peculiar brownish-red coloration. If but a small trace of boric oxide be present, the stain may be pink. The 1 The borate is supposed to be decomposable by treatment with acids sulphuric, hydro- chloric, or hydrofluoric acids ; or by a mixture of ammonium nitrate and chloride ; sulphuric and hydrochloric acid ; or sulphuric and nitric acid. 2 H. Borntriiger, Zeit. anal Chem., 39. 92, 1900 ; M. Bidaut, Compt. Mend., 76. 489, 1873 ; 80. 387, 1878 ; M. Dieulafait, ib., 85. 605, 1877 ; Ann. Chim. Phys. (5), 12. 318, 1877. 3 M. W. lies, Amer. Chem., 6. 361, 1876; H. Gilm, Ber. t II. 712, 1878. 4 Turmeric paper gives a brownish colour with alkalies, yellow with acids. 576 THE DETERMINATION OF BORON. 577 colour remains when the stain is dipped in boric acid again, or in dilute sulphuric or hydrochloric acid. The brown stain produced by alkalies changes into yellow under these conditions. If too little acid be used, there may be no coloration; if too much, the colour may be brown. If the boric acid stain be touched with a solution of potassium hydroxide, the paper becomes bluish black or bluish grey, according as much or little boric acid be present. A little hydrochloric acid will restore the red colour. As little as 0*0001 grm. of boric acid can be detected by the (pink) colour produced in this manner. The presence of oxidising agents like chlorates, chromates, iodides, 1 etc., interferes with the test by destroying the turmeric. Nitric acid is an exception. Concentrated hydrochloric acid may give a dark brown stain under the conditions of the test; ferric chloride, molyb- denum and zirconium salts give a brownish-red stain which is not coloured bluish black with potash solution. Still smaller amounts of boric oxide can be detected if needed by placing the solution under investi- gation in a small dish and evaporat- ing it to dryness in a desiccator in vacuo at a low temperature. If a few drops of an alcoholic extract of a few turmeric papers be mixed in a porcelain dish with the borate, the solution acidified with' acetic acid, and evaporated to dryness on a water bath, a reddish-brown residue will be obtained if as little as 0*0002 grm. of B. 7 3 be present, while 0*00002 grm. will produce a perceptible coloration. 2 The Boron Fluoride Test. If the powdered borate be mixed with a few drops of water and approximately three times its weight of Turner's flux, 3 and the paste be exposed on the loop of a platinum wire in the outer mantle of Bunsen's flame, the boron fluoride which is formed imparts a greenish tinge to the flame. The reaction will detect 0*01 per cent, of B 2 3 in a silicate. 4 FIG. 184. Flame test for boric oxide. 1 Chlorates and chromates may be reduced by treating the solution with solid sodium sulphite ; add hydrochloric acid, and warm the solution to drive off the excess of sulphurous acid. Filter if necessary, and boil the filtrate with a slight excess of sodium carbonate, dilute, and filter. Iodides, if present, may be removed by precipitation with silver nitrate in a solution acidified with nitric acid. The addition of 5 grms. of urea per 100 c.c. of solution inhibits effects of nitrites ; nitrates do not interfere T. M. Price and E. H. Ingersoll, Bull. U.S. Dept. Agric. (Chem.), 137. 115, 1912. 2 M. Ripper, Weinbau WeinhandL, 6. 331, 1888 ; P. Kulisch, Zeit. angew. Chem., 7. 187, 1894 ; V. Lehner and J. S. C. Wells, Journ. Amer. Chem. Soc., 21. 417, 1894 ; H. Borntrager, Zeit. anal. Chem., 29. 92, 1900 ; W. H. Low, Journ. Amer. Chem. Soc., 28. 807, 1907. _ 3 TURNER'S FLUX. Mix finely powdered calcium fluoride or potassium fluoride with 4*5 times its weight of potassium bisulphate. Hydrofluosilicic acia may be used in place of this flux if the borate is easily decomposed. 4 E. J. Chapman, Chem. News, 35. 12, 26, 36, 1877 ; C. le Neve Foster, ib., 35. 127, 1877 ; H. Kammerer, Zeit. anal. Chem., 12. 376, 1873. 37 578 A TREATISE ON CHEMICAL ANALYSIS. 301. The Determination of Boric Oxide. In 1851, Rose 1 stated that the quantitative determination of boric oxide was so very difficult that up to that time no method had been devised which gave a direct result. The problem has since been solved, but the process is difficult and laborious. Rosenbladt 2 thinks this " one of the most complicated operations in analytical chemistry." Vogt 3 estimated the boric oxide in silicates by difference after the other constituents had been determined. This method is not recommended. Several indirect methods have been proposed. 4 The isolation of a definite compound of boric oxide from the associated constituents is a necessary preliminary for the successful determination of boric oxide. The removal of silica and alumina is particularly difficult. Ferric oxide and alumina, for instance, when precipitated by ammonia carry down some boric oxide. 5 The silica also is difficult to separate without loss of boric oxide, because, if the mixture be treated with hydrofluoric and sulphuric acids in the usual manner, boron is volatilised as boron fluoride. Hence, silica may con- taminate the compound of boron which is finally weighed. 6 Before dealing with complex silicates, it will be convenient to take simpler problems ; and first, the volumetric determination of boric oxide in boric acid. Volatilisation of Boric Oxide and Borax. At the outset it is necessary to emphasise the fact that the substances now under investigation are somewhat volatile. For example, there is a difficulty in drying boric oxide, since a com- paratively large amount of boric acid is lost. For instance, the loss of boric oxide on driving off the water of crystallisation from 1 grm. of boric acid containing Water . . . 07808 07543 07492 07544 0'2885 0'2780 0-2783 grm. B 2 3 lost . . 0-1405 0-1358 01449 01508 0'0896 0'1034 01032 grm. during the expulsion of the water. 7 Aqueous solutions containing boric acid, or solutions of borax acidified with sulphuric or hydrochloric acid, lose much boric acid when evaporated. In illustra- tion, some of Tschijewski's experiments might be quoted. Solutions containing B 2 8 ,. . . . 0-6030 0-2638 01268 0'0323 ' 0'0314 0'0095 grm. B 2 3 lost . . . . 0-0522 0-0298 0'0278 0'0123 0'0114 0-0071 grm. on evaporation to dryness on a water bath. 8 Borax also during calcination loses boric oxide. Thus, Waldbott 9 found that when 0'6532 grm. of borax glass was blasted for Time . . .5 12 22 29 39 44 46 minutes. Loss . . . .0-8 1-0 2-0 2-5 3 '3 41 4'2 percent. 1 H. Rose, Ausfuhrliches Handbuch der analytischen Chemie, Braunschweig 2. 721 1851 2 T. Rosenbladt, Zeit. anal Chem., 26. 18, 1887. 3 G. Vogt, Ber. Internal. Kongress angew. Chem., 5. i, 738, 1904 ; H. Rose, I.e. 4 A. Arfvedson, Schweiggers Journ. (2), 8. 1, 1802 ; F. G. Schaifgotsch, Fogg. Ann., IO7. 427, 1859; J. J. Berzelius, ib., 2. 118, 1824 ; H. Rose, ib., 80. 26 K 1850; K. Kraut, Zeit. anal. Chem., i. 73, 1863. 5 F. Wb'hler, Liebig's Ann., 141. 268, 1867. 6 E.g. in A. Stromeyer's process (Liebig's Ann., 100. 82, 1856) every one part by weight of silica contaminating the final precipitate quadruples the calculated amount of boric oxide. 7 P. Tschijewski, Arch. Phys. Nat. (3), 12. 120, 1884 ; Bull. Soc. Chim. (2). 42. 324. 1884 ; H. Lescoeur, Ann. Chim. Phys. (6), 19. 43, 1890 ; L. Kahlenberg and 0. Schreiner, Zeit. phys. Chem., 20. 547, 1896 ; F. W. Skirrow, ib., 37. 84, 1901 ; R. Bunsen, Liebiq's Ann., ill. 257, 1859 ; G. Watson, Chem. News, 68. 199, 1893 ; 0. Hehner, Analyst, 16. 141, 1891. L. de Koningh, Journ. Amer. Chem. Soc., 19. 385, 1897 ; F. G. Schaffgotsch, Pogg. Ann , 107. 427, 1859; J. A. Rose, Beitrdge zur Kenntnis der Borsdure und uber eine direkt gewichtsanalytische Bestimmung derselben, Bonn a. Rh. (Erlangen), 1902. 9 S. Waldbott, Journ. Amer. Chem. Soc., 16. 410, 1894 ; E. Cramer Tonind Ztq 16. 155, 1892. See page 589. THE DETERMINATION OF BORON. 579 Starting with 2'1320 grms., 10 minutes' calcination over an ordinary Bunsen's burner gave no loss; 15 minutes over a moderate blast, with the crucible half covered, gave a loss of O'OOH grm. ; and over a strong blast, with the crucible open, O0073 grm. was lost. In a pottery oven fired to Seger's cone 9 (scheduled 1280) for Time 60 48 26 3 hours Loss 48-98 22-40 7 '88 1'47 percent. Hence boric oxide may also be lost when the "loss on ignition" is determined. 302. The Evaluation of Boric Acid. An aqueous solution of boric acid can be titrated with a standard solution of sodium hydroxide. Unfortunately, owing to the feeble acidic properties of this acid, the pink colour of the indicator, phenolphthalein, is developed before all the boric acid is neutralised. If, however, the solution contains sufficient glycerol, or mannitol, 1 the boric acid all reacts with the sodium hydroxide to form sodium metaborate : NaOH + H 3 B0 3 = 2H 2 + NaB0 2 before the pink of phenolphthalein appears. The polyhydric alcohol augments, 2 so to speak, the acidic qualities of the acid in question. Thomson 3 has a process for the volumetric determination of boric acid based on these phenomena. Dissolve, say, 7*5 grms. of the boric acid in about 350 c.c of water in a 500-c.c. flask, and make the solution up to the 500-c.c. mark with water. Pipette 50 c.c. into an Erlenmeyer's flask; add 50 c.c. of glycerol, 4 or a gram of mannitol, and titrate with approximately JN-sodium hydroxide solution free from carbonates, 5 using phenolphthal'ein as indicator. 6 When the 1 D. Klein, Bull. Soc. Chim. (2), I. 195, 1878 ; Compt. Rend., 46. 826, 1878 ; 99. 144, 1884; M. Copaux, ib., 127. 756, 1898; A. Lambert, ib., H>8. 1016, 1889; A. Senier and J. G. Lowe, Pharm. Journ. (3), 8. 819, 1878; W. R. Dunstan, ib. (3), 13. 257, 1882; R. Sulzer, Deut.-Amer. Apoth. Ztg., 596, 1886 ; K. Jehn, Arch. Pharm. (3), 25. 250, 1887 ; (3), 26. 495, 1888 ; E. Donath, Chem. Ztg., 17. 1826, 1893 ; E. H. Farrington, Rev. Internal. Falsif., 10. 81, 1897; G. Deniges, Journ. Pharm. Chim. (6), 6. 49, 1897; V. Gasselin, Ann. Chim. Phys. (7), 3. 1, 1884 ; C. Schwarz, Pharm. Ztg., 32. 562, 1894 ; N. Tananaeff and D. Tsukerman, Journ. Russ. Phys. Chem. Soc., 41. 1469, 1910. 2 K. Farnsteiner, Zeit. Untersuch. Nahr. Genuss., 5. 1, 1902 ; A. Beythein and H. Hempel, ib., 2. 842, 1899 ; B. Fischer, ib., 3. 17, 1900 ; G. Magnanini, Rendi Accad. Lincei, 6. 260, 1890; Gazz. Chim. JtaL, 2O. 441, 448, 1890; 21. ii. 134, 1891; Zeit. phys. Chem., 6. 58, 1890; L. Kahlenberg and 0. Schreiner, ib., 20. 547, 1896; W. Vaubel and E. Bertlet, Chem. Ztg., 29. 629, 1905 ; A. Wogrinz and J. Kittel, Chem. Ztg., 36. 433, 1912. 3 R. T. Thomson, Journ. Soc. Chem. Ind., 12. 432, 1896 ; Analyst, 21. 64, 1896 ; A. H. Allen and A. R. Tankard, ib., 29. 301, 1905 ; L. Barthe, Journ. Pharm. Chim. (5), 29. 163, 1894 ; M. Vadam, ib. (6), 8. 109, 1898 ; M. Honig and G. Spitz, Zeit. angew. Chem., 9. 549, 1896 ; G. Jorgensen, ib., 10. 5, 1897 ; N. Tananaeff and D. Tsukerman, Journ. Russ. Phys. Chem. Soc., 41. 1469, 1909. 4 GLYCEKOL SOLUTION. Glycerol, 10, water 1, by volume. The glycerol must be free from fatty acids, or the fatty acids must be neutralised just before use with y^N-sodium hydroxide solution. Glycerol usually becomes acid on keeping, possibly owing to the slow decomposition of fatty impurities. MANNITOL (M. Vadam, Chem. News, 78. 271, 1898; L. C. Jones, Amer. J. Science (4), 7. 127, 1899 ; Chem. News, 80. 65, 1899) is more convenient than glycerol. It is solid, easily handled, does not alter the bulk of the solution, and is not so liable to contamination with acids. The end point is also a little sharper. The results are quite as accurate as with glycerol. The choice is therefore a matter of convenience. 5 Carbonates interfere with the indicator and react: 2B 2 03 + Na 2 C03 = Na 2 B 4 7 + C0 2 . Hence, a little barium hydroxide is sometimes added during the preparation of the sodium hydroxide solution. See discussion, page 65. F. P. Venable and J. S. Oallison (Chem. Ztg., 14. 167, 1890) report as much as 0'06 per cent, of boric oxide in caustic soda and caustic potash sold as "pure." M. Gorges, Journ. Pharm. Chim. (6), 3. 346, 1896. 15 Other indicators have been recommended : litmus (J. L. Gay Lussac, Ann. Chim. Phys. (2), 40. 398, 1830), lacmoid (R. T. Thomson, Chem. News, 47. 123, 135, 1883), congo red 580 A TREATISE ON CHEMICAL ANALYSIS. pink colour appears, add 10 c.c. of glycerol, or another gram of mannitol. If the pink colour disappears, continue the titration with the sodium hydroxide. Repeat the addition of glycerol or mannitol and titration until the pink colour persists when more mannitol or glycerol is added. EXAMPLE. 23'7 c.c. of the ^N-solution of sodium hydroxide were required for a titration. This solution contains 20*004 grms. of NaOH per litre, which, according to the equation: NaOH + H 3 B0 3 = 2H 2 + NaB02, corresponds with 31'01 grms. of H 3 B0 3 per litre, or 1 c.c. of the standard solution is equivalent to 0'031 grm. of H 3 B0 3 . Hence, 237 x 0'031 = 0'735 grm. of H 3 B0 3 per 50 c.c. ; or 14'69 per 1000 c.c. ; or per 15 grms. of the sample. Hence, the sample has 97 '9 per cent. H 3 B0 3 . Will and Zschimmer l titrate the boric acid with baryta water ; Smith 2 adds standard manganese sulphate and titrates with standard potassium permanganate ; and Jones 3 liberated iodine from a mixture of potassium iodide and iodate in presence of boric acid and mannitol, and determined the iodine by titration with sodium thiosulphate. 303. The Evaluation of Borax. Solutions containing boric acid are neutral to jo-nitrophenol, and to methyl orange, but acid towards phenolphthalein particularly in the presence of mannitol or glycerol. Hence, if hydrochloric acid be added to an aqueous solution of borax, using jo-nitrophenol or methyl orange as indicator, 4 the solution will react acid only when all the boric acid is free and all the soda is neutralised by the hydrochloric acid. The reaction is represented : Na 2 B 4 7 + 5H 2 + 2HC1 = 2NaCl + 4H 3 B0 3 . The boric acid can then be titrated as just indicated under boric acid. The determination is made in the following manner : Dissolve 10 grms. of borax in about 300 c.c. of water, and make the solution up to 500 c.c. with distilled water freed from carbonates by boiling. Titrate 25 c.c. with approximately JN-hydrochloric acid, using a drop of ^-nitrophenol or methyl orange as indicator. Note the volume of hydrochloric acid needed for the neutralisation of the 25 c.c. pipetted from the main solution. Add the same volume of the standard hydrochloric acid to another 25 c.c. of the borax solution, add glycerol or mannitol as indicated under boric acid, and titrate, as there described, with, say, JN-sodium hydroxide, using phenolphthalein as indicator. EXAMPLE. 10 grms. of the sample made up to 500 c.c. Of this, 100 c.c., that is, 2 grms. of the sample, required 42 c.c. of the standard hydrochloric acid, and 45'6 c.c. of the standard sodium hydroxide solution, 1 c.c. of which corresponded with 0'0437 grm. of crystalline borax. Hence, 2 grms. of the sample had 45'6 x 0'0437= 1'992 ; or 99'6 per cent, of crystalline borax. A similar procedure is employed for anhydrous borax. (C. Schwarz, Pharm. Ztg., 37. 562, 1887), heliaiithine (F. Parmentier, Compt. Rend., 113. 41, 1891), orceine(M. de Luynes, Ann. Chim. Phys. (4), 6. 184, 1865), hsematoxyline (A. Guyard, Bull. Soc. Chim. (2), 40. 422, 1883), Porrier's blue (R. Engel, ib. (2), 45. 327, 1886 ; Compt. Rend., 102. 214, 262, 1886), tincture of red rose leaves (L. Barthe, Journ. Pharm. Chim. (5), 29. 163, 1894), tincture of mimosa blossoms (L. Robin, Ber. Internat. Cong. App. Chem., 8. i, 429, 1 H. Will, Arch. Pharm., 225. 1101, 1887; R. Hefemann, Pharm. Centr. (2), o. 116 1888 ; E. Zschimmer, Chem. Ztg., 2$. 442, 67, 1901. 2 E. F. Smith, Amer. Chem. Journ., 4. 279, 1883 ; Chem. News, 46. 286, 1882 ; J. Volhard, LieUg's Ann., 98. 318, 1879; C. Bodewig, Zeit. anal. Chem.. 23. 143, 1884; G. Carinelli, Gazz. Chim. ItaL, 31. i, 544, 1901. 3 L. C. Jones, Zeit. anorg. Chem., 20. 212, 1899 ; 21. 169, 1899 ; A. Stock, Compt. Rend., 130. 516, 1900 ; L. Barthe, Journ. Pharm. Chim. (5), 29. 163, 1894 ; J. Prescher, Arch. Pharm., 242. 194, 1904 ; P. Georgevic, Journ. prakt. Chem. (2), 38. 118, 1888. 4 R. T. Thomson (Chem. News, 47. 135, 1883), after comparing litmus, methyl orange, phenolacetolin, and phenolphthalein, says: "Methyl orange is by far the best indicator for the estimation of soda in borax, and is, indeed, perfect in that respect. The change in colour at the end of the experiment is very sharply denned." THE DETERMINATION OF BORON. 581 The relation between the Na 2 and the B 2 0, in borax is not always that corresponding with Na 2 0. 2B 2 3 , because some sodium metaborate Na 2 0. B.O - may be present. This would lead to high values when the borax is calculated from the amount of B 2 3 , since some should then have been calculated to NaB0 2 .4H 2 0. Jacobi's Process. Jacobi l abbreviates the process by taking advantage of the fact that in the presence of glycerol a solution of borax is acid to phenol- phthalein, while a solution of sodium metaborate is neutral. Hence, if a solution of borax containing an excess of glycerol be titrated with JN-sodium hydroxide, we have Na 2 B 4 7 + 2NaOH = 4NaB0 2 + H 2 at the neutral point. After making an allowance for any acid in the glycerol, multiply the results by 0'0175, and the product represents the amount of boric oxide in the given sample ; if multiplied by O02525, the result represents the amount of anhydrous borax; and if multiplied by 0'04775, the amount of crystalline borax. EXAMPLE. 2 grms. of " calcined " borax were dissolved in water and mixed with glycerol and phenolphthalein. 49'05 c.c. of ^N-NaOH were needed. The glycerol added required 0'4 c.c. of the alkali for neutralisation. Hence, 49'65 - 4 = 49 V 25 c.c. were needed for the titration. Hence, the 4 grms. contained 49'25 x 0-0175 = 0'862 grm. of B 2 :{ ; that is, 100 x J x 0'862 = 43'1 per cent, boric oxide or 62'2 per cent, of " calcined " borax. 304. The Evaluation of Borocalcite, Boronatrocalcite, Boracite, and Calcium Borate. Digest, say, 10 grms. of powdered borocalcite in a flask fitted with a reflux condenser 2 (fig. 185) with, say, 50 c.c. of ^N-hydrochloric acid in order to remove the carbon dioxide, liberate the boric acid from combination, and prevent loss of boric acid by volatilisation in steam during the expulsion of the carbon dioxide. Cool. Wash the condenser tube, and the contents of the flask into a 500-c.c. standard flask with freshly boiled distilled water. The insoluble, residue can be filtered off, dried, and weighed, if desired. Make the solution up to the 500-c.c. mark with water. Pipette 50 c.c. of this solution into a flask, and titrate with standard sodium hydroxide (say, |N), using jt?-nitrophenol as indicator. When the yellow colour of the indicator appears, 3 all the hydrochloric acid has been neutralised by the soda. Note the volume of the standard sodium hydroxide required for this purpose. Pipette another 50 c.c. of the solution, add the same amount of sodium hydroxide, and then titrate with sodium hydroxide and mannitol or glycerol, with phenolphthalein as indicator, as indicated under borax and boric acid. EXAMPLE. 10 grms. of the borocalcite were made up to 500 c.c., and 50 c.c. {i.e. 1 grm. of the sample) were taken for the titration. 26-6 c.c. of the soda solution were employed in the second titration. 1 c.c. of |N-sodium hydroxide corresponds with 0*0175 grm. of boric oxide, B 2 O 3 . Hence, 26-6xO'0175 = 46*55 grm., or 46'55 per cent. B 2 3 in the given sample. Schaak's Process. If appreciable quantities of iron and alumina be present, both should be separated, if possible, before the titration, since alumina will combine with the alkalies as the solution approaches the neutral point, and thus 1 K. Jacobi, Journ. Amer. Chem. Soc., 26. 91, 1904. 2 V. Kreussler, Zeit. anal. Chem., 24. 80, 1885 ; J. Walter, Dingler's Journ., 251. 637, 1884. 3 Methyl orange may be used instead of tbe nitrophenol with satisfactory results. 582 A TREATISE ON CHEMICAL ANALYSIS. give too high a number. Schaak 1 recommends the following method of separation : Dissolve the borocalcite as indicated above. Filter, and make the solution up to 500 c.c. Nearly neutralise 100 c.c. by titration with alkali, using methyl orange as indicator. Add 2 to 3 grms. of barium carbonate (free from alkali), warm on a steam bath for half an hour. Cool, filter, wash, and titrate the filtrate with alkali as indicated above. The first method may give values 1'5 per cent, too high, while Schaak's modification will furnish values 0'15 per cent, too high. Jacobi's Process^ The mineral borates may be rapidly evaluated by digesting, say, 2 grms. in hydrochloric acid, and evaporating the solution to dryness for silica. Add dilute hydro- chloric acid, boil, filter, and wash the silica ; precipitate the iron and alumina by means of ammonia ; lime by am- monium oxalate ; and magnesia by ammonium phosphate. Remove the excess of phosphate by means of pure ferric chloride ; and remove the latter by ammonia. Evaporate the solution to dryness with concentrated nitric acid, and then twice with concentrated hydro- chloric acid. Take up the residue with water, and evaporate to dryness again in a platinum dish. Heat the residue to dull redness to drive off the water and the ammonium salts. Cool, and weigh as anhydrous boric oxide, sodium (with possibly potassium) chloride, and sodium borate. Dissolve the residue in hot water, cool, and determine the soda by titration with JN-sulphuric acid, using methyl orange as indicator. Titrate also for total boric oxide with JN- sodium hydroxide, using phenolphthalein as indicator. Sum the boric oxide and the soda so obtained, and subtract the result from the weight of the residue in the platinum dish. The difference is said to be " sodium chloride," from which the Na 2 is calculated by multiplying the amount of sodium chloride by 0'5308. If potassium 3 salts and sulphates be present, their amounts are determined in separate portions. Processes like these can, of course, only give approximate results by a balancing of errors. Lunge's Process. In analysing boronatrocalcite, Lunge 4 dissolves the mineral in dilute hydrochloric or nitric acid, and precipitates the lime as calcium oxalate ; the magnesia as phosphate. The three bases lime, soda, and magnesia are 1 M. P. Schaak, Journ. Soc. Chem. Ind. , 23. 699, 1904. 2 K. Jacobi, Journ. Amer. Chem. Soc., 26. 88, 1904. 3 Or remove the boric oxide by repeated evaporation with methyl alcohol (page 589), and determine the potash and soda in the usual manner. * G. Lunge, Liebig's Ann., 138. 51, 1866; Chem. News, 15. 86, 214, 1867; K. Kraut, Liebig's Ann., 139. 52, 1866. According to Lunge, there is no need to remove the boron as fluoride, since the results with and without the removal of boron ' ' entirely coincide within the ordinary limits " of error. FIG. 185. Reflux condenser (see page 303). THE DETERMINATION OF BORON. 583 determined by adding a known volume of standard (hydrochloric or nitric) acid, and titrating the excess of acid with standard alkali, as indicated for borax. Deduct the results found for lime and magnesia, and the difference represents the soda. 1 Water is determined as the " loss on ignition." The difference represents the boric oxide and impurities. In one case, 'Kraut found 672 per cent, of sodium chloride and 4*74 per cent, of sodium sulphate 2 in a sample of boronatrocalcite. 305- The Determination of Boric Oxide in Silicates- Wherry's Process. Methods for the determination of boric oxide in silicates containing alumina and iron, by fusing the sample with alkali carbonate ; removing the silica, alumina, etc., with ammonio-zinc oxide, Schaffgotsch's or Seeman's solutions (page 638) ; and finally precipitating the boric oxide as an insoluble borate magnesium, 3 barium, 4 calcium, 5 or silver 6 borates, or potassium borofluoride, 7 are not satisfactory, because part of the boric oxide is retained very tenaciously by the precipitated alumina. Wherry 8 recommends the following modifica- tion of Schaak's process, and it can be regarded as a good method of approximation when the time needed for the more elaborate and tedious distillation process is not avail- able : Fusion. Fuse 0'3 to 0'5 grm. of the powdered silicate with ten times its weight of sodium carbonate, and take up the cold cake with 20 to 30 c.c. of dilute hydrochloric acid. Oxidise any ferrous salts which may be present by the addition of a few drops of nitric acid or bromine water. Removal of Iron and Aluminium. Heat the mixture in a round- bottomed 250-c.c. flask nearly to boiling, and add an excess of dry precipitated FIG. 186. Boiling with "splash trap." 1 Potash is said to be rarely present. 2 H. How, Chem. News, 15. 192, 1867 ; F. Wohler, ib., 16. 15, 1867. 3 C. Marignac, Zeit. anal. Chem., I. 405, 1862. 4 P. von Berg, Zeit. anal. Chem., 22. 25, 1897 ; H. N. Morse and W. M. Burton, Amer. Chem. Journ., 10. 154, 1888. 5 A. Ditte, Ann. Phys. Chim. (4), I. 549, 1875 ; Compt. Rend., 80. 490, 1875 ; A. Jo\y,ib., 100. 103, 1885. 6 A. P. J. du Menil, Berzelius 1 Jahresber., 9. 180, 1829. 7 A. Stromeyer, Liebitfs Ann., 100. 82, 1856 ; G. Kriiss and H. Moraht, #., 20O. 180, 1896 ; F Wohler ib. 141. 268, 1867 ; C. Bodewig, Chem. Neivs, 50. 49, 1884 ; Zeit. anal. Chem., 23. 149, 1884 ; C. Thaddeeff, ib., 36. 568, 1 897 : H. Rose, Pogg. Ann. , 80. 262, 1880 ; C. Hammelsberg, ib., 80. 466, 1880 ; J. J. Berzelius, ib., 2. 118, 1824. 8 E.. T. Wherry and W. H. Chapin, Journ. Amer. Chem. Soc., 36. 1687, 1908. 584 A TREATISE ON CHEMICAL ANALYSIS. calcium carbonate slowly, with constant stirring in order to precipitate the iron and aluminium hydroxides. 1 Fit a reflux condenser (figs. 128 or 185) to the flask, and boil vigorously for about 10 minutes in order to destroy bicarbonates and to expel carbon dioxide from the solution. Filter the bulky gelatinous pre- cipitate through a Biichner's funnel (page 103) at the filter pump. Wash with hot water, but keep the volume of the liquid as low as possible less than 100 c.c. Pour the hot filtered liquid into a 250-300 c.c. flask. Add a couple of grams of calcium carbonate. Fit a splash trap in the neck of the flask by means of a rubber stopper (fig. 186), and connect the flask with a suction pump. Heat the solution under reduced pressure for a few minutes. 2 When the liquid is cold, and boiling has ceased, filter the solution if the precipitate is red (ferric oxide) ; if the precipitate is white (alumina), there is no need for the filtration. The Titration. The calcium carbonate neutralises the free hydrochloric and nitric acids, but not the boric acid. Hence, the solution contains free boric acid. Titrate the solution with y^N-sodium hydroxide. Use mannitol and phenol- phthalein as indicator. If the first drop of the indicator does not strike a red colour, all the carbon dioxide has not been expelled by the boiling. In that case, acidify the solution, add an excess of calcium carbonate, and repeat the necessary operations. Faults of the Method. As indicated above, the weakness of this method is due to the assumption that aluminium and ferric hydroxides, precipitated by the calcium carbonate, are free from boric oxide. Some boric oxide is mechanically enclosed in the precipitate, and this can be removed by simply boiling with water ; but a certain proportion of the boric oxide is so intimately associated with the hydroxides that it cannot be removed by boiling water. The boric oxide can only be recovered by redissolving the .precipitate in acid, and repeating the treatment with calcium carbonate a number of times. For example, Wherry and Chapiri quote these numbers : Table LXV. Test Analyses for Boric Oxide in Minerals. Alumina Amount Per cent, boric oxide. or ferric oxide. analysis. First. Second. Third. Fourth. Fifth. Total. Per cent. Grm. Dumortiertite 63 i 0-5 3-87 1-12 0-35 O'll 5-45 Ludwigite . 40 0-3 9-93 071 071 0-52 0-22 12-09 Danburite . nil. 0-3 2377 0-29 24-06 Tourmaline . 35 0-3 8-88 0-82 0-23 9-93 Vesuvianite . 20 0-3 T75 017 ... ... 1-92 The results are in general better with small proportions of boric oxide, as we should expect. The process can therefore only be recommended when an approximate determination is needed. If results are desired with any pretence to accuracy, the distillation process must be employed. The necessary re- petition of the processes of solution and precipitation, as indicated above, makes the distillation process less troublesome and more reliable than the calcium carbonate precipitation process. J It is impossible to prevent the precipitation of some boric oxide with the iron and aluminium hydroxides. 2 If the exhaustion is conducted too rapidly, violent frothing and loss of material may render it necessary to start the analysis again. THE DETERMINATION OF BORON. 585 306. The Determination of Boric Oxide in Silicates- Distillation Process. If methyl alcohol be distilled from a mixture containing free boric acid, methyl borate, boiling at 65, distils over. The methyl borate can be saponified with alkalies, etc., and the boric acid determined by gravimetric or volumetric processes. This reaction was simultaneously employed by Gooch and by Rosenbladt in 1887, and it has since been extensively used 1 for the determina- tion of boric oxide in silicates and alumino-silicates. Ethyl alcohol can be employed, 2 but it forms ethyl borate, boiling at about 120. In consequence of the higher boiling point of ethyl borate, methyl alcohol. 3 though more expensive, is more effective than ethyl alcohol. Preparation of the Sample for Distillation. Fuse 0'5 grm. 4 of the sample with 6 grms. of sodium carbonate (page 164). When the crucible is cold, the cake 5 is transferred to a 250-c.c. flask by using 50 c.c. of water and 5 c.c. of concentrated hydrochloric acid for the purpose. This amount of acid is sufficient to neutralise the sodium carbonate used in the fusion. The flask is then fitted with a reflux condenser (fig. 185), and boiled for an hour in order to drive off the carbon dioxide. The reflux condenser is to prevent loss due to volatilisation of boric acid with the issuing steam. Aqueous solutions of boric acid must never be boiled in open vessels when the boric acid is to be determined quantitatively.** Wash the condenser tube with a little water and collect the runnings in the flask. Now add solid granular anhydrous calcium chloride to the flask. About 1 grm. of the salt per c.c. of liquid will suffice. Rotate the flask for a few minutes in order to ensure a thorough mixing of the contents. In adding the calcium chloride, keep the neck of the flask clean. The object of the calcium chloride is to prevent 1 K. Farnsteiner, Zeit. Nahr. Genuss.. 3. 1, 1900; K. Windisch, ib., 9. 641, 1905; E. Fischer, ib., 3. 17, 1900; J. Wolff, ib., 3. 600, 1900; 4. 157, 1901; J. J. Ebelmen and Bouquet, Ann. Chim. Phys. (3), 17. 54, 1846; J. J. Ebelmen, ib. (3), 16. 129, 1846; H. Sehiff, Liebicfs Ann., Suppl., 5. 154, 1867 ; H. Schitf and E. Bechi, Compt. Rend., 61. 697, 1865 ; H. Copaux, ib., 127. 756, 1888; H. Moissan, ib., 156. 1084, 1893 ; S. L. Penfeld and E. S. Sperry, Amer. J. Science (3), 34. 222, 1887; J. E. Whitfield, ib. (3), 34. 281, 1887; C. E. Cassal, Analyst, 15. 230, 1890; A. K. Reischle, Zeit. anorg. Chem., 4. Ill, 1893; A. Mandelbaum, ib. , 62. 364, 1909; K. Kraut, Zeit. anal. Chem., 36. 165,1897; A. W. Blyth, Proc. Chem. floe., 15. 51, 1899; C. Montemartini, Gaaz. Chim. Ital., 28. i. 344, 1898; G. H. Beermann, Kritische Studien iiber die neueren quantitatiren Bestimmungsmethoden der Borsdure mit Einschluss der TurmaUnanalyse, Berlin, 1898; J. Prescher, Arch. Pharm., 242. 194, 1904 ; A. Schneider and Gaab, Pharm. Centr., 37. 372, 1896 ; W. H. Low, Chem. News, 95. 52, 61, 73, 1907 ; Journ. Amer. Chem. Soc., 28. 807, 1907 ; E. T. Wherry and W. H. (Jhapin, ib., 30. 1687, 1908 ; G. W. Sargent, ib., 21. 858, 1899 ; L. de Koningh, ib., 19. 55, 1897 ; 0. von Spindler, Schweizer Wochenschrift Chem. Pharm., 43. 659, 1906; Chem. Ztg ., 29. 566, 1905 ; R. Arndt, ib., 33. 725, 1909 ; H. Copaux and G. Boiteau, Bull. Soc. Chim. (4), 5. 217, 1909 (the distillation process will not do with boric oxide in the presence of tungstates) ; G. Bertrandand H. Agulhon, ib. (4), 7. 125, 1910 ; H. Fromme, Tschermatfs Mitt., 28. 329, 1910 ; E. Polenske, Arbeit. Kaiser. Gesund., 17. 564, 1900; 19. 167, 1909 ; A. Gunther, ib., 19. 489, 1904 ; G. Raulin, Monit. Scient. (5), I. 434, 1911. 2 H. Rose, Pogg. Ann., 80. 262, 1846. 3 Avoid the use of methyl alcohol, which blackens or evolves sulphur dioxide when heated with sulphuric acid on a water bath. . 4 For substances containing a high proportion of boric oxide, 0'25 grm. may be used. The total boric oxide should be kept below 01 grm., otherwise the distillation will require a long time, the titration will require a large volume of solution, and the end point will be less distinct. If metals likely to injure the platinum crucible are present, the substance may be fused in a porcelain crucible (felspathic glaze), and the crucible and cake powdered and treated as described in the text. 5 The first residue from the Smith's process for alkalies may be dried and then used for this determination. 6 Note the violation of this rule in Jacobi's process for mineral borates (page 582). 586 A TREATISE ON CHEMICAL ANALYSIS. the hydrolysis of the methyl borate in the flask, and thus facilitate its removal. Sulphuric or phosphoric acid can also be employed as dehydrating agent. 1 The Distillation Apparatus. An apparatus similar to that indicated in FIG. 187. Distillation apparatus. fig. 188 is to be fitted up. A is a 250-c.c. distilling flask 2 containing the boric acid, etc., prepared as indicated above. B is a 500- to 800-c.c. distilling flask containing methyl alcohol free from acetone. Owing to violent bumping which occurs when methyl alcohol is boiled in glass vessels, a boiling tube 3 is necessary 1 If strong acids be used for the distillation, and if chlorides be present, chlorine may be evolved. In that case, it is well to remove the chlorine by precipitation with silver nitrate and proceed with the distillation at once. The silver nitrate, if desired, can be removed by precipita- tion with sodium hydroxide or carbonate. Similar remarks apply to the presence of nitrates, oxalates, citrates, tartrates (but not acetates) when the distillation is made from concentrated sulphuric acid in place of the calcium chloride indicated in the text. For Rosenbladt's sugges- tion to use silver sulphate in the flask, see page 547. 2 Bohemian glass should be used. Some varieties of glass contain boric oxide, and are consequently a source of danger. A Lewkowitsch's flask (J. Lewkowitsch, Journ. Chem. Soc., 55. 360, 1889 ; G. George, Chem. News, 78. 259, 1898) with a wide delivery tube may be used. 3 BOILING TUBES. This important auxiliary is a simple and effective means of preventing the bumping of most liquids during boiling, and it has the advantage of introducing no other foreign substance than glass and air. It would be almost impossible to conduct experiments with boiling methyl alcohol if some device of this character were not employed. The boiling tube here described appears to have been first suggested by D. Gernez (Compt. Rend., 86. 472, 1878 ; H. Scudder, Journ. Amer. Chem. Soc., 25. 163, 1903 ; Chem. News, 88. 242, 1903). THE DETERMINATION OF BORON. 587 each time the flask is used. C is a small Fresenius' gas washer l containing a little mercury. This acts as a safety valve should the tube leading from B to A get clogged. G can be fitted with a long exit tube to carry the vapours of alcohol away from the proximity of the flame should alcohol blow out. D is a Lendrich's condenser. 2 E is a 250-c.c. Erlenmeyer's flask to act as a receiver. It is fitted with a tube F containing enough water to act as a water seal and prevent the loss of methyl borate during the distillation. From four to six condensing flasks, E, are used, and all fit on the same stopper, G. The connections are made as indicated in the diagram. The Distillation. The flask B is heated, and methyl alcohol vapour distils into the flask A. When about 25 c.c. of alcohol have condensed in A, the latter flask is heated 3 hot enough to prevent any further condensation of alcohol vapour in A. The distillation must not proceed too rapidly, or methyl borate may be lost through the water trap. When about 100 c.c. of alcohol have condensed, change the receiver. Transfer the contents of the trap F to the flask just removed. Add a drop of >-nitrophenol to the same flask, and run in the standard soda solution until the free acid is neutralised. To titrate for boric acid, add 1 c.c. of phenol phthalein, and titrate until the pink coloration has developed. The end reaction is not very sharp, owing to the action of the alcohol on the indicator. The difference between the burette readings with the two indicators gives an approximation to the amount of boric acid in the distillate. Add to the distillate twice as much standard alkali as was required in the phenolphthalein titration for boric acid. If 5 c.c. of alkali were required for the boric acid titration, add 10 c.c. of the same alkali. The object of adding the excess of alkali is to prevent loss of boric acid 4 when the alcohol is distilled off later on. When another 100 c.c. has collected in the second Erlenmeyer flask, change the receiver, and treat the distillate as before. If less It is made as follows : Draw out a piece of glass tubing until its internal bore is from 0*5 to I'O mm. diameter. The tube is sealed (c, fig. 188) about 1 cm. from one end, a, by holding FIG. 188. Boiling tube. the tube in the flame until the sides have run together. Cut the tube the desired length- say, 16 cm. and seal the end b to prevent the entrance of liquid. The upper end of the boiling tube should reach very nearly to the top of the flask. The tube is placed open end down in the liquid. The open end a should rest on the bottom near the hottest part of the flask, and remain there during the boiling. Hence, the tube should be heavy enough to prevent its being thrown off the bottom. For liquids of high specific gravity, therefore, it may be necessary to use a piece of thick-walled capillary tubing with a wide bore (may be up to 5 mm. ) ; and for frothing liquids, and liquids of low boiling point, a narrow boiling tube is best. The action of the boiling tube is as follows : The air in ac expands and passes through the liquid in small bubbles. The air is gradually replaced by the vapour of the liquid being boiled, and the stream of bubbles is continuous so long as the temperature about the capillary is at the boiling point of the liquid. This prevents superheating and explosive boiling. The seal c should be below the surface of the liquid, even if it be necessary to bend the portion ab for the purpose and make ac nearly parallel with the bottom of the flask. This prevents the condensa- tion of vapour in ac, which would stop the stream of bubbles. The flame should be protected from draughts so as to prevent the temperature of the liquid falling below its boiling point and thus causing the capillary to fill with liquid. The boiling tube is useless if it be filled with liquid. Hence it should be cold and empty when placed in the flask. The liquid is shaken out of the boiling tube each time it is used. 1 R. Fresenius, Zeit. anal. Chem., 15. 62, 1876. 2 K. Lendrich, D.R.Q.M. 198543, 1903 ; F. Allihn, Zeit. anal. Chem., 25. 36, 1886. 3 In a paraffin bath H\ the thermometer J shows the temperature of the paraffin bath. 4 According to E. Polenske (Arbeit. Kaiser. Gesund., 17. 564, 1900), a mixture of soda and boric acid corresponding with 5Na 2 0. 4B 2 3 is stable in boiling methyl alcohol. 588 A TREATISE ON CHEMICAL ANALYSIS. than 1 c.c. y^N-sodium hydroxide be required, or 0'2 c.c. of JN-sodium hydroxide be required for the boric acid titration, the distillation is completed. If more than this amount of alkali be required, continue the distillation, a third, and may be a fourth time. If sufficient calcium chloride has been used, the second distillation is usually sufficient. 1 The following readings were made during a determination of the boric oxide in 0'5 grm. of glaze : Indicator(iN-NaOH). First distillate. Second distillate. Third distillate. Total. ;?-Nitrophenol . Phenol phthalein '27 4-1 0-3 0-4 0'2 O'l 3-2 4'6 Hence, in all, 4*6 c.c. of the ^N-NaOH solution were used for the H 3 B0 3 in the phenolphthalein titration. Add 9*2 c.c. of the JN-NaOH in order to fix the boric acid during the next stage of the operation. Removal of Methyl Alcohol from the Distillate. The distillates are now poured into a flask. Use as little water as possible for the rinsing. A boiling tube is placed in the flask, and the flask is heated. The alcohol which distils off is collected in a receiver placed below the condenser. 2 When the alcohol has distilled off, change the receiver. The residue should occupy about 25 c.c. If less, make up to approximately this volume with distilled water. Add concentrated hydrochloric acid (1:1) drop by drop until the colour of the indicator is just discharged. Put a boiling tube into the flask, and heat the solution on the steam bath for about two minutes. The flask is then connected with the filter pump and the liquid allowed to cool while the pump is in action. This decomposes any carbonates, and there is no danger of losing boric oxide. The liquid should be colourless hot or cold. If otherwise, sufficient acid is not present. Final Titration of the Boric Oxide. Neutralise any excess of acid with ^N- sodium hydroxide until the yellow tint of the nitrophenol appears. Make the solution acid with y^N-hydrochloric acid, and again neutralise with the N-alkali. One drop of the T ^N-acid should discharge the colour of the indicator. Now add 1 grm. mannitol (or 40 c.c. of glycerol) and titrate with T VN- or |N-alkali according to the amount of boric oxide present (the methyl alcohol titration shows this), as indicated under boric acid. With the sample previously discussed, the following readings were made : 23 '6 c.c. of y^N-NaOH were required. 1 c.c. of the ^N-NaOH corresponded with 0-003322 grm. B 2 3 . 3 Hence, 23'6 x 0'003322 = 0-07835 grm. B 2 3 was present per 0'5 grm. of the glaze; or the glaze contained 15*67 per cent, of boric oxide B 2 3 . Test Experiments. Determinations with artificial mixtures of known amounts of boric oxide with sodium silicate ; alum ; potassium fluoride ; ferric and ferrous salts ; arsenious, zinc, stannous, and antimonious salts ; salts of the alkalies and alkaline earths gave quite satisfactory results. 0'0350 grm. of boric oxide was 1 Instead of passing the vapour of methyl alcohol through the boric acid solution, flasks B and may be dispensed with, and a stoppered separating funnel fixed in A as in the original Gooch's apparatus. Then add liquid methyl alcohol in separate portions 10 c.c. at a time. Six additions then usually suffice (G. W. Sargent, Journ. Amer. Chem. Soc., 21. 858, 1899). 2 The alcohol may be recovered by distillation from quicklime and 100 c.c. of a concentrated solution of caustic soda. The distillate should give no reaction for boric acid. 3 The ^N-NaOH was standardised against boric acid. THE DETERMINATION OF BORON. 589 used in the following determination with three different mixtures of the substances just enumerated : First distillate (100 c.c.) required . 2'2 1'9 1-9 c c. iN-NaOH. Second distillate (1 00 c.c) required . 0'4 Q'5 1 0'3 c.c. ^N-NaOH*. Total B 2 3 found .... 0'0348 0'0346 0'0352 grm. Hence, it is inferred that the substances above enumerated did not interfere with the volatilisation of the boric oxide, nor with its accurate titration afterwards. Other Methods of collecting the Methyl Borate. instead of decomposing the distillate with sodiiyn hydroxide as indicated above, Thaddeeff 2 collected the distillate in a solution of caustic potash, and subsequently precipitated the boric acid as potassium boron uoride ; Rosenbladt 3 hydrolysed the methyl borate with magnesia; Gooch 4 used calcium oxide for the same purpose. In the two latter cases the contents of the distilling flask were acidified with nitric or acetic acid (Gooch), or a non-volatile acid sulphuric (Rosenbladt) or phosphoric (Gladding) 5 acid. The magnesia, being insoluble, does not act very rapidly, and lime is difficult to calcine to a constant weight on account of its tendency to absorb carbon dioxide and moisture. Hehner used disodium phosphate ; 6 Schneider and Gaab, 7 sodium carbonate ; Stolba, 8 borax ; and Gooch and Jones, 9 sodium tungstate (approximately 5 grms. per O'l grm. B 2 3 ). The last-named substance has the advantage of being easily obtained pure, and it is soluble in water, so that the hydrolysis proceeds comparatively quickly. 10 307. Determination of Silica and Alumina in Boro-silicates. In the analysis of silicates containing boric oxide, only part of the boric oxide is lost during the two evaporations to dryness ; the other part of the boric oxide is precipitated with the alumina and iron. This furnishes high results. 11 To eliminate the boric oxide in the silica evaporations, add methyl alcohol, or better (Jannasch) 12 methyl alcohol 13 saturated with hydrogen chloride, to the basin in which the evaporation for silica is being conducted. If the first " silica filtrate " gives an indication of boric oxide, the treatment may be repeated with the second "silica evaporation." 1 A third distillate (100 c.c.) contained the equivalent of 0'0005 grm. B 2 3 . 2 C. Thaddeeff, Zeit. anal. Chem., 36. 568, 1897. 3 T. Roseubladt, Zeit. anal. Chem., 26. 1887 ; Chem. News, 55. 18, 101, 1887- 4 F. A. Gooch, Amer. Chem. Journ., 9. 23, 1887 ; Chem. News, 55. 7, 1887 ; C. Marignac, Zeit. anal. Chem., I. 406, 1862 ; K. Kraut, ib., 36. 165, 1897 ; S. L. Penfield and E. S. Sperry, Amer. J. Science (3), 34. 220, 1887. A gram of lime, for instance, is weighed and evaporated to dryness with the methyl borate after ignition. The increase in weight represents boric oxide. The solid "crawls" badly during the evaporation, and the methyl alcohol " bumps" badly if it is allowed to boil. 5 T. S. Gladding, Journ. Amer. Chem. Soc., 20. 258, 1898 ; M. P. Schaak, Journ. Soc. Chem. Ind., 23. 699, 1904 ; H. Liihrig, Pharm. Cenir., 42. 50, 1901. C. Fresenius and 0. Popp (Zeit. offent. Chem., 3. 155, 188, 1897) used anhydrous sodium sulphate. 6 0. Hehner, Analyst, 16. 141, 1891. 7 A. Schneider and Gaab, Pharm. Centr., 7. 672, 1897. 8 F. Stolba, Journ. prakt. Chem. (2), 90. 457, 1863; R. J. Manning and W. R. Lang(/0?mi. Soc. Chem. Ind., 26. 803, 1907) precipitate the boric oxide by barium chloride in the distillate and weigh as P>a(B0 2 ) 2 . 9 F. A. Gooch and L. C. Jones, Amer. J. Science (4), 7. 34, 1899 ; Chem. News, 79. 99, 111, 1899. 10 The sodium tungstate, freed from carbonate by fusion with a little tungstic acid, is weighed in a platinum dish and then placed in the Erlenmeyer's flask F (fig. 188). After the distillation, methyl alcohol is boiled off, and the contents are transferred to the same platinum dish as was used for weighing the sodium tungstate. Evaporate to dryness, fuse as before, cool, and weigh. The increase in weight represents the boric oxide. 11 F. Wohler, Liebig's Ann., 141. 268, 1867 ; R. Fresenius and E. Hintz, Zeit. anal Chem., 28. 324, 1889 ; W. Hampe, Chem. Ztg., 15. 521, 1891. 12 P. Jannasch, Zeit. anorg. Chem., 12. 208, 1896 ; Zeit. anal. Chem., 36. 383, 1897. 13 The methyl alcohol should not blacken or evolve sulphur dioxide when heated with sulphuric acid on a water bath. CHAPTER XLI. THE DETERMINATION OF PHOSPHORUS. An accurate determination of phosphoric oxide, by whatever method it is made, . requires much skill and not a little practice. The advocate of one method often makes the other suffer in comparison, more by reason of his own want of skill in the manipulation of the method than by reason of any great advantage inherent in his own. A. A. BLAIR. 1 308. The Properties of Ammonium Phosphomolybdate. THE methods for the determination of phosphorus have probably been the subject of more investigations than any other analytical process. The subject, though clarified a little, is by no means definitely settled, since a number of contradictory statements confront the student of the subject. These can only be decided in the laboratory, but the very fact that the phenomena connected with this determination have led to a number of contradictions is a remarkable testimony to the obscurity of the reactions involved. In the valuation of phosphates by chemists selected by buyer and by seller, I have been informed that differences involving 10-15 per cent, of the total cost of the materials were at one time by no means uncommon ; and many pnalytical chemists were accordingly classed by sellers and buyers as "high" and "low" analysts. 2 The buyer naturally favoured the "low" analyst; "the seller, the "high" analyst. The differences, in most cases, were probably due to the need for standard methods of analysis, particularly in dealing with so imperfect and faulty a method of analysis as was then in vogue. In 1851, Sonneschein showed that the reaction between ammonium molybdate and orthophosphoric acid, discovered by Svanberg and Struve 3 three years earlier, could be employed for the quantitative determination of phosphorus. The precipitation of the orthophosphoric acid 4 takes place in the presence of iron, aluminium, lime, magnesia, and the alkalies. The yellow ammonium phosphomolybdate is but sparingly soluble in water. According to Chesneau, 5 - 030 grm. of the ammonium phosphomolybdate 1 A. A. Blair, Chem. News, 56. 246, 1887. 2 A. E. Davies, Chem. News, 23. 220, 1871. 3 L. Svanberg and H. Struve, Journ. prakt. Chem. (1), 44. 257, 1848 ; H. Struve, ib. (1), 54. 288, 1851 ; L. Sonneschein, ib. (1) 53. 339, 1851 ; H. Rose, Pogg. Ann., 76. 26. 1849. 4 If raeta- or pyro-phosphoric acids be present, they ought to be oxidised with, say, potassium permanganate or chromic acid. If a brown precipitate of manganese peroxide be produced, insoluble in acids, a small quantity of potassium nitrite will clear the solution. The manganese dioxide is reduced to MnO, which dissolves in the acid. Nitric acid and potassium chlorate are not always vigorous enough to convert the meta- and pyro- to, ortho-phosphoric acid. C. Meineke and E. F. Wood, Rev. Univ. Mines (3), 9. 235, 1890 ; M. von Reiss, Stahl Eisen, 9. 1025, 1889 ; 10. 1059, 1890. 5 C. Chesneau, Rev. Met., 5. 237, 1908 ; V. Eggertz, Journ. prakt. Chem. (1), 79. 496, 1860 ; F. Hundeshagen, Chem. News, 60. 169, 177, 188, 201, 205, 1889; Zeit. anal. Chem., 28. 164, 1889 ; 32. 144, 1893 ; R. Freseuius, ib., 3. 446, 1864 ; G. Jorgensen, ib., 45. 273, 1906 ; 590 THE DETERMINATION OF PHOSPHORUS. 591 dissolves in a litre of water at 15. This means that if a litre of water comes in contact with the precipitate, 0'030 grm. will pass into solution. Action of Nitric Acid. Hundeshagen represents the reaction between ammonium molybdate and, say, sodium phosphate by the equation : Na 3 P0 4 + 12(NH 4 ) 2 Mo0 4 + 26HN0 3 = (NH 4 ) 3 P0 4 . 12Mo0 3 . 2HN0 3 . H 2 + etc., where 1 grm. of phosphoric anhydride P 2 ^5 corresponds with 23 grms. of nitric acid HN0 8 . But the amount of nitric acid can be increased considerably beyond this limit without interfering with the precipitation. If, however, the quantity of the nitric acid rises above 71 grms. HN0 3 per gram of P 2 5 , the precipitate of ammonium phosphomolybdate will be partially decomposed, and the precipitation will consequently be incomplete. If over 1700 grms. of HN0 3 per gram of P 2 5 be present, no ammonium phosphomolybdate will be precipitated. Both hydrochloric and sulphuric acids act more powerfully than nitric acid. In consequence, nitric acid gives the ividest margin of safety between a 'sufficient quantity and an overdose of acid. The solvent effects of an overdose of nitric acid can be overcome by adding an excess of molybdate. The greater the excess of acid, the greater the amount of molybdate required for the complete precipitation of the phosphate. Chesneau gives the solubility of ammonium phosphomolybdate in dilute nitric acid as follows : Table LXVI. Solubility of Ammonium Phosphomolybdate in Nitric Acid. Nitric acid. Ammonium phosphomolybdate dissolved per 100 c.c. acid. Per cent. Grm. at 15. 0-0030 1 0-0371 5 0-0682 10 0-0901 This table shows that nitric acid augments the solubility of the precipitated ammonium phosphomolybdate. Action of Ammonium Nitrate. Chesneau gives the solubility of ammonium phosphomolybdate in aqueous solutions of ammonium nitrate as follows : Table LXVII. Solubility of Ammonium Phosphomolybdate in Aqueous Solutions of Ammonium Nitrate. Ammonium nitrate. Ammonium phosphomolybdate dissolved per 100 c.c. solution. Per cent. 5 10 Grm. at 15. 0-0030 0-0099 0-0113 Analyst, 34. 372, 1909 ; Mem. Acad. Roy. Soc. Danemark(7), 2. 141, 1905 ; Zeit. angew. Chem., 24. 542, 1911 ; R. Finkener, Ber., II. 1638, 1878. Some say that the solubility of the precipi- tate is decreased, and some say that the solubility is increased, by the addition of ammonium nitrate. 592 A TREATISE ON CHEMICAL ANALYSIS. This shows an increase in the solubility of ammonium phosphomolybdate in the presence of ammonium nitrate. In the absence of ammonium nitrate, the precipitate is inclined to separate in a colloidal condition which is exceedingly difficult to filter clear. Hundeshagen showed that this will occur if less than 0*5 per cent, of ammonium nitrate be present in the solution. The precipitate, even when this amount is present, settles slowly. In order to get rapidly settling precipitates, from 5 to 15 per cent, of ammonium nitrate should be present in the solution. Besides accelerating the rate of settling, ammonium nitrate also accelerates the rate of precipitation, and ensures a more complete precipitation with a smaller excess of ammonium molybdate. 1 These facts are illustrated by the following numbers. The conditions in the different experiments were other- wise the same. Table L XV II I. Effect of Ammonium Nitrate on the Precipitation of Ammonium Phosphomolybdate. Ammonium molybdate. Amount of P 2 5 precipitated. No nitrate. 10 grms. of nitrate. c.c. 50 60 70 80 90 O'llOO 0-1156 0-1155 0-1158 0-1165 0-1165 0-1160 0-1160 0-1162 0-1163 Hence, 50 c.c. of a given solution of ammonium molybdate will precipitate all the phosphoric acid in a given solution in the presence of ammonium nitrate, under conditions where 5 "6 per cent, may escape precipitation in the absence of the ammonium nitrate. Influence of Temperature. The precipitate forms rather more quickly in hot solutions. According to Hundeshagen, if the ammonium phosphomolybdate be precipitated in the cold, it separates in the form of rounded grains of different sizes, which settle very slowly. Such precipitates are inclined to choke the pores of the filter paper and retard filtration. If the precipitation occurs in hot solutions, the ammonium phosphomolybdate separates in the form of octahedral crystals (usually grouped " rosettenf ormig " ), which settle rapidly and filter easily. The precipitate formed in hot solutions is also easier to wash. If the temperature of the solution be above 60, white needles of ammonium tetra- molybdate may be produced by a prolonged heating, and this the more, the higher the temperature, and the longer the solution is heated. Ammonium tetramolybdate is soluble in water and nitric acid, but is precipitated by ammonium nitrate. When ammonium tetramolybdate is once formed, it will contaminate the precipitate even if the ammonium phosphomolybdate be dissolved and reprecipitated a second time. Hence, in order to avoid this source of error, 2 the solution should not be heated too long a time, nor at too high a temperature. 1 E. Richters, Zeit. anal. Chem., 10. 471, 1871 ; E. Stunkel, T. Wetzke, and P. Wagner, ib.,21. 353, 1882 ; F. Hundeshagen, ib., 28. 141, 1889 ; C. Gilbert, Corresponded Ver. anal. Chem., 1, 1878. 2 In illustration of the different ideas which prevail as to the time and temperature of heating required for complete precipitation, the following may be quoted : R. Fresenius (Zeit. Rammelsberg. 577 3'25 3-90 86-45 Sonneschein. 11-23 11-23 3-03 86-87 Gibbs. 3-94 3-35 3-66 89-05 Gladding. 2-44 2-44 376 91-36 THE DETERMINATION OF PHOSPHORUS. 593 Composition of the Precipitate. The actual composition of the precipitate depends upon the time of standing, temperature, and the nature of the mother liquid. This partly explains how so many different factors have been proposed for converting the weight of the precipitate into the equivalent "P 2 6 ." Thus, factors varying from 3'03 per cent. P 2 5 (Sonneschein) to 4 '39 per cent. (Debray) have been suggested. The following selection l will illustrate the different results which have been obtained for the composition of the yellow ammonium phospho- molybdate : H 2 0. NH 4 PA Mo0 3 Indeed, if the precipitation be not performed under definite conditions, there is some uncertainty as to the composition of the precipitate. Hence, the precipitate is frequently redissolved in ammonia and reprecipitated in the form of ammonium magnesium phosphate, or as ammonium phosphomolybdate. Chesneau 2 has studied this phase of the subject. He found that when ammonium nitrate is absent, the precipitate weighs less than theory requires, and this the more with solutions containing small amounts of phosphorus. When ammonium nitrate is present, the precipitate with solutions poor in phosphorus is heavier than theory requires. As a matter of fact, the precipitate is not always a single compound, but a mixture of different substances ammonium phosphomolybdate, molybdic acid, ammonium tetramolybdate, and occluded ammonium molybdate. 3 When ammonium nitrate is absent, the amount of molybdic acid in the precipitate increases with increasing concentration of the phosphoric acid, up to a certain definite limit. When ammonium nitrate is present, the amount of molybdic acid increases as the amount of phosphorus decreases. When no molybdic acid crystals are present, as will be the case when about 005 grm. of phosphorus is present, the stellate crystals of ammonium phospho- molybdate predominate. If the amount of phosphorus be reduced to 0*00027 grm., the precipitate will be contaminated with prismatic crystals of ammonium tetramolybdate. Influence of Ammonium Salts. Fresenius 4 has shown that the presence of anal. C/iem., 3. 446, 1864) 6 hrs. at 65 ; R. Fresenius (Anleitung zur quantitative chemischen Analyse, Braunschweig, i. 411, 1875) 12 hrs. at 40; 0. Abesser, W. Jani, and M. Marcker (Zeit. anal. Chem., 12. 254, 1873) 4 to 6 hrs. at 50J; C. Stiinkel, T. Wetzke, and P. Wagner (ib., 21. 353, 1882)! hr. at 80 to 90. P. Pietsch, W. Rohn, and P. Wagner, Zeit. anal. Chem., 19. 444, 1880 ; R. de Roode, Journ. Atner. Chem. Soc., 17. 43, 1895 ; M. Fleischer and K. Miiller, Journ. Landw., 2O. 96, 1875. 1 C. Rammelsberg, Ber. Berl. Akad., 573, 1877; Ber., 10. 1776, 1877 ; R. Finkener, ib., II. 1638, 1878; L. Sonneschein, Journ. prakt. Chem. (1), 53. 339, 1851; M. Seligsohn (1), 67. 470, 1856; V. Eggertz, ib. (1), 79. 490, 1860; W. Gibbs, Amer. Chem. Journ., 3. 317, 406, 1881; Proc. Amer. Acad., 18. 232, 1883; 21. 96, 1885; Chem. News, 45. 29, 1882; H. Debray, ib., 17. 183, 1868 ; Compt. Rend., 66. 702, 1868 ; D. J. Hissink and H. van der Waeiden, Chem. Weekblad, 2. 179, 1905 ; G. A. Spiess, Viertelj. Pharm., 9. 257, 1860 ; M. Nutzinger, ib., 4. 549, 1855 ; J. Konig, Landw. Ver. Stat., IO. 401, 1868 ; A. Villiers and F. Borg, Bull. Soc. Chim. (3), 9. 486, 1893; Compt. Rend., 116. 989, 1893; H. C. Babbitt, Journ. Anal. App. Chem., 7. 165, 1893; A. von Lipowitz, Fogg. Ann., 109. 135, 1860; C. Friedheim, Zeit. anorg. Chem., 4. 275, 1893; T. Salzer, Liebig's Ann., 187. 322, 1877; G. P. Baxter and R. C. Griffin, Amer. Chem. Journ., 34. 204, 1905 ; T. S. Gladding, Journ. Amer. Chem. Soc., 18. 23, 1896. 2 G. Chesneau, Rev. Met., 5. 237, 1908; Principes Theoriques et Pratiques d' Analyse minerale, Paris, 233, 1912 3 G. P. Baxter, Amer. Chem. Journ., 28. 298, 1902. 4 R. Fresenius, Zeit. anal. Chem., 3. 446, 1864 ; R. Fresenius, C. Neubauer, and E. Luck, ib., 10. 133, 1871 ; G. Konig, ib., 10. 305, 1871 ; E. Richters, Dinglcr's Journ., 199. 183,1871. 3 s 594 A TREATISE ON CHEMICAL ANALYSIS. ammonium chloride retards the precipitation, 1 and Ricbters has shown that the presence of ammonium sulphate acts in a similar manner. In these cases, a larger excess of ammonium molybdate must be added, or the results will be low. Influence of Iron Salts. If ferric salts be present, a comparatively large excess of ammonium molybdate is needed, even in the presence of ammonium nitrate ; and if too little acid be present, a reddish crust which does not readily dissolve in ammonia may be formed. This is due to the formation of a ferric phosphomolybdate. 2 According to Meineke, a larger excess of nitric acid is also needed if much iron be present in the solution, in order to prevent the precipita- tion of iron with the ammonium phosphomolybdate ; and Jiiptner 3 proposes to use 6 per cent, of tartaric acid with the molybdate solution in order to prevent the precipitation of the iron. Influence of Titanium and Vanadium. The presence of titanium retards the precipitation of ammonium phosphomolybdate. For instance, two solutions, each containing the equivalent of 0*140 grin, of ammonium phosphomolybdate, furnished, in the presence of the equivalent of Titanic acid O'OOS 0'016 grm. Ammonium phosphomolybdate . . . 0*122 0118 grm. Hence, there is a deficiency of 13 per cent, of P 2 5 even in the presence of but O'OOS grm. of titanic acid. Pattinson 4 separates the titanium 5 before pre- cipitating the phosphorus, in the following manner : Reduce the ferric to ferrous sulphate ; add a saturated solution of ammonium alum, and then an excess of ammonia. The precipitate contains aluminium phosphate, aluminium hydroxide, and titanic oxide. Fuse the ignited precipitate with sodium carbonate, and extract the fused mass with water. Sodium phosphate and aluminate dissolve, while sodium titanate remains insoluble. " As a rule, one fusion suffices to separate all the phosphoric acid from the titanic acid." Vanadium is carried down quantitatively by the phosphomolybdate precipitate, which has then different properties from the ordinary precipitate, for it is more soluble in nitric acid, and in the ordinary washing solutions. 6 Influence of Arsenic Oxide. If arsenic be present, and the temperature of the solution be too high, ammonium arsenomolybdate will contaminate the precipitate of ammonium phosphomolybdate, and, later on, the second precipitate. A temperature of 45 is generally considered a safe upper limit, 7 but if much arsenic be present, the phosphomolybdate precipitate will even then be contamin- 1 C. Meineke (Chem. Ztg., 20. 113, 1890) says that ammonium chloride does not retard the precipitation. 2 G. Chesneau, Rev. Met., 5. 237, 1908 ; G. Arth, Bull. Soc. Chim. (3), 2. 324, 1890; Chem. News, 62. 155, 1890 ; C. Meineke, Rep. Anal. Chem., 6. 304, 1886 ; E. Wolff, Zeit. anal. Chem., 3. 94, 1864 ; R. Fresenius, ib., 3. 446, 1864; P. Schweitzer, ib., 9. 85, 1870; J. V. Janovsky, ib., n. 157, 1872; S. W. Johnson, Joum. Amer. Chem. Soc., 16. 462, 1894; L. von Szell, Landw. Ver. Stat., 55. 341, 1901. 3 R Woy, Chem. Ztg., 21. 470, 1897; H. F. von Jiiptner, O&Aer. Zeit. Berg. Hutt., 42. 471, 4 J. andH. S. Pattinson, Journ. Soc. Chem. Ind., 14. 443, 1895; P. H. Waller, Joum. Amer. Chem. Soc., 20. 513, 1898. 5 In silicate analyses, titanium phosphate separates with the silica during the evaporation for silica. 6 H. Brearley and F. Ibbotson, The Analysis of Steel- Works Materials, London, 163, 1902 ; J. R. Cain and J. C. Hostetter, Journ. Ind. Eng. Chem,., 4. 250, 1912. 7 H. C. Babbitt, Journ. Amer. Chem. Soc., 6. 381, 1884 ; M. Frank and F. W. Hinrichsen, Stakl Eisen, 28. 295, 1908; M. Seligsohn, Journ. prakt. Chem. (1), 67. 481, 1856; H. Seyberth, Ber., 7. 391, 1874 ; W. Gibbs, Amer. Chem. Journ., 7. 313, 1886 ; E. D. Campbell, Journ. Anal. App. Chem., 7. 2, 1893. THE DETERMINATION OF PHOSPHORUS. 595 ated. For instance, Jorgensen found the following amounts of arsenic in the precipitated phosphomolybdate after standing 24 hours at 37 : Original solution contained . . 0'0033 0-0066 0'0143 grm. arsenic acid. Precipitate contained . . . 0'0013 0'0029 0'0040 grm. arsenic acid. It is therefore advisable to remove the arsenic, if present, either by means of hydrogen sulphide (Fresenius) or by evaporation with hydrochloric and oxalic acids (Campbell). 1 Influence of Silica. It is generally agreed that silica should be removed, since it is liable to be precipitated in the form of ammonium silicomolybdate, 2 which does not settle so quickly as the corresponding phosphorus compound. According to Isbert and Stutzer, 3 the silicomolybdate can be removed from the phospho- molybdate by washing with ice-cold water, in which the former is easily soluble, the latter almost insoluble. The silica is, however, usually removed by evapora- tion before the separation of the phosphorus. Influence of Organic Matter. Carbonaceous matters are generally supposed to hinder the precipitation of the ammonium phosphomolybdate precipitate ; 4 so also are the organic acids tartaric and oxalic. Jiiptner, however, says that this is not the case, and even recommends the addition of tartaric acid with the ammonium molybdate solution to retard the precipitation of the iron. 309. The Gravimetric Determination of Phosphorus Woy's Process. Woy's method 5 of separating phosphorus in the form of ammonium phospho- molybdate is one of the best. The phosphoric oxide in an aliquot portion 6 of the acid solution of the pyrosulphate fusion obtained in the regular course of the typical silicate analysis (page 186) may be determined by this process; or separated by Joulie's process (page 603), and determined colorimetrically. 1 R. Fresenius, Anleitungzur quantitativen chemischen Analyse, Braunschweig, I. 421, 1876 ; E. D. Campbell, Journ. Anal. App. Chem., 6. 370, 1890 ; G. Jorgensen, Mem. Acad. Roy. Soc. Danemark (7), 2. 141, 1905 ; P. Champion and H. Pellet, Bull. Soc. Chim. (2), 27. 6, 1877 ; CJiem. News, 35. 115, 1877 ; G. Maderna, Atti Accad. Lincei, 19. 15, 1910. 2 T. Petersen, Verhandl. Qeol. Reichsanst., 80, 1869; W. Knop, Chem. Centr. (2), 2. 691, 861, 1857 ; R. Fresenius, Anleitung zur quantitativen chemischen Analyse, Braunschweig, I. 411, 1903; J. H. Vogel, Repert. Anal. Chem., 7. 568, 469, 1887; E. Richters, Dingler's Journ., 199. 183, 1871 ; R. W. Atkinson, Chem. News, 35. 127, 1887. On the other hand, E. Thilo (Chem. Ztg., II. 193. 1889), E. R. E. Miiller (ib., 35. 1201, 1911), E. H. Jenkins (Journ. prakt. Chem. (2), 13. 237, 1876), G. Kennepohl (Chem. Ztg., u. 1089, 1887), C. Meineke (Repert. Anal. Chem., 6. 303, 325, 1886), J. Hanamann (Zeit. Landw. Vcrs. Ost., 3. 53, 1900), and H. C. Sherman and H. S. J. Hyde (Journ. Amer. Chem. Soc., 22. 652, 1900) consider that no notice need be t;iken of the silica. For the separation of phospho- and silico-molybdates, see P. G. Melikoff, Compt. Rend., 153. 1478, 1911. The method is to digest the precipitate by molybdic solution with " permolybdate reagent" four times during 24 hours. Filter. Destroy the hydrogen peroxide by heating, and precipitate the phosphoric acid in the filtrate as usual. The "permolybdate reagent" is made by mixing equal volumes of 30 per cent, hydrogen peroxide and 15 percent, ammonium nitromolybdate. The method is also used for the separation of phosphoric and colloidal silicic acids P. G. Melikoff and M. Becaia, ib., 154. 775, 1912 ; Chem. Ztg., 36. 617, 1912; P. G. Melikoff, Journ. Russ. Phys. Chem. Soc., 44. 605, 1912. 3 A. Isbert and A. Stutzer, Zeit. anal. Chem., 26. 583, 1887 ; Chem. News, 57. 211, 1888. 4 F. Hundeshagen, Zeit. anal. Chem., 28. 164, 1899; J. Konig, ib., 10. 305, 1871; E. Richters, Dingier* s Journ.. 199. 183, 1871 ; V. Eggertz, Journ. prakt. Chem, (1), 79. 496, 1860 : H. von Jiiptner, Oester. Zeit. Berg. Hiitt., 42. 471, 1894 ; R. Woy, Chem. Ztg., 21. 470, 1897 ; C. Reichard, ib., 27. 833, 1903; R. Hamilton, Journ. Soc. Chem. Intl., IO. 904, 1891; G. Maderna, Atti Accad. Lincei, 19. 827, 1910; A. Taram, Chem. Neios, 49. 208, 1884; M. Schild, Chimiste, 3. 25, 1912. 5 R. Woy, Chem. Ztg., 21. 441, 469, 1897; A. H. Meade, Chem. News, 101. 241, 1910; G. B. van Kampen, Cliem. Weekblad, 3. 376, 1906. 6 Or in the portion used for the determination of iron or titanium. A TREATISE ON CHEMICAL ANALYSIS. An aliquot portion, 1 say 100 c.c., of the solution of the pyrosulphate fusion (page 186) is neutralised with ammonia, and evaporated, if necessary, to about 50 c.c. 2 Add concentrated ammonia until a precipitate forms which does not disappear on standing; add 3 c.c. of concentrated nitric acid; add, say, 15 c.c. of ammonium nitrate, 3 so that the solution has between 5 and 15 per cent, of ammonium nitrate. 4 Place a thermometer in the liquid. Heat the solution to about 70. Raise the thermometer about 2 inches above the level of the liquid in the beaker, and pour rapidly into the solution, say, 20 c.c. of ammonium molybdate 5 down the stem of the thermometer, using it as a glass rod. Agitate the liquid thoroughly. The yellow precipitate separates more thoroughly in hot solutions. Keep the solution at 70 for about half an hour. Decant through, say, a 7- or a 9-cm. filter paper, and wash the precipitate 6 by decantation 7 (about six times) with a mixed solution (about 50 c.c.) of equal volumes of the ammonium molybdate solution, ammonium nitrate, and nitric acid (1 : 2), until no permanent precipitate 8 separates from a drop of the wash-liquid on standing. 9 The precipitate is now distributed between the beaker and the filter paper. Place the beaker with the precipitate beneath the funnel, and dissolve the precipitate 1 If possible, no more solution should be taken than is equivalent to 0*1 grm. of P 2 5 . If more than this amount be present, take an aliquot portion and dilute to, say, 50 c.c. 2 Most other solutions are treated in a similar manner. Arsenic is supposed to be absent. 3 AMMONIUM NITRATE SOLUTION. Dissolve 320 grms. of ammonium nitrate in water and make the solution up to a litre. 4 The following table gives the relative proportions of ammonium molybdate, nitric acid, and ammonium nitrate to be used : Table LXIX. Composition of Mother Liquid for the Precipitation of Ammonium Phosphomolybdate. P 2 5 . Ammonium molybdate. Ammonium nitrate. Nitric acid. grm. c.c. c.c. c.c. o-i 120 30 19 O'Ol 15 20 10 0-005 15 20 10 0-002 10 15 5 o-ooi 10 15 5 When 0-5 grm. of substance is in question, every per cent, of P 2 5 in the original sample requires 5 c.c. of the molybdate solution (T. P. Treadwell, Kurzes Lehrbuch der analytischen Chemie, Leipzig, 361, 1911). 5 AMMONIUM MOLYBDATE SOLUTION. Dissolve 34'34 grms. of ammonium heptamolybdate (NHJeMc-?^' 4H 2 in water and make the solution up to a litre (E). 1 c.c. is nearly equivalent to '001 grm. P 2 5 . Do not make a large stock to be stored a long time. Note that common ammonium molybdate of commerce is not (NH 4 ) 2 Mo0 4 , but (NH 4 ) 6 Mo 7 2 4. 4H 2 0. The gradual precipitation of molybdic acid is avoided by storing the solution in green glass bottles to cut off heat rays from light W. Heike, Stahl Eisen, 29. 1446, 1911. For the properties of molybdate solutions, see papers cited above and A. L. Winton, Journ. Amer. Chem. Soc. t 18. 445, 1896; M. Kupferschlager, Bull. Soc. Chim. (2), 36. 644, 1881. In washing, the yellow precipitate often has a tendency to crawl above the top of the paper. Hence, the paper should fit the funnel closely, so that the portion which crawls will not be lost. There is no need to transfer all the precipitate from the beaker to the filter paper when a second precipitation is to be made. 8 Pure water' decomposes the precipitate. If a yellow precipitate should separate in the washings on standing, all the phosphorus was not precipitated. In that case, the filtered liquid must be treated with more ammonium molybdate, filtered, and washed again. A white precipi- tate of molybdic acid or ammonium tetramolybdate can be ignored. 9 E. Raben (Zeit. anal. Uheni., 47. 546, 1908) washes until the wash-water gives no precipitate with potassium ferrocyanide. THE DETERMINATION OF PHOSPHORUS. 597 on the filter paper with dilute ammonia (2J per cent.). If all the precipitate in the beaker does not dissolve, pour more ammonia into the beaker. 1 Wash the filter paper half a dozen times with the 2J per cent, ammonia. The liquid in the beaker will now occupy about 50-100 c.c. It contains practically all the phos- phoric acid, some molybdic acid, and ammonium nitrate in ammoniacal solution. 2 The phosphorus may be determined in several different ways. ' Keprecipitation as ammonium magnesium phosphate or as ammonium phosphomolybdate will give satisfactory results. 3 310. The Reprecipitation as Ammonium Magnesium Phosphate. Add, say, 10 c.c. 4 of magnesia mixture 5 to the hot 6 ammoniacal solution of the washed ammonium phosphomolybdate in 2 '5 per cent, ammonia solution. C5* t\ oT 1 / C / b / c: ^ / ^ ^ 8 ^ - --" / ' _^ 1 - ^- *s "rms of Mgz Pg0 7 Fo ur\d 0-07 0-10 0-15 0-20 0-25 0-30 FIG. 189. Neubauer's correction chart. 0-35 The addition is made very slowly, drop by drop, with constant stirring. 7 Let the mixture stand about 3 hours. Wash and ignite the precipitate as described for the determination of magnesia (page 218), and use Table XCI. Neubau&ft Correction. It is not possible to precipitate pure ammonium 1 If the liquid in the beaker remains turbid, acidify with nitric acid, add a small crystal of citric acid, and, finally, ammonia until the liquid is alkaline. If the liquid be still turbid, filter, ignite in a platinum crucible, fuse with a little sodium carbonate, dissolve the cake in water, acidify with nitric acid, and add it to the rest of the solution. 2 For error due to phosphorus in the glass beakers, etc., see A. Vita, #ta/^ Risen, 32. 1352, 1912. 3 RECOVERY OF MOLYBDENIUM RESIDUES. R. Fresenius, Zeit. anal. Chem., 10. 204, 1871 ; F. Muck, ib., 10. 307, 1871 ; 8. 377, 1869 ; O. Maschke, ib., 12. 380, 1873 ; H. Uelsmann, ib., 16. 52, 1877 ; E. Reichardt, Arch. Pharm. (3), 2. 232, 1873 ; A. Gawalovski, Chem. News, 75. 98, 1897 ; Oester. Chem. Ztg., i. 385, 1898; A. Borntrager, Deut. Chem. Ztg., 9. 10, 1894; Chem. News, 70. 224, 1894. ' 4 As many c.c. are needed as there are centigrams of P 2 5 in the solution. 5 MAGNESIA MIXTURE. Dissolve 55 grms. of crystallised magnesium chloride ( MgCl 2 . 6H 2 0) and 70 grms. of ammonium chloride in 650 o.c. of water. When all is dissolved, make the solution up to a litre with concentrated aqueous ammonia. The solution is about E strength. M. Brassier, Ann. Chim Phys. (4), 7. 355, 1866 ; A. L. Winton, Journ. Amer. Chem. Soc., 18. 445, 1896. G. Loges (Chem. Ztg., 8. 1743, 1884) uses magnesium ammonium chloride, which is not hygroscopic like magnesium chloride : 70 grms. magnesium ammonium chloride ; 55 grms. ammonium chloride ; make up to a litre with 24 per cent, ammonia. Magnesia mixture acts slowly on glass bottles L. L. de Koninck, Chem. Ztg., 19. 450, 1895. Magnesia mixture containing magnesium sulphate instead of magnesium chloride is liable to give precipitates contaminated with basic magnesium sulphates. See page 218. 6 According to 0. Foerster (Chem. Ztg., 16. 109, 1892 ; Chem. News, 67. 143, 1893), the dse, the Otherwise, the washing of the solution should be heated before adding the magnesia mixture, precipitate is very difficult. 7 The precipitate should be white and crystalline. If it be red and flaky, the results will be inaccurate. If the phosphorus be finally weighed as magnesium pyrophosphate Mg.,P 2 7 traces of molybdic acid may contaminate the precipitate, even after two precipitations of the ammonium magnesium phosphate A. A. Blah 1 , Chem. Neirs, 56. 246, 1887 ; A Gawalovski, Oester. Chem. Ztg., I. 385, 1890. 598 A TREATISE ON CHEMICAL ANALYSIS. magnesium phosphate from the solution of the ammonium phosphomolybdate l either by nearly neutralising the solution with hydrochloric acid before adding the magnesia mixture, or by adding magnesia mixture drop by drop, with constant stirring, to the ammoniacal solution. The latter process gives the better results, and Neubauer recommends the use of a correction chart which gives the amount, in milligrams, of P 2 5 lost for different amounts (in grams) of Mg 2 P 2 7 finally obtained. Neubauer's chart is shown in fig. 189. It is used in the following manner : Suppose 0'0959 grm. of Mg 2 P 2 7 has been obtained, then 04 mgrm. of Mg 2 P 2 7 has been lost in the determination. Hence, 0-0959 + 0-0004 = 0-0963 grm. Mg 2 P 2 7 represents the weight of Mg 2 P 2 7 which would have been obtained if there had been no loss from this cause. 311. The Reprecipitation as Ammonium Phosphomolybdate. Add 20 c.c. of ammonium nitrate solution and 1 c.c. of ammonium molybdate to the ammoniacal solution of the ammonium phosphomolybdate first precipi- tated. 2 Heat the solution until gas bubbles begin to form, and add 20 c.c. of hot nitric acid gradually, with constant stirring. After standing 10 minutes, filter through a Gooch's crucible, and wash with the mixture previously employed .until no brown coloration is produced when a drop of the filtrate is brought in contact with a solution of potassium ferrocyanide. Dry 3 the yellow precipitate of (NH 4 ) 3 P0 4 .12Mo0 3 .2HN0 3 .H 2 in a current of" air at 160-180 in, say, a Paul's oven (fig. 127) to a constant weight, 4 and weigh as (NH 4 ) 3 P0 4 .12MoO B . 5 In that case, the weight of the precipitate multiplied by 0*03782 represents the weight of the P 2 ft in the sample taken for analysis. 6 312. Rapid Processes for Estimating the Ammonium Phosphomolybdate. In order to shorten the time required for the determination of phosphorus in terms of ammonium phosphomolybdate, the amount of molybdenum pre- cipitated is sometimes determined by volumetric methods. 7 In these it is 1 H. Neubauer, (Jeber die Zuverlassiglceit der Phosphor sdurebestimmung als Magnesium- pyrophosphat, Rostock, 1893; G. Jorgensen, Mem. Acad. Roy. Soc. Danemarlc (7), 2. 141, 1905 ; 0. Abesser, W. Jani, and M. Miircker, Zeit. anal. Chem., 12. 239, 1873 ; 13. Peitzsch, W. Rohn, and P. Wagner, ib., 19. 444, 1880; B. W. Kilgrove, Journ. Amer. Chem. Soc., 16. 793, 1894 ; <1. Jorgensen, Zeit. anal. Chem., 50. 337, 1911. 2 The phosphorus should not be precipitated first by the magnesia mixture, but be first precipitated by the molybdate process. This ensures a definite amount of ammonia and ammoniacal salts in the solution, and thus leads to concordant results. 3 E. Raben (Zeit. anal. Chem., 47. 546, 1908) finishes the washing with warm 70 per cent, alcohol ; then with absolute alcohol ; then with 10 c.c. of ether. He dries the precipitate at 110-120, and weighs as (NH 3 ) 3 P0 4 . 12Mo0 3 . A. Carnot, Compt. Rend., 106. 105, 1893; Chem. News, 67. 101, 1893; A. Villiers and F. Borg, ib., 67. 313, 1893 ; Compt. Rend., 106. 989, 1893; C. Reichard, Pharm. Zentralhalle, 52. 1314, 1911 ; N. von Lorenz, Oester. Chem. Ztg., 14. 1, 1911 ; G. Jorgensen, Zeit. angeiv. Chem., 24. 542, 1911 ; H. Neubauer and F.' Liicker, Zeit. anal. Chem., 51. 161, 1912. 4 The precipitate is somewhat hygroscopic, and the desiccator should have fresh concentrated sulphuric acid, not calcium chloride, otherwise the precipitate might gain in weight. 5 If the precipitate appears green, add a small crystal of ammonium nitrate and re-heat. The precipitate then becomes uniformly yellow. 6 If desired, the precipitate may be ignited to redness until the resulting 24Mo0 3 . P 2 5 has a homogeneous black colour. The ignited precipitate is not specially hygroscopic. 7 J. Macagno, Gaz. Chim. Ital., 4. 567, 1874 ; Chem. News, 31. 197, 1875 ; A. Grete, Ber., 21. 2762, 1888 ; F. A. Emmerton, Trans. Amer. Inst. Min. Eng , 15. 93, 1887 ; T. M. Drown and P. W. Shinier, ib., 10. 137, 1882 ; P. W. Shimer, ib., 17. 100, 1889 ; T. M. Drown, ib , 18. 90, 1890; C.Jones, ib., 17. 411, 1889; 18. 705, 1890; Chem. News, 62. 220, 231, 1890; D. L. THE DETERMINATION OF PHOSPHORUS. 599 assumed that the precipitated phosphomolybdate has a constant composition, and that a definite amount of molybdenum represents a definite amount of phosphorus. In Emmerton's process, the yellow precipitate is dissolved in, say, 5 c.c. of concentrated aqueous ammonia and 20 c.c. of water. Pour this mixture through the filter paper containing the precipitate, and wash the paper as indicated above. About 60 c.c. of liquid are so obtained. Add, say, 10 c.c. of concentrated sulphuric acid, 1 and at once reduce the molybdenum trioxide Mo0 3 to the green oxide Mo 2 3 or Mo 12 19 2 by means of metallic zinc (say, in the reductor, page 191). The solution, if necessary, is rapidly filtered from the undissolved zinc, and titrated with potassium permanganate, 3 when the green oxide is oxidised to Mo0 3 . The greenish-coloured solution becomes nearly colourless, and finally pink, characteristic of potassium permanganate. The oxidation is represented in symbols : 5Mo 2 3 + 6KMn0 4 = 10Mo0 3 + 3K 2 + 6MnO, or, 5Mo 12 19 + 34KMn0 4 = 60Mo0 3 + 17K 2 + 34MnO. The calculations are as follow : 1 grm. of KMn0 4 represents 1*5188 grms. of Mo0 3 if the reduction has been to Mo 2 3 ; and 1-60083 if the reduction has been to Mo 12 19 . But since the ratio of P 2 5 : Mo0 3 is as 1 : 24, every gram of Mo0 3 represents '04 108 grm. of P 2 5 . Hence, 1 grm. of KMn0 4 represents 0-06239 grm. of P 2 5 if the reduction has been to Mo 2 3 ; and 0'06576 grm. of P 2 5 if the reduction has been to Mo 12 19 . In a given titration, therefore, if 20 c.c. of potassium permanganate solution contain 1 grm. of KMn0 4 per litre, 1 c.c. of the solution represents 0-00006239 grm. of P 2 5 , or 20 c.c. represent 20x0-00006239 = 0-00125 grm. of P 2 0- in the sample undergoing titration. 4 Pemberton washes the precipitate of ammonium phosphomolybdate free from nitric acid, and dissolves the precipitate in standard alkali. The excess of alkali is titrated with standard acid, with phenolphthalein as indicator. The reaction is represented : 2(NH 4 ) 3 P0 4 . 12Mo0 3 + 46KOH = etc. 5 Osmond dissolves the ammonium phosphomolybdate, 6 collected on an Randall, ib., 97. 113, 1908; Amer. J. Science (4), 24. 313, 1907; 0. F. von der Pfordten, Zeit. anal. Chem., 23. 432, 1884 ; A. Werncke, ib., 14. 1, 1875 ; H. Pemberton, Journ. Franklin Inst. (3), 83. 184, 1882 ; Chem. Neivs, 46. 4, 1882 ; Journ. Amer. Chem. Soc., 16. 278, 1894 ; A. A. Blair and J. E. Whitfield, 17. 747, 1895 ; W. A. Noyes and E. D. Frohman, ib., 10. 533, 1894 ; W. A. Noyes and J. S. Royse, ib., 17. 129, 1895 ; C. B. Dudley andF. N. Pease, ib., 16. 224, 1894 ; R. A. Mahon, ib., 10. 792, 1897 ; 0. S. Doolittle and A. Eavenson, ib. t 10. 234, 1894 ; G. Auchy, ib., 18. 955, 1896 ; B. W. Kilgore, ib., 16. 765, 1894 ; 17. 741, 950, 1895 ; E. D. Campbell, Journ. Anal. App. Chem., I. 370, 1887 ; B. W. Kilgore, Bull. U.S. Agrw. Dept. (Chem.}, 43. 68, 1894 ; G. P. Baxter, Amer. Chem. Journ., 28. 298, 1902 ; 34. 204, 1905 ; J. G. Fairchild, Journ. Ind. Chem. Eng., 4. 520, 1912. See page 415. 1 The smaller the amount of sulphuric acid in the solution, the quicker the reduction. 2 Complete reduction to Mo 2 3 is difficult, and the reduced oxide has, in consequence, been represented by different formula Mo 2 3 (Pfordten) ; Mo 12 19 (Emmerton) ; Mo*Qfl (Blair and Whitfield). 3 The permanganate should always be standardised with a known amount of phosphorus in combination. . 4 If the solution has been standardised for a given amount of P 2 5 , these calculations are n needed. 5 C. B. Williams, Journ. Amer. Chem. Soc., 17. 924, 1895 ; A. Emmerlmg, Chem. News, 57. 15, 1888 ; Zeit. anal. Chem., 26. 244, 1887 ; F. Hundeshagen, ib., 28. 141, 1889 ; L. Fncke, Stahl Eisen, 26. 279, 1906 ; J. M. Krasser, Zeit. Nahr. Gennss., 21. 198, 1911 ; L. T. Bowser, Amer. Chem. Journ., 45. 230, 1911 ; F Hundeshagen, Zeit. offent. Chem., 17. 283 302, 322, 1911 ; M. Wagenaar, Pharm. H'eekblad, 48. 845, 1911 ; J. G. Fairchild, Journ. Washington Acad. Science, 2. 114, 191'2. 6 F. Osmond, Bull. Soc. Chim. (3), 47. 745, 1887 ; Chem. News, 56. 160, li 600 A TREATISE ON CHEMICAL ANALYSIS. asbestos pad, in a Gooch's crucible, in a hydrochloric acid solution of stannic oxide. 1 The filtered solution is made up to 100 c.c., and the intensity of the colour is compared with that of a solution containing a known amount of the phosphomolybdate. For certain purposes, a sufficiently close approximation to the amount of the precipitated ammonium phosphomolybdate can be obtained by measuring the volume of the precipitate collected in the graduated test tube of a centrifugal machine. 2 313. The Volumetric Determination of Phosphorus- Uranium Process. Uranium acetate or nitrate reacts with phosphates in aqueous solution form- ing uranium phosphate. When all the phosphorus has been transformed to uranium phosphate, any further addition of the uranium salt will give a solution which furnishes a reddish-brown coloration when brought in contact with a solution of potassium ferrocyanide. These facts are utilised in a volumetric process. The reaction takes place somewhat slowly in the cold, and, in conse- quence, the titration is generally finished with hot solutions about 90. The reaction is supposed to be slower in the presence of acetates, and hence the acetates should be kept as low as possible. But even when uranium nitrate is used for the titration, there is a considerable amount of acetate present. The amount of acetate added with uranium acetate is but a small fraction of the whole. Consequently, it makes little difference whether uranium nitrate or acetate is used in the titration. 3 When the uranium solution is to be employed for titrating calcium phosphates, bone ash, etc., it is generally standardised against a solution of calcium phosphate of known strength. The Standard Solutions. Make a standard solution of pure calcium phosphate Ca 3 (P0 4 ) 2 by dissolving 10 grms. of normal calcium phosphate in a little nitric acid, and dilute the solution to 1 litre. 4 Then dissolve 40 grms. of uranium nitrate or 35 grms. of uranium acetate in about 800 c.c. of water. 1 Dissolve 10 grms. of "tin crystals" in 80 c.c. of pure hydrochloric acid, and make the solution up to a litre. 2 V. Eggertz, Journ. prdkt. Chem. (1), 79. 496, 1860 ; H. Wedding, Stahl Eisen, 7. 118, 1887; M. Ukena, ib., 7. 407, 1887; M. A. von Reis, ib., 9. 1025, 1889; 10. 1059, 1890; K. Bormann, Zeit. angew. Chem., 3. 638, 1889; C. Reinhardt, Chem. Ztg., 15. 410, 1891 ; H. von Juptner, Oester. Zeit. Berg. Hiitt., 43. 203, 1895; J. Ohly, Chem. News, 76. 200, 1897. 3 F. Sutton, Chem. News, i. 97, 122, 1860 ; H. Neubauer, Archiv Wiss. Heilkunde, 4. 228, 1893 ; C. Pincus, Journ. praJct. Chem. (1), 76. 104, 1859 ; R. Arendt and W. Knop, ib. (1), 69. 401, 415, 1856; K. Broockmann, Sep. anal. Chem., I. 212, 1881; A. E. Haswell, ib., 2. 251, 1882 (the more uranium solution employed, the higher becomes its relative value) ; C. Mohr, Zeit. anal. Chem., 19. 150, 1880; Chem. News, 45. 248, 1882; V. Edwards, ib., 59. 159, 1889 ; G. Guerin, ib., 8. 175, 1882 ; Hep. anal. Chem., 3. 157, 1883 ; 0. Abesser, W. Jani, and M. Marcker, Zeit. anal. Chem., 12. 239, 1873; E. Kessel, ib., 8. 164, 1869; C. Leconte, Compt. Rend., 29. 55, 1849; W. Knop, Chem. Centr. (2), I. 737, 769, 1856; H. Rheineck, Dingier s Journ., 2OO. 383, 1870 ; A. Pavec, Listy's Chemikt, i. 313, 1875 ; W. Streckerand P. Schiffer, Zeit. anal. Chem., so. 495, 1911 : A. Vozarik, Zeit. vhusiol. Chem., 76. 433, 1912. 4 The strength of the solution may be confirmed, if there be any doubt of the purity of the calcium phosphate, by precipitating according to the molybdate process (page 598). J. A. Muller (Bull. Soc. Chim. (3), 25. 1000, 1901 ; Chem. News, 85. 124, 1902) prefers crystallised dicalcium phosphate Ca 2 H 2 (P0 4 ) 2 . 4H 2 on account of its stability on exposure to the air, etc. If alkaline phosphates are to be titrated, potassium phosphate is used. For this purpose, dissolve 9 '585 grms. of monopotassium phosphate KH 2 P0 4 in a litre of water. The solution is standardised by evaporating a known volume of the solution to dryness in a platinum dish. Ignite the residue at the full heat of a Bunsen flame, and weigh as KPO> Or the phosphorus in a known volume may be determined by the magnesium process (page 597). THE DETERMINATION OF PHOSPHORUS. 6oi Add a few drops of ammonia, when a slight turbidity will be produced. Then add just enough acetic acid to clear the solution, and make up to a litre. 1 Standardisation of the Uranium Solution. Take 50 c.c. of the calcium phosphate solution, and add 10 c.c. of ammonium acetate, 2 and run the uranium acetate from a burette until a drop of the clear solution gives the first sign of a permanent reddish-brown coloration with a little powdered (solid) potassium ferrocyanide 3 on a white plate. 4 The solution must be well agitated after each addition of the uranium solution. Then heat the solution to boiling, 5 when a drop will no longer react with the ferrocyanide. 6 Add more uranium solution until the brown colour with the ferrocyanide is again developed. The result of this titration enables the strength of the uranium solution to be re- presented in terms of the calcium phosphate. Evaluation of Bone Ash. Supposing that bone ash is to be analysed, dissolve between 3 and 4 grms. of the bone ash in dilute nitric acid (10 per cent.). 7 Add ammonia to the filtrate and washings until a permanent precipitate is obtained. Redissolve the precipitate 8 in a little acetic acid, add 10 c.c. of ammonium acetate, and titrate the solution (occupying, say, 500 c.c.) "with the uranium solution, as indicated above. Calculation. Suppose that 50 c.c. of standard phosphate contain 0'5 grm. of Ca 3 (P0 4 ) 2 , and n c.c. of the uranium solution are needed for the titration, obviously/ 1 c.c. of the uranium solution represents 0'6/n of Ca g (P0 4 ) 2 . If w 1 The uranium salt should be free from uranium phosphate or ferric nitrate. Let the solu- tion stand some days before it is standardised, otherwise the turbidity which sometimes develops, owing to the separation of uranium phosphate, may alter the strength of the solution. The less free acetic acid the better, since the solution is more sensitive in the absence of an excess of free acetic acid. The strength of the solution should be checked every three or four days. 2 AMMONIUM ACKTATE SOLUTION. Dissolve 154 grms. ammonium acetate in water, add 100 c.c. of acetic acid (sp. gr. r04), and dilute the solution to a litre (2E). The object of the ammonium acetate is to prevent the formation of free nitric acid. 3 Or a drop of a solution of 1 1 grms. of the same salt potassium ferrocyanide in 10 c.c. of water (E). 4 For the method of titration with a "spot" indicator, see page 334. If an excess of uranium solution be added, a known amount of the standard phosphate solution may be added, and the uranium titration continued. An allowance may be made for the phosphate added. In shaking, avoid making the solution froth, since a drop of uranium falling on the froth does not readily mix with the liquid underneath. A drop of alcohol will generally dispel the froth. 5 If the solution be heated before the greater part of the phosphoric acid has combined with the uranium, some calcium phosphate may be precipitated (R. Fresenius, C. Neubauer, and E. Luck, Zeit. anal. Chem., 10. 133, 1871 ; C. Schumann, ib., II. 382, 1872). 6 If a little uranium phosphate be removed with the rod in making the spot test, the ferro- cyanide may be " browned " before the reaction is at an end. If the "spot " test gives a brown, add another four drops of the uranium solution. If the brown does not appear the moment the glass rod is removed, the first "browning " was not the end of the titration. This supple- mentary "four-drop test" should always be made. C. Mallot (Archiv Pharm., 2. 246, 1887) uses, as indicator, an aqueous ammoniacal solution of cochineal just decolorised with nitric acid. Titrate with this solution as an internal indicator until a permanent green is obtained E. Starkenstein, Biochem. Zeit., 32. 235, 1911. 7 The insoluble earthy and sandy matters maybe filtered off in a Gooch's crucible, dried, and weighed. 8 If iron and alumina be present, phosphates of these elements will be precipitated, and these remain insoluble in the acetic acid. In that case, filter the turbid solution. Titrate the filtrate as indicated in the text. Ignite and weigh the precipitate. If less than O'Ol gnu. of the mixed phosphates be present, take half the weight as P 2 5 and add the result to the P 2 r , obtained by the titration. If over this amount of precipitate be present, the process is unreliable, or the P 2 5 may be determined in the precipitate by the molybdate process (V. Edwards, Chem. News, 73. 25, 71, 1896 ; G. H. Allibon, ib., 73. 47, 94, 1896). There is then no particular advantage in the volumetric process over, say, Woy's process. Other phosphates e.g. cerium phosphate may be precipitated if they be present E. Damour and H. St C. Deville, Instit., 69, 1858; Compt. Rend., 59. 272, 1864; J. Boussingault, Chim. Phijs. (4), 22. 457, 1871. 6O2 A TREATISE ON CHEMICAL ANALYSIS. grms. of the bone ash be employed, and this requires m c.c. of the uranium solution, obviously the w grms. of bone ash contain O'bm/n grms. of Ca 3 (P0 4 ) 2 , or '' 5m per cent, of Ca 3 (P0 4 ) 2 . 100- wn Errors. It will be noticed that a slight excess of uranium is required to produce a coloration with potassium ferrocyanide, since enough uranium must be added not only to precipitate the phosphoric oxide, but also to colour the ferrocyanide. This excess must be allowed for by working under constant conditions, and making a blank titration on a solution prepared as indicated above, for standardising the uranium, but excluding the phosphate, and always using the same volume of solution in the titration. The most important sources of error in this titration are : (1) Mistaking transient for permanent "browning" make sure, by means of the "four-drop test," that the red tint is not an illusory end point ; (2) Error due to the removal of too small a drop to colour the ferrocyanide ; (3) Error due to titrating too quickly and overreaching the end of the reaction ; and (4) Error due to the reten- tion of uranium solution by the froth, and touching the ferrocyanide with tfyis. The analysis should be made in duplicate, and no number should be accepted if it deviates appreciably from the others. It may be again emphasised that, while the agreement between the duplicates is a test of the accuracy of the work, it is not, as is sometimes supposed, an indication of the trustworthiness of the process. Some, expecting too much from this process, have given it up in disgust. Owing to the almost invariable presence of appreciable amounts of alumina and iron, it may be advisable to first separate the phosphorus by means of ammonium magnesium citrate l before titrating with uranium nitrate. If organic matter be present, it should be removed by a preliminary incineration, since organic matter interferes with the uranium reaction. Arsenic, if present, should be removed by first passing a current of sulphur dioxide through the solution ; boiling to remove the sulphur dioxide ; and precipitating the arsenic with hydrogen sulphide. 2 PURCHASING BONE ASH. According to the " English Form of the Hamburg Contract," bone ash is invoiced as follows : If the delivered bone ash contains BELOW 70 per cent. TRIBASIC PHOSPHATE OF LIME, 3 an allowance shall be made to the Buyers according to the following principle and scale : For Ash below 70 % for every 1 % less, 1^ % phosphate less to be invoiced. ,. 65 % further 1 % 2 % " 60 % 1 % 3 % ,, ,, For fractions, the allowance to be made on the same scale, for instance : 69 % invoiced as 68^ / 65 / invoiced as 62| n / 61 % invoiced as 54^ % 68-^ 67| 64^ > 61o- 60i 531 68" , 67 64 80} 60 521 671 > 66i 631 501 59| 51 67 , 65} 63 58f 59 494 661 . 64} 62^ 57| 58^ 48 66 , 64 62" 5 ^2 58 46i 651 M 63 611 551 ., and so on 1 H. Joulie, Ann. Agronom., II. 97, 1885 ; Chem. News, 52. 85, 1885 ; J. M. H. Munro, ib., 52. 86, 1885 ; L. Joulin, ib., 27. 228, 309, 314, 1873 ; C. Mene, Gom.pt. Rend, 76. 1419, 1873. 2 RECOVERY OF URANIUM RESIDUES. W. Jani, Chem. Centr. (2), 2. 231, 1871 ; W. Knop, ib. (2), 10. 161, 1865 ; F. Strohmer, Zeit. anal Chem., 17. 84, 1878 ; A. Gawalovski, ib., 15. 292, 1876 ; E. Reichardt, ib., 8. 116, 1869 ; 13. 310, 1874. 3 That is, normal calcium phosphate Ca 3 (P0 4 ) 2 . THE DETERMINATION OF PHOSPHORUS. 603 EXAMPLE. Suppose 5 grms. of bone ash are taken for analysis. The solution is made up to 250 c.c., and 50 c.c. taken for a titratioii. Each 50 c.c. represents w = 1 grm. of bone ash. Suppose m = 26'6 c.c. of the uranium solution are needed for the titiation of this 50 c.c. of solution, and that ?i = 20 c.c. of the uranium solution are needed for titrating 50 c.c. of the standard phosphate. Then, from the above formula, 100x0-5x26-6 lx2Q =661 per cent. Ca.,(P0 4 ) 2 . If the phosphate be sold at Is. 6d. per unit, a ton of the bone ash on the Hamburg scale is not invoiced at 66|x 1| = 4, 19s. 9d., but at 64f x H = 4, 17s. l^d. Of course the colour, freedom from specks, and working qualities of bone for pottery purposes are of more importance than "per cent, of normal phosphate." 314. The Volumetric Determination of Phosphorus Joulie's Magnesium Citrate Process. Dissolve the substance under investigation in dilute hydrochloric acid, or in a mixture of nitric and hydrochloric acids if pyrites be present. If necessary, filter the solution. 1 Separation of Phosphorus from Alumina and Iron as Magnesium Phosphate. A portion of the clear solution, equivalent to 0'125 or O250 grm. of the phosphate, is pipetted into a beaker. Add 10 c.c. of magnesium citrate solution 2 and an excess of ammonia. The solution should be clear and only become turbid in a few seconds, especially after stirring. 3 Let the mixture stand under a bell-jar, to prevent loss of ammonia, overnight that is, about 12-15 hours. Filter, first by decantation of the clear, and wash the precipitate with dilute ammonia (10 vols. water, 1 vol. ammonia). Four to six washings will suffice. 4 Titration of -the Magnesium Phosphate with Uranium Nitrate. Dissolve the precipitate in a 10 per cent, solution of nitric acid. Allow the washings to run into the beaker in which the precipitation was made. Wash the filter paper with acidulated water. Put the filter paper in the beaker. Add dilute ammonia (1 : 10) until the solution is slightly turbid. Clear the turbidity with one or two drops of dilute nitric acid. Heat to boiling. Add 5 c.c, of ammonium acetate solution, 5 and titrate with the uranium acetate solution as described above. 315. The Colorimetric Determination of Phosphorus. Knight's 6 colorimetric process for the determination of phosphorus is based upon the intensity of the yellow colour of the phosphomolybdate in acid solutions. This colour test is very sensitive, since 0'000025 grm. of P 2 0. 5 can be recognised in cold solutions, and 0*0000025 grm. in hot solutions (80). As with colori- 1 The removal of silica by evaporation is only necessary when the silicates are decomposed by the acid with the separation of gelatinous silica. 2 JOULIE'S MAGNESIUM CITRATE SOLUTION. - Citric acid 400 grms. is dissolved in pure magnesium carbonate 40 grms., caustic magnesia 20 grms., distilled water 500 c.c. Make the solution strongly alkaline with ammonia 600 c.c. If the solution be turbid, filter, and make the clear solution up to a litre. 3 An immediate turbidity shows that insufficient citrate is present to hold all the iron and alumina in solution. In that case, start again and use 20 c.c. of the citrate solution. It is not sufficient to simply add more of the citrate solution to the turbid solution because the precipitate of the phosphates of iron and alumina, once formed, does not readily redissolv W. Simmermacher (Chem. Ztg., 37. 145, 1913) points out that the ammonium citrate should be added slowly, and a precipitate free from iron and silica can be obtained provided all the iron be in the form of citrate before the solution becomes alkaline. 4 The wash-water then gives no precipitate with sodium phosphate. 5 SODIUM ACETATE SOLUTION. A mixture of crystalline sodium acetate 1 glacial acetic acid 50 c.c. is made up to a litre with distilled water. The object of the sodium acetate solution is to prevent the development of free nitric or hydrochloric acid in the solution. , 6 J. W. Knight, Analyst, 5. 195, 1882. 604 A TREATISE ON CHEMICAL ANALYSIS. metric processes generally, the colour is influenced by the composition of the solution in which the determination is made. 1 The principal difficulty is due to the influence of a small quantity of silica, which produces a potassium silico- molybdate with a yellow colour almost identical in tint with that of the potassium phosphomolybdate. It is therefore important to use reagents free from silica. The reagents must not be kept in glass vessels, owing to possible contamination by silica from the glass. The sokitions should be kept in paraffin or cerasine bottles, or glass bottles coated internally with cerasine or paraffin. The absence of soluble silica in the reagents should be established by a blank test. A standard solution of phosphate is made by dissolving 0*2453 grm. of pure, freshly crystallised monopotassium 2 phosphate KH 2 P0 4 in silica-free water and diluting the solution to a litre. Preserve the solution in a cerasine or paraffin bottle. Each cubic centimetre corresponds with O'OOOl gnu. P 2 5 . Standard Solution. Dilute 10 c.c. of the standard potassium phosphate with, say, 80 c.c. of silica-free water. Add 10 c.c. of nitric acid (sp. gr. 1'07), and immediately add 8 c.c. of potassium molybdate solution. 3 Make the solution up to 1000 c.c. 4 Test Solution. Acidify a convenient portion say, 25 c.c. of the solution under investigation with 5 c.c. of nitric acid (sp. gr. 1'07), add 4 c c. of potassium molybdate solution, and make the solution up to, say, 50 c.c. The Comparison. Make the two solutions up immediately, and let them stand 20 minutes in order that the yellow tint of the phosphomolybdate may develop its maximum intensity. The intensities of the colours of the two solutions are now compared in a suitable colorimeter. 5 If the colour of the test solution be too strong for comparison, an aliquot portion is diluted in the usual manner. It is important to make sure that sufficient potassium molybdate is present to react with all the phosphate, by adding a further portion of the molybdate and diluting at the same time. Let the solution contain 5 c.c. of nitric acid, and 4 c.c. of the potassium molybdate, per 50 c.c. of solution. For smaller amounts of phosphorus, use a standard phosphate solution half the strength indicated above. Errors. If the solution under investigation be turbid, it should be filtered through biscuit- ware (page 631), evaporated to dryness, and then taken up with 1 P. E. Alessandri, Pharm. Centr., 35. 170, 1865 ; C. Lepierre, Bull. Soc. Chim. (3), 15. 213, 1896 ; A. Jolles and F. Neurath, Monats. Chem., 19. 5, 1898 ; A. Jolles, Arch. Hygiene, 34. 22, 1899 ; A. Pagnoul, Ann. Agronom., 25. 549, 1899 ; 0. Schreiner, Journ. Amer. Chem. Soc., 26. 806, 1904; 25. 1056, 1903; 0. Schreiner and B. E. Brown, ib., 26. 1463, 1904; F. P. Veitch, ib., 25. 169, 1903 ; J. G. Smith, ib., 26. 897, 1904 ; T. E. Hewitt, ib., 27. 121, 1905 ; C. Estes, ib., 31. 247, 1909 ; P. Veitch, Chem. News, 89. 73, 89, 101, 1904. 2 Sodium phosphate is difficult to deal with owing to efflorescence. J. A. Miiller (Bull. Soc. Chim. (3), 25. 1000, 1901) proposes the more stable microcosmic salt NaHN"H 4 P0 4 . 4H 2 0. The crystalline CaHP0 4 . 2H 2 is quite stable in dry air or over phosphoric anhydride in a desic- cator. It is a valuable standardising agent. 3 POTASSIUM MOLYBDATE SOLUTION. Dissolve 8 grms. of the salt in 50 c.c. of water, and make the solution up to 100 c.c. with nitric acid (sp. gr. 1 2). According to Jolles and Neurath, ammonium molybdate is liable to become turbid, and is not so sensitive as sodium or potassium molybdates. The coloration develops its maximum intensity at 80 in a short time. The colour with potassium molybdate develops more quickly, and is sharper and purer in tone than with sodium molybdate. 4 N. Passerini (Gaz. Chim. ItaL, 41. i. 182, 1911) recommends adding 2 c.c. of a cold saturated solution of gallic acid to the test glasses before the other solutions are added ; J. Pouget and D. Chouchak (Bull. Soc. Chim. (4), 5. 104, 1909 ; (4), 9. 649, 1911) use strychnine sulphate with the phosphomolybdate. 5 Nessler's glasses may be used in place of the colorimeter. One Nessler's glass is filled up to the 20-c.c. mark or the 50-c.c. mark with the coloured solution under investigation. At the same time, other Nessler's glasses are filled to the same volume with solutions containing different amounts of the standard phosphate solution and the molybdate solution. By comparing the tint of the standard with the tint of the unknown solution, the amount of phosphoric acid in the solution can be determined. THE DETERMINATION OF PHOSPHORUS. 605 water and again filtered. Organic matter, if present, may produce a coloured solution. In that case, the organic matter must be destroyed or a correction must be made. Veitch destroys organic matter by ignition with magnesium nitrate and subsequent evaporation with nitric acid. Large amounts of ammonium salts nitrate and chloride intensify the colour, and hence lead to high results. Aluminium sulphate A1 2 (S0 4 ) 3 gives a dark, more or less opaque solution which spoils the determination. If less than 1 part of iron per 20 million parts of the solution be present, no particular harm can be noticed, but if larger amounts be present, the colour is intensified, and the result is vitiated unless the amount of iron be known and an allowance made for its effect. If iron and alumina be present, therefore, the phosphorus should be separated as magnesium phosphate by Joulie's process before the phosphorus determination is made. 1 Calcium nitrate, barium chloride, potassium nitrate, separately and mixed together, give no appreciable error. Nitric acid and potassium molybdate require attention. The standard and the test solutions should have approxi- mately the same amounts of these two reagents, otherwise an appreciable error may be introduced. If the nitric acid solutions be passed through ordinary filter paper, there may be an error owing to the contamination of the solutions by phosphates and silica dissolved from the filter paper. 316. The Simultaneous Determination of Phosphorus and Silica. Silica, as indicated above, is the great enemy of the colorimetric process for phosphates, since it produces a yellow-coloured silicomolybdate which was used by Jolles and Neurath for the colorimetric determination of silica. Woodman and Cay van 2 found that the colour of the silicomolybdate required from 1J to 2| hours to develop its maximum intensity, and that the colour is about one- tenth of the intensity of the corresponding phosphomolybdate. This furnishes a means of correcting the determination of phosphorus for the influence of silica, and also to determine the amount of silica in the phosphate solution. If nitric acid and potassium molybdate be added to the solution as quickly as possible, a certain intensity of colour is developed by the silica. If, however, the ammonium molybdate be added an hour before adding the nitric acid, the intensity of the colour due to the silica is but half (0'49) as great as before. Phosphates give the same intensity as before under both conditions. Let x denote the amount of P 2 5 present, and y the amount of silica present. The intensity of the colour produced by the silica is 1'25 times the intensity of the colour produced by the same amount of P 2 5 . Hence, x+I'25y will represent the joint effect of the silica and the phosphorus in terms of the phosphate standard, or . (1) The value of A is determined by the colorimetric process indicated above. If, in preparing the test solution indicated above, the addition of the nitric acid be deferred for one hour, and the comparison made 20 minutes afterwards, the colour produced by the silica will be but half its former value. Hence, 1 The magnesium phosphate is dissolved in nitric acid, etc. The magnesia precipitate (page 219) may also be evaluated by determining the amount of phosphorus it contains colori- metrically, and the corresponding amount of magnesia calculated in the usual manner. O. Schreiner and B. E. Brown, Journ. Amer. Chem. Soc., 26. 1463, 1904. 2 A. J. Woodman and L. L. Cayvan, Journ. Amer. Chem. Soc., 23. 96, 1901 ; 24. 735, 1902 ; A. G Woodman, ib.. 24. 735, 1902 ; F. P. Veitch, ib., 25. 160, 1901 ; 0. Schreiner, ib., 25. 1056, 1903 ; 26. 808. 1904 -,23. 96, 1901 ; 24. 755, 1902 ; A. T. Lincoln and P. Barker, ib., 26. 975, 1904. 606 A TREATISE ON CHEMICAL ANALYSIS. # + J(l-25y) will represent the joint effect of the silica and the phosphorus in terms of the standard, or . . . (2) Multiply this equation by 2, and subtract the result from equation (1), and we get the value of x, that is, Amount of P 2 5 = IB - A. Similarly, by subtracting one equation from the other, we get the value of y, or Amount of Si0 2 = l'6(A - B). In illustration, suppose that we find the value A for P0 4 in the solution prepared by adding the ammonium molybdate and the nitric acid simultaneously to be 27, and the value of B for P 2 5 in the solution when the nitric acid is added one hour after the ammonium molybdate to be 16. Obviously, from the above equations, Amount of Si0 2 = 16(27 - 16) = 17'6. Amount of P 2 5 =2x16-27 = 9. When the proportion of phosphorus to silica is small, Lincoln and Barker recommend the addition of, say, 5 c.c. of the standard phosphate solution to the test solution, and then conduct the process as indicated above. Due allowance is, of course, made for the phosphorus which has been added. From what has been said above, it will be obvious that the colorimetric process may be employed for the determination of small quantities of soluble silica when phosphorus is absent. 1 317. The Analysis of Bone China Bodies. The following numbers represent the ultimate percentage composition of a typical fired body, and show the kind of mixture now under investigation : Si0 2 A1 2 3 Fe 2 3 MgO CaO K 2 Na 2 P 2 5 38-0 16-0 1-2 I'O 22-0 2'5 T5 18'0 With unfired bodies, free carbon, carbon dioxide, and the loss on ignition may also have to be determined. The amount of phosphoric oxide may or may not be sufficient to combine with all the alumina and ferric oxide. If the phosphoric oxide be in excess, lime and magnesia will be precipitated as phosphates along with the iron and aluminium phosphates. It is therefore necessary to modify the scheme of analysis indicated for clays. Several processes have been devised for dealing with the problem. 2 The following method is 1 A. Jollesand F. Neurath, Zeit. angeiv. Chem., u. 315, 1898 ; R. Salvador! and G. Pellini, Gaz. Chim. ItaL, 30. i. 191, 1900. H. Hermann (Zeit. anal. Chem., 46. 318, 1907) recom- mends potassium tungstate in place of potassium molybdate as a qualitative test for colloidal (soluble) silica page 669. 2 The ACETATE PROCESS is based on the assumption that the iron and aluminium phosphates are precipitated from a slightly acetic acid solution containing ferric oxide, alumina, lime, and phosphoric oxide by means of ammonium acetate. There are several modifications. The precipitate is contaminated with calcium phosphate C. Glaser, Zeit. anal. Chem., 31. 383, 1892 ; W. Hess, Zeit. angew. Chem., 7. 679, 701, 1894 ; F. Wyatt, The Phosphates of America, New York, 150, 1892; R. T. Thomson, Journ. Soc. Chem. Ind., 5. 152, 1886; 15/868, 1896; Chem. News, 54. 152, 1886. In the so-called CAUSTIC ALKALI PROCESS, the bases other than aluminium hydroxide are separated while the alumina is held in solution by the addition of an excess of caustic soda L. Lasne, Bull. Soc. Chim,. (3), 15. 6, 118. 146, 1896; 0. von Griiber, Zeit. anal. Chem., 30. 9, 1891 ; Chem. News, 63. 146, 1891. In the OXALATE PROCESS, the solution remaining after the separation of the silica is mixed with sufficient tartaric acid to pre- vent the precipitation of alumina, etc., when ammonia is added to the solution. The lime is THE DETERMINATION OF PHOSPHORUS. 607 based on the assumption that the lime can be removed by precipitation as calcium sulphate in the presence of alcohol before the precipitation of aluminium and ferric phosphates by ammonia. 1 A great many of the details are carried out as described for clays, and there is no need to repeat full particulars. 1. Separation of the Silica. Start with about a gram of the powdered sub- stance 2 and digest it with aqua regia hydrochloric acid (sp. gr. 1'2) 10 c.c., nitric acid (sp. gr. 1-12) 5 c.c. on a water bath for about half an hour. Filter off the insoluble matter. Dry, and fuse the residue with sodium carbonate as indicated for clays (page 164). Take up the fused mass with dilute acid, mix the two solu- tions ; evaporate to dryness ; 3 take up with dilute hydrochloric acid ; filter ; wash ; again evaporate to dryness ; take up with acid ; filter ; wash until the washings are free from chlorides; ignite and weigh the separated silica (pages 168 and 169). 2. Separation of the Lime as Calcium Sulphate, Evaporate the filtrate to about 50 c.c. Mix the solution with 10 c.c. of dilute sulphuric acid (1 : 5) 4 and about 150 c.c. of 95 per cent, of alcohol. In about three hours, filter off the precipitated calcium sulphate and wash with 50 per cent, alcohol until about 10 drops of the filtrate, diluted with an equal volume of water, do not redden a few drops of a solution of methyl orange. The precipitate may be 5 ignited and weighed as calcium sulphate CaS0 4 , and the number so obtained, multiplied by 0-41186 or 0-4119, gives the equivalent of CaO ; or the calcium sulphate may be redissolved in hydrochloric acid and reprecipitated by the oxalate process. 3. Separation of Ferric Oxide, Alumina, and Phosphoric Oxide. Evaporate the then removed as calcium oxalate by the addition of ammonium oxalate. The nitrate is eva- porated to dryness in a platinum dish and the organic matter destroyed by heat. The residue is taken up with hydrochloric acid, and the alumina, etc., separated as usual L. Blum, Zeit. anal. Chem., 39. 152, 1900. H. ImmendorfF (Land. Ver. Stat., 34. 379, 1887) precipitates the lime as oxalate in a solution very slightly acidified with hydrochloric acid. 1 E. Glaser, Zeit. angew. Chem., 2. 636, 1889 ; R. Jones, ib., 4. 3, 1891 ; J. H. Vogel, ib., 4. 357, 1891 ; Chcm. Ztg., 15. 495, 1891 ; T. Meyer, ib., 14. 1730, 1890; H. H. B. Shepherd, Chem. Ncivs, 63. 251, 1891 ; W. H. Krug and K. P. McElroy, Journ. Amer. Chem. Soc., 17. 260, 1895 ; Zeit. anal. Chem., 30. 206, 1891 ; T. M. Chatard, Trans. Amer. Inst. Min. Eng., 21. 169, 1893; E. T. Teschemacher and J. D. Smith, Chem. News, 62. 84, 1890; N. Blattner and J. Brasseur, ib., 76. 150, 1897 ; H. Herzog, ib., 102. 25, 1910 ; Journ. Ind. Eng. Chem., i. 477, 1909 ; H. W. Wiley. Principles and Practice of Agricultural Analysis, Easton, Pa., 2. 224, 1908. 2 Loss ON IGNITION. For the volatilisation of phosphoric oxide during the calcination for loss on ignition, see E. Lautemann, Liebiy's Ann., 113. 240, 1800; S. Leavitt and J. A. Leclerc, Journ. Amer. Chem. Soc., 30. 391, 617, 1908. A sample calcined at low redness gave 214 per cent, of ash, and at bright redness, 2'08 per cent. Another sample gave respectively 2 '18 and 2'16 per cent. There was therefore no serious loss under the conditions of the experiment. G. Lechartier (Coinpt. Rend., 109. 727, 1890) found the loss of phosphorus and also of sulphur is not serious if the operation be properly done. For the loss of sulphur during the calcination, see also G. S. Fraps, Journ. Amer. Chem. Soc., 23. 199, 1901 ; E. Fleurent and L. Levi, Bull. Soc. Chim. (4), 9. 379, 1911. 3 If the residue has a dark brown colour, organic matter is present, and it is best destroyed by conducting the next evaporation in a platinum dish, adding a little sodium nitrite, and heating the dried residue with a naked flame till the mass fuses to a colourless, viscid liquid. The dish is kept covered to prevent loss by spurting. Digest the cold mass with hydrochloric acid, evaporate to dryness, etc. 4 Sulphates retard the complete precipitation of aluminium phosphate. According to F. P. Veitch (Journ. Amer. Chem. Soc., 21. 414, 1897), the equivalent of "1'25 grm. H 2 S0 4 prevents the complete precipitation of aluminium phosphate, and 275 grins of H 2 S0 4 gives a decided error." Hence, the amount of sulphuric acid should be kept as low as possible. 5 Sometimes, particularly if a great excess of alcohol has been added, the calcium sulphate is discoloured. Again, if much magnesia be present in the sample, magnesium sulphate will be precipitated with the calcium sulphate. In each case, the calcium sulphate should be purified. The precipitated calcium sulphate may be dissolved in hot dilute hydrochloric acid, the lime precipitated as oxalate (page 213), and the filtrate added to the main solution ; or the magnesia, if any be present, can be precipitated as ammonium magnesium phosphate in the usual manner (page 218). 608 A TREATISE ON CHEMICAL ANALYSIS. filtrates to drive off the alcohol, 1 and oxidise any iron in the filtrate by adding a drop of bromine or some other oxidising agent. Add ammonium chloride and ammonia in slight excess. Boil the solution to expel ammonia. 2 Add a little more ammonia, and boil again. While the solution is hot, filter off the pre- cipitated aluminium and ferric hydroxides and phosphates. 3 Wash with a hot (50-60) 5 per cent, solution of ammonium nitrate 4 until the washings are free from chlorine about twenty washings are needed. Ignite 5 the precipitate from the ammoniacal solution with the residue left after the treatment of the ignited silica with hydrofluoric acid (page 169). 6 Weigh. r Fuse the precipitate with potassium pyrosulphate. Cool, and take up the cake with water. Filter off the " extra silica " and proceed for the titanium as indicated on page 203, and the phosphorus by Joulie's process. 4. Separation of Magnesia and Phosphoric Oxide. The filtrate from the ammonia precipitate is evaporated to a small bulk, 8 and divided into two aliquot portions, and the magnesia determined in the one portion in the usual manner (page 218). If phosphoric acid be present in excess of that required for the formation of aluminium, titanium, and ferric phosphates, it can be determined as ammonium phosphomolybdate in the other portion (page 597). The phosphoric oxide (P. 2 5 ) here obtained is added to that separated from the ammonia precipitate, and the sum of the weights is multiplied by 2'1845. The product represents the corresponding amount of normal calcium phosphate Ca 3 (P0 4 ) 2 . The weight of the magnesia here obtained is added to that separated from the calcium sulphate precipitate. The method is summarised in the following scheme, starting from the mixed aqua regia solution and the hydrochloric acid solution from the sodium carbonate fusion : Evaporate to dryness for silica (page 167) i SiO 2 Add H 2 S0 4 and alcohol CaSO 4 (and some MgS0 4 ) Add ammonia, etc. A1 2 O 3 , Fe 2 O 3 , P 2 O 5 , etc. Divide into two portions MgO P 2 O 5 . 1 The alcohol may be recovered by connecting the flask containing the solution with a Liebig's condenser, and subsequently distilling the alcohol from soda lime. 2 The boiling is said to help to prevent the contamination of the ferric and aluminium phosphates by magnesia later on. 3 Also manganese or titanium phosphates, if titanium or manganese salts be present. In that event, the titanium and manganese can be determined in the phosphate, if worth while. 4 In washing with hot water, the precipitated iron and aluminium phosphates are partly hydrolysed, with the production of a soluble acid phosphate and an insoluble basic phosphate. Hence, the composition of the precipitate is determined by the 'method used in washing, and accordingly, the actual composition of the precipitate must be established in work making- a pretence to accuracy. 5 A prolonged blasting about 15 minutes is necessary to get a constant weight. 6 There is some danger of losing phosphoric oxide, if any be present, in evaporating the sulphuric acid to dryness, as described on page 607, and also during the pyrosulphate fusion H. Hose, Ausfiihrliches Handbuch der analytischen Chernie, Braunschweig, '2. 575, 1871. 7 Under some conditions, when the weight is very small, it can be assumed that the precipitate is a mixture of FeP0 4 and A1F0 4 , and half the weight of the precipitate is taken to represent " Fe 2 3 + A1 2 3 . " 8 If any flocculent matter should separate during the evaporation, it is filtered off, as indicated on page 218, and examined for magnesia. THEDETERM1NATION OF PHOSPHORUS. 609 The alkalies are determined on a separate sample. The results are good, but the washing of the precipitated calcium sulphate is very tedious, and, if the alcohol be not recovered, the expense is a serious item when many determinations are made. The following numbers represent three determinations with one (unfired) sample : Ca 3 (P0 4 ) 2 ..... 34-20 34 '26 34 '51 percent. . . . .12-32 12-29 12 '17 percent. It is perhaps advisable to again emphasise the fact that the precipitated aluminium and iron phosphates are progressively hydrolysed 1 by a prolonged washing. Twenty washings with a 5 per cent, solution of ammonium nitrate gave a .theoretical yield of aluminium phosphate, but fifty washings gave a slightly low result owing to the formation of a soluble phosphate and an insoluble basic phosphate. Experiments with known mixtures show that the method gives good results in the presence of calcium, magnesium, sodium, and potassium salts. If appreciable quantities of fluorides or sulphates be present, the results are poor, because the aluminium phosphate is imperfectly precipitated page 180. 1 Cold or hot alkaline acetates also dissolve aluminium phosphate slowly. 39 CHAPTER XLII. THE DETERMINATION OF SULPHUR. There is little chance of discovering any exact method for the determination of sulphates by precipitation as barium sulphate, which shall at the same time be generally applicable and require no corrections, and yet not owe its accuracy to a com- pensation of errors ; for the reason that some of the requisite conditions appear to be mutually incompatible. J. JOHNSTON AND L. H. ADAMS. 318. The Properties of Barium Sulphate. SULPHUR compounds in clays and silicates are transformed into soluble sul- phates by fusion with sodium carbonate under oxidising conditions. The acidified solution is treated with a soluble barium salt, and insoluble bariurn sulphate is precipitated. This is washed, dried, ignited in the usual manner, and weighed as BaS0 4 . Until comparatively recently, many books on analysis gave students the impression that the weight of the barium sulphate indicated the amount of sulphuric oxide in a given sample with unerring precision. It has long been certain that the results, unless definite precautions be taken, can only be approxi- mately correct, and may indeed be untrustworthy. The method, in fact, is peculiarly sensitive to modifications in the conditions of the experiment. In 1871, Glover l quoted the following analyses of seven different samples of pyrites by three different professional analysts, A, B, C : Table L XX. Test Analyses for Sulphur in Pyrites. Sample No. Per cent, of sulphur. 1. 2. 3. 4. 5. 6. 7. A . . B . C . 38-80 38-00 40-20 40-90 40-90 42-70 39-60 39-10 41-10 39-30 39-50 40-70 41-60 41-40 43-50 38-20 38-00 40-00 39-70 39-20 41-10 In some cases the difference amounts to nearly 2 per cent. Since this sort of thing is not very different from what occurs even to-day, it is necessary to examine the more important sources of error. It is generally stated that sulphuric acid will show one part of " barium " in between 80,000 and 250,000 parts of solution. According to Bottger, 2 '00000 15 gram- molecules of potassium sulphate per litre will give a sensible turbidity with 1 J. Glover. Chem. News, 23. 57, 1871. 2 F. Jackson, Journ. Amer. Chem. Soc., 25. 992, 1903 ; W. Bottger, Zeit. angew. Chem., 25. 1992, 1912. 610 THE DETERMINATION OF SULPHUR. barium chloride in about 6 hours, and '0000039 gram-molecules of barium chloride per litre will give a sensible turbidity with potassium sulphate in about 2 hours. 1. The Adsorption of Salts by the Precipitated Barium Sulphate. Turner 1 long ago pointed out that barium sulphate has a tendency to carry down other salts during its precipitation. Turner said (1829) : " The adhesion of potassium sulphate to the precipitate ensues even in a dilute solution ; and it is not prevented by the presence of other salts, such as potassium nitrate, ammonium chloride, or ammonium nitrate. The quantity of adhering salt is variable, depending apparently as well on the relative quantity of the two salts, and the strength of the solution, as on the manner and extent of edulcoration. I have known it to increase the weight of the barium sulphate by 1 per cent." A large number of observers 2 have placed similar facts on record. Salts of the alkalies and the alkaline earths, 3 silica, magnesium, cobalt, copper, iron, 4 and aluminium compounds, 5 etc., may be carried down 6 with the precipitate. In illustration, known quantities of barium chloride and sulphuric acid were mixed in the presence of the salts indicated in Table LXXI. ; the results were somewhat high under conditions where slightly low results would have been obtained in the absence of the salts. Table LXXI. Effect of Foreign Salts on the Precipitation of Barium Sulphate. BaCl 2 .2H 2 0, grm. taken. Salt added, 5 grm. BaS0 4 found. Error. 0-5046 0-5020 0-5013 5027 KG! NaCl KC10 3 0-4814 0-4931 0-4849 0-4907 -0-0004 + 0-0137 + 0-0061 + 0-0107 1 E. Turner, Phil. Trans., 119. 295, 1829 ; H. Rose, Pogg. Ann., 113. 627, 1861. 2 R. Fresenius, Zeit. anal. Chem., 8. 52, 1869 ; 9. 52, 1870 ; 19. 53, 1880 ; 16. 339, 1877 ; 30. 452, 1891; R. Fresenius and E. Hintz, ib. t 35. 170, 1896; G. Briigelmann. ib., 16. 19, 1877 ; E. Hintz and H. Weber, ib., 45. 31, 1906 ; H. Weber, ib., 45. 714, 1906 ; G. Lunge, ib., 19. 419, 1880 ; C. Meineke, ib., 38. 210, 1899 ; T. 0. Sloane, Journ. Amer. Chem. Soc., 3. 37, 1881 ; 0. W. Foulk, ib., 18. 793, 1896 ; J. Johnston and L. H. Adams, ib., 33. 829, 1911 ; L. Archbutt, Journ. Soc. Chem. Ind., g. 25, 1890; W. Smith, ib., I. 85, 1882; P. Jannasch, Journ. prakt. Chem. (2), 40. 233, 1889 ; P. Jannasch and T. W. Richards, ib. (2), 39. 321, 1889; A. Mitscherlich, ib. (1), 83. 456, 1861 ; E. Siegle, ib. (1), 69. 142, 1856 ; C. Diehl, ib. (1), 79. 430, 1860 ; R. Silberberger, Chem. Monats., 25. 220, 1904; A. Ziegeler, Chem. Centr. (3), 12. 555, 1881 ; W. Ostwald, Zeit. phys. Chem., 34. 495, 1900 ; G. Hulett, ib., 37. 385, 1901 ; 47. 357, 1904 ; F. W. Kiister and A. Thiel, Zeit. anorg. Chem., 19. 97, 1899 ; 22. 424, 1900; G. Hulett And L. H. Duschak, ib., 40. 196, 1904 ; 0. N. Heidenreich, ib., 20. 233, 1899 ; A. Fischer, ib., 42. 408, 1904 ; A. Thiel, ib., 36. 85, 1903 ; T. W. Richards, ib., I. 150, 187, 1892 ; Proc. Amer. Acad., 26. 258, 1891 ; T. W. Richards and H. G. Parker, ib., 31. 67, 1896 ; Zeit. anorg. Chem., 8. 413, 1895 ; 0. Herting, Zeii. angew. Chem., 12. 274, 1899 ; 0. Herting, Chem. Ztg., 23. 768, 1899; E. Ruppin, ib., 33. 17, 398, 1909; 34. 1201, 1910; J. F. Sacher, ib., 33. 28, 1909; O. Folin, Journ. Biol. Chem., i. 131, 1905; E. F. Teschemacher and J. D. Smith, Chem. News, 24. 61, 66, 171, 1871 ; N. Glendinning and A. Edgar, ib., 24. 140, 220, 1871; S. Wyrouboff. Bull. Soc. Chim. (3), 21. 1046, 1899; J. I. Phinney, Amer. J. Science (3), 45. 468, 1891 ;'W. S. Allen and H. B. Bishop, Internal. Cong. App. Chem., 8. i, 33, 1912. 3 According to F. Stolba (Dingier 's Journ., 168. 43, 1863), potassium sulphate is more easily removed by washing than sodium sulphate. 4 According to C. R. Gyzander (Chem. News, 93. 213, 1906), ferrous salts are not so liable to contaminate the precipitated barium sulphate as ferric salts ; hence, before precipitating the barium salt in the presence of iron, some recommend the addition of a reducing agent such as sodium thiosulphate or hydroxylamine hydrochloride. 5 E. A. Schneider, Zeit. phys. Chem., IO. 425, 1895 ; H. J. M. Creighton, Zeit. anorg. Chem., 63. 53, 1909. 6 Either by adsorption, or mechanically. 612 A TREATISE ON CHEMICAL ANALYSIS. Hence, some of the salt carried down with the precipitate is probably weighed with the barium sulphate. 1 Further experiments show that occlusion occurs in presence of all the common metallic salts, and of sulphates soluble in water ; and that the amount of this occlusion per gram-molecule is nearly the same for all the metals. Hence, all barium sulphate precipitates have a complex composition owing to the presence of occluded salts sodium nitrate, 2 ammonium salts, sodium, potassium, and ammonium chlorides, potassium and sodium sulphates, etc. Potassium salts are more likely to be occluded than sodium salts. Magnesium sulphate is scarcely occluded at all, and the error from alkaline and ammonium chlorides is very small, unless very small precipitates are in question, when this error becomes of increasing importance. In general, the greater the concentration of the alkaline salts in the mother liquid, the greater the amount of salt occluded by the precipitate. For instance, Johnston and Adams find that 350 c.c. of a solu- tion containing the equivalent of 2 grms. of barium sulphate, 2 c.c. of a 2 per cent, solution of hydrochloric acid, and Sodium chloride . . .0 5 10 20 30 50 70 100 grms. Occluded salt in precipitate . 8'2 17'4 19'0 24-4 31'4 38'4 47'6 45'8 mgrms. Barium chloride, the precipitating agent, may also be carried down with the precipitated barium sulphate about 0'15 per cent, is present in slowly formed precipitates. 3 Some of this may be removed by washing with dilute nitric or acetic acid, but another source of error may affect the work when the precipitate is subjected to acid treatment page 614. It is not always practicable to dissolve the precipitated barium sulphate and reprecipitate, as is frequently possible with most precipitates affected in a similar way. Mitscherlich 4 suggested purifying the precipitate by dissolving it in concentrated sulphuric acid, 5 and reprecipitating the barium sulphate by dilution with water. This process is by no means satisfactory, and Mar 6 obtained better results by evaporating the sulphuric acid solution of the sulphate to dryness, and washing the crystals so obtained on an asbestos pad in a Gooch's crucible with water. Ignite and weigh. The removal of silica by digesting the barium sulphate with hydrofluoric acid requires care, because barium sulphate is decomposed by the treatment. Thus Sleeper 7 obtained the following numbers on weighing 2-1563 grms. of barium sulphate after each successive evaporation with hydrofluoric acid : 2-1457 2-1295 2-1224 2-1041 2-1022 2-0800 2'0549 The barium sulphate, after the digestion, should be treated with a drop of sulphuric acid. As a matter of fact, Sleeper observed no contamination of the precipitated barium sulphate with silica when the former was precipitated from artificial mixtures of a soluble sulphate with sodium silicate. 1 Y. Kato and I. Noda, Mem. Coll. Sci. Eng. Kyoto, 2. 217, 1910 ; E. T. Allen and J. Johnston, Journ. Amer. Ghem. Soc., 32. 588, 1910 ; Journ. Ind. Eng. Chem., 2. 196, 1910. 2 This forms caustic soda on ignition. 3 A. Mitscherlich, Pogg. Ann., 55. 214, 1842 ; E. Siegle, Journ. prakt. Chem. (1), 69. 142, 1856. 4 A. Mitscherlich, Journ. prakt. Chem. (1), 83. 456, 1861 ; W. G. Mixter, Chem. News 27. 53, 1873 ; Amer. J. Science (3), 4. 90, 1872 ; F. H. Storer and A. H. Pearson, ib. (2), 48. 870, 1869. F. Stolba (Dingier* s Journ., 186. 43, 1863; Chem. News, 9. 133, 1864) proposes to purify the barium sulphate by washing it with a solution of copper acetate followed by hot water. 5 J. Nickles, Amer. J. Science (2), 39. 90, 1865 ; Chem. News, n. 125, 1865. 6 F. W. Mar, Chem. News, 63. 256, 1891 ; Amer. J. Science (3). 43. 525, 1890. 7 J. F. Sleeper, Chem. News, 69. 63, 1894. THE DETERMINATION OF SULPHUR. 613 The adsorbed salts are sometimes removed by dissolving the precipitate l in a platinum crucible or dish with 10-15 c.c. of concentrated sulphuric acid, and pouring the solution, in a thin stream, with vigorous stirring, into 350 c.c. of water. The solution is heated to facilitate the filtering of the purified barium sulphate. The soluble salts to a great extent remain in solution. Examples of two successive extractions of two different precipitates may be quoted : I. II. First extraction .... O'OlOl 0'0248 grm. sodium sulphate. Second extraction . . . . O'OOOS 0-0021 grm. sodium sulphate. The important thing is to avoid the mutual precipitation of salts from the mother liquid with the barium sulphate. The barium chloride solution should not be added suddenly, but rather in drops, gradually, with constant stirring. In illustration, the two following experiments may be cited to show the difference in the weight of the precipitate obtained when the barium chloride is added suddenly and in drops the theoretical amount of barium sulphate was 49 '33 per cent. : L II. In drops 49 '23 49 '30 Suddenly . . . . . 50*03 49 '90 These and other experiments show that the rapid addition of the precipitant gives a more impure precipitate than when the precipitant is added slowly. Indeed, if the precipitate which has been formed suddenly be fused with sodium carbonate, and, when cold, leached with water, and the clear solution of the cake be treated with silver nitrate as described on page 652, a precipitate of silver chloride is obtained corresponding, in amount, with the excess in the preceding table, cal- culated on the assumption that the excessive amount of barium sulphate is due to adsorbed barium chloride This demonstrates the adsorption of barium chloride by the precipitate formed under the conditions stated see page 72. If much hydrochloric acid be also present, it requires a greater excess of barium chloride to precipitate a given amount of sulphuric acid ; and conversely, 2 a greater excess of a soluble sulphate is required to precipitate a given amount of barium chloride. Hence, the risk of contamination of the precipitated sulphate with barium chloride is greater, the greater the amount of free hydrochloric acid present ; 3 but, within certain limits, the greater the acidity of the solution, the less the amount of salt occluded by the precipitated barium sulphate. Allen and Johnston consider that the best results are obtained when the " sulphate " solution contains 20 to 30 c.c. of 5N-hydrochloric acid per 500 c.c., 4 and the amount of "sulphuric acid" is such that 5 to 10 c.c. of normal barium chloride solution are needed for the precipitation. If barium is to be determined, the solution should contain about 0-5 per cent, hydrochloric acid before adding the precipitating agent, say ammonium sulphate (10 per cent, solution). A considerable excess of the latter is needed for complete precipitation. 1 E. Ruppin (Ghem. Ztg., 33. 17, 398, 1908; 34. 1201, 1910) boils the precipitate four times in acidulated water. J. F. Sacher, ib.,33. 218, 941, 1909 ; M. J. van't Kruys, Chem. IVeekblad, 6. 735, 1909. 1J F. W. Mar (Chem. News, 63. 256, 1891 ; Amer. J. Science (3), 41. 288, 1891 ; (3), 43. 525, 1890) and P. E. Browning (ib. (3), 45. 399, 1893) state that if a sufficient excess of sulphuric acid be present, the precipitation of barium as sulphate is not affected by the presence of up to 10 per cent, by volume of hydrochloric acid, nitric acid, or aqua regia. E. Murmann (Oestcr. Chem. Ztg., 13. 227, 1910) recommends a large excess of hydrochloric acid with a few drops of alcohol to facilitate the precipitation of the barium sulphate. 3 J. 0. Roos, Kungl. TeJcnis. Hogs. Mater ialpruf. , 34, 1896-1906 ; M. Huybrechts, Bull. Soc. Chim. Belg., 24. 177, 281, 1910. 4 0. Folin (Jour* Biochem., i. 131, 1906) prescribes, as safe limits, 1 to 4 c.c. of concen- trated hydrochloric acid per 150 c.c. of solution, that is, 4 '5 to 18 '5 c.c. of 20 per cent, acid per 350 c.c. 6 14 A TREATISE ON CHEMICAL ANALYSIS. 2. The Solubility of Barium Sulphate. Barium sulphate is but sparingly soluble in water, 1 since 100 c.c. of water at 10 dissolve 0*000197 grm. of barium sulphate, while 100 c.c. of water at 100 dissolve O'OOOSl grm. The freshly pre- cipitated sulphate is rather more soluble than the precipitate which has stood for some time in contact with the mother liquid. 2 Curiously enough, a precipitate of barium sulphate which has been filtered and washed immediately after precipitation is less easily freed from occluded salts by washing, and therefore contains more occluded salts than one which has been left to stand some time, say overnight. The difference is considerable when the amount of the salt in the solution is large. For instance, a solution containing the equivalent of 2 grms. of barium sulphate and 5 grms. of sodium sulphate furnished a precipitate containing : Occluded Na 2 S0 4 . 22'6 19'0 17'4 14'0 12'2 10'4 8'2 7'6 mgrms. Time standing . 15 min. 3 hrs. 18 hrs. 2 days 4 days 48 days 150 days 210 days. These experiments were made at 20. Experiments at 100 showed that rather less salt is occluded than at the lower temperature. The solubility of the barium sulphate is augmented by acids. 3 Thus, 100 c.c. of a solution containing hydrochloric acid dissolved the following amounts of barium sulphate : HC1 . . . . .1-82 3-65 7-29 grms. BaS0 4 ..... 0-0067 0'0089 O'OlOl grm. and nitric acid : HN0 3 . . . 315 6-31 12'61 31-52 grms. BaS0 4 . . . 0-0070 0'0107 0"0170 0-0241 grm. Dilute sulphuric acid lowers the solubility, but barium sulphate is fairly soluble in the concentrated acid. Thus, 100 parts of sulphuric acid, sp. gr. 1'32, dissolve 5-69 parts of barium sulphate, and 100 parts of sulphuric acid, sp. gr. 1*9, dissolve 15'89 parts. 4 Again, 100 c.c. of 40 per cent, hydrobromic acid dissolve 0'04 grm. of barium sulphate ; and the same amount of 40 per cent, hydriodic acid - 0016 grm. 5 Hydrogen peroxide 6 and free chlorine 7 also increase the solubility of barium sulphate. The increased solubility of barium sulphate in the more concentrated acids shows that if, say, hydrochloric acid be present in great excess, the solution should be partly neutralised before the precipitation is made. Alkaline chlorides, barium nitrate, and potassium chlorate do not appreciably affect the solubility of barium sulphate ; but alkaline nitrates, 8 phosphates, and salts of the organic acids, 10 have a marked effect on the solubility. 1 F. C. Calvert, Chem. Gaz., 13. 55, 1856; A. F. Holleman, Zeit. phys. Chem., 12. 125, 1893 ; F. Kohlrausch and F. Rose, ib., 12. 234, 1893 ; F. Kohlrausch, ib., 50. 355, 1905. 2 Page 212. W. H. Wollaston, Phil. Trans., 103. 51, 1813; W. Ostwald, Zeit. phys. Chem., 34. 495, 1900 ; G. Hulett, ib., 37. 385, 1901 ; 47. 357, 1904. 3 G. S. Fraps, Amer. Chem. Journ., 27. 283, 1902 ; H. Rose, Pogg. Ann., 95. 108, 1855 ; R. Piria, II Cimento, 5. 257, 1847 ; E. Siegle, Journ. prakt. Chem. (1), 69. 142, 1856 ; E. C Nicholson and D. S. Price, Phil. Mag. (4), n. 169, 1856 ; H. M. Noad, Journ. Chem. Soc., 9. 15, 1856 ; W. Ostwald and W. Banthisch, Journ. prakt. Chem. (2), 29. 52, 1884 ; C. Gutkowsky, Ber., 5. 330, 1872. 4 H. Struve, Zeit. anal. Chem., 9. 34, 1870 ; C. C. Selleck, Pogg. Ann., 133. 137, 1868. 5 A. R. Haslam, Chem. News, 53. 87, 1886. 6 A. Gawalovski, Zeit. Oester. Apoth. Ver., 44. 258, 1900. 7 0. L. Erdmann, Journ. prakt. Chem. (1), 75. 214, 1858. 8 M. Mitten tzwey, Journ. prakt. Chem. (1), 75. 214, 1858 ; R. Fresenius, Zeit. anal. Chem., 9. 62, 1870 ; J. J. Berzelius, Ann. Chim. Phys., 14. 374, 1820. 9 Some barium phosphate is formed, and so remains in solution. T. Scheerer, Journ. prakt. Chem. (1), 75. 113, 1858. 10 For example, alkaline citrates. J. Spiller, Journ. Chem. Soc., IO. 110, 1858 ; Chem. News, 8. 280, 1863 ; 19. 166, 1869. THE DETERMINATION OF SULPHUR. 615 The solubility of the precipitate is increased by the presence of thorium, ferric, and aluminium salts, but not so much by magnesium salts. Thus, Grams per litre 1 2*5 5 10 25 50 100 Grams of barium sulphate dissolved per 100 c.c. Ferric chloride . . . 0*0058 0*0072 0-0115 0*0123 0-0150 0*0160 0-0170 Aluminium chloride . . 0-0033 0*0043 0*0060 0'0094 0-0116 0'0170 0*0175 Magnesium chloride . . 0*0030 0'0030 0*0033 0*0033 0*0050 0*0050 0*0050 Jannasch and Richards l attributed this effect to the formation of a double ferric barium sulphate, and add that, in a solution containing ferric chloride, "an accurate determination of sulphuric acid by direct precipitation with barium chloride is not practicable." Hence, in determining sulphur in pyrites, etc., the iron should be first removed, as Lunge recommends, by precipitation with ammonia before precipitating the barium sulphate. 2 An insignificant trace of sulphur may be lost owing to adsorption by the ammonia precipitate, or by the separation of a basic ferric sulphate with the ammonia precipitate. According to Lunge, 3 this need not be feared if the solution, after treatment with a moderate excess of ammonia, be heated to about 60 or 70 (not over) for about 10 minutes. The solution should still smell strongly of ammonia. The precipitate is filtered off, and the sulphate precipitated from the filtrate as described below. With the same object in view, Tread well 4 recommends supersaturating the solution with ammonia in the cold, and afterwards raising the temperature almost to boiling with constant stirring. The precipitate is filtered and washed until the washings give no turbidity with barium chloride after standing 5 minutes. 5 The presence of copper nitrate, too, gives low results owing to the solubility of barium sulphate in an aqueous solution of copper nitrate. Hence, adds Phillips, 6 the nitric acid should be replaced by hydrochloric acid by converting the copper nitrate to chloride. The same difficulty does not then occur. 3. The Reduction of the Sulphate to Sulphide. When barium sulphate is ignited in the presence of filter paper, organic matter, etc., "an appreciable amount of barium sulphate may be reduced to sulphide, and the amount so reduced may be twice as great in a covered crucible as when ignited in an open crucible. The precipitate should be moistened with sulphuric acid, and again ignited in order to transform the sulphide back to sulphate. 7 If a Gooch's crucible be employed, Ripper 8 says that the results are inaccurate, because the adhesion of the barium sulphate to the asbestos prevents the proper washing of the precipitate. This 1 G. S. Fraps, Amer. Chem. Journ., 27. 288, 1902 ; R. Fresenius, Zeit. anal. Chem., g. 62, 1870 ; P. Jannasch and T. W. Richards, Journ. prakt. Chem. (2), 39. 321, 1889 ; 40. 233, 1889. 2 M. Dennstedt and F. Hassler. Zeit. angew. Chem., 16. 1233, 1903 ; G. Lunge and R. Stierlin, ib., 18. 446, 1656, 1921, 1905 ; G. Lunge, ib. 16. 1081, 1903 ; Zeit. anal. Chem., 19. 414, 1880 ; Journ. Amer. Chem. Soc., 17. 69, 1895 ; T. S. Gladding, ib., 16. 398, 1894 ; 18. 446, 1896 ; N. J. Lane, ib., 18. 682, 1896 ; G. von Knorre, Chem. Ind., 28. 2, 1906 ; K. Jene, Chem. Ztg., 29. 362, 1905 ; H. Mennicke, ib., 29. 4951, 1905 ; B. N. Gottlieb, ib., 29. 668, 190.'). 3 G. Lunge, Zeit. anorg. Chem., 19. 454, 1899 ; F. W. Kiister and A. Thiel, ib., 19. 97, 1899; 22. 424, 1900; 0. Herting, Zeit. angew. Chem., 12. 274, 1899; 0. Hertmg, Chem. Ztg., 23. 768, 1899 ; 0. N. Heidenreich, Zeit. anorg. Chem., 20. 233, 1900; C. R. Gyzander, Chem. News, 93. 213, 1906. 4 F. P. Tread well, Kurzes Lehrbuch der analytischen Chemic, Leipzig, 2. 385, If 5 In doubtful cases the precipitate should be ignited and fused with sodium carbonate, and the aqueous extract of the mass tested to make sure that sulphates are absent. 6 H. J. Phillips, Chem. News, 62. 239, 1890. For thorium, see E. White, Ihonum a Compounds, London, 26, 191 2. ooncu 7 C. W. Marsh, Journ. Anal. App. Chem., 3. 164, 1889 ; Chem. News, 59. 309, 11 39 ; S. F. Acree, Journ. Biol. Chem., 2. 135, 1906 ; H, Pellet, Ann. Chim. Anal, 12. 186, 318, 1907 ; P ' ^M h Ri P per', M^O-m. * 36, 1892 ; F. W. Mar, Chem. ^vs, 63. 256, 1891 ; J. Science (3), 43. 525, 1890 ; J. J. Phinney, ib, (3), 45. 468, 1891, 6i6 A TREATISE ON CHEMICAL ANALYSIS. statement is opposed to the work of Mar and Phinney, who recommend the use of Gooch's crucible. With care, the Gooch's crucible does not furnish results less accurate than filter paper. 4. Losses by Volatilisation. Barium sulphate is not appreciably decomposed at temperatures likely to be obtained during the ignition over a blast gas flame. Doeltz and Mostowitsch 1 found that a sample of pure barium sulphate lost 9 per cent, in weight when heated between 10 and 20 minutes in a platinum crucible at 1580. This temperature is, however, much higher than is obtained in ordinary blasting, and if barium sulphate should be so decomposed, the original amount of barium sulphate can be restored by moistening with sulphuric acid and recalcination for a short time. If, however, much silica be present, appreciable decomposition may occur at much lower temperatures. Allen and Johnston 2 say that, if the barium sulphate be precipitated from solutions containing alkali sulphates, it " always occludes a certain amount of ' free ' sulphuric acid which is taken up as acid sulphate of the alkali metal." Alkaline chlorides increase the amount of this impurity, and it is greater with solutions containing potassium than sodium sulphate. The "free" sulphuric acid 3 is lost by volatilisation during ignition, whereas under ideal conditions all the sulphuric acid should have been converted into barium sulphate (in the determination of sulphur). If ammonium salts are present, some ammonium sulphate is precipitated with the barium sulphate ; and since all but a trace of the occluded ammonium sulphate is driven off during ignition, the total loss by volatilisation is the resultant of these two effects. For instance, by adding ammonium sulphate to 350 c. c. of a solution containing 20 per cent, of hydrochloric acid and the equivalent of 2 grms. of barium sulphate, it was found that with : Table LXXI1. Volatilisation Losses during Ignition of Barium Sulphate. A mmnrnnm Loss in grms. .ti.IIllIlUillU.IIl chloride. Due to Due to "free" ammonium sulphate. sulphuric acid. Total. grm. 0-0112 ... 0-0111 5 0-0244 0-0039 0-0283 5 0-0244 0-0055 0-0299 10 0-0259 0-0113 0-0372 10 0-0259 0-0096 0-0355 Hence, ammonium salts and a large excess of free acid should be avoided in solutions where the sulphur is to be determined. 5. Sulphur in the Glass of Beakers and Flasks. According to Bunge, 4 baryta water and lime water can act on glass, and a thin layer of barium sulphate is often deposited on glass in which such solutions have been used. Glass made from Glauber's salt (sodium sulphate), etc., usually contains sulphates. Hence the possibility of a slight error by the solvent action of reagents on the glass vessels employed in the work. 1 F. 0. Doeltz and W. Mostowitsch, Zeit. anorg. Chem., 54. 146, 1907 ; W. Mostowitsch. Metallurgic^ 6. 450, 1909 ; W. Schiitz, ib., 8. 228, 1910. 2 E. T. Allen and J. Johnston, Journ. Amer. Chem. Soc., 12. 588, 1910 : Journ. Ind. Enq. Chem., 2. 196, 1910. 3 It arises from the free acid in the original solution, and increases with it up to a certain point, 4 C. Bunge, Zeit. anal. Chem., 52. 15, 1913. THE DETERMINATION OF SULPHUR. 6l 7 319. The Determination of Sulphates in Clays and Insoluble Silicates. The Fusion. It is first necessary to get the silicate into solution. A gram of the dry material l is fused in a platinum crucible with five to six times its weight of sodium carbonate. If iron pyrites be present, about O25 grm. of sodium nitrite 2 should also be intimately mixed with the contents of the crucible. The mixture is heated very slowly, as described for the potassium pyrosulphate fusion (page 185). Preventing Contamination with Sulphur from Flame Gases. In order to protect the contents of the crucible from sulphur com- pounds derived from the flame gases, etc., Hillebrand 3 fits the crucible into a hole in an asbestos board ; and Gumming, into a vitreous "quartz plate." The hole should be large enough to expose from one-half to two-thirds of the crucible to the flame. The asbestos will riot stand the temperature many times, and if this material be used it is preferable to employ a disc of stiff platinum foil with a hole in the centre in conjunction with the asbestos board. The whole should be inclined at an angle, as illustrated in fig. 190, so as to deflect the products of combustion to the side, and prevent their corning in contact with the mouth of the crucible. 4 One of the many forms of blast lamp on the market burning methylated spirit may be used with advantage for the fusion. There is then no need to protect the crucible as just described. Barthel's 5 lamps are quite satisfactory. Fig. 191 illustrates one in use. These lamps may be obtained equivalent to one, two, or four Bunsen's burners. Electric crucible furnaces can also be used. The Leaching of the Fused Cake. When the fusion is completed, and the crucible has cooled, the cake must be removed as described on page 165. The cake is taken up with water, not acid. Filter through a small filter paper, say 1 If sulphur, chlorine, and fluorine are to be determined, take 2 grms. 2 Both nitrite and carbonate, and indeed all the reagents, should be tested to ensure the absence of sulphur compounds (page 461). 3 W. F. Hillebrand, Bull. U.S. Geol. Sur., 176. 196, 1900; A. C. Gumming, , Proc. Roy. Soc. Edin., 32. 17, 1912. The hole in the quartz plate can be bored on a lathe with a copper tube fed with carborundum. 4 CONTAMINATION WITH SULPHUR DURING IGNITION OVER GAS FLAMES. See D. S. Price, Journ. Chem..Soc., 17. 51, 1864 ; G. L. Ulex. Deut. Ind. Ztg., 379. 1870 ; Zeit. anal. Chem., 10. FIG. 190. Fusion for sulphur. ib.,.2$. 2200, 1886 ; G. Lunge, Journ. prakt. Chem. (2), 40. 239, 1889 ; J. van Leeuwen, Rec. Tram. Pays Bets, u. 103, 1892 ; J. W. Gunning, Chem. News, 17. 161, 1868 ; R. Woy, Zeit. offent. Chem., 8. 389, 1902. CONTAMINATION WITH SULPHUR DURING EVAPORATION OVER GAS FLAMES. See 0. Binder, Chem. Ztg., 16. 254, 1892 ; K. von Meyer, Journ. prakt. Chem. (2), 42. 267, 1892. On evaporating 2 litres of distilled water down to 50 c.c. in 6 hours on a water bath, Meyer found the water contaminated with sulphur equivalent to 0'0426 grm. of barium sulphate. 5 G. Barthel, Chem. Ztg., 16. 1106, 1892 ; Zeit. anal. Chem., 30. 596, 1891 ; 31. 67, 1892. Directions for use are supplied with the burner. 6i8 A TREATISE ON CHEMICAL ANALYSIS. 9 cm., and wash with a dilute solution of sodium carbonate. If chromium be present, the solution will be yellowish. The filtrate contains the alkaline sulphates, chlorides, and silicates; and the sodium salts of chromic, vanadic, 11 FIG. 191. Spirit burner for sulphur fusions. phosphoric, arsenic, molybdic, and tungstic acids (if present). The barium carbonate and sodium zirconate, thoria, etc., if present, remain undissolved on the filter paper. 1 The filtrate is treated with barium chloride as described below. 320. The Determination of Sulphur as Barium Sulphate. The solution 2 150-250 c.c. is acidified in a 500-c.c. beaker with about 5-10 c.c. of concentrated hydrochloric acid 3 and covered to prevent loss by spurt- ing. Boil to expel carbon dioxide. If the solution be still alkaline, add more hydro- chloric acid, but avoid a great excess. Wash down the sides of the beaker and the cover. The solution should be slightly acid, for the subsequent precipitation of the barium sulphate is not so complete if much hydrochloric acid be present, 4 1 For the subsequent treatment of the residue, see page 498. There is reason to suppose that a little sulphur sometimes remains behind with the residue e.g., with thoria because in test experiments less sulphur is found than is known to be present E. White, Thorium and its Compounds, London, 26, 1912. 2 If chlorine is to be determined, make the solution up to, say, 200 c.c., and set 100 c.c. aside for the chlorine determination (page 652). The volume of the solution should be such that a gram of barium sulphate is precipitated from 150 c.c. of mother liquid. The solution should contain 10 c.c. of concentrated hydrochloric acid per 150 c.c. of solution, 4 The solution should also be free from nitrates and nitric acid for the best results. THE DETERMINATION OF SULPHUR. 619 although a large excess of hydrochloric acid gives a precipitate comparatively easy to filter. Heat the solution to boiling, and add, gradually, 1 with constant stirring, say, 15 or 25 c.c. of a hot aqueous solution of barium chloride, 2 but avoid a great excess. 3 Boil for a few minutes. 4 If a drop of the supernatant liquid on a watch-glass gives a turbidity with a drop of barium chloride, add sufficient barium chloride, slowly, drop by drop, with constant stirring, to precipitate all the sulphates in solution. 5 If no turbidity occurs, let the solution stand 2 hours in a warm place, or, say, overnight. 6 Carefully decant the clear through a (say, 7 '5 cm.) filter paper, and wash four times by decanta- tion with hot water, 7 containing 1 cubic centimetre of hydrochloric acid per litre. 8 Transfer the precipitate to the filter paper and wash with hot water until a drop of the nitrate gives no turbidity with silver nitrate. The washing of the barium sulphate requires great care, because other salts are precipitated from the solution along with the barium sulphate. Ignite the wet precipitate in a platinum crucible. The paper is folded loosely over the precipitate to prevent spattering. The paper should be smoked very gradually, and the final temperature should not exceed dull redness. The crucible is then cooled and weighed. Evaporate with a few drops of hydrofluoric acid and a drop of sulphuric acid to expel any silica present. 9 Re-ignite in an inclined, uncovered crucible over the Bunsen flame. 10 Any sulphides formed through the reducing action of the carbon of the filter paper are soon re-oxidised, and there is usually no need to add a drop of nitric acid and of sulphuric acid. If the washing has been properly done, the ignited barium sulphate will be a granular white powder ; if otherwise, the powder may sinter to a more or less hard cake. The weight of the barium sulphate, multiplied by 0-13738, gives the corresponding amount 1 Added from a burette or pipette with a capillary tip, so that about 4 minutes is needed for running in 20 c.c. 2 BARIUM CHLORIDE SOLUTION. Dissolve 122'16 grms. of barium chloride (BaCl 2 .2H 2 0) in 1000 c.c. of water (E). Here 1 c.c. corresponds with 0'104 grm. of BaCl 2 . Note that commercial barium chloride sometimes contains sulphur compounds derived from the heavy spar used in making the chloride. The heavy spar is reduced to sulphide, and some soluble unoxidised sulphur compound may be retained by the crystals of the chloride J. Pattinson and J. T. Dunn, Journ. Soc. Chem. Ind., 24. 10, 1905. Note also that rubber stoppers sometimes contaminate solutions with sulphur compounds derived from the vulcanising agent antimony sulphide. It may also be added that barium chloride prepared as a by-product in the manufacture of hydrogen peroxide may contain an impurity which reduces potassium perman- ganate L. Blum, Zeit. anal. Chem., 29. 139, 1890. 3 This plan is of very great importance when a lot of sulphate is to be precipitated. The object is to prevent, as far as possible, the mechanical inclusion of barium chloride with the precipitate. 4 When the precipitation occurs at the boiling temperature both solutions hot the precipitate is more granular, settles more quickly, niters better, and washes easier than if the solutions be cold (page 96). 5 J. Johnston and L. H. Adams (Journ. Amer. Chem. Soc., 33. 829, 1911) recommend evaporation of the whole to dryness on a steam bath immediately after the precipitation ; extract the mass with hot water ; filter through paper ; wash until the washings are free from chlorides ; ignite very carefully over a Bunsen's burner to avoid reduction ; etc. _ The weight of the precipitate is corrected by dissolving an equivalent amount of pure potassium or sodium sulphate in a medium resembling the solution to be analysed ; and the sulphate determined as just indicated. The difference between the calculated and found results is used as a correction factor for the weight of the precipitate obtained in the actual analysis. In that case it is supposed that results can be obtained within 0'05 per cent, of the total sulphur actually present in the solution. 6 If a small amount of sulphur be present, the precipitate forms very slowly. 7 Another beaker may be used for collecting the washings, to prevent having to re-filter all the liquid should the washings commence to run through turbid. 8 It is well to test the^runnings with a drop or two of sulphuric acid to make sure the barium chloride was in excess. 9 If no silica is likely to be present, the ignition is conducted in a porcelain crucible. 10 Not the blast, or S0 3 may be lost C. W. Marsh, Journ. Anal. App. Chem., 3. 164, 1889. 62O A TREATISE ON CHEMICAL ANALYSIS. of sulphur ; the weight of the barium sulphate, multiplied by 0*3430, gives the corresponding amount of S0 3 ; and multiplied by 0'586, the corresponding amount of calcium sulphate ; or Table XCII. may be used. To illustrate the results which can be expected by this process, the following numbers represent the amounts of barium sulphate obtained in eight determina- tions with one-gram samples of the same clay : 0-0191; 0-0203; 0-0201; 0-0195; 0-0190; 0'0193; 0'0195; 0-0196. The mean is 0-0196; the maximum deviation, 0'0007 grm. that is, 0-07 per cent, on the total clay. 321. The Filtration of the Barium Sulphate Precipitate. Barium sulphate is a troublesome precipitate to filter, on account of its tendency to pass through the filter paper. A close-texture paper should there- fore be used. 1 If the precipitate should run through the filter paper, refiltration is necessary. Some use a double filter paper the two papers are folded separately, and one is placed inside the other ; others use a Gooch's crucible packed with ignited asbestos 2 this, however, does not allow the barium sulphate to be subsequently purified from silica. Another objection to Gooch's crucible turns on the fact that precipitates formed in feebly acid solutions are voluminous and inclined to clog the asbestos filter. Gooch's crucible is well adapted for small precipitates. Munroe's crucibles are good. P. W. Shinier 3 recommends a filter tube fitted to a filtration flask, as indicated in fig. 192. The wide part of filter tube D must be of uniform bore about 2 '5 cm. diameter, and 5 cm. long. The free end, D, is ground smooth, not melted smooth in the flame. A is a glass rod with a cross, B, at one end for supporting a disc of piano felt, C, approximately 0-75 cm. thick. The felt is cut by means of a cork-borer or wad-cutter very slightly larger than the bore of the filter tube. Fit the tube in a filter flask and apply gentle suction. Pour some filter-paper pulp (page 179) on to the felt to make a bed, F, about 0-6 cm. thick. Compact the pulp by the aid of a " stamper," S, made by fitting a glass rod into a hole bored half way through a rubber stopper at the narrow end. Wash the filter two or three times with water. Now run the contents of the beaker containing the barium sulphate precipitate through the filter tube and wash 1 Say, M. Dreverhoff s No. 311 or 400 ; or C. Schleicher and Schiill's No. 589, blue band. - L. L. de Koninck, Lehrbuch der chemischen Analyse, Berlin, I. 388, 1904. The use of organic substances (page 179) starch paste, etc. is inadvisable on account of their reducing action when the sulphate is ignited. H. N. Warren (Chem. News, 6l. 63, 1870) recommends the coagulation of the precipitate by the addition of a few drops of an ethereal solution of pyroxylin (page 180), but we have then to fight the tendency of the coagulating precipitate to mechanically enclose some mother liquid ; J. B. Krak (Chemist Analyst, 5. 26, 1912) removes supernatant liquid when the precipitate has settled, and agitates thoroughly with 10 c.c. of a slightly acid saturated solution of sodium acetate. This curdles the precipitate and enables it to be filtered by suction. 3 P. W. Shinier, Journ. Amer. Chem. Soc., 27. 287, 1905 ; H. P. Mason, Chem. Neios, 91. 180, 1904. FIG. 192. Shimer's filtration tubes. THE DETERMINATION OF SULPHUR. 621 as usual. Then, by means of the glass rod A, push the felt out of the tube, so that felt and pulp, etc., are transferred to a weighed crucible. Remove the felt by means of a pair of forceps, and ignite the contents of the crucible in the usual manner. Any precipitate sticking to the sides of the tube is carried forward with the pulp. It is therefore important to select a filter tube with uniform sides. The tube on the right of fig. 192 represents a tube rather more easily made than that just described. It requires very little additional descrip- tion. The rubber stopper G is of such a size that it can be easily pushed up the cylinder D. The results with this arrangement are a little high, owing to the difficulties in washing large precipitates on pulp and on asbestos filters. Such filters are very liable to clog. Porous earthenware and alundum ware crucibles have recently been recommended for the filtration. The crucibles are used as described, 324, page 630. When certain filtrates have passed through porous earthenware or alundum cups, a loss of weight occurs owing to the solvent action of the solution. Hence, the use of the cups is somewhat risky if the possibility of this error be ignored. 322. The Determination of Sulphur in Pyrites, Limestones, Coals, etc. The sulphur question is of some importance in certain branches of the clay industries, not only on account of the defects produced in glazes by fuels with abnormal amounts of sulphur; but the manufacture of some modern steels is influenced by the sulphur in the bricks, derived (1) by absorption from the fuel gases, and (2) from the sulphates or pyrites in the clays themselves. The methods for the quantitative determination of sulphur in coal, clays, silicates, etc., may be conveniently grouped under four headings. 1 The selection of the process for any particular problem is to be decided by the number of determina- tions to be made, the object of the determinations, and the nature of the samples under investigation. I. Dry Fusion with Alkalies followed by Treatment with Oxidising Agents. - Eschka's Method for Coal, Coke, etc. This process 2 depends upon the oxida- tion of the sulphur of the fuel by the oxygen of the air in the presence of magnesia, 3 and sodium carbonate ; 4 and on the transformation of the products of the oxidation into sulphates by boiling ammonium nitrate. Bromine water is now generally used for the oxidation. The sulphates are then precipitated by barium chloride in the usual manner. The results are good, and the process is in common use. Details of the operation are as follow : 1 L. Campredon, Dosage du Sou/re dans Us produits de la Siddrugie, Paris, 1896. 3 A. Eschka, Oester. Zeit. Berg, Hutt., 22. Ill, 1874 ; Zeit. anal. Chcm., 13. 344, 1874 ; L. Blum, ib., 27. 445, 1888 ; 0. Herting, Chem. Ztg., 23. 768, 1899 ; G. L Heath, Journ. Amer. Chem, Soc., 20. 630, 1898 ; 21. 1127, 1899 ; S. S. Sadtler, ib., 27. 1188, 1905 ; F. U. R. Stehmann, ib., 24. 644, 1902 ; A. C. Langmuir, ib. t 22. 99, 1900; C. W. Stoddart, ib., 24. 852, 1902 ; A. A. Blair, Journ. Ind. Eng. Chem., i. 689, 1909 ; F. M. Stanton and A. C. Fieldner, Tech. Paper U.S. Bur. Mines, 8. 7, 1912 ; I. 0. Allen and I. W. Robertson, ib., 26. 3, 1912. 3 W. C. EbaughandC. B. Sprague ( Journ. Amer. Chem. Soc., 29. 1475, 1907 ; Chem. News, 96. 240, 1907) use zinc oxide in place of magnesia. W. F. K. Stock (Chem. News, 30. 211, 1874), and F. C. Garrett and E. L. Lomax (Joum. Soc. Chem. Ind., 24. 1212, 1905), prefer calcium oxide to magnesia, because it is more easily obtained free from sulphur. 4 F. Hundeshagen (Chem. Ztg., 16. 1070, 1892 ; Chem. News, 66. 169, 1892) says that sodium carbonate is not so efficient as potassium carbonate in retaining sulphur. J. 0. Handy (Journ. Anal. App. Chem., 6. 611, 1892) considers that Hundeshagen's statement is of no significance in practice ; and this agrees with the general opinion of those who use the process. 622 A TREATISE ON CHEMICAL ANALYSIS. One gram 1 of the powdered 2 sample is intimately mixed in a 30-c.c. platinum 3 crucible with 3 grms. of Eschka's mixture, 4 and about 2 grms. more of this mixture is spread as a layer over the contents of the crucible to form a kind of cover. The open crucible is placed slant- wise on a triangle and heated with a sulphur-free flame page 617. The mixture must be heated very gradually to drive off the volatile matter without risk of losing sulphur. A small escape of sulphur dioxide can be detected by its odour. The heating should never be such as to blacken the covering layer of Eschka's mixture. In about half an hour the lower half of the crucible is red-hot, and the temperature is raised until the crucible is red-hot. It is an advantage to stir the mixture occasionally with a platinum wire. When all the black particles have been burnt a condition recognised by the grey colour of the mixture in the crucible changing to a yellow or yellowish-brown colour the crucible and contents are allowed to cool. Transfer the more or less pulverulent mixture to a 200-c.c. beaker, and digest it for* about 30 minutes with 75 c.c. of hot water. Filter the clear liquid into a 300-c.c. beaker, and wash the residue twice by decantation ; then transfer it to the filter paper, and wash it with water until the beaker contains about 200 c.c. of liquid. Add a slight excess of bromine water, about 4 c.c., and enough hydrochloric acid to make the solution slightly acid. Boil, and precipi- tate the sulphates with barium chloride as described on page 618. 5 II. Dry Fusion with Alkalies and Oxidising Agents. In these processes the sample is fused with sodium carbonate and potassium nitrate, or sodium nitrite (page 617), or ammonium nitrate, 6 and the residue taken up with water and evaporated with hydrochloric acid to expel the nitric acid before treatment with barium chloride. Fusion with sodium peroxide 7 offers many advantages ; the chief disadvantage is the violence of the reaction with coal. To overcome this difficulty, Parr ignites the mixture in a closed bomb of about 30 c.c. capacity, as described below : Parr's Process for Coal and Coke. Intimately mix 0'7 grm. of finely 1 If 0-687 grm. of the sample be employed, every 0*001 grm. of barium sulphate represents 0*02 per cent, of sulphur. 2 Large grains do not burn readily. If the powder passes a 60's lawn, it is generally satisfactory. 3 A porcelain crucible can be used in place of a platinum crucible, but the combustion is then rather slower particularly with coke. 4 ESCHKA'S MIXTURE. Two parts of light calcined magnesia are intimately mixed with one part of anhydrous sodium carbonate. Three grms. of Eschka's mixture are mixed with one gram of the sample, and the whole is covered with a 2-grm. layer of the mixture. Calcined magnesia of commerce often contains sulphates. If the amount is considerable, it can be re- moved by boiling with sodium carbonate. A correction can be made for the sulphur in the sodium carbonate and magnesia by means of a blank experiment. 5 H. Fresenius (Zeit. anal. Chem., 13. 346, 1874) remarks that the process gives the total sulphur. If calcium sulphate be present in the form of gypsum, the latter can be removed by boiling with sodium carbonate, etc. F. C. Calvert, Chem. News, 24. 76, 1871. 6 S. F. and H. E. Peckham, Journ. Soc. Chem. Ind., 16. 996, 1897 ; Journ. Amer. Chem. Soc., 21. 772, 1899 ; W. Koch and F. W. Upson, ib., 31. 1355, 1909 ; H. Schreiber, ib., 32. 977, 1910; A. C. Langmuir, ib., 22. 99, 1900; J. Lidow, Journ. Muss. Phys. Chem. Ges., 31. 567, 1899 ; S. Aufrecht, Pharm. Zeit., 41. 469, 1896 ; R. Dubois, Bull. Assoc. Chim. Belg., 15. 225 1901 ; J. D. van Leeuwen, Rec. Tram. Pays Bas, n. 103, 1892. 7 S. W. Parr, Journ. Amer. Chem. Soc., 22. 646, 1900 ; C. Sundstrom, ib., 25. 184, 1903 ; J. D. Pennock and D. A. Morton, ib., 25. 1265, 1903 ; T. W. Parr, W. F. Wheeler, and R Berolzheimer, Journ. Ind. Eng. Chem., i. 689, 1909 ; F. V. Konek, Zeit. angew. Chem., 22. 516. 1903 ; T. St Warunis, Chem. Ztg., 34. 1285, 1910. H. Schillbach (Zeit. angew. Chem., 16. 1080, 1903) uses barium peroxide. THE DETERMINATION OF SULPHUR. 623 powdered coke with 11 '5 to 13 grms. of powdered and dry sodium peroxide 1 by means of a spatula in a 30-c.c. nickle crucible, cover the crucible, and insert a 3-inch fuse - under the cover, and extending well into the mixture. Support the crucible on a triangle in a dish with a layer of water J inch deep. Ignite the fuse. In three or four minutes, when the mass has cooled, place the crucible and its cover in a small beaker. Add 30 c.c. of water. The mass dissolves in about two minutes. Rinse the crucible and cover; acidify the solution with hydrochloric acid ; boil ; add barium chloride in the usual way. With coal, an intimate mixture of 0*7 grm. of finely powdered coal with 13 to 16 grms. of peroxide is compressed by means of a press or vice in a steel bomb. 3 A piece of No. 36 iron wire, 4 inches long, is inserted, one end under the mica gasket of the bomb, and touching the bomb ; the other end is above the gasket and touching the cover. A current of 4 amps, applied to cover and bomb will fuse the iron wire and start the reaction. In two minutes cool the bomb in a little water. Unscrew the cover and proceed as indicated for coke. The test results by comparison with Eschka's process are quite satisfactory. III. Oxidation by Heating the Substance in a Current of Air or Oxygen, under Ordinary or under Reduced Pressure. Combustion Process for Pyrites in Clays, etc. The determination of sulphur combined as pyrites can be effected in several different ways. The " combustion process" is as follows: A hard glass tube, A (fig. 194) length 40 cm. is drawn out, and bent at right angles as shown at B in the diagram. One end of the glass tube is connected with a wash-bottle, (7, 4 containing dilute potash solution (approximately 3N-KOH), and the other fitted with a pair of Volhard's absorption tubes, 5 D and E, each containing about 50 c.c. of an aqueous solution of potash. 6 10-20 grms. of the finely powdered and dried 1 If too little peroxide is used, the reaction is explosively violent ; if too much, the combustion is incomplete. Imperfect mixing leads to explosive or incomplete combustion. The peroxide must be rapidly weighed and brushed from the watch-glass with a glass brush. 2 The fuse is made by nitrating cotton wick with a mixture of one part of fuming nitric acid (sp. gr. 1'5) and two parts of concentrated sulphuric acid (sp. gr. 1'84) for 12 hours at 15. Wash for 12 hours in running water to remove all the acid. Dry at ordinary temperatures. When dry, digest it in a cold, nearly saturated solution of nitre for an hour. Press out the excess of solution and dry. Cut into 3 -inch lengths for use. 3 The bomb, fig. 193, is a cylinder of steel, 1 in. internal diameter; 1J in. internal depth ; sides and bottom in. thick ; flange % in. around top ; cover plate T S T in. thick and If in. diam. The cover is clamped by a small clamp fitting under the flange of the bomb and pressing down the cover by a central screw. The cover is insulated from the bomb by a mica gasket, and from the screw of the clamp by ordinary red fibre. -Charge in Bomb FIG. 193. Parr's bomb. 4 S. Schiff, Chem. Ztg., 14. 233, 1890. Almost any other form will do quite well. 5 J. Volhard, Liebig's Ann., 176. 282, 1875. The two shown in the diagram are modifica- tions by H. Fresenius (Zeit. anal. Chem., 14. 333, 1875). M. Troilius' (M. Troilius, Notes on the Chemistry of Iron, New York, 37, 1886); or E. P. Penman's (Chem. News, 93. 213, 1906), or any of the numerous other varieties of bulbs, can be used in place of Volhard's. 6 POTASH SOLUTION : 16 '8 grms. of potassium hydroxide freed from sulphates by alcohol are dissolved in water, 1 c.c. of bromine is added, and the whole made up to 100 c.c. 624 A TREATISE ON CHEMICAL ANALYSIS. (110) clay is spread in a thin layer on the bottom of a long porcelain boat, which, in turn, is placed in the combustion tube at A. The end E is connected with an aspirator, and a current of air is drawn through the apparatus. The bubbles of gas should not pass through C more rapidly than admits of their being easily counted. The combustion tube is gradually heated to redness l in the vicinity of the boat containing the powder. Any deposit which forms in the tube in the vicinity of B should be driven forward by heating the tube with a Bunsen's burner, working the flame gradually from A to B. Any sulphur dioxide formed is oxidised by the bromine to sulphuric acid. 2 At the end of the combustion about 30 minutes the absorption vessel is removed and the contents are washed FIG. 194. Determination of sulphides. into a beaker. Any sulphates remaining in the tube AB are removed by drawing water up and down the narrow end by suction at F. The solution and washings are combined, and hydrochloric acid added to neutralise the free potash and to decompose the potassium hypobromite. The whole is boiled, and, if necessary, concentrated by evaporation. A hot aqueous solution of barium chloride is added, and the barium sulphate determined in the usual manner (page 618). The clay is supposed to be free from calcium and magnesium carbonates ; otherwise, calcium or magnesium sulphates may be formed, and the results will be correspondingly low. Water-soluble calcium and magnesium sulphates, how- ever, might be determined before and after the combustion, and an allowance made for this disturbing reaction. In some cases, say "volatile sulphur" in fuels, the combustion tube is packed towards the middle with about 5 cm. of platinised quartz or asbestos, instead of 1 A Ramsay's attachment for an ordinary Bunsen's burner is convenient for this purpose. W. Ramsay, Chem. News, 49. 2, 1883. 2 Hydrogen peroxide is sometimes used in placed of bromine, and the sulphuric acid may be determined by titrating the standard potash solution used in D, E before and after the com- bustion. The clay is then supposed to be free from animoniacal compounds, and the hydrogen peroxide from sulphur compounds. M. Fleischer, Protocoll Sitz. Zentral-Moorkomm., 20. 50, 1886 ; M. Berthelot and G. Andre, Ann. Chim. Phys. (6), 25. 302, 1892 ; M. van Bemmelen, Lands. Ver. Stat., 37. 284, 1890 ; P. Jannasch, Chem. Ztg., 14. 566, 1890 ; Zeit. anorg. Chem., 6. 303, 1894 ; Journ. prakt. Chem. (2), 40. 233, 1890 ; (2) 41. 566, 1890 ; P. Jannasch, Chem., 12. 32, 178, 1873 ; F. Muck, ib., 14. 16, 1875 ; W. Remmler, ib. t 4. 370, 1865 ; W. G. Mixter, Amer. J. Science (3), 4. 90, 1871. THE DETERMINATION OF SULPHUR. 625 with copper oxide as in the combustion tubes, page 569. The substance under investigation is placed in a boat on one side of the platinised asbestos, and potash bulbs are fitted at the other end in the usual manner. The tube is heated very gradually, and a current of oxygen is passed through the system. The platinised asbestos ensures the oxidation of the sulphur dioxide, and the potash solution absorbs the sulphuric acid fumes. 1 Combustion in the Bomb Calorimeter. 2 The combustion of 07 to I'O grm. of the sample in a bomb containing 10 c.c. of water with oxygen under a pressure of 30 to 40 atmospheres 3 is said to give results " more accurate and rapidly than all other methods." The apparatus is rather expensive, and hence is only used where a comparatively large number of determinations are made. There are several bomb calorimeters on the market, and specific directions for mani- pulating the bomb are supplied with the instrument. The treatment of the residue in the bomb is similar to that indicated above when dealing with Parr's process. 4 IV. Wet Treatment with Oxidising Agents under Ordinary or under High Pressures. The sample may be treated with various oxidising agents potassium dichro- mate and fuming hydrochloric acid; 5 hydrogen peroxide, 6 nitric acid and potassium nitrate, 7 nitric acid under pressure, 8 etc. These methods are usually less convenient than the dry fusion processes. Lunge's Wet Process 9 for Pyrites. Warm 0'5 grm. of the given pyrites with about 100 c.c. of a mixture of 3 vols. of nitric acid (sp. gr. 1*4) and 1 vol. of concentrated hydrochloric acid, with occasional stirring, taking care to avoid loss by spurting. If free sulphur should separate, add cautiously a little potassium chlorate in order to oxidise the sulphur to sulphuric acid. Evapo- rate the solution to dryness on a water bath, and repeat the evaporation with 5 c.c. of concentrated hydrochloric acid until nitrous fumes cease to be evolved. Add 1 c.c. concentrated hydrochloric acid to the residue, and, in a few minutes, 100 c.c. of hot water. Filter and wash with hot water. 10 Treat the 1 C. F. Mabery, Journ. Amer. Chem. Soc., 16. 544, 1894 ; W. E. Barlow and B. Tolleus, ib., 26. 341, 1904 ; \V. G. Mixter, Amer. Chem. Journ., 2. 396, 1880 ; G. Briigelmann, Zeit. anal. Chem., 15. 1, 1876 ; 16. 1, 1877 ; A. Saner, ib., 12. 32, 1873 ; C. M. Warren, Proc. Amer. Acad., 6. 472, 1871 ; I. Bay, Compt. Rend., 146. 333, 1908 ; 0. Dammer, Zeit. angew. Chem., 22. 440, 1909 ; 0. Brunck, ib., 18. 1560, 1905 ; M. Holliger, ib., 22. 436, 1909 ; I. C. Allen and I. W. Robertson, Tech. Paper U.S. Bur. Mines, 26. 8, 1912. 2 L_ C. Allen, W. A. Jacobs, and G. A. Burrell, Bull. U.S. Bur. Mines, 19. 8, 1911 ; I. C. Allen, Min. Scient. Press, Si. 569, 1900 ; R. Arnold, R Anderson, and I. C. Allen, Bull. U.S. Geol. Sur., 398. 265, 1910 ; I. C. Allen and I. W. Robertson, Tech. Paper U.S. Bur. Mines, 26. 10, 1912 ; G. A. Filliti, Bull. Soc. Chim. (4), 21. 338, 1899; N. W. Lord, Prof. Paper U.S. Geol. Sur., 48. 174, 1906 ; D. Lohmann, Chem. Ztg. t 35. 1119, 1911. 3 If a less pressure is used, there is a risk of incomplete combustion of the sulphur. 4 If the bomb has a lead washer, the washings from the bomb are boiled for about 10 minutes with a 5 per cent solution of sodium carbonate, and then filtered, etc. This treatment decomposes any lead sulphate which may be formed. 5 K. Charitschkoff, Petroleum Zeit., 2. 714, 1907. 6 E. V. Lecocqand H. Vandervoort, Bull. Soc. Chim. Belg., 16. 181, 1902. 7 F. W. Gill and H S. Grindley, Journ. Amer. Chem. Soc., 31. 52, 1909 ; A. Goetzl, Zeit. angew. Chem., 14. 1528, 1905 ; F. C. Calvert, Chem. News, 24. 76, 1871. 8 L. Carius, LieUg's Ann., 136. 129, 1865 ; Zeit. anal. Chem., 4. 451, 1865 ; G. Anelli, Rass. Min. Met. Chim., 34. 1, 1911 ; R. Holand, Chem. Ztg., 7. 99, 130, 1893. 9 G. Lunge, Zeit. anal. Chem., 20. 417, 1881 ; Zeit. angew. Chem., 2. 473, 1889 ; Fifth Int. Congress App. Chem., 399, 1906 ; E. Hintz and H. Weber, Zeit. anal. Chem., 45. 31, 1906. 10 The residue may be dried, ignited, and weighed as a mixture of silica, silicates, and sulphates of lead, barium, calcium, etc. With calcareous or magnesian clays, there is a danger of loss as indicated above. 40 626 A TREATISE ON CHEMICAL ANALYSIS. combined filtrate and washings with a moderate excess of ammonia; heat the solution to 60 or 70 for 10 to 15 minutes. The solution should then smell strongly of ammonia. Filter, and wash until the washings give no turbidity with barium chloride after standing for a few minutes. The combined filtrates should occupy about 250 or 300 c.c. 1 Acidify the solution with hydrochloric acid, avoiding an excess. Heat the solution to boiling ; remove the flame ; add, all at once, 2 barium chloride solution with constant stirring. 3 The addition of too great an excess of barium chloride leads to high results. Let the whole stand about 40 minutes. Decant the clear, filter, wash, etc., and weigh as barium sulphate. Sulphides decomposed by Acids. When the sulphides say sulphides of the alkalies or alkaline earths can be decomposed by boiling acids, 4 grind, say, 5 grms. of the powder with water to a thin slurry. Transfer the slurry to a (250-c.c.) gas developing flask (fig. 195) fitted with a stoppered funnel, and an exit tube so that any gas generated may be passed through, say, 10 c.c. of a solution of cadmium acetate (or ammonia- cal cadmium chloride). 5 Add hydro- chloric acid to the mixture in the flask, via the stoppered funnel. Warm the contents of the flask, and drive off the hydrogen sulphide by a current of air which enters through the stoppered funnel. Add 10 c.c. of iodine solution to the cadmium acetate solution, and titrate the liberated iodine with standard sodium thiosulphate, as indicated on page 358. There is a danger of losing sulphides when ferric salts are present. The hydrogen sulphide reduces the ferric salts and is itself oxidised. Hinrichsen 6 recommends adding, say, 10 c.c. of a 20 per cent, solution of stannous chloride to the mixture in the generating flask, in order to avoid the error. Qxidisable Matter in Limestones. The disintegration of limestones is supposed by some to be related with the amount of oxidisable matters, chiefly pyrites, which they contain/ and this is determined by the following process : Add FIG. 195. Determination of sulphides. 1 If the volume is greater, concentrate by evaporation. 2 See page 613. The adsorption error (page 611) is supposed to compensate the loss caused by the solubility of the barium sulphate in the mother liquid. ooJ^ioo^m 111 ? 11 ^^?'^' 3 ?; 835 ' 1907; F> L< Crobaugh, ,/OMrw. Anal. App. Chem. 7, 280, 1893 ; T. J. Morrell, Chem. News, 28. 229, 1873. 5 CADMIUM ACETATE SOLUTION. Dissolve 25 grms of cadmium acetate in 200 c. c. of glacial acetic acid, and make the solution up to a litre. F. Weil (Compt. Rend., 102. 1487 1886) used a rd ammoniacal solution of a copper salt. After passing hydrogen sulphide through this Q?O f WaS made UP * a 8 lven volume, and an aliquot portion titrated as indicated on 352 for unprecipitated copper. The difference between the result and the amount of copper ^"F y )r tt r , e P rese Jl ted the amo nt precipitated by the hydrogen sulphide. F. W. Hinrichsen, Mitt. k. Material-pruf. Ami., 25. 321 1895 A. M. Peters, Journ. Amer. Chem. Soc., 25. 143, 1903. ' THE DETERMINATION OF SULPHUR. 627 25 c.c. of y^N-potassium permanganate to 1 grm. of the dried (110) and powdered clay in a 250-c.c. Erlenmeyer's flask. When the powder is all moistened, add 100 c.c. of a 10'per cent, solution of sulphuric acid, 1 and shake the flask. If the colour of the potassium permanganate be discharged, add a second 25 c.c. of the permanganate solution. Heat the mixture on a water bath, with occasional shaking, for about half an hour. 2 Run 10 c.c. of y^N-oxalic into the flask. 3 Let the mixture stay on the water bath a few minutes longer. All the brown manganic compounds should have passed into solution. The excess of oxalic acid is titrated with standard potassium permanganate. Three one- gram samples should be used. The results are calculated to " oxygen consumed by the limestone." A resistant limestone according to Peters' experiments, required less than 0'4 grm. of oxygen, while limestone which weathered easily required over 0'6 grm. of oxygen. 4 One gram of potassium permanganate, it will be remembered, furnishes 0-395 gram of oxygen (page 193). 323. The Volumetric Determination of Sulphates Raschig's Process. Benzidine hydrochloride C 12 H 8 (NH 2 ) 2 .2HC1 in aqueous solution is hydro- lysed in such a way that the solution behaves as if it contained neutral benzidine C 12 H S (NH 2 ) 2 and hydrochloric acid. If such a solution be titrated with soda lye, the phenolphthalein will develop a pink coloration after all the acid is neutralised by the soda. The reaction is quantitative. Suppose, further, an excess of benzidine hydrochloride be added to a solution containing, say, sodium sulphate, sparingly soluble benzidine sulphate C 12 H 8 (NH 2 ) 2 will be precipitated: C 12 H 8 (NH 2 ) 2 . 2HC1 + Na 2 S0 4 = 2NaCl + C 12 H 8 (NH 2 ) 2 . H 2 S0 4 . The excess of benzidine hydrochloride can then be determined by titration with soda lye. The difference between the titre of the benzidine hydrochloride added to the sulphate solution, and that of the benzidine hydrochloride remaining after the precipitation of the benzidine sulphate, represents the " sulphate " in the given solution. This method, due to Miiller, presupposes that the sulphate solution is exactly neutral. Since aluminium, iron, zinc, and other sulphates are not neutral, these bases should be absent. Some of the benzidine hydro- chloride is also absorbed by the precipitated sulphate, and this leads to high results. The latter difficulty is less marked if the precipitation be effected in hot solutions. Instead of proceeding by Miiller's process, it is better to follow Raschig's variation, 5 and treat the precipitated benzidine sulphate with soda lye, using 1 110 c.c. of sulphuric acid (sp. gr. T84) and 1800 c.c. of water. 2 Experiments show this to be sufficient to oxidise the pyrites, etc. 3 More if necessary. The oxalic acid is adjusted so that at least 10 c.c. is present in excess of that which has been oxidised by the permanganate. If 25 c.c. of permanganate were added, 30 c.c. of oxalic acid will suffice, provided very little permanganate was reduced ; otherwise more oxalic acid will be required. 4 Phosphoric acid can be determined, if desired, in the solution by A. L. Emery's volumetric process (Journ. Amer. Chem. Soc., 24. 895, 1902) ; the insoluble residue is reported as "clay and sand." 5 F. Raschig, Zeit. ancjeiv. Chem., 16. 617, 118, 1903; 18. 331, 1905; 19. 331, 1906; C. Friedheim and 0. Nydegger, ib. , 20. 9, 1907 ; W. Miiller, ib., 16. 653, 1017, 1903 ; Ber., 35. 1587, 1902; W. Miiller and K. Diirkes, Zeit. anal. Chem., 42. 477, 1903; G. von Knorre, ib., 49. 461, 1910; Chem. Ind., 28. 2, 1905; 0. Nydegger, Ucber die JJestimmung der Schwefelsaure dwell Benzidin, Bern, 1907; 0. Huber, Chem. Ztg., 29. 1227, 1905; K. K. Jarvinen, Ann. Acad. Uci. Fennicce, 2. iv, xvi, 1912 ; A. Heczko, Zeit. anal. Chem., 50. 748, 1912 ; 51. 1, 1912 ; K. K. Jarvinen, Ann. Acad. Scient. Fennicce, 2. A, 4, 1910 ; 16, 1911. 628 A TREATISE ON CHEMICAL ANALYSIS. phenolphthalein as indicator. The end of the reaction is represented by the equation : Cj 2 H 8 (NH 2 ) 2 . H 2 S0 4 + 2NaOH = Na 2 S0 4 + 2H 2 + C 12 H 8 (NH 2 ) 2 . The Precipitation. Pour, say, 20 c.c. of the neutral or feebly acid solution of the given sulphate, with constant stirring, into 500 c.c. of a cold solution of benzidine hydrochloride l in a 600-c.c. Erlenmeyer's flask. Stir the mixture well with a glass rod. The voluminous crystalline precipitate of benzidine sulphate is allowed to settle for about 15 minutes, and then separated by suction through a Witt's filter plate (page 103) " 2 4 cm. diameter on the upper side, and 3 '5 cm. on the lower side. The supernatant clear liquid is poured through the funnel, and the precipitate transferred by shaking -the last fifth of the liquid before it is poured into the funnel. Any precipitate which sticks to the sides of the flasks in which the precipitation was made can be washed with some of the clear filtrate, or a special wash-bottle with benzidine solution can be used. Take care that no cracks appear in the precipitate as the last drop of mother liquid drains from the filter plate. Wash the precipitate twice with 5-10 c.c. of water 3 from a wash-bottle directed so as to wash down the sulphate adhering to the sides of the funnel. As the last of the wash-water drains off, disconnect the filtration flask from the pump. 4 This must be done before the precipitate contracts to a more or less dry silver-grey skin. 5 The Transfer of the Precipitate to the Titration Flask. The funnel is then inclined at an angle of about 45, and a glass rod inserted through the stem of the funnel so as to turn the filter plate and precipitate over on to the walls of the funnel, with the plate uppermost. The filter plate is removed, and the precipitate transferred to a 250-c.c. Erlenmeyer's flask by the aid of a glass rod. The filter paper is squeezed into a roll, and also dropped into the flask. The funnel is placed against the mouth of the flask, and any adhering benzidine sulphate is washed into the flask with a vigorous jet of water. In all, about 25 c.c. of water should be used for the transfer. If any particles of the precipitate remain on the funnel, it is best to wipe them off with a small filter- 1 BENZIDINE HYDROCHLORIDE. Rub 40 grms. of benzidine thoroughly with 40 c.c. of water, and wash the "slip " into a litre flask with about 750 c.c. of water. Add 50 c.c. of concentrated hydrochloric acid (sp. gr. 1'9), and when all is dissolved to a clear brown solution, fill the flask up to the litre mark with water. If the solution be turbid, filter. For use, this solution is diluted with twenty times its volume of water. 150 c.c. will suffice for the precipitation of '1 gnn. of H 2 S0 4 . Brown flecks may separate from the solution on long standing. These do no particular harm. Benzidine costs about 2s. 6d. per kilo. 2 The filter plate is covered with two discs of moistened filter paper. The discs of filter paper are 4 '6 cm in diameter. The projecting 3 -mm. rim of the filter paper discs is carefully pressed against the sides of the funnel by means of a glass rod with a square end. Successful work depends upon the tight closing of the joint between the funnel and the filter paper. 3 The benzidine sulphate is slightly soluble in water. By proceeding in this manner, the loss from this cause can be reduced to a minimum. 4 Make sure that all the sulphuric acid has been precipitated, by testing the filtrate with barium chloride solution. No precipitate should be produced, but after standing 15 minutes a slight turbidity will probably appear owing to the decomposition of the trace of benzidine sulphate in the filtrate. The loss from this source is very small, and it is compensated by the slight excess of the T VN-NaOH one or two drops, corresponding with 0'12 to 0'2 per cent of sulphur required for colouring the phenolphthalein absorbed by the pulp of the filter paper in the flask during the titration. 5 Dried benzidine sulphate cannot be properly distributed through the water, and, in consequence, it will be but slowly decomposed by the sodium hydroxide during the titration. The results will then be low. If by accident undecomposed flakes of benzidine sulphate should appear in the liquid, boiling with a small known excess of sodium hydroxide will effect decomposition. The excess of soda can be determined by back titration with standard acid. THE DETERMINATION OF SULPHUR. 629 paper swab, and transfer the latter to the flask. The benzidine sulphate should now be all in the flask. Close the flask with a rubber stopper and shake it thoroughly, so as to form a kind of slurry precipitate, filter paper, and water quite free from lumps of benzidine sulphate. Remove the stopper from the flask and wash back any adhering particles of the sulphate with a few drops of water. The Titration. Add about 2 c.c. of phenolphthalein, 1 and titrate the solution with ^N-sodium hydroxide until the red colour of the indicator appears. Warm the flask over a flame until the colour has disappeared, and continue the titration until a faint red appears. Again heat the flask until the liquid boils, so as to expel the carbon dioxide. The end of the titration is represented by the appearance of a faint red colour in the liquid. 2 At the end of the titration one or two drops of y^N-hydrochloric acid should remove the colour, and this should not reappear after boiling two minutes. EXAMPLE. 100 c.c. of a solution of sulphate were poured into 500 c.c. of the benzidine solution, and, after filtering, etc., the solution was washed with 20 c.c. of water. The benzidine sulphate slurry required 49*15 c.c. of j^N-NaOH to develop the proper coloration. Here, 49'15 c.c. of ^N-NaOH represent 0-2408 grm. of H.,S0 4 . As a matter of fact, the solution really contained 0-2430 grm. of H 2 S0 4 . Hence, the error 0'0022 grm. approached 0'9 per cent. But closer results can be obtained after a little practice in the method. Modification in the Presence of Iron Salts. Nydegger tried the effect of slightly acidifying the solution with hydrochloric, nitric, and acetic acids ; the effect of additions of potassium, ammonium, copper, aluminium, zinc, and chromium chlorides, potassium nitrate, sodium acetate, and ammonium, zinc, manganese, cobalt, copper, aluminium, and ferrous sulphates. The results were quite satis- factory. It is generally considered that no more iron should be present than is represented by Fe : S. If ferric iron be present, some will be occluded with the precipitated benzidine sulphate and spoil the result. The ferric iron should be reduced to the ferrous condition before adding the benzidine hydrochloride. The reduction is best made with hydroxylamine hydrochloride, as described below. Satisfactory results can then be obtained with both copper and iron pyrites. For pyrites, Raschig recommends the following process : 0'8 grm. of finely powdered pyrites is placed in a dry 200-c.c. Erlenmeyer's flask with 5 c.c. of fuming nitric acid. Heat the mixture on a water bath, with a funnel in the neck of the flask to prevent loss by spurting. The pyrites should be decomposed in about half an hour. Add 30 c.c. of water, and warm the mixture a short time to dissolve the iron salts. Dilute the solution with or without the removal of the solid residue to the 100-c.c. mark in a measuring flask. Pipette 20 c.c. of this solution into a 600-c.c. beaker, and add 10 c.c. of a one per cent, solution of hydroxylamine hydrochloride, and 500 c.c. of the benzidine solution. Stir the mixture with a glass rod, and let it stand 15 minutes. Then proceed as described above. 3 Errors. With a little practice, the results obtained by this method for sulphur in pyrites, etc., rival in accuracy results with the barium chloride process. The more important sources of error are: (1) the slight solubility of the benzidine sulphate in water ; (2) slight adsorption of benzidine hydrochloride by the precipitated benzidine sulphate; (3) imperfect titration due to the "balling" of the sulphate when it is overdried ; (4) imperfect precipitation of sulphate 1 This unusually large amount of indicator is needed because some of the indicator is absorbed by the paper pulp in the flask. 2 The tint can be recognised through the fibres of the filter paper, but, if necessary, the paper can be allowed to settle to facilitate the observation of the end coloration. An intense red colour shows over- titration. 3 The number of c.c. of ^N-NaOH used represents the per cent, of sulphur in the pyrites. The other four-fifths of the solution in the measuring flask can be used for check determinations. 630 A TREATISE ON CHEMICAL ANALYSIS. through using insufficient benzidine hydrochloride ; and (5) the need for a slight excess of the ^N-NaOH solution, owing to adsorption of the indicator by the paper pulp. Some of these errors compensate one another. The time needed for the determination of sulphur in pyrites by this method is something less than one and a half hours, but the method is time-saving only when a number of determinations under similar conditions have to be made. 324. The Determination of the Soluble Salts in Clays. The "soluble salts" in clays are responsible for certain manufacturing difficulties, and consequently it may be of great importance to determine their nature and amount. It is not sufficient to boil the clay with water and remove the soluble matters by nitration through filter paper, or by settling. 1 The clay FIG. 196. Filtration through porous cone. is sometimes so finely divided that the solid cannot be separated from the liquid by filtration even through a dozen filter papers, and some clays take so long to settle that the solid matter appears to be in permanent suspension. The turbid filtrate may be evaporated to dryness on a water bath, and dried at 110. The residue may then be extracted with water, and the filtrate will usually be quite clear. 2 The most satisfactory method is to filter the solution through biscuit pottery or alundum ware (page 621), 3 paper pulp, or similar filtration media. 4 1 R. Fresenius, Anleitungzur quantitativen chemischen Analyse, Braunschweig, 2. 666, 1887 ; London, 2. 521, 1900. 2 J. Post, Chemisch-technische Analyse, Braunschweig, 2. 98, 1906. 3 L. J. Brigs(J9ttW. U.S. Agric. Dept., 19. 31, 1902) and 0. Schreiner and G. H. Failyer (ib., 31. 12, 1906) forced the solution through a Pasteur- Chamberlain water filter by means of a force-pump. J. W. Mellor, Trans. Eng. Cer. Soc., 5. 54, 1906; Pot. Gaz., 32. 1049, 1907 ; C. E. Munroe, Amer. J. Science (3), I. 336, 1871 ; Ckem. News, 24. 79, 1871 ; F. Klein, Amer. J Pharm., 83. 342, 1911 ; E. B. Forbes, Joum. Ind. Eng. Ohem., 4. 544, 1912 ; G. L. Spencer, ib., 4. 614, 1912 ; M. A. Williamson and P. A. Boeck, ib., 4. 672, 1912; H. 0. Anderson, ib.. 3. 42, 1911; L. E. Saunders, Met. Chem. Eng., g. 257, 1911; W. Pukall, er., 26. 1159, 1893 ; Chem. News, 72. 86, 1895. 4 E. Greiner, Sprech., 42. 399, 1909. THE DETERMINATION OF SULPHUR. 6 3 I A Gooch's crucible in which the bottom of the crucible has neither been per- forated nor glazed, or a biscuit crucible, can be glazed inside and outside to within half a centimetre from the bottom. This crucible is fitted up like the regular Gooch's crucible, fig. 56. Instead of using crucibles with a biscuit bottom, cones of porous earthenware can be used. These can either be fitted inside a suitable funnel, or fitted to the funnel-like neck of a Walther's filtration flask, 1 as indicated in fig. 196. The porous cone is fitted into the funnel by means of a rubber band, so as to make a tight joint. There is then no need for the perforated stopper. 2 Fig. 196 makes the arrangement clear. Riimpler's filtration cups, 3 made of thick, porous filter paper, are sometimes convenient for filtering solutions which give trouble with ordinary filter paper. Care should be taken to get the crucible 4 and filter paper to fit as indicated in the diagram, fig. 197. The inner layer in the diagram represents a section of the Riimpler's cup, and the outer layer, the crucible. To conduct an experiment : boil about 5 grms. of the finely powdered clay with about 250 c.c. of water (distilled) in a resistance glass flask for about half an hour. Restore the water lost by evaporation from time to time. After the clay has settled somewhat, pour the liquid through the biscuit earthenware cone, fig. 196. The pump is now turned on, and the cone is kept filled with liquid. The residue is washed with hot distilled water, and the filtrate is evaporated to dryness in a weighed dish. The residue is dried at, say, 110, and the amount of the residue is expressed as a percentage fraction of the original clay. The clay sometimes clogs the pores of the cone, so that filtration is extremely slow. In that case, the liquid can be filtered first through a Riimpler's cup (fig. 197), and afterwards through the biscuit cone. The washing should be conducted until a few drops of the liquid running from the clay leave no perceptible residue when evaporated to dryness on platinum foil. The result of this FIG. 197. Filtration by Riimpler's shells. experiment does not necessarily represent the whole of the soluble salts in the clay. Some may be adsorbed by the clay in such a way that they can only be removed, if at all, by an extremely prolonged washing. However, the adsorbed salts which are not removed by washing are not of importance from the point of view of "soluble salts," and the result actually obtained is the datum required. The silica, alumina, etc., can be determined in the residue in the ordinary manner (pages 167 seq.). 325. The Determination of Sulphates by the Turbidity Process. The sulphates in the clear solution, say, from the preceding operation, can be determined as barium sulphate by the method of page 617 ; or, more rapidly, if a number of determinations have to be made, by comparing the turbidity of 1 J. Walther, Pharm. Centr., 530, 1898. 2 The air in the rubber rings, under low pressure inside the flask powerful suction is apt to expand and burst the ring. Solid rubber rings do not work so well unless they are made of soft, very yielding rubber. These are not at present easily obtained. 3 A. Riimpler, Deut. Zuckerind., 29. 21, 1904. 4 The cups fit crucibles 6 "5 cm. high, 5'4 cm. upper diameter, and 3 '2 cm. diamete base (outside measurements). 632 A TREATISE ON CHEMICAL ANALYSIS. a solution containing a known amount of barium sulphate in suspension with that produced in the given solution by the addition of barium chloride. The solutions must be dilute enough to produce an opalescent, and not a settling precipitate. The opalescence depends largely on the size of the particles pre- cipitated in the solution, and this, in turn, is determined by the physical condition of the solution temperature, nature of other salts in solution, etc. The method gives fairly satisfactory results for industrial work when speed is a vital factor. The standard and the solution under investigation should not be widely different in strength, and, in consequence, a preliminary comparison should be made in test tubes. If the solution under investigation gives a very much more turbid solution than the standard, it should be diluted ; and conversely. The solutions are not to be diluted after the addition of barium chloride. 1 Standard Solution. A standard solution of calcium sulphate is made by dissolving 0*9 grm. of calcium sulphate CaS0 4 .2H 2 in a litre flask, so that 1 c.c. contains the equivalent of 0'0005 grm. S0 4 . Make 25 c.c. of this standard solution up to a litre. Pipette, say, 100 c.c. into a wide-mouthed bottle. Add O'l to 0*2 grm. of powdered crystalline barium chloride and shake vigorously at intervals of 30 minutes. Pipette 50 c.c. of this solution into a Nessler's glass surrounded by black paper or a black velvet jacket. Test Solution. Make the solution from, say, 10 grms. of clay up to 250 c.c. with water ; acidify with hydrochloric acid. Pipette, say, 100 c.c. of this solution into a tightly stoppered wide-mouthed bottle, and treat this as just described for the standard solution. Comparison. Place the test solution in a burette and run it into a Nessler's glass fitted with a black jacket, as indicated above, until the solutions in the two Nessler's glasses appear to have the same turbidity. The comparison is made with the Nessler's glasses resting upon a plate of clear glass supported over a sheet of black paper. Calculations. Suppose that the two solutions have the same turbidity when 13 c.c. have been run from the burette. Then 13 c.c. of the given solution have the same turbidity as 50 c.c. of the standard, i.e. as 0'000625 grm. S0 4 . Hence, 250 c.c. will have the equivalent of 250 x 0-000625 - =0-012 grm. S0 4 . Hence, 10 grms. of the clay have the equivalent of 0-012 grm. of S0 4 ; or, 0-12 per cent. S0 4 ; or 0'22 per cent. CaS0 4 . The most important disturbing influences arise from the introduction of side- lights during the comparison of the colours ; and the mode of precipitation hot or cold, with the solid salt or with an aqueous solution, and whether the readings be made at once or after the solution has stood some time. Con- sequently, it is necessary to follow rigorously the same procedure in preparing the standard and the test solutions. Instead of employing Nessler's glasses for the comparison, Richards and Wells use a special instrument which they call a nephelometer, for comparing the turbidity of different solutions. Hind and Jackson read the depth of the turbid liquid at which a standard sperm or wax candle flame just ceases to be visible ; the corresponding amount of S0 4 is read directly from a 1 T. W. Richards and R. 0. Wells, Amer. Chem. Journ., 31. 235, 1904; Journ. Amer. Chem. Sot., 27. 459, 1905 ; J. L. D. Hinds, ib., 18. 661, 1896 ; 22. 269, 1900 ; Ckem. News, 73. 285, 299, 1896 ; D. D. Jackson, Journ. Amer. Chem. Soc.. 23. 799. 1901 : S. W. Parr and C. H. M'Clure, ib., 26. 1139, 1904 ; S. W. Parr, W. F. Wheeler, and R. Berolzheimer, Journ. Ind. Eng. Chem., i, 689, 1909. THE DETERMINATION OF SULPHUR. 633 standard table prepared by empirical observations with solutions of known strength. 1 The cylinders are 3 -5 cm. wide, 20 cm. high, and graduated from below upwards. The cylinder is supported over the lighted candle, and the liquid with the precipitate in suspension is poured in until the tip of the candle flame just disappears. 326. The Amount of Barium Salt required to make the Soluble Sulphates in a Clay Innocuous. Boil 40 grms. of the clay 2 in a long-necked flask with a litre of water. The soluble sulphates pass into solution. Add 1 c.c. of a standard solution of barium chloride containing 10 grms. BaCl 2 per litre. Shake the contents of the flask, and pipette about 2 c.c. of the contents on to a funnel fitted with a close- packed filter paper over a test tube. This test tube is labelled No. I. Add another cubic centimetre of the barium chloride solution, shake, and transfer about 2 c.c. to a second test tube, etc., labelled No. II. This operation is repeated with a series of test tubes ranged in order so that each tube represents 1 c.c. more barium chloride solution than the tube immediately preceding, and 1 c.c. less than the tube immediately succeeding. * Add about 3 drops of dilute sulphuric acid to each test tube. That tube in which a turbidity first appears shows the tube which contains sufficient barium chloride to react with the sulphates. For instance, suppose that this is tube No. XIV., with 14 c.c. of barium chloride; and that the preceding one, No. XIIL, with 13 c.c., is clear. The end of the titration is somewhere between these two tubes, say 13*5 c.c. 3 Practice 4 shows that with the above proportions every cubic centimetre represents 1 Ib. of barium carbonate to be mixed with a ton of clay to transform the soluble sulphates into insoluble and inert barium sulphate. Hence, in the example under consideration, 13 J Ibs. of barium carbonate, or 6J Ibs. of barium chloride, 5 are required per ton of clay. The results agree fairly well with practice, although the method has many faults. If much more than 20 c.c. of barium chloride be required, the soluble salt difficulty cannot be cured by the baryta method. Seger and Cramer 6 have a more elaborate apparatus than that just indicated, but it requires less expert manipulation. The results are similar. Some clays e.g. Wetley marl give liquids very difficult to filter clear, and it is not then easy to determine the end of the barium chloride titration by the turbidity of the filtered solution. In that case, it is necessary to separate the solution from the clay by filtration through biscuit ware, as indicated page 630. 1 J. I. Hinds gives forS0 3 the relation ?/ = > 0482-ra;, where x is the cylinder reading, and y the percentage value of S0 3 sought ; and for CaO when the lime is precipitated as calcium oxalate ?/ = 0'036 -f (x - 0'3). Different cylinders, and different 'amounts of raw material, would, of course, require different formulae. 2 Air-dried, or natural undried clay, or clay dried at 110. 3 C. Beringer and J. J. Beringer, Chem. News, 59. 41, 1889 ; R. Wildenstein, Zeit. anal. Chem., i. 432, 1862. 4 B. Kerl, Handbuch der Tonwaarenindustrie, Braunschweig, 71, 523, 1906. 5 Relatively more of the insoluble carbonate is needed because of the mixing difficulty. The soluble chloride is relatively easily brought in contact with the sulphates. 6 H. A. Seger and E. Cramer, Tonind. Ztg., 18. 637, 1894; Hilfs-Gercite fur Beaufsichtigung und Vcrvollkommnung des Betriebes von Ziegeleien, Berlin, 96, 1911. CHAPTER XLIII. THE DETERMINATION OF THE HALOGENS. 327. The Detection of Fluorides. THERE is no certain direct test for fluorides in silicates. The two most important reactions available for qualitative tests are (1) the evolution of hydrogen fluoride, which occurs when a fluoride is heated with concentrated sulphuric acid. The reaction is expressed in chemical symbols : CaF 2 + H 2 S0 4 ->CaS0 4 + 2HF. The presence of the hydrogen fluoride is recognised by its action on glass the so-called " etching test." (2) If silica or a silicate be present, most of the hydrogen fluoride reacts with the silica to form silicon fluoride SiF 4 as indicated by the symbols : Si0 2 + 4HF->2H 2 + SiF 4 . The silicon fluoride will not etch glass, but it will react with a drop of water, forming hydrofluosilicic and silicic acids. Thus, in symbols: 3SiF 4 + 3H 2 0->2H 2 SiF 6 + H 2 Si0 3 . The latter gives the drop of water a turbid appearance the so-called "hanging drop test." If the silicate contains fluorides which are not attacked by the concentrated acid, it is necessary to get the fluoride in a form say, as calcium fluoride susceptible to attack before either test can be applied. The Separation of Fluorine as Calcium Fluoride from Fluo-Silicates not attacked by Sulphuric Acid. 1 Mix the fluo-silicate with about eight times its weight of sodium carbonate, fuse the mixture in a platinum crucible, and when the mass is cold extract with water. The solution contains sodium fluoride and sodium silicate. Remove the silica by adding an excess of Schaffgotsch's solution, and allow the mixture to stand overnight in a warm place. Filter, and evaporate the nitrate to a small volume. Add a little phenol phthalein, and carefully add dilute hydrochloric acid from a burette until the red colour of the solution disappears. To make sure the mass is just neutralised, and no excess of acid is present, heat the solution to boiling the red colour will reappear as the solution cools and repeat the addition of hydrochloric acid, etc., until the solution is only faintly coloured on boiling. Add an excess of a solution of calcium chloride, and again boil the solution. Filter off the precipitate, wash, dry, and ignite in a platinum crucible. Add acetic acid, and evaporate to dryness. This converts any calcium carbonate into soluble acetate. Rub up the mixture with water, and filter off the insoluble calcium fluoride. Dry and burn off the filter paper. The powder is then ready for the etching test. The Etching Test. This is available for fluorides which are free from silica or silicates, and are attacked by concentrated sulphuric acid. A small clear glass plate, free from scratches, is thoroughly cleaned and warmed. A little molten wax 2 is poured on to the warm plate, and the excess drained off, so that 1 F. P. Treadwell, Analytical Chemistry, New York, I. 352, 1903. 2 Beeswax ; or, better, melt together equal weights of carnauba wax and paraffin wax, and thoroughly mix by stirring. THE DETERMINATION OP THE HALOGENS. 635 a thin uniform layer of wax remains spread over the plate. The plate is allowed to cool in a horizontal position. While the wax is still warm, make a small *J on the wax with a pointed instrument, so as to lay bare, but not to scratch, the glass. 1 The powder under investigation is placed in a platinum crucible, 2 along with 2 or 3 c.c. of concentrated sulphuric acid. 3 Warm the upper edge of the crucible cautiously and quickly with a small flame. Press the glass plate, waxed side downwards, upon the crucible so that the cross is in the centre and the crucible is sealed to the plate when the wax cools. The crucible is supported in a hole in a piece of thick asbestos board, cut so as to fit the crucible tightly. Put two or three drops of water on the glass plate and press the end of a condenser 4 (tig. 199) down on to the plate. 5 Heat the crucible for about half an hour over a small flame 9 mm. high, and 6 mm. below the bottom of the crucible. The disposition of the apparatus will be obvious from fig. 198, where C is the condenser, B the glass plate, A the asbestos board and crucible, and D the micro-burner for heating the crucible. 6 When the mixture has been heated for about an hour, remove the condenser and plate. Warm the plate a little, and wipe off the wax. Clean the plate on both sides with polishing powder which will not scratch the glass. Examine the plate by reflected light for any etching. A test should not be considered positive unless the cross can be seen from both sides of the glass. 7 Woodman and Talbot 8 state that their process, as described above, will give FIG. 198. Etching test for fluorides. 1 The marks are best made about 1 mm. wide, and the two arms of the cross about 4 mm. long. The ends of the arms of the cross should be marked on the uncoated side of the glass with a scratching diamond or file. 2 It is best to have everything in the crucible quite dry, since the etching is not always so well defined if moisture be present. 3 The sulphuric acid must be free from fluorides. This can be determined by a blank experiment. If fluorides be present, they can be removed by diluting the acid with water and evaporating down to its former volume. 4 According to Woodman and Talbot, a suitable condenser is made from a tube resembling a carbon filter tube by stretching a piece of sheet rubber (such as is used by dentists) over the wide end. The diameter of the tube should be wider than the platinum crucible ; the narrow end is fitted with the necessary inlet and outlet tubes. 5 The etching is frequently done on the convex side of a waxed watch-glass, and a little cold water is kept in the watch-glass to prevent the wax melting, covering the cross with wax, and so stultifying the test. 6 It is advisable every now and again to put a couple of drops of water on the plate round the edge of the condenser to keep the wax from melting. 7 Sulphuric acid may corrode the glass so that the cross can be seen when the glass is breathed upon. The scouring with the polishing powder usually removes the sulphuric acid "stain." 8 A. G. Woodman and H. P. Talbot, Journ. Amer. Chem. Soc., 28. 1437, 1906 ; 29. 1362, 1907 ; C. Blarez, Chem. Neivs, 91. 39, 1905 ; A. E. Leach, Ann. Rep. Mass. State Board of Health, 36. 309, 1905 ; G. W. M. Williams, Chem. World, i. 255, 1912. 636 A TREATISE ON CHEMICAL ANALYSIS. a recognisable test, that is, the cross will be visible from both sides of the plate, with 1 : 5,000,000 parts of fluorine. The temperature is an important factor. It is not possible to estimate the amount of fluorine from the intensity of the etching, 1 but Woodman and Talbot obtained promising results by varying the temperature of the crucible. Thus, they state : Table LXXIII. Test Analyses for the Detection of Fluorine. Temperature. Distinguishes one part of fluorine per 79- 82* 113 136 173-178 ""213-218 25,000- 100,000 100,000- 1,000,000 1,000,000- 5,000,000 5,000,000-25,000,000 less than 25,000,000 This test is of very limited application so far as the silicate industries are concerned, because the complete absence of silica is essential if but small quantities of fluorine are in question. The Hanging Drop Test. About half a gram of the thoroughly dried powder under investigation is well mixed with about O'l grm. of dried precipitated silica and placed in the bottom of a test tube about 5 cm. long and 1 cm. wide. The test tube A (fig. 199) is fitted with a one-hole rubber stopper, B. The stopper carries a piece of glass tubing, (7, closed at one end, and inserted in the stopper so that the open end of the tube extends about 3 mm. below the stopper. The glass tubing C is nearly filled with two drops of water, D, from a small pipette. The stopper and everything else inside the test tube must be quite dry. Pipette 1 or 2 c.c. of concentrated sulphuric acid into the test tube, and immediately insert the stopper without dislodging the drop of water in the little tube. Place the test tube in a beaker of water. Heat the water to boiling, and, after 20-30 minutes' heat, ing, if the substance contains appreciable quantities of FIG. 199. Hanging fluorine, a heavy gelatinous ring of silicic acid will be drop test for fluorides, found at the mouth of the little tube carrying the drop of water. 2 With a little practice, or by conducting the test simultaneously with powders containing known amounts of fluoride, a rough idea can be formed how much fluorine is present, and whether the com- pound is worth a quantitative investigation by, say, Oettel and Hempel's process (page 646). Carbonates should not be present, or the stopper may be blown 1 H. Ost, Ber., 26. 152, 1893 ; but see 0. Renner, Ueber die Bestimmung des Fluors, Weida, i. Th., 1912. 2 W. Kopp (Ber. Konig. Sachs. Ges. Wiss. t 37, 1882) has modified the test. A small flask with exit tube is thoroughly dried, and a mixture of the finely powdered material with the precipitated silica is placed in the flask. Pour an excess of concentrated sulphuric acid into the flask. One end of a delivery tube is fitted to the flask, and the other end dipped into a small cylinder containing a decigram of colourless aniline (or ammonia) in 30 c.c. of water. Heat the flask to 50 or 60. A white deposit about the part of the delivery tube which dips into the aniline indicates the presence of fluorine. The white glistening crystalline solid aniline silicofluoride which forms in the liquid may be digested with a solution of caustic soda in absolute alcohol, and sodium silicofluoride will be formed. See also P. E. Browning, Amer. J. Science (4), 32. 249, 1911 ; E. Rupp, Zeit. Nahr. Genuss., 22. 496, 1911 ; A. Sartori, Chem. Ztg., 36. 229, 1912 ; A. Kickton and W. Behncke, Zeit. Nahr. Genuss., 20. 193, 1910. THE DETERMINATION OF THE HALOGENS. 637 from the tube or the tube burst. Hence, carbonates should be destroyed by calcination before the test is applied. 328. The Gravimetric Determination of Fluorine as Calcium Fluoride. The fluorides in many silicates are not decomposed by digestion with sulphuric acid, and, in consequence, it is necessary to get such fluorides into solution before the analysis can be made. When much fluorine is present the silica must be determined by a special process, because part will be volatilised as silicon tetra- fluoride when the aqueous extract of the sodium carbonate fusion is evaporated to dryness, etc. According to Deville and Fouque, 1 the " loss on ignition " can be determined without losing fluorine, since the decomposition of the calcium fluoride requires a higher temperature than for the expulsion of water. 2 Alkaline fluorides, how- ever, volatilise at a comparatively low temperature. Thus, 0*549 grm. of sodium fluoride lost 0'4 per cent, in weight when heated for about 6 minutes in the full flame of a Bunsen's burner; O2714 grm. of sodium fluoride lost 1-4 per cent, in weight under the same conditions ; while no appreciable loss could be detected after heating 15 minutes with the flame just sufficient to redden the bottom of the platinum crucible. 3 The methods for the isolation of the fluorides are rather tedious. That usually employed is based upon one proposed by Berzelius and Rose. 4 The compound is decomposed by fusion with alkaline carbonate with or without silica. Silica and the bases are precipitated from the solution of the fused cake with ammonium carbonate, ammoniacal zinc oxide (Berzelius' solution), or ammoniacal mercuric oxide (Seernann's solution), and from the resulting solution, calcium or barium fluoride, together with the oxalate, sulphate or carbonate, are precipitated from the filtrate. The carbonate, etc., is removed by acetic acid, and calcium fluoride remains. 5 The First Silica Precipitation. Two grams of the uncalcined silicate are fused in a platinum crucible with about 1 2 grms. of sodium carbonate 6 without the blast. 7 The resulting cake is leached with water, filtered, and washed. Some of the silica, sodium zirconate, barium and calcium carbonates, if present, remain on the filter paper. 8 This and all the other precipitates are reserved for the determination of silica, etc. 1 H. St C. Deville and F. Fouque, Compt. Rend., 38. 317, 1854. See page 639. 2 If there is any danger of loss of fluorine during ignition, G. Tammann (Zeit. anal. Chem., 24. 343, 1885) recommends adding a large excess of sodium carbonate or barium hydroxide. 3 S. Waldbott, Journ. Amer. Chem. Soc., 16. 418, 1894. 4 J. J. Berzelius, Pogg. Ann., I. 169, 1824 ; H. Rose, ib., 79. 115, 1850; T. M. Chatard, Trans. Amer. Inst. Min. Eng., 21. 170, 1893 ; F. Wyatt, The Phosphates of America, New York, 149, 1892; T. Korovaeff, Journ. praU. Chem. (1), 85. 442, 1862; P. Jannasch, NeuesJahr. Min., 2. 123, 1883 ; A. A. Koch, Journ. Amer. Chem. Soc., 29. 1126, 1907 ; W. F. Hillebrand, Bull. U.S. Geol. Sur., 422. 184, 1910 ; L. Fricke, Stahl Eisen, 24. 889, 1904 ; E. Zdarck, Zeit. physiol. Chem., 69. 127, 1910 ; K. Daniel, Zeit. anorg. Chem., 38. 257, 1904. 5 Small amounts of fluorine can also be determined colorimetrically with advantage page 644. 6 Calcium fluoride cannot be completely decomposed by fusion with sodium carbonate, but the decomposition is complete if the fluoride be mixed with silica or a silicate. Hence, if the fluorides be high, and the silica low, the addition of 3 grms. of silica is recommended to ensure complete decomposition (page 647). This is not usually needed for Cornish stone, nor for glazes. 7 G. Forchhammer stated that all the fluorine is not obtained by melting silicates, e.g. topaz or tourmaline, with sodium carbonate ; and Stadeler states that if the lid be removed during the fusion there will be a loss of fluorine (Journ. prakt. Chem. (1), 99. 66, 1366). 8 If sulphur, zirconium, and chlorine are to be determined, the solution from the fused cake can be made up to 250 c.c. Take 50 c.c. for the sulphur ; 100 c.c. for the fluorine ; ad 100 c.c. for the chlorine. But the solutions may be so dilute that it is advisable to use a separate sample for these determinations rather than take a correspondingly greater amount of the original sample for this fusion. 638 A TREATISE ON CHEMICAL ANALYSIS. The Second Silica Precipitation. The filtrate contains the fluorine as alkaline fluoride. Introduce 10 grms. of solid ammonium carbonate, 1 and digest the mixture for about 12 hours at about 40, whereby silica and alumina are precipitated. Filter and wash the precipitate with aqueous ammonium carbonate. The nitrate still contains some silica. The Third Silica Precipitation. Evaporate the filtrate todryness; 2 digest the mass with water ; and neutralise the solution very carefully with nitric acid, as recommended by Treadwell, 3 with phenolphthalein as indicator : Add 2N-nitric acid from a burette until the pink colour of the indicator has disappeared. Heat the solution to boiling, and the red colour reappears. When cold, again discharge the colour of the indicator; heat to boiling as before, and repeat the operations until the addition of 1-1 J c.c. of 2N-nitric acid suffices to discharge the colour. Add 5 c.c. of Seemann's solution, 4 and evaporate the mixture until the smell of ammonia has disappeared. Ammonia interferes with the subsequent precipitation of calcium fluoride. The precipitate con- tains traces of silica and mercury oxide. 5 Filter off the precipitate, and wash with water. The Precipitation of the Fluoride as Calcium Fluoride. Add nitric acid to the filtrate until the alkaline carbonate is nearly decomposed. If too much acid be present, a little more sodium carbonate must be added, so that the solution is distinctly alkaline. Boil with a large excess of an aqueous solution of calcium chloride. 6 Calcium carbonate and fluoride are precipitated. 7 The Removal of Calcium Carbonate. Collect the precipitate on a filter paper, 8 1 Not the nitrate or chloride, in order to avoid loss on evaporation. 2 The liquid is inclined to froth and spit during the evaporation, owing to the decomposition of the ammonium carbonate. Hence, the dish must be covered by a clock-glass as long as there is any danger of loss. Wash the clock-glass, and of course let the washings run into the basin. 3 F. P. Treadwell, Kurzes Lehrbuch der analytische Chemie, Leipzig, 2. 389, 1911. 4 SEEMANN'S SOLUTION. Moist, freshly precipitated mercuric oxide (about 20 grms.) is dissolved to saturation in Schaffgotsch's solution. The mercuric oxide is made by adding an aqueous solution of sodium hydroxide to a hot saturated solution of mercuric chloride, and the precipitate is washed with hot water until a portion of the precipitate volatilises on platinum foil without residue F. Seemann, Zeit. anal. Chem., 44. 343, 1905. BERZELIITS' SOLUTION. A solution of ammonio-zinc oxide, as originally recommended by T. J. Berzelius (I.e. ), is sometimes employed for the purpose as Seemann's. Precipitate an aqueous solution of pure zinc sulphate with caustic potash. Filter, wash by decantation, and dissolve the precipitate in ammonia M. Kleinstiick, ib., 50. 697, 1912. 5 And phosphorus, if all was not precipitated with the alumina. Test with a solution of ammonium molybdate. If phosphorus be present, add an .excess of silver nitrate solution. This precipitates silver phosphate and carbonate. Warm the solution in order to coagulate the precipitate, and filter. Precipitate the excess of silver nitrate with an aqueous solution of sodium chloride (2E). Boil the solution to coagulate the silver chloride, and filter. Add a little sodium carbonate until the solution reacts alkaline. The same procedure removes any chromate as silver chromate W. F. Hillebrand, Bull. U.S. Geol. Sur., 422. 186, 1910. 6 CALCIUM CHLORIDE SOLUTION. Dissolve 21'93 grms. of the crystalline salt (CaCl 2 . 6H 2 0) in 100 c.c. of water (2E). 7 G. Starck and E. Thorin (Zeit. anal. Chem., 51. 14, 1912) add a known amount of oxalic acid and precipitate the fluorine with calcium oxalate by the addition of calcium chloride in a solution slightly acidified with acetic acid. The precipitate is filtered on asbestos, dried at 210, and weighed. The calculated amount of calcium oxalate is deduced, and the difference represents calcium fluoride. G. Starck (Zeit. anorg. Chem., 70. 173, 1911; G. Starck and E. Thorin, Zeit. anal. Chem., 51. 14, 1912) adds lead chloride to precipitate lead chloro- fluoride PbFCl. This precipitate has a high molecular weight (page 332). The method has not been thoroughly tested, but it promises well. 8 Calcium fluoride alone gives a slimy precipitate which is very difficult to filter, but in the presence of calcium oxalate or carbonate it filters better. The calcium fluoride is sparingly soluble in hydrochloric and nitric acids, and almost insoluble in acetic acid. The precipitate is more soluble in the presence of ammonium salts. THE DETERMINATION OF THE HALOGENS. 639 wash with hot water, dry, and ignite slowly to faint redness, 1 in a platinum crucible. 2 Add dilute acetic acid in excess 3 of that needed to dissolve the calcium oxide and carbonate. Heat in a covered crucible on a water bath until effervescence ceases. Evaporate to dryness. Digest the residue in water very slightly acidulated with acetic acid. 4 Filter, wash, and ignite the precipitate. If much fluorine be present, the digestion with acidulated (acetic) water may be repeated. The complete removal of calcium oxide and carbonate is shown by the absence of residue when a drop or two of the filtrate is evaporated to dry- ness on a piece of platinum foil. The last filter paper and contents with the precipitated calcium fluoride are ignited gently to faint redness in a platinum crucible this temperature just suffices to burn off the filter paper the crucible and contents are then weighed. Verification of the Result. That the precipitate is really calcium fluoride should be confirmed by converting the calcium fluoride into calcium sulphate by the addition, drop by drop, of concentrated sulphuric acid. Drive off the excess of acid by evaporation, and test the fumes for hydrogen fluoride by means of a greased glass plate in the usual manner. Gradually raise the temperature to a red heat. Cool, and weigh. One gram of CaF 2 should give 17436 grms. CaS() 4 . If the weights are not concordant, as they rarely are, the impurity may be silica, calcium silicate, or calcium phosphate arising from the imperfect separation of silica and phosphorus. Calcium sulphate may also be present, Hue to the imperfect washing of the precipitate. Phosphorus can be detected by the ammonium molybdate process in the hot nitric acid solution. It is difficult to decide whether the impurity be calcium silicate or silica. In the case of the phosphate and silicate, these constituents may be decomposed by the sulphuric acid treatment ; but in the case of silica and calcium sulphate, it is possible to calculate the amount of calcium fluoride from the observed increase in weight. If the observed increase in weight be multiplied by 1*3448, we get the corresponding amount of calcium fluoride ; and by 0*6543, the corresponding amount of fluorine. The raison d'etre of the calculation will appear from the equation : CaF 2 + H 2 S0 4 ->CaS0 4 + 2HF, where 136*16 represents the molecular weight of calcium sulphate, and 78*09, the molecular weight of calcium fluoride. Hence, when the sulphuric and hydrofluoric acids have been driven off, an increase in weight of 58*07 grms. will represent 78*09 grms of calcium fluoride, or 38 grms. of fluorine. Hence, if iv represents the increase in weight obtained by the action of sulphuric 1 A slight loss of fluorine may occur by ignition of the filter paper in contact with calcium fluoride. G. Tammann, Zeit. anal. Chem., 24. 328, 1885 ; W. F. Hillobrand, Bull. U.S. Geol. Sur., 422. 186, 1910. W. Hempel (Gasanalytische Methoden, Braunschweig, 347, 1900) found that 1*5 grms. heated 15 minutes at 1000 lost 0*019 grin. ; and at a dull red heat, 0'005 giro. F. Seemann (Zeit. anal. Chem., 44. 343, 1905) noticed a loss of 0*0008 grm. when a gram was heated over a Teclu's burner at 1000 ; and F. P. Treadwell and A. A. Koch (Zeit. anal. Chem., 43. 469, 1904) found a loss of 0'0002 grm. after ignition for 10 minutes over a Teclu's burner, and .0*0009 grm. after half an hour's heating. 2 Instead of carrying this process any further, we may proceed to page 641. 3 If the precipitate were treated directly with the acetic acid, without the preliminary ignition, the proper washing of the fluoride would be extremely difficult. 4 S. L. Penfieldand J. C. Minor, Amer. J. Science (3), 47. 389, 1894 ; W. F. Hillebrand, Bull. U.S. Geol. Sur., 422. 186, 1910; F. P. Treadwell and A. A. Koch, Zeit. anal. Chem., 43. 469, 1904. According to Treadwell and Koch (I.e.), the loss of calcium fluoride per 100 c.c. of wash-water is about 0*0015 grm. With a one-gram sample, therefore, from 0*04-0*05 per cent, of fluorine will entirely escape observation " W. F. Hillebrand (I.e.). 640 A TREATISE ON CHEMICAL ANALYSIS. acid on the given sample, the corresponding amount x of fluorine will be as 58-07 : w = 3S: x\ or, Amount of fluorine = 0'6543w ', and, Amount of calcium fluoride = l'3447w. In illustration, Calcium fluoride (found) 0-0426 grm. Calcium sulphate (found) '0724 grm. Calcium sulphate (theory) . . . . . '0743 grm. No phosphates or silicates were present. Hence, it is assumed that the impurity was calcium sulphate not perfectly washed away. Consequently, the amount of fluorine in the 2-grm. sample is 0-6543 x (0'0724 - 0'0426) = 0-0195 grm. Hence, the sample has 0*98 per cent, of fluorine; or, 2 '01 per cent, of calcium fluoride. Errors. Owing to the relatively small amounts of fluorine usually present in natural silicates, great care must be taken to prevent, as far as possible, any contamination of the calcium fluoride. The chief sources of error are : (1) loss of fluorides owing to the "adsorption" by the voluminous "silica" precipitates ; l (2) the solubility of calcium fluoride in water and acetic acid ; (3) loss of calcium fluoride by volatilisation, particularly in contact with the filter papers; and (4) the contamination of the precipitated calcium fluoride with phosphates, silica, calcium silicate, or calcium sulphate. Correction of the Analysis " Total" for Fluorine. In summing up the results of a silicate analysis which includes appreciable amounts of fluorine or chlorine, the sum of the constituents will exceed the limit "100 0*5." This is due to the fact that the bases have been calculated as oxides when some of the bases were really present as fluorides or silicofluorides. For instance, a translucent white glass furnished on analysis : 2 Si0 2 . ALOj,- MnO. ZnO. MgO. CaO. Na 2 0. F. 64-8 9-3 1-1 7-0 0-2 1'9 10'8 8'1 The total reaches 103-2. But 38 grms. of fluorine are equivalent to 16 grms. of oxygen, or 1 grm. of fluorine is equivalent to 0-42105 grm. of oxygen. Consequently 8'1 grms. of fluorine are equivalent to 8'1 xO'421 = 3'4 grms. of oxygen. Hence, we append to the statement: "Total, 103'2," "less 3'4 per cent, of oxygen, corresponding with 8'1 per cent, of fluorine." 329. The Determination of Silica in the Presence of Fluorides. For reasons stated on page 172, the silica cannot be properly determined by the method of page 167 when fluorides are present, but it can be determined in the "by-products" of the fluorine determination, namely, the three precipitates indicated on pages 637 and 638. Each filter paper is spread on a watch-glass, and the precipitates rinsed into one evaporating basin with dilute hydrochloric acid. Ignite the filter papers separately, and add the ashes to the basin. Decompose the precipitates with hydrochloric acid, and evaporate to dryness on a water bath. The silica is recovered in the usual manner. The mercury volatilises on ignition. The alumina, titanic oxide, ferric oxide, magnesia, and lime 3 can be determined separately in another 1 Hence, some recommend collecting the different "silica precipitates," igniting, and again treating the residue with sodium carbonate, ammonium carbonate, etc. 2 ' ' Alumina " in the analysis includes ferric oxide. 3 In glazes, the lead, tin, etc., can be conveniently determined in a portion decomposed with sulphuric and hydrofluoric acids. THE DETERMINATION OF THE HALOGENS. 641 portion of the silicate decomposed by the hydrofluoric acid treatment, or in the nitrate from the silica. 1 330. The Gravimetric Determination of Fluorine as Potassium Fluosilicate. In Carnot's process 2 for the determination of fluoride, the silicate is digested with sulphuric acid, and the silicon tetrafluoride is passed into an aqueous solution of potassium fluoride, 3 and finally weighed as potassium fluosilicate. Carnot's process is thus based upon older methods by Fresenius, and by Wohler. The method gives good results with many of the natural fluorides used indus- trially, and also with slags and frits. All these are supposed to be decomposed by hot concentrated sulphuric acid. Daniel considers that methods depending on the volatilisation of silicon tetrafluoride are not reliable if amorphous silica, or silicates, be present, because a non-volatile fluoride may be formed which leads to low results. On the other hand, Classen 4 actually recommends : " One part of powdered quartz and 0'5 part of precipitated silica is to be intimately mixed in an agate mortar with the substance to be analysed." If there be any reason to suspect low results from the cause indicated by Daniel, the silicate can be treated as described in the preceding section. When all the silica has been separated, and the precipitate containing calcium fluoride and carbonate has been ignited, the mixture, if desired, can be treated by Carnot's process. The Preparation of the Substance for Analysis. Two grams of the substance under investigation are mixed with twice their own weight of finely ground, dry, quartz sand, 5 or rock crystal. If the sample contains other than about O'l grm. of fluorine, a proportionally less or greater amount than 2 grms. may be taken. 6 Fitting up the Apparatus. The mixture is placed in an Erlenmeyer's flask of about 150 or 250 c.c. capacity (A, fig. 200). The flask is closed with a three-hole rubber stopper, and connected with the system shown in the diagram. A Walter's gas- washing bottle 7 (7, charged with concentrated sulphuric acid, 1 H. Gilbert, Corresp. Ver. anal. Chem., 3. 1, 1880. SeeR. Fresenius and E. Hintz(Zeit. anal. Glum., 28. 324, 1889) for the determination of silica, etc., in cryolite ; Loczka, ib., 49. 328, 1910. 2 A. Carnot, Ann. Mines (9), 3. 138, 1893; Compt. Rend., 114. 750, 1892; Bull. Soc. Chim. (3), 9. 71, 1893 ; H. Lasne, ib. (2), 50. 167, 1888 ; Ann. Chim. Anal., 2. 161, 182, 1897 ; E. Coutal, ib., 2. 401, 1897 ; A. Liversidge, Chem. News, 24. 226, 1871 ; G. Harker, ib., 82. 56, 64, 1900 ; J. Schuch, Zeit. Landw. Vers. Wes. Ost., 9. 531, 1904 ; A. Kiipffer, Archiv Nat. angeiv. Chem., 14. 101, 1901 ; E. Frost and F. Balthaser, ib., 14. 292, 1901 ; Bull Assoc. Belg. Chim., 13. 453, 1899 ; H. von Kobell, Joum. prakt. Chem. (1), 92. 385, 1864 ; F. Stolba, ib. (1), 89. 129, 1863 ; F. Wohler, Pogg. Ann., 48. 87, 1839 ; J. Brandl, Liebitfs Ann., 213. 1, 1882 ; H. Schiff, Liebig's Ann. Suppl., 4. 27, 1865; R. Fresenius, Anleitung zur quantitativen chemischen Analyse, Braunschweig, i. 431, 1875 ; Zeit. anal. Chem., 5. 190, 1866 ; S. Bein, ib., 26. 733, 1887 ; E. Wrampelmeyer, ib. t 32. 550, 1893 ; J. Casares, ib., 34. 546, 1895 ; 44. 729, 1905 ; A. Gautierand P. Clausmann, Compt. Rend., 154. 1469, 1912 ; H. A. Weber, Centr. Min., 2. 504, 1902. 3 E. Deladrier (Chem. Weekblad, i. 324, 1904) uses a solution of thorium chloride, and finally weighs the precipitated thorium fluosilicate. 4 A. Classen, Ausgewahlte Methoden der analytischen Chemie, Braunschweig, 2. 430, 1903. 5 The quartz sand should be moistened with sulphuric acid, and calcined to remove organic matter, or treated by Hempel's process of purification (page 647). 8 For example, "2 grm. of fluorspar or cryolite will suffice ; 2 grms. of mineral phosphates containing from 2 to 3 per cent, of fluorine ; and 5 grms. of bone ash with about '2 per cent of fluorine can be taken for the analysis. If the silica and alumina are not to be determined, the preliminary treatment with Seemann's solution may not be needed. The ammonia must, how- ever, be eliminated. If the fluorides are completely decomposed by sulphuric acid, the treatment with sodium carbonate may be omitted if the silica is not to be determined. 7 J. Walter, Dingier 's Jo urn., 251. 367, 188-4. 41 642 A TREATISE ON CHEMICAL ANALYSIS. is also fitted with a tube B, charged with soda lime. The former is connected with the flask containing the substance under investigation by means of a tube FIG. 200. The determination of fluorine. extending to the bottom of the Erlenmeyer's flask, A. The flask A is also fitted with a stoppered funnel D, leading to a von Babo's or similar drying tube E, packed with glass beads, in order to remove any acid vapours which might rise from the flask (A}. The Emmerling's tube is connected with a tube (F) drawn out at one end. This dips 2 or 3 mm. beneath the surface of 10 c.c. of mercury contained in a small cylindrical vessel closed by a two-hole rubber stopper and fitted with a stoppered tube (, fig. 201). This vessel is shown enlarged in the sectional diagram, fig. 202. This vessel also contains about 20 c.c. of a neutral solution of potassium fluoride l 20-25 per cent. The utmost care must be taken that all the glass parts between B and G be thoroughly dried. The drying is best effected by aspirating a current of dry air through the ap- paratus while the different parts are warmed. The Decomposition of the Fluoride. Add 40 c.c. of con- centrated sulphuric acid 2 to the flask A by means of the funnel D. Aspirate a current of air through the system so that about one bubble per second rises 1 POTASSIUM FLUORIDE SOLUTION. Commercial potassium fluoride is generally acid and attacks glass. To prepare this for the work, dissolve 25 grms. in 80 c.c. of distilled water in a platinum dish ; add dilute potash solution, drop by drop, until the solution is neutral to tincture of cochineal. Add absolute alcohol, drop by drop, until the solution shows a faint turbidity. Let the precipitate settle and filter. If the solution is free from potassium fluo- silicate, it should give no precipitate when 10 c.c. is mixed with 40 c.c. of water and 50 c.c. alcohol (90 percent). ^ The sulphuric acid should be free from oxides of nitrogen, and hydrochloric and hydro- fluoric acids. This can be effected by heating the acid in a platinum basin along with a little powdered quartz. FIG. 201. Decom- position Vessel. THE DETERMINATION OF THE HALOGENS. 643 through the mercury in F. The flask A is heated in a suitable oil bath, ff, and the temperature gradually raised to 150 or 160 ( T represents a thermometer). The fluorides in A commence to decompose at about 100, and potassium fluosilicate begins to separate in the aqueous solution in G. If any bubbles of the fluosilicate stick to the walls of G, they must be broken up by gently rotating the fluid. When no new bubbles of the fluosilicate are formed, aspirate air a little more vigorously through the system for about half an hour. The Preparation of the Potassium Fluosilicate for Weighing. Decant, by means of a pipette, the fluid containing the clear gelatinous precipitate into an Erlenmeyer's flask ; wash the absorption vessel (G) and the mercury two or three times with a little distilled water. 1 The main liquid and the washings are collected in the same flask, the contents of which should not exceed 100 c.c. Shake the liquid with its own volume of absolute alcohol ; let the mixture stand 2-3 hours ; decant the clear liquid through a Gooch's crucible which has been dried at 100 and weighed; wash the precipitate on to the filter with. a mixture of equal volumes of water and alcohol, taking care to leave any globules of mercury in the flask ; continue washing until the washings give no precipitate with calcium chloride 30-40 c.c. of the dilute alcohol are usually sufficient. Dry the Gooch's crucible and contents at 100, and weigh. Weighings. A half-gram sample of purple Cornish stone, when treated as described above, furnished : Crucible and precipitate 8 '4519 gnus. Crucible alone . . .8 '4321 grms. Potassium fluosilicate . ..... 0'0198 grm. The weight of the potassium fluosilicate, multiplied by 0'5 17, gives the corre- sponding amount of fluorine; and by 1-06245, the corresponding amount of calcium fluoride. Hence, the sample contained the equivalent of 0'517 x 0'0198 = 0-0102 grm. or 2'04 percent, of fluorine; or 1-06245 x 0'0198 = 0'0210 grm. or 4 '2 per cent, of calcium fluoride. Disturbing Factors. This method gives more exact results with less trouble in manipulation than by the precipitation process (page 637). The principal sources of error are : (1) The presence of moisture in the air or the tubes, etc., through which the silicon tetrafluoride passes. This causes a decomposition of the silicon fluoride before it reaches the potassium fluoride, and thus causes low results. (2) The passage of sulphuric acid fumes along with the gases from the flask A. This would decompose the potassium fluoride in the receiving vessel and lead to the formation of potassium fluosilicate, giving high results. (3) Incomplete decomposition of the fluoride in the flask A ; and (4) the action of the potassium fluoride on the glass. To avoid the latter, Carnot recommends coating the surfaces of the flasks and tubes used in the determination with, say, copal resin. A blank experiment will soon show whether the reagents and method are trustworthy. When chlorides are present, a U-tube containing pumice saturated with anhydrous copper sulphate (page 554) must be interposed between E and F. If the mercury in G has a greenish film towards the end of an experi- ment, iodides are probably present. If so, start a new experiment with a tube of copper turnings interposed, with the object of arresting the iodine. If organic matter be present, this should be removed by calcination at a low temperature ; but since there is some danger of losing fluorine in the calcination, a tube of 1 If the glass is attacked the experiment is vitiated. Some object to the use of the " policeman," since it removes silica from the glass. The silica is formed by the action of the fluorides. 644 A TREATISE ON CHEMICAL ANALYSIS. freshly calcined quicklime may be interposed between the copper sulphate tube and G. The calcium oxide tube for the organic matter must follow, not precede, the copper sulphate tube, since hydrogen chloride in contact with calcium oxide might form some water, which is fatal to the success of an experiment. Offermann' s Volumetric Process. Offermann 1 proceeds by the method of molecule of potash KOH corresponds with an atom of fluorine F. If phenol- phthalein be used as indicator, the reaction is H 2 SiF 6 + 2KOH-K 2 SiF 6 + 2H 2 0. 331. The Colorimetric Determination of Fluorine Steiger's Process. Steiger 2 has devised a process for the determination of fluorine which is based on the well-known fact that the presence of fluorine has a powerful bleaching effect upon the yellow colour produced when a solution of a titanium salt is peroxidised by hydrogen peroxide. In this process a known volume of the solution containing the fluorine is mixed with a known amount of titanium. The tint of this solution is compared in a colorimeter with a second solution contain- ing an equivalent amount of titanium, and the bleaching effect is recorded. The extent of the bleaching enables the amount of fluorine to be computed as indicated below. Preparation of the Solution. Fuse 2 grms. of the powdered and dry sample say Cornish stone with 8 grms. of sodium carbonate, and extract the cold mass with hot water. Add 4 grms. of ammonium carbonate. Warm the mixture for a few minutes, and continue the heating on a water bath until the ammonium carbonate is destroyed, 3 and the volume of the liquid does not exceed 75 c.c. Standard Solution. Pipette 10 c.c. of the standard solution of titanium sulphate 4 into a 100-c.c. flask ; add 4 c.c. of hydrogen peroxide and 3 c.c. of con- centrated sulphuric acid. Make the solution up to the mark with water. 100 c.c. of this solution contain O'Ol grm. of Ti0. 2 Test Solution. Add 3 or 4 c.c. of hydrogen peroxide to the 75 c.c. obtained during the evaporation of the solution containing the sample under investigation. Then add 10 c.c. of the standard titanium solution. 5 Add about 4 c.c. of concentrated sulphuric acid to neutralise the sodium carbonate in the solution. 1 H. Offermann, Zeit. angew. Chem., 3. 615, 1890 ; F. Stolba, Journ. prakt. Chem. (1), 89. 129, 1863 ; Zeit. anal. Chem., 2. 396, 1863 ; J. Zellner, Monats. Chem , 18. 749, 1897 ; T. Haga and Y. Osaka, Chem. News, 71. 98, 1895 ; S. L. Penfield, ib., 39. 179, 1879 ; Amer. Chem. Journ., i. 27, 1879; H. Gilbert, Corresp. Ver. anal. Chem., 3. 114, 1880 ; A. E. Haswell, Rep. anal. Chem., 6. 223, 1886 ; G. B. van Kampen, Chem. Weekblad, 8. 856, 1911. S. Bein (Rep. anal. Chem., 6. 169, 1886) collects and weighs the precipitated silica. 2 G. Steiger, Journ. Amer. Chem. Soc., 30. 219, 1908 ; H. E. Mervin, Amer. J. Science (4), 28. 119, 1909. The intense reddish-brown colour furnished by titanic salts with dihydroxymaleic acid is also bleached by fluorine in an analogous way H. J. H. Fenton, Journ. Chem. Soc., 93. 1064, 1908. A. Gautier and P. Clausmann (Compt. Rend., 154. 1670, 1753, 1912) separate the fluorine as lead fluoride, and the lead is subsequently determined colorimetrically (page 339). 3 To prevent the formation of ammonium sulphate later on. Ammonium sulphate leads to high results, because it, like fluorine, bleaches the titanium solution. 4 STANDARD TITANIUM SULPHATE. Heat an intimate mixture of 1 grm. of titanium dioxide and 3 grms. of ammonium persulphate until the vigorous reaction has ceased. Drive off the ammonium persulphate by heat. Heat the residue with 20 c.c. of concentrated sulphuric acid (sp. gr. 1*84) until the acid fumes copiously. Pour the cold solution into 800 c.c. of water. The titanium oxide soon dissolves. Then add 57 '5 c.c. of concentrated sulphuric acid, and make the solution up to a litre. The solution contains O'OOl grm. of Ti0 2 and nearly O'l grm. of H 2 S0 4 per c.c. Verify the result by a determination of the Ti0 2 by one of the processes pages 203 et seq. 5 The hydrogen peroxide prevents the precipitation of Ti(OH) 4 by the alkaline carbonates. THE DETERMINATION OF THE HALOGENS. 645 When the solution is neutral, it acquires a light orange tint. Add a drop or two of an aqueous solution of sodium carbonate to neutralise the acid and discharge the colour. Add a drop or two of sulphuric acid to restore the colour, and then add enough concentrated sulphuric acid l to make a 3-5 per cent, solution. Make the solution up to 100 c.c. The solution contains O'Ol grm.of Ti0 2 . The Comparison. These two solutions should have the same tint, because they have the same amount of titanium per c.c. ; but if the test solution contains fluorine, it will have a paler tint owing to the bleaching effect of this element. The com- parison can be made in Nessler's glasses 6 cm. long, 2 '7 cm. diameter, placed over a white surface in diffused daylight. The comparison can also be made in a suitable colorimeter say Duboscq's colorimeter. The depths of the liquids in the two tubes are adjusted so that when the tubes are changed right to left, and left to right, the tints appear the same, or the left one appears uniformly darker. 2 The ratio Depth of fluorine solution Depth of solution with no fluorine 100 is then noted. This number is taken as an abscissa, and the corresponding ordinate of the standard curve, fig. 202, represents the amount of fluorine in the two-gram sample. For instance, if the standard solution had a depth 4'41 cm., and the test solution 3 ? 5 cm., the ratio is 4 41 100 = 85-0. The ordinate corresponding with the abscissa 85 is 0-00052. This latter number represents the amount of fluorine in the given sample. Supposing that 2 grms. of the sample be undergoing analysis, the sample would be reported to contain 0-026 per cent, of fluorine. The Standard Bleaching Curve. A standard solution of sodium fluoride is prepared from recrystallised, washed, and strongly ignited, pure sodium fluoride by 0-0025 0-0020 o-oo/s 0-00/0 0-0005 o-oooo 70 75 80 85 90 95 Apparent percentage of Ti 02 FIG. 202. Steiger's standard bleaching curve. 100 dissolving 2 '21 grms. of sodium fluoride in a litre of water. 1 c.c. of this solution is equivalent to O'OOl grm. of fluorine. A series of comparisons are made with solu- 1 That is, about 5 c.c. of the concentrated sulphuric acid per 100 c.c. of solution. 2 Since the left eye is usually more sensitive to the tint. 646 A TREATISE ON CHEMICAL ANALYSIS. tions containing 4 c.c. of hydrogen peroxide and 3 c.c. of concentrated sulphuric acid, against similar solutions containing no fluoride. In this manner, a curve similar to that shown in fig. 202 is obtained. It was found that with Fluorine . . . 0'0005 O'QOIO 0'0015 0'0020 0'0025 grin. Titanium . . . 86 '41 79 '24 74 '67 69 -60 66 '40 grms. The numbers in the last line represent the apparent amounts of titanium measured in terms of the ratio Depth of liquid containing fluorine , ~ ~ Depth of liquid free from fluorine Errors. The method is not satisfactory when the amount of fluorine exceeds 0*125 per cent., that is, 0025 grm. on a two-gram sample. If more fluorine be present, the amount of the sample taken must be lessened. If the temperature deviates appreciably from the temperature prevailing at the time the standard bleaching curve was made, the tint of the solution will appear paler, and high results will be obtained. Sodium and potassium sulphates in large amounts have a bleaching effect, but, if a large excess of sulphuric acid be present, the effect is not well marked. Aluminium sulphate intensifies the colour and thus gives low results. Quantities of silica less than O'l grm. have but little influence on the result. The method of preparing the sample for comparison eliminates most of the silica and alumina, and what remains has practically no influence on the result. Phosphoric acid bleaches the solution like fluorine, but the amount usually present in silicates does no harm. The tints are dependent on the proportion of sulphuric acid in the solution, and it is very important to adjust the solutions undergoing comparison so that they contain approximately the same proportion of sulphuric acid, both in making up the solutions for the standard bleaching curve, and in making the tests. According to Blum, 1 when but " little acid is present, .the colours cannot be properly matched. This is the case, for instance, with 0'003 grm. fluorine in 2 per cent, sulphuric acid solutions ; but with 10 per cent, of sulphuric acid a fairly satisfactory com- parison was possible with as much as 005 grm. of fluorine." As a result of his investigation, Merwin states that an " accuracy of 0'0002 grm. may be expected. The probable error is therefore not half as great as with standard gravimetric methods." 332. The Determination of Fluorine as Gaseous Silicon Fluoride Oettel and Hempel's Process. The simplest method for the determination of fluorine in silicates decom- posable by sulphuric acid, and exact enough for commercial requirements, is to treat the substance with concentrated sulphuric acid, and to measure the volume of the gaseous silicon fluoride evolved during the reaction. Oettel, and Hempel and Schleffler, 2 have devised special instruments called fluoro- meters for this purpose. Oettel's fluorometer is illustrated in fig. 203. Charging the Apparatus. It consists of a decomposition flask A (about 100 c.c.\ a (150-180 c.c.) burette , and a levelling tube C. The burette is filled with dry mercury, which is brought to zero by means of the levelling tube C. The thick- walled rubber tube connecting the levelling tube with the burette is 1 W. Blum, Bull. U.S. Geol. Sur., 422. 192, 1910. 2 F. Oettel, Z&it. anal. Chem., 25. 505, 1886 ; W. Hempel and W. Schemer, Zeit. anorg. Chem., 20. 1, 1899 ; W. E. Burk, Journ. Amer. Chem. Soc.,2^. 824, 1901 ; J. Shuch, Zeit. Landu: Vers. West. Oest., 9. 531, 1904; 0. R. Bohm, Oestr. Chem. Ztg., 10. 61, 1907; A. Jodlbauer, Zeit. Biol., 44. 259, 1903; H. Wislicenus, Zeit. agnew. Chem., 14. 706, 1901 ; Treadwell and Koch, I.e. THE DETERMINATION OF THE HALOGENS. 647 then clamped. From J to J c.c. of concentrated sulphuric acid 1 is poured on to the surface of the mercury. The decomposition flask must be perfectly dry. Transfer to the dry flask O3 to 0'5 grm. of finely powdered cryolite (or fluorspar) intimately mixed with twenty times its weight of calcined quartz. 2 Evolution of the Gas. Place the flask on the burette in the position shown in the diagram. Let the whole apparatus stand 15 minutes to assume the temperature of the room. Add 50 c.c. of concentrated sulphuric acid to the flask by means of a pipette. 3 Insert the stopper and fill the cup with mercury so as to seal the joint. The 100-c.c. bulb A will now be about half full. Open the clamp to connect the burette with the levelling tube. The pressure of the air is reduced by the descent of the mercury. Read the ther- mometer and barometer. Heat the flask by means of a small flame until the sul- phuric acid begins to boil. This should take about 20 minutes. In another 10 minutes the reaction will be complete. Raise or lower the levelling tube so as to keep the level of mercury some 10 to 15 cm. lower than the level in the burette. Shake the flask from time to time as much as possible by moving the whole stand, but without jerking the mercury from the mercury-sealed joints. Let the flask cool for about 2 hours, when it will no doubt have attained the temperature of the room. Raise the levelling tube from time to time during the cooling to counterbalance the diminished pressure of the cooling gas. Level the mercury in the bulb and level- ling tube. Read the burette, thermometer, and barometer. Suppose the following results have been obtained : Readings and Calculations. Weight of cryolite, 0'5261 grm. Temperature, < , at the beginning, 17 ; barometer at the beginning, 7 '7 54 mm. ; temperature, t v at the end, 18; barometer at the end, 757'4. Vapour pressure of the con- centrated sulphuric acid at this temperature is equivalent to 5 '4 mm. mercury. FIG. '203. Oettel's fluorometer. . l " The concentrated sulphuric acid used in the work is made by heating the concentrated acid of the laboratory in a porcelain dish with flowers of sulphur. The acid^is then poured off from the molten sulphur, and evaporated down to two-thirds its volume. W. Hempel, Methods of Gas Analysis, London, 322, 1892. 2 "The silicon dioxide required is best obtained by pulverising rock crystal and igm the powder in a combustion tube in a current of oxygen." W. Hempel, I.e. ; see also page 641. P Drawe (ZeU. angew. Chem., 25. 1371, 1912) uses 0'5-0'6 of felspar (instead of quartz) along with 5 grms. of anhydrous copper sulphate (page 554) to make sure that the moisture is removed. 3 Do not touch the flask needlessly, or it may be warmed and cause an expansion ot 648 A TREATISE ON CHEMICAL ANALYSIS. Hence the observed pressures of the gas at the beginning, p Q , and at the end, p v are respectively 757*4-5*4 = 752 mm. The observed volume of the gas is 85*9 c.c. By means of the well-known formula, Volume at t Q and p = volume at ^ and p l x l , ^ ^ > it follows that the volume of the gas at 18, 752 mm. pressure, is 85*9 c.c. ; volume of the gas at 17, 752 mm. pressure, is 85 '6 c.c. ; volume of the gas at 0, 760 mm. pressure, is 79 6 c.c. But 50 c.c. of sulphuric acid absorb very nearly 1'4 c.c. of silicon fluoride ; hence, 79*6 + 1'4 = 81 c.c. of SiF 4 were produced during the reaction. But 1 c.c. of silicon fluoride at and 760 mm. pressure represents 0*003436 grm. of fluorine. Hence, 81 c.c. represents 0*2814 grm. of fluorine per 0*5261 grm. of cryolite ; or the cryolite has 53*49 per cent, of fluorine. In illustration of the accuracy of the process, Oettel quotes the following results with calcium fluoride : Found . . . 0*2183 0*2349 0*1051 0*2792 0*2776 grm. Calculated. . . 0*2173 0*2350. 0*1050 0*2801 0'2782 grm. This is all that can be desired in commercial work. To get accurate results, it is highly important to exclude all traces of moisture from the generating apparatus. 1 333- The Analysis of Calcium Fluoride Fluorspar. The preceding process may be employed for all silicates which are completely decomposed by sulphuric acid, as indicated on page 646. Hence, calcium fluoride is easily treated by the process. In commercial analyses of fluorspar, calcium fluoride, silica, and calcium carbonate are usually determined. In special cases the amount of lead, iron, zinc, sulphur, and barium sulphate may be required. One rapid process is as follows : 1. Calcium Carbonate. Heat 2 grms. of the dried (110) material until it has constant weight. The loss in weight represents carbon dioxide. This amount of carbon dioxide, multiplied by 2*2748, represents the calcium carbonate. 2 2. Silica. Stahl 3 determines the silica in fluorspar by evaporating 2 grms. of the powdered and dry mineral (110) to dryness with 4 c.c. of hydrofluoric acid in a platinum dish, and repeating the operation twice with 2 c.c. of hydro- fluoric acid. Add a few drops of hydrofluoric acid and a few drops of ammonium hydroxide to precipitate the iron. The residue is evaporated to dryness and 1 P. Drawe (Zeit. angew. Chem., 25. 1371, 1912). 2 W. J. Waring, Chem. Eng., 4. 23, 1906 ; ib., 3. 65, 1905 ; L. Westerburg, Chem. Ztg., 26. 967, 1902 ; A. W. Gregory, Chem. News, 92. 184, 1905. For the complete analysis, the calcium fluoride is mixed with a known amount of powdered rock crystal, and the method of page 637 applied. Due allowance is made for the silica purposely added. Calcium carbonate is sometimes determined by digesting, say, 1 gram of the finely powdered sample with 10 c.c. of a 10 per cent, solution of acetic acid for about an hour on a water bath. The mixture is filtered through an ashless filter paper, washed about four times with water, and the filter paper "ashed" at as low a temperature as possible. The loss in weight would represent the amount of calcium carbonate in the given sample, were it not for the fact that calcium fluoride is slightly soluble in acetic acid. E. Bidtel (Journ. Ind. Eng. Chem., 4. 201, 1912) says that, in using this method with clean crystals of fluorspar, 0*0015 grm. is dissolved. Hence, if Weight of original sample 1 '0000 grm. Weight after treatment with acetic acid 0'9910 grm. Loss 0-0090 grm. Loss due to CaF 2 dissolve! 1 . . . . . .0*0015 grm. Calcium carbonate -. . .0*0105 grm. 3 K. F. Stahl, Journ. Amer. Chem. Soc., 18. 415, 1896. THE DETERMINATION OF THE HALOGENS. 649 ignited at a dull red heat to constant weight. 1 The loss in weight represents the silica. To correct for the calcium carbonate, multiply the amount of carbon dioxide by O'SOOO to get the corresponding decrease in weight due to the change from calcium carbonate to calcium fluoride. This result is added on to the silica just determined to get the "total silica." 2 3. Calcium Fluoride. Mix 1 grm. of the powdered and dry (110) material with sulphuric acid to form a "pasty" mass. 3 Heat the mixture over the Bunsen's burner, gradually raising the temperature, until all the sulphuric acid is expelled, and the mass is dry, and the weight constant. It is assumed that all the calcium is transformed into sulphate. Cool and weigh. Multiply the amount of carbon dioxide by O82 to get the corresponding gain in weight due to the change of the calcium carbonate to calcium sulphate. Subtract this number, and add that obtained for the silica to the result. The gain in weight due to the change of the calcium fluoride to calcium sulphate so obtained, multiplied by 1*3448, represents the amount of calcium fluoride. 4 Calculations and Weighings The following weighings and calculations illustrate the method of working : Dish with 2 grms. sample before ignition . Dish with 2 grms. sample after ignition . . 28-4491 grms. . 28-4440 grms. Carbon dioxide 0'0051 grm. Calcium carbonate (0-0051x2 -2748) 0'0116 grm. Loss due to change from carbonate to fluorine (0'0051 x 0'5) . 0'0025 grm. Gain due to change from carbonate to sulphate (0'0051 x 0'82) 0-0042 grm. Dish (2 grms. sample) before HF treatment Dish (2 grms. sample) after HF treatment . 28-4491 . 28-4443 grms. grms. grm. grm. Silica and change from carbonate to fluoride Loss on change from carbonate to fluoride 0-0048 . 0-0025 Silica 0-0073 grm. Dish (2 grms. sample) before H 2 S0 4 treatment 28'4491 Dish (2 grms. sample) after H 2 S0 4 treatment . . . . . 29 '9152 Gain on change fluoride and carbonate to sulphate . Gain on change of carbonate to sulphate Silica lost and change fluoride to sulphate .... Silica lost Gain from change fluoride to sulphate Calcium fluoride (1-4692x1 -3448) 1'9754 1-4661 0-0042 1-4619 0-0073 1 -4692 grms. grms. grms. grras. grm. grms. grms. 1 Bid tel (I.e.} mixes the powdered fluorspar in the platinum crucible with an emulsion of 1 gram of yellow mercuric oxide in water ; evaporating to dryness ; heating to a dull red heat ; cooling and weighing. The object of this treatment is to oxidise certain sulphides to sulphates, e.g., PbS + 4HgO = 4Hg + PbS0 4 . The residue is then treated for silica with hydrofluoric acid. 2 If alumina or silicates be present, the result is not quite correct. 3 Instead of using this process, the residue left after the treatment removing calcium carbonate by acetic acid (page 648) may be employed, and there is then no need to introduce a correction for calcium carbonate. 4 The calcium sulphate can be fused with about 10 grins, of sodium carbonate, and digested in a slight excess of hydrochloric acid. Any barium sulphate present will remain insoluble. This is F. Wohler's process (Pogg. Ann., 48. 87, 1839). See also R. F. Weinland and J. Alfa, Zeit. anorg. Ghem., 21. 45, 1899 ; K. Daniel, ib., 38. 257, 1904 ; R. Fresenius, Anleitung zur quantitativen c/iemischen Analyse, Braunschweig. I. 435, 1875. For barium and sulphur in fluorspar by fusion in a bomb with Parr's fusion mixture, see H. G. Martin, Journ. Ind. Eng. Chem,, I. 462, 1909. 650 A TREATISE ON CHEMICAL ANALYSIS. Consequently, the sample contained approximately 0*6 per cent, of calcium carbonate ; 0'4 per cent, of silica; and 98- 8 per cent, of calcium fluoride. 1 334. The Analysis of Sodium Silicofluoride. The fluorine cannot be determined as silicon fluoride, because, on treatment with acids, the silicon fluoride is accompanied by hydrogen fluoride, which etches glass. Thus, with concentrated sulphuric acid, Na 2 SiF 6 + H 2 S0 4 ->Na 2 S0 4 + SiF 4 + 2HF. Hence, on heating a known weight of the substance with concentrated sulphuric acid in a platinum dish, until the white fumes of sulphuric acid are given off, sodium sulphate and sulphates of the other bases present will remain. 2 The bases can be determined in the usual manner. Sodium silicofluoride can be titrated with N sodium hydroxide, using phenolphthalein as indicator. The reaction is represented : Na 2 SiF 6 + 4NaOH = 6NaF + Si0 2 + H 2 0. Commercial samples contain about 95 per cent, of sodium silicofluoride. The loss on ignition of sodium silicofluoride does not represent the " water " because, on heating, this salt decomposes into sodium fluoride and silicon fluoride: Na 2 SiF 6 ->2NaF + SiF 4 . To determine the water, use the method indicated on page 572, namely, fusion in glass tubes with lead oxide. 3 Hydrofluosilicic acid is a by-product in the manufacture of artificial manures. The acid is treated with soda or common salt in order to precipitate sodium silicofluoride. Hence the commercial salt may be contaminated with chlorides, which can be determined as described below. 335- The Properties of Silver Chloride. Chlorides are determined by adding a solution of silver nitrate to an acidified solution of the chloride. The white curdy 4 precipitate of silver chloride is washed, dried, and weighed. In 1857, Mulder 5 pointed out that the solubility of silver chloride in the mother liquid is sufficient to disturb all quantitative analyses 1 Instead of using a separate portion of the sample, the residue from the silica determination may be mixed with 4 c.c. of hydrofluoric acid and a few drops of nitric acid. Cover the crucible with a lid and keep the mass warm for about half an hour. Then evaporate to dryuess. The contents of the crucible should be white. If the contents of the crucible are not white, the iron oxide is converted into fluoride by the addition of a few drops of hydrofluoric acid, The iron, zinc, and lead fluorides are then extracted by the addition of 10 c.c. of a solution of ammonium acetate. The mixture is digested on a water bath for half an hour. Filter, and wash by decanta- tion with hot water containing a little ammonium acetate solution, and finally with hot water. Ignite, and weigh as calcium fluoride. This can be verified by converting the fluoride into sulphate. The ammonium acetate solution is made by neutralising 400 c.c. of 80 per cent, acetic acid with ammonium hydroxide, adding 20 grms. of citric acid, and making the solution up to a litre with concentrated aqueous ammonia. About 0'0012 grm. CaF 2 per gram of fluorspar is lost by the ammonium acetate treatment. 2 For the direct determination of silica and fluorine, use the process indicated on page 637. For the transformation into chlorides, use hydrochloric acid (1 acid, 2 water by volume) F. Stolba, Chem. Centr. (3), u. 595, 1880. 3 H. Rose, Ausfuhrliches Handbuch der analytischen Chemie, Braunschweig, 2. 667, 1851 ; P. Jannasch, Praktischer Leitfaden der Gewichtsanalyse, Leipzig, 357, 1 904 ; E. Hintz and H. Weber, Zeit. anal. Chem., 30. 30, 1891. 4 If but small quantities of silver chloride be present, the solution only becomes turbid. The suspended particles then settle very slowly. 5 Gr. J. Mulder, Die Silberprobirmethoden, Leipzig, 19, 311, 1859 ; Chem. News, 4. 99, 125, 137, 204, 231, 297, 321, 1861 ; 5. 2, 1862. THE DETERMINATION OF THE HALOGENS. 651 based upon the precipitation of silver chloride, and Cooke l considered that about O'OOOl grm. of silver chloride is dissolved per 100 c.c. of liquid used in washing. It is now considered to be a well-established fact that freshly precipitated silver chloride is soluble in water to the extent of about 0*000154 grm. per 100 c.c. at 21, and 0-00217 grm. at 100. 2 The solubility of freshly precipitated silver chloride steadily diminishes to about a milligram per litre when the precipitate is left standing in contact with the mother liquid. This change in the solubility is generally supposed to be due to the fact that the solubility of many salts, which are only slightly soluble in water, is affected by the size of the grain. The smaller the grain, the greater the solubility. 3 Silver chloride when first precipitated is in a very finely divided condition, and the small grains grow into larger grains at the expense of the very smallest granules, on long standing. The solution is also generally warmed, so as to coagulate any colloidal silver chloride which may be formed, and so prevent a turbid filtrate passing through the filter paper. The presence of a small excess of either silver nitrate or sodium chloride reduces the solubility of the precipitated chloride. 4 A large excess of hydro- chloric acid, alkaline chlorides or nitrates, metallic chlorides, or of silver nitrate augments the solubility of silver chloride. Thus, the solubility of silver chloride in different strengths of hydrochloric acid at 21 is, according to Whitby, as follows : Hydrochloric acid ... 1 5 10 per cent. Silver chloride . . . 0'00154 0'0002 0'0033 0'0555 grm. per litre. The lower solubility of silver iodide and bromide has led to the use of soluble iodides and bromides in place of chlorides for precipitating silver from solutions (page 333). The solubilities of the three halides in water at 20 are : AgCl. AgBr. Agl. 0-00154 0-000084 0-0000028 The silver chloride during precipitation is inclined to carry down, or " occlude " or " adsorb," silver nitrate 5 and sodium chloride. These substances are associated in some way with the precipitated chloride, so that the impurities are not removed by washing. In general, the longer the precipitate remains in contact with the mother liquid, the greater the difficulty in removing the adsorbed salts. Precipitates formed in dilute solutions are generally more amenable to washing than precipitates formed in concentrated solutions. White silver chloride dries (in the dark) to a white pulverulent mass which becomes yellow when heated. Silver chloride fuses at 260 to a transparent yellow liquid which attacks platinum very rapidly. At a high temperature, silver chloride commences to volatilise. When the molten mass is cooled, it forms a colourless or pale yellow mass. Silver chloride is easily reduced to the metal when heated in contact with organic matter, and hence the difficulty in getting rid of the filter paper used in separating the precipitate from the mother liquid. 1 J. P. Cooke, Amer. J. Science (3), 21. 220, 1881 ; Chem. News, 44. 234, 1881 ; D. Lindo, ib., 45. 193, 1882. 2 A. F. Holleman, Zeit. phys. Chem., 12. 132, 1893 ; F. Kohlrausch and F. Rose, ib., 12. 234, 1893 ; F. Kohlrausch, ib., 50. 355, 1905; W. Bottger, ib., 46. 325, 1903 ; 56. 83, 1906 ; T. W. Richards, Zeit. anarg. Chem., 6. 108, 1894 ; F. Field, Chem. Neivs, 3. 17, 1861 ; G. S. Whitby, Ber. Internal. Congress App. Chem., 7. 12, 1910 ; W. Bottger, Zeit. angew. Chem., 25. 1992, 1912. 3 W. H. Wollaston, Phil. Trans., 103. 51, 1813 ; W. Ostwald, Zeit. phys. Chem., 34. 495, 1900 ; G. Hulett, ib., 37. 385, 1901 ; 47. 357, 1904. 4 C. Hoitsema, Zeit. phys. Chem., 2O. 272, 1896. 5 J. S. Stas, (Euvres Completes, Bruxelles, i. 337, 1894. 652 A TREATISE ON CHEMICAL ANALYSIS. 336. The Gravimetric Determination of Chlorides. The silicate is fused with sodium carbonate, 1 as described for the determina- tion of sulphur (page 617). The clear aqueous solution of the fused mass, or an aliquot portion of the aqueous solution, is acidified with nitric acid in the cold. 2 Heat the solution to about 60. Gradually add, with constant stirring, 3 about 5 c.c. of an aqueous solution of silver nitrate. 4 When the precipitate has settled, add a few more drops of the silver nitrate solution, and if a precipitate be formed, more silver nitrate must be added. When the addition of silver nitrate no longer produces a precipitate, heat the mixture to 60, 5 and let the mixture settle in the dark, 6 say, overnight. Decant the cold solution through a (7'5-cm.) filter paper. Wash by decantation a few times with water just acidified with nitric acid. Transfer the precipitate to a filter paper, 7 and wash with acidulated water until a drop of the filtrate gives no turbidity with hydro- chloric acid. Dry the precipitate between 90 and 95, and separate it from the paper. Preserve the precipitate in a covered watch-glass. Incinerate the paper in a small weighed porcelain crucible not platinum. Do not allow the paper to burn with a flame. When the carbon has disappeared, the chloride will be reduced to metal. Add a few drops of nitric acid, and warm the crucible, to dissolve the metallic silver ; add a few drops of hydrochloric acid ; evaporate the solution to dryness ; and then transfer the precipitate in the watch-glass to the crucible. Heat the crucible and contents gently to about 180, but not sufficiently high to fuse the silver chloride (260). Cool, and weigh as silver chloride. If the silver chloride so obtained is quite soluble in ammonia, the work is finished ; if an insoluble residue remains, proceed as follows : Purification of the Silver Chloride Precipitate. In order to purify the precipitate from silica, alumina, etc., if any should be present, dissolve the precipitate in the crucible in aqueous ammonia ; filter the solution to separate the insoluble residue ; wash the residue with ammonia ; acidify the solution with nitric acid ; add a drop of silver nitrate ; filter and wash the precipitated silver chloride as before. This precipitate can be collected on filter paper, but a Gooch's crucible with ignited asbestos is preferable. See page 430. Calculations. The weight of the silver chloride multiplied by 0*24738 represents the corresponding amount of chlorine. The sum of the different constituents in a silicate analysis may exceed the limit 100 0*5 if appreciable amounts of chlorine be present, because some of the chlorine will have been reckoned as oxygen in evaluating the bases, etc. By the same process of reason- 1 All the reagents should be tested for possible and probable contamination with chlorides. ' 2 H. S. Washington (Manual of the Chemical Analysis of Rocks, New York, 160, 1904) prefers to decompose the silicate with a mixture of nitric and hydrofluoric acids (free from chlorides) in a platinum basin ; boil for an hour ; filter through a platinum or rubber funnel into a platinum basin ; and precipitate the silver chloride in the presence of an excess of nitric acid. 3 So as to reduce as much as possible the adsorption of salts by the precipitate. 4 SILVER NITRATE SOLUTION. Dissolve 3-4 grms. of silver nitrate in 100 c.c. of water (E). * E. Alefeld (Zeit. anal. Chem., 48. 79, 1909) adds 5 c.c. of ether to the solution before adding silver nitrate, in order to coagulate the precipitate. Gooch's crucible (packed with asbestos) is used for the filtration. It is moistened with ether before the suction is applied. The filtration may here be done immediately after the precipitation. The precipitate is washed with water. M. Whittel (Chem. Ztg., 7. 559, 1884) recommends a drop of chloroform to coagulate the precipitate. 6 White silver chloride soon becomes violet and finally dark brown, almost black, when exposed to the light. Some claim that there is then an appreciable loss, possibly owing to the formation of a sub-chloride. G. J. Mulder, Die Silberprobirmethoden. Leipzig, 59, 1859. 7 If any chloride sticks tenaciously to the beaker, dissolve it in ammonia, transfer the solution to a weighed crucible, acidify with nitric acid, evaporate to dryness, dry at 150, and weigh. Add the weight to the main mass of silver chloride. THE DETERMINATION OF THE HALOGENS. 653 ing as was employed for fluorine, every n per cent, of chlorine obtained must be multiplied by 0*2256 in order to get the equivalent amount of oxygen, and the product is appended to the analysis in this form : " Less x per cent, of oxygen, equivalent to n per cent, of chlorine." Jannasch's Filter Tube. Instead of using a Gooch's crucible, the stoppered tube as recommended by Jannasch l may be employed fig. 204. This is fitted with a layer of glass-wool, from 0'5 to I'O cm. thick, in the neck, B ; above this is a layer, (7, of asbestos such as is used for the Gooch's crucible, 0*5 to I'O cm. thick. The narrow end of the tube is fitted into the filtration flask by means of a one-hole rubber stopper, fig. 49. Water is run through the tube c FIG. 204. Jannasch's filter and weighing tube. until any particles of loose glass-wool have all been washed into the flask. Dry the tube for two or three hours at 150, and when cold (50-60) insert the stopper A, and cool in a desiccator'. Insert the stopper Z>, and weigh. Filter the silver chloride precipitate through the tube, and dry at 180-200, as indicated above. Insert the stoppers as before, and weigh. The increase in weight represents the silver chloride. The Determination of Soluble Chlorides in Clays. The chlorine present in the soluble salts in clays can be determined gravimetrically as just described ; or volumetrically, 2 as indicated on pages 76 and 79 ; or by the turbidity method. The turbidity method is conducted in a similar manner to the turbidity process (page 631) for the determination of sulphates. 3 A solution of silver nitrate must be substituted for barium chloride ; and sodium chloride substituted for calcium sulphate. 337. The Determination of Silver. The determination of silver is a reciprocal process to the determination of chlorides. In the latter case a silver salt is added to the soluble chloride ; and in the former case a soluble chloride is added to the solution containing silver. 4 In each determination the precipitated silver chloride is prepared for weighing in the same way. The weight of the silver chloride multiplied by 7526 gives the corresponding amount of silver. Benedickt and Gans' Gravimetric Process. 5 Add an excess of potassium iodide to a solution containing an excess of about 10 c.c. dilute nitric acid. 6 Any silver present is precipitated as silver iodide. If too great an excess of potassium iodide 1 P. Jannasch, Praktischer Leitfaden der Gewichtsanalyse, Leipzig, 11, 1904. Neither Jannasch's tube nor the Gooch's crucible is employed for collecting the first precipitate if the silver chloride is to be freed from silica. 2 F. Muck (Zeit. anal. Chem., 22. 222, 1883), in the presence of organic matter, evaporates the fluid to dryness and moistens the residue with a solution of potassium or sodium hydroxide free from chlorine. Then warm the mass with chlorine-free potassium permanganate until the colour of the permanganate remains green. A drop of alcohol will destroy the green colour. Filter and wash. The chlorine is determined in the filtrate in the usual manner. ' 3 T. W. Richards, Internat. Cong. App. Chem., 8. i, 423, 1913; P. A. Guye, Journ. Chim. Phys., 10. 145, 1913 ; F. Meyer and A. Stabler, Zeit. anorg. Chem., 77. 255, 1913. 4 It appears to be advantageous to precipitate the silver by the addition of a soluble bromide, since silver bromide is less soluble than silver chloride ; and still more advantageous to pre- cipitate the silver as silver iodide, since this salt is even less soluble than silver bromide. 5 R. Benedickt and L. Gans, Chem. Ztg., 16. 4, 12, 44, 1892 ; W. Hampe, ib., 18. 1899, 1894. 6 If antimony be present, the addition of tartaric acid prevents its contaminating the pre- cipitated silver iodide. 654 A TREATISE ON CHEMICAL ANALYSIS. be added, some silver iodide may remain in solution. The solution is then heated. If any lead be present, the precipitated lead iodide will be decomposed by this treatment, soluble lead nitrate will be formed, and iodine will be set free. The latter is volatilised. When the solution is free from the colour of iodine, the silver iodide is filtered off, washed, and weighed as in the case of silver chloride. The weight of silver iodide multiplied by 0*4594 gives the corresponding amount of silver. Hampe considers this " the most accurate of the wet processes for the determination of silver." It can be used for estimating the amount of silver in lead compounds. For the volumetric processes for silver, vide pages 76 and 79 ; for the cupellation process, page 326 ; and for metallic precipitation with cadmium or zinc, 1 follow the method of page 307. Whitby's Colorimetric Process. 2 This process is based on the development of a brown or yellow colour when a solution of a silver salt is heated with a little sodium hydroxide and an organic substance (dextrine, gum arabic, cellulose filter paper, starch, carie-sugar, etc.). The intensity of the coloration is propor- tional to the amount of silver present. The colour is sensitive enough to show the presence of silver in 50 c.c. of a solution containing one part of silver in 25,000,000 parts of solution; and ft is possible to estimate 0'000002 grm. of silver in 50 c.c. of solution, i.e. 0'00004 grm. of silver per litre. 3 Standard Solution. Pipette 5 c.c. of the silver solution 4 into a 150-c.c. beaker. Dilute the solution to 50 c.c. Add a few drops of a concentrated solution of cane-sugar. Immerse the beaker in a bath of boiling water for 2 minutes ; add 5 drops of an N-solution of sodium hydroxide, and heat the mixture for 20-30 seconds after the colour has appeared. 5 Cool the solution and transfer it to a burette. Test Solution. The silver solution under investigation is appropriately diluted and treated as described for the standard solution. 6 50 c.c. of the solution are transferred to a Nessler's glass. The tint of this solution should be nearly the same as the tint of the standard solution. If not, either the standard or the test solution must be diluted, and the operations (development of colour) repeated. The Comparison. The standard solution is placed in a burette and run into a second Nessler's glass until the colour is of the same intensity in both vessels. The number of c.c. of the standard solution required to produce the same intensity of colour as that of the 50 c.c. of the test solution furnishes the necessary data for the calculation of the amount of silver in the test solution (pages 200 and 204). Whitby says that ammonia should be absent, but minute traces of copper, zinC) mercury, bismuth, cadmium, and lead, in quantities insufficient to give an appreciable precipitate with sodium hydroxide, do not interfere with the result. 1 A. W. Clasen, Journ. prakt. Chem. (1), 97. 217, 1866. For reduction with hydroxylamine, A. Lainer, Monats. Chem., g. 533, 1888 ; with aluminium, N. Tarugi, Gaz. Chim. Ital., 33. ii. 223, 1904; with cobalt, C. Goldschmidt, Zeit. anal. Chem., 45. 87, 1906. 2 See page 326 ; G. S. Whitby, Ber. Internat. Congress App. Chem., 7. 12, 1910. 3 F. Jackson, Journ. Amer. Chem. Soc., 25. 992, 1903 ; W. Bottger, Zeit. angew. Chem., 25. 1992, 1912. 4 SILVER NITRATE SOLUTION. The standard silver nitrate solution should contain the equivalent of about '00001 grm. of silver per c.c. The standard solution should be approxi- mately the same strength as the test solution, and it should be prepared the same day the determination is made ; otherwise its tint deepens if it be allowed to stand for any length of time. 5 Do not heat more than about 2 minutes. 6 If no coloration appears after the solution has been heated 2 minutes, it can be assumed that silver is absent. THE DETERMINATION OF THE HALOGENS. 655 338. The Determination of Iodine. Iodine, to the extent of 0'01-0*2 per cent., is not uncommonly present in mineral phosphates. It can be detected by digesting the finely powdered phosphate with concentrated sulphuric acid in a flask so arranged that a current of air is aspirated through the flask into a solution of carbon disulphide or chloroform. The pink or violet colour of the solution shows the presence of iodine. The iodine is of no known technical importance in the silicate industries. Hence, the subject can be dismissed very quickly. Thiercelin 1 determines iodine in phosphates, etc., by digesting a large quantity of the substance, say 100 grms., with a mixture of equal parts of sulphuric acid and water in a 500-c.c. flask or retort arranged so as to conduct the vapours into potash lye contained in a suitable absorption tube, say, fig. 194. The mixture in the flask is boiled (about 30 minutes) until all the vapours of iodine have been driven into the absorption tubes. The solution of iodine in the potash is treated with an excess of sulphurous acid to convert the iodates, etc., into iodides : S0 2 + 4KOH + 1 2 = 2KI + K 2 S0 4 + 2H 2 0. The iodine absorbed by the lye can "be determined as silver iodide by adding an excess of silver nitrate to the solution neutralised with nitric acid, and then adding an excess of nitric acid, in which silver iodide is almost insoluble. 2 The precipitated silver iodide is treated as if it were silver chloride (page 652) ; but silver iodide is not so readily reduced as the chloride, and it may even be safely ignited with the filter paper. The weight of silver iodide multiplied by 0'5405 gives the corresponding amount of iodine. Sometimes iodine is determined as cuprous iodide 3 by adding a solution of ammonium cuprous chloride 4 to the solution of the iodide acidified with hydrochloric acid, and containing a sufficient excess of ammonium chloride to prevent the precipitation of cuprous chloride by the aqueous solution. The precipitated cuprous iodide is collected on a tared filter paper, or, better, in a Gooch's crucible, dried, and weighed as Cul. This method, however, offers no particular advantage over the silver iodide process. 1 M. Thiercelin, Bull. Soc. Chim. (2), 22. 435, 1874. 2 If the solution be first acidified with nitric acid, there is a risk of liberating iodine. If the solution be coloured with iodine, it is best to add a little sulphurous acid to decolorise the solution 3 F. Mohr, Zeit. anal. Chem., 12. 366, 1873 ; H. Zenger, Archiv Pharm. (3), 3. 137, 1873 ; J. Krutwig, Ber., 17. 341, 1884 ; E. Fleischer, A System of Volumetric Analysis, London, ' 4 Copper turnings are digested in a solution of cupric chloride and ammonium chloride acidified with hydrochloric acid. Sufficient ammonium chloride must be present to prevent the reagent giving a precipitate with water. CHAPTER XLIV. THE RATIONAL ANALYSIS OF CLAYS. 1 339- Clays. THIS chapter is "brought forward" from the second volume of this work in order that the analytical processes may be confined in one volume. Hence, a few explanatory words may be desirable. Kaolin is a general word applied to china clay rock, and to china clay obtained by the washing of china clay rock. The term clay not only covers all the different varieties of clay, but also a sub- stance approximating Al 2 Oo.2Si0 2 .2H 2 in composition. The last-named formula is supposed to represent a constituent common to all clays, and which has hence been termed "clay base," "clay substance," "clay proper," "clay matter," " ideal clay," etc. The word kaolinite is applied to a crystalline mineral with the empirical formula Al 2 3 .2Si0 2 .2H 2 0. We are almost sure that an amorphous or non-crystalline substance approximating the same composition exists in most, probably in all clays. This is here called clayite ; the new term is required because it possesses several properties different from kaolinite. If it were not for the dread of tnultiplying terms, a word would be coined to include both clayite and kaolinite. I use the term argillaceous matter for that heterogeneous mixture which is removed from a clay by the treatment employed in the method of " rational analysis." It includes clayite, kaolinite, and other minerals. 2 Clays are produced by the weathering and disintegration of various kinds of rocks which occur on or near the surface of the earth. If the clay occurs where it has been formed, it is called a primary or residual clay ; and if the clay has been carried elsewhere it is called a transported or secondary clay. The trans- ported clays are the washings and sweepings of the hills, which Nature has accumulated as her rubbish-heaps in convenient places. Kaolinite or clayite and quartz are common to nearly all clays, while felspar and mica are so common in clays that these four constituents can reasonably be regarded as the primary constituents of clays. The list of secondary constituents, which usually occur in relatively 'small quantities, is very extensive. In the first place, we find frag- ments of various kinds of rock granite, trachyte, rhyolite, syenite, etc. ; and in the second place, fragments of numerous minerals have been reported in clays : Anatase, andalusite, apatite, augite, bronzite, calcite, chlorite, corundum, cyanite, diaspore, dolomite, dumortierite, enstatite, epidote, fluorspar, garnet, gilbertite, glauco- phane, gypsum, hematite, hornblende, hydrated alumino-silicates related to clayite, 1 Much of the original matter in this chapter is the result of experiments made on the rationale of the method of rational analysis during the investigation of the composition of dusts for Sir Henry Cunynghame, Chairman of the Committee appointed to Inquire into the Causes and Remedies for Coal Dust Explosions in Coal Mines. The reader must thank the Chairman for permission to use such experiments as I thought fit for this chapter. Most were done by my assistants, Messrs A. D. Holdcroft and C. Edwards, during 1911. 2 J. W. Mellor, Trans. Eng. Cer. Soc., 8. 23, 1908 ; Pot. Gas., 34. 927, 1909. 656 THE RATIONAL ANALYSIS OF CLAYS. 657 hypersthene, ilmenite, leucoxene, lignite, limonite, magnesite, magnetite, nepheline, nontronite, olivene, prehnite, pyrites, rutile, scapolite, selenite, serpentine, siderite, sillimanite, spinel, staurolite, titanite (sphene), topaz, tourmaline, vivianite, zeolite, zircon, zoisite. The secondary constituents are usually present in small quantities, and in a more or less advanced state of decomposition. The rational analysis is an attempt to express the composition of a clay in terms of the primary constituent minerals. 340. The Separation of Minerals by Treatment with Chemical Reagents. Clay is a heterogeneous mixture of several different minerals. The chemical methods for estimating the minerals in clays, etc., are based on differences in the rates at which the various minerals are attacked by different reagents. The separation of the constituents of a clay which are soluble in water (page 630) is a comparatively simple example, whereas the digestion of clay with hydrofluoric acid is more complicated. Quartz, felspar, leucite, and minerals rich in silica are in general quickly decomposed by treatment with hydrofluoric acid under conditions where sillimanite, staurolite, topaz, tourma- line, zircon, and the titanium minerals are but little attacked ; while mica, hornblende, and sphene are partially decomposed. 1 Unfortunately, the differences in the rates of attack of the more important minerals in clays by known reagents are not usually great enough to allow a perfect separation. Before one mineral is completely decomposed, others will have succumbed to the attack. Hence, the products of decomposition of the one mineral will be more or less con- taminated with products derived from the partial decomposition of other minerals. Hydrochloric acid, for example, is frequently employed to remove iron oxides, and calcium, magnesium, and iron carbonates, from clays. But if the clay contains olivene, serpentine, chlorite, nepheline, epidote, leucite, and zeolites, the acid may do too much work, for these minerals are more or less decomposed by the same treatment. Similar remarks apply, mutatis mutandis, to the action of sulphuric acid. Compare with page 525. The analysis of clays, etc., by digesting the powdered sample in sulphuric acid, followed by solutions of caustic alkalies, etc., was common enough at the end of the eighteenth and at the beginning of the nineteenth centuries, as is evidenced by the analyses of Bergmann, Hochmeimer, Kirwan, Klaproth, Lampadius, Schonbauer, Vauquelin, Westrumb, etc. 2 In 1835, Forchhammer 3 showed that, when certain clays are digested in sulphuric acid, the ultimate 1 F. Fouque, Compt. Mend., 75. 1090, 1872; 79. 869, 1874; F. Fouque and M. Levy, Mintralogie Micrographique, Paris, 116, 1879 ; J. B. Mackintosh, School Mines Quart., 7. 384, 1886; Journ. Amer. Chem. Soc., 8. 210, 1886; Chem. News, 54. 102, 1886. Mackintosh believes that, the denser the mineral, the less is it attacked by the acid ; and that the rate of attack by hydrofluoric acid does not depend upon the proportion of silica, but rather on the nature of the bases. K. Obbeke, Neues Jahrb. Min., I. 455, 1881 ; J. Hazard, Zeit. anal. Chem., 23. 158, 1884 ; Chem. News, 50. 33, 1884. 2 T. Bergmann, Opuscula Physicaet Chemica, Holmiae, 2. 399, 1780 ; C. F. A. Hochmeimer, Mineraloaische Chemie, Leipzig, 1792 ; R. Kirwan, Physisch-chcmische Schriften, Berlin, 1783; M." H. Klaproth, Beitrdge zur chemischen Kenntniss der Miner alkdrper, Berlin, 1795; W. A. Lampadius, Handbuch zur chemischen Analyse der Miner alkdrper, Freiberg, 1801 ; J. A. Schonbauer, Neue analytische Methode die Mineralien und Hire Bestandtheile Richtig zu Bestimmen, Wien, 1805 ; L. N. Vauquelin, " Anleitung zur chemischen Analyse der Fossilen, Scherer's Journ., 3. 410, 1799; J. F. Westrumb, Kleine physisch-chemische Abhandlungen, 3 G.' Forchhammer, Pogg. Ann., 35. 331, 1835 ; Ann. Mines (3), 7. 517, 1835. 42 6 eg A TREATISE ON CHEMICAL ANALYSIS. composition of the soluble portion corresponds very nearly with the formula A10 2SiO .2H 2 0; while the residual insoluble portion is mainly felspar and quartz. Hence, Forchhamrner attempted to determine the mineralogical com- position of certain clays by successive treatment with different reagents. After leaching out the constituents soluble in hydrochloric acid, he extracted the residue alternately with hot sulphuric acid and an aqueous solution of sodium carbonate. This treatment was supposed to remove the argillaceous matter. We shall, however, soon see that a variety of other minerals mica, felspar, fluorspar,' etc. are more or less attacked by this treatment. Forchhammer's process has been more or less modified by Brongniart and Malaguti, Seger, 1 etc., and extensively employed principally in Germany with Brongniart's designa- tion, Vanalyse rationelle. The so-called rational analysis is a first approximation or attempt to represent the mineralogical composition of a clay in terms of the three important minerals kaolinite, felspar, and quartz and hence it can also be called a "mineralogical analysis," vide page 524. The essential features of the process are : Remove the argillaceous matter by digestion with hot concentrated sulphuric acid, and wash the residue alternately with alkaline lye and hydrochloric acid. The difference between the weight of the dried residue and the original sample represents argillaceous matter. The residual felspathic and quartz detritus is treated with hydrofluoric acid and the alumina determined. The corresponding amount of felspar is computed, and the quartz is determined by difference. One modus operandi is as follows : 341. The Rational or Mineralogical Analysis of Clays. Preliminary Treatment. If the clay contains soluble salts, they are removed by the process indicated on page 630 ; if it contains colloidal silica, this can be removed as indicated on page 668 ; and if it contains carbonates calcium, magnesium, iron or free iron oxide, these should be removed by digesting, say, 5 grms. with dilute hydrochloric acid (page 525) before treatment with sulphuric acid. The hydrochloric acid also removes colloidal aluminium and iron hydroxides, and it attacks some of the argillaceous matters, as well as apatite, haematite (powdered), magnetite, leucite, sodalite, nepheline, olivine, wollastonite, and most of the zeolites ; while scapolite, plagioclase, serpentine, chlorite, and the more compact iron oxides are but slowly attacked. Hence, Weinschenk 2 recommends chloric acid, in place of hydrochloric acid, to avoid breaking up silicates when the carbonates are decomposed. The washed residue is dried and weighed. The loss in weight represents the matters soluble in the acid. The carbon dioxide, if desired, can be determined by weight as described later, and the lime, magnesium, and iron determined in the acid solution if necessary. If the carbon dioxide be determined by weight (page 553), the washed residue, or a new sample, can be treated with sulphuric acid, etc., as described below, and an allowance made when the clay substance is determined. In illustration, Altofts' shale furnished 4'42 per cent, of carbon (page 546) and 2-44 per cent, of carbon dioxide (page 553). The solution of the shale in hydrochloric acid contained much iron, and very little lime or magnesia. Hence it was inferred that the shale contained ferrous carbonate. One part by weight of carbon dioxide corresponds with 2'633 of ferrous carbonate, and accordingly 1 A. Brongniart and J. Malaguti, Arch. Mus. Hist. Nat., 2. 219, 1841 ; J. Aron, Notizblatt, 10. 226, 1874 ; M. Finkner, ib., 3. 119 : 1867 ; C. Bischof, ib., u. 120, 1875 ; H. A. Seger, ib., 12. 245, 1876 ; Tonind. Ztg., i. 272, 1877. B ^ Wei ? Schei ? k ' Die d^teinbildenden Mineralien, Freiburg, 11, 1907 ; New York, 150, i 8. 233, 1803; H. A. Seger, Notizblatt, 12. 245. 1876; Gesammelte Schmften, Berlin, 42, 1896; Easton, Pa., i. 53, 1902; K. Langenbeck The fermstry of Pottery , Easton Pa., 9, 1895; W. Jackson and E. M. Rich, Journ. Soc. Chem. Ind., 19. 1087, 1900 ; A. Sabeck, Ohem. Ind., 25. 90 1902 THE RATIONAL ANALYSIS OF CLAYS. 663 reagents, as well as the duration and temperature of the digestion with sulphuric acid, etc. It is therefore necessary to arrange conditions such that the felspar is affected as little as possible. It is a disadvantage to have the sulphuric acid too dilute, for it will then change considerably in volume during the digestion, and particles of the clay may be left on the sides of the dish, as water evaporates, and consequently remain unattacked. 1 Sufficient acid must also be present to ensure the decomposition of all the argillaceous matters. 2. Mica, Nepheline, and Hornblende. If mica be present in the clay, as it usually is, it will be more or less attacked by the treatment. Some varieties of mica phlogopite, for instance will be almost completely decomposed along with the kaolinite. As in the case of felspar, the experimental evidence shows that the attack is dependent upon the type of mica present in the clay, the concentra- tion of the reagents, the time of heating, and the state of subdivision of the mica. Here are some results with selected samples of mica treated by the same process as that employed for- the felspars : Effect of Rational Analysis on Mica (per cent, di Size of grain (average diameter, mm.). Muscovite. Clermont Ferrant (Auvergne). Lepidolite. Rozna (Mahren). Biotite. Miask (Ural). Phlogopite. Ontario (Canada). Coarse 0-087 . 54-94 67-94 71-07 88-00 Fine 0-040 . 55-56 70-02 2 92-50 3 93-93 This shows that mica is rapidly attacked by the method of rational analysis, and part will be found in the soluble portion and part with the insoluble residue left after the sulphuric acid treatment. It might be added that similar experiments with a sample of hornblende from Moravicza (Austria), and a sample of nepheline from Vesuvius (Italy), gave : Coarse. Fine. Hornblende dissolved . . . ,.179 15 '09 per cent. Nepheline dissolved . . . . 98 "63 99 '62 3. Quartz. The acids appear to have no appreciable effect on the quartz, but quartz is affected by the caustic soda treatment. Brongniart and Malaguti substituted a solution of caustic soda for Forchhammer's sodium carbonate Ine action of caustic soda on quartz has been studied in some detail. iclis, in opposition to earlier workers, maintained that quartz is not appreciably affected by a 10 or 15 per cent, solution of caustic soda when digested on a water bath with this reagent. This observation also contradicts the later work 1 C. Bischof (1884) and H. Hecht (1895) recommended an acid 1:6; H. A. Seger C. Bischof (1904), G. Lunge (1908), and A. Sabeck (1903), 1:3; H. A. Seger Leopold (1905), and E. Berdel (1903), 1:2; and B. Zschokke (1902), 1:1. 2 Another sample gave respectively 46 '3 and 48 "2 per cent. 3 Another sample gave 60 '97 and 89 '2 per cent, respectively. * W. Michaelis, Chem. Ztg., 19. 1422, 2002, 2296, 1895; C. Rfmmelsberg, 112. 182, 1861 ; J. N. Fuchs, ib., 31. 577, 1834 ; G. Jenzsch, ib 126. 49/ ISbE > ; Berq. Hutt. Ztg., 14. 107, 1858 ; 0. Maschke, ZeiL deut. Geol. Ges., 7- 438 j. J 5 . 5 ' ^JSS Jwrn. prakt. Chem. (1), 98. 14, 1866 ; C. R. Fresenius, Anleitung zurquantitativenchemischen Analyse, Braunschweig, \.2Q7, 1873 ; 2. 338, 1877 ; H. Rose, HaMuchderanalytischenChemie, Leipzig, i. 751, 1871. 664 A TREATISE ON CHEMICAL ANALYSIS. of Lemberg and Rinne, 1 who have shown that finely divided quartz is readily dissolved by this treatment. We must therefore assume that Michaelis was dealing with comparatively coarse grains of quartz, which are fairly resistant. Lunge and Millberg have shown that the solubility of quartz depends upon the fineness of the particles, the concentration of the reagent, and the time of heating e.g. a 15 per cent, solution of caustic potash dissolves from I'O to 1'5 per cent, of finely divided quartz in an hour's digestion, and 1'8 to 2'0 per cent, after two hours' digestion under conditions where a 15 per cent, solution of sodium carbonate dissolved only mere traces of the quartz. The attack is nearly ten times more vigorous when the quartz is in an extremely fine state of division. The experiments indicated above were repeated on a sample of clear rock crystal, and on a sample of flint from Dieppe (France). The results were : Rock crystal Flint .' Coarse. 0-96 2-52 Fine. 6-40 12-10 The experimental evidence thus shows that quartz is appreciably attacked when the clay is digested with caustic soda in order to remove the so-called FIG. 206. Residue obtained during the rational analysis of shale. (The lines are ^ mm. apart when not magnified.) colloidal silica, or the silicic acids formed during the action of sulphuric acid. )f course, if the caustic soda be dilute enough, and is only in contact with the quartz a short time, the action may be reduced to a negligible minimum ; 2 but * J. Lemberg, Zeit deut. geol. Ges., 35. 560, 1883; F. Rinne, Zentr. Mm., 334, 1904; . Kaiser, ^rA no*. Ver preuss. Rheinlande, 54. 93, 1897; Zeit, Kryst., 33. 200, 1900; ^^^r^^i^ Lun g e *ndM. Schochor-Tscherny^k 3 ^. Chem. T'frl' ^S ?' ^ Un S e T and ' Millber g> -, - 393, 425, 1897 ; P. Kreiling 2 AJ ods of Chemical Analysis, London, i. ii., 585, 1908. 7 pertent solution } reC mmendS * * P r C6nt S lution of caustic soda 5 K Berdel ^ c -)> a 6 to THE RATIONAL ANALYSIS OF CLAYS. 665 the minimum may be quite appreciable if the quartz in the clay be in a very fine state of subdivision. For this reason, Kreiling, and Lunge and Millberg, returned to the use of sodium carbonate as employed by Forchhammer, in order that the quartz and felspar may be attacked as little as possible when the pro- ducts of the action of the sulphuric acid are being removed. Fig. 206 represents a microphotograph of the residue obtained during the rational analysis of a sample of shale, and the sharp angularity of the grains shows that the attack, during the removal of the " argillaceous matter," was not very marked with the larger grains. The lines in the diagram are yj^th millimetre apart when not magnified. 343. The Composition of the " Argillaceous Matter." The term Thonsubstanz (clay substance) was applied by Senft 1 in 1867 to a hydrated aluminosilicate which is supposed to occur in all clays and to which they owe their plasticity. Senft's clay substance was, in 1840, Brongniart and Malaguti's la veritable argile ; and other writers have expressed the same idea by the use of such terms as "ideal clay," "Kaolinthon," etc. (page 656). Aron and Seger applied the term to the finest fraction obtained in the elutriation of clays, and also to the constituents decomposed by the action of concentrated sulphuric acid, thus implying that the clay substance in clays can be isolated or removed by both these processes. Kaolinite and clayite usually occur in an extremely fine state of subdivision, and consequently accumulate in the finest fraction of the elutriation. But if, as is frequently the case, other finely divided -constituents occur in the clay, they too will naturally collect in the fine fraction. These con- stituents may or may not behave like kaolinite when the clay is treated with sulphuric acid, etc. Hence the use of the one term " clay substance " for totally distinct concepts must lead to confusion and misunderstanding. 2 Not only are kaolinite and clayite attacked by the treatment with sulphuric acid, etc., but the felspar, mica, and quartz, as well as pyrites, zeolites, fluorspar, hornblende, etc., present in smaller proportions are also attacked. It is therefore necessary to investigate the composition of the product of the reaction in question. The analyses of the argillaceous matter in six different Cornish china clays on the market, dried at 110, furnished the following table : Percentage Composition of the Argillaceous Matter of China Clays. Ideal clay, Maxima. Minima. Mean of six. Al 2 3 .2Si0 2 . 2H 2 0. Silica (Si0 2 ) . 46-37 45-14 45-6 45'5 Titanic oxide (Ti0 2 ) Alumina (A1 2 3 ) I'll 40-28 0'08 38-19 0*5 39-2 39 ! 5 Ferric oxide (Fe 2 3 ). Magnesia (MgO) Lime (CaO) . Potash (K 2 0) . 1-48 1-28 1-41 1-36 0'52 o-io 0-25 0-52 I'O 0-2 0-3 1-0 Soda (Na. 2 O) .... Loss 011 ignition 0-80 14-12 0-11 11-03 0'4 12-5 14-0 1 F Senft, Die Steinschutt und Erdboden, Berlin, 236, 1867. 2 E'lutriation is discussed in the second volume of this work, and I here employ the term " argillaceous matter " for that heterogeneous mixture removed by the sulphun 666 A TREATISE ON CHEMICAL ANALYSIS. These clays contained from 80 to 98 per cent, of "argillaceous matter." As a result of similar experiments on German clays, Seger 1 drew the obvious conclusion : The main constituent of certain high-grade clays is dissolved by the sulphuric acid treatment, and it corresponds in composition with Al (2 O s .2Si0^.2N 2 0. Of course, it is not fair to apply this generalisation to all the different types of clays, since with other less pure clays very great discrepancies occur see table below for an example. In spite of this, many call the matter decomposed by the sulphuric acid " clay substance," implying that this is a universal and essential constituent of all clays ; when the fact is that clay substance derived from a wide range of different types of clay is very varied in composition. 12 Since the argillaceous matter includes the more important constituents of clay rarely falling below 50 per cent, of the total constituents, and sometimes ranging up to 99 per cent. it is almost as important to get an idea of its mineral - ogical composition as it is of the clay itself. It is also better to use the general term "argillaceous matter" in preference to "clay substance," because the latter leaves the impression that we connote a substance Al 2 3 .2Si0 2 .2H 2 0, represented by analyses resembling those in the above table. A Second Approximation to the Rational Analysis. If the ultimate composition of the clay, and of the fractions which resist and succumb to the sulphuric acid treatment, be determined, it is possible to compute numbers for the argillaceous and felspathic matters which are more promising than those given by the method of page 660. The data for Al tofts shale are indicated in the following table : Analyses of Altofts Shale. Ultimate analysis. Total. Residue. Argillaceous matter. Silica . . 51-92 3877 13-15 Titanic oxide 0-87 0-04 0-83 Alumina . . . 20-08 0-88 19-20 Ferric oxide .. ' 6-40 0-06 6-34^ Magnesia . . 1-58 o-oi 1-57 I Lime . 0-57 o-oi 0-56 J-10'81 Potash 272 0-83 1-89 | Soda . 0-86 0-41 0-45J Loss' on ignition 14-61 o-oo 14-61 Totals . 99-61 41-01 58-63 All but the silica and titanic oxide in the residue (column 3) is supposed to be derived from the felspathic minerals. In typical potash felspar, K 2 O.Al 2 3 .6Si0 2 , one part of alumina corresponds with 3'53 of silica; and in typical soda felspar, Na 2 O.Al 2 3 .6Si0 2 , one part of alumina corresponds with 3-53 parts of silica. Hence, Silica in felspar = (0'83 X 3'51) + (0'41 X 3'53) = 4'37 per cent. The difference between this silica and the total silica in the residue represents the C. R. Fresenius, Journ. praJct. Chem. (1), 57. 65, 1852 ; H. A. Seger, Notizblatt, 12. 245, 286, 1876 ; C. Loeser, Kritische Beachtung einiger Untersuchungsmethoden der Kaoline und Thone, Halle a.S., 18, 1905 ; Tonitid. Ztg., 32. 1932, 1908. 2 The identification of the aluminosilicates clayite or kaolinite, halloysite, etc. will be rll risk-linns^ -I ~\7"~1,, - TT - ** ' discussed in Volume II. THE RATIONAL ANALYSIS OF CLAYS. 667 quartz. Hence, 38 '77 less 4'37 = 34'4. Assume that the other constituents in the residue all belong to the felspar. Add these together and we get 6 '6. For the reasons indicated on page 658, we know that the argillaceous matter is contaminated with carbon and siderite. Hence, we must subtract 4 '4 per cent, of carbon and 6'4 per cent, of siderite. This leaves 58'63 - 4'4 - 6*4 = 47'8 per cent, of argillaceous matter. This gives : Carbon . . . . ' . . . .- . 4'4 per cent. Ferrous carbonate . . . . . . '. 6*4 Argillaceous matter ... . . . . . 47 *8 Felspathic detritus . 6 '6 ',' Quartz ... . - . . . . 34'4 ,, which is very close to the result obtained experimentally on page 661. Computation of the Amount of Mica in the Argillaceous Matter. I have seen rational analyses of one clay from French and German laboratories reported in terms of clay substance, mica, and quartz in the former case ; and clay substance, felspar, and quartz in the latter. Some regularly assume that the argillaceous matter contains kaolinite or clayite and mica. Given the ultimate composition of the argillaceous matter, it is sometimes but not always possible to calculate the corresponding proportion of mica and kaolinite or clayite. An average mica is taken to be : Silica . . . . . . ... . 46'2 per cent. Alumina . . . 40'2 ,, Fluxes (Fe 2 3 , MgO, CaO, K 2 0, Na 2 0) . . 12'5 to 11'2 ,, Water 2'4 ,, Here, 11*2 to 12 - 5 of fluxes represent 100 parts of mica, or 1 of fluxes represents about 8 '5 parts of mica. Given, therefore, a clay whose ultimate composition is that represented in the second column of the above table, and the ultimate composition of the residue left by the sulphuric acid treatment that represented in the third column, the mica can be calculated as follows : The difference between the ultimate composition of the residue and of the whole clay obviously represents the composition of the argillaceous matter. This is shown in the last column of the above table. 1 The iron belonging to the ferrous carbonate is here included with the fluxes, which by hypothesis are associated with the argillaceous matter as mica. Hence deduct (6 - 4xO'687 = ) 4-4 of ferric oxide from the fluxes. This leaves (10'8- 4'4 = )6'4 per cent, of fluxes. Multiply the total amount of the fluxes, 6 '4, in the argillaceous matter by 8 '5, as indicated above. The result : Mica in argillaceous matter = 6 '4 x 8*5 = 54 '4 per cent. Subtract this number, 54'4, from 47 '8, and we get an absurd result. This means that the fundamental assumption that the alkalies, etc., in the argillaceous matter are wholly derived from mica is false. With some clays the results seem to be satisfactory, and the method was used extensively by Vogt 2 and by Lavezard in their studies on the clays of France. I have considered it better to give a failure rather than to select a clay with which the process is more or less satisfactory. There must always be a consider- able amount of uncertainty owing to our ignorance of the particular type of mica in the given clay. One feature in the calculation is that it reveals how 1 The microscopic examination of a clay will generally show the presence or absence of mica usually the former. 2 G. Vogt, Bull. Soc. d'Encour. VInd. Nat. (5), 2. 633, 1897 ; Contribution d V Etude des Argiles et de la Cframique, Paris, 193, 1906 ; E. Lavezard, ib., 113, 1906 ; E. Berdel, Sprech., 36. 1483, 1903 ; K. Langenbeck, The Chemistry of Pottery, Easton, Pa., 9, 1895. 668 A TREATISE ON CHEMICAL ANALYSIS. investigators, dissatisfied with Seger's method of rational analysis, are groping for further light on this subject. Virtually we have tried to rationally analyse the results of a rational analysis, and failed. Leopold's Process of Rational Analysis. This leads us to Leopold's method of conducting the rational analysis, 1 which has been used in a modified form by others. After digesting the clay with sulphuric acid and water in the usual manner, instead of treating the residue with soda, etc., Leopold dilutes the contents of the basin with water ; cools ; adds ammonia until the solution reacts alkaline ; niters ; washes with hot water ; dissolves the precipitate in hydro- chloric acid, and makes the mixture up to 250 c.c. with water. The alumina, iron oxide, and lime are determined in aliquot portions. Multiply the amounts of alumina by 6 -336 to get the kaolinite; the iron oxide can be calculated to Fe 2 3 , or to nontronite, Fe 2 3 .2Si0 2 .2H 2 ; the lime may be computed as if it were gypsum, if the ultimate analysis shows this mineral is present in appreciable quantities ; the alkalies in the filtrate from the alumina precipitate are calculated to muscovite ; and the alkalies in the residue on the filter paper, after the removal of the ammonia precipitate, are calculated to felspar. These methods are highly artificial, and the results are affected by an uncomfortable number of assumptions. The pseudonym "rational analysis" has misled many to believe that the results are as trustworthy as, or even more trustworthy than, those -furnished by a regular ultimate analysis. Sometimes, in fact, the rational analysis appears to be highly irrational. The truth is that in some cases the rational analysis is accurate, valuable, and useful ; in others, it is inaccurate, misleading, and false. 344. Free, Combined, and Colloidal Silica. If the clay under investigation contains colloidal silica, most of this will pass into solution when the clay is digested with hydrochloric acid, and all will dissolve when the argillaceous matter is determined. Colloidal silica, if not separately determined, will then be reported with the argillaceous matter. 2 Some consider that a small quantity of colloidal silica has an important influence on the working properties of the clay. This question, however, is not, at present, under discussion. The Determination of Colloidal Silica. The colloidal silica in a clay can be determined by digesting 2 grms. of the sample with 100 c.c. of a 5 per cent, solution of sodium carbonate in a platinum dish on a water bath for about an hour. Decant off the clear, 3 and repeat the treatment with a fresh solution. Filter and wash with a hot dilute solution of sodium carbonate. Add an excess of hydrochloric acid to the filtrate ; evaporate to dryness, and separate the silica as described on page 167. Or, the residue can be washed, dried, and weighed the " loss " represents the matter removed by the sodium carbonate treatment. 4 In the latter case, much of the " colloidal " alumina would dissolve and be reported with " colloidal silica." 1 A. Leopold, Magyar Chem. Folyoirait, n. 177, 1905 ; H. Bollenbach, Chem. Ind., 31. 445, 1908 ; Sprech., 41. 340, 351, 1908 ; E. Greiner, ib., 42. 399, 413, 1909. 2 W. H. Zimmer, Trans. Amer. Cer. Soc., 3. 25, 1901 ; F. G. Pence, ib., 12. 43, 1910. A. Hambloch (Chem. Ztg., 36. 1058, 1912) uses a 5 per cent, sodium hydroxide solution. 3 Care must be taken in filtration, since small particles of clav mav pass through ordinary filter papers. 4 R. Fresenius, Juurn.praU. Chem. (1), 57. 65, 1852 ; Anleitung zur quantitativen chemischen Analyse, Braunschweig; Eng. trans., London, 2. 269, 1900; W. F. Hillebrand, Bull. U.S. Geol. Sur., 176. 109, 1900 ; F. H. Hatch, Tschermak's Mitt., 7. 308, 1886 ; J. B. Harrison, cit. page 525; F. G. Pence, Trans. Amer. Cer. Soc., 12. 43, A. J. J. Browne and 1910. THE RATIONAL ANALYSIS OF CLAYS. 669 The Results. It is advisable to run a blank experiment, and to keep the stock solution of sodium carbonate in a cerasine bottle, because, as previously indicated, a solution of sodium carbonate removes silica, etc., from glass and porcelain vessels. To illustrate the results obtained by this process the following determinations are cited (all but the first number have been corrected by sub- tracting the " correction factor " 0'02) : Per cent, dissolved. Blank experiment (or correction factor) . . . 0'02 Silicic acid (precipitated and dried at 110) Silicic acid (ignited 20 minutes on blast) . Quartz (rock crystal) coarse (page 664) * . Quartz (rock crystal) fine (page 664) x Quartz sand (ignited 20 minutes on blast) T Altofts shale . . .'.--. . Glenboig clay . . .... Halloysite (Vigoux, Indre) - . 100-00 75-62 0-11 1-23 178 1-48 2-08 10-58 If no allowance is made, there is therefore a small error due to the action of the sodium carbonate solution on quartz. We are not sure if the sodium carbonate removes all the colloidal silica. The error is much greater if the sodium carbonate solution be mixed with sodium hydroxide, as in, say, Lunge's solution. Other Methods. Sjollema 2 claims that far more accurate results are obtained by boiling the sample with a 33 per cent, aqueous solution of diethylamine. Apart from the cost of the reagent, and the fact that the solution "bumps" badly when boiling, rather poor results have been obtained in my laboratory with this process. Hermann 3 recommends boiling the clay with a O'l per cent, solution of potassium paratungstate. When 10 c.c. of the solution are afterwards treated with 1 c.c. of an aqueous solution of sodium acetate 4 and 3 drops of a 5 per cent, solution of csesium chloride, a crystalline precipitate is obtained if colloidal silica be present. The process has not proved satisfactory when it is applied quantitatively. In the absence of phosphoric acid, the molyb- date process (page 603) can sometimes be used for estimating the amount of silica in the extract from the clay. Free and Combined Silica. Some analysts apply the term combined silica to the silica present in that portion of a clay which is decomposed by the sulphuric acid treatment ; and the term free silica to the silica present in that portion of the clay which is not decomposed by the sulphuric acid treatment. The idea seems to be that the clay is composed of kaolinite and sand or quartz ; that kaolinite alone is broken down by the sulphuric acid treatment ; and that the insoluble matter is sand or free silica. 5 All three assumptions are erroneous. Further, the insoluble residue includes various silicates felspar, mica, horn- blende, etc. and in these silicates, the silica, Si0 2 , is just as much "combined" as that combined in the kaolinite. Hence, the objections urged against the system of rational analysis might also be advanced here; and Ries rightly argues that the custom of reporting the " silica " as " free silica " and " combined silica " should be dropped, because the terms are misleading. 6 1 According to R. Schwarz (Zeit. anorg. Chem., 76. 422, 1912), when grains 0'04 mm. in diameter are boiled for half an hour with a 5 per cent, solution of sodium carbonate, 2 '11 per cent, of quartz dissolves ; and 2'77 per cent, of tridymite. 2 B. Sjollema, Journ. Landw. Chem., 50. 371, 1902. . 8 H. Hermann, Zeit. anal. Chem., 46. 318, 1907. 4 Dissolve 15 grms. of crystalline sodium acetate in 35 grms. of water, and 5 grms. o glacial acetic acid. 5 See T. E. Thorpe, Quantitative Chemical Analysis, London, 184, 18 6 H. Ries, Clays their Occurrence, Properties, and Uses, New York, 68, 1906. 670 A TREATISE ON CHEMICAL ANALYSIS. 345. The Composition of Felspathic and Quartz Detritus. We next investigate the composition of the residue left after the clay sab- stance has been removed. Following Seger, it is generally assumed that the residue is unweathered felspar or felspathic minerals and quartz. This assumption, in many cases, has no experimental foundation. For instance, felspar is seldom, if ever, found in the china clay from Cornwall or Devonshire, and yet half the residue in these clays would be reported as felspar by the method of rational analysis. Hussak, also, found no felspar in the majority of the kaolins he examined. 1 In spite of this, it is by no means uncommon to find ball and china clays reported with n per cent, of felspar when none is present. The felspathic residue will include all those minerals which resist, wholly or in part, the sulphuric acid treatment. Again, practically all the siliceous minerals are decomposed by hydrofluoric acid, but, as indicated on page 657, a few less important minerals will escape that acid, and, in consequence, these will be included with "quartz debris" Again, the mixed precipitate of aluminium and iron hydroxides is supposed to be alumina, and, on that assumption, its weight is multiplied by 5 '45 the amount of potash felspar corresponding with one part of alumina. Usually, a part of the potash is replaced by soda in the ideal potash felspar, and a part of the aluminium by iron. Both these factors make the above ratio deviate from 5*451. Seger used 5'41. Bollenbach 2 determines the ferric oxide separately and subtracts it from the "ammonia precipitate" before calculating the Al 2 3 .2Si0 2 . 2H 2 from the alumina. Others 3 have multiplied the amount of alumina by 9, or by 8*5, and called the felspathic detritus "mica." We have seen on page 663 that a part of the mica in the clay may appear with the argillaceous matter, and part with the felspathic detritus. In consequence, there is some uncertainty as to the nature of the minerals which make up the residue from the sulphuric acid treatment. It is seldom possible to make a quantitative estimate from the microscopic examination of the clay, because : (1) The minerals are often covered with a more or less opaque weathered crust which prevents an application of the optical tests ; and (2) The mineral grains are often too small to permit satisfactory optical tests. 4 Hence, the method of rational analysis has received but little, if any, aid from the microscope (see Vol. II.). 346. The Ultimate and Rational Analyses. Some have been so strongly impressed with the difficulties involved in con- ducting a satisfactory rational analysis that they have abandoned the operation, 5 and calculated the supposed mineral composition by a process similar to that which follows : Let it be required to calculate the proportion of the three components (1) 1 E. Hussak, Sprech., 22. 8, 1889. 2 H. Bollenbach, Sprech., 41. 340, 351, 1908. 3 A. E. Tucker (The Great Western Railway Co. v. The Carpella United China Clay Co. Ltd., 175, 1908) and R. R. Tatlock( The North British Railway Co. v. Turners Ltd., 74, 1904) consider calculating the alkalies to felspar to be the "fairest and best" way. Obviously, if we are satisfied with guessing, either way may be taken. See also W. C. Hancock. Journ. Soc. Chem. Ind., 29. 309, 1910. 4 Thus, a most experienced mineralogist reported that he was unable to say for certain whether some of the finer grains in the above-mentioned shale were mica or quartz. For the difficulty with mica and kaolinite, see J. W. Gregory, The Great Western Railway Co. v. The Carpella United China Clay Co. Ltd., 324, 1908. 5 E. R. Buckley, Report on the Clays and the Clay Industries of Wisconsin, Madison, Wis. , 267, 1901. See also Heim, Sprech., 28. 519, 547, 1895. THE RATIONAL ANALYSIS OF CLAYS. 671 kaolinite or clayite, (2) felspar, and (3) quartz directly from the ultimate analysis. Take Altofts shale in illustration. The ultimate analysis of the sample of Altofts shale, dried at 110, previously cited, furnished: Si0 2 . Ti0 2 . A1 2 3 - Fe 2 3 . MgO. CaO. K 2 0. Na 2 0. Loss on ignition. 51-92 0-87 20-08 6'40 1'58 0'57 272 0'86 14'61 The loss on ignition included 4'42 per cent, of carbon, and 2*44 per cent, of carbon dioxide. It is required to calculate the proportion of kaolinite or clayite, felspar, and quartz on the assumption that the clay contains potash and soda felspars, quartz, and clayite or kaolinite. Several different ways are possible ; not all are concordant. 1 From the formulae for potash (K 2 . A1 2 3 . 6Si0 2 ) and soda (Na 2 . A1 2 3 . 6Si0 2 ) felspar, it follows that the per cent. K 2 x 3 '83 represents the per cent, of silica in potash felspar ; and the per cent, of Na 2 x 5'8 represents the per cent, of silica in soda felspar. Hence, the amount of silica in the two felspars will be : Potash felspar . . . . . . 3 83 x 2 72 = 10 '42 per cent. Soda felspar 5'8 xO'86= 4 '99 Total silica in the felspars . ... . 15'41 ,, Again, from the formulae for the two felspars, the per cent, of potash multiplied by 1'09 represents the amount of alumina in potash felspar; and the per cent, of soda multiplied by 1 '65 gives the corresponding amount of alumina in the soda felspar. Hence, the alumina in the two felspars will be : Potash felspar . . . " . . . 1'09 x 272 = 2'96 per cent. Soda felspar . . . . . . 1'65 xO'86 = l-42 ,, Total alumina in the felspars . . A . ' . 4 '38 ,, The difference between the total alumina and the alumina in the felspars represents the alumina in the kaolinite ; and the product of the alumina in the kaolinite with 1*18 represents the amount of silica combined as kaolinite. Hence Alumina in clay . ' . . V . . . 20 '08 per cent. Alumina in felspars 4 '38 Alumina in kaolinite . . . .^ . . 15 '70 ,, The ferric oxide presents a difficulty. Is it 'to be ignored 1 Is it to be included with the "alumina" 1 ? Is it to be calculated as nontronite Fe 2 3 .2Si0 2 .2H 2 0? Or is it to be included with the clay substance ? With high-grade china clays there is no difficulty in answering the questions, because the amount of iron is negligibly small so far as the degree of accuracy of the calculation is concerned. It is usually included either with the alumina or with the argillaceous matter, 2 generally the latter. As a matter of fact, in this particular sample most of the iron was removed by digestion with hydrochloric acid (1:1), and it should there- fore be placed as a constituent by itself. The fact that the clay effervesced with evolution of carbon dioxide when treated with an acid proved that some con- stituents in the ultimate analysis were present as carbonate. The comparatively low proportion of CaO and MgO and the high proportion of '"iron." pointed to the presence of ferrous carbonate. When the necessary calculation is made, we get : Amount of carbon dioxide x 1'64 = 2'44 x 1*64 = 4'0 per cent, of FeO ; or 1 For instance, if we make an assumption about the composition of the mica, felspar, clayite, and quartz presumably in the clay, the method of calculation devised by J. W. Mellor, Trans. Eng. Cer. Soc., 7. 117, 1908, can be employed after deducting the proper allowance for siderite and carbon. 2 We have seen, page 658, that most of the iron in this particular shale was probably present as ferrous carbonate. 6 7 2 A TREATISE ON CHEMICAL ANALYSIS. 4 -0 + 2 -44 = nearly 6'5 per cent, of ferrous carbonate. Hence, since: Amount FeO x V = 4*0 x V = 4-44 per cent, of ferric oxide ; 6 -4 less 4 '4 = 2'0 per cent, of ferric oxide remains unaccounted for. Again, Silica with kaolinite (1 18 x 15 -7) . Silica with felspars IS '53 percent. 15-41 Silica with kaolinite and felspars Total silica in clay .... Silica with kaolinite and felspars . . 33 '94 ,, 51 '92 per cent. 33-94 17-98 Silica as quartz ...... Add up the percentage amounts of magnesia, 1 lime, and alkalies, with the computed amounts of silica and alumina for the felspars ; and add up the loss on ignition (7 '75), ferric oxide, and the computed alumina and silica for the kaolinite as argillaceous matters. 2 Summarising these results : ( Silica ...... 18-5 Argillaceous matter ^ Alumina and iron oxide . . .177 (Water 7'8 ( Silica 15*4 Felspathic matter -I Alumina . . . . ' U^es . . . . .. Quartz (silica) . . . Ferrous carbonate ; . Carbonaceous matter . 5-4^ 44 I . 6-Oj 44-0 25-5 18-0 6 4 4-4 A comparison of this with the actual result, as indicated in the table below, is not satisfactory. The discrepancy partly arises from the dubious assumption that the alkalies in the ultimate analysis all belong to felspathic matter, and none Comparison of the Rational Composition of Altofts Shale estimated* by different Methods. Computed from ultimate analysis Computed Observed. of matter soluble and insoluble in from ultimate analysis of sulphuric acid, whole clay. etc. Carbonaceous matter . 4-4 4-4 4'4 Ferrous carbonate 6-4 6'4 6'4 Argillaceous matter 48-6 47-8 44-0 Felspathic detritus . , ' . 5-4 6-6 25-5 Quartz debris . . 35'2 34-4 18-0 1 If calcium and magnesium carbonates were present, they would be removed by the hydro- chloric acid treatment. 2 The ' ' loss on ignition " may be employed as a check on the work, since the product of the " loss on ignition " with 7 '17 represents the corresponding amount of clayite or kaolinite. The " loss on ignition " is usually rather less than the corresponding amount of clayite, even if the amounts of carbonaceous matter and carbon dioxide be first deduced. There is a small error due to gain in weight by the oxidation of ferrous oxide. Some consider that a certain amount of alkali takes the place of H ? O in A1 2 3 . 2Si() 2 . 2H 2 0. C. F. Binns (Trans. Amer. Cer. Soc., 8. 198, 1906) assumes that a little of the alkali is adsorbed by the argillaceous matter. 3 Note the distinction between to estimate and to determine in analytical work. The former is "to judge or form an opinion from imperfect data" ; the latter, "to ascertain by definite means measurement." Sometimes it is difficult to decide where to draw the line e.g. soda, pages 238-9 ; and sometimes an estimation is more accurate than a determination. However, the distinction has been emphasised in the Courts. THE RATIONAL ANALYSIS OF CLAYS. 673 to micaceous and argillaceous matters. The method here outlined generally gives good results with high-grade clays, but with the low-grade clays, as with the rational analysis proper, the results are hopeless. Consequently, our attitude towards the method of calculation just indicated, and to the process of rational analysis, is largely determined by the particular type of clay under consideration, and the purpose for which the estimation is made. With the high-grade clays, the margin of error is small ; with low-grade clays (and possibly with Cornish stone), we can say with Bischof l the process is " worthless " (wertlos), and with Prossel, " useless " (unbrauchbar}. We must here confess our inability to prescribe a process of chemical analysis of general applicability which will enable us to deduce the exact mineralogical composition of clays. Blind faith in the routine process is sure to err. A certain amount of judgment must be exercised in adapting a process of analysis to particular clays and to particular methods of manufacture. We must decide what minerals are technically important, and a process must be devised to deal with them, not necessarily to furnish the highest degree of accuracy, but rather to serve as a guide for the best treatment of the clay to furnish the desired results. One example will suffice. Zschokke's Method of Rational Analysis. Zschokke 2 found the following combined mechanical and chemical process best suited his requirements : (1) Boil 50 grms. of the dry clay for half an hour in a porcelain dish with water, and restore the water lost by evaporation from time to time. After standing 24 hours, wash the clay on a sieve 8570 Continental mesh (say, 240 British mesh) until the washings flow through clear. Break up any lumps by rubbing on the sieve, or between the fingers, or, if necessary, on a glass plate with a rubber pestle. The residue on the sieve consists of quartz, felspar, gypsum, pyrites, etc. Dry at 110, and weigh. (2) The residue on the sieve is treated with dilute hydrochloric acid (1 : 10) and washed until the washings show no lime reaction. Filter, dry and weigh the residue. The residual substance from (1) gives the coarse-grained calcium and magnesium carbonates, gypsum, and iron oxide. (3) The washings are evaporated to dryness and weighed as in (1). The result is fine-grained quartz, felspar, calcium and magnesium carbonates, gypsum, pyrites, etc. (4) Treat the residue from (3) as described under (2). The results are subtracted from (3) and give fine-grained gypsum, felspar, quartz. The results are expressed in percentages : I. Non -plastic constituents 38 '6 A. Coarse-grained constituents 13*9 (a) Quartz, felspar . . .7*8 (b) CaC0 3 and MgC0 3 \U J V-/cV>x/3 cviivj. AM^wvyg B. Fine-grained constituents (a) Quartz, felspar (b) CaC0 3 and MgC0 3 247 11-3 13-4 II. Clay substance (by difference) 61 '4 The clays treated by this method were required for bricks fired at too low a temperature for the felspar and quartz to play a particularly important .part during the firing. Information about the size of grain of the non-plastic constituents, and the amount of the calcium and magnesium carbonates present in the clay, was particularly desired. Hence a justification of Zschokke's procedure. 1 C. Bischof, Die feuerfesten Tone, Leipzig, 103, 1904 ; B. P. Tenax (B. Prossel), Die Meingut und Porzellanfabrikation, Leipzig, 8, 1879. 2 B. Zschokke, Baumaterialienkunde, 7. 165, 1902 ; Mitt. Eidg. Matcrialpnif. Anstalt, Zurich, ii. 22, 1907. 43 674 A TREATISE ON CHEMICAL ANALYSIS. 347. Are the Ultimate and Rational Analyses Consistent? The ultimate and rational analyses of the high-grade clays, that is, clays containing a large proportion of kaolinite or clayite, usually agree very well. With these clays the proportion of felspar and quartz detritus is small. Con- sequently, the errors arising from secondary reactions in the removal of clay substance are relatively small. For instance, if a clay has 70 per cent, of argillaceous matter, 4 per cent, of felspar, and 26 per cent, of quartz, even if 20 per cent, of the felspar breaks down by the sulphuric acid 'treatment, it would only have the effect of diminishing the proportion of felspar 0'8 per cent., and of raising the clay substance accordingly. If two clays contain the same minerals, and if the sulphuric acid attacks both in the same way, the ultimate and rational analyses must necessarily agree. Disagreement can only occur when the clays contain different minerals, the same minerals in different proportions or in different states of subdivision. Cases have been recorded which apparently contradict this observation. For example : l ULTIMATE ANALYSES. Si0. 2 . A1A- FeA MgO. CaO. KA NaA A 62-40 26-51 1'14 O'Ol 0'57 0^98 8'80 B . . 62-52 25-57 0'92 O'lO 0'65 1'04 927 C . . 59-30 27-17 1'44 nil. 0'48 0'52 9'67 RATIONAL ANALYSES. Clay substance. Felspar. Quartz. A ... . 66-33 18-91 15-61 B . . . . 72-05 0-10 27-76 C 43-74 36-84 19-42 In A the proportion alkalies to felspar is 1 : 18. I am not acquainted with a felspar with alkalies and felspar in these proportions (by weight). Albite has this ratio nearly as 1:9; and orthoclase, 1 : 6. There is obviously something wrong. The case is much worse with the C clay, for it is at once obvious that 0*52 per cent, of alkalies is quite incompatible with 36-8 per cent, of felspar. See also page 672. This sort of thing is not at all uncommon, 2 although I have not met such bad examples in my own practice. We are thus driven to conclude that the method gives incompatible results, or the analyses are at fault. We naturally inquire : Was the clay substance all decomposed by the sulphuric acid treatment in the rational analysis of clays A and C ] Was some of the felspar decomposed J^y the sulphuric acid in the rational analysis of the B clay 1 Summary. Here, then, it is inferred that the method for conducting the rational analysis gives results incompatible with the ultimate analysis ; the ultimate analysis can be conducted with a great degree of precision ; hence the process of rational analysis cannot be recommended as a general method for comparing the properties of the different types of clay, although, as indicated above, the rational analysis is useful (1) for comparing the properties of high- grade clays which do not differ very materially in the nature of their constituent minerals; .and (2) it is sometimes a help in forming a rough idea what minerals are present in clays. 3 In the former case, however, the " rational composition " can be computed accurately enough from the ultimate analysis. 1 A is a raw china clay from North Carolina H. Ries, North Carolina Geol. 'Sur., 13. 62, 1897 ; B is a slip clay from Lothian, Saxony H. A. Seger, Tonind. Ztg.,.i6. 1031, 1892 ; and C clay is from Shropshire J. T. Norman, Report on some Shropshire Clays, London, 1903. 2 J. W. Mellor, Pot. Gaz., 35. 1060, 1908 ; H. Ries, Trans. Amer. lust. Min. Eng. t 28. 160, 1899. 3 Or rather the fractional separation of constituents soluble in certain menstrua. APPENDIX. The human mind is seldom satisfied, and is certainly never exercising its highest functions when it is doing the work of a calculating machine. J. C. MAXWELL. THE results of old arithmetical operations most frequently required are registered in the form of tables. The use of such tables not only prevents the wasting of time and energy on a repetition of old operations, but also conduces to more accurate work, since there is less liability to error once accurate tables have been compiled. The specific gravity-concentration tables Tables LXXIV. to LXXXIII. explain themselves. They are referred to in different parts of this book. The authority for each is indicated in the text. Table LXXXV. et seq. are used like the ordinary reference logarithm tables. For instance, at 13 '8 the vapour pressure of water is equivalent to a 11 '84 mm. column of mercury. Tables LXXXVI. to XCIV. have been specially computed for this book. They are used in the ordinary manner. For instance, page 236, a precipitate of potassium chloroplatinate weighing 0*0547 grm. will, from Table LXXXVI., be equivalent to 0*0168 grm. of potassium chloride ; and, from Table LXXXVIIL, to 0-0106 grm. of potash, K 2 0. From Table LXXXIX., 0*0059 grm. of sodium chloride is equivalent to 6'0031 grm. of soda, Na 2 0. Similarly, page 238, Table LXXXVII. tells us that 0'0272 grm. of potassium perchloride is equivalent to 0'0146 grm. of potassium chloride, and the latter, from Table LXXXVIIL, is equivalent to 0'0092 grm. of potash, K 2 0. The other tables are to be used in an analogous manner. 6;6 A TREATISE ON CHEMICAL ANALYSIS. Table LXXIV. Sulphuric Acid Specific Gravity and Concentration. G. Lunge and M. Isler, Zeit. angew. Chem., 3. 129, 1890 ; G. Lunge, Technical Chemists' Handbook, London, 1910. 100 parts by 100 parts by Specific Gravity 15. weight contain Kilo per litre. Specific Gravity 15. weight contain Kilo per litre. S0 3 . H 2 S0 4 . H 2 S0 4 . S0 3 . H 2 S0 4 . H 2 S0 4 . I'OIO 1-28 1-57 0-016 1-305 32-46 39-77 0-519 1-020 2-47 3-03 0-031 1-310 32-94 40-35 0-529 1-030 3-67 4-49 0-046 1-315 33-41 40-93 0-538 1-040 4-87 5-96 0-062 1-320 33-88 41-50 0-548 1-050 6-02 7-37 0-077 1-325 34-35 42-08 0-557 1-060 7-16 8-77 0-093 1-330 34-80 42-66 0-567 1-070 8-32 10-19 0-109 1-335 35-27 43-20 0-577 roso 9-47 11-60 0-125 1-340 35-71 43-74 0-586 1-090 10-60 12-99 0-142 1-345 36-14 44-28 0-596 1-100 11-71 14-35 0-158 1-350 36-58 44-82 0-605 1-110 12-82 15-71 0-175 1-355 37-02 45-35 0-614 1-120 13-89 17-01 0-191 1-360 37-45 45-88 0-624 1-130 14-95 18*31 0-207 1-365 37-89 46-41 0-633 1-140 16-01 19-61 0-223 1-370 38-32 46-94 0-643 1-150 17-07 20-91 0-239 1-375 38-75 47-47 0-653 1-160 18-11 22-19 0-257 1-380 39-18 48-00 0-662 1-170 19-16 23-47 0-275 1-385 39-62 48-53 0-672 1-180 20-21 24-76 0-292 1-390 40-05 49-06 0-682 1-190 21-26 26-04 0-310 1-395 40-48 49-59 0-692 1-200 22-30 27-32 0-328 1-400 40-91 50-11 0-702 1-205 22-82 27-95 0-337 1-405 41-33 50-63 0-711 1-210 23-33 28-58 0-346 1-410 41-76 51-15 0-721 1-215 23-84 29-21 0-355 1-415 42-17 51-66 0-730 1-220 24-36 29-84 0-364 1-420 42-57 52-15 0-740 1-225 24-88 30-48 0-373 1-425 42-96 52-63 0-750 1-230 25-39 31-11 0-382 1-430 43-36 53-11 0-759 1-235 . 25-88 31-70 0-391 1-435 43-75 53-59 0-769 1-240 26-35 32-28 0-400 1-440 44-14 54-07 0-779 1-245 26-83 32-86 0-409 1-445 44-53 54-55 0-789 1-250 27-29 33-43 0-418 1-450 44-92 55-03 0-798 1-255 27-76 34-00 0-426 1-455 45-31 55-50 0-808 1-260 28-22 34-57 0-435 1-460 45-69 55-97 0-817 1-265 28-69 35-14 0-444 1-465 46-07 56-43 0-827 1-270 29-15 35-71 0-454 1-470 46-45 56-90 0-837 1-275 29-62 36-29 0-462 1-475 46-83 57-37 0-846 1-280 30-10 36-87 0-472 1-480 47-21 57-83 0-856 1-285 30-57 37-45 0-481 1-485 47-57 58-28 0-866 1-290 31-04 38-03 0-490 1-490 47-95 58-74 0-876 1-295 31-52 38-61 0-500 1-495 48-34 59-22 0-886 1-300 31-99 39-19 0-510 1-500 48-73 59-70 0-896 APPENDIX. Table LXXIV. continued. 677 100 parts by 100 parts by Specific Gravity weight contain Kilo per litre. Specific Gravity weight contain Kilo per litre. 15. 15. S0 3 . H 2 S0 4 . H 2 S0 4 . S0 3 . H 2 S0 4 . H 2 S0 4 . 1-505 49-12 60-18 0-906 1-740 65-86 80-68 1-404 1-510 49-51 60-65 0-916 1-745 66-22 81-12 1-416 1-515 49-89 61-12 0-926 1-750 66-58 81-56 1-427 1-520 50-28 61-59 0-936 1-525 50-66 62-06 0-946 1-755 66-94 82-00 1-439 1-760 67-30 82-44 1-451 1-530 51-04 62-53 0-957 1-765 67-76 83-01 1-465 1-535 51-43 63-00 0-967 1-770 68-17 83-41 1-478 1-540 51-78 63-43 0-977 1-775 68-60 84-02 1-491 1-545 52-12 63-85 0-987 1-550 52-46 64-26 0-996 1-780 68-98 84-50 1-504 1-785 69-47 85-10 1-519 1-555 52-79 64-67 1-006 1-790 69-96 85-70 1-534 1-560 53-22 65-20 1-017 1-795 70-45 86-30 1-549 1-565 53-59 65-65 1-027 1-800 70-96 86-92 1-565 1-570 53-95 66-09 1-038 1-575 54-32 66-53 1-048 1-805 71-50 87-60 1-581 1-580 1-585 54-65 55-03 66-95 67-40 1-058 1-068 1-810 1-815 72-08 72-69 88-30 89-05 1-598 1-621 1-590 55-37 67-83 1-078 1-820 73-51 90-05 1-639 1-595 55-73 68-26 1-089 1-821 73-63 90-20 1-643 1-600 56-09 68-70 1-099 1-822 73-80 90-40 1-647 1-605 1-610 56-44 56-79 69-13 69-56 1-110 1-120 1-823 1-824 73-96 74-12 90-60 90-80 1-651 1-656 1-615 1-620 1-625 57-15 57-49 57-84 70-00 70-42 70-85 1-131 1-141 1-151 1-825 1-826 1-827 74-29 74-49 74-69 91-00 91-25 91-50 1-661 1-666 1-671 1-630 58-18 71-27 1-162 1-828 74-86 91-70 1-676 1-635 58-53 71-70 1-172 1-829 75-03 91-90 1-681 1-640 1-645 1-650 58-88 59-22 59-57 72-12 72-55 72-96 1-182 1-193 1-204 1-830 1-831 1-832 75-19 75-46 75-69 92-10 92-43 92-70 1-685 1-692 1-698 1-655 59-92 73-40 1-215 1-833 75-89 92-97 1-704 1-660 60-26 73-81 1-225 1-834 76-12 93-25 1-710 1-665 1-670 1-675 60-60 60-95 61-29 74-24 74-66 75-08 1-236 1-246 1-259 1-835 1-836 1837 76-35 76-57 76-90 93-56 93-80- 94-20 1-717 1-722 1-730 1-680 61-63 75-50 1-268 1-838 77-23 94-60 1-739 1-685 61-93 75-86 1-278 1-839 77-55 95-00 1-748 1-690 1-695 1-700 62-29 62-64 63-00 76-30 76-73 77-17 1-289 T301 1-312 1-840 1-8405 1-8410 78-04 * 78-33 78-69 95-60 95-95 96-30 1-759 1-765 1-784 1-705 63-35 77-60 1-323 1-8415 79-47 97-35 1-792 1-710 63-70 78-04 1-334 1-8410 80-16 98-20 1-808 1-715 1-720 1-725 64-07 64-43 64-78 78-48 78-92 79-36 1-346 1-357 1-369 1-8405 1-8400 1-8395 80-43 80-59 80-63 98-52 98-72 98-77 1-814 1-816 1-817 1-730 1-735 65-14 65-50 79-80 80-24 1-381 1-392 1-8390 1-8385 80-93 99-12 81-08 99-31 1-823 1-826 6;8 A TREATISE ON CHEMICAL ANALYSIS. Table LXXV. Nitric Acid Specific Gravity and Concentration. G. Lunge and H. Key, Zeit. angew. Chem., 4. 165, 1891 ; G. Lunge, Technical Chemists 1 Handbook, London, 1910. Percentage by Grams per Percentage by Grams per Specific weight. litre. Specific weight. litre. Gravity Gravity 15. 15. N 2 5 . HN0 3 . N 2 5 . HN0 3 . N 2 5 . HN0 3 . N 2 5 . HN0 3 . rooo 0-08 0-10 1 1 1-200 27-74 32-36 333 388 roo5 0-85 1-00 8 10 1-205 28-36 33-09 342 399 roio 1-62 1-90 16 19 1-210 28-99 33-82 351 409 1-015 2-39 2-80 24 28 1-215 29-61 34-55 360 420 1-020 3-17 3-70 33 38 1-220 30-24 35-28 369 430 1-025 3-94 4-60 40 47 1-225 30-88 36-03 378 441 1-030 4-71 5-50 49 57 1230 31-53 36-78 387 452 1-035 5-47 6-38 57 66 1-235 32-17 37-53 397 463 1-040 6-22 7-26 64 75 1-240 32-82 38-29 407 475 1-045 6-97 8-13 73 85 1-245 33-47 39-05 417 486 P050 7-71 8-99 81 94 1-250 34-13 39-82 427 498 1-055 8-43 9-84 89 104 1-255 34-78 40-58 437 509 1-060 9-15 10-68 97 113 1-260 35-44 41-34 447 521 1-065 9-87 11-51 105 123 1-265 36-09 42-10 457 533 1-070 10-57 12-33 113 132 1-270 36-75 42-87 467 544 1-075 11-27 13-15 121 141 1-275 37-41 43-64 477 556 1-080 11-96 13-95 129 151 1-280 38-07 44-41 487 568 1-085 12-64 14-74 137 160 1-285 38-73 45-18 498 581 1-090 13-31 15-33 145 169 1-290 39-39 45-95 508 593 1-095 13-99 16-32 153 179 1-295 40-05 46-72 519 605 I'lOO 14-67 17-11 161 188 1-300 40-71 47-49 529 617 1-105 15-34 17-89 170 198 1-305 41-37 48-26 540 630 rno 16-00 18-67 177 207 1-310 42-06 49-07 551 643 1-115 16-67 19-45 186 217 1-315 42-76 49-89 562 656 1-120 17-34 20-23 195 227 1-320 43-47 50-71 573 669 1-125 18-00 21-00 202 236 1-325 44-17 51-53 585 683 1-130 18-66 ' 21-77 211 246 1-330 44-89 52-37 597 697 1-135 19-32 22-54 219 256 1-335 45-62 53-22 609 710 1-140 19-98 23-31 228 266 1-340 46-35 54-07 621 725 11-45 20-64 24-08 237 276 1-345 47-08 54-93 633 739 1-150 21-29 24-84 245 286 1-350 47-82 55-79 645 753 1-155 21-94 25-60 254 296 1*355 48-57 56-66 658 768 160 22-60 26-36 262 306 1*360 49-35 57-57 671 783 165 23-25 27-12 271 316 1-365 50-13 58-48 684 798 170 23-90 27-88 279 326 1-370 50-91 59-39 698 814 175 24-54 28-63 288 336 1-375 51-69 60-30 711 829 180 25-18 29-38 297 347 1-380 52-52 61-27 725 846 185 25-83 30-13 306 357 1-385 53-35 62-24 739 862 1-190 26-47 30-88 315 367 1-390 54-20 63-23 753 879 T195 27-10 31-62 324 378 1-395 55-07 64-25 768 896 APPENDIX. Table LXXV. continued. 679 Percentage by , Grams per Percentage by Grams per Specific weight. litre. Specific weight. litre. Gravitv Gravity 15. 15. N 2 5 . HN0 3 . N 2 5 - iHN0 3 . N 2 5 . HN0 3 . N 2 5 . HN0 3 . 1-400 55-97 65-30 783 914 1-465 69-79 81-42 1023 1193 1-405 56-92 66-40 800 933 1-470 71-06 82-90 1045 1219 1-410 57-86 67-50 816 952 1-415 58-83 68-63 832 971 1-475 72-39 84-45 1068 1246 1-420 59-83 69-80 849 991 1-480 73-76 86-05 1092 1274 1-425 1-430 1-435 60-84 61-86 62-91 70-98 72-17 73-39 867 885 903 1011 1032 1053 1-485 1-490 1-495 75-18 76-80 78-52 . 87-70 89-60 91-60 1116 1144 1174 1302 1335 1369 1-440 1-445 64-01 65-13 74-68 75-98 921 941 1075 1098 1-500 1-505 80-65 82-63 94-09 96-39 1210 1244 1411 1451 1-450 66-24 77-28 961 1121 1-510 84-09 98-10 1270 1481 1-455 67-38 78-60 981 1144 1-515 84-92 99-07 1287 1501 1-460 68-56 79-98 1001 1168 1-520 85-44 ' 99-67 1299 1515 Table LXXVI. Hydrochloric Acid Specific Gravity and Concentration. G. Lunge and L. Marchlewski, Zeit. angew. Chem., 4. 133, 1891 ; G. Lunge, Technical Chemists' Handbook, London, 1910. Specific Gravity in vacuo. 100 parts by weight correspond to parts by weight of HC1. 1 litre contains g. of HC1. Specific Gravity t 15 in vacuo. 100 parts by weight correspond to parts by weight of HC1. 1 litre contains g. of HC1. 1-000 0-16 1-6 1-105 20-97 232 1-005 1-15 12 1-110 21-92 243 1-010 1-015 1-020 2-14 3-12 4-13 22 32 42 1-115 1-120 1-125 22-86 23-82 24-78 255 267 278 1-025 5-15 53 1-130 25-75 291 1-030 6-15 64 1-135 26-70 303 1-035 7-15 74 1-140 27-66 315 1-040 8-16 85 1-145 28-61 328 1-045 9-16 96 1-150 29-57 340 1-050 10-17 107 1-155 30-55 353 1-055 11-18 118 1-160 31-52 366 1-060 12-19 129 1-165 32-49 379 1-065 13-19 141 1-170 33-46 392 1-070 14-17 ! 152 1-175 34-42 404 1-075 15-16 163 1-080 1-085 1-090 1-095 1-100 * 16-15 17-13 18-11 19-06 20-01 174 186 197 209 220 1-180 1-185 1-190 1-195 1-200 35-39 36-31 37-23 38-16 39-11 418 430 443 456 469 68o A TREATISE ON CHEMICAL ANALYSIS. Table LXXVIL Hydrobromic Acid Specific Gravity and Concentration. H Topsoe, er., 3. 404, 1870. Per cent. I 2 3 4 5 6 7 8 9 HBr. 1-007 1-014 1-021 1-028 1-035 1-043 1-050 1-058 1-065 1 1-073 1-081 1-089 1-097 1-106 1-114 1-122 1-131 1-140 1-149 2 1-158 1-167 1-176 1-186 1-196 1-206 1-215 1-225 1-235 1-246 3 1-257 1-268 1-279 1-290 1-302 1-314 1-326 1-358 1-351 1-363 4 1-376 , 1-389 I ' 1-403 1-417 | 1-431 1-445 1-459 1-475 1-487 1-502 Table LXXVI 1 1. Hydrofluoric Acid Specific Gravity and Concentration. F. VVinteler, Zeit. angew. Chem., 15. 33, 1902. 1 2 3 4 5 6 7 8 9 1-003 1-007 1-011 1-014 1-018 1-023 1-027 1-030 1-035 1 1-038 1-041 1-045 1-049 1-052 1-055 1-059 1-062 1-066 1-069 2 1-072 1-076 1-079 1-082 1-086 1-089 1-092 1-095 1-098 1-101 3 1-104 1-106 1-119 1-112 1-114 1-117 1-120 1-122 1-125 1-127 4 1-130 1-133 1-136 1-138 1-141 1-143 1-146 1-149 1-152 1-157 Table LXXIX. Phosphoric Acid Specific Gravity and Concentration. H. Hager, Commentar zur Pharmacopeia germanica, Berlin, 1884. Sp. gr. Per cent. PA. Per cent. H 3 P0 4 . Sp. gr. Per cent. P 2 5 . Per cent. H 3 P0 4 . Sp- gr. Per cent. PA> Per cent. H 3 P0 4 . 1-809 68-0 93-67 1-661 59-0 81-28 1-521 50-0 68-88 1-800 67-5 92-99 1-653 58-5 80-59 1-513 49-5 68-19 1-792 67-0 92-30 1-645 58-0 79-90 1-505 49-0 67-50 1-783 66-5 91-61 1-637 57-5 79-21 1-498 48-5 66-81 1-775 66-0 90-92 1-629 57-0 78-52 1-491 48-0 66-12 1-766 65-5 90-23 1-621 56-5 77-83 1-484 ! 47-5 65-43 1-758 65-0 89-54 1-613 56-0 77-14 1-476 47-0 64-75 1-750 64-5 88-85 1-605 55-5 76-45 1-469 46-5 64-06 1-741 64-0 88-16 1-597 55-0 75-77 1-462 46-0 63-37 1-733 63-5 87-48 1-589 54-5 75-08 1-455 45-5 62-68 1-725 63-0 86-79 1-581 54-0 74-39 1-448 45-0 61-99 1-717 62-5 86-10 1-574 53-5 73-70 1-441 44-5 61-30 1-709 62-0 85-41 1-566 53-0 73-01 1-435 44-0 60-61 1-701 61-5 84-72 1-599 52-5 72-32 1-428 43-5 59-92 1-693 1-685 1-677 61-0 60-5 60-0 84-03 83-34 82-65 1-551 1-543 1-536 52-0 51-5 51-0 71-63 70-94 70-26 1-422 1-415 1-409 43-0 42-5 42-0 59-23 58-55 57-86 1-669 59-5 81-97 1-528 50-5 69-57 1-402 41-5 57-17 APPENDIX. Table LXXIX. continued. 68i Sp. gr. Per cent. P 2 5 . Per cent. H 3 P0 4 . Sp. gr. Per cent. PA. Per cent. H 3 P0 4 . Sp. gr. Per cent. P 2 5 . Per cent. H 3 P0 4 . 1-396 41-0 56-48 1-249 28-0 38-57 1-122 15-0 20-66 1-389 40-5 55-79 1-244 27-5 37-88 1-118 14-5 19-97 1-383 40-0 55-10 1-239 27-0 37-19 1-113 14-0 19-28 377 39-5 54-41 1-233 26-5 36-50 1-109 13-5 18-00 371 39-0 53-72 1-228 26-0 35-82 1-104 13-0 17-91 365 38-5 50-34 1-223 25-5 35-13 1-110 12-5 17-22 359 38-0 52-35 1-218 25-0 34-44 1-096 12-0 16-53 354 37-5 51-66 1-213 24-5 33-75 1-091 11-5 15-84 348 37-0 50-97 208 24-0 33-06 1-087 11-0 15-15 342 36-5 50-28 203 23-5 32-37 1-083 10-5 14-46 336 36-0 49-59 198 23-0 31-68 1-079 10-0 13-77 330 35-5 48-90 193 22-5 30-99 1-074 9-5 13-09 325 35-0 48-21 188 22-0 30-31 1-070 9-0 12-40 319 34-5 47-52 183 21-5 29-62 1-066 8-5 11-71 314 34-0 46-84 178 21-0 28-93 1-062 8-0 11-02 308 33-5 46-15 174 20-5 28-24 1-058 7-5 10-33 303 33-0 45-46 169 20-0 27-55 1-053 7-0 9-64 298 32-5 44-77 1-164 19-5 26-86 1-049 6-5 8-95 292 32-0 44-08 1-159 19-0 26-17 1-045 6-0 8-26 287 31-5 43-39 1-155 18-5 25-48 1-041 5-5 7-57 281 31-0 42-70 1-150 18-6 24-80 1-037 5-0 6-89 276 30-5 42-01 145 17-5 24-11 1-033 4-5 6-20 271 30-0 41-33 140 17-0 23-42 1-029 4-0 5-51 265 29-5 40-64 135 16-5 22-73 1-025 3-5 4-82 1-260 29-0 39-95 130 16-0 22-04 1-021 3-0 4-13 1-225 28-5 39-26 126 15-5 21-35 1-017 2-5 3-44 Table LXXX. Perchloric Acid. Specific Gravity and Concentration. K. van Emster, Ze.it. anorg. Ghem., 52. 270, 1907. AT 15 (WATER AT 4 UNITY). Sp. gr. Per cent. HC10 4 . Sp. gr. Per cent. HC10 4 . Sp. gr. Per cent. HC10 4 . Sp. gr. Per cent. HC10 4 . 1-005 1-00 1-090 14-56 175 26-20 260 36-03 010 1-90 1-095 15-28 180 26-82 265 36-56 015 2-77 1-100 16-00 185 27-44 270 37-08 020 3-61 1-105 16-72 190 28-05 275 37-60 025 4-43 1-110 17-45 195 28-66 280 38-10 030 5-25 1-115 18-16 200 29-26 285 38-60 035 6-07 1-120 18-88 205 29-86 290 39-10 040 6-88 1-125 19-57 210 30-45 1-295 39-60 045 7-68 1-130 20-26 215 31-04 1-300 40-10 050 8-48 1-135 20-95 220 31-61 1-305 40-59 055 9-28 1-140 21-64 225 32-18 310 41-08 060 10-06 1-145 22-32 230 32-74 315 41-56 065 10-83 1-150 22-99 235 33-29 320 42-03 070 11-58 1-155 23-65 240 33-85 325 42-49 075 12-33 1-160 24-30 245 34-40 330 42-97 1-080 13-08 1-165 24-94 250 34-95 335 43-43 1-085 13-83 1-170 25-57 255 35-49 340 43-89 682 A TREATISE ON CHEMICAL ANALYSIS. Table LXXX. continued. Sp.gr. Per cent. HC10 4 . Sp. gr. Per cent. HC10 4 . Sp. gr. Per cent, HC10 4 . Sp. gr. Per cent. HC10 4 . 345 44-35 1-430 51-71 1-515 58-17 1-600 64-50 350 44-81 1-435 52-11 1-520 58-54 1-605 64-88 355 45-26 1-440 52-54 1-525 58-91 1-610 65-26 360 45-71 1-445 52-91 1-530 59-28 1-615 65-63 365 46-16 1-450 53-31 1-535 59-66 1-620 66-01 370 46-61 1-455 53-71 1-540 60-04 1-625 66-39 375 47-05 1-460 54-11 1-545 60-41 1-630 66-76 380 47-49 1-465 54-50 1-550 60-78 1-635 67-13 385 47-93 1-470 54-89 1-555 61-15 1-640 67-51 390 48-37 1-475 55-18 1-560 61-52 1-645 67-89 1-395 48-80 1-480 55-56 1-565 61-89 1-650 68-26 1-400 49-23 1-485 55-95 1-570 62-26 1-655 68-64 1-405 49-68 490 56-32 1-575 62-63 1-660 69-02 1-410 50-10 495 56-69 1-580 63-00 1-665 69-40 1-415 50-51 500 57-06 1-585 63-37 1-670 69-77 1-420 50-91 505 57-44 1-590 63-74 1-675 70-15 1-425 51-31 510 57-81 1-595 64-12 Table LXXXL Formic Acid Specific Gravity and Concentration. G. M. Richardson and P. Alhaire, Amer. Chem. Journ., 19. 149, 1897. Grams formic acid per 100 grams of solution. Sp. 0-0 0-1 0-2 0-3 0-4 0-5 0-6 0-7 0-8 0-9 1-0020 1-0045 1-0071 1-0094 1-0116 1-0142 1-0171 1-0197 1-0222 1 1-0247 1-0272 1-0297 1-0322 1-0346 1-0371 1-0394 1-0418 1-0442 1-0465 2 0489 1-0513 1-0538 1-0562 1-0586 1-0610 1-0634 1-0657 1-0682 1-0706 3 0730 1-0754 1-0778 1-0801 1-0824 1-0848 1-0872 1-0896 1-0920 1-0941 4 0964 1-0991 1-1016 1-1039 1-1063 1-1086 1-1109 1-1131 1-1158 1-1186 5 1208 1-1224 1-1245 1-1270 1-1296 1-1321 1-1343 1-1362 1-1382 1-1402 6 1425 1-1449 1-1474 1-1494 1-1518 1-1544 1-1566 1-1585 1-1605 1-1629 7 1656 1-1678 1-1703 1-1729 1-1753 1-1770 1-1786 1-1802 1-1819 1-1838 8 1861 1-1877 1-1897 1-1915 1-1930 1-1954 1-1977 1-1995 1-2013 1-2029 9 2045 1-2060 1-2079 1-2100 1-2118 1-2141 1-2159 1-2171 1-2184 1-2203 APPENDIX. 68 3 Table LXXXI I. Acetic Acid. Specific Gravity and Concentration. A. C. Oudemans, Zeit. Chem. (2), 2. 750, 1866. Per cent. Acetic Acid CH 3 COOH. 1 2 3 4 5 6 7 8 9 0-9992 1-0007 1-0022 1-0037 1-0052 1-0067 1-0083 1-0098 1-0113 1-0127 1 1-0142 1-0157 1-0171 1-0185 ! 1-0200 1-0214 1-0228 1-0242 1-0256 1-0270 2 1-0284 1-0298 1-0311 1-0324 1-0337 1-0350 1-0363 1-0375 1-0388 1-0400 3 1-0412 1-0424 1-0436 1-0447 1 1-0459 1-0470 1-0481 1-0492 1-0502 1-0513 4 1-0523 1-0533 1-0543 1-0552 1-0562 1-0571 1-0580 1-0589 1-0598 1-0607 5 1-0615 1-0623 1-0631 1-0638 1-0646 1-0653 1-0660 1-0666 1-0673 1-0679 6 1-0685 1-0691 1-0697 1-0702 1-0707 1-0712 1-0717 1-0721 1-0725 1-0729 7 1-0733 1-0737 1-0740 1-0742 1-0744 1-0746 1-0747 1-0748 1-0748 1-0748 8 1-0748 1-0747 1-0746 1-0744 1-0742 1-0739 1-0736 1-0731 1-0726 1-0720 9 1-0713 1-0705 1-0696 1-0686 1-0674 1-0660 1-0644 1-0625 1-0604 1-0580 Table LXXXIIL Ammonia Specific Gravity and Concentration. G. Lunge and T. Wiernik, Zeit. angew. Chem., 2. 182. 1889; G. Lunge, Technical Chemists' Handbook, London, 1910. Per Grms. NH 3 Correction for Per Grms. NH 3 Correction for Sp. gr. cent. ATTT per 1000 c.c. temperature. + 1 Sp. gr. cent. AT CT per 1000 c.c. temperature. + 1 NH 3 . at 15. from 13 to 17. NH 3 . at 15. from 13 to 17. 1-000 0-00 0-0 0-00018 0-940 15-63 146-9 0-00039 0-998 0-45 4-5 0-00018 0-938 16-22 152-1 0-00040 0-996 0-91 9-1 0-00019 0-936 16-82 157-4 0-00041 0-994 1-37 13-6 0-00019 0-934 17-42 162-7 0-00041 0-992 1-84 18-2 0-00020 0-932 18-03 168-1 0-00042 0-990 2-31 22-9 0-00020 0-930 18-64 173-4 0-00042 0-988 2-80 27-7 0-00021 0-928 19-25 178-6 0-00043 0-986 3-30 32-5 0-00021 0-926 19-87 184-2 0-00044 0-984 3-80 37-4 0-00022 0-924 20-49 189-3 0-00045 0-982 4-30 42-2 0-00022 0-922 21-12 194-7 0-00046 0-980 4-80 47-0 0-00023 0-920 21-75 200-1 0-00047 0-978 5-30 51-8 0-00023 0-918 22-39 205-6 0-00048 0-976 5-80 56-6 0-00024 0-916 23-03 210-9 0-00049 0-974 6-30 61-4 0-00024 0-914 23-68 216-3 0-00050 0-972 6-80 66-1 0-00025 0-912 24-33 221-9 0-00051 0-970 7-31 70-9 0-00025 0-910 24-99 227-4 0-00052 0-968 7-82 75-7 0-00026 0-908 25-65 232-9 0-00053 0-966 8-33 80-5 0-00026 0-906 26-31 238-3 0-00054 0-964 8-84 85-2 0-00027 0-904 26-98 243-9 0-00055 0-962 9-35 89-9 0-00028 0-902 27-65 249-4 0-00056 0-960 9-91 95-1 0-00029 0-900 28-33 255-0 0-00057 0-958 10-47 100-3 0-00030 0-898 29-01 260-5 0-00058 0-956 11-03 105-4 0-00031 0-896 29-69 266-0 0-00059 0-954 11-60 110-7 0-00032 0-894 30-37 271-5 0-00060 0-952 12-17 115-9 0-00033 0-892 31-05 277-0 0-00060 0-950 12-74 121-0 0-00034 0-890 31-75 282-6 0-00061 0-948 13-31 126-2 0-00035 0-888 32-50 288-6 0-00062 0-946 13-88 131-3 0-00036 0-886 33-25 294-6 0-00063 0-944 14-46 136-5 0-00037 0-884 34-10 301-4 0-00064 9-942 15-04 141-7 0-00038 0-882 34-95 308-3 0-00065 684 A TREATISE ON CHEMICAL ANALYSIS. Table LXXXIV. Atomic Weights, International. $ co * -t d Jio II si it* Names nf II Is ^ Names of 1 x ^ |S B OI Elements. ' a M S2d 33? Elements. I If ii? <5 3 3^2 <3 +3 M -3 2 *4 37 * "^ II o Aluminium . Al 27- 27-1 Neodymium . . Nd 144 144-3 ; Antimony . Sb 120 120-2 Neon . . . Ne 20 20-2 j Argon . A 40 39-88 Nickel . Ni 59 58-68 Arsenic Barium As Ba 75 137 74-96 137-37 Niobium \ (Columbium) j Nb ) Cb ( 93-5 93-5 Beryllium ) (Glucinum) i Be ) Gl j 9 9-1 Niton (radium \ emanation) / Nt 222-4 222-4 Bismuth Bi 208 208-0 Nitrogen . . N 14 14-01 Boron . B 11 11-0 Osmium . . Os 191 1909 Bromine Br 80 79-92 Oxygen . , 16 16 Cadmium Cd 112 112-4 Palladium ... Pd 106 106-7 Caesium Cs 133 132-81 Phosphorus . P 31 31-04 Calcium Ca 40 40-07 Platinum . . Pt 195 195-2 Carbon C 12 12-00 Potassium . K 39 39-10 Cerium Ce 140 140-25 Praseodymium . | Pr 140-5 140-6 Chlorine Cl 35-5 35-46 Radium . . Ra 226-5 226-4 Chromium . Cr 52 52-0 Rhodium Rh 103 102-9 Cobalt Co 59 58-97 Rubidium . Rb 85 85-45 Copper Cu 63-5 63-57 Ruthenium . . Ru 101-5 101-7 Dysprosium . Dy 162-5 162-5 Samarium . Sm 150 150-4 Erbium Er 167-5 167-5 Scandium . . Sc 44 44-1 Europium . Eu 152 152-0 Selenium Se 79 79-2 Fluorine F 19 19-0 Silicon . Si 28 28-3 Gadolinium . Gd 157 157-3 Silver . . .. Ag 108 107-88 Gallium Ga 70 69-9 Sodium . . Na 23 23-00 Germanium . Ge 72 72-5 Strontium . Sr 87-5 87-63 Gold . Au 197 197-2 Sulphur . . S 32 32-07 Helium He 4 3-99 Tantalum Ta 181-5 181-5 Holmium Ho 163-5 163-5 Tellurium . Te 127 127-5 Hydrogen . H 1 1-008 Terbium Tb 159 159-2 Indium . . In 115 114-8 Thallium . Tl 204 204-0 Iodine I 127 126-92 Thorium Th 232 232-4 Iridium Ir 193 193-1 Thulium Tm 168-5 168-5 Iron Fe 56 55-85 Tin . Sn 118 119-0 Krypton Kr i 83 82-9 Titanium Ti 48 48-1 Lanthanum . La ! 139 139-0 Tungsten W 184 184-0 Lead . . ' . Pb 207 207-10 Uranium U 238-5 238-5 Lithium Li 7 6-94 Vanadium . V 51 51-0 Lutecium Lu 174 174-0 Xenon . . X 130 130-2 Magnesium . Mg 24 24-32 Ytterbium . Yb 172 172-0 Manganese . Mn 55 54-93 Yttrium . . Y 89 89-0 Mercury Hg 200 200-6 Zinc . . Zn 65 65-37 Molybdenum Mo 96 96-0 Zirconium . Zr 90-5 90-6 APPENDIX. 685 Table LXXXV. The Vapour Pressure of Liquid Water at Different Temperatures in Millimetres of Mercury. Based mainly 011 K. Scheel and W. Heuse, Annalen der Physik (4), 29. 723 1909 (4), 31. 715, 1910. HYDROGEN SCALE, GRAVITY NORMAL. Tempera- Millimetres of Mercury. ture. 1 2 3 4 5 6 7 8 9 -4 3-40 3-38 3-35 3-33 3-30 3-28 3-26 3-23 3-21 3-18 -3 3-67 3-64 3-62 3-59 3-56 3-53 3-51 3-48 3-46 3-43 -2 3-95 3-92 3-89 3-87 3-84 3-81 3-78 3-75 3-72 3-70 -1 4-26 4-22 4-19 4-16 4-13 4-10 4-07 4-04 4-01 3-98 -0 4-58 4-55 4-51 4-48 4-45 4-41 4-38 4-35 4-32 4-29 +0 4-58 4-61 4-65 4-68 4-72 4-75 4-79 4-82 4-86 4-89 1 4-93 4-96 5-00 5-03 5-07 5-11 5-14 - 5-18 5-22 5-26 2 5-29 5-33 5-37 5-41 5-45 5-49 5-53 5-57 5-60 5-64 3 5-69 5-73 5-77 5-81 5-85 5-89 5-93 5-97 6-02 6-06 4 6-10 6-14 6-19 6-23 6-27 6-32 6-36 6-41 6-45 6-50 5 6-54 6-59 6-64 6-68 6-73 6-78 6-82 6-87 6-92 6-97 6 7-01 7-06 7-11 7-16 7-21 7-26 7-31 7-36 7-41 7-46 7 7-51 7-57 7-62 7-67 7-72 7-78 7-83 7-88 7-94 7-99 8 8-05 8-10 8-16 8-21 8-27 8-32 8-38 8-44 8-50 8-55 9 8-61 8-67 8-73 8-79 8-85 8-91 8-97 9-03 9-09 9-15 10 9-21 9-27 9-33 9-40 9-46 9-52 9-59 9-65 9-72 9-78 11 9-85 9-91 9-98 10-04 10-11 10-18 10-25 10-31 10-38 10-45 12 10-52 10-59 10-66 10-73 10-80 10-87 10-94 11-01 11-09 11-16 13 11-23 11-31 11-38 11-46 11-53 11-61 11-68 11-76 11-84 11-91 14 11-99 12-07 12-15 12-23 12-30 12-38 12-46 12-54 12-63 12-71 15 12-79 12-87 12-96 13-04 13-12 13-21 13-29 13-38 13-46 13-55 16 13-64 13-72 13-81 13-90 13-99 14-08 14-17 14-26 14-35 14-44 17 14-53 l 14-63 14-72 14-81 14-91 15-00 15-10 15-19 15-29 15-38 ; 18 15-48 15-58 15-68 15-78 15-87 15-97 16-07 16-18 16-28 16-38 19 16-48 16-59 16-69 16-79 16-90 17-00 17-11 17-21 17-32 17-43 ; 20 17-54 17-65 17-76 17-87 17-98 18-09 18-20 18-31 18-43 18-54 21 18-66 18-77 18-89 19-00 19-12 19-24 19-35 19-47 19-59 19-71 22 19-83 ! 19-95 20-08 20-20 20-32 20-45 20-57 20-70 20-82 20-95 23 21-07 21-20 21-33 21-46 21-59 21-72 21-85 21-98 22-12 22-25 24 22-38 22-52 22-65 22-79 22-93 23-07 23-20 23-34 23-48 23-62 25 23-76 23-91 24-05 24-19 24-33 24-48 24-63 24-77 24-92 25-07 26 25-22 25-37 25-52 25-67 25-82 25-97 26-13 26-28 26-43 26-59 27 26-75 26-91 27-06 27-22 27-38 27-54 27-70 27-87 28-03 28-19 28 28-36 28-52 28-69 28-86 29-02 29-19 i 29-36 29-54 29-71 29-88 29 30-05 30-23 30-40 30-58 30-75 30-93 31-11 31-29 31-47 31-65 30 31-83 32-02 32-21 32-39 32-57 32-76 32-95 33-14 33-33 33-52 31 33-71 33-90 34-09 34-29 34-48 34-68 34-88 35-07 35-27 35-47 32 35-67 35-88 36-08 36-28 36-49 36-70 36-90 37-11 37-32 37-53 33 37-74 37-95 38-17 38-38 38-60 38-81 39-03 39-25 39-47 39-70 34 39-91 40-13 40-36 40-58 40-81 41-04 41-26 41-49 41-72 41-96 686 A TREATISE ON CHEMICAL ANALYSIS. Table LXXXV. continued. Tempera- Millimetres of Mercury. ture. 1 2 3 4 5 6 7 8 9 35 42-19 42-42 42-66 42-89 43-13 43-37 43-61 43-85 44-09 44-33 36 44-58 44-82 45-06 45-31 45-56 45-81 46-07 46-32 46-57 46-83 37 47-08 47-34 47-60 47-86 48-12 48-38 48-64 48-91 49-17 49-44 38 49-71 49-98 50-25 50-52 50-79 51-07 51-34 51-62 51-90 52-18 39 52-46 52-74 53-03 53-31 53-60 53-88 54-17 54-46 54-75 55-05 40 55-34 55-64 55-93 56-23 56-53 56-83 57-14 57-44 57-74 58-05 41 58-36 58-67 58-98 59-29 59-61 59-92 60-24 60-56 60-88 61-20 42 61-52 61-84 62-17 62-49 62-82 63-15 63-48 63-82 64-15 64-49 43 64-82 65-16 65-50 65-84 66-19 66-53 66-88 67-23 67-58 67-93 44 68-28 68-64 68-99 69-35 69-71 70-07 70-43 70-79 71-16 71-53 45 71-90 72-27 72-64 73-01 73-39 73-76 74-14 74-52 74-90 75-29 46 75-67 76-06 76-45 76-84 77-23 77-63 78-02 78-42 78-82 79-22 47 79-62 80-03 80-43 80-84 81-25 81-66 82-07 82-49 82-91 83-32 48 83-74 84-17 84-59 85-02 85-45 85-88 86-31 86-74 87-17 87-61 49 88-05 88-49 88-93 89-38 89-82 90-27 90-72 91-18 91-63 92-09 50 92-54 93-00 93-47 93-93 94-40 94-86 95-33 95-81 96-28 96-76 51 97-24 97-72 98-20 98-68 99-17 99-66 100-2 100-6 101-1 101-6 52 102-1 102-6 103-1 103-6 104-2 104-7 105-2 105-7 106-2 106-7 53 107-2 107-8 108-3 108-8 109-3 109-9 110-4 110-9 111-5 112-0 54 112-6 113-1 113-7 114-2 114-8 115-3 115-9 116-4 117-0 117-5 55 118-1 118-7 119-2 119-8 120-4 121-0 121-6 122-1 122-7 123-3 56 124-0 124-5 125-1 125-7 126-3 126-9 127-5 128-1 128-7 129-3 57 129-9 130-5 131-1 131-8 132-4 133-0 133-6 134-3 134-9 135-5 58 136-2 136-8 137-4 138-1 138-7 139-4 140-0 140-7 141-4 142-0 59 142-7 143-4 144-0 144-7 145-4 146-0 146-7 147-4 148-1 148-8 60 149-5 150-2 150-9 151-6 152-3 153-0 153-7 154-4 155-1 155-8 61 156-5 157-2 158-0 158-7 159-4 160-1 160-9 161-6 162-4 163-1 62 163-9 164-6 165-4 166-1 166-9 167-6 168-4 169-2 169-9 170-7 63 171-5 172-3 173-0 173-8 174-6 175-4 176-2 177-0 177-8 178-6 64 179-4 180-2 181-0 181-8 182-7 183-5 184-3 185-1 186-0 186-8 65 187-6 188-5 189-3 190-2 191-0 191-9 192-7 193-6 194-5 195-5 66 196-2 197-1 197-9 198-8 199-7 200-6 201-5 202-4 203-3 204-2 67 205-1 206-0 206-9 207-8 208-7 209-6 210-6 211-5 212-4 213-4 68 214-3 215-2 216-2 217-1 218-1 219-0 220-0 221-0 221-9 222-9 69 223-9 224-8 225-8 226-8 227-8 228-8 229-8 230-8 231-8 232-8 70 233-8 234-8 235-8 236-9 237-9 238-9 239-9 241-0 242-0 243-1 71 244-1 245-2 246-2 247-3 248-4 249-4 250-5 251-6 252-6 253-7 72 254-8 255-9 257-0 258-1 259-2 260-3 261-4 262-5 263-7 264-8 73 265-9 267-0 268-2 269-3 270-5 271-6 272-8 273-9 275-1 276-2 74 277-4 278-6 279-8 280-9 282-1 283-3 284-5 285-7 286-9 288-1 75 289-3 290-5 291-8 293-0 294-2 295-4 296-7 297-9 299-2 300-4 76 301-7 302-9 304-2 305-4 306-7 308-0 309-3 310-5 311-8 313-1 77 314-4 315-7 317-0 318-3 319-7 321-0 322-3 323-6 325-0 326-3 78 327-6 329-0 330-3 331-7 333-1 334-4 335-8 337-2 338-6 339-9 79 341-3 342-7 344-1 345-5 346-9 348-3 349-8 351-2 352-6 354-0 APPENDIX. Table LXXXV. continued. 687 Tempera- Millimetres of Mercury. ture. 1 2 3 4 5 6 7 8 9 80 355-5 356-9 358-4 359-8 361-3 362-7 364-2 365-7 367-1 368-6 81 370-1 371-6 373-1 374-6 376-1 377-6 379-1 380-7 382-2 383-7 82 385-2 386-8 388-3 389-9 391-5 393-0 394-6 396-2 397-7 399-3 83 400-9 402-5 404-1 405-7 407-3 408-9 410-5 412-2 413-8 415-4 84 417-1 418-7 420-3 422-0 423-7 425-4 427-1 428-7 430-4 432-1 85 433-8 435-5 437-2 438-9 440-6 442-4 444-1 445-8 447-6 449-3 86 451-1 452-8 454-6 456-4 458-1 459-9 461-7 463-5 465-3 467-1 87 468-9 470-7 472-5 474-4 476-2 478-0 479-9 481-7 483-6 485-5 88 487-3 489-2 491-1 493-0 494-9 496-8 498-7 500-6 502-5 504-4 89 506-4 508-3 510-2 512-2 514-1 516-1 518-1 520-0 522-0 524-0 90 526-0 528-0 530-0 532-0 534-0 536-1 538-1 540-1 542-1 544-2 91 546-3 548-3 550-4 552-5 554-6 556-7 558-7 560-8 563-0 565-1 92 567-2 ! 569-3 571-5 573-6 575-7 577-9 580-1 582-2 584-4 586-6 93 588-8 591-0 593-2 595-4 597-6 600-0 602-1 604-3 606-5 608-8 94 611-0 613-3 615-6 617-9 620-1 622-4 624-7 627-0 629-4 631-7 95 634-0 636-3 638-7 641-0 643-4 645-8 648-1 650-5 652-9 655-3 96 657-7 660-1 662-5 664-9 667-4 669-8 672-3 674-7 677-2 679-6 97 682-1 684-6 687-1 689-6 692-1 694-6 697-1 699-7 702-2 704-7 98 707-2 709-8 712-4 715-0 717-6 720-2 722-8 725-4 728-0 730-6 99 733-2 735-9 738-5 741-2 743-9 746-5 749-2 751-9 754-6 757-3 100 760-0 762-7 765-5 768-2 770-9 773-7 776-4 779-2 782-0 784-8 Table LXXXVI. Conversion of Grms. Potassium Chloroplatinate into Grms. Potassium Chloride. (FACTOR 0-307.) Grm. KC1. K 2 PtCl 6 . Grm. 1 2 3 4 5 6 7 8 9 0-000 0001 0001 0001 0002 0002 0002 0002 0003 0-001 0003 0003 0004 0004 0004 0005 0005 0005 0006 0006 0-002 0006 0006 0007 0007 0007 0008 0008 0008 0009 0009 0-003 0009 i 0009 0010 0010 0011 0011 0011 0012 0012 0012 0-004 0013 0013 0013 0013 0014 0014 0014 0015 0015 0015 0-005 0015 0016 0016 0016 i 0017 0017 0017 0017 0018 0018 0-006 0018 0019 0019 0019 i 0020 0020 0020 0021 0021 0021 . 0-007 0021 0022 0022 0022 0023 0023 0024 0024 0024 0025 0-008 0025 0025 0025 0026 0026 0026 0027 0027 0027 0028 0-009 0028 0028 0028 0029 0029 0029 0029 0030 0030 0030 688 A TREATISE ON CHEMICAL ANALYSIS. Table LXXXVI. K. 2 PtCl 6 to KC1. continued: K 2 PtCl 6 . Grm. KCL Grm. i 1 2 3 4 5 6 7 8 | 9 0-010 0031 0031 0031 0032 0032 0032 0032 0033 0033 0033 0-011 0034 0034 0034 0035 0035 0035 0036 0036 0036 0037 0-012 0037 0037 0037 0038 0038 0038 0039 0039 0039 0040 0-013 0040 0040 0041 0041 0041 0041 0042 0042 0042 0043 0-014 0043 0043 0044 0044 0044 0045 0045 0045 0045 0046 0-015 0046 0046 0047 0047 0047 0048 0048 0048 0049 0049 0-016 0049 0049 0050 0050 0050 0051 0051 0051 0052 0052 0-017 0052 0052 0053 0053 0053 0054 0054 0054 0055 0055 0-018 0055 0056 0056 0056 0056 0057 0057 0057 0058 0058 0-019 0058 0059 0059 0059 0060 0060 0060 0060 0061 0061 0-020 0061 0061 0062 0062 0062 0063 0063 0064 0064 0064 0-021 0065 0065 0065 0065 0066 0066 0066 0067 0067 0067 0-022 0068 0068 0068 0068 0069 0069 0069 0070 0070 0070 0-023 0070 0071 0071 0072 0072 0072 0072 0073 0073 0073 0-024 0074 0074 0074 0075 0075 0075 0076 0076 0076 0076 0-025 0077 0077 0077 0078 0078 0078 0079 0079 0079 0080 0-026 0080 0080 0080 0081 0081 0081 0082 0082 0082 0083 0-027 0083 0083 0084 ! 0084 0084 0084 0085 0085 0085 0086 0-028 0086 0086 | 0087 0087 0087 0087 0088 0088 0088 0089 0-029 0089 0089 0090 0090 0090 0091 0091 0091 0091 0092 0-030 0092 0092 0092 0093 0093 0094 0094 0094 0095 0095 0-031 0095 0095 0096 0096 0096 0097 0097 0097 0098 0098 0-032 0098 0099 0099 0099 0099 0100 0100 0100 0101 0101 0-033 0101 0102 0102 0102 0103 0103 0103 0103 0104 0104 0-034 0104 0105 0105 0105 0106 0106 0106 0107 0107 0107 0-035 0107 0108 0108 0108 0108 0109 0109 0109 0110 0110 0-036 0110 0111 0111 0111 0111 0112 0112 0113 0113 0113 0-037 0114 0114 0114 0115 0115 0115 0115 0116 0116 0116 0-038 0117 0117 0117 0118 0118 0118 0119 i 0119 0119 0119 0-039 0120 0120 0120 0121 0121 0121 0122 I 0122 0122 0123 0-040 0123 0123 0123 0124 1 0124 0124 0125 0125 0126 0126 0-041 0126 0126 0126 0127 0127 0127 0128 0128 0128 0129 0-042 0129 0129 0129 0130 ! 0130 0130 0131 0131 0131 0132 0-043 0132 0132 0133 0133 0133 0133 0134 i 0134 0134 0135 0-044 0135 0135 0136 0136 0136 0137 0137 0137 0137 0138 0-045 0138 0139 0139 0139 0140 0140 0140 0140 0141 0141 0-046 0141 0141 0142 0142 0142 0143 0143 0143 0144 0144 0-047 0144 0144 0145 0145 0145 0146 0146 0146 0147 0147 0-048 0147 0148 0148 0148 : 0148 0149 0149 0149 0150 0150 0-049 0150 0151 0151 0151 0152 0152 0152 1 0152 0153 0153 0-050 0153 0154 0154 0154 0155 0155 0155 0156 0156 0156 0-051 0156 0157 0157 0157 0158 0158 0158 0159 0159 0159 0-052 0160 0160 0160 0160 0161 0161 0161 0162 0162 0162 0-053 0163 0163 0163 0164 0164 0164 0164 0165 0165 0165 0-054 0166 0166 0166 0167 0167 0167 0168 0168 0168 0168 APPENDIX. <58 9 Table LXXXVL K 2 PtCl 6 to KCL continued. K 2 PtCl 6 . Grm. Grm. KC1. i 1 2 3 4 5 6 7 8 9 0-055 0169 0169 0170 0170 0170 0170 0171 0171 0171 0172 0-056 0172 0172 0172 0173 0173 0173 0174 01V4 0174 0175 0-057 0175 0175 0176 0176 0176 0176 0177 0177 0177 0178 0-058 0178 0178 0179 0179 0179 0179 0180 0180 0180 0181 0-059 0181 0181 0182 0182 0182 0183 0183 0183 0183 0184 0-060 0184 0184 0185 0185 0185 0186 0186 0186 0187 0187 0-061 0187 0187 0188 0188 0189 0189 0189 0190 0190 0190 0-062 0190 0191 0191 0191 0192 0192 0192 0192 0193 0193 0-063 0193 0194 0194 0194 0195 0195 0195 0195 0196 0196 0-064 0197 0197 0197 0198 0198 0198 0199 0199 0199 0199 0-065 0199 0200 0200 0200 0201 0201 0201 0202 0202 0202 0-066 0203 0203 0203 0203 0204 0204 0204 0205 0205 0205 0-067 0206 0206 ! 0206 0207 0207 0207 0207 0208 0208 0208 0-068 0209 0209 0209 0210 0210 0210 0211 0211 0211 0211 0-069 0212 0212 0212 0213 0213 0213 0214 0214 0214 0214 0-070 0215 0215 0215 0216 0216 0217 0217 0217 0217 0218 0-071 0218 0218 0218 0219 0219 0219 0219 0220 0220 0221 0-072 0221 0221 0222 0222 0222 0222 0223 0223 0223 0224 0-073 0224 | 0224 0225 0225 0225 0226 0226 0226 0227 0227 0-074 0227 0227 0228 0228 0228 0229 0229 0229 0230 0230 0-075 0230 0230 0231 0231 0231 0232 0232 0232 0233 0233 0-076 0233 0234 0234 0234 0234 0235 0235 0235 0236 0236 0-077 0236 0237 0237 0237 0238 0238 0238 0238 0239 0239 0-078 0239 0240 0240 0240 ! 0241 0241 0241 0242 0242 0242 0-079 0242 0243 0243 0243 0244 0244 0244 0245 0245 0245 0-080 0246 0246 0246 0246 0247 0247 0247 0248 0248 0248 0-081 0249 0249 0249 0249 0250 0250 0250 0251 0251 0251 0-082 0252 0252 0252 0253 0253 0253 0253 0254 0254 0254 0-083 0255 0255 0255 0256 0256 0257 0257 0257 0257 0257 0-084 0258 0258 0258 0259 0259 0259 0260 0260 0260 0261 0-085 0261 0261 0261 0262 0262 0262 0263 0263 0263 0264 0-086 0264 0264 0265 0265 0265 0265 0266 0266 0266 0267 0-087 0267 0267 0268 0268 0268 0269 0269 0269 0269 0270 0-088 0270 0270 0271 0271 0271 0272 0272 0272 0273 0273 0-089 0273 0273 0273 0274 0274 0275 0275 0275 0276 0276 0-090 0276 0277 0277 0277 0277 0278 0278 0278 0279 0279 0-091 0279 0280 0280 0280 0280 0281 0281 0281 0282 0282 0-092 0282 0283 0283 0283 0283 0284 0284 0284 0285 0285 0-093 0285 0286 0286 0286 0287 0287 0287 0288 0288 0288 0-094 0288 0289 0289 0289 0290 0290 0290 0291 0291 0291 0-095 0292 0292 0292 0292 0293 0293 0293 0294 0294 0294 0-096 0295 0295 0295 0296 0296 0296 0296 0297 0297 0297 0-097 0298 0298 0298 0299 0299 0299 0300 0300 0300 0300 0-098 0301 0301 0301 0302 0302 0302 0303 0303 0303 0304 0-099 0304 0304 0304 0305 0305 0305 0306 0306 0306 0307 1 44 690 A TREATISE ON CHEMICAL ANALYSIS. Table LXXXVI. K 2 PtCl 6 to KC1. continued. K 2 PtCl 6 . Grm. Grm. KC1. 1 2 3 4 5 6 7 8 9 0-100 0307 0307 0308 0308 0308 0309 0309 0309 0309 0309 0-101 0310 0310 0310 0311 0311 0311 0312 0312 0312 0313 0-102 0313 0313 0313 0314 0314 0314 0315 0315 0315 0316 0-103 0316 0316 0317 0317 0317 0317 0318 0318 0318 0319 0-104 0319 0319 0320 0320 0320 0321 0321 0321 0321 0322 0-105 0322 0322 0323 0323 0323 0324 0324 0324 0325 0325 0-106 0325 0325 0326 0326 0326 0327 0327 0327 0328 0328 0-107 0328 0328 0329 0329 0329 0330 0330 0330 0331 0331 0-108 0331 0332 0332 0332 0332 0333 0333 0333 0334 0334 0-109 0334 0335 0335 0335 0336 0336 0336 0337 0337 0337 0-110 0338 0338 0338 0339 0339 0339 0340 0340 0340 0340 0-111 0341 0341 0341 0342 0342 0342 0343 0343 0343 0344 ' 0-112 0344 0344 0344 0345 0345 0345 0346 0346 0346 0347 0-113 0347 0347 0348 0348 0348 0348 0349 0349 0349 0350 0-114 0350 0350 0351 0351 0351 0352 0352 0352 0352 0353 0-115 0353 0353 0354 0354 0354 0355 0355 0355 0356 0356 0-116 0356 0356 0357 0357 0357 0358 0358 0358 0359 0359 0-117 0359 0359 0360 0360 0360 0361 0361 0361 0362 0362 0-118 0362 0363 0363 0363 0363 0364 0364 0364 0365 0365 0-119 0365 0366 0366 0366 0367 0367 0367 0367 0368 0368 0-120 0368 0369 0369 0369 0370 0370 0370 0371 0371 0371 0-121 0371 0372 0372 0372 0373 0373 0373 0374 0374 0374 0-122 0375 0375 0375 0375 0376 0376 0376 0377 0377 0377 0-123 0378 0378 0378 0379 0379 0379 0379 0380 0380 0380 0-124 0381 0381 0381 0382 0382 0382 0383 0383 0383 0383 0-125 0384 0384 0384 0385 0385 0385 0386 0386 0386 0387 0-126 0387 0387 0387 0388 0388 0388 0389 0389 0389 0390 0-127 0390 0390 0391 0391 0391 0392 0392 0392 0393 0393 0-128 0393 0393 0394 0394 0394 0394 0395 0395 0395 0396 0-129 0396 0396 0397 0397 0397 0398 0398 0398 0398 0399 0-130 0399 0399 0400 0400 0400 0401 0401 0401 0402 0402 0-131 _ 0402 0402 0403 0403 0403 0404 0404 0404 0405 0405 0-132 0405 0406 0406 0406 0407 0407 0407 0407 0408 0408 0-133 0408 0409 0409 0409 0410 0410 0410 0410 0411 0411 0-134 0411 0412 0412 0412 0413 0413 0413 0414 0414 0414 0-135 0414 0415 0415 0415 0416 0416 0416 0417 0147 0417 0-136 0418 0418 0418 0418 0419 0419 0419 0420 0420 0420 0-137 0421 0421 0421 ! 0422 0422 0422 0423 0423 0423 0424 0-138 0424 0424 0424 0425 0425 0425 0426 0426 0426 0426 0-139 0427 0427 0427 0428 0428 0428 0429 0429 0429 0429 0-140 0430 0430 0430 0431 0431 0431 0432 0432 0432 0433 0-141 0433 0433 0433 0434 0434 0434 0435 0435 0435 0436 0-142 0436 0436 0437 0437 0437 0437 0438 4038 0438 0439 0-143 0439 0439 0440 0440 0440 0441 0441 0441 0441 0442 0-144 0442 0442 0443 0443 0443 0444 0444 0444 0445 0445 APPENDIX. Table LXXXVI. K 2 PtCl G to KCl.-continued. 691 K 2 PtCl 6 . Grm. Grm. KC1. 1 2 3 4 5 6 7 8 9 0-145 0445 0445 0446 0446 0446 0447 0447 0447 0448 0448 0-146 0448 0449 0449 0449 0449 0450 0450 0450 0451 0451 0-147 0451 0452 0452 0452 0453 0453 0453 0453 0454 0454 0-148 0454 0455 0455 0455 0456 0456 0456 0457 0457 0457 0-149 0457 0458 0458 0458 0459 0459 0459 0460 0460 0460 0-150 0461 0461 0461 0461 0462 0462 0462 0463 0463 0463 0-151 0464 0464 0464 0464 0465 0465 0465 0466 0466 0466 0-152 0467 0467 0467 0468 0468 0468 0468 0469 0469 0469 0-153 0470 0470 0470 0471 0471 0471 0472 0472 0472 0472 0-154 0473 0473 0473 0474 0474 0474 0475 0475 0475 0476 0-155 0476 0476 0476 0477 0477 0477 0478 0478 0478 0479 0-156 0479 0479 0480 0480 0480 0480 0481 0481 0481 0482 0-157 0482 0482 0483 0483 0483 0484 0484 0484 0484 0485 0-158 0485 0485 0486 0486 0486 0487 0487 0487 0488 0488 0-159 0488 0488 0489 0489 0489 0490 0490 0490 0491 0491 0-160 0491 0492 0492 0492 0492 0493 0493 0493 0494 0494 0-161 0494 0495 0495 0495 0495 0496 0496 0496 0497 0497 0-162 0497 0498 0498 0498 0499 0499 0499 0499 0500 0500 0-163 0500 0501 0501 0501 0502 0502 0502 0503 0503 0503 0-164 0503 0504 0504 0504 0505 0505 0505 0506 0506 0506 0-165 0507 0507 0507 0507 0508 0508 0508 0509 0509 0509 0-166 0510 0510 0510 0511 0511 0511 0511 0512 0512 0512 0-167 0513 0513 0513 0514 0514 0514 0515 0515 0515 0515 0-168 0516 0516 0516 0517 0517 0517 0518 0518 0518 0519 0-169 0519 0519 0519 0520 0520 0520 0521 0521 0521 0522 0-170 0522 0522 0523 0523 0523 0523 0524 0524 0524 0525 0-171 0525 0525 0526 0526 0526 0527 0527 0527 0527 0528 0-172 0528 0528 0529 0529 0529 0530 0530 0530 0530 0531 0-173 0531 0531 0532 0532 0532 0533 0533 0533 0534 0534 0-174 0534 0534 0535 0535 0535 0536 0536 0536 0537 0537 0-175 0537 0538 0538 0538 0538 0539 0539 0539 0540 0540 0-176 0540 0541 0541 0541 0542 0542 0542 0542 0543 0543 0-177 0543 0544 0544 0544 0545 0545 0545 0546 0546 0546 0-178 0546 0547 0547 0547 0548 0548 0548 0549 0549 0549 0-179 0550 0550 0550 0550 0551 0551 0551 0552 0552 0552 0-180 0553 0553 0553 0554 0554 0554 0554 0555 0555 0555 0-181 0556 0556 0556 0557 0557 0557 0558 0558 0558 0558 0-182 0559 0559 0559 0560 0560 0560 0561 0561 0561 0562 0-183 0562 0562 0562 0563 0563 0563 0564 0564 0564 0565 0-184 0565 0565 0565 0566 0566 0566 0567 0567 0567 0568 0-185 0568 0568 0569 0569 0569 0569 0570 0570 0570 0571 0-186 0571 0571 0572 0572 0572 0573 0573 0573 0573 0574 0-187 0574 0574 0575 0575 0575 0576 0576 0576 0577 0577 0-188 0577 0577 0578 0578 0578 0579 0579 0579 0580 0580 0-189 0580 0581 0581 0581 0581 0582 0582 0582 0583 0583 I 692 A TREATISE ON CHEMICAL ANALYSIS. Table LXXXVL K 2 PtCl 6 to KCL continued. K 2 PtCl 6 . Grm. KCL Grm. 1 2 3 4 5 6 7 8 9 0-190 0583 0584 0584 0584 0585 0585 0585 0585 0586 0586 0-191 0586 0587 0587 0587 0588 0588 0588 0589 0589 0589 0-192 0589 0590 0590 0590 0591 0591 0591 0592 0592 0592 0-193 0593 0593 0593 0593 0594 0594 0594 0595 0595 0595 0-194 0596 0596 0596 0597 0597 0597 0597 0598 0598 0598 0-195 0599 0599 0599 0600 0600 0600 0600 0601 0601 0601 0-196 0602 0602 0602 0603 0603 0603 0604 0604 0604 0604 0-197 0605 0605 0605 0606 0606 0606 0607 0607 0607 0608 0-198 0608 0608 0608 0609 0609 0609 0610 0610 0610 0611 0-199 0611 0611 0612 0612 0612 0612 0613 0613 0613 0614 0-200 0614 0614 0615 0615 0615 0616 0616 0616 0616 0617 0-201 0617 0617 0618 0618 0618 0619 0619 0619 0620 0620 0-202 0620 0620 0621 0621 0621 0622 0622 0622 0623 0623 0-203 0623 0624 0624 0624 0624 0625 0625 0625 0626 0626 0-201 0626 0627 0627 0627 0628 0628 0628 0628 0629 0629 0-205 0629 0630 0630 0630 0631 0631 0631 0631 0632 0632 0-206 0632 0633 0633 0633 0634 0634 0634 0635 0635 0635 0-207 0635 0636 0636 0636 0637 0637 0637 0638 0638 0638 0-208 0639 0639 0639 0639 0640 0640 0640 0641 0641 0641 0-209 0642 0642 0642 0643 0643 0643 0643 0644 0644 0644 0-210 0645 0645 0645 0646 0646 0646 0647 0647 0647 0647 0-211 0648 0648 0648 0649 0649 0649 0650 0650 0650 0651 0-212 0651 0651 0651 0652 0652 0652 0653 0653 0653 0654 0-213 0654 ! 0654 0655 0655 0655 0655 0656 0656 0656 0657 0-214 0657 0657 0658 0658 0658 0659 0659 0659 0659 0660 0-215 0660 0660 0661 0661 0661 0662 0662 0662 0663 0663 0-216 0663 0663 0664 0664 0664 0665 0665 0665 0666 0666 0-217 0666 0666 0667 0667 0667 0668 0668 0668 0669 0669 0-218 0669 0670 0670 0670 0670 0671 0671 0671 0672 0672 0-219 0672 0673 0673 0673 0674 0674 0674 0674 0675 0675 0-220 0675 0676 0676 0676 0677 0677 0677 0678 0678 0678 0-221 0678 0679 0679 0679 0680 0680 0680 0681 0681 0681 0-222 0682 0682 0682 0682 0683 0683 0683 0684 0684 0684 0-223 0685 0685 0685 0686 0686 0686 0686 0687 0687 0687 0-224 0688 0688 0688 0689 0689 0689 0690 0690 0690 0690 0-225 0691 0691 0691 0692 0692 0692 0693 0693 0693 0694 0-226 0694 0694 0694 0695 0695 0695 0696 0696 0696 0697 0-227 0697 0697 0698 0698 0698 0698 0699 0699 0699 0700 0-228 0700 0700 0701 0701 0701 0701 0702 0702 0702 0703 0-229 0703 0703 0704 0704 0704 0705 0705 0705 0705 0706 0-230 0706 0706 0707 0707 0707 0708 0708 0708 0709 0709 0-231 0709 0709 0710 0710 0710 0711 0711 0711 0712 0712 0-232 0712 0713 0713 0713 0713 0714 0714 0714 0715 0715 0-233 0715 0716 0716 0716 0716 0717 0717 0717 0718 0718 0-234 0718 0719 0719 0719 0720 0720 0720 0721 0721 0721 APPENDIX. 693 Table LXXXVL K 2 PtCl 6 to KCL continued. K 2 PtCl 6 . Grm. KC1. Grm. 1 2 3 4 5 6 7 8 9 0-235 0721 0722 0722 0722 0723 0723 0723 0724 0724 0724 0-236 0725 0725 0725 0725 0726 0726 0726 0727 0727 0727 0-237 0728 0728 0728 0729 0729 0729 0729 0730 0730 0730 0-238 0731 0731 0731 0732 0732 0732 0733 0733 0733 0733 0-239 0734 0734 0734 0735 0735 0735 0736 0736 0736 0736 0-240 0737 0737 0737 0738 0738 0738 0739 0739 0739 0740 0-241 0740 0740 0740 0741 0741 0741 0742 0742 0742 0743 0-242 0743 0743 0744 0744 0744 0744 0745 0745 0745 0746 0-243 0746 0746 0747 0747 0747 0748 0748 0478 0748 0749 0-244 0749 0749 0750 0750 0750 0751 0751 0751 0752 0752 0-245 0752 0752 0753 0753 0753 0754 0754 0754 0755 0755 0-246 0755 0756 0756 0756 0756 0757 0757 0757 0758 0758 0-247 0758 0759 0759 0759 0760 0760 0760 0760 0761 0761 0-248 0761 0762 0762 0762 0763 0763 0763 0764 0764 0764 0-249 0764 0765 0765 0765 0766 0766 0766 0767 0767 0767 0-250 0768 0768 0768 0768 0769 0769 0769 0770 0770 0770 0-251 0771 0771 0771 0771 0772 0772 0772 0773 0773 0773 0-252 0774 0774 0774 0775 0775 0775 0775 0776 0776 0776 0-253 0777 0777 0777 0778 0778 0778 0779 0779 0779 0779 0-254 0780 0780 0780 0781 0781 0781 0782 0782 0782 0783 0-255 0783 0783 0783 0784 0784 0784 0785 0785 0785 0786 0-256 0786 0786 0787 0787 0787 0787 0788 0788 0788 0789 0-257 0789 0789 0790 0790 0790 0791 0791 0791 0791 0792 0-258 0792 0792 0793 0793 0793 0794 0794 0794 0795 0795 0-259 0795 0795 0796 0796 0796 0797 0797 0797 0798 0798 0-260 0798 0799 0799 0799 0799 0800 0800 0800 0801 0801 0-261 0801 0802 0802 0802 0802 0803 0803 0803 0804 0804 0-262 0804 0805 0805 0805 0806 0806 0806 0806 0807 0807 0-263 0807 0808 0808 0808 0809 0809 0809 0810 0810 0810 0-264 0810 0811 0811 0811 0812 0812 0812 0813 0813 0813 0-265 0814 0814 0814 0814 0815 0815 0815 0816 0816 0816 0-266 0817 0817 0817 0818 0818 0818 0818 0819 0819 0819 0-267 0820 0820 0820 0821 0821 0821 0821 0822 0822 0822 0-268 0823 0823 0823 0824 0824 0824 0825 0825 0825 0825 0-269 0826 0826 0826 0827 0827 0827 0828 0828 0828 0829 0-270 0829 0829 0829 0830 0830 0830 0831 0831 0831 0831 0-271 0832 0832 0833 0833 0833 0834 0834 0834 0834 0835 0-272 0835 0835 0836 0836 0836 0837 0837 0837 0837 0838 0-273 0838 0838 0839 0839 0839 0840 0840 0840 0841 0841 0-274 0841 0841 0842 0842 0842 0843 0843 0843 0844 0844 0-275 0844 0845 0845 0845 0845 0846 0846 0846 0847 0847 0-276 0847 0848 0848 0848 0849 0849 0849 0849 0850 0850 ' 0-277 0850 0851 0851 0851 0852 0852 0852 0853 0853 0853 0-278 0853 0854 0854 0854 0855 0855 0855 0856 0856 0856 0-279 0857 0857 0857 0857 0858 0858 0858 0859 0859 0859 694 A TREATISE ON CHEMICAL ANALYSIS. Table LXXXVI. K 2 PtCl fi to KC1. continued. K 2 PtCl 6 . Grm. KC1, Grm. 1 2 3 4 5 6 7 8 9 0-280 0860 0860 0860 0861 0861 0861 0861 0862 0862 0862 0-281 0863 0863 0863 0864 0864 0864 0865 0865 0865 0865 0-282 0866 0866 0866 0867 0867 0867 0868 0868 0868 0869 0-283 0869 0869 0869 0870 0870 0870 0871 0871 0871 0872 0-284 0872 0872 0872 0873 0873 0873 0874 0874 0874 0875 0-285 0875 0875 0876 0876 0876 0876 0877 0877 0877 0878 0-286 0878 0878 0879 0879 0879 0880 0880 0880 0880 0881 0-287 0881 0881 0882 0882 0882 0883 0883 0883 0884 0884 0-288 0884 0884 0885 0885 0885 0886 0886 0886 0887 0887 0-289 0887 0888 0888 0888 0888 0889 0889 0889 0890 0890 0-290 0890 0891 0891 0891 0892 0892 0892 0892 0893 0893 0-291 0893 0894 0894 0894 0895 0895 0895 0896 0896 0896 0-292 0896 0897 0897 0897 0898 0898 0898 0899 0899 0899 0-293 0900 0900 0900 0900 0901 0901 0901 0902 0902 0902 0-294 0903 0903 0903 0904 0904 0904 0904 0905 0905 0905 0-295 0906 0906 0906 0907 0907 0907 0907 0908 0908 0908 0-296 0909 0909 0909 0910 0910 0910 0911 0911 0911 0911 0-297 0912 0912 0912 0913 0913 0913 0914 0914 0914 0915 0-298 0915 0915 0915 0916 0916 0916 0917 0917 0917 0918 0-299 0918 0918 0919 0919 0919 0919 0920 0920 0920 0921 0-300 0921 0921 0922 0922 0922 0923 0923 0923 0923 0924 0-301 0924 0924 0925 0925 0925 0926 0926 0926 0927 0927 0-302 0927 0927 0928 0928 0928 0929 0929 0929 0930 0930 0-303 0930 0931 0931 0931 0931 0932 0932 0932 0933 0933 0-304 0933 0934 0934 0934 0935 0935 0935 0935 0936 0936 0-305 0936 0937 0937 0937 0938 0938 0938 0938 0939 0939 0-306 0939 0940 0940 0940 0941 0941 0941 0942 0942 0942 0-307 0942 0943 0943 0943 0944 0944 0944 0945 0945 0945 0-308 0946 0946 0946 0946 0947 0947 0947 0948 0948 0948 0-309 0949 0949 0949 0950 0950 0950 0950 0951 0951 0951 0-310 0952 0952 0952 0953 0953 0953 0954 0954 0954 0955 0-311 0955 0955 0955 0956 0956 0956 0957 0957 0957 0958 0-312 0958 0958 0959 0959 0959 0959 0960 0960 0960 0961 0-313 0961 0961 0962 0962 0962 0963 0963 0963 0964 0964 0-314 0964 0965 0965 0965 0966 0966 0966 0966 0967 0967 0-315 0967 0968 0968 0968 0969 0969 0969 0969 0970 0970 0-316 0970 0970 0971 0971 0971 0972 0972 0972 0973 0973 0-317 0973 0973 0974 0974 0974 0975 0975 0975 0976 0976 0-318 0976 0977 0977 0977 0977 0978 0978 0978 0979 0979 0-319 0979 0980 0980 0980 0981 0981 0981 0981 0982 0982 0-320 0982 0983 0983 0983 0984 0984 0984 0985 0985 0985 0-321 0985 0986 0986 0986 0987 0987 0987 0988 0988 0988 0-322 0989 0989 0989 0989 0990 0990 0990 0991 0991 0991 0-323 0992 0992 0992 0993 0993 0993 0993 0994 0994 0994 0-324 0994 0995 0995 0996 0996 0996 0997 0997 0997 0997 0-325 0998 0998 0998 0999 0999 0999 1000 1000 1000 1001 APPENDIX. 695 Table LXXXVI L Conversion of Grms. of Potassium Perchlorate (KC1O 4 ) into Grms. Potassium Chloride (KC1). (FACTOR 0-538.) KC10 4 . Grm. 1 2 3 4 5 6 7 8 9 0-000 0001 0001 0002 0002 0003 0003 0004 0004 0005 0-001 0005 0006 0006 0007 0008 0008 0009 0009 0010 0010 0-002 0011 0011 0012 0012 0013 0013 0014 0015 0015 0016 0-003 0016 0017 0017 0018 0018 0019 0019 0020 0020 0021 0-004 0022 0022 0023 0023 0024 0024 0025 0025 0026 0026 0-005 0027 0027 0028 0029 0029 0030 0030 0031 0031 0032 0-006 0032 0033 0033 0034 0034 0035 0036 0036 0037 0037 0-007 0038 0038 0039 0039 0040 0040 0041 0041 0042 0043 0-008 0043 0044 0044 0045 0045 0046 0046 0047 0047 0048 0-009 0048 0049 0049 0050 0051 0051 0052 0052 0053 0053 0-010 0054 0054 0055 0055 0056 0056 0057 0058 0058 0059 0-011 0059 0060 0060 0061 0061 0062 0062 0063 0063 0064 0-012 0065 0065 0066 0066 0067 0067 0068 0068 0069 0069 0-013 0070 0070 0071 0072 0072 0073 0073 0074 0074 0075 0-014 0075 0076 0076 0077 0077 0078 0079 0079 0080 0080 0-015 0081 0081 0082 0082 0083 0083 0084 0084 0085 0086 0-016 0086 0087 0087 0088 0088 0089 0089 0090 0090 0091 0-017 0091 0092 0093 0093 0094 0094 0095 0095 0096 0096 0-018 0097 0097 0098 0098 0099 0100 0100 0101 0101 0102 0-019 0102 0103 0103 0104 0104 0105 0105 0106 0107 0107 0-020 0108 0108 0109 0109 0110 0110 0111 0111 0112 0112 0-021 0113 0114 0114 0115 0115 0116 0116 0117 0117 0118 0-022 0118 0119 0119 0120 0121 0121 0122 0122 0123 0123 0-023 0124 0124 0125 0125 0126 0126 0127 0128 0128 0129 0-024 0129 0130 0130 0131 0131 0132 0132 0133 0133 0134 0-025 0135 0135 0136 0136 0137 0137 0138 0138 0139 0139 0-026 0140 0140 0141 0141 0142 0143 0143 0144 0144 0145 0-027 0145 0146 0146 0147 0147 0148 0148 0149 0150 0150 0-028 0151 0151 0152 0152 0153 0153 0154 0154 0155 0155 0-029 0156 0157 0157 0158 0158 0159 0159 0160 0160 0161 0-030 0161 0162 0162 0163 0164 0164 0165 0165 0166 0166 0-031 0167 0167 0168 0168 0169 0169 0170 0170 0171 0172 0-032 0172 0173 0173 0174 0174 0175 0175 0176 0116 0177 0-033 0178 0178 0179 0179 0180 0180 0181 0181 0182 0182 0-034 0183 0183 0184 0184 0185 0186 0186 0187 0187 0188 0-035 0188 0189 0189 0190 0190 0191 0192 0192 0193 0193 0-036 0194 0194 0195 0195 0196 0196 0197 0197 0198 0199 0-037 0199 0200 0200 0201 0201 0202 0202 0203 0203 0204 0-038 0204 0205 0206 0206 0207 0207 0208 0208 0209 0209 0-039 0210 0210 0211 0211 0212 0213 0213 0214 0214 0215 0-040 0215 0216 0216 0217 0217 0218 0218 0219 0220 0220 0-041 0221 0221 0222 0222 0223 0223 0224 0224 0225 0225 0-042 0226 0226 0227 0228 0228 0229 0229 0230 0230 0231 0-043 0231 0232 0232 0233 0233 0234 0235 0235 0236 0236 0-044 0237 0237 0238 0238 0239 0239 0240 0240 0241 0242 696 A TREATISE ON CHEMICAL ANALYSIS. Table LXXXVII. KC1O 4 to KCL continued. KC10 4 . o 1 2 3 4 5 6 7 8 9 Grm. 0-045 0242 0243 0243 0244 0244 0245 0245 0246 0246 0247 0-046 0247 0248 0249 0249 0250 0250 0251 0251 0252 0252 0-047 0253 0253 0254 0254 0255 0256 0256 0257 0257 0258 0-048 0258 0259 ! 0259 0260 0260 0261 0261 0262 0263 0263 0-049 0264 0264 0265 0265 0266 0266 0267 0267 0268 0268 0-050 0269 0270 0270 0271 0271 0272 0272 0273 0273 0274 0-051 0274 0275 0275 0276 0277 0277 0278 0278 0279 0279 0-052 0280 0280 0281 0281 0282 0282 0283 0284 0284 0285 0-053 0285 0286 0286 0287 0287 0288 0288 0289 0289 0290 0-054 0291 0291 0292 0292 0293 0293 0294 0294 0295 0295 0-055 0296 0296 0297 0298 0298 0299 0299 0300 0300 0301 0-056 0301 0302 0302 0303 0303 0304 0305 0305 0306 0306 0-057 0307 0307 0308 0308 0309 0309 0310 0310 0311 0312 0-058 0312 0313 0313 0314 0314 0315 0315 0316 0316 0317 0-059 0317 0318 0318 0319 0320 0320 0321 0321 0322 0322 0-060 0323 0323 0324 0324 0325 0325 0326 0327 0327 0328 0-061 0328 0329 0329 0330 0330 0331 0331 0332 0332 0333 0-062 0334 0334 0335 0335 0336 0336 0337 0337 0338 0338 0-063 0339 0339 0340 0341 0341 0342 0342 0343 0343 0344 0-064 0344 0345 0345 0346 0346 0347 0348 0348 0349 0349 0-065 0350 0350 0351 0351 0351 0352 0352 0353 0354 0355 0-066 0355 0356 0356 0357 0357 0358 0359 0359 0359 0360 0-067 0360 0361 0362 0363 0363 0363 0364 0364 0365 0365 0-068 0366 0366 0367 0367 0368 0369 0369 0370 0370 0371 0-069 0371 0372 0372 0373 0373 0374 0374 0375 0376 0376 0-070 0377 0377 0378 0378 0379 0379 0380 0380 0381 0381 0-071 0382 0383 0383 0384 0384 0385 0385 0386 0386 0387 0-072 0387 0388 0388 0389 0390 0390 0391 0391 0392 0392 0-073 0393 0393 0394 0394 0395 0395 0396 0397 0397 0398 0-074 0398 0399 0399 0400 0400 0401 0401 0402 0402 0403 0-075 0404 0404 0405 0405 0406 0406 0407 0407 0408 0408 0-076 0409 0409 0410 0410 0411 0412 0412 0413 0413 0414 0-077 0414 0415 0415 0416 0416 0417 0417 0418 0419 0419 0-078 0420 0420 0421 0421 0422 0422 0423 0423 0424 0425 0-079 0425 0426 0426 0427 0427 0428 0429 0429 0430 0430 0-080 0430 0431 0431 0432 0433 0433 0434 0434 0435 0435 0-081 0436 0436 0437 0437 0438 0438 0439 0440 0440 0441 0-082 0441 0442 0442 0443 0443 0444 0445 0445 0445 0446 0-083 0447 0447 0448 0448 0449 0449 0450 0450 0451 0451 0-084 0452 0452 0453 0454 0454 0455 0455 0456 0456 0457 0-085 0457 0458 4)458 0459 0459 0460 0461 0461 0462 0462 0-086 0463 0463 0464 0464 0465 0465 0466 0467 0467 0468 0-087 0468 0469 0469 0470 0470 0471 0471 0472 0472 0473 0-088 0473 0474 0475 0475 0476 0476 0477 0477 0478 0478 0-089 0479 0479 0480 0480 0481 0482 0482 0483 0483 0484 0-090 0484 0485 0485 0486 0486 0487 0487 0488 0489 0489 0-091 0490 0490 0491 0491 0492 0492 0493 0493 0494 0494 0-092 0495 0495 0496 0497 0497 0498 0498 0499 0499 : 0500 0-093 0500 0501 0501 '0502 0503 0504 0504 0504 0505 0505 0-094 0506 0506 0507 0507 0508 0508 0509 0509 0510 0511 APPENDIX. Table LXXXVII.-KC1O 4 to KC1. continued. 697 KC10 4 . Grm. 1 2 3 4 5 6 7 8 9 0-095 0511 0512 0512 0513 0513 0514 0514 0515 0515 0516 0-096 0516 0517 0518 0518 0519 0519 0520 0520 0521 0521 0-097 0522 0522 0523 0523 0523 0524 0525 0526 0526 0527 0-098 0527 0528 0528 0529 0529 0530 0530 0531 0532 0532 0-099 0533 0533 0534 0534 0535 0535 0536 0536 0537 0537 0-100 0538 0539 0539 0540 0540 0541 0541 0542 0542 0543 0-101 0543 0544 0544 0545 0546 0546 0547 0547 0548 0548 0-102 0549 0549 0550 0550 0551 0551 0552 0553 0553 0554 0-103 0554 0555 0555 0556 0556 0557 0557 0558 0559 0559 0-104 0560 0560 0561 0561 0562 0562 0563 0563 0564 0564 0-105 0565 0566 0566 0567 0567 0568 0568 0569 0569 0570 0-106 0570 0571 0571 0572 0573 0573 0574 0574 0575 0575 0-107 0576 0576 0577 0577 0578 0578 0579 0580 0580 0581 0-108 0581 0582 0582 0583 0583 0584 0584 0585 0585 0586 0-109 0587 0587 0588 0588 0589 0589 0590 0590 0591 0591 0-110 0592 0592 0593 0594 0594 0595 0595 0596 0596 0597 0-111 0597 0598 0598 0599 0599 0600 0601 0602 0602 0603 0-112 0603 0603 0604 0604 0605 0605 0606 0606 0607 0608 0-113 0608 0608 0609 0610 0610 0611 0611 0612 0612 0613 0-114 0613 0614 0614 0615 0616 0616 0617 0617 0618 0618 0-115 0619 0619 0620 0620 0621 0621 0622 0623 0623 0624 0-116 0624 0625 0625 0626 0626 0627 0627 0628 0629 0629 0-117 0629 0630 0631 0631 0632 0632 0633 0633 0634 0634 0-118 0635 0635 0636 0636 0637 0638 0638 0639 0639 0640 0-119 0640 0641 0641 0642 0642 0643 0643 0644 0645 0645 0-120 0646 0646 0647 0647 0648 0648 0649 0649 0650 0650 0-121 0651 0652 0652 0653 0653 0654 0654 0655 0655 0656 0-122 0656 0657 0657 0658 0659 0660 0660 0661 0661 0662 0-123 0662 0663 0663 0664 0664 0665 0665 0666 0666 0667 0-124 0667 0668 0668 0669 0669 0670 0670 0671 0671 0672 0-125 0673 0673 0674 0674 0675 0675 0676 0676 0677 0677 0-126 0678 0678 0679 0680 0680 0681 0681 0682 0682 0683 0-127 0683 0684 0684 0685 0685 0686 0687 0687 0688 0688 0-128 0689 0689 0690 0690 j 0691 0691 0692 0692 0693 0694 0-129 0694 0695 0695 0696 0696 0697 0697 0698 0698 0699 0-130 0699 0700 0700 0701 0701 0702 0702 0703 0703 0704 0-131 0704 0705 0705 0706 0706 0707 0707 0708 0709 0709 0-132 0710 0710 0711 0711 0712 0712 0713 0714 0714 0715 0-133 0715 0716 0716 0717 0717 0718 0718 0719 0719 0720 0-134 0721 0721 0722 0722 0723 0723 0724 0724 0725 0725 135 0726 0726 0727 0727 0728 0729 0729 0730 0730 0731 0-136 0731 0732 0732 0733 0733 0734 0734 0735 0736 0736 0-137 0737 0738 0738 0739 0739 0740 0740 0741 0741 0742 0-138 0742 0743 0744 0744 : 0745 0745 0746 0746 0747 0747 0-139 0748 0748 0749 0749 0750 0751 0751 0752 0752 0753 0-140 0753 0754 0754 0755 0755 0756 0756 0757 0758 0758 0-141 0759 0759 0760 0760 0761 0761 0762 0762 0763 0763 0-142 0764 0765 0765 0766 0766 0767 0767 0768 0768 0769 0-143 0769 0770 0770 0771 0771 0772 0773 0773 0774 0774 0-144 0775 0775 0776 0776 0777 0777 0778 0779 0779 0780 698 A TREATISE ON CHEMICAL ANALYSIS. Table LXXXVIL KC1O 4 to KC1. continued. KC10 4 . Grm. 1 2 3 4 5 6 7 8 9 0-145 0780 0781 0781 0782 0782' 0783 0783 0784 0784 0785 0-146 0785 0786 0787 0787 0788 0788 0789 0789 0790 0790 0-147 0791 0791 0792 0793 0793 0794 0794 0795 0795 0796 0-148 0796 0797 0797 0798 0798 0799 0800 0800 0801 0801 0-149 0802 0802 0803 0803 0804 0804 0805 0805 0806 0807 0-150 0807 0808 0808 0809 0809 0810 0810 0811 0811 0811 0-151 0812 0812 0813 0813 0814 0814 0815 0815 0816 0817 0-152 0817 0818 0818 0819 0819 0820 0820 0821 0821 0822 0-153 0823 0823 0824 0825 0825 0826 0826 0827 0827 0828 0-154 0829 0829 0830 0830 0831 0831 0832 0832 0833 0833 0-155 0834 0834 0835 0836 0836 0837 0837 0838 0838 0839 0-156 0839 0840 0840 0841 0841 0842 0843 0843 0844 0844 0-157 0845 0845 0846 0846 0847 0847 0848 0848 0849 0850 0-158 0850 0851 0851 0852 0852 0853 0853 0854 0854 0855 0-159 0855 0856 0856 0857 0858 0859 0859 0860 0860 0861 0-160 0861 0862 0862 0863 0863 0864 0864 0865 0866 0866 0-161 0867 0867 0868 0868 0869 0869 0870 0870 0871 0871 0-162 0872 0873 0873 0874 0874 0875 0875 0876 0876 0877 0-163 0877 0878 0878 0879 0879 0880 0880 0881 0881 0882 0-164 0882 0883 0883 0884 0884 0885 0886 0886 0887 0887 0-165 0888 0888 0889 0890 0890 0891 0891 0892 0892 0893 0-166 0893 0894 0894 0895 0895 0896 0896 0897 0897 0898 0-167 0898 0899 0900 0900 0901 0901 0902 0902 0903 0903 0-168 0904 0904 0905 0905 0906 0907 0907 0908 0908 0909 0-169 0909 0910 0910 0911 0911 0912 0912 0913 0914 0914 0-170 0915 0915 0916 0916 0917 0917 0918 0918 0919 0919 0-171 0920 0921 0921 0922 0922 0923 0923 0924 0924 0925 0-172 0925 0926 0926 0927 0928 0928 0929 0929 0930 0930 0-173 0931 0931 0932 0932 0933 0933 0934 0935 0935 0936 0-174 0936 0937 0937 0938 0938 0939 0939 0940 0940 0941 0-175 0942 0942 0943 0943 0944 0944 0945 0945 0946 0946 0-176 0947 0947 0948 0948 0949 0949 0950 0950 0951 0952 0-177 0952 0953 0953 0954 0954 0955 0955 0956 0957 0957 0-178 0958 0959 0959 0960 0960 0961 0961 0962 0962 0962 0-179 0963 0963 0964 0964 0965 0966 0966 0967 0967 0968 0-180 0968 0969 0969 0970 0971 0971 0972 0972 0973 0973 0-181 0974 0974 0975 0975 0976 0976 0977 0978 0978 0979 0-182 0979 0980 0980 0981 0981 0982 0982 0983 0983 0984 0-183 0985 0985 0986 0986 0987 0987 0988 0988 0989 0989 0-184 0990 0990 0991 0992 0992 0993 0993 0994 0994 0995 0-185 0995 0996 0996 0997 0997 0998 0999 0999 1000 ! 1000 0-186 1001 1001 1002 1002 1003 1003 1004 1004 1004 1005 0-187 1006 1006 1007 1007 1008 1008 1009 1009 1010 1010 0-188 1011 1011 1012 1013 1013 1014 1014 1015 1015 1016 0-189 1017 1017 1018 1018 1019 1019 1020 1021 1021 1022 0-190 1022 1023 1023 1024 1024 1025 1025 1026 1027 1027 0-191 1028 1028 1029 1029 1030 1030 1031 1031 1032 1032 0-192 1033 1033 1034 1035 1035 1036 1036 1037 1037 1038 0-193 1038 1039 1039 1040 1040 1041 1042 1042 1043 1043 0-194 1044 1044 1045 1045 1046 1046 1047 1047 1048 1049 0-195 1049 1050 1050 1051 1051 1052 1052 1053 1053 1054 0-196 1054 1055 1056 1056 1057 1057 1058 1058 1059 1059 0-197 1060 1060 1061 1061 1062 1062 1063 1064 1064 1065 0-198 1065 1066 1066 1067 1067 1068 1068 1069 1070 1070 0-199 1071 1071 1072 1072 1073 1073 1073 1074 1074 1075 APPENDIX. 699 Table LXXX VI 1 1. Conversion of Grms. of Potassium Chloride (KC1) into Grms. of Potassium Oxide (K 2 O). (FACTOR 0-632.) KCL Grm. Grm. K 2 O. 1 2 3 4 5 6 7 8 9 0-000 0001 0001 0002 0003 0003 0004 0005 0005 0006 0-001 0006 0007 0008 0008 0009 0009 0010 0011 0011 0012 0-002 0013 0013 0014 0015 0015 0016 0016 0017 0018 0018 0-003 0019 0020 0020 0021 0021 0022 0023 0023 0024 0025 0-004 0025 0026 0027 0027 0028 0028 0029 0030 0030 0031 0-005 0032 0032 0033 0033 0034 0035 0035 0036 0037 0037 0-006 0038 0039 0039 0040 0040 0041 0042 0042 0043 0044 0-007 0044 0045 0046 0046 0047 0047 0048 0049 0050 0050 0-008 0051 0051 0052 0052 0053 0054 0054 0055 0056 0056 0-009 0057 0058 0058 0059 0059 0060 0060 0061 0062 0063 0-010 0063 0064 0064 0065 0066 0066 0067 0068 0069 0070 0-011 0070 0070 0071 0071 0072 0073 0073 0074 0075 0075 0-012 0076 0077 0077 0078 0078 0079 0080 0080 0081 0082 0-013 0082 0083 0083 0084 0085 0085 0086 0087 0087 0088 0-014 0088 0089 0090 0091 0091 0092 0092 0093 0094 0094 0-015 0095 0095 0096 0097 0097 0098 0099 0099 0100 0100 0-016 0101 0102 0102 0103 0104 0104 0105 0106 0106 0107 0-017 0107 0108 0109 0109 0110 0111 0111 0112 0112 0113 0-018 0114 0114 0115 0116 0116 0117 0118 0118 0119 0119 0-019 0120 0121 0121 0122 0123 0123 0124 0125 0125 0126 0-020 0126 0127 0128 0128 0129 0130 0130 0131 0131 0132 0-021 0133 0133 0134 0135 0135 0136 0137 0137 0138 0138 0-022 0139 0140 0140 0141 0142 0142 0143 0143 0144 0145 0-023 0145 0146 0147 0147 0148 0149 0150 0150 0151 0152 0-024 0152 0153 0154 0154 0155 0155 0156 0157 0157 0158 0-025 0158 0159 0159 0160 0161 0161 0162 0162 0163 0164 0-026 0165 0166 0166 0167 0167 0168 0169 0169 0170 0171 0-027 0171 0171 0172 0173 0173 0174 0174 0175 0176 0176 0-028 0177 0178 0178 0179 0179 0180 0181 0181 0182 0183 0-029 0183 0184 0185 0185 0186 0186 0187 0188 0189 0189 0-030 0190 0190 0191 0191 0192 0193 0193 0194 0195 0195 0-031 0196 0197 0197 0198 0198 0199 0200 0200 0201 0202 0-032 0202 0203 0204 0204 0205 0205 0206 0207 0207 0208 0-033 0209 0209 0210 0210 0211 0212 0212 0213 0214 0214 0-034 0215 0216 0216 0217 0217 0218 0219 0219 0220 0221 0-035 0221 0222 0222 0223 0224 0224 0225 0226 0226 0227 0-036 0228 0228 0229 0229 0230 0231 0231 0232 0233 0233 0-037 0234 0234 0235 0236 0236 0237 0238 0238 0239 0240 0-038 0240 0241 0241 0242 0243 0243 0244 0245 0245 0246 0-039 0246 0247 0248 0248 0249 0250 0250 0251 0252 0252 700 A TREATISE ON CHEMICAL ANALYSIS. Table LXXXVIII. KC1 to K 2 O. continued. KCl. Grm. Grm. K 2 0. 1 2 3 4 5 678 9 0-040 0253 0253 0254 0255 0255 0256 0257 0257 0258 0258 0-041 0259 0260 0260 0261 0262 0262 0263 0264 0264 0265 0-042 0265 0266 0267 0267 0268 0269 0269 0270 0270 0271 0-043 0272 0272 0273 0274 0274 0275 0276 0276 0277 0277 0-044 0278 0279 0279 0280 0281 0281 0282 0282 0283 0284 0-045 0284 0285 0286 0286 0287 0287 0288 0289 0289 0290 0-046 0291 0291 0292 0293 0293 0294 0294 0295 0296 0296 0-047 0297 0298 0298 0299 0299 0300 0301 0301 0302 0303 0-048 0303 0304 0305 0305 0306 0306 0307 0308 0308 0309 0-049 0310 0310 0311 0311 0312 0313 0313 i 0314 0315 0315 0-050 0316 0317 0317 0318 0318 0319 0320 0320 0321 0322 0-051 0322 0323 0323 0324 0325 0325 0326 0327 0327 0328 0-052 0329 0329 0330 0330 0331 0332 0332 0333 0334 0334 0-053 0335 0335 0336 0337 0337 0338 0339 0339 0340 0341 0-054 0341 0342 0342 0343 0344 0344 0345 0346 0346 0347 0-055 0348 0348 0349 0349 0350 0351 0351 0352 0353 0353 0-056 0354 0355 0355 0356 0356 0357 0358 0358 0359 0360 0-057 0360 0361 0361 0362 0363 0363 0364 0365 0365 0366 0-058 0366 0367 0368 0368 0369 0370 0370 0371 0372 0372 0-059 0373 0373 | 0374 0375 0375 0376 0377 0377 0378 0378 0-060 0379 0380 0380 0381 0382 0382 0383 0384 0384 0385 0-061 0385 0386 0387 0387 0388 0389 0389 0390 0390 0391 0-062 0392 0392 0393 0394 0394 0395 0396 0396 0397 0397 0-063 0398 0399 0399 0400 0401 0401 0402 0403 0403 0404 0-064 0404 0405 0406 0406 0407 0408 0408 0409 0409 0410 0-065 0411 0411 0412 0413 0413 0414 0414 0415 0416 0416 0-066 0417 0418 0418 0419 0420 0420 0421 0421 0422 0423 0-067 0423 0424 0425 0425 0426 0427 0247 0428 0428 0429 0-068 0430 0430 0431 0432 0432 0433 0433 0434 0435 0435 0-069 0436 0437 0437 0438 0439 0439 0440 0440 0441 0442 0-070 0442 0443 0444 0444 0445 0445 0446 0447 0447 0448 0-071 0449 0449 0450 0451 0451 0452 0452 0453 0454 0454 0-072 0455 0456 0456 0457 0457 0458 0459 0459 0460 0461 0-073 0461 0462 0463 0463 0464 0464 0465 0466 j 0466 0467 0-074 0468 0468 0469 0469 0470 0471 0471 0472 0473 0473 0-075 0474 0475 0475 0476 0476 0477 0478 0478 0479 0480 0-076 0480 0481 0481 0482 0483 0483 0484 0485 0485 0486 0-077 0487 0487 0488 0488 0489 0490 0490 0491 0492 0492 0-078 0493 0493 0494 0495 0495 0496 0497 0497 0498 0499 0-079 0499 0500 0500 0501 0502 0502 0503 0504 0504 0505 0-080 0506 0506 0507 0507 0508 0509 0509 0510 0511 0511 0-081 0512 0512 0513 0514 0514 0515 0516 0516 0517 0518 0-082 0518 0519 0519 0520 0521 0521 0522 0523 0523 0524 0-083 0524 0525 0526 0526 0527 0528 0528 0529 0530 0530 0-084 0531 0531 0532 0533 0533 0534 0535 0535 0536 0536 APPENDIX. Table LXXXVIIL KC1 to Q. -continued. 701 KCl. Grm. Grm. K 2 0. 1 2 3 4 5 6 . 7 8 9 0-085 0537 0538 0538 0539 0540 0540 0541 0542 0542 0543 0-086 0543 0544 0545 0545 0546 0547 0547 0548 0548 0549 0-087 0550 0550 0551 0552 0552 0553 0554 0554 0555 0555 0-088 0556 0557 | 0557 0558 0559 0559 0560 0560 0561 0562 0-089 0562 0563 0564 0564 0565 0566 0566 ! 0567 0567 0568 0-090 0569 0569 0570 0571 0571 0572 0572 0573 0574 0574 0-091 0575 0576 0576 0577 0578 0578 0579 0579 0580 0581 0-092 0581 0582 0583 0583 0584 0585 0585 0586 0586 0587 0-093 0588 0588 0589 0590 0590 0591 0591 0592 0593 0593 0-094 0594 0595 0595 0596 0597 0597 0598 0598 0599 0600 0-095 0600 0601 0602 0602 0603 0603 0604 0605 0605 0606 0-096 0607 0607 0608 0609 0610 0610 0610 0611 0612 0612 0-097 0613 0614 0615 0615 0616 0617 0617 0618 0619 0619 0-098 0619 0620 0621 0621 0622 0622 0623 0624 0624 0625 0-099 0626 0626 0627 0627 0628 0629 0629 0630 0631 0631 0-100 0632 0633 0633 0634 0634 0635 0635 0636 0637 0637 0-101 0638 0638 0639 0639 0640 0641 0642 0643 0643 0644 0-102 0645 0645 0646 0647 0647 0648 ' 0648 0649 0650 0650 0-103 0651 0652 0652 0653 0653 0654 0654 0655 0656 0657 0-104 0657 0658 0659 0659 0660 0660 0661 0662 0662 0663 0-105 0664 0664 0665 0666 0666 0667 0668 0669 0669 0670 0-106 0670 0671 0671 0672 0672 0673 0673 0674 0675 0676 0-107 0676 0677 0678 0678 0679 0679 0680 0681 0681 0682 0-108 0683 0683 0684 0684 0685 0686 0686 0687 0688 0688 0-109 0689 0690 0690 0691 0691 0692 0693 0693 0694 0695 0-110 0695 0696 0696 0697 0698 0698 0699 0700 0700 0701 0-111 0702 0702 0703 0703 0704 0705 0705 0706 0707 0707 0-112 0708 0708 0709 0710 0710 0711 0712 0712 0713 0714 0-113 0714 0715 0715 0716 0717 0717 0718 0718 0719 0719 0-114 0720 0721 0722 0722 0723 0724 0724 0725 0726 0726 0-115 0727 0727 0728 0729 0729 0730 ! 0731 0731 0732 0732 0-116 0733 0734 0734 0735 0736 0736 0737 0738 0738 0739 0-117 0739 0740 0741 0741 0742 0743 0743 0744 0744 ! 0745 0-118 0746 0746 0747 0748 0748 0749 0750 0750 0751 0751 0-119 0752 0753 0753 0754 0755 0755 0756 0757 0758 0758 0-120 0758 0759 0760 0760 0761 0762 0762 0763 0763 0764 0-121 0765 0765 0766 0767 0767 0768 0769 0769 0770 0770 0-122 0771 0772 0772 0773 0774 0774 0775 0775 0776 0777 0-123 0-124 0778 0784 0779 0785 0779 0786 0780 0786 0781 0787 0781 ': 0782 0788 0788 0782 0789 0783 0790 0784 0790 0-125 0-126 0791 0796 0791 0797 0792 0797 0792 0798 0793 0799 0793 0794 0799 0800 0794 0800 0795 0801 0795 0802 0-127 0-128 0-129 0802 0809 0815 0803 0809 0816 0803 0810 0816 0804 0811 0817 0805 0811 0817 0805 0806 0812 0812 0818 0819 0807 0813 0819 0807 ! 0808 0814 0814 0820 0821 702 A TREATISE ON CHEMICAL ANALYSIS. Table LXXXVIIL KC1 to K 9 O.- continued. KCl. Grm. Grm. K 2 0. 1 2 3 4 5 6 7 8 9 0-130 0821 0822 0822 0823 0824 0824 0825 0826 0826 0827 0-131 0827 0828 0829 0829 0830 0831 0831 0832 0833 0833 0-132 0834 0835 0835 0836 0836 0837 0838 0838 0839 0840 0-133 0840 0841 0842 0843 0843 0844 0844 0845 0846 0846 0-134 0847 0847 0848 0848 0849 0850 0850 0851 0852 0852 0-135 0853 0854 0854 0855 0856 0856 0857 0858 0858 0859 0-136 0860 0861 0862 0862 0863 0864 0864 0865 0865 0866 0-137 0866 0867 0867 0868 0868 0869 0869 0870 0871 0871 0-138 0872 0873 0873 0874 0874 0875 0876 0876 0877 0878 0-139 0878 0879 0880 0880 0881 0881 0882 0883 0883 0884 0-140 0885 0885 0886 0887 0887 0888 0889 0889 0890 0890 0'141 0891 0892 0892 0893 0894 0894 0895 0896 0896 0897 0-142 0897 0898 0899 0899 0900 0901 0901 0902 0902 0903 0-143 0904 0904 0905 0906 0906 0907 0908 0908 0909 0909 0-144 0910 0910 0911 0912 0912 0913 0914 0914 0915 0915 0-145 0916 0917 0918 0918 0919 0920 0920 0921 0921 0922 0-146 0923 0923 0924 0925 0925 0926 0927 0927 0928 0928 0-147 0929 0930 0930 0931 0932 0932 0933 0933 0934 0935 0-148 0935 0936 0937 0937 0938 0939 0939 0940 0940 0941 0-149 0942 0942 0943 0944 0944 0945 0945 0946 0947 0947 0-150 0948 0949 0949 0950 0951 0951 0952 0952 0953 0954 0-151 0954 0955 0956 0956 0957 0957 0958 0959 0959 0960 0-152 0961 0961 0962 0963 0963 ' 0964 0964 0965 0966 0966 0-153 0967 0968 0968 0969 0969 0970 0971 0971 0972 0973 0-154 0973 0974 0975 0975 0976 0976 0977 0978 0978 0979 0-155 0980 0980 0981 0981 0982 0983 0983 0984 0985 0985 0-156 0986 0987 0987 0988 0988 0989 0990 0991 0991 0992 0-157 0992 0993 0994 0994 0995 0995 0996 0997 0997 0998 0-158 0999 0999 1000 1000 1001 1002 1002 1003 1004 1004 0-159 1005 1006 1006 1007 1007 1008 1009 1009 1010 1011 0-160 1011 1012 1012 1013 1014 1014 1015 1016 1016 1017 APPENDIX. 703 Table LXXXIX. Conversion of Grms. of Sodium Chloride (NaCl) into Grms. of Sodium Oxide (Na 2 O). (FACTOR 0-530.) Grin. Grm. Na 2 0. Nad. 1 2 3 4 5 6 7 8 9 0-000 0000 0001 0001 0002 0003 0003 0004 0004 0004 0005 0-001 0005 0006 0006 0007 0007 0008 0008 0009 0010 0010 0-002 0011 0011 0012 0012 0013 0013 0014 0014 0015 0015 0-003 0016 0016 0017 0017 0018 0019 0019 0020 0021 0021 0-004 0022 0022 0023 0023 0024 0024 0025 0025 0026 0026 0-005 0027 0027 0028 0028 0029 0029 0030 0030 0031 0031 0-006 0032 0032 0033 0033 0034 0034 0035 0036 0036 0037 0-007 0037 0038 0038 0039 0039 0040 0040 0041 0041 0042 0-008 0042 0043 0043 0044 0045 0045 0046 0046 0047 0047 0-009 0048 0048 0049 0049 0050 0050 0051 0051 0052 0052 0-010 0053 0054 0054 0055 0055 0056 0056 0057 0057 0058 0-011 0058 0059 0059 0060 0060 0061 i 0061 0062 0062 0063 0-012 0064 0064 0065 0065 0066 0066 0067 0067 0068 0068 0-013 0069 0069 0070 0070 0071 0072 0072 0073 0073 0074 0-014 0074 0075 0075 0076 0076 0077 0077 0078 0078 0079 0-015 0080 0080 0081 0081 0082 0082 0083 0083 0084 0084 0-016 0085 0085 0086 0086 0087 0087 0088 0089 0089 0090 0-017 0090 0091 0091 0092 0092 0093 0093 0094 0094 0095 0-018 0095 0096 0096 0097 0098 0098 0099 0099 0100 0100 0-019 0101 0101 0102 0103 0103 0104 0104 0104 0105 0105 0-020 0106 0107 0107 0108 0108 0109 0109 0110 0110 0111 0-021 0111 0112 0112 0113 0113 0114 0114 0115 0116 0116 0-022 0117 0117 0118 0119 0119 0119 0120 0120 0121 0121 0-023 0122 0122 0123 0123 0124 0125 0125 0126 0126 0127 0-024 0127 i 0128 0128 0129 0129 0130 0130 0131 0131 0132 0-025 0133 0133 0134 0134 0135 0135 0136 0136 0137 0137 0-026 0138 0138 0139 0139 0140 ,0140 0141 0142 0142 0143 0-027 0143 0144 0144 0145 0145 0146 0146 0147 0147 0148 0-028 0148 0149 0149 0150 0151 0151 0152 0152 0153 0153 0-029 0154 0154 0155 0155 0156 0156 0157 0157 0158 0158 0-030 0159 0160 0160 0161 0161 0162 0162 0163 0163 0164 0-031 0164 0165 0165 0166 0166 0167 0167 0168 0169 0169 0-032 0170 0170 0171 0171 ! 0172 0172 0173 0173 0174 0174 0-033 0175 0176 0176 0176 0177 0178 0178 0179 0179 0180 0-034 0180 0181 0181 0182 0182 0183 0183 0184 0184 0185 0-035 0186 0186 0187 0187 0188 0188 0189 0189 0190 0190 0-036 0191 0191 0192 0192 0193 0193 0194 0195 0195 0196 0-037 0196 0197 0197 0198 0198 0199 0199 0200 0200 0201 0-038 0201 i 0202 0202 0203 0204 0204 0205 0205 0206 0206 0-039 0207 0207 0208 0208 0209 0209 0210 0210 0211 0211 704 A TREATISE ON CHEMICAL ANALYSIS. Table LXXXIX. NaCl to Na 2 O. continued. Grm. NaCl. Grm. Na 2 O. 1 2 3 4 5 6 7 8 9 0-040 0212 0213 0213 0214 0214 0215 0215 0216 0216 0217 0-041 0217 0218 0218 0219 0219 0220 0220 0221 0222 0222 0-042 0223 0223 0224 0224 0225 0225 0226 0226 0227 0227 0-043 0228 0228 0229 0229 0230 0231 0231 0232 0232 0233 0-044 0233 0234 0234 0235 0235 0236 0236 0237 0237 0238 0-045 0239 0239 0240 0240 0241 0241 0242 0242 0243 0243 0-046 0244 0244 0245 0245 0246 0246 0247 0248 0248 0249 0-047 0249 0250 0250 0251 0251 0252 0252 0253 0253 0254 0-048 0254 0255 0255 0256 0257 0257 0258 0258 0259 0259 0-049 0260 0260 0261 0261 0262 0262 0263 0263 0264 0264 0-050 0265 0265 0266 0267 0267 0268 0268 0269 0269 0270 0-051 0270 0271 0271 0272 0272 0273 0273 0274 0275 0275 0-052 0276 0276 0277 0277 0278 0278 i 0279 0279 0280 0280 0-053 0281 0281 0282 0282 0283 0284 0284 0285 0285 0286 0-054 0286 0287 0288 0288 0289 0289 0290 0290 0291 0292 0-055 0292 0293 0293 0294 0294 0295 0295 0296 0296 0297 0-056 0297 0297 0298 0298 0299 0299 t 0300 0301 0301 0302 0-057 0302 0303 0303 0304 0304 0305 0305 0306 0306 ; 0307 0-058 0307 0309 0309 0310 0310 0310 0311 0311 0312 0312 0-059 0313 0313 0314 0314 0315 0315 ! 0316 0316 0317 0317 0-060 0318 0319 0319 0320 0320 0321 0321 0322 0322 0323 0-061 0323 0324 0324 0325 0325 0326 0326 0327 0328 0328 0-062 0329 0329 0330 0330 0331 0331 0332 0332 0333 0333 0-063 0334 0334 0335 0335 0336 0337 \ 0337 0338 0338 0339 0-064 0339 j 0340 0340 0341 0341 0342 i 0342 0343 0343 0344 0-065 0345 0345 0346 0346 0347 0347 0348 0348 0349 0349 0-066 0350 0350 0351 0351 0352 0352 0353 0354 0354 j 0355 0-067 0355 0356 0356 0357 0357 0358 0358 0359 0359 0360 0-068 0360 0361 0361 0362 0363 0363 0364 0364 0365 0365 0-069 0366 0366 0367 0367 0368 0368 0369 0369 0370 0370 0-070 0371 0372 0372 0373 0373 0374 0374 0375 0375 0376 0-071 0376 0377 0377 0378 0378 0379 0379 0380 0381 0381 0-072 0382 ! 0382 0383 0383 0384 0384 0385 0385 0386 0386 0-073 0387 0387 0388 0388 0389 0390 0390 0391 0391 0392 0-074 0392 0393 0393 0394 0394 0395 0395 0396 0396 0397 0-075 0398 0398 0399 0399 0400 0400 ; 0401 0401 0402 0402 0-076 0403 0403 0404 0404 0405 0405 0406 i 0407 0407 0408 0-077 0408 0409 0409 0410 0410 0411 0411 0412 0412 0413 0-078 0413 0414 0414 0415 0416 0416 0417 0417 0418 0418 0-079 0419 0419 0420 0420 0421 0421 0422 0422 0423 0423 0-080 0424 0425 0425 0426 0426 0427 0427 0428 0428 0429 0-081 0429 0430 0430 0431 0431 0432 0432 0433 0434 0434 0-082 0435 0435 0436 0436 0437 0437 0438 0438 0439 0439 0-083 0440 0440 0441 0441 0442 0443 0443 0444 0444 0445 0-084 0445 0446 0446 0447 0447 0448 0448 0449 0449 0450 j i APPENDIX. Table LXXXIX.-NaCl to Na 2 O. -continued. 705 Grm. NaCl. Grm. Na 2 0. 1 2 3 4 5 6 7 8 9 0-085 0451 0451 0452 0452 0453 0453 0454 0454 0455 0455 0-086 0456 0456 0457 0457 0458 0458 0459 0460 0460 0461 0-087 0461 0462 0462 0463 0463 0464 0464 0465 0465 0466 0-088 0466 0467 0467 0468 0469 0469 0470 0470 0471 0471 0-089 0472 0472 0473 0473 0474 0474 0475 0475 0476 0476 0-090 0477 0478 0478 0479 0479 0480 0480 0481 0481 0482 0-091 0482 0483 0483 0484 0484 0485 0485 0486 0487 0487 0-092 0488 0488 0489 0489 0490 0490 0491 0491 0492 0492 0-093 0493 0493 0494 0494 0495 0496 0496 0497 0497 0498 0-094 0498 0499 0499 0500 0500 0501 0501 0502 0502 0503 0-095 0504 0504 0505 0505 0506 0506 0507 0507 0508 0508 0-096 0509 0509 0510 0510 0511 0511 0512 0513 0513 0514 0-097 0514 0515 0515 0516 0516 0517 0517 0518 0518 0519 0-098 0519 0520 0520 0521 0522 0522 0523 0523 0524 0524 0-099 0525 0525 0526 0526 0527 0527 0528 0528 0529 0529 0-100 0530 0531 0531 0532 0532 0533 0533 0534 0534 0535 0-101 0535 0536 0536 0537 0537 0538 0538 0539 0540 0540 0-102 0541 0541 0542 0542 0543 0543 0544 0544 0545 0545 0-103 0546 0546 0547 0548 0548 0549 0549 0550 0550 0551 0-104 0551 0552 0552 0553 0553 0554 0554 0555 0555 0556 0-105 0557 0557 0558 0558 0559 0559 0560 0560 0661 0561 0-106 0562 0562 0563 0563 0564 0564 0565 0566 0566 0567 0-107 0567 0568 0568 0569 0569 0570 0570 0571 0571 0572 0-108 0572 0573 0573 0574 0575 0575 0576 0576 0577 0577 0-109 0578 0578 0579 0579 0580 0580 0581 0581 0582 0582 0-110 0583 0583 0584 0585 0585 0586 0586 0587 0587 0588 0-111 0588 0589 0589 0590 0591 0591 0592 0592 0593 0594 0-112 0594 0595 0595 0596 0596 0597 0597 0598 0598 0599 0-113 0599 0600 0600 0601 0601 0602 0602 0603 0603 0604 0-114 0604 0605 0605 0606 0606 0607 0607 0608 0608 0609 0-115 0610 0611 0611 0612 0612 0613 0613 0614 0614 0615 0-116 0615 0616 0616 0617 0617 0618 0618 0619 0619 0620 0-117 0620 0621 0621 0622 0622 0623 0623 0624 0624 0625 0-118 0625 0626 0626 0627 0628 0628 0629 0629 0630 0630 0-119 0631 0631 j 0632 0632 0633 0633 0634 0634 0635 0635 0-120 0636 0637 0637 0638 0638 0639 0639 0640 0640 0641 0-121 0641 0642 0642 0643 0643 0644 0645 0645 0646 0646 0-122 0647 0647 i 0648 0648 0649 0649 0650 0650 0651 0651 0-123 0652 0652 0653 0653 0654 0654 0655 0656 0656 0657 0-124 0657 0658 0658 0659 0659 0660 0660 0661 0661 0662 0-125 0663 0663 0664 0664 0665 0665 0666 0666 0667 0667 0-126 0668 0668 0669 0669 0670 0670 0671 0672 0672 0673 0-127 0673 0674 0674 0675 0675 0676 0676 0677 0677 ! 0678 0-128 0678 0679 0680 0681 0681 0682 0682 0683 0683 0684 0-129 0684 0685 0685 0686 0686 0687 0687 0688 0688 0689 45 ;o6 A TREATISE ON CHEMICAL ANALYSIS. Table LXXXIX. NaCl to Na 2 O. continued. Grm. Grm. Na 2 O. NaCl. 1 2 3 4 5 6 7 8 9 0-130 0689 0690 0690 0691 0691 0692 0692 0693 0693 0694 0-131 0694 0695 0695 0696 0696 0697 0697 0698 0698 0699 0-132 0700 0700 0701 0701 0702 0702 0703 0703 0704 0704 0-133 0705 0705 0706 0706 0707 0707 0708 0709 0709 0710 0-134 0710 0711 0711 0712 0712 0713 0713 0714 0714 0715 0-135 0716 0716 0717 0717 0718 0718 0719 0719 0720 0720 0-136 0721 0721 0722 0722 0723 0723 0724 0724 0725 0726 0-137 0726 0727 0727 0727 0728 0728 0729 0729 0730 0731 0-138 0731 0732 0732 0733 0733 0734 0734 0735 0735 0736 0-139 0736 0737 0737 0738 0739 0739 0740 0740 0741 0741 0-140 0742 0742 0743 0743 0744 0744 0745 0745 0746 0746 0-141 0747 0747 0748 0748 0749 0749 0750 0751 0751 0752 0-142 0752 0753 0753 0754 0754 0755 0755 0756 0756 0757 0-143 0757 0758 0758 0759 0759 0760 0761 0761 0762 0762 0-144 0763 0763 0764 0764 0765 0765 0766 0766 0767 0768 0-145 0768 0769 0770 0770 0771 0771 0772 0772 0773 0773 0-146 0774 0774 0775 0775 0776 0776 0777 0778 0778 0779 0-147 0779 0780 0780 0781 0781 0782 ! 0782 0783 0783 0784 0-148 0784 0785 0785 0786 0786 0787 | 0788 0788 0789 0789 0-149 0790 0790 0791 0791 0792 0792 0793 0793 0794 0794 0-150 0795 0795 0796 0797 0797 0798 0798 0799 0799 0800 0-151 0800 0801 0801 0802 0802 0803 0803 0804 0805 0805 0-152 0806 0806 0807 0807 0808 0808 0809 0809 0810 0810 0-153 0811 0811 0812 0813 0813 0814 | 0814 0815 0815 0816 0-154 0816 0817 0817 0818 0818 0819 0819 0820 0820 0821 0-155 0821 0822 0823 0823 0824 0824 0285 0825 0826 0826 0-156 0827 0827 0828 0828 0829 0829 0830 0831 0831 0832 0-157 0832 0833 0833 0834 0834 0835 0835 0836 0836 0837 0-158 0837 0838 0838 0839 0840 0840 0841 0841 0842 0842 0-159 0843 0843 0844 0844 0845 0845 0846 0846 0847 0848 0-160 0848 0849 0849 0850 0850 0851 0851 0852 0852 0853 0-161 0853 0854 0854 0855 0856 0856 0857 0857 0858 0859 0-162 0859 0859 0860 0860 0861 0861 0862 0862 0863 0863 0-163 0864 0864 0865 0866 0866 0867 0867 0868 0868 0869 0-164 0869 0870 0870 0871 0871 0872 0872 0873 0874 0874 0-165 0875 0875 0876 0876 0877 0877 0878 0878 0879 0879 0-166 0880 0880 0881 0881 0882 0882 0883 0884 0884 0885 0-167 0885 0886 0886 0887 0887 0888 0888 0889 0889 0890 0-168 0890 0891 0891 0892 0892 0893 0894 0894 0895 0895 0-169 0896 0896 0897 0897 0898 0898 0899 0899 0900 0900 0-170 0901 0902 0902 0903 0903 0904 0904 0905 0905 0906 0-171 0906 0907 0907 0908 0908 0909 0909 0910 0911 0911 0-172 0912 0912 0913 0913 0914 0914 0915 0915 0916 0916 0-173 0917 0917 0918 0918 0919 0919 0920 0921 0921 0922 0-174 0922 0923 0923 0924 0924 0925 0925 0926 0926 0927 APPENDIX. Table LXXXIX. NaCl to Na 2 O. continued. 707 Gnn. Grm. Na 2 O. NaCl. 1 2 3 4 5 6 7 8 9 0-175 0928 0928 0929 0929 0930 0930 0931 0931 0932 0932 0-176 0933 0933 0934 0934 0935 0935 0936 0937 0937 0938 0-177 0938 0939 0939 0940 0940 . 0941 0941 0942 0942 0943 0-178 0943 0944 0944 0945 0945 0946 0947 0948 0948 0949 0-179 0949 0950 0950 0951 0951 0952 0952 0953 0953 | 0954 0-180 0954 0955 0955 0956 0956 0957 0957 0958 0958 0959 0-181 0959 0960 0960 0961 0961 0962 0962 0963 0964 0964 0-182 0965 0965 0966 0966 ; 0967 0967 0968 0968 0969 0969 0-183 0970 0970 0971 0971 0972 0973 0973 0974 0974 0975 0-184 0975 0976 0976 0977 0977 0978 0978 0979 0979 0980 0-185 0981 0981 0982 0982 0983 0983 0984 0984 0985 0985 0-186 0986 0986 0987 0987 0988 0988 0989 0990 0990 0991 0-187 0991 0992 0992 0993 0993 0994 0994 0995 0995 0996 0-188 0996 0997 0997 0998 0999 0999 1000 1000 1001 1001 0-189 1002 1002 1003 1003 1003 1004 1004 1005 1005 1006 0-190 1007 1007 1008 1008 1009 1009 1010 1010 1011 1012 0-191 1012 1013 1013 1014 1014 1015 1015 1016 1017 1017 0-192 1018 1018 1019 1019 1020 1020 1021 1021 1022 1022 0-193 1023 1023 1024 1025 1025 1026 1026. 1027 1027 1028 0-194 1028 1029 1029 1030 1030 1031 1031 1032 1032 1033 0-195 1034 1034 1035 1035 1036 1036 1037 1037 1038 1038 0-196 1039 1039 1040 1040 1041 1041 1042 1042 1043 1044 0-197 1044 1045 1045 1046 1046 1047 1047 1048 1048 1049 0-198 1049 1050 1051 1051 1052 1052 1053 1053 1054 1054 0-199 1055 1055 1056 1056 1056 1057 1057 1058 1058 1059 A TREATISE ON CHEMICAL ANALYSIS. Table XC. Conversion of Grms. of Magnesium Pyrophosphate (Mg 2 P 2 O 7 ) into Grms. of Magnesium Oxide (MgO). (FACTOR 0'362.) Grm. Grm. MgO. Mg 2 P 2 7 . 1 2 3 4 5 6 7 8 9 0-000 0001 0001 0001 0002 0002 0003 0003 0003 0-001 0004 0004 0004 0005 0005 0005 0006 0006 0007 0007 0-002 0007 0008 0008 0008 0009 0009 0009 0010 0010 0010 0-003 0011 0011 0012 0012 0012 0013 0013 0013 0014 0014 0-004 0014 0015 0015 0015 0016 0016 0017 0017 0017 0018 0-005 0018 0018 0019 0019 0020 0020 0020 0021 0021 0021 0-006 0022 0022 0022 0023 0023 0024 0024 0024 0025 0025 0-007 0025 0026 0026 0026 0027 0027 0028 0028 0028 0029 0-008 0029 0029 0029 0030 0030 0031 0031 0031 0032 0032 0-009 0033 0033 0033 0034 0034 0034 0035 0035 0035 0036 0-010 0036 0037 0037 0037 0038 0038 0038 0039 0039 0039 0-011 0040 0040 0041 0041 0041 0042 0042 0042 0043 0043 0-012 0043 0044 0044 0045 0045 0045 0046 0046 0046 0047 0-013 0047 0047 0048 0048 0049 0049 0049 0050 0050 0050 0-014 0051 0051 0051 0052 0052 0052 0053 0053 0054 0054 0-015 0054 0055 0055 0055 0056 0056 0056 0057 0057 0058 0-016 0058 0058 0059 0059 0059 0060 0060 0060 0061 0061 0-017 0062 0062 0062 0063 0063 0063 0064 0064 0064 0065 0-018 0065 0066 0066 0066 0067 0067 0067 0068 0068 0068 0-019 0069 0069 0070 0070 0070 0071 0071 0071 0072 0072 0-020 0072 0073 0073 0073 0074 0074 0075 0075 0075 0076 0-021 0076 0076 0077 0077 0078 0078 0078 0079 0079 0079 0-022 0080 0080 0080 0081 0081 0081 0082 0082 0083 0083 0-023 0083 0084 0084 0084 0085 0085 0085 0086 0086 0087 0-024 0087 0087 0088 0088 0088 0089 0089 0089 0090 0090 0-025 0091 0091 0091 0092 0092 0092 0093 0093 0093 0094 0-026 0094 0094 0095 0095 0096 0096 0096 0097 0097 0097 0-027 0098 0098 0098 0099 0099 0100 0100 0100 0101 0101 0-028 0101 0102 0102 0102 0103 0103 0104 0104 0104 0105 0-029 0105 0105 0106 0106 0107 0107 0107 0108 0108 0108 0-030 0109 0109 0109 0110 0110 0110 0111 0111 0111 0112 0-031 0112 0113 0113 0113 0114 0114 0114 0115 0115 0115 0-032 0116 0116 0117 0117 0117 0118 0118 0118 0119 0119 0-033 0119 0120 0120 0121 0121 0121 0122 0122 0122 0123 0-034 0123 0123 0123 0124 0125 0125 0125 0126 0126 0126 0035 0127 0127 0127 0128 0128 0129 0129 0129 0130 0130 0-036 0130 0131 0131 0131 0132 0132 0132 0133 0133 0134 0-037 0134 0134 0135 0135 0135 0136 0136 0136 0137 0137 0-038 0138 0138 0138 0139 0139 0139 0140 0140 0140 0141 0-039 0141 0142 0142 0142 0143 0143 0143 0144 0144 0144 APPENDIX. 709 Table XC. Mg 2 P 2 O 7 to MgO. continued. Grm. Mg 2 P 2 7 . Grm. MgO. 1 2 3 4 5 6 7 8 9 0-040 0145 0145 0146 0146 0146 0147 0147 0147 0148 0148 0-041 0148 0149 0149 0149 0150 0150 0151 0151 0151 0152 0-042 0152 0152 0153 0153 0153 0154 0154 0155 0155 0155 0-043 0156 0156 0156 0157 0157 0157 0158 0158 0159 0159 0-044 0159 0160 0160 0160 0161 0161 0161 0162 0162 0163 0-045 0163 0163 0164 0164 0164 0165 0165 0165 0166 0166 0-046 0167 0167 0167 0168 0168 0168 0169 0169 0169 0170 0-047 0170 0171 0171 0171 0172 0172 0172 0173 0173 0173 0-048 0174 0174 0174 0175 0175 0176 0176 0176 0177 0177 0-049 0177 0178 0178 0178 0179 0179 0180 0180 0180 0181 0-050 0181 0181 0182 0182 0182 0183 0183 0184 0184 0184 0-051 0185 0185 0185 0186 0186 0186 0187 0187 0188 0188 0-052 0188 0189 0189 0189 0190 0190 0190 0191 0191 0191 0-053 0192 0192 0193 0193 0193 0194 0194 0194 0195 0195 0-054 0195 0196 0196 0197 0197 0197 0198 0198 0198 0199 0-055 0199 0199 0200 0200 0201 0201 0201 0202 0202 0202 0-056 0203 0203 0203 0204 0204 0205 0205 0205 0206 0206 0-057 0206 0207 0207 0207 0208 0208 0209 0209 0209 0210 0-058 0210 0210 0211 0211 0211 0212 0212 0212 0213 0213 0-059 0214 0214 0214 0215 0215 0215 0216 0216 0216 0217 0-060 0217 0218 0218 0218 0219 0219 0219 0220 0220 0220 0-061 0221 0221 0222 0222 0222 0223 0223 0223 0224 0224 0-062 0224 0225 0225 0226 0226 0226 0227 0227 0227 0228 0-063 0228 0228 0229 0229 0230 0230 0230 0231 0231 0231 0-064 0232 0232 0232 0233 0233 0234 0234 0234 0235 0235 0-065 0235 0236 0236 0236 0237 0237 0237 0238 0238 0238 0-066 0239 0239 0240 0240 0240 0241 0241 0242 0242 0242 0-067 0243 0243 0243 0244 0244 0244 0245 0245 0245 0246 0-068 0246 0247 0247 : 0247 0248 0248 0248 0249 0249 0249 0-069 0250 0250 0251 i 0251 0251 0252 0252 0252 0253 0253 0-070 0253 0254 0254 0254 0255 0255 0256 0256 0256 0257 0-071 0257 0257 0258 0258 0258 0259 0259 0260 0260 0260 0-072 0261 0261 0261 0262 0262 0262 0263 0263 0264 0264 0-073 0264 0265 0265 0265 0266 0266 0266 0267 0267 0268 0-074 0268 0268 0269 0269 0269 0270 0270 0270 0271 0271 0-075 0272 0272 0272 0273 0273 0273 0274 0274 0274 0275 0-076 0275 0275 0276 0276 0277 0277 0277 0278 0278 0278 0-077 0279 0279 0279 0280 0280 0281 0281 0281 0282 0282 0-078 0282 0283 0283 0283 0284 0284 0285 0285 0285 0286 0-079 0286 0286 0287 0287 0287 0288 0288 0289 0289 0289 0-080 0290 0290 0290 0291 0291 0291 0292 0292 0292 0293 0-081 0293 0294 0294 0294 0295 0295 0295 0296 0296 0296 0-082 0297 0297 0298 0298 0298 0299 0299 0299 0300 0300 0-083 0300 0301 0301 0302 0302 0302 0303 0303 0303 0304 0-084 0304 0304 0305 0305 0306 0306 0306 ! 0307 0307 0307 710 A TREATISE ON CHEMICAL ANALYSIS. Table XC. Mg 2 P 2 O 7 to MgO. continued. Gr-. Grm. MgO. Mg 2 P 2 7 . 1 2 3 4 5 6 7 8 9 0-085 0308 0308 0308 0309 0309 0310 0310 0310 0311 0311 0-086 0311 0312 0312 0312 0313 0313 0313 0314 0314 0315 0-087 0315 0315 0316 0316 0316 0317 0317 0317 0318 0318 0-088 0319 0319 0319 0320 0320 0320 0321 0321 0321 0322 0-089 0322 0323 0323 0323 0324 0324 0324 0325 0325 0325 0-090 0326 0326 0326 0327 0327 0328 0328 0328 0329 0329 0-091 0329 0330 0330 0331 0331 0331 0332 0332 0332 0333 0-092 0333 0333 0334 0334 0334 0335 0335 0336 0336 0336 0-093 0337 0337 0337 0338 0338 0338 0339 0339 0340 0340 0-094 0340 0341 0341 0341 0342 0342 0342 0343 0343 0344 0-095 0344 0344 0345 0345 0345 0346 0346 0346 0347 0347 0-096 0348 0348 0348 0349 0349 0349 0350 0350 0350 0351 0-097 0351 0352 0352 0352 0353 0353 0353 0354 0354 0354 0-098 0355 0355 0355 0356 0356 0357 0357 0357 0358 0358 0-099 0358 0359 0359 0359 0360 0360 0361 0361 0361 0362 0-100 0362 0362 0363 0363 0363 0364 0364 0365 0365 0365 0-101 0366 0366 i 0366 0367 0367 0367 0368 0368 0369 0369 0-102 0369 0370 0370 0370 0371 0371 0371 0372 0372 0372 0-103 0373 0373 0374. 0374 0374 0375 0375 0375 0376 0376 0-104 0376 0377 0377 0378 0378 0378 0379 0379 0379 0380 0-105 0380 0380 0381 0381 0382 0382 0382 0383 0383 0383 0-106 0384 0384 0384 0385 0385 0386 0386 0386 0387 0387 0-107 0387 0388 0388 0388 0389 0389 0389 0390 0390 0391 0-108 0391 0391 0392 0392 0392 0393 0393 0394 0394 0394 0-109 0395 0395 0395 0396 0396 0396 0397 0397 0397 0398 0-110 0398 0399 0399 0399 0400 0400 0400 0401 0401 ' 0401 0-111 0402 0402 0403 0403 i 0403 0404 0404 0404 0405 0405 0-112 0405 0406 0406 0407 i 0407 0407 0408 0408 0408 0409 0-113 0409 0409 0410 0410 0411 0411 0411 0412 0412 0412 0-114 0413 0413 0413 0414 0414 0414 0415 0415 0416 0416 0-115 0416 0417 0417 0417 0418 0418 0418 0419 0419 0420 0-116 0420 0420 0421 0421 0421 0422 0422 0422 0423 0423 0-117 0424 0424 0424 0425 0425 0425 0426 0426 0426 0427 0-118 0427 0428 0428 0428 0429 0429 0429 0430 0430 0430 0-119 0431 0431 0432 0432 0432 0433 0433 0433 0434 0434 0-120 0434 0435 0435 0435 0436 0436 0437 0437 0437 0438 0-121 0438 0438 0439 0439 r 0439 0440 0440 0441 0441 0441 0-122 0442 0442 0442 0443 0443 0443 0444 0444 0445 0445 0-123 0445 0446 0446 0446 0447 0447 0447 0448 0448 i 0449 0-124 0449 0449 0450 0450 ! 0450 0451 0451 0451 0452 0452 0-125 0453 0453 0453 0454 0454 0454 0455 0455 0455 0456 0-126 0456 0456 0457 0457 0458 0458 0458 0459 0459 0459 0-127 0460 0460 0460 0461 0461 0462 0462 0462 0463 0463 0-128 0463 0464 0464 0464 0465 0465 0466 0466 0466 0467 0-129 0467 0467 0468 0468 0468 0469 0469 0470 0470 0470 i APPENDIX. 711 Table XC. Mg 2 P 2 O r to MgO. continued. Grm. Grm. MgO. Mg 2 P 2 7 . 1 2 3 4 5 6 7 8 9 0-130 0471 0471 0471 0472 0472 0472 0473 0473 0473 0474 0-131 0474 0475 0475 0475 0476 0476 0476 0477 0477 0477 0-132 0478 0478 0479 0479 0479 0480 0480 0480 0481 0481 0-133 0481 0482 0482 0483 0483 0483 0484 0484 0484 0485 0-134 0485 0485 0486 0486 0487 0487 0487 0488 0488 0488 0-135 0489 0489 0489 0490 0490 0491 0491 0491 0492 0492 0-136 0492 0493 0493 0493 0494 0494 0494 0495 0495 0496 0-137 0496 0496 0497 0497 0497 0498 0498 0498 0499 0499 0-138 0500 0500 0500 0501 0501 0501 0502 0502 0502 0503 0-139 0503 0504 0504 0504 0505 0505 0505 0506 0506 0506 0-140 0507 0507 0508 0508 0508 0509 0509 0509 0510 0510 0-141 0510 0511 0511 0512 0512 0512 0513 0513 0513 0514 0-142 0514 0515 0515 0515 0516 0516 0516 0517 0517 0518 0-143 0518 0518 0519 0519 0519 0520 0520 0520 0521 0521 0-144 0521 0522 0522 0523 0523 0523 0524 0524 0524 0525 0-145 0525 0526 0526 0526 0527 0527 0527 0528 0528 0528 0-146 0529 0529 0529 0530 0530 0531 0531 0531 0532 0532 0-147 0532 0533 0533 0533 0534 0534 0534 0535 0535 0535 0-148 0536 0536 0536 0537 0537 0538 0538 0538 0539 0539 0-149 0539 0540 0540 0540 0541 0541 0542 0542 0542 0543 0-150 0543 0543 0544 0544 0544 0545 0545 0546 0546 0546 0-151 0547 0547 0547 0548 0548 0548 0549 0549 0550 0550 0-152 0550 0551 0551 0551 0551 0552 0552 0553 0553 0553 0-153 0554 0554 0555 0555 0555 0556 0556 0556 0557 0557 0-154 0557 0558 0558 0559 0559 0559 0560 0560 0560 0561 0-155 0561 0561 0562 0562 0563 0563 0563 0564 0564 0564 156 0565 0565 0565 0566 0566 0567 0567 0567 0568 0568 0-157 0568 0569 0569 0569 0570 0570 0571 0571 0571 0572 0-158 0572 0572 0573 0573 0573 0574 0574 0574 0575 0575 0-159 0576 0576 0576 0577 0577 0577 0578 0578 0578 0579 0-160 0579 0580 0580 0580 0581 0581 0581 0582 0582 0582 0-161 0583 0583 0584 0584 0584 0585 0585 0585 0586 0586 0-162 0586 0587 0587 0588 0588 0588 0589 0589 0589 0590 0-163 0590 0590 0591 0591 0592 0592 0592 0593 0593 0593 0-164 0594 0594 0594 0595 0595 0595 0596 0596 0597 0597 0-165 0597 0598 0598 0598 0599 0599 0599 0600 0600 0601 0-166 0601 0601 0602 0602 0602 0603 0603 0603 0604 0604 0-167 0605 0605 0605 0606 0606 0606 0607 0607 0607 0608 0-168 0608 0609 0609 0609 0610 0610 0610 0611 0611 0611 0-169 0612 0612 0613 0613 0613 0614 0614 0614 0615 0615 0-170 0615 0616 0616 0616 0617 0617 0618 0618 0618 0619 0-171 0619 0619 0620 0620 0620 0621 0621 0622 0622 0622 0-172 0623 0623 0623 0624 0624 0624 0625 0625 0626 ( 0626 0-173 0626 0627 0627 0627 0628 0628 0628 0629 0629 0630 0-174 0630 0630 0631 0631 0631 0632 0632 0633 0633 0633 712 A TREATISE ON CHEMICAL ANALYSIS. Table XC. Mg 2 P 2 O 7 to MgO. continued. Grm. Grm. MgO. Mg 2 P 2 7 . 1 2 3 4 5 6 7 8 9 0-175 0634 0634 0634 0635 0635 0635 0636 0636 0636 0637 0-176 0637 0637 0638 0638 0639 0639 0639 0640 0640 0640 0-177 0641 0641 0641 0642 0642 0643 0643 0643 0644 0644 0-178 0645 0645 0645 0646 0646 0646 0647 0647 0647 0648 0-179 0648 0649 0649 0649 0650 0650 0650 0651 0651 0651 0-180 0652 0652 0652 0653 0653 0653 0654 0654 0654 0655 0-181 0655 0656 0656 0656 0657 0657 0657 0658 0658 0658 0-182 0659 0659 0660 0660 0660 0661 0661 0661 0662 0662 0-183 0662 0663 0663 0664 0664 0664 0665 0665 0665 0666 0-184 0666 0666 0667 0667 0668 0668 0668 0669 0669 0669 0-185 0670 0670 0670 0671 0671 0672 0672 0672 0673 0673 0-186 0673 0674 0674 0674 0675 0675 0675 0676 0676 0677 0-187 0677 0677 0678 0678 0678 0679 0679 0679 0680 0680 0-188 0681 0681 0681 0682 0682 0682 0683 0683 0683 0684 0-189 0684 0685 0685 0685 0686 0686 0686 0687 0687 0687 0-190 0688 0688 0689 0689 0689 0690 0690 0690 0691 0691 0-191 0691 0692 0692 0693 0693 0693 0694 0694 0694 0695 0-192 0695 0695 0696 0696 0696 0697 0697 0698 0698 ! 0698 0-193 0699 0699 0699 0700 0700 0700 0701 0701 0702 0702 0-194 0702 0703 0703 0703 0704 0704 0704 0705 0705 0706 0-195 0706 0706 0707 0707 0707 0708 0708 0708 0709 0709 0-196 0710 0710 0710 0711 0711 0711 0712 0712 0712 0713 0-197 0713 0714 0714 0714 0715 0715 0715 0716 0716 0716 0-198 0717 0717 0717 0718 0718 0719 0719 0719 0720 0720 0-199 0720 0721 0721 0721 0722 0722 0723 0723 0723 0724 0-200 0724 0724 , 0725 0725 0725 0726 0726 0727 0727 0727 0-201 0728 0728 0728 0729 0729 0729 0730 0730 0731 0731 0-202 0731 0732 0732 0732 0733 0733 0733 0734 0734 0734 0-203 0735 0735 0736 0736 0736 0737 0737 0737 0738 0738 0-204 0738 0739 0739 0740 0740 0740 0741 0741 0741 0742 0-205 0742 0742 0743 0743 0744 0744 0744 0745 0745 0745 0-206 0746 0746 0746 0747 0747 0748 0748 0748 0749 0749 0-207 0749 0750 0750 0750 0751 0751 0751 0752 0752 0753 0-208 0753 0753 0754 0754 0754 0755 0755 0755 0756 0756 0-209 0757 0757 0757 0758 0758 0758 0759 0759 0759 0760 0-210 0760 0760 0761 0761 0761 0762 0762 0763 0763 0763 0-211 0764 0764 0765 0765 0765 0766 0766 0766 0767 0767 0-212 0767 0768 0768 0769 0769 0769 0770 0770 0770 0771 0-213 0771 0771 0772 0772 0773 0773 0773 0774 0774 0774 0-214 0775 0775 0775 0776 0776 0776 0777 0777 0778 0778 0-215 0778 0779 0779 0779 0780 0780 0780 0781 0781 0782 0-216 0782 0782 0783 0783 0783 0784 0784 0784 0785 0785 0-217 0786 0786 0786 0787 0787 0787 0788 0788 0788 0789 0-218 0789 0790 0790 0790 0791 0791 0791 0792 0792 0792 0-219 0793 0793 0794 0794 0794 0794 0795 0795 0795 0796 APPENDIX. Table XC. Mg 2 P 2 O 7 to MgO. continued. 713 Grm. Mg 2 P 2 7 Grm. MgO. 1 2 3 4 5 6 7 8 9 0-220 0796 0797 0797 0797 0798 0798 0799 0799 0799 0800 0-221 0800 0800 0801 0801 0801 0802 0802 0803 0803 v/Ov/U 0803 0-222 0804 0804 0804 0805 0805 0805 0806 0806 0807 0807 0-223 0807 0808 0808 0808 0809 0809 0809 0810 0810 0811 0-224 0811 0811 0812 0812 0812 0813 0813 0813 0814 0814 0-225 0815 0815 0815 0816 0816 0816 0817 0817 0817 0818 0-226 0818 0818 0819 0819 0820 0820 0820 0821 0821 0821 0-227 0822 0822 0822 0823 0823 0824 0824 0824 0825 0825 0-228 0825 0826 0826 0826 0827 0827 0828 0828 0828 0829 0-229 0829 0829 0830 0830 0830 0831 0831 0831 0832 0832 0-230 0833 0833 0833 0834 0834 0834 0835 0835 0835 0836 0-231 0836 0837 0837 0837 0838 0838 0838 0839 0839 0839 0-232 0840 0840 0841 0841 0841 0842 0842 0842 0843 0843 0-233 0843 0844 0844 0845 0845 0845 0846 0846 0846 0847 0-234 0847 0847 0848 0848 0849 0849 0849 0850 0850 0850 0-235 0851 0851 0851 0852 0852 0853 0853 0853 0854 0854 0-236 0854 0855 0855 0855 0856 0856 0856 0857 0857 0858 0-237 0858 0858 0859 0859 0859 0860 0860 0860 0861 0861 0-238 0862 0862 0862 0863 0863 0863 | 0864 0864 0864 0865 0-239 0865 0866 0866 0866 0867 0867 0867 0868 0868 0868 0-240 0869 0869 0870 0870 0870 0871 0871 0871 0872 0872 0-241 0872 0873 0873 0874 0874 0874 0875 0875 0875 0876 0-242 0876 0876 0877 0877 0877 0878 0878 0879 0879 0879 0-243 0880 0880 0880 0881 0881 0881 0882 0882 0883 0883 0-244 0883 0884 0884 0884 0885 0885 0885 0886 0886 0887 0-245 0887 0887 0888 0888 0888 0889 0889 0889 0890 0890 0-246 0891 0891 0891 0892 0892 0892 0893 0893 0893 0894 0-247 0894 0895 0895 0895 0896 0896 0896 0897 0897 0897 0-248 0898 0898 0898 0899 0899 0900 0900 0900 0901 0901 0-249 0901 0902 0902 0902 0903 0903 0904 0904 0904 0905 0-250 0905 0905 0906 0906 0906 0907 0907 0908 0908 0908 0-251 0909 0909 0909 0910 0910 0910 0911 0911 0912 0912 0-252 0912 0913 0913 0913 0914 0914 0914 0915 0915 0915 0-253 0916 0916 0917 0917 0917 0918 0918 0918 0919 0919 0-254 0919 0920 0920 0920 0921 0921 0921 0922 0922 0923 0-255 0923 0923 0924 0924 0924 0925 0925 0925 0926 0926 0-256 0927 0927 0927 0928 0928 0928 0929 0929 0929 0930 0-257 0930 0931 0931 0931 0932 0932 0932 0933 0933 0934 0-258 0934 0934 0935 0935 0935 0936 0936 0936 0937 0937 0-259 0938 0938 0938 0939 0939 0939 0940 0940 0940 0941 0-260 0941 0942 0942 0942 0943 0943 0943 0944 0944 0944 0-261 0945 0945 0946 0946 0946 0947 0947 0947 0948 0948 0-262 0948 0949 0949 0950 0950 0950 0951 0951 0951 0952 0-263 0952 0952 0953 0953 0954 0954 0954 0955 0955 0955 0-264 0956 0956 0956 0957 0957 0957 0958 0958 0959 0959 A TREATISE ON CHEMICAL ANALYSIS. Table XC. Mg 2 P 2 O 7 to MgO. continued. Grm. Grm. MgO. Mg 2 P 2 7 . 1 2 3 4 5 6 7 8 9 0-265 0959 0960 0960 0960 0961 0961 0961 0962 0962 0963 0-266 0963 0963 0964 0964 0964 0965 0965 0965 0966 0966 0-267 0966 0967 0967 0967 0968 0968 0968 0969 0969 0969 0-268 0970 0970 0971 0971 0971 0972 0972 0972 0973 0973 0-269 0973 0974 0974 0975 0975 0975 0976 0976 0976 0977 0-270 0977 0977 0978 0978 0979 0979 0979 0980 0980 0980 0-271 0981 0981 0981 0982 0982 0983 0983 0983 0984 0984 0-272 0984 0985 0985 0985 0986 0986 0987 0987 0987 0988 0-273 0988 0988 0989 0989 0989 0990 0990 0991 0991 0991 0-274 0992 0992 0992 0993 0993 0993 0994 0994 0995 0995 0-275 0996 0996 0997 0997 0997 0998 0998 0998 0999 0999 0-276 0999 1000 1000 1000 1001 1001 1001 1002 1002 1002 0-277 1003 1003 1003 1004 1004 1005 1005 1005 1006 1006 0-278 1006 1007 1007 1007 1008 1008 1009 1009 1009 1010 0-279 1010 1010 1011 1011 1011 1012 1012 1013 1013 1013 0-280 1014 1014 1014 1015 1015 1015 1016 1016 1017 1017 0-281 1017 1018 1018 1018 1019 1019 1019 1020 1020 1020 0-282 1021 1021 1022 1022 1022 1023 1023 1023 1024 1024 0-283 1024 1025 1025 1026 1026 1026 1027 1027 1027 1028 0-284 .1028 1028 1029 1029 1030 1030 1030 1031 1031 1031 0-285 1032 1032 1032 1033 1033 1034 1034 1034 1035 1035 0-286 1035 1036 1036 1036 1037 1037 1037 1038 1038 1039 0-287 1039 1039 1040 1040 1040 1041 1041 1041 1042 1042 0-288 1043 1043 1043 1044 1044 1044 1045 1045 1045 1046 0-289 1046 1047 1047 1047 1048 1048 1048 1049 1049 1049 0-290 1050 1050 1051 1051 1051 1052 1052 1052 1053 1053 0-291 1053 1054 1054 1055 1055 1055 1056 1056 1056 1057 0-292 1057 1057 1058 1058 1058 1059 1059 1060 1060 1060 0-293 1061 1061 1061 1062 1062 1062 1063 1063 1064 1064 0-294 1064 1065 1065 1065 1066 1066 1066 1067 1067 1068 0-295 1068 1068 1069 1069 1069 1070 1070 1070 1071 1071 0-296 1072 1072 1072 1073 1073 1073 1074 1074 1074 1075 0-297 1075 1076 1076 1076 1077 1077 1077 1078 1078 1078 0-298 1079 1079 1079 i 1080 1080 1081 1081 1081 1082 1082 0-299 1082 1083 1083 1083 1084 1084 1085 1085 1085 1086 APPENDIX. 715 Table XCL Conversion of Grms. of Magnesium Pyrophosphate (Mg 2 P 2 O 7 ) into Grms. Phosphorus Pentoxide (P 2 O 5 ). (FACTOR 0'638.) Grm. Grm. P 2 5 . Mg 2 P 2 7 . i 1 2 3 4 5 6 7 8 9 0-000 0001 0001 0002 0003 0003 0004 0004 0005 0006 0-001 0006 0007 0008 0008 0009 0010 0010 0011 0011 0012 0-002 0013 0013 0014 0015 0015 0016 0017 0017 0018 0019 0-003 0019 0020 0020 0021 0022 0022 0023 0024 0024 0025 0-004 0026 0026 0027 0027 0028 0029 0029 0030 0031 0031 0-005 0032 0033 0033 0034 0034 0035 0036 0036 0037 0038 0-006 0038 1 0039 0040 0040 0041 0041 0042 0043 0043 0044 0-007 0045 0045 0046 0047 0047 0048 0048 0049 0050 0050 0-008 0051 1 0052 0052 0053 0054 0055 0055 0056 0056 0057 0-009 0057 0058 0059 0059 0060 0061 0061 0062 0063 0063 0-010 0064 0064 0065 0066 0066 0067 0068 0068 0069 0070 0-011 0070 ' 0071 0071 0072 0073 0073 0074 0075 0075 0076 0-012 0077 0077 0078 0078 0079 0080 0080 0081 0082 0082 0-013 0083 0084 0084 0085 0085 0086 0087 0087 0088 0089 0-014 0089 0090 0091 0091 0092 0093 0093 0094 0094 0095 0-015 0096 0096 0097 0097 0098 0099 0100 0100 0101 0101 0-016 0102 0103 0103 0104 0105 0105 0106 0107 0107 0108 0-017 0108 0109 0110 0110 0111 0112 0112 0113 0114 0114 0-018 0115 0115 0116 0117 0117 0118 0119 0119 0120 0121 0-019 0121 0122 0122 0123 0124 0124 0125 0126 0126 0127 0-020 0128 0128 0129 0130 0130 0131 0131 0132 0133 0133 0-021 0134 0135 0135 0136 0137 0137 0138 0138 0139 0140 0-022 0140 : 0141 0142 0143 0143 0144 0144 0145 0145 0146 0-023 0147 0147 0148 0149 0149 0150 0151 0151 0152 0152 0-024 0153 0154 0154 0155 0156 0156 0157 0158 0158 0159 0-025 0160 0160 0161 0161 0162 0163 0163 0164 0165 0165 0-026 0166 0167 0167 0168 0168 0169 0170 0170 0171 0172 0-027 0172 0173 0174 0174 0175 0175 0176 0177 0177 0178 0-028 0179 0179 0180 0181 0181 0182 0182 0183 0184 0184 0029 0185 0186 0186 0187 0188 0189 0189 0189 0190 0191 0-030 0191 0192 0193 0193 0194 0195 0195 0196 0197 0197 0-031 0198 0198 0199 0200 0200 0201 0202 0202 0203 0204 0-032 0204 0205 0205 0206 0207 0207 0208 0209 0209 0210 0-033 0211 0211 0212 0212 0213 0214 0214 0215 0216 0216 0-034 0217 0218 0218 0219 0219 0220 0221 0221 0222 0223 0-035 0223 0224 ! 0225 0225 0226 0226 0227 0228 0228 0229 0-036 0230 0230 0231 0232 0232 0233 0234 0234 0235 0236 0-037 0236 0237 0237 0238 0239 0239 0240 0241 0241 0242 0-038 0242 0243 0244 0244 0245 0246 0246 0247 0248 0248 0-039 0249 0249 0250 0251 0251 0252 0253 0263 0264 0255 7 i6 A TREATISE ON CHEMICAL ANALYSIS. Table XCL Mg 2 P 2 O 7 to P 2 O 5 . continued. Grm. Mg 2 P 2 7 . Grm. P 2 O 6 . 1 2 3 4 5 6 7 8 9 0-040 0255 0256 0256 0257 0258 0258 0259 0260 0260 0261 0-041 0262 0262 0263 0263 0264 0265 0265 0266 0267 0267 0-042 0268 0269 0269 0270 0271 0271 0272 0272 0273 0274 0-043 0274 0275 0276 0276 0277 0278 0278 0279 0279 0280 0-044 0281 0281 0282 0283 0283 0284 0285 0285 0286 0286 0-045 0287 0288 0288 0289 0290 0290 0291 0292 0292 0293 0-046 0293 0294 0295 0295 0296 0297 0297 0298 0299 0299 0-047 0300 0300 0301 0302 0302 0303 0304 0304 0305 0306 0-048 0306 0307 0308 0308 0309 0309 0310 0311 0311 0312 0-049 0313 0313 0314 0315 0315 0316 0316 0317 0318 0318 0-050 0319 0320 0320 0321 0322 0322 0323 0323 0324 0325 0-051 0325 0326 0327 0327 0328 0329 0329 0330 0330 0331 0-052 0332 0332 0333 0334 0334 0335 0336 0336 0337 0338 0-053 0338 0339 0339 0340 0341 0341 0342 0343 0343 0344 0-054 0345 0345 0346 0347 0348 0348 0348 0349 0350 0350 0-055 0351 0352 0352 0353 0353 0354 0355 0355 0356 0357 0-056 0357 0358 0359 0359 0360 0360 0361 0362 0362 0363 0-057 0364 0364 0365 0366 0366 0367 0367 0368 0369 0369 0-058 0370 0371 0371 0372 0373 0373 0374 0375 0375 0376 0-059 0376 0377 0378 0378 0379 0380 0380 0381 0382 0382 0-060 0383 0383 0384 0385 0385 0386 0387 0387 0388 0389 0-061 0389 0390 0390 0391 0392 0392 0393 0394 0394 0395 0-062 0396 0396 0397 0397 0398 0399 0399 0400 0401 0401 0-063 0402 0403 0403 0404 0404 0405 0406 0406 0407 0408 0-064 0408 0409 0410 0410 0411 0412 0412 0413 0413 0414 0-065 0415 0415 0416 0417 0417 0418 0419 0419 0420 0420 0-066 0421 0422 0422 0423 0424 0424 0425 0426 0426 0427 0-067 0427 0428 0429 0429 0430 0431 0431 0432 0433 0433 0-068 0434 0434 0435 0436 0436 0437 0438 0438 0439 0440 0-069 0440 0441 0441 0442 0443 0443 0444 0445 0445 0446 0-070 0447 0447 0448 0449 0449 0450 0450 0451 0452 0452 0-071 0453 0454 0454 0455 0456 0456 0457 0457 0458 0459 0-072 0459 0460 0461 0461 0462 0463 0463 0464 0464 0465 0-073 0466 0466 0467 0468 0468 0469 0470 0470 0471 0471 0-074 0472 0473 0473 0474 0475 0475 0476 0477 0477 0478 0-075 0479 0479 0480 0481 0482 0482 0482 0483 0484 0484 0-076 0485 0486 0486 0487 0487 0488 0489 0489 0490 0491 0-077 0491 0492 0493 0493 0494 0494 0495 0496 0496 0497 0-078 0498 0498 0499 0500 0500 0501 0501 0502 0503 0503 0-079 0504 0505 0505 0506 0507 0507 0508 0508 0509 0510 0-080 0510 0511 0512 0512 0513 0514 0514 0515 0516 0516 0-081 0517 0517 0518 0519 0519 0520 0521 0521 0522 0523 0-082 0523 0524 0524 0525 0526 0526 0527 0528 0528 0529 0-083 0530 0530 0531 0531 0532 0533 0533 0534 0535 0535 0-084 0536 0537 0537 0538 0538 0539 0540 0540 0541 0542 APPENDIX. 717 Table XCL Mg 2 P 2 O 7 to PA. continued. Grm. Mg 2 P 2 7 . Grm. P 1 6 . 1 2 3 4 5 6 7 8 9 0-085 0542 0543 0544 0544 0545 0545 0546 0547 0547 0548 0-086 0549 0549 0550 0551 0551 0552 0553 0553 0554 0554 0-087 0555 0556 0556 0557 0558 0558 0559 0560 0560 0561 0-088 0561 0562 0563 0563 0564 0565 0565 0566 0567 0567 0-089 0568 0568 0569 0570 0570 0571 0572 0572 0573 0574 0-090 0574 0575 0575 0576 0577 0577 0578 0579 0579 0580 0-091 0581 0581 0582 0582 0583 0584 0584 0585 0586 0586 0-092 0587 0588 0588 0589 0590 0590 0591 0591 0592 0593 0-093 0593 0594 0595 0595 0596 0597 0597 0598 0598 0599 0-094 0600 0600 0601 0602 0602 0603 0603 0604 0605 0605 0-095 0606 0607 0607 0608 0609 0609 0610 0611 0611 0612 0-096 0612 0613 0614 0614 0615 0616 0616 0617 0618 0618 0-097 0619 0619 0620 0621 0621 0622 0623 0623 0624 0625 0-098 0625 0626 0627 0627 0628 0628 0629 0630 0630 0631 0-099 0632 0632 0633 0634 0634 0635 0635 0636 0637 0637 0-100 0638 0639 0639 0640 0641 0641 0642 0642 0643 0644 0-101 0644 0645 0646 0646 0647 0648 0649 0649 0650 0650 0-102 0651 0651 0652 0653 0654 0654 0655 0656 0656 0657 0-103 0657 0658 0658 0659 0660 0660 0661 0662 0662 0663 0-104 0664 0664 0665 0665 0666 0667 0667 0668 0669 0669 0-105 0670 0671 0671 0672 0672 0673 0674 0674 0675 0676 0-106 0676 0677 0678 0678 0679 0679 0680 0681 0681 0682 0-107 0683 0683 0684 0685 0685 0686 0686 0687 0688 0688 0-108 0689 0690 0690 0691 0692 0693 0693 0694 0694 0695 0-109 0695 0696 0697 0698 0698 0699 0699 0700 0701 0701 0-110 0702 0702 0703 0704 0704 0705 0706 0706 0707 0708 0-111 0708 0709 0709 0710 0711 0711 0712 0713 0713 0714 0-112 0715 0715 0716 0716 0717 0718 0718 0719 0720 0720 0-113 0721 0722 0722 0723 0723 0724 0725 0725 0726 0727 0-114 0727 0728 0729 0729 0730 0731 0731 0732 0732 0733 0-115 0734 0734 0735 0736 0736 0737 0738 0738 0739 0739 0-116 0740 0741 0741 0742 0743 0743 0744 0745 0745 0746 0-117 0746 0747 0748 0748 0749 0750 0750 0751 0752 0752 0-118 0753 0753 0754 0755 0755 0756 0757 0757 0758 0759 0-119 0759 0760 0760 0761 0762 0762 0763 0764 0764 0765 0-120 0766 0766 0767 0768 0769 0769 0770 0770 0771 0771 0-121 0772 0773 0773 0774 0775 0775 0776 0776 0777 0778 0-122 0778 0779 0780 0780 0781 0782 0782 0783 0783 0784 0-123 0785 0785 0786 0787 0787 0788 0789 0789 0790 0790 0-124 0791 0792 0792 0793 0794 0794 0795 0796 0796 0797 0-125 0798 0798 0799 0799 0800 0801 0801 0802 0803 0803 0-126 0804 0805 0805 0806 0806 0807 0808 0808 0809 0810 0-127 0810 0811 0812 0812 0813 0813 0814 0815 0815 0816 0-128 0817 0817 0818 0819 0819 0820 0820 0821 0822 0822 0-129 0823 0823 0824 0825 0825 0826 0827 0827 0828 0829 7 i8 A TREATISE ON CHEMICAL ANALYSIS. Table XCL Mg 2 P 2 O 7 to P 2 O 5 . continued. Grni. Mg 2 P 2 7 . Grm. P 2 O 5 . 1 2 3 4 5 6 7 8 9 0-130 0829 0830 0831 0831 0832 0833 0833 0834 0835 0835 0-131 0836 0836 0837 0838 0838 0839 0840 0840 0841 0842 0-132 0842 0843 0843 0844 0845 0845 0846 0847 0847 0848 0-133 0849 0849 0850 0850 0851 0852 0852 0853 0854 0854 0-134 0855 0856 0856 0857 0857 0858 0859 0859 0860 0860 0-135 0861 0862 0863 0863 0864 0864 0865 0866 0866 0867 0-136 0868 0868 0869 0870 0870 0871 0872 0872 0873 0873 0-137 0874 0875 0875 0876 0877 0877 0878 0879 0879 0880 0-138 0880 0881 0882 0882 0883 0884 0884 0885 0886 0886 0-139 0887 0887 0888 0889 0889 0890 0891 0891 0892 0893 0-140 0893 0894 0894 0895 0896 0896 0897 0898 0898 0899 0-141 0900 0900 0901 0901 0902 0903 0903 0904 0905 0905 0-142 0906 0907 0907 0908 0909 0909 0910 0910 0911 0912 0-143 0912 0913 0914 0914 0915 0916 0916 0917 0917 0918 0-144 0919 0919 0920 0921 0921 0922 0923 0923 0924 0924 0-145 0925 0926 0926 0927 0928 0928 0929 0930 0930 0931 0-146 0931 0932 0933 0933 0934 0935 0935 0936 0937 0937 0-147 0938 0938 0939 0940 0940 0941 0942 0942 0943 0944 0-148 0944 0945 0946 0946 0947 0947 0948 0949 0949 0950 0149 0951 0951 0952 0953 0953 0954 0954 0955 0956 0956 0-150 0957 0957 0958 0959 0960 0960 0961 0961 0962 0963 0-151 0963 0964 0965 0965 0966 0967 0967 0968 0968 0969 0-152 0970 0970 0971 0972 0972 0973 0974 0974 0975 0976 0-153 0976 0977 0977 0978 0979 0979 0980 0981 0982 0982 0-154 0983 0983 0984 0984 0985 0986 0986 0987 0988 0988 0-155 0989 0990 0990 0991 0991 0992 0993 0993 0994 0995 0-156 0995 0996 0997 0997 0998 0998 0999 1000 1000 1001 0-157 1002 1002 1003 1004 1004 1005 ! 1005 1006 1007 1007 0-158 1008 1009 1009 1010 1011 1011 1012 1013 1013 1014 0-159 1014 1015 1016 1016 1017 1018 1018 1019 1020 1020 0-160 1021 1021 1022 1023 1023 1024 1025 1025 1026 1027 0-161 1027 1028 1028 1029 1030 1030 1031 1032 1032 1033 0-162 1034 1034 1035 1035 1036 1037 1037 1038 1039 1039 0-163 1040 1041 1041 1042 1042 1043 1044 1044 1045 1046 0-164 1046 1047 1048 1048 1049 1050 1050 1051 1051 1052 0-165 1053 1053 1054 1055 1055 1056 1057 1057 1058 1058 0-166 1059 1060 1060 1061 1062 1062 1063 1064 1064 1065 0-167 1065 1066 1067 1067 1068 1069 1069 1070 1071 1071 0-168 1072 1072 1073 1074 1074 1075 1076 1076 1077 1078 0-169 1078 1079 1079 1080 1081 1081 1082 1083 1083 1084 0-170 1085 1085 1086 1087 1087 1088 1088 1089 1090 1090 0-171 1091 1092 1092 1093 1094 1094 1095 1095 1096 1097 0-172 1097 1098 1099 1099 1100 1101 1101 1102 1102 1103 0-173 1104 1104 1105 1106 1106 1107 1108 1108 1109 1109 0-174 1110 1111 1111 1112 1113 1113 1114 1115 1115 1116 APPENDIX. 719 Table XCI. Mg 2 P 2 O 7 to P 2 O 5 . continued. Grm. Mg 2 P 2 7 . Grm. P 2 5 . 1 2 3 4 5 6 7 8 9 0-175 1117 1117 1118 1118 1119 1120 1120 1121 1122 1122 0-176 1123 1124 1124 1125 1125 1126 1127 1127 1128 1129 0-177 1129 1130 1131 1131 1132 1132 1133 1134 1134 1135 0-178 1136 1136 1137 1138 1138 1139 1139 1140 1141 1141 0-179 1142 1143 1143 1144 1145 1145 1146 1146 1147 1148 0-180 1148 1149 1150 1150 1151 1152 1152 1153 1154 1154 0-181 1155 1155 1156 1157 1157 1158 1159 1159 1160 1161 0-182 1161 1162 1162 1163 1164 1164 1165 1666 1166 1167 0-183 1168 1168 1169 1169 1170 1171 1171 1172 1173 1173 0-184 1174 1175 1175 1176 1176 1177 1178 1178 1179 1180 0-185 1180 1181 1182 1182 1183 1183 1184 1185 1185 1186 0-186 1187 1187 1188 1189 1189 1190 1191 1191 1192 1192 0-187 1193 1194 1194 1195 1195 1196 1197 1198 1198 1199 0-188 1199 1200 1201 1201 1202 1202 1203 1204 1205 1205 0189 1206 1206 1207 1208 1208 1209 1210 1210 1211 1212 0-190 1212 1213 1213 1214 1215 1215 1216 1217 1217 1218 0-191 1219 1220 1220 1221 1221 1222 1223 1223 1224 1224 0-192 1225 1226 1226 1227 1228 1228 1229 1229 1230 1231 0-193 1231 1232 1233 1233 1234 1234 1235 1236 1236 1237 0-194 1238 1238 1239 j 1240 1240 1241 1241 1242 1243 1243 0-195 1244 1245 1245 1246 1247 1247 1248 1249 1249 1250 0-196 1250 1251 1252 1252 1253 1254 1254 1255 1256 1256 0-197 1257 1257 1258 1259 1259 1260 1261 1261 1262 1263 0-198 1263 1264 1264 1265 1266 1266 1267 1268 1268 1269 0-199 1270 1270 1271 1272 1272 1273 1273 1274 1275 1275 720 A TREATISE ON CHEMICAL ANALYSIS. Table XCI L Conversion of Grms. of Barium Sulphate into Grms. of Sulphur Trioxide. (FACTOR 0-343.) Grm. BaS0 4 . Grm. SO 3 . 1 2 3 4 5 6 7 8 9 0-000 0001 0001 0001 0002 0002 0002 0003 0003 0-001 0003 0004 0004 0004 0005 0005 0005 0006 0006 0007 0-002 0007 0007 0008 0008 0008 0009 0009 0009 0010 0010 0-003 0010 0011 0011 0011 0012 0012 0012 0013 0013 0013 0-004 0014 0014 0014 0015 0015 0015 0016 0016 0016 0017 0-005 0017 0017 0018 0018 0019 0019 0019 0020 0020 0020 0-006 0021 0021 0021 0022 0022 0022 0023 0023 0023 0024 0-007 0024 0024 0025 0025 0026 0026 0026 0027 0027 0027 0-008 0028 0028 0028 0028 0029 0029 0029 0030 0030 0031 0-009 0031 0031 0032 0032 0032 0033 0033 0033 0034 0034 0-010 0034 0035 0035 0035 0036 0036 0036 0037 0037 0037 0-011 0038 0038 0038 0039 0039 0039 0040 0040 0040 0041 0-012 0041 0042 0042 0042 0043 0043 0043 0044 0044 0044 0-013 0045 0045 0045 0046 0046 0046 0047 0047 0047 0048 0-014 0048 0048 0049 0049 0049 0050 0050 0050 0051 0051 0-015 0051 0052 0052 0052 0053 0053 0054 0054 0054 0055 0-016 0055 0055 0056 0056 0056 0057 0057 0057 0058 0058 0-017 0058 0059 0059 0059 0060 0060 0060 0061 0061 0061 0-018 0062 0062 0062 0063 0063 0063 0064 0064 0064 0065 0-019 0065 0066 0066 0066 0067 0067 0067 0068 0068 0068 0-020 0069 0069 0069 0070 0070 0070 0071 0071 0071 0072 0-021 0072 0072 0073 0073 0073 0074 0074 0074 0075 0075 0-022 0075 0076 0076 0076 0077 0077 0078 0078 0078 0079 0-023 0079 0079 0080 0080 0080 0081 0081 0081 0082 0082 0-024 0082 0083 0083 0083 0084 0084 0084 0085 0085 0085 0-025 0086 0086 0086 0087 0087 0087 0088 0088 0088 0089 0-026 0089 0090 0090 0090 0091 0091 0091 0092 0092 0092 0-027 0093 0093 0093 0094 0094 0094 0095 0095 0095 0096 0-028 0096 0096 0097 0097 0097 0098 0098 0098 0099 0099 0-029 0099 0100 0100 0100 0101 0101 0102 0102 0102 0103 0-030 0103 0103 0104 0104 0104 0105 0105 0105 0106 0106 0-031 0106 0107 0107 0107 0108 0108 0108 0109 0109 0109 0-032 0110 0110 0110 0111 0111 0112 0112 0112 0113 0113 0-033 0113 0114 0114 0114 0115 0115 0115 0116 0116 0116 0-034 0117 0117 0117 0118 0118 0118 0119 0119 0119 0120 0-035 0120 0120 0121 0121 0121 0122 0122 0122 0123 0123 0-036 0123 0124 0124 0125 0125 0125 0126 0126 0126 0127 0-037 0127 0127 0128 0128 0128 0129 0129 0129 0130 0130 0038 0130 0131 0131 0131 0132 0132 0132 0133 0133 0133 0-039 0134 0134 0134 0135 0135 0135 0136 0136 0136 0137 1 APPENDIX. 721 Table XCIL BaSO 4 to SO* continued. Grm. BaS0 4 . Grm. S0 3 . 1 2 3 4 5 6 7 8 9 0-040 0137 0138 0138 0138 0139 0139 0139 0140 0140 0140 0-041 0141 0141 0141 0142 0142 0142 0143 0143 0143 0144 0-042 0144 0144 0145 0145 0145 0146 0146 0146 0147 0147 0-043 0147 0148 0148 0149 0149 0149 0150 0150 0150 0151 0-044 0151 0151 0152 0152 0152 0153 0153 0153 0154 0154 0-045 0154 0155 0155 0155 0156 0156 0156 0157 0157 0157 0-046 0158 0158 0158 0159 0159 0159 0160 0160 0161 0161 0-047 0161 0162 0162 0162 0163 0163 0163 0164 0164 0164 0-048 0165 0165 0165 0166 0166 0166 0167 0167 0167 0168 0-049 0168 0169 0169 0169 0169 0170 0170 0170 0171 0171 0-050 0172 0172 0172 0173 0173 0173 0174 0174 0174 0175 0-051 0175 0175 0176 0176 0176 '0177 0177 0177 0178 0178 0-052 0178 0179 0179 0179 0180 0180 0180 0181 0181 0181 0-053 0182 0182 0182 0183 0183 0184 0184 0184 0185 0185 0-054 0185 0186 0186 0186 0187 0187 0187 0188 0188 0188 0-055 0189 0189 0189 0190 0190 0190 0191 0191 0191 0192 0-056 0192 0192 0193 0193 0193 0194 0194 0194 0195 0195 0-057 0196 0196 0196 0197 0197 0197 0198 0198 0198 0199 0-058 0199 0199 0200 0200 0200 0201 0201 0201 0202 0202 0-059 0202 0203 0203 0203 0204 0204 0204 0205 0205 0205 0-060 0206 0206 0206 0207 0207 0208 0208 0208 0209 0209 0-061 0209 0210 0210 0210 0211 0211 0211 0212 0212 0212 0-062 0213 0213 0213 0214 0214 0214 0215 0215 0215 0216 0-063 0216 0216 0217 0217 0218 0218 0218 0219 0219 0219 0-064 0220 0220 0220 0221 0221 0221 0222 0222 0222 0223 0-065 0223 0223 0224 0224 0224 0225 0225 0225 0226 0226 0-066 0226 0227 0227 0227 0228 0228 0228 0229 0229 0229 0-067 0230 0230 0230 0231 0231 0232 0232 0232 0233 0233 0-068 0233 0234 0234 0234 0235 0235 0235 0236 0236 0236 0-069 0237 0237 0237 0238 0238 0238 0239 0239 0239 0240 0-070 0240 0240 0241 0241 0241 0242 0242 0243 0243 0243 0-071 0244 0244 0244 0245 0245 0245 0246 0246 0246 0247 0-072 0247 0247 0248 0248 0248 0249 0249 0249 0250 0250 0-073 0250 0251 0251 0251 0252 0252 0252 0253 0253 0253 0-074 0254 0254 0255 0255 0255 0256 0256 0256 0257 0257 0-075 0257 0258 0258 0258 0259 0259 0259 0260 0260 0260 0-076 0261 0261 0261 0262 0262 0262 0263 0263 0263 0264 0-077 0264 0264 0265 0265 0265 0266 0266 0267 0267 0267 0-078 0268 0268 0268 0269 0269 0269 0270 0270 0270 0271 0-079 0271 0271 0272 0272 0272 0273 0273 0273 0274 0274 0-080 0274 0275 0275 0275 0276 0276 0276 0277 0277 0277 0-081 0278 0278 0279 0279 0279 0280 0280 0280 0281 0281 0-082 0281 0282 0282 0282 0283 0283 0283 0284 0284 0284 0-083 0285 0285 0285 0286 0286 0286 0287 0287 0287 0288 0-084 0288 0288 0289 0289 0289 0290 0290 0291 0291 0291 46 722 A TREATISE ON CHEMICAL ANALYSIS. Table XCIL BaSCX to SO*. continued. Grm. BaS0 4 . Grm. S0 a . 1 2 3 4 5 6 8 9 0-085 0292 0292 0292 0293 0293 0293 0294 0294 0294 0295 0-086 0295 0295 0296 0296 0296 0297 0297 0297 0298 0298 0-087 0298 0299 0299 0299 0300 0300 0300 0301 0301 0301 0-088 0302 0302 0303 0303 0303 0304 0304 0304 0305 0305 0-089 0305 0306 0306 0306 0307 0307 0307 0308 0308 0308 0-090 0309 0309 0309 0310 0310 0310 0311 0311 0311 0312 0-091 0312 0312 0313 0313 0314 0314 0314 0315 0315 0315 0-092 0316 0316 0316 0317 0317 0317 0318 0318 0318 0319 0-093 0319 0319 0320 0320 0320 0321 0321 0321 0322 0322 0-094 0322 0323 0323 0323 0324 0324 0324 0325 0325 0326 0-095 0326 0326 0327 0327 0327 0328 0328 0328 0329 0329 0-096 0329 0330 0330 0330 0331 0331 0331 0332 0332 0332 0-097 0333 0333 0333 0334 0334 0334 0335 0335 0335 0336 0-098 0336 0336 0337 0337 0338 0338 0338 0339 0339 0339 0-099 0340 0340 0340 0341 0341 0341 0342 0342 0342 0343 0-100 0343 0343 0344 0344 0344 0345 0345 0345 0346 0346 0-101 0346 0347 0347 0347 0348 0348 0348 0349 0349 0350 0-102 0350 0350 0351 0351 0351 0352 0352 0352 0353 0353 0-103 0353 0354 0354 0354 0355 0355 0355 0356 0356 0356 0-104 0357 0357 0357 0358 0358 0358 0359 0359 0359 0360 0-105 0360 0360 0361 0361 0362 0362 0362 0363 0363 0363 0-106 0364 0364 0364 0365 0365 0365 0366 0366 0366 0367 0-107 0367 0367 0368 0368 0368 0369 0369 0369 0370 0370 0-108 0370 0371 0371 0371 0372 0372 0372 0373 0373 0374 0-109 0374 0374 0375 0375 0375 0376 0376 0376 0377 0377 0-110 0377 0378 0378 0378 0379 0379 0379 0380 0380 0380 0-111 0381 0381 0381 0382 0382 0382 0383 0383 0383 0384 0-112 0384 0385 0385 0385 0386 0386 0386 0387 0387 0387 0-113 0388 0388 0388 0389 0389 0389 0390 0390 0390 0391 0-114 0391 0391 0392 0392 0392 0393 0393 0393 0394 0394 0-115 0394 0395 0395 0395 0396 0396 0397 0397 0398 0398 0-116 0398 0398 0399 0399 0399 0400 0400 0400 0401 0401 0-117 0401 0402 0402 0402 0403 0403 0403 0404 0404 0404 0-118 0405 0405 0405 0406 0406 0406 0407 0407 0407 0408 0-119 0408 0409 0409 0409 0410 0410 0410 0411 0411 0411 0-120 0412 0412 0412 0413 0413 0413 0414 0414 0414 0415 0-121 0415 0415 0416 0416 0416 0417 0417 0417 0418 0418 0-122 0418 0419 0419 0419 0420 0420 0421 0421 0421 0422 0-123 0422 0422 0423 0423 0423 0424 0424 0424 0425 0425 0-124 0425 0426 0426 0426 0427 0427 0427 0428 0428 0428 0-125 0429 0429 0429 0430 0430 0430 0431 0431 0431 0432 0-126 0432 0433 0433 0433 0434 0434 0434 0435 0435 0435 0-127 0436 0436 0436 0437 0437 0437 0438 0438 0438 0439 0-128 0439 0439 0440 0440 i 0440 0441 0441 0441 0442 0442 0-129 0442 0443 0443 0443 0444 0444 0445 0445 0445 0446 APPENDIX. Table XCII. BaSO 4 to SO 3 . continued. 723 Grm. BaS0 4 . Grm. S0 8 . 1 2 3 4 5 6 7 8 9 0-130 0446 0446 0447 0447 0447 0448 0448 0448 0449 0449 0-131 0449 0450 0450 0450 0451 0451 0451 0452 0452 0452 0-132 0453 0453 0453 0454 0454 0454 0455 0455 0456 0456 0-133 0456 0457 0457 0457 0458 0458 0458 0459 0459 0459 0-134 0460 0460 0460 0461 0461 0461 0462 0462 0462 0463 0-135 0463 0463 0464 0464 0464 0465 0465 0465 0466 0466 0-136 0466 0467 0467 0468 0468 0468 0469 0469 0469 0470 0-137 0470 0470 0471 0471 0471 0472 0472 0472 0473 0473 0-138 0473 0474 0474 0474 0475 0475 0475 0476 0476 0476 0-139 0477 0477 0477 0478 0478 0478 0479 0479 0480 0480 0-140 0480 0481 0481 0481 0482 0482 0482 0483 0483 0483 0-141 0484 0484 0484 0485 0485 0485 0486 0486 0486 0487 0-142 0487 0487 0488 0488 0488 0489 0489 0489 0490 0490 0-143 0490 0491 0491 0492 0492 0492 0493 0493 0493 0494 0-144 0494 0494 0495 0495 0495 0496 0496 0496 0497 0497 0-145 0497 0498 0498 0498 0499 0499 0499 0500 0500 0500 0-146 0501 0501 0501 0502 0502 0502 0503 0503 0504 0504 0-147 0504 0505 0505 ; 0505 0506 0506 0506 0507 0507 0507 0-148 0508 0508 0508 0509 0509 0509 0510 0510 0510 0511 0-149 0511 0511 0512 0512 0512 0513 0513 0513 0514 0514 0-150 0515 0515 0515 0516 0516 0516 0517 0517 0517 0518 0-151 0518 0518 0519 0519 0519 0520 0520 0520 0521 0521 0-152 0521 0522 0522 0522 0523 0523 0523 0524 0524 0524 0-153 0525 0525 0525 0526 0526 0527 0527 0527 0528 0528 0-154 0528 0529 0529 0529 0530 0530 0530 0531 0531 0531 0-155 0532 0532 0532 0533 0533 0533 0534 0534 0534 0535 0-156 0535 ! 0535 0536 0536 0536 0537 0537 0537 0538 0538 0-157 0539 0539 0539 0540 0540 0540 0541 0541 0541 0542 0-158 0542 ! 0542 0543 0543 0543 0544 0544 0544 0545 0545 0-159 0546 0546 0546 0547 0547 0548 0548 0548 0549 . 0549 0-160 0549 0550 0550 0550 0551 0551 0551 0552 0552 0552 0-161 0553 0553 0553 0554 0554 0554 0555 0555 0555 0556 0-162 0556 0556 0557 0557 0557 0558 0558 0558 0559 0559 0-163 0559 0560 0560 0561 0561 0561 0562 0562 0562 0563 0-164 0563 0563 0564 0564 0564 0565 0565 0565 0566 0566 0-165 0566 0567 0567 0567 0568 0568 0568 0569 0569 0569 0-166 0570 0570 0570 0571 0571 0572 0572 0572 0573 0573 0-167 0573 0574 0574 0574 0575 0575 0575 0576 0576 0576 0-168 0577 ! 0577 0577 0578 0578 0578 0579 0579 0579 0580 0-169 0580 0580 0581 0581 0581 0582 0582 0583 0583 0583 0-170 0584 0584 0584 0585 0585 0585 0586 0586 0586 0587 0-171 0587 0587 0588 0588 0588 0589 0589 0589 0590 0590 0-172 0590 0591 0591 0591 0592 0592 0592 0593 0593 0593 0-173 0594 0594 0595 0595 0595 0596 0596 0596 0597 0597 0-174 0597 0598 0598 0598 0599 0599 0599 0600 0600 0600 724 A TREATISE ON CHEMICAL ANALYSIS. Table XCIL BaSO 4 to SO* continued. Grm. Grm. S0 3 . BaS0 4 . 1 2 3 4 5 6 7 8 9 0-175 0601 0601 0601 0602 0602 0602 0603 0603 0603 0604 0-176 0604 0604 0605 0605 0605 0606 0606 0607 0607 0607 0-177 0608 0608 0608 0609 0609 0609 0610 0610 0610 0611 0-178 0611 0611 0612 0612 0612 0613 0613 0613 0614 0614 0-179 0614 0615 0615 0615 0616 0616 0616 0617 0617 0617 0-180 0618 0618 0619 0619 0619 0620 0620 0620 0621 0621 0-181 0621 0622 0622 0622 0623 0623 0623 0624 0624 [ 0624 0-182 0625 0625 0625 0626 0626 0626 0627 0627 0627 0628 0-183 0628 0628 0629 0629 0629 0630 0630 0631 0631 0631 0-184 0632 0632 0632 0633 0633 0633 0634 0634 0634 0635 0-185 0635 0635 0636 0636 0636 0637 0637 0637 0638 0638 0-186 0638 0639 0639 0639 0640 0640 0640 0641 0641 0641 0-187 0642 0642 0643 0643 0643 0644 0644 0644 0645 0645 '0-188 0645 0646 0646 0646 0647 0647 0647 0648 0648 0648 0-189 0649 0649 0649 0650 0650 0650 0651 0651 0651 0652 0-190 0652 0652 0653 0653 0654 0654 0654 0655 0655 0655 0-191 0656 0656 0656 0657 0657 0657 0658 0658 0658 0659 0-192 0659 0659 0660 0660 0660 0660 0661 0661 0661 0662 0-193 0662 0662 0663 0663 0663 0664 0664 0664 0665 0665 0-194 0665 0666 0666 0666 0667 0667 0667 0668 0668 0669 0-195 0669 0669 0670 0670 0670 0671 0671 0671 0672 0672 0-196 0672 0673 0673 0673 0674 0674 0674 0675 0675 0675 0-197 0676 0676 0676 0677 0677 0677 0678 0678 0678 0679 0-198 0879 0679 0680 0680 0681 0681 0681 0682 0682 0682 0-199 0683 0683 0683 0684 0684 0684 0684 0685 0685 0685 0-200 0686 0686 0686 0687 0687 0687 0688 0688 0688 0689 0-201 0689 0689 0690 0690 0690 0691 0691 0691 0692 0692 0-202 0692 0693 0693 0693 0694 0694 0695 0695 0695 0696 0-203 0696 0696 0697 0697 0697 0698 0698 0698 0699 0699 0-204 0699 0700 0700 0700 0701 0701 0701 0702 0702 0702 0-205 0703 0703 0703 0704 0704 0704 0705 0705 0705 0706 0-206 0706 0707 0707 0707 0708 0708 0708 0709 0709 0709 0-207 0710 0710 0710 0711 0711 0711 0712 0712 0712 0713 0-208 0713 0713 0714 0714 0714 0715 0715 0715 0716 0716 0-209 0716 0717 0717 0717 0718 0718 0719 0719 0719 0720 0-210 0720 0720 0721 0721 0721 0722 0722 0722 0723 0723 0-211 0723 0724 0724 0724 0725 0725 0725 0726 0726 0726 0-212 0727 0727 0727 0728 0728 0728 0729 0729 0730 0730 0-213 0730 0731 0731 0731 0732 0732 0732 0733 0733 0733 0-214 0734 0734 0734 0735 0735 0735 0736 0736 0736 0737 0-215 0737 0737 0738 0738 0738 0739 0739 0739 0740 0740 0-216 0740 0741 0741 0742 0742 0742 0743 0743 0743 0744 0-217 0744 0744 0745 0745 0745 0746 0746 0746 0747 0747 0-218 0747 0748 0748 0748 0749 0749 0749 0750 0750 0750 0-219 0751 0751 0751 0752 0752 0752 0753 0753 0754 0754 APPENDIX. 725 Table XCIL BaSO 4 to SO* continued. Grm. BaS0 4 . Grm. S0 8 . 1 2 3 4 5 6 7 8 9 0-220 0-221 0754 0758 0755 0758 0755 0758 0755 0759 0756 0759 0756 0759 0756 0760 0757 0760 0757 0760 0757 0761 0-222 0761 0761 0762 0762 0762 0763 0763 0763 0764 0764 0-223 0764 0765 0765 0766 0766 0766 0767 0767 0767 0768 0-224 0768 0768 0769 0769 0769 0770 0770 0770 0771 0771 0-225 0771 0772 0772 0772 0773 0773 0773 0774 0774 0774 0-226 0775 0775 0775 0776 0776 0776 0777 0777 0778 0778 0-227 0778 0779 0779 0779 0780 0780 0780 0781 0781 0781 0-228 0782 0782 0782 0783 0783 0783 0784 0784 0784 0785 0-229 0785 0785 0786 0786 0786 0787 0787 0787 0788 0788 0-230 0788 0789 0789 0789 0790 0790 0790 0791 0791 0791 0-231 0792 0792 0792 0793 0793 0793 0794 0794 0794 0795 0-232 0795 0795 0796 0796 0796 0797 0797 0797 0798 ! 0798 0-233 0798 0799 0799 0799 0800 0800 0800 0801 0801 0801 0-234 0802 0802 0802 0803 0803 0803 0804 0804 0804 0805 0-235 0805 0805 0806 0806 0806 0807 0807 0807 0808 0808 0-236 0808 0809 0809 0810 0810 0810 0811 0811 0811 0812 0-237 0812 0812 0813 0813 0813 0814 0814 0814 0815 0815 0-238 0815 0816 0816 0816 0817 0817 0817 0818 0818 0818 0-239 0819 0819 0819 0820 0820 0820 0821 0821 0822 0822 . 0-240 0822 0823 0823 0823 0824 0824 0824 0825 0825 0825 0-241 0826 0826 0826 0827 0827 0827 0828 0828 0828 0829 0-242 0829 0829 0830 0830 0830 0831 0831 0831 0832 0832 0-243 0832 0833 0833 0834 0834 0834 0835 0835 0835 0836 0-244 0836 0836 0837 0837 0837 0838 0838 0838 0839 0839 0-245 0839 0840 0840 0840 0841 0841 0841 0842 0842 0842 0-246 0843 0843 0843 0844 0844 0844 0845 0845 0846 0846 0-247 0846 0847 0847 0847 0848 0848 0848 0849 0849 0849 0-248 0850 0850 0850 0851 0851 0851 0852 0852 0853 0853 0-249 0853 0853 0854 0854 0854 0855 0855 0855 0856 0856 0-250 0857 0857 0857 0858 0858 0858 0859 0859 0859 0860 0-251 0860 0860 0861 0861 0861 0862 0862 0862 0863 0863 0-252 0863 0864 0864 0864 0865 0865 0865 0866 0866 0866 0-253 0867 0867 0867 0868 0868 0868 0869 0869 0869 0870 0-254 0870 0870 0871 0871 0871 0872 0872 0872 0873 0873 0255 0874 0874 0874 0875 0875 0875 0876 0876 0876 0877 0-256 0877 0877 0878 0878 0878 0879 0879 0879 0880 0880 0-257 0881 0881 0881 0882 0882 0882 0883 0883 0883 0884 0-258 0884 0884 0885 0885 0885 0886 0886 0886 0887 0887 0-259 0887 0888 0888 0888 0889 0889 0889 0890 0890 0890 0-260 0891 0891 0891 0892 0892 0893 0893 0893 0894 0894 0-261 0894 0895 0895 0895 0896 0896 0896 0897 0897 0897 0-262 0898 0898 0898 0899 0899 0899 0900 0900 0900 0901 0-263 0901 0901 0902 0902 0902 0903 0903 0903 0904 0904 0-264 0905 0905 0905 0906 0906 0906 0907 0907 0907 0908 7 26 A TREATISE ON CHEMICAL ANALYSIS. Table XCIL BaSO 4 to SO 3 . continued. Grm. Grm. S0 3 . BaS0 4 . 1 2 3 4 5 6 7 8 9 0-265 0908 0908 0909 0909 0909 0910 0910 0910 0911 0911 0-266 0911 0912 0912 0912 0913 0913 0913 0914 0914 0914 0-267 0915 0915 0915 0916 0916 0917 0917 0917 0918 0918 0-268 0918 0919 0919 0919 0920 0920 0920 0921 i 0921 0921 0-269 0922 0922 0922 0923 0923 0923 0924 0924 0924 0925 0-270 0925 0925 0926 0926 0926 0927 0927 0928 0928 0928 0-271 0929 0929 0929 0930 0930 0930 0931 0931 0931 0932 0-272 0932 0932 0933 0933 0933 0934 0934 0934 0935 0935 0-273 0935 0936 0936 0936 0937 0937 0937 0938 0938 0938 0-274 0939 0939 0940 0940 0940 0941 0941 0941 0942 0942 0-275 0942 0943 0943 0943 0944 0944 0944 0945 0945 0945 0-276 0946 0946 0946 0947 0947 0947 0948 0948 0948 0949 0-277 0949 0949 0950 0950 0950 0951 0951 0952 0952 0952 0-278 0953 0953 0953 0954 0954 0954 0955 0955 0955 0956 0-279 0956 0956 0957 0957 0957 0958 0958 0958 0959 0959 0-280 0960 0960 0961 0961 0961 0962 0962 0962 0963 0963 0-281 0963 0964 0964 0964 0965 0965 0965 0966 0966 0966 0-282 0967 0967 0967 0968 0968 0968 0969 0969 0969 0970 0-283 0970 0970 0971 0971 0971 0972 0972 0972 0973 0973 . 0-284 0973 0974 0974 0974 0975 0975 0975 0976 0976 0977 0-285 0977 0977 0978 0978 0978 0979 0979 0979 0980 0980 0-286 0980 0981 0981 0981 0982 0982 0982 0983 0983 0983 0-287 0984 0984 0984 0985 0985 0985 0986 0986 0986 0987 0-288 0987 0987 0988 0988 0988 0989 0989 0990 0990 0990 0-289 0991 0991 0991 0992 0992 0992 0993 0993 0993 0994 0-290 0994 0994 0995 0995 0995 0996 0996 0997 0998 0998 0-291 0998 0999 0999 0999 1000 1000 1000 1001 1001 1001 0-292 1002 1002 1002 1003 1003 1003 1004 1004 1004 1005 0-293 1005 1005 1006 1006 1006 1007 1007 1007 1008 1008 0-294 1008 1009 1009 1009 1010 1010 1010 1011 1011 1012 0-295 1012 1012 1013 1013 1013 1014 . 1014 1014 1015 1015 0-296 1015 1016 1016 1016 1017 1017 1017 1018 1018 1018 0-297 1019 1019 1019 1020 1020 1020 1021 1021 1021 1022 0-298 1022 1022 1023 1023 1024 1024 1024 1025 1025 1025 0-299 1026 1026 1026 1027 1 1027 1027 1028 1028 1028 1029 APPENDIX. 727 Table XCIII. Conversion of Grms. of Lead Sulphate into Grms. of Lead Monoxide. (FACTOK 0-736.) Grm. PbO. Grm. PbS0 4 . 1 2 3 4 5 6 7 8 9 0-00 0007 0015 0022 0029 0037 0044 0051 0059 0066 0-01 0073 0081 0088 0096 0103 0110 0118 0125 0132 0140 0-02 0147 0154 0162 0169 0177 0184 0191 0199 0206 0213 0-03 0221 0228 0235 0243 0250 0258 0265 0272 0280 0287 0-04 0294 | 0302 0309 0316 0324 0331 0339 0346 0353 0361 0-05 0368 0375 0383 0390 0397 0405 0412 0419 0427 0434 0-06 0442 0449 0456 0464 0471 0478 0486 0493 0500 0508 0-07 0515 0522 0530 0537 0544 0552 0559 0567 0574 0581 0-08 0589 0596 0603 0611 0618 0626 0633 0640 0648 0655 0-09 0662 0670 0677 0684 0692 0699 0706 0714 0721 0729 0-10 0736 0743 0751 0758 0765 0772 0780 0787 0795 0802 0-11 0810 0817 0824 0832 0839 0846 0854 0861 0868 0876 0-12 0883 0891 0898 0905 0913 0920 0927 0935 0942 0949 0-13 0957 0964 0971 0979 0986 0994 1001 1008 1016 1023 0-14 1030 1038 1045 1052 1060 1067 1074 1082 1089 1097 0-15 1104 1111 1119 1126 1133 1141 1148 1155 1163 1170 0-16 1178 1185 1192 1200 1207 1214 1222 1229 1236 1244 0-17 1251 1258 1266 1273 1281 1288 1295 1303 1310 1317 0-18 1324 1332 1339 1347 1354 1362 1369 1376 1384 1391 0-19 1398 1406 1413 1420 1428 1435 1442 1450 1457 1465 0-20 1472 1479 1487 1494 1501 1509 ! 1516 1523 1531 1538 0-21 1546 1553 1560 1568 1575 1582 1590 1597 1604 1612 022 1619 1626 1634 1641 1649 1656 1663 1671 1678 1685 0-23 1693 1700 1707 1715 1722 1730 1737 1744 1752 1759 0-24 1766 1774 1781 1788 1796 1803 1810 1818 1825 1833 0-25 1840 1847 1855 1862 1869 1877 1884 1891 1899 1906 0-26 1914 1921 1928 1936 1943 1950 1958 1965 1972 1980 0-27 1987 1994 2002 2009 2017 2024 2031 2039 2046 2053 0-28 2061 2068 2075 2083 2090 2098 2105 2112 2120 2127 0-29 2134 2142 2149 2156 2164 2171 2178 2186 2193 2201 0-30 2208 2215 2223 2230 2237 2245 2252 2259 2267 2274 0-31 2282 2289 2296 2304 2311 2318 2326 2333 2340 2348 0-32 2355 2362 2370 2377 2385 2392 2399 2407 2414 2421 0-33 2429 2436 2443 2451 2458 2466 2473 2480 2488 2495 0-34 2502 2510 2517 2524 2532 2539 2546 2554 2561 2569 0-35 2576 2583 2591 2598 2605 2613 2620 2627 2635 2642 0-36 0-37 2650 2723 2657 2730 2664 2738 2672 2745 2679 2753 2686 2760 2694 2767 2701 2775 2708 2782 2716 2789 0-38 0-39 2797 2870 2804 2878 2811 2885 2819 ! 2826 2892 2900 2834 2907 2841 2914 2848 2922 2856 2929 2863 2937 728 A TREATISE ON CHEMICAL ANALYSIS. Table XCIIL PbSO 4 to PbO. continued. Grm. Grm. PbO. PbS0 4 . 1 2 3 4 5 6 7 8 9 040 2944 2951 2959 2966 2973 2981 2988 2995 3002 3010 0-41 3017 3025 3032 3040 3047 3054 3062 3069 3076 3084 042 3091 3098 3106 3113 3121 3128 3135 3143 3150 3157 043 3165 3172 3179 3187 3194 3202 3209 3216 3224 3231 044 3238 3246 3253 3260 3268 3275 | 3282 3290 3297 3304 045 3312 3319 3327 3334 3341 3349 3356 3363 3371 3378 046 3386 3393 3400 3408 3415 3422 3430 3437 3444 3452 047 3459 3466 3474 3481 3489 3496 3503 3511 3518 3525 048 3533 3540 3547 3555 3562 3570 3577 3584 3592 3599 049 3606 3614 3621 3628 3635 3643 3650 3658 3665 3673 0-50 3680 3687 3695 3702 3709 3717 3724 3731 3739 3746 0-51 3754 3761 3768 3776 3783 3790 3798 3805 3812 3820 0-52 3827 3834 3842 3849 3857 3864 3871 3879 3886 3893 0-53 3901 3908 3915 3923 3930 3938 3945 3952 3960 3967 0-54 3974 3982 3989 3996 4004 4011 4018 4026 4033 4041 0-55 4048 4055 4062 4070 4077 4085 4092 4099 4107 4114 0-56 4122 4129 4136 4144 4151 4158 4166 4173 4180 4188 0-57 4195 4202 4210 4217 4225 4232 4239 4247 4254 4261 0-58 4269 4276 4283 4291 4298 4306 4313 4320 4328 4335 0-59 4342 4350 4357 4364 4372 4379 4386 4394 4401 4409 0-60 4416 4423 4431 4438 4445 4453 4460 4467 4475 4482 0-61 4490 4497 4504 4512 4519 4526 4534 4541 4548 4556 0-62 4563 4570 4578 4585 4593 4600 4607 4615 4622 4629 0-63 4637 4644 4651 4659 4666 4674 4681 4688 4696 4703 0-64 4710 ! 4718 4725 4732 4740 4747 *4754 4762 4769 4777 0-65 4784 4791 4799 4806 4813 4821 4828 4835 4843 4850 0-66 4858 4865 4872 4880 4887 4894 4902 4909 4916 4924 0-67 4931 4938 4946 4953 4961 4968 4975 4983 4990 4997 0-68 5005 5012 5019 5027 5034 5042 5049 5056 5064 5071 0-69 5078 5086 5093 5100 5108 5115 5122 5130 5137 5145 0-70 5152 5159 5167 5174 5181 5189 5196 5203 5211 5218 0-71 5226 5233 5240 5248 5255 5262 5270 5277 5284 5292 0-72 5299 5306 5314 5321 5329 5336 5343 5351 5358 5365 0-73 5373 5380 5387 5394 5402 5410 5417 5424 5432 5439 0-74 5446 5454 5461 5468 5476 5483 5490 5498 5505 5513 0-75 5520 5527 5535 5542 5549 5557 5564 5571 5579 5586 0-76 5594 5601 5608 5616 5623 5630 5638 5645 5652 5660 0-77 5667 5674 5682 5689 5697 5704 5711 5719 5726 5733 0-78 5741 5748 5755 5763 5770 5778 5785 5792 5800 5807 0-79 5814 5822 5829 5836 5844 5851 5858 5866 5873 5881 0-80 5888 5895 5903 5910 5917 5925 5932 5939 5947 5954 0-81 5962 5969 5976 5984 5991 5998 6006 6013 6020 6028 0-82 6035 6042 6050 6057 6065 6072 6079 6087 6094 6101 083 6109 6116 6123 6131 6138 6146 6153 6160 6168 6175 0-84 6182 6190 6197 6204 6212 6219 6226 6234 6241 6249 0-85 6256 6263 6271 6278 6285 6293 6300 6307 6314 6322 0-86 6330 6337 6344 6351 6359 6366 6373 6381 6388 6396 0-87 6403 6410 6418 6425 6433 6440 6447 6455 6462 6469 0-88 6477 6484 6491 6499 6506 6514 6521 6528 6536 6543 0-89 6550 6558 6565 6572 6580 6587 6594 6602 6609 6617 APPENDIX. Table XCIII. - PbSO 4 to PbO. continued. 729 Grm. Grm. PbO. PbS0 4 . 1 2 3 4 5 6 7 8 9 0-90 6624 6631 6639 6646 6653 6661 6668 6675 6683 6690 0-91 6698 6705 6712 6720 6727 6734 6742 6749 6756 6764 0-92 6771 6778 6786 6793 6801 6808 6815 6822 6830 6837 0-93 6845 6852 6859 6867 6874 6882 6889 6896 6903 6911 0-94 6918 6926 6933 6940 6948 6955 6962 6970 6977 6985 095 6992 6999 7007 7014 7021 7029 7036 7043 7051 7058 0-96 7066 7073 7080 7088 7095 7102 7110 7117 7124 7132 0-97 7139 7146 7154 7161 7168 7176 7183 7191 7198 7205 0-98 7213 7220 7227 7235 7242 7250 7257 7264 7272 7279 0-99 7286 7294 7301 7308 7316 7323 7330 7338 7345 7353 1-00 7360 7367 7374 7382 7389 7396 7404 7411 7418 7426 1-01 7434 7441 7448 7455 7462 7470 7477 7484 7492 7499 1-02 7507 7514 7522 7529 7536 7543 7551 7558 7566 7573 1-03 7581 7588 7595 7603 7610 7618 7625 7632 7640 7647 1-04 7654 7662 7669 7677 7684 7691 7698 7706 7713 7720 1-05 7728 7735 7742 7749 7757 7764 7771 7779 7786 7794 1-06 7802 7809 7816 7823 7831 7838 7846 7853 7860 7867 1-07 7875 7882 7890 7897 7905 7912 7919 7927 7934 7941 1-08 7949 7956 7964 7971 7979 7986 7993 8001 8008 8015 1-09 8022 8030 8038 8045 8052 8060 8067 8074 8082 8089 1-10 8096 8103 8111 8118 8126 8133 8140 8147 8155 8162 1-11 8170 8177 8184 8191 8199 8207 8214 8221 8228 8236 1-12 8243 8251 8258 8265 8272 8280 8287 8294 8302 8310 1-13 8317 8324 8331 8339 8347 8355 8362 8369 8376 8383 1-14 8390 8397 8405 8413 8420 8428 8435 8442 8450 8457 1-15 8464 8471 8479 8486 8493 8501 8508 8515 8523 8530 1-16 8538 8545 8552 8560 8567 8574 8582 8589 8596 8604 1-17 8611 8619 8626 8633 8640 8648 8655 8662 8670 8677 1-18 8685 8692 8699 8706 8714 8721 8729 8737 8744 8751 1-19 8758 8766 8773 8781 8788 8796 8803 8810 8818 8825 1-20 8832 8840 8847 8854 8861 8869 8876 8883 8891 8898 1-21 8906 8913 8920 8927 8935 8942 8950 8957 8964 8972 1-22 8979 8987 8994 9001 9009 9016 9023 9031 9038 9045 1-23 9053 9060 9067 9075 9082 9090 9097 9104 9111 9119 1-24 9126 9134 9141 9149 9156 9164 9171 9178 9186 9193 1-25 1-26 9200 9274 9208 9281 9215 9288 9223 9295 9230 9303 9237 9310 9244 9318 9252 9325 9259 9332 9266 9340 1-27 9347 9354 9361 9369 9376 9384 9391 9398 9406 9413 1-28 1-29 9421 9494 9428 9501 9435 9509 9443 9516 9450 9524 9457 9531 9465 9538 9472 9546 9480 9553 9487 9560 1-30 1-31 1-32 1-33 1-34 9568 9642 9715 9789 9862 9575 9649 9722 9796 9870 9583 9657 9730 9803 9877 9590 9664 9737 9810 9884 9598 9671 9744 9818 9892 9605 9678 9751 9825 9899 9612 9686 9759 9833 9906 9620 9693 9767 9840 9913 9627 9700 9774 9847 9920 9634 9707 9781 9854 9928 1-35 9936 9943 9950 9957 9965 9972 9980 9988 9995 1-0001 730 A TREATISE ON CHEMICAL ANALYSIS. Table XCIV. Solvents for Precipitates in Munroe's Crucible. O. D. Swett, Journ. Amer. Chem. Soc., 31. 928, 1909. (Salts are applied in aqueous solution unless otherwise specified. ) 1. Water. 22. 2. Alcohol. 23. 3. Carbon disulphide. 24. 4. Sulphuric acid with nitric acid or 25. nitrates. 26. 5. Sulphuric acid, concentrated or fuming, 27. with ammonium chloride. 28. 6. Nitric acid. 29. 7. Carbon dioxide in aqueous solution. 30. 8. Acetic acid. 31. 9. Oxalic acid. 32. 10. Hydrochloric acid with ammonium 33. chloride or oxalic acid. 34. 11. Hydrofluoric acid. 35. 12. Potassium hydroxide. 36. 13. Sodium hydroxide. 37. 14. Ammonium hydroxide. 15. Potassium sulphide. 16. Potassium sulphide, yellow. 40. 17. Sodium sulphide. 18. Sodium sulphide, yellow. 42. 19. Ammonium sulphide. 43. 20. Ammonium sulphide, yellow. 44. 21. Potassium bisulphite. 45. Sodium thiosulphate. Ammonium sulphate. Ammonium nitrate. Sodium hydrogen phosphate. Ammonium oxalate. Ammonium acetate, alkaline. Ammonium tartrate, alkaline. Potassium carbonate. Sodium carbonate. Potassium chlorate. Ammonium carbonate. Potassium chloride. Potassium iodide. Sodium chloride. Ammonium chloride. Ammonium fluoride, dry. Calcium chloride. Magnesium chloride. Potassium cyanide. Ferrous sulphate. Silver nitrate. Lead acetate. Mercuric nitrate. Ferric acetate. Solvents for Precipitates in Condition for Weighing. (Numbers indicate correspondingly numbered solvents in the preceding list. Hyphens indicate successive treatments ; commas indicate alternative treatments. Abbreviations: h hot, c concentrated, d dilute.) Precipitates. Solvents Precipitates. Solvents. Aluminium oxide . 12ch-l-10, 13ch-l-10 Barium chromate . 6, 10 Ammonium arseno- 14. Barium silicofluoride 36 molybdate. Barium sulphate 4h, 5h Ammonium chloride Ih. Bismuth 6d Ammonium magnes- 6, 10. Bismuth carbonate 6, 10 ium arsenate. (basic). Ammonium phos- Ih, 12, 13, 25, 26, 29, Bismuth chromate 6 phomolybdate. 30,36 (basic). Ammonium chloro- Ih Bismuth nitrate 6 platinate. (basic). Antimony 6-1-10 Bismuth oxide 10 Antimony pentasul- 14h Bismuth oxychloride lOc phide. Bismuth sulphide . 6ch-l-3 Antimony tetroxide lOch Cadmium carbonate 6d, 14 Antimony trisul- lOc, 12d, 15, 19 Cadmium oxide 4, 5, 6, 10, 14 phide. Cadmium sulphide . 4dh, 6, 10 Arsenious sulphide . 12, 13, 15, 17, 21, 29, Calcium carbonate . 6d, lOd 30 Calcium carbonate, Ih, 6d, lOd Barium carbonate . 7, 24, 36 ignited. Barium carbonate, Ih, 6d Calcium fluoride 4-1, lOc ignited. Calcium oxalate 6,22 APPENDIX. 731 Table XCIV. continued. Precipitates. Solvents. Precipitates. Solvents. Calcium sulphate . 23dh Mercuric oxide 6, 10 Chromium oxide . 6ch + 31dry + 42 or Mercuric sulphide . 15 + 12 43, 31dry h+1, Mercurous chloride . 6ch-l 42dryh + 6 Mercurous chromate 6ch Cobalt . 6 Mercurous phosphate 6 Cobalt hydroxide . j 23, 24, 36 Metastannic acid 10, 13 Cobaltous sulphate . ; Ih Nickel . 6c Cobaltous sulphide . 4c, 5, 8c, lOc Nickel oxide . 10 Cupric hydroxide . ! 6, 10, 14 Nickelous hydroxide lOd, 14, 23, 24, 36 Cupric sulphide . 6h, 40 Nickelous subsul- 6 Cuprous oxide . 4 phide. Cuprous sulphide 6 Nickelous sulphate . Ih Cuprous thiocyan- 4 Nickelous sulphide . 6 ate. Palladium iodide 14 Gold 4c Platinum sulphide . 16, 18 Gold sulphides 16,40 Potassium chloride . Ih Iron acetate (basic) . | 10 Potassium cobaltic Ih Iron arsenate . . 6, 10 nitrite. Iron formate (basic) 10 Potassium fluoborate 1, 2d Iron hydroxide . 6d, lOd. Potassium chloro- Ih, 12h Iron oxide . . lOc platinate. Iron phosphate . 10, 45 Potassium sulphate Ih Iron succinate (basic) 4, 5, 6, 10 Silica 11, 12h, 13h, 29h, Iron sulphide . . 4, 10 30h, 37 -ignition Lead arsenate 6 Silver 6d Lead carbonate 6d Silver chloride 14h Lead chloride . Ih Silver cyanide . 14,40 Lead chromate 6, 12 Silver iodide . 34c, 22, 40, 44ch Lead oxalate . 6d Silver phosphate 6, 14 Lead oxide 6d Silver sulphide 6ch-l-3 Lead phosphate . 6 Sodium carbonate . Ih Lead sulphate . 24, 27, 28, lOch, 6ch, Sodium chloride Ih 22, 12h, 13h, 14h Sodium chloroplatin- Ih, 2 Lead sulphide 6c, lOc ate. Magnesium oxide . 4d, 6d, lOd Sodium sulphate Ih Magnesium phos- 6, 10 Stannic acid . 10-1 o IT * phate. Stannic oxide 36dry -ignition, 10- 1 Magnesium pyro- j 6, 10 arsenate. Stannic phosphate . Stannic sulphide 12 lOch, 12, 15, 17 Magnesium pyro- 6, 8, 10, 6ch, lOch phosphate. (hydrous). Stannous sulphide lOc Magnesium sulphate Manganese dioxide . Manganese sesqui- oxide. Ih 5h + 9 44.9,4 + 41 (hydrous). Strontium carbonate Strontium sulphate Uranyl pyroarsenate 7, 24, 36 33c, 35c, 38c, 39c 6 Manganese sulphide Manganous pyro- phosphate. Manganous sulphate 4d, 6d, lOd 4ch, 5h, 6ch, lOch 1 Uranyl pyrophos- phate. Zinc oxide Zinc sulphide . 4, 6, 10 lOdh 4, 6, 10 THE LIBRARY. THE following books will be found useful on the shelves of the library attached to the laboratory. There are numerous other excellent works of general interest which have not been included in this list. MINERAL AND ROCK ANALYSIS. W. F. HILLEBRAND, The Analysis of Silicate and Carbonate Rocks, Washington, 1910. No analyst who has to deal with silicates, and certainly no writer on the subject, can fail to acknowledge his indebtedness to this American chemist. The book describes the methods used by the chemists of the U.S. Geological Survey. The first edition: Some Principles and Methods of Rock Analysis, Washington, 1900, has been the means of raising the standard of silicate analyses all round. H. S. WASHINGTON, Manual of the Chemical Analysis of Rocks, New York, 1904. This excellent American book is based on an earlier edition of Hillebrand's brochure, and it gives more elaborate details of manipulation and the precautions to be observed to secure accurate results by those not specially conversant with silicate analyses. M. DITTRICH, Anleitung zur Gesteinanalyse, Leipzig, 1905. This also is an excellent book from the German point of view, and it covers much the same ground as Washington's manual. GENERAL ANALYSIS. C. R. FRESENIUS, Anleitung zur quantitativen chemischen Analyse, Braunschweig, 1875-87. This is one of the classics, for it has played an important part in modern analytical chemistry. Although the latest edition is over twenty years old, there are few analytical laboratories without their ' ' Fresenius. " This work has been translated into many European languages. L. L. DB KONINCK, Lehrbuch der qualitativen und quantitativen chemischen Analyse, Berlin, 1904. In zwei Banden. This is a kind of modernised " Fresenius," and it is very useful when awkward separa- tions are being planned. Several alternative processes are usually given. Analysts are indebted to the Belgian chemist for many new ideas. A. CLASSEN, Ausgewahlte Methoden der analytischen Chemie, Braunschweig, 1901-3. In zwei Banden. This large two- volume German book describes the analytical methods for all the elements. It can scarcely be said to supplant; but rather does it supplement " Fresenius." This is usually the second book consulted after ' ' Fresenius " when a new separation has to be taken in hand. L. E. RIVOT, Docimasie. Traite $ Analyse des Substances Minerales, Paris, 1886. This five-volume French work is a small encyclopaedia on the subject, and is accordingly a reference book not so much used now as formerly. 732 THE LIBRARY. 733 A. CARNOT, Traite d' Analyse des Substances Minerales, Paris, 1898. In four volumes. The fourth volume of this bulky French work is not yet published. There are some good things in the three volumes which have appeared, although the great lapse of time between the first and later volumes has not tended to make the work attractive. It is a modernised "Rivot." F. PETERS, Handbuch der analytischen Chemie, Heidelberg, 1912. This promises to be a comprehensive work of fourteen volumes, to be published in parts. A start has been made with the volume to be devoted to Antimony, Arsenic, and Tin. It will probably be many years before the work approaches completion, but once completed', it will be as useful as the celebrated ' ' Gmelin " published by the same firm. The Germans alone appear to possess the great patience required to give the world a book such as is here contemplated. A. RUDISULE, Nachweis, Bestimmung und Trennung der chemischen Elemente, Bern, 1913. This promises to be a comprehensive work in eleven volumes. Vol. I., devoted to Arsenic, Antimony, Tin, Tellurium, and Selenium, has been published. It seems as if the work will clash with "Peters" previously mentioned. Either or both works, when completed, will be indispensable for reference. C. FRIEDHEIM, Leitfaden fur die quantitative chemische Analyse, unter Mitberuck- sichtigung von Massanalyse, Gasanalyse, und Elektrolyse, Berlin, 1905. This is the 1905 edition of the old classical Leitfaden of C. F. Rammelsberg (1845), and it describes many difficult separations not commonly found in other manuals. P. JANNASCH, Praktischer Leitfaden der Gewichtsanalyse, Leipzig, 1904. This German book contains many novel processes not found described in other text- books, and, like the preceding volume, it is useful when devising a plan for an unfamiliar separation. 'F. TREAD WELL, Kurzes Lehrbuch der analytischen Chemie, Leipzig, 1911. In zwei Banden. This well-known German book is equally familiar in its American dress. It is one of the best modern books for advanced exercises, and it is a very useful book of reference. The two volumes deal with both qualitative and quantitative analysis. G. CHESNEAU, Principes Theoriques et Pratiques d Analyse Minerale, Paris, 1912. This French book, occupying 641 pages, is a general treatise on quantitative inorganic analysis. It also contains some applications of physical chemistry to analysis very neatly expressed. There is no danger of the student's attention being distracted by references to authorities. The chapter on phosphorus is particularly good. G. LUNGE und E. BERL, Chemisch-technische Untersuchungsmethoden, Berlin, 1910. In vier Banden. Although it is generally considered that the day of comprehensive text-books is past, and that the specialist requires monographs written by experts in his own subject, this work may be regarded as a library of small monographs. Each subject is treated by a specialist in that particular theme. The cost appears high if the whole set be purchased, because the mineral chemist will find he has also bought monographs on subjects he never crosses, and conversely. An English translation is in course of publication. T. B. STILLMAN, Engineering Chemistry: A Manual of Quantitative Chemical Analysis for the use of Students, Chemists, and Engineers, Easton, Pa., 1910. This book treats of the analysis of coal, gases, metals, ores, water, oils, paints, and several commercial products. A. MEURICE, Cours d 1 Analyse Quantitative, Paris, 1908. This French book of 829 pages is mainly devoted to the analysis of technical products. Chapter III., for instance, deals with Cements, Mortars, Limestones, Dolomites, etc. ; and Chapter IV. with Refractory Products, Sands, and Clays. The succeeding chapters deal with minerals to be treated for copper, iron, cobalt, nickel, manganese, etc. 734 A TREATISE ON CHEMICAL ANALYSIS. QUALITATIVE ANALYSIS. C. R. FRESENIUS, Anleitung zur qualitativen chemischen Analyse, Braunschweig, 1895. This book has been through nearly twenty editions since its first publication at Bonn in 1841, and it has also been translated into all the important European languages. It is still a standard of reference. A. B. PRESCOTT and 0. C. JOHNSON, Qualitative Chemical Analysis, New York, 1910. A detailed study of the reactions of the different elements, and the elaboration of schemes for the treatment of complex mixtures. J. STIEGLITZ, The Elements of Qualitative Chemical Analysis, New York, 1911. Two vols. This work is largely occupied with descriptions of the physical chemistry of qualitative analysis. VOLUMETRIC ANALYSIS. H. BECKURTS, Die Methoden der Massanalyse, Braunschweig, 1910. This is the latest edition of C. F. Mohr's classical Lehrbuch der chemisch-analytischen Titriermethoden (1855), and it is very thorough. The Germans excel in books of reference of this kind. A. CLASSEN, Theorie und Praxis der Massanalyse, Leipzig, 1912. The unique feature of this book is the theoretical discussions which precede the directions for manipulation. Classen was the editor of the 1888 edition of Mohr's Lehrbuch. F. SUTTON, A Systematic Handbook of Volumetric Analysis, London, 1911. This volume has been through numerous editions since its publication in 1863, and it has proved a valuable work of reference when a volumetric process for a particular deter- mination is sought. E. KNECHT and E. HIBBERT, New Reduction Methods in Volumetric Analysis, London, 1910. This book describes some applications of titanous chloride in volumetric analysis. The titanous chloride process for titrating ferric iron directly is described on p. 49. SPECIAL TREATISES AND MONOGRAPHS. F. A. GOOCH, Methqds in Chemical Analysis, New York, 1912. Analysts generally are under many obligations to Prof. Gooch for new apparatus, new methods, and for the revision and improvement of many of the older processes. It is a great convenience to have Gooch's papers, abridged by himself, collected into one volume. W. CROOKES, Select Methods in Chemical Analysis (chiefly Inorganic), London, 1905. This book, compiled by the editor of the Chemical News, is strongest when consulted for the rarer elements. A. H. Low, Technical Methods of Ore Analysis, New York, 1906. This American book is useful when rapid technical processes are required. It special- ises chiefly in evaluating ores. H. BREARLEY and F. IBBOTSON, The Analysis of Steel- Works Materials, London, 1902. Although designed for metallurgical work, this book contains many useful hints on the separation of the different metals. A. A. BLAIR, The Chemical Analysis of Iron, Philadelphia, 1908. Like Brearley and Ibbotson's book, this is also a useful book to consult for unfamiliar separations, and it is a most popular book of its kind. THE LIBRARY. 735 H. BREARLEY, The Analytical Chemistry of Uranium, London, 1903. Brearley's monograph reviews the analytical chemistry of uranium, and it is a pity that more monographs of a similar kind are not available for other metals. R. J. MEYER and 0. HAUSER, Die Analyse der seltenen Erde und der Erdsduren, Stuttgart, 1902. This monograph treats of the analysis of minerals containing the rare earths, thorium, zirconium, titanium, niobium, and tantalum. L. MOSER, Die Bestimmungsmethoden des Wismuths, Stuttgart, 1909. The author has tried in this monograph to give as complete as possible a review of the analytical chemistry of bismuth. H. NISSENSON, Die Untersuchungsmethoden des Zinks, Stuttgart, 1907. The author of this monograph has collected those methods employed for the determina- tion of zinc, and more particularly those applicable to the technically important zinc ores. L. PARRY, The Assay of Tin and Antimony, London, 1906. The writer of this monograph is mainly concerned with the methods for the separation and analytical determination of tin and antimony working under industrial conditions against time. A. STABLER, Handbuch der Arbeitsmethoden in der anorganischen Chemie, Leipzig, 1913. Fiinf Bande. The first volume of this handbook deals with the mechanical operations, and the equipment of chemical laboratories. The other volumes have not yet appeared. ELECTRO- ANALYSIS. E. T. SMITH, Electro-analysis, Philadelphia, 1907. F. M. PERKIN, Practical Methods of Electrochemistry, London, 1905. A. FISCHER, Electroanalytische Schnellmethoden, Stuttgart, 1908. A. HOLLARD et L. BERTIAUX, Analyse des Metaux par Electrolyse, Paris, 1909. CEMENTS. F. R. VON ARLT, Laboratoriumsbuch fur die Zementindustrie, Halle a. S., 1910. F. B. GATEHOUSE, A Handbook for Cement Works Chemists, London, 1908. J. MALETTE, Analyse Chimique des Chaux et Ciments, Paris, 1911. E. LEDUC and G. CHENU, Chaux, Ciments, Pldtres, Paris, 1912. R. K. MEADE, Portland Cement : Its Composition, Raw Materials, Manufacture, Testing, and Analysis, Easton, Pa., 1906. ASSAYING. C. H. FULTON, A Manual of Fire Assaying, New York, 1907. C. and J. J. BERINGER, A Text-book of Assaying, London, 1908. P. DE P. RICKETTS and E. H. MILLER, Notes on Assaying, New York, 1908. C. H. ARON, Assaying, San Francisco, 1900. T. K. ROSE, The Metallurgy of Gold, London, 1906. Although the first eighteen chapters are concerned with the metallurgy of gold the nineteenth and twentieth chapters are devoted to the assay of gold ores and bulhon. is an expert on this subject. 736 A TREATISE ON CHEMICAL ANALYSIS. WATER. W. P. MASON, Examination of Water (Chemical and Bacteriological), New York, 1910. H. B. STOCKS, Water Analysis for Sanitary and Technical Purposes, London, 1912. AMERICAN PUBLIC HEALTH ASSOCIATION, Standard Methods for the Examination of Water and Sewage, New York, 1912. FUEL AND GAS. E. E. SOMERMEIER, Coal : its Composition, Analysis, Utilisation, and Valuation, New York, 1912. J. H. COSTE, The Calorific Power of Gas : A Treatise on Calorific Standards and Calorimetry, London, 1911. J. S. HALDANE, Methods of Air Analysis, London, 1912. W. HEMPEL, Gasanalytische Methode, Braunschweig, 1900 ; English edition, 1902. H. FRANZEN, Gasanalytische Uebungen. Ein Hilfsbuch fur das gasanalytische Praktikum, Leipzig, 1907. This booklet describes some exercises in the manipulation of gases. There is an English translation. THEORETICAL. W. OSTWALD, Die wissenschaftlichen Grundlagen der analytischen Chemie, Leipzig, 1910. This is an extremely interesting book on the theory of analytical chemistry, and, when Hrst published in 1894, it drew the serious attention of analysts to the significance and importance of the theory of their art. An English translation is available. G. CHESNEAU, Principes Theoriques des Methodes d'A nalyse Minerale, fondees sur les Reactions Chimiques, Paris, 1906. Like the preceding volume, this work treats of the theory, but whereas the former uses the language of the ionic hypothesis, Chesneau's book does not. An American translation is available. PERIODICALS. Zeitschrift fur analytische Chemie. H. and W. Fresenius and E. Hintz, Wies- baden ; 18s. per ann. The Analyst. Organ of the Society of Public Analysts, W. I. Sykes, London ; 21s. per ann. Annales de Chimie Analytique appliquee a I 'Industrie, a la Pharmacie et a la Biologic, C. Crinon, Paris ; 1 2s. per ann. INDEX OF NAMES. AARS, L. A., 449. Abegg, R., 46. Abel, A. A., 275. Abel, E., 289, 348. Abelspies, J. F. C., 270. Abesser, 0., 216, 593, 598, 600. Abich, H., 121, 162, 178. Abraham, K., 197. Acree, S. F., 615. Adams, L. H., 611, 619. Adams, M., 311. Adie, R. H., 540, 541. Adrian!, A., 319. Agulhan, H., 585. Albert!, R., 538. Alcock, F. H., 306. Alderskrou, H. B. von, 225. Aldis, W. G., 3. Alefeld, E., 652. Alessandri, P. E., 604. Alexander, H. H., 333. Alexandron, W., 197. Alex!, C.,442. Alfa, J., 649. Alhaire, P., 682. Alibigoff, G., 488. Allart, A., 310. Allen, A H., 46, 63, 160, 169, 234, 435, 579. Allen, E. T., 124, 155, 178, 182, 184, 435, 444, 612, 616. Allen, I. C., 621, 625. Allen, O. D., 239. Allen, W. S., 611. Allibon, G. H., 601. Allihn, F., 42, 100, 102, 110, 277, 587. Allison, A., 477. Aloy, J., 204. Alt, H., 364. Alvarez, E. P., 481. Amberg, R., 563. Anderch, H. C., 169. Andersen, A. C., 72, 193. Anderson, A. A., 533, 536. Anderson, H. 0., 1, 630. Anderson, R., 625. Anderson, W. C., 551. Andre, G., 624. Andrews, L. W., 35, 40, 78, 201, 286, 288, 399. Anelli, G. , 625. Angenot, H., 267, 410. 462, 434, 287, Antany, M., 319. Anthony, M., 387, 425. Antony, V., 339, 429. Apel, see Tietyens, L. Archbutt, L., 476, 611. Archetti, A., 526. Archibald, E. H., 231, 520. Archimedes, 22. Arendt, R., 600. Arfvedson, A., 578. Argall, P., 135. Arndt, R., 585. Arnold, C., 148. Arnold, H., 437. Arnold, R., 625. Aron, J., 531, 658. Arth, G., 177, 594. Arthur, W., 399. Ashley, H. E., 170, 460, 461. Ashoff, K., 624. Aston, 16. Atack, F. W., 415. Atkinson, R. W., 124, 167, 192, 376, 454,' 595 Atterberg, A., 232, 234. Attfield, J., 183, 305. Atwater, W. 0., 36. Auchy, G., 106, 378, 380, 480, 547, 550, 564, 568, 599. Auerbach, E. B. , 472. Aufrecht, S., 622. Auger, V.,492. Auraann, C., 237. Aupperle, J. A., 551, 560. Austin, A., 355. Austin, M., 106, 218, 284, 366, 372, 374. Austin, P. T., 88, 161, 183, 192, 207, 314. Autenrieth,W., 540. Avery, C. E., 90, 161, 293, 462. Avery, S., 195. Avery, S., and Beans, H. T., 290, 293. Azzarello, E., 280. BABBIT, H. C., 593, 594. Babo, T. von, 161. Bachinskii, A. , 5. Bachmeyer, W., 190. Backstrom, H., 501. Bagley, E., 408. Bahr, J. F., 504. Bailey, E.G., 134. Bailey, E. H. S., 529. 737 47 738 A TREATISE ON CHEMICAL ANALYSIS. Bailey, G. H., 224, 498, 499. Bailey, H. S., 100, 272. Bailey, see Spooner. Bailey, T., 429. Bain, J. W., 207, 209. Baker, B. F., 522. Baker, E. L., 541. Baker, J. T., 141. Ball, W. 0., 240. Ballhauser, F., 641. Balli, B., 60. Belthaser, K., 522. Bannister, C. 0., 333. Bansen, H., 138. Banthisch, W., 614. Barba, W. P., 104. Barfoed, C., 279, 296, 308. Barker, C. R., 214. Barker, P., 605. Barlow, W. E., 625. Barnebey, 0. L., 454. Barnes, S. K., 449. Barral, J. A., 136. Barreswil, L. C., 484. Barthe, L., 579, 580. Barthel, G , 617. Bartonec, H., 405. Baskerville, C., 203, 207. Bassadonna, M., 195. Basset, H. , 478. Batey, J. P., 289, 290, 291. Battandier, M., 102. Baubigny, M., 273,357, 360, 364, 373, 388, 389, 480, 517. Baud, E., 446. Baudisch, 0., 455. Bauer, E., 91. Bauer, K. L., 17. Bauer, 0., 312. Bauer, T. , 169, 175. Baumhauer, E. H. von, 386. Baur, R., 555. Baxter, G. P., 118, 179, 215, 452, 485, 593, 599. Bay, J., 625. Bayerlein, H., 475, 538. Bayley, T,, 94, 390, 396. Beans, H. T., see Avery, S. Beard, M., 455. Becaia, M., 595. Bechamp, A., 115, 428. Bechi, E., 585. Beck, K., 315. Beck, 0., 270, 284. Beck, P., 323, Becker, F., 4, 270. Becker, G. F., 23. Beckurts, H., 31, 280, 538. Bedford, M. H., 161. Beebe, A. L., 187. Beerman, G. H., 585. Behncke, W., 636. Beilstein, F., 114, 239, 358, 364. Bein, S., 641, 644. Belden, W., 499. Bell, C. A., 40. Bell, J. C., 250. Belluci, I., 456. Belohoubek, A., 491. Belton, F. G., 317. Bemmelen, J. M. von, 173, 308, 380, 641. Bemmelen, M. von, 624. Benas, T., 311. Bender, 0., 136. Benedict, F. G., 59, 117. Benedict, R., 169, 330. Benedict, S. R., 385, 386. Benedikt, R., 35, 653. Benelli, T., 319, 339. Benkert, A. L., 348. Benner, R. C., 20, 103, 113, 335. Bennet, J. F., 292. Benneville, J. S. de, 408. Benoit, 18. Benz, E., 501, 505, 506, 510. Berchem, 0., 56. Berdel, E., 663, 664, 667. Berg, P. von, 358, 364, 583. Bergdolt, A., 215. Berger, H. W., 114. Bergeron, M., 355. Berglund, E., 276. Bergmann, F., 395. Bergmann, J., 35. Bergmann, T., 163, 657. Beringer, A. , 362. Beringer C., and Beringer, J. J., 291, 312, 353, 633. Beringer, J. J. , see Beringer, C. Berju, G., 528. Berl, E., 551. Berlin, N. J., 501. Bernhardt, P., 107. Bernheim, R., 540. Bernoulli, F. A., 408. Berolzheimer, R., 622, 632. Berthelot, M., 195, 295, 624. Berthier, P., 162, 425, 499, 568. Berthollet, C. L., 116, 552. Bertiaux, L., 256, 280. Bertin, M., 287. Bertlet, E., 579. Bertrand, A., 237. Bertrand, G., 359, 382, 585. Bertrand, 0., 245 Berzelius, J. J., 3, 16, 23, 89, 160, 161, 162, 218, 226, 228, 278, 308, 408, 483, 499, 502, 508, 578, 583, 614, 637, 638. Besson, A. A., 91. Bettel, W., 425, 426, 429, 475. Bettendorff, A., 480. Bettendorf, 0., 483. Bettink, W. H., 397, 398. Betzel, R., 190. Beutell, A., 35, 59, 66. Bevan, E. J., 72, 546. Beyne, E., 360, 517. Beythein, A., 579. Bialobrzesky, M., 291. Bicard, 73. Bicks, T., 624. Bidauf, M., 576. Bidtel, E., 648, 649. INDEX OF NAMES. 739 Biggs, R., 197. Biltz, 47. Biltz, E., 117. Biltz, H., 455. Biltz, \V., 496. Binder, 0., 44, 138, 61-7. Binns, C. F.. 672. Birch, W. C., 188, 190, 452. Birnbaum, H., 156. Biscaro, G., 80. Bischof, C., 658, 663, 672. Bischof, G., 339, 355. Bischoff, M., 528. Bishop, H. B., 280, 611. Bislee, H., 212. Biilmann, E., 540. Black, J. W., 364. Blackman, P., 91. Blaese, C. von, 239. Blair, A. A., 92, 110, 113, 116, 191, 209, 382, 415, 456, 551, 590, 591, 599, 621. Blakeley, A. G., 318. Blalock, T. L., 36,40. Blanc, M. C., 72. Blank, R., 529. Blarez, C., 635. Blasdale, W. C., 213, 258, 521. Blattner, N., 607, Blau, G., 196. Bleeker, J. B., 124. Bley, H., 181. Bleyer, B., 449. Blochmann, R., 58, 146. Blomstrand, W., 390, 419, 501. Bloor. W. R., 170, 180, Blount, B., 165, 169, 175, 176, 250, 350, 521, 550. Bloxam, C. L., 270, 278. Blum, L., 102, 178, 218, 367, 368, 373, 374, 378, 547,607, 619, 621. Blum, W., 196, 382, 646. Blunt, J. P., 194. Blunt, T. P., 326, 355. Blyth, A. W., 585. Blyth, W. B., 135. Bobierre, A. , 287. Bodemann, T., 396. Bodenbender, A., 528. Bodewig, C., 184, 462, 580, 583. Bodlander, G., 381. Bodmann, G., 505. Boeck, P. A., 630. Boetticher, H., 488. Bogoluboff, P., 394. Bohm, C. R., 497. Bohm, 0. R., 510, 646. Bohrisch, P., 271. Bois, H. W. du, 378, 505. Boisbaudran, L. de, 506. Boiteau, G., 585. Bolis, A., 216. Bollenbach, H., 476, 668, 670. Boiling, R., 376. Bolm, F., 236, 563. Bolton, H.C.,525. Boltwood, B. B., 256. Boltz, G. E., 95. Bong, G., 162. Boujean, E., 162, 222, 226. Bonsdarff, W., 182. Boot, M. J. C., 40. Borchevs, W., 555. Borda, 11. Borg, F., 234, 305, 354, 593, 598. Borgmann, E., 94, 287. Bormann, K., 600. Bornemann, E., 551. Borneraann, K., 209. Borneman, G., 427. Borntrager, A., 103, 488. Borntrager, H., 409, 415, 461, 552, 576, 577, 597. Bosart, L. W. , 462. Bosek, 0., 295. Boshart, K., 449. Bostock, C. H., 188. i Bostock, G. H., 91. Bottger, R., 102, 234. Bbttger, W., 274, 315, 373, 374, 520, 610, 651, 654. i Bottomley, J. T., 155, 396. Boucher, C., 330. Boudouard, 0., 506, 509. Bouquet, J. P., 276. Bouquet, see Ebelraen, J. J. Bourdon, K., 55. Boursingault, J., 116. Bousfield, W. R.,66. Boussingault, J., 214, 224, 317, 515, 601. Boutzoureauo, B., 442. Bowser, L. T., 215, 541, 554, 599. Boyd, R. C., 366, 374. Boyer, J. W. , 203. Boyle, J. J., 382. Bradley, W. M., 103. Bragard, M., 364, 366, 367. Brakes, J., 203, 207. Brandenberg, R., 528. Brandt J., 641. Brandt, L., 195, 281, 452, 454. Brasseur, J., 607. Brassier, M., 597. Brauer, E., 3. Braun, C. D., 456. Braun, F. W., 120, 132. Braun, W., 390. Brauner, A., 389. Brauner, B., 276, 278, 295, 389, 390, 500. Brauner, R., 283, 284. Bray, W. C., 182, 199, 275, 351, 457, 483, 484, 517. Brearley, H., 311, 312, 361, 363, 378, 382, 386, 399, 401, 402, 408, 415, 469, 470, 490, 594. Breazeale, F. , 549. Breckenbridge, J. E., 237. Breneman, A., 9. Brewer, M. M., 107. Briant, L., 219. Bricout, G., 506. Bridgman, H. L., 133. Briggs, L. J., 630. Brinton, P. H. M. P., 232, 236, 382. Brislee, F. J., 195. 740 A TREATISE ON CHEMICAL ANALYSIS. Bristow, H. W., 138. Britton, J. B., 475. Brongniart, A., 658. Broockmann, K., 600. Brookman, K., 219. Brooks, C. J., 309. Brown, B. E., 604, 605. Brown, E. 0., 351. Brown, J., 451, 573. Brown, T. M., 191, 292, 296, 461. Browne, A. J. J., 525 668. Browne, D. H., 167. Browne, F., 559 Browning, P. E., 226, 294, 301, 351, 497, 514, 613. Browning, P. E., and Goodmann, 418. Brubaker, H. W., 554. Brudny, V., 324. Briigelmann, G., 77, 78 611, 625. Brugnatelli, T., 244. Bruhns, G., 51, 196. Brunck, 0., 107 213 214, 361, 389, 394, 625. Bruni, G., 274. Brunner, A., 160, 382. Brunner, C., 296, 364, 475 546, 547. Briinnich, J. C., 283. Brunton, D. W., 135. Brunton, J. D., 130, 132. Brush, G. J.,570. Brush, J., 505. Bruyn, L. de, 351. Bryant, E. G., 305, 354. Bube, K., 215. Biichner, E., 103. Biichner, G., 307. 357. Buckley, B. G., 231. Buckley, E. R., 670. Budden, E. R., 339. Buff, H., 8. Bull, J. C., 317, 333. Bullnheimer, K., 641. Bullnheimer, H., 411. Billow, C., 279. Billow, K., 341. Bunge, C., 14. Bunge, G., 616. Bunge, P., 4. Bunsen, R., 17, 39, 96, 98, 155, 172, 175, 183, 220, 229, 231, 239, 276, 278, 282, 283, 286, 296, 308, 309, 437, 505, 578. Bunzel, H. N., 491, 492. Burgass, R., 455. Burgemeister, A., 100. Burger, 0., 540. Burgess, L. L., 540. Burghardt, C. A., 161, 266, 475. Burgstaller, A., 77. Burk, W. E., 641, 646. Burkhardt, G., 555. Burrell, G. A., 625. Burton, W. M., 583. Busse, E., 350. CADY, H. P., 451. Cahen, E., 348, 378. Cain, J. R., 201, 382, 471, 480, 481, 484, 594. Cain, R., 484. Calker, F. J. P. van, 42. Cairns, F. A., 546 Cairns, F. I., 462. Calberia, E., 569. Caldwell, J. C., 104. Callisoii, J. S., 579. Calvert, F. C., 475, 614, 622, 625. Cameron, A., 175, 176. Cameron, F, K., 539, 549. Camilla, S., 378. Campagne, E., 481. Campari, G., 429. Campbell, A. C., 310, 455. Campbell, D., 216, 220. Campbell, E de M., 480, 491. Campbell, E. D., 182, 270, 399, 444, 594, 595, 599 Campbell, F. H., 573. Campredon, G., see Campredon, L. Campredon, L., and Campredon, G. , 409. Canby, R. C., 292. Canon, H. , 517. Cantoni. H,, 195, 280, 351. Carinelli, G., 580. Carius, L., 625. Carpari, see Fresenius, R. Carrasco, 0., 551. Capilli, E., 389. Carles, P. , 339. Carmichael, H., 103, 435. Carnegie, D., 40, 190. Carnelly, T., 355. Carnot, A., 177, 296, 305, 357, 364, 380, 390, 418, 429, 457, 458, 478, 495, 589, 641. Carpenter, F. C., 80. Carpiaux, E., 551. Carron, E. C., 524. Carson, C. M., 275. Carter, T. L., 434. Carulla, F. J. R., 188. Casamajor, P., 29, 103, 104. Casares, J., 641. Caspari, R., 236, 237. Cassal, C. E., 585. Cassel, H. R. , 429. Castendyck, C., 418. Cathrein, A., 208. Causse, H., 299. Caven, R. M., 305. Cayeuse, L. , 525. Cayvan, L. L., 605. Cazeneuve, P., 454, 474. Celichowski, K., 540, 541. Cerdan del Campoy, A., 386. Cesaro, G., 159. Chabaud, A., 55. Challinor, R. W., 396. Chamberlain, R. T., 573. Chamel, G., 477. Champion, E. C., 270. Champion, P., 595. Chance, E. M., 318. Chancel, G., 468. Chancel, M. F. , 495. Chandler, C. F., 161, 499. Chapin, W. H., 583, 585. INDEX OF NAMES. 741 Chapman, A. 0., 188. Chapman, E. J., 577. Chappius, P., 30. Charlton, H. W., 77. Charpentier, P., 77. Chatard, T. M., 162, 179, 207, 332, 382, 408, 415, 607, 637. Chatelier, H. le, 163, 529. Chautems, J., 280. Chauvenet. 270. , Cheladze, N., 444, 449. Cheney, M. S., 395. Chesneau, G., 273, 419, 420, 503, 563, 593, 594. Chester, A. H.,.462. Chevala, A., 323. Chevron, L., 284. Chikashige, M., 226. Chism, R. E., 345. Chouchak, D., 604. Christensen, A. , 301. Christomanos, A. K., 218, 244, 474. Charitschkoff, K., 455, 625. Chumanoff, S., 58. Church, A. H., 110. Chydenius, J. J., 502, 510. Claason, E., 483. Claesson. P., 551. Clark, A'. B., 296. Clark, A. N., 214. Clark, C. M., 12. Clark, E., 475. Clark, F. W., 296, 299. Clark, J., 280, 296, 305, 366, 474, 480. Clark, S. G., 191. Clark, W. W., 467. Clarke, F. A., 161. Clarke, F. G., 491. Clarke, F. W., 161, 250, 296, 475. Clarke, J., 473, 506. Clarke, T., 499. Clarkson, T., 132. Clasen, A. W., 305, 306, 654. Classen, A., 31, 46, 195, 256, 261, 280, 312, 320, 335, 354, 373, 461, 500, 554. 555, 641. Classen, E., 472. Claud, F. 0., 23. Claud, T. C. , 348. Glaus, A., 279, 351. Glaus, C., 201, 437. Clausmann, P., 641, 644. Clemence, A. B., 102. Clermont, P. de, 178, 373, 429. Cleve, P. T., 495, 501, 506. Cloedt, E. von, 348. Cloiz, S., 276. Clowes, E. S., 373. Clowes, F., 12, 39, 305, 521. Gloubry, G. de, 311. Clouet, J., 474. Clouet, P., 480. Cobbett, 40. Cobenzl, A., 409. Cocheteux, A., 451. Cochins, F., 117. Cochrane, M. H., 100. Cockburn, T., 364. Cohen, E., 169. Cohen, J. B., 196. Cohn, A., 201, 396. Cohn, A. L, 60. Cohn, R., 200. Coleman, J. B., 12, 113, 521. Collan, U., 617. Colle, H.,323. Collens, E., 197. Collet, E., 413. Collier, P., 228. Collins, S. H., 249. Collins, S. W., 280. Collins, W. D., 536. Collitt, B., 196. Colson, A., 116. Commaille, A. , 305. Connell, A., 504. O'Connor, H., 110. Conroy, J. T., 115. Contamine, G., 232, 234. Contat, A., 189. Convert, A. , 87. Cook, J. B., 90. Cooke, J. P., 23, 103, 462, 651. Cooper, A. J., 355. Copaux, H., 393, 398, 585. Copaux, M.,579. Coppalle, A., 319. Coram, H. C., 58. Coren winder, B. C., 232, 234. Corleis, A., 546. Corminbeuf, H., 162, 483. Cornette, P., 177. Cossa, A., 213, 239. Coste, J. H., 570. Coutal, E., 641. Cowper, R., 144. Cox, A. J., 244. Craig, C., 176. Craig, D., 160. Craig, G., 175. Cramer, E., 59, 72, 527, 528, 555, 578, 633. Crawford, W. G., 203, 486. Creighton, H. J. M., 611. Cremer, H., 405, 408, 409. Cribb, C. H., 553. Cripps, R. A., 573. Crismer, L., 479. Crobough, F. L., 190, 626. Crookes, W., 18, 23, 97, 239, 437, 496, 497. Cross, C. F., 72, 177, 179, 546. Crosse, A, F., 434, 565. Cruess, W., 258. Cruser, F. van Dyke, 412, 415. Cumenge, E., 488. dimming, A. C., 192, 345, 617. Cunningham, M., 540. Curr, H. J. C., 343. Curry, B. E. , 290. Curtis, J. S., 328. Curtma-n, C. 0., 540. Curtmann, L. J., 177, 180, 210, 363, 51 /. Cushman, A, S., 40, 312. Czerwek, A. , 296. Czudnowiez, C., 465, 480. 742 A TREATISE ON CHEMICAL ANALYSIS. DAKIN, H. D., 366, 374. Dale, B., 195. Dale, R, S., 98. Damaskin, N., 200. Dammer, O. , 625. Damour, E., 506, 525, 601. Dancer, W.,298. Daniel, K., 637, 641, 649, 660. Danne, H. A., 573. Danner, E. W., 196, 280. Danziger, J. L., 334, 367, 385. Darmstadt, M., 515. Darroch, J., 267, 299. Dart, A. C., 436. Darton. H. N., 197. Das, T. W..98. Davidsohn, I., 502. Davidson, R., 187. Davies, A. E. , 590. Davies, J., 200, 553. Davies, J. H., 352. Davies, J. L., 305. Davis, J. T., 496. Davis, W. A., 232, 237, 319. Davison, A. L., 358. Davy, H., 162, 215. Daw, F. W., 378. Dawson, H. M., 182. Day, A. L., 124, 155. Day, W. C., 474, 475. Dean, G. W., 460, 461. Debray, H., 437, 593. Deckert, H., 195. Dede, L., 372, 390, 524. Dederichs, W., 276. Defacqz, E., 406, 410. Defect, F. W., 59. Dehms, F., 355, 396. Deiss, E., 457. Delacharlonny, P. M., 225. Delafantaine, M., 502, 508. Delardrier, E., 641. Delepine, M., 115. Delfl's, H., 364. Delisle, A., 551. Dernarcay, M. , 305. Demichel, A., 36. Demorest, D. J., 238, 299, 382, 481. Deniges, G.. 445, 579. Dennant, J.', 229. Dennis, L. M., 94, 499, 510. Dennstedt, M., 547, 615. Derrick, W. H., 268, 270. Desaga, P., 110. Descharmes, P., 124. Descroizilles, E. A. H., 45. Deshays, 382. Desi, E. D., 410, 416. Desvergnes, L., 408. Dettloff, A., 70, 551. Deus, A., 364. Deussen, E., 185, 466. Deville, H. St C., 162, 184, 308, 372, 437, 506, 520, 525, 601, 637. Dewey, F. P., 296, 436. Dexter, W. P., 409. Diamant, J., 234. Dibbitts, B. C., 118, 178, 315, 316, 363. Dieckmann, T., 411. Diehl, C., 611. Diehl, W., 323, 325. Diethelm, B., 35. Dietl, J., 42. Dietrich, E., 555. Dietrich, K., 61. Dieulafait, M., 576. Digby, K. E., 330. Dirvell, P. J., 425, 429. Disselhorst, H., 29. Ditte, A., 278, 279, 295, 481, 583. Dittmar, H., 18, 26, 78, 94, 234, 240, 241, 372, 475. Dittrich, W., 16, 163, 177, 179, 208, 247, 360, 373, 384, 462, 466, 474, 495 496, 498, 504, 571. Ditz, H., 396. Divers, E., 150, 212, 226, 440, 441. Divine, R. E., 179, 479. Dixon, A. B., 563. Dixon, W. A., 142. Dliss, E., 378. Doat, J., 237. Dbbereiner, J. W., 162. Doelter, C., 462, 525. Doeltz, F. 0., 616. Dohler, E., 364, 412, 413. Donath, E., 244, 270, 278, 341, 307, 373, 378, 399, 410, 443, 461, 475, 495, 551, 568, 579. Donovan, W., 461. Doolittle, 0. S., 599. Dootson, F. W., 396. Doring, T., 223, 224. Dormaar, J. M. M., 312. Dott, D. B., 272, 318, 312. Dougherty, G. T., 402. Dougherty, J. T. , 386. Douzard, E., 58, 117. Dover, M. V., 351. Draper, C., 62, 533. Draper, H. N., 363. Drawe, P., 319. 647, 648. Drechsel, E., 77, 78, 244, 547. Dreverhoif, M., 88, 388, 620. Droixhe, A., 284. Drossbach, P., 501, 504, 510. Drossbach, J. P., 512. Drown, T. M., 102, 176, 271, 598. Drushel, W. A., 226, 541. Dubin, H., 180, 210,363. Dubois, F., 499. Dubois, H. W., 452. Dubois, P. C., 475. Dubois, R. , 60, 622. Dubose, A., 77. Dubovitz, H., 94. Ducretet, E., 55. Ducru, 0. , 280. Dudley, C. B., 190, 249, 599. Duerr, G., 47. Dufert, F. W., 55. Duflos, A., 389. Dufty, L., 382, 551. Dulin, R. S., 351. INDEX OF NAMES. 743 Dumas, J. B. , 308. Duncan, J. B., 301, 303. Dunlap, E. E., 323. Dunn, J. T., 93, 619. Dunnington, F. P., 53, 205, 316, 322. Dunstan, W. R., 579. Duparc, L. , 474. Dupasquier, M., 305. Dupre, A., 81, 195, 200, 287. Dupre, F., 194. Dupre, F. T. B., 233. Dupre, L. W., 141. Dupre, P. V., 573. Diirkes, K., 627. Duschak, L. H., 516, 611. Dutta, J. M., 481. Duval, R., 136. Duvillier, E., 240. EARLY, W., 462. Eavenson, A., 599. Ebaugh, W. C., 621. Ebeling, A., 189. Ebelmen, J. J., 520. Ebelmen, J. J., and Bouquet, 585. Eberhard, 0., 73. Eckardt, M., 413. Eckstadt, 286. Eckenroth, A., 305. Eddy, E. A., 226, 244. Eder, J. M., 462. Edgar, A., 611. Edgar, A. J. M., 121. Edgar, G., 191, 467, 480, 481, 482, 483, 488. Edwards, A. M., 183. Edwards. V., 600, 601. Eggertz, C. G., 233, 236. Eggertz, V., 590, 593, 595, 600. Eggertz, V. von, 355, 362, 372. Egleston, T. , 443. Ehrenhofer, W., 551. Eiloart, A., 42, 377. Eitel, W., 571. Eldridge, G. F., 271. Eliot, C. W., 547. Elliott, A. H., 546. Ellis, C. S.,55. Elsdon, G. B., 340. Eisner, F., 55. Eisner, L., 114, 428. Emery, A. L., 627. Emmerling, A., 144, 599. Emmerton, F. A., 379, 598. Emster, K. van, 681, 682. Engel, R., 308, 580. Engelbach, T., 519. " Engels, C., 312. Engstrom, N., 501. Enright, B., 215, 522. Erdmann, 0. L., 35, 116, 317, 345, 390, 540, 614. Erlenmeyer, E., 7, 48, 112, 568. Errera, G., 228. Eschka, A., 345, 621. Esilman, A., 460. Estes, C., 604. Esteve, E.. 271. Euler, H., 182, 413. Evans, W. P., 110. Evre, M. St, 390. Exner,'F. F., 337, 338. FABER, P., 205, 206. Fahlberg, C., 367, 368, 369. Failyer, G. H., 146, 539, 630. Fairchild, J. G., 335, 599. Fairley, M., 204. Fairlie, A. M., 351. Falk, F. A , 77. Falk, M. J., 367. Faraday, M., 17, 121. Farnsteiner, K., 579, 585. Farrington, E. H., 579. Farsoe, V. , 323. . Fasal, J., 311. Fauch, A., 138. Faust, J. K., 102, 377. Favre, C., 234. Favre, P. A., 118. Fay, H., 441. Feichtinger, G., 528, 529. Feigenberg, B., 536. Fellenberg Rivier, L. R. von, 162. Feller, T., 98. Fels, J.,475. Fenton, H. J. H., 203, 468, 486, 644. Ferguson. W. C., 94. Fernekes,' G., 3?,0, 351. Ferrer, J., 385. Fery, C.,201. Fessenden, R., 244. Fichter, F., 302. Fiebag, P., 44. Fieber, R., 475. Fiechter, A., 234. Field, F., 223, 275, 284, 348, 354, 363, 651 Fieldner, A. C., 621. Filliti, G. A., 625. Finkner, M., 658. Finn, A. K, 488, 492. Finkener, R., 231, 232, 234, 501, 546, 555, 591, 593. Fischer, A., 256, 296, 312, 611. Fischer, B., 579. Fischer, C. G., 111. Fischer, E. , 280. Fischer, E. F., 46. Fischer, F., 135. Fischer, H., 270, 540, 541. Fischer, J., 336. Fischer, L. A., 31, 39. Fischer, W. M., 378. Fischer, N. W., 188, 389. Fischer, R., 585. Fischer, T. , 483. Fisher, H., 284. Fittig, R., 178. Flath, J., 364. Flaiolet, M., 351. Fleck, H., 323. SffiiV 6 2 5 8 66,103,228, 250 , 399, 517, 655. Fleischer, M., 593, 624. 744 A TREATISE ON CHEMICAL ANALYSIS. Fleissner, H., 66. Fleitmann, T. , 89. Fleming, E. P., 244. Fleming, W., 55. Fleurent, E., 607. Fliickiger, P. A., 533. Fliirscheim, B., 167. Fock, A., 575. Foerster, F., 144, 301. Foerster, 0., 261, 286, 597. Fohr, C. F. , 328. Fohr, K. F. , 167. Folin, 0., 611, 613. Folkard, C. W., 23. Follenius, 0., 357, 451, 466, 491. Foord, C., 39. Foord, G., 94, 169. Foote, H. W., 423, 499. Foote, 0., 244. Forbes, E. B., 630. Forbes, D., 207, 328. Forbes, W. R., 104. Forchhammer, G., 525, 568, 637, 657. Ford, A. P. , 376. Ford, W. E., 423. Forder, S. W., 528. Forestier, H., 323. Formhals, H., 299 Forster, F., 312, 437. Fbrster, 0., 42, 61. Fortini, V., 524. Foster, C. leN., 577. Fouks, F. J., 405, 408. Foulk, 0. W., 611. Foullon, H. P., 488. Fouque, F., 637, 657. Fouret, R., 266. Fox, J. J., 317, 364. Fraenhel, A., 311. Frailong, R., 55. Francis, E. E., 100. Frank, H., 190, 416. Frank, M., 594. Frankel, A., 311. Frankel, E., 517. Frankel, L. F., 162. Franz, B., 208. Fraps, G. S., 607, 614, 615. Frazer, A. , 573. Frehse, M., 322. Freidburg, L. H., 161. Friedheim, C., 167, 280, 282, 283, 284, 289, 297, 311, 318, 343, 410, 411, 413, 417, 418, 479, 593, 127. Fremy, E., 528. French, E., 55. French, W., 201. Frerichs, F., 517. Fresenius, C., 589. Fresenius, C. R., 39, 159, 216, 234, 236, 284, 448, 471, 663, 666. Fresenius, H., 232, 236, 278, 395, 475, 538, 555, 622, 623. Fresenius, R., 25, 58, 118, 119, 144, 149, 167, 178, 194, 208, 211, 214, 218, 219, 220, 229, 232, 233, 235, 246, 278, 282, 283, 285, 297, 310, 315, 320, 373, 374, 381, 387, 388, 391, 399, 406, 427, 428, 451, 455, 475, 477, 488, 489, 491, 509, 510, 514, 515, 517, 520, 548, 554, 559, 589, 590, 592, 593, 594, 595, 597, 601, 611, 614, 615, 617, 630, 641, 649, 668. Fresenius, R., and Carpari, 81, 89, 102. Fresenius, W., 31, 46, 144, 251, 285, 517. Freund, S., 495, 504. Freundlich, H., 336. Frevert, H. L., 452. Frey, 0., 529. Fribourg, C., 207. Fricke, L , 599, 637. Friedel, C., 117, 488. Friedenthal, H., 60, 201. Friedrich, K., 43, 434. Friedrichs, see Greiner, E. 1 Friend, J. A. N., 452. I Friswell, R. J., 144, 148. Fritchle, 0. P., 488. , Fritzsche, A., 214. Froehde, A., 270. Frohman, E. D., 415. Fromme, H., 466, 585. Frost, 0. J., 292. I Froysell, H. H.,195. Friihling, R. , 102, 103, 527. Fuchs, F., 555. Fuchs, J. N., 663. Fuchs, J. N. von, 226, 470. Fuller, G., 53. Fulton, C. H., 434, 443. Funk, R., 357. Funk, W., 359, 361, 364, 388, 389. Fyfe, A., 280. GAAB, see Schneider, A. Gage, R. B., 466. Gaither, E. W., 554. Galbraith, W., 470, 476. Galetti, M., 367. Gallety, J. G., 305. Gallo, G., 189. Gamboa, R. L. y, 516. Gans, L., 653. Gardiner, A. D., 364. Gardner, W. M., 193, 195, 196. Gamier, L., 351 Garratt, F., 481. Garrett, F. C., 521. Garret, T. H., 197. Garrigues, W. E. , 350. Garside, T., 116. Gasselin, V., 579. Gastaldi, C., 270. Gastine, G.. 286. Gauhe, F., 389, 391, 392. Gauss, 11. Gautier, A., 203, 206, 641, 644. Gawalovski, A., 8, 35, 36, 42, 44, 55, 70, 89, 94, 98, 128, 143, 244, 277, 320, 527, 555, 597, 602, 614. Gay Lussac, J. L., 45, 47, 305, 307, 579. Gaze, R. , 66. Geffeken, G., 533. Geibel, W., 115, 372. Geisow, H , 498. INDEX OF NAMES. 745 Geissler, C., 43, 136. Geissler, K, 320. Gemmel, A., 188, 189. Genth, F. A , 389, 475, 480. Gentil, B. G., 77. George, G., 586. Georuevic, P., 580. Gerdes, R., 59. Gerhard, F.. 150. Gerlach, G.T., 65, 155. Gerland, B, \V., 480. Gerlinger, P., 351. Germain, E., 551. Gernez, D., 178, 586. Ghilian, A., 378. Gibb, A., 280, 282. Gibb, T., 280. Gibbs, W., 102, 161, 170, 211, 218, 297, 362, 363, 374, 388, 389, 390, 392, 395, 408, 411, 418, 437, 469, 476, 478, 488, 495, 506, 593, 594. Gilbert, A. , 266. Gilbert, C., 100, 592. Gilbert, H., 641, 644. ' Gilbert, J. P., 173, 176. Gilbert, K., 540. Gilbert, R. D., 480. Giles, H. B., 47, 48, 49, 51, 115, 161, 266, 418, 420. Gilette, C. E., 124. Gill, F. W.,625. Gilm, H., 576. Gintl, W. F., 568. Gintl, W. H., 188. Giraud, H., 299. Gladding, T. S., 216, 218, 589, 593, 615. Gladstone, J. H., 188, 200, 396. Glaser, C., 475, 503, 547, 606. Glaser, E., 607- Glaser, F., 60, 489. Glasmann, B., 192, 467. Glazer, C., 501. Glendinning, N., 121, 611. Glenn, W., 36, 129, 130, 136. Glinka, F., 162. Glixelli, S., 360. Glover, J., 610. Gmelin, L., 546. Gneiwosz, S., 58. Gobel, see Hodes. Gockel, H., 34, 35, 39, 58, 189, 288, 460, 559. Godetfray, F., 239. Godwin, B., 198. Goettsch, H. M., 286, 491. Goetze, 34. Goetzl, A., 563, 625. Goldberg, A., 62, 63. Goldschmidt, 314. Goldschmidt, C., 654. Goldschmidt, V., 425. Goldschmidt, V. M., 176. Goldstein, E. 114. Goldstein, J., 568. Gomberg, M., 551. Gooch, F. A., 33, 104, 107, 183, 189, 191, 196, 201, 216, 218, 226, 244, 280, 283, 294, 301, 337, 351, 372, 374, 411, 441, 449, 452, 457, 467, 480, 520, 536, 552, 560, 570, 589. Goodmann, see Browning, P. E. Gorceix, H , 501. Gore, G., 161, 162. Gore, H. C., 117. Gorges, M., 579. Gorgeu, A., 372, 380. Gortner, R. A., 371, 382. Goske, A., 47, 102. Gottlieb, B. N., 615. Gottschalk, V. H., 184. Goutal, E., 562. Goyder, G. A., 328. Gozdorf, G. A., 328. Grabe, A., 441. Grager, 224. Graham, T., 435. Grandeau, L., 237. Grassi, L., 456. Gray, J., 266. Gray, J. S. C., 102. Green, H. H., 47. Gregoire, A., 551. Gregory, A. W., 484, 648. Gregory, J. W., 201, 670. Greiner, E., 43, 55, 631, 661, 668. Greiner, E., and Friedrichs, 59, 148. Grete, A.? 598. Griffin, C. E., 480, 491, 492. Griffin, J. J., 9, 94. Griffin, R. C., 593. Griffiths, A. B., 116. Grimm, C. , 39. Grindley, H. S., 625. Groll, F., 195. Grosjean, B. J., 103. Gross, A., 288. Grossmann, H., 394 : 399. Grossmann, J., 246. Griiber, 0. von, 606. Gruner, A. , 87. Guerreau, M. , 328. Guess, H. A., 350. Guichard, M. , 55. Guichard, P., 415. Guiot, H., 373. Guirin, G., 600. Gunning, J. W., 244, 525, 617. Gunther, A., 585. Gunther, C. G., 121. Gustavson, G., 559. Gutbier, A., 158, 296, 441, 443. Guthrie, F. B.,214. Gutkowsky, C., 614. Guttmann, L. F., 9, 575. Guyard, A., 179, 180, 377, 379, 488, 580. Gwiggnes, A., 90. Gyory, S., 301. Gy zander, C. R., 611, 615. HAAS, H., 268, 410, 422. Haber, F., 448, 449. Habermann, J., 150, 268, 391, 493. Habich, R., 378. Haefcke, H., 231, 234. Haeffely, E., 308, 314. 746 A TREATISE ON CHEMICAL ANALYSIS. Haen, E. de, 351. Haga, T., 644. Hagen, H., 280, 538. Eager, H., 149, 212, 320, 475, 680. Hake.H. W., 156. Haldeman, F. M., 94. Hale, F. E., 287. Hall, E. J., 367, 368. Hall, H. M., 150. Hall, L. A., 539. Hall, R. D., 409, 419, 420, 502. Hall, R. W., 114. Hall, V. J., 177. Hallmann, C., 296. Hallopeau, L. A., 410. Hambloch, A , 668. Hambly, F. J. , 882. Hamburger, E. W., 192. Hamel, F., 237. Hamilton, E. M., 77. Hamilton, E. W., 192. Hamilton, R., 169, 595. Hampe, W., 150, 268, 283,285, 350, 362, 364. 376, 378, 463, 589, 653. Hanamann, J., 595. Hancock, D., 291. Hancock, W. C , 670. Handy, J. 0., 220, 446, 621. Hanriot, M., 457. Hanson, W., 473. Hanus, T., 479. Harcourt, A. G. F., 339, 340. Harcourt, A. V., 197. Hardy, H., 339. Harker, G., 641. Harpe, C. de la, 213. Harper, D. K, 183, 449. Harpf, A., 66. Harris, H. B., 399. Harrison, J. B., 525, 668. Harrison, F. W., 452. Hart, E., 169, 185, 461. Hart, F., 528, 529. Hart, P., 90, 475. Hart, W. B., 141. Harting, P., 339. Hartley, W. K, 66, 305. Hartmann, L., 55. Hartmann, 0., 5. Hartwell, B. L., 182, 510. Hasenclever, P., 479. Hasenclever, R., 28. Haslam, A. R., 614. Hassel, C., 360,373. Hassler, F., 547, 615. Hassreidter, V., 279, 359, 366, 386. Haswell, A. E., 453, 600, 644. Hatch, F. H., 668. Hattensauer, G. , 367. Hauenschild, H.j 528. Hauer, C.' R. van, 418, 472. Hauer, K. van, 357. Haupt, M., 351. Hauser, 0., 419, 423, 486, 497, 501, 504. Hausermann, C., 475. Haushofer, K., 525, 661. Havens, F. S., 449, 457. Hay wood, J. K., 290. Hazard, J., 657. Hazen, A., 81. Headden, W. P., 219, 418, 420. Heath, F. H., 351. Heath, G. L., 355, 621. Hebebrand. A., 113. Heberlein, E., 279. Hebre, E., 390. Hecht, H., 663. Heczko, A., 627. Heerman, P., 76. Hefelmann, R., 213. Hefemann, R., 580. Hehner, 0., 578, 589. Heide von der, 55. Heidenhain, H., 546, 550. Heidenreich, M., 312. Heidenreich, 0., 162, 172. Heidenreich, 0. N., 611, 615. Heifer, A., 215. Heike, W., 378, 481. Heilborn, W., 394. Heim, M., 670. Heimann, E., 280. Heintz, W., 215, 218. Heldt, W.,527, 529. Helmhacker, R., 410. Helpert, S., 411. Hempel, H., 405, 579. Hempel, W., 89, 117, 118, 123, 162, 169, 215, 244, 381, 475, 538, 551, 639, 646, 647. Henderson, G. C. , 305. Henderson, W. H., 410, 418. Hendrick, J., 201, 527, 551. Henkel, T., 59. Henneberg, W., 499. Hennig, R., 8. Henning, G. F., 561. Hensen, C., 481. Hentschel, G., 521. Henz, F., 296, 297, 309, 312, 336. Hepter, J., 208. Heraeus, W. C., 107, 115, 219, 372. Herapath, T. L., 200. Herdsmann, W. H., 250. Heriot, M., 197. Herman, J , 166. Hermann, H., 606, 669. Hermann, R., 408, 419, 462, 495, 504. Herrenschmidt, H., 389. Herschel, J. F. W., 387, 470. Herstein, B., 412. Herting, 0., 409, 611, 615, 621. Herz, W., 182, 211, 362, 388. Herzfeld, H., 504. Herzsfeld, J., 497. Herzka, E., 70. Herzog, H., 318, 607. Hess, W., 606. Hess, W. H., 182, 214, 444. Hessert, J., 554, 559. Heublein, 0., 72. Heuse, W., 685. Hewitt, J. T., 451, 465. Hewitt, T. E., 604. Heyer, M., 528. INDEX OF NAMES. 747 Heygendorff, von, 50, 58, 113. Heyl, P.,226. Hibbert, E., 58, 187, 189. Hicks, H. B., 502. Hicks, J. G., 305. Hildebrand, E. C. , 98. Hildebrand, W. F., 167. Hildebrandt, H., 63. Hilger, A., 268, 410, 420. Hill, A. E., 78, 551, 561. Hill, C. A., 339. Hillebrand, W. F., 84, 112, 117, 122, 124, 174, 185, 186, 203, 206, 207, 209, 221, 235, 246, 251, 371, 382, 434, 435, 462, 465, 466, 467, 472, 473, 480, 488, 489, 490, 492/498, 502, 514, 560, 570, 617, 637, 638, 639, 668. Hillemann, A., 641. Hinden, F., 160. Hinds, J. I., 633. Hinds, J. J. D.,271. Hinds, J. L. D. , 632. Hinrichs, G. D., 15, 16. Hinrichsen, F. W., 180, 189, 409, 411, 594, 626. Hintz, E., 448, 475, 488, 491, 501, 504, 505, 509, 510, 511, 589, 611, 625, 641, 650. Hirsch, R., 103. Hirschel, W., 59. Hirschsohn, E., 47. Hirschwald, J., 661. Hisinger, W., 361. Hissink, U. J., 593. Hlasiwetz, H., 547. Hobling, V., 55. Hochmeimer, C. F. A., 657. Hochstatter, R. W., 150. Hodes and Gobel, 111. Hodge. B., 280. Hodges, A. D., 135. Hodtke, 0., 455. Hoff, J. H. van't, 529, 530. Hoffmann, L. W., 150. Hoffmann, L., 425, 426. Hoffmann, M., 271, 275, 280, 283, 305, 308. Hoffmann, R., 161. Hofmann, A. W., 275. Hofmann, H. 0., 270. Hofmann, K. A., 540. Hogartb, J. W., 60. Hogg, T. W., 192, 476. Hoglund, M., 501. Hohlrouch, F., 315. Hohnel, M., 475. Hoitsema, C., 77, 651. Holand, R., 625. Holdcroft, A. D., 117, 169, 172, 574. Holde, D., 9. Holland, P., 169, 207, 223,382. Hollard, A., 280, 335, 398. Hollard, B., 256. Holleman, A. F., 212, 234, 235, 614, 651. Holliger, M., 625. Holloway, C. T., 345. Holloway, G. T., 252, 384. Holmblad, M., 528. Holthof, C., 161, 372. Holverscheit, R.,472, 484. Holzmann, M., 501, 504. Hommel, W., 416. Honig, M., 579. Hood, C. T., 451. Hooper, H. E., 292. Hopkins, A. J., 196. Hopkins, C. G., 251, 391. Hopkins, C. P., 66. Hoppe-Seyler, F., 525. Horkheimer, P., 498. Hormuth, L., 110. Horn, D. W., 40, 474. Horn van der Bos, J. L. M. van der, 514, 517. Homes, R. , 525. Horsley, G. F., 98. Horvath, A., 324. Hostetter, J. C., 481, 484, 594. 1'HOte, L., 355. Hottinger, R., 10. Hovenstadt, H., 144. How, H., 363, 373, 583. Howard, C. D., 51. Howe, J. L , 342. Hubbard, R. A., 179. Huber, 0., 627. Hubert, A. E. von, 355. Hiibner, E., 197. Hudig, J., 60. Hufschmidt, F. , 280. Hugershoff, F., 44. Huldschinsky, E., 364, 396. Hulett, G., 212, 611, 614,651. Hulett, G. A., 114. Hulsebock, C. J. van L., 55, 58. Hume, W. F., 525. Hundeshagen, F., 212, 519, 521, 590, 592, 595, 621. Hunt, T. S., 475, 525. Huntley, G. N., 570. Hurff, G. B., 192. Hussah, E., 670. Hutchin, H. W., 405, 408. Hutchin, W. W., 408. Hutchinson, C. C., 273. Huttner, C.,431. Huybrechts, M., 613. Hyde, H. S. J., 595. IBBOTSON, F., 179, 191, 311, 312, 382, 386, 394, 401, 408, 415, 470, 491, 594. Iklee, G., 528. lies, M. W., 161, 268, 355, 576. Ilinsky, M., 385,393, 455. Immendarff, H., 607. Ingen, D. A. van, 367. Ingersoll, E. H., 377. Inhelder, A., 296, 306, 312. Irby, J. R. M., 195. Isbert, A., 595. Islam, H., 550. Isler, M., 676. Itallie, L. van, 454. Ivanicki, A., 394. JACK, W. E., 453, 460. Jackson, D D., 632. Jackson, E., 203, 204. 748 A TREATISE ON CHEMICAL ANALYSIS. Jackson, F., 339, 610, 654. Jackson, W., 662. Jacob, E., 94. Jacobi, K., 581, 582. Jacobs, W. A., 625. Jacoby, R., 501. Jacquelin, V. A., 355. Jager, B., 138. Jager, E., 555. Jago, W., 103. Jagnaux, R., 437. Jahoda, R., 189. Jambar, J., 218. James, C., 510. James, G. A., 123. James, Z. A., 155. Jamieson, G. S., 290, 299, 399. Janda, F., 136. Jani, W., 216, 593, 598, 600, 602. Jannasch, P., 97, 112, 160, 162, 172, 183, 280, 330, 348, 425, 430, 441, 443, 475, 479, 566, 571, 589, 611, 615, 624, 637, 641. 650, 653. Jannettaz, E., 525. Janovsky. J. V., 594. Jarvinen/K. K., 218, 627. Jawein, L., 358, 366. Jean, F. 234, 329, 376, 377. Jeannel, A., 361. Jeannel, M.. 183. Jeanneret, B., 497, 499. Jefferson, A. M. , 182, 444. Jeffery, J. H., 452. Jehn, C., 463. Jehn, K., 579. Jeller, R., 461, 495. Jene, K., 615. Jenkins, E. H., 595. Jenzsch, G., 519, 663. Jervis, H., 165, 179, 183, 192, 383, 399, 477. Jewett, J., 361 Jewitt, F. F., 142. Jhmori, T., 12. Job, A., 511. Job, R., 311. Jodlbauer, A., 641. Johanusen, 0., 545. John, A. D. St, 177. Johnson, C. M., 399, 401, 412, 469, 551, 563. Johnson, E. M., 55. Johnson, E. S., 55, 60 Johnson, F. M. G., 118. Johnson, G. S., 7, 58. Johnson, H. H., 239. Johnson, P., 129. Johnson, S. W., 594. Johnston, J., 124, 178, 611, 612, 616, 619. Johnstone, W., 87. Jolles, A., 200, 311, 604, 606. Joly, A., 437, 583. Jones, C., 98, 190, 197, 244, 501, 598. Jones, G. C., 452. Jones, J., 475. Jones, L. C., 579, 580, 589. Jones, R., 607. Jong, M. de, 314. Jong, R. de, 456. Jordis, E., 169, 170, 175. Jordis, F., 245. Jordon, W. J., 345. Jorgensen, G., 216, 579, 598, 590, 595. Joshua, W. P., 323. Joulie, H., 602. Joulin, L., 319, 602. Jowitschitsch, M. Z., 479. Joy, C. A., 496. Juette, M., 363 Julian, F., 91. Julien, A. A., 354. Jiiptner, H. F. von, 8, 38, 44, 77, 184, 194, 196, 249, 250, 386, 461, 546, 595, 600. Jurich, K. W., 447. Just, J., 530. Jutsum, S. C., 548. KACHLER, M., 42, 103. Kahan, B., 517. Kahlenberg, L., 316, 534, 578, 579. Kaiser, E., 664. Kaiser, J. A., 280. Kallauner, 0., 524. Kalman, S., 475. Kamm, 0., 540. Kammerer, H., 330, 366, 372, 478, 517, 577. Kampen, G. B. van, 595, 644. Karasglanoff, L., 378. Karslake, W. J., 484. Karsten, C. J. B., 525. Karsten, G., 183. Kassner, 0., 473. Kastner, L., 441. Kato, Y., 612. Kayser, H., 114. Kayser, R., 226, 475. Kaysser, A., 378, 461. Keen, W. H., 367. Kehrmann, F., 167, 411. Keiser, E. H., 528. Keler, H. von, 178, 201. Keller, E., 119, 433, 440, 442, 443. Keller, K. , 305. Kelly, A. A., 165. Kempf, R., 22, 25. Kempf, T., 189. Kendall, E. C , 351. Kennepohl, G., 595. Kennicut, L. P., 475. Kerl, B., 633. Kern, E. F., 191, 457, 489, 490, 491. Kern, S., 102, 116, 161, 188, 304, 305, 409, 475. Kernot, G., 317. Kerr, J. F., 90. Kershaw, J. B. C., 256. Kersten, C., 509. Kessel, E., 218; 600. Kessler, F., 361, 372, 374, 451, 453. Kessler, M., 169 Kettembeil, W., 364, 369. Kettler. E., 214. Kickton, A., 636. Kiefer, H. E. , 528. Kietreiber, F., 275. Kilgare, B. W., 599. INDEX OF NAMES. 749 Kilgrave, B. W., 598. Kinder, H., 194, 195, 364, 451. King, V. L., 455. Kinnear, H. B., 522. Kippenberger, C., 42, 44. Kirchhoff, G , 231, 239. Kirchline, F. 0., 560. Kirkland, J. B., 364. Kirsclmick, C., 367. Kirwan, R., 657. Kittel, J., 579. Klaproth, M. H., 161, 162, 499, 657. Klason, P., 441, 442. Klaudie, J., 301. Klaye, A., 364. Klecki, V. von, 484. Klein, C. A., 319. Klein, D., 579. Klein, F., 630. Kleine, A., 280, 546. Kleinstiick, M., 638. Klenker, 0., 296. Klosmann, H., J49. Knapp, F., 528, 529. Knauer, A., 55. Knausz, C., 525. Knecht, E., 58, 187, 189, 415. Kneff, C. W., 131. Knieder, 396. -Knight, N., 124, 175, 213, 521. Knobukow, N. von, 190, 337, 357. Knofler, 0.. 55, 56, 168. Knop, W., 595, 600, 602. Knopp, W., 161, 226. Knorr, A. E , 77. Knorre, C. von, 386, 393. Knorre, G. von, 299, 350, 408, 455, 473, 511, 615, 627. Knorre, J. von, 373. Knosel, T., 240. Knox, J., 278. Kniipfer, 0., 78. Kobell, F. von, 124, 406, 513, 641. Kober, P. A., 104, 352. Koch, A. A., 350, 351, 637, 639, 646. Koch, R. F., 162. Koch, W., 622. Koehler, 0., 284. Kohler, F., 33. Kohler, 0., 275. Kohlrausch, F., 12, 18, 212, 614, 651. Kohn, C. A., 195. Kolb, A., 299. Kolbe, H., 237, 381, 554. Kollock, L. G., 358. Konek, F. V., 622. Konek, F. von, 267. Konig, C. R., 345. Konig, G. , 593. Konig, J., 593, 595. Konig, R., 55. Koninck, L. L. de, 23, 30, 43, 44, 45, 78, 80, 87, 89, 100, 114, 148, 163, 166, 167, 201, 219, 220, 234, 276, 281, 285, 288, 318, 348, 360, 367, 372, 379, 382, 389, 393, 428, 455, 465, 473, 540, 547, 597, 620. Koningh, L. de, 578, 585. Konowaloff, M., 182. Kopfer, F., 547. Kopp, W., 636. Koppe, P., 378. Koppel, I., 390. Koppeschaar, W. F., 522. Korn, 0., 497. Kornor, J. A., 659. Korovaeff, T., 637. Kortright, F. L., 168, 510. Kosinenko, W., 528. Kosmann, B., 149. Koss, M., 510. Kottmayer, G., 35. Koukline, E. V., 437. Krak, J. B , 620. Krai, H., 118. Kramer, C., 361. Krasser, J. M., 599. Krauch, C., 141. Krause, G., 233, 240. Krauskoff, F. C., 534. Krauss, C., 392, 393. Kraut, K., 89, 215, 237, 394, 497, 547, 578, 582, 585, 589. Krawczynski, S., 55. Kraze, F., 440. Kreider, U. A., 237. Kreider, J. L., 553. Kreiling, P. , 664. Kreitling, P. , 35. Kretschy, M., 228. Kretzchimar, M., 60. Kreusler, U., 88, 188, 551. Kreusler, V., 98, 581. Krieger, B. , 460. Kriescher, W. C., 165. Kritschewsky, L., 188. Kronig, A., 188. Kroupa, R., 128, 333. Krug, W. H., 607. Kruger, M., 215. Kriiger, P., 237. Kruiys, M. J. van, 522. Kruss, H. , sec Kriiss, G. Kriiss, G., 200, 201, 421, 425, 508, 510, 555, 583. Kriiss, G., and Kriiss, H., 82. Krutwig, J., 326, 451, 655. Kruys, M. J. van, 60, 613. Kiibel, W., 216. Kuhlmann, F., 160. Kiihn, B., 339, 357. Kuhn, 0., 10, 25. Kuhnlenz, F. A., 626. Kulisch, P., 577. Kupferschlager, M., 596. Kiipffer, A. von, 194, 641. Kurschner, F., 271. Kusnetzoff, P., 33. Kiister, F. D., 54. Kiister, F. W., 66, 72, 279, 286, 527, 611, 615. Kuzirian," S. B. , 520, 552. LAAR, C., 484. Laby, T. H., 386. Lachand, M., 324. 750 A TREATISE ON CHEMICAL ANALYSIS. Lachs, H., 201. Ladd, E., 460. Ladenburg, A. , 288. Lainer, A., 425, 654. Lamar, W. R., 536. Lambert, A. , 579. Lampadius, W. A., 396, 657. Land, W. J., 94. Landecker, M., 419, 420, 422, 503. Landris, E. K., 228. Landsiedl, A., 551. Lane, N. J., 55, 615. Lang, J., 295. Lang, W. K., 275, 589. Langbeck, H. W., 62. Lange, K. , 472. Lange, T., 505. Lange, W., 88. Langenbeck, K., 662, 667. Langley, J. W., 547. Langley, R. 514. Langley, R. W., 366, 423. Langmuir, A. C., 366, 457, 488, 621, 622. Langmuir, F. L., 473, 474, 478. Lapierre, C., 324. Lapique, L., 200. Larsen, G., 276. Larssen, F. 55. Lasne, H., 219, 495, 606, 641. Lassayne, J. L., 339. Lathe, F. E., 351. Laudrin, E. , 528, 529. Laufer, E., 661. Laurent, C., 31. Lautemann, E., 607. Lavezard, E., 667. Lavoisier, A. L., 144. Lea, M. C., 39, 103, 437. Leach, E. A., 635. Lean, B., 288. Leather, J. W., 200. Leavitt, S., 607. Lebrasseur, A., 44. Lechartier, G., 607. Leclerc, A., 162, 207, 382, 566. Leclerc. J. A. , 607. Lecocg, E. V. , 625. Leconte, C., 490, 600. Lecrenier, A., 321, 387, 428. Ledebur, A., 195, 374, 376, 378, 382, 457. Lee, R. H., 390. Leeds, A. R., 94, 100, 117, 178, 462. Leent, F. H. van, 236, 308, 540. Leeuwen, J. D. von, 617, 622. Lefevre, C. , 284. Leffler, R. L., 362, 476. Lefort, J., 408, 417. Lehman, F., 162. Lehmann, G. W., 292. Lehmann, R., 499. Lehner, V., 123, 203, 273, 368, 435, 443, 486, 577. Lehnering, P., 451. Lehnkering, P., 187, 194. Leidie, E., 437. Leidler, P., 425. Lejeurre, M., 43, 44. Lemberg, J., 661, 664. Lendrich, K., 110, 587. Lenfield, L, 505. Lenoble, E., 73. Lenssen, E., 212, 311, 316, 451. Lenz, W., 149. Leon, J. C. y, 446. Leonhard, A., 466. Leopold, A., 663, 668. Lepez, C., 308. Lepierre, C., 604. Lepinay, J. M. de, 30. LeRoy, G. A., 55. Lescoeur, H., 578. Lesser, E., 275, 296. Lesser, H. , 283. Lestelle, H., 73. Leuba, A., 474, 476. Levi, L. E., 9, 607. Levoir, L. C., 527, 529. Levol, A., 79, 283, 317, 425. Levy, A. G., 550. Levy, L., 203, 207, 484. Levy, M., 657. Lewis, E. A., 246, 312. Lewite, A., 423, 486. Lewkowitsch, J., 586. Ley, H., 201. Leybold, W., 44. Leysaht, H., 457. Libermann, L., 339. Lidoff, A. P., 573. Lidow, J., 622. Lieben, A., 617. Liebeimann, L. , 179. Liebig, J. von, 392. Liebknecht, 0., 505. Liebreich, von, 324. Liebrich, A., 339. Liebschutz, M., 308. Lienau, H., 446. Liesse, C., 213, 522. Lincoln, A. T. 5 42, 605. Lindemann, 0., 77, 465, 480. Lindet, L., 208. Lindgren, J. M., 394. Lindhcrst, M., 660. Lindo, D., 161, 165, 232, 651. Lindt, M., 58. Lindt, 0., 624. Ling, A. R., 250. Linnemann, E., 160, 362, 499, 508. Lipowitz, A. von, 593. Lipp, F., 409. Lissner, A., 573. List, K., 159, 228. Litterscheid, F. M., 351. Little, A. F. V., 348. Little, H. F. V., 378, 382. Liversidge, A., 641. Ljubavin, N. N., 560. Locke, H. J., 162, 566, 571. Loczka, J., 641. Loeser, C., 666. Loges, G., 597. Lohhofer, W., 72. Lohmann, D., 625. INDEX OF NAMES. 751 Lohmaun, J., 441. Lohofer, W. T., 76. Lohse, 0., 102. Lomox, E. L., 621. Long, J. C., 316. Longchamp, M., 212. Longi, A., 378. Lonner, C., 286. Lonnes, C., 287. Lorom, H. T., 266, 270. Lord, N. W., 625. Lorenz, N. von, 598. Lorenz, R., 557. Lb'sekann, G., 366, 488. Loven, J. M., 182. Lovibond, J. W., 84, 340. Loviton, L. 275. Low, A. H., 226, 275, 292, 305, 334, 351, 352, 354, 367, 376. Low, H. D., 188. Low, 0., 305. Low, W. H,, 311, 577, 585. Lowe, F. A., 130. Lowe, J., 316, 323, 347, 545, 548, 551. Lowe, J. G., 579. Lowe, W. F., 178, 435. Lb'wenstein, E., 157, 575. Lowenthal, J., 308, 309, 310, 311, 316. 451. Lowry, T. M., 66. Luboldt, R., 196. Lucanus, B., 427. Lucas, M., 339, 355, 398. Lucas, T. 479. Lucchesi, A., 425. Luck, E., 218, 220, 593, 601. Liicker, F., 589. Luckow, C., 256, 395. Liidert, H. , 373. Ludewig, W., 169, 175. Ludwig, E., 172, 175, 572. Ludwig, R., 280. Luechese, L., 475. Luff, A. P. , 372. Luhrig, H., 589. Lunden, H., 31, 249. Lunge, C., 76. Lunge, G. 9, 35, 62, 64, 72, 90, 173, 178, 195, 215, 250, 288, 376, 527, 538, 551, 555, 556, 557, 582, 611, 615, 617, 625, 663, 664, 676, 678, 679, 683. Lupp, A., 402. Luther, R., 10, 18, 40. Lutshinnskii, I. I., 455. Lutz, 0., 520. Lux, F., 323. Luynes, de, 60. Luynes, M. de, 580. Luynes, V. de, 265. Lyons, R. E., 441. Lyte, F. M., 50, 228, 367. MAASS, T. A. , 445. Maass, W., 423. Mabery, C. F., 547, 625. Macagno, J., 598. Macazzini, A., 330. MacDougall, F. H., 541. Macfarlane, T , 102. Machiedo, L., 514, 515. Maclvor, R. W. E., 218, 284, 443, 460 475 Mac Kay, G. M. J., 351. Mackenzie, G. L., 268. Mackintosh, J. B., 187, 362, 545, 657. Mackintosh, J. C., 275. Maclaurin, J. S., 461. Macleod, J., 345. Maori, V., 461. Maderna, G., 595. Mager, W., 292. Magnanini, G. , 201. 579. Magri, G., 387. Mahler, P., 562. Mahon, R. A., 599. Mahon, R. W., 367. Mai, J., 479. Maillard, L., 484. Majewski, L., 267. Makinen, E., 222. Makius, G. H., 434. Malaguti, J., 658. Mallard, E., 161. Mallet, J. W., 182, 293, 408. Mallinckrodt, E., 87. Mallot, C., 601. Manasse, 0., 480. Manchot, W., 451. Mandelbaum, A.. 585. Mangold, C., 9. Manley, J. J., 6, 11. Mann, C., 77, 451, 465, 473, 551. Mannhardt, H., 58. Manning, C. R., 117. Manning, R. J., 589. Manselle, E., 529. Mansier, M., 87, 89, 179. Mar, F. W., 612, 613,615. Marburg, J., 298. Marc, R., 66. Marchand, R. F., 345. Marchese, C., 323. Marchlewski, L. P., 551, 556, 624, 679. Marcker, M., 593, 598, 600. Margosches, B. M., 223, 287. Marguerite, F., 198, 407. Marie, C., 336. Marignac, C., 161, 189, 407, 409, 410, 419, 421, 499, 508, 583, 589. Markoonikoff', V., 219. Marriott, W. M'K., 201. Marrs, L. E., 382, 466. Marsh, C. W., 615, 619. Marshall, H., 110,382. Martin, H. G., 649. Martin, M., 273. Martinon, M., 484. Martius, L., 113. Marx, C., 46. Maschke, 0., 406, 597, 663. Maskelyne, N. S., 160. Mason, C. D., 165. Mason, H. P., 102, 620. Mason, W. P., 527. Massignon, J., 474. Masson, L, 573. 75 2 A TREATISE ON CHEMICAL ANALYSIS. Masson, 0., 47. Mastbaum, H., 221. Mathieu, L., 286,287. Matignou, C., 407, 481, 484. Matthews, J. A., 367. Matthews, J. M., 457, 499. Matzurke, G., 361. Maumene, E. J., 391, 396. Mauzelius, E., 124. May, W. C., 335. Mayard, M., 528. Mayencon, M., 162. Mayer, A., 552. Mayer, E. W., 378. Mayer, F., 362, 469, 569. Mayer, 0., 201. Mayer, T., 366. Mayer, W., 563. Mayerhofer, J., 278. Mayrhoffer, J., 524. Mays, K., 60. Mazeon, K. N. , 429. M' Arthur, J., 234, 240, 241. M 'Bride, R. S., 342. M 'Bride, S., 194. M'Cabe, C. R., 118, 119. M'Caffrey, C. T., 212. M'Clure, C. H., 632. M'Coy, L. R. W., 115, 276, 278, 292, 294. M'Coy, H. N., 290, 491, 492. M'Cracken, R. F., 382. M'Crae, J., 182. M'Crudden, K. H., 213. M'Culloch, N., 399. M'Elroy, K. P., 607. M'Kenna, A. C. F., 179. M'Kenna, A. G., 372, 374, 409. M'Kenna, A. J., 190. M'Kenna, C. T., 122. M'Mahon, J. H., 33. M'Namara, W., 333. M'Neil, H. C., 573. Mdivani, B., 409. Meade, 57. Meade, A. H., 595. Meade, J. K., 119, 330, 350. Meade, R. K., 55, 94, 519, 522. Mecklenburg, W. , 496. Medri, L., 270. Medway, H. E., 337. Meeker, G. H., 176. Meiklejohn, C. A., 267. Meimberg, E., 421, 422. Meinecke, C., 34. Meineke, C., 58, 170, 172, 175, 178, 194, 200, 286, 287, 288, 290, 361, 372, 373, 378, 393, 453, 590, 594, 595, 611. Meissner, C., 35, 43, 44. Meister, J., 148. Meitzendorff, 351. Meker, G., Ill, 115. Meldrum, R., 172. Melikoff, P. G., 595. Melliss, M., 499. Mellor, J. W., 60, 84, 117, 122, 169, 172, 200, 231, 247, 274, 308, 323, 360, 379, 397, 444, 452, 486, 487, 574, 630, 662, 671, 674. Mellquist, H., 441, 442. Meloche, C. C., 368. Memrainger, C. G., 116. Alendeleeff, IX, 30. Alene, C., 310, 602. Mengin, A. H., 305. Mengin, J. H., 305, 307. Alenil, A. P. J. du, 583. Menneke, F. A., 175. Mennicke, H., 615. Menschutkin, N., 46. Mercier, A., 234. Merck, E., 141. Merrill, L. T., 305. Mertens, K. H., 320. Mervin, H. E., 644. Merwin, H. E., 203. Merz, V., 498. Meschezerski, J., 517. Messinger, J. J., 547. Metzger, F. D. 420. Metzger, F. J., 382, 408, 409, 466, 510, 511. Metzger, P., 128. Metzl, A., 296. Metzner, G., 441. Metzyer, P., 55. Meyer, A., 59. Meyer, A. V., 167. Meyer, E. von, 617. Meyer, L., 493. Meyer, 0. von, 425. Meyer, R., 417. Meyer, R. J., 419, 496, 497, 501, 511. Meyer, T., 607. Meyer, V., 116, 575. Michel, E., 528. Michel, F., 461. Michaelis, P., 280, 282, 283, 284, 297, 343, 411, 413, 441. Michaelis, W., 462, 663. Middel, T.,6. Milbauer, J., 34, 355. Milford, L. R., 188, 534. Millberg, C., 684. Miller, E. H., 190, 270, 351, 358, 367, 368, 412, 415, 416, 434, 436. Miller, I., 455. Miller, W. H., 6, 17, 23. Milobendski, T., 38. Milon, H., 266. Minar, J. C. , 639. Minozzi, A., 271. Mitchell, J., 270. Mitchell, W. L., 189. Mitscherlich, A., 160, 177, 184, 188, 214, 224, 225, 235, 239, 317, 462, 475, 515, 520, 547, 611, 612. Mitscherlich, E. A., 540, 541. Mittasch, A., 361, 516. Mittentzwey, M., 189, 614. Mixer, C. T., 378. Mixter, C. T., 452. Mixter, W. G., 612, 624, 625. Mohr, 33. Mohr, C., 36, 218, 600. Mohr, C. F., 288. INDEX OF NAMES. 753 Mohr, F., 31, 44, 46, 58, 61, 70 f 79, 89, 160, 195, 228, 234, 286, 290, 305, 310, 330, 354, 381, 418, 462, 655. Moir, J., 429. Moissan, H., 547, 563, 585. Moissenet, I.., 307. Molar, C., 16. Moldenhauer, F., 168, 320, 368, 369. Mollins, J. de, 98. Monnet, P., 481. Montemartini, C. , 585. Moody, H. R., 177. Moore, R. B., 455. Moore, T., 266, 270, 311, 362, 374, 386, 399. Moraht, H., 200, 201, 583. Morawski, T., 378. Morgan, F. H., 58, 310, 455. Morgan, G. T., 188, 280, 555. Morgan, J. J., 209, 355, 409. Moride, M., 287. Morley, E. J., 118. Morozewicz, J., 232. Morrell, T. J., 626. Morrell, T. T., 200, 396, 539. Morse, H. N., 36, 40, 196, 474, 475, 516,583. Morton, D. A., 178, 622. Morton, H., 82. Mosander, G., 506. Moser, L., 351, 353, 514, 515. Mostowitsch, W., 616. Moulin, A., 474. Moyaux, L., 187, 195. Muck, F., 94, 102, 183, 220, 373, 597, 624, 653. Muencke, R., 42, 148. Mueneke, C., 91. Muhlhaeuser, 0., 562, 563, 564. Muhlhauser, 0., 365. Muhs, G., 182, 211. Muir, M. M. P., 46, 189, 355. Mukerjee, B. M., 59. Mulder, A., 36. Mulder, G. H., 225. Mulder, G. J. , 650, 652. Mulder, J., 38. Miiller, A., 72, 160, 196, 286, 355, 396, 451, 474, 497, 5)1, 547, 661, 663. Miiller, E., 188, 194, 302, 378. Miiller, E. R. E., 166, 595. Miiller, F., 410. Miiller, F. C. G. , 382. Miiller, G., 42, 56. Miiller, J., 367. Miiller, J. A., 268, 270, 296, 311, 333, 475, 600, 664. Muller, K., 593. Miiller, L., 291. Miiller, M., 441, 443. Muller, W., 268, 391, 627. Munktell, J. H., 88. Munro, J. M. H., 602. Munroe, C. E., 107, 374, 630. Murmann, E., 103, 106, 113, 179, 212, 213, 214, 238, 285, 319, 364, 367, 373, 390, 534, 613. Musculas, M., 286. Musset, b\, 35. Musset, Z ., 288. Muter, J., 46. Muthmann, W., 441, 511. Miitnianski, M., 286. Mylius, F., 44, 144, 286, 357 431, 437, 457. NAME, R. G. von, 350. Namias, S. 409, 475. Napier, J. 434. Naquin, W. P., 107. Natanson,J., 201. Naumann, G., 364. Naumann, K., 62, 63. Naylor, A. R., 193, 195. Neher, E. 275. Neher, F., 276, 278.' O'Neill, G., 475. I Neisch, A. C., 510. Neitzel, E.. 66. Nelsmann, H., 310. Nestler, A., 53. Neubauer, C., 593, 601. Neubauer, H., 107, 215, 218, 598, 600. Neujean, A., 319. Neumann, A., 534. Neumann, B., 207, 256, 336. Neumann, G., 192. Neurath, F., 604, 606. Neustadl, L., 517. Newberry, S. B., 522. Newlands, J. A. R , 7. Newton, H. D., 189. Niccoli, L., 429. Nichols, W. R., 178. Nicholson, E., 226, 555. Nicholson, E. C., 614. Nickles, C., 98. Nickles, J. 612. Nicolardot, P., 406. Nihoul, K, 80. Nilson, C. F., 278. Nilson, L. F., 233, 236, 421, 501, 506, 510. Nissenson, H., 301, 336, 364, 369, 455. Noad, H. M., 252, 614. Noda, I., 612. Nolly, H. de, 563. Norblad, J. A., 488. Norman, J. T., 674. Norris, G. L., 374, 457. Norris, J. F., 441. North, B.,193, 195,196. Norton, J. T., 352, 495. Novoluy, K.,72. Noyes, A. A., 182, 275, 316, 457, 483, 484, 517. Noyes, W. A., 206, 241, 378, 415, 599. Nutzinger, M., 593. Nydegger, 0., 475, 627. OBBEKE, K., 657. Oddo, G., 529. Oddy, R. W., 196. Odling, W., 280. Oeffinger, H., 212. Oehmichen, P., 435. Oertling, L., 4. Oerum, H. P. T., 200. 48 754 A TREATISE ON CHEMICAL ANALYSIS. Oettel, F.. 42 262, 270, 364, 646. Offerhaus,'0., 560. Offermann, H. , 644. Ogilvie, J. P., 96. Ogilvie, T. E., 216, 218. Ohl, W., 395. Ohly, J., 497, 600. Olsen, J. C., 373. Oordt, G. van, 448, 449. Opificus, L., 240, 322, 323. Ordway, J., 75. Orlowsky, J. von, 351. Orthey, M., 378. Osaka, Y., 644. Osborne, N. S., 32, 38, 3i. Osborne, T. B., 203, 418. Osmond, F., 382, 599. Ost, H., 31, 178, 312, 636. Ostwald, W.. 10. 18, 39, 40, 94, 212, 529, 611, 614, 651. Otto, R., 149. Oudemans, A. 0., 683. Oudesluys, C. L., 475. PADT, L., 87. Padoa, M., 274. Pagasogli, G., 10. Page, R. W., 358. Pagliani, S., 200. Fagnoul, A. 604. Paguireff, von, 212. Palmer, C. S., 179. Pamfil, G. P., 203. Panajotow, G., 275, 306. Panayeff, J. von, 497. Pannertz. B., 55. Pannsetz, F.. 55. Panpushka, S., 228. Pan ten, G., 355. Pappada, N. 96. Parker, H. G., 611. Parker, J. S., 454. Parmentier, F., 580. Parnell, E. W., 216, 284, 495. Parr, S. W., 267, 350, 385, 394, 562, 564, 566, 622, 632, Parr, T. W., 622. Parry, J., 355, 409. Parsons, C. L., 148, 448, 449, 527. Partheil, A. 322, 323. Passerini, N". , 604. Passon, M., 522. Patera, A., 305, 48 S. Patten, H. P., 177. Patterson, G. W., 475. Pattinson, H. S., 367. Pattinson, H. S., see Pattinson, J. Pattinson, J., 93, 250, 619. Pattinson, J., and Pattinson, H. S., 275, 372, 374, 376, 594. Pattison, M. M., 506. Paul,T., 104, 297, 324. Pavec, A., 600. Pawolleck, B., 58, 478. Pay en, A., 355. Payne, H. L., 29, 40. Peake, W. A., 546. Pearce, R., 292. Pearce, S. H., 135. Pearson, A. H.. 478, 612. Pease, F. K, 190, 599. Pechard, E., 417. Peckham, H. E., see Peckham, S. F. Peckham, S. F., 176. Peckham, S. F., and Peckham, H. E. 622. Pederson, H., 394. Peirce, A. W., 441. Peitzsch, B., 593, 598. Peligot, M., 231, 232. Pelleni, G., 442, 606. Pellet, H., 73, 80, 207, 474, 595, 615. Pellet, K., 310. Pelouse, J., 200. Pelouze, T. J., 339. Pemberton, H., 599. Pence, F. G., 668. Penfield, S. L., 103, 183, 423, 449, 570, 585 589, 639, 644. Pennington, M. E.. 419, 420. Pennock, J. D., 178, 622. Penny, F., 453. Peppel, S. V., 170. Perkin, F. M., 148, 256, 452, 540. Perkins, G. E., 354. Perl, L., 475. Perman, E^-P., 352, 623. Personne, M. J., 344. Persoz, J., 554, 559. Peters, A. M., 626. Peters, A. W., 351. Peters, C. A., 215, 452. Peters, S., 382. Petersen, E., 51. Petersen, T., 595. Petreciolli, 0., 296. Petren, J., 441. Petrzilka, H., 116. Pettenhofer, M., 551. Petterson, 0., 556. Petzholdt, A., 520, 521. Pfaff, C. H., 339, 347, 354. Pfaff, F., 525. Pfaff, S., 525. Pfeil, K., 162. Pfennig, F., 399. Pfordten, F. 0. von der, 189. Pfordten, 0. van der, 149, 408, 413, 599. Pforzheim, 5. Pfliiger, E., 47. Phelps, M. A., 226, 280, 283. Phelps, R., 546. Phillips, F. 0., 43, 160, 226, 475. Phillips, H. J., 615. Phillips, J. A., 175, 185. Philossophoff, P., 528. Phinney, J. I., 6 11. Phiiiney, J. J., 615. Phipson, T. L., 305. Picard, P., 382. Pickering, E. C., 18. Pickering, S. U., 290. Pickering, U. S., 167, 352, 372. Pierre, C. St, 428. INDEX OF NAMES. 755 Piesse, 0. H., 103. Pieszczek, E., 322. Piloty, O., 280, 284. Pinagel, A., 167, 410, 418. Pincus, C., 600. Piucrua, E., 457. Pinnow, J., 286. Pipereaut, P., 364. Piria, R., 614. Pisani, F., 189, 351, 363, 415, 488, 490, 497. Plaats, J. D. van der, 118. Planes, M., 348. Plato, W.. 282. Flatten, F., 282. Platz, J., 364. Pleisch, 307. Pleissner, M., 315. Pohl, R., 208, 469, 498. Poleck, T., 473. Pollaci, F., 287. Pollard, W. B., 267. Polenske, E., 585, 587. Polstorff, K., 279. Poole, A. C. D., 43. Pope, F. J., 209. Popel, M., 528. Popp, 0., 215, 218, 506, 589. Post, J., 630. Pott, C., 285. Potyka, J., 161. Pouget, J., 604. Pouget, M., 366, 367. Powell, N. S., 525. Poynting, J. H., 11. Pozzi-Escot, M. E., 58, 373, 407, 429. . Pratt, J. H., 462. Pratt, J. W., 116. Precht, H., 231, 232, 234, 235, 237, 240, 241. Preller, I., 524. Prescher, J., 580, 585. Prescot, A. B., 249. Prescot, A. B., and Sullivan, 339. Preusser, J., 409, 410. Prilram, R., 42, 183. Price, D. S., 614, 617. Price, T. M., 577. Pringel, A., 167. Pringsheim, H. H., 267. Prinzl, A., 33. Priar, T., 419. Prister, A., 429. Privozink, E., 617. PHwoznik, E., 426, 435. Proctor, B. S., 6. Proctor, H. R., 14, 23, 55, 87. Prossel, B., 672. Prost, E., 359, 367, 641. Prothiere, E., 278. Puckner, W. A., 103, 291. Pukall, W., 630. Puller, R E. 0., 282, 283, 284. Pullman, 0, S., 491. Pulsifer, H. B.. 200. Purgotti, A., 187. Purgotti, E., 427. Piischel, A., 60. Pyliala, E., 455. QUENEAU, A. L., 563. Quennessen, L,, 111, 115. Quennssen, M., 437. RAAB, H., 373. Raalte, A. van, 314. Raaschou, P. E., 90. Raben, E., 596, 598. Radau, C., 472. Radlowski, K. von, 21i. Raffa, E., 218, 283. Raikow, P. N., 372. Ramage, H., 382. Rammelsberg, 56, 57. Rammelsberg, C., 160, 162, 278, 284, 305, 409, 410, 415, 462, 472, 480, 488, 504, 583, 593, 663. Ramsay, 16. Ramsay, W., 624. Randall, D. L., 190, 452, 599. Randall, W. W., 372. Ransome, F. L., 465. Rapalje, W. S., 373. Raquet, M., 517. Raschig, F., 193, 218, 627. Rath, G. von, 341, 480. Rathke, B., 440, 441. Ratner. C., 296. 307. Raulin, G. , 399, 585. Rawlins, H. J. B., 310, 311. Rawson, S. G., 515. Raymond, E., 55. Rebbufat, 0., 527. Reckleben, H., 148. Reddrop, J., 146, 382. Redpath, G. C., 367. Reed, S. A., 130, 135. Reese, C. L., 196. Regnault, V., 118. Reich, E., 292. Reich, K., 208, 292. Reichard, C., 203, 418, 484, 595, 598. Reichardt, E., 9, 478, 597, 602. Reichel, F., 283, 284. Reid, A. F., 34, 59. Reidenbach, R., 163, 215. Reik, R., 516. Reindel, F., 367. Reinhardt, C., 100, 165, 226, 286, 372, 378, 451, 475, 557, 600. Reinitzer, B., 33, 81, 362, 469. Reinsch, E. H. E., 273. Reis, A. von, 547. Reis, M. A. von, 283, 312, 399, 557, 600. Reischauer, C. , 568. Reischle, A. K., 585. Reiss, M. von, 590. Reitmar, 0., 188. Remele, A., 390, 462, 488. Remmler, W., 624. Remont, A., 116. Rempel, H., 277. Rempel, R., 65. Renard, A., 367, 368. Rennor, 0., 636. Rennie, H. W., 268, 270. 756 A TREATISE ON CHEMICAL ANALYSIS. Resch, F., 58. Reuss, W., 149. Reuter, M., 296. Revaud, A., 278. Rey, H., 9, 35, 678. Reymann, S., 372. Reynolds, E. .!., 155. Reynolds, W. G., 441. Rheineck, H., 362, 488, 600. Riban, J., 201, 256,278, 425. Rice, E. W., 61. Rich, E. M , 662. Richard, R. H., 478. Richards, -E. S., 395. Richards, J. W., 254, 328, 434, 525. Richards, T. W., 17, 31, 40, 97, 104, 107, 213, 520, 611, 615, 632, 651. Richardson, C., 250. Richardson, F. W., 473. Richardson, G. M., 682. Richardson, W. D., 250, 343. Richarz, F., 10. Richmond, H. D., 20, 372. Richter, W., 527. Richters, E., 592, 593, 595. Rickmann, R., 306. Rideal, S., 474, 475. Ridout, R. H., 58. Rieckher, 280. Rieke, R., 190. Ries, H., 669, 674. Riggs, R. B., 362, 473. Riley, E., 169, 203. Rimbach, E., 500. Rinne, F., 664. Ripper, M., 577, 615. Risdale, C. H., and Risdale, N. D., 250. Risdale, N. D., see Risdale, C. H. Ritthausen, H.. 560. Rivat, G., 287. Rivett, A. C. D., 5*3. Rivot, L. E., 350, 410, 437. Roberts, C. F., 195. Roberts, C. M., 129. Roberts, E. J., 506. Roberts, R. W., 573. Robertson, I. W., 621, 625. Robin, H., 517. Robin, L., 580. Robinson, H. L., 112, 468, 506. Robinson, W. 0., 449. Rocholl, F., 160 Rocholl, H., 162, 460. Rodwell, G. F., 187, 315, 320. Roer, E., 209. Rogers, H. F., 184. Rogers, H. R., 339, 340. Rogers, R. E., and Rogers, W. M., 546, 547. Rogers, W. M. , see Rogers, R. E. Rohland, P., 232, 235, 529. Rohmer, M., 280, 299. Rohn, E., 441. Rohn, W., 593, 598. Rohrbeck, 33. Roland, L., 551. Rollin, G., 299. Rontgen, A., 641. Roode, R. de, 564, 593. Roos, J. 0., 613. Rosanotf, M. A., 78. Roscoe, H. E., 481, 483. Rose, F., 212, 614, 651. Rose, H., 115, 160, 161, 178, 184, 225, 226 229, 270, 276, 279, 282, 283, 284, 296, 308 309, 317, 318, 320, 345, 348, 357, 371, 373 389, 390, 409, 412, 417, 420, 440, 441, 451 462, 470, 475, 488, 504, 509, 514, 515, 520 554, 578, 583, 585, 596, 608, 611, 614. 637 650, 660, 662, 663. Rose, J. A., 578. Rose, T. K., 20, 77, 429, 434, 435. Rose, J. D., 55. Roseblum, S., 474, 475. Rosenbladt, T., 390, 578, 589. Rosenhain, W., 116. Rosenheim, A., 200, 201, 364, 390, 396, 418 443, 502, 510. Rosenheim, R., 441. Rosenlecher, H., 9. Rosenstein, M., 351. Roseiithal, G., 361. Rosenthaler, L., 294. Ross, W. H., 103. Rossing, A., 296. Rossler, C., 305, 399. Rossler, H., 434. Rossler, L., 425. Rossler, 0., 552. Rost, C. 0., 382. Roth, J., 520, 525. Rothe, J. W., 457. Rothenbach, F.. 418. Rothmund, V., 77. Rottger, F., 232. Roussin, A., 305. Rowell, H. W., 301. Royse, J. S., 599. Rubies, S. Pina de, 462. Rubricus, H., 373, 382. Rudnick, P., 44. Rudoif, G., 362. Rudorff, C., 101. Riidorff, F. , 260, 262. Ruediger, A. P., 451. Rueginberg, M., 416. Rueprecht, A., 4. Ruer, R., 501. Ruff, 0., 419, 421. Riihl, F., 479. Riimpler, A., 351, 361. Ruoss, 73. Rupp, E., 162, 215, 323, 367, 399, 429, 466 636. Ruppert, F., 517. Ruppin, E., 611, 613. Riirup, L., 374, 378. Russell, T. H., 119. Russell, W. J., 391. Russmann, A., 517. SABECK, A., 659, 663, 664. Sacher, J. F., 33, 34, 323, 332, 333, 611, 613 Sadtler, G. P., 540. Sad tier, S. P., 390. INDEX OF NAMES. 757 Sadtler, S. S., 559, 568, 621. Safarik, A., 418. Saint Pierre, C., 115. Salkowski, E., 363. Salm, E. , 60. Salvador!, R., 606. Salzer, T., 593. Sammet, C. F., 279. Sammis, J. L., 55. Samter, V., 136, 502. Sanchez, J. A., 310, 311, 312, 455. Sand, H. J. S., 244, 312, 335. Sander, C., 59. Sariger, G. A., 78. Saniter, E. H., 372, 374, 376, 475. Santos, J. R., 293. Sargent, C. M., 102. Sargent, G. W. , 377, 399, 457, 585, 588. Sarkar, A. C. , 481. Sarnstrom, C. G., 378. Sartori, A., 636. Sartorius, F., 4, 5. Saner, A., 55, 58, 103, 144, 547, 624, 625. Saner, E., 563. Saul, J. E., 429. Saunders, L. E., 630. Savillier, M., 359. Schaak, M. P., 582, 589. Schabur, K., 453. Schach, J., 641. Schseffer, A., 119. Schafer, 441. Schafer, W., 61. Schaffgotsch, F. C., 244. Schaffgotsch, F. G., 226, 578. Schafhautl, M., 525. Scheel, K., 29, 685. Scheele, K. W., 407. Scheen, 0., 312. Scheerer, T., 159, 203, 212, 213, 296, 362, 465, 499, 504, 614, 663. Scheffler, W. , 646. Scheibler, C., 40, 197, 408, 425, 555. Scheibler, F., 407. Scheitz, P., 61. Schellbach, P., 34. Schenck, F. C., 527. Schenke, V.,237. Scheringa, K., 97. Sclierubel, E. F., 324, 343. Scheurer-Kestner, A., 115, 311, 314. Schieffelin, W. J., 536. Schiff, H., 228, 278, 585, 641. Schiff, S., 55, 56, 551, 623. Schiffer, P., 600. Schild, M., 595. Schillbach, H., 622. Schiller, E., 419, 421. Schilling, H., 364. Schimidzu, T., 150. Schindler, C., 333. Schihdler, L., 415. Schirm, E., 179, 455. Schirmeister, H. , 209. Schleicher, C., and Schiills, 61, 620. Schleier, M., 455. Schlesinger, B. E., 497. Schloesing, T., 237. Schlosser, W., 29, 30, 32, 35, 38, 39. Schmatolla, 0., 289. Schmelck, L., 113. Schmidt, C., 525. Schmidt, E , 299, 301. Schmidt, H., 218. Schmidt, M. R., 382. Schmidt, W. , 55. Schmorger, M., 221. Schneider, A., and Gaab, 585, 589. Schneider, Ed., 97. Schneider, E. A., 611. Schneider, F. C., 280. Schneider, L., 382, 409, 511. Schneider, R., 372. Schober, J. B. , 77. Schobig, E., 390. Schoch, C., 521. Schochor-Tscherny, M.. 664. Schceffer, J. A., 321. Schoffel, R., 378, 409, 545. Scholer, G., 551. Scholes, S. R., 144, 285. Scholl, A , 58. Scholtz, M., 64. Scholz. H. A., 267, 317. Schonbauer, J. A., 657. Schonn, H., 203, 204, 406. Schott, F., 528, 529. Schottlander, P., 26. Schranz, W., 194. Schreiber, H., 622 Schreiner, 0., 82, 146, 539, 578, 579, 604, 605, 630. Schroder, K., 58, 196, 453. Schroeder, J. , 534. Schrotter, A., 214, 568. Schrotter, A. R. von, 553. Schubert, A., 500. Schubert, F., 55. Schiick, B., 394, 399. Schulatschenko, A., 528. Schnlerud, L., 478. Schiills, see Schleicher, C. Schulte, W., 304, 305. Schulten, M. de, 357. Schultze, C. H., 215. Schultze, E. H., 521. Schultze, H., 201. Schultze, J., 364. Schulz, F., 369. Schulze, A., 38. Schulze, F.,361, 555. Schumann, C., 218, 601. Schiitz, W., 214, 616. Schiitze, R., 561. Schiitzenberger, P., 116, 506. Schuz, J., 102, 103. Schwarz, C., 579, 580. Schwarz, C. L. H., 286, 310. Schwarz, H., 476, 563. Schwarz, R., 669. Schwarzberg, P., 387. Schwarzenberg, P., 471. Schweder, T., 395. Schweitzer, A., 511. 758 A TREATISE ON CHEMICAL ANALYSIS. Schweitzer, H., 8, 9. Schweitzer, P., 515, 517, 594. Schwezoff, B., 287. Schwirkus, G., 5. Schwirkus, P., 3, 6. Scott, W. L., 245. Scudder, H., 586. Seaman, H. J., 116. Seamon, W. H., 188, 367. Sebel, J., 59. Seemann, F., 638, 639. Seemann, L., 425. Sefstrom, G. 5 483. Seger, H., 59, 72, 527, 528. Seger, H. A., 633, 658, 662, 663, 666, 674. Sehnal, J., 315. Seidel, T., 280. Seiler, F., 200. Seitter, E., 193. Seligman, R., 150. Seligmann, R., 454. Seligsohn, M., 413, 593, 594. Sell, W. J., 473, 478. Sellach, C. S., 315. Selleck, C. C., 614. Senft, F., 665. Senier, A., 579. Serullas, M:, 237, 534. Sestini, F., 55. Setlik, B., 410. Seward, H., 167. Seyberth, H., 594. Seyda, A., 201. Shade, J. W., 141. Sharpies, S. P., 297. Shedd, 0. M., 541. Shengel, J. C., 305. Shepherd, H. H. B., 607. Sherman, H. C., 595. Sherman, H. S., 119. Shilton, A. J., 200. Shinier, P. W., 166, 176, 190, 560, 593, 620. Shinn, F. L., 441. Shiver, F. S., 237. Shuch, J., 646. Shutt, F. T., 77. Sidersky, D., 55, 555. Siedler, P., 301. Siegle, E., 611, 612, 614. Siemens, F., 144. Silberberger, R., 611. Silveira, A. A. da, 499. Simmermacher, W., 603. Simmonds, C., 125. Simon, A., 59. Simon, C.. 91. v Simond, F., 196. Simpson, E. S., 266, 419, 421, 423. Sinnatt, F., 287. Sipb'cz, L., 572. Sire, G., 55. Sjollema, B., 232, 669. Skertchly, W. P., 575. Skey, W., 167, 169, 183, 201, 385, 396, 456. Skimose, M., 440, 441. Skinner, F. F., 59. Skinner, W. W., 536. Skirrow, F. W., 578. Skrabal, A., 194, 195, 451, 517. Slawik, P., 378, 471, 481. Sleeper, J. F., 612. Sloane, T. 0., 611. Smeaton, W. G., 82. Smith, E. E., 200. Smith, E. F., 162, 189, 226, 256, 305, 34{ 358, 409, 413, 416, 419, 420, 475, 50$ 580. Smith, E. L., 59. Smith, F., 283. Smith, H. P., 373, 460. Smith, J. C., 103. Smith, J. D., 607, 611. Smith, J. D. A., 355. Smith, J. G., 604. Smith, J. L., 10, 100, 161, 162, 163, 184 223, 240, 475, 560. Smith, L., 418. Smith, R. G., 96. Smith, R, 0., 335, 337, 338. Smith, T., 301. Smith, T. 0., 60, 290, 510. Smith, W., 50, 72, 528, 611. Smith, W. A., 66. Smith, W. R., 73. Smithen, F. W., 156. Snelling, W. 0., 44, 107. Snyder, F. T., 133. Soderlund, A. M., 55. Sofianopoulos, A. J., 228. Sonneschein, L., 590, 593. Sonstadt, E., 114, 212, 234, 429. Sorensen, S. P. L., 72, 193. Sorge, 58. Sorray, P. de, 371. Souchay, A., 168, 183, 212, 214, 320, 480. Soukup, A., 455. Source, M. de la, 88. Soxhlet, F., 59, 102. Spanjer, 0., 529. Spears, E. B., 182, 457, 484. Spectator, 118. Speller, F. N., 457. Spelta, E., 442. Spencer, A. E., 499. Spence, G. L., 630. Spengel, A., 428. Sperry, E. S., 585, 589. Speter, M., 496. Spiegel, L., 62, 445. Spielmann, 566. Spiess, G. A., 593. Spiller, J., 363, 435, 614. Spindler, 0. von, 40, 585. Spitz, G., 579. Sponholz, E., see Sponholz, K. Sponhozl, K., and Sponhol?,, E., 533. Spooner, and Bailey, 156. Sprague, C. B., 621. Spring, L. V. W., 394. Spring, W., 340, 551. Spuller, J. , 475. Squibb, E. R., 55. Stadel, W., 315. Stadler, G., 547. INDEX OF NAMES. 759 Staehler, A., 107. Stahl, K. F., 169, 648. Stahl, W., 330. Stammer, C., 315, 355. Stanek, V., 355. Stanford, E. C. C., 251, 568. Stanger, W. H., 175, 521. Stanichitch, M., 382. Stanton, F. M., 621. Starck, G., 226, 638. Starkenstein, E., 601. Stas, J. S., 10, 144, 150, 224, 241, 288, 651. Stead, J. E., 280, 475, 495. Steen, 0., 305, 318. Steffan, A., 471. Stefko, V., 475. Stegmiiller, P., 315. Stehman, J. V. II., 560. Stehmann, F. U. R., 621. Steiger, G., 82, 124, 203, 204, 644. Stein, A., 55. Stein, S., 8. Steinle, 0., 128. Steinlen, R. L., 223. Steuer, B., 529. Stewart, C. M., 118. Stierlen, R., 615. Stillman, J. M., 244. Stillmann, T. B., 278. Stingl, J., 378. Stbber, F., 197. Stock, A., 280, 284, 580. Stock, W. F., 453, 460. Stock, W. F. K., 621. Stock, W. T. K., 189. Stockmann, C., 165, 361. Stocks, H. B., 286. Stoddart, C. W., 621. Stoddard, J. T., 128. Stoermar, M., 171, 173, 187. Stotfart, J. F., 100. Stoicoff, A., 453. Stokes, H. N., 201. 465. Stoklasa, J., 156. Stolba, F., 79, 102, 107, 115, 157, 161, 162, 166, 229, 305, 330, 441, 554, 569, 589, 611, 612, 641, 644, 650. Stolberg, C., 213, 522. Stone, F. B., 348. Stone, G. C., 177, 278, 366, 367, 369, 378. Stone, W. E., 527. Stookay, L. B., 480. Storch, A., 190. Storch, L., 188, 488. Storer, F. H., 330, 475, 478, 484, 547, 552, 612. Strecker, W., 600. Streit, G., 208. Strigel, A., 237. Strohmer, F., 602. Stromeyer, A., 310, 311, 389, 578, 583. Stromeyer, C. E., 136. Stromeyer, F., 286, 495, 514. Stroschein, E., 59, 94. Struckmann, C., 525. Struve, H., 218, 315, 590, 614. Stuart, A. P., 169. Stuart, A. T., 77. Stuhl, M., 94. Stull, W. N., 87. Stumpf, J., 55. Stunkel, C., 216. Stunkel, E., 592. Stutzer, A., 59, 188, 595. Sugiura, K., 352. Suida, W., 162, 184, 462. Sullivan, see Prescot, A. B. Sulzer, R., 579. Sundstrom, C., 622. Sundstrom, K. J. , 521. Surr, G., 237, 529, 540. Sutton, F., 44, 46, 600. Svanberg, L., 590. Swarts, T., 143. Swett, 0. D., 107, 730. Sworn, S. A., 354. Szathmary, L. von, 617. Szell, L. von, 594. Szterkher, E., 323. TAGGART, W. T., 413. Talbot, H. P., 167, 177, 635. Talbot, J. H., 410. Tambon, J., 370. Tamm, A., 595. Tamm, H., 350, 362, 366, 475. Tammann, G., 637, 639. Tananaeff, N., 579. Tankard, A. R., 579. Tarugi, N., 305, 654. Tasilly, E., 201. Tate, G., 328, 475. Tatlock, R. R,, 38, 200, 235, 250, 670. Taurel, M., 446. Taylor, C. E., 420. Taylor, E. R., 102, 297. Taylor, W. E., 182. Tcharviani, and Wonder, M., 499. Teclu, K, 110. Teed, F. L., 339. Tempany, H. A., 96. Tenax, B. P., 672. Terrell, A., 195, 377, 405, 467, 469. Teschemacher, E. T., 607, 611. Texter, 0., 219. Thaddeeff, C., 583, 589. Theile, J., 295, 297. Thiel, A., 305, 364, 611, 615. Thiele, E., 44. Thiele, G., 148. Thiele, H., 35, 66, 106, 195, 278. Thiercelin, M. , 655. Thiesen, M., 5, 29. Thill, J.,176. Thilmany, A., 55. Thilo, E., 595. Thomas, A., 386. Thomas, N. W., 451. Thomas, W. S., 378. Thomaschewski, P., 102. Thompson, G. W., 319, 321. Thompson, J. P., 457. Thomson, J. E.,226. Thomson, A., 200. 760 A TREATISE ON CHEMICAL ANALYSIS. Thomson, R. T., 62, 63, 76, 579, 580, 606. Thomson, T., 228, 504. Thorin, E., 638. Thorner, W., 555, 557. Thornton, W. M., 112, 207, 208. Thorpe, T. E., 11, 125, 382, 669. Thoulet, J., 177. Thresh, T. C., 348. Thugutt, S. J., 124. Thuringer, V., 394, 395. Tichborne, C. R. C., 100. Tietrjens, L., and Apel, 234. Tighe, A., 421. Tillmans, J., 72. Timby, T. G., 461. Tischkoff, P., 372. Tissandier, G., 319. Tollens, B., 46, 221. Tolleus, B., 625. Tolmacz, B., 59. Tomei, M., 528. Tomiczek, F., 276, 278, 295. Tommasi, D., 354. Tompkins, H. R., 475. Tookey, C., 305. Topf, G.,289, 290, 311,323. Topsoe, H., 680. Torrey, J. } 342. Trabert, A., 245. Trautmann, W., 472, 483, 488, 497. Treadwell, F. P., 10, 31, 39, 112, 115, 119, 157. 182, 195, 199, 232, 290, 292, 335, 342, 385, 395, 396, 412, 462, 480, 552, 615, 634, 638, 646. Trenkner, M., 435. Treubert, F. , 425. Tribe. A., 188. Trickett, A. B., 180. Trillat, A., 339. Trillat, J. A.,234. Trnke, R., 234. Troilius, M. , 623. Troost, L., 184. Trowbridge, P. F., 541. Truchott, P., 144, 169, 497, 615. Tscheihwile, P. A., 196. Tschenschmer, E., 139. Tschijewski, P., 578. Tschilikin, M., 409. Tschischikow, A., 520. Tschugajeff, L., 385, 394. Tsukakoski, U., 351. Tsukerman, D., 579. Tucker, A. E., 670. Turner, E., 23, 611. UEHLING, E. A., 54. Uelsmann, H., 89, 597. Uhlig, E. C., 448. Ukena, M., 376, 600. Ulbricht, R., 8. Ulex, G., 235, 617. Ulke, T., 215. Ullgren, C., 283, 305, 354. Ullgren, E. , 546. Ullmann, H. M., 203. TJpson, F. W.,622. Urbain, G., 508. Utz, F., 211, 335. VACHER, A., 61. Vadam, M., 579. Vandervoort, H., 625. Vandevyver, M., 36. Vanier, G. P., 59, 169. Vanino, L.,193, 425. Vanquelin, L. K, 45, 657. Varenne, E., 390. Vaubel, W., 579. Veazey, B. H., 32, 38, 39. Veitch, F. P., 180, 604, 605, 607. Veksin, P., 474. Veley, V. H., 528. Venable, F. P., 455, 499, 579. Verbeck, A. T. H., 18. Verda, A., 200. Verneuil, A., 506. Vernier, 34. Vernon, M., 200. Verwer, H., 195. Verwey, A., 222, 224. Vesterberg, A., 525. Vezin, H. A., 133. Victor, C., 310. Victor, E., 351. Videgren, E. V., 351, 354. Vignal, H., 476. Vignon, L., 307, 342, 455. Vila, A. A., 364. Ville, C., 103. Villiers, A., 234, 305, 354, 387, 593, 598. Virgile, J. F., 283. Virgili, J. Fages y, 231. Vita, A., 364. Vitali, D., 351. Voelcker, A., 252. Vogel, A., 51, 60, 167, 200, 617. Vogel, H., 231, 234, 385, 396, 454. Vogel, H. A. von, 552. Vogel, J. H., 595, 607. Vogel, 0., 557. Vogt, G., 578, 667. Vogt, H. L., 207. Vogt, J. H. L., 465. Vogther, M., 197. Vogtherr, H., 475. Vohl, H., 15. Voigt, K., 366. Voit, C., 118. Volcker, A., 547. Volhard, J., 35, 46, 77, 94, 189, 215, 279, 342, 350, 378, 426, 451, 493, 547, 554, 559, 580, 623. Volker, A., 373. Voorkees, E. B. , and L. A. , 54. Vortmann, G., 278, 296, 336, 343, 351. Vosatka, J., 42, 43. Vozarik, A., 600. Vries, H. J. F. de, 107, 227, 236. Vry, J. E. de, 381. Vulpius, G., 478. Vulte, H. T., 475. INDEX OF NAMES. 7 6l WAAGE, P., 372, 478. Waegner, A., 511. Waerden, H. von der, 593. Wagenaar, M., 599. Wagener, E. M. van, 40. Wagener, R., 396. Wagmeister, 355. Wagner, A., 189, 200, 355, 454. Wagner, F., 216. Wagner, J., 31, 35, 42, 300, 617. Wagner, P., 592, 593, 598. Wagner, R. von, 478. Wait, C. E., 150. Waitz, E., 194, 290. Waldbott, S., 578, 637. Walfiscz, A., 58. Walker, J., 3, 279. Walker, J. W., 351. Walker, M. S., 196. Walker, P. H., 173, 318, 324. Walker, R. H., 145. Walland, H., 215. Wallensteiner, J., 44. Waller, E., 55, 240, 475, 536. Waller, P. H., 594. Wallis, T., 196. Walter, J., 581, 641. Walters, H. E., 382. Walther, J., 631. Walton, J. H., 203, 206, 244, 267, 317. Walz, J., 36. Warburg, E., 12. Ward, H. L., 427. Warder, R. B., 72. Waring, W. G., 177, 359, 360, 364, 369. Waring, W. J., 648. Warington, R., 97, 103, 177, 216, 339, 340, 546, 555 Warren, C. H., 420. Warren, C. M., 508, 547, 625. Warren, H. N., 107, 179, 188, 219, 232, 234, 266, 271, 296, 363,620. Warren, R. D., 118. Wartha, V., 61, 148, 184. Warmris, T. St, 622. Warwick, A. W., 329. Washburn, E. W., 290, 291. Washington, H. S., 16, 246, 251, 652. Wasowicz, V., 624. Watel, E., 474. Waterhouse, C. B.. 209. Waters, C. E., 94.' Watson, G., 96, 578. Watson, H., 18. Watson, W. G., 231. Watts, F., 96. Watts, H. F., 409, 483. Wdowiszewski, G. W., 207, 309. Wdowiszewski, H., 394, 408, 451, 563, 565. Webber, M., 136. Weber, H., 226, 333, 451, 488, 501, 504, 505, 511, 611, 625, 650. Weber, H. A., 162, 641. Weber, H. C., 240. Weber, H. C. P., 241. Weber, R., 144, 218. Wedding, H., 600. Wedekind, E., 169, 423, 498, 500 Wegelin, G., 188. Wegscheider, R., 141, 390. Weidmann, W. 0. , 373 Weigel, 0., 274. Weigelt, C., 112. Weil, F., 90, 626. Weingarten, P., 571. Weinheber, M., 441, 443. Weinland, 649. Weinschenk, B., 658. Weinschenk, E., 545. Weinstein, B. 20. Weinstein, L., 42. Weisberg, J., 527. Weiss, L., 419, 420, 422, 497, 499, 503 511. Weisskoph, P., 102 Welch, J. 0,15. Weld, F. C., 134. Weldon, W.,376. Weller, A., 84, 203, 299 Wells, H. A., 189. Wells, H. C., 239. Wells, H. L., 408, 409, 499. Wells, J. S., 268. Wells, J. S. C., 162, 577. Wells, R. C., 204, 632. Wenger, P., 449. Wense, W., 237. Werneke, A., 415, 599. Werner, A., 182. Werner, H., 200. Werther, G., 162, 207, 224, 228, 239, 462, 484. Worthier, G., 204, 484. I Wessely, L., 373. '.. West, B., 227. West, see Zuckschwerdt. West-Knight, J., 603. Westerburg, L., 648. Weston, R. S., 382. Westmoreland, J. W., 351. Westrumb, J. F., 657. Wettengel, E. B , 312. Wetzke, T.,216, 592, 593. Whatmough, W. H., 288. Wheeler, W. F., 622, 632. Wherry, E. T. , 583, 585. Whitby, G. S., 651, 654. Whitcomb, W. H., 316. White, A. H. , 528. White, E., 141, 510, 615, 618. White, J., 250. White, J. T., 547. Whitehead, C., 427. Whitfield, J. E., 415, 585, 599. Whitman, H. A., 318. Whittel, M.,652. Whittemore, C. F., 510. | Whittlesey, W., 161, 462. | Wiborg, J., 551. Wick, A. von, 89. Wick, A. von, 89. Wicke, W., 87. Widemann, M., 547. Widmar, J., 547. Widmer, J., 568. 762 A TREATISE ON CHEMICAL ANALYSIS. Wiegand, E., 189. Wiernik, T., 683. Wiggart, F., 399. Wilber, F., 161. Wilber, F. A., 207. Wilbur. C. A., 161,462. Wildenstein, A., 103. Wildenstein, R., 633. Wiley, H. W., 240, 343, 607. Wilkie, J. M., 339, 340. Will, H., 381, 580. Willenz, M., 351. Williams, C. B., 37, 39, 599. Williams, C. P., 317. Williams, F., 376. Williams, G. W. M., 635. Williams, J. E., 305. Williams, R., 351. Williamson, J. A., 141. Williamson, M. A., 630. Willot, F. J., 454. Willstatter, R., 457. Wilm, T., 295, 341, 480. Wilson, J. B., 173. Wilson, S. R., 454. Windisch, K., 106, 585. Winiwarter, E. von, 360. Winkle, F., 115. Winkler, A., 527. Winkler, C., 46, 72, 113, 118, 149, 172, 378, 396, 399, 425, 435, 488, 506, 527, 528, 529, 548. Winkler, L. W., SO, 228, 339, 355. Winteler, F., 680. Wintan, A. L., 233, 235, 596, 597. Wirth, F., 501, 510. Wislicenus, W., 47, 646. Witt, 0. N., 100, 103, 506. Witter, W., 434. Wittsteiri, G. C., 114, 172, 212, 283, 286, 295, 296, 322, 568. Wityn, A., 237. Witz, 0. N., 484. Wogrinz, A., 579. Wohler, F., 276, 280, 392, 420, 499, 578, 583, 589, 641, 649. Wohler, L., 428. Wolbling, H., 437. Wolf, C. G. L., '201. Wolf, J., 312. Wolfbauer, F., 250. Wolff, 397. Wolff, C. H., 82. Wolff, E., 594. Wolff, F. H., 105. Wolff, J., 585. Wolff, N., 35, 372, 378. Wollaston, W. H., 53, 212, 614, 651. Wollny, R., 55, 146, 147. Wolter, L., 409. Wood, 283, 284. Wood, E. F., 590. Wood, E. S., 324. Wood, T. B., 540, 541. Woods, C. D., 36. Woodman, A. G., 605, 635. Woodman, A. J., 605. Woodman, D., 323. Woodward, C. J., 54. Wormser, S., 528, 529. Worthington, W. H., 161, 460 Woudstra, H. W., 339. Woy, R<, 594, 595, 617. Wrampelmeyer, E., 641. Wright, C. R., 372, 376. Wright, C. R. A., 300. Wright, L. T., 136. Wrightson, F.,395. Wrinkle, L. F. J., 178, 181. Wroblewski, A., 286. Wuite, J. P., 228. Wiilfing, E. A., 226, 465, 661. Wunder, M., 395, 444, 449, 452, 497, 499. Wunder, M., see Tcharviani. Wunder, W. , 394. Wunderlich, E., 661. Wyrtz, H., 124. Wyatt, F.,600, 637. Wyrouboff, G., 357, 367, 506. Wyrouboff, S., 611. Wysor, R. J., 458. YARDLEY, H. B., 219. Yoder, P. A., 165. Young, A. V. E., 177. Young, R. F., 522. Young, S. W., 290, 311. Young, W. G., 80. Youtz, L. A., 282, 296, 297, 299. Yvon, M., 334. ZAHNEY, L. W., 103. Zaleski, S., 125. Zanette, J. C., 215. Zanetti, J. E., 452. Zdarck, E., 637. Zega, A., 200. Zeise, W. C., 240. Zellner, J., 644. Zengelis, C., 244, 311. Zenger, H., 655. Zenker, F. E., 412. Zerban, F., 107. Zettnor, E., 408. Zettnow, E., 278. Ziegeler, A., 611. Zimmer, W. H., 668. Zimmermann, A., 495. Zimmerman, C., 226, 397, 451, 488, 490 491. Zohls, A., 267. Zohren, T., 9. Zons, F. W., 510. Zopfehen, H., 100. Zschiesche, H., 506. Zschimmer, E., 580. Zschokke, B., 663, 664, 673. Zuckschwerdt and West, 234. Zulkowsky, C., 528, 529. Zulkowsky, K., 103, 120, 122, 179, 287, 367 479, 624. Zunino, V., 529. INDEX OF SUBJECTS. ABSOKPTION tubes or bulbs, Bender's 560. Berl's, 548, 551. Hill's, 561. . Landsiedl's, 548, 551. Miiller's, 547. errors in weighing, 550. Acetic acid, 319. determination of, 321. After-flow of glass apparatus, 38. Air, correction of weights for buoyancy, 22. washing and drying, 150. Alcohol purification, 240. recovery of, 240. Alkali, 76. Alkalies (see Potassium and Sodium). determination of, 222. isolation of, 519. loss during evaporation, 224. separation as chlorides, Berzelius, 226. Smith, 222. volatilisation of, 225. Alkaline salts, removal in analysis, 183. sulphates to chlorides, transformation of, 226. Alloy, Habermann's, for hydrogen, 392. Alumina (see Aluminium hydroxide). calcined, analysis, 447. determination of, 177, 179, 499, 508. effect of barium on, 186. effect of fluorides on, 180. effect of magnesia on, 181. effect of manganese on, 360. effect of sulphates on, 180. effect of zinc on, 359. in borates, 583. in borosilicates, 589. in fluorides, 640. hydrated analysis of, 447. precipitates, 178, 210. volatilisation during evaporation, 171. Aluminium, 308, 394, 395, 415, 455, 456, 473, 517. determination of, Hess and Campbell's process, 444. in phosphates, 607. hydroxide, 183. solubility in ammonia, 183, 470. metallic precipitation of, 305. phosphate, 607, 608. effect sulphates on precipitation, 607. reduction of iron by, 188. Aluminium, separation from phosphorus, 603. iron, Chancel's thiosulphate, 495. iron and titanium, 208. titanium, 456. zirconium, 495, 496. Alumino-silicates, 656. Alundum, analysis of, 447. filtration cups, 621, 631. Ammeter, 253. Ammonia action on glass vessels, 480. for alumina determinations, 178. solubility of aluminium hydroxide in, 183, 470. chromium hydroxide in, 470, 479. silica in, 183. Ammonium acetate solution, 576, 600.' and sulphate solution, 485. bisulphite, 191, 350. reduction of iron, 191. solution, 445. chloride, 224. solution, 182. dichromate solution, 516. molybdate solution, 332, 596. standard, 333, 416. monosulphide, 341. nitrate, 183. solution, 355, 596. phosphomolybdate, 590. properties, 590. salts, removal, 218, 224. removal in analysis, 183. sodium phosphate solution, 374. thiosulphate solution, 293. uranate properties, 489. Ampere, 253. Analyses, analoid process, 250. clays, rapid, 460. duplicate errors in, 550. errors (see Errors). mechanicalised, 250. object of, 243. of clays, rational and ultimate, 670. of silicates, abbreviated schemes, 241. complete scheme, 242. permitted errors, 245. rapid, 244. reporting, 251. standard processes, 250. tables, 54. Anatase, 656. Andalusite, 656. 763 764 A TREATISE ON CHEMICAL ANALYSIS. Anhydrite, 530. Aniline dye, detection of, 322. Anode, 253. Antimony, 318, 406. arsenic, and tin separation, 280. chloride volatilisation on boiling, 271- compounds evaluation, 306. detection in enamels, 306. determination, 177, 296. as sulphide, Clarke and Henz's process, 296. effect molybdenum, 299. tungsten, 299. . precipitation by tin, 305. volumetric, 304. Gyory's bromate process, 301. - Weller's iodine, 299. oxychloride, 405. precipitation by, 304. Tookay's process, 305. pentasulphide, 297. separation from gold, 428. platinum, 428. tungsten, 410. Talbot's process, 410. tin, 275, 297. electrolytic process, 299. Panajotow's process, 306. sulphide, 272, 278. tri-sulphide, properties, 295. Antimonious salts, oxidation, 299. Apatite, 656, 658. Argillaceous matter, 656. composition of, 665. in clays, determination, 659. Arsenic, 211, 415, 473, 481, 618. chloride, volatilisation on boiling, 271. determination, 177. as arsenic sulphide, 282. as magnesium pyroarsenate, 283. as silver arsenate, 282, 292. as tin arsenate, 314. gravimetric, 284. elimination from solutions, 300. Mohr's volumetric, 290. oxide, 594. Pierce's volumetric process, 292. salts action hydrogen sulphide, 276. separation from gold, 428. platinum, 428. tungsten, 411. Kehrmann's process, 411. molybdenum, 413. tin, antimony, 280. uranium, 490. sulphide, 276, 278, 280. trisulphide, 282. volumetric determination, 304. Arsenious oxide, evaluation of, 293. Asbestos for Gooch's crucibles, 104. Augite, 656. BALANCE, 1. assay, 435. effect of unequal arms, 21. sensibility of, 12, 13. theory, 5. Balance, use, 6. Barite analysis, 531. Barium, 211, 245, 415. borate, 583. carbonate, 471. separations, 470. chloride, 227. solution, 619. chromate properties, 477. correction in alumina determination, 186. determination, 535, 516, 517. in fluorspar, 649. effect on calcium precipitate, 513. magnesium precipitate, 513. phosphate, 614. separation calcium, 514, 515. chromate process, 515. Stromeyer and Rose, 514. strontium, 514, 515. chromate process, 515. indirect process, 516. Stromeyer and Rose, 514. spectrum, 513. sulphate, 227, 322. absorption salts by., 611. determination, 531. filtration, 620. loss on ignition, 616. properties, 610. purification of precipitated, 613. reduction, 615. Barthel's blast lamps, 617. Barytes analysis, 531. Basic acetate separation, 183, 362. effect of chromium on, 469. theory, 361. benzoate separation, 361. formate separation, 361. succinate separation, 361. Baths, liquids for vapour, 574. Meyer's vapour or oil, 575. Treadwell's, 462. oil for, 462. Baudisch's cupferron reagent, 456. Bauxite, analysis, 446. Bead forceps, 328. Benzidine hydrochloride, 628. Beryllium, 445. detection, 448. determination, 177, 179, 449. Parsons' and Barnes' process, 449. separation from yttrium, 508. zirconium, 508. Berzelius' solution, 638. Biotite, 633. Bismuth, 325, 353. determination, 177. colorimetric, 348. formate process, 348. Jannasch's basic hydroxide process, 348. Lowe's basic nitrate process, 347. nitrate, standard solution, 349. oxide, effect in permanganate titration, 190. separation from cadmium, 347, 348. copper, 347, 348. INDEX OF SUBJECTS. 765 Bismuth, separation from lead, 347. sulphide, 277, 278. Boiling tubes, 586. Bone ash, 602. evaluating, 602. china bodies, analysis, 606. Boracite, evaluation, 581. Borax, evaluation, 580. Jacobi's process, 581. loss boric oxide during calcination, 578. Borers, Fraenkel's, 138. Meyer's, 138. Boric acid, evaluation of. 578. loss boric oxide during calcination, 578. turmeric test, 576. oxide detection, 576. alcohol flame, 576. boronfluoride test, 577. glycerol flame, 576. tumeric test, 5.76. determination, 578. distillation process, 585. Stromeyer's test, 578. - Vogt's process, 578. Wherry's process, 583. volumetric, 579, 580. Jacobi's process, 581, 582. Jones' process, 581. - Schaak's process, 581. Smith's process, 580. Will and Zschimmer's process, 580. in potassium hydroxide, 579. in sodium hydroxide, 579. loss during evaporation of solutions, 578. retention by alumina, 583. volatilisation of, 578. Boring, 138. Borocalcite evaluation, 581. Boronatrocalcite evaluation, 581. Bottles reagents, 142. Jevitt's, 142. Smart's, 143. Bromine, 372, 478. water, 372. Bronzite, 656. Brown-stone, Japanese, 384. Burette corks, 43. floats, 35. Scheibler's, 197. stands, 42. Burettes, 196. automatic filling, 55. calibration, 39. Burner, Argand's, 111. Bun sen's, 110. Meker's, 111. blast, 112. micro, 111. Button brush, 328. CADMIUM, 394, 415, 445. acetate solution, 626. carbonate separation, 671. determination, 177. Cadmium, determination as sulphate, 357. Berg's volumetric, 358. electrolytic, 358. ferrocyanides, 368. metallic precipitation by, 305. separation from copper, 275. zinc, 273. sulphate properties, 357. sulphide, 278, 280. Caesium chloride, 533, 536. determination of, 239. spectrum, 533. Calcine, evaluation of, 314. Calcite, 656. Calcium, 415. and magnesium carbonates determina- tion, 522. Newberry's process, 522^ borate, 583. borate, evaluation, 581. carbonate, 47, 70, 223. chloride for drying gases, 548. solution, 638. determination, 519, 522, 524. effect barium on, 513. strontium on, 513. gravimetric. 213. in phosphates, 607. influence of magnesium on, 212. fluoride, 634, 637. analysis, 648. oxalate properties, 211. phosphate, standard solution, 600. separation from barium, 514. chromate process, 515. Stromeyer and Rose, 514. magnesium, indirect, 244. strontium, 514, 515. chromate process, 515. Stromeyer and Rose, 514. spectrum, 513. sulphate, 325. analysis, 529. volumetric, Kraut, 215. zincate, 359. Calculations, 52. Calibration of burettes, 39. of flasks, 35. of pipettes, 39. Carbon determination, 385, 545, 559, 563. Berthier's process, 568. combustion in bomb, 562, 623. Parr's process, 562. dry combustion process, 559. - - Shimer's 559, 563, 568. Lowe's process, 545. Mackintosh's process, 545. Morgan's wet combustion process, 555. Schoffel's process, 545. wet combustion process, 546. dioxide, detection, 552. determination, 321. Lunge and Marchlewski's pro- cess, 556. rapid volumetric, 552. ;66 A TREATISE ON CHEMICAL ANALYSIS. Carbon dioxide determination, Scheibler and Dietrich's process, 555. volumetric, 554. removal from air, 547. washing, 192. filtration tubes, 545. Carbonates, 552. analysis, 552, 553, 554, 555, 556. titration, direct, 69. back, 70. mixed hydroxides and, 72. Carboraridum analysis, 562. grinding, 562. Cathode, 253. Cement paper to glass, 143. Cements, analysis, 521. Ceria separation, 50. Cerium, 308, 457, 496, 500, 501. determination, 179, 506. volumetric permanganate, 511. iodates, 496. oxalate, 501. effect of uranium on solubility, 504. separation from didymium, 506. lanthanum, 506. thorium, 505. yttrium, 504. zirconium, 504. China clay, 656. composition of, 665. rock, 656. Chlorides, conversion to nitrates, 427. . determination, 652. turbidity process, 653. soluble, 653. volumetric determination of, Mohr's pro- cess, 79. Yolhard's process, 76. Chlorine, 637. determination, 519, 618, 652. as silver bromide, 653. as silver chloride, 652. as silver iodide, 653. removal from carbon dioxide, 547. Chlorite, 656, 657, 658. Chromates, absorption by filter paper, 477. Chrome iron ore (see Chromite). Chromic oxides, analysis, 474. Chromite, 504. analysis, 474. Chromium, 211, 308, 394, 395, 415, 456, 461, 473, 618. delicacy of ammonia test, 177. detection, 468, 484. determination, 171, 179. as barium chromate, 477. as oxide, 479. colorimetric, 473. colorimetric diphenylcarbazide, 474. volumetric, 476, 481. hydroxide, solubility in ammonia, 470, hydroxides, 480. ores, valuation, 494. separation, 472. cobalt, 469. Chromium separation, Knorre's process, 473. manganese, 384, 469. nickel, 469. vanadium, 483. zinc, 469. Clay base, 656. beds, evaluating, 140. sampling, 138. china, composition of, 665. determination of, 525. in limestones, 526. ideal, 656, 665. making soluble salts innocuous, 633. matter, 656. primary, 656. proper, 656. residual, 656. secondary, 656. substance, 656, 665. transported, 656. true, 665. Clayite, 656. Clays, 656. action hydrochloric acid, 657. hydrofluoric acid, 657. sulphuric acid, 661. deflocculation, 659. determination soluble salts in, 630. estimation argillaceous matter, 659. felspar, 660. quartz, 660. mica in, 667. opening for analysis, 164. organic matter, 659 rational analysis, 658. soluble chlorides in, 653. Cleaning glass apparatus, 36. Coal, determination carbon in, 562. sulphur in, 621. Cobalt, 211, 308, 413, 445, 455, 456, 457. detection, 386. Danziger's test, 386. Ilinski and Knorre's test, 386. Skey's test, 386. Vogel's test, 386. determination, colorimetric process, 397. volumetric, 400. effect on ammonia precipitate, 387. barium carbonate process, 387. basic acetate process, 387. isolation of, 387. ores, analysis, 399. oxide, evaluation of, 399. separation, 469. nickel, 390. Branck's glyoxime process, 394. electrolytic process, 395. Fischer's nitrite process, 390. Ilinski and Knorre's nitroso- naphthol process, 394. Liebig's cyanide process, 393. - sulphate, standard solution, 398. sulphide, 273. properties, 388. Cochineal, 63. Cocks for burettes, 43. INDEX OF SUBJECTS. 767 Colloidal precipitates, 95. silica, 525. Colonis, analysis, 265. Colorimeter, 82. dipping, Duboscq's, 82. Weller's, 84. Colorimetry, 82. errors, 85. Columbium, see Niobium, 6. Condenser, GockePs, 288. Congo red, 580. Conversion factors, 332. Cooling box, 72. Copper, 211, 308, 415, 457. carbonate, 354. determination, 177, 321. ammonia colorimetric, 355. ferrocyanide colorimetric, 355. Cupferron process, 350, 456. DeHaen's volumetric iodine process, 351. electrolytic, 258. metallic precipitation by alu- minium, 354. cadmium, 354. nitrosonaphthol process, 350, 394. Rivot's thiocyanate process, 350. volumetric thiocyanate process, 350. oxalate, 427. oxide, evaluation, 354. precipitation by, 304. reduction iron by, 188. separation, cadmium, 275, 350. gold oxalates, 427. palladium, 439. uranium, 490. - sulphate and pumice, 554. standard solution, 356. sulphide, 277, 278, 279. zinc couple, reduction iron by, 188. Corundum, 656. Crucible, Brunck's, 107. Gibbs', 391. Gooch's, 104. asbestos for, 104. use of, 104. combustion, 560. - Heraeus', 158. Munroe's 107. Rose's, 391. Shimer's combustion, 560. filtration, 630. Richards' 104. Vollers', 104. copper, 475. gilded platinum, 475. gold, 475. nickel, 475. platinum, 475. silver, 475. Cryolite, 641. analysis of, 448. Cupboards, fume, 118. Treadwell's 119. Cupellation, 431. Cupels, 327. Cupferron, 456. Cuprous thiocyanate properties, 351 Cyanite, 656. DANBURITE, 584. Decipia, 501. Deflocculation clays, 659. Desiccating agents, 117, 118. Desiccators, 1 1 6. Hempel's, 117. Scheibler's, 116. Diaspore, 656. Dicyanodiamidine, 395. Didymium, 506. detection, 507. determination, 179, 506. separation, cerium, 506. thorium, 505. sulphate, 507. Dimethylglyoxime solution, 395. Diphenylcarbazide, 474. Dish for electrolysis, Classen's, 257. Dolomite, 552, 656. analysis, 521. Drainage error, glass apparatus, 38. Dreverhoff s filter papers, 88. Drying tubes (see Absorption tubes). Taiibers', 563. Dumortierite, 584, 656. Duplicate analyses, 171, 248. Dysprosia, 500. ELECTRIFICATION during weighing, 551, Eleetro-analysis, 253. essential factors in, 254. general apparatus, 256. iterature, 256. Electrolyte, 253. Enamels, analysis, 265. Eustatite, 656. Epidote, 656, 657. Erbia, 500. Errors, accidental, 247. analytical, 171, 245, 249. constant, 247. in duplicate analysis, 240. in silica determination, 171. in volumetric analysis, 67. in weighing absorption tubes, 550: parallax, 32. personal, 249. Eschka's mixture, 622. Estrichgips, 530. Ethyl borate, 585. Europa, 500. Evaporating funnel, Meyer's, 167. Excess of reagent in analysis, 182. FACTORS, conversion, 332. Feathering in cupellation, 327. Felspar, action sulphuric acid, 662. in clays, estimation, 660. Felspathic detritus, 660, 670. estimation of, 660, 670. Ferric alum, 77. chloride, loss by evaporation, 167. standard solution, 308. hydroxide, 183. 768 A TREATISE ON CHEMICAL ANALYSIS. Ferric hydroxide absorption, nickel salts, 387. iron, separation, ferrous,- 471. oxide, 325. precipitate, retention of zinc by, 360. potassium sulphate solution, 201, 206. Ferrous iron, separation, ferric, 471. oxide, determination, 461. disturbing factors, 464. Eder's process, 462. Ferruginous precipitates, ignition of, 184. Filter papers, 87, 90. absorption barium salts by, 72. chromates by, 477. lead salts by, 340. ash, composition of, 89. combustion in Jiiptner's muffle, 184. Swedish, 88. M. DreverhofFs. 88. J. H. Munktell's, 88. C. Schleicher and Schiills', 88. tared, 100. toughened, 100. - plates, Witt's, 103. stands, 91. tube, Fresenius', 102. Jannasch's, 653. Koninck's, 102. Mason's, 102. Wagner's, 107. Zopfchen's, 100. for carbon, 101. Filtration, 87, 90. accelerated, Bunsen's, 98. by pressure, 98. colloidal sulphides, 275. crucible, Brunck's, 107. Gooch's, 104. Munroe's 107. Neubauer's, 107. Richards', 104. Vollers', 104. solvents for Munroe's, 108. special felts for, 106. cups, alundum, 731. earthenware, 631 Riimpler's, 631. - flasks, Walther's, 631. funnels, 90. gelatinous precipitates, 363. Koninck's apparatus, 101. of fine suspended matter, 620, 630. through discs, 103. Carmichael's, 103. Casmejor's, 103. Kaehler's, 103. Smith's, 103. tubes, carbon, 545. Fischer's salt, 390. Flashing in cupellation, 328 Flasks, Bolton's, 70. calibration of, 35. Corleis', 546. Erlenmeyer's, 70. Herzka's, 70. Lewkowitsch's, 586. standard, 47, 48. Biltz's, 47. Flasks, standard, Giles', 47. - Goske's, 47. Hirschsohn's, 47. Pfliiger's, 47. Walther's filtration, 631. Floats, burette, 35. Flooring plaster, 529. Fluoride determination, 519, 637. alumina in, 640. as calcium fluoride, 637. as lead chlorofluoride, 638. as potassium fluoride, 641. as silicon fluoride, 646. Carnot's process, 641. silica in, 640. Steiger's colorimetric, 644. loss by ignition fluorides, 639. removal from carbon dioxide, 547. separation as calcium fluoride, 634. Fluorides, detection, 634. etching test, 634. hanging drop test, 636. Fluorometer, Oettel's, 646. Fluorspar, 641, 656. analysis, 648. Flux, Clark's, 474. Kayser's, 475. Rose's, 270. tribasic, 475. - Turner's, 577. Fluxes for opening chromates, 474. Formaldehyde, 322. Freezing in cupellation, 327. | Fresenius' tower, 547. Fuel, volatile sulphur in, 624. Fumaric acid solution, 510. Funnels, Biichner's, 103. Hirsch's, 103. Funnels for filtration, 90. Fusion for cupellation, 327. mixture, 163. Parr's, 564. GADOLINIA, 501. Galena, 268. analysis of, 328. decomposition for analysis, 329, 330. Garnet, 656. Gas generators, 147. - Friswell's, 147. - Kipp's, 147. Koninck's, 148. Perkin's, 148. Wartha's, 147. Gases in rocks, 573. Glaucophane, 656. Gault clays, analysis, 518. Generators, gas, 147. Friswell's, 147. Kipp's, 147. Koninck's, 148. Perkin's, 148. Wartha's, 147. Gilbertite, 656. Glass, action reagents on, 144, 616. analysis, 265. vessels, action ammonia on, 480. INDEX OF SUBJECTS. Glass vessels, action reagents on* 371. attack by reagents, 364. Glazes, analysis, 265. Glucinum, see Beryllium. Glycerol, 322, 578. Gold, 457. assaying, 431. colours, analysis, 429. cupellation, silver and, 431. detection, 429. determination, 177, 439. colorimetric, 429. cupellation process, 326. "liquid," 431. lustres, analysis, 431. precipitation, 425. hydrazine hyrdochloride, 427. oxalic acid, 427. sulphurous acid, 426. ~ precipitating agents, 441. separation, antimony, 428. arsenic, 428. copper, 427. Mylius' ether process, 431. silver, 431. silver, platinum, and, by cupellation process, 435. - tin, 428. sulphide, 277, 278. properties, 425. Gooch and Eddy's solution, 227. Grab samples, 127. Granite, 656. Graphite, analysis, 568. crucibles, analysis, 568. Grinding, 120. carborundum, 562. dangers of fine, 123. effect on ferrous silicates, 465. Guard tubes, 58. Binder's valve, 188. Bunsen's valve, 188. Contat's, 189. - Gockel's, 189. Kempfs, 189. Schiebler's, 189. Gypsum, 656. analysis, 529. HjEMATOXYLINE, 580. Heating, 110. gas, 110. electrical, 112. Heavy spar, analysis, 531. Helianthine, 580. Hematite, 656, 658. Holmia, 500. Hornblende, 656. action sulphuric acid, 663. Hot funnel, 324. plates, 112. Hydrochloric acid, action on clay, 657. standard, 47. temperature correction, 49. Hydrochloroplatinic acid, 240. Hydrofluoric acid, 169, 226, 462. action on clays, 657. 769 effect on permanganate Hydrofluoric acid, titration, 465. Hydrogen peroxide, 177, 205, 332, 449. solution, 205. preparation of, 391, 392. sulphide, 147. action on arsenic salts, 276. arsenic free, 149. precipitation by, 276. theory of, 272. washing and drying, 150. Hydrolysis, 181. Hydroxides titration, 65. mixed carbonates and, 72. Hydroxylamine hydrochloride, 322. Hypers thene, 657. ILMENITE, 657. Indicator correction, 80, 292, 334, 369. Indicators, 46, 60. external, 334. indigo solution, 302. "spot test, "334. titration with spot test, 454. Indigo solution as indicator, 302. Indirect separations, 229. magnesium and calcium, 244. potassium and sodium, 227. Inquartation, 432. lodates, test for in iodides, 287. Iodine, 288. determination, 655. as cuprous iodide, 655. as silver iodide, 655. indicator for free, 287. solutions, standard, 288, 291, 312. preservation, 289. titration with solutions of, 300. Iridium, 428. determination, 436, 438. Iron, 308, 353, 394, 395, 415, 444, 517, 594. delicacy of ammonia test, 177. determination, 177, 179, 187, 321, 499, 508. Baudisch's cupferron process, 455. colorimetric process, 200. effect of titanium, 189. Ilinsky and Knorre's process, 455. in phosphates, 607. volumetric, 483. Marguerite's, 198. Penny's dichromatc, 453. permanganate, 451. Reinhardt's 451. effect of manganese sul- phate, 451. 452. acid, effect on precipitation molybdenum sul- phide, 412. metallic precipitation by, 305. ores, analysis, 460. oxides, analysis, 460. phosphate, 608. precipitating agents, 455. separation aluminium, Chancel's theo- sulphate, 495. 49 770 A TREATISE ON CHEMICAL ANALYSIS. Iron, separation, cobalt, 469. ether process, 456. manganese, 469. - nickel, 388, 469. of phosphorus, 603. titanium, 495. - - --- and aluminium, 208. -- uranium, 492. zinc, 469. standardising permanganate with, 195. sulphide, 273. JOULIE'S solution, 603. KAOLIN, 656. Kaolinite, 656. Koninck's reagent, 541. LABELS, 143. Lacmoid, 580. Lamp, Barthel's blast, 617. Lanthania, 500, 501. Lanthanum, 179. - detection, 507. determination, 506. oxalate, effect uranium on solubility, 504. - separation from cerium, 506. thorium, 505. - sulphate, 507. Lawning, 125. Lawiis, 125, 126. scale of sieves, 126. - silk, 125. standard, 126. Lead, 211, 353, 394, 640. - acetate solution, 415. , . carbonate (for silicate fusions), 565. chlorofluoride, 638. - - chromate, 484. analysis of, 324. determination, 177. -- as chromate, 325. as molybdate, 331. - -- as oxalate, 326. -- as sulphate, 317. -- colorimetric process, 339. electrolytic process, 334. precipitation by zinc, 305. - volumetric molybdate, 333. dioxide, 335. - -- frits, solubility test, 330. metallic, determination of, 322. molybdate, 414. -- properties, 331. monoxide, determination of, 322 oxalate, 427. - oxides, 325. - peroxide, determination of, 323. - precipitation by, 304. separation from bismuth, 318. cadmium, 318. copper, 318. - sesquioxide, 323. - sulphate, 324. -- properties, 315. -- solubility in stannous chloride, 314. Lead, sulphide, 273, 276, 277. -- solubility in calcium chloride, 32< -- volatilisation of. 320. - uranate, 484. - vanadate, 484. Lepidolite, 663 Leucite, 657, 658. Leucoxene, 657. Lignite, 657. Lime (see Calcium). determination, Passon's process, 522. free in quicklime, mortars, etc.,' 527. -- determination, 527, 528. - Friihling's process, 527. Stone and Schenck's process - Winkler's process, 527. Limestones, 170. - analysis, 518, 521. determination oxidisable matter in, 626. sulphur in 626. - mineralogical analysis, 524. phosphatic, 519. Limonite, 657. Liquid gold, analysis, 431. Lithium chloride, solubility in amyl alcohol 536. pyridine, 534. determination, 533. Gooch's process, 536. Kahlenberg and KrauskofPs process 534. - in silicates, determination, 239. perchlorate, 534. spectrum, 533. Litmus, 60, 579, 580. Litre, 30. expanded, 31. Mohr's, 31. normal, 31. Loss on ignition, 157. determination of, 157. Ludurgite, 584. Lunge's solution, 525, 659. Lutecia, 500. MAGNESIA, determination, 180. mixture, 597. /^lagnesite, 552, 65,7. Magnesites, analysis, 519^521. Magnesium, 415. ammonium arsenate, 283. in phosphate, properties of, 215. and calcium carbonates, determination 522. : Newberry's process, 522. indirect determination, 244. borate, 583. citrate solution, 603. determination, 182, 519, 524. effect of barium, 221, 513. calcium, 221. manganese, 221. silica, 221. strontium on, 513. in phosphates, 608. pyrophosphate, 597. INDEX OF SUBJECTS. 771 Magnesium, pyrophosphate, gravimetric, 218. reduction of iron by, 187. zinc alloys for reduction iron, 188. Magneting, 121. Magnetite, 657, 658. Manganese, 211, 367, 394, 395, 415, 455, 456, 474. delicacy of ammonia test, 177. determination, 177, 508, 512. bromine process, 372. Pattinson's volumetric, 376. phosphate process, 374. sulphide process, 373. Volhard's volumetric, 377. - Volhard-Fischer's volumetric, 380. Weller's colorimetric, 382. dioxide evaluation, 381. Mohr's process, 381. earths, analysis, 384. effect on analysis of silicates, 371. ores, 384. valuation, 384. phosphate, 608. separation, 469. from chromium, 384. from cobalt, 389. from nickel, 389. sulphate, standard solution, 382. sulphide, 273. properties, 373. Manganous salts, effect on permanganate titra- tion, 466. Mannitol, 578. Marls, analysis, 521. mineralogical, analysis, 524. Marking porcelain, etc., 165. Meniscus, 32. errors, 32. Mercuric chloride solution, 453. sulphate solution, 189. sulphide properties, 342. Mercurous antimonic tungstate, 410. nitrate solution, 408. test papers, 192. Mercury, 394, 415, 445. chloride, volatilisation on boiling, 271. determination, Erdmannand Marchand's distillation process, 344. Esohka-Holloway's distillation pro- cess, 345. Rath's process, 341. - Volhard's sulphide process, 342. separation from bismuth, 341. cadmium, 341. copper, 341. lead, 341. sulphide, 279. Meta phosphates, 590. Metallic precipitations, 304. Methyl alcohol, 322, 585. berate, 585. ; orange, 62, 580. Mica, action sulphuric acid, 663. in clays, 667. Mimosa blossoms, tincture, 580. Mineralogical analysis, clays, 656. limestones, 524. Mineralogical analysis, marls, 524. Moisture, hygroscopic, 155. determination, 156. Molybdenum, 211, 445, 456 469, 470, 473 481, 497, 618. action on determination of antimony, 299. detection, 406. determination, 177. as lead molybdate, 414. as oxide, 413. as sulphide, 412, 414. effect of iron on, 412. volumetric, 483. permanganate, 415. oxide, volatilisation of, 413. residues, recovery, 597. salts, action hydrogen sulphide, 412. separation of arsenic, 413. tin, 413. tungsten, 416. Hommel's process, 416. Pechard's process, 417. Rose's process, 417. sulphide, 277, 278 properties, 411. Molybdic oxide, 407. salts, reduction, 415. Monazite, 510. sand, 505. Mortars, Abich's, 121. agate, 122. power driven, Carling's, 122. M'Kenna's, 122. analysis of, 527. diamond, 121. steel, 121. MunktelPs filter papers, 88. Muscovite, 663. NAPLES yellow, evaluation of, 326. Neodidymia, 500, 501. Nepheline, 657, 658. action sulphuric acid, 663. Nessler's tubes, 85. Nickel, 211, 308, 415, 445, 455, 456, 457. detection, 386. Grossmann and Schiick's test, 395. Parr's test, 386. Tschugajeffs test, 386. determination, 179. colorimetric process, 399. dicyanodiamidine process, 395. volumetric cyanide process, 400. effect on ammonia precipitate, 387. barium carbonate precipitate, 387. basic acetate precipitate, 387. isolation of, 387. ores, analysis, 399. oxalate, 427. oxide, evaluation of, 399. salts, absorption by ferric oxide, 887. separation, 469. from cobalt, 390. Brunck's glyoxime process, 394. electrolytic process, 395. * Fischer's nitrate process, 390. 772 A TREATISE ON CHEMICAL ANALYSIS. Nickel, separation from cobalt, Liebig's cyanide process, 393. Ilinski and Knorre's nitroso- naphthol process, 394. iron, 388. palladium, 395. sulphide, 273. Niobium, 405, 496, 497, 511. detection, 407. determination, 418, 503, 504. Simpson's process, 418. separation, tantalum and, 421. < Marignac's process, 421. specific gravity process, 423. Weiss and Landecker's pro- cess, 422. Nitre analysis, 539. ^-Nitrophenol, 65, 681. Nitroso-B-napthol solution, 394, 455. Nontronite, 657. Normal solutions, 45. OHM, 253. Oil for baths, 462. Olivene, 657, 658. Opening chromites, 474. glazes, etc., for analysis, 266. minerals with Mo, W, Nb, Ta, 405. rare earth minerals, 502. silicate, 441, 497. for barium, 517. for ferrous iron, 461. tantaliferous minerals, 418. zircons, 499. Organic acids on precipitation manganese, etc., retarding action, 363, 373. matter, determination, 520, 621. effect on determination ferrous salts, 465. in clays, 659. in limestones, 520. Orthophosphates, 590. Osmium, 428. determination, 437, 438. Oven, Paul's drying, 297. Oxidation of ferrous salts, 362. Oxidisable matter in limestones, determina- tion, 626. PALLADIUM, 426, 428. determination, 439. separation, copper and, 439. nickel, 395. Paper, cement glass to, 143. pulp, 179. size, 143. varnish, 143. Parallax errors, 32. Parting, 432. Pearlash, analysis, 539. Perhydrol, 205, 322. Permanganate titration, effect of hydrofluoric acid on, 465. manganese salts, 466. organic matter, 465. titanium, 466. Permolybdate reagent, 595. Peroxide fusion mixture, 244, Phenolacetolin, 580. Phenolphthalein, 62, 580. Phenylhydrazine, 322. bisulphite solution, 445. Phlogopite, 663. Phosphates, analysis, 606. acetate process, 606. calcium sulphate process, 606. caustic alkali process, 606. oxalate process, 606. determination, 520, 590. Phosphomolybdates and silicomolybdates, separation, 595. Phosphorus, 211, 415, 473. . determination, 520, 590. action ammonium nitrate, 591. ammonium salts, 593. arsenic oxide, 594. nitric acid, 591 organic matter, 595. silica, 595. titanium salts, 594. vanadium salts, 594. as ammonium magnesium phos- phate, 597. Neubauer's correc- tion, 597. as ammonium phosphomolybdate, 598. colorimetric, 603. Knight's, 603. effect of temperature, 592. Emmerton's process, 599. in phosphates, 608. Joulie's volumetric process, 603. magnesium citrate process, 603. Osmond's process, 599. Pemberton's process, 599. rapid processes, 598. silica and, 605. uranium volumetric process, 600. Woy's process, 595. loss on ignition, 607. separation, iron and alumina, 603. tungsten, 411. Kehrmann's process, 411. silica, 595. Pipette, Beale's filter, 73. Carnot's ether, 458. for weighing corrosive liquids, Berl's, Lunge's, 9. Holde's, Rother's ether, 458. Pipettes, automatic filling, 55, 58. calibration, 39. Plagioclase, 658. Plaster of Paris, analysis. 529. anhydrite, 530. Frey's process, 529. Platinum, 426, 427, 428, 457. alloys, analysis, 437. apparatus, 114. loss of, 114. crucibles, 165, 184. action phosphates on, 219. attack by selenides, 441. INDEX OF SUBJECTS. 773 Platinum crucibles, correction for dissolved 186. corrosion of, 114. occlusion flame gases by, 372. permeability to flame gases, 184. Venvey's, 224. determination, 177, 435, 436, 438. felt for filtration, 107. precipitation, 425. recovery, 240. separation, antimony, 428. arsenic, 428. silver, 431. silver and gold by cupellation pro- cess, 435. tin, 428. sulphide, 277, 278. properties, 425. Pole paper, 253. Policeman, 91. Porcelain, action reagents on, 144. Porrier's blue, 580. Potassium (see Alkalies). acid fluoride fusions, 185. aluminium sulphate solution, 201. bisulphate fusions, 184. borofluoride, 583. bromate, standard solution, 302. carbonate 69. analysis, 539. chloride solubility in amyl alcohol, 536. pyridine, 534. chloroplatinate, properties of, 231. standard solution, 539. chromate, 79. cobaltinitrite, 540. cyanide, standard solution, 401. determination, 535, 537, 538. cobaltinitrite process, 540. colorimetric, 539. volumetric, 540. dichromate, standard solution, 453. ferricyanide indicator, 454. ferrocyanide solution, 355. standard solution, 368. fluoride solution, 643. hydroxide solution, 548, 623. iodide, 287. solution, 349, 539. test for iodates in, 287. molybdate solution, 603. nitrate, 461. permanganate, burettes for, 197. preservation of, 196. standardisation with iron, 195. sodium oxalate, 193. phosphate, standard solution, 604. pyrosulphate fusions, 184. separation sodium, indirect, 227. potassium chloroplatinate pro- cess, 234. potassium perchlorate process, 237. spectrum, 533. sulphate solution, 509. thiocyanate, 77. solution, 201. Potassium thiocyanate, test ferric salts, 189. Praseodidymia, 500, 501. Precipitates, absorption salts by, 96 colloidal, 95. barium sulphate, 620. gelatinous, 96. filtration of, 179, 183, 479. Gooch's cone, 183. washing, 87. gelatinous, 308. theory of, 95. Prehuite, 657. Pressure bottles and flasks, 277, 412. Pulverisation, 120. Pulveriser, Braun's disc, 120. Purple of cassius, analysis, 429. Pyridine, 535. Pyrites, 657. analysis, 610. determination sulphur, 621. Pyrolusite, analysis, 384. Pyrophosphates, 590. QUARTATION, 432. Quartz, 641, 647. action sulphuric acid, 663. debris, 660, 670. estimation of, 660, 670. determination, 525. in clays, estimation, 660. RADIATOR, 112. Ramsay's cap for Bunsen's burner, 268. Rare earths, 599. determination, 502. minerals, analysis, 503. separation, uranium, 488. Rational analysis (see Mineral ogical analysis). clays, 658. Zschokke's method, 673. Leopold's process, 668. v. ultimate analysis clays, 670. consistency of, 673. Reagent bottles, 142. Reagents, 141. action on glass, 144. porcelain, 144. equivalent system, 146. testing, 141. Red earths, analysis, 460. lead, analysis of, 322. for silicate fusions, 565. rose leaves, tincture, 580. Reduction ferric salts by stannous chloride, 453. of iron by magnesium, 187. by zinc (see Zinc). for permanganate titration, 187. Reductor, 190. Jones', 190. Reinhardt's solution, 453. Rhodium, 428. determination, 439. Rhyolite, 656. Riders, 3, 10. Jiime sampling, 131. Roasting for cupellation, 327. 774 A TREATISE ON CHEMICAL ANALYSIS. Rotating electrode, electro-analysis, 337. Rubidium chloride, 533, 536. - determination of, 239. - spectrum, 533. Riimpler's filtration cups, 631. Ruthenium, 428. determination, 437, 438. Rutile, 504, 657. SALTPETRE, analysis, 539. - Chili, evaluation, 538. Samaria, 501. Sampler for liquids, 128. - Metzger's, 128. - split -stream, 133. -- Clarkson's, 133. Braun's, 133. -- Taylor and Brunton's, 131. time, 134. - Brunton's, 134. -- Snyder's, 134. -- Vezin's, 134. Samples, dispatch of, 136. - receiving, 136. Sampling, 127. by channelling, 131. - by coning, 129. - by quartering, 129. - by riffle, 131. - by split shovel, 131. - clay beds, 138. - reducing bulk during, 134. - grain size during, 134. Scandia, 501. Scapolite, 657, 658. Schaffgotsch's solution. 226. Scheelite, 504. Schleicher and Schiill's filter papers, 88. Screening, 125. Sealed tube operations, 493. Seemann's solution, 638. Selenite, 657. Selenium, 426, 4?,7. -- compounds, volatilisation, 441. - detection, 439. determination, 177, 441. -- as sulphide, 440. effect of copper, 440. sulphurous acid process, 442. volumetric, 441. - precipitating agents, 441 separation, tellurium, 441, 442. l sulphide, 277, 278, 440. Separation of sulphides soluble in ammonium sulphide, 277. --- sodium sulphide, 277. Serpentine, 657, 658. Siderite, 552, 657. Sieving, 125. Silica, 594. colloidal, 668. -- detection of, 669. determination of, 526, 668. combined, 669. dehydration of, 175, 176. - determination, 167. -- effect fluorides, 172. Silica, determination, effect alumina, 173. borates, 172. lime, 174. magnesia, 174. zinc on, 359. errors of, 171. in boro-silicates, 589. in fluorides, 640. in phosphates, 607. phosphorus and, 605. theory of, 172. free, 669. gelatinous, 165. - ignition, 168, 171, 173. residues, 170, 272. separation from tungsten, 409. ammonia process, 409. hydrofluoric process, 409. potassium bisulphite process, 409. phosphorus, 595. solubility in potassium bisulphate, 186. sodium bisulphate, 505. volatilisation during evaporation, 167. Silicates, opening up, 160. Silicon carbide, estimation, 565. Silicomolybdates and phosphomolybdates, separation, 595. Sillimanite, 657. Siloxicon, 566. analysis, 566. Silver arsenate, 293. borate, 583. chloride properties, 650. determination, Benedict and Gans, 653. : cupellation process, 326, 654. metallic precipitation, 654. turbidity method, 76, 326, 654. Whitby's colorimetric process, 654. nitrate, ammoniacal, 73. solution, 293, 326, 652, 654. -- standard, 77, 79. phosphate, 638. precipitation by, 304. separation gold, cupellation process, 432. platinum, cupellation process, 435. sulphate solution, 547. volumetric determination, Volhard's pro- cess, 77. Size for paper, 143. Slide rule, 52. Soda lime, 548. Sodelite, 658. Sodium (see Alkalies). acetate solution, 363, 603. carbonate, 69, 163, 560. chloride, solubility in amyl alcohol, 536. pyridine, 534. chloroplatinate, properties of, 231. determination, 239, 535, 537. as sodium antimoniate, 239. Berzelius, 226. hydroxide solution, 557. free carbonates, 65, 579. standard, 65. INDEX OF SUBJECTS. 775 Sodium, raonosulphide, 277. nitrate, analysis, 538. oxalate, 193. Sorensen's, 193. phosphate, 603. solution, 485. separation potassium, indirect process, 227. silicate, analysis of, 75. silicofluoride, analysis, 650. spectrum, 533. stannate, evaluation of, 314. - thiosulphate, 289. anhydrous, 289. standard solution, 290, 352. preservation, 290. tungstate, 589. uranate, analysis, 492. zirconate, 497. Solubility, meaning on analysis, 274. test for lead frits, etc. , 330. Soluble chlorides in clays, determination, 653. salts, 326. determination of, 321. in clays, 630. in clays, making innocuous, 633. sulphates, determination, 532. Solvents for use with Munroe's crucible, 108. Specific gravity, indirect separations by, 424. of solutions, 51. Spectrum analysis, 513. barium, 513. c.nesium, 533. calcium, 513. lithium, 533. potassium, 533. rubidium, 533. sodium, 533. strontium, 513. Sphene, 657. Spinel, 657. Split shovel, 131. Spot test indicators, 334. Squared paper, 53. Standard flasks, 47, 48. Biltz's, 47. - Giles', 47. Goske's, 47. Hirschsohn's, 47. Pfliiger's, 47. solutions, adjustment, 50. temperature corrections, 29, 49. Stands for burettes, 42. Stannic acids, 307. salts to stan nous, reduction, 312. Stannous chloride solution, 453. standard, 300. Starch, 286. iodide, 286, 352. Staurolite, 657. Stoppers for reagent bottles, 142. loosening fixed, 142. Strontium, 214, 245, 415. Strontium determination, 515, 516. effect on calcium precipitate, 513. magnesium precipitate, 513. Strontium, separation barium, 514, 515. chromate process, 515. indirect process, 510. Stromeyer and Rose, 514. calcium, 514, 515. chromate process, 515. Stromeyer and Rose, 514. spectrum, 513. Sulphides, 465. determination sulphur in, 624, 626. removal free sulphur from, 297, 342. Sulphites, determination of, 321. Sulphur, 637. contamination with during evaporation, ignition, 617. determination, 519, 531, 610. as barium sulphate, 618. combustion in bomb, 625. combustion process, 623. dry fusion process, 621, 622. Eschka's method, 621. in fluorspar, 649. in silicates, 617. - in sulphides, 624, 626. oxidation in air, 622. Parr's process, 622. volumetric, Raschig's process, 627. turbidity process, 631. volatile, 624. volumetric, Miiller's process, 627. wet process, 625. dioxide, 192, 445. determination of, 321. loss on ignition, 607. removal from metallic sulphides. 297, 342. Sulphuric acid, action on clays, 661. felspar, 662. hornblende, 663. mica, 663. nepheline, 663. quartz, 663. Swedish filter papers, 88. Syenite, 656. TABLES, analysis, 54. Talc, 571. Tannin, 479. solution, 333. Tantalic oxide, solubility in sodium bisulphate, 505. Tantalite, 504. Tantalum, 405, 496, 497, 517. detection, 407. determination, 418, 503, 504. Simpson's process, 418. fluoride, volatilisation of, 421. in clays, 405. separation, niobium and, 421. Marignac's process, 421. specific gravity process, 423. Weiss and Landecker's process, 422. Tartaric acid, destruction of, 208, 309. Tellurium and selenium, separation, 441, 442. detection, 439., 77 6 A TREATISE ON CHEMICAL ANALYSIS. Tellurium, precipitating agents, 441, 443. sulphide, 277. Temperature and volume, 28. correction standard solutions, 29, 49. for standard solutions, 30, 49, 199. Terbia, 500. Terra alba, analysis, 529. x Thioacetic acid, 412. Thiosemicarbazide, 480. Thoria, separation, 501. Thorianite, 510. Thorium, 444, 457, 495, 496, 500. determination, 505. as thorium thiosulphate, 501. Metzger's process, 510. hydroxide, properties, 506. iodates, 496. minerals, evaluation, 510. oxide, properties, 506. separation cerium, 505. didymium, 505. from yttrium, 504. from zirconium, 504. lanthanum, 505. sulphate, 614, 618. Tin, 318, 405, 517, 640. antimoniate, 308. alienate, 308, 405. ash, evaluation of, 314. chloride, volatilisation on boiling, 271. compounds, evaluation of, 314. determination, 177, 503. as sulphide, 309. electrolytic process, 312. Lowenthal's hydroxide process, 308. precipitation by cadmium, 307. zinc, 307. volumetric, Lenssen's iodine, 311. Lowenthal's, 311. Mene's ferric chloride, 310. ore opening, 266. oxide, 268. evaluation, 314. phosphate, 405. phosphomolybdate, 405. separation antimony and arsenic, 275. - fronTgold, 428. platinum, 428. from tungsten, 409. Donath and Miiller's process, 409. Rose and Rammelsberg's process, 409. Talbot's process, 410. molybdenum, 413. titanium, 268. sulphide, 277, 278, 309. volatilisation of, 309. Tintometer, 82. Lovibond's, 84. Titanic oxide, 169. Titanite, 657. Titanium, 444, 457, 517, 594. detection, 468, 486. determination, 169, 179, 181, 500, 503, 504. colorimetric, 203, 486. Titanium, determination, colorimetric Weller's, 203. effect of zirconium as, 209. gravimetric, Blair, 209. Gooch, 207. in permanganate titration iron, 189, 466. phosphate, 498, 608. salts, action hydrogen sulphide, 276. hydrolysis, 181. separation, 268. aluminium, 208, 456. cobalt, 469. iron, 208, 495. manganese, 469. nickel, 469. uranium, 492. zinc, 469. zirconium, 495, 496. sulphate solution, 205. 644. standard solution, 644. Titration, back, 70. direct, 69. minor, Lupp's, 402. mixed hydroxides and carbonates, 72. spot test indicators, 454. Tongs, platinum, 113. Blairs, 113. Topaz, 657. Tourmaline, 584, 585, 657. Trachyte, 656. Triangles, 113. Coleman's, 113. Hebebrand's, 113. nickel, 113. pipe clay, 113. platinum, 113. porcelain, 113. quartz, 113. Schmelck's, 113. Tungsten, 405, 473, 618. arseniate, 405. detection, 406. Defasqz's test, 406. determination, 503. > as trioxide, 407. - Berzelius' process, 408. effect on determination of antimony, 299. phosphate, 405. phosphomolybdate, 405. separation from antimony, 410. Talbot's process, 410. from arsenic, 411. Kehrmann's process, 411. from molybdenum, 416, 417. Hommel's process, 416. Pochard's process, 417. Rose's process, 417. from phosphorus, 411. Kehrmann's process, 411. from silica, 409. ammonia process, 409. hydrofluoric acid process, 409. potassium bisulphate process, 409. from tin, Donath and Miiller's pro- cess, 409. INDEX OF SUBJECTS. 777 Tungsten, separation from tin, Rose and Rammelsberg's process, 409. Talbot's process,. 410. vanadium and, 418. Tungstic oxide, 407. reduction of, 410. volatilisation of, 409. Turbidimetry, 85. Turbidity methods of analysis, 85. Turmeric test boric acid, 497, 597. zirconium, 497. ULTIMATE v. rational analysis clays, 670. consistency of, 673. Uranium, 211, 308, 415, 456, 457. acetate solution, 600. determination, 488. as ferrocyanide, 490. as oxide, 489, as phosphate, 490. effect of phosphoric acid, 491. volumetric permanganate, 491 . effect on solubility cerium oxalate, 504. lanthanum oxalate, 504. nitrate solution, 369, 600. oxides, analysis, 492. residues, recovery, 602. separation, 501. arsenic, 490. cobalt, 469. copper, 490. iron, 492. manganese, 469. nickel, 469. rare earths, 488. - titanium, 492. vanadium, 492. zinc, 469. VANADIUM, 415, 456, 457, 474, 594, 618. detection, 468. determination, 177. as lead vanadate, 484. colorimetric, 484. Gregory's, 484. mercury vanadate, 472. volumetric, 480, 481, 482, 483. Cain 'and Hosteller's process, 484. effect on permanganate titration, 465. determination of iron, 467. titanium, 468. in permanganate titration iron, 189. pentoxide, 480. separation chromium, 483. cobalt, 469. manganese, 469. nickel, 469. tungsten, 418. uranium, 492. zinc, 459. tetra-oxide, 480. Varnish for paper, 143. Vesuvianite, 584. Vivianite, 657. Volt, 253. Voltmeter, 253. Volume and temperature, 28. Volumetric process, analysis, 28, 45. WADS, analysis, 384. Wagner's solution, 522. Wash-bottles, 93. gas, Drechsel's, 547. hot water, 94. with Beutell's valve, 95. With Bunsen's valve, 94. with Griffin's valve, 94. with Waters' valve, 94. Washing precipitates, 87. theory, 95. Water combined, 157. density and temperature, 29. determination, 570. Brush and Penfield's process, 570. Danne's process, 573. fractional dehydration minerals, 573. high temperature, 571. -in talc, 571. Jannasch's process, 571. - distilled, 150. glass, analysis of, 75. volume and temperature, 29. Wax for etching test, 624. Weighing absorption tubes, errors in, 550. accuracy of, 15. bottles, Guttmann's 575. Levi's, 9. Reinhardt's, 100. by double vibrations, 11. by tares, 11. correction for buoyancy, 22. corrosive substances, 9. double, 11. electrification during, 551. hygroscopic substances, 8. volatile substances, 8. tubes, 8. Weights, estimation by measurement, 328. calibration, 17. correction of, 16. glass, 8. quartz, 8. Weston's cap for Bunsen burner, 268. White lead, 325. analysis, 319. drying, 319. Witherite, analysis, 531. Wolframite, 405, 504. Wollastonite, 658. XYLONITE paper, 8. YTTRIA earths, separation, 501. Yttrium, 496, 500. determination, 509. separation beryllium, 508. cerium, 504. thorium, 504. zirconium, 508. ZEOLITE, 657. Zinc, 211, 394, 395, 415, 445, 455, 456. aluminium alloys for reduction iron, 188. A TREATISE ON CHEMICAL ANALYSIS. Zinc, amalgamated, reduction of iron by, 187. delicacy of ammonia test, 177. determination as phosphate, 366. as sulphide, 360, 364. volumetric ferrocyanide, 367. effect of cadmium, 368. effect on silica determination, 359. - grey, 370. hydroxide, separation with, 471. metallic precipitation by, 305. ores, analysis, 359. oxalate, 427. oxide, 325. emulsion, 379. evaluation, 369. reduction of iron by, 187. retention by ferric oxide precipitate, 360. salts, action of hydrogen sulphide on, 360. separation, 469. basic acetate process, 364. cadmium, 273. cobalt, 364. manganese, 364. Zinc, separation nickel, 364. silicates, analysis, 359. sulphide, 273, 276. volati properties of, 360, 364. ilisation of, 366. Zircon, 657. Zirconium, 209, 444, 457, 518, 618, 637. detection, 497, 498. turmeric test, 497. determination, 499. as phosphate, 496, 497, 498. iodate, 496. minerals, 508. phosphate, 496. separation, 501. aluminium, 495, 496. beryllium, 508. cerium, 504. iron, 495, 496. . titanium, 495, 496. thorium, 504. yttrium, 508. volatilisation of, 500. Zoisite, 657. PRINTED BY NEILL AND CO., LTD., EDINBURGH. UNIVERSITY OF CALIFORNIA LIBRARY This book is DUE on the last date stamped below. DEC 22 1947 DEC 29 1947 8 1953 \ Maj'56BC C'D LD FEB27 1959 20Jan'50MB LD 21-100TO-12,'46(A2012si6)4120 94 Mf UNIVERSITY OF CALIFORNIA LIBRARY