METHODS IN CHEMICAL ANALYSIS ORIGINATED OR DEVELOPED IN THE KENT CHEMICAL LABORATORY OF YALE UNIVERSITY COMPILED BY FRANK AUSTIN GOOCH PROFESSOR OF CHEMISTRY AND DIRECTOR OF THE KENT CHEMICAL LABORATORY IN YALE UNIVERSITY FIRST EDITION NEW YORK JOHN WILEY & SONS LONDON: CHAPMAN & HALL, LIMITED COPYRIGHT, 1912, BY FRANK AUSTIN GOOCH Stanhope flfress F. H. GILSON COMPANY BOSTON, U.S.A. PREFATORY NOTE THE object of this volume is to present concisely the principal results reached by workers in the Kent Chemical Laboratory of Yale University in the investigation and develop- ment of methods in chemical analysis. In the account of processes, modified or original, only proved procedure and immediately related experimental data are, as a rule, given. For further details in respect to the elaboration of processes, the discussion of difficulties, and the experimental illustration of the effects of varying the prescribed procedure, references are given to the original sources from which this summary has been compiled. To his colleagues, Professors Philip E. Browning, R. G. Van Name, and W. A. Drushel, the compiler is much, indebted for valuable criticism of the proof sheets. 575454 CONTENTS. CHAPTER I. APPLIANCES AND GENERAL PROCEDURE. Mechanical Processes. The determination of products gaseous at ordinary temperatures by loss of weight, i. The distillation and condensation of vola- tile products, 3. The distillation and absorption of volatile products, 4; ap- paratus with ground joints, 4; apparatus with sealed joints, 5. The removal of volatile products from material to be reserved for treatment. The preven- tion of mechanical loss from solution in reactions evolving gaseous products, 6. The transfer of liquids and gases under pressure, 7. A convenient form of rotary shaker, 9. The purification of precipitates by solution and reprecipita- tion, 10. Electrolytic processes. The rotating cathode, n. The filtering crucible, 13; electrolysis and subsequent filtration, 13; electrolysis with filtra- tion, 16; electrolysis with continuous filtration, 17. The fixation of chlorine on the silver anode, 20; hydrochloric acid, silver anode, and platinum cathode, 20; sodium chloride, silver anode, and mercury cathode, 22. lodometric processes. The standardization of iodine solutions by the action of metallic silver, 27. Arsenic trioxide as an iodometric standard, 29. The starch indicator for free iodine, 29. Standard tartar emetic, 38. Processes of oxidation. Arsenic trioxide as a standard, 41 ; standardization without iodine, 41; standardization with the aid of iodine, 42. The gravimetric standardization of permanganate, 42. The loss of oxygen in oxidation by permanganate, 42; concentration of acid, 42; hydrochloric acid with ferrous salts, 48; hydrochloric acid with oxalic acid, 50; effect of other chlorides, 52. Acidimetry and Alkalimetry. The use of succinic acid as a standard, 54. Organic acids and acid anhydrides as standards, 56. The use of the iodide- iodate mixture and the estimation of iodine evolved, 59; determination of free acids, 59; determination of alkali hydroxides and carbonates, 60; determina- tion of acids liberated in hydrolysis, 61. The use of the bromide-bromate mixture and estimation of the bromine evolved, 70. Reaction of iodine with alkali hydroxides, 70. CHAPTER II. THE ALKALI METALS. Sodium. The detection of sodium, 74; the estimation of sodium as the pyrosulphate, 79. Potassium. The spectroscopic detection and determina- tion of potassium, 80; detection of potassium, 80; determination of potas- sium, 83. The separation and determination of potassium as the perchlorate, vi CONTENTS 88. The estimation of potassium as the pyrosulphate, 92. The volumetric estimation of potassium as the cobalti-nitrite, 93; potassium in the pure salt, 94; potassium in fertilizers, 96; potassium in soils, 97; potassium in animal fluids, urine, blood, lymph, milk, 98. Rubidium and Caesium. The spectro- scopic determination of rubidium, 102. Estimation of caesium and rubidium as the acid sulphates, 106. CHAPTER III. COPPER; SILVER; GOLD. Copper. The gravimetric determination of copper as the sulphocyanate, IO8; separation of copper from bismuth, antimony, tin, and arsenic, 112. The determination of copper as cuprous iodide and separation from cadmium, 114. The electrolytic determination of copper, 116. The iodometric estima- tion of copper, 1 1 8. The determination of copper by titration of the pre- cipitated oxalate with potassium permanganate, 123; precipitation of copper oxalate, 125; solubility of copper oxalate, 125; prevention of supersaturation, 129; precipitation in presence of acetic acid, 130; separations by Peters' pro- cedure, 131; separations by the method of desiccation, 132; separations in presence of acetic acid, 134; determination of copper associated with lead, 135. Silver. The gravimetric determination of silver as the chromate, 136. The electrolytic determination of silver, 138. The iodometric estimation of silver, based upon the use of potassium chromate as a precipitant, 140. The iodo- metric determination of silver, based upon the reducing action of potassium arsenite, 143. Gold. The electrolytic determination of gold, 145. The iodometric estimation of small amounts of gold, 146. The colorimetric deter- mination of small amounts of gold, 150. CHAPTER IV. BERYLLIUM; MAGNESIUM; CALCIUM; STRONTIUM; BARIUM. Beryllium. Ammonium beryllium phosphate, 153. The conversion of beryllium chloride to beryllium oxide, 153. The separation of beryllium chloride from ferric oxide, 154. Magnesium. The determination of mag- nesium by precipitation and ignition of ammonium-magnesium carbonate, 154. The determination of magnesium as the pyrophosphate, 156. The arsenate process for the separation of magnesium and the alkalies, 158. Calcium, Strontium, Barium. The detection of barium and strontium associated with calcium, and lead, 160. The separation of barium, strontium, and calcium by the action of amyl alcohol on the nitrates, 162; detection of strontium and calcium, 163; separation and estimation of strontium and calcium, 164; esti- mation of barium and calcium, 166; estimation of barium and strontium together, and of calcium, 166. The separation of barium and strontium by the action of amyl alcohol on the bromides, 167. The estimation of barium as the sulphate, 168; in presence of hydrochloric acid, 168; in presence of nitric acid and aqua regia, 170; purification of precipitated barium sulphate, CONTENTS Vii 172. The estimation of barium as the chloride, 174; precipitation by ether- hydrochloric acid mixture, 174; precipitation by acetyl chloride in acetone, 175; separation from calcium and magnesium, 177. The precipitation of barium bromide by ether-hydrobromic acid mixture, 179. The estimation of calcium, strontium, and barium precipitated as oxalates, 180; gravimetric determination of strontium and barium, 180; titration of oxalates by per- manganate, 181. CHAPTER V. ZINC; CADMIUM; MERCURY. Zinc. The estimation of zinc as the pyrophosphate, 185. The conversion of zinc chloride to zinc oxide, 186. The electrolytic determination of zinc, 186. The estimation of zinc by precipitation as the oxalate and titration with potas- sium permanganate, 187. Cadmium. The estimation of cadmium as the oxide, 1 88; precipitation as carbonate, 188; precipitation as hydroxide, 189. The estimation as the pyrophosphate, 190. The electrolytic determination of cadmium, 191; deposition from the sulphuric acid solution, 191; deposition from solutions containing acetates, 192; deposition from solutions containing cyanides, 193; deposition from solutions containing pyrophosphates or ortho- phosphates, 194. Mercury. The gravimetric determination of mercury as mercurous oxalate, 195. The determination of mercury by titration with sodium thiosulphate, 196. The estimation of mercury by precipitation as mercurous oxalate and titration of the excess of precipitant with permanganate, 197. The titration of mercurous salts with potassium permanganate, 198. CHAPTER VI. BORON; ALUMINIUM; LANTHANUM; THALLIUM. Boron. The gravimetric determination of boric acid, 201 ; the use of calcium oxide as a retainer, 201 ; the use of sodium tungstate as a retainer, 204. The acidimetric estimation of boric acid, 205; neutralization of stronger acids, 206; strengthening of boric acid by mannite, 208. The iodometric deter- mination of boric acid, 210. Aluminium. The determination of aluminium by precipitation with ether-hydrochloric acid/ 214; separation of alumin- ium from iron, 214; determination of aluminium and beryllium, 216; deter- mination of aluminium and zinc, 216; determination of aluminium and copper, 217; separation of aluminium from mercury and bismuth, 217. Lanthanum. The estimation of lanthanum precipitated as the oxalate, 218. Thallium. The determination of thallium as the acid sulphate and as the neutral sulphate, 219. The gravimetric estimation of thallium precipitated as thallic hydroxide by potassium ferricyanide and potassium hydroxide, 220. The gravimetric estimation of thallium as the chromate, 221. The iodometric estimation of thallium by precipitation with potassium dichromate and determination of the excess of the precipitant, 222. The estimation of thallium by the action of potassium ferricyanide in alkaline solution and of potassium permanganate in acid solution upon the ferrocyanide produced, 223. Viii CONTENTS CHAPTER VII. CARBON; SILICON; TITANIUM; ZIRCONIUM; CERIUM; TIN; LEAD. Carbon. The determination of carbon dioxide in carbonates by loss, 225 ; expulsion of carbon dioxide by the action of acid, 225; expulsion of carbon dioxide by ignition, 226. The precipitation and gravimetric determination of carbon dioxide, 228. The iodometric determination of carbon dioxide, 231; carbon dioxide in carbonates, 232. The combustion of organic substances in the wet way, 234; carbon content by the permanganate process, 234; carbon content by oxidation with chromic acid, 236; carbon dioxide evolved and oxygen used, 239. Silicon. The detection of silicon in silicates and fluo- silicates, 241. Titanium. The determination of titanic acid by reduction and titration with potassium permanganate, 242. Zirconium. The separa- tion of zirconium from iron by volatilization of the latter in hydrogen chloride, 244. Cerium. The separation of cerium from other cerium earths by the action of bromine upon the mixed hydroxides in presence of an alkali hydrox- ide, 244. The iodometric estimation of cerium, 246; digestion process, 246; distillation process, 247. The estimation of cerium oxalate by potassium per- manganate, 248. The estimation of cerium in presence of other rare earths by the action of potassium ferricyanide in alkaline solution and potassium per- manganate in acid solution, 249. Tin. The electrolytic determination of tin, 251. Lead. The detection of lead, 252. The electrolytic determination of lead as the dioxide, 252. The estimation of lead by precipitation as oxalate and titration with potassium permanganate, 254. CHAPTER VIII. NITROGEN; PHOSPHORUS; ARSENIC; ANTIMONY; BISMUTH; VANADIUM. Nitrogen. The determination of nitrogen liberated by action of sodium hypobromite upon ammonia compounds and derivatives, 256. The estima- tion of nitrates by expulsion of nitrogen pentoxide on ignition, 256. The estimation of nitrates by reduction with a ferrous salt and titration of the residual unoxidized salt, 258. The estimation of nitrates by reduction with ferrous chloride and measurement of the nitrogen dioxide evolved, 260. The iodometric determination of nitrates, 263; action of manganous chloride in hydrochloric acid, 263; distillation with phosphoric acid and potassium iodide and determination of iodine in the distillate, 266; decomposition by antimony trichloride, determination of oxidation of residue and iodine in the distillate, 268. The iodometric determination of nitrites, 269. The estimation of ni- trites, and of nitrites and nitrates in one operation, 271; determination of nitrites, 271 ; determination of nitrites and nitrates, 272. The estimation of nitrates and chlorates in one operation, 273. The qualitative separation and CONTENTS i3t detection of ferrocyanides, ferricyanides and sulphocyanates, 275; the ferro- cyanogen ion, 275; the ferricyanogen ion, 275; the ferrocyanogen ion, the ferricyanogen ion and the sulphocyanogen ion in mixtures, 276. The gravi- metric determination of sulphocyanates, 276. The volumetric estimation of sulphocyanates by potassium permanganate, 279. Phosphorus. The determination of phosphoric acid by precipitation as ammonium magnesium phosphate and weighing as magnesium pyrophosphate, 282. The iodometric determination of phosphorus in iron, 283. The estimation of phosphoric acid and phosphorus precipitated as ammonium phospho-molybdate, 285. The determination of phosphoric acid by precipitation as uranyl phosphate and estimation of the uranium volumetrically, 286. Arsenic, Antimony, and Tin. The determination of arsenic by precipitation as ammonium magnesium arsenate and weighing as magnesium pyroarsenate, 288. The iodometric esti- mation of arsenic acid, 291; reduction by hydriodic acid and oxidation by iodine in alkaline solution, 291; reduction by hydriodic acid and titration of iodine liberated, 295. The detection and approximative estimation of minute quantities of arsenic in copper, 301. The separation of arsenic from copper by precipitation as ammonium magnesium arsenate, 305. The iodometric deter- mination of antimonic acid, and of antimonic acid and arsenic acid, 308. The separation of antimony from arsenic by the simultaneous action of hydrochloric acid and hydriodic acid, and the estimation of antimony in the residue, 311. The detection of arsenic, and of antimony with tin, in mixtures containing compounds of these elements, 312; action of hydrochloric acid and potassium iodide, 313; action of hydrochloric acid and potassium bromide, 316. The iodometric determination of arsenic and antimony, and associated copper, 318. The estimation of arsenic, antimony, and tin in the lower condition of oxida- tion by the action of potassium ferricyanide in alkaline solution and potassium permanganate in acid solution, 322; determination of antimony in anti- monious condition, 323; determination of tin in stannous condition, 323; determination of arsenic in arsenious condition, 324. The estimation of arsenic acid and antimonic acid associated with vanadic acid, 325. Vanadium. The gravimetric estimation of vanadic acid based on liberation of iodine and absorption of that element by silver, 325. The precipitation of ammonium vanadate by ammonium chloride, 326. The estimation of vanadium as silver vanadate, 328. The estimation of vanadic acid by the action of the halogen acids, 330; the action of hydrochloric acid, 330; the action of hydrobromic acid, 335; the action of hydriodic acid, 337. The determination of vanadic acid by reduction in acid solution and reoxidation by iodine in alkaline solu- tion, 341; reduction by organic acids, 341; reduction by hydriodic acid, 343; reduction by hydrobromic acid, 345. The use of the Jones reductor in the estimation of vanadic acid, 346; regulation of reduction by the use of silver sulphate, 348; registration of reduction by use of ferric sulphate, 349. The estimation of vanadic and arsenic acids and of vanadic and antimonic acids in presence of one another, 350. The estimation of vanadic acid associated with chromium, with molybdenum, and with iron, 352. The estimation of vanadium in the tetroxide condition by the action of potassium ferricyanide in alkaline solution and potassium permanganate in acid solution, 352. CONTENTS CHAPTER IX. OXYGEN; SULPHUR; SELENIUM; TELLURIUM. Oxygen. The iodometric determination of oxygen in air and in aqueous solution, 355; determination of oxygen in air, 355; determination of dissolved oxygen, 360. The estimation of oxidizers by the gravimetric determination of iodine set free in reaction, 361; potassium permanganate, hydrogen dioxide, potassium dichromate, ferric chloride, 362. Sulphur. The detection of sul- phides, sulphates, sulphites, and thiosulphates in presence of one another, 363. The iodometric determination of thiosulphates, 364. The iodometric deter- mination of sulphites in alkaline solution, 366. The determination of dithionic acid and dithionates, 369. The determination of persulphates, 370; arsenate- iodide method, 370; method of Le Blanc and Eckardt, 371 ; method of Grutz- ner, 372; method of Mondolfo, 374; method of Namias, 374. Selenium. The gravimetric estimation of selenious acid by liberation of iodine and ab- sorption of that element by silver, 375. The gravimetric determination of selenious acid by precipitation of selenium, 376. The iodometric determina- tion of selenious acid by methods based upon the action of potassium iodide in presence of acid, 377; the contact method, 377; the distillation method, 379; the differential method, treatment of the residue, 380. The determination of selenious acid by potassium permanganate, 382. The determination of sele- nious acid by the direct action of sodium thiosulphate, according to the method of Norris and Fay, 383. The iodometric determination of selenic acid by the action of the halogen acids, 385; reduction by hydrochloric acid, with distilla- tion, 385; reduction of hydrobromic acid, with distillation, 386; reduction by hydriodic acid, with distillation, 388; reduction by hydriodic acid, differential method, 388. The separation of selenium from tellurium by procedure based upon the difference in volatility of the bromides, 390. Tellurium. The gravimetric estimation of tellurous acid by the liberation of iodine and absorp- tion of that element by silver, 394. The determination of tellurous acid by oxidation with potassium permanganate, 394; oxidation in presence of a chlo- ride, 396; oxidation in presence of a bromide, 397. The determination of tellurous acid by the precipitation of tellurous iodide, 398. The iodometric estimation of tellurous acid, 399. The iodometric determination of telluric acid, 401. The precipitation of tellurium dioxide and the separation of tel- lurium from selenium, 402. CHAPTER X. CHROMIUM; MOLYBDENUM; URANIUM. Chromium. The estimation of chromium as silver chromate, 406. The iodometric determination of chromic acid, 407. The iodometric estimation of chromic acid and vanadic acid, 409. The estimation of chromic acid and vanadic acid by reductions and oxidations, 411. The volumetric estimation of chromium in the chromic condition, 413. Molybdenum. The gravimetric estimation of molybdic acid by liberation of iodine and absorption of that element by silver, 414. The iodometric estimation of molybdic acid, 415; the CONTENTS XI digestion method, 415; distillation process, 416; reoxidation of the residue by iodine, 420; reoxidation of the residue by permanganate, 421. The estima- tion of molybdic acid reduced in the Jones reductor, 424. The determination of molybdic acid and vanadic acid by reductions and oxidations, 427. Ura- nium. The determination of uranium by the aid of the Jones reductor, 430. CHAPTER XI. FLUORINE; CHLORINE; BROMINE; IODINE. Fluorine. The detection of fluorine, 432. The acidimetric estimation of fluosilicic acid, 432. The iodometric estimation of fluosilicic acid, 435. The estimation of fluorine evolved as silicon fluoride, 436; elimination of silicon fluoride at high temperatures, 436; iodometric determination of fluorine in fluorides, 439. Chlorine, Bromine, Iodine. The detection of iodine, bromine, and chlorine in presence of one another, 440. The determination of free chlorine and free bromine by liberation of iodine and absorption of that element by silver, 443. The gravimetric determination of iodine by absorption by metallic silver, 444; free iodine, 444; iodine in iodides, 446. The determina- tion of halogens in benzol derivatives by the use of metallic potassium, 447. The direct determination of chlorine in mixtures of alkali chlorides and iodides, 449; use of ferric sulphate, 449; the nitrite method, 451. The direct deter- mination of bromine (and chlorine) in mixtures of alkali bromides (and chlo- rides) with iodides, 452. The application of iodic acid to the analysis of iodides, 454. The iodometric determination of iodine in haloid salts, 457. The deter- mination of the halogens by the electrolytic reduction of silver in mixed silver salts, 459; silver chloride and silver bromide, 460; silver iodide by itself, and in mixture with silver chloride or silver bromide, 461. The estimation of chlorates by reduction with ferrous sulphate, 462. The iodometric estimation of chlorates, 463. The detection of alkali perchlorates associated with chlo- rides, chlorates, and nitrates, 465. The iodometric determination of perchlo- rates, 467. The estimation of bromates by reduction with ferrous sulphate, 471. The iodometric estimation of bromates, 471; reduction by hydriodic acid, 471; reduction by arsenious acid, 474; reduction by arsenate-iodide mixture, 475. CHAPTER XII. ' MANGANESE; NICKEL; COBALT; IRON. Manganese. The determination of manganese as the sulphate, 477. The determination of manganese as oxide, 478. The determination of manganese separated as the carbonate, 481. The determination of manganese precipi- tated as ammonium manganese phosphate and weighed as manganese pyro- phosphate, 482. The electrolytic determination of manganese, 485. The determination of manganese precipitated by the chlorate process, 487. Nickel (Cobalt). The electrolytic determination of nickel with the rotating cathode, 489. The estimation of nickel by precipitation as the oxalate and titration with potassium permanganate, 490. The detection of nickel in presence of XU CONTENTS cobalt, 491. The separation of nickel and cobalt by the etherial solution of hydrochloric acid, 492. Iron. The determination of iron in the ferric state by reduction with sodium thiosulphate and titration of the excess of the latter with . >dine, 492. The standardization of permanganate in iron analysis, 495. The benavior of ferric chloride in the Jones reductor, 497. The effect of nitric acid in the titration of a ferrous salt by potassium permanganate, 498. The permanganate estimation of iron in presence of titanium, 499. The esti- mation of iron by potassium permanganate after reduction with titanous sul- phate, 502. Separations of iron by volatilization in gaseous hydrogen chloride, 504; iron and aluminium, 506; iron and beryllium, 507; iron and chromium, 507; iron and zirconium, 508. The estimation of iron and vanadium in presence of each other, 508. The estimation of ferric iron, vanadic acid, and chromic acid in presence of one another, 510. METHODS IN CHEMICAL ANALYSIS CHAPTER I. APPLIANCES AND GENERAL PROCEDURE. MECHANICAL PROCESSES. The Determination of Products Gaseous at Ordinary Temperatures by Loss of Weight. Various forms of apparatus have been designed for determin- ing, by loss of weight, reaction products which are gaseous at ordinary temperatures, but many of these are cumbersome or require skill in glass blowing for their construction. A form of apparatus described by Kreider* is light and easily | made from three test tubes, modified and fitted as shown in the figure. The test tube, A, serves as the reaction chamber. B is perforated with a hole about I cm. in diameter and fits tightly within A; and C, so selected that it fits loosely within B, is drawn out to a small capillary tube. When the apparatus is to be used, the capillary of C, which has been fitted as described, is pushed through the hole of B, packed loosely with cotton; B is filled to the depth of from 6 cm. to 8 cm. (about two-thirds of its contents) with granular calcium chlo- ride; and B and C are adjusted as shown. To the test tube, C, is fitted a one-holed stopper, through which passes a short glass tube which is to be closed by a rubber cap and plug. Upon removing the plug, and applying suction to the short tube, the reagent employed to liberate the volatile product to be deter- mined is drawn up through this capillary until C is sufficiently filled. Upon replacing the plug the reagent remains within C, held by atmospheric pressure. Gentle pressure upon the cap * J. Lehn Kreider, Am. Jour. Sci., [4], xix, 188. i 2 METHODS TO CHEMICAL ANALYSIS expels a drop of liquid from the capillary, and upon the release of the cap a little air is drawn in to allow for expansion of air in the large tube without loss of liquid during subsequent handling. The tubes A and B are so selected that very little of the product evolved can escape between them, and, in case they fit very loosely, a ring of paraffin melted into the mouth of A, about B, by means of a hot wire, seals the joint securely. A very con- venient way to attach the paraffin is to melt it between A and another tube, which fits A, as does B, and may be removed by a turning motion, leaving the ring into which B will fit. Very little heating is then required to make a tight joint. If care be used in taking apart A and B, at the close of an experiment, such a ring of paraffin remains in place and may be used many times without replacement, being remelted by a touch of the hot wire before every new experiment. In making a determination, the substance under examination is weighed and placed in the bottom of A. The reagent to be employed, 10 cm. 3 to 15 cm. 3 , is drawn into C, and held there in the manner described. The test tube A is slipped over B, and the joint is sealed with paraffin, as has been shown. The appa- ratus is wiped, placed on the balance and weighed. Upon removing the cap from the small tube in C, the reagent runs from C into A. The volatile product, forced upward through the drying column of calcium chloride, escapes through the annular space between B and C. When action ceases, a current of dry air is forced through C, to remove all the volatile product, the cap is replaced, and the apparatus is weighed. The loss of weight represents the volatile product. Hydrogen by Loss. Metal taken, grm. Hydrogen found, grm. Error, grm. Magnesium Zinc O. IOOO 0. IOOO O. IOOO O. IOOO O. IOOO O. 200O O . 2OOO O. 2OOO . 20OO O. 2OOO 0.0087 0.0085 0.0084 o . 0084 o . 0083 0.0061 0.0062 0.0062 o . 0060 0.0061 +0.0003 -j-o.oooi o.oooo 0.0000 O.OOOI o oooo +O.OOOI -j-o.oooi O.OOOI o.oooo APPLIANCES AND GENERAL PROCEDURE 3 Tests of this apparatus in the determination of carbon dioxide in carbonates, and of nitrogen in urea and in ammonium salts, are described later. In the preceding table are given, results of experiments made to determine thus the weights of hydrogen liberated by the action of magnesium and zinc upon dilute hydrochloric acid. The Distillation and Condensation of Volatile Products. The rapid evaporation of liquid charged with soluble or in- soluble matter is apt to carry mechanically to the distillate some material which should remain in the residue. A form Fig. 2. of apparatus elsewhere described* and shown in the accom- panying figure (Fig. 2) solves the problem successfully. The retort, made of a pipette, bent as shown, with stoppered funnel * Gooch, Am. Chera. Jour., ix, 28. 4 METHODS IN CHEMICAL ANALYSIS sealed on or attached by a rubber joint, is fitted to an upright condenser which, in turn, is connected by a stopper to a thistle tube, fitted tightly to the receiver by means of a stopper per- forated or grooved to permit the passage of air. For work to be described the apparatus has been modified by substituting for the perforated or grooved stopper a tight stopper carrying a bulbed trap.* In making a distillation, the liquid is introduced by the funnel, the glass cock is closed, the water started through the condenser, and the retort, not more than half filled and inclined backward, is carefully heated. For the heating a paraffin bath is in many cases most convenient, and it is advantageous to lower the retort into the paraffin, already heated to a temperature considerably above the boiling point of the liquid, so that evaporation may take place rapidly and often without actual boiling. The diam- eter of the gooseneck should be at least 0.7 cm. to prevent the formation of bubbles within it. The use of this apparatus in the determination of boric acid is described elsewhere. The Distillation and Absorption of Volatile Products. Apparatus with A convenient apparatus for the distillation and Ground joints, absorption of volatile products f is easily constructed, with glass joints throughout, by sealing together a separating fun- nel A, a Voit flask B, a Drexel wash bottle C, and a bulbed trap g, as shown in the figure. Upon the side of the distillation flask B is pasted or etched a gradu- ated scale, by means of which the volume of liquid within the flask may be known at any time. The separating funnel is con- .nected with a Kipp generator set up for the delivery of carbon di- oxide, hydrogen, hydrogen chlo- Fl S- 3- r ide or other suitable gas. The flask serves as the retort,. the wash bottle properly charged as * See trap of Fig. 7, p. 6. t F. A. Gooch and John T. Norton, Jr., Am. Jour. Sci., [4], vi, 168, APPLIANCES AND GENERAL PROCEDURE 5 the receiver, and the products of distillation are swept forward by the generator gas, which may serve as a reagent or simply as a medium for aiding the transfer of products from the retort to the receiver. This apparatus has served a useful purpose in processes to be described for the determination of molybdenum, vanadium, and iodine liberated from the iodide-iodate mixture by acids free or evolved. A similar device adapted to double distillation is shown in Fig. 4. will be given later.* Fig. 4- Application of this apparatus Fig- 5- Apparatus with Fig. 5 shows a convenient device by Ed gar, f put Sealed joints, together without ground joints, for the distillation and absorption of volatile products. The distillation retort, simi- lar in design to that of Fig. 2, consists of a modified pipette, * F. A. Gooch and A. W. Peirce, Am. Jour. Sci., [4], i, 181. f Graham Edgar, Am. Jour. Sci., [4], xxvii, 174. 6 METHODS IN CHEMICAL ANALYSIS with the inlet tube bent upward and sealed to a separator/ funnel while the outlet tube, expanded to a small bulb, is bent upward and then downward to enter the absorption flask. A slow current of hydrogen, or other suitable gas, is made to enter at the bottom of the retort to stir the liquid so that a very small volume may be distilled without danger of " bumping." The Removal of Volatile Products without Loss of Non-volatile Material Reserved for Treatment. In processes which involve the elimination of a volatile reagent or product of reaction from a boiling solution, it is often essential to prevent losses by spattering or by mechanical transfer of non- volatile material in the steam. In many such processes the simple device shown in the figure is effective in preventing appreciable error by loss.* A flask, preferably of the Erlenmeyer shape, with a broad bottom, permits boiling of the liquid in a shallow layer favorable to the checking of explosive ebullition. A two-bulbed trap, made by cutting short an ordinary calcium chloride drying tube and hung with the large opening downward, ob- structs the steam while permitting sufficient relief of pressure and thus serves to catch and return to the liquid particles of the non-volatile matter thrown upward. The Prevention of Mechanical Loss of Solution in Reactions Evolving Gaseous Products. The danger of mechanical loss in reactions ac- companied by effervescence (as in the neutraliza- tion of carbonates by strong acids) or by formation of spontaneously volatile product (as in the libera- tion of iodine to be subsequently titrated) may be minimized by making use of a trapped reaction chamber. For this purpose the apparatus shown in the figure is serviceable.! It consists of a Drexel washing bottle with a separatory funnel sealed to the inlet tube, and a Will and Varrentrapp absorp- Flg ' 7> tion apparatus joined to the outlet tube. The reaction is brought * F. A. Gooch and P. E. Browning, Am. Jour. Sci., [3], xxxix, 197. f F. A. Gooch and C. F. Walker, Am. Jour. Sci., [4], iii, 293. u APPLIANCES AND GENERAL PROCEDURE 7 about by admitting the appropriate reagents through the funnel tube to the solution to be acted upon in the cylinder, so that all volatile products must escape through the properly charged absorption bulbs. The Transfer of Liquids and Gases under Pressure. A simple form of force pump, with Bunsen valves of special construction, has been described by Kreider.* Valve. In forcing a liquid or gas indifferent to rubber from one vessel to another, the ordinary Bunsen valve is apt to collapse in such a way as to permit a back flow. Kreider finds that a stout glass tube of desirable size, sealed at one end and drawn out with an opening in the constriction, as indi- cated in the accompanying figure, and a piece of rubber tubing containing a smooth slit placed over it, makes a valve in which collapse is impossible. A valve similar in appearance to the one here described has been previously used ; but the similarity is confined to the appearance, as will be evident from the following description: The con- striction should not be greater than is necessary to leave a small space between the glass and the rubber when the latter is loosely drawn over it ; but it should be long enough to permit a slit of about a centimeter's length in the rubber to close tightly, or about twice the length of the slit. A slit one centimeter long will be found to open under very slight pressure, and, to accomplish its purpose, it is only required to close sufficiently for the external pressure to force the rubber against the opening in the tube. This opening should be carefully rounded and a little higher rather than any lower than the surrounding glass, and is better made before seal- ing the end, in order to keep the tube perfectly straight. The rubber should fit tightly about the larger parts of the glass tube and be put on with care to have the smoothly cut slit straight, and loose enough to close tightly. If the slit is placed about 90 from the opening in the tube, sufficient space will remain to permit the escape of the gas or liquid, but the moment the pres- sure outside becomes greater than that within, the rubber will be pressed tightly over this opening and thus a return made im- possible. When dry the valve does not resist high pressure per- fectly; but when wet, or better, when both glass and rubber, * D. A. Kreider, Am. Jour. Sci., [3], 1, 132. 8 METHODS IN CHEMICAL ANALYSIS including the slit, are moistened with glycerin, a nearly perfect vacuum may be retained for several days. The valve thus lubricated with glycerin, when used as a protection in an am- monia wash bottle, will prevent absolutely the access of am- monia to the mouth, and if made according to the directions will act with very little pressure. Placed in the connection between the vacuum flask and water pump ordinarily used in filtration, it has been found a valuable check on the valve of the pump, and when the latter fails this device prevents the back flow of water into the filtrate. In processes which necessitate the use of a partial vacuum, this valve may be employed to hold the vacuum in continual readiness. Force Pump. By adjusting two of the valves just described to the opposite extremities of a T-tube, with the horizontal limb enlarged or sealed to a larger tube so as to permit the attachment of a large and stout piece of rubber tubing closed at one end, as shown in Fig. 9, a convenient and powerful little force pump is obtained. A stout T-tube of small-bore is cut off short at the two ends at right angles to one another; to one is sealed a tube just large enough to permit the insertion of a valve; to the other, a large tapering tube, slightly lipped so as to hold a piece of rubber tubing firmly and allow of tying the latter if necessary. Of the third end of the tube, a valve like that shown in Fig. 8 is made. The com- pressing rubber should not be of greater length than Fig. 9. the hand is able to cover completely, and may be closed with a glass stopper selected to fit tightly. Providing the space through the T-part is kept at a minimum compared with that of the compressing rubber, rapid pumping will be found possible and the power limited only by the strength of the user's grip. The apparatus may be quickly constructed of materials always at hand. Originally it was made, in about fifteen minutes, of a T-tube to which the necessary enlargements were connected by rubber tubing and the unused space filled by a glass rod. The valves may be inserted directly into the ends of the compressing rubber, but the form shown in the figure is more serviceable. By attaching the lower end to a tapering tube as shown, the pump is easily inserted into a perforated stopper of any size. APPLIANCES AND GENERAL PROCEDURE 9 The pump has been found serviceable in various applications. For filling burettes it is better than a siphon, the stoppers of the standard solution bottles being provided with two holes, through one of which the delivery tube passes, while to the other the pump is applied by the adapter shown in the figure. It may be applied to a Kipp generator in which higher pressure is momen- tarily required. In various other ways it has been found to be a useful piece of apparatus. A Convenient Form of Rotary Shaker. An apparatus designed by Perkins* serves admirably for put- ting a liquid into rotary motion for the purpose of securing gentle but thorough agitation. The container is an Erlenmeyer Fig. 10. flask. This is suspended in a retort clamp held loosely in another clamp, which in turn is also loosely held by another clamp firmly attached to the upright rod, the whole forming a system of loose joints at right angles, which permits oscillatory movement of the flask. Motion is given to the flask by a wire crank attached eccentrically to the rotating table driven by the motor. The use of this apparatus in the absorption of iodine, free or liberated in reaction, by metallic silver will be described later. * Claude C. Perkins, Am. Jour. Sci.. [4], xxviii, 33. 10 METHODS IN CHEMICAL ANALYSIS The Purification of Precipitates by Solution and Reprecipitation. In many processes of analytical chemistry, the preparation of substances in pure condition is brought about by precipita- tion, solution and reprecipitation ; and sometimes this cycle of operations must be repeated. When a precipitate, gathered upon a filter, is easily acted upon by the appropriate solvent, the process of dissolving the precipitate from the filter is simple; but when the precipitate is refractory toward solvents or difficult to attack on account of its physical condition, as is the case with many gelatinous precipitates, the proper handling of the precipi- tate involves some inconvenience and delay. In meeting such difficulties, it is advantageous to place within the ordinary paper filter, before filtering, a movable lining of platinum gauze upon which the precipitate rests for the most part and with which it may be re- moved.* The simplest form of this device is easily made by cutting platinum gauze to the shape shown in the accompanying figure. In ordinary use, this piece of gauze, folded to make a cone of a little less than 60, and held by pincers at the point of overlapping, is placed Fig. ii. within this filter and allowed to fit itself closely by the natural spring of the gauze when released. Upon filters so prepared a precipitate may be collected and washed as usual; and, at the end of the operation, the cone with nearly all the precipitate may be transferred (conveniently by means of ivory-pointed pincers) to a dish or beaker for suitable treatment. The small amounts of the precipitate which have passed through the gauze, being somewhat protected by the gauze against the compacting action of filtration and washing, are generally removable with ease from the filter by a jet of the washing liquid. After washing, the gauze may be replaced within the same filter and serve for a second collection of the precipitate, to be subsequently dissolved, in case double precipi- tation and solution are desirable. The final collection of the precipitate is, of course, made upon paper without the gauze lin- ing, when precipitate and filter are to be ignited. * Gooch, Am. Jour. Sci., [4], xx, n. APPLIANCES AND GENERAL PROCEDURE II This device has proved to be very serviceable in handling such precipitates as ferric hydroxide, aluminium hydroxide and the basic acetates. Precipitates collected upon asbestos in the perforated crucible are frequently removable without difficulty by allowing a suitable solvent to percolate precipitate and felt; but in case the precipitate is pasty solution in this man- ner may be unpleasantly slow. In such cases, it is convenient to remove the greater part of the precipitate, collected and washed in the usual manner, upon a disk of platinum foil, perforated, fitted with a wire handle, as shown in the figure, and placed upon the asbestos felt before the transfer of the precipitate to the crucible. To make such a disk, shown in Fig. 12, is the work of a few moments only; and by its use pasty Fi precipitates, such as cuprous sulphocyanate or the sulphides of the metals, are easily handled for solution. ELECTROLYTIC PROCESSES. The Rotating Cathode. The rotating cathode, previously utilized in the arts in the manufacture of seamless copper tubing (by electroplating with currents of low electromotive force and continuous replenish- ment of the bath by the use of a soluble copper anode) , and by von Klobukow* for slow stirring of the electrolytic bath in analy- sis, has been applied in rapid motion by Gooch and Medwayf- to analytical processes in which the object is to remove metals completely and expeditiously from solution. An ordinary 20 cm. 3; platinum crucible is used as the cathode, and this is rotated at a speed of from 600 to 800 revolutions by means of a small, inexpen- sive electric motor fastened so that its shaft is vertical. Upon this shaft the crucible is fixed by pressing it over a rubber stopper bored centrally and fitted tightly on the end of the shaft. To secure electrical connection between crucible and shaft, a narrow strip of sheet platinum is soldered to the shaft and then bent * Jour, prakt. Chem. (N. F.), xxxiii, 473. f F. A. Gooch and H. E. Medway, Am. Jour. Sci., [4], xv, 320 12 METHODS IN CHEMICAL ANALYSIS upward along the sides of the stopper, thus putting the shaft in contact with the inside of the crucible when the last is pressed over the stopper. The shaft is made in two parts as a matter of convenience in removing the cruci- ble, and is joined, with care to make a good contact between the two pieces of shafting, by a rubber con- nector of sufficient thickness to pre- vent the crucible from wabbling when rotated. The solution to be electrolyzed is placed in a beaker upon a small ad- justable stand, so that the crucible may be dipped into the liquid to any desired depth. A platinum plate is employed as an anode, and this is connected to the positive pole of a series of storage batteries, while the negative pole of this series is con- nected to the bearing in which the shaft rotates, thus allowing the cur- rent to go from the platinum plate through the solution to the crucible, up the shaft of the motor, and back to the batteries. The power to run the motor may be conveniently taken from the incandescent light circuit. The stand carrying the beaker is raised until the liquid covers about two-thirds of the crucible adjusted to the shaft, thus giving a cathode surface of about 30 cm. 2 . The anode is introduced and the motor started. The wires from the storage batteries are connected and the current allowed to pass through the solution. The duration of the electrolysis is varied according to the strength of current used, but in each case, after the deposit is nearly com- plete, the current from the batteries is shut off, the motor stopped, the beaker, platinum anode and crucible carefully washed with a fine jet of water, the motor again started, and the current allowed to pass for the remaining time. When the deposition is complete the crucible is removed and washed, first with water, then with alcohol, and finally is dried by passing it over a flame. In subsequent study of the material and shape of the rotating Fig. 13. APPLIANCES AND GENERAL PROCEDURE 13 cathode, Medway* has shown that a silver crucible may, with some economy and without sacrifice of accuracy, be substituted for the platinum crucible, at least in the determination of copper; that neither nickel nor aluminium is a suitable metal for use as the cathode; and that a rotating disk of platinum is inferior to the crucible for use as the cathode. The results of experimental tests of the rotating cathode in the determination of various metals copper, silver, nickel, cadmium, tin, gold, zinc are given under the headings of these metals. The Filtering Crucible in Electrolytic Analysis. The rapidity with which a metal or oxide may be thrown upon the electrode and thereafter handled successfully in the ordinary processes of electrolytic analysis depends upon keeping to con- ditions under which deposits are compact and adherent. It is for the purpose of getting adherent deposits that in modern rapid processes use is made of rotating electrodes,! of apparatus so arranged that gases evolved or introduced shall stir the liquid, { and of the agitating action of a magnetic field. The use of these methods is, however, limited to those cases in which attainable conditions and the nature of the processes are such that the deposits may be handled and washed without loss of material from the electrode. Plainly, the range of con- ditions and processes may be very much extended by the adop- tion of means for handling easily and safely electrolytic deposits more or less loose. Gooch and Beyer || have made use of devices for this purpose, in which the filtering crucible of platinum or of porcelain is adapted to use as an electrolytic cell. Electrolysis and Fig. 14 shows a convenient form of apparatus for Subsequent such use in electrolytic analysis. The crucible (A), fitted in the usual manner with an asbestos felt (a), serves as an electrode (e) } the surface of which is very much * Am. Jour. Sci., [4], xviii, 180. t V. Klobukow, Jour, prakt. Chem. (N. F.), xxxiii, 473. Gooch and Med- way, Am. Jour. Sci., [4], xv, 320. (See page u.) Exner, Jour. Am. Chem. Soc., xxv, 896. t Levoir, Zeit. anal. Chem., xxviii, 63. Richards, Jour. Am. Chem. Soc., xxvi, 530. Frary, Zeit. Elektrochem., xiii, 308. Jour. Am. Chem. Soc., xxix, 1592. || F. A. Gooch and F. B. Beyer, Am. Jour. Sci., [4], xxv, 249. METHODS IN CHEMICAL ANALYSIS A F====== L Fig. 14- increased by a layer of pieces of platinum foil (b) within the crucible and in contact with its walls. The joint between cap and crucible is made water-tight by a thin rubber band (F). The capacity of the cell is made con- veniently ample by attaching to the cruci- ble, by means of a close-fitting, thin rubber band (E), a glass chamber (C) easily made from a wide, short test tube. The second electrode (/) is introduced from above through the glass funnel (D), which serves to prevent spattering of the liquid during the electrolysis, and hangs within the glass chamber. The cell, held by a clamp, may be kept cool during action by immersing it in water contained in a cooler, as indicated in Fig. 15. Electrical connection is made with the crucible by means of a platinum triangle (c), bent as shown and held tightly against the outer wall of the crucible by a rubber band (d). Fig. 15 shows, on the left, the apparatus adjusted for work. In using the apparatus, the crucible, fitted with asbestos and containing clippings of platinum foil, is capped, ignited and weighed. The glass chamber with the wide rubber band folded back against itself is set upon the crucible and the band is snapped into place. The other adjustments are made in the manner shown. The electrolyte is introduced and the current turned on. After the expiration of time enough to complete the electrolysis, the cooler is lowered and arrangements are made to draw off the liquid in the cell. If the process is such that no harm can follow the stopping of the current before removing the liquid, the upper electrode and funnel are washed and removed, the cap and band are slipped off, and the apparatus is set in the holder of the filter- ing flask as for an ordinary filtration. The liquid is drawn through the felt to the flask, the chamber washed down, and re- moved from the crucible, and the deposit is well washed. The crucible and contents are dried and weighed, the increase over the original weight being, of course, the weight of the deposit. The details of experiments made to test this form and use of the apparatus are given in the table. Copper sulphate strongly APPLIANCES AND GENERAL PROCEDURE 15 acidulated with sulphuric acid was the electrolyte. Deposition was completed in the times given, and the ferrocyanide test ap- plied to the whole nitrate showed the absence of copper in every case. The apparatus and deposit were washed first with water and finally with alcohol. It was noticed that, though the filtrate contained no copper, the washings did sometimes contain a bare trace. When the filtrate was allowed to stand after treatment with potassium ferrocyanide it turned blue rapidly, and this Fig. is- action, which indicated probably the presence of hydrogen diox- ide or of persulphuric acid produced in the electrolysis o the sul- phuric acid, suggests that the liquid should be drawn from the deposit as quickly as may be after the current is cut off. In the first two experiments no special care was taken in this respect, and in these experiments the results are a trifle higher than those of the other experiments, in which the manipulation was quickly made. Obviously this process of electrolytic analysis is fairly rapid, easily executed, and accurate; but the desirability of quickly i6 METHODS IN CHEMICAL ANALYSIS removing the liquid from the deposit after stopping the current is evident. Electrolysis with Filtration after Interruption of the Current. CuSO4.5HjO taken. grm. Volume of liquid. cm. 1 H 2 S0 4 cm.* Current. Time, min. Theory for copper. grm. Copper found. grm. Error, grm. Amp. Volt. 0.5038 50 5 i: 5 7 41 0.1283 0. 1290* +0.0007 0.5010 50 5 U 5 7 40 I o. 1276 0.1282* +0 . 0006 0.5009 50 & s \ 2 4 5 7 4! o. 1276 0.1279* +0.0003 0.5005 50 ^' 5 ft 5 7 0.1275 0.1277* + O.OOO2 0.5047 50 f 5 i: 5 7 40 \ o. 1285 0.1286* +0.0001 0.5039 50 '' 5 i: 5 7 ,11 ^0.128^ ,, 0.1285* + O.OOO2 o . 5030 50 5 5 4 o. 1281 O.I282t +O.OOOI 4 7 4 ) * No copper in filtrate or in washings. t Trace of copper in washings. Electrolysis Hollowing are results obtained as in the preceding with Filtration, process excepting the single point that the liquid was drawn off while the current was still running. In these experi- ments the nitration was effected by removing the cooler, taking off the cap and band from the crucible, and quickly swinging into place the filtration apparatus shown at the right in Fig. 15. The liquid was then drawn through the crucible and replaced by wash Electrolysis and Filtration without Interruption of the Current. CuS04.sHO taken. grm. Volume of liquid. cm. 3 H 2 S0 4 (i : i). cm.* Current. Time, min. Theory for copper. grm. Copper found. grm. Error, grm. Amp. Volt. o . 5030 * 50 5 i; 5 7 5} o. 1281 0.1278! 0.0003 0.5008 50 5 i: 5 7 4! 0.1275 0.1275} 0.0000 0.5024 50 5 i: 5 7 40 1 0.1280 0.1277! -0.0003 0.5014 50 5 i: 5 7 21 0.1277 o. 1276* O.OOOI 0.5018 50 S t: 5 7 40 } 0.1278 o 1278* o.oooo No copper in filtrate or in washings. t Trace of copper in filtrate. J Trace of copper in washings. APPLIANCES AND GENERAL PROCEDURE IJ water until the current ceased to flow because there was no electrolyte to carry it. The apparatus was washed with water and finally with alcohol, and the crucible and contents were dried for periods of ten minutes at ioo-no, to constant weight. I. When a deposit is so loosely adherent as to be Electrolysis ,. , , , with continuous moved by the liquid, it may be compacted upon the Filtration. filtering felt by keeping the liquid in process of filtra- tion and constant motion through the cell to the receiver. The adjustment of apparatus for this purpose is shown in Fig. 1 6. Fig. 16. Here the electrolytic cell rests in the crucible holder fitted to a separating funnel used as a receiver and connected into the vac- uum pump. A stopcock in the tube of the crucible holder is convenient but not necessary. The manner of using the apparatus is simple. First, the weighed crucible, fitted in the usual manner with an asbestos felt and containing the platinum clippings, is adjusted to the glass chamber. The cell is pressed into the platinum triangle and set into the holder. The funnel which carries the wire i8 METHODS IN CHEMICAL ANALYSIS electrode is put in place. The cell is charged with the electrolyte and the current is turned on. The electrolysis begins and, under regulated action of the vacuum pump, the liquid is drawn through to the receiver at a convenient rate. Usually, before the upper electrode is uncovered the stopcock is closed, the suction pump disconnected, and the liquid drawn off from the receiver and returned to the electrolytic cell. The pump is again connected, the stopcock is opened and nitration begins again. Should the deposit be noticeably loose, it may be compacted by allowing the cell to drain completely under the action of the suction pump. The electrolyte is thus kept in circulation, and loose particles of the deposit are held upon the filtering layer. From time to time, the process of emptying the receiver and filling the cell is repeated. When the electrolysis is complete, as shown by proper testing of the filtrate, the liquid is drawn through the crucible and replaced by water from above until the current no longer flows. The electrodes are disconnected, the extension chamber easily slipped off, and the washing of the crucible and its contents continued sufficiently, with care, should the deposit be spongy, to give time enough in the washing to properly soak out absorbed material. The crucible and contents are dried, ignited, and weighed as usual. This method of manipu- lation was also put to the test in the electrolysis of copper sul- phate. Experimental details are given in the table. Electrolysis with Continuous Filtration. CuSO4.sH 2 O taken. grm. Volume of liquid. cm. 3 H 2 S0 4 (i : i). cm. s Current. Time, min. Theory for copper. grm. Copper found, grm. Error, grm. Amp. Volt. 0-5013 50 5 l: 5 7 41 0.1277 O.I28of +0.0003 0.5003 50 5 ): 5 7 41 0.1274 0.1276! +O.OOO2 0.5015 50 5 i: 5 7 4f 0.1277 0.1279* +O.OOO2 0.5001 50 5 i: 5 7 si 20) 0.1274 0.1274* O.OOOO 0.5041 50 5 i: 5 7 41 o. 1284 0.1285! + O.OOOI * No copper in filtrate or washings. t Trace of copper in filtrate. J Trace of copper in washings. The results show that there is no difficulty in getting accur- ate results while maintaining continuous filtration during the APPLIANCES AND GENERAL PROCEDURE Fig. 17. process, and that the time needed to complete the action is somewhat shortened when the liquid is kept in circulation by filtering. II. Another form of apparatus for electrolysis with continuous nitration, in which a porcelain filtering crucible replaces the platinum filter cru- cible, is shown in Fig. 17. In this appara- tus, it is necessary to make the connection from above with the electrode inside the crucible, and this is accomplished by a linked platinum wire, as shown. In put- ting together and using this apparatus, a finely perforated disk of platinum foil (c) is laid upon the more coarsely perforated bottom of the porcelain crucible (A). Upon this disk the asbestos felt (a) is de- posited in the usual manner. Platinum clippings (b) form a layer of suitable thick- ness above the asbestos, and upon this layer, and in contact with it, is placed another perforated disk of platinum foil to which is attached a twisted wire (e) so linked that it may be folded within the crucible. This apparatus is ignited and weighed, and to it is adjusted, as shown, a cham- ber to hold the electrolyte. The other electrode (/), inclosed within a funnel (D) made from a thistle tube, is introduced in the manner indicated. This apparatus is adapted only to use in the method of continuous filtration, and it is used exactly as in Process I. Experimental details are given in the table. By either of the processes described, reasonably rapid and accurate electrolytic determinations may be made without the use of rotating motors or special stirring apparatus, and without large and expensive apparatus of platinum. The use of the fil- tering crucible as a part of the electrolytic cell makes possible the utilization of operations and conditions in which the deposit may lack the degree of adhesiveness necessary in ordinary electro- lytic processes. The application of the processes to the more difficult determinations of manganese and lead as the dioxides formed upon the anode in very imperfectly adherent condition will be described later. 20 METHODS IN CHEMICAL ANALYSIS Electrolysis with Continuous Filtration: the Use of the Porcelain Crucible. CuSO 4 .sH 2 taken. grm. Volume of liquid. cm. 3 H 2 S0 4 (i :i). cm. 3 Current. Time. min. Theory for copper. grm. Copper found. grm. Error. grm. Amp. Volt. ( 2 6 5) 0.5025 50 5 <3 8 15 o. 1280 O.I277f 0.0003 (4 10 10 ) I ( 2 6 5) 0.5009 50 5 3 8 15 o. 1276 0.1279* +o . 0003 1 U 10 15) 0.5025 50 6 i: 6 10 4! o. 1280 0.1278* O.OO02 0.5011 50 5 i: 6 10 41 o. 1276 o.i27 3 f 0.0003 0.5013 50 5 k 6 10 41 0.1277 o. 1276$ o.oooi * No copper in filtrate or in washings. f Trace of copper in filtrate. I Trace of copper in washings. The Fixation of Chlorine on the Silver Anode. Hydrochloric From a consideration of the apparently very exact Anode andPiati- resu ^ ts obtained by many investigators* in the de- num Cathode, termination of the chlorine in chlorides by fixation of that element upon an anode plated with silver, it would seem that nothing could be simpler than the accurate determination of the chlorine in hydrochloric acid by procedure advocated for the treatment of metallic chlorides. That such is not the case, however, has been shown by Gooch and Read.f In an experi- mental study of the electrolysis of hydrochloric acid with use of the silver anode and platinum cathode, it is shown that silver oxide is formed at the anode and must be decomposed by heating to a high temperature (incipient redness, at the tip of the Bunsen flame), not simply dried over a steam radiator |; that silver de- posited from the double cyanide solution upon platinum gauze to make the silver anode always includes more or less alkali salt, which is lost from the anode surface attacked by chlorine during electrolysis; that to avoid contamination of the silver anode by nonvolatile material it should be plated from a solution of silver * Smith, Jour. Chem. Soc., xxv, 890. Myers, ibid., xxvi, 1124. Withrow, ibid., xxviii, 1350. Hildebrand, ibid., xxix, 447. McCutcheon, ibid., xxix, 1445. Lukens and Smith, ibid., xxix, 1455. Lukens and McCutcheon, ibid., xxix, 1460. t F. A. Gooch and H. L. Read, Am. Jour. Sci., [4], xxviii, 544. t Smith's Electro-analysis (1907), page 305. APPLIANCES AND GENERAL PROCEDURE 21 oxalate in ammonium hydroxide; and that the silver anode is attacked and dissolved by oxygen-chlorine acids produced chiefly toward the end of the electrolysis. The Electrolysis of Hydrochloric Acid with a Silver Anode Plated in the Oxalate Solution. Increase Apparent Apparent Chlorine taken in HC1. since they both stand in the order named as the first two dextrins produced from starch, as shown by saliva digestion of starch, as also by malt-extract digestion of starch. Starch both in solid state and in solution tends to pass through these stages of hy- drolysis. Germ growth rapidly appears in solutions of pure amidulin and pure erythrodextrin, with the destruction of these bodies to form dextrins lower in the series. Pure starch causes no red color, nor loss of iodine, in alkaline titration of arsenite solution or of tartar emetic. If any purplish tinge occasionally occurs, it is no hindrance to the reading and causes no appreciable loss of iodine. With an impure starch, the reading from the first permanent color, whether red or blue, is nearest to the correct value. The readings may be compared with plain iodine readings and a cor- rection applied, since the loss for a constant quantity of starch proves to be constant in the titration of 20 cm. 3 to 50 cm. 3 of arsenite solution. Titration should be made at considerable dilution e.g., 150 cm. 3 to 200 cm. 3 (since the production of red is at a minimum and the loss of iodine small at high dilutions) and in presence of a suitable amount of potassium iodide to ren- der the reaction delicate. Whenever it is practicable, however, the best method of using an impure starch is to make the titra- tion without it up to the appearance of the yellow tinge of free iodine and then to add the starch. In this way the color comes out a clear blue and the exact adjustment may be easily made by alternate additions of a drop or two of n/io arsenite and w/io iodine. APPLIANCES AND GENERAL PROCEDURE 33 Hale * experimented with various preparations of Ordinary Prepa- starch. The starch solution was made by grinding 5 grm. of starch paste with a few cubic centimeters of cold water with the addition of o.oi grm. of mercuric iodide, pouring into a liter of boiling water, and boiling five to ten min- utes. Only the clear supernatant liquid was used.f Preparation i n 25 cm. 3 of cold water 5 grm. of pure starch with Potassium . , .. ..... , iodide. were ground with 2 grm. of potassium iodide and the mixture was poured into 75 cm. 3 of boiling water and boiled, the beaker being protected with asbestos. The mass became mucilaginous. After fifteen minutes the volume was in- creased to 500 cm. 3 , and the boiling was continued for forty-five minutes. The solution was filtered, forty-eight hours being re- quired, leaving a residue of jellylike consistency upon the filter. This method was suggested by the extreme delicacy which the presence of potassium iodide gives to the starch reaction, and by certain statements made in the literature, one that concentrated potassium iodide causes starch to swell and dissolve,! another that a solution of starch made by a somewhat similar method would keep a year without fermentation. Heating with In 70 cm. 3 of pure glycerin 5 grm. of potato starch ForT^o- were heated at a temperature of i85-i9O C. for half dextrin. an hour with constant stirring. The starch dissolved and the solution turned through yellow to a deep red. The solu- tion was cooled to 120 C. and poured slowly and continuously into 200 cm. 3 of alcohol. The precipitate was thoroughly stirred, settled, and filtered while warm. It filtered readily and was washed with alcohol until the filtrate came through colorless. The colorless residue of amorphous amylodextrin was then dis- solved in 500 cm. 3 of water heated to 6o-7O C. Soluble starch * n a ^ tt: ^ e co ^ water 2 g rm - f starch and 0.5 grm. by saliva Diges- of acid potassium carbonate were ground together tion (Amiduiin). an mixed with 200 cm. 3 of water kept boiling for a few minutes. The mixture was cooled to 4O-45 and treated with 10 cm. 3 of filtered saliva neutralized byo.i percent hydro- chloric acid with the use of a slip of litmus paper (first dipped in * Loc. cit. t Gastine's formula, Zeit. anal. Chem., xxviii, 339. J Payen, Compt. rend., Ixi, 512. Zeit. anal. Chem., xxv, 37. 34 METHODS IN CHEMICAL ANALYSIS acetic acid and then washed) as indicator. When, in the course of three or four minutes, the solution had become entirely clear, it was boiled for ten minutes. In this process the addition of the alkali hinders the action from going beyond the first step of digestion: the boiling at the end destroys any further action of the saliva. Starch cellulose, which is said to produce a feeble red or brownish color with iodine,* is digested and destroyed by the saliva. In the presence of a suitable amount of potassium iodide each of the first three preparations gives a sharp indication with a single drop of n/io iodine; the amidulin is only a little less deli- cate. In the titration of 50 cm. 3 of n/io arsenite, enough potas- sium iodide is present in the n/io iodine to give a sharp reaction at a volume of 125 cm. 3 This is shown in the following state- ment. For comparison, a titration made with a large amount (25 cm. 3 ) of ordinary impure starch is included. Titrations with the Different Preparations of Starch. Volume i25Tcm. 3 . Starch solution. cm. W/IO Nearly ^o! 0i w/io I sol. Pure Impure Color. KI ordi- Amylo- Amid- ordi- starch. nary dextrin. ulin. nary starch. starch. cm. 1 cm. s 5 49.38 I Permanent purplish. 5 49.40 I ( Good blue. 5 49.40 I Good blue. 50 49.40 2 Slow-fading purplish. 5 49.42 . 2 Good blue. 5 . . 1-5 Permanent purplish. So 49.40 1-5 Deep blue. So 49.42 5 Permanent purplish. So 49-44 . 5 Purple. 50 49-55 25 Abundant red. ! In a series of parallel titrations, designated A and B in the next table, readings were made with the blue developed by the Kl-starch indicator and with plain iodine, alternately, in order to eliminate accidental errors. Starch was added subsequently to corroborate the very pale iodine reading. The corrected readings were found by subtracting one drop from the actual * Beilstein (First Edition), i, 1082, line 17. APPLIANCES AND GENERAL PROCEDURE 35 reading. To render the plain iodine readings sharp a crystal of potassium iodide was added in the first two titrations. Titrations -with and without Starch. Volume 125 cm. 3 Nearly M/IO I. Direct readings. Nearly w/io I. Corrected readings. M/IO I sol. Absolute M/IO I SOl. Absolute errors in A, . . in B. H/IO As 2 3 . By iodine color, A. ByKI- starch blue, B. A. B. amount calculated from 50 cm. 3 A. B. cm. 8 cm. 3 cm. s cm. 3 cm. 3 cm. 3 cm. 8 cm. 3 i drop* i drop 5 4-94* 4.96 4.92 4-94 4-94 O.O2 o.oo 10 9.88 9.90 9.86 9.88 9.88 O.O2 o.oo 15 14-83 14-83 14.81 14.81 14.82 O.OI O.OI 20 19.78 19.78 19.76 19.76 19.76 0.00 o.oo 30 29.64 29.64 29.62 29.62 29-63 O.OI O.OI 40 39-55 39-55 39-53 39-53 39-51 + 0.02 +O.O2 So 49.41 49.41 49-39 49-39 49-39 o.oo 0.00 * A crystal of KI was added. It is to be noted that the readings with plain iodine and with pure starch agree exactly except for the first two titrations, and here there is only a difference of a drop. The absolute errors are interesting, as they show how the absolute values fluctuate about a standard set by the 5O-cm. 3 readings. This fluctuation is lim- ited to a drop plus or minus. The statement has been made that starch from different sources has a varying power of absorbing iodine, e.g., that potato starch absorbs three times as much as rice starch.* To discover whether this fact has any bearing upon the use of the starch in- dicator, and at the same time to learn whether pure starch solu- tions made in the ordinary way (Gastine)f would give as delicate readings as that boiled with potassium iodide, solutions were made from pure potato starch, pure rice starch, pure arrowroot starch and a pure (so-called) soluble starch of unknown origin. The results of titration with these solutions as indicators are shown in the following statement. It is at once seen that these values are coincident within a drop, and that all the starch solutions are within the limits set by a plain iodine reading on the one hand and a potassium iodide * Girard, Ann. Chim., [6], xiii, 275. t Loc. cit. METHODS IN CHEMICAL ANALYSIS starch reading on the other, though there was no dilution of the standard solutions. Starch from Various Sources. No extra dilution. Volume about no cm. 8 W/IO As 2 0,. cm.* Nearly /io I sol. cm. 8 KHCO 3 satu- rated cm. 8 Starch solution. cm. 8 Color. 50 50 5 SO 50 CQ {49." I I 49 13 ) 49-13 49.14 49-15 49-15 40 1 3 5 5 5 5 5 e i (Ordinary pure potato.) < i (Ordinary pure rice.) i (Ordinary pure soluble.) i (Ordinary pure arrow- root.) i (KI pure starch.) Pale blue. Deep blue. Blue, slightly purplish. Blue, slightly purplish. Deep blue. Good blue. Yellow If the starch is pure the amount of it used, within reasonable limits, is without effect upon the titration of n/io arsenite by n/io iodine, and the same may be said of the amount of acid carbonate, as shown in the following table. Variation in the Amount of Starch. M/IO AsjOa. cm. J Nearly / 10 1 sol. cm. 8 KHCO 3 saturated. cm. 8 Starch solution.* cm. 8 Color. 10 9.82 5 i-5 Deep blue, purplish. IO 9.82 5 5 Deep blue. 10 jp.Bal 19.24) 5 1 Pale blue, purplish. Deep blue. 10 9.82 5 15 Deep blue. IO 9.84 5 20 Deep blue. 10 9-85 5 25 Deep blue. 10 9.82 5 5 Deep blue. 10 9.82 10 5 Deep blue. 10 9-83 15 -5 Deep blue. IO 9.82 20 5 Deep blue, purplish. 10 9.81 25 5 Deep blue. * Ordinary preparation; pure potato starch. With pure starch, the end reaction of the titration of tartar emetic by iodine is likewise a pure blue. Following are Hale's results. The average of the 10 cm. 3 readings (absolute) multiplied by five equals 47.73. The absolute 5O-cm. 3 reading (47.75-0.02) equals 47.73. Evidently even tartar emetic causes no loss on pure starch, for the 5O-cm. 3 reading agrees with the plain iodine reading for the same amount and with the io-cm. 3 titrations. APPLIANCES AND GENERAL PROCEDURE 37 End Reaction with Tartar Emetic. Volume. /io tartar emetic. Nearly w/io I sol. Starch solution, KHCO 3 . Color. pure potato. cm. 3 cm. 1 cm.* cm. 8 cm. 100 10 9.58 i-5 10 Blue, no red. IOO 10 9.58 i-5 10 Blue. 75 IO 9-S 6 IO Yellow. 125 50 47-75 i-5 25 Blue (purplish tinge). 125 50 47-75 25 Yellow. Hale* emphasizes the need of potassium iodide in suitable amount to bring out the delicacy of the blue end reaction, and the further necessity of not exceeding a suitable proportion, in order that the starch blue compound may not be modified by transformation to starch red. The proportions of potassium iodide and iodine entering into starch blue and starch red were carefully studied. According to experimental results, it would appear that the group KI.I 4 is characteristic of starch blue and the group KI.I 2 of starch red ; that in the presence of a sufficiently concentrated solution of potassium iodide the group KI.I 4 changes by addition of KI to 2KI.I 2 , while dilution with water tends to bring about the reverse action. Wholly apart from the consideration of theory, the influence of iodides upon the delicacy of the starch blue test for iodine is a matter of considerable practical importance analytically. Ex- perience shows that 0.3 grm. of potassium iodide is sufficient in volumes not exceeding 300 cm. 3 to make readings by the starch indicator sharp, and that under these conditions indications are quite as delicate at ordinary room temperatures as at the tem- perature of ice water. The addition of more potassium iodide renders the reading no sharper within that range of dilution, though the total volume has considerable proportionate influence upon the amount of iodine needed to bring out the indication. In the absence of potassium iodide in sufficient amount, the in- fluence of temperature is very noticeable. Hydrochloric acid either in small amount or large amount does not render the read- ing sharp in absence of potassium iodide. These results are in agreement with the work of Lonnes.f * Am. Chem. Jour., xxviii, 450. t Zeit. anal. Chem., 33, 409- METHODS IN CHEMICAL ANALYSIS Effect of Temperature when Potassium Iodide is Restricted. Volume. Potassium iodide not exceeding Hydro- chloride acid. (Sp.gr. 1. 12.) Pure potato starch solution. Temper- ature. M/IOOO iodine. Color. 0.3 grm. cm.* cm. 3 cm.* cm.* 100 i crystal 2 23 0-45 Faint blue. 200 i crystal 2 23 0.65 Faint blue. 300 i crystal 2 23 0.80 Faint blue. 400 300 i crystal i crystal 2 2 23 23 0.80 Faint blue. Faint blue. 300 i crystal 2 5 0.70 Faint blue. 300 2 23 9. 20 Faint blue. 300 2 5 I 'CO Faint blue 300 I 2 23 A ie Faint blue 300 IO 2 23 4OO Faint blue The influence of different amounts of potassium iodide in bringing out the starch iodide indication of iodine set free by reaction with gold chloride is of interest in this connection. At volumes lying between the limits of 25 cm. 3 and 50 cm. 3 o.i grm. of potassium iodide is an appropriate amount; at a volume of 15 cm. 3 , o.oi grm. to 0.05 grm. of the iodide will do the work; and at lower dilutions even less of the iodide is effective.* Variation of Standard. Standard Tartar Emetic. Gruenerf has shov/n that tartar emetic solutions containing about 1 6 grm. of tartar emetic, 20 grm. to 30 grm. of tartaric acid and I cm. 3 of concentrated hydro- chloric acid to the liter, will keep from five to twelve months without any change in strength. There is no deposit of anti- monious oxide under these conditions, no oxidation and no signs of fungous growth. Gruener determined the strength of his tartar emetic solutions by titration with a decinormal iodine solution, standardized by decinormal arsenite. The mean of twenty-nine determinations showed 43.95 per cent of antimo- nious oxide in tartar emetic. Theory required 43.37 per cent (Sb = i20, KSbOC4H 4 O 6 4H2O = 332). The cause of this dis- crepancy between arsenite and tartar emetic solutions made up as standards according to the accepted molecular formulas is a matter of considerable interest. One suggested explanation of * See page 146. t Hippolyte Gruener, Am. Jour. Sci., [3], xlvi, 206. APPLIANCES AND GENERAL PROCEDURE 39 this difference is that the end reaction between starch and iodine is delayed until an excess of iodine is present. Hale* has shown, however, that a pure starch solution gives a sharp end reaction with both tartar emetic and arsenite solutions, and that while with impure starch there is a loss of iodine accompanied by the production of reddish hues in titrating tartar emetic, as shown by the difference between the readings made in the presence of potassium iodide by the yellow color of iodine and by the blue of starch iodide, yet it is no greater than in titrating arsenite solution in the presence of an impure starch. If only an impure starch is available, the reading should be made without starch, for the presence of potassium iodide renders very sharp the yel- low color of the first excess of free iodine. This first reading may be afterwards corroborated by adding the starch solution, which will then give only a pure blue color. The above dis- crepancy must then be due to some other cause than delay of formation of the starch iodide. Halef has given proof that this discrepancy is due to the ease with which tartar emetic loses its water of crystallization, and that in order to get a salt of the exact composition, KSbOC 4 H 4 O6.^H 2 O (mol. wt. 332.15), certain conditions must be very closely observed. Molecular weights of tartar emetic variously prepared, calcu- lated from the results of iodometric titration, run in a series from that of crystalline tartar emetic almost to that the salt which has lost i .5 molecules of water, passing through all intermediate stages, but never surely resting at any one spot. Two important stages are reached : when all the water of crystallization is gone, the anhydrous state ; and when 0.5 molecule of water further has been lost, the first anhydride stage. The variation of molecular weight is shown in the following table. The condition most easily and definitely reached is that of the hydrous crystalline salt. The greatest error met with in the recrystallized salt, if air-dried, is about +0.2 per cent, calcu- lated on the ratio of antimony to tartar emetic, and that after standing in fine condition in a closed bottle for several weeks. Drying in the air bath does not yield a product, even with the most finely divided preparation, that precipitated by alcohol, which is sufficiently uniform to serve the purpose of a standard. * F. E. Hale, Am. Jour. Sci., [4], xiii, 379. t F. E. Hale, Jour. Am. Chem. Soc., xxiv, 828. METHODS IN CHEMICAL ANALYSIS Hale recommends, therefore, the preparation of tartar emetic, in medium-sized crystals (fa to J inch in diameter) , rather than in the minutely crystallized condition. Variations in Molecular Weight. 49.29 cm. 3 (absolute amount) of w/io iodine solution = 50 cm. 3 of w/io ar- senic trioxide solution. Preparation. n/io tartar emetic solu- tion. cm. W/IO iodine solu- tion. cm. 3 Molec- ular weight. Anti- mony. per cent. Remarks. Freshly crystallized CO 49. 29 6 i ' O 49.56 330.34 Crystals kept several weeks .... ( 5o 49.58 330.21 36.35 Tartar emetic. grm. Medium-sized crystals, air- dried . . 0.5 20 82 "2 3O - S Crystals, 4 days over ^SCV. . . . 0-5 ^v u 29.95 329.13 Crystals 7 days over H2SO4. o 5 20 08 328 77 Crystals, 16 days over H2SO4. o 5 ~y * y JQ 16 o / / 327 l< Crystals, 4 hours at 95-i3o. . . Crystals, 7? hours at io4-i3o.. 0-5 0-5 o w * 30.42 30.51 O t 3 324.50 323.I5 37-14 (Anhydrous.) Crystals at I04-i3o j -5 31.27 3I5.29 ( ... 314. 15 (H 2 O anhy- dride.) Crystals. at i6o-i65 0.5 31.88 309 . 28 Crystals, 2 hours at i6o-i65. . 3L-93 308.80 305-I5 (H 2 O anhy- dride all hydroxyls gone.) Preparation. Enough tartar emetic is dissolved in about 300 cm. 3 of boiling water to make a concentrated but not a saturated solution. This hot solution is filtered into flat crystal- lizing dishes and allowed to crystallize over night. The crystal- lization should not be too rapid. The crystals are filtered off by suction, washed twice with distilled water, kept under suction for about five or ten minutes more, and then air-dried from one and a half to four hours at room temperature, not above 25 and preferably lower, in a clear, dry air. The crystals may be con- sidered dry an hour or two after they cease to show the slightest tendency to cling to a glass rod used as a stirrer. Tartar emetic may be prepared in this manner, which, with good starch, shows practically the theoretical value when tested APPLIANCES AND GENERAL PROCEDURE against a standard arsenite solution by titration of both with iodine. As has been pointed out, if pure starch is not available the first reading should be made without starch in presence of potassium iodide, and this corroborated after adding the starch solution, which will then give only a pure blue. A comparison of several preparations of tartar emetic with standard arsenite is given in the table. Comparison of Tartar Emetic with Standard Arsenite. .2 I ^ ll n Crystals, air- i*3 ' o jj o> c5 a a J2 "S w/io solution. dried. la Rl O 8-3 d 'S.2.2 .2 a) 1 Color. SP *J^ Jl u ^S K 12 5* H OT (2* 55 hours, temp. cm. 3 cm. 3 cm. 3 cm. 3 cm. 3 Tartar emetic 1 ii I9 -2 4 5Q 49-32 2S Medium blue. Tartar emetic I 4 i9-24 5 49.42 2 5 Medium blue. Tartar emetic II.. . Arsenite i9-24 5 49-43 49 43 25 5" Medium blue. Medium blue. Tartar emetic III. . Tartar emetic IV . . 3 4 i9-24 5o 49.58 49.58 25 2 5 Deep blue. Medium blue. Arsenite . ... 50 49.61 5 Medium blue. PROCESSES OF OXIDATION. Arsenic Trioxide as a Standard. Carefully sublimed and anhydrous arsenic trioxide serves admi- rably for the standardization of potassium permanganate for use in direct oxidations, as well as in the processes of iodometric analysis. It is generally best to work with standard arsenite solution prepared as previously described,* and it is convenient, though not necessary, to have at hand a solution of iodine, f also standardized against the arsenite solution. standardization The standardization of the permanganate solu- without iodine, tion, made up to an approximate value, is easily accomplished by running a suitable amount of it into a solution of potassium iodide contained in a reaction bottle { and acidi- fied with dilute sulphuric acid, neutralizing with acid potassium * See page 29. f Ibid. I See page 6. 42 METHODS IN CHEMICAL ANALYSIS carbonate, and titrating by the standard arsenite the iodine set free by the permanganate. In this operation it is best to dis- pense with the starch indicator usually employed to fix the end reaction. The vanishing point of the color of free iodine is itself sufficiently definite, even at a dilution of 300 cm. 3 , and the dis- appearance of color is much sharper than that of the blue starch iodide.* standardization ^ a solution of iodine standardized against the with the Aid of arsenite solution is at hand, the process just de- scribed may be modified so that there is no danger of overrunning the end-point in a single titration. In this pro- cedure, an excess of the standard arsenite, taken in known amount, is added at once after the reaction of the permangan- ate upon the iodide in presence of acid, and is followed by the acid carbonate. The excess of arsenite is determined by titration with iodine in presence of starch. Should too much iodine be added in the titration, it is, of course, only necessary to add another measured amount of arsenite and then to repeat the titration by iodine. f The Gravimetric Standardization of Permanganate. The standardization of a permanganate solution may be made gravimetrically by adding a suitable amount of the solution to an excess of potassium iodide acidified with hydrochloric acid, shaking the mixture with a weighed amount of specially prepared silver in an atmosphere of hydrogen, collecting upon asbestos in the perforated crucible the residue of silver and silver iodide, drying, and weighing, according to the procedure previously described.}: The Loss of Oxygen in Oxidations by Potassium Permanganate. Concentration A statement made many years ago, that in the of Acid. interaction of oxalic acid and potassium perman- ganate free oxygen is always a product, met with adverse criti- cism ; but subsequently a similar effect was noted by Brauner 11 * Gooch and Peters, Am. Jour. Sci., [4], viii, 125. f Gooch and Gilbert, Am. Jour. Sci., [4], xv, 390. j See page 27. Francis Jones, Jour. Chem. Soc., 1878, 95. || Jour. Chem. Soc. (1891), 238. APPLIANCES AND GENERAL PROCEDURE 43 in the action of permanganate upon tellurous acid dissolved in sulphuric acid, with the additional observation that the evolu- tion of oxygen is proportional to the amount of sulphuric acid employed, and that in alkaline solution little evidence of such an effect appears. Recognizing that the production of permanganic acid, free oxygen and ozone, by the action of strong sulphuric acid upon permanganate in absence of oxidizable material, is a common phenomenon, and that the formation of a precipitate consisting largely of hydrated manganese dioxide by the action of hot dilute sulphuric acid upon the permanganate in aqueous- solution is likewise well known, Gooch and Danner* have in- vestigated the action of sulphuric acid in different concentrations upon permanganate in solution with a view to determining how far such action may be directly or indirectly responsible for the liberation of free oxygen in processes of oxidation. In certain experiments tubes of suitable size and length, hold- ing from 100 cm. 3 to 200 cm. 3 , were sealed at one end, filled com- pletely with the mixtures of acid and permanganate, inverted > and allowed to stand with the lower and open end submerged in liquid of the exact composition of that which filled them. The details and results of these experiments are recorded in the state- ment below. Time elapsed. Gas from 100 cm. J Appearance. Time elapsed. Gas from ico cm. 3 Appearance. A. B. H 2 S04 [i : i]=5o per cent. H 2 SO4 [i : i]=25 per cent. S rain. o.i cm. 1 No change. S min. Small bubble. No change. """ I hour. i.i cm.* No change. i day. 14 cm. 1 Red brown. , ' 3 days. 4 days. 15. 3 cm.* 15. 6 cm.* Light brown. Light brown. 3 days. 9.6 cm. 8 | Reddish purple. Turbid. 7 days. Brown, turbid. ( Reddish pink. 8 days. 16 cm.* (Clearing by precipi- \ tation. 7 days. 15 days. 15. i cm.* < 18 cm.* Clearing by pre- cipitation. Nearly clear. 15 days. 17. 3 cm.* Clear, straw-colored. 17 days. 35 days. 17.4 cm.* 17.5 cm.* Clear, straw-colored. Clear, straw : colored. 35 days. 18.4 cm. 3 | Clear and color- less. C. D. H 2 SO 4 [i : i] = i2.s per cent. H 2 SO4 (i : 1 1=6.25 per cent. i hour. Small bubble. No change. i hour. Small bubble. No change. i day. 3 days. 14 days. 37 days. 44 days. Bubble. Bubble larger. 7.1 cm. 1 ii cm.* 12 cm.* No change. No change. Color lighter. Color lighter. Color lighter. i day. 3 days. 14 days. 37 days. 44 days. Bubble. Bubble larger, i cm. 3 3 cm. 3 5 cm.* No change. No change. No change. No change. Little change. * F. A. Gooch and E. W. Danner, Am. Jour. Sci., [3], xliv, 301. 44 METHODS IN CHEMICAL ANALYSIS In other experiments note was made of changes in color and formation of precipitates in ioo-cm. 3 portions of liquid containing 10 cm. 3 of decinormal permanganate and varying proportions of acid during five days' standing, and the degree of decomposition of the permanganate was finally determined by adding a small excess of oxalic acid to the mixtures contained in Erlenmeyer flasks, heating to about 80 C., and titrating with permanganate the residual oxalic acid. Percentage H 2 SO?[i:i]. Time elapsed. Percentage of KMnO, decomposed. i day. 2 days. 3 days. 4 days. 5 days. Color ( Color unchanged. 1 Color Color Color \ unchanged. Slight * 1 unchanged. unchanged. unchanged. Slight sediment. j. 3.0 (^sediment. Slight scum. *> { Color unchanged. Color unchanged. Color unchanged. ! Color unchanged. Slight Color unchanged. Slight 7.4 sediment. sediment. 30 { Color unchanged. Color unchanged. Color unchanged. Reddish tinge. Slight Reddish tinge. - 6.9 sediment. 4 o | Color unchanged. Color unchanged. f Tinged ( with red- l^dish brown 1 Reddish j brown. Reddish brown. - 39-2 ( Color Color Reddish Reddish Red 50 j unchanged. unchanged. i brown. brown. brown. ^ 57-4 , Color Color Reddish Sherry Reddish ) o j redder. redder. brown. brown. olive. | 58.9 ( Color Color Sherry Reddish Reddish AY -r 70 ( redder. . redder. brown. olive. olive. 01. 1 In the first five experiments little change of tint was noted upon the addition of the oxalic acid to the cold solution, but in the last two experiments the reddish-olive color became at once distinctly red presumably because the higher sulphate of man- ganese was attacked in the cold by the oxalic acid (as Brauner has shown) , and so the natural color of the permanganate was permitted to assert itself. The extreme decomposition that which took place in the last experiment, in which 70 per cent of the [i : i] acid was present corresponds nearly to the reduc- tion of the entire amount of permanganate present to the condition of oxidation of MnO2 which is known to exist in com- bination with sulphuric acid in the form of a higher manganic sulphate. It is to be noted that the separation of the insoluble higher oxide took place only when the percentage of acid was low. APPLIANCES AND GENERAL PROCEDURE 45 In still another series of experiments the solution of perman- ganate was mixed with sulphuric acid previously diluted with an equal volume of water, and cooled; and after the lapse of time indicated oxalic acid was added in quantity a little more than sufficient to bleach the permanganate. The solution was warmed to about 80 C., and the residual oxalic acid titrated by gradual addition of more permanganate. The difference between the amount of permanganate needed under the conditions to destroy the known amount of oxalic acid, and that used in the deter- mination of the standard, should measure the oxygen lost and the permanganate decomposed under the action of the sulphuric acid. The results and details of these experiments are given below. Residual Permanganate Reduced by Oxalic Acid at 80. H 2 S0 4 [i : ij. cm. 3 Water, cm. 3 KMnO 4 in deci- normal solution. cm. 3 Percentage of H,S0 4 li:i). in solution during action. Percentage of KMnO 4 decomposed.. A. Treated immediately. 2' 8 IO IO o 4 6 IO 20 o 6 4 IO 30 5 8 2 10 40 1.6 10 IO 50 i. 9 B. Treated after standing eight hours at ordinary temperature. 2 8 IO 10 -0.3 4 6 IO 20 0.3 6 4 10 30 !-3 8 2 10 40 5-3 IO 10 50 15 7 C. Treated after standing five days at ordinary temperature. 2 8 10 IO 4 o 4 6 10 20 21 6 6 4 1C 30 49 7 8 2 IO 40 55 9 IO IO 50 564 D. Treated after standing one and one-half hours at 8o-9o C. 2 8 10 IO i 3 4 6 IO 20 43-8 6 4 IO 3 35 9 8 2 IO 40 49.1 10 10 50 55-3 4 6 METHODS IN CHEMICAL ANALYSIS It is obvious that the decomposition of the permanganate in- creases directly in each series of experiments with the increase in the proportion of sulphuric acid, that the amount of decom- position is greater as the time of action is extended, and that increase of temperature heightens the change. It is noted in particular that the presence of ten per cent of [i : i] sulphuric acid induces at the ordinary temperature no immediate decom- position of the permanganate, none in eight hours, and a break- ing down amounting to four per cent in five days; and that the presence of fifty per cent of acid of the same strength occasions the decomposition of about two per cent at once, fifteen per cent in eight hours, and more than half the entire amount of perman- ganate in the course of five days. It is evident, also, that twenty per cent of the [i : i] acid produces no appreciable effect at ordinary temperatures and under exposures of a few hours only. The effect of heating the mixture of acid and permanganate to So C. for an hour and a half is closely comparable with that brought about by the five days' action at the ordinary tempera- ture. It is, of course, probable that some decomposition of the permanganate by the sulphuric acid is brought about after the addition of the oxalic acid during the warming of the mixture up to the temperature at which the oxalic acid and permanganate interact. Residual Permanganate Reduced by Ferrous Sulphate. H 2 S0 4 [i : I), cm.* Water. cm. KMnO 4 in deci- normal solution. cm.* Percentage of H 2 SO<|i:il in solution during action. Percentage of KMn0 4 decomposed. A. Treated at once: Volume, 20 cm. 3 2 8 10 IO o.o 4 6 IO 20 O.2 6 4 IO 3 O.I 8 2 IO 40 O. I 10 10 50 0-3 B. Treated at once: Volume, 100 cm. 3 IO 80 10 10 O. I 20 70 10 20 O.I 30 60 IO 30 o.o 40 SO IO 40 0-5 50 60 40 30 10 IO i-3 3-o 70 20 IO 70 5-o 80 IO IO 80 3-3 90 10 90 8.1 APPLIANCES AND GENERAL PROCEDURE 47 Experiments in which ferrous sulphate is used at the ordinary temperature to effect the reduction of the residual permanganate, instead of oxalic acid at the higher temperature, show a lower de- gree of decomposition, as is natural. The increase in the amount of decomposition as the proportions of sulphuric acid [i : i] are advanced beyond 50 per cent by volume is striking. The results of these experiments appear in the accompanying table. It appears, therefore, that when potassium permanganate and sulphuric acid are brought into solution together under these con- ditions, there is likelihood of a reduction of the former, which is greater as the strength of the acid is increased, as the tempera- ture is raised, and as the duration of action is extended. It ap- pears further, at least when the acid is not present in proportion greater than 50 per cent of the [i : i] mixture, that in the early stages of the action the oxygen lost to the permanganate is liber- ated, and that later on the decomposition of the permanganate results in the precipitation of manganese as a higher oxide or in the formation of a higher sulphate. The first effect of the mutual action of the acid and the permanganate is to set free perman- ganic acid, which, being unstable, breaks up with the results described. The bearing of these observations and inferences upon the question of the action of potassium permanganate during oxida- tions carried on in the presence of sulphuric acid is obvious ; for, if the aqueous acid is able to liberate permanganic acid in such proportions as to be spontaneously unstable, it is reasonable to presume that any reducing substance present at the time of such action may, by virtue of its attractive action upon the oxygen of many more molecules of the permanganic acid than would be necessary to supply the exact amount needed for perfect oxi- dation, tend to increase the general instability of the already unstable molecules and so set up a far-reaching decomposition. These considerations throw light upon the phenomena observed by Brauner* in the oxidation of tellurous oxide in presence of sulphuric acid; and the fact that the liberation of free oxygen in this special ca.se is more noticeable than in the oxidation of ferrous salts or oxalic acid, for example, is explicable in the light of Brauner's observation that the attraction of tellurous oxide for oxygen is greatly inferior to that of these substances for * Jour. Chem. Soc., 1891, 238. 48 METHODS IN CHEMICAL ANALYSIS oxygen not sufficient, in fact, to break up so unstable a sub- stance as manganic sulphate, which is at once reduced by ferrous salts or oxalic acid. The practical lesson to be drawn from the investigation is the desirability of keeping the acid present in oxidations effected by the agency of permanganate at the lowest limit consistent with perfect oxidation; the time of digestion with an excess of per- manganate as small as may be; and the temperature, if possible, not above the ordinary room temperature. As to the correlative question of the liberation of oxygen dur- ing oxidations by potassium permanganate in alkaline solution, experience in the collection of the gas liberated in oxidations effected in presence of acid leads to distrust of the evidence of such experiments unless the amount of gas liberated is consider- able. While, on the one hand, small quantities of liberated gas may be so completely absorbed as not to appear free at all, it often happens, on the other hand, that the simple admixture of unlike liquids such, for example, as a solution of potassium per- manganate with sulphuric acid of strength insufficient to liberate oxygen may bring about a very appreciable liberation of dis- solved gases. So far as appears, however, the affirmation that oxygen is liberated in oxidations by potassium permanganate in alkaline solutions rests upon evidence of that nature only. Hydrochloric Lowenthal and Lenssen* were the first to show Acid with Fer- that the titration of a ferrous salt by potassium per- manganate in the presence of hydrochloric acid, according to the process of Margueritte,t is vitiated by the evolu- tion of chlorine outside the main reaction, and to point out that a remedy for the difficulty is to be found in the titration of the ferrous salt in divided portions, other equal volumes of the ferrous solution being added to the liquid in which the first titration is accomplished until the amount of iron indicated by successive titrations becomes constant. KesslerJ showed the restraining influence of certain sulphates, of manganous sulphate in par- ticular, upon the irregular and undesirable interaction of the permanganate and hydrochloric acid, and Zimmermann, in * Zeit. anal. Chem., i, 329. t Ann. Chim. Phys., [3], xviii, 244. f Ann. Phys. Chem., cxviii, 48; cxix, 225-226. Ann. Chem., ccxiii, 302. APPLIANCES AND GENERAL PROCEDURE 49 apparent ignorance of Kessler's forgotten proposal, advocated the introduction of a manganous salt, best the sulphate, into the ferrous salt to be determined, thus accomplishing the purpose of the empirical procedure of Lowenthal and Lenssen. The pro- tective influence of the manganous salt turns apparently, as Zimmermann suggested, upon the initiation of Guyard 's reaction, according to which the permanganate and manganous salt in- teract to form a higher oxide of manganese capable of oxidizing the ferrous salt, but slow to act upon the hydrochloric acid.* According to Volhard,f the reaction of Guyard is favored and hastened by heat and concentration of the solution, while it is delayed by acidity and dilution ; but even in solutions containing very little manganous salt and a considerable quantity of free acid, the faint rose color developed by the careful addition of permanganate ultimately vanishes until every trace of the man- ganous salt is precipitated. When a considerable amount of the salt is present interaction follows immediately the introduction of the permanganate. In titrations of a ferrous salt by permanganate, Zimmermann advocates the use of 4 grm. of manganous sulphate uniformly. In putting this matter to the test, Gooch and Peters { have found that as much as 5 grm. of manganous sulphate may be present in 135 cm. 3 of the liquid, containing about 5 cm. 3 of hydrochloric acid of full strength, without interfering with the regularity of the titration; and the effect is trivial even when the amount of manganous sulphate reaches 10 grm. In all cases, however, in which the larger amounts of manganous salt are present, the end reaction is marked by the advent of a brownish-red precipi- tate rather than the clear pink of the soluble permanganate ; and it is obvious that in case a substance to be oxidized were not active enough to act with rapidity upon the product of the Guyard reaction, difficulty might follow the failure to adjust the conditions more particularly. Regularity of action is also noted when manganous chloride is substituted for the sulphate, and in this respect the results accord with those of Zimmermann and differ from those of Wagner. * See also Manchot, Ann. Chem. (1902), cccxxv, 105. t Ann. Chem. (1879), cxcviii, 318. t F. A. Gooch and C. A. Peter?, Am. Jour. Sci., [4], vii, 461. Zeit. physikal. Chem., 28, 33. METHODS IN CHEMICAL ANALYSIS Total volume at beginning of titration. HC1. (Sp. gr. 1.09). FeCl 2 . KMnO 4 H/IO. MnSO 4 .sH 2 O. MnCl 2 .4H 2 O. cm.* cm. 8 cm. 1 cm. J grm. grrn. 135 IO 25 21 . 70 I 135 IO 25 21 .70 3 135 10 25 21.70 5 135 IO 25 21-75 7 135 IO 25 21-75 IO US 2O 25 21-75 10 175 5 25 21-75 IO 135 IO 25 11 . 70 I 135 IO 25 21 . 70 2 145 20 25 21 .70 2 155 30 25 21-75 3 165 40 25 2I.7O 4 Hydrochloric Acid with Oxalic Acid. It has been stated* that hydrochloric acid inter- feres in no way with the titration of oxalic acid by permanganate. Gooch and Peters find, however, that in such titrations there is a small though real waste of per- manganate proportionate to the amount of hydrochloric acid present. This fact is brought out clearly in the comparison of experimental results in the following table. Temperature at Beginning about 80 C. Approximate volume at beginning of titration. H 2 SO< [i : i]. HC1. (Sp. gr. 1.09.) Ammonium oxalate n/io. KMn0 4 . M/IO. Variation from mean of A taken as standard. cm. 1 cm. 1 cm. 8 cm. 8 cm. s cm.* A. 200 5 50 47-50 o.oo 2OO 5 .... 50 47-50 o.oo 200 10 .... 50 47-50 o.oo 2OO 10 .... 5 47-5 0.00 200 25 50 47-50 o.oo 200 25 5 47-50 o.oo B. 150 IO 2.5 25 23.80 +0.05 15 IO 2.5 25 23.90 +0.15 ; 150 10 5-o 25 23.90 +0.15 I 15 IO IO.O 25 24.00 +0.25 500 5 .... 25 23.80 +0.05 500 IO IO.O 25 24.00 +0.25 500 10 IO.O 25 24.10 +0.35 * Fleischer, Volumetric Analysis, Trans. Muir., p. 71 . Zimmermann, loc. cit. APPLIANCES AND GENERAL PROCEDURE Temperature 2o-26 C. Volume at begin- ning of titration. H 2 SO 4 [i : i]. cm.* HC1. (Sp. gr. 1.09) cm.* Ammo- nium oxalate, M/IO. cm.* KMnO 4 . W/IO. cm.* MnSO 4 .- 5 H 2 0. grm. MnCl,.- 4H 2 0, grm. Variation from standard. cm.* IO 23 oo OO4O -f-O. 15 I tO IO 2C 23 QO OI2O -f-O. 15 IO 2Z 23.80 +O.O5 130 IO 2Z 23 . 75 .O4OO +O.OO 130 IO 23 76 .O5OO +O.OI 130 IO 25 23. 70 . IOOO 0.05 IO 23 7$ 2OOO o oo I3O IO 2? 24 20 O2OO -f-o 45 I3O IO 25 2 3 95 .O2OO -|-o. 20 I3O IO 23.80 .0400 I3O 20 25 23 . 75 .0400 o.oo 130 3 25 23.75 .0400 o.oo I3O IO 25 23.75 I .OOOO o.oo IO 23 7 1 ? 2 OOOO o oo 130 IO 2C 23 7^ 3 oooo o oo I3O I 23 72 I OOOO o 03 I3O I 2C 23 . 74 2 .OOOO O.OI I 3 130 130 I 2 3 > 10 10 o d 4 = 20.09 cm. 3 KMnO4.) J 4 , 1 .>>*> e t? 3! : s fc 75 'g o a 10 HCI. 10 FeCl 3 . ij .! me of dige M| "d* -0 l"o 2 o> o -H 1 II o; 8 o a 35 ||| K ^ la M o o a W M ^ cm. 3 cm.* cm. 8 cm. 3 c. min. cm. 3 cm. 3 cm. 3 IOO 9.90 9.90 50 60 none none 9.90 18.88 8.69 IOO 9.90 9.90 5 60 none none 9.90 18.87 8.68 IOO 9.90 9.90 50 60 none none 9.90 18.80 8.61 IOO 9.90 9.90 50 60 none none 9.90 18.81 8.62 IOO 9.90 .... 9.90 50 60 none none 9.90 18.80 8.61 IOO 9.90 9.90 50 60 none none 9.90 18.82 8.63 IOO 9.90 9.90 5 30 none none 9.90 18.77 8.58 IOO 9.90 9.90 50 3 none doubtful 9.90 18.70 8-51 IOO 9.90 9.90 50 15 none very faint 9.90 18.70 8-51 IOO 9.90 9.90 50 15 none very faint 9.90 18.68 8-49 IOO 9.90 9.90 50 60 none none 9.90 18.87 8.68 IOO 9.90 9.90 50 60 none none 9.90 18.85 8.66 IOO 9.90 9.90 50 60 none none 9.90 18.81 8.62 IOO 9.90 9.90 50 60 none none 9.90 18.81 8.62 IOO 9.90 9.90 50 60 none none 9.90 18.87 8.68 IOO 9.90 9.90 50 60 none none 9.90 18.85 8.66 IOO 9.90 9.90 5 30 none none 9.90 18.80 8.61 IOO 9.90 9.90 3 none doubtful 9.90 18.72 8-53 IOO 9.90 9.90 50 15 none very faint 9.90 18.65 8.46 IOO 9.90 9.90 50 15 none very faint 9.90 18.67 8.48 With cadmium chloride and gold chloride, the apparently catalytic effect is due entirely to chlorine retained in solution, while with chromic chloride and platinic chloride increased effects are due partly to chlorine retained in solution and partly to the total reduction of residual oxides of manganese. ACIDIMETRY AND ALKALIMETRY. The Use of Succinic Acid as the Standard in Neutralization Processes. Methods for preparation and the use of succinic acid as a stand- ard in the titration of an alkali hydroxide have been studied by Phelps and Hubbard.* Succinic acid was prepared by hydrolysis of the pure ethyl ester, by hydration of the pure anhydride, by recrystallization of the commercial acid from hot water, and by recrystallization of the commercial acid from hot water * I. K. Phelps and J. L. Hubbard, Am. Jou'-. Sci., [4], xxiii, 211. APPLIANCES AND GENERAL PROCEDURE 55 containing nitric acid. The product obtained by each of these methods dries to a constant weight in air, and loses nothing on further standing over sulphuric acid. From ethyl succinic ester, boiling at 2I3.3-2I3.5 C. under a barometric pressure of 749 mm., pure succinic acid was obtained by boiling it for four hours on a return condenser with nitric acid and water in these proportions: 20 cm. 3 of succinic ester, 200 cm. 3 of water, three drops of nitric acid. The solution was evaporated to crystallization, and, after filtering off from the mother liquor, the solid product was recrystallized from distilled water. The crystals dried carefully in the open air to constant weight melted in an open capillary tube at 182.3. Ammonium Hydroxide against Succinic Acid. Succinic acid, grm. HC1 value of succinic acid. Found, grm. Theory, grm. Error, grm. Acid from ester. 0.2000 0.1235 0.1235 o.oooo O. 2OOO 0.1235 0.1235 o.oooo O.2OOO O.I23S 0.1235 o.oooo O.2OOO 0.1235 0.1235 0.0000 Acid from anhydride. 0:2000 O.I23I 0.1235 O.OOO4 0.2000 0.1233 0.1235 O.OOO2 O.2OOO 0.1234 0.1235 O.OOOI O.2OOO 0.1234 0.1235 o.oooi Acid recrystallized from water. O.2OOO O.I23I 0.1235 O.OOO4 O.2OOO O.I23I 0-1235 O.OOO4 0.2000 O.I23I 0.1235 0.0004 0.2000 , O.I23I 0.1235 O.OOO4 O. 2OOO O.I23I 0.1235 O.OOO4 O.2OOO O.I23I 0.1235 O.OOO4 Acid recrystallized from water containing nitric acid. O.2OOO 0.2000 0.1233 0.1233 0.1235 0.1235 O.OOO2 O.OOO2 For the preparation of the acid from succinic anhydride the anhydride was purified by recrystallizing from absolute alcohol 56 METHODS IN CHEMICAL ANALYSIS until the crystals obtained, after carefully drying, melted sharply at H9C. The pure anhydride obtained in this manner was converted to the acid by dissolving it in distilled water heated to the boiling point. The crystals formed on cooling the solution were filtered off and dried in air and over sulphuric acid. The melting point of the product was 182.8. Preparations made by dissolving the acid of commerce in dis- tilled water at the boiling point, crystallizing by cooling, and drying in the air, melted at 181.7. The acid made by dissolv- ing the succinic acid of commerce in boiling water, adding nitric acid, crystallizing, and drying, melted at 182.3. Tests of these preparations of succinic acid were made by titrating solutions of weighed amounts by approximately n/io ammonium hydroxide, with cochineal as an indicator, the ammo- nium hydroxide having been previously standardized against ap- proximately n/io hydrochloric acid, the exact strength of which had been determined gravimetrically by precipitation with special precautions by silver nitrate. The results are given on page 55. Organic Acids and Acid Anhydrides as Standards in Neutraliza- tion Processes. Phelps and Weed* have shown that, with phenolphthalein as an indicator, pure sodium hydroxide in solution and pure barium hydroxide in solution may be determined very exactly by titra- tion against weighed amounts of succinic acid, succinic an- hydride, malonic acid, benzoic acid, phthalic acid and phthalic anhydride used as standards. Following are experimental results (PP- 57, 58). Phelps and Weed point out that these organic acids and acid anhydrides, in pure state, make standards in alkalimetry and acidimetry, as accurate as the best previous standard, hydro- chloric acid determined gravimetrically as silver chloride. The most serviceable are those most readily soluble in water, suc- cinic and malonic acids. Phelps and Weedf point out, also, that these organic acids and anhydrides since they may be used to fix the standards of alkali hydroxides, and these to fix the standard of hydrochloric * I. K. Phelps and L. H. Weed, Am. Jour. Sci., [4], xxvi, 138. t Am. Jour. Sci., [4], xxvi, 143. APPLIANCES AND GENERAL PROCEDURE 57 Sodium Hydroxide and Barium Hydroxide against Succinic Acid and Succinic Anhydride. Succinic acid. grm. Succinic anhydride. grm. HC1 value of NaOH used. grm. HCl value of BaO 2 H 2 used. grm. Theory in terms of HCl. grm. Error in terms of HCl. grm. fo 2OOO O 1236 O.I23S +O.OOOI 1 o 2OOO o 1238 o. 1235 +0.0003 I -\ O 2OOO O 1237 o. 1235 +O . OOO2 O 2OOO o. 1236 0.1235 +O.OOOI |^O 2OOO o. 1236 o. 1235 -f-o.oooi fo 2OOO O 1237 O. 123"? +O.OOO2 2 { O 2OOO O 1237 O. I23i? +O.OOO2 1 O 2OOO O 1237 O.I235 +O.OO02 t O 2OOO i u . tvw o 1237 O.I235 +O.OOO2 ( O 2OOO o. 1237 O.I235 +O . OOO2 f O . 2OOO 1 O . 2OOO | O.2OOO {o . 2OOO O 2OOO O 14.^8 0.1238 0.1237 0.1235 0.1236 0.1235 0.1235 0-1235 0.1235 o. 1458 +o . 0003 +O . OOO2 o.oooo +O.OOOI o.oooo o 2000 o 1458 o. 1458 o.oooo O 2OOO o 14.^0 o. 1458 -J-O.OOOI O. 2OOO o. 1458 0.1458 o.oooo O. 2OOO o. 1457 0.1458 O.OOOI O. 2OOO o. 1456 o. 1458 O.OOO2 O. 2OOO o. 1459 0.1458 +O.OOOI O 2OOO O 14^8 o 1458 o oooo 1. Freshly made the ester and dried over sulphuric acid. 2. Dried for a year over sulphuric acid. 3. Dried for a year over calcium chloride. Sodium Hydroxide and Barium Hydroxide against Malonic Acid* Malonic acid, grm. HCl value of NaOH used. grm. HCl value of BaO 2 H 2 used. grm. Theory in terms of HCl. grm. Error in terms of HCl. grm. O.20OO o. 1404 o. 1402 -}-O.OOO2 O. 2OOO O. 1403 o. 1402 -J-O.OOOI O. 2OOO o 1402 o 1402 o oooo O. 2OOO o 1401 o 1402 o oooi O.2OOO o 1401 o 1402 o oooi O.2OOO o 1400 o 1402 O OOO2 O.2OOO O. 2OOO o. 1402 o 1400 0.1402 o 1402 o.oooo O OOO2 * Prepared from malonic ester by hydrolysis, recrystallized from water, and dried over sulphuric acid. METHODS IN CHEMICAL ANALYSIS Sodium Hydroxide and Barium Hydroxide against Benzoic Acid.* Benzoic acid, grm. HC1 value of NaOH used. grm. HC1 value of BaO 2 H 2 used. grm. Theory in terms of HC1. grm. Error in terms of HC1. grm. O.2OOO o 2000 0.0598 O.O5Q9 0.0597 0.0597 +O.OOOI +O.OOO2 O. 2OOO O. 2OOO 0.0597 0.0598 0.0597 0.0597 0.0000 +O.OOOI O 2OOO o otjoS O o<07 +o oooi O 2OOO o o 235- J Gooch and Walker, cf. page 454. Kjeldahl, Zeit. anal. Chem., xxii, 366; Furry, Am. Chem. Jour., vi, 341; Groger, Zeit. angew. Chem., 1894, 52. 6o METHODS IN CHEMICAL ANALYSIS by the same method. The only difficulty with the Groger process lies in the fact that in dilute solutions, as shown by Furry,* the end-point of the final reaction between iodine and sodium thiosulphate is somewhat obscured by a peculiar back- play of color due to a continuous slow liberation of iodine in the system. When the suitably concentrated iodide-iodate mixture (i grm. of KI and 0.166 grm. of KIO 3 in 50 cm. 3 ) is acted upon by a definite quantity of approximately decinormal hydrochloric acid in a total volume not exceeding 100 cm. 3 , the iodine set free, as determined by titration with n/io sodium thiosulphate (standardized against iodine of strength fixed by titration against weighed arsenious oxide), measures very exactly, according to Phelps and Weed,f the acid taken. The recorded results of experiments, in which the end reaction was brought about by iodine added directly or set free by addition of more standard acid after the bleaching by thiosulphate, are exceedingly good. Volume about 100 cm. 3 HC1 values. HC1 solution used. gnu. NajSjOa solution used. grm. Iodine solution used. grm. HC1 found, grm. Error in HC1. grm. o. 1074 O IO7< O IO7< -J-O . OOOI o. 1074 w -*- w / O o. 1074 \J . A v/ ^ o. 1074 O. OOOO 0.1074 0.1075, 0.1075 +0.0001 0.0520 0.0520 O.O52O o.oooo 0.1560 0.1562 0.1562 +O.OO02 0.0645 0.0712 0.0068 o . 0644 O.OOOI 0.0968 0.1068 O.OIOI 0.0967 0.0001 0.0645 0.0712 0.0067 o . 0645 o.oooo 0.0968 0.1068 O.OIOI 0.0967 O.OOOI 0.0484 0-0534 o . 0050 o . 0484 o.oooo 0.0645 0.0748 0.0104 0.0644 O.OOOI 0.0484 0.0534 o . 0050 o . 0484 o.oooo 0.0645 0.0712 0.0066 o . 0646 +O.OOOI Determination Alkali hydroxides may be determined by the Hyllroxides and ac ti n of an excess of standard hydrochloric acid or Carbonates. sulphuric acid, the excess being determined by esti- mation of the iodine set free by the iodide-iodate mixture at suit- able concentrations. Alkali carbonates may also be similarly * Am. Chem. Jour., vi, 341. t Am. Jour. Sci., [4], xxvi, 143. APPLIANCES AND GENERAL PROCEDURE 6l determined by treatment with an excess of sulphuric acid, boil- ing to remove carbon dioxide from the solution containing the nonvolatile acid, and determination of iodine liberated by the action of the iodide-iodate mixture in a total volume of about 100 cm. 3 Results obtained by this treatment of sodium hy- droxide first treated with carbon dioxide are given in the follow- ing statement, and, for comparison, results obtained by titration of the excess of standard acid by standard sodium hydroxide. Volume not Exceeding 100 cm. 3 HC1 values. Treatment with CO 2 . min. NaOH solution used. grm. H 2 S0 4 solution used. gnu. NaOH solution to coloration. grm. Na,S,0, solution used. grm. Iodine solution to coloration. grm. Difference in terms of HC1. grm. je O IOI2 o. 1306 O.O2Q3 -|-O OOOI 3Q O IOI2 o. 1219 0.0205 -|-O OOO2 jr O 1349 o. 1524 0.0172 -|-o 0003 2Q O I 34Q o 1568 o 0216 -f"O OOO^ T I? O IOI2 o 1306 O 0302 O 0013 -f-o 0005 2Q O IOI2 o 1306 o 0302 O 0013 -r~O OOO^ 15 35 0.1349 0.1349 0.1742 0.1742 0.0392 0.0552 O . 0004 0.0162 +o . 0005 +0.0003 Determination of When certain salts are brought into association Adds Liberated with water, a tendency to the formation of more Hydrolysis. ^ as { c products and free acid in consequence of the hydrolytic action of water becomes evident. Such action goes on until an equilibrium is reached between the factors and products of reaction. Action between Iodide-iodate Mixture and Certain Salts. In the presence of the iodide-iodate mixture, free acid, a product of the hydrolytic action, may be constantly removed, and so the hydrolysis may proceed further, the iodine set free in reaction of the iodide-iodate mixture upon the free acid being obviously a measure of the degree of hydrolytic action. Such action may proceed to the complete hydrolysis of the salt or may cease at an earlier stage, depending upon the nature of the salt. The behavior of certain salts in presence of the iodide-iodate mixture has been studied experimentally by Moody.* In Moody's experiments, a suitable amount of the salt to be tested is put * Seth E. Moody, Am. Jour. Sci., [4], xx, 181; xxii, 176, 379, 483. METHODS IN CHEMICAL ANALYSIS into the Voit flask (B) of the apparatus shown in Fig. 3,* and 10 cm. 3 of an iodide-iodate mixture (i grm. KI. : 0.3 grm. KIO 3 ) are added, the Drexel flask (C) and trap are charged with a solution containing 3 grm. of potassium iodide, hydrogen is passed from the generator through the apparatus, and the con- tents of the Voit flask are heated until (according to circum- stances) all or nearly all the liberated iodine is collected in the Drexel flask. The free iodine, whether in the receiver or re- maining in small amount in the. distilling flask, is titrated with sodium thiosulphate. Aluminium Sulphate. Proceeding in this manner, Moody found that aluminium sulphate, though only partially hydrolyzed at ordinary temperatures, is completely broken up by heating, according to the reaction A1 2 (S0 4 ) 3 + 5 KI + KI0 3 + 3H 2 = 2 A1(OH) 3 + 3 K 2 SO 4 + 312, and that aluminium chloride behaves similarly. The following results were obtained with potassium alum, the potassium sulphate not being susceptible to hydrolytic action. Approx. n/io aluminium potassium alum. KIO,. KI. Time in minutes. Approx. M/IO Na 2 S 2 O,. A1 2 3 calculated from iodine found. A1 2 3 precipi- tated and weighed. Difference. cm. s grm. grm* cm. 8 grm. grm. grm. 2 S o-3 .0 3 24-55 0.0410 0.0414 0.0004 25 3-3 .0 3 24.60 0.0411 0.0416 0.0005 25 o-3 .0 25 24.50 o . 0409 0.0414 0.0005 25 o-3 .0 30 24.70 0.0413 0.0416 0.0003 25 0.3 .0 35 24.50 o . 0409 0.0415 0.0006 25 o-3 .0 30 24-55 0.0410 0.0415 0.0005 25 o-3 .0 25 24.50 0.0409 0.0415 0.0006 Upon the presumption that the aluminium salt taken is per- fectly neutral, the reaction affords means for determining alu- minium iodometrically. Upon repeating the experiment with similar amounts of an ammonium alum it was found that the amounts of iodine liberated were in the average somewhat excessive, corresponding to 0.0006 grm. of A1 2 O 3 more than the theory called for. The process is, therefore, less exact in the presence of ammonium salts. In fact, * See page 4. APPLIANCES AND GENERAL PROCEDURE as will be shown, ammonium sulphate in the amounts taken may be completely hydrolyzed in the course of three hours. Chromic Sulphate, Tin Chloride. Chromic sulphate, taken in the form of chrome alum, undergoes complete hydrolysis in presence of the iodide-iodate mixture with precipitation of chromic hydroxide, as does the double potassium tin chloride, SnCl 4 .2KCl. Iron Sulphate. Ferric sulphate reacts like aluminium sul- phate according to the equation Fe 2 (S0 4 ) 3 + 5 KI + KI0 3 + 3 H 2 O = 2 Fe(OH) + 3 K 2 SO 4 +3X2. The hydrolysis of ferrous sulphate is accompanied by oxidation of the ferrous hydroxide at the expense of the iodate, as follows : 3 FeS0 4 + 5 KI + KIO 3 + 3 H 2 O = 3 Fe(OH) 2 + 3 K 2 SO 4 + 3 I 2 . 6 Fe(OH) 2 + KI0 3 + 3 H 2 O = 6 Fe(OH) 3 + KI. From the experimental tests it appears, however, that the hydrolysis of ferrous sulphate is complete in the presence of the iodide-iodate mixture, and that the iodine set free is an exact measure of the SO 4 ion present and of the iron in the ferrous sul- phate of ideal composition. Iodine Volume. KI. KIO 3 . Time in minutes. Approx. /io Na 2 S 2 O 3 . Iodine found. value of FeS0 4 Difference. taken. cm. 3 grm. grm. cm.* gr 111. grm. .gnu. 40 I .O 0-45 30 26.67 0-3324 0.3322 +0.0002 40 I .O 0-45 30 26.68 0.3325 0.3322 + 0.0003 40 1.0 0-45 30 26.65 0.3321 0.3322 o.oooi 40 1.0 0-45 30 26.67 0.3324 0.3322 +O.OOO2 40 1.0 0.45 3 26.66 0.3323 0.3322 + O.OOOI Cobalt Sulphate. Cobalt sulphate when similarly hydrolyzed is likewise oxidized at the expense of the iodate, the reactions following similar lines. 3 CoS0 4 + 5 KI + KI0 3 + 3 H 2 = 3 Co(OH) 2 + 3 K 2 SO 4 + 3 I 2 . 6 Co(OH) 2 + KIO 3 + 3 H 2 O = 6 Co(OH) 3 + KI. The iodine value obtained in test experiments, upon the hypoth- esis that cobaltous sulphate is completely hydrolyzed, is closely 6 4 METHODS IN CHEMICAL ANALYSIS comparable with the iodine equivalent of the cobalt found by the electrolytic deposition of the metal. Volume. KI. KIO 3 . Time in hours. Approx. M/IO Iodine value found. Iodine value of CoS0 4 Difference. 812 3- taken. cm.* gnu . grm. cm. 3 grm. grm. grm. 40 .0 o.45 4 17.80 0.2244 0.2242 + O.0002 40 .0 0-45 32 17.78 0.2242 0.2242 o . oooo 40 .0 0.45 3f 17-75 0.2238 o. 2242 0.0004 40 .0 0-45 4 17.79 0.2243 0.2242 + 0.0001 40 .0 0-45 4 17.79 0.2243 0.2242 +O.OOOI 40 .0 0.45 4 17.78 o. 2242 o. 2242 o . oooo Nickel Sulphate. Nickel sulphate, like cobalt sulphate, is hydrolyzed completely, after a considerable time, in the presence of the iodide-iodate mixture,' likewise yielding iodine, which may be collected similarly and estimated as a measure of the nickel present. Nickelous hydroxide formed in the reaction remains, however, unoxidized by potassium iodate in neutral solution, and therefore the following equation shows the final products : 3 NiS0 4 + 5 KI + KI0 3 + 3 H 2 = 3 Ni(OH) 2 + 3 K 2 SO 4 + 3 ^* Volume. KI. KIO 3 . Time in hours. Approx. K/IO Na 2 S 2 3 . Iodine found. Iodine value of NiS0 4 taken. Difference. cm. s grm. grm. cm. 8 grm. grm. grm. 40 .O 0-45 3 17.87 0.2254 0.2255 O.OOOI 40 .0 0-45 3 17.88 0.2256 0.2255 +0.0001 40 .O 0.45 3 17.84 0.2250 0.2255 0.0005 40 .O 0-45 3 17.87 0.2254 0-2255 O.OOOI 40 .O 0.45 3 17.83 0.2249 0.2255 0.0006 Thus it appears that nickel sulphate may be completely hydro- lyzed in the presence of the iodide-iodate mixture, and that the nickel of nickel sulphate of ideal composition can be estimated from the amount of iodine liberated in the action of that salt upon the iodide-iodate mixture. Zinc Sulphate. Like the sulphates of nickel, cobalt, iron, aluminium and chromium, zinc sulphate is hydrolyzed in pres- ence of the iodide-iodate mixture, but unlike the sulphates of the other elements mentioned the reaction stops short of complete APPLIANCES AND GENERAL PROCEDURE hydrolysis, as shown in the results of the table. The reaction may be expressed by the equation 15 ZnSO 4 + 20 KI + 4 KIO 3 + 12 H 2 O = 3 Zn 5 (OH) 8 SO 4 + 12 K 2 SO 4 + 12 I 2 . Volume. KI. KIO 3 . Time in hours. Approx. n/io Na 2 S 2 3 . Iodine found. Equivalent of SO, found. SO 3 present. cm- 3 grm. grm. cm. 3 grm. grm. grm. 40 I .O 0-45 * 27-8 0-3557 0.1123 0.1408 40 I .0 0-45 i 28.0 0.3582 0.1130 0.1408 40 I.O 0-45 3 2 7 .8 0-3557 O.II23 0.1408 40 1 .0 0-45 3 2 7 .8 0-3557 o . 1 1 23 0.1408 The mean percentage of hydrolysis here found is 79.90. It appears that a basic sulphate containing 5 of Zn to one of SO 4 is formed, and so definitely that from the iodine liberated the zinc content may be calculated with accuracy. Ammonium Sulphate. When solutions of ammonium sul- phate are subjected to heat the salt is hydrolyzed , and as the acid is increased, either as a direct product of this hydrolysis or by addition, further dissociation is inhibited. The decrease in dis- sociation is dependent upon the increase of the acid, and when sufficient acid is present further hydrolysis is entirely prevented. The amount of hydrolysis is, however, small under the most favorable conditions.* Moody has studied the effect of the iodide-iodate mixture in eliminating the acid as it is produced. In experiments made in the manner described above it was found to be impossible to collect the iodine in the Drexel flask used as a receiver when charged with potassium iodide only, although it was evident that much iodine came over. It appeared, upon investigation, that ammonium iodide and ammonium iodate were formed by reaction in the receiver between the liberated iodine and the ammonia also volatilized, and to obviate the difficulty sulphuric acid was added to the contents of the receiver into which the dis- tillate was passed. Under these conditions iodine is obtained in amount corresponding to that which should be eliminated when the ammonium sulphate is entirely hydrolyzed. * Bruck, Dissertation, Giessen, 1903. 66 METHODS IN CHEMICAL ANALYSIS This is shown in the subjoined table: Volume. KI. KIO 3 . Time in hours. H 2 S0 4 (i*D in the receiver. Approx. K/IO Na,SA. Iodine found. Iodine value of (NH4),S0 4 taken. Difference. cm.* grm. grm. cm. 3 cm. 3 grm. grm. grm. 35 .0 0.30 3 40 38.25 0.4769 0-4773 0.0004 35 .0 0.30 3 40 38.25 0.4769 0-4773 0.0004 35 .0 0.30 3 40 38.30 -4775 0-4773 +O . OOO2 35 ,o 0.30 3 40 38.25 0.4769 0-4773 O.OOO4 35 .0 0.30 3 40 38.23 0.4766 0-4773 O.O007 In another series of experiments, the apparatus of Fig. 4* was used. In these experiments the mixture was boiled in the first Voit flask and the distillate passed from the first flask, V 1 , through the second flask, V 2 , containing 50 cm. 3 of n/io H 2 SO 4 to take up the ammonia, and then into the receiver containing potassium iodide without acid. The sulphuric acid remaining free in the second Voit flask at the end of the operation was determined by titration of the iodine liberated upon the addition of theiodide- iodate mixture, the difference between this amount of iodine and the iodine equivalent of the sulphuric acid used being the meas- ure of the ammonia absorbed. The iodine passing to the receiver was determined as usual by titration with sodium thiosulphate. The results of these experiments, given below, show that the sulphuric acid neutralized in the Voit flask is a measure of the ammonia, while the iodine in the Drexel flask corresponds to the sulphuric acid of the ammonium sulphate. Similar results were obtained with ammonium chloride. Volume, 45 cm. 3 ; KI, i grm.; KI0 3 , 0.6 grm.; Time, j to 3% hours. Iodine equivalent of ammonia absorbed in Voit flask. Iodine estimated in Drexel flask. Iodine value of (NH 4 ) 2 S0 4 taken. Approx. w/io Na 2 S 2 3 . I. Difference. Approx. H/IO Na 2 SjO,. I. Difference. gnu. cm.* grm. grm. cm.* grrn. grm. 0-4773 38.15 0-4757 0.0016 38.23 0.4767 0.0006 0-4773 38.20 0.4763 O.OOIO 38.25 0-4769 0.0004 0-4773 38.15 0-4757 0.0016 38.20 0.4763 O.OOIO 0-4773 38.20 0.4763 O.OOIO 38.27 0.4771 O.OOO2 0-4773 38.17 0-4759 0.0014 38.20 0.4763 O.OOIO 0-4773 38.15 0-4757 0.0016 38.20 0.4763 O.OOIO 0-4773 38.20 0.4763 O.OOIO 38.25 0.4769 0.0004 * See page 5. APPLIANCES AND GENERAL PROCEDURE 67 This procedure is not presented as an analytical method for determining ammonia or the acid-ion, but to show that the effects of hydrolysis must not be ignored when ammonium salts are heated in solution with the iodide-iodate mixture. Alums- Basic Moody has shown * how the phenomena of hy- Aiumina and drolysis may be applied to the determination of basic alumina and free acid in the analysis of alums and commercial aluminium sulphates, which, beside aluminium sul- phate, may contain ferrous sulphate, ferric sulphate, and zinc sulphate as impurities. Potassium sulphate and sodium sul- phate, if present, do not set free iodine from the boiling solution containing the iodide-iodate mixture. The determinations of the ferrous iron, the ferric iron, the zinc, and the ammonia furnish data from which the equivalent amounts of sulphuric acid, to be taken into account in the reckoning of the free acid or basic alumina, may be calculated. The behavior of these commercial products toward the iodide-iodate mixture affords, therefore, an easy method of determining basic alumina or free acid, as the case may be. Following are the details of treatment : I. A sample of 15 grm. is weighed and treated with water. The solution is filtered and made up to I liter. The material which does not dissolve is dried at 1 00 and weighed as insoluble material. II. Of the solution, a portion of 25 cm. 3 is titrated directly with standard potassium permanganate to find the amount of iron in the ferrous salt, and from this is calculated the ferrous oxide. III. Of the solution, another portion of 25 cm. 3 is treated with zinc to reduce the ferric salt and then titrated with permanganate to give the total iron. From the difference between the total iron and the ferrous iron is calculated the ferric oxide. IV. Of the solution, a portion of 25 cm. 3 is diluted to 50 cm. 8 , treated with 3 grm. of sodium acetate and I cm. 3 of acetic acid, and electrolyzed with the use of the rotating cathode f by a current of about 2 amperes for 30 minutes. The deposit of zinc, including some iron, is washed with alcohol, dried and weighed. * Am. Jour. Sci., [4], xxii, 483. t See page n. 68 METHODS IN CHEMICAL ANALYSIS The solution of the deposit in sulphuric acid is titrated with permanganate, and the amount of iron thus found is deducted from the total weight of the deposit to give the amount of zinc. From the zinc is calculated the zinc oxide. V. Of the solution, a portion of 25 cm. 3 is drawn from a bu- rette into the Voit flask of the distillation apparatus, a solution (10 cm. 3 ) containing 0.3 grm. of potassium iodate and I grm. of potassium iodide is added, the mixture boiled, and the iodine, collected in the receiver charged with water containing 3 grm. of potassium iodide (and acidified with sulphuric acid in case the substance contains ammonia), is titrated with sodium thiosul- phate. The iodine set free corresponds to the various oxides, to am- monia and to sulphuric acid in the following proportions : A1 2 3 61; Fe 2 O 3 61; FeO a Is 5ZnO 81; NH 3 I; H 2 SO 4 a Is it is the total iodine. VI. Of the solution, a portion of 25 cm. 3 is treated in an open beaker with I grm. of potassium iodide and 0.3 grm. of potassium iodate, the mixture is boiled until nearly all free iodine is ex- pelled, the precipitate is filtered on paper, ignited carefully, and weighed as A1 2 O 3 , Fe 2 O 3 and ZnO, the total oxides. Given the total iodine liberated in the distillation process, the weight of the total oxides obtained in the parallel boiling process, the ferrous oxide and ferric oxide by the permanganate titrations and the zinc oxide deduced from the corrected electrolytic deter- mination, the total alumina and the basic alumina or free acid (as the case may be) are easily calculated. Total oxides (ferric oxide + ferrous oxide + zinc oxide) = total alumina. /6X I2 ^97\ Qr ( 7454 ) x total alumina = iodine equivalent to total alumina. APPLIANCES AND GENERAL PROCEDURE 6 9 or (4.767) X ferric oxide = iodine liberated by ferric sulphate. /2Xi26. 9 7\ or ( iodine liberated by ferrous sulphate. /8 X l26 -97\ or ( 249 6) x zinc oxide = 4 / 5 X 814 / -- i iodine liberated by zinc sulphate. (7.469) X ammonia = iodine liberated by ammonium sulphate. Total iodine (iodine corresponding to total alumina, ferric sulphate, ferrous sulphate, zinc sulphate, ammonium sulphate) = differential iodine. Differential iodine (if positive) Differential iodine (if negative) X = basic alumina " The results of analyses of four specimens of alums are given in the following table : Percentage Composition. AUOs. FeO.* ZnO. NHs. Insoluble material. Total. Combined. Basic. No. I: (i) 14.48 J 3-49 0-99 0-43 3-70 None. 0.61 (2) 14.28 I3-46 0.82 0.44 3-70 None. 0.61 No. II: d) 15-94 14.21 i-73 0-43 1-47 None. 0.21 (2) I5-90 14-34 1-56 0-45 1-34 None. O.2I No. Ill: (i) J5-59 14.81 0.78 0-34 0.73 None. 0.71 (2) 15-97 14.80 1.17 0.36 0.82 None. 0.71 D: (i) 16.59 15.24 1-35 0.24 O.II None. 0.61 (2) 16.37 15.18 1.19 0.24 0.19 None. 0.61 With a trace of 70 METHODS IN CHEMICAL ANALYSIS The Use of the Bromide-bromate Mixture and the Estimation of the Bromine Evolved. The reaction of a mixture of potassium bromide and potassium bromate upon aluminium sulphate has been studied by Gooch and Osborne* with the aid of the apparatus previously described. f Assuming that the acidic ion is entirely liberated from the aluminium salt, the reaction should follow the equation 2 KA1(S0 4 ) 2 + 5 KBr + KBrO 3 + 3 H 2 O = 2 Al(OH), + 4 K 2 S0 4 + 3 Br 2 . The precipitation proves to be complete, or nearly so; but the process of hydrolysis does not go easily to the point of forming aluminium hydroxide. A reasonable excess of the bromide-bro- mate mixture is able in a moderate time to carry the hydrolysis of aluminium sulphate to a fairly definite point corresponding nearly to the removal of five-sixths of the acidic ion, while the iodide-iodate mixture under similar conditions of action removes practically all the acidic ion. With a very large increase in the concentration of the bro- mide and bromate and prolonged boiling, the completion of the hydrolysis to the point of liberating bromine equivalent to the entire amount of the acidic ion is very nearly realized. Like the iodide-iodate mixture, the bromide-bromate mix- ture is a very delicate indicator of free acid; 0.00018 grm. of sulphuric acid proving to be sufficient to liberate bromine from the bromide-bromate mixture when boiled in the Voit flask under the experimental conditions. Some experiments to test the effect of a mixture of potassium chloride and potassium chlorate upon aluminium sulphate indi- cated that the hydrolysis under otherwise similar conditions is very slight compared with that produced by the bromide- bromate mixture or by the iodide-iodate mixture. The Reaction of Iodine with Alkali Hydroxides. When the solution of a metallic hydroxide is acted upon by iodine at a temperature high enough to decompose the small amounts of hypoiodites that might otherwise be present, the final * F. A. Gooch and R. W. Osborne, Am. Jour. Sci., [4], xxiv, 167. t See pages 4, 61. APPLIANCES AND GENERAL PROCEDURE 71 action results in the formation of an exactly neutral mixture of iodate and iodide according to the equation 6 KOH + 3 I 2 = RI0 3 + 5 RI + 3 H 2 O. In a process described elsewhere* for the determination of carbon dioxide, Phelps f has applied this reaction to the deter- mination of barium hydroxide. In this process, standard iodine is made to act in a suitable apparatus upon the barium hydroxide and the excess of it is determined. Instead of determining, according to the procedure of Phelps, the amount of iodine left over in the action of an excess of that element upon the alkali hydroxide, Walker and Gillespiet pre- fer to expel, by boiling, all free iodine remaining after action and to determine the amount of iodine liberated by acting with an acid from the residual mixture of iodate and iodide. According to the procedure recommended, an excess of deci- normal iodine is drawn into an Erlenmeyer beaker and the de- sired amount of alkali hydroxide is run in rapidly. The beaker, closed by a little trap, made of a calcium chloride drying tube, to prevent appreciable loss by spattering, is placed over a low flame, and the contents boiled until the last trace of the excess of iodine has been volatilized from the solution and the trap. The volume is carefully regulated before and during the boiling, being kept as small as possible, usually amounting to about 100 cm. 3 at the start and 35 cm. 3 at the close. In the case of barium hy- droxide, care has to be taken to keep the dilution sufficient to prevent the separation of the crystalline barium iodate, which is soluble with difficulty. To steady the ebullition a little spiral of platinum is introduced into the beaker. After the boiling, the colorless solution, containing a neutral mixture of iodate and iodide, is cooled in running water, and treated with 10 cm. 3 of dilute acid in the case of barium hydroxide, dilute [i : 3] hydro- chloric acid, to save the inconvenience of working in the pres- ence of precipitated barium sulphate; with potassium hydrox- ide, dilute [i : 3] sulphuric acid. The liberated iodine is titrated * See page 231. f Am. Jour. Sci., [4], ii, 70. $ Claude F. Walker and David H. M. Gillespie, Am. Jour. Sci., [4], vi, 455- See Fig. 6. 7 2 METHODS IN CHEMICAL ANALYSIS directly with sodium thiosulphate, in the presence of starch. Tests of this method in comparison with that of Phelps showed a fair agreement, though the results of both methods were in- variably lower by a small, nearly constant amount than those ob- tained by gravimetric estimations and by the Groger process.* This error of the Phelps process and its modification is probably due to the action of atmospheric carbon dioxide on the hydroxide solution during the short time it is exposed. While it affects considerably the value of the method as a means of accurately determining the absolute amount of hydroxide present in a given volume of solution, it apparently does not so seriously affect the accuracy of the differential method founded on the original Phelps process or its modification. Analyses ofn/io Hydrochloric Acid Solution. (By adding to excess of w/io Ba(OH) 2 , boiling with excess of iodine to colorlessness, and acidifying the residue.) Ba(OH) 2 HClby HC1 taken. Ba(OH) 2 taken. neutralized by HC1 found. Groger Variation. HC1. method. cm.* grm. grm. grm. gnu . grm. IS 0.17 0.1128 o . 0480 0.0481 +O.OOOI IS 0.17 0.1118 0.0475 0.0481 0.0006 IS 0.17 O.III2 0.0473 0.0481 0.0008 25 o. 26 0.1860 0.0791 0.0801 o.ooio 25 0.26 0.1866 0.0794 0.0801 0.0007 35 0.34 0.2634 O. 1 1 2O O. II2I o.oooi 35 0.34 0.2603 0.1107 O.II2I 0.0014 Analyses ofn/io Hydrochloric Acid Solution. (By adding to excess of w/io KOH, boiling with excess of iodine to color- lessness, and acidifying the residue.) KOH HClby HC1 taken. KOH taken. neutralized HC1 found. Gioger Variation. by HC1. method. cm. s grm. grm* grm. grm. grm. 2O 0.14 0.0972 0.0633 O . 0641 0.0008 2O 0.14 0.0975 . 0634 O . 0641 0.0007 25 0.14 O.I222 0.0795 O . 0801 O.OOO6 25 0.14 O.I2O7 0.0785 o . 0801 O.OOI6 In applying the Walker-Gillespie modification to the indirect determination of hydrochloric acid and sulphuric acid, the acid * See page 59. APPLIANCES AND GENERAL PROCEDURE 73 solution to be analyzed is drawn into an Erlenmeyer beaker, an excess of decinormal iodine is added, and a measured excess of standardized alkali is introduced. The beaker is trapped, the solution boiled, and the residue cooled, acidulated, and titrated with thiosulphate in presence of starch. Test results of the process are given in the accompanying tables. Analyses ofn/io Sulphuric Acid Solution. (By adding to excess of n/io Ba(OH) 2 , boiling with excess of iodine to decoloration, and acidifying the residue.) H 2 SO< taken. Ba(OH) 2 taken. Ba(OH) 2 neutralized by H,SO. H,S0 4 found. H 2 SO< by Groger method. Variation. cm. 3 grin. grm. grm. grm. grm. (i) 10 O. 21 o . 0884 O . 0506 o . 0496 +O.OOIO (2) 10 O. 21 0.0880 . 0503 o . 0496 +0.0007 (3) is 0.30 0.1328 0-0754 0-0745 +0.0009 (4) IS 0.30 0.1313 0.0751 0-0745 +0.0006 (5) 25 0-43 0.2168 0.1239 o . i 240 o.oooi (6) 30 0-43 o . 2600 o. 1481 0.1489 0.0008 The reaction between iodine and the hydroxides of potassium and barium in hot solution is, therefore, regular and complete under analytical conditions, and not appreciably affected by the mass action of considerable excesses of iodine. It is best applied in analysis by treating the alkali with an excess of iodine, removing this excess by boiling, and estimating the iodine in the residue. While mechanical difficulties and the interfering action of carbon dioxide may affect the extreme accuracy of the process as a direct means for analyzing alkalies, it may be indirectly applied with fair accuracy to the analysis of various acids and possibly to other compounds. The reaction between iodine and alkali carbonates, on the contrary, is irregular and cannot be made the basis of an analytical process. CHAPTER II. THE ALKALI METALS. SODIUM. The Detection of Sodium. THE application of the spectroscopic method to the detection of potassium and sodium in ordinary analysis is unsatisfactory on account of its failure, except in delicate quantitative compari- sons, to give any idea as to the quantity of either element indi- cated; and, since the most minute quantity of either element is sufficient to produce its characteristic line in the spectroscope, and many of the reagents employed in analysis contain a trace of alkali, the spectroscopic indication is often misleading. While to the careful observer the presence or absence of potassium in appreciable amount is revealed, the evidence as to the quantity of the ubiquitous element sodium is practically worthless. Kreider and Breckenridge * have developed a method for the detection of sodium based upon the utilization of the perchlorate method for the preliminary separation of potassium. The in- solubility of potassium perchlorate and the easy solubility of sodium perchlorate in 97 per cent alcohol afford means for the separation of these elements as well as for the identification of the former.f By converting to the chloride the sodium per- chlorate in the alcoholic filtrate from the precipitated potassium salt, exceedingly small amounts of sodium may be detected. For this purpose the passing of gaseous hydrochloric acid into the alcoholic solution of the sodium perchlorate, kept cool, has proved most effectual, the dehydrating effect of the acid upon the alcohol greatly increasing the insolubility of the sodium chloride. The delicacy of this test for sodium is shown in the results of the table. By the use of 10 cm. 3 of 97 per cent alcohol with gaseous hydrochloric acid 0.0003 g rm - f sodium oxide can be found with. * D. Albert Kreider and J. E. Breckenridge, Am. Jour. Sci., [4], ii, 263. t See page 88. 74 THE ALKALI METALS 75 certainty, and when the alcohol is allowed to become saturated with the gas even 0.00006 grm. The quantity of alcohol, 10 cm. 3 , is sufficient for all purposes, since this amount will dissolve about 2 grm. of sodium perchlorate ; but even in 40 cm. 3 0.0002 grm. of sodium oxide may be seen distinctly. It is evident that this method may also be applied to the quantitative determination of sodium. Test for Sodium. NaClO 4 taken, grm. Na 2 O equivalent, grm. 97 per cent alcohol. cm. s Indication. O.OIOO 0.00250 IO Very strong. 0.0050 O.OOI25 IO Strong. 0.0040 O.OOIOO IO Strong. 0.0030 0.00075 IO Strong. 0.0030 0.00075 10 Good. O.OO2O o . 00050 10 Good. O.OO2O o . 00050 10 Good. O.OOIO 0.00025 IO Good. 0.0005 O.OOOI2 IO Trace. 0.0003 o . 00006 IO Trace. 0.0001 o . 00003 IO None. 0.0000 . 00000 IO None. O.OOIO 0.00025 40 Distinct. In the following table are recorded the results of experiments in which the perchlorates of sodium and potassium were sepa- rated by 97 per cent alcohol and the sodium test made upon the alcoholic solution. The sodium and potassium salts were drawn from separate standard solutions of the purified perchlorates. Separation of Potassium with Test for Sodium. KC10 4 taken. grrn. K 2 O equivalent. grm. NaC10 4 taken. grm. Na 2 equivalent. grm. Indication for potassium. Indication for sodium. 0.0500 0.01699 0.0500 0.01250 Strong. Strong. O.O2OO 0.00680 O.O2OO 0.00500 Strong. Strong. O.OIOO 0.00340 O.OIOO 0.00250 Strong. Strong. 0.0050 0.00170 0.0050 0.00125 Strong. Strong. O.OO4O 0.00136 o . 0040 O.OOIOO Good. Good. O.OO3O O.OOIO2 O.OO3O 0.00075 Good. Good. O.OO2O O.OOO68 O.OO2O 0.0005 Good. Good. O.OOIO 0.00034 O.OOIO 0.00025 Good. Good. 0.0005 0.00017 0.0005 O.OOOI2 Trace. Trace. 0.0003 O.OOOIO 0.0003 O.OOOO7 Trace. Trace. O . OOOI o . 00003 0.0001 O.OOOO3 Faintest trace. None. o . oooo o.ooooo O.OIOO O.OO25O None. Strong. O.OIOO 0.00340 o.oooo 0.00000 Strong. None. 76 METHODS IN CHEMICAL ANALYSIS After evaporating to dryness on the steam bath, the residue was treated with 10 cm. 3 of 97 per cent alcohol, the insoluble potas- sium perchlorate was removed by filtering through a dry paper filter and dry funnel into a dry test tube, and the filtrate saturated with gaseous hydrochloric acid. The results show that sodium and potassium perchlorates when associated in any proportion may be separated by 97 per cent alcohol with exactness, and that the sodium may be indicated in the filtrate with great delicacy and certainty by the action of gaseous hydrochloric acid. For the conversion of other salts of sodium and potassium to perchlorates for the purpose of making the test for sodium, it is obvious that perchloric acid free from sodium must be employed. The perchloric acid prepared according to the method described later,* by heating sodium chlorate to form the perchlorate, destroying any residual chlorate by treatment with the strongest hydrochloric acid, separating sodium chloride by filtration on asbestos in the filtering crucible, and removing the excess of hydrochloric acid by evaporation, while answering perfectly well for the detection of potassium, is inapplicable to the test for sodium, because of the small amount of this element which the acid always contains on account of the partial solubility of sodium chloride in hydrochloric acid. This perchloric acid must, therefore, be purified by distillation, and to prevent loss by de- composition the process must be carried on under diminished pressure. To obviate violent ebullition, only acid previously concentrated to the fuming point should be subjected to the distillation process, and this only in small amounts. Rubber stoppers or connectors are not to be used where the acid may condense upon them and flow back into the flask, since oxidizable matter carried back causes explosions which vary in force and seriousness according to circumstances. To construct a suitable apparatus a strong side-neck flask is selected, the bottom covered to the depth of I cm. 3 with fine chips of porcelain, and into the neck is sealed a stoppered funnel reaching well into the bulb. The stopcock of this funnel is carefully cleansed and lubricated with metaphosphoric acid ob- tained by boiling sirupy orthophosphoric acid until the tem- perature of 350 C. has been attained. The side neck of the * Page 89. THE ALKALI METALS 77 flask is inclined upward for a short distance before being bent into the receiver, with which it is connected by a rubber stopper through which the tube extends for a safe distance. An ordinary bottle of 250 cm. 3 capacity serves for a receiver. This is closed by a doubly perforated stopper, and through one of the per- forations the adapter from the condenser is inserted, while through the other connection is made with a small glass bulb and absorption tube, filled with stick potash to take up any chlorine evolved in the inevitable slight decomposition of the acid, which in turn is connected with an automatic mercury pump. The whole flask is surrounded by a cylinder of thin sheet iron closed below, while the upper opening is protected by an asbestos cover in order that the heat may be uniformly applied up to the point at which condensed acid flows into the receiver. In making a distillation, 3 cm. 3 or 4 cm. 3 of the concentrated acid are admitted to the flask through the stoppered funnel, the pump is started, and when the pressure has been reduced to about 8 mm. the heat is raised to about 130 and the dis- tillation begins. During the process the pressure is kept at about 3 mm. to 5 mm., and the acid is admitted at about the same rate at which the distilled product drops from the con- denser. Under the conditions described, the process will yield per hour from 25 cm. 3 to 40 cm. 3 of the dihydrate of perchloric acid, the most concentrated form in which the acid is stable. Of this product o.i grm. of potassium oxide requires for pre- cipitation 0.16 cm. 3 . With pure perchloric acid at hand the separation of the sodium salt from a mixture of the pure chlorides of sodium and potassium, preparatory to making the sodium test, requires only treatment of the mixture with perchloric acid, evaporation on the water bath until the white fumes of perchloric acid appear, treatment of the residue with 97 per cent alcohol, and filtration. In general, however, it is necessary to remove certain interfering substances before applying the method. While potassium may be safely tested for in the presence of other bases and acids, except ammo- nium, caesium and rubidium, and sulphuric acid, there are many elements the insolubility of whose chlorides in alcohol necessitates their removal before testing for sodium. But among the common alkalies ammonium is the only one whose presence is METHODS IN CHEMICAL ANALYSIS objectionable. Lithium does not affect the test for either potas- sium or sodium. In the experiments made with potassium and sodium salts asso- ciated with salts of other common elements, the results of which are recorded below, the following treatment was adopted. The several groups of bases were successively removed in the ordinary way: Hydrogen sulphide in ammoniacal solution removed the lead, mercury, copper and zinc. Barium and calcium were re- moved by ammonium carbonate, the final filtrate being evapo- rated and the residue ignited to volatilize ammonium salts. The residue was dissolved and treated with barium hydroxide for the removal of magnesium, and, after filtering, the barium was again removed by ammonium carbonate, and the filtrate evaporated. The residue was then ignited as before, treated with 10 cm. 3 of boiling water and the solution filtered in order to remove the organic matter usually found at this stage of the treatment. To the filtrate was added o.i to 0.5 cm. 3 of pure perchloric acid, about 1.7 sp.gr., according to the amount of residue, and the mixture was evaporated over the steam bath until the white fumes of perchloric acid appeared. When the quantity of so- dium is large it is safer to evaporate several times in order to secure a complete conversion to the perchlorate. Potassium and Sodium in Mixtures of Salts of Other Elements. Pb, Cu, Al, Fe, Zn, Ba Ca and Mg KjO taken. Na 2 O taken. Indication for , Indication for as nitrates. potassium. sodium. grm. gnu. grm. 0.0500 of each. o.oooo o.oooo Faintest trace. Trace. 0.0500 of each. 0.0017 O.OOI2 Good. Good. o.iooo of each. 0.0000 0.0000 Faintest trace. Trace. o.iooo of each. 0.0000 0.0005 Faintest trace. Good. The fact that minute traces of sodium and potassium are found in the blank tests is to be expected from the delicacy of the method when it is remembered that but very few of the so-called chemically pure reagents are absolutely free from so- dium, and that even distilled water kept in glass vessels con- tains a trace of the alkali elements. However, the indication for sodium in the blank tests appeared only as a cloudiness and after complete saturation. When the quantity of sodium oxide THE ALKALI METALS 79 present is not less than 0.0005 grm. the precipitate appears in granular form and before the alcohol is completely saturated. The method is all that could be desired for the qualitative deter- mination of sodium. The Estimation of Sodium as the Pyrosulphate. Sodium sulphate in water solution may be recovered by simple evaporation and ignition. When free sulphuric acid is also present the acid sulphate is first formed, then, as the tempera- ture rises, the pyrosulphate, and gradually, at red heat, the neutral sulphate. In quantitative estimations it is usual to weigh in the form of neutral sulphate and to hasten conversion to this condition by making the final ignition in the atmosphere pro- duced by ammonium carbonate either projected into the hot cru- cible,* or placed in the crucible before the application of heat.f In a study of the behavior of the sulphates of the alkali ele- ments, Browning J has shown that by holding the temperature between 250 and 270 during the heating of sodium sulphate with the excess of sulphuric acid, sodium pyrosulphate, Na2S2Oy, may be formed with a degree of exactness which makes it possible to estimate sodium as that salt. Results of this procedure are given in the table, and, for com- parison, the results obtained in the usual process of weighing as the neutral sulphate after moistening with ammonium hydroxide and then igniting strongly. Sodium as the Pyrosulphate. NaCl taken, grm. NmSA calculated, grm. Na,S 2 T found. grm. Error, grm. Na,SO 4 calculated. grm. Na s SO 4 found. grm . Error. giro. 0.1042 o 1028 0.1978 O I(K2 0.1972 O ICK2 O.OOO6 O OOOO 0.1266 0.1254 O.OOI2 0.1093 0.1402 0.2075 O.2662 o . 2065 0.2651 O.OOIO O.OOII 0.1328 0.1703 0.1320 0.1696 O.OOOS 0.0007 Salts of potassium also yield when similarly treated the pyrosulphate, while salts of caesium and rubidium remain * Fresenius, Quant. Anal., trans, Cohn, pages 161, 165. t Treadwell, Anal. Chem., trans. Hall (1911), page 42. t Philip E. Browning, Am. Jour. Sci., [4], xii, 301. See page 92. 8o METHODS IN CHEMICAL ANALYSIS under the conditions of temperature in the condition of acid sulphates.* Lithium salts apparently do not yield the acid sulphate or the pyrosulphate in stable form. POTASSIUM. The Spectroscopic Detection and Determination of Potassium. Bunsen and Kirchoff originally determined the delicacy of the spectroscopic test for potassium by exploding in a darkened room a mixture of potassium chlorate with milk sugar, and observing the amount of finely divided chloride which it was necessary to diffuse through the given space in order to bring out unmistakably the spectrum of the metal. These investiga- tors were able to state that the presence of no more than ^^-^ of a milligram of the potassium salt is sufficient to give to the flame the characteristic spectrum of the element. By similar methods, the delicacy of the tests for lithium carbonate and sodium chlorate were shown to be a thousand times and three thousand times as delicate respectively. The practical detection of lithium and sodium spectroscopically is extremely easy and satisfactory, the only difficulty being that the exceeding delicacy of the sodium test and the ubiquitousness of sodium salts often make a decision doubtful as to whether that element is present appreciably in the substance under examination, or by accident. With potassium the case is different. Detection of Gooch and Hart f have succeeded in showing that Potassium. while the simple method in vogue for developing the luminosity of lithium and sodium the dipping of a single loop of platinum wire in the liquid or solid substance, and the placing of the loop in the Bunsen flame is unsatisfactory because so great a proportion of the material is dispersed before the heat of the flame effects the dissociation of the salt, much better results may be obtained by making use of more powerful flames and substituting for the single loop the hollow coils of platinum wire first recommended by Truchot J for the quantitative determina- tion of lithium. Such coils are easily made by winding the wire somewhat obliquely about a rod of suitable size, pressing the * See page 106. f F. A. Gooch and T. S. Hart, Am. Jour. Sci., [3], xlii, 448. J Compt. rend., Ixxviii, 1022. THE ALKALI METALS 81 coils close together, and gathering the free ends into a twisted handle. The size of the coils is adjustable without difficulty, so that each coil may be made to hold almost exactly any appro- priate amount, and to take up this amount with very little varia- tion in successive fillings, provided only that the precaution be taken in the process of filling to plunge the coil while hot into the liquid, and to keep its axis inclined obliquely to the surface of the liquid while withdrawing it. The coils, after use, may be conveniently cleaned by heating them in the flame of an annular burner beneath which is burned in a small lamp a 5 per cent solution of chloroform in alcohol, the products of combustion of the alcohol and chloroform being conveyed to the interior of the flame from below by a glass funnel fitted by a cork to the tube of the burner. This arrangement of apparatus gives a hot, colorless flame through which hydro- chloric acid is constantly diffused in condition to clean the wires completely and without attention. How closely the capacity of such coils may be adjusted, and how uniformly they may be filled, is shown in the figures of the accompanying record. Capacity of Coils. I. grm. II. grm. III. grm. IV. grm. V. grm. VI. . grm. Weight of filled coil 0.1996 o. 1996 0.1996 0.1996 0.1996 0.1986 O.OOIO 0.2780 0.2780 0.2780 0.2780 0.2781 o. 2760 O.OO2O2 0.2794 0.2794 0.2794 0.2794 0.2794 o. 2764 0.0030 0.2844 0.2845 0.2844 0.2845 o. 2844 o . 2804 o . 00404 0-3572 0.3571 0-3572 0-3571 0-3571 0.3521 o . 00504 0.3296 0.3296 0.3298 0.3298 0.3296 0.3100 o 01968 Weight of filled coil. . Weight of filled coil Weight of filled coil Weight of filled coil Weight of empty coil Weight of contents (mean).. It is plain that these coils afford simple means of taking up known amounts of material in solution. By gentle heating the liquid may be evaporated and the solid material left thinly spread and in condition to be acted upon with effect when brought to the flame. The evaporation may be conducted with little danger of loss of material by holding the handle of the coil across the flame with the coil proper at a safe distance outside; or, prefer- ably, by exposing the coils over a hot radiator, the handles resting upon a flat asbestos ring. In test experiments with such coils, use was made of a Muncke burner giving a flame 3 cm. wide at the base and 20 cm. in height. 82 METHODS IN CHEMICAL ANALYSIS The coil was introduced, after thorough drying, just within the outer mantle, on the side next the spectroscope, with the axis transverse to the slit of the spectroscope and the handle across the body of the flame. The spectroscope used was a well-made single-prism instrument with adjustable slit and scale. Observa- tions were made in the ordinary diffused light of the laboratory, with the eye in use shielded, the eye not in use covered, and the scale illuminated to the lowest degree of visibility. Upon experimenting with the apparatus described, it was found that coils holding 0.02 grm. of water, measuring 2 mm. in diam- eter by i cm. in length, made of No. 28 wire (0.32 mm. in diam- eter), and wound in about thirty turns, were well adapted to the purpose. With these coils and the flame adjusted to a height of 20 cm., 7 ^ mg. of potassium to the coilful produces a line distinctly visible with a slit of 0.18 mm., and y^V^ m g- with a slit of 0.23 mm.; and it is evident that this practical method of producing the spectrum of potassium gives results of a delicacy approaching that indicated in the experiments of Bunsen and Kirchhoff. These determinations were made with pure potassium chlo- ride carefully prepared from the chlorate, but in practical analysis it almost always happens that sodium is also present. Experiments were therefore made to determine the influence of varying amounts of the latter upon the visibility of the potas- sium line. The dilution of the potassium chloride was adjusted nearly to the last limit of visibility, so that a coilful of the liquid should contain 7 ^ mg., or yoW m g- f the element, according as the slit was 0.18 mm. or 0.23 mm. wide; to this solution were added weighed amounts of pure sodium chloride twice repre- cipitated and washed by hydrochloric acid; and the spectro- scopic tests were carried out as before, the sodium line being kept within the field of view with the potassium line. It is plain from the results (p. 83) that a considerable amount of sodium may be present in the flame, when the sodium line is in full view in the spectrum, and the slit adjusted to nearly the lowest limit of visibility of pure potassium, without inter- fering with the appearance of the potassium line, but that a quantity of sodium amounting to a hundred times that of the potassium may be sufficient to overpower entirely the spectrum of the potassium. The inference is plain that the proportion of .THE ALKALI METALS sodium to potassium should not be permitted to reach 100:1 when it is desirable to bring out the full delicacy of the spec- troscopic test with the sodium line in the field of view. When too great a proportion of sodium is present ,, its influence may be moderated by throwing the sodium line out of view, if the instrument in use possesses the necessary adjustment; other- wise, it is easy to effect a partial separation of the sodium chlor- ide from the potassium chloride, before bringing the solution to the "test, by precipitating with alcohol, and experience shows that the delicacy of the test for potassium is not impaired materially by such treatment of the mixed chlorides. Effect of Sodium upon the Potassium Line, Weight of K in a coil- ful. mg. Weight of Na in a coilful. mg. Ratio of Na:K. Width of slit. mm. Number of trials. Characteristic of line. O.OOIO o.oooo o : 0.23 3 Visible. O.OOIO O.OO2O 2 0.23 3 Visible. O.OOIO O.OIOO 10 0.23 3 Visible. O.OOIO 0.0200 ' 2O 0.23 3 Visible. O.OOIO 0.0400 40 0.23 3 Visible. O.OOIO O.O5OO SO 0.23 4 Very faint or none. O.OOIO O. IOOO IOO 0.23 3 None. O.OOIO O.2OOO 2OO 0.23 3 None. 0.0014 O.OOOO O 0.18 3 Visible. 0.0014 0.0560 40 0.18 3 Visible. 0.0014 0.0700 50 0.18 3 Visible. 0.0014 0.1400 IOO 0.18 2 Visible. 0.0014 0.1400 IOO 0.18 2 None. Certain experiments, in which the same method of manipula- tion was applied to the determination of potassium salts other than the chloride, indicate that the test is less delicate in the case of the sulphate, but rather more delicate in the case of the carbonate, the red line of potassium showing unmistakably in the latter case when only ^Vs of a milligram of potassium was present. Determination It has also been shown by Gooch and Hart * that of Potassium, ^e quantitative determination of small amounts of potassium may be successfully accomplished by bringing an un- known solution of potassium chloride to the measured volume * Loc. cit. 84 METHODS IN CHEMICAL ANALYSIS at which the residue left after evaporating the contents of a coil gives a line of the same strength as that produced by the residue of a coilful of a standard solution of potassium chloride and so determining the concentration of the unknown solution of finally measured volume. It is convenient to use several coils adjusted to the same capacity, and to clean, fill, dry and ignite them before the spectroscope in the manner previously de- scribed. From time to time the capacity of the coils should be readjusted, or else final comparison tests should be made with a single coil. It is essential that the eye of the observer should be kept as nearly as possible in the same condition of sensitiveness and in the same position in making the compari- sons, and best to hold the eye at the observing telescope during the entire interval between the exposures, shading it carefully, and to light the comparison scale of the spectroscope to the faintest possible visibility sufficient to fix exactly the position in which the line is to be sought. It is important, too, that the trials of the test and standard should come as closely together as possible in point of time. The observations of a series should be made by the same individual, the preparation and exposure of the wires being made by another. It is not possible to attain the best results in such work single-handed. The dilution of the test solution is made conveniently, and with sufficient accuracy, in ioo-cm. 3 cylinders graduated to half cubic centi- meters. It is advantageous to take a standard solution which corresponds to the presence of yjg- mg. of potassium to the coil- ful, and set the slit at a width sufficient to give lines for com- parison bright enough to be visible without much effort. The mode of proceeding is to dilute the test solution until the line given by the potassium contained in a coilful is of the same brightness as that given by the same quantity of the standard solution. From the final volume of the test-solution the quantity of potassium present in it is directly calculable; for, since any given volume of the test solution at its final dilution contains exactly the same amount of potassium as the same volume of the standard solution, it is only necessary to multiply the number expressing the volume in cubic centimeters of the test solution by that of the weight in grams of the potassium contained in one cubic centimeter of the standard in order to obtain the weight in grams of potassium in the whole test solution. THE ALKALI METALS The following is the record of the comparison of two unknown test solutions of pure potassium chloride with a standard solu- tion containing o.oooi grm. of the same salt to I cm. 3 Determination of Potassium in Pure Potassium Chloride. Experiment I. Experiment II. Volume of test Characteristic of line Volume of test Characteristic of line solution. compared with solution. compared with standard. standard. cm. cm. 1 2O Stronger. 30 Stronger. 50 Stronger. 60 Stronger. IOO Stronger. 82 Weaker. no Stronger. 70 Stronger. 1 20 Stronger. 7 6 Stronger. 150 Like. 78 Stronger. 2OO Weaker. 80 Like. 1 60 Weaker. 150 Like. (150 X o.oooi = 0.0150) Potassium found. Potassium taken, Limits on either side. . Error.. 0.0150 grm. 0.0150 grm. 0.0120 grm. 0.0160 grm. o.oooo grm. (80 X o . oooi = o . 0080) Potassium found 0.0080 grm. Potassium taken 0.0080 grm. T '^ -t.1 ( 0.0078 grm. Limits on either side < I 0.0082 grm. Error o . oooo grm. These results with the pure potassium salt show a degree of accuracy quite unexpected. In the former no attempt was made to approximate as closely as possible to the limits of dilution on both sides of the condition of equal brightness in test and stand- ard, but in the latter great care was taken in this respect and the .possible error does not exceed two and a half per cent of the en- tire amount of potassium involved. Similar experiments with potassium chloride in presence of varying amounts of sodium chloride, the sodium line being turned from the field of view, led to a recognition of the fact that sodium chloride tends to increase the brilliance of the potassium line, the maximum strengthen- ing effect of about 20 per cent occurring when the amount of sodium chloride stands to that of the potassium in the ratio of 10 : 1, a phenomenon due to the chemical action of sodium dissociated in the flame. The effect of ammonium chloride, and of hydrochloric acid, in destroying the potassium light is well known, and is due, presumably, in very large degree, to 86 METHODS IN CHEMICAL ANALYSIS prevention of the dissociation of the potassium chloride. The dissociated sodium should naturally reenforce the disintegrating action of heat upon the potassium chloride. The complication introduced by the presence of any certain amount of the sodium salt in the test may be obviated by the addition of the same amount of the sodium salt to the standard, and experience shows that an unknown amount of the sodium salt in the test may be matched with a degree of accuracy suffi- cient for the end in view. The determination of potassium in the presence of sodium is performed, therefore, in three stages: first, the test solution is diluted until its potassium line matches approximately that of the standard made to contain in I cm. 3 o.oooi grm. of potassium and o.ooio grm. of sodium chloride; secondly, sodium chloride is added to the solution thus diluted until the sodium lines of test and standard are brought to equal- ity ; and, finally, the potassium lines of test solution and standard solution are again brought into comparison. Following are the records of experiments made in this manner. Determination of Potassium in Presence of Sodium. Experiment I. Parti. Part II. Part III. Volume of test solu- tion Width of slit. Character- istic of potassium line as com- NaCl in 100 cm. 8 of test solu- Width of slit. Character- istic of sodium line as compared Volume of test solu- Width of slit. Character- istic of potassium line as com- pared with tion. with pared with cm.* mm. standard. grm. mm. standard. cm. 3 mm. standard. 30 70 100 0.23 0.23 0.23 Stronger. Stronger. Weaker. 0.01* 0.03 O.O5 OO GO OO M M M 6 d d Weaker. Weaker. Weaker. 108 108 0.23 0.23 'Weaker. Stronger. (Weaker. 0.08 0.09 0.18 0.18 Weaker. Weaker. 109 0.23 } Stronger. /Like. O.IO 0.18 Like. * Originally present. v The test solution having been accidently over-diluted, its strength was increased by the addition of o.ooio grm. of potassium, and this amount was added in the computation below to that originally in the test solution. (109 X o.oooi = 0.0109) Potassium found 0.0109 grm. Potassium taken. o.ono grm. Error o . oooi grm. =0.9 per cent. THE ALKALI METALS Determination of Potassium in Presence of Sodium. Experiment II. Part I. Part II. Part III. Volume of test solu- Width of slit. Character- istic of potassium line as com- NaCl in 100 cm. 3 of test solu- Width of slit. Character- istic of sodium line as compared Volume of test solu- tion. Width of slit. Character- istic of potassium line as com- pared with tion. with pared with cm. 3 mm. standard. grm. mm. standard. cm. 8 mm. standard. 40 100 0.23 0.23 Stronger. Stronger. 0.025* 0.050 0.18 0.18 Weaker. Weaker. 1 60 1 80 0.23 0.23 Stronger. Stronger. 160 0.23 Weaker. 0.085 0.18 Weaker. I QO 0.23 Stronger. O.IOO 0.18 Weaker. 2OO 0.23 Stronger. O.IIO 0.18 Like. 205 0.23 Weaker. 2IO 0.23 Weaker. * Originally present. / 205 X o.oooi =0.0205 ) \ > mean = 0.02021; 1 \2OO X O.OOOI =O.O2OO ) / Potassium found 0.02025 g rm - Potassium taken 0.02000 grm. Error 0.00025 grm. = 1.25 per cent. Experiment III. Parti. Part II. Part III. Volume of test solu- tion. Width of slit. Character- istic of potassium line as com- NaCl in ico cm. 3 of test solu- Width of slit. Character- istic of sodium line as compared Volume of test solu- tion. Width of slit. Character- istic of potassium line as com- cm. 3 mm. pared with standard. tion, grm. mm. with standard. cm. 3 mm. pared with standard. 40 80 IOO 0.23 0.23 0.23 Stronger. Stronger. Stronger. 0.045* 0.082 0.18 0.18 Weaker. Like. no 120 130 6.23 0.23 0.23 Stronger. Stronger. Like. no 0.23 Like. * Originally present. (130 X o.oooi = 0.0130) Potassium found 0.0130 grm. Potassium taken 0.0140 grm. Error o.ooio grm. = 7 per cent. 88 METHODS IN CHEMICAL ANALYSIS Determination of Potassium in Presence of Sodium. Experiment IV. Part I. Part II. Part III. Volume of test solu- tion. Width of slit. Character- istic of potassium line as com- pared with NaCl in 100 cm. 3 of test solu- tion. Width of slit. Character- istic of sodium line as compared with Volume of test solu- tion. Width of slit. Character- istic of potassium line as com- pared with cm.* mm. standard. grm. mm. standard. cm.* mm. standard. First. 7rt 022 Stronger o 18 Weaker 90 100 0.23 0.23 Stronger. ( Weaker. \Like. 0.07 0.09 O.IO 0.80 0.18 0.18 Weaker. Weaker. (Like. < Stronger. ( Stronger. 100 1 2O 130 140 0.23 0.23 0.23 0.23 Stronger. Stronger. Stronger. ( Stronger. I Weaker. Second. 120 0.23 Stronger. 140 0.23 Stronger. j Stronger. % I 5 0.23 { Weaker. 1 60 0.23 Weaker. * Originally present. First. Second. (140 X o.oooi = 0.0140) (150 X o.oooi = 0.0150) Potassium found ... 0.0180 grm. 0.0150 Potassium taken. . . 0.0150 grm. 0.0150 Error o.ooio grm. = 7 per cent o.oooo Though not accurate in the highest degree when considerable amounts of potassium are to be estimated, the method is reason- ably applicable to the determination of small quantities of that element. The Separation and Determination of Potassium as the Perchlorate. Kreider's method* for the preparation of perchloric acid has greatly facilitated the use of the perchlorate method for the estimation of potassium. This method consisting essentially in the regulated heating of sodium chlorate (readily obtained in the market) , treatment of the residue with strong hydrochloric acid to yield a precipitate of sodium chloride and a solution of perchloric acid containing a small amount of sodium chloride, * D. Albert Kreider, Am. Jour. Sci., [3], xlix, 443. THE ALKALI METALS 89 and filtration of the mixture upon asbestos may be detailed as follows : A convenient quantity of sodium chlorate, from 100 to 300 grm., is melted in a glass retort or round-bottomed flask and gradually raised to a temperature at which oxygen is freely but not too rapidly evolved, and kept at this temperature for one and a half or two hours, until the thickening of the mass indi- cates the conversion of the chlorate to chloride and perchlorate; or, the retort may be connected with a gasometer and the end of the reaction determined by the volume of oxygen expelled, according to the equation 2 NaClO 3 = NaCl + NaClO 4 + O 2 . The product thus obtained is washed from the retort to a capa- cious evaporating dish, where it is treated with sufficient hydro- chloric acid to effect the complete reduction of the residual chlorate, which, if the ignition has been carefully conducted with well-distributed heat, will be present in but small amount. It is then evaporated to dry ness on the steam bath, or more quickly over a direct flame, and with but little attention until a point near to dry ness has been reached. Then stirring is found of great advantage in facilitating the volatilization of the remain- ing liquid and in breaking up the mass of salt; otherwise the perchlorate seems to solidify with a certain amount of water, and removal from the dish, without moistening and reheating, is impossible. After trituration of the residue in a porcelain mortar, an ex- cess of the strongest hydrochloric acid is added to the dry salt, preferably in a tall beaker, where there is less surface for the escape of hydrochloric acid and from which the acid may be decanted without disturbing the precipitated chloride. If the salt has been reduced to a very fine powder, by stirring ener- getically for a minute, the hydrochloric acid will set free the perchloric acid and precipitate the sodium as chloride, which in a few minutes settles, leaving a clear solution of the perchloric acid with the excess of hydrochloric acid. The clear supernatant liquid is then decanted upon asbestos in a perforated crucible, through which it may be rapidly drawn with the aid of suction, and the residue is again treated with the strongest hydrochloric acid. The liquid is again decanted, the salt is finally brought 90 METHODS IN CHEMICAL ANALYSIS upon the filter, where it is washed with a little strong hydro- chloric acid. A large platinum cone will be found more con- venient than the crucible, because of its greater capacity and filtering surface. The filtrate, containing the perchloric acid with the excess of hydrochloric acid and the small per cent of sodium chloride which is soluble in the latter, is then evaporated over the steam bath till all hydrochloric acid is expelled and the heavy white fumes of perchloric acid appear, when it is ready for use in potassium determinations. Evidently the acid is not chemically pure be- cause the sodium chloride is not absolutely insoluble in hydro- chloric acid ; but a test with silver nitrate proves that the sodium, together with any other bases which may have gone through the filter, has been completely converted into perchlorate, and, unless the original chlorate contained potassium or the acid had been exposed to the fumes of ammonia, the residue of evaporation, which does not exceed 0.04 grm. in weight to I cm., 3 is easily and completely soluble in 97 per cent alcohol. Perchloric acid thus prepared was found to contain 0.9831 grm. of free anhydrous acid in I cm. 3 . Should the sodium chlorate used in the process contain potas- sium as an impurity, the mixture of sodium perchlorate and chloride, after being treated with hydrochloric acid for the reduc- tion of the residual chlorate, is reduced to a fine powder, and well digested with 97 per cent alcohol, which dissolves the sodium perchlorate, but leaves the chloride as well as any potassium salt insoluble. The alcoholic solution of the perchlorate is then dis- tilled from a large flask until the perchlorate begins to crystallize, when the heat is removed and the contents quickly emptied into an evaporating dish. The mixture is evaporated to dryness on the steam bath and the residue is treated with strong hydro- chloric acid for the separation of the perchloric acid in the manner described above. In applying perchloric acid, prepared by Kreider's method, to the determination of potassium according to the treatment suggested by Caspari,* very satisfactory results were obtained. Briefly, the method is as follows; The substance, free from sul- phuric acid, is evaporated to the expulsion of free hydrochloric acid. The residue, stirred with 20 cm. 3 of hot water and then * Zeit. angew. Chem., 1893, 68. THE ALKALI METALS treated with perchloric acid in quantity not less than one and one-half times that required by the bases present, is evaporated with frequent stirring to a thick, sirup-like consistency, again dis- solved in hot water and evaporated with continued stirring until all hydrochloric acid has been expelled and the fumes of per- chloric acid appear. Further loss of perchloric acid is to be com- pensated for by addition of more. The cold mass is then well stirred with about 20 cm. 3 of wash alcohol 97 per cent alcohol containing 0.2 per cent by weight of pure perchloric acid with precautions against reducing the potassium perchlorate crystals to too fine a powder. After settling, the alcohol is decanted on the asbestos filter and the residue is again washed with about the same amount of wash alcohol. The residual salt, freed from alcohol by gently heating, is dissolved in 10 cm. 3 of hot water and a little perchloric acid. The solution is evaporated once more with stirring until fumes of perchloric acid rise. The precipitate is treated with I cm. 3 of wash alcohol, transferred to the asbestos, preferably by a policeman, and washed with pure alcohol ; the whole process requiring about 50 to 70 cm. 3 of alco- hol. It is then dried at about 130 C. and weighed. The substitution of a perforated crucible for the truncated pipette employed by Caspari is advantageous; and asbestos capable of forming a close, compact felt should be selected , inasmuch as the perchlorate is in part unavoidably reduced, during the necessary stirring, to so fine a condition that it tends to run through the filter when under pressure. A number of determinations made of potassium unmixed with other bases or nonvolatile acids are recorded in the following table : Potassium in the Pure Salt. KC1 taken. Volume of filtrate. KC1O 4 found. Error on, KC1O 4 . Error on KC1. Error on K,O. grm. cm. 3 grm. grm. grm. grm. 0. IOOO 54 0.1851 0.0008 0.0004 0.0003 O. IOOO 58 0.1854 0.0005 O.OOO2 O.OOO2 O. IOOO 5i 0.1859 o.oooo O . OOOO O . OOOO 0.1000 50 0.1854 0.0005 O.OOO2 O.OOO2 O.IOOO 48 0.1859 0.0000 0.0000 O.OOOO O. IOOO 52 0.1854 0.0005 0.0002 O . OOO2 As Caspari has pointed out, sulphuric acid must be removed by precipitation as barium sulphate before the treatment with 9 2 METHODS IN CHEMICAL ANALYSIS perchloric acid is attempted, and unless the precipitation is made in a strongly acid solution, some potassium is carried down with the barium. Phosphoric acid need not be previously removed; but to secure a nearly complete separation of this acid from the potassium, a considerable excess of perchloric acid should be left upon the potassium perchlorate before it is treated with the alcohol. When these conditions are carefully complied with, fairly good results may justly be expected. Below are given a number of the results obtained. Potassium in Mixtures of Salts. Compounds taken, grin. Volume of filtrate. cm. 8 KC10 4 found. grm. Error on KC1O 4 . grm. Error on KC1. grm. Error on K 2 0. grm. KCl -o.ioool CaCO 3 =o.i3 MgSO 4 = o.i 3 Fe 2 Cl 6 =0.05 ^ Al 2 {S0 4 ) 3 = o.o 5 MnO 2 =o.o5 | HNa 2 PO4.i2H 2 = o. 4 o J 50 82 80 80 92 60 0.1887 0.1875 0.1861 0.1843 0.1839 0.1854 +0.0028 +O.OOl6 +0.0002 0.0016 0.0020 O.OOO5 +0.0014 +0.0008 +O.OOOI 0.0008 O.OOIO O.OOO2 +0.0009* +0.0005* +0.000lf 0.0005! 0.0006 t O.OOO2f * The residue showed phosphoric acid plainly when tested, t Only traces of phosphoric acid found in the residue. In the last three experiments, in which the amount of perchloric acid employed was about three times that required to unite with the bases present, the phosphoric acid subsequently found with the potassium was hardly enough to appreciably affect the weight. The Estimation of Potassium as the Pyrosulphate. Browning* has shown that potassium sulphate, like sodium sulphate f when treated with sulphuric acid and submitted to a temperature ranging between 250 and 270, takes very defi- nitely the form of the pyrosulphate, K 2 S 2 O 7 , and that potassium may be estimated as that salt. Under similar conditions caesium and rubidium remain in the form of acid sulphates. Results of the procedure are given below in comparison with the results obtained upon moistening the pyrosulphate with ammonium hydroxide and igniting strongly to form the neutral sulphate. * Philip E. Browning, Am. Jour. Sci., [4], xii, 301. t See page 79. THE ALKALI METALS 93 Potassium Sulphates by Ignition. KCl taken. grin. K,S 2 0, calculated. grm. K 2 S 2 7 found. grm. Error, grm. K 2 S0 4 calculated. grm. K 2 S0 4 found. grm. Error, grm. O 2172 O 37O4. o 3608 o 0006 o . i 706 O.II92 O.IO74 o 1006 0.2909 0.2032 0.1830 o 1868 0.2886 O.2O22 0.1823 o 1860 -0.0023 O.OOIO 0.0007 o 0008 0.1993 0-1393 0.1972 0.1381 O.OO2I O.OOI2 The Volumetric Estimation of Potassium as the Cobalti-nitrite. The use of sodium cobalti-nitrite to estimate potassium has been described by R. H. Adie and T. B. Wood,* who show results fairly accurate and favorably comparable with those obtained by the platinic chloride gravimetric method. In the process worked out by these investigators a solution of a potassium salt containing the equivalent of 0.5 per cent to I per cent of K 2 O is acidified with acetic acid and precipitated by an excess of sodium cobalti-nitrite.t The mixture is allowed to stand at least a few hours, preferably over night, and is then filtered through a perforated crucible fitted with an asbestos felt. The precipitate is washed with 10 per ctent acetic acid. According to Sutton, it is important that the precipitation should be made in a solu- tion containing the equivalent of 0.5 per cent to I per cent of K 2 O, since in solutions of lower concentration the precipitate conies down in a condition in which it is apt to run through the filter in washing. The precipitate is then decomposed by boiling in dilute sodium hydroxide, and the cobalt is removed as the hydroxide by filtration. The nitrites, which are a measure of the potassium in the precipitate, are estimated by titrating with standard potassium permanganate. Adie and Wood found by analysis that the composition of the potassium salt precipitated in presence of the excess of sodium cobalti-nitrite is represented by the formula K 2 NaCo(NO 2 )6.H 2 O, and that in their method a cubic centimeter of strictly n/io potassium permanganate is equivalent to 0.000785 grm. of K 2 O. This process has been studied by DrushelJ with a view to determining the best conditions for precipitating and filtering * Jour. Chem. Soc., Ixxvii, 1076; Button's Vol. Anal., 9th ed., page 62. f Ibid. J W. A. Drushel. Am. Jour. Sci.. [4], xxfv, 433. 94 METHODS IN CHEMICAL ANALYSIS the potassium cobalti-nitrite, and to shortening the work by oxidizing the precipitated cobalti-nitrite with potassium per- manganate without the preliminary decomposition of the precipi- tate and removal of cobalt recommended by Adie and Wood, the excess of permanganate being reduced by standard oxalic acid, and the remaining oxalic acid titrated to color. In this treatment trivalent cobalt is reduced to the bivalent condition; the oxygen thus made available is equivalent to one-twelfth of that necessary to oxidize the nitrites. The factor used, there- fore, in calculating the results from the direct titration should be twelve-elevenths of that given by Adie and Wood; that is, in titrating the precipitate without first separating the cobalt one cubic centimeter of strictly n/io potassium permanganate is equivalent to 0.000857 g rm - f K 2 O. By repeated experiments it was found that difficulty in filtra- tion, as well as the necessity for allowing the precipitate to stand over night, may be avoided by evaporating the mixture nearly to dryness on the steam bath after adding the sodium cobalti- nitrite solution in considerable excess. Upon cooling the pasty residue becomes hard and dry. When treated with cold water the excess of sodium cobalti-nitrite dissolves, and the insoluble portion may be collected and freely washed without showing a tendency to pass through the filter. Potassium in the The application of the cobalti-nitrite method as Pure Salt. worked out by Drushel is as follows: The solution of a potassium salt, containing not more than 0.2 grm. K 2 O and free from ammonium salt, is treated with a rather large excess of sodium cobalti-nitrite solution acidified with acetic acid, and evaporated to a pasty condition over the steam bath. It is then cooled, treated with 50 cm. 3 to 100 cm. 3 of cold water and stirred until the excess of sodium cobalti-nitrite is dissolved, allowed to settle, and decanted through a perforated crucible fitted with an asbestos felt. The precipitate is washed two or three times by decantation, after which it is transferred to the crucible and thoroughly washed with cold water.* In the mean- time a measured excess of standard potassium permanganate is diluted to ten times its volume and heated nearly to boiling. * It was found later that a half-saturated sodium chloride solution is pref- erable to cold water for washing the precipitate, since it permits the use of a coarser asbestos felt in filtering without danger of loss, THE ALKALI METALS 95 Into this the precipitate and felt are transferred and stirred, after which the crucible is also put into the solution, since particles of the precipitate stick persistently to the sides of the crucible. After the oxidation has proceeded five or six minutes manganese hydroxide separates out and the color of the solution darkens. At this point 5 cm. 3 to 25 cm. 3 of sulphuric acid [1:7] are added, and the solution, after stirring, is allowed to stand a few minutes. Then a measured amount of standard oxalic acid, containing 50 cm. 3 of strong sulphuric acid per liter, is run in from a burette, with care to add an excess. The temperature is maintained a little below the boiling point until the solution becomes color- less and the manganese hydroxide has completely dissolved. Titration is then effected by permanganate in the usual manner. The results of the experimental tests with potassium chloride alone and in presence of salts. of the calcium group are given below. Potassium in Pure Potassium Chloride. KoO taken as KC1. grtn. K 2 O found. Error in K 2 O. Gravimetrically. grm. Volumetrically. grm. Gravimetrically . grm. Volumetrically. grm. 0.0237 0.0237 0-0354 0.0474 0.0048 0.0024 0.0005 0.0015 0.0355 0.0240 0.0243 0-0359 0.0478 0.0048 0.0024 0.0238 0.0242 0-0355 0.0471 0.0050 0.0023 o . 0006 O.OOI7 0-0355 +0 . 0003 +O.OOO6 +O.OOO4 +0.0004 0.0000 o.oooo +O.OOOI +o . 0005 o.oooo 0.0003 +O.OOO2 +O.OOOI +0.0001 +0.0002 o.oooo In the first six experiments of this series the precipitate was dried at 115, weighed, and then treated with permanganate. Potassium in Mixtures of Salts. CaCl 2 . grm. 3S: grm. BaCl 2 . taken. grm. Sr(N0 3 ) 2 . grm. K 2 O taken, grm. K 2 O found, grm. Error, grm. O 2OOO O 2OOO o 0005 o 0007 +O OOO2 o 3000 o 5000 o 0237 o 0234 o 0003 o 5000 I OOOO o 0829 o 0824 o 0005 o 5000 I .OOOO 0.5000 O.O7II o 0737 +o 0026 0.5000 o . 5000 I .OOOO I .OOOO 0.5000 0.5000 0.5000 0.0474 0.0237 0.0493 0.0251 +O.OOI9 +O.OOI4 0.5000 I .OOOO 0.07II 0.0713 +O.OOO2 9 6 METHODS IN CHEMICAL ANALYSIS The salts of calcium and magnesium 4o not influence the accuracy of the process, while the presence of salts of barium and strontium tends to high results. Potassium in The method* is applicable to the estimation of Fertilizers. potassium in fertilizers. The process, as laid down, is as follows : Ten grams of the fertilizer are placed in a 5-cm. 3 flask, 300 cm. 3 of water added, the contents boiled for 30 minutes, and ammonia water added to slight alkalinity. Enough am- monium oxalate is added to precipitate all the calcium, and, after cooling, the solution is made up to the mark on the neck of the flask and well shaken. The solution is then filtered through a dry filter into a dry flask, and 5O-cm. 3 portions of the filtrate are transferred with a pipette to platinum dishes, for estimation by the cobalti-nitrite method. After evaporating these portions to half their volume over the steam bath, I cm. 3 of sulphuric acid [i: i] is added and the evaporation is continued as far as possible over the steam bath, and finally over a low flame. After the danger of spattering is over, the flame is in- creased and the charred organic matter is burned off, finally, over the blast lamp. The potassium sulphate is dissolved by adding a little water and heating over the steam bath, and the potassium is estimated as in the previously described treatment of the pure potassium salt.f Potassium in Mixed Fertilizers. K Z O by platinum chloride method. K 2 O by cobalti- Water-soluble Number. nitrite method. PzO 5 in sample. Per cent. Per cent. Per cent. Per cent. I 5-22 S-l8 5-i8 4.16 2 6-53 6.56 6.56 3.10 3 2.23 2.24 2.24 7.82 4 8.68 8.64 8.78 o 94 5 6.37 6.42 ' 6.38 6.62 6 6.08 6.13 6.13 S-6i 7 4.08 4.02 4.02 3-iS 8 4.62 4.66 4.67 2-43 9 1.68 1.67 1.77 6.03 * W. A. Drushel, Am. Jour. Sci., [4], xxiv, 437. t See page 94. THE ALKALI METALS 97 In the preceding table are given results obtained by the method with nine fertilizers, and, for comparison, results (by two ana- lysts) by the platinic chloride method. Potassium in In applying the cobalti-nitrite method to the esti- Soiis. mation of potassium in soils, the general procedure may be outlined as follows : * A weighed amount of dry soil is extracted with an excess of hydrochloric acid over the steam bath. The excess of acid is removed from the extract by evaporation. The bases which might interfere with the process are removed with sodium car- bonate or ammonium hydroxide and ammonium oxalate. Am- monium salts and organic matter are removed by ignition. Potassium in Soils. Character of soil. Soil taken, grm. K 2 O found. Platinum chloride method. grm. Cobalti-nitrite method. grm. Per cent. Clay Cfay Loam Loam Gravel Gravel. . . . Clay gravel . . (d) ](2) I (0 2-5 2.5 2.5 2.5 2.5 2.5 2.5 2-5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 0.0028 0.0035 O. II , 0.14 0.14 0-39 0-37 0-37 0.30 0.27 0.30 0.24 0.23 0.23 0.17 0.18 o. 19 0.18 0.20 0.19 0.18 0.18 0.16 0.18 0.0035 0.0093 0.0075 0.0058 '((I).... 1(3).. O.OIOO 0.0092 d)... " 0.0074 0.0068 v (s) (d) ... o . 0060 0.0058 ( (i).. o . 0042 1(2).. 0.0045 ( (i) 0.0047 0.0044 0.0050 0.0046 {d).. 0.0048 Si.; (3).. 0.0045 o . 0040 0.0044 (4).. (s) Small amounts of phosphoric acid do not interfere. The resi- due is dissolved in a little water and a few drops of acetic acid, * W. A. Drushel, Am. Jour. Sci., [4], xxvi, 329. 98 METHODS IN CHEMICAL ANALYSIS and the mixture evaporated with an excess of sodium cobalti- ni trite to a pasty condition, stirred up with cold water, and filtered upon asbestos in a perforated crucible. The precipitated potassium sodium cobalti-nitrite is washed with a half-satu- rated solution of sodium chloride, and treated with an excess of permanganate in hot dilute solution. The color of the per- manganate is destroyed by an excess of standard acidulated oxalic acid, and the excess of oxalic acid titrated to color with permanganate. The test analyses show an excellent agreement with one another and with the results of the platinum chloride method. By using 10 grm. of soil for each estimation it should be possible to attain a higher degree of accuracy. Potassium in Drushel has also studied the application of the uri^f^oo^r cobalti-nitrite method to the determination of potas- Lymph, Milk, sium in animal fluids.* Of the constituents of urine, ammonia and the organic substances, especially urea, are the only ones which should interfere with the volumetric method as previously described. To remove these without the loss of potassium is apparently the only new problem in connection with the estimation of potassium in urine, and this is accomplished by the following procedure: Aliquot portions of urine of 10 to 50 cm. 3 each are measured with pipettes or a burette into small platinum evaporating dishes, and evaporated to dryness over the steam bath in a good draft hood. The residues are best treated by acting upon them with 5 cm. 3 to 10 cm. 3 of a 9: i nitric-sulphuric acid mixture in an evaporating dish kept covered until the first violent oxidation is over, evaporation to dryness, and ignition. By this treatment the ignition of the residue from 50 cm. 3 of urine may be readily made in 30 minutes without loss of mate- rial. The residue thus prepared is treated with a little water and a few drops of acetic acid to dissolve the alkalies, and from this point the process is carried out as in the application of the cobalti-nitrite method to the estimation of potassium in pure salts, as previously described. The results obtained in the application of this method to a number of specimens of human urine are given in the following table. * W. A. Drushel, Am. Jour. Sci., [4], xxvi, 555. THE ALKALI METALS 99 Potassium in Urine. Urine taken. cm. 3 Specific gravity. Volume in 24 hours. cm.* K found. K in 34 hours, grm. Platinum chloride method.* grin. Cobalti- nitrite method. grm. 10 10 IO IO 25 25 25 20 20 20 20 2O 2O 2O 25 50 SO 25 1.025 95 950 QIO 1130 1500 0.0293 0.0292 2. 7 8 2-77 2. 7 8 2-77 2.81 2.84 2.81 3-44 3-42 3-47 3-74 3-74 3-74 3-74 2-55 2.52 2-53 2-54 1.025 0.0293 0.0292 0.0740 o . 0740 0.0757 0.0764 o . 0663 0.0662 0.0747 1.025 1.024 0.0752 o . 0663 0.0662 I.OI8 0.0425 0.0839 0.0843 0.0424 Modified Lindo-Gladding method, after removal of P,O S An additional difficulty presents itself in the presence of a large amount of protein material which cannot be removed by coagulation and nitration without a considerable loss of potassium. This is particularly true of the blood, where m<5st of the potassium is intimately associated with the protein of the corpuscles. It is necessary therefore to decompose protein material by oxidation. For this purpose the nitric- sulphuric acid mixture works less satisfactorily than treatment with concentrated nitric acid, digestion on* the steam bath, dilution, treatment of aliquot portions by evaporation, gentle ignition, addition of sulphuric acid and a final ignition; or liquid bromine may be substituted for nitric acid in the first oxidation. Results obtained for potassium in circulating fluids are given in the following table. 100 METHODS IN CHEMICAL ANALYSIS Potassium in Blood and Lymph. K 2 O found. Nature of fluid. Amount taken. grm. Platinum chloride method.* grm. Cobalti- nitrite method. grm. Per cent. (-10.89 0.0227 0.0227 0.21 1 II . 21 o . 0228 o 20 Defibrinated pig's bloodf. j 20.33 0.0391 0.19 j 10. 16 0.0203 0. 21 1 10-85 O.O2II o . 20 L 11.03 0.0236 O. 21 {30.00 0.0174 O.OI74 0.058 30 oo O OI 7O o 060 Sheep's blood} o ^^ Q w w* / y o . 0181 _ /: 30.00 O . OIoI O . OOO 30.00 0.0181 0.060 30.00 0.0180 0.060 Serum of dog's blood} ( IO.II < 10.04 O.OO24 0.0024 0.0024 0.024 0.024 ( 10.07 0.0023 o . 034 f 10.28 O.OOlS 0.0018 0.018 10.01 0.0019 0.019 Dog's lymph} J IO OO O.OO2O O.O20 j 10.03 O.OOIQ O.OI9 IO. 12 O.OOI9 O.OI9 [10.32 0.0022 O.OO22 O.O2I * Modified Lindo- Gladding method, after removal of Ca, Fe, and P 2 O 6 . t Oxidized by bromine. J Oxidized by nitric acid. In the estimation of potassium in milk suitable amounts are evaporated to dryness, oxidized with concentrated nitric acid, again evaporated to dryness, and ignited gently until nearly all organic matter has been burnt. The residue is mois- tened with concentrated sulphuric acid and again ignited. In the residue thus obtained the potassium may be estimated by the cobalti-nitrite method as applied to pure salts of po- tassium. Results obtained by the method in the analysis of cow's milk are given on the following page. The modifications introduced by Drushel into the cobalti- nitrite process evaporation nearly to dryness and oxidation of the nitrite without previous removal of the cobalt add THE ALKALI METALS Tot greatly to its usefulness. The necessity of long standing is avoided, the precipitate may be filtered and washed without trouble and the manipulation previous to titration is much shortened. Potassium in Milk. K 2 O found. Milk taken. Platinum Cobalti- chloride method.* nitrite method. Per cent. grm. grm. grm. 25-8 0.0413 o. 16 25.8 0.0432 0.17 2 5 .8 0.0428 0.17 51-6 0.0833 0.16 25-7 o 0454 0.18 25-7 0.0457 0.18 2< 7 o 04x1 0.18 * Lindo-Gladding method, after removal of Ca and P2O 6 . For small amounts of potassium fairly accurate results are obtained by using the permanganate factor calculated from Adie and Wood's formula for potassium sodium cobalti-nitrite. Sutton has suggested that more accurate results may be secured by obtaining a factor empirically from a pure potassium salt. The results recorded above were obtained, however, by using the theoretical factor calculated from the formula of Adie and Wood, K2NaCo(NO2)6.HoO, their analyses of potassium sodium cobalti- nitrite having been verified* by the analysis of a carefully pre- pared salt. The chief sources of error in the method appear to be the slight solubility of the potassium sodium cobalti-nitrite, one part in 25,000 to 30,000 parts of water at room temperature, and the tendency of the precipitate to include traces of sodium cobalti- nitrite. The method requires less time and labor than the chloroplati- nate method, and is applicable in the presence of substances which form no insoluble cobalti-nitrites and which neither oxidize oxalic acid nor reduce potassium permanganate. * Am. Jour. Sci., [4], xxvi, 562. '102 ; MEfHODS'IN CHEMICAL ANALYSIS RUBIDIUM AND CAESIUM. The Spectroscopic Determination of Rubidium. The method of manipulation previously described for the Spectroscopic determination of small amounts of potassium has been adapted by Gooch and Phinney* to the similar determina- tion of rubidium. In the work upon potassium the observations of the red line were made in the ordinary laboratory in diffused light, but pre- liminary experimentation upon the rubidium spectrum immedi- ately developed the fact that the blue lines are better to work by in the case of this element, and that a dark room becomes a necessity. For the experiments described pure rubidium chloride was prepared by many fractional precipitations by alcohol out of aqueous solutions, and in settling the question as to the coils which should be used the choice fell upon the size holding 0.02 grm. of water and made of the No. 28 wire, the superior stiffness of these and consequent constancy in capacity giving them the advantage over smaller coils of finer wire, though the latter are capable of bringing out greater sensitiveness of the reaction. It was found, for example, that under the most favorable conditions as to height of flame and width of slit, 0.0002 mg. of rubidium chloride produced the blue lines at the last limit of visibility when the larger and heavier coil was in the flame; with a coil holding 0.006 grm. of water and made of very fine wire the more immediate volatilization of the chloride so increased the delicacy of the Spectroscopic reaction that it was possible to see the lines from 0.00005 mg. of the salt. These figures serve as an indica- tion of the possible delicacy of this method of producing spectra, but it should be remembered that all eyes do not see the rubidium lines with equal ease. In comparative tests of brightness it was found best to employ as the standard the lines given by amounts of the chloride not exceeding 0.0005 m g- to 0.0007 m g-> to set the slit at a width of 0.2 mm. and to bring the coils to the flame in sets of three the first, usually a standard, serving to fix the position of the lines so that the comparative distinctness of the lines given by the other two might be the more readily determined. When that dilution had been found at which the test was barely brighter * F. A. Gooch and J. I. Phinney, Am. Jour. Sci., [3], xliv, 392. THE ALKALI METALS than the standard and that dilution at which the test was barely weaker than the standard, it was assumed that the mean of the numbers of cubic centimeters representing these two volumes might be taken as the volume at which the test and standard lines were equal. The amount of rubidium in the test solution was then calculated by multiplying the volume in cubic centi- meters by the number of coilfuls in I cm. 3 and the product by the amount of rubidium contained in a coilful of the stand- ard solution. The results of two experiments with pure rubidium chloride are given below. Determination of Rubidium in Pure Rubidium Chloride. Experiment I. Standard. Rubidium in a coilful (A cm.*). Test (known to contain 10 mg. Rb). Volume in cm. 8 Line of test compared with standard. mg. 0.0005 340 Brighter. 0.0005 370 Equally bright. 0.0005 370 Brighter. 0.0005 39 Weaker. 0.0005 390 Weaker. Found, 37 * 39 X 50 X 0.0005 = 9-5 nig. Taken 10.0 mg. Error 0.5 mg. = 5 per cent. Experiment II. Standard. Rubidium in a coilful (^j cm. 3 ). Test (known to contain 10 mg. Rb). Volume in cm. 3 Line of test compared with standard. mg. 0.0005 300 Brighter. 0.0005 360 Equally bright. 0.0005 380 Brighter. 0.0005 380 ' Brighter. 0.0005 390 Brighter. 0.0005 400 Weaker. 0.0005 410 Weaker. Found, 39 + 4 X 50 X 0.0005 = 9. 875 rag. Taken , 10 o Error o. 125 mg. = i . 25 per cent. 104 METHODS IN CHEMICAL ANALYSIS These results make it plain that when the comparison is made between solutions of pure rubidium chloride the spectroscopic method is capable of yielding fair approximations to truth. In the practical determination of rubidium, however, the ques- tion of the effect of the presence of sodium and potassium which naturally accompany it is of importance. It appears from practical tests that within limits the presence of sodium in the flame increases the brilliance of the rubidium spectrum. The brightness of the lines is raised under the con- ditions by a maximum of 50 per cent by the presence of sodium up to 40 per cent of the weight of the rubidium, and increase in the amount of sodium does not further influence the bright- ness of the lines until the proportion of sodium to rubidium is as ten to one; or, speaking broadly, the dissociating effect of sodium upon the rubidium chloride (to which the brightening noted is to be attributed) does not appear to be materially different whether one or a score of molecules of sodium chlo- ride are present to one of the rubidium chloride. But when the proportion of sodium to rubidium much exceeds ten to one the glare of light diffused through the entire spectrum (though the sodium line itself may be cut off) begins to affect the vision, and as the increase advances ultimately extinguishes the rubidium lines. It appears also that potassium chloride produces an effect simi- lar to that of sodium chloride, the brightness of the rubidium line increasing by a maximum of 50 per cent when the potassium is present to between two-thirds and twice the amount of the rubid- ium; while the presence of potassium in the proportion five to one influences the visibility unfavorably, and in the proportion of thirty to one extinguishes the rubidium line in the glare of light. It is necessary therefore either to effect the separation of the rubidium from sodium and potassium, or else to bring test and standard to the same condition as regards the presence of these elements, before any reasonable degree of accuracy can be expected in the spectroscopic determination of rubidium as it ordinarily occurs in nature. The separation from sodium is easily accomplished by the conversion of the salts to the form of chloro- platinates; but for the quantitative separation of rubidium from potassium there is no good method known. The practical value of the spectroscopic reaction of rubidium for purposes of quanti- THE ALKALI METALS 105 tative analysis depends, therefore, upon matching potassium lines as well as the rubidium lines (following the method outlined in the determination of potassium in presence of sodium), and so bringing the lines of test and standard equally under the influence of potassium. It has been shown that there is no difficulty in matching solutions of potassium by means of the red line, but the convenience of using the spectroscope without readjust- ment throughout an entire experiment makes a comparison by means of the blue line highly desirable and this has been found to be feasible. The details of a determination of rubid- ium in presence of a permissible amount of potassium are given in the following statement. Determination of Rubidium in Presence of Potassium. Standard solution containing (&cm.*), Test solution containing 8 mg. rubidium and no potassium. to the coilful Step i. Step 2. Step 3. Step 4. Step 5. Preliminary test for Rb. Preliminary matching of K line. Rematching of Rb line. Readjustment of K line. Final matching of Rb line. Test at 20 cm. 1 gave Rb line like standard. Test at 20 cm. 8 gave K line like standard when i mg. of K had been added. Test at 35 cm.* gave Rb lines like standard. Test at 35 cm.* gave K line like standard when 2 mg. were pres- ent. Test at 35 cm. gave Rb line like standard. Found, 35X50X0.0005 =0.875 mg. Taken =0.8 mg. Error =0.075 mg. = g.4 per cent. When the amount of potassium present is so great as to vitiate the test for rubidium a precipitation by alcohol may be utilized to remove the excessive amount of the potassium salt. The mixed chlorides are dissolved in the least possible quantity of water and treated with absolute alcohol; the precipitate is filtered off and washed with alcohol; the filtrate and washings are evaporated and the residue dissolved in a known volume of water is ready for the spectroscopic test. Results of experi- ments conducted in this manner follow. io6 METHODS IN CHEMICAL ANALYSIS Rubidium taken in the form of chloride. Potassium taken in the form of chloride. Rubidium found. Absolute error. Percentage error. mg. grm. mg. mg. per cent. I O.I 0.8 O.2 20 2 O.I i-7 o-3 15 I O. I 0.9 O.I IO The error of the process is manifestly large, and only roughly approximate results can be hoped for when large amounts of rubidium are dealt with; but, in view of the fact that the only alternative is an indirect process, even this great error may not be prohibitive in the estimation of very small amounts of rubidium.* The Estimation of Caesium and Rubidium as the Acid Sulphates. Browning f has shown that by holding the temperature between 250 and 270 during treatment with sulphuric acid suitable salts of caesium and rubidium may be brought with a fair degree of certainty to the condition of the acid sulphates, which by treat- ment with ammonia and ignition at red heat yield the neutral sulphates. Salts of potassium and sodium, however, when heated at the same range of temperature, yield pyrosulphates reasonably stable under the conditions.! Lithium salts when treated simi- larly gave no evidence of the existence of a stable acid sulphate or pyrosulphate. Details of the experiments with the salts of rubidium and caesium are given below. A weighed amount of caesium nitrate was placed in a pre- viously weighed platinum crucible and treated with an excess of sulphuric acid. The crucible was then placed upon a steam bath until the water and nitric acid were largely expelled, and then removed to a radiator, consisting of a porcelain crucible fitted with a pipe-stem triangle so arranged that the bottom of the platinum crucible was about midway between the top and bottom of the porcelain crucible. This improvised radiator was set in an iron ring and a thermometer placed so that the mer- cury bulb was on a level with the bottom and close to the side of the platinum crucible. An ordinary Bunsen burner * For an example of the practical use of this method, see The Excretion of Rubidium, Mendel and Slosson, Am. Jour. Physiol., xvi, 152. t Philip E. Browning, Am. Jour. Sci., [4], xii, 301. t See pages 79, 92. THE ALKALI METALS 107 served as the source of heat and the temperature was kept so far as possible between 250 C. and 270 C. After the fuming of sulphuric acid had ceased, the crucible and contents were removed to a desiccator, cooled and weighed. This process of heating was continued for half-hour periods until the weights were constant. The results given show that by regulating the heat at a temperature between 250 C. and 270 C. caesium may Caesium Sulphates by Ignition. CsNO, taken. grm. CsHSO 4 calculated. grm. First constant weight. grm. Second constant weight. grm. Error on CsHSO 4 . grm. Cs 2 S0 4 calculated. grm. Cs 2 S0 4 found. grm. Error on Cs 2 S0 4 . grm. O2OI 3 O 2O?4 O 2O2O -j-o 0007 O I7o6 o 2013 O 2OIO o 0003 OIO32 o 1217 O 1 2OI o 0016 0.1032 0.1218 o 1214 O.I2I7 0-1437 o 143? 0.1252 0.1458 o 1430 0.1222 +0.0005 +0.0021 o 0005 0.0961 0.1130 o . 0948 o. 1118 0.0013 O.OOI2 o 1214 O 143? o 1422 o 0013 OT T CQ o 1 3 ?6 O I 33O o 0026 O IO?6 O 1 24? o 1272 o 1248 +o 0003 o 1056 O 1 24.? o 1 252 +o 0007 be brought with a fair degree of certainty to the condition of the acid sulphate. As a check upon the results the acid sulphate was, in a few cases, treated with a little ammonium hydroxide, the excess of this was removed upon a steam bath and the neutral sulphate was obtained by ignition at a red heat to a constant weight. These determinations agree fairly well with the theory. The same procedure was followed with rubidium, a pure ru- bidium chloride having been chosen as the starting point. The results are given below. No tendency was observed on the part of these elements to hold sulphuric acid in excess of the amount necessary for the formation of the acid sulphate. Rubidium Sulphates by Ignition. RbCl taken. grm. RbHSO 4 calculated. grm. RbHSO 4 found. grm. Error, grm. Rb 2 S0 4 calculated. grm. Rb 2 SO 4 found. grin. Error* grm. o. 1252 o 1889 o 1878 O OOII O.I2I2 O.I23O o. 1829 o 1856 o. 1840 o 1850 +O.OOII o 0006 0.1460 0.1460 6.OOOO 0.1230 0.1610 0.1360 0.1856 0.2430 0.2052 0.1858 0.2416 0.2032 +O.OOO2 O.OOI4 O.OO2O 0.1357 0.1777 0.1501 0.1350 0.1772 0.1490 0.0067 0.0005 O.OOII CHAPTER III. COPPER; SILVER; GOLD. COPPER. The Gravimetric Determination of Copper as the Sulphocyanate. As early as 1854 attention was drawn by Rivot* to the pos- sibility of estimating copper gravimetrically by weighing as cuprous sulphocyanate, and to the advantages which the process afforded in separating copper from other metals. Rivot's pro- cedure consisted in dissolving the substance to be analyzed in hydrochloric acid, reducing the copper with hypophosphorous or sulphurous acid, and precipitating with potassium sulpho- cyanate. The precipitate, dried at a moderate temperature, was weighed as cuprous sulphocyanate and then as a control converted by ignition with sulphur into cuprous sulphide and weighed in that condition. In spite of the evident advantages for certain purposes, Rivot's method, in its original form, has never come into general use, the chief reason for this being apparently the difficulty and inaccuracy attendant upon the weighing of the precipitate upon dried paper filters, a process which can hardly be depended upon unless managed with extreme care. Van Namef has shown, however, that the process is accurate and easily managed if attention is given to the necessary con- ditions of concentration and acidity, and the precipitated cuprous sulphocyanate is filtered and weighed upon asbestos in the per- forated crucible. The table contains results obtained as follows: A suitable quantity of a standard copper sulphate solution was run from a burette, diluted to a convenient volume, a few cubic centimeters of a concentrated solution of ammonium bisulphite! added, and the copper precipitated by an excess of ammonium sulpho- * Compt. rend., xxxviii, 868. f R. G. Van Name, Am. Jour. Sci., [4], x, 451; xiii, 20. % Prepared by saturating strong aqueous ammonia with sulphur dioxide. 108 COPPER; SILVER; GOLD 109 cyanate. After allowing the mixture to stand for a few minutes or hours, according to the amount of free acid present, the pre- cipitate was collected upon asbestos in a weighed crucible, washed with cold water and dried at 1 10 until no further loss of weight took place. Copper weighed as Cuprous Sulphocyanate. Cu taken. H 2 S0 4 cone. HNH 4 SO 3 sat. sol. NH 4 SCN approx. n/io. Final volume. Time of standing. Cu found. Error. grin* cm. 1 cm. 8 cm. 8 cm. 8 hours. gnu . grm. 0.0795 None. 5 13 68 1 0.0795 0.0000 0.0795 None. 3 13 66 48 0.0793 O.OOO2 0.0795 None. 3 25 78 0.0796 +O.OOOI 0.0795 None. 3 25 78 12 0.0796 +O.OOOI 0.0795 i-5 10 13 85 12 0.0792 0.0003 0.0795 i-5 8 13 105 48 0.0785 o.ooio 0.0795 i-5 3 25 85 4 0.0783 0.0012 0.0795 i-5 5 25 85 21 0.0795 0.0000 0.0795 S 5 25 85 3 0.0797 +O.OOO2 HCl cone. cm. 1 0.0795 IO 5 25 IOO 20 0.0795 0.0000 0.0795 25 IO 25 IOO 28 0.0784 O.OOII Larger amounts of copper may also be estimated in the same way, as the following table shows, but with a crucible of the ordinary size the process is more rapid and convenient when a smaller weight of copper is taken. Copper weighed as Cuprous Sulphocyanate. Cu taken. HjSO 4 cone. NH 4 SCN approx. w/io. Final volume. Cu found. Error. gnu. cm. s cm. 8 cm. 8 gnu . grm. 0.3175 0.3I7S 0.3175 None. None. None. 60 60 60 500 500 500 0.3176 0.3177 0.3176 -f-o.oooi +O.OOO2 + O.OOOI 0.3175 10 IOO 500 0.3175 . o.oooo HCl cone. cm. 8 0.3175 20 IOO 500 0.3165 o.ooio In solutions containing free acid the precipitation of the cop- per is greatly retarded, and the mixture should be allowed to stand for several hours, or, if the amount of acid is considerable, for at least twenty-four hours before filtering. Precipitation from 110 METHODS IN CHEMICAL ANALYSIS a warm solution is permissible, but boiling the liquid causes the precipitate to turn brown with gradual loss in weight, and is therefore to be avoided. The only difficulty which is likely to be encountered in the use of this method is a tendency, which sometimes appears, for traces of the precipitate to pass through the filter during the last stages of the washing. This tendency is most marked with precipitates from concentrated solutions containing little or no free acid. It may be reduced to an insignificant amount or entirely eliminated by employing one or more of the following expedients: (i) pre- cipitating in dilute solution ; (2) precipitating in the presence of free acid ; (3) filtering and washing under light pressure, using a rather dense (not thick) asbestos mat; (4) washing with a decinormal solution of ammonium sulphocyanate. The last expedient is of little or no use when the precipitate is to be directly weighed, but is very satisfactory in separating copper from other substances. Precipitation in acid solution is the most effective method of obtaining precipitates which are easily filtered, but must be used with caution, for the errors from incom- plete precipitation may easily exceed the mechanical losses which the acidity was employed to prevent. The effect of hydrochloric acid of various concentrations upon the completeness of the precipitation was studied in a series of experiments,* in which the principal stress was laid upon deter- mining the amounts of copper left in solution rather than the weights of the precipitates. The procedure was as follows: After filtering off the precipitate the copper in the filtrate was determined by evaporating the solution with nitric acid to a small bulk, heating in a platinum crucible over a radiator to expel sulphuric acid and decompose interfering substances, dis- solving the residue in nitric acid, filtering, electrolyzing and weighing the copper. The electrolytic deposit was then redis- solved and the copper estimated more accurately by a colori- metric method based on comparison of the intensity of the brown color produced upon the sample by potassium ferrocyanide with that of a variable standard of known copper content. The results of this series of experiments may be summarized as follows : Allowance must be made for the amount of hydro- chloric acid used up by interaction with the ammonium bisulphite * Am. Jour. Sci., [4], xiii, 20. COPPER; SILVER; GOLD III solution forming ammonium chloride and sulphur dioxide. One cubic centimeter of a bisulphate solution, prepared by saturating strong aqueous ammonia with sulphur dioxide, may neutralize by this reaction about nine-tenths of a cubic centimeter of hydro- chloric acid, sp. gr. 1.18. In order that the amount of copper left in solution may not exceed o.i mg. per 100 cm. 3 of nitrate, the concentration of effective hydrochloric acid, i.e., that remain- ing after interaction with the bisulphite, stated in volume per cent of the concentrated acid (cubic centimeters of acid of sp. gr. 1.18 for 100 cm. 3 final volume of solution), should not be above 0.8 when' the excess of sulphocyanate employed is small (20 per cent above the theory), but may be as high as 3 per cent if ten times the theoretical amount of sulphocyanate be employed. A suitable degree of acidity for precipitating copper under ordinary conditions is given by 0.5 to i.o per cent of hydrochloric acid, expressed as above, using from five to ten times the theoretical amount of sulphocyanate. As far as could be judged from a limited number of determina- tions made in the presence of sulphuric acid, the above holds true for the equivalent amount of sulphuric acid. Cu,(SCN), taken. Volume of liquid. cm. 3 HC1 (sp. gr. 1.18). cm. NH 4 SCN. grni. Cu in filtrate, grm. 0-3 2OO 0.06035 0-3 2OO 0.00040 0.25 2OO 2OO IO. 1.52* 0.0050 0.00018 o-3 2OO . 0.767 0.00007 o-3 2OO o. 19! 0.00004 0-3 2OO 2 .... 0.0019 0-3 0-3 2OO 200 200 2 2 2 2.5 1-77** 0.19 0.0013 0.0009 0.0006 NH 4 C1. grm. 0-3 200 IO 0.0031 0-3 0-3 200 200 IO ft 0.19 0.00013 0.00045 0-3 2OO I 0.19 0.00005 * Solution w/io In respect to NH 4 SCN. t Solution w/2o in respect to NH 4 SCN. J Solution w/8o in respect to NH 4 SCN. ** HC1 and NH 4 SCN present in equivalent amounts. ft Solution approximately w/io in respect to NH 4 C1. 112 METHODS IN CHEMICAL ANALYSIS The effect of varying concentrations of different reagents in- volved in the process upon the solubility of cuprous sulpho- cyanate is shown in a rough way by the preceding table. As no stirring was employed the figures have no absolute value, but serve merely to give an idea of the relative magnitude of the solubilities in question. Weighed amounts of cuprous sulphocyanate prepared by pre- cipitation in the usual way, thoroughly washed, and dried at 105, were allowed to stand in the solutions to be tested from 40 to 50 hours. After careful filtering through asbestos the copper in the clear filtrate was estimated by electrolysis, or, in cases where the amount was small, by the colorimetric method referred to above. The solubility in presence of either hydrochloric acid, ammo- nium chloride or a large amount of ammonium sulphocyanate is considerable. It is lowest in dilute solutions of ammonium sulphocyanate, and the presence of a small amount of this salt lessens the solubility in hydrochloric acid, and in solutions of ammonium chloride. Separation of From the nature of the process it is evident that Copper from j t j s mucn i ess likely to be interfered with by the Bismuth, Anti- J . . mony, Tin and presence of other metals than the other gravimetric Arsenic. methods for copper, and may, therefore, be directly applied with good results in many cases where the use of the electrolytic or the oxide method would involve a previous sepa- ration. Van Name* has tested the method for the separation of copper from bismuth, antimony, tin and arsenic. Having copper present with these metals in a solution con- taining free hydrochloric acid, tartaric acid was added to aid in preventing the formation of insoluble products of hydrolytic action, and the copper then precipitated as cuprous sulphocya- nate in the usual way. All the determinations were allowed to stand fifteen hours or more before filtering to insure completeness of precipitation. The filtering was performed upon asbestos in a perforated crucible. The precipitate was thoroughly washed with cold water and dried at 105 to a constant weight. In the following table are results obtained by this procedure. The acidity was kept within the limits shown above to be safe, * Am. Jour. Sci., [4], xiii, 138. COPPER; SILVER; GOLD and the amount of sulphocyanate used was in most cases about ten times the theory. Copper in Presence of Bismuth, Tin, Antimony, and Arsenic. Final Volume 200 cm. 3 . Cu. taken. grm. Bi. grm. HC1 (sp. gr. about I.I7). cm. Tartar! c acid. grm. HNH 4 SO 3 sat. sol. cm. 1 NH 4 SCN approx. W/IO. cm. s Cu r (SCN) 2 found. grm. Calcu- lated asCu. grm. Error, grm. 0.0793 O.2 6 2 60 0.1504 0.0786 0.0007 0.0793 O.I 6 2 IOO 0.1512 o . 0790 -0.0003 0.0793 0-3 6 2 125 O.I5I5 0.0792 o.oooi 0.0793 O.2 6 2 125 0.1518 0.0793 o.oooo 0.0793 O.2 6 2 125 0.1519 0.0794 +O.OOOI 0.0793* 0.2 6 2 230 0.1519 0.0794 +O.OOOI Sn taken as SnCU+HCl. grm. 0.0793 O.2 5 I 2 40 0.1502 0.0785 0.0008 0.0793 O.2 6 I 2 125 0.1514 0.0791 O.OOO2 0.0793 O.2 5 I 2 130 0.1516 0.0792 o.oooi Taken as SnCl,+HCl. grm. 0.0793 0.2 6 I 2 125 0.1529 0.0799 +0.0006 As. grm. 0.0793 O. 2 6 I 2 125 0.1523 0.0796 +0.0003 Sb. grm. 0.0793 O.2 6 2 2 125 0.1518 0.0793 0.0000 As, Bi, Sb, Sn of each. grm. 0.0795 O. I 6 2 2 130 0.1523 0.0796 +0.0001 0.0795 O.I 6 2 2 130 0.1525 0.0797 +0.0002 * Final volume 300 cm. * If bismuth is present in considerable amount, a good deal of hydrochloric acid is needed, and there is danger that interaction with the precipitants may reduce the acidity to the point where hydrolysis and precipitation of the bismuth begins. In such cases preliminary blank tests must be carried out to determine the minimum concentration of hydrochloric acid which may be employed under the conditions. With antimony the effective- ness of the tartaric acid is so great that this difficulty does not arise if enough tartaric acid is used. Tin in the stannous con- dition sometimes forms a slight precipitate of sulphur on stand- 114 METHODS IN CHEMICAL ANALYSIS ing in contact with the bisulphite, and it is, therefore, advisable to oxidize it at the outset to the stannic state. It is evidently possible to estimate copper by this method in the presence of bismuth, antimony, tin and arsenic, either separately or in any combination. To separate copper from unknown quantities of bismuth, or from mixtures containing bismuth, the following procedure is recommended: Having the copper and bismuth in hydrochloric acid solution, add tartaric acid, and, after diluting if necessary, determine by blank tests with small aliquot portions of the solution how much ammonium bisulphite can be added to the whole without precipitating the bismuth. Then, keeping the bisulphite well within this limit, carry out the precipitation of the copper as already described, using a considerable excess of ammonium sulphocyanate. Where bismuth is absent, antimony and tin may be treated in the same way, but the latitude possible in the adjustment of the conditions is so much greater with these metals that preliminary tests will seldom be needed. For the separation from arsenic no special precautions are required. The Determination of Copper as Cuprous Iodide and Separation from Cadmium. The separation of copper from cadmium by the precipitation of the cuprous iodide by appropriate means has long been known. Pisani* mentions the fact that potassium iodide can be used to effect precipitations, and claims that a satisfactory separation can be made in this way. Flajolotf, stating that potassium iodide cannot be used as a precipitant on account of the solu- bility of cuprous iodide in that reagent, and that hydriodic acid cannot be employed if nitric acid is present, recommends that the solution containing copper be brought to acidity with sul- phuric acid, that a considerable excess of sulphurous acid be added, and that the precipitation be effected by hydriodic acid. Kohnerf reviews the various methods for the separation of copper from cadmium and states that the iodide method is impracticable on account of the solubility of cuprous iodide both in excess of hydriodic acid and in potassium iodide. * Compt. rend., xlvii, 294. t Jour, prakt. Chem., Ixi, 105. J Zeit. anal. Chem., xxvii, 203; Jour. Anal. Chem., iii, 339. COPPER; SILVER; GOLD Browning* has fixed conditions under which accurate results may be obtained by this method. According to the best procedure shown, the sulphates of copper and cadmium, in amount not exceeding 0.25 grm. of each metal, are dissolved in 25 cm. 3 of water and treated with I grm. or 2 grm. of potassium iodide, f The mixture is evaporated to dry- ness to expel iodine and then treated with 100 cm. 3 of water. Filtration is made under gentle pressure upon the asbestos felt in the perforated crucible. It is advisable, on account of the tendency of cuprous iodide to pass through the filter, to use a fairly thick felt and to keep it moist and under pressure during the filtration. The precipitate is washed thoroughly with either hot or cold water, dried at I2O-I5O and weighed as cuprous iodide. Copper Weighed as Cuprous Iodide. Copper taken, grm. Copper found (weighed as C\i 2 I 2 ) grm. Error on copper, grm. KI used, grm. Final volume of liquid. cm. 1 0.1194 0.1196 +O.OOO2 I IOO 0.1191 o. 1194 +0.0003 I IOO 0.1193 O.II93 O.OOOO 2 IOO 0.0049 0.0045 O.OO04 2 IOO 0.0051 0.0047 O.0004 2 IOO 0.1195 O.II95 0.0000 3 IOO 0.1192 0.1188 0.0004 4 IOO Copper Separated from Cadmium and Weighed as Cuprous Iodide. Copper taken. Copper found (weighed as Cu 2 I 2 ). Error on copper. Cadmium taken. Cadmium found (weighed as CdO). Error on cadmium. .:'! KI used. Found volume of liquid. grm. grm. grm. grm. grm. grm. grm. cm. 0.2383 0.2386 +0.0003 O . 0484 o . 0490 +0.0006 2 IOO o. 1192 o. 1185 0.0007 0.2439 0.2430 +O.OOOI 2 IOO 0.1193 0.1194 +O.OOOI 0.1942 0.1942 O.OOOO 2 IOO O. I2OI O. I2OI o . oooo 0.2426 o. 2428 +O.OOO2 2 IOO O.II93 0.1193 O.OOOO 0.2436 0.2433 0.0003 2 IOO 0.0239 0.0238 O.OOOI 0.1934 0.1932 O.OOO2 I IOO 0.0236 0.0239 +0.0003 o. 1942 0.1936 0.0006 I IOO 0.0239 0.0242 +0.0003 0.1444 o. 1442 0.0002 I IOO 0.0238 0.0238 0.0000 0.1467 0.1461 O.OOO6 I IOO 1 * Philip E. Browning, Am. Jour. Sc., [3], xlvi, 280. t For the results of study of the effect of excess of potassium iodide, free acid and concentration upon the solubility of cuprous iodide, see page 121. Il6 METHODS IN CHEMICAL ANALYSIS . The table contains results of experiments made according to this procedure with pure copper sulphate, and with copper sul- phate and cadmium sulphate in mixture. The Electrolytic Determination of Copper. Gooch and Medway* have applied the rotating cathode to the rapid electrolytic determination of copper, making use of the apparatus shown in Fig. I3.f The deposition of copper from a solution of the sulphate was first attempted, and the procedure was as follows: The solution, 50 cm. 3 in volume, was placed in a i5O-cm. 3 beaker and acidulated to give better conductivity. The stand carrying the beaker was raised until the liquid covered about two-thirds of the crucible adjusted to the shaft, thus giving a cathode surface of about 30 cm. 2 The anode was introduced and the motor started. The wires from the storage batteries were connected and the cur- rent was allowed to pass through the solution. The duration of the electrolysis was varied according to the strength of current used; but in each case, after the deposit was nearly complete, the current from the batteries was shut off, the motor stopped, and the sides of the beaker, the platinum anode and the crucible were carefully washed with a fine jet of water, the motor was again started and the current allowed to pass for the remaining time. When the deposit was complete the crucible was removed and washed, first with water, then with alcohol, and finally was dried by passing it over a flame. Sulphuric acid (6 or 7 drops of the dilute acid 1:3) was generally used to acidulate the solution, since it was found that the copper was deposited in less time with sulphuric acid than with nitric acid present. Experiments in which small amounts of nitric acid (6 to 9 drops of the dilute acid 1:3) were used show that the copper may also be deposited completely in pres- ence of this acid. The following tables show the results of a series of experiments made as described. The standard of the solution of copper sul- phate was fixed by the usual slow method of electrolytic analysis. * F. A. Gooch and H. E. Medway, Am. Jour. Sci., [4], xv, 320. t See page 12. COPPER; SILVER; GOLD Solution of CuSOt Acidulated with Copper taken, grm. Copper found, grm. Error, grm. Current. amp. N. D. 100 Time, min. 0.0651 0.0652 +0.0001 0.8 2.7 25 0.0651 0.0652 +0.0001 0.8 2.7 15 0.0651 0.0651 0.0000 i 3-3 10 0.0651 o . 0649 O.OOO2 i 3-3 IO 0.0651 o . 0648 0.0003 i 3-3 IO o. 1272 0.1272 o.oooo 2-5 8-3 15 o. 1272 0.1271 O.OOOI 2-5 8-3 15 0.1272 o. 1271 O.OOOI 2-5 8-3 IS 0.1272 0.1270 O.OOO2 3 10 13 o. 1272 0.1268 0.0004 3 10 12 0.2548 0.2548 0.0000 3 IO 2O 0.2548 0.2548 0.0000 4 13-3 2O 0.2548 0.2550 +O.OOO2 4 13.3 2O 0.2548 0.2546 O.OOO2 4 !3-3 15 0.2548 0.2545 0.0003 4 13-3 15 Solution of CuSO^ Acidulated with HN0 3 . Copper taken. grm. Copper found, grm. Error, grm. Current, amp. N. D. m Time, min. 0.0651 . 0648 0.0003 I 3-3 35 0.0651 0.0652 +O.OOOI 0.8 2-7 30 0.0651 o . 0650 O.OOOI 0.8 2-7 2 S 0.0651 o . 0649 O.OOO2 i 3-3 25 0.0651 0.0650 O.OOOI 0.8 2-7 25 0.0651 0.0652 +O.OOOI i 3-3 35 0.0651 o . 0648 0.0003 i 3-3 30 0.0651 o . 0650 O.OOOI i-5 5 25 0.0651 0.0650 O.OOOI !-5 5 25 0.0651 0.0647 0.0004 1.8 6 20 It has been shown also by Medway* that a crucible of silver may be substituted for the platinum crucible as the rotating cathode, in the deposition of copper from the acidulated sulphate solution, with results that leave little to be desired on the score of accuracy. Deposition upon the Silver Crucible. Copper taken, grm. Copper found, gnu. Error, grm. Current, amp. N. D. m Time, min. 0.1088 0.1086 0.0002 2 6.6 15 o.ioss o. 1090 +0.0002 2 6.6 15 0.1088 0.1084 0.0004 1-5 5 15 0.1088 0.1085 0.0003 2 6.6 is 0.1088 o. 1080 O.OOOS 2 6.6 15 0.1041 o. 1041 O.OOOO 2 6.6 15 o. 1041 o. 1046 +0.0005 2 6.6 15 0.1041 0.1039 0.0002 2 6.6 15 Am. Jour. Sci., [4], xviii, 180. Il8 METHODS IN CHEMICAL ANALYSIS To remove the copper from the crucible, the deposit is rubbed off as much as possible and the rest may be dissolved in a strong boiling solution of hydrochloric acid with but trifling loss of silver, as is shown in the statement of the results of two experi- ments given below: I. II. Weight of crucible before treatment ? 6 0080 36 0062 Weight of crucible after treatment 36.0x562 36 0041 Loss of silver o . 002 7 O.OO2I It appears that the silver crucible may, with some economy and without sacrifice of accuracy, be substituted for the platinum crucible used as a rotating cathode in the electrolytic determi- nation of copper. The lodometric Estimation of Copper. When potassium iodide is added to a suitable solution of a cupric salt, cuprous iodide is precipitated, while iodine equiva- lent to the amount of iodine fixed in the cuprous iodide is liber- ated. This reaction has been made the basis of an iodometric method for the determination of copper, the first suggestion of such a method having apparently been made by De Haen in 1854. In this process cupric sulphate was treated in solution with potassium iodide and the free iodine determined by sul- phurous acid according to Bunsen. From the amount of iodine thus found the copper was calculated, according to the equation 2 CuSO 4 + 4 KI = 2 K 2 SO 4 + Cu 2 I 2 + I 2 . This method was mentioned in the following year by Mohr,* with the modification suggested by Schwarz that the free iodine be determined by sodium thiosulphate instead of by sulphurous acid. E. O. Brown, f apparently without knowledge of De Haen's previous work, proposed, in 1857, similar procedure, and in 1868 the method with slight modification was presented again by Rtimpler.t Concerning the utility of the method opinions have * Titrirmethode, page 387. t Jour. Chem. Soc., x, 65. } Jour, prakt. Chem., cv, 193. COPPER; SILVER; GOLD varied. Mohr never favored it. As late as 1877, Mohr,* after quoting Meidinger to the effect that cuprous iodide freshly pre- cipitated and washed is capable of taking up iodine, and Carl Mohr's criticism that potassium iodide acts upon cuprous iodide according to the concentration, states that the method is not exact and has nowhere found practical application. On the other hand, Freseniusf recommends the method for the deter- mination of small amounts of copper, noting that ferric salts and other substances capable of setting free iodine from an acidi- fied solution of potassium iodide must not be present, and indi- cates the most favorable procedure. The copper salt treated, he says, should be the sulphate, preferably in neutral solution, though a moderate amount of sulphuric acid is not objectionable. Much free sulphuric acid and all free nitric acid should be neu- tralized by sodium carbonate, and the precipitate dissolved in acetic acid, an excess of which does no harm in the iodometric process. Of recent writers, some have favored the method, while others have commented upon it unfavorably. Low t has been out- spoken in praise, to the extent of declaring a preference for this method in the most accurate technical work over all other methods, even the electrolytic method. According to Low's earlier modification, metallic copper is dissolved in nitric acid, the solution is freed from nitrogen oxides by boiling, a considerable amount of zinc acetate is added, and in the solution having a volume of 50 cm. 3 an excess of solid potassium iodide is dissolved. Zinc acetate is preferred to sodium acetate to take up the free nitric acid. It is said that an excess of potassium iodide is necessary to insure rapidity of action and is harmless. According to the later modification of this method , Low prepares the cupric salt by dissolving the metal in nitric acid (sp. gr. about 1.20), boils the solution, adds bromine water to destroy the nitrogen oxides, boils to expel the bromine, treats with ammonium hydroxide in excess, adds acetic acid and boils again if necessary to get a clear solution. The advantage of using an excess of potassium iodide is emphasized, and the state- ment is made that unless an excess of this reagent is present the * Titrirmethode, 5 Aufl., 288. t Quant. Anal., 6te Aufl., 335, 1875. | Jour. Am. Chem. Soc., 18, 468; 24, 1083, 120 METHODS IN CHEMICAL ANALYSIS . reaction does not proceed to completion until the titration of the free iodine takes place. Low recommends the use of I grm. of potassium iodide, an excess of 0.6 grm., for every 0.075 grm. of copper. As a result of elaborate study, Moser* has reached the con- clusion that the reaction by which cuprous iodide is formed from potassium iodide and cupric sulphate in neutral solution is complete at very high concentration of the solution; that the completeness of the reaction is greatly affected by the volume of liquid ; that the amount of potassium iodide employed is almost without influence either in neutral solution or in acid solution; and that the presence of free sulphuric acid even in large amounts or of hydrochloric acid present in amount equivalent to the cupric sulphate is advantageous. Moser recommends, there- fore, the addition of sulphuric acid for the purpose of bringing the reaction to completion. According to Fernekes and Koch,f an excess of acetic acid does not influence titrations, while a certain amount of potassium iodide 1.5 grm. to 2 grm. for 0.0038 grm. of copper, and 2.5 grm. for 0.0939 grm. of copper must be added to bring about complete action in a volume of 100 cm. 3 Quite recently Cantoni and RosensteinJ have tested the reac- tion between potassium iodide and a cupric salt under various conditions ; but these investigators do not give the absolute val- ues of the amounts of copper taken and found, merely recording the relative effects of varying conditions. From the record of their results it would appear that a fivefold increase of the mini- mum amount of potassium iodide added to portions of 100 cm. 3 of solution containing the same amount of copper salt is without influence upon the result; that increase of volume from loocm. 3 to 350 cm. 3 , other conditions being the same, may affect the results by as much as 5 per cent of their value. The authors conclude that the method gives good results under properly con- trolled conditions. So evidence and opinions as to the effect of various conditions in the process are contradictory, the chief matters of difference being the influence of an excess of potassium iodide used as the * Zeit. anal. Chem., xliii, 597. f Jour. Am. Chem. Soc., xxvii, 1229. j Bull. Soc. Chim., [3], xxxv, 1067-73. COPPER; SILVER; GOLD 121 precipitant, the dilution at which the precipitation should take place, and the effect of acids upon the formation of the cuprous iodide. These points have, therefore, been carefully investigated ex- perimentally by Gooch and Heath,* with the results summarized below. As to the use of potassium iodide in effecting the precipitation of cuprous iodide, it appears that the excess present has within limits an influence upon the result; that beyond the limits the addition of potassium iodide has no appreciable effect; and that the absolute amount of potassium iodide required increases with the dilution. An excess of potassium iodide ranging from 0.6 grm. to I grm. in a volume of 50 cm. 3 , and from 3 grm. to 5 grm. in a volume of 100 cm. 3 , will precipitate completely 0.0020 grm. of copper. In the practical application of these facts it must be borne in mind that it is the excess of potassium iodide and not the full amount added which is important. So it is reasonable to fix upon 2 grm. as the uniform amount of potassium iodide suitable for the precipitation of cuprous iodide equivalent to 0.2 grm. of copper in a volume of 50 cm. 3 ; and upon 5 grm. as the amount of potassium iodide suitable in neutral solutions having a volume of 100 cm. 3 . In a study of the effect of free acid upon potassium iodide alone, it appears that no more than 2 cm. 3 of concentrated sul- phuric acid or hydrochloric acid may be present with 2 grm. of potassium iodide in 50 cm. 3 of solution without liberating an appreciable amount of iodine, and the presence of I cm. 3 of pure nitric acid causes error. The tendency to liberate iodine is manifestly less at the higher dilution, and it appears that in a volume of 100 cm. 3 of solution containing 5 grm. of potassium iodide 3 cm. 3 of concentrated sulphuric acid, hydrochloric acid or nitric acid free from nitrogen oxides may safely be present. Acetic acid of 50 per cent strength may apparently make up half the solution at either dilution. When either sulphuric acid, hydrochloric acid or nitric acid is present, obviously the higher dilution is preferable. In the determination of o.ooio grm. of copper by titration of the iodine set free in a volume of 100 cm. 3 in presence of 5 grm. of potassium iodide, it appears that so much as 50 cm. 3 of 50 per cent acetic acid, 3 cm. 3 of sulphuric * F. A. Gooch and F. H. Heath, Am. Jour. Sci., [4], xxiv, 67. 122 METHODS IN CHEMICAL ANALYSIS lodometric Determination of Copper. KI. Volume. Copper taken as Cu(NO,) 2 . Acid. Copper found. Error. Present. Approx- imate At begin- ning of At end of titra- excess. titration. tion. grrn. grm. grm. cm. s cm.* cm. s grm. grm. Final volume between no cm. 3 and 120 cm. 3 . H a SO 4 cone. 0.1200 O.OQOO O.O9OO 5-o 5-o S-o 4-5 4-5 4-5 2-5 3-0 3-5 100 100 IOO 119 114 114 O. I2OO 0.0903 0.0905 o.oooo +0.0003 4-0.0005 HC1 cone. 0.0900 5-o 4-5 2.O IOO 117 0.0897 0.0003 0.1200 O.O9OO S-o 5-o 4-5 4-5 2.0 3-o IOO IOO 119 114 0.1195 o . 0901 0.0005 + 0.0001 O.I2OO 5-0 4-5 3-o IOO 119 O. I2OO o.oooo O.I2OO 0.0900 S-o S-o 4-5 4-5 3-5 4-o IOO IOO ' 119 114 0.1197 o . 0903 0.0003 +0.0003 HNO, cone. 0.09.00 O.IO5O 0.0900 S-o S-o S-o 4-5 4-5 4-5 i .0 i-5 2-5 IOO IOO IOO 114 117 114 0.0900 0.1051 0.0901 o.oooo + O.OOOI + O.OOOI 50 per cent HC 2 H 3 2 . O.I2OO 0.0900 0.1050 S-o S-o S-o 4-5 4-5 4-5 3-0 S-o IO.O IOO IOO IOO 119 114 117 0.1195 0.0898 0.1048 0.0005 O.OOO2 O.OOO2 Final volume increased by titration to 132 cm. 3 and 150 cm. 3 ; KI corre- spondingly increased. H 2 SO 4 cone. 0.2218 0.3231 7-o 8.0 6.0 6.4 2 3 IOO IOO 135 150 0.2214 0.3226 0.0004 -0.0005 HC1 cone. 0.2023 0.2581 7-0 7-8 6.0 6.7 2 3 IOO IOO 132 141 0.2016 0.2574 0.0007 0.0007 HNOj cone, purified. 0.2023 0.2520 8.0 IO.O 7-o 8-5 I 3 IOO IOO 132 148 0.2017 0.2512 0.0006 0.0008 HC 2 H 3 0, 50 per cent. 0.2125 0.2064 7-5 8.0 8 IOO IOO 133 132 o. 2119 o. 2058 0.0006 0.0009 COPPER; SILVER; GOLD 123 acid, 3 cm. 3 of hydrochloric acid, or 3 cm. 3 of nitric acid (free from nitrogen oxides) may be present without appreciable influ- ence upon the indications of the process. The best general procedure in determining by the iodometric method amounts of copper not exceeding about 0.3 grm. seems to be covered by the following directions: The solution of the cupric salt, containing no more than 3 cm. 3 of concentrated sulphuric acid, hydrochloric acid or nitric acid (free from nitro- gen oxides), or 25 cm. 3 of 50 per cent acetic acid, is made up tc* a volume of 100 cm. 3 , 5 grm. of iodate-free potassium iodide added, and the titration of the free iodine made by sodium thio- sulphate in the usual manner with the use of the starch indicator at the end. The n/io sodium thiosulphate used in the titration adds appreciably to the initial 100 cm. 3 of solution when much copper is estimated, and in case the end reaction has not appeared when so much as 25 cm. 3 of the thiosulphate has been added,. 2 grm. to 3 grm. more of potassium iodide should be added before continuing the titration. The error of the process, properly conducted, should not exceed' a few tenths of a milligram in terms of copper. Results obtained by this procedure are given in the table. The Determination of Copper by Titration of the Precipitated Oxalate with Potassium Permanganate. Upon the well-known fact that copper oxalate is insoluble in water and scarcely attacked by moderate amounts of dilute nitric acid, Bournemann* has based an approximative method for sepa- rating copper from cadmium, by precipitation as the oxalate in. the presence of nitric acid, and estimating the copper, after igni- tion, by any of the well-known gravimetric methods. Classen \ describes a method for the separation of metals as oxalates by- adding to the solution of the salt of the metals a dilute solution of potassium oxalate [i : 6] and concentrated acetic acid to 80 per cent of the total volume. Regarding copper salts in par- ticular, Classen states that precipitation takes place only in dilute solution, and then not completely. According to the experience of Peters, J the precipitation of * Chem. Ztg., xxiii, 565. t Ber. Dtsch. Chem. Ges., x. b, 1316. t Charles A. Peters, Am. Jour. Sci., [4], x, 359. 124 METHODS IN CHEMICAL ANALYSIS copper oxalate from solutions saturated with oxalic acid and containing at least o.oi grm. of copper in 50 cm. 3 of liquid may be made practically complete, the filtrate in such cases giving no blue color with ammonia, in a test tube viewed lengthwise, and only a faint brown color when the filtrate is neutralized, made acid with acetic acid, and tested with potassium ferrocyanide. Peters shows that copper may be determined quantitatively as the oxalate, by precipitation with oxalic acid and titration of the precipitate by potassium permanganate, and separated from certain other metals in the presence of nitric acid, by the addi- tion of considerable amounts of oxalic acid, provided that the amount of copper present in solution exceeds a certain mini- mum value. It is shown that when the amount of copper present falls below the minimum either precipitation does not take place or it is incomplete. It is noted that the minimum is variable with the concentration of the precipitant, oxalic acid, and to some extent dependent upon the condition of the pre- cipitant, the minimum being smaller when the oxalic acid is added in crystalline form rather than in solution to the liquid containing the copper salt. Peters' observations in respect to the effect of concentration and condition of the oxalic acid in solution may be summarized in the following statement: Minimum amount of copper, taken as the Amount of oxalic acid used. be present in order that nearly complete precipitation may Crystalline. In solution. Volume of liquid. take place. grm. grm. grm. cm. O.OIO 5 5* 50 0.025 2 3-5 SO 0.040 I 50 0.050 0-5 50 Saturated solution added to the copper salt dissolved in the least amount of water. It is noted that when a saturated solution of oxalic acid, con- taining o.i grm. of oxalic acid to I cm. 3 , is slowly added to a drop of the copper solution containing 0.0003 g rm - of copper, the precipitated oxalate first formed dissolves completely in a volume of 5 cm. 3 of the precipitant. When no added nitric acid is present precipitates formed in the hot solutions at a volume of 50 cm. 3 may be filtered, either COPPER; SILVER; GOLD 125 at once or after cooling, without loss; but the condition of the precipitate is better in the presence of nitric acid. When the nitric acid is added the mixture must stand before filtration best over night. Results are unsatisfactory when ammonium nitrate is present. Precipitation of According to the procedure recommended by Pe- Copper oxaiate. tergj CO pper sulphate dissolved in approximately 50 cm. 3 of hot water, to which nitric acid is added to the amount of 5 cm. 3 when certain separations are to be effected, is treated in a small beaker with crystallized oxalic acid, 2 grm. to 3 grm., and allowed to stand over night. The precipitate is filtered off on asbestos and washed two or three times with small amounts of cold water and, still in the crucible, is returned to the beaker, 5 cm. 3 or 10 cm. 3 of dilute sulphuric acid [i : i] are added, to- gether with a convenient amount of water, and, after heating the liquid to boiling, the oxalic acid is titrated with permanganate, the oxaiate of copper dissolving readily as fast as the excess of oxalic acid is removed by the permanganate. The precipitate may also be dissolved in 10 cm. 3 of strong hydrochloric acid and titrated at 3O-5O after the addition of 0.5 grm. of manganous chloride. Experimental results are given below. Permanganate Titration of Copper Oxaiate. CuO taken as CuSO 4 . grm. Oxalic acid. grm. HNO 3 (sp. gr. 1.40). cm. 3 Volume at precipitation. cm. 3 CuO found, grm. Error, grm.' o. 1860 o-5 50 0.1864 +0.0004* 0.1860 0-5 50 0.1866 +O.OOO6 0.1860 o-S 5 0.1866 +0.0006 0.1860 I.O 50 0.1866 +0 . 0006 0.0398 I.O 50 0.0391 0.0007 0.1860 o-5 5-o 55 0.1859 O.OOOI 0.1860 o-5 5-o 55 0.1860 o.oooo o. 1990 2.O 5-o 55 o. 1989 O.OOOI o. 1990 3-o 5-o 55 o. 1990 o.oooo * The titration was made in the hydrochloric acid solution containing manganous chloride. Solubility of The fact noted by Peters, that small amounts of Copper oxaiate. precipitated copper oxaiate may be redissolved in a sufficient excess of the precipitant, points to an appreciable de- gree of solubility of the precipitate in the solution of oxalic acid. The observation that very considerable amounts of copper oxa- 126 METHODS IN CHEMICAL ANALYSIS late fail to come down at all until a certain minimum of the copper salt is present, although precipitation is nearly complete when that minimum is reached, indicates supersaturation of the so- lution by copper oxalate; while the capacity of the liquid for supersaturation is apparently limited to some extent by increase in concentration of the oxalic acid. The solubility coefficient of the copper oxalate under the conditions is made up, therefore, of at least two factors, of which one depends upon the normal solubility in the solution of oxalic acid which constitutes the medium of precipitation, while the other depends upon the solu- bility due to supersaturation. In order that small amounts of copper may be precipitated, it is necessary to eliminate, or at least to limit, the capacity of the medium for supersaturation; and in order that large amounts, as well as small amounts, of copper may be determined with the highest degree of accuracy, it is necessary to reduce to the lowest point the normal solubility of the oxalate under the conditions of precipitation. Gooch and Ward * have studied the conditions under which small as well as large amounts of copper may be determined by the oxalate method. It is to be noted, in. the first place, that the character of precipitated copper oxalate depends upon the con- ditions of precipitation. When oxalic acid is added to a cold concentrated solution of a salt of copper, the copper oxalate precipitated is of extreme fineness and tends to pass through the closest filters. The precipitate formed in hot solution is, on the other hand, crystalline and easily separated by filtration of this liquid. The solubility of the precipitate, as well as the ease with which it may be separated from the liquid, turns upon the conditions of precipitation and treatment. Throughout a series of experiments the error of the determination increases with the dilution. That the errors found in titration actually represent approximately the losses in copper for the smaller volumes, is shown by the electrolytic determination of copper in the filtrates from the precipitated oxalate. For a volume of 10 cm. 3 , the average error in the titration of the oxalate precipitated, either from the solution of the sulphate or from a solution of the nitrate, is 0.0002 grm. ; for 50 cm. 3 it is o.ooi I grm. ; for 100 cm. 3 , 0.0053 grm.; for 200 cm. 3 , 0.0203 grm. For similar concentrations of the copper salt and of the oxalic acid, the deficiency in the copper * F. A. Gooch and H. L. Ward, Am. Jour. Sci., [4], xxvii, 448. COPPER; SILVER; GOLD 127 indicated by titration of the precipitated oxalate increases more rapidly than the dilution, a fact which suggests some specific action of water, perhaps hydration affecting the solubility, or hydrolysis affecting the composition, of the copper oxalate. That time and temperature are not essential factors in the pre- cipitation of the oxalate at moderate dilution from solutions of the neutral salt, was shown by Peters in the following experi- ments which indicate also that the precipitates, whether thrown down in hot solution or in cold solution, possess after long standing the same degree of insolubility. Effects of Temperature at Precipitation and Filtration ; Filtration after Stand- ing over Night. Copper taken. Volume. Oxalic acid. Copper found. Error. Precipitation. Filtration. grni. cm. s grm. grm. grm. 0.0502 50 2.0 0.0491 O.OOII Hot. Cold. 0.0502 5 2.O 0.0492 O.OOIO Hot. Hot. 0.0502 50 2.O o . 0490 O.OOI2 Cold. Hot. 0.0502 50 2.O 0.0491 O.OOII Cold. Cold. If any part of the apparent loss of copper oxalate precipitated from solutions of oxalic acid is due to hydrolysis of the normal oxalate, and formation of a basic oxalate as the product of hydro- lytic action, it should be possible to obviate such apparent loss by increasing the active acidity of the solution and thus inhibit- ing hydrolysis, provided the solubility of the normal oxalate is not made greater thereby. Experiment shows that beyond a reasonable degree of concentration the results are not affected by the use of oxalic acid up to the point of saturation of the solu- tion, but that the apparent error is actually diminished by the presence of even very small amounts of sulphuric acid or nitric acid in the liquid, while, within reasonable limits, the addition of more acid produces no further effect. At the higher dilution the effect of the active acid is marked. At a volume of 100 cm. 3 the average error of deficiency shown in absence of the stronger acids is cut in two by the addition of o.i cm. 3 to 5 cm. 3 of nitric acid, or of 0.5 cm. 3 to 2 cm. 3 of sulphuric acid. At the smaller vol- ume of 50 cm. 3 the effect is not so marked, but it is still obvious. These results favor strongly the hypothesis that copper oxalate 128 METHODS IN CHEMICAL ANALYSIS is increasingly subject to hydrolysis as dilution increases, and that the tendency to form a basic salt may be checked by the presence of the stronger acids in suitable amounts. Even very large amounts of nitric acid produce a surprisingly small increase in the apparent solubility of the oxalate. Losses due to solubility of copper oxalate may evidently be kept at low limits by restricting the volume of the solution of oxalic acid in which precipitation takes place; but too high concentration is likely to introduce error due to mechanical in- clusion of oxalic acid in the precipitate. The natural alternative to a close restriction of the volume of the aqueous solution is the limitation of the solvent power of a larger volume of liquid by partially substituting for water some other miscible liquid less capable of dissolving the precipitated oxalate. In testing the effect of alcohol, suggested by Gibbs,* it was found that the presence of that medium, either with or without nitric acid, im- proves the results obtained at similar dilutions of the oxalic acid solution, and to about the same degree whether with or without nitric acid. So it would seem, if the effect of nitric acid is to prevent the formation of a basic salt, that alcohol not only diminishes the normal solubility, but checks hydrolytic action as well. In a volume of 100 cm. 3 containing 20 per cent of alco- hol the error approximates o.ooio grm.; and for a volume of 50 cm. 3 containing 50 per cent alcohol the error is reduced to 0.0003. The effect of nitric acid accompanying the alcohol is not marked. In further experiments it was found that the addition of acetic acid, as proposed by Classen, f is even more effective than the use of alcohol, or of alcohol with nitric acid, but that when consid- erable amounts of copper are present the precipitates formed in solutions containing acetic acid are apt to be very finely divided and consequently difficult to filter. A better condition of the precipitate is obtained, however, if, with the acetic acid, there is also present a moderate amount of nitric acid. It appears that acetic acid, when present to the amount of 25 per cent of the liquid, produces in volumes of 100 cm. 3 about the same effect as alcohol, and, when present to the amount of 50 per cent, diminishes still further the solvent power of the medium for the oxalate. * Am. Jour. Sci., [2], xliv, 214. f Ber. Dtsch. chem. Ges., x, b, 1316. COPPER; SILVER; GOLD 129 The additional presence of nitric acid to 10 per cent of the entire volume does not materially affect the solubility. Sulphuric acid to 10 per cent of the volume of the liquid is without apparent effect upon the solubility of copper oxalate, provided the oxalic acid is also present in the proportion of 4 grm. to 100 cm. 3 of the liquid. Treatment by oxalic acid in a medium consisting of acetic acid of half strength, with or without nitric to the extent of 10 per cent by volume, is plainly the best of the procedures studied for the complete precipitation of copper oxalate in ideal condition. Prevention of Gooch and Ward* have made use of various means Supersaturation. m ^ e effort to break up supersaturation of the pre- cipitating medium when only small amounts of copper oxalate are present. The supersaturated solution was frozen, and the mass melted, following procedure which has been found to be successful in hastening the deposition of small amounts of am- monium magnesium arsenate;f the supersaturated solution was evaporated to dryness, and the residue extracted with water; alcohol was added to the solution of the copper salt before at- tempting precipitation by oxalic acid ; acetic acid of 50 per cent strength was used as the medium in which precipitation was attempted by oxalic acid. From the experimental results it appears that by precipitating at a volume of 50 cm. 3 , freezing, melting, and boiling, the condition of supersaturation may be broken up, the oxalate obtained being soluble in the proportion of about o.oon grm. to 50 cm. 3 of liquid; that by precipitation at a volume of 50 cm. 3 , evaporation to dryness, and extraction with the same volume of water, the copper may be recovered to an amount within about 0.0004 g rm - f that taken; that treat- ment by oxalic acid in 50 per cent alcohol fails to precipitate about 0.0020 grm. of copper from amounts less than 0.0200 grm., while for amounts exceeding that limit the copper is nearly all re- covered; and that in volumes of 50 cm. 3 or 100 cm. 3 , consisting of 50 per cent acetic acid, the copper oxalate is thrown down completely, the presence of nitric acid to the extent of 10 per cent making the nitration more effective without influencing the solubility, while even at a volume of 150 cm. 3 the precipitation is complete provided the acetic acid makes up two-thirds of the volume. Acetic acid appears, therefore, to be most effective in * Loc. cit. t Gooch and Phelps, see page 290. 130 METHODS IN CHEMICAL ANALYSIS Presence of Acetic Acid. breaking up the condition of supersaturation as well as in dimin- ishing the degree of normal solubility in the medium of pre- cipitation. Precipitation in According to the procedure recommended by Gooch and Ward, for small amounts as well as the larger amounts of copper, oxalic acid, 2 grm. or 4 grm., is added to the copper salt dissolved in 50 cm. 3 or 100 cm. 3 , respectively, of the 50 per cent solution of acetic acid containing 5 per cent to 10 per cent of nitric acid to induce a favorable con- dition for crystallization. After standing over night in contact with the solution, the precipitate is collected upon asbestos in a perforated crucible and washed carefully with small amounts of water. The crucible with its contents is placed in a beaker and covered with about 200 cm. 3 of hot water containing 25 cm. 3 of dilute sulphuric acid [i 13], and the solution is titrated with n/io potassium permanganate. Results of this procedure are given in the table. Filtration of Copper Oxalate Precipitated in Solutions Containing 50 per cent Acetic Acid and 5 to 10 per cent Nitric Acid. Copper taken. grm. Total volume. cm. 3 Oxalic acid. grm. Acetic acid. cm. 3 Nitric acid. cm. 8 Copper found, grm. Error, grm. Volume 50 cm. 3 . O.OOIO 50 2 25 5 O.OOIO o.oooo O.OO2O SO 2 25 5 O.OO2I +O.OOOI 0.0031 SO 2 25 5 0.0033 + O.OOO2 O.OO4I 50 2 25 5 o . 0042 +0.0001 O.OO5I 50 2 25 5 0.0049 0.0002 O.OIO2 50 2 25 5 0.0103 + O.OOOI o . 0204 50 2 25 5 o . 0204 o.oooo 0.05II 50 2 25 5 0.0512 +O . OOOI Volume 100 cm. 3 . 0.0031 IOO 4 50 5 0.0031 0.0000 0.0041 IOO 4 50 5 0.0041 0.0000 0.0051 IOO 4 So 5 0.0051 o.oooo 0.0102 IOO 4 50 5 0.0103 +O.OOOI o . 0204 IOO 4 So 5 0.0196 -0.0008* 0.05II IOO 4 SO 5 0.0510 o.oooi 0.05II no 4 50 10 o . 0506 0.0005 0.05II IOO 4 50 10 0.0510 o.oooi 0.1530 IOO 4 50 10 0.1529 o.oooi 0.1530 IOO 4 SO 10 0.1530 o.oooo Filtration imperfect. COPPER; SILVER; GOLD Separations by Peters has shown * that copper exceeding the mini- Peters' Proce- mum amount of o.oi grm. in 50 cm. 3 (that is, in dure from Cad- . . . . , mium, Arsenic, amount sufficient to be precipitated | according to iron, Tin, zinc, ^is procedure), may be successfully separated from cadmium, arsenic, and iron taken as ferric nitrate. When iron is present as the sulphate the results are low. The separation from tin in the stannous form is fairly good for amounts of that element not exceeding o.i grm. In presence of the stannic salt the losses are considerable. Attempted separations from bis- muth and antimony were unsuccessful, and the separation from zinc unsatisfactory on account of the tendency of zinc oxalate to fall with the copper oxalate. Experimental results are given in the table. Permanganate Titration of Copper Oxalate in Separations. CuO taken as CuSO 4 . grm. Element from which copper was separated. grm. Oxalic acid. grm. HNO 3 (sp. gr. 1.40) . cm. 8 Volume at precipita- tion. cm. 3 CuO found. grm. Error, grm. Cadmium. CdO taken / as CdSO 4 . o. 1990 O.IO 2.0 5-o 60 0.1983 0.0007 0.1990 O.2O 2.0 5-o 65 0.1987 0.0003 o. 1990 0.30 2.O 5-o 70 0.1987 0.0003 o. 1990 0.40 2.O 5-o 75 0.1994 +0.0004 0.1990 0.50 2.O S-o 80 0.1996 +0.0006 Arsenic. As 2 O 3 taken as Na 3 AsO 3 . ,. 0.1990 O.IO 2.0 55 0.1991 +0.0001 0.1990 O.2O 2.0 60 0.1987 0.0003 o. 1990 0.50 2.O 75 0.1986 0.0004 0.1990 O.2O 2.O 5-o 60 0.1994 +o . 0004 0.1990 O.2O 2.O 5-o 75 0.1992 +O.OOO2 0.1990 0.60 2.O 5-o 85 0.1995 +0.0005 AssO, taken as H 2 KAs0 4 . 0.1990 O.IO 2.O . . . 60 0.1985 0.0005 0.1990 O.2O 2.O 70 0.1990 0.0000 0.1990 O.IO 2.0 5-o 65 0.1990 0.0000 0.1990 O.2O 2.0 S-o 75 0.1992 +O.OOO2 0.1990 0.30 . 2.0 5-o 85 0.1985 0.0005 o . 2030 0.30 3-o S-o 85 0.2026 0.0004 See page 125. t See page 124. I 3 2 METHODS IN CHEMICAL ANALYSIS Permanganate Titration of Copper Oxalate in Separations. CuO taken as CuSO 4 . Element from which copper was separated. Oxalic acid. HNO 3 (sp. gr. 1.40). Volume at precipita- tion. CuO found. Error. grm. grm. grm. cm. 8 cm. 8 grm. grm. Iron. Fe 2 O 3 taken as Fe(NO 3 )3. 0.1990 0.136 2.O 50 60 0.1987 -0.0003 0.1990 o. 272 2.0 5-o 60 0.1983 0.0007 0.1990 0.364 2.0 5-o 60 0.1988 O.OO02 o. 1990 0-544 2.O 5-o 65 0.1971 0.0019 Tin. Sn taken as SnCU-t-HCl. Cu found. 0.1990 . 0468 2.O 5-o 55 0.1979 o.oon 0.1990 0.0936 2.0 S-o 60 o . 2004 +0.0014 0.1990 0.0936 2.O S-o 60 o. 1992 -f-O . OOO2 0.1990 0.0936 2.O 5-0 60 0.1995 +o . 0005 Sn taken as SnCl 4 . 0.1990 O. IO 2.O 5-o 55 0.1979 o.oon 0.1990 O.IO 2.O 55 0.1959 0.0031 0.1990 O.2O 2.O 5.0 55 0.1974 0.0016 0.1990 0.50 2.O 5-o 60 0.1955 -0.0035 Zinc. ZnO taken as ZnSO 4 . 0.1990 0.028 2.0 5-o 60 o. 2007 +0.0017 0.1990 0.057 2.O 5-o 65 o . 2008 +0.0018 0.1990 0.057 2.O 5-o 65 o . 2008 +0.0018 0.1990 0.085 2.O 5-0 70 0.2035 +0.0045 Separations by Ward* has studied the effect of evaporation to tho Method of dryness, to break up supersaturation, in separations Desiccation. Q j- CO pp er f rO m cadmium, arsenic, and iron. In this process the boiling solution is treated with oxalic acid and nitric acid. The liquid and precipitate are evaporated to dry- ness, and the residue is extracted with cold water containing not too much nitric acid and, in separations from iron, more oxalic acid. The residual oxalate is filtered .off on asbestos in the perforated crucible, washed with small amounts of water, treated * H. L. Ward, Am. Jour. Sci., [4], xxxiii, 423. COPPER; SILVER; GOLD 133 with boiling water containing sulphuric acid, and titrated with permanganate. The details of the preferred treatment and the results are given in the accompanying tables. Copper from Cadmium by Desiccation Process. Copper present. Cad- mium present. Nitric acid at precipi- tation. Volume at pre- cipita- tion. Oxalic acid. Liquid used in extrac- tion. Nitric acid in extrac- tion. Copper found. Error. grm. grm. cm.* cm. 3 gnu cm.* cm.* grm. grm. 0.0051 O. IO 5 5 4 50 5 0.0048 0.0003 0.0514 O.OI 5 SO 4 50 5 0.0507 0.0007 0.0514 0.06 5 50 4 SO 5 o . 0506 0.0008 o . 0504 O. IO 5 SO 4 SO 5 0.0502 O.OOO2 0.0514 O. IO 5 SO 4 50 5 . 0508 o.ooo5 0.0514 O.2O 5 50 4 50 5 O . 0508 0.0006 0.0516 0.30 5 50 4 50 5 0.0507 0.0009 0.1542 O.2O 5 SO 4 50 5 0.1537 0.0005 Copper from Arsenic by Desiccation Process. Copper present. Arsenic as arsenate. Volume at precip- itation. Oxahc acid. Liquid used in extraction. Nitric acid in extraction. Copper found. Error. grm. grm. cm.* grm. cm.* cm.* grm. grm. 0.0051 O. IO 50 50 2 0.0047 0.0004 . 0504 0.05 50 SO 2 0.0499 0.0005 0.0504 0.05 50 50 2 0.0501 -0.0003 0.0504 O. IO 50 50 2 o . 0503 o.oooi 0.0504 O. IO 50 50 2 0.0497 0.0007 o . 0504 O.2O 50 50 2 o . 0498 0.0006 0-1533 O. 2O 50 50 2 0.1528 0.0005 Copper from Iron by Desiccation Process. Volume Oxalic Liquid Nitric Oxalic 'Copper present. Iron present. at pre- cipita- acid in precipi- used in extrac- acid in extrac- acid add- 2d in ex- Copper found. Error. ' tion. tation. tion. tion. traction grm. grm. cm.* grm. cm.* cm.* grm. grm. grm. 0.0504 0.0393 50 50 2 2 O . 0500 0.0004 0.0504 0.0393 50 50 2 2 0.0501 0.0003 0.0504 0.0393 50 50 2 2 0.0499 0.0005 0.0504 0.0786 50 50 2 2 0.0499 -0.0005 0.0511 O. IOOO* 50 50 2 3 o . 0506 0.0005 0-1533 O. IOOO* 50 50 2 3 0.1527 O.OOO6 More than o.i grm. of iron apparently occasions greater solubility of the copper oxalate. 134 METHODS IN CHEMICAL ANALYSIS Separations in Presence of Acetic Acid. WARD* investigated also the application of the process described above for precipitating copper oxa- late in presence of 50 per cent acetic acid, and of nitric acid up to 10 per cent to favor crystallization, to separa- tions of copper from cadmium, arsenic, iron and zinc. Details and results are given in the tables. *, Copper from Cadmium in 50 per cent Acetic Acid. Copper present. Cadmium present. Volume at precipi- tation. Oxalic acid. Acetic acid. Nitric acid. Copper found. Error. grin. grin* cm. 3 grm. cm. 3 cm. 3 grin. grm. 0.0051 O.2O IOO 4 50 5 0.0049 O.OOO2 0.0051 0.30 IOO 4 50 5 0.0053 + O.OOO2 0.0255 O. 2O IOO 4 50 10 0.0257 +O.OOO2 0.0510 0.20 IOO 4 50 10 0.0512 +0 . 0002 0.0511 0.30 IOO 4 5 IO 0.0520 +0.0009 0.1533 0.30 IOO 4 50 IO 0.1556 +0.0023 0.1629 0.30 IOO 4 50* IO o. 1636 +O.OOO7 * The acetic acid was added after precipitation. This apparently makes a sharper separation of the cadmium from large amounts of copper. Copper from Arsenic in 50 per cent Acetic Acid. Copper present. Arsenic present as arsenate. Volume at precipi- tation. Oxalic acid. Acetic acid. Nitric acid. Copper found. Error. grm. grm. cm. 3 grm. cm. 3 cm. 3 grm. grm. 0.0051 0.30 IOO 4 50 0.4 0.0054 +0.0003 0.0511 0.30 IOO 4 50 0.4 . 0505 0.0006 0.1530 0.20 IOO 4 50 10 0.1530 o.oooo O.I53S 0.30 IOO 5 50 IO 0.1530 0.0005 Copper from Iron in 50 per cent Acetic Acid. Copper present. Iron present. Volume at precipi- tation. Oxalic acid. Acetic acid. Nitric acid. Copper found. Error. grm. grm. grm. grm. grm. grm. grm. grm. 0.0510 0.188 IOO 4 50 0.0514 +0.0004 0.0510 0.188 IOO 4 50 2 0.0511 + O.OOOI 0.0510 0.188 IOO 4 50 5 o . 0499 O.OOII 0.0511 O. IOO IOO 4 SO IO o . 0489 O.OO22 0.0510 0.188 IOO 4 50 10 0.0487 0.0023. Much free nitric acid obviously occasions loss of copper oxa- late. Ward t advocates as the best procedure in separating * Loc. cit. COPPER; SILVER; GOLD 135 copper from iron the nearly complete precipitation of the copper by oxalic acid added to the boiling solution devoid of other free acid, with the subsequent addition of acetic acid in amount equal to twice the volume of the solution, to complete the precipita- tion, after cooling. In this manner the results of the following table were obtained. Copper from Iron by Precipitation in Water Solution with Subsequent Addition of Acetic Acid. Copper present. grm. Iron present. grm. Volume at precipita- tion. cm. 3 Oxalic acid. grm. Acetic acid. cm. 1 Copper found, grm. Error, grm. O.OO5I 0.31 50 6 IOO 0.0049 O.OOO2 0.0051 0-45 50 6 IOO 0.0046 0.0005 0.0543 0-15 50 6 IOO 0.0544 +0.0001 0.0543 0.21 50 6 IOO 0.0542 o.oooi 0.0543 0.31 50 6 IOO 0.0546 +0.0003 0.0543 0.45 IOO 12 2OO 0.0538 0.0005 0.1629 0.45 50 6 IOO o. 1649 +O.OO2O 0.1629 o.4S IOO 12 2OO o. 1629 o.oooo When copper and iron are -found together in acid solution, the free acid should either be removed by evaporation or neutralized by potassium hydroxide and the solution made faintly acid with acetic acid before precipitating the copper as oxalate. Ward has shown that the presence of moderate amounts of potassium sul- phate, nitrate or chloride does not affect appreciably the ana- lytical results. The presence of ammonium salts, must, however, be avoided, on account of the tendency of copper oxalate to form a soluble double ammonium oxalate. Determination of The oxalate of lead, though fairly soluble in nitric Copper Assod- acid, falls with copper oxalate from the nitric acid solution, and so the separation of those elements by precipitation of copper oxalate from the acid solution is not feasible. Ward* has shown, however, that lead may be first precipitated completely as lead sulphate, and that then, either with or without previous removal of the precipitated sulphate, the copper may be determined by precipitation and titration of the oxalate. To the solution of lead and copper in the form of nitrates are added an equal volume of acetic acid and 3 cm. 3 - 5 cm. 3 of sulphuric acid. The precipitated sulphate may be * Loc. cit. METHODS IN CHEMICAL ANALYSIS filtered off and weighed and the copper estimated in the filtrate by precipitation and titration as already described ; or, if a deter- mination of copper only is desired, the precipitation of the copper oxalate may be effected without removing the lead sulphate, sul- phate and oxalate being filtered off together, treated with boiling dilute sulphuric acid and titrated to color with permanganate. Results of each procedure are given below. Copper and Lead. Copper present. grm. Lead present. grm. Sul- phuric acid. cm.' Acetic acid. cm.* Volume cm. 3 Oxalic acid. grm. Copper found. grm. Error, grm. Lead found. grm. Error, grm. Titration of copper oxalate precipitated after separation of lead sulphate. 0.0511 0.0511 0.0511 o . 0500 O. IOOO O. IOOO 3 5 5 O O O to o to no IOO IOO 2 2 2 0.0513 0.0508 0.0508 +O.OOO2 -0.0003 -0.0003 0.0499 o . 0996 0.0997 O.OOOI 0.0004 0.0003 Titration of copper oxalate without separation of lead sulphate. 0.0051 0.0511 0.0511 0.0543 0.0511 0.0511 O. IO22 0.1086 0.1533 0.1533 0.30 0.10 0.25 0.30 0.30 0.40 0.30 0.25 O.2O O.2O 5 5 5 10 3 3 10 5 5 5 50 50 50 50 50 50 50 50 5 50 IOO IOO IOO IOO IOO IOO IOO IOO IOO IOO 4 2 2 4 2 4 2 4 2 2 0.0052 o . 0508 O.O5II 0-0537 0.0509 o . 0508 0.1018 o. 1081 0.1527 0.1530 + O.OOOI 0.0003 o . oooo 0.0006 O.OOO2 0.0003 0.0004 0.0005 0.0006 0.0003 SILVER. The Gravimetric Determination of Silver as the Chromate. The precipitation of silver chromate from the solution of a soluble chromate made faintly acid with acetic acid may be carried to completion by the addition of silver nitrate in con- siderable excess. The exact determination of the chromium of a soluble chromate or dichromate may therefore be effected by treat- ing with silver nitrate the solution of either salt, adding ammonia to alkalinity and then acetic acid to faint acidity, transferring the precipitate and washing it in the filtering crucible with a dilute solution of silver nitrate until foreign material other than that reagent has been removed, finishing the washing with a small amount of water applied judiciously in portions, and COPPER; SILVER; GOLD 137 weighing the dried or gently ignited residue of silver chromate.* The success of this process turns upon keeping the chromium at the moment of precipitation essentially in the form of chromate rather than dichromate, and in taking care that an excess of silver nitrate shall be present nearly to the end of the washing. Gooch and Bosworthf have investigated the conditions under which, in reversal of the process just described, silver may be precipitated completely as the chromate, and find that from solutions of silver nitrate alone or containing free nitric acid, potassium chromate precipitates silver chromate completely, provided enough potassium chromate is present to take up the nitric acid with formation of potassium nitrate or dichromate, as well as to form the silver salt. The precipitate, filtered at once or brought to better crystalline condition by dissolving it in ammonia and boiling the solution to small volume, may be transferred to the asbestos filter by dilute potassium chromate Silver Weighed as Silver Chromate. Silver taken as AgNO 3 . K 2 CrO 4 used. HNO 3 . Ag 2 CrO 4 weighed. Silver found. Error in terms of silver. Volume Volume of Weight. of Weight. Volume. Weight. solution. solution. cm. 3 grm. cm. 3 grm. cm. 3 grm. grm. grm. grm. Precipitation by K 2 CrO 4 . is 0.1652 50 0.3 0.2536 o. 1649 0.0003 IO O. IIOI 50 0.3 0.1693 O.IIOI O.OOOO 25 0.1437 50 0.3 O. 22OO 0.1436 o.oooi 25 0.1437 50 0.3 O.22IO 0.1437 0.0000 Precipitation by K 2 CrO 4 in presence of HNO 3 . 25 0.1355 50 o-9 IO 0.182 o. 2091 0.1360 +0.0005 25 0.1355 50 0.9 IO 0.182 0.2081 0.1353 O.OOO2 25 0.1358 50 0.9 IO 0.182 0.2090 0.1360 +O.OOO2 25 0.1355 50 0.9 10 0.182 0.2075 0.1349 0.0006 25 0.1355 50 0.9 10 0.182 0.2090 0.1360 +0.0005 Precipitation by K^CrOi in presence of HNOs, treatment with NH^OH, and boiling to a volume of 10 to 15 cm. 3 . 25 0.1348 50 0.6 IO 0.063 o. 2076 0.1350 +O.OOO2 25 0.1348 50 0.6 10 0.063 0.2068 0.1344 0.0004 25 0.1348 50 0.6 10 0.063 0.2072 0.1347 O.OOOI 25 0.1348 50 0.6 10 0.063 0.2074 0.1348 o.oooo 25 0.1348 50 0.6 IO 0.063 0.2070 o . 1346 O.OOO2 * See page 406. t F. A. Gooch and Rowland S. Bosworth, Am. Jour. Sci., [4], xxvii, 241. 138 METHODS IN CHEMICAL ANALYSIS and washed by small portions of water without appreciable loss. The weight of the residue of silver chromate, dried at gentle heat, may be taken as a measure of the silver originally present. The Electrolytic Determination of Silver. Silver may be determined by deposition of the metal upon the rotary cathode * from an ammoniacal cyanide solution in pres- ence of a few grams of ammonium sulphate, the method of ma- nipulation being precisely similar to that employed in the depo- sition of copper as previously described.! The following table records determinations made in this man- ner I with a solution of silver nitrate standardized as the chloride. Deposition of Silver from Ammoniacal Cyanide Solution. Silver taken, gnu. Silver found, grin. Error, grm. Current, amp. N. D. 100 . Time, min. 0.0968 o . 0966 O.OOO2 1.8 6 IS 0.0968 0.0967 O.OOOI 1.9 6-3 15 0.0968 o . 0965 -0.0003 1.8 6 15 0.0968 0.0969 +O.OOOI 2 6.7 IO 0.0968 0.0965 0.0003 3 10 8 0.1898 o. 1901 +0.0003 2.5 8-3 10 0.1898 0.1898 o.oooo 2.5 8-3 10 0.1898 o . 1900 +0.0002 3 10 IO 0.1898 0.1893 0.0005 2-5 8-3 IO Gooch and Feiser have determined silver by depositing it from an ammoniacal solution of the oxalate, as was done by Gooch and Read|| in the preparation of silver-plated electrodes, in order to avoid all possible contamination of the silver deposit by nonvolatile material. In testing this process, measured amounts of the silver nitrate solution (25 cm. 3 or 50 cm. 3 ) were drawn from a burette into a small beaker and treated with ammonium oxalate to complete precipitation. The silver oxa- late was dissolved in a slight excess of ammonia, and this solution, diluted to 100 cm. 3 , was electrolyzed with the rotary cathode and a current of 0.25 amp.-i.5 amp. and 4-7 volts. The cathode * See Fig. 13. t See page 116. t Gooch and Medway, Am. Jour. Sci., [4], xv, 320. F. A. Gooch and J. P. Feiser, Am. Jour. Sci., [4], xxxi, 109. II F. A. Gooch and H. L. Read, Am. Jour. Sci., [4], xxviii, 544. COPPER; SILVER; GOLD 139 with the deposited silver was dried cautiously over a low flame and thereafter ignited to incipient redness. The details of individual experiments are given in the table. Electrolysis of Silver Nitrate Dissolved in Ammonium Oxalate and Ammonia* Current. Agin AgNO 3 taken. Ag found. Error. Time. Revolu- tions per Amp. N?K. Volt. mm. grm. grm. grm. min. A crucible used as cathode. 0.2687 0.2685 0.0002* i -5-i 5-3-3 6-7 25 500 0.2687 0.2687 0.0000* i -5-i 5-3-3 6-7 30 500 0.2687 0.2684 O.OOO3* i-o-5 3-3-1-7 6-7 30 450 0.2687 0.2685 O.OOO2* 1-0.5 3-3-1-7 4-6 30 450 0.3183 0.3181 O.OOO2* i-5-i 5-3-3 4-6 20 450 0.3183 0.3178 0.0005t i -5-i 5-3-3 4-6 10 450 Gauze disks used as cathode. 0.3183 0.3179 0.0004* 1-0.5 0.5-0.25 4-6 20 400 0.3183 0.3182 O.OOOI* 1-0.25 0.5-0.10 6-8 25 400 0.3183 0.3181 O.OOO2* 1-0.25 0.5-0.10 5-8 25 400 0.3183 0.3180 0.0003* 0.75-0.25 0.4-0.10 5-7 40 45 0.3183 0.3180 0.0003* 1-0.25 0.5-0.10 4-6 40 450 0.3183 0.3176 0.0007! 0.75-0.25 O . 4-0 . TO 6 20 450 Gauze cone used as cathode. 0,2687 0.2686 O.OOOI* i. 5-i 3-2 4-6 25 500 0.2687 o . 2683 0.0004* 1-25 2-0.5 6-7 30 450 0.2687 o. 2684 -0.0003* i-5 2-1 4-6 25 450 0.2687 0.2686 O.OOOI* 1-0.25 2-0.5 4-6 25 450 0.5375 0.5373 O.OOO2* i -5-i 3-2 6-7 25 450 0.5375 0.5371 0.0004* i -5-i 3-2 6-7 25 500 * Deposition complete, as shown by H 2 S test, t Deposition incomplete, as shown by HjS test. The details of other experiments in which the silver was first precipitated as silver chloride and then deposited from the solu- tion in ammonia and ammonium oxalate are given in the follow- ing table. In these experiments, in which the solutions were more strongly ammoniacal than those of the .experiments of the preceding series, the deposits were dark and spongy, but they became lighter in color and more compact upon drying. 140 METHODS IN CHEMICAL ANALYSIS Electrolysis of Silver Chloride Dissolved in Ammonium Oxalate and Ammonia. Current. Agin AgNO 3 taken. Ag found. Error. Time. tions per Amp. Approx. N. D. 100 . Volt. mm. grm. grm. grm. mm. A crucible used as cathode. 0.3191 0.3191 0.3187 0.3189 0.0004 O.OOO2 i -5-i i -5-i 5-33 5-33 5-7 4-6 20 30 500 500 Gauze disks used as cathode. 0.3191 0.3185 0.0006* i -5-i 0-75-0-5 5-7 15 500 0.3191 0.3189 O.OOO2 i. 5-i o.75-o.5 S-7 25 500 0.3191 0.3190 o.oooi i -5-i 0-75-0-5 4-6 35 500 * Time of electrolysis short. It appears, therefore, that silver may be deposited from an ammoniacal solution of the oxalate, in presence of ammonium nitrate or ammonium chloride, in pure condition and in form suitable for quantitative estimation. When the current density is high the deposit is apt to be voluminous, shrinking considerably upon drying, and this phe- nomenon was especially notable in deposits upon the compara- tively small and smooth surface of the crucible. The best form of apparatus for this process appears to be the gauze cone set point downward, and so placed with relation to an annular plati- num band used as the anode that the end of the axis, where the mechanical effect of rotation is least, shall not receive much of the deposit. Experiments made with stationary gauze electrodes were not successful, nor were those made with a dish cathode and stirring anode. The lodometric Estimation of Silver, based upon the Use of Potassium Chromate as a Precipitant. With proper precautions, silver may be precipitated and accu- rately estimated as silver .chromate.* The addition of a sufficient excess of potassium chromate to a solution of silver nitrate, even iri the presence of small amounts of nitric acid, brings about a complete precipitation of the silver as silver chromate, and * See page 136. COPPER; SILVER; GOLD 141 the precipitate thus obtained may be transferred to the asbestos filter by means of a dilute solution of potassium chromate and washed with small amounts of water without appreciable loss of silver chromate. Upon the basis of such exact precipitation of silver chromate by potassium chromate, Gooch and Bosworth* have accomplished the iodometric estimation of silver, both by the determination of the chromic acid ion of the potassium chromate which remains after the precipitation of the silver salt by a known amount of standard potassium chromate, and by the estimation of the chromic acid ion of the precipitated and washed silver chromate. Precipitation of Silver Chromate and Determination of the Excess of the Pre- cipitant. Silver taken. K 2 CrO<. NaNO 3 present. grm. Na 2 S 2 3 used. cm. 8 . Silver found. grm. Error in terms of silver. grm. Volume of solution. cm. 3 Weight, grm. Volume. cm. s . Weight, grm. 20 O.I26l 0.3039 26.54 o. 1262 +O.OOOI 25 0.1576 0.3039 22.61 0.1575 o.oooi 15 o . 0946 0.3293 30-5I o . 0946 o.oooo IS o . 0946 0.3293 30.60 o . 0940 0.0006 15 o . 0946 0.3293 30.57 0.0941 0.0005 15 o . 0946 0.3293 30.55 0.0945 O.OOOI 19.98 o. 1260 0.3293 32.20 0.1255 -0.0005 2O 0.1261 0.3293 32-09 o . i 263 +O.OOO2 2O o. 1261 0.3293 32.10 0.1263 +O.OOO2 25 0.1576 0.3293 27.30 0.1576 o.oooo 10 O. IIOI 37 0.2436 20.47 o. 1107 +0.0006 IO O.IIOI 37 0.2436 20.53 0.1103 +O.OOO2 IO O.IIOI 25 0.1647 9.07 0.1097 0.0004 IO O.IIOI 27 0.1778 1 1. 02 o. 1096 -0.0005 15 0.0862 30 0.1974 17.14 0.0859 0.0003 15 0.0862 3 0.1974 17-05 0.0865 +0.0003 15 0.0862 30 0.1974 17.14 0.0859 0.0003 25 0.1437 50 0.3294 10 28.53 0.1435 O.OOO2 According to the first procedure a known amount of standard potassium chromate in excess of the amount needed to precipi- tate the silver is added to the solution of silver nitrate. The pre- cipitate is dissolved in ammonia and reprecipitation brought about by boiling to a volume of 10 cm. 3 -i5 cm. 3 . The second, crystal- line precipitate is filtered upon asbestos and washed with the least possible amount of water applied in small portions successively. The filtrate is treated with potassium iodide and acidified with * F. A. Gooch and Rowland S. Bosworth, Am. Jour. Sci., [4], xxvii, 302. 142 METHODS IN CHEMICAL ANALYSIS sulphuric acid. The iodine set free is titrated with sodium thio- sulphate. The difference between the silver value of the iodine thus found and that of the potassium dichromate used is taken as the measure of the silver present. In the table are given the details of experiments performed in accordance with this procedure. Inasmuch as relatively large amounts of potassium chromate are necessary to bring about complete precipitation of the silver chromate in the presence of nitric acid, the above procedure is less adapted to the determination of silver in a solution con- taining that acid than the second method whereby the precipi- tated and washed silver chromate is determined. In this second method the silver solution containing free nitric acid is treated with potassium chromate in excess of the amount necessary to take up the nitric acid with formation of potassium dichromate. The precipitate is dissolved in ammonia and reprecipitation is effected by boiling to a volume of 10 cm. 3 -i5 cm. 3 . The second, crystalline precipitate is transferred to an asbestos filter by means of a dilute solution of potassium chromate, washed with the least possible amount of water applied in small portions successively, and dissolved in a few cm. 3 of a strong solution of potassium iodide. The solution in potassium iodide is diluted and acidified with sulphuric acid. The iodine set free is titrated with sodium thiosulphate and taken as the measure of the silver present. Results of this procedure are given in the following table. lodometric Determination of Precipitated Silver Chromate. Silver taken. HNO 3 present. K 2 CrO 4 , weight. Na 2 S 2 3 used. Silver found. Error in terms of silver. Volume of solution. Weight. cm.* grrn. grm. grm. cm. 3 grm. grm. 25 0.1348 0.063 O.6o lS-35 0.1342 O.O006 2O o. 1078 0.063 0.65 14.67 0.1073 . 0005 IS o . 0808 0.063 0.65 II .06 o . 0809 + O.OOOI 15 0.0808 0.063 0.65 10.96 o . 0802 O.OOO6 20 o. 1078 0.063 0.65 14.67 0.1073 -0.0005 3 0.1618 0.063 0-75 22.02 0.1610 0.0008 20 0.1078 0.063 0.65 14.71 0.1075 0.0003 25 0.1348 0.063 0.65 I8.4I o.i347 o.oooi 25 0-1348 0.063 0.65 I8.4I o.i347 O.OOOI 20 0.1078 0.063 0.65 14-74 o. 1078 O . 0000 COPPER; SILVER; GOLD 143 The lodometric Determination of Silver Based upon the Reducing Action of Potassium Ar senile. It is well known that an ammoniacal solution of silver arsenite deposits metallic silver when the ammonia is evaporated by boil- ing. During this reaction, by which the silver salt is reduced, the arsenious acid becomes oxidized to the higher condition of oxidation, where arsenic has a valence of five, according to the equation 2 Ag 2 O + As 2 O 3 = As 2 O 5 + 4 Ag. Bosworth* has shown that this reaction, is quantitative and capable of serving as the basis of a reliable iodometric method for the determination of silver. To the solution of silver taken as the nitrate is added a known volume of a standard potassium arsenite solution in excess of the amount necessary to reduce the silver salt present. Ammonia is added in sufficient quantity to dissolve the precipitate formed, or 25 cm. 3 of a saturated solu- tion of sodium bicarbonate may be used instead of the ammo- nia. The resulting solution is diluted to 100 cm. 3 , and boiled. The solution, out of which metallic silver separates, is cooled, slightly acidified, and then made alkaline with sodium bicarbo- nate. The excess of potassium arsenite is titrated with n/io iodine. The silver value of the iodine used is subtracted from that of the potassium arsenite originally taken, and the result used as a measure of the silver present. Results of this procedure are given in the table. Reduction of Silver by Arsenite and Titration of the Excess. Silver taken, grrn. KH 2 AsO 3 added. I used. Silver found, grm. Error in terms of silver. gnu. cm. 3 Silver value, grrn. cm. 3 Silver value, grm. Use of NH 4 OH with filtering. 0.1054 20 O. 2OOO 8-53 0.0949 0.1051 0.0003 0.1054 20 O . 2OOO 8.52 o . 0948 0.1052 O.OOO2 0.1054 30 0.3000 17.42 0.1939 o. 1061 +o . 0007 0.1159 2O 0.2000 7.60 o . 0846 0.1154 0.0005 0.1054 21 O.2IOO 9-37 0.1043 0.1057 +0.0003 0.1054 2O O. 2OOO 8.48 0.0944 o. 1056 +0 . 0002 * Rowland S. Bosworth, Am. Jour. Sci. [4], xxviii, 287. 144 METHODS IN CHEMICAL ANALYSIS Reduction of Silver by Arsenite and Titration of the Excess. Silver taken, grm. KH2AS03 added. I used. Silver found, grm. 3rror in terms of silver. grm. cm. 3 Silver value, grm. cm. 3 Silver value, grm. Use of NaHCO 3 with filtering. 0.1054 15 0.1618 5.65 0.0571 0.1047 0.0007 0.1054 23 0.2481 14.14 0.1430 0.1051 -0.0003 o. 1054 12 0.1295 2.40 0.0243 0.1052 O.OO02 0.1054 15 0.1618 5.60 0.0566 0.1052 O.O002 0.1054 15 0.1618 5-55 0.0561 0.1057 +0.0003 0.1054 20 0.2158 10.91 0.1104 0.1054 o.oooo 0.2635 35 0.3776 n-33 0.1146 0.2630 0.0005 Use of NH 4 OH. Titration carried on in presence of the precipitate. 0.1054 20 O.2OOO 8-55 0.0952 0.1048 0.0006 0.1054 20 O.2OOO 8.50 0.0946 0.1054 o.oooo 0.1054 23 0.2300 11.28 0.1256 o. 1044 O.OOIO 0.1054 2O O.2OOO 8-45 0.0941 0.1059 +0.0005 0.1054 2O O.2OOO 8.48 0.0944 o. 1056 +O.OOO2 Use of NaHCO 3 . Titration carried on in presence of the precipitate. 0.1054 18 o. 1800 6.80 0.0757 0.1043 o.oon 0.1054 17 O.I7OO 5-8l 0.0647 0.1053 o.oooi 0.1054 15 0.1500 4.00 o . 0445 0-1055 +O.OOOI 0.1054 21 O.2IOO 9-45 0.1052 0.1048 0.0006 0.1054 25 0.2500 13.00 0.1447 0.1053 O.OOOI 0.1054 31 0.3100 18.40 o . 2048 0.1052 O.OOO2 Use of NaHCO 3 . 2 grm. of NaNO 3 present. Titration carried on in presence of precipitate. 0.0949 21 0.2100 IO.42 0.1160 o . 0940 0.0009 0.1054 21 O. 2IOO 9-43 o. 1050 o. 1050 0.0004 0.1265 2O O. 2OOO 6.60 0.0735 0.1265 o.oooo 0.1686 21 O.2IOO 3.8o 0.0432 0.1678 0.0008 0.1054 15 0.1500 4.08 0.0454 0.1046 0.0008 This process proves to be applicable to the determination of the silver in freshly precipitated silver chloride, so that its range is thus extended to the determination of silver in many mixtures. According to the procedure outlined, the freshly precipitated silver chloride is acted upon by ammonia until dissolved. The COPPER; SILVER; GOLD solution is diluted to 100 cm. 3 and the reduction is effected, adding an excess of standard arsenite and boiling the mixture. The excess of arsenite is titrated according to the method de- scribed above. Results of this procedure, including separations from copper and lead, are given in the following table. Reduction of Silver Chloride after Separations. Silver taken. grm. KH 2 AsO 3 added. I used. Silver found, gnu. Error in terms of silver. grui. cm. Silver value. grm. cm. 8 Silver value, grm. Reduction of precipitated AgCl. 0.1017 is 0.1619 5-4 0.0599 O.IO2O +0.0003 0.1017 15 o. 1619 5-44 0.0603 o. 1016 O.OOOI 0.1017 15 0.1619 5-40 0.0599 0.1020 +0.0003 0.1017 15 o. 1619 5-42 o . 0601 0.1018 +0.0001 0.1017 17 0.1834 7-44 0.0825 0.1009 0.0008 Reduction of AgCl precipitated in the presence of 0.09 grm. of copper. 0.1017 15 0.1619 5-41 0.0600 o. 1019 +O.OOO2 0.1017 is o. 1619 5-44 0.0603 0.1016 O.OOOI 0.1017 is 0.1619 5-39 0.0598 O.IO2I +0.0004 Reduction of AgCl precipitated in the presence of 0.2 grm. of lead. O. I22O o. 1108 16 15 0.1726 0.1619 4.57 4.60 0.0507 0.0510 0.1219 0.1109 O.OOOI + 0.0001 Reduction of AgCl precipitated from a solution containing 0.09 grm. of copper and 0.2 grm. of lead. 0.1017 15 0.1619 5-45 o . 0604 0.1015 O.OOO2 GOLD. The Electrolytic Determination of Gold. Medway* has shown that from an ammoniacal cyanide solu- tion gold may be deposited rapidly and in good form upon the rotating crucible used as the cathode. * H. E. Medway, Am. Jour. Sci., [4], xviii, 56. 146 METHODS IN CHEMICAL ANALYSIS Deposition from an Ammoniacal Cyanide Solution. Gold taken. Gold found. Error. Current. N. D. 100 . Time. grm. grm. grm. amp. mm. 0.0695 o . 0694 0.0001 2 6.6 3 0.0695 o . 0696 +O.OOOI 2 * 6.6 3 0.0598 0.0598 o.oooo 0-5 1.8 3 0.0598 0.0598 o.oooo 0-5 1.8 3 0.0598 0-05975 0.00005 I 3-3 25 The lodometric Estimation of Small Amounts of Gold. Gooch and Morley* have shown that when potassium iodide reacts at suitable concentration upon small amounts of gold tri- chloride in solution, the reaction takes place regularly and in accordance with the theory that two molecules of the thiosul- phate are the equivalent of two atoms of iodine and one atom of gold. AuCl 3 + 2 KI + 2 Na 2 S 2 O 3 = AuCl + Na 2 S 4 O 6 + 2 KC1 + 2 Nal. The reduction of the auric salt, with the consequent liberation of iodine, is, however, conditioned by the volume of the solution, the mass of the iodine present, and the time of action. The following statement, in which each result is the average of several titrations in close agreement, shows the effect upon the immediate evolution of iodine brought about by adding varying amounts of water to the gold solution before introducing the iodide, and the effect of different amounts of iodide at different dilutions. Volume Potassium iodide. Gold chloride. before the addition of the thio- sulphate. o.oi grm. 0.02 grm. 0.05 grm. o.i grm. 0.2 grm. grm. cm. 8 0.8l 0.81 0.81 0.82 0.84 0.00087 15 S Jc\ 0.77 0.78 0.80 0.81 0.8l 11 25 .3 ^ 8 g 0.74 0.72 0.78 0.79 0.80 50 'Q w ~ 0.61 0.61 0.68 o. 76 0.79 " IOO w is 2 0-45 0.49 0.60 0.72 0-75 " 2OO "5 c It is evident that for the smaller amounts of iodide the libera- tion of iodine decreases rapidly with the dilution. The larger amounts at the highest concentration show readings a trifle * F. A. Gooch and Frederick H. Morley, Am. Jour. Sci., [4], viii, 261. COPPER; SILVER; GOLD 147 above the normal perhaps because the well-known effect of concentrated solutions of a soluble iodide upon the delicacy of the starch end-color begins to appear. At volumes lying be- tween the limit of 25 cm. 3 and 50 cm. 3 , o.i grm. of potassium iodide is an appropriate amount to use; at a volume of 15 cm. 3 , o.oi grm. to 0.05 grm. of the iodide will do the work; and at lower dilutions, as will appear in the tabular statements to follow, even less of the iodide is effective. In carrying out the process of analysis, a convenient amount of the solution of gold chloride is drawn from a burette, potassium iodide is introduced in amounts always several times the theoret- ical equivalent of the gold, and more than enough to dissolve the aurous iodide precipitated at first, a sufficiency of clear starch indicator is added, the starch blue bleached by standardized thiosulphate, and standardized iodine added until the liquid assumes a faint, rose color. Experimental results are given in the following table: Solutions Approximately n/ioo Gold chloride =0.8710 grm. to i liter. Sodium thiosulphate, nearly n/ioo, =1.7012 grm. to i liter. Iodine, nearly w/ioo, =1.3697 grm. to i liter. Volume at beginning of titration, approximately 50 cm. 3 . AuCl 3 taken. cm. 3 KI taken. grm. Na 2 S 2 O 3 used. cm. 3 Gold found, grm. Theory for gold. grm. Error, grm. 5 0.05 4.02 0.00426 0.00435 0.00009 5 0.05 4.01 0.00425 0.00435 o.oooio 5 0.05 4.06 0.00431 0.00435 0.00004 5 0.05 4.07 0.00432 0.00435 0.00003 5 0.05 4-*04 0.00428 0.00435 0.00007 10 0.08 8.17 0.00867 0.00871 0.00004 10 0.08 8.15 o . 00864 0.00871 0.00007 10 O.o8 8.16 o . 00865 0.00871 0.00006 10 O.o8 8.15 o . 00864 0.00871 0.00007 10 O.o8 8.19 o . 00869 0.00871 O.OOOO2 10 O.o8 8.46 0.00897 0.00871 +0 . 00026 10 0.08 8.24 o . 00874 0.00871 -j-o . 00003 When approximately centinormal solutions of gold, iodine and thiosulphate are used, an error of o.oi cm. 3 in reading the volume corresponds to an error of o.ooooi grm. of gold. It is not to be expected that such readings can be trusted ordinarily to a higher degree of accuracy than 0.02 cm. 3 . In case all three solutions should be read to this limit of accuracy with the errors of all 148 METHODS IN CHEMICAL ANALYSIS lying in the same direction, the summation of error would corre- spond to 0.00006 grm. of gold. Errors in reading are, of course, reduced when n/iooo solu- tions are employed, but the use of n/iooo iodine necessitates a correction of o.i cm. 3 for volumes not exceeding 30 cm. 3 , that being the amount necessary to bring out the rose color in blank tests containing no gold. After the introduction of o.i cm. 3 of n/iooo iodine into a mixture of potassium iodide and starch indicator of volume not exceeding 30 cm. 3 , a single drop of the gold solution equivalent to 0.000002 grm. of gold gives a distinct rose color; before such adjustment of the solution five drops equivalent to o.ooooio grm. of gold must be added to develop the same color. The following table gives the data of tests with such solutions. Solutions Approximately n/iooo Gold chloride =0.0871 grm. to i liter. Sodium thiosulphate, nearly w/iooo =0.17012 grm. to i liter. Iodine, nearly w/iooo =0.13697 grm. to i liter. AuClj taken. cm. s KI taken, grm. Na 2 S 2 O 3 used. cm. 3 Gold taken, grm. Gold found, grm. Error, grm. 10 O.OI 8-39 0.000871 o . 000890 +0.000019 .9 O.OI 7-45 0.000784 0.000790 +0 . 000006 8 O.OI 6.30 0.000697 0.000668 0.000029 7 0.008 5-So O.OOo6lO 0.000583 0.000027 6 0.008 5-12 0.000523 o . 000543 + O.OOOO20 5 0.005 4-23 0.000435 o . 000449 +0.000014 4 0.005 3.38 o . 000348 0.000358 +O.OOOOIO 3 0.003 2-55 0.000261 O.OOO27O +0.000009 2 0.003 1.71 0.000174 0.000181 +0.000007 I 0.003 0.90 0.000087 0.000095 +0.000008 In the practical application of the process to the determination of gold, the elementary form of that metal is the natural starting point. To get the metal into solution with chlorine water or mixed hydrochloric and nitric acids is an easy matter, but the removal of the excess of the oxidizer by evaporation without reducing some auric chloride to the aurous form is difficult. Free chlorine may, however, be removed from a solution of auric chloride, without reducing the auric salt, by treatment of the solution with ammonia in excess, boiling gently, acidifying with hydrochloric acid, and heating if necessary to redissolve the pre- cipitate by ammonia, again treating with ammonia and heating. COPPER; SILVER; GOLD 149 and once more acidifying. On the second addition of ammonia no precipitation usually takes place with these small amounts of gold. The following table contains determinations made with a solu- tion of pure gold leaf. lodometric Determination of Gold, Gold chloride made by dissolving 0.0104 g rm - of pure gold in the manner described and diluting to 200 cm. 3 . Sodium thiosulphate, nearly w/iooo, =0.17012 grm. to i liter. Iodine, nearly w/iooo =0.13697 grm. to i liter. Potassium iodide =iogrm. to i liter. Portions were treated with the potassium iodide without previous dilu- tion. AuClj taken. cm. KI taken, grm. Na 2 S 2 O 3 used. cm. 3 Gold taken, grm. Gold found, grm. Error, grm. I 0.005 o-55 0.000052 0.000058 +0 . 000006 I 0.005 0-55 0.000052 0.000058 +0.000006 2 0.005 i. 06 O.OOOIO4 O.OOOII2 -J-O.OOOOOS 2 0.005 1.08 O.OOOIO4 O.OOOII4 +O.OOOOIO 5 O.OI 2.45 0.000260 O.OOO26O O . OOOOOO 5 O.OI 2.50 O.OOO26o O.OOO265 +o . 000005 5 O.OI 2-45 o . 000260 O.OOO26O O . OOOOOO 5 O.OI 2.50 O.OOO26O O.OOO265 +0.000005 5 O.OI 2.50 0.000260 0.000265 +0.000005 10 O.O2 4.86 0.000520 0.000515 0.000005 10 O.O2 4-85 0.000520 0.000517 0.000003 IO O.O2 4.90 0.000520 O.OOO52O . OOOOOO IO O.O2 4.80 0.000520 O.OOO5I2 0.000008 IO O.O2 4.84 0.000520 O.OOO5I6 0.000004 To show the range and error of the process, the results of these and other experiments recorded by Gooch and Morley may be summarized as follows: Range of the Process. Strength of solutions. Number of determina- tions. Gold taken. Error, average. Extremes. Iodine. Thiosul- phate. Gold in in i cm. 3 rng. mg. mg. mg. II 8.71 -4-35 n/ioo W/IOO 0.871 0.05 ( +0.03 ( O.I 20 0.871-0.087 n/ioo n/ioo 0.0871 +0.02 ( +0.06 ( O.O2 IO 0.871-0.087 n/iooo W/IOOO 0.0871 +0.004 ( +O.O2O ( O.O29 14 0.520-0.052 n/iooo tt/IOOO 0.052 +0.002 ( +0.01 { 0.008 150 METHODS IN CHEMICAL ANALYSIS It is plain that the average experimental errors, due to all causes, do not very much exceed the errors which might natu- rally be expected to arise from errors of reading. In repeating this work, Maxson * has obtained results of a similar order of accuracy. Maxson has also studied the possi- bility of reduction of the aurous iodide formed in the process and finds in periods much exceeding those required for the analytical operation no evidence of further action other than the for- mation of the aurous salt. Thus, aurous iodide, obtained by treating a solution of auric chloride, containing 0.0125 grm. of gold, with potassium iodide according to the directions of Gooch and Morley, adding starch and bleaching the starch iodide with sodium thiosulphate, shows no color of starch blue after the in- terval of an hour. Inasmuch as an interval of ten minutes is enough for the complete manipulation of a single determination, it is plain that the stability of the aurous iodide does not figure in the accuracy of the determination of the small amounts of gold for which the process was designed. The Colorimetric Determination of Small Amounts of Gold. A colorimetric method for the estimation of small amounts of gold has been based by Maxson f upon the coloration exhibited by suspensions of red colloidal gold. Blake J has shown that acetylene is the most suitable reagent for effecting the reduction of the auric salt with production of the red colloidal gold, the treatment consisting in drying the chloride at 170, dissolving in ether, and pouring the etheral solution into water containing ether and saturated with acetylene gas. For the purposes of this method the simpler treatment with an aqueous solution of acetylene proved to be adequate. The procedure consists in preparing a red colloidal suspension containing a known amount of gold in a given volume and by means of measured amounts of this solution matching the color developed in a similar solution containing in measured volume the gold to be determined. For making the comparisons of color a modified form of the apparatus proposed by Penfield for the colorimetric estimation * Ralph N. Maxson, Am. Jour. Sci., [4], xvi, 155; xvii, 466. t Ralph Nelson Maxson, Am. Jour. Sci., [4] xxi, 270. J Am. Jour. Sci., [4], xvi, 381. COPPER; SILVER; GOLD 151 of titanium is used. This consists of comparison tubes set ver- tically in a dark box and illuminated from below. A mirror suit- ably situated beneath the box containing the tubes gives efficient illumination. Such an apparatus is cheaply and easily procured. With tubes having a diameter of I cm. and a length of 13 cm., and accurately graduated to hold 10 cm. 3 , this simple colorimeter is capable of determining accurately very small amounts of gold. The standard suspension is made by treating in a measuring flask a measured amount of a standardized solution of ordinary undried gold chloride with an aqueous solution of acetylene and, after the full development of color, making up to the mark. A solution of pure gold to be examined is treated similarly with aqueous acetylene and made up to known volume. When small amounts are handled the volume of the solution should not exceed a few cubic centimeters, and only a small amount of the aqueous solution of acetylene should be added; otherwise the coloration may be partially or totally inhibited. If traces of electrolyte are present, the coagulation of the red gold may sometimes be avoided by the addition of a few drops of ether to the cold solution. It is a well-known fact that small amounts of electrolyte will rapidly change red gold to the blue modification. It is necessary, therefore, to conduct the comparisons in a room reasonably free from fumes, and to have all containing vessels free from soluble material. Flasks which have been treated with steam for a few minutes give the best results. Red suspensions contained in such flasks may show no trace of blue after an interval of several weeks. The comparison of colors is carried out in the following man- ner: A measured amount of the suspension is drawn off into the left-hand tube and diluted to the mark with water ; a suitable amount of water is then placed in the right-hand tube and the standard suspension drawn into the tube until the colors are seen to be identical. The amount of water to be used can be determined by preliminary experiment. The positions of the tubes are reversed before the final reading, and the mean taken. In the table below is the record of experiments made to determine the range of amounts of gold capable of accurate estimation in such an apparatus with tubes of the dimensions described above. The comparisons were made with a red sus- 152 METHODS IN CHEMICAL ANALYSIS pension, prepared by careful dilution of a more concentrated standard suspension, which contained 0.0000107 grm. of metal in i cm. 3 . Colorimetric Estimation of Gold. Gold suspension taken. cm.* Gold suspension used. cm. 3 Gold used, gnu . Gold found. grm. Error. grm. 9-50 8.00 7.00 6.00 5.00 9-05 7-59 6.89 5-83 4.84 0.000102 O.OOOO86 0.000075 O.OOOO65 o . 000054 0.000097 0.000082 0.000074 o . 000063 0.000052 0.000005 0.000004 O.OOOOOI O.OOOO02 O.OOOOO2 4.00 3-88 o . 000043 o . 000042 O.OOOOOI 3.00 2.OO 2-47 1.82 0.000032 O.OOOO22 0.000027 O.OOO02O 0.000005 O.OOOOO2 I .OO 0-93 O.OOOOII O.OOOOIO O.OOOOOI The intensity of the color in the experiments ranged from a deep red to a faint pink. Further comparisons made with sus- pensions more dilute than those described above gave errors of magnitude increasing with the dilution. The amounts handled here are, then, the minimum quantities that can be accurately estimated with the apparatus described. It is obvious that if larger amounts of the metal are to be determined, tubes of greater dimensions should be used. The application of such a method for the determination of gold naturally starts with that element. The weighed amount of metal, contained in a clean porcelain crucible, can be readily brought into solution with the aid of chlorine water or aqua regia and the excess of the solvent evaporated off upon the water bath. CHAPTER IV. BERYLLIUM; MAGNESIUM; CALCIUM; STRONTIUM; BARIUM. BERYLLIUM. Ammonium Beryllium Phosphate. BERYLLIUM, like calcium, strontium and barium, cannot be es- timated by precipitation as a double ammonium phosphate and subsequent ignition, as are magnesium, zinc and cadmium. The precipitated ammonium beryllium phosphate, as has been shown by Austin,* always contains triberyllium phosphate. The Conversion of Beryllium Chloride to Beryllium Oxide. Havens f has shown that small amounts of beryllium chloride may be easily converted to the oxide, without precipitation and filtration, by treatment with nitric acid and ignition. The solution of the chloride is evaporated just to dryness on a radiator, care being taken not to heat to the volatilizing point of the salt, a few drops of concentrated nitric acid are added, the liquid is evaporated, and the residue heated gently at first and finally in the flame of the blast lamp. This conversion of beryllium chloride to beryllium nitrate may be carried on in platinum without attacking that metal appreciably, provided care be taken to remove all excess of hydrochloric acid and to add the nitric acid to the dry residue. Conversion of Beryllium Chloride to the Oxide. BeO taken in solution as BeCla. BeO found. Error. grin. grm. gnu* 0.0483 o . 0483 0.1076 0.0481 0.0483 0.1085 0.0002 o.oooo +0.0009 * Martha Austin, Am. Jour. Sci., [4], viii, 207. t F. S. Havens, Am. Jour. Sci., [4], iv, 112. 153 154 METHODS IN CHEMICAL ANALYSIS The Separation of Beryllium Oxide from Ferric Oxide. Havens and Way* have separated beryllium oxide from ferric oxide by volatilization of the latter in a stream of gaseous hydro- gen chloride charged with a little free chlorine, with care to avoid mechanical loss through too rapid volatilization of the iron.f MAGNESIUM. The Determination of Magnesium by Precipitation and Ignition of Ammonium Magnesium Carbonate. According to Schaffgotsch,J the very concentrated solution of the sulphates, nitrates or chlorides of magnesium, sodium and potassium is treated with a concentrated solution of ammo- nium carbonate. The voluminous precipitate which first falls is acted upon by an excess of the precipitant, sometimes dissolving completely, and crystalline ammonium magnesium carbonate, MgCO3.(NH 4 )2CO3.4H 2 O, is soon formed; after standing twenty- four hours the precipitate is filtered off, washed with the con- centrated ammoniacal solution of ammonium carbonate, dried and strongly ignited. In the absence of salts of potassium, the residue is weighed at once as magnesium oxide, and from the fil- trate sodium salts are recovered by evaporation. When a salt of potassium is originally present, with or without a salt of sodium, the ignited magnesium oxide is to be washed out and again ignited before weighing, and the washings are to be added to the filtrate containing the greater part of the alkalies. Gooch and Eddy have shown that ammonium magnesium carbonate is noticeably soluble in Schaffgotsch's solution of full strength, and rather more so in the same reagent of half strength, and that an exact separation of magnesium from the alkalies, in solutions of reasonable volume, cannot be made without modi- fication of the method. By suitable addition of alcohol, how- ever, it has been found possible to make the precipitation com- plete and to effect the separation of magnesium from small amounts of alkali salt in one operation ; when considerable quan- * Franke Stuart Havens and Arthur Fitch Way, Am. Jour. Sci., [4], viii, 217. t See page 507. t Ann. Phys., civ, 482 (1858). F. A. Gooch and Ernest A. Eddy, Am. Jour. Sci., [4], xxv, 444. MAGNESIUM 155 titles of alkali salt are present the separation may be made by two treatments. The solution containing the salts of magnesium and the alkalies is brought to a volume of about 50 cm. 3 and an equal amount of absolute alcohol is added, precipitation is made by addition of 50 cm. 3 of the saturated ammoniacal ammonium carbonate solu- tion containing 50 per cent alcohol, and the mixture is allowed to stand twenty minutes after stirring for five minutes. If the amount of alkali salt originally present is small, not exceeding o.i grm., the precipitate may be collected on asbestos in a perfo- rated crucible, washed with the precipitant, dried, ignited, and weighed as magnesium oxide. When the amount of alkali salt originally present is larger, the precipitate may be freed from traces of the alkali salt by pouring off the supernatant liquid through the prepared asbestos filter, dissolving the precipitate, and precipitating ammonium magnesium carbonate as at first. This second precipitate, collected upon the filter originally used, leaves upon ignition practically pure magnesium oxide. The accompanying figures show excellent results. Precipitation of Magnesium Ammonium Carbonate and Weighing of the Oxide. i i .2 _i 'o a 1 NaCl taken KCI taken. NH 4 C1 tak MgO weigh Error MgO Jfc *l| gctfS Volume of water solution. Volume of alcohol added. Volume of precipitant Volume of s tion used i washing. grm. grm. grm. grm. grm. grm. grm. cm. 3 cm. s cm. 3 cm.* Single precipitation. 0.1444 0.1443 o.oooi o . oooo 50 50 50 0.1444 o . 1440 0.0004 O.OO02 50 50 50 . .. o . 1444 O. I 0.1445 +O.OOOI O.OOOI 50 50 50 50 0.1444 O.I 0.1444 o.oooo O.OOOI 50 50 50 50 0.1444 O.I 0.1445 +O.OOOI O.OOO2 50 50 50 50 0.1444 O. I . . . 0.1449 +0.0005 O.OOOI 50 50 5 50 0.1444 O. 2 o . 1449 +0.0005 0.0002 50 50 50 50 0.1444 0.2 o. 1461 +0.0017 . OOOO 50 50 50 50 0.1444 3-o 0.1444 o.oooo O.OOOI 50 50 5 50- 0.1444 3-0 0.1447 +0.0003 O.OOO2 50 50 50 50 Double precipitation. 0.1444 0.1444 0.2 O.2 o. 1446 0.1442 +O . OOO2 O.OO02 O.OOO2 0.0002 50 50 50 50 50 50 50 50 156 METHODS IN CHEMICAL ANALYSIS The Determination of Magnesium as the Pyrophosphate. As Neubauer* had previously pointed out, the ideal ammonium magnesium phosphate, NH 4 MgPO 4 , which yields upon ignition the pyrophosphate, Mg 2 P2O 7 , may, according to conditions of precipitation, be contaminated by the trimagnesic phosphate, Mg 3 P2O 8 , or by a double phosphate, (NH 4 )4Mg(PO4)2, which upon ignition leaves magnesium metaphosphate. Gooch and Austin f have emphasized the effect of ammonium salts in bringing about the contamination of the precipitate which results in the forma- tion of the metaphosphate on ignition, and have also pointed out that the use of strongly ammoniacal solutions and strongly ammoniacal wash-water is distinctly disadvantageous, as well as inconvenient. Having shown that as little as o.oooi grm. of magnesium oxide may be precipitated in 500 cm. 3 of faintly ammoniacal solution (even when containing as much as 60 grm. of ammonium chloride or 100 cm. 3 of a saturated solution of ammonium oxalate), the authors recommend the use of faintly ammoniacal solutions and wash -water, and, to prevent as much as may be the formation of the phosphate containing excess of the ammonium ion, restriction of the amount of ammonium chlo- ride present. When ammonium salts are present in quantity, as is the case in the ordinary course of analysis, the precipitate first thrown down by addition of ammonium sodium phosphate and ammonia in faint but distinct excess is settled, and the supernatant liquid is poured off through the filter used sub- sequently in collecting the precipitate. The precipitate is dis- solved in the least possible amount of hydrochloric acid and thrown down again from the diluted solution by ammonia in slight excess. For safety, a little ammonium sodium phosphate may also be added. The precipitate is filtered off and washed with faintly but distinctly ammoniacal water, and, to avoid reduction, the ignition is made slowly and carefully so that all ammonia is expelled before the temperature is raised to redness. When filtrations are made on asbestos in the perforated crucible, as was done in the experiments the results of which are recorded below, it is well to cap the crucible and moisten the * Zeit. Angew. Chem., 1896, 439. t F. A. Gooch and Martha Austin, Am. Jour. Sci., [4], vii, 187. MAGNESIUM 157 precipitate upon the felt with a drop of a saturated solution of ammonium nitrate before proceeding to dry and ignite. The accompanying figures show the accuracy which may be expected when precipitations are made in presence of varying amounts of ammonium salts. Effect of Ammonium Salts in Respect to the Constitution of the Precipitate. Mg 2 P 2 7 correspond- ing to Mg(N0 3 ) 2 M g2 P 2 7 found. Error in terms of Mg 2 P 2 7 . Error in terms of MgO. NH 4 C1 present. HNa- NH 4 P0 4 4 H 2 Volume. taken. I. II. I. II. grm. grm. grm. grm. grm. grm. grm. cm. 3 cm. J 0-S3II 0.5312 +O.OOOI O.OOOO * * 2-5 150 IOO 0-53U 0-53" O.OOOO O.OOOO * * 2-5 150 IOO 0-53II 0.5346 +0.0035 +0.0013 2 2 2-5 150 IOO 0-53II 0.5348 +0.0037 +0.0014 2 2 2-5 150 IOO 0-53II 0.5383 +0.0072 +0.0026 5 5 2.5 ISO IOO 0-53II 0.5368 +0.0057 +O.OO2I 5 5 2.5 150 IOO 0-53II 0.5376 +0.0065 +0.0023 IO IO 2.5 2OO IOO 0-53II 0-5395 +0.0084 +0.0030 IO IO 2.5 2OO IOO 0-53II 0.5396 .+0.0085 +0.0031 60 5 2.5 250 IOO 0.53H 0.5389 +0.0078 +0.0028 60 5 2.5 250 IOO * Probably less than i grm. Gooch and Austin point out that good results are obtained in one precipitation by the method of Wolcott Gibbs, proposed many years ago.* According to this method the boiling solution of the magnesium salt is treated with ammonium sodium phos- phate, and ammonia is added after cooling. Even in presence of considerable amounts of ammonium chloride this process yields a phosphate of nearly ideal constitution if only the boiling be prolonged from three to five minutes. The greater part of the ammonium magnesium phosphate about 90 per cent forms in this process before free ammonia is added, and the ammonium which enters the phosphate thus formed is derived from the micro- cosmic salt, which must become correspondingly acidic. Under these conditions, the tendency to form an insoluble ammonium magnesium phosphate, richer in ammonium and poorer in mag- nesia than the normal salt, is slight. In the process of Gibbs, as well as in the modified precipitation process, the use of the faintly ammoniacal solution and wash-water is sufficient and advantageous. * Am. Jour. Sci., [3], v, 114. 158 METHODS IN CHEMICAL ANALYSIS The Ar senate Process for the Separation of Magnesium and the Alkalies. Browning and Drushel* have taken advantage of the insolu- bility of ammonium magnesium arsenate in ammoniacal solution, with the reducibility of arsenic acid by hydrobromic acid and volatility of arsenious bromide, to perfect a method for the sepa- ration of magnesia from the alkali metals, and the determination of these elements. From a solution of the chlorides of magne- sium and potassium or sodium, the magnesium may be precipi- tated in a distinctly but not strongly ammoniacal solution by 40 per cent to 80 per cent excess of ammonium arsenate, with brisk stirring. When only a small amount of magnesium is present in a relatively large amount of solution, the precipitate forms slowly and becomes complete only on long standing. The precipitation of amounts of magnesium so small as not to be precipitated immediately by ammonium arsenate may be brought about by freezing the solution, a process previously shown to be applicable f in the precipitation of small amounts of arsenic acid by magnesia mixture. Precipitation may also be hastened by adding alcohol amounting to 15 per cent to 20 per cent of the mix- ture and filtering as soon as the precipitate settles completely. The precipitate is collected under moderate pressure in an ignited and weighed perforated crucible containing a close felt of fine as- bestos. It is washed with 40 cm. 3 to 50 cm. 3 of ammoniacal water, after which it is dried at 125 to 140 and carefully ignited and weighed as magnesium pyroarsenate. It is shown elsewhere { that arsenic acid may be reduced and volatilized by the action of hydrochloric acid and potassium bromide. The removal of the arsenic acid from the alkali salts is easily accomplished by the similar procedure of treating the mixture with hydrobromic acid or with ammonium bromide and hydrochloric acid and evaporating in an open dish under a good hood. The complete method as recommended for the estimation of magnesium and its removal from the alkalies, and the sub- sequent estimation of the alkalies, is as follows : The magnesium is precipitated in a distinctly but not strongly ammoniacal solution by a 40 per cent to 80 per cent excess of * Philip E. Browning and W. A. Drushel, Am. Jour. Sci., [4], xxiii, 293. t See page 290. i See page 316. MAGNESIUM ammonium arsenate. The completeness of the precipitation may be hastened by freezing the solution in an ice-and-salt mix- ture or by adding alcohol to about 15 per cent to 20 per cent of the total volume of the solution, which may vary from 100 cm. 3 to 250 cm. 3 according to the amounts of salt present. The mag- nesium arsenate obtained is filtered on an asbestos felt contained in a perforated platinum crucible, the crucible and felt having been previously ignited and weighed, and is dried, ignited and weighed as the pyroarsenate. Experimental results follow. Magnesium and the Alkalies. (NH 4 ),AsO used calculated as As 2 Oa. grm. Dilu- tion. cm. 3 NaCl or KC1 converted to Na 2 SO 4 or K 2 SO 4 and calculated as Na,O or K 2 O. MgCl 2 converted into Mg 2 As 2 O 7 and calculated as MgO. Taken, grm. Found, grm. Error, grm. Taken, grm. Found, grm. Error, grm. Precipitate stood 12 to 24 hours. O. I IOO 0.1194 o. 1191 -0.0003 0.0199 0.0197 O.OOO2 O. 2 150 0.1194 o. 1196 + O.O002 0.0399 0.0397 O.OOO2 0-45 250 0.1194 0.1195 + O.OOOI o . 0998 o . 0998 o . oooo o-45 250 0.1194 0.1194 o.oooo 0.0998 0.0997 o.oooi 0-45 250 0.2389 0.2385 0.0004 o . 0998 0.0999 +O.OOOI 0.4 250 0.0478 0.0481 +0.0003 o. 1198 0.1193 0.0005 o-35 250 0.0956 0.0957 +O . OOOI o . 0998 o . 0996 O.OOO2 o-35 250 0.0956 0.0957 +0.0001 o . 0998 0.0994 0.0004 o-4S 250 0.0909 0.0915 +0.0006 0.0998 0.0993 0.0005 O. I IOO 0-0545 o . 0549 +0.0004 0.0006 o . 0004 O.OOO2 Precipitation hastened by alcohol. O. I IOO o. 1181 o. 1184 +0.0003 o . 0040 0.0038 O.OOO2 O. I IOO 0.1181 0.1184 +0.0003 o . 0040 0.0038 O.OO02 O. I IOO 0.0040 0.0038 O.OOO2 Precipitation hastened by freezing. Oil IOO 0.1181 o. 1184 +0.0003 o . 0040 o . 0040 O.OOOO 0.2 IOO 0.1181 0.1183 +O.OOO2 0.0040 0.0039 O.OOOI 0-45 250 0.1181 0.1179, O.OOO2 O. IOO2 o. 1004 + O.OOO2 , The nitrate is transferred from the filter flask to a platinum dish, and after the addition of 10 cm. 3 of hydrochloric acid (sp. gr. 1. 20) and about the same amount of hydrobromic acid (sp. gr. 1.3), or I to 3 grm. of ammonium bromide, is evaporated to dry- 160 METHODS IN CHEMICAL ANALYSIS ness under a draft hood. The residue is gently ignited to re- move the ammonium salts and transferred to a weighed platinum crucible with a small amount of water. A little sulphuric acid [i : i] is added, and the solution evaporated to remove the water and excess of sulphuric acid, by placing the crucible on a triangle in a porcelain crucible used as a radiator. After the sulphuric acid has ceased to fume, the crucible is removed from the radi- ator, and after ignition at the full heat of the Bunsen burner the alkali is weighed as the normal sulphate. CALCIUM; STRONTIUM; BARIUM. The Detection of Barium and Strontium, Associated with Calcium and Lead. In the ordinary procedure of qualitative analysis the alkali earth elements are usually separated by ammonium carbonate, after hydrogen sulphide and ammonium hydroxide have been used to remove the greater number of the bases. It has long been observed that a considerable part of the alkali earth, espe- cially barium and strontium, fails to appear when the ammo- nium carbonate is added. The reasons given for this loss have been the oxidation of hydrogen sulphide or other sulphides to sulphates and the consequent precipitation of the alkali earth sulphates, the formation in alkaline solution of carbonates and the consequent precipitation of the carbonates, and the tendency of the large amounts of ammonium salts which collect during the analysis to interfere with the precipitation of the alkali earth carbonates by ammonium carbonate. Various precautions have been suggested to avoid these sources of error, such as the prompt removal of the excess of hydrogen sulphide by boiling, the use of freshly prepared hydroxides free from carbonate, and the removal of the ammonium salts by ignition before attempting to precipi- tate the alkali earth carbonates. To obviate these difficulties, Browning and Blumenthal * have suggested the precipitation of the insoluble sulphates after the removal by hydrochloric acid of mercury in the mercurous condi- tion, silver, and that amount of lead which may be precipitated; removal of the insoluble lead sulphate by treatment of the * Philip E. Browning and Philip L. Blumenthal, Am. Jour. Sci., [4], xxxii, 246. CALCIUM; STRONTIUM; BARIUM l6l precipitate with ammonium acetate; reduction of the remaining insoluble sulphates by ignition with carbon; treatment of the residue with acetic acid ; and testing of the solution for barium, strontium and calcium present as soluble acetates. The follow- ing method is suggested: The solution (about 10 cm. 3 ), which may contain mercury in the mercurous condition, silver, lead, barium, strontium and calcium, besides other elements, is treated with hydrochloric acid in faint excess and the precipitated chlo- rides are filtered off. To the filtrate are added about 5 grm. of ammonium acetate, and a 10 per cent solution of ammonium sulphate to complete precipitation. After gentle warming, the alkali earth sulphates are filtered off and washed with a saturated solution of ammonium acetate until the washings give no test for lead by hydrogen sulphide. The filtrate and washings are reserved for treatment by the ordinary course of analysis. To the precipitated sulphates on the paper a small amount of pure sugar carbon is added, the paper is rolled up, and the mass placed either in a porcelain crucible with a cover, or in a closed glass tube, and heated to full redness for a few minutes. The fused mass is treated with about 5 cm. 3 of 50 per cent acetic acid and warmed, when, if the alkali earth elements are present, an odor of hydrogen sulphide will generally be evident. The extract is thrown upon a filter and the residue washed with about 5 cm. 3 of water. The filtrate containing acid and water is treated with a few drops of a solution of potassium chromate to test for barium. The barium chromate is removed by filtration, and the filtrate boiled with sodium carbonate to precipitate stron- tium and calcium as the carbonates. If the precipitate of the carbonates is very small, it may be dissolved in hydrochloric acid and tested spectroscopically. If, however, it is not too minute in quantity, it should be dissolved in nitric acid after careful washing, and the strontium and calcium separated by dehydra- tion with amyl alcohol. The results follow in the table. All tests for strontium and calcium were confirmed by the spectroscope. From these results it would appear that these tests for barium arid strontium are effective to at least a milligram of each ele- ment and may with advantage precede the group precipitation by hydrogen sulphide in the ordinary course of qualitative analysis. 162 METHODS IN CHEMICAL ANALYSIS Pb present, grin. Ba present, grm. Sr present, gnu. Ca present, grin. Indications. 0.0500 0.0500 0.0500 0.0500 Good tests for all. 0.0250 0.0250 0.0250 0.0250 j Good tests f or Pb, Ba and Sr. Ca faint. O.OIOO O.OIOO O.OIOO 0.0100 1 Pb and Ba good. Sr fair. Ca doubtful. 0.0050 0.0050 o . 0050 0.0050 1 Pb and Ba good. Sr fair. O.OOIO O.OOIO O.OOIO O.OOIO < Pb good. Ba fair. Sr faint. O. IOOO 0.0050 Pb and Ba good. O . IOOO O.OOIO Pb good. Ba faint. The Separation of Barium, Strontium and Calcium by the Action of Amyl Alcohol on the Nitrates. Following in general known procedure for the separation of sodium and potassium from lithium by the use of amyl alcohol,* Browning t has developed methods for the detection of calcium and strontium in association, the separation and estimation of strontium associated with calcium, and the separation of barium associated with calcium or with strontium. Experiments with pure strontium nitrate show that when the dry salt is treated by dissolving in a few drops of water, adding amyl alcohol, boiling until the water is expelled and the boiling point rises to the normal boiling temperature of the alcohol (i28-l3O), filtering upon asbestos in the filtering crucible, washing with small amounts of previously boiled amyl alcohol, and heating to 150 in an air bath, nearly the entire original amount is recovered. The quantity which remains in solution in the dehydrated alcohol amounts very regularly to 0.0008 grm. of the nitrate or 0.0004 grm. of the oxide for every 10 cm. 3 of liquid. When calcium nitrate in water solution is similarly treated by boiling with amyl alcohol, the salt passes into solution with the exception of minute portions, not exceeding altogether 0.0003 g rm - or 0.0004 grm., which separate on the surface of the container. This very slight residue (apparently the calcium salt of an acid formed by the action of nitric acid upon amyl alcohol), when dried, dissolved in dilute nitric acid, and again treated with amyl * Gooch, Am. Chem. Jour., ix, 33. f P, E. Browning, Am. Jour. Sci., [3], xliii, 50. CALCIUM; STRONTIUM; BARIUM I6 3 Detection of Strontium and Calcium. alcohol, separates out in the boiling, but, if ignited and then dissolved in a drop of dilute nitric acid, is not precipitated by subsequent boiling with amyl alcohol. To detect strontium and calcium associated in the form of nitrates, the mixture, not exceeding 0.2 grm. (that being the limit of the solubility of calcium nitrate in 5 cm. 3 of amyl alcohol) is put into a test tube and dissolved in a few drops of water, 5 cm. 3 of amyl alcohol are added, and the boiling is carried on until the normal boiling point of the alcohol, I28-I3O, is reached. If strontium is present to the amount of o.ooi grm. of the oxide, a very decided separation takes place. If the amount is smaller, it cannot be readily distinguished from the residual spots deposited on the bottom of the tube by the calcium salt. The alcohol containing the calcium salt dissolved is decanted upon a dry filter paper in a dry funnel and the residue washed in the tube with about 5 cm. 3 of absolute ethyl alcohol, this also being filtered into the tube containing the amyl alcohol. The filtrate is reserved to be tested for calcium. The residue, if so small that it may be a calcium deposit, is dried gently, ignited by agitating the tube over a flame, and dissolved in a drop of dilute nitric acid. To this solution 5 cm. 3 of amyl alcohol are added and the boiling is repeated. Any amount of strontium above 0.0005 rm - of tne oxide separates out distinctly, while the slight calcium residue does not reappear, as is shown in the accompanying record. SrO taken, grm. ' Ca(NO 3 ) 2 taken, grm. . Deposit after first boiling. Deposit after second boiling. O.I Trace. None. O.2 Slight. None. 0.2 Slight. None. 0.2 Distinct. Faintest trace. 0.0003 . . Faint trace. Faint trace. 0.0003 Faint trace. Faintest trace. 0.0005 Distinct. Distinct. 0.0005 Distinct. Distinct. O.OOIO Distinct. Distinct. O.OOIO O.2 Distinct. Distinct. 0.0005 O.I Distinct. Faintest trace. 0.0007 O.I Distinct. Faint trace. 0.0008 0.05 Distinct. Distinct. The test for calcium is made upon the filtrate and washings after the first boiling, by adding to the clear liquid about 2 cm. 3 164 METHODS IN CHEMICAL ANALYSIS of dilute sulphuric acid. In five minutes or less, calcium, if present to an amount exceeding o.oooi grm., appears as a light, flocky precipitate, different in character and easily distinguishable from the faint cloudiness, gathering to a minute and granular precipitate, which results from the presence of a trace of stron- tium salt not precipitated in the boiling process. Following are the tests of this method. Sr(NO 3 )j taken, grm. CaO taken, grm. Result. O.OOI O.I O.2 Faint granular cloudiness. Faint granular cloudiness. Faint granular cloudiness. O.I O. I O.I 0.2 O.OOI 0.0005 O.OOO2 O.OOOI 0.00005 O.OOO5 O.OOO2 O.OOOI O.OOO2 Decided flocky floating masses. Decided flocky floating masses. Decided flocky floating masses. Plain flocky floating masses. Faint flocky floating masses. Decided flocky floating masses. Decided flocky floating masses. Plain flocky floating masses. Decided flocky floating masses. Strontium and Calcium. For the quantitative estimation of strontium and Separation and . Estimation of calcium,* the dry nitrates of these elements are dissolved in the least possible amount of water contained in a small beaker (50 cm. 3 to 100 cm. 3 ), a suitable amount of amyl alcohol (10 cm. 3 to 30 cm. 3 ) is added, the beaker is heated upon a wide piece of asbestos board so that inflammable fumes may not reach the flame below, and the mixture is boiled with a thermometer inserted until the normal boiling point of the alcohol (128 to 130) is reached. From the precipitated strontium nitrate the alcoholic solution is decanted through a weighed filtering crucible and asbestos felt, and the residue, dried at a gentle heat over a radiator, is dissolved in a drop or two of dilute nitric acid. Amyl alcohol is again added and the boiling repeated, experience having shown that the residue of the first boiling is apt to retain an appreciable amount of the calcium salt. The precipitate of the second boiling is filtered off upon the felt through which the solution had previously been decanted, washed with amyl alcohol, dried at 150, and weighed as strontium nitrate Sr(NO 3 )2. Correction for the solubility of strontium nitrate in amyl alcohol is made according to the * P. E. Browning, Am. Jour. Sci., [3], xliii, 50; and xliv, 462. CALCIUM; STRONTIUM; BARIUM quantity of that reagent left after the boilings, 0.0008 grm. of the nitrate or 0.0004 grm. of the oxide for every 10 cm. 3 of amyl alcohol decanted or filtered off, exclusive of that used in washing, in which process no appreciable amount of the stron- tium salt is dissolved. The calcium in the filtrate is determined as the sulphate by evaporation of the solution, ignition, treatment with sulphuric acid, and a final ignition. From the apparent amount of calcium sulphate found a correction of 0.0005 g rm - for the included stron- tium sulphate is to be subtracted. Results of experimental tests are given in the tabular statement. Separation and Estimation of Stronium and Calcium. Final volumes 25 cm. 3 ; two treatments. SrO taken, grm. SrO found.* grm. Error, grm. CaO taken, grm. CaO found, f grm. Error, grm. 0.0148 0-0155 +0.0007 0.0256 0.0254 0.0002 0.0183 0.0183 O.OOOO o. 1030 0.1015 0.0015 0.0364 0.0366 +O.O002 0.0516 % 0.0511 0.0005 0.0365 0.0365 O.OOOO 0.0515 0.0513 O.OOO2 0.0493 o . 0494 +0.0001 0.0515 0.0502 0.0013 0.0497 0.0497 . 0000 0.0519 0.0511 O.OOOS 0.0497 0.0503 +0.0006 o . 0249 0.0245 O.OOO4 0.0729 0.0732 +o . 0003 0.0257 0.0251 O.OOO6 0.0730 0.0732 +O.OO02 0.0255 0-0255 O.OOOO 0.0744 0.0744 O.OOOO 0.0258 0.0260 +0.0002 0.0912 O.O9IO O.OOO2 0.1286 o. 1276 o.ooio * Corrected by addition of 0.0020 grm. for two treatments in final volumes of 25 cm. 3 , t Corrected by subtraction of 0.0035 grm. for SrSO 4 included in CaSO 4 obtained by evaporation and ignition. Final volumes 8 cm. 3 ; two treatments. SrO taken, grm. SrO found.* grm. Error, grm. CaO taken, grm. CaO found. f grm. Error, grm. 0.0570 0.0571 +O.OOOI 0-0534 0.0536 +0.0002 0-0573 0-0573 O.OOOO 0.0534 0-0539 +0.0005 0.0285 0.0280 0.0005 0.0272 0.0272 O.OOOO 0.0568 0.0566 O.OOO2 0-0535 0-0533 O.OOO2 0.0568 0.0567 0.0001 0-0533 0.0531 O.OOO2 0.0288 0.0286 +O . OOO2 0.0271 0.0268 0.0003 o. 1420 o. 1422 +O.OOO2 0.0535 o . 0540 +0.0005 O.I4I9 o. 1422 +0.0003 0.0665 0.0665 O.OOOO O.H35 0.1138 0.0003 o. 1066 o. 1066 O.OOOO O.II37 0.1132 0.0005 0.1064 0.1066 +O.OOO2 * Corrected by addition of 0.0006 grm. for two treatments in final volumes of 8 cm. 3 . t Corrected for o.ooio grm. of SrSO 4 included in CaS0 4 obtained by evaporation and ignition. i66 METHODS IN CHEMICAL ANALYSIS Estimation of Barium and Calcium. Procedure similar to that employed in the sepa- ration and estimation of strontium and calcium by the action of amyl alcohol on the nitrates may be applied to the separation of barium and calcium,* but, barium nitrate being almost entirely insoluble in the amyl alcohol, there is in this case no advantage in keeping the volume of alcohol at the lowest point. A convenient volume at the beginning of the dehydration is 30 cm. 3 , and the results of one treatment are fully as satisfactory as those of the double treatment. In separating barium from calcium by this method, the dry nitrates are treated in ioo-cm. 3 beakers by dissolving in a few drops of water, adding 30 cm. 3 of amyl alcohol, boiling until the normal boiling point of the alcohol is reached (128 to 130), filtering by means of the perforated filtering crucible fitted with the asbestos felt, washing with previously boiled amyl alcohol, drying at 150 and weighing. The test results show the exactness of the method. Estimation of Barium and Calcium. BaO taken, grin. BaO found, grm. Error, grm. CaO taken, grm. CaO found, grm. Error, grm. o. 1410 o. 1406 0.0004 O.OII2 O.OII4 +O.OOO2 0.1300 O.I30I +0.0001 0.0926 0.0926 O.OOOO 0.1043 0.1049 +O.OOO6 0.0741 0.0736 -0.0005 0.0781 0.0781 o.oooo 0.0556 0-0554 O.OOO2 0.0525 0.0526 +O.OOOI 0.0373 0.0372 o.oooi Estimation of The dry nitrates of the three elements, barium, Barium and strontium and calcium, are dissolved in the least pos- Strontium ., , . -11 / together, and sible amount of water, a suitable amount (15 cm. 3 of Calcium. O ^Q cm. 3 ) of amyl alcohol is added, and the mixture boiled until the normal boiling temperature of the alcohol is reached. The alcoholic solution is decanted from the precipi- tated barium nitrate and strontium nitrate through a weighed filtering crucible and asbestos felt. The residue is dried over a radiator, dissolved in a drop or two of dilute nitric acid, and again treated as before with amyl alcohol. The precipitate of the second boiling is filtered off upon the asbestos felt previously used, washed with amyl alcohol, dried at 150 and weighed. The calcium is determined as the sulphate in the combined * P. E. Browning, Am. Jour. Sci., [3], xliii, 314. CALCIUM; STRONTIUM; BARIUM I6 7 filtrates and washings. Results, corrected for the solubility of the strontium nitrate (0.0008 grm. to 10 cm. 3 of alcohol decanted or filtered, exclusive of washings) and for the contamination of the calcium sulphate by strontium sulphate (0.0005 grm. for IO cm. 3 of alcohol), are given in the table. Estimation of Barium and Strontium, and Calcium. Ba(NO 3 ) 2 Ba(N0 3 ) 2 Error and Sr(N0 3 ) 2 taken. and Sr(N0 3 ) 2 found and corrected. Error in nitrates. averaged and calculated as oxide. CaO taken. CaO found. Error. grm. grm. grm. grm. grm. grm. grm. 0.3941 0-3945 +0 . 0004* +O.OOO2 0.0283 0.0277 0.0006 0.1436 0.1442 +0.0006* +0.0003 0.0568 0.0558 o.ooio 0.3163 0.3IS2 O.OOII* 0.0006 0.0284 0.0274 O.OOIO 0.1978 0.1987 +0.0009* +0 . 0005 0.0285 0.0280 0.0005 0.1948 0.1932 o.ooi6f 0.0008 0.0833 0.0835 +O.OOO2 0.1971 0.1971 o.oooo* O.OOOO o . 0830 0.0817 0.0013 0.1973 0.1960 0.0013* 0.0007 0.0830 0.0824! 0.0006 0.1959 0.1970 +O.OOII* +0.0005 0.0830 0.0819 O.OOII . 0.1971 0.1963 o.oooSf o . 0004 0.0834 0.0831^ 0.0003 * Final volume in each of two treatments, 30 cm. s . t Final volume in each of two treatments, 15 cm. 3 . J CaS04 precipitated and filtered: in other experiments obtained by evaporation. The Separation of Barium and Strontium by the Action of Amy I Alcohol on the Bromides * Methods upon which dependence can be placed for the separa- tion of barium and strontium are few in number. The differences in solubility of the bromides of barium and strontium in amyl alcohol provide a method for a comparatively good separation and, when properly corrected, an exact determination of these elements. Anhydrous barium bromide dissolves in amyl alcohol to the extent of about 0.0013 grm. in 10 cm. 3 , while the same quantity of amyl alcohol will take into solution approximately 0.2 grm. of strontium bromide. When a mixture of the dry salts is dissolved in water, and the water removed by boiling in a suitable amount of amyl alcohol, the barium bromide becomes nearly insoluble, while nearly all the strontium bromide goes into solution. The insoluble barium bromide cannot, however, be fil- tered off, washed, and dried to constant weight without decom- position, so it becomes necessary to determine the barium in * Philip E. Browning, Am. Jour. Sci., [3], xliv, 459. i68 METHODS IN CHEMICAL ANALYSIS some other form. Moreover, a single treatment by boiling does not remove the strontium completely from the precipitated barium bromide. To separate barium and strontium taken as the bromides, therefore, the mixed salts are treated in a beaker by dissolving in the least possible amount of water, adding 10 cm. 3 of amyl alcohol, and boiling until the temperature rises to the normal boiling point of the alcohol (128 to 130). The solution is decanted through a weighed filtering crucible and asbestos felt. The gently dried residue is dissolved in the minimum amount of dilute hydrobromic acid and again boiled with 10 cm. 3 of amyl alcohol. The precipitate is filtered off upon the asbestos felt previously used and dissolved from the felt with water. From the solution the barium is precipitated as the sulphate, which is dried, ignited and weighed. The strontium is precipitated from the united filtrates and washings by sulphuric acid after the addition of ethyl alcohol to insure thorough mixture. The re- sults of tests of the method are given below. Estimation of Barium and Strontium. BaO taken, grm. BaO found.* grm. Error, grm. SrO taken, grm. SrO found. t grm. Error, grm. O.I2I2 0. 1219 +0.0007 0.1068 0.1071 +0 . 0003 o. 1215 0. 1219 +0.0004 0.0358 0-0359 +O.OOOI O.I22O O.I22I +O.OOOI 0.0353 0.0347 0.0006 O. 1212 O. I22O +0.0008 0.0363 0.0358 0.0005 o. 1219 O.I22I +O.OOO2 0.0361 0.0354 0.0007 O.I2II 0.1218 +O.OOO7 o. 1126 o. 1116 O.OOIO O.I3I9 0.1319 . 0000 0.0577 o . 0586 +0.0009 o . 0496 0.0492 0.0004 0.0574 0.0579 +0.0005 * Corrected for 0.0025 BaO corresponding to the bromide dissolved in two treatments with 10 cm. 3 of amyl alcohol. t Corrected by subtraction of 0.0040 grm. for the barium sulphate corresponding to dissolved barium bromide. The method is rapid, and, while the correction to be applied, owing to the solubility of the barium salt, is large, it is definite. The Estimation of Barium as the Sulphate. in Presence of ^ n t ^ ie ordinary mode of precipitating barium as Hydrochloric barium sulphate, three conditions are carefully ob- served, absence of excess of acid, slow mixing of the reagents, and standing twelve hours, or until the precipitate has completely subsided before filtration. Usually, in this CALCIUM; STRONTIUM; BARIUM 169 process, the precipitate is thrown out in a finely divided, milky condition and settles very slowly. Mar* has observed, however, that the presence of hydrochloric acid influences the form in which the sulphate is deposited without affecting the complete- ness of precipitation, provided a sufficient excess of sulphuric acid is also present. From a solution of 0.5 grm. of barium chloride in 400 cm. 3 of water the precipitate appears immediately upon the addition of sulphuric acid, settling slowly, and this condition prevails also, even in hot solutions, when only one or two cubic centimeters of hydrochloric acid have been previously added. With 10 cm. 3 to 15 cm. 3 of strong hydrochloric acid in the solution heated to 85 or 90, the precipitate settles clear in ten or twelve minutes, and is in excellent condition for filtration. When the solution contains 50 cm. 3 of the acid, the precipitate settles clear in five minutes. Upon adding the sulphuric acid to such very acid solutions, no precipitate shows for a moment, but then it sepa- rates in beautiful crystalline condition and falls almost immedi- ately. It can be safely filtered with or without pressure in ten minutes. In an instance cited, 2 grm. of barium chlqride were precipitated in the presence of 30 cm. 3 of hydrochloric acid, the precipitate was allowed to settle clear, and was then filtered and washed, the whole operation being completed in seven minutes. This rapid subsidence of the precipitate is seen in hot solutions only, 75 being the lowest temperature compatible with the at- tainment of good results, and 85 to 90 better. Quantitative experiments, quoted below, show that precipi- tation is practically complete in 400 cm. 3 of solution when sul- phuric acid is added to the amount of 10 cm. 3 of the I : 3 dilute acid (sp. gr. 1.28 one of acid in a total volume of four) in presence of hydrochloric acid in amounts up to 150 cm. 3 of the concentrated acid, sp. gr. 1.20. Considerable amounts of barium are precipitated at once, but when only a few milligrams are present complete formation of the precipitate requires more time. Two or three hours are in every case sufficient. In filtering on asbestos in the perforated crucible, as was done, care must be taken to use a very close felt, on account of the very minutely crystalline nature of the precipitate. * F. W. Mar, Am. Jour. Sci., [3], xli, 288. METHODS IN CHEMICAL ANALYSIS Precipitation in Presence of Hydrochloric Acid. Bad,. 2H a O taken. grm. Total volume. cm. HC1 [sp.gr. 1.20). cm.* Dilute H 2 S0 4 (sp.gr. 1.28). cm. Time between precipita- tion and filtration. min. BaSO 4 found, grm. *rror. grm. 0.0050 400 IS IO u ( 0.0023 ( 0.0043 0.0025 0.0005 0.0050 400 15 10 5 0.0031 0.0017 0.0050 400 15 IO 10 O.0040 0.0008 O.OIOO 400 IS IO IO 0.0078 0.0017 0.0100 400 IS IO 15 0.0085 o.ooio O.OIOO 400 IS 10 30 0.0083 O.OOI2 O.OIOO 400 IS 10 60 0.0087 0.0008 . 0.0030 400 is IO 1 20 O.OO24 0.0005 0.0050 400 15 10 150 0.0046 O.OOO2 , 0.5014 400 15 10 10 0.4785 0.0003 0.2227 400 15 IO IO O.2I22 0.0005 0.5003 400 is IO IO 0-4773 0.0005 0.5046 400 15 10 10 0.4814 0.0005 0.5016 400 15 IO IO 0.4888 O.OOO2 0.5004 400 150 IO IO 0.4779 . 0000 0.5001 400 150 IO IO 0.4776 o.oooo in Presence of Browning* has investigated with similar results Nitric Acid or the effect of free nitric acid and aqua regia (3 : I Aqua Regia. mixture of hydrochloric acid and nitric acid) upon the precipitation of barium as the sulphate in presence of an excess of sulphuric acid. In a total volume of 100 cm. 3 contain- ing 10 cm. 3 of dilute sulphuric acid [1:3] the barium sulphate falls with an average loss, after six hours' standing, of less than o.ooio grm. in the presence of amounts of nitric acid up to 25 per cent of the entire volume. In aqua regia the solubility of barium sulphate is even less. Following are experimental results of these methods of treatment. In this connection, the effect of the presence of a considerable amount of free nitric acid on the precipitation of barium as sul- phate in cases where certain substances are present which under ordinary conditions tend to hold up the precipitate, is of inter- est. Freseniusf has demonstrated this property in the case of ammonium nitrate, Scheerer and RubeJ have shown that meta- * Philip E. Browning, Am. Jour. Sci., [3], xlv, 399. t Zeit. anal. Chem., ix, 62. } Jour, prakt. Chem., Ixxv, 113-116. CALCIUM; STRONTIUM; BARIUM Precipitation in Presence of Nitric Acid. 171 BaSO* equivalent Ba(NO,) 2 taken. grm. BaS0 4 found. grm. Error in terms of BaSO 4 . grm. Averages, grm. Time between precipitation and filtration. hours. Per cent by volume of strong HNO 3 . Total volume. cm. 0.3540 0.2336 0.0004 | 12 5 IOO 0.2489 0.2483 0.0006 12 5 IOO 0.2495 0.2489 O.OOO6 v o . OOOO 12 5 IOO 0.2492 0.2482 O.OOIO J 12 5 IOO 0.2486 o . 2483 -0.0003 > O.OOO2 6 5 IOO 0.2490 o . 2490 o.oooo 6 5 IOO 0.2555 0.2546 0.0009 I A i 5 IOO 0.2538 0.2534 0.0004 / "~" O . OOOO i 5 IOO 0.2486 0.2477 0.0009 12 25 IOO 0.2491 0.2490 o.oooi 12 25 IOO 0.2494 o . 2484 O.OOIO 12 25 100 0.2538 0.2535 -0.0003 f 0.0008 12 25 IOO o. 2492 o . 2484 0.0008 12 25 IOO 0.2487 0.2471 0.0016 12 25 IOO 0.3414 0.3407 0.0007 12 25 IOO 0.2489 0.2485 0.2481 0.2478 0.0008 0.0007 > 0.0007 6 6 25 25 IOO IOO Precipitation in Presence of Aqua Regia. BaSO 4 equivalent to Ba(N0 3 ) 2 taken. BaSO 4 found. Error in terms of BaSO 4 . Averages. Time between precipitation and filtration. Per cent by volume of strong aqua regia. Total volume* grm. grm. grm. grm. hours. cm. 1 0-2539 0-2534 -0.0005 j 12 5 IOO 0.2540 0-2538 O.OOO2 0.0002 12 5 IOO 0.2490 o . 2490 O.OOOO ) 12 5 IOO 0.2491 0.2488 0.3419 0.2492 o . 2484 0.3421 +O.OOOI 0.0004 + 0.0002 > o.oooi 12 6 6 5 5 5 IOO- IOO IOO 0.2491 0.2485 0.0006 1 12 25 IOO o. 1701 o . i 708 0.1697 0.1705 O.OOO4 -0.0003 ^-0.0003 12 12 25 25 IOO 100- 0.1710 o. 1710 O.OOOO j 12 25 IOO- 0.3415 0.3410 0.0005 ) 6 25 IOO- 0.3418 0.3418 O.OOOO > 0.0003 6 25 IOO. 0.3412 0.3405 0.0007 0.0007 I 25 100- phosphoric acid acts similarly, and Spiller* notes the same gen- eral effect where alkali citrates are present. Browning shows that these salts cause no apparent interference with the precipi- * Chem. News, viii, 280, 281. 172 METHODS IN CHEMICAL ANALYSIS tation of barium in the presence of nitric acid amounting to one- tenth by volume of the entire liquid. The barium sulphate precipitated under such circumstances is, however, contaminated with foreign salts present and must be purified in order that the amount of barium actually present may be correctly indicated. The precipitate collected on paper and ignited, is, therefore, purified by dissolving it in sulphuric acid and recrystallizing according to the method of Mar, to be described.* Results of this treatment are given in the tabular statement. Purification of the Precipitate. BaS0 4 Apparent Impurity present to the amount of 5 grm. eauivalent toBa(NO 3 ) 2 taken. amount of BaS0 4 found. BaSO 4 after purification. Error after purification . Percentage of strong HN0 3 by volume. grm. grm. grm. grm. Ammonium nitrate o 1710 o 1800 O 1702 O 0008 IO Ammonium nitrate o. 341 < o . 3440 0.3410 O 0005 IO Ammonium citrate 0.3412 0.3442 0.3407 0.0005 IO Sodium citrate 0.1360 0.1730 0.13.66 +O.OOO6 IO Metaphosphoric acid 0.3461 0-35II 0.3470 +0.0009 10 Purification of When barium is precipitated as the sulphate the Precipitated Ba- tendency of the precipitate to include foreign matter, Tmm Sulphate. present, is very marked. It has been the custom to attempt the purification of barium sulphate contaminated by alkali salts by digesting in hydrochloric acid the washed precipi- tate. Phinneyf has shown, however, that dilute hydrochloric acid alone dissolves barium sulphate itself, while mixtures of hydrochloric acid with enough sulphuric acid to prevent such solvent action do not completely remove the impurity; and Mart has shown that the presence of hydrochloric acid, even in large excess, does not prevent contamination of the precipitate by alkali salts. After trying ineffectually the purification of the Impure barium sulphate by solution in strong sulphuric acid and reprecipi tation by water, Mar experimented with the crystalli- zation of barium sulphate from its solution. The contaminated precipitate is dissolved in hot concentrated sulphuric acid and * This page. t J. I. Phinney, Am. Jour. Sci., [3], xlv, 468. t Am. Jour. Sci., [3], xli, 293. Loc. cit. CALCIUM; STRONTIUM; BARIUM 173 recovered from solution in crystalline form by evaporation of the acid. The crystallized sulphate is then washed upon a felt of asbestos in the filtering crucible, ignited and weighed. The evaporation may be effected over a radiator or by means of a ring burner; in either case, the process requires several hours. The operation may, however, be completed safely in a half-hour by the aid of the Hempel evaporating burner. Examples of the efficiency of the method are given in the tabular statements. Degree of Contamination Found in Precipitated Barium Sulphate. BaCl 2 .2H 2 O taken. grm. BaSO 4 found, grm. Error, grm. HC1 in solution. cm. 3 Alkaline salts present. 0.5092 0.5032 +0.0169 no KC1O 3 3 grm. 0.5027 0.4907 +0.0107 IO KC1O 3 3 grm. 0.5026 0.4944 +0.0154 IOO KC1 5 grm. 0.5045 0-4939 +O.OI22 10 KC1 5 grm. o . 5020 0.4931 +0.0137 10 KC1 5 grm. 0.5013 o . 4849 +O.OO6I 10 NaCl 5 grm. Slow Evaporation over Radiator or by Ring Burner. BaCl 2 .2H 2 O taken, grm. BaSO 4 found, grm. Error, grm. 0.5029 0.4796 O.OOO6 0.5008 0.4783 +O.OOOI 0.5038 0.4810 +O.OOOI 0.5087 0.4861 +0.0003 0.5025 0-4795 +0.0006 Rapid Evaporation by Hempel Burner. BaCl 2 .2H 2 taken, grm. BaSO 4 found, grm. Error, grm. o . 5050 0.4824 +O . OOO2 0.5069 0.4838 O.OOOO 0.5041 0.4825 +0.0021 0.5021 0.4812 +O.OOI8 0.4033 0.4801 0.0005 The results of applying this method to the purification of barium sulphate precipitated in presence of nitric acid from solutions containing citrates or a metaphosphate are given on page 172. 174 METHODS IN CHEMICAL ANALYSIS The Estimation of Barium as the Chloride. precipitation by It has long been known that barium chloride is chtoric^ckT" insoluble to a marked degree in hydrochloric acid, Mixture. but the difficulty of filtering off the strong acid and washing the precipitate with strong acid prevented the early use of this characteristic of the chloride for the quantitative estima- tion of barium. The treatment of strong acid filtrates by means of the asbestos felt in the filtering crucible is now an easy matter, and the limits of insolubility of barium chloride in hydrochloric acid have been studied by Mar* with a view to developing a simple method for the separation of barium from calcium and magnesium. It has been shown that barium chloride is soluble to an extent not exceeding one part in 20,000 in pure, concen- trated hydrochloric acid, the solubility increasing very rapidly with the diminution in the strength of the acid, while in con- centrated hydrochloric acid containing ether the solubility falls to an amount not exceeding one part in about 120,000. To utilize this fact for the separation of barium from calcium and magnesium, Mar dissolves the chlorides of the earths in the least possible amount of boiling water and precipitates by 25 cm. 3 of concentrated hydrochloric acid with the addition of 5 cm. 3 of absolute ether after cooling. The acid is added drop by drop at first, as the precipitate is thus obtained in a coarse crystalline condition, filters very quickly, and is less liable to include foreign matter. After standing a few minutes the precipitate is filtered on an asbestos felt in a perforated crucible, washed with hydro- chloric acid containing about 10 per cent of ether, and dried at I5O-2OO. The method is accurate and rapid, and possesses the further advantage, when a number of determinations are to be made, that the precipitate may be dissolved off the felt by a little water, and, after ignition, the crucible and felt used again without reweighing. The felt upon which a half-dozen precipitates are thus treated may not change by so much as o.oooi grm. in the process. The fumes of the strong acid cause no inconvenience if the filtration is performed in front of a good flue. The figures of analysis, given below, indicate the accuracy of the process when applied to the pure barium salt. * F. W. Mar, Am. Jour. Sci., [3], xliii, 521. CALCIUM; STRONTIUM; BARIUM The Pure Barium Salt. 175 BaCl 2 .2H 2 0. HC1. Ether. BaCl 2 . Error. grrn. cm. 8 cm. s grm. grm. 0.5008 50 IO 0.4267 0.0002 0.5002 5 IO 0.4257 0.0007 0.4999 5 IO 0.4252 0.0009 0.4999 50 IO 0.4258 0.0003 0.5003 25 25 0.4259 0.0005 0.5002 25 5 0.4262 0.0002 0.5099 25 5 0-4344 0.0003 0.5003 25 5 0.4261 -0.0003 Following are figures which show the results obtained in sepa- rating and determining barium when associated with magnesium and with calcium in mixtures of the chlorides. Separation of Barium from Calcium and Magnesium. BaCl 2 .2H 2 O. Cad,. HC1. Ether. BaCl 2 . Error. grm. grm. cm. 8 cm. s gnu* grm. 0.5001 0-5 50 10 0.4250 0.0013 0.4999 0-5 50 10 0.4250 o.oon 0.5005 0.5 25 25 0.4260 0.0006 0.5002 0.42 25 5 0.4258 0.0004 0.5001 0-5 25 5 0.4255 0.0008 0.5005 0.5 25 5 0.4251 0.0015 0.5001 0.5 25 5 0.4254 0.0009 0.5001 0.5 25 5 0.4258 0.0005 0.5003 0.5 25 5 0.4261 0.0004 O. IOO2 3-o 25 5 o . 0842 O.OOI2 O.OIO7 3-0 25 5 o . 0080 0.0005 BaCl 2 .2H 2 0. MgCl 3 .6HO 2 . HC1. Ether. BaCl 2 . Error. grm. grm. cm. 8 cm. J grm. grm. 0.4999 0-5 25 5 0.4253 0.0007 0.5000 0-5 25 5 0.4257 0.0005 O.IOO2 3-o 25 5 o . 0844 o.ooio O.OIOO 3-o 25 5 0.0077 0.0008 Precipitation by Gooch and Boynton* have given procedure for the Acetyi Chloride precipitation of barium chloride from water solution and its separation from calcium and magnesium by the use of acetyl chloride to decompose the water of the solution according to the reaction CH 3 COC1 + H 2 = CH 3 COOH + HC1. * F. A. Gooch and C. N. Boynton, Am. Jour. Sci., [4], xxxi, 212. 176 METHODS IN CHEMICAL ANALYSIS Inconvenient violence of the reaction is moderated by the addi- tion of acetone which mixes in all proportions with both acetyl chloride and water, and by itself exerts no appreciable solvent action upon barium chloride. When a mixture of acetone and acetyl chloride, preferably 4:1, is added slowly to a very concentrated solution of barium chloride in water, the water is attacked at once, hydrogen chlo- ride is liberated, and precipitation begins immediately. If the temperature is kept down during the process by immersing in cool running water the vessel in which reaction takes place, no more than a mere trace of barium can be detected by sulphuric acid in the residue left after evaporating the liquid separated from the precipitate by filtration through asbestos. When, how- ever, the temperature is allowed to rise, in consequence of the heat liberated in the reaction, an appreciable amount of barium may be found by sulphuric acid in the nitrate. It appears that when the acetone-acetyl chloride mixture [4:1] acts upon the cooled concentrated water solution of barium chloride the pre- cipitate is the hydrous chloride, BaCl 2 .2H 2 O, only the water in excess of that needed to form the hydrous salt being immedi- ately attacked ; that acetyl chloride by itself produces only slight dehydration of this salt without marked solubility; and that prolonged action of an acetone-acetyl chloride mixture [2:1] results in appreciable dehydration and considerably increased solubility of the salt. When the acetone-acetyl chloride mixture is added without cooling to the water solution of barium chloride the heat of reaction favors dehydration of the hydrous salt, and the anhydrous salt may go into solution to the amount of sev- eral milligrams in 10 cm. 3 of the precipitating mixture. Upon filtering the mixture and treating the filtrate with acetone, with acetyl chloride, or with the acetone-acetyl chloride mixture, the dissolved anhydrous salt is not thrown out of solution, but the addition of a drop of water is sufficient to induce immediate precipitation in the form of the hydrous salt. The best conditions for the quantitative precipitation of barium chloride by the acetone-acetyl chloride mixture are found in the use of minimum amounts of water, the preservation of ordinarily low temperature, a liberal proportion of acetone, and not too prolonged digestion of the precipitate in the excess of the pre- cipitant. The salt to be analyzed is weighed out into a small CALCIUM; STRONTIUM; BARIUM 177 beaker and dissolved in I cm. 3 of water. The beaker is cooled by immersion in a water bath, preferably supplied with running water at a temperature of about 15. To the cooled solution, constantly shaken, the acetone-acetyl chloride mixture is added from a dropping funnel at the rate of five drops to the second. The precipitate is filtered off upon asbestos in a perforated cru- cible, dried, or ignited, and weighed as the anhydrous chloride, BaCU. The best conditions studied for the handling of O.I grm. of hydrous barium chloride are the solution of the salt in I cm. 3 of water, treatment with 30 cm. 3 of the 4 : 1 mixture of acetone and acetyl chloride, washing with acetone, and drying in the air bath at 135 or at low redness. Following are the results of experimental tests of the method applied to pure barium chloride. The Pure Barium Salt. BaCl 2 taken as BaCl 2 .2H 2 O. grm. BaCl 2 found. grm. Error. grm* Water to dissolve BaCl 2 .2H 2 O. cm.* Amount of mixture and composition by volume. To precipitate. To wash. 0.0859 0.0859 o.oooof 5 cm. 3 2:1 10 cm. 3 2: 0.0861 0.0854 0.0007! 5 cm. 3 2: 10 cm. 3 2: 0.0861 0.0862 0.0858 0.0854 0.0003! 0.0008* 5 cm. 3 2: 6 cm. 3 2: 10 cm. 3 2: 10 cm. 3 2: 0.0857 0.0854 0.0003* 6 cm. 3 2: 10 cm. 3 2: 0.0858 o . 0860 +O.OOO2* 6 cm. 3 2: 30 cm. 3 4:1 0.0860 0.0859 O.OOOI* 6 cm. 3 2: 30 cm. 3 4:1 0.0853 0.0850 0.0003* 6 cm. 2: Acetone. 0.0854 0.0848 0.0006* 6 cm. 2: Acetone. 0.0852 0.0851 0.0001* 6 cm. 2: Acetone. 0.0857 0.0856 o.oooif 6 cm. 2: Acetone. 0.0852 o . 0845 0.0007! 6 cm. 2: Acetone. 0.0855 0.0852 0.0003! 6 cm. 2: Acetone. 0.0862 0.0862 o.oooof 30 cm. 4: Acetone. 0.0868 0.0868 o.oooo! 30 cm. 4: Acetone. * Ignited at low redness, t Dried at 135 for i$ hours. The application of these conditions to the separa- tion of barium from moderate amounts of calcium and magnesium proves to be easily feasible. When acetone is added to the concentrated solution of calcium chloride or magnesium chloride in water two liquid layers are formed, the Separation from Calcium and Magnesium. i 7 8 METHODS IN CHEMICAL ANALYSIS acetone above and the aqueous layer below; but the addition of a few drops of acetyl chloride renders the liquids miscible, while further addition causes no precipitation. When the 4: 1 mixture of acetone and acetyl chloride is added at the rate of five drops in the second to the solution containing no more than 0.5 grm. of the calcium and magnesium salts, barium chloride is precipitated while calcium chloride and magnesium chloride are dissolved; but when the soluble chloride is present in the proportion of Separation of Barium from Calcium. BaCl 2 taken as BaCl 2 .2H 2 O. CaCl 2 .2H 2 O taken. BaCl 2 found. Error. Water used to dissolve salts. Amount of mixture [4:1! used. grm. grm. grm. grm. cm. 1 cm. 0.0859 O. IOOO 0.0859 o.oooo* 30 0.0867 o. 1040 0.0867 o.oooo* 30 0.0868 O. IO22 0.0868 0.0000* 30 0.0865 O. IO2O 0.0865 o.oooo* 30 0.0868 O.IOI7 o . 0869 +O.OOOI* 30 o . 0864 o. 1016 0.0861 0.0003* 30 . 0866 0.3025 0.0867 +O.OOOI* i 30 o . 0859 0.5025 o . 0859 0.0000* 2 30 0.0860 I.OO2O 0.0878 +0.0018* 3 30 0.0859 I .OO2O 0.0855 o.ooo4f 2 30 0.0864 1.0035 0.0867 +0.0003! 2 30 * The precipitant was added at first at the rate of five drops in the second, t The precipitant was added at the rate of two drops in the second at the outset and later of five drops in the second. Separation of Barium from Magnesium. BaCl 2 taken as BaCl 2 .2H 2 O. MgCl 2 .6H 2 taken. BaCl 2 found. Error. Water used to dissolve salts. Amount of mixture [4 : i] used. grm. grm. grm. grm. cm.". cm.s 0.0858 O. IOOO 0.0857 O.OOOI* 30 0.0869 0.1025 0.0870 +O.OOOI* 30 0.0858 0.1025 0.0858 o.oooo* 30 0.0862 O.IOIO 0.0863 +0.0001* 30 0.0858 0.1006 0.0860 +O.OO02* 30 o . 0860 O. IO2O 0.0859 O.OOOI* 30 0.0860 O. IOIO 0.0862 +O.OOO2* 30 0.0865 0.3010 0.0867 +0.0002* \ 30 0.0864 0.5000 0.0867 +0.0003* 2 3 0.0868 1.0015 0.0878 +O.OOIO* 3 30 0.0853 I. 0010 0.0854 +o.oooif 3 30 * The precipitant was added at the rate of five drops in the second. t The precipitant was added at first at the rate of two drops in the second and later of five drops in the second. CALCIUM; STRONTIUM; BARIUM 179 i.o grm. to o.i grm. of the barium chloride, the rate of addition of the precipitating mixture should not be greater than two drops in the second at the start in order to avoid inclusion of the soluble salt in the insoluble barium salt. Even in such cases the mixture may be added at the rate of five drops in the second, after the greater part of the barium is down. The experimental results obtained in the separation of o.i grm. of the barium salt from 0.5 grm. of calcium and magnesium salts are excellent. When the 4 : 1 mixture of acetone and acetyl chloride is added to the concentrated water solution of o. I grm. of strontium chlo- ride a partial precipitation of the hydrous chloride, SrCl2.2H2O, takes place. The action of a 4 : 1 mixture of acetone and acetyl chloride upon the concentrated solution of the chlorides affords easy and exact means for the separation and estimation of barium asso- ciated with calcium and magnesium. It is not recommended for the separation of barium from strontium. The Precipitation of Barium Bromide by Ether-Hydrobromic Acid Mixture. Thorne * has shown that barium bromide dissolved in the least possible amount of water is completely precipitated by a mixture of concentrated hydrobromic acid and ether in equal parts, and that the precipitate may be obtained of normal constitution, BaBr2, and weighed as such if, after filtering upon asbestos in the perforated crucible, it is treated with ammonium bromide, and then gradually heated to 250. Following are the results of test experiments made in this manner. The Pure Barium Salt. BaBr 2 .2H 2 O taken. grm. HBrand ether [i : i]. cm.* BaBr 2 found, grm. BaBr 2 calculated, grm. Error, grin. O.2OO8 30 0.1793 0.1790 +o . 0003 o. 2041 30 0.1822 0.1820 +O.OOO2 o . 2047 30 o. 1821 0.1825 0.0004 0.2171 30 0.1937 0.1936 +0.0001 0.3101 30 0.2768 0.2765 +0.0003 0-503S 30 0.4496 0.4490 +0.0006 0.5015 30 0.4476 0.4473 +o . 0003 * Norman C. Thorne, Am. Jour. Sci., [4], xviii, 441. i8o METHODS IN CHEMICAL ANALYSIS Barium is precipitated completely either by hydrobromic aci'd or by hydrochloric acid* in mixture with ether, the precipitate falling as bromide or chloride in proportions according with the relative amounts of these acids present. In the presence of a great excess of hydrobromic acid the salt precipitated will be essentially bromide even if the original salt is taken in the form of the chloride. Salts of calcium and of magnesium remain in solution. Separation of Barium from Calcium and Magnesium. BaCl 2 .2H 2 O taken. grni. CaCO 3 . grm. MgCO 3 . grm. HBr and ether [i:ij. cm. 3 BaBr 2 found. cm. 3 Theory as BaBr 2 . grm. Error in BaBr 2 . grm. 0.2253 o 2088 30 3O 0.2744 O 2< / ?8 0.2741 o 2<\4o +0.0003 O OOO2 o. 3273 3O O 3Q7 1 ? o 3082 o 0007 0.3177 3O 0.3864 o 386 < o oooi 0.5041 0.5000 3O 0.6134 0.6143 o 0009 o . 5083 0.5000 30 0.6185 0.6191 o 0006 o 5046 o 5000 3O o 6130 o 6136 -j-o 0003 0.5022 0.5018 O ^OO7 0.5000 0.5000 o 3000 30 30 7Q 0.6110 0.6106 o 6087 0.6104 0.6108 o 6002 +0.0006 0.0002 o ooo ^ o ^048 o 3000 ?o o 614.4. o 6142 +O OOO2 The Estimation of Calcium, Strontium and Barium, Precipitated as Oxalates. The very high degree of insolubility which makes possible the well-known and exact process for the determination of calcium by precipitation as oxalate from ammoniacal water solutions, and weighing, is not directly applicable to strontium and barium on account of the greater solubility of the oxalates of those elements. Strontium oxalate is soluble in 12,000 parts of water, f while one part of barium oxalate dissolves in less than 3000 parts of cold water. % Gravimetric Peters has shown, however, that strontium salts S e str^dum n ma y be precipitated by ammonium oxalate with and Barium. practical completeness in a solution containing one- fifth of its volume of 85 per cent alcohol, and with approximate * Mar, page 174. t Souchay and Lenssen, Ann. Chem., cii, 35. t Souchay and Lenssen, Ann. Chem., xc, 102. Am. Jour. Sci., [4], xii, 223. CALCIUM; STRONTIUM; BARIUM 181 completeness from water solutions at a dilution not exceeding 250 cm. 3 in presence of an amount of ammonium oxalate sev- eral times larger than that required theoretically; and that ba- rium oxalate is precipitated almost completely from a solution containing one- third of its volume of 85 per cent alcohol. The precipitates thrown, down hot and allowed to stand over night may be filtered off on asbestos in the perforated crucible, ignited a few minutes in the flame of a Bunsen burner and weighed as carbonate, or, after treatment with sulphuric acid, as sulphate. The results thus obtained are fairly accurate, as shown. Precipitation in Approximately 17 per cent Alcohol. SrO taken as Sr(N0 3 ) 2 . SrO calculated from SrCO 3 found. Difference. grm. grm. grm. O. 1 1 20 O. II2O O.III3 o. 1116 0.0007 0.0004 0-2435 0.2425 O.OOIO Precipitation in Approximately 25 per cent Alcohol. BaO taken as Ba(N0 3 ) 2 . BaO found as BaC0 3 . Difference. grm. grm. grm. 0.2912 o. 2912 0.2912 0.2909 0.2901 o . 2901 0.0003 O.OOII O.OOII Titration of Peters * has also shown that calcium, strontium and pofaL7um b per- barium may be accurately estimated by titration of manganate. the oxalates in hydrochloric acid solution by potas- sium permanganate in presence of a manganous salt.f Calcium oxalate is precipitated from the boiling hot solution with ammonium oxalate, and allowed to stand twelve hours. The supernatant liquid is decanted upon asbestos in the filter- ing crucible. The precipitate is washed two or three times, by decantation, with 50 cm. 3 -ioo cm. 3 of cold water and brought on the felt with care to avoid extended washing with hot water after all the precipitant, ammonium oxalate, has been removed. The crucible containing the precipitate is returned to the beaker, * Am. Jour. Sci., [4], xii, 216. t See page 50. 182 METHODS IN CHEMICAL ANALYSIS 100 cm. 3 -2OO cm. 3 of water added, together with 5 cm. 3 -io cm. 3 of strong hydrochloric acid and 0.5 grm.-i.o grm of manganous chloride, and the oxalic acid titrated at a temperature of 35-45. The results given are obviously excellent, and show that calcium, taken as the oxalate, may be estimated by potassium perman- ganate in the presence of hydrochloric acid and a manganous salt. Precipitation from Water Solution. CaO taken as CaCl 2 . Ammonium oxalate. Volume at precipitation. CaO found. Error. grm. grm. cm.* .grm. grm. 0.0656 0-3 IOO 0.0657 -f-O.OOOl . 0656 0-3 100 0.0656 o.oooo o . 0656 0-3 ' 150 0.0658 +O.OOO2 0.0656 0-3 IOO 0-0655 o.oooi 0.0985 0-5 175 0.0981 0.0004 O.I3I3 0.6 150 O.I3I5 +O.OOO2 O.I3I3 0.6 2OO O.I3IS +O.OOO2 Peters showed also that sulphuric acid may be employed in place of hydrochloric acid and manganous chloride when the dilution at titration is sufficient. In precipitating strontium as the oxalate in alcoholic solution, ammonium oxalate is added to the hot solution with 85 per cent alcohol amounting to one-fifth to one-third of the total volume, the mixture is allowed to stand over night, and the clear liquid decanted upon an asbestos filter. The precipitate is washed with a mixture of equal parts of 85 per cent alcohol and water, transferred to the filter, dried in the filtering crucible over a flame to free it from alcohol, returned to the beaker previously Precipitation from Alcoholic Solution. Volume during titration 150 to 250 cm. 3 . SrO taken as Sr(N0 3 ) 2 . Ammonium oxalate. Volume at precipita- tion. Proportion of 85 per cent alcohol. Acid present during titration. SrO found. Error. grm. grm. cm. 3 grm. grm. 0.0974 0.0974 0-4 0-4 IOO IOO t HC1 HC1 0.0973 0.0983 O.OOOI +o . 0009 0.0974 0-4 IOO HC1 0.0975 + O.OOOI 0.0974 0.8 IOO I HC1 0.0981 +O.OOO7 o. 1948 0.4 2OO \ HC1 0.1943 O.OOO5 o. 1948 0.8 2OO 3 HC1 o. 1942 O.OOO6 CALCIUM; STRONTIUM; BARIUM dried, treated with 5 cm. 3 -io cm. 3 of hydrochloric acid and 0.5 grm.-i.o grm. of a manganous salt, and the liberated oxalic acid is titrated by permanganate. The results obtained by this method are accurate. Precipitation from Water Solution. SrO taken as SrClj. Ammonium oxalate. Volume at precipitation. Acid present during SrO found. " Error. grm. grm. cm. s grm. grm. 0.0974 0.8 100 HC1 0.0971 0.0003 0.0974 0.8 100 HC1 0.0980 +0.0006 0.0974 0.8 100 HC1 0.0975 +O.OOOI 0.0974 0.8 IOO HC1 0.0980 +0.0006 0.0974 0.8 IOO HC1 0.0973 o.oooi 0.0974 0.8 IOO HC1 0.0978 +0.0004 By precipitating in the water solution in presence of a con- siderable excess of ammonium oxalate and washing with small amounts of water (30 cm. 3 -4O cm. 3 ) applied judiciously, the loss by solubility may be made practically inappreciable, and in this case, there being no alcohol present to affect the titration, the precipitate need not be dried before treatment with perman- ganate. The results above show that o.i grm. of the strontium salt, calculated as the oxide, may be estimated as the oxalate with accuracy when precipitated in 100 cm. 3 of water by a suf- ficient excess of ammonium oxalate. Precipitation from Water Solution. SrO taken as Sr(N0 8 ) z . grin* Ammonium oxalate. grm. Volume at precipitation. cm. 8 Acid present during titration. SrO found, grm. Error. grrn. 0.0974 0-5 IOO H 2 S0 4 0.0966 0.0008 0.0974 0-5 IOO H 2 S0 4 0.0985 +O.OOII 0.0974 0-5 IOO H 2 SO 4 0.0977 +0.0003 0.0974 0-5 IOO H 2 SO 4 0.0963 o.oon 0.0974 0.8 IOO H 2 SO 4 0.0981 +0.0007 0.0974 0.8 IOO H 2 SO 4 0.0966 0.0008 0.0974 i .0 IOO H 2 S0 4 0.0965 0.0009 0.0974 2.O IOO H 2 SO 4 o . 0963 o.oon 0.0974 2.O IOO H 2 S0 4 0.0970 0.0004 0.0778 0-5 IOO H 2 SO 4 0.0792 +0.0014 0.0778 0-5 IOO H 2 SO 4 0.0767 O.OOII 0.0778 o-5 IOO H 2 S0 4 0.0776 O.OOO2 0.0778 o-5 IOO H 2 S0 4 0.0776 O.OOO2 0.0974 0.8 250 H 2 S0 4 0.0973 + 0.0001 0.0974 2.O 250 H 2 SO 4 0.0975 o.oooi 184 METHODS IN CHEMICAL ANALYSIS Amounts of strontium oxalate approximately equivalent to O.I grin, of strontium oxide may be successfully precipitated without alcohol and titrated in a volume of 200 cm. 3 ~3OO cm. 3 in presence of sulphuric acid, as shown by the results which are given in the preceding table. To precipitate barium as the oxalate, ammonium oxalate is added to a solution of a barium salt, containing 30 per cent of its volume of alcohol, and after standing over night the precipitate is filtered on asbestos, washed by decantation with 100 cm. 3 -2OO cm. 3 of water containing 30 per cent of its volume of alcohol, and dried over a flame to insure the removal of alcohol. The crucible containing the precipitate is returned to the beaker, also previously dried over a flame, 100 cm. 3 -2OO cm. 3 of water, 5 cm. 3 10 cm. 3 of strong hydrochloric acid, and 0.5 grm. i.o grm. of manganous chloride are added, and the solution is titrated at 35-45 with permanganate. The results of the experiments given show that barium, either as the nitrate or chloride, may be estimated in the manner described with a fair degree of accuracy. Precipitation from Alcoholic Solution. BaO taken as Ba(N0 3 ) 2 . grm. Ammonium oxalate. grm. Volume at precipitation. cm.* Acid present during titration. BaO found, grm. Error, gnu. 0.1165 0.2 TOO HC1 0.1177 +O.OOI2 0.1165 O.2 100 HC1 0.1170 +0.0005 0.1165 O.2 IOO HC1 0.1164 O.OOOI 0.1165 0.2 IOO HC1 0.1151 O.OOI4 0.1165 O.2 IOO HC1 0.1165 O.OOOO 0.1165 O.2 IOO HC1 0.1176 +O.OOII 0.1165 O.2 IOO HC1 0.1164 O.OOOI 0.2330 0-4 IOO HC1 0.2319 o.oon 0.2330 0.4 IOO HC1 0.2335 +o . 0005 0.2330 0.4 IOO HC1 0.2342 +O.OOI2 BaO taken as BaCU. . 0942 0.4 IOO HC1 0.0952 +0.0010 0.0942 0.4 IOO HC1 0.0939 0.0003 0.0942 0.4 IOO HC1 o . 0941 O.OOOI 0.1884 0.4 IOO HC1 0.1893 +0.0009 0.1884 0.4 IOO HC1 0.1892 +o . 0008 Barium oxalate cannot be successfully titrated in presence of sulphuric acid on account of the great insolubility of barium sulphate and its protecting influence upon undecomposed barium oxalate. CHAPTER V. ZINC; CADMIUM; MERCURY. ZINC. The Estimation of Zinc as the Pyrophosphate. IN studying the determination of zinc by the method which involves precipitation as ammonium zinc phosphate and weigh- ing as the ignited pyrophosphate, Austin* has shown that, as in the case of the similar precipitation of manganese, f the presence of a definite excess of ammonium salt during the precipi- tation is essential to the formation of the ideal salt, NH4ZnPO4, uncontaminated by the tri basic phosphate, Zn 3 (PO 4 ) 2 , while too much tends to produce a double salt too rich in ammonia. The condition of the ammonium zinc phosphate most nearly approximating to the ideal is obtained by precipitating in pres- ence of ammonium chloride in large amount. Microcosmic salt is added until the solution (100 cm. 3 to 200 cm. 3 ) contain- ing the ammonium salt is alkaline, and the whole is heated until the mass subsides in crystalline condition. The amount of ammonium chloride should be 20 grm. if the filtration is to be made as soon as the solution cools. One-half this amount will do if the liquid stands a number of hours. Larger amounts tend to give a salt too rich in ammonia. The time of standing seems to be a less important factor than either the excess of microcosmic salt or the amount of ammonium chloride. When the solutions are made finally faintly acid with acetic acid, according to the method of Langmuir,{ the results are low. Following are the results obtained by the method described. In a subsequent article it is made plain that Dakin's || pro- posal to wash with a I per cent solution of ammonium phosphate, followed by alcohol, leads to erroneous results. * Martha Austin, Am. Jour. Sci., [4], viii, 210. t See page 483. t Jour. Am. Chem. Soc., xxi, 115. Am. Jour. Sci., [4], xiv, 156. II Chem. News, Ixxxii, 101; Ixxxiii, 37. 185 186 METHODS IN CHEMICAL ANALYSIS Estimation as Zinc Pyrophosphate. Zn 2 P 2 7 correspond- ing to ZnS0 4 taken. Found. Error. Error in terms of zinc. Zn 2 P,O 7 correspond- ing to Zn left in the filtrate. HNaNH<- PO<. 4 H 2 0. NH 4 C1. Time of standing. grm. grm. grm. grm. gnu. grm. grm. hrs. 0-6355 0-6335 O.OO2O 0.0008 none 4-47 10 16 0-6355 0.6381 +0.0026 +O.OOIO none 4-47 20 i 0-6355 0.6379 +0.0024 +0.0009 none 4-47 20 2 0-6355 0.6386 +0.0031 +0.0012 none 4-47 20 i 0-6355 0.6393 +0.0038 +0.0014 none 4-47 20 I 0.6367 0-6355 +O.OOI2 +0.0005 none 4-47 30 16 The Conversion of Zinc Chloride to Zinc Oxide. Havens* has shown that zinc chloride may be quantitatively converted to zinc oxide by treatment with nitric acid, evapora- tion of the excess of acid, and ignition of the residue. The solution of zinc chloride is evaporated in porcelain, best with a gentle current of air playing upon the surface of the liquid to avoid spattering, and treated repeatedly with nitric acid, added in small portions with intermediate evaporations. The residue is finally ignited to convert the nitrate to oxide. Results are given in the table. Conversion of Chloride to Oxide. ZnO taken as chloride. ZnO found. Error. grm. grm. grm. .0.1019 O.IOIO O. IIOO o. 1016 o. 1007 0.1095 0.0003 0.0003 0.0005 The Electrolytic Determination of Zinc. The deposition of zinc upon the rotating crucible f succeeds best, according to Medway,t when the zinc salt preferably the sulphate is dissolved in 50 cm. 3 of water to which 4 grm. of potassium oxalate are added. The presence of ammonium salts appears to retard the complete deposition of the metal. * F. S. Havens, Am. Jour. Sci., [4], vi, 45. f See Fig. 13, page 12. J H. E. Medway, Am. Jour. Sci., [4], xviii, 56. ZINC 187 Deposition on the Rotating Cathode. Zinc taken, grm. Zinc found, grm. Error, grm. Current, amp. N. D.. Time, min. 0-0553 0.0556 +0.0003 2-5 8-3 25 0-0553 0-0553 o.oooo 2-5 8-3 25 0-0553 0.0552 O.OOOI 2-5 8-3 25 0.0993 0.0995 +O.OOO2 2-5 8-3 30 0.0993 0.0994 +0.0001 2 6.6 25 0.0993 0.0991 0.0002 2 6.6 25 In determining zinc by electrolysis with stationary electrodes, it has been found that, when the attempt is made to remove the zinc from the platinum upon which it has been deposited, a coating of platinum black is left, some of the zinc having amalgamated with the platinum. Only by dissolving the zinc, heating the crucible to redness and finally making another appli- cation of acid can this black coating be conveniently removed. In order to avoid this formation of platinum black it has been found necessary to coat the platinum with copper and deposit the zinc upon this. The zinc and copper may then be easily re- moved together by acid. In depositing the zinc upon a rotating cathode, however, it is found to be unnecessary to coat the plati- num with copper, since the zinc can be removed without any appearance of platinum black. In depositing zinc upon the rotating cathode from an acetate solution containing a salt of iron, Moody* has found iron de- posited with the zinc.f The Estimation of Zinc by Precipitation as the Oxalate and Titra- tion with Potassium Permanganate. Wardt has shown that zinc may be accurately estimated by precipitation as oxalate and titration with potassium perman- ganate. To the boiling water solution of the zinc salt oxalic acid is added, followed by acetic acid in large amount. The pre- cipitate is filtered upon asbestos in the perforated crucible, and washed with small amounts of water. Crucible and precipitate * Seth E. Moody, Am. Jour. Sci., [4], xxii, 484. t See page 67. t H. L. Ward, Am. Jour. Sci., [4], xxxiii, 334. i88 METHODS IN CHEMICAL ANALYSIS are treated with dilute sulphuric acid, the solution heated to boiling, and the free oxalic acid titrated by permanganate. Results are given in the table. Determination of Zinc. Zinc taken as acetate. Volume at precipitation. Oxalic acid. Acetic acid added. Zinc found. Error. grin. cm. 3 grm. cm. 3 grm. grm. 0-0055 IOO 2 IOO 0.0056 +O.OOOI 0.0274 IOO 2 IOO 0.0276 +0.0002 0.0548 50 2 50 0-0553 +o . 0005 0.0548 IOO 2 IOO 0.0550 -f-O . OOO2 0.1370 IOO 2 IOO 0.1372 +O.OOO2 CADMIUM. The Estimation of Cadmium as the Oxide. Precipitation as Various objections have been made to that method Carbonate. f or estimating cadmium which involves precipita- tion as carbonate, ignition, and weighing as oxide. Browning and Jones * have shown that when the carbonate is filtered upon an asbestos felt in a perforated crucible previously ignited and weighed, danger of reduction is obviated and the process is sim- plified and placed in the category of good analytical methods. To the solution of the cadmium salt in about 300 cm. 3 of hot water is added a 10 per cent solution of potassium carbonate, drop by drop and with constant stirring. The solution, with the precipitate in suspension, is boiled for about fifteen min- utes. The precipitated cadmium carbonate becomes granular and settles quickly, and is then filtered upon asbestos in the per- forated crucible, ignited and weighed. Results obtained by this method, and quoted in A of the following table, show a slight plus error due, as was shown experimentally, to inclusion of alkali salt, but indicate that the carbonate method can be suc- cessfully and simply applied to the quantitative estimation of cadmium. The work of Flora f fully substantiates these results, giving by similar procedure the analytical figures in B. * Philip E. Browning and L. C. Jones, Am. Jour. Sci., [4], ii, 269. t Charles P. Flora, Am. Jour. Sci., [4], xx, 456. CADMIUM 189 Precipitation as Cadmium Carbonate. CdO taken. CdO found. Error. grm. grm. grm. A. o. 1140 0.1143 +0.0003 0.1142 0.1137 0.0005 o. 1141 0.1148 +O.OOO7 0.1141 o. 1148 +O.OOO7 o. 1142 o. 1146 +o . 0004 0.1143 O.II47 +o . 0004 0.1143 O.II44 +0.0001 0.1139 o. 1146 +0.0007 o. 1270 o. 1272 +0.0002 0.1279 0.1283 +o . 0004 0.1272 o. 1281 +0.0009 o. 1278 o. 1281 +o . 0003 0.2556 o. 2561 +0.0005 0.2550 0.2547 0.0003 0.1272 0.1279 +0.0007 o. 1281 0.1288 +0.0007 0.1274 0.1278 +o . 0004 0.1284 o. 1290 +o . 0006 0.1271 0.1277 +0.0006 o. 1278 o. 1285 +0.0007 0.2555 0.2555 o.oooo B. 0.1277 0.1275 O.OOO2 0.1277 o. 1280 +0.0003 0.1277 o. 1272 0.0005 0.1399 0.1391 0.0008 0.1399 0.1399 o.oooo 0.1703 o. 1700 0.0003 0.1703 o . i 700 0.0003 0.2129 0.2128 O.OOOI 0.2129 0.2128 o.oooi 0.2554 0.2554 o.oooo Precipitation as Flora* has also tested the similar method by Hydroxide. which cadmium is precipitated as hydroxide and weighed as oxide. To the boiling solution of the cadmium salt, about 300 cm. 3 in volume, a 10 per cent solution of potassium hydroxide is added drop by drop. After boiling about fifteen minutes the precipitate settles quickly in a semigranular state and is filtered on a weighed asbestos felt in the perforated cru- cible, washed, ignited and weighed. The results are lower than those of the carbonate method. * Loc. cit. METHODS IN CHEMICAL ANALYSIS Precipitation as Cadmium Hydroxide. CdO taken, gnu. CdO found, gnu. Error, grm. 0.1277 0.1277 o.oooo 0.1277 o. 1270 0.0007 0.1277 o . i 260 0.0017 0.1277 0.1286 +o . 0009 0.1362 0.1350 O.OOI2 0.1399 0.1389 O.OOIO 0.1703 0.1697 0.0006 0.1703 0.1693 O.OOIO 0.1703 o. 1699 0.0004 0.1788 o. 1802 +0.0014 o. 2129 0.2139 +O.OOIO 0.2129 0.2128 o.oooi While these figures show that fair results may be obtained, the hydroxide method is not comparable with the carbonate method as to accuracy or convenience. The precipitate does not take the same granular form; it is hard to filter, difficult to wash, and is removed with difficulty from the beaker in which precipitation takes place. The Estimation of Cadmium as the Pyrophosphate. It has been shown by Austin * that cadmium may be estimated with accuracy as the pyrophosphate. The precipitate obtained by making alkaline with microcosmic salt the nearly neutral solution, containing ammonium chloride in the proportion of ten grams to one hundred cubic centimeters, is allowed to stand several hours, then filtered off on asbestos in the filtering crucible, dried, ignited and weighed. The cadmium separates out from the solution as a beautiful crystalline mass of cadmium ammo- nium phosphate of ideal constitution. The conditions must, however, be preserved with care; there must be no excess of ammonia, no free acid, and no excess of ammonium salt beyond the quantity indicated, while that amount is necessary. A criticism of this method by Miller and Pagef was shown in a later article % to be without foundation. * Martha Austin, Am. Jour. Sci., [4], viii, 214. t School of Mines Quarterly, xxii, 391. I Am. Jour. Sci., [4], xiv, 156. CADMIUM Estimation as Cadmium Pyro phosphate. Cd 2 p 2 o 7 correspond- ing to CdCl 2 Found. Error. Error in terms of cadmium. Cd 2 P 2 7 correspond- ing to Cd found in HNaNH 4 - P0 4 .4H 2 O. NH 4 C1. Time of standing. taken. the filtrate. grin. grm. grin. grm. grm. grm. grm. hrs. 0.6972 0.6976 +o . 0004 +O.OOO2 trace 4-5 IO 16 0.6972 o . 6969 -0.0003 O.OO02 trace 4-5 10 18 0.6972 0.6962 o.ooio 0.0006 trace 4-5 IO 16 The Electrolytic Determination of Cadmium. The deposition of cadmium as the metal, upon the rotating cathode,* has been studied by Medwayf for the sulphate solu- tion, and by Flora { for solutions containing sulphuric acid, acetates, cyanides, pyrophosphates, phosphates, oxalates, for- mates, tartrates, free nitric acid, urea, formaldehyde or acetal- dehyde. Deposition from Cadmium taken as the sulphate to the amount of the Sulphuric o.2 grm. approximately and dissolved in 50 cm. 3 of lon ' water containing 10 drops of dilute sulphuric acid may be successfully deposited upon the crucible rotating at the rate of 650700 revolutions per minute. To avoid solvent action after stopping the current, dilute ammonia is added drop by drop to faint alkalinity, while the current is still passing and after complete deposition of the metal. That this procedure is satisfactory the following results of Medwayll show. Deposition from the Solution of the Sulphate. Cadmium taken. grm. Cadmium found. grm. Error, grm. Current, amp. N. D. 100 . Time, min. 0.1088 o. 1083 0.0005 2 6.6 IS 0.1088 0.1085 0.0003 2 6.6 15 0.1088 o. 1092 +0 . 0004 i-5 5 15 0.1088 0.1090 +0 . 0002 2 6.6 15 0.1088 0.1093 +0.0005 i. 5 5 12 0.1088 0.1093 +o . 0005 2 6.6 10 0.1088 0.1087 O.OOOI 2 6.6 IO * See Fig. 13, page 12. f H. E. Medway, Am. Jour. Sci., [4], xviii, 56. J Charles P. Flora, Am. Jour. Sci., [4], xx, 268 et seq., and 292 et seq. For descriptions and results, see pages 191 to 195. II Am. Jour. Sci., [4], xviii, 56. 192 METHODS IN CHEMICAL ANALYSIS In a detailed study of the use of the rotating cathode for the estimation of cadmium, Flora * calls attention to the fact that the dilution of the solution submitted to electrolysis is of great importance. It is advisable, in order to avoid mechanical loss, to deposit not more than 0.25 grm. of the metal upon the cathode, while even smaller quantities are to be preferred. The current density must also be kept within limits ; for otherwise a spongy deposit may result. Cadmium seems to be especially apt to form these spongy, unweighable deposits, and the greatest diffi- culties come from this behavior of the metal. The best con- ditions for the deposition in presence of sulphuric acid may be briefly summarized as follows: Cadmium sulphate, equivalent to not more than 0.25 grm. of the metal, is dissolved in 45 cm.* to 50 cm. 3 of water ; ten to fifteen drops of dilute sulphuric acid are added ; and the solution subjected to electrolysis at a normal current density (N. D.ioo) ranging between 3 amp. and 9 amp. per 100 cm. 2 of surface. It is not necessary to heat the liquid, as the passage of the current soon heats it sufficiently. When electrolysis is complete, the excess of sulphuric acid may be de- stroyed with a slight excess of ammonia water, the current broken, and the cathode removed, thoroughly rinsed with water and alco- hol, and dried by waving over a free flame. If the deposit is not spongy the drying is a matter of only a few moments, and there is no danger of oxidizing the metallic deposit. If it is preferred, the current may be reduced by interposed resistance, the rotation stopped, and the liquid readily siphoned without danger of in- juring the metallic coating. This process is also available when the cadmium is taken as the chloride if the volume of the solution does not exceed 45 crnAf Deposition from * n a vomme f 6o cm - 3 to 6 5 cm - 3 > containing 0.5 Solutions con- grm. to 2 grm. of sodium acetate and a small amount taining Acetates. of potass i um su lphate to so regulate the conductivity that the normal current density shall not much exceed 3 amp., the deposition of not more than 0.15 grm. of cadmium, taken as the sulphate, proceeds rapidly and satisfactorily. The deposit under the conditions is rather crystalline, fairly compact, and easily washed, so that the method forms one of the very best * Charles P. Flora, Am. Jour. Sci., [4], xx, 268-276, 392-396, 454~455- t Flora, loc. cit., page 392. CADMIUM 193 where the cadmium is taken in the form of the sulphate. At greater concentrations the precipitate shows a tendency to spongi- ness, and it fails absolutely when the cadmium is introduced as the chloride.* Deposition from Solution Containing Alkali Acetate. Cadmium taken. NaOC 2 H 3 O. K 2 S0 4 . Current. N. D.. E.M.P. Time. Cadmium found. Error. grm. grm. grm. amp. amp. volts. min. grm. grm. 0.1118 0-5 I .O 1.0 3-0 8.0 2O O. II2I +0.0003 0.1491 i-5 0-5 0.9 2.7 8.0 15 0.1494 +0.0003 0.1491 . i-5 o-5 0.9 2.7 8.0 15 o . 1496 +o . 0005 Cadmium taken. NaOH. K 2 S0 4 . Current. N.D. 100 . E.M.F. Time. Cadmium found. Error. grm. grm. grm. amp. amp. volts. min. grm. grm. O. 1491 0-5 1-25 3-75 8.0 IO 0.1496 +0.0005 0.1491 '.2* 0-5 0.8 2.4 8.0 15 0.1491 O . OOOO 0.1491 0.2* 0-5 0.8 2.4 8.0 15 0.1493 + O.OO02 0.1223 o-S* O.2 i .0 3-o 12. O 20 0.1223 O . OOOO 0.1223 0-5* O. 2 i .0 3-o 12.0 20 0.1223 0.0000 * Neutralized by a slight excess of acetic acid. Deposition from deposition of cadmium from a solution of the solutions con- double cyanide has proved to be very satisfactory, les ' and the results with the rotating cathode are in com- plete accordance with previous work on this method. The range of conditions of current and quantity of electrolyte is broad; the deposit is a beautiful silvery plate which dries very quickly, and is so compact as to be rubbed off only with difficulty; and although the complete deposition of the metal is slower than it is from solutions containing sulphates or acetates, it is suffi- ciently rapid. Care should be taken to avoid foaming of the solution, as this retards somewhat the deposition of the final traces of cadmium. Generally, a volume of 65 cm. 3 to 70 cm. 3 is found most satisfactory. In experiments to test the method the cadmium sulphate was treated with sodium hydroxide and the precipitate was dissolved in potassium cyanide. The follow- ing results were obtained. f * Flora, loc. cit., page 392. t Ibid., page 272. IQ4 METHODS IN CHEMICAL ANALYSIS Deposition from Solution Containing Alkali Cyanide. Cadmium taken. NaOH. KCN. Current. N. D. 1M . E.M.P. Time. Cadmium found. Error. grm. grm. grm. amp. amp. volts. min. grm. grni. 0.1491 i-5 o-5 2-5 7-5 8 35 0.1498 +0 . 0007 0.1491 i .0 o-S 2.5-4.5 7.5-I3-5 8 30 o. 1490 O.OOOI 0.1223 i-5 I.O 2-5 7-5 8 35 0.1225 +O.OOO2 Satisfactory results are also found by this method when the cadmium is taken as the nitrate or as the chloride * the best dilu- tion being 60 cm. 3 to 65 cm. 3 The time required is a trifle longer than in the estimation of cadmium sulphate by this method, and with the current density necessary to hasten the deposition a considerable tendency to foam is manifest. Deposition from Brand f has recommended the use of a solution Solutions Con- containing sodium pyrophosphate for the electro- taining Pyro- . KJ phosphates or lytic estimation of metals, among others, cadmium; Ortho P hos P hates. anc j tne fitness of this solution for use with the rotat- ing cathode has also been studied by Flora. | When the cadmium, taken as sulphate or chloride, is precipi- tated by sodium pyrophosphate, the precipitate dissolved in an excess of ammonium hydroxide, phosphoric acid, sulphuric acid, or hydrochloric acid, and the solution submitted to electrolysis at a volume of 60 cm. 3 , fairly accurate results may be obtained ; but neither is the method so accurate as those previously de- scribed, nor are the conditions so flexible. Practically the same thing may be said of the use of the rotating cathode in the elec- trolysis of cadmium orthophosphate dissolved in phosphoric acid according to the method recommended by Smith. Flora || has also studied the behavior of solutions containing oxalates, for- mates, tartrates, urea, formaldehyde and acetaldehyde, nitrates and free nitric acid. The results of this study and of the work previously described may be summarized as follows: Summary. Cadmium taken in the form of the sulphate may be very accurately and satisfactorily estimated by deposition * Flora, loc. cit., page 393, and page 454. t Zeit. anal. Chem., xxviii, 581. J Loc. cit., page 273. Am. Chem. Jour., xii, 329. II Loc. cit. MERCURY 195 upon the rotating cathode from solutions containing sulphuric acid, sodium acetate and acetic acid, or potassium cyanide; but little less satisfactorily from solutions containing urea, formal- dehyde or acetaldehyde ; and also, with proper precautions, from solutions containing pyrophosphates, phosphates, tartaric acid or formic acid. From solutions containing oxalates or oxalic acid, ammonium tartrate or potassium formate satisfactory de- posits are not obtained. When taken as the chloride, cadmium does not permit such a wide range of conditions. Nevertheless, from solutions of the chloride containing sulphuric acid or potassium cyanide, or the pyrophosphates, the metal is deposited in a form comparable with that obtained when cadmium sulphate is taken. Solutions of the chloride of cadmium to which is added hydrogen disodic phosphate give less desirable results; while solutions containing urea, formaldehyde or acetaldehyde give deposits free from spon- giness only after careful regulation of the conditions. In solu- tions containing the oxalates, oxalic acid, the formates and the tartrates, acetates, formic acid and tartaric acid the results were negative. Cadmium nitrate is in general ill-fitted for elec- trolytic estimation, the cyanide solution being the only one from which satisfactory results were obtained. From solutions con- taining one per cent or more of free nitric acid, the cadmium is not deposited by the current. MERCURY. The Gravimetric Determination of Mercury as Mercurous Oxalate. Mercury taken in the form of mercurous nitrate may be esti- mated as mercurous oxalate precipitated by ammonium oxalate and dried over sulphuric acid, as has been shown by Peters.* It is necessary, however, to control the acidity, dilution, and pres- ence of mercuric salts. It appears that 5 cm. 3 of dilute nitric acid, sp. gr. 1.15, may be present in a volume of 100 cm. 3 , and that 5 cm. 3 of the acid will prevent precipitation of small amounts of mercuric salt, o.oioo grm. to 0.0200 grm., calculated as mer- cury, depending upon the amount of ammonium oxalate present in excess. According to the procedure recommended by Peters, mercurous nitrate dissolved in 100 cm. 3 of water containing 2 per * C. A. Peters, Am. Jour. Sci., [4], ix, 405. 196 METHODS IN CHEMICAL ANALYSIS cent to 5 per cent of dilute nitric acid, sp. gr. 1.15, is precipitated by the addition of ammonium oxalate in slight excess with stir- ring. It is an easy matter to keep the excess of the precipitant within the limits of I cm. 3 to 2 cm. 3 of the n/io solution, because the mercurous oxalate, when properly stirred, settles rapidly. The precipitate is collected on asbestos in a perforated crucible, washed two or three times with cold water, and dried to constant weight over sulphuric acid at the ordinary temperature, since mercurous oxalate is slowly decomposed at temperatures in the vicinity of 100. Mercury in the mercuric form may be safely present to the amount of 10 per cent of the mercurous salt. Results of experiments according to this procedure, in which the drying was effected by exposure for fifteen hours or less at ordinary temperatures over sulphuric acid, are given in the accompanying table. Mercurous Oxalate Dried over Sulphuric Acid. Hg taken as Hg,(NO,),. Hg present as Hg(NO,),. Excess of H/IO ammonium oxalate. HNO 3 (sp.gr. 1. 15.) Volume at precipita- tion. Hg found. Error. grm. grm. cm. 3 cm. 8 cm. s grm. grm. o. 1217 2-4 IOO 0.1217 O.OOOO O.I2I7 2-4 IOO o. 1217 O.OOOO O. 1122 2-4 IOO o. 1124 +O.OOO2 O.II22 0.0067 0-93 2 IOO 0.1130 +0.0008 O.II22 0.0067 0-93 2 IOO O. III2 O.OOIO O.II22 0.0067 4.40 2 IOO o. 1124 +O.OOO2 O.II22 0.0135 0.72 4 IOO o. 1125 +o . 0003 0.2244 0.0071 1.68 4 IOO 0.2253 +o . 0009 O.2244 0.0071 2.46 4 IOO o. 2241 0.0003 O.2244 0.0048 0-54 4 2OO 0.2248 +O.OOO4 O.2244 0.0048 2.44 4 200 0.2245 +0.0001 The Determination of Mercury by Titration with Sodium Thio- sulphate. *. The method proposed by Scherer* for the determination of mercury by titration of mercury salts with sodium thiosulphate has been studied by Norton. f It is shown in the estimation of mercury taken in the form of mercuric chloride the reaction pro- ceeds definitely according to the equation 3 HgCl 2 +2 Na 2 S 2 3 +3 H 2 O = 2 HgS.HgCl,+2 Na 2 SO 4 +4 HC1. * Lehrbuch der Chemie, i, 513. t John T. Norton, Am. Jour. Sci., (4], x, 48. MERCURY 197 This method yields accurate results if carried out under certain fixed conditions. These conditions, which must be closely ad- hered to, are as follows: The solution containing the mercury in the form of mercuric chloride is diluted to 100 cm. 3 and heated to a temperature of 60 C. The sodium thiosulphate in ^ O th normal solution is run in from a burette until the white pre- cipitate, 2 HgS.HgCl 2 , first formed begins to take on a brownish tinge due to incipient formation of the sulphide. The solution is then diluted with cold water, some asbestos fiber added to coagu- late the precipitate, and the whole is quickly thrown on the filter. After careful washing, potassium iodide is added to the filtrate and the excess thiosulphate is titrated with iodine in presence of starch. The duration of the process need not exceed fifteen minutes. It is worthy of note that there is no necessity of using any hydrochloric acid in addition to that formed in the reaction. Results obtained by this procedure are given in the table. Titration of Mercuric Chloride. HgCl 2 taken as Hg. grm. Volume at beginning. cm. 3 Temper- ature. C. Na 2 S 2 3 in excess. cm.a HgCl 2 found as Hg. grill* Error, grm. 0.0759 TOO 60 3-06 0.0766 +0.0007 0.0384 IOO 60 2.8l 0.0387 +0 . 0003 o. 1498 IOO 60 I . I 0.1500 +0.0008 0.1503 IOO 60 1.63 0.1506 +o . 0003 0.1479 IOO 60 2.41 o. 1480 +O.OOOI o. 1489 IOO 60 2. 12 0.1503 +0.0014 o. 2244 IOO 60 2.63 0.2259 +0.0015 o. 1490 IOO 60 2-33 0.1484 0.0006 0.0758 IOO 60 2. 0.0762 +o . 0004 0.0383 IOO 60 2.53 0.0379 0.0004 In applying Scherers process to mercurous nitrate and mer- curic nitrate, Norton was unable to discover conditions of definite action. The Estimation of Mercury by Precipitation as Mercurous Oxalate and Titration of the Excess of Precipitant with Permanganate. Peters* has shown that mercury taken as mercurous nitrate may be estimated by precipitating it as the oxalate and deter- mining by titration with potassium permanganate the excess of * C. A. Peters, Am. Jour. Sci., [4], ix, 401. 198 METHODS IN CHEMICAL ANALYSIS ammonium oxalate used as the precipitant.* The operation is successful in presence of 2 per cent to 5 per cent of dilute nitric acid, sp. gr. 1.15, in 100 cm. 3 of the mixture; and mercuric salt to the extent of 10 per cent of the amount of the mercurous salt may be present provided the nitric acid amounts to 2 per cent, and the excess of ammonium oxalate used to effect precipitation is not too great. The precipitate settles well when properly stirred, and the excess of the precipitant need not exceed the safe limit of I cm. 3 or 2 cm. 3 . Following are results obtained by the permanganate titration of the filtrate from the precipitated mercurous oxalate. Precipitation by Ammonium Oxalate and Titration of the Excess. Hg taken as Hg 2 (N0 3 ) 2 . grin. Hg present as Hg(NO,),. grin. Excess of w/io ammo- nium oxalate. cm. 1 HN0 3 (sp. gr. I.IS). cm. 1 Volume at precipita- tion. cm. 8 Hg found. gnu . Error. O.I2I7 0.0067 0.86 2 IOO 0.1232 0.0015 0.1217 0.0067 0.92 2 IOO O. I22O +0.0003 o. 1217 0.0067 0.97 2 IOO 0.1218 -j-o.oooi o. 1217 0.0067 0.90 2 IOO o. 1224 +0.0007 O.I2I7 0.0067 0.91 2 IOO 0.1213 0.0004 O.I22I 0.0067 3-92 2 IOO 0.1237 +0.0016 O.I2I7 0.0067 3-93 2 IOO 0.1235 +0.0018 O.I242 0.0067 8-75 2 IOO o. 1263 +O.OO2I o. 1217 0.0134 0.87 2 IOO 0.1230 +0.0013 O.I2I7 0.0134 0.86 2 IOO 0.1232 +0.0015 O.I2I7 0.0134 3-93 4 IOO o. 1218 +O.OOOI O.I2I7 0.0134 3-90 4 IOO O.I2II 0.0006 0.2244 . 0.0067 1.88 4 115 0.2244 o . oooo O.2244 0.0067 1.91 4 130 0.2240 0.0004 0.2244 0.0141 2.98 4 IOO 0.2230 0.0014 0.2244 0.0141 2.94 4 IOO 0.2241 0.0003 The Titration of Mercurous Salts with Potassium Permanganate. If a solution of a mercurous salt, such as mercurous sulphate, in dilute sulphuric acid, is titrated with potassium permanganate in the usual manner, the bleaching of the color is rapid at first, but long before the oxidation is complete the solution assumes a golden-yellow color and on standing the brown oxides of man- ganese are precipitated. For this reason no definite end reac- tion is obtainable. This difficulty may, however, be avoided, as Randall t has shown, if the permanganate solution is added in * See page 195. f D. L. Randall, Am. Jour. Sci., [4], xxiii, 137. MERCURY 199 excess, the color then bleached with a standard ferrous sulphate solution, and the end-point finally reached by a few drops of permanganate. Under these conditions the end reaction is per- fectly sharp and the oxidation of the mercurous salt complete. According to Randall's procedure, the mercurous sulphate or nitrate in solution in water containing a suitable excess of sul- phuric acid, or nitric acid free from oxides of nitrogen, is diluted and treated with n/2O potassium permanganate until the mix- ture, colored brown by the oxides of manganese, takes on a distinctly red tint. Ferrous sulphate in n/io solution is added in amount sufficient to clear the solution, and the titration is immediately completed with permanganate. The n/2O solu- tions are preferred to the usual n/io solutions on account of the high equivalent of the mercurous salt. Even with n/2O solu- tion, o.i cm. 3 of the permanganate is equivalent to about o.ooio grm. of mercury. For the same reason the titrations are made with all possible care and accuracy. Following are experimental results obtained with a mercurous sulphate solution prepared by shaking up an excess of the salt in water acidified with sulphuric acid, allowing the mixture to stand twenty-four hours, and filtering through asbestos. The solution was standardized by weighing the mercurous chloride precipitated by sodium chlo- ride and dried in a vacuum. Titration of Mercurous Sulphate. Hg 2 S0 4 sol. cm.* H 2 S0 4 [i : i]. cm.* ' KMn0 4 approx. n/2o. cm. 8 FeSO 4 = cm. 8 = KMnO 4 cm. 8 KMnO 4 final. cm. 8 Hg found. grm. Hg. (theory). grm. Error, grm. Volume 150 cm. 3 . IOO 5 14.70 10 10.90 3.80 0.0346 0-0354 0.0008 IOO 5 14.72 10 10.90 3.82 0.0347 0.0354 0.0007 IOO 5 14.70 10 10.90 3.80 0.0346 0.0354 0.0008 IOO 5 14.71 10 10.90 3.81 0.0346 0.0354 0.0008 H g2 S0 4 sol. Volume 500 cm. 3 . 505.8 5 29.20 IO 10.90 18.30 0.1668 0.1666 +O.OOO2 500.2 5 32.16 13 14.17 17.99 0.1639 0.1648 0.0009 510-1 499-2 5 5 29.29 28.95 IO IO 10.90 10.90 18.39 18.05 0.1676 0.1645 0.1680 0.1644 0.0004 +0.0001 200 METHODS IN CHEMICAL ANALYSIS For practical purposes the application of the method to mer- curous nitrate is of much greater importance. Results obtained with mercurous nitrate in solution prepared by dissolving the crystallized salt in water containing enough pure nitric acid to prevent the formation of basic salts, and, as an additional pre- caution, passing a current of hydrogen, washed by alkaline per- manganate and alkaline pyrogallol, for twelve hours through the solution to remove nitrous acid are given below. Titration of Mercurous Nitrate. Volume 200 cm. 3 . TTa (NC& 2 HNO 3 . Approx. H/2O KMn0 4 . FeSO 4 . = KMnO 4 KMnO 4 final. Hg found. Hg theory. Error. cm. 3 cm. 3 cm. 3 cm. 8 cm. 3 cm. 3 grm. grm. grm. 25 5 49-99 9.^ 10.64 39-35 0.3586 0-3594 O.OOoS 25 5 49-68 9-50 10.36 39-32 0.3583 0-3594 O.OOII 25 5 49-73 9-50 10.36 39-37 0.3588 0-3594 O.O006 25 5 49.89 9.67 10.54 39-35 0.3586 0-3594 O.OOOS 25 5 49.70 9-50 10.36 39-34 0.3585 0-3594 O.OOO9 H 2 S0 4 [i : i]. 25 5 50.29 IO.OO 10.90 39-39 0.3589 0-3594 0.0005 CHAPTER VI. BORON; ALUMINIUM; LANTHANUM; THALLIUM. BORON. The Gravimetric Determination of Boric Acid. The use of ^ HE P rocess f determining boric acid by distil- Caicium Oxide lation with methyl alcohol, evaporation of the dis- as a Retainer. t jn ate on ca l c i um oxide, and ignition of the residue,* has been reexamined by Gooch and Jones f in the light of the experience of many investigators.t It is shown that difficulties arising when nitric acid is present in the retort may be obviated by limiting the amount of that acid by the use of phenolphthalein as an indicator at the outset of the distillation. The addition of a drop of the acid and another of the indicator should be re- peated once or twice during the distillation to insure the replace- ment of the acid volatilized from the salt slightly decomposed in the process. The effect of much nitric acid is bad, not only because it neutralizes the calcium oxide when it passes to the distillate, but because when it is used the dried mixture of calcium hydroxide and borate containing nitrate, nitrite and organic matter is likely to puff explosively if ignition is begun as soon as the residue is dry. If the residue is heated gradually and as strongly as possible over a radiator before the flame is actually applied to the crucible, no such action takes place. That good results may be obtained with small amounts of calcium oxide, provided care as to the use of nitric acid and the conditions of ignition be taken, is shown by the figures of the original description and by the following results of experiments in which phenolphthalein was employed as an indicator and the residue heated strongly over the radiator before actual ignition. * Gooch, Am. Chem. Jour., ix, 23. t F. A. Gooch and Louis Cleveland Jones, Am. Jour. Sci., [4], vii, 34. t Penfield, Am. Jour. Sci., [3], xxxiv, 222; Kraut, Zeit. anal, Chem., xxxvi, 3; Moissan, Compt. rend., cxvi, 1084. 201 202 METHODS IN CHEMICAL ANALYSIS Distillation with Nitric Acid. CaO taken. B 2 O 3 taken. B 2 O 3 found. Error. grin. grm. grm. grm. 2.3405 i . 7620 2.1757 2.5656 0.1788 0.1790 . 0.1824 0.1788 0.1792 0.1785 0.1840 0.1786 +9.0004 0.0005 +0.0016 O.OO02 No difficulty exists in respect to explosive effects when acetic acid is used in place of nitric acid, though even in this case it is safer to use the radiator in the first stages of heating, thus avoid- ing the danger of mechanical loss by too rapid ignition. Following are determinations made by this method with the use of acetic acid. Distillation with Acetic Acid. CaO taken. B 2 O 3 taken. B 2 O a found. Error. grm. grm. grm. grm. 0.9977 I.O22O o . 2065 0.2067 O . 2062 0.2070 0.0003 +0.0003 I-37I7 I.I3IO 0.2077 0.1791 0.2075 0.1795 O.OOO2 +O.OOO4 The results of the preceding table, as well as those of the inves- tigators mentioned, are a sufficient answer to the criticism* that acetic acid and nitric acid do not liberate boric acid in the dis- tillation process so that good results may be obtained. In fact, it has been shown f that the prolonged action of carbonic acid is adequate to bring about complete volatility of boric acid with methyl alcohol. The number of distillations required depends, of course, upon the amount of boric acid to be volatilized. To remove 0.2 grm. of the anhydride completely to the distillate, it was shown in the original description of the method J that six io-cm. 3 charges of methyl alcohol suffice. It is also shown that the difficulty in bringing calcium oxide to a constant weight before and after absorption of boric acid has been magnified unduly. Thus the following table shows the * Reischle, Zeit. anal. Chem., xxvi, 512. f Jones, Am. Jour. Sci., [4], v, 442. J Loc. cit. Thaddeef, Zeit. anal. Chem., xxxvi, .568. BORON 203 series of weights taken in several experiments in bringing calcium oxide to a constant weight in a 5O-cm. 3 platinum crucible ignited over a blast lamp, as well as the weight taken after adding a known amount of standard boric acid solution to the slaked oxide, evaporating, and igniting. The results recorded are those of experiments made on days not moist beyond the average, and with the greatest care to approach the limit of accuracy with which calcium oxide and the boric acid held thereby can be weighed under ordinarily favorable conditions. The first weight of calcium oxide recorded under each experiment was taken after a strong ignition over the blast lamp for about one-half hour. The succeeding weights were taken after similar ignition of five minutes. In all cases the crucible was left to stand a definite period in a sulphuric acid dessicator, and, after the approximate value had once been obtained, the weights of the preceding weigh- ing were replaced on the balance before the crucible was taken from the dessicator. The average of the weights bracketed is the weight taken as constant for the calculations. Ignition of Calcium Oxide and Calcium Bofate. CaO taken, grm. B 2 O 3 taken, grm. CaO+B 2 O 3 taken. grm. CaO+B 2 3 found. grm. Error, grm. (0.9505 (i Ho. 9493 [o.9493 (0.9493 \ 0.2095 I . 1588 i'::S?}'-'* +0.0003 ( 4 ; .lii'- 1315 0.2150 L3465 1-3499 1-3474) 1.3475 > 1.3475 1.3476 ) +O.OOIO (0.8028 0.9205 (3) (0:8024 1- 8 24 0.1184 0.9208 0^206 I " 9206 +0.0002 1 C 2 . 6980 ( 4 )J 2-6975 i 2 6o73 I f U.' 6973 1 2 - 6973 0.2073 2.9046 2 . 9043 +O.OOO2 Obviously calcium oxide may be weighed with accuracy, with or without boric acid ; but the fact remains that a less hygroscopic absorbent one requiring less care in the handling is to be desired. 204 METHODS IN CHEMICAL ANALYSIS The Use of Gooch and Jones find that sodium tungstate, fused 1 8 " a sn 'g nt excess of tungstic acid over that con- Retainer. tained in the normal tungstate (to insure its freedom from carbonate), may be used with good results as an absorbent for boric acid. This substance is definite in weight, not hydro- scopic, soluble in water, and recoverable in its original weight after evaporation and ignition. According to the procedure advocated, use is made of the apparatus originally proposed, so modified that the Erlenmeyer flask used as a receiver is fitted tightly to the condenser and trapped with water bulbs.* The retort is made very easily from a i5O-cm. 3 pipette and has the special advantage that particles of the residue spattering during distillation are easily washed from the walls of the vessel by a slight rotary motion of the retort. Special care should be taken to give the tungstate ample time for contact with the distillate before exposing the latter to atmospheric evaporation. The distillate is received, therefore, in a dilute solution of sodium tungstate which is placed in the receiver cooled by ice and trapped with water. The mixture is well stirred, allowed to stand one-half hour after the distilla- tion is complete, evaporated to small volume in a large dish, and transferred to the crucible in which the tungstate was originally weighed. After thorough drying the residue is ignited to fusion and weighed. When acetic acid is employed in the retort, care must be taken in the ignition to expose the fused mass freely to the air (by causing it to flow upon the sides of the crucible) until the color of the cooled tungstate is white, in order that the re- ducing effect of the acetate may be eliminated. In the experi- ments recorded in the following table the tungstate was used in the receiver to retain the boric acid, distilled as usual with methyl alcohol from the borates treated with acetic acid, nitric acid or sulphuric acid, in amounts regulated by the use of phenolphtha- lein as an indicator. Excessive use of acid is disadvantageous, and this is especially true in the case of sulphuric acid; for, if this acid is carried over with the methyl alcohol, as it is at 100 if present in appreciable excess, a part of it, at least, is held per- manently by the tungstate to increase the apparent weight of the boric acid estimated. The number of distillations necessary depends upon the amount of boric acid to be volatilized. Six * See Fig. 2, page 3. BORON 205 charges of 10 cm. 3 each of methyl alcohol are enough to transfer 0.2 grm. of boric anhydride to the distillate. Distillation of Six 10 cm. 3 Portions of Methyl Alcohol. Na 2 WO 4 +WO 3 taken. B 2 O 3 taken. B 2 O 3 found. Error. gnu. grm. grm. grm. With nitric acid! 8.5516 4-9639 8.0033 0.1582 0.1329 0.1267 0.1572 0.1323 o. 1256 O.OOIO O.OOO6 O.OOII With acetic acid. 4.9658 0.1434 o. 1418 0.0016 6.0289 0.1431 0.1433 4-O.OOO2 4.6797 0.1589 0.1587 O.OOO2 4.0013 0.1433 0.1422 O.OOII With sulphuric acid. 6.3439 0.1582 0.1579 -0.0003 8.8227 0.1582 0.1577 -0.0005 10. 1516 0.1265 o. 1264 O.OOOI 6.5738 0.1392 0.1390 O.OOO2 The Acidimetric Estimation of Boric Acid. When boric acid and mannite are mixed in solution, a peculiar compound of strongly acid properties is the result. This com- pound decomposes carbonates, and its acid taste is comparable to that of citric acid, much stronger than that of boric acid alone. Magnanini * has found that the product of such a mixture of boric acid and mannite solutions shows greater electrical conductivity and a lower freezing point than a similar molecular solution of either substance alone. Other polyatomic alcohols (but all to a less degree than mannite) and some organic acids show this peculiar property of combining chemically with boric acid to increase its acid qualities. f Of the reaction between boric acid and glycerin, Thomson, t Barthe, and J6rgensen|| have taken * Gaz. Chim., xx, 428-440; xxi, 134-145. f Klein, Compt. rend. Ixxxvi, 826; xcix, 144. Lambert, ibid., cviii, 1016 1017. t J. S. C. I., xv, 432. J. Pharm. Chim., xxix, 163. II Zeit. angew. Chem. (1897), 5. 206 METHODS IN CHEMICAL ANALYSIS advantage to develop methods for the volumetric estimation of boric acid, glycerin being used to form a combination with boric acid sufficiently acidic to give an acid reaction when used with a sensitive indicator and make possible its titration with an alkali solution; but Honig and Spitz* show that (in the method of Jorgensen) a very large amount of glycerin must be used to pre- vent the appearance of the indication of alkalinity with phenol- phthalein before all the boric acid is neutralized according to the equation 2 NaOH + B 2 O 3 = 2 NaOBO + H 2 O. Vadam f uses mannite, which, as he finds, gives sharper indications with litmus. The solution must be boiled to decompose bicarbonates while the volatilization of boric acid is prevented by the use of a return condenser; and silica must be removed by the process of Berzelius, and the solution neutralized to stronger acids, before a titration of the boric acid can be made. Many indicators said to be insensible to free boric acid have been used to indicate the neutralization of the stronger acids. Honig and Spitz, J and Thomson, use methyl orange; Morse and Burton|| use tropaeolin oo; while Vadam IT makes use of litmus. Neutralization Finding all these indicators to be more or less of stronger affected by boric acid, Jones** has had recourse to the well-known reaction according to which a stronger acid liberates regularly iodine from a mixture of iodide and iodate, which is the solution of this difficulty. If both the iodide and iodate are in excess of the acid, the entire amount of free acid will be neutralized and the corresponding amount of iodine liberated according to the following equation : 5 KI + KIO 3 + 6 HC1 = 6 KC1 + 3 H 2 O + 3 I 2 . This liberated iodine may be removed by sodium thiosulphate and a solution obtained which is absolutely neutral, containing potassium iodide, iodate and tetrathionate. The statements made by P. Georgevicft an d Furry, jj that boric acid present in * Zeit. angew. Chem. (1896), 549. t J. Pharm. Chim. [6], viii, 109-111. J Zeit. anorg. Chem., xviii, 549. J. S. C. I., xv, 432. || Am. Chem. Jour., x, 154. 1 J. Pharm. Chim. [6], viii, 109-111. ** L. C. Jones, Am. Jour. Sci., [4], vii, 147. tt J prakt. Chem., xxxviii, 118. tt Am. Chem. Jour., vi, 341. BORON 207 moderate amount in solution has not the slightest action on a mixture of iodide and iodate, have been found to apply to solu- tions containing not more than o.i grm. of B 2 C>3 to 25 cm. 3 . Therefore, when this acid is liberated by an excess of a stronger acid and the iodine set free destroyed by thiosulphate, it remains free in solution to be titrated in any convenient manner. Follow- ing along the lines suggested by the above reactions, Jones * has developed a volumetric process for the estimation of boric acid. The solution in which boric acid is present to an amount not exceeding o. I grm. in 25 cm. 3 is made slightly acid to litmus by hydrochloric acid and treated with 5 cm. 3 of a solution (10 per cent) of barium chloride. An amount of iodate and iodide of potassium sufficient to liberate iodine at least equivalent to the excess of hydrochloric acid in the acidified solution is mixed with starch in a separate beaker, the iodine which is usually thrown out by this mixture being just bleached by a dilute solution of thiosulphate. To the now neutral solution of iodide and iodate a single drop of the solution to be analyzed is transferred by a glass rod. If a blue coloration is developed, the solution is known to be acidic with hydrochloric acid, and all the boric acid is in free condition. The amount of iodide and iodate necessary depends upon the acidity of the solution containing boric acid. Usually a mixture of 10 cm. 3 of a 25 per cent solution of iodide with the same amount of a saturated solution of iodate is sufficient. Any large excess of hydrochloric acid should be neutralized by sodium hydroxide before the iodide and iodate mixture is added. After the addition of the iodide and iodate solution and starch to the boric acid solution, the liberated iodine is carefully bleached by thiosulphate. Excess of thiosulphate in reasonable amount does not seem to be detrimental, but in practice the starch iodide color is clearly bleached, and then no more is added. Soluble carbonates prevent a definite indication of the neutral point by thiosulphate and starch iodide, therefore the barium chloride was added to transform them to insoluble barium car- bonate in the action of the iodide-iodate mixture. The mixture of iodide and iodate is not added to the solution to be analyzed until after it has been made distinctly acidic, for the reason that, when the neutral point is approached in the addition of * Am. Jour. Sci., [4], viii, 129. 208 METHODS IN CHEMICAL ANALYSIS hydrochloric acid, the starch iodide thrown out locally by the acid is not bleached again by the small amount of sodium borate remaining undecomposed and thus obscures the neutral point, strengthening The solution, after the bleaching of iodine by of Boric Acid thiosulphate, is colorless and contains only starch, neutral chloride, potassium tetrathionate, iodide and iodate, and all the boric acid present in uncombined condition. The carbonate lies out of the sphere of action in insoluble form as barium carbonate. A few drops of the indicator, phenol- phthalein, are now added, and n/$ sodium hydroxide is run in until a strong red coloration is produced. A pinch of mannite, I grm. or 2 grm., is then added, which bleaches the phenol- phthalein coloration, and the alkali solution again run in to a faint indication, which, if permanent on the addition of more mannite, may be taken as the reading point. The combination of boric acid and mannite liberates immedi- ately in the presence of iodide and iodate about half the amount of iodine required on the theory that B 2 O 3 acts with the neu- tralizing power of metaboric acid, HOBO. If no mannite is present phenolphthalein gives an alkaline indication when only about one-half the amount of alkali theoretically necessary to form the metaborate, NaOBO, has been added. Obviously, then, the starch iodide coloration will not appear on the addition of mannite, if the free boric acid has been neutralized to phenol- phthalein by alkali, and the remainder of the alkali is added to complete neutralization immediately after the addition of mannite. The end reaction with phenolphthalein is sharp and the small amount of carbonate present in the standard solution of alkali is precipitated by the barium chloride already in the solution. The calculation must, therefore, be based on the amount of free alkali hydroxide used, according to the following representation : B 2 O 3 + 2 NaOH = 2 NaOBO + H 2 O. The best results and the most definite indications are obtained in cold solution of a volume not greater than 50 cm. 3 . This fact accords with the observations of Magnanini* that the relative electrical conductivity of the boromannite solution is decreased by dilution and elevation of the temperature. When silicates are * Gaz. Chim., xx, 428, and xxi, 134. BORON 209 present in solution, the silicon dioxide is liberated by the excess of hydrochloric acid, and this oxide, whether in hydrous or an- hydrous condition, neither affects the indication with iodine nor phenolphthalein, nor does it form with mannite a compound of acidic properties. The presence of fluorides is not detrimental. Ammonium salts do interfere with the indication given by phenolphthalein, but they may be removed by boiling with potassium hydroxide in excess, or an indicator which is not affected by them may be used. The following table contains the results of a series of analyses in which the boric acid was first drawn into an excess of sodium hydroxide, then estimated according to the method described. Titration of Boromannite Solution with Standard Alkali. B 2 O 3 sol. taken, cm.' NaOH sol. required. cm. 3 B 2 O S taken, grm. B 2 O 3 found, grm. Errors on B 2 O 3 . grm. 21-95 21. 02 0.1571 0.1577 +o . 0006 20.68 19.65 0.1479 0.1474 0.0005 20.73 19.63 0.1483 0.1473 O.OOIO 23-05 23.71 0.1776 0.1777 +0.0001 23.10 23.80 0.1780 0.1783 +0.0003 22.76 23-35 0.1754 0.1750 0.0004 24.08 24.78 0-1855 0.1857 +O.OOO2 22.00 22.50 0.1695 0.1686 O.OOOQ 20.78 21.28 0.1601 0.1595 O.OOO6 Practical tests of the method upon crude calcium borate and colemanite are given below. The finely ground minerals were dissolved in hydrochloric acid and the analyses proceeded with as above described. A determination of boric acid by this process can be completed in five minutes and the results are obviously accurate within the limits of ordinary analysis. Analysis of Crude Borate of Lime. Ca borate taken. B 2 O 3 found. B 2 3 . grm. grm. Per cent. 0.4016 0.4044 0.4000 0.2289 0.2302 0.2285 56.99 56.92 57-11 210 METHODS IN CHEMICAL ANALYSIS Analysis of Colemanite. Mineral taken, grm. B 2 O, found, grm. B 2 3 . Per cent. Average. Per cent. 0.4034 O . 2064 51.15 ] o . 4070 o . 2069 50.80 0.6004 0.6006 0.3054 0.3056 50.86 50.89 |- 50.99 0-5059 0.2592 5I-24 0.5092 0.2592 50.89 J The usually interfering substances, fluorine, silica and car- bon dioxide, have no detrimental influence on the results of this process. The lodometric Determination of Boric Acid. In studying the strongly acidic compound formed when boric acid and mannite are associated in solution, Jones* finds that the acid developed is, under certain definite conditions, suffi- ciently strong to liberate, quantitatively, from a mixture of potassium iodide and iodate, the amount of iodine required on the supposition that each molecule of metaboric acid (HOBO) acts in a manner similar to a univalent mineral acid under the same conditions. 5KI +KIO 3 + 6HOBO = 3 I 2 + 6 KOBO + 3 H 2 O. Obviously, this reaction depends upon the behavior of the acidic boromannite compound as an acid stronger than acetic, tartaric or citric acid ; for these acids have been found by Furry f to be incapable of liberating iodine regularly from a mixture of iodide and iodate. Conditions which tend to increase the acidic activity of this compound are high concentrations and moderately low tempera tures.f It has not been found possible under any conditions to rely upon the immediate liberation of the full amount of iodine; a certain period of time is required for the completion of the reaction. When the solution is of small volume and saturated with mannite, the reaction goes tp the end most quickly * Am. Jour. Sci., [4], viii, 127. t Am. Chem. Jour., vi, 341. J Magnanini, Gaz. Chim., xx, 428; xxi, 134, 1016, 1017. BORON '. 211 sometimes almost immediately ; but if the solution of boric acid is too concentrated nearly saturated the boric acid alone throws out some iodine from the iodide-iodate mixture added to destroy other free acid, and on bleaching with thiosulphate a starting point is obtained at which some of the boric acid has already entered into combination. The amount of iodine thus liberated by the boric acid is, however, not large, and if, upon the addition of the iodide and iodate, the iodine thrown out by the free hydrochloric acid present is immediately bleached by thiosulphate and the analysis proceeded with from this as the neutral point, even in concentrated solutions the error is almost inappreciable. If, however, considerable time intervenes be- tween the adding of the iodide and iodate and the determination of the neutral point by thiosulphate, iodine equivalent to as much as several milligrams of boric acid may be liberated. This difficulty was not met with in those experiments in which the iodide and iodate were added to solutions of concentration like that of the standard solution used (7.738 grm. per liter), but in an attempt to estimate the boric acid in colemanite, where the solution was kept as concentrated as possible, hoping in this way to decrease the time required for the complete liberation of iodine, low values were obtained; that is, a false starting point was used. At the time of adding the iodide and iodate the vol- ume should not be less than 25 cm. 3 for each decigram of boric anhydride (B 2 Oi) present, and should not be much greater than two or three times that amount. At lower concentrations of the boric acid, even though the liquid be saturated with mannite, the necessary time of standing is prolonged and the effect of car- bon dioxide upon the iodide and iodate is increased ; for carbon dioxide, whether derived from the atmosphere or existing dis- solved in the solution, upon standing slowly liberates iodine. The effect of carbon dioxide is, however, small, and in the time required for the completion of the process has never been found equivalent to more than a single drop of the solution of thiosul- phate used. Even if the material to be analyzed contains carbonates, after acidifying in concentrated solution and shaking vigorously the small amount of uncombined carbon dioxide remaining has an almost inappreciable effect upon the results. The length of time required for the liberation of the theoretical amount of iodine in a solution of 25 cm. 3 to 50 cm. 3 to each 212 METHODS IN CHEMICAL ANALYSIS o.i grm. of boric anhydride is 20 to 45 minutes, and at the end of 45 minutes' standing in a solution saturated with man- nite the reaction may be considered complete. During this period, however, it is well to keep the solution cool zero tem- perature will do no harm and to insure thorough mixture by occasional shaking. As free iodine would tend to escape upon standing unless kept in a closed flask, it is convenient, immedi- ately after the addition of mannite, to treat with an excess of the standard solution of thiosulphate, 8 cm. 3 or iocm. 3 more than the amount required to bleach the iodine liberated, and at the expiration of 40 to 60 minutes to titrate back with n/io iodine. The strength of the thiosulphate solution found most convenient is n/5, while the use of iodine of one-half this strength, n/io, diminishes the error of reading correspondingly. In solutions of the volume recommended the addition of starch to give the indication with iodine is unnecessary, since a single drop of one- twentieth normal iodine in excess is sufficient to give a strong lemon coloration. Procedure in Summary. The procedure recommended is as follows : The borate is dissolved in as small volume and as little hydrochloric acid as possible, with shaking to remove free carbon dioxide and adjustment of volume so that at the time of adding potassium iodide and iodate there shall be approximately 25 cm. 3 - 50 cm. 3 of solution for each decigram of boric anhydride present. The greater part of the excess of hydrochloric acid in the solution is destroyed by sodium hydroxide, with the use of litmus paper as an indicator, leaving the solution distinctly acid in reaction. Potassium iodide (3 cm. 3 -5 cm. 3 of a 40 per cent solution) and iodate (5 cm. 3 -io cm. 3 of a 5 per cent solution) are added in excess of that required to liberate iodine in an amount corresponding to the hydrochloric acid and the boric acid present. The iodine liberated by the free hydrochloric acid is bleached by a small amount of a strong solution of thiosulphate, and, after agitating to insure thorough mixture, iodine is added to faint coloration. Sufficient mannite is now used to saturate the solution about 10 grm.-i5 grm. for a volume of 50 cm. 3 and sodium thiosul- phate is added in standard solution 8 cm. 3 -io cm. 3 in excess of that required to bleach the iodine immediately thrown out by the mannite. The solution is again brought to saturation, if nec- essary, by mannite, and, after standing in a cool place for 40-60 BORON 213 minutes, titrated with decinormal iodine to determine the excess of thiosulphate present. The results of experiments upon pure boric acid, crude calcium borate and crystallized colemanite are given in the accompany- ing tables. Pure Boric Acid. B 2 3 * taken. cm .3 Thiosul- phate f taken. cm. 3 lodinej taken. cm. 3 Time of standing. min. Volume. cm. 8 B 2 3 taken. grm. B 2 O 3 found, grm. Error, grm. 28.00 32.00 1.88 30 28 0.2165 0.2168 +0.0003 27.03 32.00 4-37 27 27 0.2090 o. 2081 0.0009 27.02 31-97 4.04 60 27 0.2089 o . 2090 +0 . OOOI 27.06 32.04 3.88 60 50-60 0.2093 O. 2IOI +0.0008 27.02 32.02 4.40 60 50-60 0.2089 o. 2081 0.0008 27.04 31.72 3-39 60 50-60 0.2091 o . 2096 +0.0005 27.01 3 J -53 2.88 1 2O 50-60 0.2089 O. 2IOO +O.OOII 26.05 31.01 4.01 1 80 50-60 0.2014 0.2025 +O.OOII 27.00 31.00 2. 12 3 50-60 0.2088 o . 2089 +0.0001 27.00 32.00 4-05 30 50-60 0.2088 O.2O92 +0.0004 26.01 32.02 6. 20 30 50-60 O. 2OII 0.2018 +o . 0007 27.03 31.01 2.21 48 50-60 O.2O9O 0.2087 0.0003 27.05 31.89 3-8l 45 50-60 O.2092 0.2093 +O . OOOI 26.07 31.02 4.14 40 50-60 o. 2016 O. 2O2O +0.0004 27.00 32.04 4-30 40 60 0.2088 o. 2086 O.OOO2 * 7.773 grm. per liter. t 0.198 normal. Calcium Borate. J 0.0996 normal. Mineral. Thiosul- phate taken. Iodine taken. Time of standing. Volume of solutions. B 2 O 3 found. Per cent. grm. cm. 3 cm. 3 min. cm. 3 grm. 0.4015 0.4010 35-05 35-34 4-75 5-23 60 1 2O 40 45 O. 2280 0.2283 56.92 56.94 Colemanite. o . 4002 32 .00 5-50 90 50 o 2043 51.04 0.2513 32 .01 7.36 60 40 o 1279 50.91 0.4007 33 03 7.72 So 65 o 2036 50.81 These results show little variation, and in the case of cole- manite correspond closely to the theory 50.97 per cent. The process is convenient, generally applicable, and accurate within the ordinary limits of analysis. 214 METHODS IN CHEMICAL ANALYSIS ALUMINIUM. The Determination of Aluminium by Precipitation with Ether- Hydrochloric Acid. Crude aluminium chloride may be freed from every trace of a ferric salt by dissolving it in the least possible amount of water, saturating the cooled solution with gaseous hydrochloric acid, filtering upon asbestos in a filtering crucible or cone, and wash- ing the crystalline precipitate with the strongest hydrochloric acid. Prepared in this way the salt gives no trace of color when dissolved in water and tested with potassium sulphocyanate, but the degree of insolubility is not sufficient for the pur- poses of analysis. Gooch and Havens* have found that in a mixture of hydrochloric acid of highest concentration and ether in equal parts the solubility of aluminium chloride amounts approximately to 5 parts of the hydrous salt, A1C1 3 .6H 2 O, cor- responding to I part of the oxide, A^Oa, in 125,000 parts of the mixture. Pure aqueous hydrochloric acid of full strength mixes per- fectly with its own volume of anhydrous ether, but the addition to this mixture of any very considerable amount of a solution of ferric chloride in strong hydrochloric acid occasions the sepa- ration of a greenish etherial solution of the ferric salt upon the surface of the acid. The addition of more aqueous acid does not change the conditions essentially, but more ether renders the acid and the oily solution completely miscible. For the separa- tion of insoluble aluminium chloride from certain small amounts of soluble ferric chloride, the mixture of the strongest aqueous hydrochloric acid and ether in equal parts serves the purpose excellently; when larger amounts of ferric chloride are to be dissolved, ether must be added proportionately in order to pre- vent the separation of the etherial solution of ferric chloride from the rest of the liquid. separation of The quantitative procedure as finally developed Aluminium by Gooch and Havens for the separation of alumin- ium from iron, and the determination of aluminium as the oxide, is as follows : The concentrated aqueous solution of the salts is mixed with a suitable volume of strongest aqueous * F. A. Gooch and F. S. Havens, Am. Jour. Sci., [4], ii, 416. ALUMINIUM hydrochloric acid enough to make the entire volume approxi- mately 15 cm. 3 to 25 cm. 3 . This mixture is saturated with gase- ous hydrochloric acid while kept cool by immersing in running water the receptacle containing it. A volume of ether equal to the volume of the liquid is introduced, and the cooled etherial mixture is treated with gaseous hydrochloric acid to saturation. The precipitated crystalline chloride is collected upon asbestos in a perforated crucible, washed with a previously prepared mix- ture of hydrochloric acid and ether carefully saturated at 15, dried a half-hour at 150, covered with a layer of pure mercury oxide* (about I grm.), and ignited carefully under a good ven- tilating flue, finally with the blast lamp. The gaseous hydrochloric acid is most conveniently produced in regulated current by treating massive ammonium chloride with strong sulphuric acid in the Kipp generator. A platinum dish hung in an inverted bell jar, provided with inlet and outlet tubes through which the current of water for cooling is passed, makes a convenient container for the solution to be saturated with the gas. The filtration is made upon asbestos in a perfo- rated crucible. The filtrate and washings are caught directly in a crucible (placed under the bell jar of the filter pump) in which the subsequent evaporation is to be effected. The heating of the strongly acid solution must be gradual and conducted with care to prevent mechanical loss by a too violent evolution of the gaseous acid. Results are given in the table. Precipitation of Aluminium Chloride in Presence of Ferric Chloride. A1 2 O 3 taken in solution as the chloride. Al 2 Os found by ignition with HgO. FejOs present as chloride. Final volume. Error. grm. grm. grm. cm. s grm. 0.0761 0.0758 2 5 0.0063 0.0761 0.0754 25 0.0007 0.0761 0.0751 .... 25 o.ooio 0.0761 0.0757 0.15 25-30 0.0004 0.0761 0.0756 0.15 25-30 0.0005 0.0761 0-0755 0.15 25-30 0.0006 0.0761 0.0755 0.15 25-30 0.0006 * Loc. cit., p. 419. 2l6 METHODS IN CHEMICAL ANALYSIS Determination By similar procedure Havens* has effected the sep- of Aluminium aration of aluminium from beryllium. The beryl- Beryllium. jj um ma ^ ^ recovere( j j n ^6 filtrate from the alu- minium chloride by precipitation as hydroxide with ammonia after nearly complete evaporation of the acid, and weighed as oxide after ignition. Or, the filtrate may be evaporated just to dryness on a radiator, care being taken not to heat to the volatilizing point of the beryllium chloride ; a few drops of strong nitric acid added; the liquid evaporated, best with a current of air playing on the surface; and the residue heated gently at first, to break up the nitrate, and finally over the blast lamp. Results of this procedure are given below. Aluminium and Beryllium. Al 2 Oj taken in solution as the A1 2 3 . Error. Final volume. BeO taken in solution as the BeO found. Error. chloride. chloride. grm. grm. grm. cm. 8 grm. grm. grm. 0.1059 0.1058 O.OOOI 12 0.0198 O . 0204 +0.0006* O.IOS3 o. 1044 0.0009 12 0.0194 0.0196 -fo.0002* 0.1065 o. 1059 0.0006 12 0.0197 0.0205 +0.0008* 0.1068 o. 1060 0.0008 12 0.0199 0.0207 +0.0008* 0.1049 o. 1047 O.OOO2 12 0.0198 0.0208 +O.OOIO* o. 1060 , 0.1057 0.0003 12 0.0977 O . 0969 -0.0008* o. 1064 0.1063 O.OOOI 12 0.1085 o. 1084 O.OOOI* o. 1046 o. 1038 0.0008 30 0.1083 o. 1087 +0.0004* 0.1051 o. 1048 0.0003 30 o. 1071 0.1078 +0.0007* 0.1076 0.1075 O.OOOI 30 0.1086 0.1094 +o.ooo8f * By the evaporation process, t By the precipitation process. Determination Havensf effected similarly the separation of alu- of Aluminium minium from zinc, the zinc chloride in the filtrate be- andZmc. -^ conver t ec i to the oxide by evaporation, repeated treatments with nitric acid in porcelain, and ignition. On ac- count of the danger to platinum, the evaporations and treat- ments with nitric acid are made in porcelain. In these processes, however, the porcelain is somewhat attacked, and correction must be made for a slight contamination of the residual oxide. Results are given on the following page. * Franke S. Havens, Am. Jour. Sci. [4], iv, HI. f Franke Stuart Havens, Am. Jour. Sci., [4], vi, 45. ALUMINIUM 217 Aluminium and Zinc. A1 2 0, taken as the A1 2 3 found. Error. ZnO taken. ZnO found. Error. Error corrected.* Final volume. chloride. grm. grm. grm. grm. grm. grm. grm. cm. 3 o 0^62 O OOOO O IIIO 0.0580 0.0577 0.0003 0.1034 0.0572 0.0572 O . OOOO o. 1014 0.1027 +0.0013 0.0007 12 0.0563 0.0550 -0.0013 0.1026 o. 1038 +0.0012 0.0008 16 0.0577 0.0576 O.OOOI O.IOOO O.IOI4 +O.OOI4 O.OOO6 16 o-559 0.0558 O.OOOI O.IO2O 0.1035 + O.OOI5 0.0005 16 0.0563 0.0556 0.0007 0.2024 o . 2046 +O.OO22 +O . OOO2 20 O.IIII o. 1107 0.0004 0.2092 0.2116 +O.OO24 +0 . 0004 20 * Corrected by the amount of material found in the evaporation of similar amounts of the strong acids in porcelain. Determination The process applies also to the separation of alu- of Aluminium minium from copper, as shown by the results given.* and copper. The copper in the filtrates of these determinations was converted to the oxide through the sulphate because this operation may be conducted safely in platinum. Aluminium and Copper. A1 2 O 3 taken as chloride. grm. A1 2 O 3 found, grm. Error, grm. CuO taken, grm. CuO found, grm. Error, grm. 0.0437 . 043 2 O 0005 0.0359 O.O35Q O OOOO O O34on CO ntent of the organic matter may be estimated by absorbing in a measured amount of standardized barium hydroxide the carbon dioxide produced, acting upon the residual hydroxide with standard iodine in excess, and deter- mining by decinormal arsenite the excess of iodine. According to this process, the organic material is weighed out and intro- duced into an evolution flask with 10 cm. 3 to 15 cm. 3 of pure water (freed from carbon dioxide by boiling down one-third and kept in stoppered bottles), and the evolution flask is connected with an absorption flask charged with a volume of standardized barium hydroxide 3 cm. 3 to 5 cm. 3 in excess of the amount re- quired to precipitate the carbon dioxide to be determined.! The whole system is then evacuated with the water pump to a pressure of 200 mm. to 225 mm. and the boiling flask warmed. * I. K. Phelps, Am. Jour. Sci., [4], iv, 372. t See, for the general arrangement, the more elaborate apparatus of Fig. 19 .page 237. CARBON 235 An excess of potassium permanganate solution (prepared by dis- solving in water, acidulating with sulphuric acid and boiling until free from carbon dioxide) is then run in through the funnel tube and the mixture warmed again. The carbon dioxide is set free by sulphuric acid [i 13], either at once or after introducing an excess of pure sodium hydroxide and boiling, and driven completely to the absorption flask by boiling for five minutes. During the passage of the gas the absorption flask is shaken frequently and kept cool by standing in a dish of water and by pouring cold water over it from time to time. During the boil- ing, the vacuum in the flasks may be tested by opening momen- tarily the stopcock of the funnel tube and noting the direction of the flow of water contained in the funnel. After the boiling, the atmospheric pressure is restored by allowing air, purified from carbon dioxide by passage through potash bulbs, to enter through the funnel tube of the boiling flask. The flasks are disconnected, the stopper of the absorption flask with its attach- ments is removed and carefully washed free from barium hydrox- ide, and a second stopper, provided with a separating funnel and a Will and Varrentrapp absorption apparatus containing water to serve as a trap, is inserted into the mouth of the absorp- tion flask. The contents of the flask are brought to the boiling point, decinormal iodine solution is run in through the funnel tube in sufficient quantity to destroy the larger part of the excess of barium hydroxide, the mixture again heated to boiling, and iodine run in again to permanent red coloration. After cooling, the excess of iodine is determined by titration with decinormal arsenite. The difference between the barium hydroxide taken and that equivalent to the iodine which has disappeared is the Oxidation with Permanganate in Acid Solution. Ammonium oxalate taken, grm. BaO,H, taken, grm. BaO 2 H, remaining. grm. CO 2 found, grin. CO, calculated. grm. Error on COj. grrn. 0.2522 0.7267 o . i i 70 0.1565 0.1561 +o . 0004 0.2542 0.7267 O.III3 0.1579 0.1574 +o . 0005 0.5020 i -4535 0.2417 0.3110 0.3108 +O.OOO2 0.5058 1-3954 0.1753 0.3131 0.3131 0.0000 1.0033 2.6163 0.1955 0.6213 0.6211 +0.0002 1.0003 2.5951 0.1836 0.6189 0.6192 -0.0003 I.OOIO 2.6163 0.2037 0.6192 0.6197 0.0005 236 METHODS IN CHEMICAL ANALYSIS Oxidation with Permanganate in Alkaline Solution. Barium formate taken. BaO 2 H 2 taken. BaO 2 H 2 remaining. CO 2 found. CO, calculated. Error on COj. grm. grm grm grm. grm. grm. 0.5001 0.9302 0.1745 0.1939 0.1935 +O . 0004 0.5033 0.9012 o. 1402 0.1953 0.1947 +0.0006 I . OOO2 1.6861 0.1793 0.3867 0.3870 0.0003 I . 0059 1.6279 0.1093 0.3897 0.3892 +O . 0005 1-3750 2.2529 o. 1820 0.5315 0.5320 0.0005 1.5028 2.4419 0.1754 0.5816 0.5814 + 0.0002 Tartar emetic taken.* grm. BaO 2 H 2 taken, grm. BaO,H 2 remaining. grm. CO 2 found, grm. C0 2 calculated. grm. Error on COj. grm. 0.5051 i . 2450 0.1709 0.2756 0.2751 +0.0005 0.5030 i .2226 O.IS36 0.2743 0.2739 +0 . 0004 0.7509 1-7355 o. 1401 0.4094 0.4091 +0.0003 0.7541 i 7430 o. 1410 0.4111 0.4107 +0.0004 i .0018 2.3456 0.2187 0.5458 0.5456 + O.OOO2 1.0005 2.2435 o. 1196 0.5451 0.5450 +O.OOOI * Dried at 100. measure of the carbon dioxide. The figures given show the result of the permanganate oxidation when applied to ammo- nium oxalate in acid solution, and to barium formate and tartar emetic in alkaline solution. Obviously, certain organic substances oxidizable by perman- ganate may be analyzed by the process outlined above, but, as noted by Wanklyn and Cooper* and by others, potassium permanganate, whether in acid or alkaline solution, fails to oxi- dize completely certain organic substances (acetates and carbo- hydrates, for example) even at the boiling temperature. Carbon Content It is well known, however, that a mixture of con- whhCnromix: centrated sulphuric and chromic acids has a much Acid. wider field of action in oxidizing organic compounds, and Phelpsf has studied the application of this method also. The apparatus recommended for the process, and shown in the ac- companying figure, may be described as follows: A thick-walled liter flask with round bottom, serving as an oxidizing chamber, is closed by a rubber stopper with two per- * Phil. Mag., [5], vii, 138. t Loc. cit. CARBON 237 1-ig. n forations. Through one of these passes the tube of a separating funnel reaching nearly to the bottom of the flask and drawn out at the lower end . A disk of platinum foil, through which the tube of the funnel passes, is hung in the neck of the flask, nearly closing it, and held in place by an attached platinum wire, the end of which is squeezed under the rubber stopper. Through the second hole of the stopper passes an exit tube 0.7 cm. 3 in bore. This tube, expanded just above the stop- per to a small bulb which serves to prevent mechanical loss of the solid contents of the flask during the boil- ing, is joined by means of a rubber connector (provided with a screw pinchcock) to the inlet tube of the absorption flask, which is an ordinary 500 cm. 3 round -bottomed flask. This flask is also closed by a rubber stopper with two perforations, through one of which passes the inlet tube and through the other the exit tube, which is also enlarged to a small bulb just above the stop- per and is closed by a rubber connector and screw pinchcock. The ground-glass stopper of the funnel tube is carefully cleaned and lubricated with a thick solution of metaphosphoric acid. A partial vacuum is easily obtained by boiling water in the evolution flask and the barium hydroxide solution in the absorp- tion flask at the same time, closing the flasks, and cooling both flasks being connected and ready for a determination. In making a determination, the organic substance is weighed out in a counterbalanced bulb, so thin that it may be easily broken later and made with a wide mouth for convenience in introducing the solid substance. After the substance is weighed, the mouth of the bulb is sealed by heating in a small blowpipe flame and the tube introduced into the evolution flask, together with an amount of pure potassium dichromate, known to be in excess of that required to oxidize the organic substance. The flasks are connected, as already described, with an appropriate amount of barium hydroxide solution in the absorption flask and 10 cm. 3 of pure water in the evolution flask, and the vacuum 238 METHODS IN CHEMICAL ANALYSIS obtained (as described above) by boiling the contents of both flasks until the water in the evolution flask has decreased to 2 cm. 3 or 3 cm. 3 , and cooling. The tube containing the organic sub- stance is then broken by shaking the flask, and 20 cm. 3 of concen- trated sulphuric acid, previously purified from organic material by heating to the fuming point with a few crystals of potas- sium dichromate, are run in through the funnel tube. While still hot, the acid is shaken in the flask violently, the platinum foil hung in the neck serving to protect the rubber stopper. The flask is warmed to approximately 105, the highest temperature to which, as shown by Cross and Bevan,* such a mixture of chromic and sulphuric acids may be safely heated without the disengagement of oxygen gas. Water is then run in until the crystals of chromic anhydride have disappeared and the danger of the evolution of oxygen is past. The solution is heated to its boiling point, with care to keep the outward pressure less than the inward pressure the relation being easily observed by opening momentarily the stopcock of the funnel tube and noting the direction of the flow of water contained in the funnel. The flask is shaken and heated alternately for five minutes to bring about the oxidation of small amounts of carbon monoxide, originally produced. Then more water (60 cm. 3 to 70 cm. 3 ) is introduced through the funnel and the stopcock between the Oxidation with Chromic Acid. Substance taken. Ba0 2 H 2 taken. Ba0 2 H 2 found. CO, found. C0 2 calculated. Error on C0 2 . grm. grm. grm. grm. grm. grm. Analysis of ammonium oxalate. 0.5009 1-3534 o. 1469 0.3097 0.3101 0.0004 0.5006 1.3400 0.1308 0.3103 0.3099 +o . 0004 0.5005 1.3400 0-1343 0.3094 0.3098 0.0004 I .0002 2 . 5460 0.1347 0.6188 0.6192 0.0004 I .0010 2.5192 ' o. 1094 0.6185 0.6197 0.0012 Analysis of cane sugar. O. 2OOI 1.3926 0.1905 0.3085 O^OSS 0.0003 0.2000 1.3926 0.1936 0.3077 0.3086 0.0009 O.2OOI 1.3926 0.1857 0.3097 0.3088 + 0.0009 O.2OI4 1.3400 0.1279 0.3III 0.3108 +0.0003 * Jour. Chem. Soc., liii, 889. CARBON 239 boiling and absorption flasks is opened to admit the carbon dioxide to the latter, which is kept cool and shaken as before. The contents of the evolution flask are then heated to boiling and a slow current of air, freed from carbon dioxide by passage through potash bulbs, is allowed .to enter through the funnel tube to keep the liquid from undue bumping. The boiling is con- tinued for fifteen minutes, after which the excess of barium hy- droxide is determined iodometrically and the carbon dioxide calculated. The results of experiments with ammonium oxalate and with cane sugar, one of the more difficultly oxidizable substances, are given in the preceding table. The results are evidently very satisfactory. carbon Dioxide Taking advantage of the fact that at 105 the Evolved and mixture of chromic and sulphuric acids does not oxygen Used. evo i ve OX ygen, Phelps* has been able to determine the oxygen used in the combustion of certain organic substances. The knowledge of this amount of oxygen and of that contained in the products of oxidation gives, by difference, the oxygen content of the original substance. In this operation a known weight of the substance is first treated with a known weight of pure potassium dichromate and 20 cm. 3 of concentrated and purified sulphuric acid, according to the process described above for the determination of carbon as carbon dioxide. The residual liquid, which, after dilution and boiling for the removal and estimation of the carbon dioxide, should have a volume of 60 cm. 3 to 80 cm. 3 , is washed into the Voit flask of the distillation and absorption apparatus figured and previously described.! The absorption chamber is charged with a solution of arsenious oxide in sodium hydroxide, the for- mer being in slight excess of the amount required to take up the chlorine to be evolved by the chromate and the latter in quantity more than sufficient to neutralize the acid which may be volatil- ized to the receiver in the later operation. The apparatus is connected, hydrochloric acid (35 cm. 3 of the strongest acid) introduced through the stoppered funnel, a slow current of carbon dioxide started through the system, and the liquid in the flask slowly boiled down, for a period of five or six hours, * Am. Jour. Sci., [4], ii, 379. t See Fig. 3, page 4. 240 METHODS IN CHEMICAL ANALYSIS until the volume is 30 cm. 3 to 40 cm. 3 . After cooling and dis- connecting the apparatus, the solution in the receiver is made acid with sulphuric acid and then alkaline with acid potassium carbonate. The residual arsenite is determined by titration with decinormal iodine. Determination of Carbon Dioxide Evolved and Oxygen Used. Oxygen Substance taken. C0 2 found. Error on C0 2 . K 2 Cr 2 7 taken. As 2 O 3 taken. As 2 3 found. Oxygen used. required by Error on oxygen. theory. grm.. grm. grm. grm. grm. grm. grm. grm. grm. Analysis of ammonium oxalate. I.OI22 I .OOI9 0.6265 0.6212 O.OOOI +0.0010 2.0009 2.O002 1.3002 I-3SI7 o . oooo o . 0440 0.1160 0.1147 0.1139 0.1128 +O.OO2I +O.OOI9 Analysis of phthalic acid. O.IOO2 0.1093 0.2138 0.2324 +O.OOI4 +O.OO07 2.OOI2 2. OOOO 1.2004 I.I03I 0.0814 0.0634 0.1456 0.1582 0.1448 0.1580 +0.0008 +O.OOO2 Analysis of cane sugar. 0.2025 0.4012 0.3117 0.6166 0.0008 0.0024 3 . oooo 5.0000 i . 7002 2.3022 0.0796 0.0366 0.2275 0.4495 0.2273 0.4502 +0 . 0002 0.0007 Analysis of paper. 0.3034 0.4523 0.4932 0.7334 O.OOIO -0.0033 3.5015 5-0035 I .4017 I . 8000 0.0879 0.0710 0.3539 0.5368 0.3598 0.5358 0.0005 +O.OOIO Analysis of tartar emetic. 0.5057 1.0099 0.2671 0.5321 0.0009 0.0030 2.5018 3-5003 I .7000 1.7520 0.0766 0.0198 0.1459 0.2911 0.1462 0.2919 -0.0003 0.0008 Analysis of barium formate. 1.0079 1.5014 0.3906 0.5814 +0.0006 +0.0005 3.0026 3.0010 2 . 2002 I . 8080 o . 0496 o . 0890 0.1423 0.2118 0.1422 0.2118 +O.OOOI o . oooo When the boiling begins the chromate is gradually reduced with evolution of chlorine, but if the evaporation of water is pushed too rapidly the sulphuric acid may reach a degree of con- SILICON 241 centration such that oxygen is evolved instead of chlorine in the process of reduction. Sometimes, during the reduction, chloro- chromic anhydride is visibly volatilized to the receiver, but inas- much as it is there reduced and registered by the arsenite this transfer is of no moment. The results of experiments in which the carbon of various organic substances was first determined by oxidation according to the process previously described,* the amount of potassium dichromate employed being accurately known, and the residual dichromate found by reduction with hydrochloric acid by the process just outlined, are given in the tabular statement. From these results, it will be seen that the process works with accuracy for a great variety of organic substances. It was found impossible, however, to determine the elements in compounds which, like ether and naphthalene, are at the same time volatile and hard to oxidize completely. SILICON. The Detection of Silicon in Silicates and Fluo silicates. The formation of silicon fluoride by the action of hydrofluoric acid or a fluoride and sulphuric acid upon a silicate is often applied to the detection of silica, the silicon fluoride giving with water a white precipitate of silicic acid. The usual proce- dure in making this test is to allow the gaseous silicon fluoride to come in contact with a moistened glass rod, but the condensa- tion of steam or sulphuric acid on the rod often makes the results uncertain. Browningf has found that when moistened black paper is brought in contact with the fumes of the gaseous fluoride the deposit of silica is very easily detected. According to the procedure recommended, the reaction is made to take place in a small lead cup about one centimeter in diameter and depth, made by running the melted metal into a mold, covered by a flat piece of lead with a small hole in the center, as shown in Fig. 20. Into Flg - 20> this cup is put the fluosilicate, or the silicate with a small amount of finely powdered calcium fluoride, generally about o.i grm., and a few drops of concentrated sulphuric acid are added. Upon the * See page 236. t Philip E. Browning, Am. Jour. Sci., [4], xxxii, 249. 242 METHODS IN CHEMICAL ANALYSIS upper side of the cover a piece of moistened black filter paper is placed and upon this a small moistened pad of ordinary filter paper to keep the black paper moist during heating upon a steam bath. At the conclusion of the heating of about ten minutes, a white deposit is found on the under side of the black paper, over the opening in the cover, if silica is present in ap- preciable amount. Tests of this procedure are given in the table. Silicate Tests. Material tested, gnu . Approximate per cent of SiO 2 . CaF 2 present, grin. Result. o 1000 SiOa IOO O.IOOO Nothing. o 1000 SiO2 IOO O IOOO Very good. o oioo SiO2 IOO O. IOOO Very good. o 0050 SiO2 IOO O. IOOO Very good. o ooio SiO2 IOO O.IOOO Trace. o.oioo Kaolinite 0.0050 Kaolinite o ooio Kaolinite 46 46 46 O. IOOO O. IOOO O. IOOO Very good. Very good. Trace. o oioo Gadolinite 24 O. IOOO Very good. o 0050 Gadolinite 24 O. IOOO Trace. o.oioo Lepidolite 50 O. 2OOO Good. Fluosilicate Tests. o 0050 Na2SiFe Very good. o ooio Na2SiFg Very good. TITANIUM. The Determination of Titanic Acid by Reduction and Titration with Potassium Permanganate. The estimation of titanic acid by reduction with zinc and direct titration leads to low results even when precautions are taken to avoid atmospheric oxidation during titration.* New- tonf has shown, however, that it is possible to determine titanic acid successfully by reducing with zinc in an atmosphere of hydrogen, adding an excess of ferric sulphate, and titrating the resulting and equivalent ferrous salt with potassium perman- * Cf. Pisani, Compt. rend., lix, 298; Marignac, Zeit. anal. Chem., vii, 112; Wells and Mitchell, Jour. Am. Chem. Soc., xvii, 878. t H. D. Newton, Am. Jour. Sci., [4], xxv, 130. TITANIUM 243 ganate; the reaction between the salts of titanium and iron taking place according to the equation: Ti 2 (SO 4 ) 3 + Fe 2 (SO 4 ) 3 = 2 Ti(SO 4 ) 2 + 2 FeSO 4 . According to this procedure, titanic acid dissolved in con- centrated sulphuric acid is introduced into a ioo-cm. 3 flask and the solution is diluted with water until it contains 10 per cent of sulphuric acid, this strength being sufficient to hold the titanic acid in solution, while insufficient to reoxidize reduced titanium oxide. Zinc is added in suitable amount, and a rubber stopper carrying a delivery tube and a small separating funnel is inserted in the neck of the flask. After the air has been driven from the flask by the hydrogen evolved, the delivery tube is dipped under water and the stopcock of the funnel is closed. Gentle heat is applied until all the zinc is dissolved, and the solution is cooled. An excess of ferric sulphate is passed into the flask through the separating funnel and followed at once by cold, freshly distilled water until the flask is filled to the neck. The contents of the flask are poured into more cold distilled water and the ferrous salt produced by action of the ferric salt upon the reduced titanium oxide is titrated by n/io permanganate. Test results are given in the table. Permanganate Titration of Titanic Acid Reduced by Zinc. KMn0 4 . cm. 3 TiO 2 taken, grm. TiO 2 found, grm. Error, grm. 6.50 0.0520 0.0523 +o . 0003 6.52 0.0520 0.0524 +O.0004 6-45 0.0520 0.0519 o.oooi 6.50 0.0520 0.0523 +o . 0003 6.48 0.0520 0.0521 +O.OOOI 6.42 0.0520 0.0518 O.OOO2 6.50 0.0520 0.0523 +0.0003 19-95 0.1596 0.1599 +0.0003 19-93 0.1596 0.1598 +O.OOO2 19-95 o. 1596 0-1599 +0.0003 19.90 o. 1596 0.1595 o.oooi 19-95 o. 1596 0-1599 +0.0003 19-95 o. 1596 0.1599 +0.0003 19.90 0.1596 0-1595 O.OOOI 19.85 0.1596 0.1591 -0.0005 19-95 0.1596 0.1599 +0.0003 19.88 0.1596 0.1594 O.OOO2 19-95 0.1596 0.1599 +0.0003 19-95 0.1596 0.1599 +0.0003 244 METHODS IN CHEMICAL ANALYSIS ZIRCONIUM. The Separation of Zirconium from Iron by Volatilization of the Latter in Hydrogen Chloride. It has been shown by Havens and Way* that zirconium oxide may be separated from ferric oxide by the volatilization of ferric chloride in an atmosphere of hydrogen chloride, containing a little chlorine, at a temperature of 200 to 300. f CERIUM. The Separation of Cerium from Other Cerium Earths by the Action of Bromine upon the Mixed Hydroxides in Presence of an Alkali Hydroxide. One of the best known processes for the separation of cerium from lanthanum and didymium is that of Mosander.J This process consists in passing chlorine gas jnto a mixture of the hydroxides suspended in a distinct excess of a fixed alkali hy- droxide, until the solution is saturated and the reaction of the liquid is no longer alkaline to litmus. Under these conditions nearly all the cerium remains undissolved as the eerie hydroxide, while the other cerium earths go largely into solution. In treat- ing mixed material, the residue of eerie hydroxide generally retains some of the cerium earths, so that the treatment with chlorine must be repeated. Two disadvantages associated with this method, therefore, are the preparation and use of chlorine gas, and the solvent action of the hydrochloric acid formed in the reaction upon the eerie hydroxide 2 Ce(OH) 3 + C1 2 = 2 CeO 2 + 2 HC1 + 2 H 2 O. Browning and Roberts have shown that by substituting bro- mine for chlorine in the Mosander process about 50 per cent of the other cerium earths can be separated from eerie hydroxide in one treatment, and that after three treatments practically all the other cerium earths are removed without any solvent action * Franke Stuart Havens and Arthur Fitch Way, Am. Jour. Sci., [4], viii, 217. t See page 508. { Jour, prakt. Chem., xxx, 267. Philip E. Browning and Edwin J. Roberts, Am. Jour. Sci, [4], xxix, 45. CERIUM 245 upon the eerie hydroxide. The advantages of the method are, the convenience in the use of the bromine, and the apparent lack of tendency of the hydrobromic acid to dissolve the eerie hy- droxide. The procedure is as follows : The mixed hydroxides are precip- itated with a slight excess of sodium hydroxide or potassium hy- droxide, and, suspended in the alkaline solution, are treated with liquid bromine or bromine water in distinct excess, and the mixture is placed upon a steam bath until the greater part of the free bromine is expelled. The residue is then filtered off, washed, and treated as before. This process is repeated. In the test experiments upon a mixture composed of about 50 per cent of cerium oxide and 50 per cent of cerium earth oxides other than cerium oxide, the filtrate after each treatment was found to contain the amounts of cerium earth oxides, free from cerium, indicated in the table. The residue from the last treat- ment on being dissolved in acid showed only faint didymium bands. In another experiment a larger amount of material, 10 grm., was subjected to a fourth and fifth treatment with bromine, the fourth treatment yielding a small fraction of a gram of oxides other than cerium oxide, and the fifth only a few milligrams. In both cases these oxides were free from cerium. The oxides from the first filtrates were much lighter in color than those obtained from the last, which, of course, indicates that the lanthanum is dissolved by the action of the bromine more readily than the didymium. The quantitative results follow in the table. Separations by Bromine and Alkali Hydroxide. Mixed oxides taken. Oxides found in first filtrate. Oxides found in second filtrate. Oxides found in third filtrate. Total oxides found. grm. grm. grm. grm. grm. I .OOOO 0.3310 0.0720 0.0190 0.4420 I .OOOO I. 0000 I .OOOO IO.OOOO IO.OOOO 0.2900 0.2250 0.2750 3.1360 3-4590 O. IOIO o . i 290 0.0860 1.0050 0.5240 o . 0420 o . 0640 o . 0740 0.5930 0.8560 0.4330 0.4180 0.4350 4-7340 4.8390 The action of iodine is similar to that of chlorine and bromine, but it is too incomplete to be of practical value. 246 METHODS IN CHEMICAL ANALYSIS The lodometric Estimation of Cerium. Anhydrous cerium dioxide, prepared by the ignition of the oxalate or hydroxide, is very slowly acted on by acids, especially when pure. For this reason the method which Bunsen de- scribed has remained the only one adapted to the satisfactory volumetric estimation of the ignited dioxide.* According to this method, the substance to be determined is weighed out in a glass flask of 10 to 15 cubic centimeters' capacity, a few crys- tals of potassium iodide are added, and the neck of the flask is drawn out by the aid of a blowpipe to a narrow opening. The flask is filled almost to the narrowing of the neck with hydro- chloric acid which is free from chlorine or iron chloride, and a little sodium carbonate is added in order to displace the last trace of air by carbon dioxide. The flask is then closed by sealing off the neck in the blowpipe and warmed in a water bath until the cerium compound is completely dissolved, and the quantity of iodine set free is determined by iodometric analysis. Browning (with Hanford and Hall)f has shown that good results may be obtained by digesting cerium dioxide with potas- sium iodide and hydrochloric acid in a glass-stoppered bottle and determining by sodium thiosulphate the iodine set free, or by distilling and estimating the iodine passing to the receiver , in accordance with the reaction: 8HCl+2KI = 2 CeCl 3 + 2 KC1 + 4 H 2 O + I 2 . Digestion According to the first procedure, weighed portions Process. o f t h e p ure ce rium dioxide are placed in small glass- stoppered bottles of about 100 cm. 3 capacity, together with I grm. of potassium iodide free from iodate, and a few drops of water to dissolve the iodide. A current of carbon dioxide is passed into the bottle for about five minutes to expel the air, 10 cm. 3 of pure strong hydrochloric acid are added, the stopper is inserted and the bottle heated gently, upon a steam radiator, for about one hour, until the dioxide dissolves completely and the iodine is set free. After cooling the bottle, to prevent loss 'of iodine upon removing the stopper, the contents are carefully washed into about 400 cm. 3 of water and titrated with standard * Ann. Chem., cv, 49. t Philip E. Browning, with G. A. Hanford and F. J. Hall, Am. Jour. Sci., [4], viii, 452. CERIUM 2 47 sodium thiosulphate to determine the amount of iodine liberated. Results of this procedure, corrected by 0.04 cm. 3 of n/io iodine, which is the amount of iodine set free in blank determinations, are given in the accompanying table. Digestion Process. CeO 2 taken, grm. CeO 2 found. grin. Error, grm. O.IOOO 0.0994 0.0006 O. 1032 0.1034 +O.OOO2 o. 1016 o. 1017 +O.OOOI 0.1054 o. 1041 0.0013 O.2OIO O.2O2I +O.OOII o. 1104 O.IIO9 -j-o.0005 o. 1914 0.1907 0.0007 o. 1604 0.1603 O.OOOI 0.2146 0.2145 O.OOOI o. 1108 0.1099 0.0009 0.1346 0-1347 +O.OOOI 0.1540 0-1534 0.0006 0.1976 0.1968 0.0008 0.1230 o . i 240 +0.0010 0.1199 O. I2OI +0.0003 0.1524 0.1528 +0.0004 O. 1212 O.I2II O.OOOI 0.1528 0.1543 +0.0015 Distillation According to a second procedure, portions of ce- Process. rium dioxide were weighed out into the retort of a distillation apparatus* consisting of a Voit flask, serving as the retort, sealed to the inlet tube of a Drexel wash bottle, used as a receiver, the outlet tube of which was trapped by sealing on Will and Varrentrapp absorption bulbs. In the retort are placed the cerium dioxide, 15 cm. 3 of water, I grm. of potassium iodide, and 10 cm. 3 of pure strong hydrochloric acid. In the receiver are 100 cm. 3 of water and 2 grm. to 3 grm. of potassium iodide, and in the bulbs a dilute solution of potassium iodide. Before adding the hydrochloric acid a current of carbon dioxide is passed through the apparatus for some minutes. After adding the acid, the liquid is boiled in the current of carbon dioxide to a volume of 15 cm. 3 , and when the free iodine has almost com- pletely left the retort and passed into the receiver the apparatus is allowed to cool. The iodine in the receiver is titrated directly * See Fig. 3, page 4. 248 METHODS IN CHEMICAL ANALYSIS with sodium thiosulphate, and that in the retort after dilution of the residue to about 400 crn. 3 , the amount in the retort seldom exceeding the equivalent of a few drops of n/io iodine solution. Results are given in the table. Distillation Process. CeO 2 taken, grm. CeO 2 found, grm. Error, grm. o. 1028 O.2O6O 0.1013 0.2055 0.0015 0.0005 o. 2014 o. 1716 0.2012 o. 1711 0.0002 0.0005 0.0974 0.1600 0.1268 0.1276 o. 1620 0.0972 0.1587 0.1254 0.1268 o. 1612 O.O002 -0.0013 O.OOI4 O.OOOS O.OOOS o. 1016 0.1548 O. IOII 0.1543 O.OOO5 0.0005 0.1352 0.1342 o.ooio The Estimation of Cerium Oxalate by Potassium Permanganate. Stolba * has stated that cerium oxalate may be estimated volu- metrically in the same manner as calcium oxalate by treating the washed precipitate, suspended in warm water, to which a mod- erate amount of sulphuric acid has been added, with potassium permanganate; that as the titration proceeds the precipitate disappears and the end reaction is sharp; and that the perman- ganate does not oxidize the cerium from the lower to the higher condition. Browning and Lynch f have presented experimental proof of the correctness of this statement, and have shown that the cerium may be determined either by titration of the precipitated cerium oxalate or by titration of the excess of ammonium oxalate left over in the precipitation. To the cerium salt in solution in 100 cm. 3 to 200 cm. 3 of water is added a definite amount of a standardized solution of ammo- nium oxalate in excess; the whole is warmed to induce a crys- * Sitzungsber. d. kgl. bohm. Gesellsch. d. Wissenschaften v. 4 Juli, 1879; Zeit. anal. Chem., xix, 194. t Philip E. Browning and Leo A. Lynch, Am. Jour. Sci., [4], viii, 457. CERIUM 249 talline condition of. the precipitate. The precipitate is filtered off on paper, carefully washed, and dissolved by passing 10 cm. 3 of hot [1:3] sulphuric acid repeatedly through the filter. The filtrate and washings are made up to about 500 cm. 3 and warmed to 70 or 80, and the oxalic acid in solution is titrated with per- manganate. The filtrate from the cerium oxalate, containing the excess of ammonium oxalate, is diluted to 500 cm. 3 , acidified with 10 cm. 3 of dilute [i : 3] sulphuric acid, I grm. of manganous sulphate is added to insure regularity of action in presence of hydrochloric acid, and the oxalic acid is titrated by perman- ganate. Results obtained in this manner are given in the table. Permanganate Titration of Precipitate and of Excess of Precipitant. Amount taken, calculated as CeCl 3 . grm. Treatment of precipitate. Treatment of filtrate. Amount found, calculated as CeCl 3 . grm. Error, calculated as CeCl 3 . grm. Amount found, calculated as CeCl 3 . grm. Error, calcu- lated as CeCl 3 . grm. Precipitation in neutral solution. o. 1091 o. 1103 -{-O.OOI2 0.1091 0.1087 O.OOO4 O.IO87 0.0004 0.1364 0.1373 +o . 0009 O.I39I +O.OO27 0.1364 0.1367 +0.0003 0.1367 +0.0003 0.2182 O.22O2 +O . OO2O 0.2206 +O.OO24 Precipitation in acid solution. 0.1091 o 1087 o 0004 0.1519 0.1364 0.2182 0.1535 0.1367 0.2183 +0.0016 +0.0003 +O.OOOI 0.1535 0.1367 0.2183 +0.0016 +o . 0003 +O.OOOO The Estimation of Cerium in the Presence of Other Rare Earths by the Action of Potassium Ferricyanide in Alkaline Solution and Potassium Permanganate in Acid Solution. Browning and Palmer* have shown that the oxidation of ce- rium from the cerous to the eerie condition may be effected by potassium ferricyanide in alkaline solution, registered in the * Philip E. Browning and Howard E. Palmer, Am. Jour. Sci., [4], xxvi, 83. 25 METHODS IN CHEMICAL ANALYSIS amount of potassium ferrocyanide formed according to the follow- ing equation, 2 K 3 FeC 6 N 6 + Ce 2 3 + 2 KOH = 2 K4FeC 6 N 6 + H 2 O -f 2 CeO 2 , and subsequently determined by titration with potassium per- manganate in acid solution. The procedure is as follows : To the cerous sulphate in solution 20 cm. 3 of a solution containing 20 grm. of potassium ferricyanide to the liter are added, and potassium hydroxide to complete precipitation. The precipitated hydroxide is filtered off, and the filtrate and washings, amounting in volume to from 200 cm. 3 to 250 cm. 3 , after being made distinctly acid with dilute sul- phuric acid, are titrated with a standard solution of potassium permanganate until the presence of the permanganate color shows the oxidation of all the ferrocyanide to ferricyanide* according to the equation : 5 K4FeC 6 N 6 + KMnO 4 + 4 H 2 SO 4 = 5 KsFeCeNe + 3 K 2 S0 4 -f- MnSO 4 + 4 H 2 O. From this equation and the preceding one the amount of cerium present may be readily calculated. Each day before the ferricyanide is used a portion of 20 cm. 3 of the solution is acidified and titrated with the permanganate, and the correction indicated, generally from one to three drops, is subtracted from the amount of the permanganate used in actual determinations. Oxidation by Ferricyanide and Titration of the Reduced Ferricyanide. Ce taken, calcu- lated as Ce 2 O 3 . grm. Ce found, calcu- lated as Ce 2 O s . grm. Error, grm. 0.1834 0.1819 0.0015 0.1376 0.1380 +0.0004 0.1834 o. 1829 0.0005 0.1834 o. 1829 0.0005 0.1834 0.1834 o.oooo 0.1376 0.1385 +o . 0009 0.1376 0.1371 -0.0005 0.1376 0-1374 O.OO02 0.1376 0.1380 +0.0004 0.1834 0.1824 o.ooio 0.1326 0.1335 +0.0009 0.1326 0.1328 +O.OO02 * Button's Vol. Anal. Qth ed., page 209. TIN 25* Oxidation by Ferricyanide and Titration of the Reduced Ferricyanide. Ce taken, calcu- lated as Ce 2 O 3 . grm. Ce found, calcu- lated as Ce 2 O 3 . grm. Error, grm. Other rare earths present, calculated as oxides. grm. 0.1328 0.1335 +0.0007 o. ThO 2 . 0.1327 0.1322 0.0005 o. ThO 2 . O.O266 0.0275 +0 . OOOQ o. ThO 2 . 0.0267 'O.O272 -j-o . 0005 o. ThO 2 . 0.1324 0.1326 -j-O . OOO2 o. Y 2 3 . 0.1326 0.1323 0.0003 o. Y 2 O 3 . O.O266 0.0264 O.OOO2 o. Y 2 3 . 0.0264 O.O27I +0.0005 o.i Y 2 3 . 0.1376 0.1370 0.0006 0.15 La 2 O 3 +Di 2 Os. O.IIOI o. 1091 O.OOIO 0.15 La 2 3 +Di 2 O 3 . 0.1324 0.1332 +0.0008 0.03 ZrO 2 . All the various operations in this process are carried on with- out warming the solution. The filtrations and washings are made under gentle pressure, and require on an average not more than fifteen to thirty minutes. In the preceding tables are results obtained with cerium alone and in presence of sails of other rare earths. This method presents no difficulties in manipulation and is especially adapted to the rapid estimation of cerium in rare earth mixtures. TIN. The Electrolytic Determination of Tin. From a solution of stannous ammonium chloride in a satu- rated solution of ammonium oxalate,* Medway has precipitated the tin successfully upon the rotating crucible. f Deposition of Tin on the Rotating Cathode. Tin taken. Tin found. Error. Current. Time. N. D. 100 . grm. grm. grm. Amp. mm. O . 0804 o . 0802 O.OOO2 2-5 8-3 20 o . 0804 o . 0800 O.OOO4 2 6.6 20 o. 1607 o. 1610 +0.0003 2-5 8-3 2O 0. 1607 0.1603 0.0004 2-5 8-3 2O 0.1607 o. 1607 o.oooo 3-5 u. 6 15 * H. E. Medway, Am. Jour. Sci., [4], xviii, 56. f See Fig. 13, page 12. 252 METHODS IN CHEMICAL ANALYSIS LEAD. The Detection of Lead. It has been shown by Browning and Blumenthal* that lead may be separated as the sulphate, in association with the alkali earth sulphates, and tested for in the solution obtained, by treat- ment of these insoluble sulphates with ammonium acetate, f The Electrolytic Determination of Lead as the Dioxide. In depositing lead dioxide electrolytically, solutions contain- ing nitric acid are employed; precautions must be taken in regard to concentration of acid, strength of current and tem- perature; and the liquid is siphoned off before interruption of the current. t With the rotating cathode making 600 revolutions a minute and a sand-blasted platinum dish for the anode, Exner obtained in ten to fifteen minutes adherent deposits with a current N. D.ioo = 10 amp. and 4.5 volts acting upon 125 cm. 3 of solution containing 20 cm. 3 of concentrated nitric acid. To obviate the necessity of large and expensive apparatus of platinum, Gooch and Beyer || have experimented with the filter- ing crucible used as a cathode in the devices previously described. *[f In preliminary trials of electrolysis in the closed cell with sub- sequent filtration** it was found that when the concentration of nitric acid amounted to 10 cm. 3 in 60 cm. 3 of liquid, with a current of 4 amperes (N. D.ioo = 10 amp.) and 6 volts, two sources of error appeared. In the first place, the deposition of metallic lead upon the cathode was often noticeable; and secondly, it appeared to be impossible to make the precipitation of lead dioxide com- plete so long as that substance was allowed to float in the liquid. Similar results were obtained in experiments in which urea was added to the liquid for the purpose of obviating the solvent action * Philip E. Browning and Philip L. Blumenthal, Am. Jour. Sci., [4], xxxii,' 246. t See page 442. J Smith, Electro-analysis, page 105, edition of 1911. Jour. Am. Chem. Soc., xxv, 904. || F. A. Gooch and F. B. Beyer, Am. Jour. Sci., [4], xxvii, 59. ^ See pages 13 to 20. ** See Fig. 15, page 15. LEAD 253 of dissolved oxides of nitrogen upon lead dioxide. In the experi- ments with this form of apparatus, the stirring of the asbestos felt by gas evolved upon the bottom of the crucible used as an anode, as well as the deposition of oxide on the outer surface of the crucible, was prevented by taking the precaution to moisten the asbestos, from the outside, with a drop of nitrobenzene which, being insoluble in water, prevents the contact of the aqueous electrolyte with the electrode surface underneath the asbestos. An increase of nitric acid to the proportion of 30 cm. 3 in 100 cm. 3 of solution served to prevent the deposition of lead upon the cathode, but to prevent the re-solution of lead dioxide it was found to be necessary to use the process of continuous filtration, so that the deposit might be compacted upon the felt, and after deposition was complete to replace the acid liquid by a solution of ammonium nitrate without interruption of the current. After washing out the nitric acid with the solution of ammonium nitrate, the final washing was completed with water. The form of apparatus employed, shown in Figure 1 6, and the manner of using are described on page 17. In the table are given the results of experiments following this procedure, and, for com- parison, the result of an experiment in which it was found that, though electrolysis was continued by the circulating process until the filtrate contained no lead, traces of lead dioxide went into solution after the current had been diminished by the gradual dilution with water used to replace the electrolyte. Tests for lead in filtrates and washings were made by neutralizing with ammonium hydroxide and adding ammonium sulphide, or acetic acid and potassium chromate. From the results of the experiments described, it appears that good analytical results in the deposition of lead dioxide may be obtained with the filtering crucible used as an electrolytic cell if nitric acid be present to the proportion of 30 cm. 3 of the con- centrated acid in 100 cm. 3 of solution, the liquid kept in con- tinuous filtration until the electrolysis of the lead salt is complete, the acidic liquid replaced by a solution of ammonium nitrate so that the electric current passing shall not fall off until the nitric acid has been removed, the final washings made with water, and the deposit weighed after drying at 200. The time required for the complete deposition of 0.15 grm. of lead dioxide under the conditions described is about two hours. 254 METHODS IN CHEMICAL ANALYSIS Electrolysis with Continuous Filtration. Pb(NO s ),. grm. Volume. cm. 3 HN0 3 cone. cm. 8 Current. Time, min. PbO 2 found. grm. Theory for PbO 2 . grm. Error, grm. Amp. N.D.ux,. Volt. A. With no ammonium nitrate in electrolyte or in wash-water. 0.2023 50 15 2 4 5 IO 4 5 5 130 0.1460 0.1436 0.0024 B. With ammonium nitrate in electrolyte and in wash- water. O.2O22 50 15 2 5 4 40 4 .10 5 IOO o-i459 o. 1462 -f 0.0003 0.2014 5 IS 2 5 4 S 4 IO 5 US Q-I4S4 o. 1458 +o . 0004 O.2OOI SO 15 2 S 4 5 4 10 5 n5 0.1444 0.1442 O.OO02 o . 2006 50 15 2 5 4 5 4 10 S H5 o. 1448 0.1446 O.OOO2 o . 2046 50 15 2 S 4 5 4 IO S US 0.1477 0.1472 0.0003 C. With ammonium nitrate in wash-water only. O.2O2O 50 15 2 5 4 5 4 IO 5 H5 0.1458 0.1460 +o . 0003 0.2037 50 15 2 5 4 5 4 4 IO 5 H5 0.1470 0-1473 +o . 0003 The Estimation of Lead by Precipitation as Oxalate and Titration with Potassium Permanganate. Many investigators have made use of precipitation as the oxalate for the estimation of lead. Ward* has shown that the addition of considerable amounts of acetic acid favors complete- ness of precipitation, whether ammonium oxalate or oxalic acid is used as the precipitant. Precipitation is effected in the boiling solution, the precipitated oxalate is collected on asbestos in the perforated crucible and washed with small amounts of water. The oxalic acid is set free by treatment of the washed precipitate with warm dilute sulphuric acid and titrated with permanganate. When the acetic acid present does not exceed one-fourth of the volume, precipitation is not quite complete, but if half the volume at precipitation is made up of glacial acetic acid, even * H. L. Ward, Am. Jour. Sci., [4], xxxiii, 334. LEAD 255 in presence of moderate amounts of acetates, the results of titra- tion are accurate. The details of the preferred treatment are given in the table. Precipitation and Titration of Lead Oxalate. Lead taken as nitrate. grm. Volume at precipitation. cm. Acetic acid. cm.' Ammonium oxalate. grm. Lead found, grm. Error, grm. Precipitation by ammonium oxalate. 0.0050 IOO 50 4 o . 0048 O.OOO2 0.0050 IOO 50 4 0.0045 0.0005 0.0250 IOO 50 4 0.0256 +0 . 0006 0.0250 IOO 50 4 0.0250 . 0000 0.0500 IOO 50 4 0.0505 +o . 0005 O. IOOO 200 IOO 8 0.1002 +O.OOO2 Precipitation by oxalic acid. 0.0050 50 25 I O.OO5O O . OOOO 0.0250 50 25 I 0.0256 + O.OO06 0.1000 IOO 50 2 O. IOO2 +0 . OOO2 In presence of 2 grm. of ammonium acetate or potassium acetate. 0.1000 IOO 50 2 O.IOOO . OOOO O.IOOO IOO 50 2 0.0997 0.0003 O.IOOO IOO 50 2 O.IOOO o.oooo CHAPTER VIII. NITROGEN; PHOSPHORUS; ARSENIC; ANTIMONY; BISMUTH; VANADIUM. NITROGEN. The Determination of Nitrogen Liberated by Action of Sodium Hypobromite upon Ammonia Compounds and Derivatives. THE apparatus designed by Kreider* for the determination of volatile products by loss is well suited to the determination of the nitrogen liberated from urea, ammonium oxalate, ammonium chloride, etc., by the action of sodium hypobromite. Results obtained by the use of this apparatus, described and figured elsewhere,! are given in the accompanying table. Determination of Nitrogen. Taken, grm. Found, grm. Error, grm. f O.IOOO o . 0469 +0.0003 I O.IOOO 0.0467 +0.0001 Urea 4 O. IOOO 0.0467 +O.OOOI O.IOOO o . 0468 +O.OOO2 ( O.IOOO 0.0467 -j-o.oooi O.IOOO 0.0204 +0.0007 O.IOOO 0.0197 o.oooo Ammonium oxalatc - O.IOOO 0.0198 +O.OOOI O. IOOO 0.0198 +O.OOOI O.IOOO 0.0196 O.OOOI O. IOOO 0.0264 +O.OOO2 O.IOOO 0.0265 +0.0003 Ammonium chloride. O.IOOO 0.0261 O.OOOI O.IOOO 0.0263 +O.OOOI O.IOOO 0.0261 O.OOOI The Estimation of Nitrates by Expulsion of Nitrogen Pentoxide on Ignition. For the determination of the nitrogen pentoxide combined in nitrates which leave definite oxides on ignition, and estimation of the containing nitrates, various fluxes have been used to aid * J. Lehn Kreider, Am. Jour. Sci., [4], xix, 188. t Seepage I. 256 NITROGEN 257 in the expulsion of the volatile oxide and to conserve the residual oxide in definite form for weighing. Borax, silicon dioxide, potas- sium dichromate and sodium metaphosphate, which have been employed thus, as well as in the similar determination of car- bon dioxide in carbonates* all present certain disadvantages. In sodium paratungstate of composition corresponding approxi- mately to the formula 5Na 2 O.i2WO 3 , or Nai Wi 2 O 4 i, Gooch and Kuzirianf find a material very easily prepared, stable in fusion, and well suited for use as a flux in the rapid determination of nitrates by loss on ignition. The sodium paratungstate is pre- pared by dehydrating and fusing over the blast lamp a known weight of normal sodium tungstate, Na 2 WO 4 .2H 2 O, adding an equal weight of tungsten trioxide, WO 3 (previously ignited with care to remove all ammonia and to insure complete oxidation), and heating to clear fusion. The cooled mass, which is very easily pulverized, is ground and bottled. From this material, kept over sulphuric acid (though not more than ordinarily hygro- scopic), portions are weighed for the analytical determinations. Approximately half the weight of the paratungstate is tungsten trioxide (molecular weight 232), and this should be capable of expelling nitrogen pentoxide (molecular weight 108.02) to an amount one-half its own weight. The weights of paratungstate to be used are approximately four times the weight of nitrogen pentoxide to be expelled. It is a good practice to weigh a plati- Analysis of C. P. Nitrates of Commerce, after Drying. Nitrate taken, grm. Na 10 W 12 41 taken, grm. Loss on ignition, grm. Theory lor N'jOj. grin. Error, grm. KN0 3 . 0.5000 5 2668 o 2670 O.OOO2 0.5000 5 o . 2678 o 2670 + O.OOO8 0.5000 5 0.2674 o 2670 -|-O.OOO4 0.5000 5 0.2672 o 2670 + 0.0002 0.5000 S 0.2675 o 2670 + 0.0005 Sr(NO,)j. 0.5000 2 0.2544 0.2543 + O OOOI 0.5000 3 o 2546 o 2543 +0.0003 Ba(N0 3 ) 2 . o. 5000 3 o 2073 o 2067 +0.0006 o 5000 3 o 2076 o 2067 +o 0009 * See page 226. f F. A. Gooch and S. B. Kuzirian, Am. Jour. Sci., [4], xxxi, 497. 258 METHODS IN CHEMICAL ANALYSIS num crucible, introduce the dried nitrate and weigh again, add a suitable amount of the prepared sodium paratungstate, stir carefully with a platinum wire with care to avoid mechanical loss, and weigh again. The crucible is then heated over a Bun- sen burner, first at very low heat and then to fusion of the mix- ture for five minutes, cooled in a desiccator over sulphuric acid, weighed, and reignited to test the constancy of weight. The constant weight is usually got in the first ignition. In the table are given the results of the estimation of the nitrogen pentoxide in nitrates according to the procedure described. The Estimation of Nitrates by Reduction with a Ferrous Salt and Titration of the Residual Unoxidized Salt. In the methods for the quantitative estimation of nitrates which depend upon the reduction in presence of acid by a ferrous salt and the determination of the amount of oxidation produced, scrupulous care is necessary that the atmosphere in contact with the ferrous salt while the nitrogen dioxide is present shall be free from oxygen. This fact was recognized by Fresenius,* who modified the original process of Pelouzef by filling the flask with carbon dioxide or hydrogen at the outset. Ederf used carbon dioxide similarly. Holland's method, roughly described, con- sists in boiling, until the air is expelled, the solution of the nitrate in a flask provided with a doubly bent exit tube, rubber- join ted and fitted with a pinchcock; then admitting through the tube as the flask cools a mixture of ferrous salt and strong hydrochloric acid, heating the mixture on a water bath, and, finally, titrating the resulting ferric salt with stan- nous chloride. All of these methods give high results, either on account of the oxygen invariably present in carbon dioxide as produced in the laboratory, or because of slow leakage through the rubber con- nections during the long heating, or because the nitrogen dioxide is not driven out completely from the solution of the iron salts, and acts with atmospheric oxygen to oxidize the ferrous salt during the titration. * Zeit. anal. Chem., i, 32. t Ann. Chim., [3], xx, 120. t Zeit. anal. Chem., xvi, 267. Chem. News, xvii, 219. NITROGEN 259 Phelps* has shown, however, that the oxidation of ferrous sulphate by a nitrate in presence of fairly strong hydrochloric acid may be accomplished quantitatively by the aid of the ap- paratus to be described, the standard of the solution of the ferrous salt and the amount remaining after the action being determined either iodometrically or by titration with potassium permanganate . This apparatus consists of a 25O-cm. 3 boiling flask, closed with a rubber stopper carrying a stoppered funnel of 50 cm. 3 capacity to serve as an inlet tube and a glass tube of 0.8 cm. bore to serve as an outlet tube. The inlet tube is constricted at the lower end, and the outlet tube is enlarged just above the stopper to a small bulb (to prevent mechanical loss during the boiling) and bent twice at right angles. The flask is supported above a Bunsen burner and the outlet tube dips under mercury contained in a test tube. The nitrate to be analyzed is introduced with water into the flask, the stem of the separating funnel being left full of water, the outlet tube is adjusted to just touch the surface of the mer- cury in the trap, and air is expelled from the flask by boiling the solution to small volume. An amount of standardized ferrous sulphate solution known to be in excess is introduced into the separating funnel, the outlet tube plunged a centimeter or two deep into the mercury (which is readily accomplished by chang- ing the position of the flask on the wire gauze, provided that the gauze is depressed well at the center and the flask is set well up on the higher part at the beginning of the operation), and then the flame withdrawn until diminution of pressure sufficient to draw the ferrous solution into the flask is made evident by the rise of the mercury in the outlet tube. By applying and with- drawing the flame and by regulating the rate of inflow of the solution, the ferrous salt may be introduced without admitting air, and the funnel washed carefully with an amount of con- centrated hydrochloric acid nearly enough to equal the total volume of the liquid in the flask. After the pressure has been restored in the apparatus by heating the flask, the exit tube is again raised to the surface of the mercury and the solution in the flask boiled to a volume of 10 cm. 3 to 15 cm. 3 . The excess of acid is then nearly neutralized by introducing sodium carbonate * I. K. Phelps, Am. Jour. Sci., [4], xiv, 440. 260 METHODS IN CHEMICAL ANALYSIS in solution, the carbon dioxide evolved assisting in maintaining the pressure in the apparatus so that the condensed liquid in the test tube which may contain oxidized nitrogen dioxide shall not be returned to the iron solution. The flask is cooled and the ferrous salt remaining determined by titration with potassium permanganate after dilution with 600 cm. 3 of water and addition of 2 grm. to 3 grm. of crystallized manganous chloride, or iodo- metrically by first introducing Rochelle salt (3 grm.) in solution and then neutralizing with acid potassium carbonate, after which are added in succession a saturated solution of acid potassium carbonate, iodine, starch paste, standard arsenious acid solu- tion to the bleaching of the starch blue, and, finally, iodine to coloration. It was shown experimentally that prolonged boiling after the dark compound of nitrogen dioxide with the ferrous salt is broken up is essential and that ammonium salts must be absent if the highest accuracy is desired. Experimental tests of the method as outlined are given in the table. Reduction of Nitrate by Ferrous Sulphate: Titration of Excess. KNO 3 taken. Oxygen value of ferrous salt taken. Oxygen value of ferrous salt found. Error on oxygen. KN0 3 found. Error on KNO 3 . grm. grm. grm. grm. grm. , grm. 0.0500 0.01823 0.00621 +0.00015 o . 0506 +0 . 0006 0.0500 0.01865 0.00681 0.00003 0.0499 o.oooi 0.0500 0.01954 0.00768 o.ooooo O.O50O o.oooo O. IOOO 0.02881 0.00507 +0.00001 O. IOOO o.oooo O. IOOO 0.02822 0.00441 +0.00008 0.1003 +0.0003 0.2500 0.06453 0.00512 +0.00009 0.2503 +0.0003 0.5000 0.13394 0.01524 +0.00005 0.5002 + O.OOO2 0.5000 O.I22IO 0.00340 +0.00005 0.5002 +O.OOO2 The Estimation of Nitrates by Reduction with Ferrous Chloride and Measurement of the Nitrogen Dioxide Evolved. In consequence of the fact that analytical methods in which nitrates are estimated by the amount of nitrogen dioxide evolved in the reaction with ferrous salts in presence of acid give un- expectedly low results, the form of apparatus to be used in such processes and the conditions affecting accuracy have been made the object of study by Roberts.* As the result of such * Charlotte F. Roberts, Am. Jour. Sci., [3], xlvi, 126. NITROGEN 261 study, it is pointed out that to insure rapid action, the hydro- chloric acid used should be fairly strong, and that in the measure- ment of the volume of nitrogen dioxide (NO), swept along by carbon dioxide and collected over sodium hydroxide, the best analytical results are obtained when the gas is passed through a solution of potassium iodide to break up, before the measurement is made, any higher, soluble oxides of nitrogen which may have been formed in the process, notably when the ferrous salt is present in only small excess and when the reaction of decomposi- tion takes place in the hot solution. Note is made of the fact that the action of traces of intermixed air upon the nitrogen dioxide is to form nitrogen trioxide which is soluble, and that the nitrogen of the air is added in exactly the proportion according to which the nitrogen dioxide is removed with the disappearing oxygen. When, however, the gas carrying nitrogen trioxide meets with potassium iodide before measurement, the original amount of nitrogen dioxide is regenerated. Nevertheless, ex- perience shows that the danger of error introduced by the use 'of potassium iodide when traces of air are present is small in comparison with the danger of error due to the presence of higher oxides of nitrogen produced in the main reaction. More- over, there is always a small counterbalancing error due to the solubility of nitrogen dioxide in the solution of sodium hydroxide over which it is measured. The apparatus found to be most satisfactory for this work consists of a small tubulated retort, upon the neck of which is fitted a small condenser to prevent loss of liquid during the dis- tillation. Into the tubulature of this retort is fitted tightly, by a carefully ground joint, a tube drawn out so as to dip below the surface of the liquid, and fitted with carefully ground stop- cocks, as shown in the figure, and so branched above as to make it possible to transmit carbon dioxide through the ap- paratus, or to admit any liquid without introducing air. The condenser is joined to a Will and Varrentrapp bulb contain- ing a solution of potassium iodide as a trap, and this in turn is connected by thick vacuum tubing with a Hempel gas burette charged with a strong solution of sodium hydroxide. Carbon dioxide is generated in a Kipp's apparatus by action of boiled hydrochloric acid upon boiled marble, and the liquid is charged with cuprous chloride, following Warrington's device, to take up 262 METHODS IN CHEMICAL ANALYSIS traces of dissolved oxygen. Notwithstanding all precautions, however, the gas from the generator is never so pure that a hundred cubic centimeters of it will not leave a tiny bubble when shaken with a solution of caustic soda. In using this apparatus, the nitrate (about o.i grin, of potas- sium nitrate) is introduced into the retort, generally in the dry condition, carbon dioxide is passed through the apparatus until the gas collected over sodium hydroxide leaves only the minute bubble which the gas from the generator alone has been found to Fig. 21. give, and 40 cubic centimeters of a boiled solution of ferrous chloride in hydrochloric acid are admitted through the funnel tube, after shutting off the carbon dioxide and lowering the leveling tube of the Hempel burette. With the stopcocks arranged as in sketch, the liquid is then slowly heated to boiling and the process, continued until the reaction of the ferrous salt upon the nitrate is apparently complete, when the carbon dioxide is again passed through the apparatus to secure complete removal of the nitrogen dioxide, the absorption of the carbon dioxide being hastened by inclining and shaking the burette at intervals. The volume of the gas under existing barometric and thermometric NITROGEN 263 conditions is noted and from this the weight of the nitrate may be calculated. Results obtained in this manner are given in the table. Reduction by Ferrous Chloride: Measurement of Nitrogen Dioxide. KNO, taken, grtn. KNO 3 found, grin. Error, grm. O.IOOO 0.0990 0.0010 O.IOOO 0.1005 +0.0005 O.IOOO 0.0992 0.0008 O.IOOO 0.0994 0.0006 O.IOOO 0.1008 +0.0008 O.IOOO o . 0989 O.OOII Action of Manganous Chloride in Hydrochloric Acid. The lodometric Determination of Nitrates. Noting that a saturated solution of manganous chloride in concentrated hydrochloric acid acts like the solution of ferrous chloride in hydrochloric acid in inducing the easy decomposition of nitrates, with the difference, however, that all products of oxidation may be distilled, while the metal chloride reverts to its original form, Gooch and Gruener * have applied this reagent to the quantita- tive estimation of nitric acid as well as in the qualitative test. The solution of manganous chloride in hydrochloric acid acts but slowly upon nitrates at the ordinary temperature, but upon warming the decomposition of the nitrate begins at once with the formation of a higher chloride of manganese and liberation of nitrogen dioxide. Ultimately, if heating is continued, chlorine of the higher chloride is evolved and manganous chloride remains. During the process of heating the color of the solution passes from the original characteristic green through darker shades to black, and returns by the reverse changes to the original tint. The decomposition of the nitrate extends under the conditions to the last traces, but the breaking up of the nitrates, with the formation of the higher chloride, does not take place completely in the presence of water amounting to more than a half of the volume of strong acid, and an action already established in strong acid is reversed by the addition of a large amount of water. Chlorates, peroxides and other substances which liberate oxygen * F. A. Gooch and H. W. Gruener, Am. Jour. Sci., [3], xliv, 117. 264 METHODS IN CHEMICAL ANALYSIS or chlorine when in contact with strong hydrochloric acid induce similar phenomena, but in the absence of such other substances the reaction serves to detect nitrates when present in fairly small amounts (perhaps one part in sixty thousand), as shown in the accompanying table : KNOa taken. MnCl 2 .4H 7 O in strong HC1. Color developed. grrn. cm. 3 O.OIOOO IO Black. o . 00500 5 Black. O.OOIOO 5 Dark brown. o . 00050 5 Dark green. O.OOO25 5 Deepened tint. O.OOOI5 5 Deepened tint. o . 00005 5 None. O . OOOOO 5 None. In applying this reaction to the quantitative estimation of nitrates, the nitrate to be estimated is treated, in an atmosphere of carbon dioxide, with a saturated solution of crystallized Fig. 22. manganous chloride in concentrated hydrochloric acid, the vola- tile products of action chlorine, nitrogen dioxide and perhaps nitrosyl chloride are passed into a solution of potassium iodide, and the iodine set free is titrated by sodium thiosulphate. The NITROGEN 265 operation is conducted in an apparatus made wholly of glass where, by any possibility, rubber connections might be acted upon. The retort used was a pipette bent and fitted as shown in Fig. 22. To the retort are sealed Will and Varrentrapp nitrogen bulbs, the outlet tube of which is drawn out so that it may be pushed well within the inlet tube of the second receiver a Will and Varrentrapp absorption flask and held in place by an outside rubber connector. The third receiver acts simply as a trap to exclude air from the absorption apparatus proper. In conducting the experiment the receivers were charged with solutions of potassium iodide, the first containing three grams, the second one gram, and the third only a fraction of a gram for every tenth of a gram of nitrate used. The first receiver was cooled in water during the subsequent process of distillation. Decomposition of Nitrate by Hydrochloric Acid and Manganese Chloride: Titra- tion of Iodine Set Free by Volatile Products. KNO 3 taken, grm. MnCl2 mixture. cm. 3 KNO 3 found, grm. Error in terms of KNO 3 . grm. Error in terms of HNO 3 . grm. O . 2038 . 2053 o. 1032 o. 1017 2O 2O 10 IO 0.2047 0.2057 0.1035 0.1004 +0 . OOOQ +o . 0004 +0.0003 0.0013 +0.0005 +o . 0003 +O.OOO2 O.OOOS o. 1049 IO o. 1040 O.OOOO O.OOOO o. 1027 0.0524 IO IO 0.1023 0.0526 0.0004 +O.OOO2 0.0003 +O.OOOI 0.0513 IO O . 05 1 2 o.oooi O.OOOI 0.0354 10 0.0350 0.0004 0.0003 0.0232 0.0107 0.0127 10 5 5 O.O23O 0.0106 0.0130 O.O002 O.OOO4 +0.0003 O.OOOI O.OOOI +O.OOO2 0.0145 5 0.0143 O.0002 O.OOOI 0.0053 o . 0043 0.0014 5 5 5 0.0052 0.0047 0.0018 O.OOOI +o . 0004 -j-o . 0004 O.OOOI +0.0003 +0.0003 o.oooo 5 o.oooo O.OOOO o.oooo The nitrate and the manganous mixture following it are intro- duced by applying gentle suction to the end of the absorption train. The current of carbon dioxide is started immediately after putting in the manganous mixture. After a suitable time has elapsed for the removal of air, heat is applied to the retort and the distillation is continued until nearly all the liquid has passed over. Finally, the contents of the receivers are united, 266 METHODS IN CHEMICAL ANALYSIS the washing of the bulbs was effected easily and expeditiously by passing the wash-water directly through retort and receiver, the introduction of the manganese chloride into the distillate being not at all prejudicial to the accuracy of the titration. The estimation of free iodine is made by titration with sodium thio- sulphate as soon as may be after admitting air to the distillate,, in order that traces of dissolved nitric oxide may not be reoxi- dized and again react upon the iodide present to liberate more iodine. The results of the experiments conducted in this man- ner are given in the table. Various attempts to utilize the reduction of ar- Distillation with Phosphoric Acid seme acid brought about upon heating a mixture and Potassium o f standard potassium iodide, potassium arsenate Iodide and . . . . . Determination and sulphuric acid with the nitrate, as in the similar of iodine in process for estimating chlorates,* have proved to be futile. The decomposition of the last traces of ni- trates by the action of potassium iodide and sulphuric acid does not occur except at concentrations so great that the sulphuric acid itself liberates iodine from iodides. Gruenert has shown, however, that when phosphoric acid is substituted for sulphuric acid the iodine evolved in the distillation of such a mixture may be taken as the measure of nitrates present in small amounts, provided the concentration of the residue is not carried so far as to bring about reduction of nitric acid to ammonia! nor the tem- perature so raised by removal of water and elevation of the boiling point of the phosphoric acid as to cause dissociation of hydriodic acid. In Gruener's experiments a small retort was used, the neck of which was bent downward about two inches from the body, so that the retort itself might be tipped backward, allowing the unbent portion of the retort to run upward, thus guarding against loss from spattering. Into the tubulature of the retort was ground a glass tube drawn out at both ends to serve as a perfo- rated stopper for the entrance of carbon dioxide. The neck was passed through a rubber stopper into a side-neck Erlenmeyer flask, the exit tube of which was prolonged and dropped into a side-neck test tube used as a trap. The retort was covered with a * See page 463. t Hippolyte Gruener, Am. Jour. Sci., [3], xlvi, 42. \ Chapman, Jour. Chem. Soc., xx, 166 (1867). NITROGEN 267 simply contrived hood which kept the upper parts warm and pre- vented the iodine from settling anywhere. In the retort was placed the nitrate with an excess of potassium iodide, and in the receiver a known amount of decinormal solution of arsenious oxide strongly alkaline with hydrogen sodium carbonate and diluted to a con- venient bulk. The trap contained nothing but water. The method, so far as it is applicable, may be summed up as follows: The nitrate, not to exceed the equivalent of 0.05 grm. of potassium nitrate, is introduced into the retort, with ten times its weight of potassium iodide, and 17 cm. 3 to 20 cm. 3 of phos- phoric acid, of specific gravity 1.43. All water used should be recently boiled. Carbon dioxide is passed from a generator set up with materials carefully boiled and containing cuprous chloride to take up the oxygen from any traces of air. The neck of the retort passes into a receiver containing a known amount of decinormal arsenious oxide, alkaline with a good excess of hydrogen sodium carbonate and diluted to a convenient bulk. To this flask is attached for additional safety a simple trap con- taining water. The solution in the retort is boiled until it is clear that no more iodine remains, when the receiver, after proper washing and addition of the liquid in the trap, is titrated with iodine to find the amount of arsenious oxide still left. This gives the measure of the iodine evolved and consequently of the nitrate present, according to the equation : 2 HNO 3 + 6HI = 4H 2 O + 2 NO + 3 I 2 . The details of test determinations are given in the table : Decomposition by Phosphoric Acid and Iodide: Estimation of Iodine Set Free* KNO 3 taken. KI taken. Found. Specific grav- ity of solu- tion of Amount of solution used. Error, KNO 3 . Error, HNO,. phosphoric acid. grin. grm. grm. cm.* grm. grm. 0.0500 I 0.0500 43 17 o.oooo 0.0000 O.O2OO o-5 O.O2OI 43 17 +0.0001 +O.OOOI O.O2OO 0.0198 43 17 0.0002 o.oooi 0.0250 0.0250 43 17 0.0000 o . oooo 0.0300 0.0307 43 17 +0.0007 +0.0004 0.0300 0.0312 43 17 +O.OOI2 +0.0007 0.0350 0-0353 43 17 +0.0003 +O.OOO2 o . 0400 o . 0409 35 2O +0.0009 +o . 0006 0.0450 o . 0444 35 2O 0.0006 0.0004 0.0500 o . 0499 37 2O o.oooi o.oooi 268 METHODS IN CHEMICAL ANALYSIS The process is good for estimating nitrates in quantities not exceeding the equivalent of 0.04 grm. or 0.05 grm. of potassium nitrate. With quantities of nitrate above 0.05 grm. it is not safe, inasmuch as with a moderate amount of water present some nitric acid distils over undecomposed and with little water present other complications arise. To register the action of nitrates, Gruener* tried Decomposition the effect of antimony trichloride in hydrochloric acid and showed that the reaction proceeds mainly Determination according to the equation, of Oxidation in Residue and 3SbCl 3 + 2 HNO 3 + 6HC1 = 3 SbCl 5 + 2 NO + 5 H 2 O, of Iodine in Distillate. and in smaller degree according to the equation, SbCl 5 + 2 NO = SbCl 3 + 2 NOC1. Nitrosyl chloride fails to oxidize arsenious oxide in alkaline solution, breaking up hydrolytically into hydrochloric acid and nitrous acid, but from acidulated potassium iodide out of contact with air it sets free quantitatively an amount of iodine correspond- ing to the chlorine. These reactions may, therefore, be ap- plied together to the estimation of the nitrate, by noting both the iodine evolved in the action of nitrosyl chloride upon potas- sium iodide in the distillate and the degree of oxidation of the previously standardized antimony salt in the residue. The pro- cedure is as follows: Into a diminutive retort made from a pipette, shaped like a Liebig's drier f and connected by a sliding joint covered by rubber with a Kjeldahl tube used as a receiver, and so placed that carbon dioxide passing through the apparatus shall enter from below the dry nitrate is introduced, and washed down with a few drops of recently boiled water, or, if more liquid is required, with hydrochloric acid. From a burette a definite amount of antimonious chloride solution, somewhat in excess of the nitrate taken, is added. The receiver is charged with potassium iodide in recently boiled water and is joined to a trap filled with water. After carbon dioxide has been passed through the apparatus for about ten minutes, the solution is warmed in a bath at even tempera- ture (iO3-iO7) to insure the safety of the retort, to keep the * Am. Jour. Sci., [3], xlvi, 47. t See Fig. 5, page 5. NITROGEN 269 antimony pentachloride from breaking up, to retain the bulk of the acid in the retort, and to prevent mechanical loss. After fif- teen minutes' digestion the contents of the receiver and trap are washed out and at once titrated with sodium thiosulphate or neutralized and titrated with standard arsenite. The residue in the retort is treated exactly as was the antimonious chloride when it was standardized, viz., by dissolving in hydrochloric acid, adding tartaric acid, diluting, nearly neutralizing with sodium hydroxide with careful cooling to prevent action of the tartaric acid upon antimony pentachloride, treating with an excess of acid sodium carbonate, and titrating with decinormal iodine in presence of starch. Below are given the results of experimental tests. Decomposition by Antimony Trichloride: Determination of Oxidation in Residue and of Iodine in Distillate. KN0 3 taken. grm. KNO 3 from SbCl 5 in residue. grm. KNO 3 from I in receiver. grm. Entire KNO 3 found. grm. Error in KNO 3 . grm. Error in HNO 3 . grm. 0.0222 0.0213 O.OO2O* 0.0233 +O.QOII +O.OOO7 0.0336 0.0307 O.OO26* 0.0333 0.0003 O.OOO2 0.0470 o . 0436 0.0045* 0.0471 +O.OOOI +O.OOOI 0-0553 0.0497 0.0057* 0.0554 +0.0001 +0.0001 o . 0664 0.0673 0.0076* 0.0679 +0.0015 +0.0009 0.0759 0.0670 0.0082* 0.0752 0.0007 0.0004 0.0837 0.0730 O.OIO3* 0.0841 +0.0004 +O.OOO2 0.0934 0.0842 O.OII3* 0.0955 +O.OO2I +O.OOI3', 0.1034 O.OQO2 0.0134* o. 1036 +O.OOO2 +O.OOOI O.O262 0.0235 O.OO24* 0.0259 0.0003 O.OOO2 O.OI27 0.0123 0.0007* 0.0130 +0.0003 +O . OOO2 0.0065 0.0064 0.0003* 0.0067 +O.OOO2 +O.OOOI O.OO26 O.OO22 O.OOOI* 0.0023 0.0003 0.0002 0.1232 o. 1129 0.0098* 0.1227 0.0005 0.0003 . I 540 0.1394 0.0146* 0.1540 O.OOOO O.OOOO- 0.1878 0.1655 O.O2IO* 0.1865 0.0013 0.0008 o . 0530 0.0481 O.OO52t 0.0533 +0.0003 +O.OOO2 0.0547 o . 0484 0.0065t o . 0549 +O.OOO2 + O.OOOI 0.0541 0.0474 0.0063! 0.0537 0.0004 O.OOO2 Found by thiosulphate. t Found by arsenite after neutralization. The lodometric Determination of Nitrites. The apparatus (consisting of a boiling flask fitted with a stopper which carries a stoppered funnel and outlet tube dipping in a mercury trap) previously used by Phelps* for the determina- tion of nitric acid has been applied by him in the determination * See page 259. 270 METHODS IN CHEMICAL ANALYSIS of nitrites.* In this method the nitrite is reduced by the action of potassium iodide and arsenious acid in acid solution and meas- ured by titration of the arsenite remaining, after neutralization. An amount of standard arsenite solution, slightly in excess of that required to take up the iodine to be set free later by the nitrous acid, and 25 cm. 3 of a concentrated solution of sodium carbonate, are placed in the flask. The stem of the stoppered funnel is completely filled with water, the rubber stopper inserted tightly and the contents of the flask boiled until all air is expelled, a process requiring an active boiling of 5-8 minutes. The flame is then removed, the outlet tube is plunged deep into the mercury, the flask is cooled with ice water, and enough sulphuric acid [1:3] (7 cm. 3 ) is sucked in through the funnel tube to nearly decompose the sodium carbonate previously added and liberate carbon dioxide to balance the atmospheric pressure. When the inward and outward pressures have been equalized the outlet tube is raised so that the end shall dip in the water layer con- densed above the mercury in the trap, the acid on the walls of the funnel and in the tube is washed into the flask, and the nitrite solution to be analyzed is run in through the funnel with 2 grm. of potassium iodide. Sulphuric acid [i 13] is then added in amount (5 cm. 3 ) sufficient to acidify the contents of the flask, and potassium carbonate is then added, in solution, to alkalinity or until free iodine has been taken up. The mixture is boiled for five minutes to expel nitrogen dioxide and then cooled, and the residual arsenite is titrated with decinormal iodine in presence of starch. In making the various additions of liquid to the flask, care is of course taken to avoid all introduction of air. When the sulphuric acid is added to the alkaline solution containing the arsenite, iodide and nitrite, iodine is set free locally, but this is at once acted upon by the alkaline arsenite, so that finally, when the acid reaction is reached, there is only a small amount still free, and the possibility of a loss of iodine by volatilization is reduced to a minimum. The table gives the record of experiments made in this manner upon a solution of commercial sodium nitrite, standardized by treatment with potassium permanganate and oxalic acid in acid solution, according to the procedure of Kinnicut and Nef,f the * I. K. Phelps, Am. Jour. Sci., [4], xvii, 198. t Am. Chem. Jour., v, 388. NITROGEN 271 natural error of which is one of deficiency, as is evidenced by the odor of nitrogen oxides observed when even a very dilute solution of a nitrite is acidified. Decomposition by Iodide and Ar senile in Acid Solution: Titration of Residual Ar senile in Alkaline Solution. NaNO 2 taken. Oxygen value of As 2 Os taken. Oxygen value of As 2 O 3 found. Error on oxygen. Error on NaNO 2 . grm. grm. grm. grm. grm. 0.0958 O.OI2OO O . 00064 +O.OOO25 +O.OOII 0.0958 O.OI2OO o . 00066 +O.OO024 +O.OOIO o. 1916 0.03200 o . 00965 +0.00017 +0.0007 o. 1916 0.03200 o . 00965 +0.00017 +0.0007 0.3832 o . 05600 O.OII2O +0.00043 +0.0018 0.3832 o . 05600 O.OIIlS +0.00045 +0.0019 0.6716 o . 08000 O.OOl6o +0.00076 +0.0033 0.6716 o . 08000 0.00158 +0.00078 +0.0034 o. 1916 0.03280 O.OIOO3 +0.00062 +0.0027 The Estimation of Nitrites, and of Nitrites and Nitrates in One Operation. Determination By the action of manganous chloride in hydro- of Nitrites. chloric acid upon a nitrite, passing the products of action into potassium iodide, and collecting the residual nitrogen dioxide, Roberts* has been able to estimate the nitrite both from the iodine set free and from the volume of nitrogen dioxide evolved. The operation is conducted in a slightly modified form of the apparatus employed in the estimation of nitrates. f This consists of a retort, an absorption system charged with potassium iodide, a Hempel burette used for the collection and measurement of the residual gas over sodium hydroxide,! and a carbon dioxide generator for sweeping the gas to the burette. In making the analysis, the air must be thoroughly driven out of the apparatus before the nitrite is introduced, as the carbon dioxide, passing over the solution, decomposes it. Accordingly, carbon dioxide is first passed through the apparatus for some time, then the nitrite is introduced through the funnel tube and rinsed in with * Charlotte F. Roberts, Am. Jour. Sci., [3], xlvi, 231. t See page 260. t See page 262. 272 METHODS IN CHEMICAL ANALYSIS a little water, followed by the manganous chloride solution, care being taken that the water shall not exceed one-third of the total volume of the liquid, according to the precaution shown to be necessary by Gooch and Gruener.* Working in this way with a solution of specially prepared sodium nitrite, the following results were obtained : Decomposition by Hydrochloric Acid and Manganese Chloride: Titration of Iodine Set Free: Measurement of Nitrogen Dioxide Liberated. Volume taken. NaNO 2 determined by KMnO 4 .* NaN0 2 reckoned from NO. NaNO 2 reckoned from iodine. cm. 8 grm. grm. grm. 10 10 IS 0.0463 0.0460 0.0704 0.0456 o . 0460 0.0708 0.0450 o . 0470 O.O722 15 IS 0.0701 0.0688 o . 0704 o . 0696 0.0722 0.0695 Determination of Nitrites and Nitrates. * Process of Kinnicut and Nef . Roberts f has also shown that in treating a mix- ture of nitrite and nitrate according to the method just described for the determination of nitrites, the measure of the nitrogen dioxide and the estimation of liberated iodine afford data for the calculation of the nitrite and nitrate in the mixture. Representing the weight of nitric oxide found by a, and the weight of iodine found by b, and letting x equal the amount of nitric acid operated upon, and y the amount of nitrous acid, 30.01 , 30.01 x + - -y = a, 63.02 47.02^ and whence 380.76 . 126.92 , - x H y = b; 63.02 47.02 * x = 0.248 b 1.051 a, y = 2.35 a -0.1856. Results calculated from data furnished by experiments made in the manner described are given below. * See page 263. f Loc. cit. NITROGEN 273 Nitrite and Nitrate t in One Operation. NaNO 2 taken, grm. NaNO 2 found, grm. Error, grm. KNO 3 taken, grm. KNOj found, grm. Error, grm. 0.0702 0.0718 +O.OOI6 O. IOOO O.IOOO 0.0000 0.0702 0.0712 -j-o.ooio O. IOOO 0.0999 0.0001 0.0702 0.0710 +0.0008 O. IOOO 0.1004 +0.0004 0.0702 0.0698 0.0004 O.IOOO O. IOI2 +O.OOI2 o . 0468 0.0453 0.0015 O.IOOO 0.0994 0.0006 o . 0468 o . 0444 0.0022 0.0500 0.0513 +0.0013 In calculating these results atomic weights were used, which differ somewhat from those now in vogue, but the differences thus introduced are not significant where the inevitable irregu- larities are so considerable. The Estimation of Nitrates and Chlorates in One Operation. A method for the determination of chlorates which has long been in common use consists in the treatment of those compounds with hydrochloric acid, the passing of evolved chlorine into potassium iodide, and the determination of liberated iodine by titration with sodium thiosulphate. This method is analogous to the method proposed by Gooch and Gruener* for the determi- nation of nitrates, excepting that in the latter case the presence of manganese chloride is essential. In the case of the nitrate, however, there is a second product, nitrogen dioxide, which may be collected and measured, as in the process of treating a nitrate with ferrous chloride in hydrochloric acid, described by Roberts, f By combining the determination of the iodine evolved by the action of the products of decomposition upon potassium iodide with the measurement of nitrogen dioxide evolved, Roberts J has been able to effect the estimation of chlorates and nitrates in a single operation involving distillation with a solution of manga- nous chloride in hydrochloric acid, the amount of nitrogen dioxide found giving the amount of nitrate, and the iodine liberated meas- uring both nitrate and chlorate. The operation is carried out with a slightly modified form of * See page 263. f See page 260. J Charlotte F. Roberts, Am. Jour. Sci., [3], xlvi, 231. 274 METHODS IN CHEMICAL ANALYSIS the apparatus employed in the process* for the estimation of nitrates to which reference has been made. In this apparatus, Fig. 21, a small retort fitted with a hollow ground-glass stopper prolonged beneath in a tube, and joined above with two branching tubes, one for the admission of carbon dioxide, and the other, attached to a funnel tube with stopcock, for the admission of liquids without introduction of air, is con- nected with a small condenser, which in turn is attached to an absorption apparatus containing potassiuni iodide, and this with a Hempel burette containing a strong solution of sodium hy- droxide. In treating the mixture of chlorate and nitrate, two Will and Varrentrapp bulbs and generally a Geissler bulb con- taining potassium iodide are employed as the absorption system to make sure that no chlorine shall escape. The mixture of chlorate and nitrate is introduced into the retort, the air is driven out by carbon dioxide, and then the solution of manganous chloride in hydrochloric acid is added through the funnel tube. The liquid becomes dark at once, but a short heating suffices to restore it to its original clear, light- green color. When this has been accomplished, a current of carbon dioxide is passed through the apparatus, the bulbed tubes are disconnected, and their contents titrated with sodium thio- sulphate. The volume of the gas collected in the burette is noted, and the existing barometric and thermometric conditions, from which the weight of nitrate may be calculated. Following are the results obtained in tests of the method. Nitrates and Chlorates in One Operation. KC1O 3 taken, grm. KC1O 3 found, grm. Error. grm KN0 3 taken, grm. KN0 3 found, grm. Error, grm. O. IOOO 0.0990 O.OOIO O. IOOO 0.0995 0.0005 0.0500 o . 0484 0.0016 0.0500 o . 0498 O.O002 0.0500 o . 0496 0.0004 0.0500 0-0515 +0.0015 0.0500 o . 0494 0.0006 0.0500 o . 0508 +0.0008 0.0500 0.0493 O.OOO7 O. IOOO 0.0987 0.0013 O. IOOO 0.0995 0.0005 O.IOOO 0.1007 +0.0007 O.IOOO o . 0980 O.OO2O 0.0300 0.0305 +0.0005 O.IOOO 0.0990 O.OOIO O.IOOO 0.1006 +0.0006 0.0300 0.0293 O.OOO7 See page 261. NITROGEN 275 The Qualitative. Separation and Detection of Ferrocyanides, Ferri- cyanides and Sulphocyanates. The ordinary method of testing for ferrocyanides, ferricyanides and sulphocyanates by means of ferric and ferrous salts leaves little to be desired in point of delicacy when the substances are not present together. When, however, a sulphocyanate and ferrocyanide occur together the colors tend to mask each other, and various methods to obviate the difficulty have been sug- gested, such as bleaching the red ferric sulphocyanate by mer- curic chloride, and distilling the sulphocyanic acid before testing for that acid. In testing for a ferrocyanide in the presence of a ferricyanide, the formation of the deep-blue color with the ferric salt or ferrous salt has generally been considered of sufficient delicacy for all practical purposes. Browning and Palmer* have attempted the separation of these substances from one another, as well as their detection. The Ferro- Potassium ferrocyanide has long been mentioned cyanogen ion. as a precipitant of the ferrocyanides of the rare earth elements, cerium, thorium, yttrium, zirconium, etc., while it is also known that ferricyanides of these elements are soluble. These facts suggested the 'use of some member of the above- mentioned group as a precipitant of the ferrocyanogen ion, and selection was made of a soluble salt of thorium as perhaps the most satisfactory and available. Experience shows that upon adding a few drops of a 10 per cent solution of thorium nitrate to the solution of a ferrocyanide faintly acidified with acetic acid it is possible to detect by the cloudiness produced so little as I part of the ferrocyanide in 500,000 parts of solution. Alkali acetates tend to decompose the thorium ferrocyanide into soluble products, but the difficulty may be overcome by addition of thorium salt or hydrochloric acid. Neither potassium ferricya- nide nor potassium sulphocyanate to the amount of o.i grm. in 10 cm. 3 interferes with this test. The Ferricyano- In making the choice of a precipitant for the ferri- genion. cyanogen ion with a view to subsequent testing for the sulphocyanogen ion by the ferric salt, some reagent giving a colorless solution is preferable. Salts of the elements zinc and cadmium meet this condition, and cadmium salts prove to be the * Philip E. Browning and Howard E. Palmer, Am. Jour. Sci., [4], xxiii, 448. 276 METHODS IN CHEMICAL ANALYSIS more delicate. It is found that o.oooi grm. of the ferricyanide may be readily detected in from 5 cm. 3 to 10 cm. 3 of water acidified with acetic acid even when o.i grm. of potassium sul- phocyanate is present. Both thorium ferrocyanide and cadmium ferricyanide present difficulty in filtering on account of finely divided condition; but this difficulty is met by mixing with the precipitate fine-shredded asbestos and shaking. The Ferrocyano- ^ ne me thod recommended for the separation and gen ion, the detection of the ferrocyanogen, ferricyanogen and fonT^rthfsui- 8111 ? 110 ^^ ^ 11 ions is as follows: phocyanogen ion I. The solution to be tested, preferably dilute ixtures. ant j a k out ^ cm 3 to 10 cm. 3 in volume, is acidified faintly with acetic acid or hydrochloric acid and treated with a soluble thorium salt to complete precipitation. To the liquid and suspended thorium ferrocyanide finely shredded asbestos is added. The whole is agitated and thrown on a filter, and the precipitate is washed with a little water. The washed precipitate is decomposed by strong sodium hydroxide on the filter, the clear filtrate is acidified with hydrochloric acid and the test for the ferrocyanogen ion is made with ferric chloride. II. The filtrate from the thorium ferrocyanide is treated with a soluble cadmium salt to complete precipitation of the cadmium ferricyanide, which, after the addition of the asbestos, is filtered and washed. The cadmium ferricyanide on the filter is decom- posed by sodium or potassium hydroxide, and the solution is filtered and tested with a ferrous salt. III. The filtrate from the cadmium ferricyanide is acidified with hydrochloric acid and treated with ferric chloride, which gives the red ferric sulphocyanate. The table on page 277 gives results of practical tests of this procedure. The Gravimetric Determination of Sulphocyanates. Van Name* has shown that while the sulphocyanate of silver, unlike that of copper, is readily soluble in an excess of ammonium or alkali sulphocyanates, which for this reason may not be used to precipitate silver for gravimetric estimation, the reverse process, the precipitation of a soluble sulphocyanate by an excess * R. G. Van Name, Am. Jour. Sci., [4], x, 454. NITROGEN 277 of silver nitrate, furnishes a convenient means of standardizing sulphocyanate solutions and in general for estimating sulpho- cyanic acid. K 4 FeC 6 N 6 present. grm. K 3 FeC 6 N 6 present. grm. KSCN present. grm. Indication. Tests for K 4 FeC 6 N 6 only. O.OOIO 0.0005 O.OOO2 0.0001 O.I O.I O. I O. I O. I O.I 0. I 0. I Distinct. Distinct. Distinct. Distinct. Tests for K 3 FeC 6 N 6 only. O.I O.I O.I O. I O.OOIO 0.0005 O.OOO2 . OOOI O. I O. I 0. I 0. I Distinct. Fairly distinct. Faint. Very faint. Tests for KSCN only. O.I O.I O. I O. I O.I O.I O.I O.I O.OOIO 0.0005 O.OOO2 O.OOOI Distinct. Distinct. Distinct. Distinct. Tests for K 4 FeC 6 N 6 , K 3 FeC 6 N 6 and KSCN. OIOO 0050 OOIO O.OIOO 0.0050 O.OOIO OIOO 0050 OOIO Good Good Good tests for tests for tests for K K K 4FeC 6 N 6 , 4FeC 6 N 6 , 4FeC 6 N 6 , K 3 FeC 6 N 6 , K 3 FeC 6 N 6 , K 3 FeC 6 N 6 , KSCN. KSCN. KSCN. Tests of mixtures unknown to analyst. O OOIO O OOIO Found K 4 FeC b N 6 , K 3 FeC 6 N 6 . O OOIO O OOIO Found K 4 FeC 6 N 6 , KSCN. O.OOIO O.OOIO O.OOIO Found K 4 FeC 6 N 6 , K 3 FeC 6 N 6 , KSCN. When freshly precipitated the sulphocyanate of silver re- sembles the chloride in appearance, but when allowed to stand a few hours becomes finely granular and is very easily filtered and washed. It may be safely dried upon an asbestos filter at 110 to 120 to a constant weight corresponding to the theoretical constitution; but at a somewhat higher temperature is decom- posed, leaving a residue of silver sulphide. 2 7 8 METHODS IN CHEMICAL ANALYSIS To the neutral solution of the sulphocyanate in approximately 100 cm. 3 of water, silver nitrate in solution is added in excess. The precipitate is collected upon asbestos in a platinum crucible, washed with cold water and dried to a constant weight at 115, the drying requiring usually between two and three hours. The filtering is facilitated by allowing a few hours for the precipitate to settle; but this is by no means essential, as it is easy with a little care to obtain a clear filtrate even when the filtering is performed at once. In the following table are results obtained by this procedure with like volumes of a solution of pure ammonium sulphocyanate free from chloride. Gravimetric Determination of Sulphocyanates . Final Volume of Liquid 150 cm. 3 . NH 4 SCN. cm. 3 AgNO 3 . cm. 3 Excess of AgNO 3 . cm. 3 AgSCN found, grm. 25 25 25 25 25 25 25 25 25 25 25-3 25-3 25-4 25-4 30-4 0.15 0.15 0.25 0.25 5-25 0.4372 0.4376 0-4373 0-4375 0.4382 o . 4366 0.4381 0-4373 0.4372 0.4369 Rough excess. Rough excess. Rough excess. Rough excess. Rough excess. Mean 0.4374 The mean of the weights of silver sulphocyanate, 0.4374 g rm - is equivalent to 0.2006 grm. of ammonium sulphocyanate for every 25 cm. 3 of solution. Four titrations of the same solution by Volhard's method, against a silver nitrate solution whose standard had been fixed by gravimetric determination as silver chloride, gave as a mean result 0.2003 grm. of ammonium sulpho- cyanate for 25 cm. 3 of solution. The agreement between these two values is within the possible error of the Volhard standard. It is, therefore, evident that the standard of a sulphocyanate solution, free from chloride, obtained in the above way may safely be employed for the estimation of unknown amounts of silver by Volhard's method, as well as for other purposes. NITROGEN 279 Volumetric Estimation of Sulphocyanates by Potassium Permanganate. When a solution of a sulphocyanate is acidified with sulphuric acid and titrated with potassium permanganate in the usual manner, the end-point is sharp, but the results, calculated from the equation HSCN+3 O +H 2 O = HCN+H 2 SO 4 , are invariably low, the magnitude of the error varying greatly, as the following tables show, with the time occupied in titrating, temperature, dilution, and amount of shaking. The principal Taken for each experiment: 50 cm. 3 of approximately n/6$ NH 4 SCN, equiva- lent to 46.61 cm. 3 of the KMnO 4 solution. H 2 S0 4 [1:1]. cm. s Concentra- tion of NH 284 METHODS IN CHEMICAL ANALYSIS The precipitate is washed with 10 per cent nitric acid and then with I per cent potassium nitrate. The asbestos felt is trans- ferred to a i5O-cm. 3 flask or narrow-based Erlenmeyer. The precipitation flask and cork are thoroughly washed with a mix- ture of 5 cm. 3 of ammonia and 10 cm. 3 of water, and the wash- ings are allowed to rinse the sides of the perforated crucible standing on a small funnel and so to run into the flask. Strong hydrochloric acid is added, 25 cm. 3 , and, when the phosphorus does not exceed 0.0060 grm., 0.5 grm. of potassium iodide; but when more phosphorus is present a little more potassium iodide is needed. Experience has shown that the iodide present should not exceed the amount theoretically necessary by more than a half-gram. The flask is trapped loosely with a short bulbed tube hung in the neck.* The liquid is boiled down from a total volume of 40 cm. 3 to just 25 cm. 3 , easily marked by two strips of paper pasted on opposite sides of the flask. If the solution is boiled farther, the molybdic acid is likely to be reduced beyond the degree of oxi- dation indicated by the symbol Mo 2 O 5 . The residue is cooled and transferred to a stoppered bottle (shown in Fig. 2) fitted with a separatory funnel and a trap filled with a solution of potassium iodide. Through the stoppered fun- nel- are added in solution I grm. of tartaric acid, enough sodium hydroxide to nearly neutralize the free acid, followed by acid sodium carbonate to complete the neutralization, and a measured amount of iodine in excess of that required for the oxidation. After neutralization the iodine color in the solution should perceptibly fade within fifteen minutes ; but for complete oxida- tion the bottle should be set aside, out of sunlight, for an hour and a half, and then the excess of the iodine is titrated with a stand- ard solution of arsenious acid. Since there is a slight tendency on the part of the iodine to form a little iodate during the long digestion, it is wise to acidulate the solution in each case slightly with dilute hydrochloric acid after the titration with the arsenic solution, and then to deter- mine by sodium thiosulphate the trace of iodine which has taken the form of iodate. In the following table of test experiments the absolute errors in terms of phosphorus are given; and the percentage errors, * See. Fig. 6, page 6. PHOSPHORUS 285 between the phosphorus taken and the phosphorus found, re- ferred to 10 grm. of material (the maximum amount of high- grade iron or steel usually taken for analysis), are also added. Reduction of Phosphomolybdate by Hydriodic Acid: lodometic Determination of Reduced Molybdic Acid. Amount of P taken, gnn. Amount of P found, grm. Error on P. grm. Error of P. Per cent, grm. Neutralized by 0.002727 o. 001812 o . 000909 0.003508 0.005454 o. 001818 0.003636 o . 000909 0.000363 0.008180 0.002778 0.001743 0.000914 0.003262 0.005417 o. 001861 0.003716 o . 000988 0.000289 0.008179 +0.000051 0.000069 +0.000005 0.000246 0.000037 +o . 000043 +0.000080 +0.000079 0.000074 O.OOOOOI +O.OOO5 0.0007 +o . 00005 O.OO2* 0.0003 +0.0004 +o . 0008 +O.OOO8 O.OOO7 O.OOOOI NaHCO 3 . NaOH+NaHCO 3 . NaOH+NaHC0 3 . NaOH+NaHCO 3 . NaHCO 3 . NaOH+NaHCO 3 . NaHCO 3 . NaHC0 3 . NaOH+NaHCO 3 . NaHCO 3 . Obviously accidental. The Estimation of Phosphoric Acid and Phosphorus Precipitated as Ammonium Phosphomolybdate. The method studied by Randall * for the estimation of molyb- dic acid by the aid of the zinc reductor, the receiving flask charged with ferric alum, and the permanganate titration, has been ap- plied by himf to the estimation of phosphorus in iron, precipi- tated in the form of ammonium phosphomolybdate. According to Randall's procedure, the phosphomolybdate, precipitated in a flask and shaken in the usual manner, is allowed to settle, then filtered on asbestos in a perforated crucible, and washed with a solution of ammonium acid sulphate (15 cm. 3 ammonia, 25 cm. 3 sulphuric acid, I liter water). The flask is washed out with a solution of 20 cm. 3 of water and 5 cm. 3 of ammonia, and this is poured on the asbestos in the crucible. The molybdenum solution is acidified with 10 cm. 3 of strong sulphuric acid and passed through the reductor into the ferric alum solution, pre- ceded by 100 cm. 3 of hot water and followed by 200 cm. 3 of the hot dilute acid with 100 cm. 3 of water, the reduced solution being titrated immediately with approximately tenth normal permanga- nate. The results are calculated on the assumption that the am- * See page 424. f D. L. Randall, Am. iour. Sci., [4], xxiv, 315. 286 METHODS IN CHEMICAL ANALYSIS monium phosphomolybdate contains phosphorus and molybde- num in the proportion given by the symbol (NH 4 )3i2MoO 3 PO4, and that the reduction proceeds to the condition represented by the symbol Mo 2 O 3 . In the following table are shown results obtained by this procedure applied to pure ferric nitrate and a known amount of microcosmic salt. Determination of Phosphorus by Titration of Reduced Phosphomolybdate. P taken, grm. P found, grm. Error, grm. . 003645 0.003673 +0.000028 0.003645 0.003697 +0.000052 0.003645 0.003638 0.000007 . 003645 0.003726 +O.OOOo8l 0.003645 o . 003630 0.000015 0.003645 0.003661 +O.OOOOI6 The Determination of Phosphoric Acid by Precipitation as Uranyl Phosphate and Estimation of the Uranium Volumetrically. The method of estimating uranium by reduction in the zinc redactor and oxidation with permanganate* has been applied by Pulmanf to the determination of uranic oxide, and so of phosphoric acid, precipitated as ammonium uranyl phosphate. This precipitate is filtered with extreme difficulty, but with an asbestos felt coated with the finer floating particles of the par- tially settled emulsion of prepared asbestos, it is possible to obtain filtrates from the ammonium uranyl phosphate which are perfectly clear, though the process of filtering and washing is slow on account of the compactness of the felt surface and the gelatinous nature of the precipitate. The process as worked out for the determination of the phos- phoric acid is as follows: A measured amount of a standard phosphate solution (containing about 4.7 grm. of microcosmic salt per liter) is drawn into a beaker, and a solution containing 12 grm. of ammonium acetate, formed by neutralizing about 10 cm. 3 of ammonium hydroxide (0.90 sp. gr.) with acetic acid * See page 430. t O. S. Pulman, Jr., Am. Jour. Sci., [4], xvi, 229. PHOSPHORUS 287 (50 per cent), and from 2 cm. 3 to 4 cm. 3 of free acetic acid is added. The total volume is made up to about 150 cm. 3 and the solution heated nearly to boiling. The ammonium uranyl phosphate is then precipitated by slowly adding an excess of uranium nitrate, with stirring, and the mixture is boiled gently for about twenty minutes, allowed to settle, and filtered on a tight felt of asbestos. The precipitating beaker and the pre- cipitate are washed thoroughly with a dilute solution of ammo- nium acetate containing a little free acetic acid (to overcome the tendency of the precipitate to pass through the filter), and the crucible containing the precipitate is placed in a glass funnel. Enough dilute sulphuric acid [1:5] is then added to dissolve the precipitate and thoroughly wash out all the soluble uranium salt from the asbestos, the solution being caught below as it passes through the crucible and funnel in the beaker used for the precipitation. The solution is made up to a volume of from 100 cm. 3 to 150 cm. 3 with dilute sulphuric acid [1:5], heated to boiling. A few cubic centimeters of warm dilute sulphuric acid [1:5] are passed through the reductor and followed by the uranium solution, a few cubic centimeters more of the dilute sulphuric acid, and 250 cm. 3 of hot water. The contents of the flask are then poured into a porcelain dish, diluted with 200 cm. 3 of hot water, and titrated with a n/io solution of potassium permanganate. Results obtained by this method are shown in the following table: Reduction of Uranyl Phosphate and Titration with Permanganate. P*0 6 taken. UO 3 corre- sponding to P 2 O 5 taken. H 2 SO 4 (1.84). Dilution at reduc- tion. KMnO 4 . UO 3 found. Error on U0 3 . Error on P 2 6 . grin. grm. cm. 8 cm. 8 cm. 3 grm. grm. grin* o . 0404 o. 1630 25 ISO II .06 0.1632 +O.OOQ2 +0.00005 o . 0404 o. 1630 25 150 11.03 0.1628 0.0002 0.00005 O.O226 O.O9I2 20 1 2O 6.14 o . 0906 O.OOO6 0.00015 O.O226 O.O9I2 20 1 2O 6.17 0.0911 ^o.oooi O.OOOO2 0.0719 o. 2902 25 ISO 19.62 0.2896 O.OOO6 0.00015 0.0719 0.2902 25 ISO 19.61 0.2894 0.0008 O.OOO2O The process for the determination of uranium by the reductor depends upon the fact that any reduction of uranium lower than uranous oxide and such reduction undoubtedly takes place 288 METHODS IN CHEMICAL ANALYSIS in the reductor is corrected by exposure to the air, the lower oxide being rapidly oxidized to exactly the uranous state, while the uranous salts are stable enough to be estimated before they are oxidized appreciably by atmospheric action. ARSENIC, ANTIMONY AND TIN. The Determination of Arsenic by Precipitation as Ammonium Magnesium Ar senate and Weighing as Magnesium Pyroarsenate. The striking analogy between the phosphates and the arse- nates led Levol* to undertake the separation of an ammonium arsenate corresponding to the ammonium magnesium phos- phate, the composition of which Berzelius had given. Levol states that ammonium magnesium arsenate of the composition NH 4 MgAsO4.ioH 2 O is obtained by adding a solution of a double ammonium magnesium salt to arsenic acid, that it is a salt possessing about the same degree of solubility in water, in am- moniacal water, and in ammoniacal water containing magnesium salt, as the corresponding phosphate, and that at red heat, after carefully drying, it yields magnesium pyroarsenate. Several sources of error in this process have been pointed out by many investigators. First, the low indications obtained when the pyroarsenate is weighed suggest a loss of arsenic during ignition, in consequence of the reducing action of ammonia evolved in the process. f Another possible source of error is the solubility of ammonium magnesium arsenate in ammoniacal w r ater and solu- tions of ammonium salts.J There is also the possibility that the constitution of ammonium magnesium arsenate as precipitated may not be ideal, in consequence of the action of ammonium salts known to be influential in determining the constitution of the analogous ammonium phosphates of magnesium and other elements. * Ann. Chim., [3], xvii, 501. t Wach and Rose, Schweigger, Jour. Ch. Phys., lix, 297. Reichel, Ann. Phys., Ixxvi, 20. Rammelsberg, Ber. Dtsch. chem. Ges., xiv, 279. Kaiser^ Zeit. anal. Chem., xiv, 250. t Rose, Zeit. anal. Chem., iii, 206. Wood, Am. Jour. Sci., (3], vi, 368. Brauner, Zeit. anal. Chem., xvi, 57. Neubauer, Zeit. anorg. Chem., ii, 45; Zeit. angew. Chem., 1896, 435; Jour. Am. Chem. Soc., xvi, 289. Gooch and Austin, Am. Jour. Sci., [4], vi, 233; vii, 187; viii, 206. ARSENIC, ANTIMONY AND TIN 289 The conditions to be observed in applying the method to the determination of arsenic have, therefore, been carefully studied by Austin.* It is shown in the first place by special tests that the presence of ammonium chloride does produce solubility of the precipitate thrown down by magnesia mixture,! but that this solvent effect may be overcome even when the ammonium chloride present amounts to as much as 60 grm. in 300 cm. 3 of total volume, by a sufficiency of the magnesia mixture; and, further, that the ammonium magnesium arsenate once precipitated may be safely washed with small amounts (25 cm. 3 to 50 cm. 3 ) of faintly ammo- niacal water. It appears also that the ammonium salt induces the formation of an ammonium magnesium arsenate too rich in ammonia to give the pyroarsenate, Mg 2 As 2 O7, on ignition, even when the precipitation is complete. It is found, however, that a suitable increase in the amount of magnesia mixture present at precipi- tation may bring about the formation of an ammonium magne- sium arsenate of ideal constitution, even in presence of a consid- erable amount of the ammonium salt. According to the most favorable procedure, the slightly acid solution of the arsenate containing no ammonium salts is added drop by drop to the distinctly ammoniacal magnesia mixture, and a little more ammonia is added. The precipitate is filtered off as soon as it subsides, on asbestos in the per- forated crucible and with use of the filtrate to effect the trans- fer, and washed with about 25 cm. 3 of faintly ammoniacal water applied in small portions. After careful drying, the residue is ignited with caution and weighed as magnesium pyroarsenate, Mg 2 As 2 O 7 . When no ammonium salts are present an excess of about 30 cm. 3 of magnesia mixture in a total volume of 200 cm. 3 is sufficient to form the arsenate in ideal condition. If ammonium salts are present and conditions prevent their removal, the * Martha Austin, Am. Jour. Sci., [4], ix, 55. t The magnesia mixture is prepared by dissolving no grm. of the crys- tallized magnesium chloride in a small volume of water, filtering, and adding to it 58 grm. of ammonium chloride (purified in solution by adding bromine water and bleaching with ammonia), filtering, diluting to a volume of 2 liters, and adding enough ammonia 10 c.c. to make the solution smell distinctly of ammonia. 2QO METHODS IN CHEMICAL ANALYSIS slightly acidulated arsenate should be added gradually to a very large excess (150 cm. 3 ) of the magnesia mixture. Results obtained by the procedure are given in this table. Determination of Arsenic as Magnesium Pyr 'oar senate. Mg 2 As 2 O 7 corresponding to As 2 O 5 . Magnesia mixture. cm.* NH 4 C1. grm. Taken, grin. Found, grm. Error, grm. In absence of ammonium salts. o . 7843 0.7830 0.0013 50 0.7843 o . 7849 0.0006 50 0.7843 0.7841 O.OOO2 50 . 0.7843 0.7843 . 0000 50 In presence of ammonium salts. 0.7843 0.7763 0.0080 75 IO 0.7843 0.7762 0.0081 75 IO 0.7843 0.7832 O.OOII IOO IO o . 7843 0.7838 0.0005 100 10 0.7843 0.7784 -0.0059 IOO 20 0.7843 0.7810 -0.0033 IOO 20 0.7843 o . 7849 -f-o . 0006 150 60 0.7843 o . 7846 +o . 0003 150 60 In no one of the many precipitates tested by silver nitrate for included chlorides was more than a trace found. Precipitation of Small Amounts of the Arsenate. Mg 2 As 2 7 . Magnesia mixture. NH 4 C1. Taken. Found. Error. cm. 8 grm. grm. grm. grm. 25 0.0015 0.0014 O.OOOI 2 5 0.0015 0.0017 +O.OOO2 25 . 0.0077 0.0078 +O.OOOI 50 0.0077 0.0076 O.OOOI 25 IO 0.0015 0.0013 O.OOO2 25 10 0.0077 0.0070 0.0007 To effect the precipitation of amounts of arsenate so small as not to be at once precipitable, recourse may be taken to a proc- ARSENIC, ANTIMONY AND TIN 291 ess of freezing and melting, in which the magnesium arsenate is made insoluble in the freezing and continues to be insoluble when the medium melts. This is best effected by putting the solution in a platinum dish, surrounding the dish with a mix- ture of ice and salt until the mass is solid, and then allowing the mass to melt at the room temperature. Results obtained in this manner by Gooch and Phelps* are given in the table. The lodometric Estimation of Arsenic Acid. Holthoff's development of Mohr's suggestion relative to the reduction of arsenic acid to the lower condition of oxidation by the action of sulphurous acid,f with the demonstration that arsenic acid can be evaporated even to dryness in presence of hydrochloric acid without danger of significant volatilization, has placed the analysis of ordinary compounds of arsenic within the scope of Mohr's classical and exact method of determination by titration with iodine. As Holthoff left the method, it is satis- factory so far as regards accuracy, and as modified by McCay,{ who substitutes for the four hours' digestion heating for one hour in a pressure bottle, is eminently successful. Gooch and Brown- ing! have still further shortened the process of reduction of arsenic acid by making use of hydriodic acid as the active agent instead of sulphurous acid. A method for the determination of iodine in Reduction by ,,.,,, haloid salts based upon the action of arsenic acid, and Oxidation m ^ e presence of sulphuric acid, according to the by Iodine in Alkaline Solu- equation, tion ' H 3 As0 4 + 2 HI = HsAsOs + H 2 O + I 2 , the iodine being completely volatilized, but leaving behind in the arsenious acid produced by the action the record of the amount of hydriodic acid originally present, is described else- where. || This reaction is in the present case utilized conversely, and potassium iodide in excess, in presence of sulphuric acid, is employed to bring about the reduction of the arsenic acid to arsenious acid, which may be determined, after neutralization, by * F. A. Gooch and M. A. Phelps, Am. Jour. Sci., [4], xxii, 492. t Zeit. anal. Chem., xxiii, 378. t Am. Chem. Jour., vii, 373. F. A. Gooch and P. E. Browning, Am. Jour. Sci., [3], xl, 66. II See page 457. 2Q2 METHODS IN CHEMICAL ANALYSIS the iodine method. The conditions of the methods are different, in that in the former the hydriodic acid is entirely broken up by the action of the arsenic acid, and the iodine volatilizes easily; while in the latter some hydriodic acid must remain in solution until a very low degree of concentration is reached, and remaining must exhibit its characteristic proneness to retain free iodine. It is found in practice that when a solution made up to contain sulphuric acid, an arsenate, and potassium iodide to an amount somewhat in excess of that theoretically demanded to effect the conversion of the arsenic acid to arsenious acid, is boiled, iodine is evolved. The color of the liquid passes from the dark red when the iodine is abundant through the various gradations of tint to a canary yellow, and then, as the sulphuric acid reaches a degree of concentration sufficient to determine by its own specific action the liberation of iodine, the color again darkens. If the process of concentration is continued, and much arsenic is present, crystals of arsenious iodide separate and form more abundantly on cooling. If evaporation is pushed still farther arsenious iodide begins to volatilize, and at the point where the sulphuric acid fumes the liquid loses all color and the arsenic has vanished more or less completely. In one experiment con- ducted in this manner it was found, by the method to be described later, that of 0.3861 grm. of arsenic pentoxide originally present with I grm. of potassium iodide and 10 cm. 3 of sulphuric acid [i : i] the equivalent of 0.1524 grm. remained. In another simi- lar experiment, in which, however, only a few milligrams of ar- senic oxide were involved, not a trace of arsenic remained at the end. It is obvious that two points in this course of action demand attention : First, means must be used for removing the remnant of free iodine which is withheld by the hydriodic acid, or of rendering it harmless in the titration process to follow; and, secondly, the degree to which the solution may be concentrated without loss of arsenic must be fixed. In the converse of this process, the marked influence of the amount of sulphuric acid present upon the degree of concentration necessary to expel the iodine was particularly noted. In the present case, the effect of the proportion of sulphuric acid in solutions containing definite amounts of potassium iodide and potassium arsenate is like- ARSENIC, ANTIMONY AND TIN 293 wise of first importance. In studying the effects of concentration, the solution was made up to about 100 cm. 3 and concentrated by boiling until the color was faintest; then, to determine provi- sionally, and for preliminary purposes, the point at which vola- tilization of arsenic was likely to occur, the concentration was continued until the arsenious iodide began to separate. The results are tabulated as follows: KI. As 2 5 . H 2 S0 4 [i : i]. Volume when color was lightest. Volume when AsI 3 appeared. grm. grm. cm.3 cm. 3 cm.j I o. 1900 20 80 33 I o. 1900 15 65 25 I o. 1900 IO 40 19 I 0.1900 5 30 II The amount of sulphuric acid which, considering rapidity in concentrating to the proper point, ease in neutralizing the acid previous to titration, and general convenience in manipulation, seemed to be best is 10 cm. 3 of the [i: i] mixture. The most suitable limit of concentration of the solution appears to be 40 cm. 3 . It is manifest from the phenomena described that when much hydriodic acid remains in the solution the last portions of free iodine cannot be completely removed by heat without volatiliza- tion of the arsenic. It was found that upon adding approximately n/ioo sulphurous acid drop by drop to the hot concentrated solution the point at which the color vanished could be deter- mined without difficulty, but that if the solution was permitted to stand a single minute the color of iodine returned, developed by the action of air upon the hot hydriodic acid. By diluting the solution with cold water as soon as the sulphurous acid has done its work and immediately neutralizing with potassium car- bonate, reversion of arsenious acid to arsenic acid is precluded, magnesia mixture producing in the solution no precipitate of the ammonium magnesium arsenate. The process as recommended by Gooch and Browning may be summarized briefly as follows : To the arsenate taken in solu- tion in a 2OO-cm. 3 Erlenmeyer flask are added potassium iodide in excess of the amount needed according to the equation to complete the reduction, and 10 cm. 3 of [1:1] sulphuric acid 294 ^METHODS IN CHEMICAL ANALYSIS The liquid is diluted to about 100 cm. 3 and boiled rapidly (with the precaution of trapping with a two-bulbed tube hung with the large end downward)* until the volume diminishes to 40 cm. 3 , shown by a mark upon the flask. The color of free iodine is bleached by cautious additions of sulphurous acid (corresponding roughly to centinormal iodine) and the solution is instantly diluted with water, nearly neutralized with potas- sium carbonate, and completely with the acid carbonate. The liquid is cooled and titrated as usual with iodine, using starch as an indicator. The whole operation is easily completed in a half-hour. Results obtained by the procedure are given in the table. Reduction by Hydriodic Acid: Oxidation by Iodine in Alkaline Solution. KI taken, gnu. H 2 S0 4 [i : i] taken. cm.j As 2 O 6 taken, grin. As 2 O 5 found, grin. Error, grm. i-5 IO 0.3861 0.3862 +O.OOOI i-5 IO 0.3862 0.3856 O.OOo6 i-S IO 0.3861 0.3862 +0.0001 i-S 10 0.3860 0.3862 +O.OO02 . i-S IO 0.3863 0.3862 o.oooi i-.S IO 0.3862 0.3862 o.oooo I IO 0.1927 0.1922 -0.0005 I IO 0.1928 0.1922 0.0006 I IO 0.1930 0.1925 0.0005 I IO 0.1930 0.1927 0.0003 I IO 0.1936 o. 1929 0.0007 I IO o. 1929 0.1928 o.oooi I 10 0.0383 o . 0380 0.0003 I 10 0.0383 0.0385 +O.OOO2 o-5 IO o . 0383 0.0384 +O.OOOI 0.4 IO o . 0383 0.0385 +O.O002 o-3 IO o . 0383 0.0386 +0.0003 O.2 IO o . 0383 0.0384 + O.OOOI O.2 10 0.0076 o . 0074 O.O002 O.2 IO 0.0076 0.0074 0.0002 0. 2 IO o . 0038 0.0034 O.0004 O.2 10 o . 0038 0.0034 O.OO04 1 In subsequent work by Gooch and Morris f it is shown that the process may be shortened by restricting the volume at which heating begins so that the boiling need not be extended beyond five or six minutes. f * See Fig. 6, page 6. t F. A. Gooch and Julia C. Morris, Am. Jour. Sci., [4], x, 151. . ARSENIC, ANTIMONY AND TIN 295 According to this slight modification, the solution of the arse- nate is heated in a {rapped Erlenmeyer flask* with potassium iodide to an amount about 0.5 grm. in excess of the amount theoretically required and 10 cm. 3 of sulphuric acid of half strengthen a total volume of 50 cm. 3 to 75cm. 3 . The liquid is boiled till the iodine vapors are no longer visible in the flask above the liquid, the iodine color in the still hot liquid is bleached by the cautious addition of sulphurous acid, the whole is diluted with cold water, and cooled quickly. The solution is nearly neutralized with potassium hydroxide and the neutraliza- tion is completed with acid potassium carbonate. The reduced acid is titrated with iodine after adding the starch indicator. By this procedure the results of the following table were obtained. Reduction by Hydriodic Acid and Sulphurous Acid: Oxidation by Iodine in Alkaline Solution. Volume. cm. 1 H 3 O 3 AsO taken, grm. HjO 3 AsO found, grm. Error, grm. 35 0.1559 0.1559 o.oooo 35 0-1559 0.1560 +O.OOOI 40 0.1559 0.1559 0.0000 65 0.1559 0.1559 0.0000 5 0.2495 0.2499 +0.0004 50 0-2557 0.2449 0.0008 60 0.3119 0-3H7 O.OOO2 60 0.3119 0.3120 +O.OOOI 75 0.3119 0.3124 +0.0005 75 0.3119 0.3132 +0.0013 75 0.3119 0.3121 +O.OOO2 75 0.3119 0-3H5 0.0004 75 0.3119 0.3124 +o . 0005 Reduction by j/he process just described for the reduction and Hydriodic Acid: J Titrationof estimation of arsenic acid, depending upon the re- iodine Liberated. mova i by volatilization of all but the last traces of liberated iodine, and the conversion of this minute residue by sulphurous acid, involves no secondary reactions of a sort likely to influence the main effect. It is exact and rapid. The method of Williamson, f brought forward more recently, depends upon the conversion of the liberated iodine to hydri- odic acid. The interaction at ordinary temperatures of a suit- * See Fig. 6, page 6. f Jour. Soc. Dyers and Colorists, 1896, 86-89. 296 METHODS IN CHEMICAL ANALYSIS ably strong acid, hydrochloric or sulphuric acid, upon the mix- ture of the arsenate and iodide, sets free iodine, and the liberated iodine is converted to hydriodic acid by the action of sodium thiosulphate, the end-point being the disappearance of the iodine color. According to Williamson's directions, 25-cm. 3 portions of the solution of the arsenate are treated with potassium iodide and mixed with an equal volume of hydrochloric acid of sp. gr. 1. 1 6. The precaution is recommended that the strength of the solution of the arsenate shall not exceed the decinormal value, in order that the dilution consequent upon titration by the thio- sulphate may not be too great; the reducing action brought about by the action of the strong acid upon the arsenate and iodide being reversible upon the dilution of liquid with water. This procedure thus limits the process to the determination of about o.i 8 grm. of arsenic acid in 25 cm. 3 of the solution to be treated with an equal volume of hydrochloric acid of sp. gr. 1.16. Obviously, however, the process should, so far as the reduction is concerned, be applicable to larger amounts of arsenic, provided the strength of the acid is kept up proportionately. It is essen- tial that the liquid at the end of the titration should contain approximately 10 per cent of its mass of absolute hydrochloric acid or about one-third of its volume of the aqueous acid of sp. gr. 1.16. The arsenic acid is measured either by the amount of stand- ard thiosulphate required to bleach the iodine or by the amount of iodine required afterward to reoxidize the arsenious acid, after neutralizing with acid potassium carbonate. If the former alter- native is followed, the end reaction must be the disappearance of the yellow color of the iodine, since in solutions so strongly acid it is impossible to place dependence upon the starch indicator; in using the latter alternative, the starch indicator is, of course, permissible and preferable. In the direct titration of the iodine by thiosulphate two sources of error present themselves as possibilities: first, the excessive liberation of iodine by the action of air upon, the strongly acidu- lated iodide; and second, the liability of the thiosulphate,* if present' even in momentary or local excess during the process of titration, to break down under the action of strong acid, thus * Norton, see page 364. ARSENIC, ANTIMONY AND TIN 297 changing its capacity to convert iodine to hydriodic acid. The latter contingency should be remote in proportion to the caution used in adding the thiosulphate and in keeping the liquid well stirred; the former must of necessity vary with the acidity of the solution containing the iodide, the time of exposure to atmos- pheric action, and the degree of contact with the air incidental to stirring. How far each of these possibilities is likely to inter- fere in the practical conduct of an ordinary analysis has been in- vestigated by Gooch and Morris.* The effects likely to result simply from the strong acidifica- tion of the solution containing potassium iodide, and their varia- tion for conditions of dilution representing the beginning and the end of a titration on the lines laid down, are shown in the following table. The solution of potassium iodide was diluted as indicated before the addition of the acid, and the iodine set free was titrated by thiosulphate. Effect of Concentration of Acid and Time of Action upon Potassium Iodide. HC1 (sp. gr. 1.16) taken. cm. 8 KI taken, giro. Total volume. cm. 3 Na 2 S 2 3 added at once. In terms of H 3 3 AsO. grm. Na 2 S 2 O 3 added after 5 minutes. In terms of H 3 3 AsO. grin. Na 2 S 2 O 3 added after stirring 5 minutes. In terms of H 3 O 3 AsO. grm. 2C 2 CO 0.0013 2< 2 7C O . 0004 2C 2 50 0.0035 25 2 75 0.0019 2C 2 co O 0042 25 2 vc O OO2I 50 50 CO 2 2 2 IOO 150 IOO 0.0017 o . 0004 o . 003 ^ 50 2 ISO 0.0019 co 2 IOO O OO3 C. CO 2 ISO o 0014 The concentration of acid and the time before titration are, obviously, the essential factors. The absolute amount of acid present and the stirring seem to make little difference. As to the action of the hydrochloric acid on small amounts of the thiosulphate, there is the evidence of the experiments de- tailed in the following statements, in which I cm. 3 , 2 cm. 3 and 5 * F. A. Gooch and Julia C. Morris, Am. Jour. Sci., [4], x, 151. 298 METHODS IN CHEMICAL ANALYSIS Effect of Concentration of Acid upon Thio sulphate. Iodine to Error of Iodine to Error of HC1 (sp. gr. Volume before Na 2 S 2 O 3 nearly K/IO. In color with- out dilution. . titration without color after diluting titration after 1.16). titration. terms of H 3 O 3 AsO. In terms of H 3 3 AsO. dilution. In terms of H 3 3 AsO. to 75 cm. 3 In terms of H 3 O 3 AsO. dilution. In terms of H 3 3 AsO. cm. 8 cm. 8 cm. 3 grm. grm. grin. grm. grm. 25 26 I 0.0071 0.0062 0.0009 0.0071 O . OOOO 25 50 I 0.0071 0.0071 O . OOOO 0.0071 . OOOO 25* 50 I 0.0071 0.0079 +0.0008 o . 0079 +0.0008 50 2 0.0141 0.0146 +0.0005 0.0146 +0.0005 25* 50 2 0.0141 0.0157 +0.0016 0.0157 +0.0016 25 30 5 0.0353 0.0336 0.0017 0.0374 +0.0024 2 5 50 5 0.0353 0-0359 +0.0006 0-0359 +o . 0006 25* 50 5 0.0353 0.0411 +0.0058 0.0411 +0.0058 In these experiments the acid stood in contact with the thiosulphate 5 minutes before titration. cm. 3 of nearly n/io thiosulphate were exposed to the action of 25 cm. 3 hydrochloric acid (sp. gr. 1. 1 6), without dilution or diluted with an equal volume of water, and titrated with nearly n/io iodine. The condition of acidity when the volume of 50 cm. 3 contains 25 cm. 3 of hydrochloric acid (sp. gr. 1.16) is that of the beginning of titration of Williamson's process. In order that the effect of error due to such action upon the determination of arsenic acid may appear immediately, the thiosulphate and iodine used are expressed in terms of that acid. Time Effect of Acid upon Thiosulphate. HC1 (sp. gr. 1.16). cm. 3 KI. grm. Volume. cm. 8 Na 2 S 2 3 nearly /io. In terms of H 3 O 3 AsO. cm. 1 grm. Iodine in terms of H 3 O.,AsO, at once. grm. Iodine in terms of H 3 3 AsO, after 5 min. grm. Error in terms of H 3 3 AsO. grm. 25 25 25 25 25 25 25 25 25 25 25 25 2 2 2 2 2 2 2 2 2 2 2 2 50 75 50 75 So 75 50 75- 50 75 50 75 I I 2 2 5 5 I I 2 2 5 5 0.0071 0.0071 0.0141 0.0141 0.0353 0-0353 0.0071 0.0071 0.0141 0.0141 0-0353 0.0353 0.0057 0.0071 0.0131 0.0143 O . 03 2 2 0.0357 0.0014 O.OOOO O.OOIO +0.0002 O.OO2I +O.OOO4 -0.0043 O.OO04 O.O025 O.OOO2 O.OO4I +0.0008 . . O.OO28 0.0067 0.0116 0.0139 0.0314 0.0361 ARSENIC, ANTIMONY AND TIN 299 The two sources of error due to the action of hydrochloric acid, the liberation of iodine and the decomposition of the thio- sulphate, naturally tend to neutralize one another, but the completeness of such neutralization must be largely a matter of chance in the varying conditions of actual analysis. The experi- ments of the preceding table, in which n/io thiosulphate, to the amount of I cm. 3 , 2 cm. 3 and 5 cm. 3 , was added to the liquid, 50 cm. 3 and 75 cm. 3 , containing 25 cm. 3 acid, and titrated with iodine at once, and after five minutes, were made to test the effects for the conditions of dilution prevailing at the beginning and at the end of a titration. It is clear that under the conditions covered by the experi- ments of the two preceding tables the decomposition of the thiosulphate is likely to occur in greater or less degree, and that when the acid of sp. gr. 1.16 is not much diluted the products of decomposition are not oxidized by the iodine completely. The latter observation is quite in harmony with the fact that sulphur dioxide bleaches iodine in strong hydrochloric acid only slowly and incompletely. In such cases dilution favors further action of the iodine, but results obtained by titration with iodine in the acid solution diluted with an equal amount of water are unmodified by further dilution. In the following tables are recorded actual determinations of arsenic according to Williamson's process. To each 25 cm. 3 of the arsenate were added I, 2 or 3 grm. of potassium iodide and 25 cm. 3 hydrochloric acid (sp. gr. 1.16). The iodine was bleached Williamson's Procedure. HC1. KI. Volume at beginning of titration. Volume at end of titration. H 2 KAsO 4 in terms of H 3 O 3 AsO. H 3 3 AsO found. Error. cm. 8 grm. cm. 3 cm. 3 grm. grm. grm. 25 2 50 51 O.CX362 o . 0085 +0.0023 25 2 50 52 0.0125 0.0156 +0.0031 25 2 SO 55 0.0312 0.0350 +0.0038 25 2 50 55 0.0624 . 0666 +0.0042 2-5 2 50 73 0.1559 0.1588 +0.0029 25 2 50 73 0-1559 0.1587 +0.0028 25 2 50 73 0.1559 O.I5QI +0.0032 25 2 50 73 O.I5S9 0.1595 +0.0036 25 3 50 73 0.1559 0-1595 +0.0036 25 i 50 73 0.1559 0.1581 +O.OO22 25 2 50 73 0.1559 0.1581 +0.0022 25 2 50 73 0.1559 0.1588 +0.0029 300 METHODS IN CHEMICAL ANALYSIS by nearly decinormal thiosulphate without addition of the starch indicator, which loses all delicacy in the presence of strong acid. The time occupied by each titration was about five minutes. The standards of the arsenate were determined by the vapori- zation process,* the purity of reagents employed in that process having been proved by trying the process in the estimation of a solution of arsenic acid made by oxidizing pure decinormal arse- nious acid by iodine. The range of error in these results is from +0.0023 grm. to +0.0042 grm. with a mean of +0.0031 grm. not very different from what might be expected from the effect of the interaction of the strong hydrochloric acid and the iodide alone. The counter effect due to the decomposition of the thiosulphate is not large, yet it is probably real, as will appear in the sequel. In the following series of determinations, made with new solutions and new standards throughout, the arsenic acid was determined in two ways: (I) The iodine set free by 25 cm. 3 of hydrochloric acid (sp. gr. 1.16) and 3 grm. potassium iodide, the solution having a total volume of 50 cm. 3 at beginning and of 75 cm. 3 at the end, was titrated by sodium thiosulphate; (II) The arsenious acid remaining after the first titration by sodium thiosulphate was titrated, after being neutralized with a.cid potassium carbonate by iodine, in the presence of the starch indicator. Titration of Iodine Liberated and Titration of Ar senile Produced. HjKAs0 4 taken in terms of HsOsAsO. HAAsO found by the thiosulphate. Error. H 3 O 3 AsO found by titration of H 3 O 3 As with iodine. Error. grm. grm grm. grm. grm. 0.1767 0.1708 +0.0031 0.1776 +0-0009 0.1767 o. 1708 +0.0031 0.1777 +O.OOi:o 0.1767 0.179$ +O.OO28 0.1785 +O.OOl8 0.1767 0.179? +0.0026 0.1785 +0.0018 0.1767 0.1701- +0.0027 0.1780 +0.0013 0.1767 0.1798 +0.0031 0.1785 +o.ooc8 The average error of the first operation is 0.0029 grm., not far from that of the previous series; the error of the second operation, the titration of the arsenious acid, amounts on the average to 0.0014 g rm - I* 1 tne second operation the error due * See page 291. ARSENIC, ANTIMONY AND TIN 301 to over use of the thiosulphate by iodine set free outside the main reaction is obviously eliminated. The tetrathionate present after neutralization with acid potassium carbonate is unaffected by iodine, as was found by titrating 25 cm. 3 of n/io iodine mixed with 25 cm. 3 hydrochloric acid (sp. gr. 1.16) by the thio- sulphate, neutralizing with acid potassium carbonate,* adding starch and getting the starch blue with a single drop of n/io iodine. The average error of this process (0.0014), therefore, is probably due to the products of decomposition of the thiosul- phate in the first operation. From the foregoing experiments it is clear that an arbitrary- correction of about 0.0030 grm. must be deducted from the indications of Williamson's process of direct titration by thiosul- phate, made with the greatest care under the conditions men- tioned ; and that a correction varying from one-half that amount (0.0015 rm -) to nothing (according to the amount of arsenious acid present), when the determination is made by iodine after neutralization with acid potassium carbonate. After making these arbitrary corrections in the results of the preceding table,, the individual variations fall within reasonable limits. On the other hand, the vaporization process, in which the arsenate is reduced by boiling with sulphuric acid and potas- sium iodide in the manner described,! gives indications reasonably regular and accurate without the application of an arbitrary correction. The Detection and Approximative Estimation of Minute Quantities of Arsenic in Copper. Sanger's successful application of the Berzelius-Marsh process to the quantitative determination of arsenic in wall papers and fabrics,| by the comparison of test mirrors with standard mirrors carefully prepared under the conditions of the test, opens the way, naturally, to the similar estimation of minute amounts of * It is worthy of note that, as was found, it is not possible to substitute an alkali hydroxide for the carbonate in the early stages of the process of neutral- ization, on account of the decomposing effect of the former reagent upon the tetrathionate. This effect is in proportion to the heating of the solution, but is never wholly absent even when ice is intermixed with the liquid and the greatest care taken to prevent a rise of temperature. t See pages 29 1, 294. J Am. Chem. Jour., xiii, 431. 302 METHODS IN CHEMICAL ANALYSIS arsenic in any substances which may be submitted to the process immediately or after suitable preparation. Gooch and Moseley * have studied the application of Sanger's process to the deter- mination of traces of arsenic in copper. It has been shown by Headden and Sadler f that the presence of copper in the Marsh generator is instrumental in holding back the arsenic. It is obvious, therefore, that means must be em- ployed for the complete removal of the copper from the arsenic before the solution of the latter is put into the reduction flask, and there is no method by which arsenic may be removed from copper easily, and without loss, aside from those methods which depend upon the volatility of arsenious chloride from solution in strong hydrochloric acid. Of such methods, on the score of rapidity in execution, accessibility of pure materials, and com- pactness of apparatus, preference is to be given to that process which is based upon the simultaneous action of strong hydro- chloric acid and potassium bromide upon the salt of arsenic. J To get the copper into condition for the application of the process of separation from arsenic, it is sufficient to dissolve an amount not exceeding I grm. in nitric acid somewhat diluted with water, to add to the solution 2 cm 3 or 3 cm. 3 of strong sulphuric acid, and to evaporate the liquid until fumes of the sulphuric acid are disengaged abundantly. A single treatment of this sort serves to remove the nitric acid so completely that no interfer- ence with the normal action of the Marsh apparatus is apparent in the subsequent operation. The residue after concentration is diluted with water to about 5 cm. 3 and washed into the dis- tillation flask with an amount of the strongest hydrochloric acid (sp. gr. 1. 20) equal to that of the remainder of the liquid. It is desirable that the entire volume of the liquid should not much exceed 10 cm. 3 . The flask, which has a capacity of 40 or 50 cm. 3 , is inclined at an angle of about 45 and joined by means of a pure rubber stopper to a bent pipette which serves as a distilla- tion tube. The lower end of the vertical limb of the pipette dips beneath the surface of about 5 cm. 3 of hydrochloric acid of half-strength contained in a test tube which is cooled and sup- ported by water nearly filling an Erlenmeyer flask. A gram of * F. A. Gooch and H. P. Moseley, Am. Jour. Sci., [3], xlviii, 292. t Am. Chem. Jour., vii, 342. t See page 316. ARSENIC, ANTIMONY AND TIN 303 potassium bromide is introduced, and the distillation (which may be completed in three or four minutes) is pushed nearly to dry- ness. The flask is washed out, another portion of potassium bromide is introduced, and the first distillate is introduced and redistilled as before, excepting that the condensation is this time effected in pure water. This second operation serves merely to hold back traces of copper carried over in the first distillation, but the addition of the potassium bromide in the second distil- lation is quite as necessary as in the first, since the bromine liberated in the process has the effect of reoxidizing the arsenic in the receiver and so making that element nonvolatile under the conditions until the reducing agent is again introduced. The free bromine in the final distillate must be reconverted to hydro- bromic acid before the contents of the receiver may be introduced into the reduction flask, and this effect may be most easily and unobjectionably accomplished by the addition of a little stannous chloride dissolved in hydrochloric acid of half-strength and puri- fied from arsenic by prolonged boiling. Incidentally and simul- taneously the arsenic is reduced to the arsenious form, and, though Sanger has shown that minute amounts of arsenic are completely eliminated from the solution in the reduction flask when that element is introduced in the higher form of oxidation, it is our experience that the rapidity of elimination of the arsenic is so increased by the introduction of the small amount of stan- nous chloride needed to bleach the bromine that the mirror appears in from five to ten minutes and is practically complete in half an hour, especially if the precaution is taken to add a little more stannous chloride, according to Schmidt's suggestion,* after the operation has been in progress about twenty minutes. Schmidt has shown that the addition of stannous chloride to the Marsh apparatus in action not only does not effect the reten- tion of arsenic, as many other metallic salts do, but actually brings about the final evolution in the form of the hydride of that portion of the arsenic which may have been deposited during the process in elementary form upon the zinc. The Sanger apparatus is used in form essentially unchanged; but the zinc in the reserve generator is coated with copper by the action of a solution of copper sulphate, since in this way it is made more sensitive to the action of the dilute sulphuric acid, * Zeit. anorg. Chem., i, 353. 304 METHODS IN CHEMICAL ANALYSIS while the presence of copper (which is of course out of the ques- tion in the reduction flask) can be of no disadvantage in the reserve generator and may even serve a useful end in fixing traces of arsenic if the zinc and acid employed are not absolutely free from that element. In the formation of the mirror, too, it has proved to be an advantage to inclose the portion of the glass tube to be heated in a short thin tube of iron or nickel slightly larger than the glass tube and kept from contact with it except at the ends, which are notched and bent inward. By keeping the outer tube of metal at a low red heat it is possible to diminish the tendency of the arsenic* particularly when the amounts are fairly large, to form a double mirror corresponding to the allo- tropic conditions of the arsenic. It is necessary, moreover, to substitute hydrochloric acid for the sulphuric acid usually em- ployed in the reduction flask; but, though the opinion is current that hydrochloric acid introduces difficulties in the Marsh test, no evidence of the formation of a zinc mirror in the ignition tube or of other unfavorable action due to the use of pure hydrochloric acid has been noted in this operation. It is, of course, obvi- ous that the hydrochloric acid used must be arsenic free. Tests of this process were made with copper-prepared by elec- trolyzing in ammoniacal solution the purest copper sulphate obtainable and stopping the deposition before the solution had become exhausted. This copper, in which no arsenic was found, was dissolved in nitric acid, arsenic in the higher condition of oxidation was added, and the process of the separation of the arsenic from the copper and conversion to the mirror carried out in the manner described. A rsenic in Copper. Copper taken. Arsenic taken. Mirror estimated (by comparison with standard mirror). Error. gnu. mgrm. mgrm. mgrm. None. None. None. None. 0.7 None. None. None. o-5 0.5 0.005 O.OII 0.003 0.013 0.002 + O.002 0-35 0.3 O.O2O 0.030 0.015 0.030 0.005 None. 0.43 0.040 0.035 0.005 0.44 0.050 0.040 o.oio ARSENIC, ANTIMONY AND TIN 305 It is plain from the results given above that the method is capable of detecting sharply minute amounts of arsenic in copper and of effecting the estimation of quantities less than 0.05 mgrm. with some approximation to accuracy. There is, as Sanger has pointed out, a good deal of variation even in standard mirrors made with all possible care and pre- caution, and in the estimation of mirrors containing as much as 0.05 mgrm. of arsenic the uncertainty of comparison as well as the actual variation of the mirror is considerable. When a sample of copper is under test which may contain more than 0.05 mgrm. of arsenic, it is desirable to introduce into the reduction flask the measured solution containing the arsenic gradually and in definite portions, and to judge by the formation of the mirror in an interval of ten minutes after the introduction of a portion of this test solution whether it is wiser to add the entire solution or to estimate the arsenic in the entire solution from that found in an aliquot portion. Results of the analysis of several samples of commercial electrolytic copper are appended. The last two represented, presumably, the very purest electrolytically refined copper obtainable commercially. Arsenic in Commercial Copper. Copper taken, giro. Arsenic found, mgrrn. Percentage of arsenic. Sample A o. 3 O.OI nor the tin added later to effect the reduction of the arsenic, will be precipitated by hydrogen sulphide under the existing conditions of temperature and acidity. The results of experiments are recorded in the accompanying table. Tests with Hydrochloric Acid and Potassium Bromide. Arsenic taken as H 3 As0 4 . grm. Antimony taken as H 3 SbO. grm. Tin taken as SnCl 4 . grm. Precipitation by H 2 S in successive distillates after treatment with SnCl 2 . Precipitation by HjS in the residue dis- solved in water. I None None I-X None Fuint coloration O OOOI f I. Found. ) None O.OOIO 1 II. None. f ( I. Found. ( None O OIOO ( II. None. j j I-II. Found. / None O IOOO | III. None. ( j MIL Found, i Fciint coloration 0.4000 { IV. None. f j I-VI. Found ( Orange I .0000 )VII. None. J j I-X. Found. / precipitation.* Orange O 4000 1 XI. None. J I None precipitation.* Larpe O.OOOI 0.4000 i I-II. Found. ) Larc'e O .OOOI O.OOOI I III. None. J j I. Found. ) Distinct color O OOIO O OOOI 1 II. None. { i I. Found. i Distinct color O.OIOO O.OOOI 1 II. None. ) ( I-II. Found. ) Distinct orange o 4000 1 III. None. J I None O OOOI o 4000 i I-II. Found. I T .ftrat* O.OOOI O.OOOI I III. None. J t I. Found. J Distinct color 1 II. None. ( * Subsequently identified as antimony sulphide by depositing the metal on platinum. * Zeit. anal. Chem., xxix, 192. 318 METHODS IN CHEMICAL ANALYSIS The lodometric Determination of Arsenic and Antimony, and of Associated Copper. The determination of arsenic and antimony in the filtrate from cuprous iodide after the titration of free iodine by sodium thiosulphate has been studied by Heath* who has thus made use of processes described elsewhere f for the reduction of arsenic acid and antimonic acid by the action of potassium iodide and sulphuric acid. The reactions involved in these processes may be expressed by the general symbols, M 2 O 5 + 4 HI = M 2 O 3 + H 2 O + 2 I 2 and M 2 O 3 + 2 I 2 + 2 K 2 O = M 2 O 5 + 4 KI. Trials of the method were made with tartar emetic and potas- sium arsenate. The tartar emetic was dissolved in water and the antimony oxidized to the higher condition by means of standard iodine solution in presence of sodium or potassium bicarbonate, the amount of iodine required being taken as a measure of the amount of antimony used. The solution of antimony thus obtained, or the solution of arsenic taken as potassium arsenate, was acidified and a known volume of a standard solution of copper nitrate was added. The copper was determined iodometrically with precautions recommended else- where, t During the determination of copper, mineral acids must not be present on account of their tendency to bring about reduction of the higher salts of arsenic and antimony by action of the excess of potassium iodide used in throwing out the copper. Such action causes high results on copper and low results on arsenic and antimony. A mixture of acetic acid and potassium iodide reduces the higher salts slowly. Tartaric acid is somewhat irregular in action and tends to cause an interfering precipi- tation of acid potassium tartrate. The action of citric acid is, however, satisfactory if the solution is not allowed to stand. The experimental results show that in the iodometric determination of copper associated with arsenic there must be no delay in titrat- ing the iodine. Antimonic acid is not reduced appreciably, in a reasonable time, under similar conditions. * F. H. Heath, Am. Jour. Sci., [4], xxv, 513. t See pages 29 1, 308. I See pages 118, 123. ARSENIC, ANTIMONY AND TIN 319 In the course of preliminary work the fact developed that tetrathionic acid, which results from titration of free iodine by sodium thiosulphate in the copper determination, makes trouble in the subsequent operation. For, when the solution is boiled, after addition of sulphuric acid, the tetrathionic acid decomposes to give hydrogen sulphide and free sulphur, and sulphides of antimony and arsenic may be precipitated. It was found, how- ever, that if liquid bromine is added to the cold solution in suffi- cient quantity to decompose all the excess of potassium iodide present, and the solution then boiled, there is very little subse- quent trouble on account of tetrathionic acid. The procedure for the determination of copper and arsenic or of copper and antimony may be outlined as follows : To the solution containing the copper and also the arsenic or anti- mony in the higher condition of oxidation, add I grm. to 2 grm. of citric acid. To precipitate amounts of copper not exceeding 0.3 grm. in a volume of 50 cm. 3 , add 3 grm. of potassium iodide; in a volume of 100 cm. 3 , add 5 grm. of potassium iodide. Titrate the free iodine with n/io sodium thiosulphate. The reactions involved in the copper determination are 2 CuSO 4 + 4 KI -> 2 K 2 SO 4 + Cu 2 I 2 + I 2 and I 2 + 2 Na 2 S 2 O 3 = 2 Nal + Na 2 S 4 O 6 . Filter off the cuprous iodide on asbestos. To the filtrate add I cm. 3 of liquid bromine and boil the solution in an Erlenmeyer flask, using a trap to prevent loss by spattering. If, after boiling for a short time and allowing the large amount of free iodine to volatilize, the solution does not become clear, cool it, add a little more bromine (0.5 cm. 3 ) and boil again. When the solution has become clear, concentrate it somewhat (to about 60 cm. 3 ) to expel excess of bromine. Dilute to about 100 cm. 3 , add 2 grm. of potassium iodide and boil to a volume of 50 cm. 3 Cool the solution, bleach the free iodine by adding sulphurous acid,* using starch as indicator. Dilute to 100 cm. 3 , add n/io iodine to color and just bleach by careful addition of dilute sul- phurous acid from a pipette. Neutralize the solution with sodium or potassium bicarbonate and titrate the arsenic or antimony with standard iodine solution in the usual way. * Compare page 295. 320 METHODS IN CHEMICAL ANALYSIS Following are tables showing results obtained by this procedure Copper and Arsenic. KI used. giro. Volume at end of precipi- tation. cm. 3 Cu taken. grm. Cu found. grm. Error inCu. grm. Liquid bromine used. cm. 3 KI used. grm. As taken. grm. As found. grm. Error in As. grm. 3 SO 0.0700 o . 0700 O.OOOO I .0 2 0.1238 0.1231 O.OOO/ 3 50 0.0700 0.0693 O.OOO7 I.O 2 0.1238 0.1231 0.0007 3 50 0.0875 0.0869 O.OOO6 1 .0 2 0.1238 0.1239 +0.0001 4 50 o . 0700 0.0698 O.OOO2 (0.61 (0.4 [ I 0.1238 0.1235 -0.0003 4 45 0.0700 0.0703 +O.O003 i .0 2 0.1238 0.1247 +o . 0009 4 50 0.0700 O.C700 0.0000 i .0 2 0.1238 0.1235 -0.0003 5 70 o. 1400 0.1407 +0.0007 (I.O) (0.5 ) 2 0.1238 0.1233 +o . 0005 5 SO 0.0910 0.0907 0.0003 I.O ( 0.5 J 2 0.1238 0.1239 +O.OOOI 4 So 0.0700 0.0703 +0.0003 1 .0 2 0.1238 0.1237 o.oooi 4 50 0.0875 0.0879 +0.0004 1 .0 2 0.1238 0.1234 0.0004 5 65 o. 1400 O.I4IO +O.OOIO I.O 3 0.1238 0.1239 +O.OOOI 3 30 0.0875 0.0860 0.0015 0.8 2 0.0495 0.0493 O.OOO2 Copper and Antimony. KI used. grm. Volume atendol precipi- tation. cm. 3 Cu taken. grm. Cu found. grm. Error in Cu. grm. Liquid bromine used. cm. 3 KI used. grm. Sb taken. grm. Sb found. grm. Error inSb. grm. 4 85 0.0700 0.0703 +0.0003 (I.O) (0.5 ) 2 0.1417 o. 1421 +0.0004 4 80 0.0700 0.0701 +O.OOOI I.O I 0.5 J 2 0.1727 0.1725 O.OOO2 4 80 0.0875 0.0869 0.0006 (I.O 1 0.5 ( 2 0.1286 o. 1289 +0.0003 5 80 0.0735 0.0739 +0 . 0004 I.O 1 0.5 i 2 0.1641 0.1645 +o . 0004 5 90 0.1050 0.1050 0.0000 I.O 1 0.5 f 2 0.1378 0.1372 0.0006 4 75 0.0875 0.0874 O.OOOI (I.O) {0.5 ( 2 0.1329 0.1326 0.0003 6 "5 o . 0700 0.0707 +o . 0007 1.3 (0.5 f 2 0.2474 0.2477 +0.0003 8 1 20 0-1575 0.1571 0.0004 Ull 2 0.1419 0.1413 0.0006 From the results obtained it seems possible by this method to separate and determine copper and arsenic, or copper and antimony, with errors of only a few tenths of a milligram. It is also possible to determine the sum of arsenic and antimony ARSENIC, ANTIMONY AND TIN 321 5 CO ^0 00 CO NO 10 M 00 CN o o o o o 8 gji; g o o o o o o M*^* 4 - M o o o o o o o O o .2 1 1 1 1 1 1 1 1 I 1 & - O ^" *> ON O 8 CO a NO 8- w^'d g O NO t^ ^ O ON NO NO CN1 to s to W ^" ^" CO * 10 Tj" * =|||;g 05 to O to cs O CN cs CN l J tf-g . W W w 2 S 2 8 corrected, average. cm. 3 cm. 3 gnu. grm. grin. gnu. 12-5 25 2 0.1223 0.1218 "] 12.5 25 2 0.1223 0.1218 1 O. I2IQ 12.5 25 2 0.1225 O.I22O j 12.5 25 2 O.I22Q 0.1220 J 374 METHODS IN CHEMICAL ANALYSIS Method of Marshall * is authority for the statement that per- Mondoifo. sulphates liberate iodine from potassium iodide, and that the action is hastened by heat and affected little by the addition of dilute sulphuric acid. Upon this reaction Mondolfo f has based a method for the estimation of persulphates, which consists in heating a mixture containing the persulphate and potassium iodide in a stoppered bottle for ten minutes at 60 to 80, and titrating by thiosulphate the iodine set free. According to Peters and Moody, J however, the reduction of persulphate is incomplete even when the volume of the mixture is restricted to 25 cm. 3 , and the digestion continued thirty minutes with the addition of small amounts of sulphuric acid. Under the best conditions the amount of persulphate found was 0.1207 g rm - in an amount of solution for which the arsenate-iodide process indicated 0.1222 grm., and the LeBlanc and Eckardt process 0.1218 grm. in the average. Method of The process proposed by Namias, without knowl- Namias. e( j ge o f Mondolfo's method differs from the method of Mondolfo in the particular that the reaction is carried out at the ordinary temperature. The mixture of persulphate and potassium iodide in solution is allowed to stand eleven hours in a stoppered bottle and the iodine set free is titrated with thio- sulphate. In experiments conducted by Peters and Moody the reduction of the persulphate under the conditions described was incomplete, the color of iodine returning upon longer standing after the titration. Under the best conditions the value of 0.1208 grm. was found for the persulphate contained in an amount of solution for which the arsenate-iodide method indi- cated o.i 222 grm. and the LeBlanc and Eckardt process 0.1218 grm. in the average. Comparison To compare the values obtained for the persul- of Methods. phate solution the averages of the results obtained by the different methods, together with the average of all the experiments, are given in the table. The process of Mondolfo and the process of Namias, both of which involve the liberation of iodine from potassium iodide and * Jour. Chem. Soc. 59, 771. f Chem. Ztg., 23, 699. t Loc. cit. Loc. cit. SELENIUM 375 the titration of that iodine by thiosulphate, give results which are practically identical and lower* than those obtained by the other three methods. Process. Number of experiments. Average of results. ' grm. Mondolfo 6 O.I2O7 Namias 8 o 1208 LeBlanc and Eckardt 12 o 1217 Griitzner (corrected) 4 O. I2IQ Arsenate-iodide method ... 4 0. 1222 The process of LeBlanc and Eckardt in which the persulphate is reduced by a ferrous salt, the process of Griitzner, in which an arsenite solution is the reducing agent, and the arsenate- iodide method, in which the persulphate is determined by the difference between the amount of iodine in an iodide added and the amount necessary to oxidize the arsenite remaining after boiling the solution, are all in close agreement and all higher than those obtained by the process of Namias or that of Mondolfo. The process of LeBlanc and Eckardt is simple, rapid and con- venient. The method of Griitzner is advantageous in that the ordinary arsenite solution is the standard for the process, though requiring the application of a correction. The arsenate-iodide method, introduced as a control, is accurate but less simple than the other methods. SELENIUM. The Gravimetric Estimation of Selenious Acid by Liberation of Iodine and Absorption of that Element by Silver. When selenious acid is treated in acidulated solution with potassium iodide and shaken with silver in an atmosphere of hydrogen,* the reaction proceeds to the complete reduction of the element, according to the equation SeO 2 + 4 KI + 4 HC1 = 4 KC1 + 2 H 2 O + Se + 2 I 2 . The increase in weight of the insoluble material represents the iodine evolved plus the selenium of the selenium oxide. f Tests * See page 444. f Claude C. Perkins, Am. Jour. Sci., [4], xxix, 338. 376 METHODS IN CHEMICAL ANALYSIS made in the application of this process to selenium dioxide twice crystallized from nitric acid and resublimed over manganese dioxide are given in the table. Weighing of Silver Iodide and Selenium. Ag taken. Se taken. Increase. Calculated Se. Error. grrn. grin. grm. grm. grm. 2.0133 O . 0050 0.0365 o . 0049 O.OOOI 2.0133 2 - 5639 2.5639 3.0018 0.0075 0.0126 o . 0428 0.0504 0.0529 0.0894 0.3178 0.3799 O 0071 O.OI2I 0.0429 0.0501 O.0004 0.0005 + 0.0001 -0.0003 The Gravimetric Determination of Selenious Acid by Precipitation of Selenium. For the gravimetric determination of selenious acid it is usual to precipitate the selenium with sulphurous acid in presence of hydrochloric acid and to weigh the elementary selenium. Pre- cipitation by this method, however, is slow and incomplete in many cases, so that it is always necessary to treat the nitrate a second time with sulphurous acid and to digest for some time. Adopting the idea from volumetric methods for the determina- tion of selenium * in which an iodide in acid solution is used to reduce the selenious acid, Peirce f effects the reduction by the use of potassium iodide in large excess. When potassium iodide is added in slight excess to the solution of selenious acid acidified with hydrochloric acid the selenium is precipitated in the form of a red powder. Boiling for ten minutes removes most of the liberated iodine and changes the selenium into the black modification. This may be collected upon an asbestos felt, washed, dried at 100 to a constant weight, and weighed. If the selenium amounts to less than o.i grm. the results accord well with theory. When the amount is larger, the selenium is apt to assume on boiling a pasty, molten condi- tion which makes filtering and washing impossible. In this condition the selenium holds iodine. By using the potassium iodide in large excess, however, the pasty condition may be modified, and by effecting the reduction in large volumes of solu- tion the danger of inclusion is lessened. * See pages 377, 379. f A. W. Peirce, Am. Jour. Sci., [4], i, 416. SELENIUM 377 According to this procedure, it is sufficient to dilute the solu- tion containing selenious acid or a selenite to 400 cm. 3 before acidifying with hydrochloric acid and then to add potassium iodide to an amount about three grams in excess of that actually required. Boiling from ten to twenty minutes will change the selenium to the black modification and remove most of the iodine. The process of precipitation and filtering can be completed in half an hour. The selenium is dried at 100 to a constant weight. When the selenium occurs in the higher form of oxidation the reduction follows the same course, though iodine is not liberated until the solution is quite warm ; but at the end of the usual time of boiling the action is complete. Results obtained by this procedure are given in the table. Reduction of Selenium. Se taken as SeO 2 . grm. Se found, grrn. KI. grm. Volume. cm. 3 Error, grm. 0.2853 0.2861 7 QOO +0.0008 0.3189 0.3192 8 400 +0.0003 0.3318 0.3324 7 500 +0.0006 0.3798 0.3805 7 500 +0.0007 0.4252 0.4259 7 350 +0.0007 0.4430 0-4434 10 450 +0.0004 Se taken as SeO 3 . 0.1063 0.1065 5 500 +0.0002 0.1063 0.1062 5 375 O.OOOI O.2OIO 0.2017 5 350 +0.0007 0.3H5 0.3126 6 500 + 0.001 1 The lodometric Determination of Selenious Acid by Methods Based upon the Action of Potassium Iodide in Presence of Acid. The contact According to the method of Muthman and Schae- Method. f er ^* wn en selenious acid is brought into contact with potassium iodide in an acidulated solution, iodine and selenium are liberated in elementary condition, the former being directly determinable by titration with sodium thiosulphate after addi- tion of starch. On account of the difficulty in determining the exact point in the titration at which the starch blue disappears * Ber. Dtsch. chem. Ges., xxvi, 1008. 378 METHODS IN CHEMICAL ANALYSIS from the liquid in which the finely divided and opalescent selenium is held in suspension, the process was recommended by the authors for use only when great accuracy is not es- sential. Evidently if the reaction between the acidulated iodide and selenious acid is simple and complete, the process should be capable of improvement by removing the selenium before the titration is attempted. This has been done by Gooch and Reynolds * by filtration by means of the vacuum pump upon a thick felt of asbestos in a perforated crucible or cone of large filtering surface. With a properly prepared filter of this descrip- tion there is no difficulty in separating the selenium in a very few moments so completely that it is possible to determine the iodide remaining dissolved in the excess of potassium iodide with all the accuracy characteristic of this most exact of titration processes. When, however, the difficulty of determining the end-reaction in the titration of the iodine by the thiosulphate is overcome, it becomes apparent that the reaction itself is likely to be incomplete. Even when the potassium iodide is used in moderate excess in presence of a large proportion of hydrochloric acid, and with addition of the thiosulphate previous to the filtra- tion in order that there may be every opportunity for the iodine and thiosulphate to interact, the results show marked deficiency in the reduction of the selenious acid. Either the reaction according to the equation SeO 2 + 4 HI = Se + 2 H 2 O + 2l 2 is incomplete or else there is formed between the selenium and iodine some combination, such as was noted by Hautefeuille f in the interaction between iodine and hydrogen selenide. In further study of this reaction, and following the suggestion of Peirce's experience in the gravimetric determination of sele- nium, Norton t has found that the accuracy of the process is very much increased for small amounts of selenium by the use of relatively very large amounts of potassium iodide. The process is still inaccurate when large amounts of selenium are present. This is shown by the results of the accompanying table. * F. A. Gooch and W. G. Reynolds, Am. Jour. Sci., [3], 1, 255. f Compt. rend., Ixviii, 1554. J J. T. Norton, Jr., Am. Jour. Sci., [4], vii, 292. SELENIUM 379 Contact Method. SeO 2 used. KI. Volume of solution. HC1 sp. gr. 1. 12. SeO 2 found. Error. grm. grm. cm. 3 cm. 3 grm. grm. 0.0553 IO 150 IO 0.0558 +0.0005 0.0574 5 ISO IO 0.0567 0.0007 0.0683 5 150 IO 0.0683 o.oooo 0.0487 5 ISO 10 o . 0484 0.0003 0.2617 10 150 10 0.2589 0.0028 The Distillation The reaction between selenious acid, potassium Method. iodide, and hydrochloric acid, in the sense of the equation given above, may be pushed further toward completion by submitting the mixture to distillation. For this purpose Gooch and Reynolds * make use of the apparatus previously described and figured, f The distillation flask is a Voit gas- wash- ing flask, and this is sealed to the inlet tube of a Drexel wash- bottle used as a receiver, to the outlet tube of which is sealed a Will and Varrentrapp absorption apparatus to serve as a trap. The mixture to be distilled, containing not more than 0.2 grm. of selenium dioxide, is introduced into the flask, a solution of 3 grm. of potassium iodide in 100 cm. 3 of water is put into the receiver and trap, and during the distillation a slow current of carbon dioxide is passed through the apparatus to keep the boil- ing regular. Naturally, the acidified solution of the iodide in the flask retains with great tenacity traces of dissolved iodine, so that, in order to determine all the iodine liberated in the reac- tion, the residue in the flask as well as the distillate in the receiver and trap must be titrated with sodium thiosulphate. Results are fairly good, though a little deficient, for amounts of selenium dioxide up to 0.2 grm.; but when the amount of the dioxide reaches 0.5 grm. it is found that the sum total of iodine in the ^distillate and in solution in the residue falls far below the theory based upon the assumption that the products are selenium, iodine, and water. The details of treatment and the results for the smaller amounts of selenium afe recorded in the table. The selenium in the residue is, for the smaller amounts, left, after the boiling, in fine dense crystalline condition, so that it does not interfere with the titration of the free iodine; but for the larger amount, 0.5 grm., it is in pasty form adhering to the flask. * Am. Jour. Sci., [3], i, 256. f See Fig. 3, page 4. METHODS IN CHEMICAL ANALYSIS Examination proves that the selenium holds iodine, which is liberated slowly to water and more rapidly to an aqueous solution of potassium iodide; but the error thus introduced is allowable up to the limit of 0.2 grm. of selenium dioxide. Distillation Method. SeO 2 taken. KI in flask. HClin flask (sp. gr. 1.20). Total volume boiled. Time in minutes. SeO 2 found. Error. gnu. grin. cm. 3 cm. 1 grm. grm. 0.0499 i 5 60 5 0.0497 O.OOO2 0.0499 i 5 60 5 0.0497 O.OO02 0.0499 i 5 60 10 o . 0496 O.OO03 O . 2OOO 3 5 60 10 0.1995 0.0005 O.2OOO 3 5 60 10 0.1991 0.0009 0.2023 3 5 60 IO 0.2018 0.0005 Method: Treat- ment of the Residue. Peirce states * that the range of the process may be much ex- tended by the use of very large amounts of potassium iodide. In this case, however, more iodine is retained in the residue and the difficulty resulting from atmospheric action when the acidu- lated residue is exposed becomes magnified. Differential When a solution of an arsenate, potassium iodide, and sulphuric acid is boiled under defined conditions! arsenic acid is reduced to arsenious acid with liber- ation of iodine. When the arsenic acid is in excess the whole of the iodine is evolved and the arsenious acid produced is its exact measure. Upon making the solution alkaline with acid potassium carbonate, the arsenious acid may be reoxidized by standard iodine, and the amount of iodine thus used is the exact equivalent of that set free in the reduction process. Gooch and Peirce J have shown that when selenious acid is present during the reac- tion between arsenic acid and the iodide, selenium is reduced, and the subsequent estimation of arsenious acid in the residue will be less than should be produced by the iodide by an amount equivalent to the selenious acid present, the reduction taking place according to the equation SeO 2 + 4HI = Se + 2 H 2 O + 2 I 2 . * See page ^76. t See page 457. J F. A. Gooch and A. W. Peirce, Am. Jour. Sci., [4], i, 31. SELENIUM 381 According to the procedure worked out, the selenious acid to be determined is put into an Erlenmeyer flask of 300 cm. 3 capacity; a known amount of standardized potassium iodide (somewhat in excess of that theoretically required) is added ; and a solution containing about 2 grm. of pure di-hydrogen potassium arsenate with 20 cm. 3 of sulphuric acid of half-strength is introduced. During the boiling the mixture is protected from ordinary me- chanical loss by a trap * (consisting of a two-bulbed drying tube, cut short and hung loosely with the wide end downward in the mouth of the flask) and by the introduction of a few bits of por- celain. The liquid is boiled until the volume decreases, accord- ing to indicating marks on the flask, from 100 cm. 3 or more to 35 cm. 3 , concentration to about this lower limit having been found to be necessary for the completion of the reaction. The residue is cooled, the acid is nearly neutralized with potassium hydroxide, acid potassium carbonate is added until it is present to the amount of 20 cm. 3 of its saturated solution in excess of the quantity needed for complete neutralization, and, after the addition of starch, standard iodine is introduced until the starch-blue appears. Differential Method. Initial volume. Final volume. H 2 SO 4 half- strength. Di-hydro- gen- potassium arsenate. KI taken. 800, taken. Se0 2 found. sSfSOff cm.* cm. 3 cm. 8 grm. grm. grin. grm. grm. IOO 35 20 2 1.3277 o. 1280 0.1275 0.0005 IOO 35 20 2 1.0429 O . 0998 . 0994 0.0004 IOO 35 20 2 1.0887 O. 1024 o. 1028 +0.0004 IOO 35 20 2 1.0405 o. 1036 0.1028 O.OOO& IOO 35 2O 2 1.0721 o. 1030 o. 1029 o.oooi IOO 35 2O 2 0.9958 0.1273 o. 1272 O.OOOI 125 35 2O 2 2.0828 0.1997 O . 2OOO +o . 0003 125 35 2O 2 2.2272 O.2IIO 0.2II3 +0.0003. 125 35 2O 2 2.1535 O.2O67 0.2069 + O.OOO2 150 40 2O 2 2-6554 0.2560 0.2549 o.ooii 175 35 20 2 3.2428 0.3IIO 0.3II8 +0.0008 175 35 2O 2 3.2428 0.3085 0.3083 0.0002 The iodine introduced measures the arsenious acid (and so the quantity of iodine set free by the arsenic acid), and the difference between it and the iodine originally present in the * See Fig. 6, page 6. 382 METHODS IN CHEMICAL ANALYSIS form of the iodide represents the amount set free by the seleni- ous acid. The preceding table comprises the details and results of a series of determinations made in the manner outlined. The Determination of Selenious Acid by Potassium Permanganate'. In the action of potassium permanganate upon selenious acid, whether in a solution acidified with sulphuric acid or made alka- line by caustic soda, the reduction of the permanganate does not proceed to the lowest degree of oxidation of the manganese, the selenious acid being unable to reduce the higher hydroxides which separate. When the permanganate is introduced into the acid- ified solution the color vanishes, leaving a clear colorless liquid, but as more is added the solution becomes yellow and deepens gradually in color to a reddish brown, until turbidity due to the deposition of a brown hydroxide of manganese ensues, and finally the characteristic color of the permanganate is plainly distin- guishable. The exact point at which precipitation of the man- ganic hydroxide begins depends upon the dilution, acidity, and temperature of the solution. In employing the reaction quanti- tatively it is necessary, according to Gooch and demons,* to add the permanganate in distinct excess, to destroy the surplus by means of standard oxalic acid added to the solution acidified with sulphuric acid, and then to determine the excess of oxalic acid in the warmed solution by addition of more permanganate. The difference between the amount of permanganate actually used and that required to oxidize the known amount of oxalic acid introduced is the measure of the selenious acid acted upon, pro- vided the amount of sulphuric acid present in the final titration and the temperature are adjusted, to prevent on the one hand interference with the end-reaction by precipitation of manganese hydroxide, according to Guyard's reaction, and on the other hand to obviate evolution of oxygen outside the main reaction.! According to the method of treatment prescribed, the solution of selenium dioxide in 100 cm. 3 of water containing 10 cm. 3 of sul- phuric acid of half-strength is heated to 75, an approximately decinormal standardized solution of potassium permanganate is added until the characteristic color predominates over that of the * F. A. Gooch and C. F. demons, Am. Jour. Sci., [3], 1, 51. t See page 47. SELENIUM 383 brown hydroxide deposited during the oxidation, oxalic acid in solution of known strength is introduced until the excess of per- manganate has been destroyed and the insoluble hydroxide dis- solved, and, finally, at a temperature of 50 or less, permanganate is added to coloration. The final volume varies between 250 cm. 3 and 350 cm. 3 , and the sulphuric acid (absolute) between about five per cent at the start and one and a half or two per cent at the end. The determination of large amounts of selenious acid by this method is somewhat less advantageous than would be the case if the reduction of the permanganate proceeded further in the first action. One hundred cubic centimeters of a standard solu- tion is as much as can be conveniently handled in a single process of titration, and that volume of decinormal permanganate (which is about as strong as the standard solution should be when accu- rate work is expected) is capable of oxidizing about 0.25 grm. of selenium dioxide. In the table are given the results of practical tests of this method. Permanganate Oxidation. Oxygen Oxygen SeO 2 taken. equivalent to permanganate equivalent to oxalic acid SeO 2 found. Error. used. used. grm. grm. grm. grm. grm. O.IOOO 0.03506 0.02065 O.IOOI +O . OOOI O.IOOO 0.03519 0.02073 0.1004 +O.OOO4 O. IOOO 0.03706 0.02255 0.1007 -f-o . 0007 O.IOOO 0.03853 0.02422 0.0994 O.OOO6 O.IOOO 0.03512 0.02065 o. 1005 +0.0005 O. 2OOO 0.06124 0.03256 0.1994 O.OOO6 O. 2OII o . 06069 0.03177 o . 2008 0.0003 0.2004 0.06072 0.03177 O. 2OIO +0.0006 O.2O2O o . 06083 0.03185 O.2OI2 0.0008 0.2038 0.06106 0.03185 0.2028 o.ooio The Determination of Selenious Acid by the Direct Action of Sodium Thiosulphate, According to the Method of Norris and Fay. In the method of Norris and Fay * for the iodometric deter- mination of selenious acid, advantage is taken of a direct and unique action of sodium thiosulphate upon selenium dioxide in the presence of hydrochloric acid, four molecules of sodium * Am. Chem. Jour., xviii, 703. METHODS IN CHEMICAL ANALYSIS thiosulphate acting upon one molecule of selenious acid.* The method, which consists in treating the solution of selenious acid in ice water, in the presence of hydrochloric acid, with an excess of a n/io solution of sodium thiosulphate and titrating back the excess of the thiosulphate with iodine, involves the addi- tion of an excess of the thiosulphate to the solution of selenious and hydrochloric acids, and thus establishes conditions which demand care as to the relation of the acid, thiosulphate, degree of dilution, and temperature. Norton f shows that with precautions noted the process of Norris and Fay is simple, rapid and accurate; without ttfem, as the experimental results given below indicate, errors of consider- able amount may enter. Reduction by Thiosulphate and Titration of Excess. Amount of SeO 2 taken. gnu. HC1 (sp. gr. 1. 12). cm.* Excess of Na 2 S 2 3 . cm. s SeO 2 taken, grm. Error. gnn. Volume at beginning, 200 cm. 3 0.1042 5 24.16 o. 1041 o.oooi 0.0611 10 13-3 0.0611 0.0000 0.0850 10 21.9 0.0828 0.0022 0.0757 25 13.07 0.0749 0.0008 0.0540 25 21 .02 0.0522 0.0018 Volume at beginning, 400 cm. 3 0.0616 IO 2.28 0.0625 +0.0009 0.0628 10 7. II 0.0631 +0.0003 o . 0508 10 11.4 0.0511 +0.0003 0.0587 10 12.8 0.0594 +0.0007 0.0807 IO 15.3 0.0813 +0.0006 0.0633 IO 20.85 o . 0638 +0.0005 0.0682 25 I. II 0.0685 +0.0003 0.0779 25 1-35 0.0788 +0.0009 0.0465 25 18.93 o . 0469 +o . 0004 It is recommended to so adjust conditions that no more than 20 cm. 3 of n/io thiosulphate shall be present in excess. If this limit be placed upon the thiosulphate, 5 cm. 3 of hydrochloric acid (sp. gr. 1.12) may safely be present in a volume of 200 cm. 3 at the beginning, or 10 cm. 3 of the acid in a volume of 400 cm. a The presence of 5 cm. 3 of the acid in 400 cm. 3 of solution is really sufficient to bring about the reaction. * The complete reaction is not stated. t J. T. Norton, Jr., Am. Jour. Sci., [4], vii, 287. SELENIUM 3 8 5 The I odometric. Determination of Selenic Acid by the Action of the Halogen Acids. Reduction by It has long been known that selenic acid is re- Add'^hDis- diicible by hydrochloric acid with evolution of chlo- tiiiation. rine, but the reaction was regarded as more or less uncertain until Petterson showed * that conditions of action may be secured under which the reduction proceeds regularly accord- ing to the equation SeO 3 + 2 HC1 = SeO 2 + H 2 O + C1 2 . The chlorine evolved may be estimated iodometrically and taken as the measure of the selenic acid originally present or of the selenious acid produced. According to this method of determination, it is only necessary to boil a solution of selenic acid in hydrochloric acid of moderate concentration, and if the solution is not too dilute the reduction is obtained in a few moments. Gooch and Evans | have determined the limits within which a successful determination of the selenic acid may be expected. It is shown that so long as the volume of the hydrochloric acid, sp. gr. 1. 20, does not amount to more than 10 per cent of the entire liquid no chlorine whatever is evolved, and that only when the percentage of this acid rises as high as thirty does the chlorine evolved during boiling for five minutes approach the theoreti- cal yield. Care must be taken, however, not to prolong the boil- ing after the solution reaches a concentration corresponding to hydrochloric acid of half-strength ; for under such conditions easily attained in boiling down mixtures of selenious acid and hydrochloric acid over-reduction may take place and selenium appear visibly in the distillate. Obviously it is advantageous, in attempting the practical reduction of selenic acid, to begin the distillation with acid of strength sufficient to insure the evolution of chlorine in quantity at the outset, and it has been found best to start with a mixture one- third of which is the strongest aqueous hydrochloric acid, sp. gr. 1.20. With solutions so constituted the reduction goes on rapidly. Good results may be expected when the mixture, containing one-third of its volume of the strongest aqueous hydrochloric acid at the beginning, is boiled until the * Zeit. anal. Chem. xii, 287. t F. A. Gooch and P. S. Evans, Jr., Am. Jour. Sci., [3], 1, 400. 3 86 METHODS IN CHEMICAL ANALYSIS chlorine is expelled, care being taken that the volume of the liquid shall not become less than two-thirds of the original volume. The apparatus * made by sealing to the outlet tube of a Voit wash bottle (used as a retort) to the inlet tube of a Drexel wash bottle (charged with potassium iodide and used as a receiver) with a set of Will and Varrentrapp bulbs (sealed to the receiver, to serve as a trap) is convenient for the operation. A current of carbon dioxide aids in carrying the chlorine to the receiver and in pro- moting quiet boiling. From solutions having a total volume of 75 cm. 3 at the outset and containing 25 cm. 3 of the strongest aqueous hydrochloric acid (sp. gr. 1. 20), the entire amount of chlorine corresponding to the reduction of 0.2 grm. of selenic acid to selenious acid is liberated in ten minutes. The iodine in the receiver is estimated by thio- sulphate. The details of experiments made under this procedure with selenic acid obtained by oxidizing with permanganate pure selenium dioxidef are given in the table. Reduction by Hydrochloric Acid. SeQa taken. Total volume at the outset. HCl (sp. gr. 1.20) present. Time in minutes. SeO 3 found. Error. cm. 3 cm. 8 grm. 0.0572 75 25 IO 0.0568 0.0004 o 0572 75 25 10 0.0569 -0.0003 o. 1144 75 25 . 10 0.1143 o.oooi o. 1144 75 25 10 0.1137 0.0007 0.1144 75 25 IO 0.1147 +0.0003 0.2288 75 25 IO 0.2233 0.0005 0.2288 75 25 IO 0.2279 0.0009 Reduction by When acted upon by sulphuric acid and potassium Hydrobromic bromide in solution, selenic acid liberates bromine Acid, witn Distillation. in proportion to the excess of acid, the amount of bromide, and the temperature. SeO 3 + 2 HBr = SeO 2 -f H 2 O + Br 2 . When such a solution is boiled the bromine is evolved and may be collected in potassium iodide, and the iodine thus set free may be determined by standard sodium thiosulphate and taken as the measure of the bromine distilled. * See Fig. 3, page 4. t See page 382. SELENIUM 387 Gooch and Scoville * have shown that the applicability of the reaction to quantitative purposes turns upon the adjustment of the proportions of the reagents used. The apparatus shown in Fig. 3 t is convenient for the distillation process. When the proportions of sulphuric acid, potassium bromide, and selenic acid are favorable, the bromine liberated is removed rapidly to the distillate, leaving the residue perfectly colorless, but as the distillation is continued the liquid residue again takes on color and more iodine is set free by the action of the distillate upon potassium iodide, while selenium is plainly visible in the receiver. When the amount of potassium bromide is large, its effect is to retain bromine in the liquid so obstinately that no period of colorlessness intervenes before the second stage of color arrives; when its amount is small, while that of the sulphuric acid is also small, the reduction of the selenic acid and the evolu- tion of the bromine progress slowly ; and the interval of colorless- ness is prolonged when the amount of bromide is small, while that of the acid is comparatively large. The proportions found best in handling 0.25 grm. of selenic acid, or less, are an initial volume of 60 cm. 3 containing 20 cm. 3 of sulphuric acid of half-strength, with I grm. of potassium bromide. Under these conditions it is found that the reduction is almost theoretically exact when the distillation is continued until the recoloration of the boiling liquid is distinctly recognizable ; and this point corresponds in practice very closely to a concentration of volume to 35 cm. 3 . In the fol- lowing table are given the results of experiments made under these conditions of action. Reduction by Hydrobromic Acid. SeO 3 taken as H 2 SeO*. HoSO 4 of half- strength. KBr taken. Initial volume. Final volume. SeO 3 calculated. Error. grm. cm. 8 grm. cm." cm. grm. grm. o . 0590 2O 60 35 0.0588 0.0002 o . 0590 20 60 35 0.0591 -f-o.oooi 0.0614 20 60 35 0.0616 +0.0002 0.0614 20 60 35 0.0607 0.0007 o. 1180 20 60 35 0.1177 0.0003 0.1180 20 60 35 o. ii 80 o . oooo 0-1534 20 60 35 0.1527 0.0007 0.2349 20 60 35 0.2350 +O.OOOI * F. A. Gooch and W. S. Scoville, Am. Jour. Sci., [3], 1, 402. f See page 4. 3 88 METHODS IN CHEMICAL ANALYSIS Reduction by The determination of selenic acid by acting with hydrochloric acid and potassium iodide and estimat- with Distillation. mg ^ {od{ne liberated has been stu died by Gooch and Reynolds.* While the simple contact of selenic acid and potassium iodide in solution acidified with hydrochloric acid does not produce a regular liberation of iodine, it is possible by sub- mitting such mixtures to distillation, when the amounts of selenic acid present are not too large, to bring about a definite reaction in which the products are selenium, water and iodine, according to the equation Se0 3 + 6HI =Se + 3H 2 + 3 I 2 . In applying this reaction to analytical purposes it is convenient to make use of an apparatus and procedure previously described f and to treat, preferably, not more than 0.2 grm. of the selenic oxide with I grm. to 3 grm. of potassium iodide, 5 cm. 3 of con- centrated hydrochloric acid in a total volume of 60 cm. 3 , and to continue the boiling for ten minutes. The results of experiments made in this manner with selenic acid obtained in solution by oxidizing known amounts of selenium dioxide by potassium permanganate J are given in the table. Reduction by Hydriodic Acid. SeO 3 taken. Klin flask. HClin flask (sp. gr. 1.20). Total volume boiled. Time in minutes. SeO 3 found. Error. grm. grm. cm. 3 cm .3 grm. grm. 0-0593 i 5 60 5 0.0593 0.0000 0-0593 i 5 60 5 0.0591 0.0002 0-0593 3 5 60 10 0.0596 +o . 0003 0.1779 3 5 60 10 o. 1769 O.OOIO 0.1779 3 5 60 IO o. 1780 +0.0001 0.1779 3 5 60 10 o. 1764 -0.0015 Reduction by In a mixture made up of a selenate, an arsenate, Di^rentLf"* 1 ' potassium iodide, and sulphuric acid, the arsenic acid Method. attacks the hydriodic acid before all of the selenic acid is reduced. In order to apply to selenic acid the differential method of determination, the selenic acid must first be reduced to * F. A. Gooch and W. G. Reynolds, Am. Jour. Sci., [3], 1, 258. t See page 379. J See page 382. SELENIUM 389 the condition of selenious acid. Ordinarily, the simplest mode of reducing selenic acid is by boiling it in solution with hydrochloric acid of definite strength,* but in this case the presence of hydro- chloric acid is precluded on account of the consequent volatiliza- tion of arsenious chloride during the process of concentration in the subsequent treatment with the iodide. It is possible, how- ever, to make use of reduction by hydrobromic acid, since arse- nious bromide is not appreciably volatile under the conditions. Gooch and Peircef have shown that the determination of selenic acid may therefore be accomplished by first reducing it to selenious acid by the bromide process and then treating the residue by the differential method for the determination of selenious acid.J Differential Method. SeO 2 taken as H 2 SeO 4 . grm. KI used in second reduction. grm. SeO 2 found, grm. Error, grm. 0.0378 O . 6306 0.0380 +0.0003 0.0378 0.5643 0.0374 0.0004 0.0516 0.7136 0.0517 +O.OOOI o . 0503 0.7302 0.0508 +o . 0005 0.0541 0.6671 0.0544 -fo . 0003 o. 1007 3277 O.IOII +O.OOO4 o. 1008 .3277 O.IOII +0.0003 0.1007 .2082 0.1005 O.OOO2 o. 1007 .1684 0.1016 +o . 0009 o. 1007 .0522 0.0999 O.OOOS 0.1009 .2679 o. 1005 0.0004 0.1031 1 .1119 0.1032 +0.0001 0.1870 I .8720 0.1879 +o . 0009 0.2014 1-99*5 O.2O2O +0.0006 o. 2016 2.0745 0.2025 +0.0009 0.2059 1.8687 o . 2064 +o . 0005 The procedure is as follows: The solution containing the selenate to be determined is put in an Erlenmeyer flask of 300 cm. 3 capacity with I grm. of potassium bromide and sulphuric acid amounting to 20 cm. 3 of the acid of half-strength. The so- lution amounting to 60 cm. 3 or 100 cm. 3 is boiled until the color- less solution left when the bromine vanishes begins to color again. Experience shows that the reappearance of the brownish color is very easily seen and that it is not safe to conclude that the free * See page 385. f F. A. Gooch and A. W. Peirce, Am. Jour. Sci., [4], i, 33. J See page 380. 390 METHODS IN CHEMICAL ANALYSIS bromine has been eliminated, under the conditions of dilution and proportion, until this stage of concentration which corresponds to a volume of about 35 cm. 3 has been reached ; but the dis- tillation should not be pushed beyond the point at which the returning color is noted. When this condition has been reached the solution is cooled and treated exactly in the manner described for the reduction of selenious acid. The neutralization by acid potassium carbonate, after the final boiling, generally occasions the precipitation of manganous carbonate, but the precipitate does not interfere in the slightest with the titration which follows. The preceding table comprises determinations made to test the accuracy of the iodometric determination of selenic acid by the combined processes of reduction. The Separation of Selenium from Tellurium by Procedure Based upon the Difference in Volatility of the Bromides. When small amounts of selenic acid are boiled in aqueous solu- tion with potassium iodide and hydrochloric acid, selenium is precipitated, while the iodine set free simultaneously may be estimated in the distillate and residue, and taken as the measure of the selenic acid originally present.* If the iodide is omitted from the mixture, so that the hydrochloric acid alone shall be the reducer, the reduction proceeds only to the point of formation of selenious acid, provided the boiling is not continued after the hydrochloric acid has reached the condition of half-strength at which it boils unchanged under normal atmospheric pressure. A solution of selenic acid, potassium bromide, and sulphuric acid, of regulated dilution and proportions, also yields under de- fined conditions selenious acid as the product of reduction. When, however, the ebullition of a solution of selenious acid in hydrochloric acid is continued after the acid has reached the condition of half -strength, traces of selenium appear in the receiver and connecting tubes, the distillate sets free iodine from potas- sium iodide, and it is evident that the selenious acid is under- going further reduction ; and the same effects are produced when the boiling of the mixture of sulphuric acid, potassium bromide, and selenious acid is pressed beyond the point at which the solu- tion begins to be colored. Obviously, under certain conditions * See page 388. SELENIUM 391 of concentration, selenium tetrachloride and selenium tetrabro- mide, respectively, are forming from the acid; and the appear- ance of the elementary selenium is due to partial decomposition of the halogen salts. Phenomena of a similar nature are seen when an aqueous solution of selenious acid, phosphoric acid, and sodium chloride is submitted to distillation : that is to say, there comes a time in the process of boiling such mixtures when the appearance of elementary selenium and the action of the distil- late upon potassium iodide make evident the volatilization and partial decomposition of the selenium compounds of the halo- gens, and the further continuance of the treatment results in the more or less complete removal of the selenium compounds to the distillate. From the mixture containing the phosphoric acid, selenious acid and sodium chloride only a partial volatilization of the selenium chloride takes place. The volatilization of selenium bromide, however, produced by the reaction between phosphoric acid, selenious acid and potassium bromide, may be made complete. Upon this reaction Gooch and Peirce * have based a process for the separation of selenium and tellurium, taking advantage of the volatility of selenium tetrabromide and the non-volatility of tellurium tetrabromide under definite conditions. The distillation apparatus used in the process of separation is shown in Fig. 4-f It consists of two Voit flasks, a Drexel bottle, and Will and Varrentrapp bulbs, connected by sealed or ground joints, as shown in the figure. The operation is conducted as follows: In the first Voit flask, V 1 , selenium dioxide and tellurium dioxide are dissolved in potas- sium hydroxide, the alkali is neutralized and the precipitate thus formed is redissolved by phosphoric acid, added in excess to the amount of 20 cm. 3 of the acid of sp. gr. 1.70. To the solution is added i grm. of potassium bromide with enough water to make the entire volume of the solution 50 cm. 3 The second flask, V 2 , contains 10 cm. 3 of water, and the Drexel bottle and trap are charged with a solution of potassium iodide. Carbon dioxide is passed through the apparatus and the solution in V 1 is boiled until the volume has diminished to 15 cm. 3 , the flask and connecting tube being cloaked with a mantle of asbestos board and gently * F. A. Gooch and A. W. Peirce, Am. Jour. Sci., [4], i, 181. t See page 5. 39 2 METHODS IN CHEMICAL ANALYSIS flamed toward the last to remove traces of the selenium bromides held back mechanically by the oily tellurium compound which collects. After cooling, the first flask V 1 is removed; I grm. of potassium iodide and 5 cm. 3 of hydrochloric acid are added to the contents of the second flask, V 2 ; the current of carbon dioxide is again started through the apparatus; the mixture is boiled ten minutes; and the iodine in the flask, receiver and trap, determined by titration with sodium thiosulphate, is taken as the measure of the selenium dioxide.* Results of this procedure are given in the table. Double Distillation. Te0 2 taken. grm. KBr taken. grm. H 3 P0 4 (sp. gr. 1.70) taken. cm. 3 Final volume. cm. 8 SeO 2 taken, grm. SeO 2 found. grm. Error, grin. 2O . 15 o . 0366 0.0372 +o . 0006 20 15 0.0366 0.0377 +O.OOII 2O 15 o. 1098 o. 1090 O.OOOS 20 15 o. 1098 O.IIOI + O.OOO3 O.I 20 15 0.0733 0-0735 + O.OOO2 0. 2O 15 0.0997 0.0995 O.OOO2 O. 2O 15 o. 1004 o. 1003 O.OOOI O. 2O 15 0.0916 0.0914 0.0002 0. 20 15 0.0997 0.0995 O.OOO2 0. 20 15 O. IOIO o. 1014 -J-O.OOO4- O. 2O 15 o. 1015 0.1008 O.OOO7 O. 2O IS o. 1019 O. IO22 + 0.0003 O. 2O IS O. IOIO O. IOI2 + O.OOO2 0. 2O 15 O. IOO2 O. IOOO O.OO02 0. 20 15 0.1006 o. 1004 O.OOO2 O. 2O 15 o. 1006 O.IOOI 0.0005 The phenomena of the distillation are very characteristic. When selenious acid is present without tellurous acid, the solu- tion boils quietly in the first flask until the volume of liquid has decreased to about 30 cm. 3 , when traces of red selenium begin to deposit in the tube joining the first and second flask. When the volume has further diminished to about 25 cm. 3 the liquid begins to take on color, darkens rapidly, and evolves bromine, which at once attacks the selenium previously deposited. The greater part of the bromine is absorbed in the second flask, V 2 , but a trace finds its way to the Drexel bottle, in which it sets free a slight amount of iodine from the iodide. As the operation * See page 379. SELENIUM 393 progresses, an orange-yellow crystalline solid, presumably sele- nium tetrabromide for the most part, appears in the tube where the selenium has been, while a dark oily liquid, consisting largely, no doubt, of the monobromide, condenses in drops upon the walls of the flask and returns to form a floating layer upon the hot liquid. Finally, when the volume has diminished to 15 cm. 3 , the liquid has become perfectly clear and colorless, white fumes of hydrobromic acid are evolved, and the tube between the two flasks has been cleared. At this point the second flask, V 2 , contains (besides a trace of selenium corresponding to the slight amount of bromine which has escaped to the Drexel bottle) the colorless selenious acid regenerated by the action of the water and free bromine upon the mixed selenium bromides. The contents of this flask may now be treated with potassium iodide and hydrochloric acid as directed above * and the iodine in the receiver, including, of course, the small amount set free by the bromine which reaches the receiver in the first stage of the process, and the small amount remaining in the flask, measure the selenium dioxide acted upon. When tellurium dioxide is subjected without the selenium dioxide to similar treatment the phenomena are different. The solution containing the tellurous acid, potassium bromide, and phosphoric acid, in the proportions used in the experiments with selenious acid, colors at about the same degree of concentration at which the solution containing the selenious acid began to darken. As the concentration progresses, the color deepens, ruby red crystals (probably hydrated tellurium tetrabromide) form, which accumulate upon the walls of the flask and turn yellow, and when the volume of the solution is diminished to 15 cm. 3 a green vapor begins to distil. During the process no iodine is set free in the Drexel bottle, and upon stopping the boiling and adding potassium iodide to V 2 no iodine is liberated, even when the boiling has gone so far that a trace of the green vapor has condensed and run into the water in the flask. When the tellurium dioxide and selenium dioxide are both present the characteristic phenomena occur together the evo- lution of bromine, coloring of the liquid, distillation of selenium bromides, crystallization of tellurium tetrabromide, and volatiliza- tion of the selenium compounds. * See page 379. 394 METHODS IN CHEMICAL ANALYSIS TELLURIUM. The Gravimetric Estimation of Tellurous Acid by Liberation of Iodine and Absorption of that Element by Silver. Tellurous acid reacts with potassium iodide in presence of hydrochloric acid and silver * in an atmosphere of hydrogen according to the reaction TeO 2 + 4 KI + 4 HC1 = 4 KC1 + 2 H 2 O + Te + 2 I 2 . The increase in weight of insoluble material represents both iodine and tellurium. Tests made by Perkins | with tellurium dioxide prepared from the basic nitrate gave the following results. Weighing of Silver Iodide and Tellurium. Ag taken, grni. Te taken, grm. Increase, grm. Calculated Te. grm. Error, grm. 2.0152 0.0330 0.1654 0.0332 +O.OOO2 2.0152 O . 0990 Q-493 1 o . 0989 0.0001 2.0815 0.0528 0.2635 0.0529 +O.OOOI 2.0815 o . 0660 0.3294 0.0661 + O.OOOI 2.0815 O . 0990 0.4948 o . 0993 +o . 0003 2.1693 0.1650 0.8240 0.1654 +o . 0004 3.0126 o. 1650 0.8258 0.1657 +0.0007 3.0126 o . 0660 0.3302 o . 0663 +0.0003 The Determination of Tellurous Acid by Oxidation with Potassium Permanganate. The estimation of tellurous acid by oxidation with excess of potassium permanganate (either in acid or alkaline solution), destruction of the higher oxides of manganese or the manganate by standard oxalic acid in presence of sulphuric acid, and titra- tion of the residual oxalic acid by more permanganate has been shown by Brauner J to be feasible, but the tendency of the 'permanganate to throw off too much oxygen when the oxidation is made in solutions strongly acidified with sulphuric acid (as must be the case if the tellurous oxide is to be held perma- nently in solution by sulphuric acid) necessitates the application of a considerable correction. Gooch and Danner have shown, * See page 444. f Claude C. Perkins, Am. Jour. Sci., [4], xxix, 540. t Jour. Chem. Soc., 1891, 238. F. A. Gooch and E. W. Danner, Am. Jour. Sci., [3], xliv, 301. TELLURIUM 395 however, that when the tellurous oxide is first dissolved in an alkali hydroxide and the solution is made acid to a limited degree with sulphuric acid, either before or after oxidation by the permanganate, no correction appears to be necessary. According to the first procedure, the alkaline solution of the oxide is diluted to 100 cm. 3 , a measured amount of standardized permanganate is added in excess, sulphuric acid [i : i] is intro- duced to an amount not exceeding by more than 5 cm. 3 that needed for neutralization, standardized oxalic acid is measured in to an amount more than sufficient to destroy the manganic oxide and permanganate left after the oxidation, and the surplus of oxalic acid is titrated by permanganate. According to the second procedure, the alkaline solution of the oxide is treated with sulphuric acid [i : i] until the precipitate first thrown down is just redissolved, and I cm. 3 more of the [i : i] acid is added. To the solution thus acidulated is measured in potassium permanganate in excess, and then standardized oxalic acid to an amount a little more than sufficient to destroy the permanganate remaining; the liquid is warmed to 80, and the excess of oxalic acid is titrated by permanganate. The results of experiments made according to these procedures are given below. Oxidation in Alkaline Solution. TeO 2 taken, grin. TeO 2 found, grm. Error, grm. Mean error, grm. 0.1200 0.1199 O.OOOI 1 0.0783 0.0783 . 0000 0.0931 O. IIOO o . 0938 o. 1116 +0.0007 I +0.0016 1 l . +0.0006 o . 0904 o . 0907 +0.0003 0.1065 0.1077 +O.OOI2 J Oxidation in Acid Solution. TeO 2 taken, grm. TeO 2 found, grm. Error, grm. Mean error, grm. 0.0910 0.0912 +O.OOO2 "1 0.0910 O . 0908 O.OOO2 0.0911 0.0913 0.0922 0.0913 +O.OOII I O.OOOO | +0.0003 0.0912 0.0913 +0.0001 0.0914 0.0921 +0.0007 J 39 6 METHODS IN CHEMICAL ANALYSIS Oxidation in ^ n tne presence of free hydrochloric acid the action Presence of a of the permanganate upon tellurous acid has been shown by Brauner * to be irregular and excessive, and the irregularity cannot be corrected (as in the titration of ferrous salts in presence of hydrochloric acid) by the addition of a manganous salt according to the well-known procedure of Kessler | and Zimmermann.t Gooch and Peters have pointed out that there should be nothing to prevent the accurate deter- mination of tellurium in tellurous compounds in the presence of chlorides by the permanganate process, provided the first oxi- dation is made in alkaline solution and the second oxidation is carried out with such precautions as are necessary to a correct determination of oxalic acid by permanganate in presence of hydrochloric acid; for the special danger of over-action on the part of the permanganate cannot exist while the solution is alkaline, and has passed when the tellurite has become a tellurate and before the solution is made acid. It has been shown that the presence of a manganous salt is necessary and sufficient to secure regularity of action when oxalic acid is titrated in presence of a considerable amount of hydrochloric acid. When the amount is no more than would be formed in the decomposition of a gram or two of halogen salt of tellurium the disturbing effect under ordinary conditions of work is probably inappreciable, but even in such a case it is better to oxidize in the presence of a manganous salt for the reason that the titration of the oxalic acid may then be made at the ordinary atmospheric temperature. According to the procedure recommended, the alkali hydroxide solution of tellurous oxide containing the alkali chloride is treated with standardized potassium permanganate until the character- istic permanganate color is visible; standardized ammonium oxalate is introduced in excess of the quantity required to reduce the remaining permanganate, manganate, and higher oxides; and enough sulphuric acid [i : i] is added to neutralize the alkali hydroxide and leave an excess of about 5 cm. 3 . Then the final titration with permanganate may be made either after heating * Jour. Chem. Soc., 1891, 241. t Ann. Phys., cxviii, 48; cxix, 225, 226. I Ann. Chem., ccxiii, 302. F. A. Gooch and C. A. Peters, Am. Jour. Sci., [4], viii, 122. TELLURIUM 397 the solution to 80 or at the ordinary temperature after the ad dition of 0.5 grm. to I grm. of manganous chloride. Experimental results of the procedure are given below: Permanganate Oxidation in Alkaline Solution: Treatment with Oxalic Acid Permanganate Titration in Acid Solution. Volume at beginning, 150 cm. 3 : Te = i27.5. TeO 2 taken, grm. NaCl. grm. H 2 SO 4 1 : i. cm. J MnCl 2 .4H 2 0. grm. TeO 2 found, grm. Error, grm. Temperature of titration, 6o-8o C. O.IOOO 0.4 5 0.1003 +0.0003 O.IOOO 0.4 5 O.IOOO 0.0000 O.IOOO 0.4 5 0.1004 +0.0004 O.IOOO i .0 5 . . . 0.1003 +0.0003 0.0650 i .0 5 0.0653 +0.0003 Temperature of titration, 2o-26 C. 0.0700 0.4 5-7 I.O 0.0705 +0.0005 0.0700 0.4 5-7 I.O 0.0698 0.0002 0.0700 0.4 5-7 0-5 0.0701 +0 . 0001 O.IOOO 0.4 5-7 0-5 o. 1008 +0.0008 Oxidation in Fairly good determinations of tellurous acid may Presence of a be made similarly in the presence of a bromide, pro- vided the titration is made at the atmospheric temperature in the presence of a sufficiency (0.5 grm. to I grm.) of a manganous salt and of an excess of sulphuric acid limited to Permanganate Oxidation in Alkaline Solution: Treatment with Oxalic Acid: Permanganate Titration in Acid Solution. Volume at beginning, 150 cm. 3 : Te = i27.5. Temperature of titration, 24-26 C. TeO 2 taken. NaCl. KBr. H 2 S0 4 12.5 per cent. MnCl 2 . 4 H 2 0. TeO 2 found. Error. grm. grm. grm. cm. 8 grm. grm. grm. 0.0650 o-5 I I.O 0.0661 +O.OOII 0.0650 0-5 I I .O 0.0647 0.0003 O.IOOO o-5 I I .O O. IOO2 +O.OOO2 0.3000 o-5 5 o-5 O.3OIO +0.0010 0.0650 0-5 o-5 i i .0 0.0661 +O.OOII 398 METHODS IN CHEMICAL ANALYSIS 5 cm. 3 of the 12.5 per cent mixture, in a volume of 150 cm 3 . At higher temperatures and higher concentrations of acid, bro- mine is liberated by the permanganate. The experimental results are given in the table. The Determination of Tellurous Acid by the Precipitation of Tellurous Iodide. Hydriodic acid and tellurous acid interact with the formation of tellurium tetraiodide, converted by water into an oxyiodide and by excess of alkali iodides to soluble double salts. Gooch and Morgan * have observed that when potassium iodide is added to a cold solution of tellurous acid containing at least one-fourth of its volume of strong sulphuric acid, no tendency to form a double salt becomes apparent until the potassium iodide amounts to more than enough to convert all the tellurous acid present into tellurium tetraiodide according to the equation H 2 Te0 3 + 4 H 2 S0 4 + 4 KI = TeI 4 + 4 KHSO 4 + 3 H 2 O. The tellurium tetraiodide thus formed is extremely insoluble in sulphuric acid of one-fourth strength, though soluble in excess of potassium iodide and acted upon by water with formation of tellurium oxyiodide and hydriodic acid. It is produced at first in the condition of a finely divided dark brown precipitate which upon agitation of the liquid gathers in curdy masses and settles, leaving the liquid clear. By taking advantage of this tendency to curd, it is possible to determine without great difficulty the exact point during the gradual addition of potassium iodide when the precipitation of the tellurium iodide is complete. Upon this property Gooch and Morgan have based a simple titrimetric method for the direct determination of small amounts of tellu- rous acid. According to the procedure described, tellurous oxide is dis- solved in a very little of a strong solution of potassium hydroxide, and dilute sulphuric acid is added carefully until the tellurous acid which is precipitated upon neutralization of the alkali hydroxide is just redissolved. To this solution, contained in an Erlenmeyer flask, sulphuric acid [i : i] is added in such amount that the liquid shall contain, after the subsequent addition of potassium iodide in solution, at least one-fourth of its volume of * F. A. Gooch and W. C. Morgan, Am. Jour. Sci., [4], ii, 271. TELLURIUM 399 strong sulphuric acid. The flask is placed upon a pane of window glass supported upon strips of wood about I cm. above the level of a work table covered with white paper. A solution of approx- imately decinormal potassium iodide, free from iodate and care- fully standardized in terms of iodine by a method to be described,* is introduced gradually from a burette into the middle of the Erlenmeyer beaker. As the drops of the potassium iodide touch the liquid the precipitation forms at the center and travels in rings toward the outer walls of the beaker. When the liquid be- comes so opaque that the effect of the potassium iodide is dis- tinguished with difficulty, the beaker is rotated and the curded precipitate permitted to settle ; and then the process of titration is continued as before until precipitation ceases. With an Erlen- meyer flask 10 cm. in diameter across the bottom and a final volume of liquid amounting to not more than 100 cm. 3 the pre- cipitation is easily followed. The results of a series of determinations made according to the method described are recorded in the following table : Precipitation of Tellurous Iodide. Te = i2 7 * Final volume. cm. 3 Strongest H 2 SO 4 present. cm. 3 Iodine value of KI used. grm. TeOj taken, grm. TeO 2 found, gnu. Error, grm. 50 I? o . 0706 0.0223 O.O22I O.OOO2 5 17 o . 0764 0.0244 0.0239 0.0005 50 17 O.ISQI o . 0496 0.0499 +0.0003 60 17 0.1655 0.0517 0.0519 +O.OOO2 60 17 0.1578 o . 0498 0.0494 O.OOO4 80 3 0.1591 o . 0498 0.0499 +0.0001 100 30 0.3179 O. IOOI 0.0997 0.0004 100 30 0.3186 o. 1008 0.0999 0.0009 IOO 30 0.3208 O. IOII 0.1005 0.0006 IOO 30 0.3208 O. IOIO 0.1005 -0.0005 Determined by permanganate oxidations and reductions by hydrobromic acid (see p. 402). The lodometric Estimation of Tellurous Acid. The determination of tellurous acid by oxidation of the alkali hydroxide solution with potassium permanganate, reduction of residual permanganate and higher oxides of manganese with oxalic acid in presence of sulphuric acid, and titration of the excess * See page 457. 400 METHODS IN CHEMICAL ANALYSIS of oxalic acid by permanganate, is not feasible in presence of an iodide, because upon acidifying the mixture, iodine is at once set free. Potassium permanganate and the higher oxides of manganese are, however, completely and rapidly reduced by an excess of potassium iodide upon the addition of acid, and the iodine liber- ated is a measure of the permanganate. Norris and Fay * have utilized this reaction in an excellent iodometric method for the de- termination of tellurous acid. This method consists in treating the alkaline solution of tellurous oxide with standard permanga- nate until the meniscus of the liquid shows a deep pink color, then diluting the solution with ice water, adding potassium iodide and sulphuric acid, and titrating with sodium thiosulphate. The difference between the amount of iodine thus found and the amount found by treating similarly the same amount of perman- ganate, taken by itself, is the measure of the tellurous acid. It is plain that any agent capable of converting the iodine to hydriodic acid without at the same time reducing telluric acid should be capable of measuring the excess of the permanganate, and so the amount of tellurous acid originally present. Gooch and Peters f make use of the standard arsenite, employed also in standardizing the permanganate,! to take up the free iodine. According to this procedure, the solution of tellurous oxide in alkali hydroxide is added to a solution of potassium iodide; standardized permanganate is run in until the green color of the manganate appears (about 30 cm. 3 of the n/io solution for every o.i grm. of TeC>2); dilute sulphuric acid is introduced in slight excess, followed, after the iodine has separated, by an excess of acid potassium carbonate ; and the iodine is titrated to vanishing color (without starch) by standard arsenite. It is evident that when the solution is acidified more than enough iodide to com- plete the reduction of the manganese oxides should be present, or else that the arsenious acid should be present in suitable amount before the sulphuric acid is put in. This latter, procedure may be used in case, for any reason, it is preferred not to introduce more iodide into the solution than was present originally, as, * Am. Chem. Jour., xx, 278. f F. A. Gooch and C. A. Peters, Am. Jour. Sci., [4], viii, 125. j See page 41. TELLURIUM 401 for example, when a direct determination of the iodine present is to follow. Experimental results follow in the table: Permanganate Oxidation in Alkaline Solution and lodometric Determination of the Excess. TeOj taken. NaCl. KBr. KI. Total volume at end. NaOH present during oxidation. TeO 2 found. Error. gnn. grm. grm. gnu. cm. 8 grm. grm. grm. 0.1000 o-S 1 60 O.I 0.1005 +0.0005 0.1000 o-5 1 60 O.I O. IOOI +0.0001 O.IOOO . o-S 160 O.I 0.1003 +0.0003 O. IOOO i .0 250 O.I o. 1007 +0.0007 O.2OOO I.O 250 0.2 0.1997 -0.0003 O. IOOO 0.5 0-5 o-S 250 O.I O. IOOO o.oooo 0.2100 1.0 1.0 I.O 22 5 0.2 0.2105 +0.0005 O.IOOO o-S 1 60 I.O O. IOII +O.OOII 0.2000 I.O 300 2.O o . 2009 +0.0009 The lodometric Determination of Telluric Acid. Gooch and Rowland * have shown that telluric acid may be reduced by the action of potassium bromide and sulphuric acid to the condition of tellurous acid and estimated by deter- mining the iodine liberated by the bromine set free in the operation. According to the method demonstrated, the alkali tellurate is introduced into the apparatus for distillation with 3 grm. of potassium bromide, care being taken to insure in the 50 cm. 3 or more of liquid the presence of 10 cm. 3 of sulphuric acid of half strength. A current of carbon dioxide is passed through the apparatus, and the solution is boiled to set free the bromine, which is absorbed in potassium iodide and estimated by standard sodium thiosulphate. The distillation apparatus f consists of a Voit gas-washing flask which is joined by a sealed joint to the inlet tube of a Drexel washing bottle. To the outlet tube of the Drexel bottle is sealed a Will and Varrentrapp absorption apparatus. The * F. A. Gooch and J. Rowland, Am. Jour. Sci., [3], xlviii, 375. t See Fig. 3, page 4. 4O2 METHODS IN CHEMICAL ANALYSIS washing bottle and attached bulbs contain a solution of 3 grm. of potassium iodide, and the former is kept cool by standing it during the distillation in a vessel of cold water. The formation of tellurium tetrabromide in the concentrated acid liquid makes it impossible to tell by the color when all the bromine has been distilled, but the evidence of the experiments goes to show that the boiling of the liquid from a volume of 50 cm. 3 to 25 cm. 3 is sufficient, while concentration from 100 cm. 3 to 20 cm. 3 apparently does no harm. The tellurite used in testing this procedure was made from tellurous oxide which had been shown by titration with perman- ganate to have an equivalent weight of about 159, which corre- sponds to an atomic weight of 127 for the element tellurium. Results are given in the table. Reduction by Hydrobromic Acid and Estimation of Bromine Set Free. Initial volume. cm. Final volume, cm.* TeO 2 taken, grm. TeO 2 found, grm. Error grm. 5 2O O.OIO2 0.0098 O.OOO4 50 20 O.OIO2 0.0099 0.0003 5 2O O.OI02 o . 0098 0.0004 5 20 O.OI02 0.0098 0.0004 100 40 O. IOOO 0.0994 O.OOO6 80 40 O.IOOI O. 1001 o.oooo IOO 2O O.IOO2 0. IOOI O.OOOI 50 2O O. IOOO 0.1003 +o . 0003 50 25 0.50II 0.5008 -0.0003 50 25 0.5002 o . 5006 +o . 0004 50 25 O.5OOO 0.4998 O.OOO2 50 2O 0.5000 0.4994 0.0006 The Precipitation of Tellurium Dioxide and the Separation of Tellurium from Selenium. Browning and Flint * have shown that fairly accurate and con- cordant estimations of tellurium may be obtained by acting upon the alkaline solution of a tellurite with hydrochloric acid, am- monium hydroxide, and acetic acid, sucessively, and weighing the precipitate as tellurium dioxide, the best of all the forms in which * Philip E. Browning and William R. Flint, Am. Jour. Sci., [4], xxviii, 112. TELLURIUM 403 tellurium has been weighed. It is unaffected by the air, is anhy- drous, is not hydroscopic, and can easily be obtained in pure condition. It can be heated to any temperature under that of low redness without any danger of volatilization. All processes for the estimation of tellurium in which the tellurium is precipitated and weighed in elementary condition are open to the objections, first, that there is more or less difficulty in securing completeness of precipitation owing to the rapid in- crease of free acid * in the solution; and, second, that the product is extremely susceptible to oxidation. On the other hand, those methods in which compounds decomposable by heat are trans- formed to the dioxide by ignition are generally both tedious by reason of the length of time required (as, for example, the basic nitrate process as described by Norris f) and, what is more to the point, liable to errors caused not only by lack of constancy of composition but also by the volatilization of the product to be weighed. The process of Browning and Flint presents there- fore special advantages in the determination of tellurium. According to the preferred procedure set forth, the material is dissolved in hydrochloric acid or in a 10 per cent solution of potassium hydroxide, about 2 cm. 3 for 0.2 grm. of dioxide. From the solution acidified with hydrochloric acid and diluted with boiling water to a volume of 200 cm. 3 the finely crystalline tellurium dioxide is precipitated by the careful addition of dilute ammonia in faint excess followed by the faintest possible excess of acetic acid. If these simple operations are properly carried out, the pre- cipitate will become crystalline by the time the alkali hydroxide is in excess ; the addition of a few drops of acetic acid causes the precipitation to become entirely quantitative when the solution has cooled, so that no tellurium will be found in the filtrate by stannous chloride. The precipitate can be transferred, and safely and rapidly washed with cold water, and dried to constant weight at about 105 (or even under low redness) in a quarter of an hour. Furthermore, the filtration can be performed at the end of half an hour, or after twenty-four hours, as most convenient. * Crane, Am. Chem. Jour., xxiii, 409. See also Lenher and Homburger, Jour. Am. Chem. Soc., xxx, 387. t Jour. Am. Chem. Soc., xxviii, 1675. 404 METHODS IN CHEMICAL ANALYSIS Test results are given below : The Solution in HCl Diluted with Boiling Water, and Treated with Ammonia and Acetic Acid. TeO 2 taken, grni. TeO 2 found, gnu. Error, grm. 0.2002 0.2000 0.0002 0.20IQ O.2OI7 O.OOO2 0.2904 O.2OO6 O.2OO2 O.2OO4 O.OOO2 O.OOO2 O.2OTI O. 2OIO O.OOOI 0.2003 0.2003 0.0000 The Solution in KOH Acidified with HCl, Diluted with Boiling Water, Treated with Ammonia and Acetic Acid. 2 TeO 2 .HNO 3 taken.* TeO 2 theory: Te taken as 127.5. TeO 2 found. Error. grm. grm. grm. grm. 0.2502 o . 2089 0.2083 O.OOo6 0.2524 0.2108 O. 2IIO +0.0002 0.2505 o. 2092 o . 2089 0.0003 0.2528 O. 2III 0.2106 0.0005 0.2531 0.2113 0.2106 0.0007 0.5008 0.4182 0.4182 0.0000 0.5010 0.4183 0.4175 0.0008 0.5005 0.4179 0.4178 O.OOOI Dissolved in KOH. Separation from Selenium. If hydrochloric acid solutions of tellurium and selenium diox- ides be mixed, abundantly diluted with boiling hot water, and the operation of the above described process properly applied, only the tellurium is precipitated, the selenium remaining en- tirely in solution in the filtrate. This not only provides a simple and rapid preparative process for the purification of tellurium from selenium, but also makes possible the estimation of tellu- rium directly in the presence of the latter element. The dilution should be made, however, with boiling hot water, as cold water induces a flocky precipitation and inclusion of selenious acid. Results are given in the table. TELLURIUM 405 Separation of Tellurium from Selenium TeO 2 taken. grm. SeO 2 taken, grm. TeO 2 found, grm. Error, grm. The solution faintly acid with HC1, diluted cold. 0.2000 o. 2015 0.2038 O.I O.I O I . 2002 o. 2016 o . 2040 +O.OOO2 +O.OOOI +0 0002 The solution faintly acid with HC1, diluted hot and treated with NH 4 OH and HOC 2 H 3 O. 0.2028 0.05 O.2OIQ O.OOO9 0.2024 0.05 o. 2024 O OOOO o . 2003 O.I 0.1992 o.oon o . 2009 O.I 0.2003 0.0006 CHAPTER X. CHROMIUM; MOLYBDENUM; URANIUM. CHROMIUM. The Estimation of Chromium as Silver Chromate. It has been shown by Autenrieth * that when chromic acid is added to a boiling solution of silver nitrate, or when a soluble chromate or dichromate is added to a solution of silver nitrate previously acidified with nitric acid, or when silver chromate is treated with nitric acid, silver dichromate is formed; and that, on the other hand, it is silver chromate which is precipitated when silver nitrate in excess is added to a solution of a soluble dichromate, cold or hot, the reaction proceeding according to the equation 4 AgN0 3 + K 2 Cr 2 7 + H 2 O = 2 Ag 2 CrO 4 + 2 KNO 3 + 2 HNO 3 . The characteristics of both silver dichromate and silver chro- mate have recently been summarized and further studied by Margosches.t The solubility of silver dichromate in water and in ordinary solutions is such as to preclude the use of this sub- stance as the final product of a quantitative process depending upon precipitation. The solubility of silver chromate in a moderately large volume of water is also considerable, and the solvent action of free acid, even acetic acid in quantity, is marked. Gooch and Weed % have found, however, that the precipitation of silver chromate is practically complete in a solution only faintly acid with acetic acid and in presence of a large excess of silver nitrate. If such a precipitate is collected in the filtering crucible and washed with a dilute solution of silver nitrate until no other impurities remain, silver chromate does not dissolve, and the excess of silver nitrate may be removed by the cautious use of water without appreciable effect upon the precipitate. * Ber. Dtsch. chem. Ges., xxxv, 2057. t Zeit. anorg. Chem., xli, 68; 1, 231. t F. A. Gooch and L. H. Weed, Am. Jour. Sci., [4], xxvi, 85. .406 CHROMIUM 407 From the results of test experiments it is apparent that accu- rate determinations of chromium taken as the chromate or dichromate may be secured by precipitating silver chromate in presence of an excess of silver nitrate, making the solution am- moniacal and then faintly acid with acetic acid, transferring the precipitate, after standing half an hour, to the filtering crucible, washing with a dilute solution of silver nitrate, and, after other soluble impurities have been removed, finishing the washing with small amounts of water applied in successive portions. The results are given in the table. In no case did the filtrate, with the washings, show by the lead acetate test the presence of a chromate. Precipitation of Silver Chromate. Ag 2 CrO 4 . K 2 Cr 2 7 AgNO 3 used Volume at taken. in precipitation. precipitation. Found. Theory. Error. gnn. grm. cm. 3 grm. grm. grm. 0.0921 0.4248 TOO o. 2072 O. 2076 0.0004 0.0921 0.4248 IOO 0.2073 O. 2076 0.0003 0.0921 0.4248 IOO 0.2075 o. 2076 o.oooi 0.0921 O . 4248 IOO o. 2074 o. 2076 O.OOO2 0.0921 0.4248 100 0.2075 o. 2076 o.oooi 0.0921 . 4248 IOO 0.2073 o. 2076 0.0003 0.0921 0.4248 ICO 0.2073 o. 2076 0.0003* o 0921 0.4248 IOO 0.2075 o. 2076 O.OOOI* 0.0921 0.4248 IOO O.2O8O o. 2076 -fo . 0004! 0.0921 0.4248 TOO 0.2070 0.2076 o.ooo6f 0.5801 3 150 1.3087 1.3082 +0.0005 0-7352 3 2OO 1.6573 1-6574 o.oooif * The precipitation was made in presence of 5 grm. of N H^NOj. t The precipitation was made in presence of 5 grm. of XaNO 2 . t An excess of I cm. 3 of 40 per cent acetic acid was added before filtering. The lodometric Determination of Chromic Acid. Kessler * has shown that arsenious acid may be estimated by treating it, in the presence of hydrochloric acid, with an excess of a chromate solution of known strength, by which treatment the arsenious acid is oxidized and the chromic acid reduced, and determining the excess of chromic acid by adding an excess of a ferrous salt and titrating with chromic acid until a drop taken from the solution fails to give a blue color with a ferricyanide. * Ann. Phys., xcv, 204. 408 METHODS IN CHEMICAL ANALYSIS The amount of the chromate originally used less the excess de- termined by the ferrous salt gives the amount of the chromate used for the oxidation, from which may be calculated the amount of arsenious acid originally present. Despite the use of a ferrous salt and the numerous steps involved in the manipulation, Kessler claims very satisfactory results for his method. Browning * gives results of experiments showing that Kessler's reaction may be used in the reverse process for the determination of chromic acid, the arsenious acid being used in excess, according to the reaction 4 CrO 3 + 3 As 2 O 3 + (*)As 2 O 3 = 2. Cr 2 O 3 + 3 As 2 O 5 + (*) As 2 O 3 . According to the procedure indicated, an excess of n/io arsenite is added to the cold solution of chromic acid acidulated with 10 cm. 3 of dilute hydrochloric acid or sulphuric acid [i 13], the total volume being less than 100 cm. 3 After a few minutes, when the solution has taken on the bluish green color character- istic of chromic salts, the solution is treated with acid potassium carbonate or acid sodium carbonate in excess (about 5 grm.). To the alkaline solution n/io iodine is added in excess, the mix- ture is allowed to stand, with frequent shaking, for about half an hour, residual iodine is taken up with n/io arsenite, and the excess of the last is titrated by n/io iodine in presence of starch. The long period of standing is made necessary by the tendency of precipitated chromic hydroxide to hold arsenious acid and thus delay the oxidation by iodine. The precipitation of the chromic hydroxide may be obviated by addition of Rochelle salt before the neutralization, but in this event the dark green color taken on by the solution makes the end-reaction of the starch iodide difficult to determine with great accuracy. In the presence of a ferric salt the brown color of the precipi- tate also makes the determination of the end-reaction difficult unless the precipitate is allowed to settle after each addition of iodine. Edgar f has shown, however, that this difficulty may be obviated by adding sirupy phosphoric acid (3 cm. 3 to 5 cm. 3 ) before the neutralization, so that the iron precipitate is white and the starch blue conies out against the pale green of the pre- cipitate. Results of test determinations are given in the table. * Philip E. Browning, Am. Jour. Sci., [4], i, 35. t See page 511. CHROMIUM 409 Reduction by Standard Ar senile. CrO 3 taken, grin. CrO 3 found, grm. Error, grm. Remarks. O. IOOI o. 1004 +O . 0003 0.1005 0.1004 O.OOOI The iodine acted 20 minutes. o. 1006 o. 1007 +O.OOOI The iodine acted 20 minutes. O.IOO| O. IOII +0.0007 The iodine acted 20 minutes. o. 1009 O. IOOQ o . oooo The iodine acted 2 hours. O. IO02 0.1003 +O.OOOI The iodine acted 2 hours. O. IOII o. 1004 0.0007 Rochelle- salt used. o. 1007 o. 1007 o . oooo Rochelle salt used. 0.0401 0-0395 0.0006 o . 0402 o . 0388 0.0014 0.5 grm. ferric alum present. O. IOOI o. 1018 +o. 1017 o. 1009 o. 1007 O.OOO2 o. 1007 O. IOII +0.0004 i grm. ferric alum present. 0.1005 o. 1017 +O.OOI2 0.1004 O. IOIO +o . 0006 O. IOOO 0.1032 +0.0032 Rochelle salt used. 0.1005 o. 1006 +O.OOOI i grm. ferric alum present. The lodometric Estimation of Chromic Acid and Vanadic Acid. That vanadic acid and chromic acid may be accurately esti- mated in presence of one another by taking advantage of the differential reducing actions of hydrobromic and hydriodic acids has been shown by Edgar.* In carrying out the operation, the alkali salts of the chromic and vanadic acid are put in the Voit flask of the distillation apparatus previously described,! one or two grams of potassium bromide are added, the flask is connected with the absorption apparatus containing a solution of potassium iodide made alka- line with sodium carbonate or sodium hydroxide, and the whole apparatus is filled with hydrogen gas. Fifteen to twenty cubic centimeters of concentrated hydrochloric acid are added through the separatory funnel and the solution is boiled for ten minutes, an interval of time found to be enough for the completion of the reduction. A slow current of hydrogen is maintained to avoid back suction of the liquid from the Drexel bottle. The apparatus is disconnected, the Voit flask placed in a beaker containing cold water, and the alkaline solution in the absorp- tion apparatus cooled by running water. The contents of the * Graham Edgar, Am. Jour. Sci., [4], xxvi, 333. t See Fig. 3, page 4. 4io METHODS IN CHEMICAL ANALYSIS trap are washed into the Drexel bottle and the solution therein is made slightly acid with hydrochloric acid. The liberated iodine is titrated with approximately n/io sodium thiosulphate and the color is brought back by a drop or two of n/io iodine solution, after the addition of starch. Alkaline potassium iodide is again placed in the absorption apparatus and the latter connected with the Voit flask. The current of hydrogen is turned on and, after the air has been ex- pelled, the apparatus is disconnected momentarily, one or two grams of potassium iodide are added to the solution in the Voit flask, and connections made again. Through the separatory fun- nel 10 cm. 3 to 15 cm. 3 of concentrated hydrochloric acid and 3 cm. 3 of sirupy phosphoric acid are added and the solution in the reduction flask is boiled to a volume of 10 cm. 3 to 12 cm. 3 . The absorption apparatus is removed and cooled, hydrochloric acid is added and the liberated iodine titrated with approximately n/io sodium thiosulphate. Double Treatment with Hydrobromic Acid and with Hydriodic Acid. V 2 O 6 taken NaV0 3 . CrO 3 taken as K 2 Cr 2 7 . I. Titration. Na 2 S 2 3 w/ioXi.031. II. Titration. Na 2 S 2 3 w/ioXi.o3i. Error on V 2 O 5 . Error on CrO 3 . Rrm. grm. cm. 3 cm. 3 grm. grm. O 1^23 16. 20 16. 22 ( I) ( o.oooo * ^-o^O (II) I +0.0002 O 152? 16. IQ 16. 20 ( I) ( o.oooi w O * x y (II) ( 0.0000 ( I) j o.oooi o . 203 i 21 . cm. s Pi grm. grm. grm. IO 2* 0.4960 0.2425 400 300 40 0.9536 0.2425 0.0000 IO 2* 0.4970 0.2429 400 300 40 0-9534 o. 2624 O.OOO5 IO 2* 0.4942 0.2416 400 300 30 0.9509 0.2418 +O.0002 10 2* 2 0.4969 0.2429 . . 400 300 30 0-9559 0.2431 +O . OOO2 IO 2* 3 0.4956 0.2423 400 350 30 0.9546 o. 2428 +0.0005 IO 2* 3 0.4969 0.2429 400 350 23 0.9662 0.2432 +0.0003 IO 2] 3 0.4949 o. 2419 400 300 27 0.9523 o. 2422 +o . 0003 10 at 5 0.4970 0.2429 400 250 55 0-9559 0.2431 +O.OO02 10 2 t 5 0-4955 o. 2422 400 300 30 0.9524 0.2422 O.OOOO IO at 5 0.4967 0.2428 400 300 33 o . 9546 o. 2428 o.oooo IO at 6 0.4964 0.2427 400 300 30 0.9550 0.2429 +O.OO02 * The iron was added in the form of iron alum. t The iron was added as FeSO 4 oxidized by HNO a . See Fig. 6, page 6. CHLORINE; BROMINE; IODINE 451 escaping vapor, silver nitrate is added in excess to the contents of the flask, the precipitate is settled, collected in a perforated crucible' on asbestos, washed, dried, and weighed as silver chloride. Tests of the method, with determinations in blank that is, experiments from which the iodine was purposely omitted - are detailed in the tabular statement. The Nitrite With pure sodium nitrite at hand there is probably Method. no ser i ous objection to introducing that substance directly into the solution, but if impurities are present it is desir- able to generate the gas outside the solution. For a generator, two straight drying tubes are connected by a rubber tube and set up after the fashion of the von Babo generator, and the rapidity of the current is regulated to a rate of five or six bubbles to the second by changing the relative elevation of the gener- ator tubes. The iodine separates immediately upon the intro- duction of the nitrous fumes and escapes upon boiling, leaving the solution colorless in a very short time. The litmus test must, however, be relied upon to indicate the removal of the iodine. According to the method, as developed, the solution of the chloride and iodide contained in an Erlenmeyer flask is diluted to 400 cm. 3 , 10 cm. 3 of sulphuric acid of half strength are added, and the gas from 2 grm. of sodium nitrite acted upon by dilute sulphuric acid (generated in simple apparatus, such as is described above) is passed with reasonable rapidity into the agitated solu- tion. The liquid is boiled until colorless, and still further until litmus paper placed in the steam gives no reaction for iodine after an exposure of two minutes. The contents of the flask are Distillation with Nitrite: Determination of Chloride in the Residue. NaN0 2 H 2 S0 4 [i : i]. used in gener- KC1. = HC1. KI. Initial volume. Final volume. Time in min- AgCl found. HC1 found. Error. ator. utes. cm. 3 grm. grm. grm. grm. cm. 8 cm. 3 grm. grm. grm.' 10 2 0-4953 0.2421 I 400 350 2O 0.9524 o. 2422 + O.OOOI 10 2 0.49750.2432 I 400 350 16 0-9573 0,2434+0.0002 10 2 0.4956 0.2423 I 300 250 15 0.9530 0.2423 o.oooo 10 2 0-4973 0.2431 I 300 250 IS 0.9550 o. 24290.0002 10 2 0.4964 0.2427 I 300 250 15 0.9550 0.2429I+0.0002 IO 2 0.4969 0.2429 I 300 250 15 0.9567 0.2433 +0.0004 452 METHODS IN CHEMICAL ANALYSIS treated with silver nitrate, and the precipitated chloride is settled, collected on asbestos in the perforated crucible, washed, dried and weighed. The results of test experiments are given in the preceding table. The Direct Determination of Bromine (and Chlorine) in Mixtures of Alkali Bromides (and Chlorides) with Iodides. The methods elaborated by Gooch and Mar * for the direct determination of chlorine in mixtures of alkali chlorides and iodides have been studied by Gooch and Ensign f with a view to similar application in the determination of bromine in mix- tures of alkali bromides and iodides. The conditions found suitable in the separation of chlorine from iodine prove to be in- appropriate to the separation of bromine from iodine. The use of a ferric salt to eliminate iodine was not found to be practi- cable under any conditions. Even at the high dilution ranging from 650 cm. 3 to 500 cm. 3 bromine was likewise set free even when the concentration of the sulphuric acid present was very low. On the other hand the nitrite process, only fairly successful when the sulphuric acid present is restricted to 5 cm. 3 of the [i : i] acid in a final volume of 500 cm. 3 , is established as trustworthy when the sulphuric acid present is held within the limits of 2 cm. 3 to 4 cm. 3 of the [i : i] mixture. When the quantity of sulphuric acid is still further diminished there is evidently a slight tend- ency to show an apparent excess of bromide, due in all probabil- ity to the retention of a little combined iodine in the solution. The best proportion for practical use is probably 3 cm. 3 of the half and half acid in an initial volume not less than 600 cm. 3 , and this proportion proves to be applicable to the separation of iodine from an iodide associated with chloride as well as bromide. The method may be briefly summarized as follows: The neutral solution containing the bromide and iodide is diluted to 600 cm. 3 or 700 cm. 3 (instead of 400 cm. 3 , which was found to be a sufficient dilution in the case of the separation of chlorine from iodine); I cm. 3 to 1.5 cm. 3 of strong sulphuric acid, or, better, 2 cm. 3 to 3 cm. 3 of the [i : i] mixture (instead of the 10 cm. 3 * See pages 449, 451. t F. A. Gooch and J. R. Ensign, Am. Jour. Sci., [3], xl, 145. CHLORINE; BROMINE; IODINE 453 employed in the chlorine separation) are added, a sufficient amount of pure sodium or potassium nitrite is introduced (or, if it is preferred, the gas generated by the action of dilute sul- phuric acid upon the ordinary nitrite and introduced from the outside) ; and the liquid is boiled, after trapping the flask, until the color has vanished and the escaping steam no longer gives to red litmus paper the color characteristic of iodine. The residual liquid is treated with an excess of silver nitrate and the precipitated bromide filtered off, washed, dried and weighed. The process of boiling need not extend beyond half an hour, or a little more, and care should be taken that the volume of the liquid shall never be less than 500 cm. 3 . The process has been tested for quantities of the potassium bromide and iodide not much larger than 0.5 grm. each. The presence of 0.5 grm. of potassium chloride does not affect the sharpness of the separation. The results are given below : Separation of Iodine from Bromine. H 2 S0 4 [i : i). KI. NaNO 2 in the liquid. KBr = HBr taken. Initial volume. Final volume. Time in min- utes. AgBr = HBr found. Error in HBr. cm. 3 grm. grm. grm. grm.. cm.* cm." grin. grm. grm. 3 0-5 0-35 0.5508 0.3745 650 500 30 0.8689 0-3744 O.OOOI 3 o-5 0-35 3-5513 0.3747 650 500 30 0.8694 0.3746 O.OOOI 3 o-5 o-35 0.5513 0.3747 650 500 30 0.8699 0.3748 +O.OOOI 3 0-5 o-3S 0.3005 o . 2042 650 500 30 0.4746 0.2045 +0.0003 3 0-5 0-35 0.2759 0.1875 650 500 30 0.4358 0.1878 +0.0003 3 0-5 i-75 0.5513 0.3747 650 500 30 0.8705 0.3750 +0.0003 3 o.S i-75 0.55100.3746 650 500 30 0.8707 0-3751 +0.0005 1 H 2 S0 4 [I :i|. KI. NaN0 2 used in gener- KBr = HBr taken. Initial volume. Final volume. Time in min- AgBr = HBr found. Error in HJr. ator. utes. cm. 3 grm. grm. grm. grm. cm. 3 cm. 3 grm. grm. grm. 2 o-5 2 0.5366 0.3647 650 500 30 0.84780.3654 +O.OOO7 2 0-5 2 0.5369 0.3650 650 500 30 0.8472,0.3651 + O.OOOI 2 o-5 2 0.5515 0.3747 650 500 30 O. 8687!o. 3742 O.OOO5 3 0-5 2 0.5371 0.3652 650 500 30 0.84590.3644 O.OOOS 3 0-5 2 0.5365 0.3647 650 500 30 0.84650.3647 O . OOOO 3 0-5 2 0.5368 0.3649 650 500 30 0.84860.3656 +0.0007 3 0.5 2 0.5364 0.3646 650 500 30 0.8471 0.3650 +0.0004 3 0-5 2 0.5505 0.3742 650 500 30 0.86900.3744 +O.O002 3 o.5 2 0.0576 0.0391 650 500 30 0.09150.0394 +0.0003 3 o-S 2 0.0552 0.0375 650 500 30 0.08830.0380 +o . 0005 454 METHODS IN CHEMICAL ANALYSIS Separation of Iodine from Bromine and Chlorine. H f S0 4 [i : ij. KI. NaNO 2 KBr. HC1. Theory for AgCl+ AgBr. Found AgCl+ AgBr. Error in silver salt. Error calculated asHBr. Error calculated as HC1. cm. 3 grin. grin. gnu. grm. grm. grm. grm. grm. grm . 3 3 0.5 0-5 0-35 o-35 0.5517 0.55H 0.4981 0.4980 1.8280 1.8268 1.8262 1-8253 0.0018 -0.0015 0.0008 0.0005 0.0006 0.0004 The Application of lodic Acid to the Analysis of Iodides. lodic acid may be easily and completely reduced by an excess of hydriodic acid with the liberation of iodine according to the equation : HI0 3 + 5 HI = 3 I 2 + 3 H 2 0. To apply this reaction to the quantitative estimation of iodic acid, it is only necessary to add to the free iodic acid or soluble iodate, in suitably concentrated solution, an excess of a solu- ble iodide, to acidify best with dilute sulphuric acid and to titrate with sodium thiosulphate the iodine thus set free, one-sixth of the iodine found being credited to the iodic acid. It has been shown by Riegler * that this reaction may be also applied to the quantitative estimation of iodides, the iodine set free upon the addition of a known excess of iodic acid to the iodide solution being removed by petroleum ether, and the residual iodic acid determined as described above. Gooch and Walker f have studied the limit of applicability of this reaction and have developed a direct method for the quan- titative estimation of iodides, dependent upon the action of iodic acid or an iodate in the presence of free sulphuric acid, neutral- ization of the solution by means of an acid carbonate, and titra- tion of the free iodine by arsenious acid five-sixths of the iodine thus found being credited to the iodide to be estimated. It has been found that by fulfilling certain necessary conditions the proposed method is entirely successful, so far as concerns the estimation of iodine in iodide solutions free from large amounts of chlorides or bromides. The degree of dilution at the time when the mixture of iodide and iodate is acidified has an important influence upon the com- * Zeit. anal. Chem., xxxv, 305. t F. A. Gooch and C. F. Walker, Am. Jour. Sci., [4], iii, 293. CHLORINE; BROMINE; IODINE 455 pleteness of the reaction. Thus, the mean error of test deter- minations in which the volume at the time of the reaction does not exceed 150 cm. 3 is practically nothing, while the errors at volumes of 300 cm. 3 and 500 cm. 3 amount to 0.0016 grm. and 0.0028 grm. respectively. The doubling of the amount of sul- phuric acid used in acidifying does not increase the amount of iodine liberated at the highest dilution. The plain inference is that the interaction between the iodide and iodate should be brought about in a volume of liquid not much exceeding 150 cm. 3 . The apparatus employed is a reaction bottle * of 500 cm. 3 or 1000 cm. 3 capacity, according to requirements, with stopcock and thistle-tube fused to the inlet tube and a Will and Varrentrapp absorption trap sealed to the outlet. The iodide for the test is drawn from a burette into the bottle and carefully washed down, potassium iodate in excess of the amount theoretically necessary is added, and the volume of the liquid is adjusted to 150 cm. 3 . The stopper with the thistle-tube and trap is put in place and the .trap is half filled by means of a pipette with a 5 per cent solu- tion of potassium iodide. Five cubic centimeters of [i : 3] sul- phuric acid are added through the thistle-tube and washed down, the stopcock is closed, and the solution gently agitated, if neces- sary, to insure a complete separation of iodine. Potassium bicarbonate in saturated solution to an amount about 10 cm. 3 in excess of that required to neutralize 5 cm. 3 of dilute [1:3] sulphuric acid is poured into the thistle-tube and allowed to flow into the bottle slowly enough to avoid a too violent evolu- tion of gas. The stopcock is closed and the solution agitated by giving to the bottle a rotary motion, at the same time keep- ing the bottom pressed down upon the work-table, to prevent a possible splashing of the iodide out of the trap into the acid solution. When the neutralization of the solution has been com- pleted, the bottle is shaken until the last trace of violet vapor has been absorbed in the liquid. The greater part of the solu- tion in the trap is then run back into the bottle, the stopper is removed, and the tube and trap are carefully washed, the washings being added to the bulk of the solution. Decinormal arsenious acid is introduced from a burette to the bleaching point, starch emulsion added, and the solution titrated back with decinormal iodine (usually only a few drops) to coloration. * Shown in Fig. 7, page 6. 456 METHODS IN CHEMICAL ANALYSIS The results of experiments made in the manner described upon portions of a solution of potassium iodide standardized by the arsenate method * are given in the accompanying table. Analysis of Pure Potassium Iodide. KI taken, grm. KI found, grm. Error, grm. 0.0814 0.0816 +O . OOO2 0.0814 0.0813 0.0001 0.0814 0.0805 O.OOOQ 0.0815 o . 0809 O.OO06 0.0814 o . 0808 O.OOO6 0.0814 0.0806 O.OOOS 0.0814 0.0812 O.O002 0.1628 0.1624 0.0004 0.1628 0.1617 o.oon 0.1628 0.1621 0.0007 0.1628 0.1619 0.0009 0.1628 0.1624 0.0004 0.1628 0.1621 0.0007 0.1628 0.1626 O.OOO2 0.2442 0.2451 +o . 0009 0.2442 0.2442 o.oooo o . 2442 0.2439 0.0003 0.3256 0.3258 +O.OOO2 0.3256 0.3256 O.OOOO 0.3256 0.3258 +O . 0002 0.3256 0.3272 +0.0016 0.3256 0.3256 o.oooo 0.4071 0.4076 +0.0005 0.4071 o . 4080 +o . 0009 0.4071 0.4073 +0.0002 The presence of any considerable amount of chloride or bromide, resulting no doubt in the formation of iodine chloride or iodine bromide, is prejudicial to the accuracy of the process. This is shown in the table following : Effects of Chloride and Bromide. KI taken, grm. KI found, grm. Error, grm. NaCl taken, grm. KBr taken, grm. 0.0772 0.0795 +0.0023 O.2 0.0772 0.0784 +O.OOT2 O.2 0.0771 0.0823 +0.0052 0-5 0.0773 0.0819 +o . 0046 o-5 0.1544 0.1588 +o . 0044 0-5 . . . 0.1544 0.1590 +o . 0046 0-5 0.0772 o . 0802 + 0.0030 O.2 0.0773 0.0853 +o . 0080 0.2 0.0772 0.0873 +O.OIOI . . . 0-5 0.0772 0.0861 +0.0089 0-5 0.1544 o. 1646 +O.OIO2 0-5 0.1543 0.1626 +o . 0083 0-5 * See page 457- CHLORINE; BROMINE; IODINE 457 It is plain that the value of the process in the determination of iodine in an iodide is restricted of necessity to those cases in which it is known that chlorides or bromides are not present to any considerable extent. For determining the standard of a solution of nearly pure potassium iodide, employed in so many laboratory processes, it is useful. The lodometric Determination of Iodine in Haloid Salts. The determination of iodine in a mixture of alkali chloride, bromide and iodide has been made the subject of investigation by Gooch and Browning.* In this work it was shown that the iodine of the iodide may be all liberated, under denned conditions, by the combined action of an arsenate and sulphuric acid, and its amount registered quantitatively by the amount of arsenious oxide produced. f Under similar conditions, the presence of as much as 0.5 grm. of sodium chloride brings about no formation of arsenious oxide, but does induce a loss of that substance by volatilization as arsenic chloride, proportionate to the amounts of both these substances. The effect of potassium bromide is to produce trifling reduction of arsenic .acid without volatility. Due correction of the amounts of iodine indicated by determi- nation of the arsenious oxide in the residue may be made by adding to the indicated amount 0.008 of the product of the weight of chlorine present in chlorides by the weight of iodine, and sub- tracting 0.0024 of the weight of bromine in bromides. The mode of proceeding in the determination of iodine in a mixture of alkali chlorides, bromides and iodides, according to this method, may be briefly summarized as follows: The substance (which should not contain of chloride more than an amount corresponding to 0.5 grm. of sodium chloride, nor of bromide more than corresponds to 0.5 grm. of potassium bromide, nor of iodide much more than the equivalent of 0.5 grm. of potassium iodide) is dissolved in water in an Erlenmeyer beaker of 300 cm. 3 capacity, and to the solution are added 2 grm. of dihydrogen potassium arsenate dissolved in water, 20 cm. 3 of a mixture of sulphuric acid and water in equal volumes, and enough water to increase the total volume to 100 cm. 3 , or a little * F. A. Gooch and P. E. Browning, Am. Jour. Sci., [3], xxxix, 188; xlv, 334. t For the reaction, see pages 291, 463. 458 METHODS IN CHEMICAL ANALYSIS more. A platinum spiral is introduced, a trap made of a straight two-bulb drying-tube cut off short is hung with the larger end downward in the neck of the flask,* and the liquid is boiled until the level reaches the mark put upon the flask to indicate a volume of 35 cm. 3 . Great care should be taken not to press the con- centration beyond this point on account of the double danger of losing arsenious chloride and of setting up reduction of the arse- nate by the bromide. On the other hand, though 35 cm. 3 is the ideal volume to be attained, failure to concentrate below 40 cm. 3 introduces no appreciable error. The liquid remaining is cooled and nearly neutralized by sodium hydroxide (ammonia is not equally good), neutralization is completed by hydrogen potas- sium carbonate, an excess of 20 cm. 3 of the saturated solution of the latter is added, and the arsenious oxide in solution is titrated by standard iodine in the presence of starch. Reduction of Ar senate and Determination of Arsenious Oxide Produced. v& cm.* H 2 KAsO 4 . grtn. NaCl. KBr. Final volume cm. 8 Theory for iodine. grtn. Iodine found. gnu. Error found. gl'111. Error corrected. 2O 2 35 o . 4080 0.4079 O.OOOI O.OOOI 20 2 . . . 35 . 0.4091 o . 4086 0.0005 0.0005 20 2 35 0.4083 o . 4086 +o . 0003 +0.0003 2O 2 35 0.0400 0.0396 O.OOOI 0.0004 2O 2 35 0.0400 0.0391 0.0009 0.0009 20 2 35 0.0400 0.0400 o.oooo o . oooo 2O 2 35 0:0400 o . 0401 +0.0001 +O.OOOI 2O 2 35 0.0040 0.0037 0.0003 -0.0003 2O 2 35 o . 0040 0.0038 O.OOO2 O.OOO2 20 2 0-5 . . . 35 0.4077 o . 4066 O.OOII O.OOOI 2O 2 0-5 35 0.4082 0.4073 0.0009 +O.OOOI 2O 2 o-5 35 o . 4086 0.4073 0.0013 -0.0003 2O 2 0.5 35 0.0400 o . 0402 +0 . 0002 +O.OOOI 20 2 o-5 35 0.0400 0.0395 o . 0005 0.0004 2O 2 o-S 35 0.0040 0.0037 0.0003 O.OOO2 2O 2 o-S .... 35 0.0040 0.0037 0.0003 O.OOO2 2O 2 0-5 35 0.4082 0.4092 +0.0010 +O . OOO2 20 2 o.S 35 0.4138 0.4136 O.O002 O.OOIO 2O 2 0-5 35 0.4083 0.4099 +0.0016 +0.0008. 2O 2 0-5 35 0.0400 0.0410 +0.0010 +O.OOO2 20 2 o-S 35 0.0400 o . 0404 -j-o . 0004 0.0004 2O 2 0-5 35 o . 0040 0.0048 +0.0008 0.0000 2O 2 o-S 35 0.0040 o . 0049 +0.0009 +O.OOOI 2O 2 o-5 o-S 35 0.4087 o . 4083 0.0004 O.OOO2 20 2 o-5 o-5 35 0.4112 0.4111 O.OOOI +0.0001 2O 2 0-5 o-5 35 0.4083 0.4079 0.0004 O.OOO2 * See Fig. 6, page 6. CHLORINE; BROMINE; IODINE 459 With ordinary care the method is rapid, reliable and easily executed, and the error is small. . In analyses requiring extreme accuracy all but accidental errors may be eliminated from the results by adding (algebraically) to the amount of iodine indicated an amount i = (0.008 X wt. Cl X wt. I) - (0.0024 X wt. Br). Results obtained by this process are given in uncorrected and corrected forms. Results * obtained in a comparison of titrations of arsenious oxide in the residue with estimations of the iodine expelled under the prescribed conditions and collected in the distillate show close agreement. Comparison of A rsenious Oxide in Residue and Iodine Expelled. Iodine taken in form of KI. giro. Iodine found from As 2 O 3 in residue. gnu. Iodine found in distillate by As 2 O 3 . grm. Iodine found in distillate, by Na 2 S 2 3 grin. Error in residue. grm. Error in distillate. grm. O 4O^4 o 40^2 o 0002 O 4CX7 O 4O^ O OOO2 o 40^4 o 40^2 O OOO2 o 4054 o 40^2 O OOO2 o 4042 o 4046 o 4046 -|-o 0004 -|~o 0004. o . 4050 . 405 2 o . 4040 -J-O OOO2 o ooio o . 4050 0.4058 0.4052 0.4052 0.4039 o . 405 i +0 . 0002 o 0006 O.OOII o 0007 o . 4054 o . 4046 o . 405 i 0.0008 o 0003 o 4042 o 4046 O 4O3Q -j-o 0004 o ooo 3 o . 40 ^ ? o 4052 O 4O ^ X Tf 10 M Tj" CO to t^oo ON 00 g g 8 88 8 8 w & d dodo o do d d + + + 1 + 1 1 1 1 1 ^ o M i 8 oo H CO ^ H o o "^'S-Q i *k*M O M t> CM ON OO CO t>. t^. O> 00 00 OO OO OO M M M M M d odd d <* * -2:g-K k>l \ HH rS **^ (-H HH ?*\ |. s |g M x M > xx M to C^ CO vO 1 O CS vo O 00 M M M M W M 1 i O O Q O O O O o d o d do o 88 d d 888 odd ++++ 1 1 1 1 I 1 1 1 > <* oo A g ON M (5* H d d S A **** lOOO *O ^J" CO ^* M t~. O^ ^ ^O VQ ^ V^ $ >, ON ON ON ON ON ON ON COS CO fO CO m t d d d d odd d d 6 d d ; 5> -* |S -g R i 1 i d d o o d o d d MJHh + 1 + odd 1 1 1 3BB8 o d d o 1 ++ 1 OO t^. ** o Q G M t^ g|^S 5>: - X 5> 11 i d d MANGANESE 481 crucible, the composition of the oxide remaining is generally very closely that of the oxide Mn 3 O 4 . The estimation of man- ganese in the form of the manganoso-manganic oxide, Mn 3 O4, is by no means to be considered utterly untrustworthy when the process is conducted in the manner described, though it must be recognized that an irregular result may occur occasionally. The danger of accepting such an irregularity as a correct indication may be eliminated to a very considerable extent if the precau- tion is taken invariably to moisten the ignited oxide with nitric acid, and ignite again. Concordant results thus obtained may be taken with a fair degree of confidence. Results of this procedure, as well as of the treatment described for producing in succession the oxides MnO2, Mn 2 O 3 , Mn 3 O4, are given in the preceding tabular statement. The Determination of Manganese Separated as the Carbonate. Austin * has shown that the precipitation as carbonate in presence of large amounts of ammonium chloride, according to Guyard,f and estimation as sulphate | or oxide offer reliable means for the determination of manganese. The manganese salt dissolved with a considerable amount of ammonium chloride (about 10 grm.) in 200 cm. 3 of boiling water is treated with ammonium carbonate in excess. The liquid is kept warm until the precipitate subsides and is then filtered off either upon paper or upon asbestos in a perforated crucible. The crucible and precipitate collected upon asbestos are ignited in the oxidizing flame of a powerful burner || to give the oxide Mn 3 O 4 which is weighed as such; the precipitate collected upon paper may be ignited to the oxide and then converted by treatment with a few drops of sulphuric acid and gentle heating to the sulphate. Weighing as the carbonate is not feasible; for, as Rose has correctly stated, U" evolution of carbon dioxide and oxidation of the residue begin before the water is thoroughly removed. Experimental results are given in the table. * Martha Austin, Am. Jour. Sci., [4], v, 382. t Hugo Tamm, Chem. News, xxvi, 37. | See page 477. See page 478. || See page 479. Tf Ann. Phys., Ixxxiv, 52. 482 METHODS IN CHEMICAL ANALYSIS Precipitation as Carbonate: Conversion to Oxide and to Sulphate. NH 4 C1. Mn 3 4 taken. Mn 3 4 found. v Error. MnSO 4 taken. MnSO 4 found. Error. grin. grm. grm. grm. grm. grm. IO 0.1776 0.1770 0.0006 IO 0.1776 0.1788 + O.OOI2 10 0.1776 0.1770 O.OOO6 10 0.1776 0.1774 O.OOO2 IO 0.2478 o . 2463 O.OOI5 0.4905 0.4903 O.OO02 IO O. II2I 0. IIIO O.OOII 0.2219 0.2225 +O.OOO6 10 O.I58l 0.1584 +o . 0003 0.2128 0.3126 O.OOO2 IO 0.1699 0.1672 0.0027 0-3344 0-3355 0.0009 The Determination of Manganese Precipitated as Ammonium Manganese Phosphate and Weighed as Manganese Pyrophosphate. By Gibbs' original method manganese phosphate was precipi- tated by hydrogen disodium phosphate in large excess above the quantity required to cause the precipitation. The flocky white precipitate was dissolved either in sulphuric or hydrochloric acid, and precipitated again at the boiling temperature by ammonia in excess. This semi-gelatinous precipitate, on boiling or long standing even in the cold, becomes crystalline, the crystals form- ing beautiful talcose scales which have a pearly luster and a pale rose color. The precipitate was filtered off, washed with hot water, dried and ignited. The results obtained by Gibbs' students agree closely with the theory for the pyrophosphate. When a manganous salt is precipitated in the cold by an excess of an alkaline phosphate, it falls, as Heintz * has shown, in the form of the trimanganous phosphate of the formula Mn 3 P 2 O 8 . This same phosphate constitutes the greater part of the precipi- tate which forms when a manganous salt reacts in the cold, in the presence of ammonium chloride, with microcosmic salt and ammonia in slight excess, but boiling or even subsequent stand- ing may effect a more or less complete conversion of the man- ganese phosphate to the ammonium manganese phosphate. The success of the analytical process in which manganese is weighed as the pyrophosphate turns, therefore, upon the change of the trimanganous phosphate, Mn 3 P 2 Os to the ammonium * Ann. Phys., Ixxiv, 449. MANGANESE 483 manganese phosphate, NH 4 MnPO4. Gooch and Austin * have shown that the presence of a large amount of ammonium salt is essential to the formation of the precipitate of ideal constitu- tion. Apparently the proportion of ammonium chloride present to ammonium manganese phosphate formed should be at least 40 : i, or, speaking approximately, 200 molecules of ammonium chloride must be present in the liquid to every molecule of the phosphate formed ; and the ammonium chloride may be increased almost to the point of saturation of the liquid without causing more than a trifling solubility of the ammonium manganese phosphate in the presence of an excess of the precipitant. Fur- thermore, the precipitate may be washed with perfect safety with pure water as well as with slightly ammoniacal water, or with ammoniacal water containing ammonium nitrate, if the filtration is performed rapidly and the precipitate gathered in small space, as is the case when the phosphate is collected on asbestos in the perforated crucible. The finely granular pre- cipitate which may be obtained by slow action of dilute ammonia added gradually to the hot solution of the manganese salt appar- ently includes a portion of unconverted phosphate which resists the replacement of the manganese by ammonium. On the other hand, the precipitate of flocky condition thrown down in the cold passes easily to the silky and crystalline condition when heated with the proper amount of ammonium salt, and possesses a constitution approaching the ideal under such conditions. The conversion of the flocky manganous phosphate is so rapid that the precipitation may be carried on safely in glass vessels. If the ammonium chloride in the solution were to be included in the precipitate it would volatilize entirely during the ignition, leaving no trace unless, possibly, a portion of its chlorine were to combine with the manganese. Tests for chlorine in the residue of pyrophosphate have resulted negatively, no more than a mere trace being found in any case, so that the contam- inating effect of the ammonium chloride proves to be insignifi- cant and the responsibility for the excess of weight above the theory must apparently rest with the included microcosmic salt. In the practical determination of manganese by the phosphate method of Gibbs, therefore, the presence of large amounts of * F. A. Gooch and Martha Austin, Am. Jour. Sci., [4], vi, 233. 4 8 4 METHODS IN CHEMICAL ANALYSIS ammonium chloride is strongly advocated. Good results may be obtained most easily and surely by the following procedure: The slightly acid solution (in platinum or glass), containing in a volume of 200 cm. 3 an amount of manganese not more than enough to make 0.4 grm. of the pyrophosphate, 20 grm. of ammonium chlo- ride, and 5 cm. 3 to 10 cm. 3 of a cold saturated solution of micro- cosmic salt, is precipitated in the cold by the careful addition of dilute ammonia in only slight excess. The mixture is heated until the precipitate becomes silky and crystalline, the whole is allowed to stand and cool half an hour, the precipitate is col- lected upon asbestos in a perforated platinum crucible, washed (best with slightly ammoniacal water), dried at gentle heat, and ignited as usual. By this process determinations of the larger amounts of manganese 0.4 grm. of the pyrophosphate approximate rather more closely to the theoretical values than do those of the smaller amounts 0.2 grm. In either case the average error should not exceed o.ooio grm. in terms of manganese. Results obtained by this procedure are given in the table. Determination as Manganese Pyrophosphate. Mi^PjO; equivalent to MnO 2 . Error in terms of Mn 2 P 2 7 . grm. Error in terms of manganese. grm. Saturated solution of HNaNH 4 PO<. cm. s NH 4 C1. grm. Total volume. cm. J Manganese in the filtrate. Taken, grin. Found, grm. In platinum. 0.1885 0.1903 +0.0018 +0.0007 5 20 200 None. 0.1885 0.1910 +0.0025 +O.OOIO 5 20 2OO None. 0.1885 0.1913 +0.0028 +0.001 1 5 20 2OO None. 0.1885 0.1911 +0.0026 +0.0010 5 20 200 None. 0.3770 0.3776 +o . 0006 +O.OOO2 5 2O 2OO None. 0.3770 0.3773 +0.0003 +O.OOOI 5 2O 2OO None. 0.3770 0.3778 +0.0008 +0.0003 5 2O 2OO None. 0.3770 0-3783 +0.0013 +0.0005 5 2O 200 None. In glass. 0.1885 0.1904 +0.0019 +0.0007 5 20 2OO None. 0.1885 0.1898 +0.0013 +o . 0005 5 20 200 None. 0.3770 0.3767 0.0003 o.oooi 5 2O 200 None. 0.3770 0.3784 +0.0014 +0.0005 5 20 2OO None. MANGANESE 485 The Electrolytic Determination of Manganese. For the electro-deposition of manganese as the dioxide various processes have been described.* With stationary electrodes solutions containing nitric acid, sulphuric acid, acetic acid, formic acid with or without a formate, or ammonium acetate alone, with chrome alum, or with acetone, have been employed.! For use with the rotating anode, solutions containing ammonium acetate with chrome alum or alcohol have been advocated.! In all these processes hydrated manganese dioxide is deposited upon a large anode which is preferably a roughened platinum dish of considerable capacity. Gooch and Beyer have described experiments made to test the utility of the electrolytic filtering cell in the determination of manganese as the dioxide. The procedure adopted was the simplest. Portions, 50 cm. 3 each, of a solution of pure man- ganous sulphate, standardized by evaporation of measured por- tions and gentle ignition of the residue over a radiator, || were treated, in each case, with six drops (0.17 cm. 3 ) of concentrated sulphuric acid, and electrolyzed in the filtering cell with a current of 2 amperes (N.D.ioo= 5 amp.) and 20-10 volts, the voltage decreasing as the solution became heated. In one set of experiments the process of continuous filtration during electrolysis, for which the adjustment of apparatus is shown in Fig. 16, was employed. If The time required for the deposition of 0.1860 grm. of the dioxide was one hour and three- quarters. In a second set of experiments the closed cell, shown in Fig. 15, was used during the electrolysis, and the adjustment for filtration made subsequently as previously described.** A period varying from two hours and ten minutes to two hours and fifty minutes is required in the latter process. Tests with hydro- gen dioxide and ammonia showed that the deposition was com- plete in the process of continuous filtration and practically so in the closed-cell process. The closed-cell process naturally * Smith's Electro-analysis, edition of 1907, page 134 et seq. t Ibid. % Koester, Zeit. Elektrochem., x, 553. F. A. Gooch and F. B. Beyer, Am. Jour. Sci., [4], xxvii, 62. || Am. Jour. Sci., [4], v. 209. H See page 17. ** See page 15. 4 86 METHODS IN CHEMICAL ANALYSIS requires less attention during the electrolysis, and so it is advan- tageous to run the process for a period, perhaps two hours, with the closed cell, and then to adjust the apparatus for filtration during further electrolytic action, in order that floating particles of the dioxide may be drawn to the felt and completeness of pre- cipitation may be assured. In this way the advantage of the circulating process may be obtained with less attention to manip- ulation than is required when the filtration is continuous from the start. The deposit is washed with water after interruption of the current, first dried at 200 for ten or fifteen minutes and weighed, and thereafter ignited to low redness in the spreading flame of a large burner.* Results of experiments with the cell arranged for continuous filtration, and of experiments in which the closed cell was used until the electrolysis was nearly over,, are given in the accompanying table: Deposition of Manganese Dioxide. Solution of MnS0 4 taken. cm. 2 S0 4 concen- trated. cm. 8 Current. Theory Mn0 2 . grm. MnO 2 weighed as MnO 2 . grm. MnO 2 weighed as Mn 3 O 4 . grm. Error, grm. Amp. N. D.JOO. Volt. Electrolysis with continuous filtration. 5 0.17 2 t: 2O 12 o 1860 o 1862 -(-O OOO2 0.1858 O OOO2 50 0.17 2 C 2O 12 0.1860 0.1856 o . 0004 0.1856 o . 0004 5 o. 17 2 2O~I2 o 1860 o 1843 o 0017 o 1872 +O OOI2 Electrolysis in closed cell with subsequent filtration. CO 0.17 2 5' 2O I 2 o 1860 o 1860 o oooo o I&ZT, o 0007 50 o. 17 2 5" 2O~IO o 1860 o 1856 o 0004 0.1856 o 0003 co O 17 2 1 8 10 o 1860 o 18^3 o 0007 0.1858 O.OOO2 The results are evidently as good as could be expected of any process which involves the weighing of a manganese oxide brought to condition by heating. The degree of oxidation of the oxide thrown down under the conditions approximates closely * Am. Jour. Sci., [4], v, 214. MANGANESE 487 to that of the ideal oxide represented by the symbol MnO 2 .H 2 O, formerly assigned by Riidorff * to the electrolytically formed oxide, and differs in that respect from that of the electrolytically deposited oxide which was studied by Groeger.f The Determination of Manganese Precipitated by the Chlorate Process.. Gooch and Austin { recommend the substitution of sodium chlorate for potassium chlorate in the precipitation of manganese from the nitric acid solution, the greater solubility of the sodium salt and the consequent rapidity with which its decomposition takes place making its use advantageous. According to the treatment outlined, the manganous nitrate, free from chlorides and sulphates, is dissolved in concentrated nitric acid (85 cm. 3 ), and treated with sodium chlorate (5 grm.). The mixture is boiled five minutes, more nitric acid (15 cm. 8 ) and a few crystals of sodium chlorate are added, and the heat- ing is discontinued as soon as the liquid boils again. The in- solubility of the precipitate, if this procedure is followed, is so great that no more than insignificant traces, never exceeding o.oooi grm., may be recovered from the filtrate after filtering on asbestos upon the perforated cone or crucible and washing with water. On the other hand, prolonged boiling after the last addi- tion of chlorate results in considerable losses of manganese (from o.ooio grm. to 0.0030 grm.), due to the reducing effect of lower oxides of nitrogen formed (as is always the case in boiling nitric acid) after the chlorine dioxide has been expelled. An excess of chlorate at the end of the operation seems to be essential and a slight yellow color in the solution, due to chlorine dioxide, is a favorable indication. It is best to filter the undiluted solu- tion under pressure upon asbestos on a perforated cone having a filtering surface of about 40 cm. 2 . Dilution before filtration tends to increase the solubility of the manganese. While manganese is very completely precipitated in the chlo- rate process conducted with the precautions indicated, the condi- tion of oxidation cannot be assumed to be that of the dioxide, and the indications of any process which rests upon the assumption * Zeit. angew. Chem., 1892, 6. t Zeit. angew. Chem., 1895, 253. J F. A. Gooch and Martha Austin, Am. Jour. Sci., [4], v, 260. 488 METHODS IN CHEMICAL ANALYSIS that the oxygen value of the manganese compound precipitated is that of the dioxide is likely to be erroneous. If, therefore, the chlorate method is used to separate manganese, precautions must be taken to secure a definite condition of oxidation before apply- ing a process of determination which depends upon the oxygen value of the precipitated oxide. Manganese Oxide by Chlorate Process. Mn taken. Mn found upon the hypothesis that MnO, is the oxide finally obtained. Error. giro. grm. grm. By reduction with potassium iodide and titration of free iodine by thiosulphate. o . 0643 0.0637 0.0006 0.0643 0.0642 o.oooi 0.0643 o . 0642 O.OOOI 0.0651 0.0651 o.oooo 0.1125 O. II2I 0.0004 0.1125 O.II2I 0.0004 0.1125 O. I I 2O 0.0005 0.1214 O . I 2O6 0.0008 0.1214 o. 1207 0.0007 o. 1214 0.1223 +0.0009 0.1214 O.I2I4 0.0000 By reduction with arsenious oxide and titration of the excess by iodine in presence of Rochelle salt. 0.1213 0. 1212 O.OOOI 0.1213 0. I2OI O.OOI2 0.1213 0.1213 o . i 203 . I 2O8 o.ooio O.OOO5 Gooch and Austin show that a definite condition of oxidation may be secured by dissolving the precipitate in hydrochloric acid, diluting a little, adding sulphuric acid, evaporating until no more hydrochloric acid remains, and treating the manganous sulphate as follows, according to the method of Wright and Menke:* The solution of manganous sulphate (not exceeding 0.5 grm.), very nearly neutralized by potassium carbonate, is * Jour. Chem. Soc., xxxvii, 36. NICKEL 489 mixed with a solution of zinc sulphate (2 grm.), and a freshly made and carefully filtered dilute solution of potassium perman- ganate (1.5 grm.)- The liquid now amounting to about 500 cm. 3 is heated to 80 and acid potassium carbonate is added in quan- tity a little more than enough to neutralize the free acid. The loose precipitate is collected upon asbestos and after careful washing is returned to the flask in which the precipitation was made. The oxygen value of the oxide thus obtained, now that of the oxide MnO 2 , may be determined by any appropriate method of treatment. The preceding table contains the results of two different methods for determining the oxygen value of manganese dioxide separated by the chlorate method and brought to ideal condition by the procedure described. NICKEL (COBALT). The Electrolytic Determination of Nickel with the Rotating Cathode. Gooch and Medway* determine nickel by deposition upon the rotating cathode, f from 50 cm. 3 of solution containing one- half its volume of concentrated ammonia and a gram of am- monium sulphate. It should be especially noted that the solution must be kept within the limit of volume indicated above, as further dilution lengthens the time necessary for complete deposition. Deposition of Nickel. Nickel taken.^ gnu. Nickel found, grm. Error, grm. Current, amp. N.D. 100 . Time, min. 0.0954 0.0954 0.0000 i-5 5 30 0.0954 0.0953 o.oooi 3 10 25 0.0954 0.0956 +0.0002 3 10 25 0.0954 0.0953 o.oooi 3-5 11.7 20 0.0954 0.0955 +O.OOOI 3-5 ii. 7 2O 0.1738 0.1736 O.OOO2 3-5 11.7 25 0.1738 0.1740 +0.0002 3-5 11.7 25 0.1738 o . i 740 +0.0002 4 13-3 25 0.1738 0.1737 O.OOOI 4 13-3 25 0.1738 0.1738 o.oooo 4 13-3 25 * F. A. Gooch and H. E. Medway, Am. Jour. Sci., [4], xv, 323. t See page n. 490 METHODS IN CHEMICAL ANALYSIS The Estimation of Nickel by Precipitation as the Oxalate and Titration with Potassium Permanganate. Classen * has shown that nickel may be completely precipi- tated by treating the soluble nickel potassium oxalate with a large amount of acetic acid, and estimated by igniting the oxalate and weighing the oxide. Ward f has adapted this process to the volumetric determination of nickel by titration of the precipi- tated oxalate with permanganate. Finding that the precipitate formed by precipitation with potassium oxalate tends to include the alkali oxalate and that the nickel oxalate thrown down from the boiling acetic acid solution of a nickel salt falls in finely divided condition and is difficult to filter, Ward makes the first precipitation in water solution and adds the acetic acid after- ward. The nickel salt is dissolved in a definite amount of water, crystallized oxalic acid is added to the boiling solution, and an equal volume of acetic acid is added. The precipitate is allowed to settle over night, filtered on asbestos in a perforated crucible, and washed with small amount of water. The crucible is placed in a beaker containing about 25 cm. 3 of dilute [1:3] sulphuric acid, and heat is applied to dissolve the oxalate. The vol- ume of the solution is made up to about 200 cm. 3 with water, and cobalt sulphate is added until the green color of the nickel salt is neutralized and a slight pink tinge appears. The addition of the cobalt salt, recommended by Gibbs, is necessary to secure a definite end point. The solution is heated to boiling and the titration is made in the usual manner with permanganate. Re- sults of this procedure are given in the table. Precipitation by Oxalic Acid. Nickel taken as the sulphate. Volume of water solution at precipitation. Oxalic acid. Acetic acid added. Nickel found. Error. grm. cm.* grm. grin. grm. grm. 0.0050 IOO 2 IOO 0.0054 +0.0004 0.0257 100 2 IOO 0.0258 +O.OOO7 0.0503 IOO 2 IOO O.O5O2 O.OOOI 0.1257 IOO 2 IOO O.I27I +0.0014 * Zeit. anal. Chem., xvi, 470. t H. L. Ward, Am. Jour. Sci., [4], xxxiii, 336, NICKEL 491 The Detection of Nickel in Presence of Cobalt. Browning and Hartwell * have modified advantageously the method of Clarke f for the separation of nickel and cobalt, according to which cobalt precipitated as the ferricyanide re- mains insoluble in ammonium hydroxide while the nickel salt is dissolved. The method as modified may be described as follows: To not more than o.i grm. of the salts of the two elements in about 5 cm. 3 of water a few drops of a saturated solution of alum are added, free mineral acid is neutralized with ammonium hydroxide, and the solution is made faintly acid with acetic acid. To this solution is added about 0.5 grm. of solid potassium ferricyanide with shaking to effect the solution of this reagent and the pre- cipitation of the nickel and cobalt salts. The precipitate is treated with about 5 cm. 3 of strong ammonium hydroxide and the mixture is filtered. To the filtrate, which should have no reddish color, a piece of sodium or potassium hydroxide about the size of a pea is added and the mixture is boiled. The appear- Detection of Nickel. CoS0 4 .7H 2 gnu* NiS0 4 .7H 2 grm. KA1(S0 4 ) 2 saturated solution. cm. KjFeC.N,. grm. NH 4 OH (concen- trated). cm. NaOH, solid. grm. Result. .... O.OIOO 2 0-5 5 < I Heavy precipitate. O.OO5O 2 0-5 5 < I Heavy precipitate* .... 0.0010 0.0003 2 2 o-S o-S 5 5 < I Heavy- precipitate. Distinct. .... O.OOOI 2 o.S 5 Plain. O. IO 2 o. o.oooo o . 4498 o . 4494 O.0004 o . 5086 o . 5085 o.oooi 0.4462 0-4457 0.0005 0.4226 0.4222 0.0004 0.5170 0-5165 0.0005 The standard of potassium permanganate as determined from pure iron is identical with that obtained in these experiments with the special ammonium oxalate, but the standard obtained IRON 497 in the former way would under ordinary conditions be more satis- factory for work in iron analysis. A simple and rapid method, then, for standardizing the potassium permanganate solution is to determine its strength, first, by comparison with electrolytic iron in the manner above described. Then by reduction and titration with the permanganate the exact amount of iron in each cubic centimeter of the ferric chloride solution may be determined. This being ascertained, the solution of ferric chloride may be employed at any time for the standardization of potassium permanganate. The Behavior of Ferric Chloride in the Jones Reductor. The column of amalgamated zinc as applied in the Jones reductor * has proved very effective in the reduction of ferric sulphate preparatory to the estimation of the ferrous salt by 'potassium permanganate. The impression has prevailed that the salt of iron acted upon by the amalgamated zinc must be the sulphate and that chlorides and nitrates must not be present even in small amounts. Randall | has shown, however, that it is possible to reduce ferric chloride in the zinc reductor and to determine the iron with success by potassium permanganate, provided the titration is carried on in the presence of manganous sulphate and in solutions sufficiently dilute. A small excess of hydrochloric acid has no influence on the result in dilute solu- tions, and in a volume of one liter the excess may amount to as much as 25 cm. 3 of the strongest acid. According to the method described, the procedure is to first run 100 cm. 3 of warm dilute 2.5 per cent sulphuric acid through the reductor charged with amalgamated 2O-mesh zinc, next to pass in the iron solution diluted with 100 cm. 3 of warm 2.5 per cent sulphuric acid and then to wash down with 200 cm. 3 of the warm dilute acid followed by 100 cm. 3 of hot water. The receiving flask of the reductor is kept in a vessel containing run- ning tap water, so that the solution is cooled as fast as it is reduced. Manganous sulphate, I grm. to 5 grm., is added before the titration with permanganate. Following are results obtained by this method, and, for comparison, the results obtained upon the same solution of ferric chloride standardized by evaporating * See page 437. t D. L. Randall, Am. Jour. Sci., [4], xxi, 128. 498 METHODS IN CHEMICAL ANALYSIS measured amounts with 10 cm. 3 of sulphuric acid to the fuming point of the acid, passing the solution through the reductor and titrating with permanganate. Reduction After Conversion of Chloride to Sulphate. Fed,. H 2 S0 4 [i:i.] Volume at titration. KMnO 4 . Fe found. Variation from average. cm. 3 cm.* cm. s cm. 8 grin. grm. 75 75 75 25 25 25 750 750 750 70.75 70.83 70.83 0.4867] 0-4873 I 0.4873 f 0.4872* r -0.0005 J +O.OOOI ] +O.OOOI 75 25 75 70.88 0.4876; (.+0.0004 Reduction of the Chloride. Fed,. cm. 3 Fe taken.* grm. H 2 S0 4 2.5 per cent. cm. 3 HC1 cone. cm. 8 Volume at titration. cm. 8 MnSO 4 . grm. KMnO 4 . cm. 8 Fe found, grm. *rror. grm. 75 0.4872 IOO o 750 70.81 0.4871 O.OOOI 75 0.4872 IOO o 750 70.75 0.4867 -0.0005 75 0.4872 IOO o 750 70.83 0.4873 +O.OOOI 75 0.4872 IOO 750 70.82 0.4872 o.oooo 75 0.4872 IOO o 750 70.83 0.4873 +O.OOOI IOO 0.6497 IOO o 750 94-43 0.6497 o.oooo IOO 0.6497 . IOO o 750 94-44 o . 6498 +0.0001 IOO 0-6497 IOO IO 750 94-44 0.6498 +O.OOOI IOO 0.6497 IOO 20 1000 94-53 0.6503 +o . 0006 IOO 0.6497 IOO 25 IOOO 94-53 o . 6503 +o . 0006 IOO 0.6497 IOO 25 IOOO 25 94-53 o . 6503 +0.0006 IOO 0.6497 IOO 25 IOOO 25 94.48 0.6500 +0.0003 IOO 0.6497 IOO 25 IOOO 5-oo 94-49 0.6501 +o . 0004 * Mean of results obtained by titration after the conversion of the chloride to sulphate. The Effect of Nitric Acid in the Titration of a Ferrous Salt by Potassium Permanganate. In Schneider's * method for the determination of manganese, permanganic acid is titrated with hydrogen peroxide in the pres- ence of nitric acid, and Ibbotson f and Brearley in their modifi- cation of this process recommend the use of standard ferrous ammonium sulphate instead of hydrogen peroxide. Blair t recommends the use of ferrous ammonium sulphate. Obviously it is of interest to know the exact effect of nitric acid upon solu- * Ding. Pol. Jour. f cclxix, 224. f Chem. News, Ixxxiv, 247. | Blair, Jour. Am. Chem. Soc., xxvi, 793. Also, Chemical Analysis of Iron. IRON 499 tions of the ferrous salt undergoing oxidation by permanganate, and an investigation of this matter by Randall * has shown that when more than 10 per cent by volume (20 cm. 3 in 200 cm. 3 ) of the concentrated acid is present oxidation of the ferrous salt takes place, as is made plainly evident by the change of color of the solution, low results, and uncertain end reaction. The error due to oxidation of the ferrous salt by nitric acid is, how- ever, in part counterbalanced by the reoxidation by perman- ganate of any nitrous acid which may have been produced. If the titration is made without unnecessary delay the presence of as much as 5 per cent by volume of nitric acid has no appreci- able effect upon the estimation of ferrous iron. Following are results of the titration of two different solutions of ferrous sul- phate in presence of varying amounts of nitric acid : Permanganate Titration in Presence of Nitric Acid. FeSO 4 . cm. 3 Dilution. cm. 3 H 2 S0 4 [i : i.] cm. 3 HNO 3 . cm. 3 Approx. n/io KMn0 4 . cm. 3 25 2OO 5 13.37 25 2OO 5 o 13-39 25 2OO 5 o I3.4I 25 2OO 5 o 13.38 T * 25 2OO 5 o 13.38 25 200 o 3 13.38 25 200 o 5 13.40 25 200 5 13.41 25 2OO 5 13.40 125 2OO 5 13.38 fas 2OO 5 o 13-47 25 200 5 o 13.50 1 25 200 5 o 13-49 H^ 25 2OO 5 13-53 25 2OO o 10 13.50 [20 2OO o 20 13.51 200 o 30 13 oo The Permanganate Estimation of Iron in Presence of Titanium. For analytical purposes, a ferric salt in solution is most easily and conveniently reduced to the ferrous condition by the action of zinc ; and where many determinations of iron are to be made, the use of the well-known Jones reductorf yields accurate results * D. L. Randall, Am. Jour. Sci., [4], xxiii, 139. t See page 347. 500 METHODS IN CHEMICAL ANALYSIS very rapidly. The use of zinc, whether in the flask or in the redactor, has, however, been precluded when the ferric salt is accompanied by titanic acid, for this substance is reduced with the iron and subsequently oxidized by the permanganate in the titration process. When, therefore, titanium is present with the iron, it has been customary to have recourse to other methods of reduction. In this event, either hydrogen sulphide or sulphur dioxide is substituted for the zinc to bring about the reduction of the ferric salt, since titanic acid is not reduced by these re- agents ; but the removal of the excess of hydrogen sulphide or of sulphur dioxide from solution without oxidation of the ferrous salt is not an easy or rapid process. Gooch and Newton * have studied the problem of adapting the ordinary convenient process of reducing the ferric salt by zinc to the estimation of iron in presence of titanium. It is obvious that to solve this problem it is only necessary to find and employ some reagent which shall be neutral toward the ferrous salt but capable of reoxidizing the titanium compounds formed by the reducing action of the zinc, and shall have no action on the per- manganate. Compounds of silver, copper or bismuth oxidize easily the reduced titanium salt; but the use of a compound of silver is precluded by the fact that it oxidizes also the ferrous salt to some extent as well as the titanium salt. Cupric salts and pure bismuth oxide prove, however, to be without action upon the ferrous salt. It is found that the violet color of the solu- tion containing the titanium compound produced by the action of zinc upon the titanium sulphate is discharged upon adding a little cupric sulphate to the solution and heating, and, after filtering, a drop of potassium permanganate gives its character- istic rose tint to the solution. It is found also that when cupric oxide is added to a similarly reduced solution of the titanium salt the characteristic color vanishes on shaking the solution. The following table contains the results obtained in titrating with potassium permanganate the ferrous salt left after reducing by zinc in small flasks carefully measured amounts of ferric sulphate and titanium sulphate, treating the mixture thus obtained with cupric sulphate or with cupric oxide, and filtering off the reduced copper and cuprous salt: * F. A. Gooch and H. D. Newton, Am. Jour. Sci., [4], xxiii, 365. . IRON 501 Oxidation of the Titanous Sulphate by Cupric Compounds: Titration of Ferrous Sulphate by Permanganate. Fe 2 O 3 taken, grm. TiO 2 taken, grm. Fe 2 O 3 found, grm. Error, grm. O.I37S 0.1375 0.1375 O. I O.I O.I 0.1378 0.1374 0.1377 +0.0003 } O.OOOI > +0.0002 ) Treated with CuSO 4 . 0.1375 0-1375 0.1375 O. I O.I O. 2 0.1378 0.1378 0.1382 +0.0003 ) +0.0003 > +0.0007 ) Treated with CuO. Similar experiments in which bismuth oxide was substituted for the copper oxide are also recorded. To the measured amount of ferric sulphate and titanium sulphate contained in a small flask, provided as usual with the funnel valve, zinc is added and the reduction effected in the ordinary manner. The titanium salt appears to act catalytically in this process, so that reduction goes on more easily and with less expenditure of zinc than in the similar reduction of the ferric salt taken by itself. After the zinc has disappeared, the solution, of characteristic violet color, is cooled in the flask, treated with a little bismuth oxide, gently shaken, filtered from the excess of bismuth oxide and the precipi- tated bismuth into about a liter of cold water, and titrated with standard potassium permanganate. Results are given in the following table: Oxidation of Titanous Sulphate by Bismuth Oxide: Titration of Ferrous Sulphate by Permanganate. Ferric sulphate. cm. 8 TiO 2 . grm. KMn0 4 . cm. a Fe-jOa taken. grrn. Fe 2 O 3 found, grm. Error, grm. IO 0.04 12.84 0.0993 0.0992 O.OOOI IO 0.06 12.85 0.0993 0.0993 O.OOOO 10 0.08 12.90 0.0993 0.0997 +0.0004 10 O. 12.90 0.0993 0.0997 +0.0004 10 O. 12.89 0.0993 o . 0996 +O.OOO3 IO 0. 12.85 0.0993 0.0993 O.OOOO IO 0. 12.80 0.0993 o . 0989 0.0004 IO O. 12.90 0.0993 0.0997 +0.0004 10 O.2 12.90 0.0993 0.0997 +0.0004 10 O. 2 12.89 o . 0993 0.0996 +0.0003 10 0.2 12.90 0.0993 0.0997 +0.0004 20 O. I 25.70 0.1986 0.1986 O.OOOO 502 METHODS IN CHEMICAL ANALYSIS It is evident that either cupric sulphate, cupric oxide, or bismuth oxide may be used to reoxidize the salt of titanium reduced by zinc, without affecting appreciably the ferrous salt in solution. Similar results are obtained when the ferric sulphate solution is passed through the column of amalgamated zinc in the Jones reductor.* The flask is kept cool in running water, a small amount of bismuth oxide added, the flask shaken and allowed to stand a few minutes, and the mixture filtered with the aid of the suction pump. In the cold solution free from dissolved oxygen there is little danger of reoxidation of ferrous sulphate, as has been shown by Peters and Moody. f The experimental results are given in the table. The Zinc Reductor: Bismuth Oxide: Permanganate. Ferric sulphate. cm. TiOj. grm. KMn0 4 . cm. 1 FezOa taken, grin. Fe,O, found, grm. Error, grm. 40 O.OI 46.80 0.3943 0-3943 O.OOOO 40 O.O2 46.79 0.3943 0.3942 O.OOOI 40 0.04 46.80 0.3943 0-3943 0.0000 40 0.06 46.83 0.3943 0.3946 +0.0003 40 o. 46.75 0.3943 0-3939 0.0004 40 o. 46.82 0.3943 0-3945 +O.OOO2 40 0. 46.78 0.3943 0.3941 O.OOO2 40 0. 46.80 0.3943 0-3943 0.0000 40 o. 46.75 0.3943 0-3939 0.0004 40 o. 46.80 0-3943 0-3943 O.OOOO 40 0.2 46.77 0-3943 0.3940 0.0003 40 0.2 46.81 0-3943 0-3944 +0.0001 40 O.2 46.85 0-3943 0-3947 +0.0004 The Estimation of Iron by Potassium Permanganate after Reduc- tion with Titanous Sulphate. Knecht \ was the first to recommend the use of titanium sesquioxide and its salts in volumetric operations where a rapid and powerful reducing agent is required, and later in collabo- ration with E. Hibbert published a method for the direct titra- tion of ferric chloride by a standard solution of titanous chloride, * See page 347. t See page 371. J Ber. Dtsch. Chem. Ges. xxvi, 166. Ibid., xl, 3819. IRON 503 using potassium sulphocyanate as an indicator, the reaction between the two salts taking place according to the following equation : FeCl 3 + TiCl 3 = FeCl 2 + TiCl 4 . According to these investigators the method yields excellent results, and rapidly. The only precautions necessary are that the solution of titanous chloride, being naturally very sensitive to the action of atmospheric oxygen, must, after having been boiled with marble to expel occluded oxygen, be kept under a constant pressure of hydrogen. It has been found, however, that even with such precautions the standard of the solution grad- ually changes and must be checked from time to time against known amounts of ferric iron. The proposal has therefore been made by Newton,* to reduce the iron by titanous sulphate, oxidize the excess of titanous sulphate, and titrate the remaining ferrous salt by permanganate. f A solution of titanous sulphate of convenient strength may be made up by mixing 20 grm. of commercial titanic acid with three times its own weight of a mixture of sodium and potassium carbonates and fusing in a platinum crucible, treating the melt (after being finely ground) in a platinum dish with hot concen- trated sulphuric acid, cooling, diluting a little, filtering through asbestos, treating with zinc until reduction is accomplished, and, while zinc is still left in the flask, filtering the solution quickly through a platinum cone into about two liters of freshly distilled water contained in a small reservoir connected with burette and hydrogen generator. To determine ferric sulphate in solution it is only necessary to add in the cold an excess of titanous sulphate, prepared as de- scribed, destroy this excess by treating with a little bismuth oxide, t filter from the excess of bismuth oxide into about a liter of cool distilled water, and titrate with permanganate. If an appreciable amount of hydrochloric acid is present in the solution of the ferric salt it is advisable to evaporate to dryness and con- vert the chloride to sulphate by use of concentrated sulphuric acid. Upon addition of the sulphuric acid a white pasty mass is formed which rapidly goes into solution on diluting with water * H. D. Newton, Am. Jour. Sci., [4], xxv, 343. t See page 500. t See page 501. 504 METHODS IN CHEMICAL ANALYSIS and warming. As titanous sulphate prepared in the manner described above always contains some iron, it is necessary to make a correction for this by treating with bismuth oxide an amount of the solution equal to that used, filtering, and running in potassium permanganate to color. This correction should not amount to more than o.i cm. 3 when working with 0.3 grm. of ferric oxide. If these simple precautions be taken, ferric iron may be determined with rapidity and exactness by reduction with titanous sulphate, treatment with bismuth oxide, and titration with potassium permanganate. Test results are given in the table. Permanganate Titration after Reduction with Titanous Sulphate. Fe 2 (S0 4 ) 3 . cm.* KMnO 4 . cm .s Fe 2 O 3 taken. grm. Fe 2 O 3 found, grm. Error, grm. 2O I3-46 0.1063 o. 1064 +O.OOOI 2O I3-42 0.1063 o. 1060 0.0003 20 13-44 0.1063 O.IO62 0.0001 20 13-44 0.1063 O.IO62 O.OOOI 20 I3-4I o . 1063 o. 1060 -0.0003 30 20. l8 0.1594 0.1594 o.oooo 30 2O.2O 0-1594 0.1596 +O.OOO2 30 2O. 2O 0.1594 0.1596 +0.0002 30 2O.22 0.1594 0.1598 +0.0004 30 2O. IQ 0.1594 0.1595 +O.OOOI 40 26.92 0.2127 O.2I27 O.OOOO 40 26.90 0.2127 9.2125 O.OO02 40 26.90 0.2127 0.2125 0.0002 40 26.93 0.2127 0.2128 -fo.oooi 40 26.90 0.2127 O.2I25 O.OO02 Separations of Iron by Volatilization in Gaseous Hydrogen Chloride. Metallic iron is easily acted upon by an excess of chlorine at moderately elevated temperatures with the formation of ferric chloride and by hydrochloric acid gas with formation of ferrous chloride. Out of contact with air, or moisture, both chlorides may be volatilized at appropriate temperatures the ferric chlo- ride below 200 C. ; the ferrous chloride at a bright red heat. If water vapor, or oxygen, or air, be present during the heating, both chlorides are partially decomposed with the formation of non-volatile residues, ferric oxide or ferric oxychloride. When IRON 505 ferric oxide is submitted to the action of hydrochloric acid gas at about 200 the greater part of the iron sublimes,* as ferric chloride, but a residue remains. At the outset the ferric oxide volatilizes quickly and abundantly in the form of the greenish vapor of ferric chloride, and if the operation is interrupted at this stage the residue which remains is nearly black, insoluble in water, slightly soluble in cold hydrochloric acid, and readily soluble in hot hydrochloric acid with the formation of ferric chloride. It is probably something analogous to the oxychlo- ride which Rousseau f identified as the product of the action of water upon ferric chloride at 275 to 300. This dark residue yields to the action of the hydrochloric acid at 180 to 200 only slowly; but ultimately the residue is essentially ferrous chloride. Little volatilization occurs within the range of temperature from 200 to 500. If the temperature of the oxide is 450 to 500 when the brisk current of acid begins to act, the whole mass of oxide is rapidly converted and volatilizes without residue, the production of the ferrous chloride (formed by dissociation of ferric chloride) being apparently kept at a minimum by the adjustment of equilibrium in the atmosphere of ferric chloride and chlorine resulting from the partial dissociation. If dissociation of ferric chloride to ferrous chloride is the cause of the formation of a residue at 200, the temperature of slow action, the introduction of chlorine into the atmosphere of hydrochloric acid gas should change the condi- tion of equilibrium and enable the ferric chloride to volatilize without dissociation. Gooch and Havens J find, as a matter of fact, that if a little manganese dioxide is added to the contents of the hydrogen chloride generator, so that the gas may carry with it a little chlorine, every trace of ferric oxide is volatilized from the boat at 180 to 200. The residue of ferrous chloride found at 1 80 to 200 when the hydrochloric acid is used alone is like- wise volatilized at the same temperature when the admixture of chlorine is made. These facts, that ferric oxide is completely volatile in hydro- chloric acid gas applied at once at a temperature of 450 to 500, and at 180 to 200 if the acid carries a little chlorine, open * Moyer, Jour. Am. Chem. Soc., xviii, 1029. t Compt. rend., cxvi, 118. J F. A. Gooch and Franke Stuart Havens, Am. Jour. Sci., [4], vii, 370. 506 METHODS IN CHEMICAL ANALYSIS the way to many analytical separations of iron from substances not volatile under these conditions. The separation of the iron oxide from various oxides proves to be complete at 450 to 500 if the mixture is submitted at once to the action of hydrochloric acid gas, or at 180 to 200 when chlorine is mixed with the hydrochloric acid. The temperature of red heat employed by Deville * is unnecessary if the mixed oxides are submitted at once to the action of hydrochloric acid at 450 to 500 without previous gentle heating in the acid atmosphere. The mixture of chlorine and hydrochloric acid is to be preferred, however, not only because the temperature of the reaction is lower but because it needs no regulation, while the danger of error arising from the liability of ferric chloride to dissociate, or from deficiency of oxidation in the oxide treated, or from mechanical loss due to too rapid volatilization is avoided. According to the procedure adopted, the mixed oxides, put in a porcelain boat which is placed in a wide combustion tube heated in a small furnace, are submitted to the action of dry hydrochloric acid gas, generated by dropping sulphuric acid upon a mixture of strong hydrochloric acid, common salt, and a small amount of manganese dioxide. - The gas is admitted at one end of the com- bustion tube and passed out at the other through a water trap, while the required temperature, best 200 to 300, is maintained by regulating the burners of the furnace. The time of action varies somewhat with the condition of the oxide to be volatilized and the temperature. Generally an hour's heating at 200 proves sufficient for the complete removal of o.i grm. of iron. At higher temperatures the action is more rapid; but the non- volatile oxide is liable to mechanical loss if the volatilization of the iron is too rapid. It is better, therefore, to make use of a lower temperature until the volatilization of iron is nearly com- plete, and then to raise the heat for a few minutes to insure the removal of the last traces of the volatile chloride. Results of this procedure follow. iron and Separations of iron oxide and aluminium oxide Aluminium. j^y fa e p roce dure outlined above are shown in the following table :f * Ann. Chim., [3], xxxviii, 23. t Gooch and Havens, loc. cit. IRON Volatilization of Ferric Chloride. 507 Fe 2 3 taken. grm. A1 2 3 taken. grm. A1 2 3 found. grm. Error, grm. Time, hours. Temperature. c. Atmosphere. 0.1000 0.2000 O.IO2O 0.2145 O IOOO O.IOI5 O. 1006 0.1015 0.1008 o.oooo o.oooo o.oooo +0.0002 o oooo \ I I 1 4 450-500 450-500 450-500 450-500 180200 HC1. HC1. HC1. HC1. HCl+Cla O.IOOO 0.1072 0.2045 o. 1050 o . 2008 0.1032 0.1013 0.1032 0.1023 O. 1007 0.1087 0.1032 o. 1015 0.1033 O.IOI9 o. 1006 O.IO87 o . oooo +O.OOO2 -j-o.oooi 0.0004 O.OOOI o.oooo I ly if 180-200 180-200 180-200 450-500 450-500 4ON 00 oOO 1 + 1 ++ 1 + 1 + 1 1 + OOOOOOOOOOOO O O O v> ^o O OOOOOOOOOOOO 4- + 1 ++ 1 1 ++ O^O O^iow loiorot^-N toO O w OOO\O IOVOM IOIOM l^t^f^O O O rOfO^fOfOt^ OOOMMMOOwOOO 666666666666 OOOWMMOOMOOO 666666666666 OOOOOOOOOOOO + 1 ++ 1 1 + 1 1 1 1 MHHMMMVOUDMlOCOfOI-l 666666666666 00 00 00 00 00 O O\00 MMMMMIOIOMIOCOPOM MMMMMOOMO^S^M 666666666666 IRON 513 brown ferric hydroxide, is in large measure obviated, the blue of the starch iodide being quite clear against the pale green chromic hydroxide. After standing for from fifteen to twenty minutes the solution is made alkaline with sodium bicarbonate and an excess of standard iodine solution added. This is allowed to stand in a stoppered flask for from one-half hour to an hour, the excess of iodine being then removed with arsenious acid and the solution titrated to color with iodine after the addition of starch. According to Edgar's procedure the first step in the calculation is the reduction to terms of tenth normal solution of the figures of titration obtained in the processes (I), (II) and (III). It is evident that the subtraction of the titration figure of (III) from that of (I) gives the number of cubic centimeters corresponding to reduction of the vanadium pentoxide to tetroxide, while the subtraction of this result from that of (II) gives the number equivalent to the reduction of the ferric salt. By multiplying these figures by the amounts of vanadic acid, chromic acid and ferric oxide equivalent to I cm. 3 of the n/io reagent the weights of these substances present are obtained. An example of this procedure is given below: Titratiom. cm. 8 H/IO factor. cm.* (I) 31.15 Xi.ioo =34.26 (tt/io)^V 2 O 5 +CrO,. (II) 23.4 Xi.ioo =25.74 (w/io)~V 2 O 5 +Fe 2 O,. (Ill) 30.00 8.74X1.000 =21.26 (n/io)~CrO 3 . 34.26 21.26 =13.00 =3=V 2 O 5 . 25.7413.00 =12.74 = Fe 2 O 3 . 21 . 26X0.003334 (factor for CrO 3 ) =0.0709 grm. CrO 3 found. 13.00X0.00912 (factor for V 2 O 6 ) =0.1185 grm. V 2 O 6 found. 12.74X0.00799 (factor for Fe 2 O 3 ) = 0.1018 grm. Fe 2 O 3 found. Results obtained by this process are given in the table on the preceding page. The calculation given above is that of the first determination of the table. INDEX OF AUTHORS* PAGE Ashley, R. Harmon. Dithionic acid and dithionates, determination of. 369 Sulphites, estimation of, in alkaline solution 366 Austin, Martha. Arsenic, determination of, as magnesium pyroarsenate 288 Beryllium, not determinable as pyrophosphate 153 Cadmium, estimation of, as pyrophosphate 190 Magnesium, determination of, as pyrophosphate (with F. A. Gooch) 156 Manganese, determination of separated as carbonate 481 as oxide (with F. A. Gooch) 478 by chlorate process (with F. A. Gooch) 487 as pyrophosphate (with F. A. Gooch) . 482 as sulphate (with F. A. Gooch) 477 Phosphoric acid, determination of, as magnesium pyrophosphate (with F. A. Gooch) 282 Zinc, estimation of, as pyrophosphate 185 Beyer, F. B. Filtering crucible in electrolytic analysis (with F. A. Gooch^ 13 Lead, electrolytic determination of (with F. A. Gooch) . . 252 Manganese, electrolytic determination of (with F. A. Gooch) 485 Blake, J. C. Bromates, iodometric estimation of (with F. A. Gooch). . 471 Gold, red colloidal 150 Blumenthal, Philip L. Barium and strontium, detection of, associated with calcium and lead (with P. E. Browning) 160 Lead, detection of, in sulphates (with P. E. Browning) 252 Bosworth, Rowland S. Silver, gravimetric determination of, as chro- mate (with F. A. Gooch) 136 iodometric determination of, precipitated as chromate (with F. A. Gooch). . . . 140 iodometric determination of, reduced by arsenite 143 Boynton, C. N. Barium, estimation of, precipitated by acetyl chloride in acetone (with F. A. Gooch) 175 separation of, from calcium and magnesium (with F. A. Gooch) 177 Breckenridge, J. E. Perchloric acid, preparation of (with D. A. Kreider) 76 Sodium, detection of (with D. A. Kreider) 74 Brooks, F. T. Chlorine, bromine and iodine, detection of (with F. A. Gooch) 440 Brown, James. Oxidations by permanganate in presence of chlorides . . 52 Browning, Philip E. Arsenic acid, iodometric estimation of (with F. A. Gooch) 291 Arsenic, antimony and tin, estimation of, by ferricyanide and permanganate (with H. E. Palmer) 322 Barium, precipitation of, as sulphate 170 Barium and calcium, estimation of, separation by amyl alcohol on nitrates 166 * Workers in the Kent Chemical Laboratory. 515 516 INDEX OF AUTHORS PAGE Browning, Philip E. Barium and strontium, detection of, associated with calcium and lead (with P. L. Blumenthal) 160 separation of, by amyl alcohol on bromides. . . 167 Barium, strontium and calcium, separation of, by amyl alcohol on nitrates 162 Barium with strontium, and calcium, separation of, and estimation by amyl alcohol on nitrates. . 166 . Cadmium, estimation of, as oxide (with L. C. Jones) 188 Caesium and rubidium, estimation of, as acid sulphates 106 Cerium, estimation of, by ferricyanide and per- manganate (with H. E. Palmer) 249 separation of, from cerium earths (with E. J. Roberts) 244 iodometric estimation of, digestion process '(with Hanford and Hall).. . . 246 distillation proc- ess (with Hanford and Hall) 247 Cerium oxalate, estimation of, by permanganate (with L. A. Lynch) 248 Chromic acid, iodometric determination of 407 Copper, determination of, as cuprous iodide 114 separation from cadmium as cuprous iodide 115 Ferricyanides, detection of (with H. E. Palmer) . . 275 Ferrocyanides, detection of (with H. E. Palmer) . 275 Fluorine, detection of 432 Iodine, iodometric determination of, in haloid salts (with F. A. Gooch) 457 Lead, detection of, in sulphates (with P. L. Blu- menthal) 252 Magnesium, separation of, from alkalies (with W. A. Drushel) 158 Nickel, detection of, in presence of cobalt (with J. B. Hartwell) . 491 Potassium, estimation of, as pyrosulphate 92 Silicon, detection of, in silicates and fluosilicates 241 Sodium, estimation of, as pyrosulphate 79 Strontium and calcium, detection of 163 separation of, and esti- mation 166 Sulphides, sulphates, sulphites and thiosulphates, detection of (with Ernest Howe) 363 Sulphocyanates, detection of (with H. E. Palmer) 276 Tellurium, separation of, from selenium (with W. R. Flint) 402 Tellurium dioxide, precipitation of (with W. R. Flint) 402 Thallium, determination of, as sulphates 219 estimation of, by precipitation as thallic hydroxide (with H. E. Palmer) 220 by ferricyanide and per- manganate (with H. E. Palmer) 223 INDEX OF AUTHORS 517 PAGE Browning, Philip E. Thallium, determination of, as chromate (with G. P. Hutchins) 221 iodometrically (with G. P. Hutchins) 222 Vanadium, estimation of. as silver vanadate (with H. E. Palmer) 328 Vanadic acid, ipdometric estimation of, reduced by organic acids (with R. J. Goodman) 341 by hydriodic acid 343 by hyrlrobromic acid 345 Volatile products, removal of, without loss of non- volatile material (with F. A. Gooch) 6 demons, C. F. Selenious acid, determination of, by permanganate (with F. A. Gooch) 382 Curtis, F. W. Vanadic acid, reduction of, by hydrochloric acid (with F. A. Gooch) 334 hydrobromic acid (with F. A. Gooch) 335 hydriodic acid (with F. A. Gooch) 337 Danner, E. W. Antimony, separation of, from arsenic, and estimation (with F. A. Gooch) . . 311 Oxygen, loss of, in oxidations by permanganate (with F. A. Gooch) 43 Tellurous Acid, determination of, by permanganate (with F. A. Gooch) 394 Drushel, W. A. Lanthanum, estimation of, precipitated as oxalate. . . . 218 Magnesium, separation of, from alkalies (with P. E. Browning) 158 Potassium, estimation of, as cobalti-nitrite 93 in the pure salt 94 in mixtures of salts. ... 95 in fertilizers 96 in soils 97 in animal fluids 98 Eddy, Ernest A. Magnesium, determination of, as oxide (with F. A. Gooch) . . 154 Edgar, Graham. Chromic acid and vanadic acid, estimation of 409 Iron and vanadium, estimation of 508 Ferric iron, vanadic acid and chromic acid, estima- tion of 510 Molybdic acid and vanadic acid, determination of, by reductions and oxidations 427 Vanadic acid and antimonic acid, estimation of. ... 350 and arsenic acid, estimation of 350 . reduction of, by zinc, with use of ferric sulphate (with F. A. Gooch) 349 Volatile products, distillation and absorption of 5 Ensign, J. R. Bromine (and chlorine), determination of, in mixtures of alkali bromides (and chlorides) with iodides (with F. A. Gooch) 452 Evans, P. S. Jr. Selenic acid, iodometric determination of, reduced by hydrochloric acid (with F. A. Gooch) 385 Fairbanks, Charlotte. Halogens, determination of, by electrolytic re- duction of silver salts (with F. A. Gooch) .... 459 Molybdic acid, iodometric estimation of (with F. A. Gooch) 415 distillation process (with F. A. Gooch) 416 reoxidation of residue (with F. A. Gooch) 420 518 INDEX OF AUTHORS PAGR Fairbanks, Charlotte. Phosphorus, iodometric determination of, in iron 283 Feiser, J. P. Silver, electrolytic determination of, in oxalate solution (with F. A. Gooch) 138 Flint, William R. Tellurium, separation of, from selenium (with P. E. Browning) 402 Tellurium dioxide, precipitation of (with P. E. Browning) 402 Flora, Charles P. Cadmium, estimation of, as oxide 188 electrolytic estimation of 191 from sulphuric acid solution 191 from solutions of acetates 192 from solutions of cyanides 193 from solutions of pyrophosphates and orthophosphate 194 Gilbert, R. D. Ammonium vanadate, precipitation of, by ammonium chloride (with F. A. Gooch) 326 Vanadic acid, estimation of, by Jones reductor and silver sulphate (with F. A. Gooch) 346 Gillespie, David H. M. Alkali hydroxide, determination of, by reaction with iodine (with C. F. Walker) 71 Gooch, F. A. Aluminium, estimation of, by precipitation as chloride (with F. S. Havens) 214 separation of, from iron (with F. S. Havens) 214 Aluminium sulphate, hydrolysis of, in bromide-bromate mixture (with R. W. Osborne) 70 Antimonic acid and arsenic acid, iodometric determina- tion of (with H. W. Gruener) 308 Antimony, separation of, from arsenic, and estimation (with E. W. Danner) 311 Arsenic, antimony and tin detection of (with B. Hodge) . 312 (with I. K. Phelps). 316 Arsenic acid, iodometric estimation of (with P. E. Browning) 291 (with J. C. Morris) 294 estimation of small amounts, precipitated as ammonium magnesium arsenate (with M. A. Phelps)- 290 estimation of minute amounts, in copper (with H. P. Moseley) 301 separation of, from copper, as ammonium magnesium arsenate (with M. A. Phelps) 303 Barium, estimation of, as chloride (with C. N. Boynton) 175 separated from calcium and magnesium (with C. N. Boynton) 177 Boric acid, gravimetric determination of (with L. C. Jones) 201 with sodium tungstate as re- tainer (with L. C. Jones) . . 204 Bromates, iodometric estimation of (with J. C. Blake) . . 471 Bromine (and chlorine), determination of, in mixtures of alkali bromides (and chlorides) with iodides (with y. R. Ensign) " 452 Carbon dioxide, determination of, by ignition (with S. B. Kuzirian) 226 precipitation and gravimetric estimation of (with I. K. Phelps) 228 Chlorates, iodometric estimation of (with C. G. Smith) . . 463 Chlorine, determination of. in mixtures of alkali chlorides and iodides (with F. W. Mar) 449 fixation of, on silver anode (with H. L. Read) . . 20 INDEX OF AUTHORS 519 PAGE Gooch, F. A. Chlorine, bromine and iodine, detection of (with F. T. Brooks) 440 Chromium, estimation of, as silver chromate (with L. H. Weed) 406 Copper, determination of, by titration of oxalate (with H. L. Ward) .....".. 126 electrolytic determination of, by rotating cathode (with H. E. Medway) . . 116 iodometric estimation of (with F. H. Heath) .... 121 Filtering crucible in electrolytic analysis (with F. E. Beyer) 13 Gaseous products, evolution of, without mechanical loss (with C. F. Walker) 6 Halogens, determination of, by electrolytic reduction of silver salts (with C. Fairbanks) 459 Iodides, analysis of, by iodic acid (with C. F. Walker) . . . 454 gravimetric determination of, by absorption in silver (with C. C. Perkins) 444 Iodine, iodometric determination of, in haloid salts (with P. E. Browning) 457 standardization of, by silver (with C. C. Perkins) 27 Iron, estimation of, in presence of titanium (with H. D. Newton) . 500 separations of, by volatilization in hydrogen chlo- ride (with F. E. Havens) 504 Lead, electrolytic determination of (with F. B. Beyer). . 252 Magnesium, determination of, as oxide (with E. A. Eddy) 154 as pyrophosphate (with M. Austin) 156 Manganese, determination of, by chlorate process (with ' M. Austin) 487 by electrolysis (with F. B. Beyer)... . 485 as oxide (with M.Austin) 478 as pyrophosphate (with M. Austin) 482 as sulphate (with M. Aus- tin) 477 Molybdic acid, iodometric estimation of, by distillation process (with C. Fairbanks) 416 (with J. T. Norton, Jr.) 418 by iodine oxidation of residue (with C. Fairbanks) .... 416 by permanganate oxidation of residue, (with O. S. Pulman, Jr.) 422 Nitrates, estimation of, by ignition (with S. B. Kuzirian) 256 iodometric estimation of (with H. W. Gruener), 263, 266, 268 Nitrates and chlorates, estimation of (with H. W. Gruener) 273 Oxygen, loss of, in permanganate titrations (with L. W. Danner) 43 (with C. A. Peters) . .48, 50 Perchlorates, detection of, in association with chlorides, chlorates and nitrates (with D. A. Kreider) 465 Phosphoric acid, determination of, as magnesium pyro- phosphate (with M. Austin) 282 Potassium, spectroscopic detection and determination of (with T. S. Hart) 80 Precipitates, purification of 10 Rotating cathode (with H. E. Medway) 1 1 520 INDEX OF AUTHORS PACK Gooch, F. A. Rubidium, spectroscopic determination of (with J. I. Phinney) 102 Selenic acid, iodometric determination of, reduction by hydrochloric acid (with P. S. Evans, Jr.) 385 by hydrobromic acid (with W. S. Scoville).. 386 by hydriodic acid (with VV. C. Reynolds).. . 388 by differential method (with A. W. Peirce). . 380 Selenious acid, determination of, by permanganate (with C. F. demons) 382 iodometric determination of (with W. G. Reynolds). 378 by differential method (with A. W. Peirce) . 380 Selenium and tellurium, separation of, by difference in volatility of bromides (with A. W. Peirce) 390 Silver, gravimetric determination of, as chromate (with R. S. Bosworth) 136 electrolytic determination of, in oxalate solution (with J. P. Feiser) 138 iodometric determination of, as chromate (with R. S. Bosworth) 140 Telluric acid, iodometric determination (with J. How- land) 401 Teliurous acid, determination of, by permanganate (with E. W. Danner) 394 in presence of bro- mide (with C. A. Peters) 397 in presence of chloride (with C.A.Peters).. 396 by precipitation as iodide (with W. C. Morgan). . . 398 iodometric (with C. A. Peters) 399 Vanadic acid, estimation of, precipitated by ammonium chloride (with R. D. Gilbert). ........ 326 reduced by hydrochloric acid (with L. B. Stookey) 330 by hydrochloric acid (with R. W. Curtis) . 334 by hydrobromic acid (with R.W. Curtis) 335 by hydriodic acid (with R.W. Curtis) 337 by Jones reductor, with silver sul- phate (with R. D. Gilbert) 346 by zinc, with ferric sulphate (with Graham Edgar) 349 Volatile products, removal of, without loss of non- volatile material (with P. E. Brown- ing) 5 INDEX OF AUTHORS 525 PAGE Phinney, J. I. Barium sulphate, purification of 172 Rubidium, spectroscopic determination of (with F. A. Gooch) 102 Pulman, O. S., Jr. Molybdic acid, iodometric estimation of, by reduc- tion with hydriodic acid and reoxidation by per- manganate (with F. A. Gooch) ^ 422 Phosphoric acid, estimation of, precipitated as uranyl phosphate 286 Uranium, determination of, by aid of the Jones reductor 430 Randall, D. L. Ferric chloride, behavior of, in Jones reductor 497 Ferrous sulphate, permanganate titration of, in presence of nitric acid 499^ Mercurous salts, titration of, by permanganate 198 Molybdic acid, reduction of, in Jones reductor. 424 Phosphoric acid, volumetric estimation of, precipitated as ammonium phosphomolybdate 285 Read, H. L. Fixation of chlorine on silver anode (with F. A. Gooch) . . 20 Reynolds, W. G. Selenic acid, iodometric determination of (with F. A. Gooch) 388 Selenious acid, iodometric determination of (with F. A. Gooch) 378 Roberts, Charlotte F. Nitrates, estimation of 260 Nitrates and chlorates, estimation of 273. Nitrates and nitrites, estimation of 271 Nitrites, estimation of 271 Permanganate standardization, in iron analysis 495 Roberts, Edwin J. Cerium, separation of, from cerium earths (with P. E. Browning) 244, Scoville, W. S. Selenic acid, iodometric determination of (with F. A. Gooch) 386 Smith, C. G. Chlorates, iodometric determination of (with F. A. Gooch) 463 Stookey, L. B. Vanadic acid, reduction of, by hydrochloric acid, and estimation (with F. A. Gooch) 330 Thorne, Norman C. Barium, separation of, as bromide, from calcium and magnesium 180 Barium bromide, precipitation of, by ether-hydro- bromic acid . 1 79 Van Name, R. G. Copper, gravimetric determination of, as sulpho- cyanate. 108 separation of, as sulphocyanate, from bis- muth, antimony, tin and arsenic 112 Sulphocyanates, gravimetric determination of 276 volumetric determination of 279* Walker, Claude F. Alkali hydroxides, determination of, by reaction with iodine (with D. H. M. Gillespie) 70 Gaseous products, removal of, without mechani- cal loss (with F. A. Gooch) 6 Iodides, analysis of, by iodic acid (with F. A. Gooch) 454 Ward, H. L. Copper, determination of, by titration of oxalate (with F. A. Gooch) 126 associated with lead 135 separated from cadmium, arsenic and iron 132 Lead, estimation of, by titration of oxalate 254, Nickel, estimation of, by titration of oxalate 490 Zinc, estimation of, by titration of oxalate 187 Way, Arthur Fitch. Separations by volatilization in hydrogen chloride, iron and beryllium (with F. S. Havens) 507 iron and chromium (with F. S. Havens) 507 iron and zirconium (with F. S. Havens) 508 526 INDEX OF AUTHORS PAGE Weed, L. H. Alkali hydroxides and carbonates, iodometric determina- tion of (with I. K. Phelps) 60 Chromium, estimation of, as silver chromate (with F. A. Gooch) 406 Organic acids and acid anhydrides, as standards in neutralization processes (with I. K. Phelps) 56 as standards in iodometric processes (with I. K. Phelps) , 56, 59 INDEX OF SUBJECTS PAGE Acidimetry 54 Acids, determination of, by neutralization . . 54, 56 by iodide-iodate mixture 59, 7 2 liberated in hydrolysis 61 standards in neutralization processes. 56, 57, 58 Acid anhydrides, as standards in neutralization processes 54 59 succinic anhydride 57 phthalic anhydride 58 Alkalimetry ^ 54 Alkali carbonates, iodometric determination of 60 Alkali hydroxides, determination of, by neutralization 54, 56 by succinic acid 54 by organic acids 56 by acid anhydrides 56 iodometric, by iodide-iodate mix- ture 59 by iodine 70 Aluminium, determination of, precipitated by ether-hydrochloric acid. . 214 separation of, from beryllium 216 bismuth 217 copper 217 iron 214 mercury ... 217 zinc 216 separation of, from iron, by gaseous hydrogen chloride 506 Aluminium sulphate, hydrolysis of, in bromide-bromate mixture 70 in iodide-iodate mixture 62 Alums, analysis of, determination of basic alumina, or free acid 67 Ammonium sulphate, hydrolysis of, in iodide-iodate mixture 65 Antimonic acid, iodometric determination of 308 and arsenic acid, estimation of 308 and vanadic acid, estimation of 325, 350 Antimony, detection of, by hydrochloric acid and bromide 316 by hydrochloric acid and iodide 313 estimation of, by ferricyanide and permanganate, associated with arsenic and tin .322, 323 iodometric, associated with copper 318 separation of, from arsenic, and determination of, by hydro- chloric acid and iodide 311 from copper precipitated as sulphocyanate .... 1 12 Arsenic, detection of, by hydrochloric acid and bromide 316 by hydrochloric acid and iodide 313 determination of, as magnesium pyroarsenate 288 small amounts precipitated by freezing.. 290 minute quantities in copper 301 by ferricyanide and permanganate, asso- ciated with antimony and tin 322, 324 iodometric, associated with copper 318 separation of, from antimony, by hydrochloric acid and iodide. 311 from copper precipitated as oxalate.i3i, 133, 134, 135 as sulphocyanate 112 527 528 INDEX OF SUBJECTS PAGE Arsenic acid, estimation of, iodometric 291 , 294 reduction by hydriodic acid, with titration of liberated iodine 295 and antimonic acid, estimation of 308 and vanadic acid, estimation of 325, 350 Arsenic trioxide, use of, as an iodometric standard 29 in standardizing permanganate, without iodine. 411 with iodine. ... 412 Barium, detection of, associated with calcium and lead 160 estimation of, as sulphate in presence of hydrochloric acid 168 nitric acid and aqua regia 170 gravimetric estimation of, precipitated as oxalate from alcohol 180 precipitation of, by acetyl-chloride in acetone, and estimation. . 175 by ether-hydrobromic acid, and estimation .... 179 by ether-hydrochloric acid, and estimation 174 separation of, from calcium, and estimation by action of amyl alcohol on the nitrates 162, 166 from calcium and magnesium, and estimation, by acetylchloride in acetone 177 by ether hydrobromic acid 180 by ether hydrochloric acid 175 from strontium, and estimation, by action of amyl alcohol in the bromides 167 volumetric estimation of, precipitated as oxalate from alcohol . . 184 Barium with strontium and calcium, estimation of, by action of amyl alcohol on the nitrates 166 Barium, strontium and calcium, separation of, by action of amyl alcohol on the nitrates 162 Barium sulphate, purification of, after precipitation 172 Beryllium, conversion of chloride to oxide 1 53 not determinable by precipitation as ammonium beryllium phosphate : 153 separation from aluminium, by ether-hydrochloric acid. ... 216 from iron, by gaseous hydrogen chloride 154, 507 Bismuth, separation of, from aluminium, by, ether-hydrochloric acid, and determination 217 from copper precipitated as sulphocyanate . ... 112 Boric acid, acidimetric estimation of 205 neutralization of stronger acids 206 strengthening of acidity by mannite.. 208 gravimetric estimation, with calcium oxide as retainer 201 with sodium tungstate as retainer . . . 204 iodometric determination 210 Bromates, iodometric estimation of, reduction by arsenate-iodide mixture 475 by arsenious acid 474 by hydriodic acid 471 Bromide-bromate mixture, reaction of, with aluminium sulphate 70 Bromine, determination of, in benzol derivatives 447 gravimetric determination of, by liberation of iodine and ab- sorption of that element by silver 443 Bromine and chlorine, determination of, in alkali, bromides, chlorides, and iodides 452 Bromine, chlorine and iodine, detection of 440 Cadmium, electrolytic determination of, with the rotating cathode 191 from solutions of acetates 192 cyanides 193 INDEX OF SUBJECTS 529 PAGE Cadmium, electrolytic determination of, in sulphuric acid 191 in pyrophosphates and ortho- phosphates 194 estimation of, as oxide, precipitated as carbonate 188 as hydroxide 189 as pyrophosphate 190 separation of, from copper precipitated as oxalate 131 , 133, 134 Caesium, estimation of, as acid sulphate 106 Calcium, detection of, with strontium . . . 163 separation of, from barium, and estimation, by the action of amyl alcohol on the nitrates 162, 166 from strontium, and estimation, by the action of amyl alcohol on the nitrates 162, 164 from strontium and barium, and estimation, by amyl alcohol on the nitrates 162, 166 volumetric estimation of, precipitated as oxalate. ;-. 181 Carbon, determination of, in organic substances, by combustion in the wet way 234 by permanganate 234 by chromic acid 236 Carbon dioxide, determination of, in carbonates, by loss in action of acid 225 by ignition with sodium paratungstate 226 iodometric 231, 232 in oxidations by permanganate, iodo- metric 234 in oxidations by chromic acid, iodo- metric 236 precipitation of, and gravimetric determination 228 Cerium, iodometric estimation of, by digestion process 246 by distillation process 247 separation of, from cerium earths, by bromine and alkali hydroxide 244 volumetric estimation of, by ferricyanide and permanganate . . 249 precipitated as oxalate 248 Chlorates, estimation of, iodometric 463 reduced by ferrous sulphate 462 Chlorates and nitrates, determination of, iodometric and gas-volumetric 273 Chlorine, determination of, in alkali chlorides and iodides, by distilla- tion with ferric sulphate 449 with nitrite 451 in benzol derivatives 448 with bromine, in alkali chlorides and iodides 452 fixation of, on silver anode 20 gravimetric determination of, by liberation of iodine and ab- sorption of that element by silver 443 bromine and iodine, detection of 440 Chromium, gravimetric estimation of, as silver chromate 406 separation of, from iron by gaseous hydrogen chloride 508 in chromic condition, volumetric estimation of 413 Chromic acid, estimation of, with ferric iron and vanadic acid 510 iodometric estimation of 407 and vanadic acid 409, 41 1 Chromic sulphate, hydrolysis of, in iodide-iodate mixture 63 Cobalt, separation of, from nickel, by ether-hydrochloric acid 492 Cobalt sulphate, hydrolysis of, in iodide-iodate mixture 63 Copper, determination of, as cuprous iodide 114 by titration of precipitated oxalate 123, 125 electrolytic determination of, on rotating cathode 116 on rotating cathode of silver 117 gravimetric determination of, as sulphocyanate 108 530 INDEX OF SUBJECTS PAGE Copper, iodometric estimation of 1 18 iodometric estimation of, associated with arsenic 318 with antimony 318 precipitation as oxalate 125 separation of, from cadmium, by precipitation as cuprous iodide 1 14 as oxalate, associated with lead 135 from arsenic, cadmium, iron, tin, zinc 131 by desiccation process . 132 by acetic acid process. . 134 as sulphocyanate, from antimony, arsenic, bis- muth, and tin 112 Copper oxalate, solubility of 125 prevention of supersaturation of solutions of 129 Dithionic acid and dithionates, determination of 369 Electrolysis with the filtering crucible, with filtration 16 continuous 17 subsequent 13 Electrolytic processes 1 1 Ferricyanides, detection of, with ferrocyanides and sulphocyanides . . . .275, 276 Ferric chloride, behavior of, in the Jones reductor 497 Ferric sulphate, hydrolysis of, in iodide-iodate mixture 63 Ferrocyanides, detection of, with ferricyanides and sulphocyanides. . .275, 276 Ferrous sulphate, hydrolysis of, in iodide-iodate mixture 63 titration of, in presence of nitric acid 498 Filtering crucible, in electrolytic analysis 13 Fluorine, detection of 432 iodometric determination of 439 evolved as silicon fluoride, estimation of 436 Fluosilicic acid, acidimetric estimation of 432 in alcoholic solution 433 in water solution 434 iodometric estimation of 435 Force pump (Kreider) 8 Gaseous products, determination of, by loss I prevention of mechanical loss in evolution of 6 transfer under pressure 7 Gold, determination of, colorimetric 150 electrolytic 145 iodometric (small amounts) 146 Halogens, determination of, by electrolytic reduction of silver salts. . . . 459 in benzol derivatives 447 Hydrochloric acid, electrolysis of, with silver anode 20 Hydrolysis, in bromide-bromate mixture 70 in iodide-iodate mixture 61 Hydrogen, determination of, by loss 2 lodic acid, use of, in analysis of iodides 454 Iodides, determination of, by liberation of iodine and absorption of that element by silver 446 by use of iodic acid 454 Iodide-iodate mixture, reaction of, with salts alums 67 aluminium sulphate ... 62 ammonium sulphate. . . 65 chromic sulphate 63 cobalt sulphate 63 iron sulphates 63 nickel sulphate 64 INDEX OF SUBJECTS 531 PAGE lodide-iodate mixture, reaction of, with salts stannic sulphate 63 zinc sulphate 64 use of, in determination of free acids 59 of acids liberated by hydrolysis 61 of alkali hydroxides and carbonates .... 60 Iodine, gravimetric determination of, by absorption in silver 444 reaction of, with alkali hydroxides 70 standardization of, by arsenic trioxide 29 by silver 27 in haloid salts, iodometric determination of 457 in iodides, gravimetric determination of, by liberation and ab- sorption in silver 446 bromine and chlorine, detection of 440 Iodometric processes 27 Iron, estimation of, in presence of titanium 499 separation of, from copper precipitated as oxalate ... .131, 132, 133, 134 by gaseous hydrogen chloride 504 from aluminium 506 from beryllium 507 from chromium 507 from zirconium 508 (ferric), estimation of, in presence of vanadium 508 vanadic acid and chromic acid 510 reduction of titanous sulphate, and estimation by per- manganate 502 volumetric determination, by thiosulphate and iodine .... 492 (ferrous;, permanganate titration of, in presence of chlorides. . . .48, 497 of nitric acid. . . . 498 analysis, standardization of permanganate in 495 chloride, behavior of, in Jones reductor 497 Lanthanum, estimation of, by permanganate titration, precipitated as oxalate 218 Lead, detection of, separated as sulphate 252 electrolytic determination of, as dioxide 252 estimation of, by permanganate titration of oxalate 254 Liquids, transfer of, under pressure 7 Magnesium, determination of, by precipitation as ammonium magne- sium carbonate and ignition 154 as pyrophosphate 156 separation of, from alkalies, by arsenate process 158 Manganese, determination of, as oxide 478 as pyrophosphate 482 as sulphate 477 electrolytic determination of ' 485 precipitation of, by chlorate process, and determination .... 487 separation of, as carbonate, and determination as oxide .... 481 Mechanical processes I Mercury, determination of, by permanganate titration after precipi tation with ammonium oxalate 1 97 of mercurous salts 198 by titration with thiosulphate and iodine. . . 196 gravimetric, as mercurous oxalate 195 separation of, from aluminium by ether-hydrochloric acid, and determination 217 Molybdic acid, estimation of, by reduction in Jones reductor, use of ferric alum, and permanganate titration of residue, 424, 426 gravimetric estimation of, by liberation of iodine and absorption of that element by silver 414 532 INDEX OF SUBJECTS PAGE Molybdic acid, iodometric estimation of 415 by digestion method 415 by distillation process 416 by reduction with hydriodic acid and reoxidation of residue, by iodine 420 by reduction with hydriodic acid and reoxidation of residue, by permanganate 42 1 and vanadic acid, determination of, by reductions and oxidations 427 Neutralization processes, with use of acid anhydrides as standards 56 organic acids as standards 56 succinic acid, as standard 54 Nickel, detection of, in presence of cobalt 491 determination of, electrolytic 489 by precipitation as oxalate, and permanganate titration -.: 49 separation of, from cobalt, by ether-hydrochloric acid 492 Nickel sulphate, hydrolysis of, in iodide-iodate mixture 64 Nitrates, estimation of, by ignition with sodium paratungstate 256 by reduction with ferrous chloride and gas- volumetric estimation of nitrogen dioxide .... 260 by reduction with ferrous sulphate and per- manganate titration 258 iodometric estimation of, by action of manganous chloride. . . . 263 by distillation with antimony tri- chloride . 268 by distillation with iodide and phosphoric acid 266 Nitrates and chlorates, iodometric and gas- volumetric estimation of .... 273 Nitrates and nitrites, iodometric and gas volumetric estimation of 272 Nitrites, gas-volumetric determination of, with use of manganous chloride ...'.. 271 iodometric determination of, with use of manganous chloride. . 271 with potassium iodide and arsenite. . 269 Nitrites and nitrates, iodometric and gas- volumetric determination of. . 272 Nitrogen, determination of, liberated by hypobromite . 256 Organic acids and anhydrides, use of, as standards in neutralization processes 54 benzoic acid 58 malonic acid 57 phthaiic acid 58 succinic acid 57 phthaiic anhydride 58 succinic anhydride 57 Organic substances, combustion of, in the wet way 234 by chromic acid 236 by permanganate 234 Oxidation processes 41 Oxidizers, gravimetric determination of, by liberation of iodine and absorption of that element by silver 361 Oxygen, amount used in oxidation by chromic acid, and indirect estima- tion of oxygen content of organic substances 239 iodometric estimation of, in air 355 in water solution 360 loss of, in oxidations by permanganate 4 2 INDEX OF SUBJECTS 533 PAGE Perchlorates, detection of, associated with chlorides, chlorates and nitrates 465 iodometric determination of 467 Perchloric acid, preparation of, for potassium determination 88 for sodium test 76 Permanganate, standardization of 41, 42, 362, 495 Persulphates, determination of, by arsenate-iodide method 370 by method of Griitzner 372 by method of LeBlanc and Eckardt. ... 371 by method of Mondolfo 374 by method of Namias 374 Phosphoric acid, estimation of, as magnesium pyrophosphate 282 by permanganate, precipitated as uranyl phosphate 286 iodometric estimation of, precipitated as ammonium phosphomolybdate 285 Phosphorus in iron, iodometric determination of, precipitated as ammo- nium phosphomolybdate 283 Potassium, detection of, spectroscopic 80 effect of sodium salts on 82 determination of, spectroscopic 83 in presence of sodium salts 85 estimation of, as cobalti-nitrite by gravimetric process. ... 95 by volumetric process 93, 95 in animal fluids (blood, lymph, milk, urine) . . 98 in fertilizers 96 in mixtures of salts 95 in pure salt. 94 in soils 97 separation of, and estimation as perchlorate 88 as pyrosulphate 92 Potassium permanganate, reaction of, as affected by concentration of acid, time and temperature. . . .42-48 upon ferrous salt, with hydro- chloric acid 48, 50 upon oxalic acid, with hydrochlo- ric acid 50 with other chlo- rides 52 standardization of, by arsenic trioxide 41 with iodine 42 by liberation of iodine and absorption of that ele- ment by silver 42, 362 in iron analysis 495 Precipitates, purification of, by solution and reprecipitation I o Rotary shaker 9 Rotating cathode ii Rubidium, estimation of, as acid sulphate 106 spectroscopic determination of, in pure salt 102 in presence of potassium salt 105 effect of potassium on 104 effect of sodium on 104 Selenic acid, iodometric determination of, by action of hydrochloric acid by action of hydrobromic acid 3* by action of hydriodic acid . . 388 534 INDEX OF SUBJECTS PACK Selenious acid, gravimetric determination of, by precipitation of selenium 376 by liberation of iodine and absorption of that ele- ment by silver 375 iodometric estimation of 377 by contact method 377 by differential method 380 by distillation method 379 volumetric estimation of, by permanganate 382 by thiosulphate 383 Selenium, separation of, from tellurium, and estimation, by differential volatility, of bromides 390 Silicon, detection of, in silicates and fluosilicates 241 Silicon fluoride, estimation of, eliminated at high temperature 436 Silver, electrolytic determination of 138 gravimetric determination of, as chromate 136 iodometric estimation of, by precipitation as chromate 140 reduced by arsenite from the chloride . . 143 anode, use of, for fixation of chlorine 20 chloride, electrolytic reduction of 460 bromide, electrolytic reduction of 460 iodide, electrolytic reduction of 461 Sodium, detection of, by hydrochloric acid in alcohol 74 after separation of potassium perchlorate 75 in mixtures of salts of other elements 78 estimation of, as pyrosulphate 79 Sodium chloride, electrolysis of, with mercury cathode 22, 26 electrolytic analysis of, alkalinity of inner cell 23 transfer of silver to cathode ... 24 Stannic chloride, hydrolysis of, in iodide-iodate mixture 63 Starch indicator, colors of, with free iodine 29, 32 delicacy of, in presence of potassium iodide 35, 38 effect of temperature upon 38 effects of varying amounts of 36 end reaction of, with tartar emetic 37 from various sources 36 loss of iodine in action of < 30, 36 preparation of 33 Strontium, detection of, associated with calcium and lead 160 gravimetric estimation of, precipitated as oxalate 180 volumetric estimation of, precipitated as oxalate 182, 183 Strontium and barium, separation of, and determination, by action of amyl alcohol on the bromides 167 Strontium and calcium, detection of, by action of amyl alcohol on the nitrates 163 separation of, and estimation, by action of amyl alcohol on the nitrates 164 Succinic acid, use of, as standard, in iodometric processes 56 in neutralization processes 54 Sulphides, detection of, associated with sulphites, sulphates and thio- sulphates 363 Sulphites, detection of, associated with sulphides, sulphates and thio- sulphates 363 iodometric determination of, in alkaline solution 366 Sulphocyanates, detection of, with ferricyanides and ferrocyanites . . . .275, 276 gravimetric determination of 276 volumetric estimation of, by permanganate 279 Tartar emetic, preparation of 36, 40 titration of, by iodine 38 INDEX OF SUBJECTS 535 PAGE Tartar, emetic use of, as an iodometric standard 40 variation in composition of 40 Telluric acid, iodometric estimation of 401 Tellurium, gravimetric estimation of, as dioxide precipitated by ammonia and acetic acid 402 separation of, from selenium 404 Tellurous acid, estimation of, by titrimetric precipitation, as tellurous iodide 398 gravimetric estimation of, by liberation of iodine and absorption of that element by silver 394 iodometric estimation of 399 volumetric estimation of, by permanganate 394 in presence of chloride 396 in presence of bromide 397 Thallium, determination of, as acid sulphate and as neutral sulphate .... 219 gravimetric estimation of, as chromate 221 precipitated as thallic hydroxide by potassium ferricyanide and potassium hydroxide 220 iodometric estimation of, precipitated by potassium dichro- mate 222 Thiosulphates, detection of, in association with sulphides, sulphites and sulphates 363 iodometric estimation of, effects of acid, concentration and temperature 364 Tin, detection of, associated with arsenic, by hydrochloric acid, and iodide 313 and bromide . . 316 electrolytic determination of 251 estimation of, associated with arsenic and antimony, by ferricyanide and permanganate 322, 323 separation of, from copper precipitated as oxalate 131 from copper precipitated as sulphocyanate 112 Tin chloride, hydrolysis of, in iodide-iodate mixture 63 Titanium, determination of, by reduction and titration with permanga- nate 242 Uranium, determination of, by Jones reductor 430 Valve (Kreider) 7 Vanadic acid, estimation of, gravimetric, by action of hydriodic acid and absorption of iodine by silver 325 by precipitation as ammonium vanadate. . . 326 by precipitation as silver meta- vanadate .. 328 in association with antimonic acid 350 with arsenic acid 350 with chromic acid 411 with iron 508 with iron and chromic acid 510 with chromium, iron and molybdenum 352 by permanganate after reduction by zinc 346 with use of ferric sulphate 349 with use of silver sulphate 348 iodometric estimation of, by action of hydrochloric acid . . 330 by action of hydrobromic acid 335, 345 by action of hydriodic acid . . .337, 343 by reduction with organic acids and reoxidation by iodine ... 341 536 INDEX OF SUBJECTS PAGE Vanadium in tetroxide condition, estimation of, by ferricyanide and permanganate 352 Volatile products, distillation of, and absorption 4, 5 and condensation 3 removal of, without mechanical loss of non-volatile material 6 Zinc, electrolytic determination of 186 estimation of, as pyrophosphate 185 precipitation of, as oxalate, and estimation 187 separation of, from aluminium 216 from copper precipitated as oxalate 131 Zinc chloride, conversion of, to oxide 186 Zinc sulphate, hydrolysis of, in iodide-iodate mixture 64 Zirconium, separation of, from iron volatilized in hydrogen chloride. .244, 508 14 DAY USE RETURN TO DESK FROM WHICH BORROWED LOAN DEPT. This book is due on the last date stamped below, or on the date to which renewed. Renewed books are subject to immediate recall. Cl^ . -- r- H 4J&n 62M J REC'D LD DEC 71961 rtMnl'tf*"ll /yJUl Wrsrf X c c X "^ \> ^ <& 6 jf^ 6 , 6Ma^0 OH REC'D LD APR 2 2 '65 -9PM LD 21A-50m-8.'61 (Cl795slO)476B General Library University of California Berkeley