iiiiiti HI H M U I 1 H > Hljl HI) ii flit fill ill!!!!! I iiii iiiiiiitii i i ill lul mi mrmiiui !:!n!iii;!!!ii; m im I I ! ii! 1 1 1 1 ' 1 1 . i n I illllM I iSiltiinn Itlt ililj tin! Ml jl iltjj'iiiHii i't [I MM H I l l jj h I! I if! tiH i M niill i ( "III H i'llifil 11 Hi m ill f mi ji HIM i hill lit!! H! I Hi liil; ii IHlHllKllllltl nil tl! !' HH H'M MH H H l!!!ifi!i!Hl!i!m!!!!!!!H H j |}|j l il|ilji!jili!!i!l GIFT OF Dean Frank U Probert Mining Dept. Sold by Book Department MINING AND SCIENTIFIC PRESS 42O MARKET STREET SAN FRANCISCO Catalogue of Technical Books on request Cx THE CYANIDE HANDBOOK Published by the McGraw-Hill Boole Company " to the Bo ok Departments of the McGraw Publishing Company Hill Publishing- Company Publishers of Books for- Elec trical Worl d The Engineering and Mining Journal Engineering Record Power and The Engineer Electric Railway Journal American Machinist Metallurgical and Chemical Engineering THE Cyanide Handbook BY J. E. CLENNELL, B.Sc. (LONDON) Associate of the Institution of Mining and Metallurgy; Associate of the Chemical, Metallurgical and Mining Society of South Africa; Author of "Chemistry of Cyanide Solutions," etc.; Cyanide Chemist to the Creston-Colorada Company FIRST EDITION SECOND IMPRESSION Corrected McGRAW-HILL BOOK COMPANY 239 WEST 39TH STREET, NEW YORK 6 BOUVERIE STREET, LONDON, E.G. 1910 Copyright, 1910, by the McGRAW-HiLL BOOK COMPANY GIFT 0% FRANK H rtQ8Et 7% Plimpton Press Norwood Mass. U.S.A. INTRODUCTION NOTWITHSTANDING the fact that so much has been written about the cyanide process since its successful establishment in South Africa in 1890, there still appear to be some aspects of the subject which are orly slightly touched upon, or quite inadequately treated, by the mi^iy able exponents of the process, and many facts of great value to the practical worker can only be gathered by long and painful search through technical literature. The various mechanical developments of the process have been dealt with by different writers too numerous to mention, and books have been published in which this side of the question is handled in a fairly adequate manner. The chemical aspect of the process has, how- ever, been treated only in a fragmentary and imperfect manner, although valuable contributions to our knowledge of this subject have appeared from time to time, chiefly in the form of papers read before such bodies as the Chemical and Metallurgical Society of South Africa, the Institution of Mining and Metallurgy, and the Society of Chemical Industry. The aim of the present work is to bring together such scattered information in a handy and accessible form, which will appeal to all workers interested in the process, whether as managers of plants, foremen, chemists, or assayers. Depending as it does on the application of chemical principles, a very important section of the work connected with a cyanide plant is carried on in the labora- tory; this department will therefore receive a larger share of attention than is usually given to it in books of this description. The discussion of the chemical properties, reactions, methods of manufacture, and analysis of cyanide itself, which have been scarcely touched upon in other similar books, will here be dealt with in sufficient detail. An outline will be given of the latest and most approved practice in modern cyanide plants, but without going into minute details of working, which are liable to constant change. CONTENTS PAGE PART I. GENERAL INTRODUCTION v SECTION I. EARLY HISTORY OF THE CYANIDE PROCESS .... 3 (A) Discovery of Cyanogen and its Chief Compounds ... 3 (B) Early Observations on the Solubility of the Precious Metals in Cyanide 10 (C) Introduction of the MacArthur-Forrest Process .... 24 SECTION II. OUTLINE OF OPERATIONS IN THE CYANIDE PROCESS . 33 (A) The Dissolving Process ' 33 (B) The Precipitation Process 38 (C) The Smelting Process 41 PART II. CHEMISTRY SECTION I. CYANOGEN AND ITS COMPOUNDS WITH NON-METALS . . 47 (A) Cyanogen 47 (B) Hydrocyanic Acid 51 (C) Compounds of Cyanogen with Non-metals 56 SECTION II. SIMPLE METALLIC CYANIDES 62 Details Respecting Certain Metallic Cyanides 64 Cyanides of Alkaline Metals 65 Cyanides of the Alkaline Earth Metals 68 Cyanides and Double Cyanides of Zinc 69 Cyanides of the Iron Group 71 Cyanides of the Precious Metals 71 Cyanides of Copper * . 76 Cyanides of Mercury 81 Cyanides of Lead 83 SECTION III. METALLIC COMPOUNDS OF COMPLEX CYANOGEN RADICALS 84 (A) Metallic Compounds of the Radical RC 6 N 6 84 (B) Metallic Compounds of Radicals in which Cyanogen is Com- bined with Oxygen or a Similar Element 95 SECTION IV. CHEMISTRY OF THE DISSOLVING PROCESS .... 102 SECTION V. CHEMISTRY OF THE PRECIPITATION AND SMELTING PROCESSES 117 SECTION VI. MANUFACTURE OF CYANIDE 127 (A) Production of Cyanides from Refuse Animal Matter . . 127 vii viii CONTENTS PAGE (B) Production of Cyanides from Atmospheric Nitrogen . . . 128 (C) Production of Cyanides from Ammonia or Ammonium Com- pounds 130 PART III. PREPARATORY TREATMENT OF ORE FOR CYANIDING SECTION I. CRUSHING MACHINERY 135 (A) Rock-Breakers 135 (B) Stamps 137 SECTION II. GRINDING MACHINERY 146 (A) Rolls 146 (B) Grinding Mills 148 (C) Ball Mills 152 (D) Tube Mills 154 SECTION III. MECHANICAL HANDLING OF MATERIAL FOR CYANIDE TREATMENT 159 (A) Handling of Old Accumulations 159 (B) Handling of Current Mill-Product 165 (C) Transfer of Material to Higher Levels 167 (D) Preparation of Ore for Direct Cyanide Treatment . . . 170 SECTION IV. HYDRAULIC CLASSIFICATION, SEPARATION, AND SETTLE- MENT OF SLIMES 172 (A) Hydraulic Classification by Means of Pointed Boxes . . 172 (B) Separation and Settlement of Slimes 176 SECTION V. AMALGAMATION AND CONCENTRATION 179 (A) Amalgamation in Relation to Cyanide Treatment . . . 179 (B) Concentration in Relation to Cyanide Treatment . . . 185 PART IV. THE DISSOLVING PROCESS SECTION I. PERCOLATION 197 (A) Leaching Vats 197 (B) The Leaching Process ' 200 (C) Double Treatment 205 (D) Discharging of Treated Material 206 SECTION II. AGITATION 208 (A) Agitation by Mechanical Stirrers 208 (B) Agitation by Circulating the Pulp and Injection of Air . . 210 (C) Andrew F. Crosse's Slime Treatment Process 211 (D) Treatment of Slimes by Settlement and Decantation . . . 212 SECTION III. FILTRATION BY PRESSURE AND SUCTION 216 (A) Advantages of the System 216 (B) Filter-Presses 216 (C) Filtration by Suction 221 SECTION IV. HANDLING OF SOLUTIONS 235 (A) Sumps and Storage Tanks 235 CONTENTS ix PAGE (B) Settling and Clarifying Tanks .... .... 236 (C) Pumps 237 (D) Piping for Conveying Solutions 237 (E) Arrangements for Heating Solutions 239 PART V. THE PRECIPITATION AND SMELTING PROCESS SECTION I. ZINC-BOX CONSTRUCTION AND PRACTICE 243 (A) Zinc-Box Construction 243 (B) Circular Vats for Zinc Precipitation 245 (C) Cutting and Preparing the Zinc Shavings 246 (D) Lead-Zinc Couple 246 (E) Charging the Zinc into the Boxes 247 (F) Conditions which Influence Precipitation . . . . ; . 248 (G) Difficulties in Zinc Precipitation 252 (H) Non-Accumulation of Zinc in Solutions 256 (I) Tests for Regulating the Working of the Boxes .... 256 SECTION II. CLEAN-UP OF ZINC-BOXES 258 (A) Preliminary Operations 258 (B) Separation of Precipitate from Coarse Zino 258 (C) Removal of Precipitate from Box 259 (D) Settlement and Sifting of Precipitate 259 SECTION III. ACID TREATMENT AND ROASTING OF PRECIPITATE . . 261 (A) Acid Treatment 261 (B) Roasting of Zinc-Gold Precipitate 264 SECTION IV. FLUXING, SMELTING, AND REFINING OF THE PRECIPI- TATE 266 (A) Fluxes for Zinc-Gold Precipitate 266 (B) Mixing of Precipitate and Fluxes 267 (C) The Fusion Process p 268 (D) Use of Special Fluxes ' 270 (E) Treatment of Matte 271 (F) Smelting of Zinc-Gold Precipitate with Litharge . ... 272 (G) Nature and Properties of Cyanide Bullion - . 274 (H) Refining of Bullion 275 PART VI. SPECIAL MODIFICATION OF THE CYANIDE PROCESS SECTION I. DIRECT TREATMENT AFTER DRY OR WET CRUSHING . 279 SECTION II. CRUSHING WITH CYANIDE SOLUTION 281 SECTION III. ROASTING BEFORE CYANIDE TREATMENT .... 286 Furnaces with Revolving Rabbles 286 Revolving Cylindrical Furnaces 288 SECTION IV. USE OF AUXILIARY DISSOLVING AGENTS 293 Oxidizers in Conjunction with Cyanide 293 The Bromocyanide Process 295 X CONTENTS PAGE Bromocyanide Practice in Western Australia 295 Use of Mercuric Chloride 297 SECTION V. ELECTROLYTIC PRECIPITATION PROCESSES 298 The Siemens-Halske Process 298 Later Developments of Electrolytic Process 300 Electrolytic Process for Gold-Copper Ores 301 Other Electrolytic Processes 303 Solution and Electrolytic Precipitation in the same Vessel . . 304 SECTION VI. OTHER PRECIPITATION PROCESSES 306 SECTION VII. TREATMENT OF CUPRIFEROUS ORES 310 PART VII. ASSAYING SECTION I. SAMPLING 317 (A) Sampling of Ores and Similar Material in Bulk . . . . 317 (B) Sampling of Sand and Similar Material 324 SECTION II. ASSAY OF A TYPICAL SILICEOUS ORE 333 (A) Crucible Fire Assay 333 (B) Cupellation and Parting 340 SECTION III. SPECIAL METHODS OF ASSAY FOR PARTICULAR ORES AND PRODUCTS 349 (A) The Scorification Assay 349 (B) Crucible Fusions for Various Classes of Ore 354 Class I. Siliceous Ores 355 Class II. Basic Ores 358 Class III. Pyritic Ores 359 Class IV. Lead-Zinc Ores 365 Class V. Arsenical Ores 365 Class VI. Antimonial Ores 366 Class VII. Cupriferous Ores 367 Class VIII. Telluride Ores 369 (C) Assay of Certain Refractory Materials 371 SECTION IV. ASSAY OF GOLD AND SILVER BULLION 378 (A) Fire Assay Method 378 (B) Volumetric Methods 389 PART VIII. ANALYTICAL OPERATIONS SECTION I. ANALYSIS OF ORES AND SIMILAR SILICEOUS MATERIAL. 395 Preliminary Operations 395 General Outline of Analysis 396 Estimation of Various Ingredients 397 SECTION II. ANALYSIS OF BULLION AND OTHER MAINLY METALLIC PRODUCTS 430 Sampling and Preliminary Operations 430 Estimation of Precious Metals 431 Estimation of Base Metals 433 Non-Metals and Negative Radicals 434 CONTENTS xi PAGE SECTION III. ANALYSIS OF CYANIDE SOLUTIONS AFTER USE IN ORE TREATMENT 437 Tests used in Daily Routine of Cyanide Plant 438 Estimation of Cyanogen Compounds 440 Estimation of Metals 442 Estimation of Acid Radicals other than Cyanogen Compounds . 448 Alkali Determinations 451 Organic Matter 451 Special Tests 452 SECTION IV. ANALYSIS OF COMMERCIAL CYANIDE 455 Sampling and Preparation 455 Estimation of Acid Radicals 456 Estimation of Metallic Radicals 459 Moisture and Insoluble Matter 461 Calculation of Results 462 SECTION V. ANALYSIS OF SUNDRY MATERIALS 463 (A) Lime 463 (B) Coal 465 (C) Water 467 PART IX. METALLURGICAL TESTS SECTION I. DENSITY DETERMINATIONS 477 SECTION II. ALKALI CONSUMPTION TESTS 483 SECTION III. SCREENING AND HYDRAULIC SEPARATION .... 485 SECTION IV. CONCENTRATION AND AMALGAMATION TESTS . . . 487 SECTION V. CYANIDE EXTRACTION AND CONSUMPTION TESTS . . 490 Standard Preliminary Cyanide Test 490 Tests to determine Conditions of Treatment 491 Tabulation of Results 495 SECTION VI. USEFUL FORMULAE 496 Conversion of Metric Weights and Measures 498 Atomic Weights (1909) 499 PART I GENERAL SECTION I EARLY HISTORY OF THE CYANIDE PROCESS (A) DISCOVERY OF CYANOGEN AND ITS CHIEF COMPOUNDS Early Use of Cyanogen Compounds. The cyanide process is a technical application of certain chemical reactions displayed by a group of substances known as the " cyanogen compounds." The beginnings of its history must therefore be traced in the beginnings of our knowledge of those compounds. Mixtures con- taining one or other of these substances, such as could easily be obtained from natural sources, have probably been known and used from very ancient times without any accurate knowledge of their composition. Thus it is stated by Hoefer 1 that prussic acid was known to the ancient Egyptians, who used it for poisoning those who had been guilty of divulging the sacred mysteries of the priests. An extract of laurel leaves seems to have been used for similar purposes in the Middle Ages. Discovery of Prussian Blue. In 1704 the substance now known as Prussian blue was accidentally discovered by Diesbach 2 whilst endeavoring to prepare a red coloring matter or " lake " by mixing a decoction of cochineal with alum, green vitriol, and fixed alkali: Using alkali, which had previously been used for the distillation of animal matter, instead of a red precipitate, he obtained a fine blue color. The discovery was communicated to Dippel, who improved and simplified the process; but it was kept secret until 1724, when Woodward 3 published a method of making Prussian blue by fusing dried blood and alkali, and treating the extract with alum, green vitriol, and spirit of salt. It was soon found that other animal matters might be used instead of blood, and many chemists 1 "Histoire de Chimie," I, 226. 2 Macquer, " Dictionnaire de Chimie," I, 265. See also "Miscellanea Bero- linensis ad Incrementum Scientiarum," I, 377 (1710). 3 "Philosophical Transactions," XXXIII, 15-24 (1724). 3 : THE CYANIDE HANDBOOK occupied themselves with the preparation and examination of this remarkable coloring matter. Macquer's Researches on Prussian Blue. The first attempt at a scientific examination of it appears to be that of Macquer 1 (1752). He came to the conclusion that it consists of iron charged with an excess of inflammable matter furnished by the " phlogisti- cated alkali " used for precipitating it. He also showed that the peculiar principle which imparted the blue color might be removed from Prussian blue and transferred to an alkali, which latter thus lost its alkaline properties and yielded a solution from which pure Prussian blue could be obtained again by addition of an iron salt and an acid. He had thus obtained a ferrocyanide. Prussic Acid. Other investigators followed, notably Berg- man and Guyton, the latter of whom gave the name " acide prussique" to the substance to which the color of Prussian blue was supposed to be due. In 1782 important researches on this substance were made by Scheele, 2 who succeeded in isolating prussic acid by distilling the extract (obtained by fusing animal matter and alkali) with excess of sulphuric acid; he showed that pure Prussian blue could be obtained from the distillate by adding to it an iron salt and a few drops of sulphuric acid. He concluded that the "coloring material" was composed of aerial acid (i.e., carbonic acid) and an inflammable substance. He also showed that the neutral salt obtained by the action of alkali on Prussian blue (i.e., ferrocyanide of potassium or sodium) is a triple salt consisting of alkali, a little iron, and the coloring material. Scheele' s Observations on the Cyanides of Gold and Silver. But the most interesting observations from our present point of view are those which he made on the cyanides and double cyanides of gold and silver, and attention may be drawn to the following remarkable passages: " The calx of iron is not the only one which has the property of fixing the coloring material, but this also happens with gold, silver, copper, and perhaps with several metallic calces, for if the solutions of the same are precipitated with the ' liquor prsecipitans ' and again dissolved by pouring in more liquor, the solution remains clear in 1 "Memoire de 1' Academic des Sciences": see Macquer's " Dictionnaire de Chimie," I, 265-274 (1778). 2 "Sammtliche Physische und Chemische Werke," II, 321-348 ("Versuche liber die Farbende Materie im Berlinerblau"). EARLY HISTORY OF THE CYANIDE PROCESS 5 the open air, and the calx is not again precipitated by the aerial acid." l Translated into modern language this passage means that when solutions of gold, silver, and copper salts are precipitated by a cyanide solution and the precipitate redissolved by excess of cyanide, a solution is formed which remains clear on exposure to air, and from which the metallic oxide is not thrown down by carbonic acid. "If some of this liquor" (i.e., the alkaline extract of animal matter) "be dropped into a completely saturated gold solution, the gold is thrown down as a white precipitate, but if much of the liquor be added, it is again dissolved. This solution is colorless like water; the precipitate is, however, insoluble in acids. Silver is thrown down white, like cheese; if more liquor be added the precipitate is again dissolved." 2 He had thus prepared the double cyanides of gold and silver, on the formation of which the extraction of the precious metals by cyanide entirely depends. Berthollet on Composition of Prussic Acid. Berthollet 3 (1787) repeated and extended the observations of Scheele, showed that the views previously held, that ammonia and carbonic acid are contained in prussic acid, were erroneous, and gave the correct composition of the acid as follows: "Hence I conclude that hydrogen and nitrogen exist in prussic acid, that they are combined with carbon, and that when oxygen is added the necessary condi- tions for forming carbonate of ammonia are present." He was not able, to determine the proportions of the constituents, however. Berthollet also seems to have obtained chloride of cyanogen, but made no exact investigation of it. He says (loc. cit., p. 35) : "If oxygenated muriatic acid" (i.e., chlorine) "be mixed with prussic acid prepared by Scheele's method, the former is converted into muriatic acid and the latter acquires a much sharper odor, and appears to have become more volatile: its affinity for alkalis is still further diminished. It forms no Prussian blue with iron solutions, but a green precipitate which becomes blue on exposure to light or mixture with sulphurous acid." *IUd., p. 347. 2 Ibid., p. 339. 3 "Annales de Chimie," I, 30 (1790): "Extrait d'un M^moire sur 1'Acida Prussique." 6 THE CYANIDE HANDBOOK Artificial Formation of Cyanides. Clouet, 1 in 1791, published a method of obtaining cyanogen compounds which is the basis of many methods that have been used or proposed for the manufac- ture of cyanide. He passed ammonia gas through a porcelain tube containing charcoal heated to redness, and received the issuing gases in vessels containing a solution of ferrous sulphate. He found that the product exhibited the properties recognized by chemists as belonging to the coloring matter of Prussian blue. Cyanides in Nature. Vauquelin 2 (1803) showed that prussic acid is present ready formed in certain vegetable substances occurring in nature. Solubility of Gold in Cyanides. Hagen, in 1805 ("Unter- suchungen," I, 665), is reported to have made the statement that gold is dissolved, not only by free chlorine and aqua regia, but by a solution of prussiate of potash. I have not been able to confirm this reference, but, if correct, it would seem to be the earliest state- ment of the solubility of metallic gold in cyanide solutions. The existence of a soluble " prussiate " of gold was of course well known, having been discovered, as already mentioned, by Scheele; but in all references to this substance which I have been able to find it is formed by acting on salts of gold and not on the metal itself. Proust on Cyanides and F err o cyanides. In the following year (1806) Proust published an extensive investigation of the proper- ties of the prussiates, 3 showed that iron in the " triple prussiates " (i.e., ferrocyanides) exists in the condition of the "black oxide" (i.e., in the ferrous condition), and requires the addition of a solu- tion of the red oxide (i.e., a ferric salt) in order to give Prussian blue. With regard to the reactions of cyanides and ferrocyanides with gold, he remarks: t( ~ ,, (Triple prussiate, Nothing; [Simple " White precipitate; which becomes a fine yellow. On heating the mixture, the pre- cipitate does not explode; it is a true prussiate of gold. Heated in a retort, it gives water, empyreumatic oil pretty abundantly, 1 "Annales de Chimie" [1] XI, 30-35: "Memoire sur la Composition de la Matiere Colorante du Bleu de Prusse." 2 "Annales de Chimie et de Physique," XLV, 206. 3 "Annales de Chimie," LX, 185-224, 225-252: "Fails pour Servir a 1'His- toire des Prussiates." EARLY HISTORY OF THE CYANIDE PROCESS 7 a carbonaceous gas which burns with a blue flame, and leaves a residue of gold mixed with charcoal powder." Cyanogen Obtained. Proust also tried the effe'ct of heating cyanide of mercury, and found that it gave off inflammable matter which burnt with a red and blue flame with a yellowish border. He formed the erroneous conclusion, however, that the gas con- sisted of a mixture of prussic acid and "gaseous oxide." Researches on Prussic Acid. Further researches were made by Ittner 1 (1809), who succeeded in obtaining prussic acid by the action of hydrochloric acid on the cyanide of mercury. By carry- ing out this process and condensing the distillate in a mixture of ice and salt, Gay-Lussac, 2 in 1811, succeeded in obtaining anhy- drous prussic acid, both in the solid and liquid form, and deter- mined its boiling-point, 26.5 C, melting-point, 15 C, and specific gravity, .70583. Electrolysis of Ferrocyanides. R. Porrett, 3 in 1814, published a paper " on the nature of the salts termed triple prussiates." He electrolyzed the "triple prussiate of soda" (i.e., sodium ferrocya- nide), and observed that ferrous oxide and prussic acid are obtained at the positive pole and only soda at the negative pole. Hence he concluded that the iron and the elements of prussic acid must together constitute a peculiar acid, which he succeeded in isolating by treating a solution of barium ferrocyanide with the exactly equivalent quantity of sulphuric acid. He also discovered (in 1808) the acid of the class of substances now known as " thiocya- nates" or "sulphocyanides"; and suggested the use of these com- pounds for the quantitative estimation of copper, when added together with a reducing agent. Gay-Lussac : Analysis of Prussic Acid and Cyanogen. In 1815 Gay-Lussac 4 published his researches on prussic acid, in which he describes the analysis of prussic acid by exploding the vapor with oxygen in a vessel standing over mercury, and also by passing the vapor over heated iron. He determined the quantita- tive composition of mercuric cyanide, and showed that the gas produced by heating this compound consists only of carbon and .# 1 Ittner, "Beitrage sur Geschichte der Blausaure." 'Gay-Lussac, "Note sur 1'Acide Prussique": " Annales de Chimie," LXXVII, 128. a Porrett, "Philosophical Transactions," XXVI, 527. 4 Gay-Lussac, "Recherches sur 1'Acide Prussique": "Annales de Chimie," XCV, 156. 8 THE CYANIDE HANDBOOK nitrogen, in the proportion (approximately) of 6 parts of the former to 7 of the latter. He proposed the name " Cyanogene " for the new gas, and having shown that prussic acid is a compound of equal volumes of cyanogen and hydrogen, he suggested for it the name "acide hydrocyanique." The discovery of cyanogen is of great importance in the history of chemistry, being the first example of the isolation of a " com- pound radical," that is to say, of a group of elements associated together and acting in a number of chemical changes as though they were a single element. Davy's Researches on Cyanogen. The observations of Gay- Lussac were confirmed and extended by Sir H. Davy, 1 in 1816, who electrolyzed anhydrous hydrocyanic acid, showed that no water is formed by the combustion of cyanogen in oxygen, and also decom- posed cyanogen by the electric spark, and found that carbon was separated and a volume of nitrogen liberated equal to that of the cyanogen taken. He also prepared iodide of cyanogen by heating iodine with mercuric cyanide. From this time onward many chemists occupied themselves with the study of cyanogen and its compounds, and we can here only allude to a few of the more important discoveries. Cyanic Acid. In 1818, Vauquelin found that one of the products obtained when cyanogen gas is dissolved in water is an acid, which is decomposed by stronger acids into carbonic acid and an ammonia salt. The name cyanic acid was given to this sub- stance; it was found to contain cyanogen and oxygen, and gave salts, known as cyanates, containing different metals in combina- tion with these elements. Synthesis of Urea. F. Wohler 2 noted that when cyanogen gas is passed into solution of ammonia various bodies are formed, including " a characteristic crystalline substance, which, however, does not appear to be ammonium cyanate." In 1828 3 he identified this substance as urea, and showed that it can be formed by direct transformation of ammonium cyanate, which has the same per- centage composition. This observation is very interesting, as being the first example of the formation of an organic body from 1 Davy, "Gilb. Ann.," LIV, 383. "Journ. of Science and the Arts," I, 288: "On the Prussic Basis and Acid." 2 Wohler, "Pogg. Ann.," Ill, 177 (1825). 8 "Pogg. Ann.," XII, 253. EARLY HISTORY OF THE CYANIDE PROCESS 9 purely inorganic materials, for it had already been shown by Clouet (1791) that cyanides could be formed from ammonia passed over red-hot carbon. It is also one of the earliest observed exam- ples of isomerism, or of two substances having identical che'mical composition but different properties, the reaction being NH 4 -CNO = CO(NH 2 ) 2 . In 1831, J. Pelouze 1 showed that potassium formate is one of the products obtained by evaporating an aqueous solution of potassium cyanide, this being another instance of the artificial production of organic compounds. Manufacture of Cyanides. In 1834, F. and E. Rodgers 2 described various methods of manufacturing cyanides, including that universally practised until a few years ago, which is given by them in the following terms: "Cyanuret of potassium may be prepared by exposing a mixture of anhydrous carbonate of potash and anhydrous ferrocyanuret of potassium to a moderate red heat in a covered crucible for about 20 minutes." Action of Acids on Ferrocyanides. In 1835, Everitt 3 gave the correct explanation of the reaction taking place when potassium ferrocyanide is distilled with dilute sulphuric acid, the equation (expressed in modern symbols) being: 3H 2 SO 4 + K 4 FeCy 6 = 3KHSO 4 + KFeCy 3 + 3HCy. obtaining the white or yellowish insoluble salt (KFeCy 3 ) now known as Everitt's salt. Haloid Cyanogen Compounds. Bromide of cyanogen was discovered and investigated by Serullas 4 (1827), who also studied and determined the composition of the analogous chloride of cyanogen (first observed by Berthollet: see above), and redis- covered and investigated the iodide of cyanogen, first obtained by Davy (1816). Other Investigators. Further important researches on cyano- gen and its compounds were made by Liebig, Bunsen, Berzelius, Kuhlmann, Berthelot, Joannis, Playfair, and many others, some of which will be referred to later in dealing with the chemistry 1 Pelouze, " Annales de Chimie et de Physique," XLVIII, 395. 2 Rodgers, "Philosophical Magazine," (3) IV, 91: "On Certain Metallic Cyanurets." 3 Everitt, "Philosophical Magazine," (3) VI, 97. * Serullas, "Annales de Chimie et de Physique," (2) XXVII, 184; XXXV, 291, 337; XXXVIII, 370. 10 THE CYANIDE HANDBOOK of cyanide. We shall now proceed to describe those observations which more immediately led up to the discovery of the cyanide process. (B) EARLY OBSERVATIONS ON THE SOLUBILITY OF THE PRECIOUS METALS IN CYANIDE Elkingtoris Patent. The earliest definite statement of the solubility of metallic gold and silver in cyanide solutions appears to be that contained in the patent taken out Sept. 25, 1840, by J. R. and H. Elkington (Brit, patent, No. 8447, of 1840), which was based on information supplied to them by Dr. Alexander Wright, of Birmingham, who in his turn appears to have been led to the discovery by the observations of Scheele, already quoted. The following are extracts from Elkington's specification: " We take oxide of gold prepared by any of the known methods, or metallic gold in fine division, and dissolve the same in a solution of prussiate of potash or soda; to about two ounces of gold con- verted into oxide we add two pounds of prussiate of potash dis- solved in one gallon of water, and boil the same for half an hour; the solution is then ready for use." . . ."We claim, therefore, in respect of this branch of our invention, the method of coating, covering, or plating the metals above mentioned with gold by the use, first, of an oxide of gold or metallic gold in fine division dis- solved in prussiate of potash or any soluble prussiate, and also by the use of the oxide of gold dissolved in any analogous salt; and secondly, by combining the action of a galvanic current, with the use of a solution of gold as above described, giving the prefer- ence to a solution of gold in the prussiate of potash as above par- ticularly described." It will be obvious, from the quotations already given from early writers (Proust, Porrett, and others), that the expression " prussiate of potash " signifies the cyanide and not the ferrocya- nide of potassium; and in fact it appears to be used in this sense as late as 1880 (see Clark's specification, U. S. Patent, No. 229,586, quoted below). Bagration: Solubility of Gold in Cyanide. In 1843, Prince Pierre Bagration l published his paper " On the Property which Potassium Cyanide and Ferrocyanide Possess of Dissolving 1 Bagration, "Bull, de I'Acad. Imple de St. Pet." (1843), II, 136. See ab- stract in "Journ. fur. prakt. Chem.," XXXI, 367 (1844). EARLY HISTORY OF THE CYANIDE PROCESS 11 Metals." In using gold plates as anodes in electro-gilding opera- tions, he had observed that the gold continued to dissolve in the cyanide solution, even when the electric current was interrupted. He therefore tried the experiment of suspending a plate of gold so as to be partially immersed in a solution of potassium cyanide, and found that " after the lapse of about 3 days the part which dipped into the liquid was almost completely dissolved. The strongest action had taken place in the upper part where the solution and the plate were in contact with the atmospheric air." He found that the action was promoted by warming the solution, and that " silver and copper in the form of very thin plates or wires likewise dissolve " [in cyanide solutions]. The paper concludes as follows: "My researches lead me to suppose that hydrocyanic acid at the moment of formation also possesses this property [of dissolving gold]; but it is certain that in future cyanides of potassium must be reckoned among the number of the solvents of gold, and that one must beware of using gold or silver vessels for operations which require the use of these salts." He noticed that ferrocyanide of potassium is also a solvent, though in a very much less degree. Glassford and Napier on Gold and Silver Cyanides. In 1844, a paper was published by C. F. O. Glassford and J. Napier, 1 "On the Cyanides of the Metals and Their Combinations with Cyanide of Potassium," giving considerable information about the preparation and properties of the cyanides of gold and silver, and the double cyanides which these form, respectively, with cyanide of potassium; but, strangely enough, the solubility of the metals without the aid of the electric current is not referred to, although the fact is mentioned that when two plates of gold dipping in cyanide solution are connected with an electric battery, the gold dissolves much more rapidly from one plate than it is deposited on the other. This paper contains I believe, the earliest reference to the volumetric method of estimating cyanides by means of silver nitrate solution, now universally used in testing cyanide solutions (loc. cit., p. 70). Eisner: Influence of Oxygen on the Solution of Metals in Cyanide. In 1846, L. Eisner, 2 in a paper called " Observations on * Glassford and Napier, "Philosophical Magazine," (3) XXV, 56-71 (paper read before Chemical Society, Feb. 19 and March 4, 1844). ? Eisner, "Journ. fur prakt. Chem.," XXXVII, 441, 12 THE CYANIDE HANDBOOK the Behavior of Reguline Metals in an Aqueous Solution of Potassium Cyanide," showed that the oxygen of the air played an essential part in the solution of gold, silver, and some other metals in cyanide solutions. In this paper he remarks: "It is a fact known for some years past, that metallic gold, silver, copper, iron dissolve in aqueous solutions of potassium cyanide, even without the assistance of the galvanic current, at the ordinary temperature of the air (about 12-15 C) " . . . "It was very probable that the solution of the last-named metals (Au, Ag, Cd), as well as that of Cu, Fe, Zn, and Ni, had only taken place through the agency of oxygen, so that if this supposition were correct, oxygen must have been removed from the layer of atmos- pheric air still present above the column of liquid." Eisner describes experiments showing that the amount of oxygen in tubes containing gold and silver, standing over cyanide solution, was gradually reduced, and that the air which remained extinguished a lighted match when the latter was slowly intro- duced. When such tubes were opened under water or mercury, these liquids rose in them to a certain height. Mercury gradually rose in a tube containing cyanide, silver, and air, inverted over it. Finely divided metallic silver placed in a well-stoppered tube filled with cyanide solution showed very slight action, only a little silver cyanide being obtained on acidulating with HC1; but after filtering, the liquid passing through gave a considerable precipitate with this acid, showing that further action had taken place on exposure to air. Eisner formulated no equation to explain the combined action of cyanides and oxygen on gold or silver, although the usually accepted expression of this reaction goes by his name. He classi- fies the metals according to their action on cyanide, as follows: 1. Insoluble: Pt, Hg, Sn. 2. Soluble with decomposition of water: Fe, Cu, Zn, Ni. 3. Soluble with aid of atmospheric air, without decomposition of water: Au, Ag, Cd. Early Cyanide Extraction Test. In 1849 there appeared an account of what seems to have been an early attempt to utilize the solubility of gold in cyanide compounds for metallurgical purposes. This is a description of experiments made in 1848 by Dr. Duflos, 1 entitled, " Entgoldungsversuche der Reichensteiner Arsenik- i "Journ. fur prakt. Chcm.," XL VIII, C5-70 (1849). EARLY HISTORY OF THE CYANIDE PROCESS 13 abbrande." In consequence of the statements of Bagration (quoted above), a quantity of the material was extracted with a very dilute solution of yellow prussiate of potash. The liquid leaching through was digested with hydrated sulphide of iron for the purpose of converting any cyanide of gold and potassium which might be present into sulphide of gold and ferrocyanide of potas- sium. The residue was burnt, after washing and drying, but the ash was found to contain no gold. Tests made with a mixture of chloride of lime solution and yellow prussiate gave no better result. " Liebig's potassium cyanide was not tried, not so much on account of its high price, as because the considerable amount of ferrous oxide present in the ' Abbrande,' which would form yellow prussiate with the potassium cyanide, already indicated that an unsatisfactory result was to be expected a priori." If this test had not been omitted, the experimenter might very possibly have forestalled Messrs. Mac Arthur and Forrest by nearly 40 years ! Faraday's Researches on Gold Leaf. In 1856, Faraday 1 delivered a lecture before the Royal Society entitled, "Experi- mental Relations of Gold and Other Metals to Light," in which he showed that extremely thin films of gold, capable of transmitting light, could be produced by acting upon gold-leaf, adhering to a plate of glass, with a weak solution of cyanide of potassium. The method of obtaining such films was as follows: "If a clean plate of glass be breathed upon and then brought carefully upon a leaf of gold, the latter will adhere to it; if distilled water be immediately applied to the edge of the leaf, it will pass between the glass and the gold and the latter will be perfectly stretched; if the water be then drained out, the gold-leaf will be left well extended, smooth, and adhering to the glass. If, after the water is poured off, a weak solution of cyanide be introduced beneath the gold, the latter will gradually become thinner and thinner; but at any moment the process may be stopped, the cyanide washed away by water, and the attenuated gold film left on the glass." Faraday also emphasized the importance of air in this reaction, and gives the following explanation of it (loc. cit., p. 147) : " Air-voltaic circles are formed in these cases and the gold is dissolved almost exclusively under their influence. When one piece of gold-leaf was placed on the surface of a solution of cyanide of potassium, and another, moistened on both sides, was placed i Faraday, "Philosophical Transactions," CXLVII, 145-181 (1857). 14 THE CYANIDE HANDBOOK under the surface., both dissolved; but twelve minutes sufficed for the solution of the first, whilst above twelve hours were required for the submerged piece. In weaker solutions, and with silver also, the same results were obtained; from sixty to a hundred-fold as much time being required for the disappearance of the sub- merged metal as for that which, floating, was in contact both with the air and the solvent." Again (loc. cit., p. 173): "A piece of ruby paper" (colored by finely divided gold) " immersed in a strong solution of cyanide of potassium suffered a very slow action, if any, and remained unal- tered in color; being brought out into the air, the gold very gradu- ally dissolved, becoming first blue." Use of Cyanide in Amalgamation. The extraction of gold by means of mercury had long been known. Rose 1 states that it has been in use for the last 2000 years; but it had frequently been observed that the surface of the mercury became coated with films of various nature, which impaired its effectiveness for amalgama- tion. It was found that these films, consisting chiefly of grease or of oxides and carbonates of copper and other base metals, could be removed, and the efficiency of the mercury restored, by treatment with various chemicals, such as nitric acid, caustic soda, sal- ammoniac, etc. Among others, cyanide of potassium was found to be very effective for this purpose, and appears to have been already in frequent use as an aid to amalgamation in America during the 60's of the last century. This practice was, of course, carried out in ignorance of the fact that cyanide was a solvent of gold, and thus cannot in any sense be considered to be an anticipa- tion of the cyanide process. One or two writers, however, drew attention to the solubility of gold in cyanide, only mentioning the fact as a drawback to its use as an aid to amalgamation. Thus, Prof. Henry Wurtz, in a memoir read to the American Association for the Advancement of Science, at Buffalo, Aug. 1, 1866, 2 says: " Early in my investigations I found that no supposition of ex- traneous films or coatings of air, oxide of iron, talc, or other foreign matter (though undeniably often present and even occasionally active in this way), was available in the general explanation of the phenomena; for it was quickly found that the most powerful 1 "Metallurgy of Gold," 4th edition, p. 91. 2 See "American Journ. of Mining," JV 354 (Dec. 7, 1867): "On a Theory of ( fold Genesis." EARLY HISTORY OF THE CYANIDE PROCESS 15 chemical agents, save those which act by dissolving and removing a superficial film of the gold itself, gave no aid to quicksilver." The following note is added: " To this category belongs cyanide of potassium, as I need not inform chemists. Despite the highly deleterious, and in fact dan- gerous, nature (in inexperienced hands) of this agent, it would appear that attempts are now being made to palm it off upon mining men, especially in California, as a new and highly valuable adjunct in amalgamation. As a writer upon this subject, I feel therefore called upon to take this opportunity to say that many experiments were made by myself three or four years since with the cyanide, in the preliminary amalgamation of copper plates, with the idea that it might be of value, in connection with the amalgam of sodium, for this purpose; and it is possible that it might be, were it not for one vital obstacle, namely, that the coating thus formed, unlike that produced by the aid of sodium alone, tarnishes with great readiness. Alone, the cyanide is very feeble in its action in this way, and much inferior to nitric acid, the common agent used. I find, in fact, that ammonia, mixed with sal-ammoniac (indeed almost anything which will dis- solve copper or its oxides), is equally available. The cyanide has been much tested in Colorado, but mostly abandoned. The new project, apparently started in California, of using it in the battery and the pan, is one which could scarcely interest any but those occupied in the vending of miners' nostrums; as the known solubility of both gold and mercury in it would alone be fatal in the view of an expert. In the pan, particularly, another destructive agency would be set up, namely, a solvent action upon the sulphides of the baser metals, and a secondary precipitation of the latter upon the quicksilver. The subject, however, is not worth the space that would be required for its discussion here." Rae's Patent: Cyanide in Conjunction with Electric Current. In the following year we come to the patent of Julio H. Rae (U. S. Patent, No. 61,866; Feb. 5, 1867) for an "Improved Mode of Treating Auriferous and Argentiferous Ores." This is the first distinct attempt to apply cyanide as a solvent of gold in ores, though it may perhaps be questioned whether Rae was aware of the solubility of gold in cyanide of potassium without the aid of an electric current. He gives a figure of an agitation apparatus, the object being to dissolve the gold and silver, and precipitate them again on a cathode of copper or suitable material. The process is thus described : "This invention consists in treating auriferous and argentifer- ous ores with a current of electricity or galvanism for the purpose of separating the precious metals from the gangue. In connection with the electric current suitable liquids or chemical preparations, such, for instance, as cyanide of potassium, are used, in such a manner that by the combined action of the electricity and of the chemicals, the metal contained in the ore is first reduced to a state 16 THE CYANIDE HANDBOOK of solution, and afterward collected and deposited in a pure state, and that the precious metals can be extracted from the disinte- grated rock or ore at a very small expense and with little trouble or loss of time." The agitation apparatus consisted of a vertical shaft, to which was attached a metallic cage, the bottom of the shaft resting on a plate of metal (by preference platina) at the bottom of a jar of glass or other suitable material. The shaft was to be connected with the positive pole of a battery, so that shaft, cage, and plate together formed the anode, while the other pole of the battery was to be connected with a thin slip or coil of copper or suitable material (suspended in the liquid above the ore) and "forming a base on which the precious metals are deposited. By the action of the electric current the action of the chemicals on the metals con- tained in the rock is materially facilitated and a perfect solution thereof is effected." The use of the expression " materially facilitated " would seem to imply that Rae supposed cyanide of potassium (the only chemical mentioned in the specification) would have some action, even without the electric current. Cyanide as a Solvent of Metallic Sulphides. The use of cyanide of potassium as a solvent in the treatment of ores is again men- tioned in 1870 by H. C. Hahn 1 (" Schwefel-metalle gegen Cyan- kalium"), who remarks that "Cyanide of potassium is not only a good solvent for Ag, Cu, etc. (Cl, I, Br, O), but also for several sulphides; thus, CuS (as already known), AuS and AgS dissolve easily; ZnS and FeS after some digestion in strong cyanide solution." He also notes that these sulphides dissolve both in the pure condition and in ores. He describes experiments showing the solubility of silver sulphide in cyanide, and notes that the separation of silver from copper, by means of sulphureted hydro- gen, from solutions in which the metals are present as soluble double cyanides, is imperfect, owing to the incomplete precipita- tion of the silver sulphide. He refers to the application of cyanide in ore treatment as follows: "With regard to AuS, I will mention that its solubility in KCy has already found practical application. The owner of the Union Mine in Amador County, California, treats the concentrates, which contain 200 to 6000 dollars gold per ton, with quicksilver and a strong solution of cyanide of potassium, iHahn, "Berg- und Huttenm. Zeit.," XXIX, 66 (Feb. 25, 1870). EARLY HISTORY OF THE CYANIDE PROCESS 17 and extracts 23 per cent, more of the total gold contents than formerly by ordinary amalgamation." Hahn seems to have ascribed the action of cyanide to the solubility of gold sulphide rather than of metallic gold; and it is evident also that the practical application referred to was one in which the object was to aid amalgamation, and not to dissolve the gold in cyanide. Skey on Relative Solubility of Metals in Cyanide. The objec- tions to the use of cyanide as an aid to amalgamation are again referred to by W. Skey, 1 in an article, read Jan. 29, 1876, "On the Electromotive Order of Certain Metals in Cyanide of Potassium, with Reference to the Use of this Salt in Milling Gold." He noted that cyanide had been used with marked effect in preventing the " flouring " of mercury during the treatment of blanketings at the Thames Gold-Field, New Zealand, and ascribes the action to the solvent effect of cyanide on a film of mercurial salts produced by the action of soluble ferric salts, which "either oxidize or chloro- dize the surface of any mercury which they may be in contact with." To produce this effect, the cyanide must dissolve a portion of the mercury. " Moreover, in thus contemplating the contingencies entailed or risked by the use of any alkaline cyanide in such milling operations, it must be remembered that both gold and silver are not absolutely insoluble in these cyanides. Now the loss of mer- cury in this way may not be serious, but if gold or even silver be thus lost (that is, by its solution), even in much less quantity than mercury well could be, the loss then may be serious. Whether the loss of metal certain to be entailed by the use of cyanide of potassium falls upon the mercury, or upon the gold or silver of these blanketings, conjointly or separately, depends entirely upon this relative affinity of these metals for this salt, or in other words it depends upon their electromotive order therein." He shows the commonly accepted view, that mercury is positive to gold and silver in cyanide solutions, to be erroneous, and that mercury is in reality very decidedly negative; thus, metallic gold in contact with a solution of mercuric cyanide would rapidly dissolve and mercury be reduced. He gives the following as the elec- tromotive order of metals in potassic cyanide, from negative to positive : i Skey, "Trans. Proc. N. Z. Inst./ VIII, 334-337. 18 THE CYANIDE HANDBOOK Carbon Lead Platinum Gold Iron Silver Arsenic Tin Antimony Copper Mercury + Zinc. " Most, if not all, the sulphides or other ores occurring in nature are negative to the whole series. Any of these metals will generally precipitate the ones named above it from its cyanide solution"; thus, gold precipitates, mercury and silver gold, taking its place in the liquid. " Thus, it appears, a loss of gold by solution of it must frequently happen whenever cyanide of potassium is em- ployed to assist in the amalgamation of blanketings, or other auriferous stuff. In fact, all that loss of metal occasioned by its solution, and most of which is, as we have seen, a necessity in- volved in the working of the process itself, falls upon the gold and silver present, the mercury being positively protected from the action of this salt by these more valuable metals." They also think the loss would be greater if the operation were conducted in iron pans, and also greater with a strong cyanide solution than a weak one, and suggests a method of minimizing the loss by passing the waste liquor in a thin stream over copper plates (presumably with a view to precipitating the dissolved gold and silver). Dixon on Treatment of Pyritic Ores with Cyanogen Compounds and Oxidizers. In 1877, W. A. Dixon 1 published a paper " On a Method of Extracting Gold, Silver and Other Metals from pyrites," which was read before the Royal Society of New South Wales, Aug 1, 1877. After giving an account of various methods tried for the treatment of ores containing sulphides of Cu, Pb, Sb, Fe, As, etc., and rich in gold and silver, and noting that amalga- mation and chlorination gave unsatisfactory results, he describes a number of experiments made with cyanides and ferrocyanides as solvents, both alone and in conjunction with various oxidizing agents. With regard to cyanide, he says : " Prince Bagration and Eisner [Watt's " Diet " : " Cyanides of Gold "] have observed that precipitated gold is soluble in cyanide of potassium if exposed to the air; and the latter says also in ferro- cyanide of potassium. A patent was applied for in America in 1868 i Dixon, "Proc. Royal Soc. N. S. W," XI, 93-111. EARLY HISTORY OF THE CYANIDE PROCESS 19 for the use of cyanide of potassium for the extraction of gold from its ores, but I have no particulars of the process." (This possibly refers to Rae's patent, mentioned above.)* " It seemed to me, how- ever, that the high price of this salt, its instability when exposed to the air and in solution, and its extremely poisonous properties, precluded its use for this purpose. On trying the reaction between precipitated gold and cyanide of potassium, I found that it was extremely slow if the gold was at all dense. In presence of alkaline oxidizing agents, however, I found that the solution of gold was sufficiently rapid. Thus, on standing overnight, the quantity of gold and cyanide of potassium solutions being similar in each case, with the cyanide alone, traces only of gold were dissolved, but with the addition of calcium hypochlorite, ferrocyanide (ferricyanide?) of potassium, or binoxide of manganese, all the gold was dissolved; with chromate of potassium, a small quantity; with permanganate of potassium, none. With ferrocyanide of potassium alone I did not obtain any gold in solution after standing some days, but I thought that with suitable oxidizing agents it might be obtained in solution according to the equation 4Au + 2K 4 FeCy 6 + 7O + 4H 2 O = 4AuCy 3 + Fe 2 O 3 + 8KHO. In the cold, however, with the exception of ferricyanide of potassium, none of the above oxidizing agents had any effect, but heated to 212 F., the reaction with all of them was sufficiently rapid, and I found that this was also the case with permanganate." Precipitation on Silver and Copper. For precipitation, he suggests first filtering the hot solution through finely divided metal- lic silver to recover the gold, and then precipitating the silver as sulphide; he found, however, " that copper in any form precipitated both gold and silver from the solution, or at all events that these metals (i.e., gold and silver) were not dissolved until the copper had all gone into solution; also, that if the copper was present as sulphide, the silver was transformed into sulphide, which is insolu- ble." To remove dissolved copper, he suggests digesting the alkaline solution with ferrous hydrate, which would at the same time con- vert any cyanide present into ferrocyanide, " which has the impor- tant advantages of being exceedingly stable and non-poisonous." Extraction Tests on Ores. Various experiments are then detailed, on roasted arsenical pyrites, and other pyritic ores, using 20 THE CYANIDE HANDBOOK mixtures of ferrocyanide with different oxidizers in hot solutions, " A portion of roasted arsenical pyrites was digested at 212 for twelve hours with \ oz. ferrocyanide of potassium, 32 gr. oxide of manganese (20 Ib. per ton), and sufficient water, made alkaline with soda, to make a cream: the solution yielded 9 oz., 8 dwt., 19 gr. gold per ton, leaving 1 oz., 9 dwt., 15 gr. This was the best result obtained with this pyrites, the yield with other oxidizing agents and by more prolonged digestion being all somewhat lower." In another case the material was previously roasted with salt, and extracted with acid to remove copper. The material then contained Gold, 8 oz., Odwt., 19 gr.; Silver, 49 " 11 " 5 " and yielded from 3 oz. 12 dwt. to 5 oz. 1 dwt. of gold and from 46 oz. to 46 oz. 3 dwt. of silver per ton. " This showed that all the silver which had been converted into chloride during the roasting was obtained in solution as cyanide. With the gold, on the other hand, all the results showed that with complex pyrites a portion only could be obtained in solution, either in mercury or in water as cyanide or chloride, whilst none could be obtained as sulphide. 77 These results did not appear to Dixon to be sufficiently satisfac- tory to warrant further investigation, and the remainder of the article is devoted to the discussion of smelting processes, etc. It would seem, however, that additional tests on the lines of the exper- iments detailed might probably have led to the discovery of a workable cyanide process. Solubility of Silver in Cyanide. Some interesting remarks on the solubility of silver in cyanide solutions are to be found in Percy's " Metallurgy " 1 (published in 1880), although the explana- tion he gives of the reaction would seem to be quite erroneous. He states that " metallic silver dissolves in a hot aqueous solution of cyanide of potassium, but slowly and sparingly, as shown by the following experiment by R. Smith. A single silver leaf was heated during several hours in a concentrated aqueous solution of cyanide of potassium, and it was found that only a small quantity of silver had been dissolved. Christomanos, on the other hand, asserts that pure silver dissolves readily in a hot solution of this 'salt (Fresenius, Zeit. fiir Anal. Chem., VII, 301; 1868). In this case, hydrogen must be evolved according to the following equation, 1 J. Percy, "Metallurgy: Silver and Gold," Part I, p. 115. EARLY HISTORY OF THE CYANIDE PROCESS 21 unless the silver is oxidized by oxygen derived from any cyanate of potash that might be present; though in the experiment by R. Smith, above recorded, no evolution of hydrogen was observed: 4KCy + 2Ag + H 2 O = 2KCyAgCy + K 2 O + 2H." Cyanide as an Aid to Amalgamation. In 1880-81 three patents were taken out in the United States, all of them involving the use of cyanide or some cyanogen compound in ore treatment; but the chemical being added not as a solvent but as an aid to amalgama- tion, these processes have no direct bearing on our subject, though they are of some interest as showing the state of metallurgical knowledge at the time, shortly before the introduction of the cya- nide process. So far from utilizing the solvent action of cyanide, the latter would be actually a drawback in any attempted practical application of these patents. The specifications referred to are: (1) U. S. patent, No. 229,586: Thos. C. Clark: Filed Dec. 27, 1879; issued July 6, 1880: "Extracting Precious Metals from Ores." (2) U. S. patent, No. 236,424: Hiram W. Faucett: Filed July 13, 1880; issued Jan. 11, 1881: "Process of Treating Ore." (3) U. S. patent, No. 244,080: John F. Sanders: Filed April 16, 1881; issued July 12, 1881: "Composition for Dissolving the Coating of Gold in Ore." Clark's Process. In Clark's process the ore is roasted and thrown while hot into a bath formed of a solution of salt, prussiate of potash, and caustic soda or caustic potash, the object being to effect desulphurization and disintegration, so as to bring the precious metals into a suitable form for amalgamation by freeing them from the union and influence of baser metals. Faucett Process. In Faucett's process the crushed ore is to be heated and subjected, under pressure, to the action of " disinte- grating chemicals " in solution in a closed vessel, after which it is to be further pulverized and amalgamated. The pressure is effected by the steam generated by contact of the hot ore with the chemical solution. The chemicals contained in the solution are chloride of sodium, nitrate of potash, cyanide of sodium, sulphate of protoxide of iron, and sulphate of copper, with or without admixture of hydrofluoric acid, fluoride of potassium, or fluoride of sodium. Sanders' Process. In Sanders' process, a mixture of cyanide of potassium and phosphoric acid is used to remove " the coatings 22 . THE CYANIDE HANDBOOK that envelop gold in the ore and that consist usually of various metallic oxides and of silver. " . . . " After agitation the mixture above mentioned will be found " to have dissolved the oxides and the sulphurous coatings of the ore, and the agitation of the barrel or vessel removes the dissolved impurities, leaving the gold free and exposed, and permitting it to be amalgamated by the addition of quicksilver in the usual manner." He also makes the following disclaimer: "I am aware that cyanides have already been used in the extraction of gold; also that gold-bearing ores have been dis- integrated in the presence of heat by various chemicals. This I do not claim." The mixture claimed consists of about 16 parts of cyanide of potassium to 1 part of glacial phosphoric acid, with sufficient water to form a thick pulp with the raw gravel. This would undoubtedly dissolve in some cases considerable amounts of gold, but that circumstance seems to have entirely escaped the inventor's observation. Simpson's Patent: Cyanide with other Chemicals as Gold Solvent. A patent which has much more direct bearing on the subject and which may legitimately be regarded as in some respects an anticipation of the process of MacArthur and Forrest is that of Jerome W. Simpson: U. S. patent, No. 323,222: Filed Oct. 20, 1884; issued July 28, 1885: "Process of Extracting Gold, Silver, and Copper from their Ores." The crushed ore is to be treated with certain salts in solution adapted to combine chemically with the metal and form therewith a soluble salt. After thorough agitation, the solid matter is allowed to settle and a piece or plate of zinc is to be suspended in the clear liquid, "which causes the metal dissolved in the salt solution to be precipitated thereon, from which it can be removed by scraping or by dissolving the zinc in sulphuric or hydrochloric acid. The precipitated metal may then be melted into a button." The salt solution referred to consists of " one pound of cyanide of potassium, one ounce carbonate of ammonia, one-half ounce chloride of sodium, and 16 quarts of water, or other quantities in about the same proportions." It would appear that the car- bonate of ammonia was added as a solvent for copper, and the chloride of sodium as a solvent for silver; for in the case of ores containing only gold and copper he omits the chloride of sodium, and for ores rich in silver he employs a proportionately larger EARLY HISTORY OF THE CYANIDE PROCESS 23 quantity of this salt. He also disclaims the exclusive use of cya- nide as a solvent, as follows: " I am aware that cyanide of potassium, when used in connec- tion with an electric current, has been used for dissolving metal, and also that zinc has been employed as a precipitant, and the use of these I do not wish to be understood as claiming broadly." He also disclaims the exclusive use of carbonate of ammonia but claims: " 1. The process of separating gold and silver from their ores which consists in subjecting the ore to the action of a solution of cyanide of potassium and carbonate of ammonia, and subsequently precipitating the dissolved metal, substantially as set forth. "2. The process of separating metals from their ores, to wit: subjecting the ore to the action of a solution of cyanide of potas- sium, carbonate of ammonia, and chloride of sodium, and subse- quently precipitating the dissolved metals." The addition of these small quantities of other chemicals almost suggests that they were introduced because Simpson doubted the validity of a patent based on the use of cyanide alone. With regard to this process it may be remarked that the solution specified is inconveniently strong in cyanide, and that the precipi- tation process, if carried out as described, would be very inefficient. It is nevertheless possible that if the process had been introduced on a working scale it might eventually have been so modified as to become a practical success. Alleged Anticipations of the Cyanide Process. Caveats cover- ing the use of cyanides for extracting precious metals are alleged to have been taken out by Endlich and Muhlenberger, in 1885, and by Louis Jan in, Jr., in 1886, but no definite particulars are available with regard to these. It is rather curious to note that both Endlich and Janin write in a disparaging way about the pros- pects of cyanide as a commercially successful solvent of gold and silver (see Eng. and Min. Journ., L, 685 and LI, 86). That the practical difficulties in the way of the application of cyanide as a solvent had not been overcome at the time of the MacArthur-For- rest patents, may be judged from the following remarks of Janin (ibid., L, 685), from a letter dated Nov. 24, 1890: "To sum up the disadvantages of the cyanide process in a nutshell, I will say that to extract a reasonable percentage of the gold and silver, so great an excess of cyanide is required that the 24 THE CYANIDE HANDBOOK extraction is no longer economical or profitable. Moreover, if this excess is used, it is impossible to completely precipitate by the means advocated by the MacArthur-Forrest people. In fact, as a means of extracting gold I can more conscientiously recommend aqua regia than potassium cyanide." (C) INTRODUCTION OF THE MACARTHUR-FORREST PROCESS Researches on Gold Solvents by the MacArthur-Forrest Syndicate. We have now traced the history of cyanide in its application to the treatment of gold and silver as far as the year 1886. In this year experiments were being made in Glasgow by a research syndicate consisting of J. S. MacArthur, R. W. Forrest, W. Forrest, and G. Morton, for the purpose of investigating, and if possible developing into a commercial success, a method of ore treatment known as the "Cassel Gold-Extracting Process." An interesting account of the steps which led to the successful application of cyanide in the treatment of ores as a result of these experiments is given by J. S. MacArthur, 1 in a paper read before the Scottish Section of the Society of Chemical Industry, March 7, 1905, from which a few extracts may here be given. The process under inves- tigation depended on the solvent action of chlorine generated electrolytic ally in an alkaline solution. It was found, however, that such solvents attacked the base metals in the ore in preference to the gold. Attempts were made to dissolve the gold and limit the action on base metals by adding chlorine or bromine, without the electric current, and introducing a salt which would not absorb Cl or Br, nor precipitate gold, but which would precipitate base metal compounds. Bleaching-powder, borax, and bicarbonate of soda were found to partially answer this purpose, but even with this addition the action of the halogen was directed rather against the base metal than the gold. The efforts of the syndicate were therefore directed to finding a gold solvent that would not be a base-metal solvent. This condition naturally pointed to some alka- line or neutral solution. In making these tests it was decided not to rely on the method of taking the difference between original assay of ore and final assay of residue as indicating gold extracted, but in all cases the gold was to be precipitated from its solution and i "Journ. Soc. Chem. Ind.," XXIV, 311: "Gold Extraction by Cyanide: a Retrospect." EARLY HISTORY OF THE CYANIDE PROCESS 25 recovered in a visible form, so that it could be actually handled and weighed, sulphureted hydrogen being used as the precipitant. Cyanide Extraction Trials. " Among the various solvents on our program for trial, we had included potassium cyanide, and in November, 1886, we tried the effect of it on the tailings of one of the Indian gold mines, and, as usual, treated the solution with H 2 S for recovery of the gold; and getting none, we passed on to our next solvent, meanwhile observing our rule of labeling the residue and laying it aside. We had neglected to notice that H 2 S did not precipitate gold from its solution in cyanide, and thus our experi- ment was for the time literally relegated to the shelf. About eleven months after, I had occasion to devise a rapid method of gauging approximately the gold contents in weak solutions of gold chloride, and used for the purpose tin chloride to produce the well-known 'purple of Cassius.' One solution that I had to test contained mercury, and using potassium cyanide to separate the mercury and the gold, I was apprised of the fact that H 2 S did not precipitate gold from its cyanide solution. Instantly my mind reverted to the experiment carried out nearly a year before, and I saw that it might have been successful without the success being recognized. Immediately a sample of rich concentrates from a Californian mine was treated, and on this occasion we examined the residue rather than the solution, and found a high percentage of extraction. A sample of poor concentrates from India was now treated, and again a high extraction was obtained. The results were startling. We unearthed the residues from the old experiments (all our work was done in duplicate), and to our intense satisfaction we found that they too had transferred their gold to the cyanide solution. There was now no doubt about the importance of the discovery, and at once a provisional specification was drafted and lodged." Mac Arthur-Forrest First Patent. The provisional specifi- cation of the English patent, No. 14,174, was lodged Oct. 19, 1887, by J. S. MacArthur, R. W. Forrest and W. Forrest, and the complete specification taken out July 16, 1888, entitled: "Improvements in Obtaining Gold and Silver from Ores and other Compounds." The essential points in which this patent differs from those of previous inventors above alluded to, are: 1. The solution of the gold and silver is to be effected by means of a liquid to which a cyanide alone is added, without the aid of an electric current or of other chemicals. 26 THE CYANIDE HANDBOOK 2. The patentees emphasize the use of dilute in preference to strong solutions for accomplishing their purpose. 3. Certain definite relations are specified between the quanti- ties of ore, values in gold and silver, and strength and quantity of solution, which, however, were not adhered to in practice and cannot be regarded as being essential to the process. 4. Cyanogen gas is mentioned (probably in error) as one of the solvents claimed. The following extracts from the specification will illustrate these points: (1) Nature of the Solvent. "In carrying out the invention, the ore or other compound in a powdered state is treated with a solution containing cyanogen or cyanide (such as cyanide of potassium, or of sodium, or of calcium), or other substance or compound containing or yielding cyanogen." (2) Use of dilute solutions. " In practice, we find the best results are obtained with a very dilute solution, or a solution con- taining or yielding an extremely small quantity of cyanogen or a cyanide, such dilute solution having a selective action such as to dissolve the gold or silver in preference to the baser metals." (3) Relation between quantities of ore, solution, etc. " In preparing the solution we proportion the cyanogen to the quantity of gold or silver, or gold and silver, estimated by assay or other- wise to be in the ore or compound under treatment, the quantity of a cyanide or cyanogen-yielding substance or compound being reckoned according to its cyanogen. ... In dealing with ores or compounds containing, per ton, 20 ounces or less of gold or silver, or gold and silver, we generally use a quantity of cyanide the cyan- ogen of which is equal in weight to from one to four parts in every thousand parts of the ore or compound, and we dissolve the cyanide in a quantity of water of about half the weight of the ore. In the case of richer ores or compounds, whilst increasing the quantity of cyanide to suit the greater quantity of gold or silver, we also increase the quantity of water so as to keep the solution dilute." (4) Use of cyanogen ,gas. "In using free cyanogen, the cyanogen obtained as a gas in any well-known way is led into water to form the solution to be used in our process; or any suitable known mode of setting cyanogen free in solution may be employed." Reference is made in the specification to the use of agitation, EARLY HISTORY OF THE CYANIDE PROCESS 27 pressure, and increased temperature as aids to extraction, but no claims are made in respect to these. No special means of precipitation is described in this patent, but the following statement is made: " When all or nearly all the gold or silver is dissolved, the solution is drawn off from the ore or undissolved residue, and is treated in any suitable known way, as, for example, with zinc, for recovering gold and silver. The residu- ary cyanogen compounds may also be treated by known means for regeneration or reconversion into a condition in which they can be used for treating fresh charges of ores or compounds." The solvents enumerated are: " Any cyanide soluble in water, . . . such as ammonium, barium, calcium, potassium or sodium cyanide, or a mixture of any two or more of them. Or any mix- ture of materials may be taken which will by mutual action form cyanogen or a substance or substances containing or yielding cyanogen." Original Claim. The first claim made was: "The process of obtaining gold and silver from ores and other compounds, con- sisting in dissolving them out by treating the powdered ore or compound with a solution containing cyanogen or a cyanide or cyanogen-yielding substance, substantially as hereinbefore de- scribed." Amended Claim. In consequence of litigation, this was modified so as to apply only to dilute solutions and the amended claim (Aug. 20, 1895) read as follows: "Having now particu- larly described and ascertained the nature of our said invention, and in what manner the same is to be performed, we declare that we do not claim generally the use of solutions of any strength, but what we claim is: " 1. The process of obtaining gold and silver from ores and other compounds, consisting in dissolving them out by treating the powdered ore or compound with a dilute solution containing cyanogen or a cyanide or cyanogen-yielding substance, substan- tially as hereinbefore described, and subject to the above disclaiming note: 1 Mac Arthur-Forrest Second Patent. On July 14, 1888, another patent was applied for by Mac Arthur and Forrest: Eng. patent No. 10,223, entitled, "Improvements in Extracting Gold and Silver from Ores and other Compounds." Use of Alkalis. After referring to their previous patent, they 28 THE CYANIDE HANDBOOK describe the use of alkalis, such as potash or lime, for neutralizing the ore previous to cyanide treatment; also various methods of treatment by agitation, grinding in pan-mills, and percolation in tanks with permeable false bottoms, are indicated. The method of precipitation is then described as follows: Zinc Precipitation. " The separated solution is next made to pass through a mass of metallic zinc in a state of fine division. We find that the best results are obtained in this part of the process when the zinc has been freshly divided by mechanical or other means, so that its surfaces are as purely metallic as possible; and, further, when the quantity or mass of zinc employed is such that the solution has, in passing through it, ample opportunity for being thoroughly acted on. " The zinc to be used is reduced to the desired state of fine divi- sion by any suitable means. The degree of division is, preferably, such as is obtainable by shaving or by cutting thin strips or very small pieces or grains, by means of a turning tool, circular saw or other suitable tool, from cakes or blocks of zinc of convenient size. By another method, the zinc is brought into the desired state of fine division by passing it in a molten state through a fine sieve and allowing it to fall into water. In order to obtain the best results the finely divided zinc should be used as soon as possible after it has been produced." The use of zinc in this form is also suggested as applicable to other gold- and silver-extracting processes, in which the metals are dissolved, not as cyanides but as chlorides, bromides, thiosul- phates or sulphates. " After separation of the solution the precious metals may be separated from zinc by distillation. Or the larger portion of the precious metals may be separated from zinc by sieving (by prefer- ence under water), when, with a suitable sieve, the greater part of the precious metals will pass through, the greater part of the zinc being left on the sieve." The claims made are: (1) "In processes for extracting gold and silver from ores or other compounds by means of a cyanide or cyanogen compound, the preparatory treatment of the ores or compounds with an alkali or alkaline earth, substantially as and for the purposes hereinbefore described. (2) " In precipitating gold and silver from cyanide, chloride, bromide, thiosulphate, or other similar solutions by means of zinc, EARLY HISTORY OF THE CYANIDE PROCESS 29 the employment of the zinc as freshly prepared in a state of fine division, substantially as hereinbefore described. (3) " The process for extracting and recovering gold and silver from ores and other compounds, consisting in first treating same with an alkali or alkaline earth, then extracting the gold by means of a cyanide or cyanogen compound, and finally precipitating the gold and silver by means of zinc as freshly prepared in a state of fine division, all substantially as hereinbefore described." It will be noted that no broad claim to the use of zinc as a precipitant is made, which would of course be untenable, but only to the special form of the metal found suitable in this process, the discovery of which is described by MacArthur in the paper which we have already referred to. 1 Discovery of Precipitation by Zinc Turnings. " We knew well that zinc precipitated gold from its cyanide solution; but it re- mained to make this reaction industrially applicable. We used various forms of finely-divided zinc with more or less advantage; but a picture of some fine zinc shavings, bought with other things in a shilling box of chemicals in my boyish days, haunted my mind, and repeatedly I described it to one of the works foremen without effect, until one day, when making a zinc case for packing cyanide, he made a shaving by a sharp tool and came asking me if this was what I wanted. My reply was ' Yes/ and in half an hour he had prepared the first bundle of zinc shavings for gold precipitation the pioneer bundle of hundreds of tons of this flimsy but useful material." American Patents: (Mac Arthur-Forrest). During the year 1888, patents were taken out in most of the important mining countries of the world, covering the points included in the two just discussed. The American patents are: (1) U. S. Patent, No. 403,202, May 14, 1889; referring to the dissolving process; and (2) U. S. Patent, No. 418,137, Dec. 24, 1889; referring to the precipitation process. The former of these differs in several respects from the British patent: it specifies more particularly the classes of ore which had not previously been satisfactorily or profitably treated, and to which the cyanide process may be applied with advantage; and it also emphasizes the points in which the process differs from those of Rae, Simpson, and other inventors, as follows: > J. S. MacArthur, " Journ. Soc. Chem. Ind.," XXIV, 313. 30 THE CYANIDE HANDBOOK " By treating the ores with the dilute and simple solution of a cyanide, the gold or silver is, or the gold and silver are, obtained in solution, while any base metals in the ores are left undissolved, except to a practically inappreciable extent; whereas, when a cyanide is used in conjunction with an electric current, or in con- junction with another chemically-active agent, such as carbonate of ammonium or chloride of sodium or phosphoric acid, or when the solution contains too much cyanide, not only is there a greater expenditure of chemicals in the first instance, but the base metals are dissolved to a large extent along with the gold or silver, and for their subsequent separation involve extra expense, which is saved by our process." Further particulars are given as to the nature of the containing vessel, and as to mechanical aids to extrac- tion. The proportions of ore, solution, etc., are the same as in the British patent, but a maximum strength is given and the claim is limited to solutions not exceeding this, being as follows: "The process of separating precious metal from ore containing base metal, which process consists in subjecting the powdered ore to the action of a cyanide solution containing cyanogen in the pro- portion not exceeding eight parts of cyanogen to one thousand parts of water." (This would be equivalent to 2 per cent. KCN.) Among the methods of recovering gold and silver from the solution, the process of treating it with sodium amalgam is referred to. Introduction of the Process. Arrangements were now made in the chief mining countries to introduce the process on a practical basis: the first works on a commercial scale seem to have been those at Karangahake, New Zealand, established in 1889. 1 In the early part of 1890, the Cassel Company established an experimental plant near the old Salisbury Battery, at the Natal Spruit, near Johannesburg, Transvaal, where tailings and concen- trates were treated in agitation vats, the contents of which were afterwards discharged on to suction filters and the liquid drawn off precipitated by zinc shavings. In April, 1890, these works were in regular operation, and the results were so successful that a plant on a commercial scale was established in the latter part of the year at the Robinson Mine, Johannesburg. i Large scale tests were made on ore from the Crown Mines (New Zealand) in 1888. EARLY HISTORY OF THE CYANIDE PROCESS 31 Importance of the Cyanide Process in Metallurgy. Since then the process has spread to every gold and silver mining dis- trict in the world, with few exceptions, and although not equally successful with every class of ore, it may fairly be claimed that its introduction has revolutionized the metallurgy of the precious metals. During the ten years from 1886 to 1896, the world's production of gold was doubled, and it is stated that in the period 1896-1906 it again doubled, mainly through the introduction of new metallurgical methods in which the use of cyanide has played a leading part. " In 1889 the world's consumption of cyanide did not exceed 50 tons per annum. In 1905 the consumption was nearly 10,000 tons per annum, of which the Transvaal gold field took about one-third." 1 A royalty was at first charged by the African Gold Recovery Company, representing the Cassel Company (the owners of the MacArthur-Forrest patents) in South Africa. The validity of the patent was, however, disputed, and the patents were finally set aside in February, 1896 after an appeal to the High Court of the Transvaal. It must be admitted, however, that although the patents may have been technically invalid, the application of the process in such a form as to be commercially successful was chiefly due to the energy and enterprise of MacArthur, Forrest, and their associates. Further Developments. The improvements since introduced have chiefly taken the form of applying well-known mechanical devices previously used in other connections, as adjuncts in the working of the process. Such alterations in the chemical treatment as have been suggested have as a rule found only a partial and temporary application; they have generally taken the form of additions of oxidizers or other reagents to the ore or solution: these so-called improvements will be referred to later in a special section dealing with them. We may, however, enumerate the following as being, perhaps, the most interesting develop- ments : 1893. The Electric Precipitation Process. 1895. The Bromocyanide Process. Slimes Treatment by Decantation. Aeration of Ore and Slimes. i Frankland, "Journ. Soc. Chem. Ind.," XXVI, 175. 32 THE CYANIDE HANDBOOK 1 897. Zinc-lead Precipitation Process. 1898. Slimes Treatment by Filter Presses. " Roasting Previous to Cyanide Treatment. 1902. Lead Smelting Applied to Zinc Precipitate. Economical Fine Grinding of Ore as a preliminary to Cyanide Treatment. SECTION II OUTLINE OF OPERATIONS IN THE CYANIDE PROCESS (A) THE DISSOLVING PROCESS Stages of Cyanide Treatment. Having now traced the history of cyanide treatment as far as its establishment on a practical basis by MacArthur and Forrest, we shall proceed to give a brief summary of the actual operations involved in the application of the process, leaving the details for discussion in later chapters. As in all metallurgical processes for the treatment of ores, the material must undergo a preliminary crushing or disintegration to render the metallic particles accessible to the solvent. After the ore has been thus obtained in a suitable condition, the treatment takes place in three distinct stages, which may be described as: 1. The dissolving process; 2. The precipitation process; 3. The smelting process. Mechanical Difficulties in Treatment of Crushed Ore. The method of treatment to be adopted in the dissolving process, and the plant and appliances used, are determined largely by the physical nature of the material to be dealt with. In crushing any kind of ore, whatever appliance be used, certain parts will be reduced to a finer state of division than others; the product is never homogeneous, but consists of particles of all sizes, from the largest that will pass the crushing apparatus to the minutest grains. Sands and Slimes. At an early stage in the history of the process it was found that the treatment of the crushed material as a single product was often unsatisfactory. The methods adopted, though varying greatly in detail, may be roughly classified under two heads Agitation and Percolation. In the agitation process it was found that the coarser and heavier particles showed a tendency to settle and remain stationary in corners or other parts 33 34 THE CYANIDE HANDBOOK of the vessel where the agitation was least effective, and so largely escaped treatment. In the percolation system the finer material showed a tendency to separate in certain regions, forming bands or masses of clay-like material, practically impervious to the solution, and which thus also escaped effective treatment. Hence the suggestion naturally arose that the best method of handling the material would be to separate the coarser and heavier portion of the crushed ore (technically described as " sands ") from the finer and lighter portion (known as "slimes"), and to treat the former by percolation and the latter by agitation. This idea has been almost universally carried out until within the last few years, when the alternative suggestion of crushing the whole of the material so fine that it may be treated satisfactorily by agitation has been rapidly gaining ground. Amalgamation. In most cases the ore, after crushing and before cyanide treatment, undergoes some form of amalgamation or treatment with mercury, the usual method being to pass the ore as it leaves the stamp-batteries, together with a sufficient quantity of water, in a thin stream over sheets of copper coated with mer- cury. The latter combines with and arrests the coarser particles of precious metal, extracting in this way a percentage of the values which varies largely with different kinds of ore, and may perhaps average about 60 per cent. Hydraulic Separation. The stream leaving the amalgamated plates then passes to some form of hydraulic separator, the most usual being that known as " Spitzlutte," consisting essentially of a pointed box with a central partition, and an outlet at the bottom near which a jet of water is introduced. The pulp fed in at the top on one side passes downward beneath the partition, where it meets an ascending stream of water; the lighter particles are carried away and overflow on the opposite side at the top of the box, while the coarser and heavier particles fall through the outlet at the bottom. The pulp may be passed through a succession of such boxes to give a sufficiently complete separation of sands and slimes. Collection of Sands for Treatment. When the sands are to be treated by percolation they are led, usually, into a collecting-vat, where they are evenly distributed by means of a hose or by some mechanical device. As the vat fills, the surplus water, containing a further quantity of slime, is continuously drawn off. When the vat has been thus filled with sand, the contents are frequently OUTLINE OF OPERATIONS IN THE CYANIDE PROCESS 35 prepared for cyanide treatment by the addition of alkali (generally lime) , and a wash of water or very weak solution is given in order to dissolve the alkali and distribute it to every part of the charge. In some cases the cyanide treatment is partially carried out in the collecting-vats, but usually the sand is transferred through dis- charge doors at the bottom of the collecting-vat into filter-tanks placed underneath. These tanks are nearly always circular, and constructed of wood or iron. The bottom is covered by a wooden framework on which rests a filter composed of cocoanut matting or some similar material covered by thick canvas. Percolation Process. The solution is led by iron pipes from the storage-tank to the top of the filter-tank and allowed to flow on until the charge of sand is completely covered. It is then left standing for a longer or shorter period, and drawn off by opening a valve in a pipe beneath the filter bottom; the liquid thus drained off passes either direct to the precipitation boxes or, more gener- ally, to a settling-tank, where any suspended matter is allowed to subside, and from whence the clear liquid is drawn off for precipita- tion. A number of washes in succession are given in this way, the quantity of liquid used, strength of solution, time of contact and other details being varied according to the nature of the material. The general practice may be described somewhat as follows: 1. One or two washes of very weak (alkaline) solution, for neutralizing acid substances in the ore. 2. The strong solution say one-fifth the weight of charge usually about .25 per cent. KCy; left standing 6 to 12 hours or more, then drained as completely as possible. 3. A number of washes, successively weaker in cyanide down to .1 per cent. KCy. 4. A final wash of water, or the weakest solution available. The total amount of solution used would be from 1 to 1^ times the weight of the dry sand in the charge; total time of treatment, 4 to 8 days. These quantities refer to tailings from ordinary siliceous ores after amalgamation. When the ore is treated " direct" (i.e., without amalgamation), or when rich or refractory material is treated, the quantities and strength of solution and the time of contact may vary within very wide limits. Collection of Slimes for Treatment. The slimes carried off by the overflow from the spitzlutten and collecting-vats are carried, generally by means of wooden launders, to a suitable vessel, gener- 36 THE CYANIDE HANDBOOK ally a large pointed box (spitzkasten), where they undergo partial settlement, in order to remove superfluous water. Lime is gener- ally introduced either in the battery itself or into the stream of pulp leaving the battery, which has the effect of causing the fine particles of ore quickly to coagulate and settle when the rapid flow of the liquid is arrested. The clear water from the spitzkasten may be pumped back to the battery for further use, while the thickened pulp is drawn off from the openings at the bottom of the spitzkas- ten and passes to the agitation tanks, where it is generally settled to remove a further quantity of water, and then treated with dilute cyanide solution, with alternate agitation and settlement. Agitation Process. The agitation is produced either by mechanical stirrers mounted on a vertical revolving axis, or by injecting air under pressure, and circulating the pulp by passing it continuously from the bottom of the tank through a centrifugal or other pump, which throws it back into the same tank. The strength of cyanide necessary in treating slimes is in general considerably less than that required for sands; the values, being mostly in minute particles, dissolve rapidly, especially when arrangements are made for effective aeration. Separation of Dissolved Values. The chief difficulties are encountered in the mechanical separation of the solution carrying the dissolved gold and silver, from the residue. Two distinct methods have been extensively adopted for this purpose; these are: 1. The decantation process; 2. The filter-press process. Decantation Process. In the decantation process, which is applicable chiefly to very low-grade material, such as the Rand battery slimes, the material is treated in very large tanks by alternate agitation and settlement; the clear settled solution is drawn off, by means of a jointed pipe, from the surface down to the level of the slime-pulp. Fresh solution is then added and the operation repeated two or three times. Sometimes the pulp is finally transferred to a large and relatively deep tank, where the pressure of a high column of liquid causes the slime to settle with a reduced percentage of water; after several successive charges have been added, the clear settled solution is drawn off as far as possible, and the residue, carrying say 30-40 per cent, of moisture, drawn off at the conical bottom of the large tank. As it is not possible, OUTLINE OF OPERATIONS IN THE CYANIDE PROCESS 37 from economical considerations, to give more than one or two successive agitations and decantations on the same charge, the percentage of extraction by this method is usually somewhat low, and with richer material the alternative process of filter-pressing has been found more advantageous. Filter-Press Process. In the filter-press process, after the slime-pulp has had a preliminary agitation with cyanide solution, it is transferred to a suitable receiver, whence it is forced by means of compressed air into the filter-presses, consisting of a number of chambers enclosed by metal frames and plates, so arranged that when the chambers are filled with the slime-pulp, liquid (either cyanide solution or water) may be forced through them under pressure, thence passing through filter-cloths into channels con- nected with the outlet, from which it is conveyed to a settling- tank and thence to the precipitation boxes. Various types of press, operating by pressure or by suction, have been introduced, and will be described in more detail in a special section. In general, it may be said that the filter-press process involves more expense for labor, power, etc., per ton of material treated, than the decan- tation process, but that the percentage of extraction is higher, and the time of treatment less; also less liquid is required, an important consideration in regions where water is scarce. After several washes of solution and water have been thus forced through the press, the latter is opened, and the cakes of slimes, containing a much lower percentage of moisture than the residues of the decan- tation process, are removed and discharged. Fine Crushing for Cyanide Treatment. In the system now coming into general use the ore is crushed rather coarsely in the battery, with water or dilute cyanide solution containing sufficient lime to make it slightly alkaline. The pulp then passes, with or without previous amalgamation, through a " tube " or " flint " mill, consisting of a revolving steel cylinder with a suitable lining of some hard material, and about half filled with more or less rounded flint pebbles. The ore is effectively and economically ground in this apparatus; the cylinder is set in a slightly inclined position, and the pulp passing out through an opening at the lower end goes to the spitzkasten, whence the coarser portion is returned to the upper end of the tube-mill to be reground. The fine portion, con- sisting almost entirely of particles small enough to pass a sieve of 150 holes to the- linear inch, passes, either direct or after ama.1- 38 THE CYANIDE HANDBOOK gamation, to the agitation tanks, the whole of it being treated as slimo and passed through filter-presses. It will be seen that the treatment is thus much simplified; the high initial cost of leaching and decantation plants is avoided, with a higher percentage of extraction and an increased quantity of ore treated per month. Extraction by Cyanide Process. The extraction of gold and silver varies, according to the nature of the material, from under 70 to over 95 per cent, of the total value : many ores hitherto con- sidered refractory have been found to yield a high percentage after fine grinding. Disposal of Residues. The residues after treatment are dis- charged either by shoveling into trucks, or by various mechanical devices, or by sluicing. (B) THE PRECIPITATION PROCESS Clarification of Solutions before Precipitation. The solutions drawn off from the leaching-tanks, or from slime-treatment tanks or filter-presses, usually contain a certain amount of siliceous and other matter in suspension, which it is desirable to remove by a preliminary operation. This is done by settlement in special tanks, or by passing the liquor through small filter-presses, so that only perfectly clear liquid enters the precipitation boxes. Zinc-Boxes. These boxes usually consist of a number of compartments so arranged that the solution flows upward in each compartment through a mass of zinc shavings, then over a parti- tion and downward through a narrow space to the bottom of the next compartment, where it again ascends through a mass of zinc shavings. The shavings are supported on perforated remov- able trays, so that any precipitate which may become detached may fall to the bottom of the box. Lead-Zinc Couple. The efficiency of precipitation is fre- quently increased by previously immersing the shavings in lead acetate, or by allowing a strong solution of lead acetate to drip slowly into the liquid at the head of the boxes. This produces a deposit of finely-divided lead on the zinc and probably sets up an electrical action which brings about the rapid replacement of gold or silver by zinc in the cyanide solution. Conditions for Good Precipitation. The essential points for effective precipitation are : OUTLINE OF OPERATIONS IN THE CYANIDE PROCESS 39 1. Sufficient surface, secured by cutting the zinc into long nar- row strips. 2. Uniform distribution in the boxes, so as to allow no spaces or channels where the solution could pass without coming into effective contact with the zinc. 3. Unimpeded uniform flow of liquid : the shavings must not be so fine, or so tightly packed, that the liquid cannot freely pass. 4. Sufficient time of contact. The flow of liquid through the box must not be too rapid. 5. A sufficient quantity of free cyanide in the solution to be precipitated. 6. A slight but decided alkalinity: Insufficient cyanide or alkali lead to the formation of white deposits, probably consisting of cyanide, ferrocyanide, or hydrate of zinc, which interfere very seriously with the precipitation of the metals. 7. Absence of suspended matter, such as finely divided silica, ore slimes, etc., which would settle on the zinc and also impede precipitation. 8. Absence of large accumulations of soluble salts of the base metals, particularly of copper, which sometimes forms a firmly adherent layer on the surface and completely prevents the replace- ment of gold by zinc. Disposal of Precipitated Solution. The solution leaving the zinc-boxes passes usually to storage-tanks or sumps, where it is made up to the required strength by addition of fresh cyanide, and pumped back as required for the treatment of fresh charges of ore. Accumulation of Zinc in Solution. It might be supposed that, as the zinc is continually dissolving in the precipitation boxes, and the same liquid is repeatedly used in the treatment, the liquid would become gradually saturated with salts of zinc. It is gener- ally found, however, owing to reactions which will be discussed later, that the accumulation of zinc ceases after the solutions have been in use for some time, and that the fresh quantities dissolved are compensated by the introduction of fresh water and the re- moval of old solution in the moisture of the discharged residues. Alternative Methods of Precipitation. Other methods of pre- cipitation have frequently been suggested, since the zinc process, particularly in its earlier forms, presented certain obvious draw- backs. Only two of these, however, have found any extensive application, viz.: 40 THE CYANIDE HANDBOOK 1. Electric precipitation; 2. Zinc-dust precipitation. The former has now been abandoned in favor of improved forms of the zinc process in all but a few localities, and the latter may be said to be still on its trial. Electric Precipitation. The electric process was devised by the late Dr. Wernher Siemens, and was introduced in Russia by A. von Gernet about the same time that the Mac Arthur-Forrest process was introduced in South Africa. Subsequently (in 1893), von Gernet, representing the firm of Siemens & Halske, of Berlin, introduced the process in the Transvaal, and for several years it was extensively used. The method then employed consisted in electrolyzing the solutions with anodes of iron and cathodes of sheet lead, the gold and silver being deposited on the latter, while the anodes were gradually consumed with formation of soluble ferrocyanides, Prussian blue, and other products. The cathodes were removed from time to time and cupeled for recovery of the gold and silver. Advantages and Disadvantages of the Process. The process presented the advantage that weaker cyanide solutions could be precipitated than was at that time possible with the zinc process; the dirty and troublesome operation of cleaning up the zinc-boxes was avoided, and the bullion was obtained in a purer and more marketable form. As, however, a very large area was needed for efficient precipitation, the boxes required were much larger, and in general, the initial outlay, and expenses for maintenance and skilled labor, particularly in the cupellation of the bullion, out- weighed these advantages. When improved forms of the zinc process, particularly the use of the zinc-lead couple, were intro- duced, the electric process was no longer able to hold its own, though a modification of it, which will be described later, is still in use in Mexico. Zinc-Dust Precipitation. The use of zinc-dust for precipita- tion was first suggested by Sulman and Teed (1895), in conjunc- tion with their bromocyanide process. It has been used to a considerable extent in America in conjunction with other forms of the dissolving process. The material used is a by-product obtained in the distillation of commercial zinc, and contains a certain amount of oxide, which is sometimes removed by preliminary treatment with ammonia or an ammonium salt. Being already in OUTLINE OF OPERATIONS IN THE CYANIDE PROCESS 41 a very fine state of division, it presents a large surface and hence forms an efficient precipitant. It is generally agitated with the solution in a special tank, the agitation being obtained by forcing in compressed air. It has been stated, however, that the parti- cles of zinc become coated with gold in such a manner that the zinc is largely protected during the subsequent acid treatment, and hence a low grade of bullion is produced. Consumption of Zinc. The consumption of zinc, when used in the form of shavings, amounts generally to 0.3 Ib. per ton of ore treated; when in the form of dust, the consumption is some- times about the same, but often much higher. The precipitate obtained by zinc-dust is more uniform in composition and may conveniently be pressed into cakes prior to smelting. (C) THE SMELTING PROCESS Nature of Zinc-Gold Precipitate. The material collected when the zinc-boxes are periodically cleaned up consists of finely divided gold and silver, mixed with, and perhaps to some extent actually alloyed with, metallic zinc, together with greater or smaller quan- tities of such impurities as lead, copper, iron, and various insoluble metallic salts, such as carbonates, sulphates, complex cyanogen compounds, silica, alumina, etc. This is removed from the unde- composed zinc shavings as much as possible by rubbing on a sieve; the coarse zinc remaining on the sieve is returned to the boxes, and the remainder, usually, is passed over a finer sieve, the pro- ducts remaining on and passing this sieve being described as " shorts " and " fines " respectively. These are often submitted to separate treatment. Acid Treatment. Formerly, the zinc-gold precipitate was usually simply dried, and smelted direct with suitable fluxes. It is now customary, however, to treat it with dilute sulphuric acid, to remove as much as possible of the metallic zinc and other im- purities soluble in this acid. This treatment is generally carried out in a wooden vat, and is aided by stirring by hand or mechani- cally, and sometimes by heating with steam or otherwise. The vat should be provided with a cover and the operation conducted in a good draft, so that the very poisonous fumes evolved may be rapidly carried off. When the action of the acid is complete the precipitate is allowed to settle and the clear liquid decanted. One or two washes are given with hot water to remove zinc sulphate 42 THE CYANIDE HANDBOOK and other soluble salts, and frequently the residue is put through a small filter-press, after which the cakes are partially dried and smelted with fluxes. Roasting of Zinc-Precipitate. Another method often adopted instead of, or in conjunction with, acid treatment is to roast the precipitate in a small reverberatory furnace. In some cases the precipitate is intimately mixed with niter or other oxidizing agent previous to roasting. This operation may be carried out either before or after acid treatment. It is, however, supposed to generally involve considerable losses of gold and silver. Smelting. The precipitate, after undergoing one or other of these preliminary operations, is then mixed with a flux consisting of borax, soda, some siliceous material, and sometimes fluor-spar and an oxidizing agent. The fusion is generally made in graphite crucibles, but preferably with a clay lining. Various types of reverberatory and other furnaces have been used in place of crucibles; tilting furnaces, operating in the manner of a Bessemer converter, have also been used. The proportion of fluxes is varied according to the nature of the material to be smelted, and is of course determined largely by the preliminary treatment which it has undergone. Where an oxidizer has to be added manganese dioxide has been found to possess some advantages over niter. Casting the Bullion. Immediately before pouring, the bullion is stirred and a sample taken. The slag formed consists chiefly of silicates of zinc and other metals still present after the acid or other treatment. Where the fusion has been properly conducted, the bullion contains 900 to 950 parts gold and silver per 1000. The slag and any other by-products of the smelting process should be tested for gold and silver; they frequently contain sufficient values to justify working over again by special processes. Lead Smelting of Zinc Precipitate. Another method of treat- ing the zinc precipitate, which has been adopted with much success in South Africa and other countries, is the process of smelting with lead, introduced by P. S. Tavener. The precipitate (with or with- out previous acid treatment and filter-pressing) is mixed with litharge, powdered coal, siliceous slag or sand, and a little scrap- iron. On melting in a suitable furnace, the gold and silver are obtained as an alloy with lead; the slag, carrying the zinc, etc., is tapped off, and the lead bullion cast into bars. These are then OUTLINE OF OPERATIONS IN THE CYANIDE PROCESS 43 cupeled on a bone-ash "test," and the litharge formed collected for use in a subsequent operation. The cupeled metal is then broken up and remelted with borax or other flux in clay or graph- ite crucibles. The cost of this process is said to be very much less than that of other methods of smelting; practically all by-products are used up, and the comparatively rich slags from the ordinary fusion methods can be utilized as fluxes in the lead-smelting process. The slag produced in the fusion for lead bullion often contains less than 1 oz. gold per ton. Refining with Oxygen. The suggestion has also been made to refine zinc precipitate by passing oxygen gas or air through the molten material, as in the refining of gold by chlorine. A number of experiments were made with this object by T. K. Rose, 1 and some attempts have been made to apply the principle in practice, though we are not aware that it has been regularly adopted as yet. i Rose, Trans, I. M. M., XIV, 378. PART II CHEMISTRY SECTION I CYANOGEN AND ITS COMPOUNDS WITH NON-METALS WE shall here give a summary of the modes of formation, physical properties, and principal reactions of the more important cyanogen compounds, but we shall confine the discussion prin- cipally to the inorganic compounds of the cyanogen radical. Only such organic derivatives as may be of some interest and importance in connection with the cyanide process or with the manufacture of cyanide will be referred to. For convenience, we may subdivide this discussion of cyan- ogen and its compounds as follows: Section I (pages 47-61). (a) Cyanogen. (6) Hydrocyanic Acid. (c) Compounds of Cyanogen with Non-metals. Section II (pages 62-83). (a) Simple Metallic Cyanides, including easily decomposed Double Cyanides. Section III (pages 84-101). (a) Metallic Compounds of Complex Cyanogen Radicals. (6) Cyanates, Thiocyanates, and Related Compounds. For more detailed information respecting these compounds the reader is referred to such works as Watt's " Dictionary of Chem- istry," revised by Morley and Muir, or to the articles by Joannis in Fremy's "Encyclopaedia." (A) CYANOGEN (CN) (Molecular formula, C 2 N 2 ) Formation. 1. By dry distillation of mercuric cyanide: Hg(CN) 2 = Hg + (CN),. A portion of the cyanogen polymerizes, forming the brown solid substance paracyanogen (CN) X . 47 48 THE CYANIDE HANDBOOK 2. By dry distillation of silver cyanide: 2AgCN = 2Ag + (CN) 2 . 3. By dry distillation of mercuric chloride with a cyanide or ferrocyanide : HgCl 2 + MCN = HgCl + MCI + CN. The gas obtained is in some cases mixed with nitrogen 4. By action of copper salts on cyanides in solution 2CuSO 4 + 4KCN = Cu 2 (CN) 2 + 2K 2 SO 4 + (CN) 2 . 5. Cyanogen cannot be formed by direct union of its elements, but it is a product of certain decompositions of compounds of carbon and nitrogen at a high temperature. Hence it is one of the products obtained by the destructive distillation of coal in the manufacture of illuminating gas. Physical Properties. A colorless gas, with a characteristic sharp odor resembling bitter almonds and prussic acid; it is very irritating to the eyes, and poisonous. Density, 26 (H == 1), or 1.8064 (air = 1). Easily condensed to a transparent, colorless, very mobile liquid, boiling under ordinary pressure at 20.7 C., having sp. gr. 0.866 at 17.2 C., and refractive index 1.316. Liquid cyanogen is a non-conductor of electricity; it has no action on most metallic or non-metallic elements, but dissolves iodine and phosphorus. Cyanogen gas dissolves to some extent in water and alcohol. Chemical Characteristics. Cyanogen in its chemical relation- ships may be regarded from two different points of view. (a) In many reactions it plays the part of a compound radical; that is to say, it behaves as though the nitrogen and carbon con- stituted a single element (often represented by the symbol Cy). In this respect it is in many ways analogous to the elements of the halogen group. Compare (CN) 2 C1 2 Br 2 I 2 KCN KC1 KBr KI AgCN AgCl AgBr Agl Hg(CN) 2 HgCl 2 HgBr 2 HgI 2 (6) From another point of view it belongs to the group of organic substances known as nitrites, which may be regarded a's CYANOGEN AND ITS COMPOUNDS WITH NON-METALS 49 derived from organic acids by the substitution of the radical CN for C0 2 H. Compare Oxalic acid (CO 2 H) 2 Cyanogen (CN) 2 (oxalo-nitrile) Formic acid (H - CO 2 H) Hydrocyanic acid (H CN) (formo-nitrile) " ' Acetic acid (CH 3 CO 2 H) Aceto nitrile (CH 3 CN) Its constitution may be represented by the symbol: Reactions. 1. Decomposed by electric arc into carbon and nitrogen. Not decomposed by heat alone. 2. Burns (in air or oxygen) with a characteristic purplish (peach-colored) flame, forming carbon-dioxide and liberating nitrogen : (CN) 2 + 2O 2 = N 2 + 2CO 2 . 3. Combines with hydrogen when heated with it in a closed tube or under the influence of the electric current: (CN) 2 + H 2 = 2HCN. Some paracyanogen is also formed. 4. The alkali metals combine slowly with cyanogen at ordi- nary temperatures, and with incandescence on warming: (CN) 2 + K 2 = 2KCN. Zinc combines slowly on heating. Iron decomposes it into C and N at a red heat. Hg, Cu, Au, and Pt have no direct action. 5. Certain non-metals (S, P, I) may be vaporized unchanged in an atmosphere of CN. Perfectly dry Cl has no action on it, but in presence of moisture a yellowish oil is deposited, ap- parently a mixture of chlorides of C and N, and also a white, solid, aromatic substance. 6. A solution of cyanogen in water gradually undergoes decomposition, turning first yellow, then brown, and depositing flakes of a brown substance. The first action is probably the formation of cyanic arid hydrocyanic acids: 2CN + H 2 O = HCNO + HCN. Further decomposition then takes place, giving ammonium oxalate, ammonium carbonate, urea, and azulmic acid (the insol- 50 THE CYANIDE HANDBOOK uble brown substance), 1 to which the formula C 4 H 5 N 5 O has been assigned. 7. Dry H 2 S gas does not combine with cyanogen. In presence of water two yellow crystalline substances are formed having the composition : (CN) 2 H 2 S and (CN) 2 (H 2 S) 2 . The first is formed when CN is in excess; the second when H 2 S is in excess. Solutions of these compounds give precipitates with various metallic salts. 8. Concentrated aqueous HC1 converts cyanogen into oxamide and ammonium oxalate, thus acting as a "hydrating agent": 2CN + 2H 2 O = CONH 2 (oxamide). CONH 2 2CN + 4H 2 O = CO 2 NH 4 (ammonium oxalate). CO 2 NH 4 Hydriodic acid gives the same products in the cold, but at a high temperature gives ammonia and ethane, thus: 2CN + 12HI = 2NH 3 + C 2 H 6 + 6I 2 . 9. Alkalis absorb cyanogen and give reactions similar to those of water: Cy 2 + 2KOH = KCy + KCyO + H 2 O. [This reaction is analogous to that of chlorine: C1 2 + 2KOH = KC1 + KC10 + H 2 O.] On treatment of the solution with strong acids, ammonium salts are formed and carbonic acid evolved (by decomposition of the cyanate) : KCNO + 2HC1 + H 2 O = KC1 + NH 4 C1 + CO 2 . A slight decomposition occurs without addition of acids, according to the reaction KCNO + KOH + H 2 O = K 2 CO 3 + NH 3 . When excess of cyanogen is used, the solution becomes brown and deposits an alkaline azulmate. HgO acts like caustic alkalis: if a current of cyanogen be passed into water containing mercuric oxide in suspension, it i Wohler, "Pogg. Ann.," Ill, 177. Richardson, "Philosophical Magazine," XII, 339. CYANOGEN AND ITS COMPOUNDS WITH NON-METALS 51 yields azulmic acid, ammonium carbonate, and mercuric cyanide and oxycyanide. 10. Cyanogen passed over potassium carbonate heated to redness gives a cyanate and cyanide: Cy 2 + K 2 C0 3 - KCy + KCyO + CO 2 . 11. Dry ammonia gas gradually combines, depositing a brown substance (hydrazulmin) : 2(CN) 2 + 2NH 3 = C 4 H6N 6 . Aqueous ammonia gives azulmic acid, oxamide, oxamic acid, and ammonium oxalate. PARACYANOGEN (CN) 4 1 Preparation. This polymer of cyanogen is obtained by heat- ing mercuric cyanide in sealed tubes to about 440 C.; also by decomposition of hydrazulmin (see above) : C 4 H6N 6 = (CN) 4 + 2NH 3 . Properties. A blackish-brown, light, spongy substance, in- soluble in water, soluble in H 2 SO 4 , and precipitated by water from the solution in this acid. It absorbs and condenses gases. (B) HYDROCYANIC ACID (HCy) (Prussic Acid; Hydric Cyanide) Preparation. 1. By heating mercuric cyanide with hydro- chloric acid: HgCyj + 2HC1 = HgCl2 + 2HCy The vapor may be rendered anhydrous by passing through tubes containing CaCO 3 and CaCl 2 , which remove HC1 and H 2 respectively. Only about of the theoretical quantity is ob- tained, owing to HgCl 2 forming a compound with HCN. The yield may be increased by adding NH 4 C1 to the mixture. 2. By passing H 2 S over HgCy 2 or AgCy, heated in a tube to 30 - 40 C. HgS + 2HCy Lead carbonate is used to absorb excess of H 2 S. 3. By action of dilute sulphuric acid on potassium ferrocyanide (the usual method of preparation) : 3H 2 SO 4 = 3KHSO 4 + KFeCy 3 + 3HCy. 1 The formula CN - C = C - CN has been assigned for paracyanogen. I i N = N 52 THE CYANIDE HANDBOOK The most suitable proportions are: Potassium ferrocyanide 8 parts Sulphuric acid (cone.) 9 " Water 12 " The water and acid should be mixed before hand, and cooled before adding to the powdered ferrocyanide. The distilling- flask is heated on a sand bath and the distillate collected in a cooled condenser. 4. By the action of acids on almost any metallic cyanide: e.g., KCy + HC1 = KC1 + HCy. This class of reaction has been utilized for preparing dilute solutions of hydrocyanic acid of known strength for medicinal purposes. (a) AgCy + HC1 - AgCl + HCy (Evcritt). (b} PbCy 2 + H2SO 4 = PbSO 4 + 2HCy (Thomson). (c) Tartaric acid acting on KCy yields a nearly insoluble acid potassium tartratc, together with HCy (Clarke). 5. When a series of electric sparks are passed through a mix- ture of acetylene and nitrogen, there is a gradual combination. Thus: C 2 H 2 + N 2 = 2HCN. Part of the acetylene decomposes with separation of carbon, unless hydrogen be also added. This reaction is of great interest, as it shows how cyanide compounds may be formed synthetically from their elements: acetylene is produced by direct union of carbon and hydrogen, when electric sparks are passed between carbon-points in an atmosphere of hydrogen. 6. By passing a current of dry ammonia gas over carbon con- tained in a porcelain tube heated to redness: C + 2NH 3 = NH 4 CN + H 2 . By passing the vapors through warm dilute sulphuric acid HCy is liberated: NH 4 CN + H 2 SO 4 = NH 4 HSO 4 + IICN. A similar reaction takes place when ammonia and carbon monoxide are passed through a redrhot tube: CO + 2NH 3 = NH 4 CN + H 2 O. 7. By dehydration of ammonium formate: HCO 2 NH 4 = 2H 2 + HCN. CYANOGEN AND ITS COMPOUNDS WITH NON-METALS 53 This is a general reaction for forming the "nitriles" of the corresponding organic acids; thus, ammonium oxalate yields cyanogen, and ammonium acetate yields acetonitrile when acted on by a dehydrating agent. 8. By destructive distillation of ammonium oxalate or oxam- lde: (a) (C0 2 NH 4 ) 2 = 3H 2 + O + 2HCN. (6) (CO 2 NH 2 ) 2 = H 2 O + 3O + 2HCN. 9. Hydrocyanic acid occurs ready formed in, or is a product of the decomposition of, many vegetable substances. The ker- nels of many fruits, such as peaches, cherries, almonds, etc., and certain leaves, as those of the cherry-laurel, contain a substance known as amygdalin, which is decomposed by water under the influences of a vegetable ferment (emulsin), the latter undergoing no apparent change in the reaction. This reaction results in the formation of benzaldehyde, glucose, and hydrocyanic acid: CaoHwNOu + 2H 2 O = C 6 H 5 COH + 2C 6 H 12 O 6 + HCN. Amygdalin Benzaldehyde Glucose 10. Hydrocyanic acid occurs as one of the products in many reactions of organic bodies; e.g., in the action of ammonia on chloroform: HCN . in the combustion of methylamine: CH3NH 2 + O 2 = 2H 2 O + HCN; and in the action of alkalis on nitrobenzene. The anhydrous acid may be prepared by gently heating the aqueous acid, passing the vapors through vessels containing cal- cium chloride, and condensing the distillate by a freezing mixture. Physical Properties. A volatile colorless liquid; extremely poisonous. Boiling-point + 26.1 C. When allowed to evaporate in the air, a portion solidifies in translucent orthorhombic prisms, having a melting-point 14 C. Density of liquid, 0.7058 at 7 C. 0.6969 at 18 C. Density of vapor, .97 to .90 [air = 1] 13.5 [H = 1]. Very soluble in water. The addition of a small quantity of water to the anhydrous acid raises the boiling-point and lowers the melting-point. The solution of HCN in water is accompanied by condensation and lowering of temperature (an unusual effect). 54 THE CYANIDE HANDBOOK A mixture of 60 per cent. HCN and 40 per cent. H 2 O has the min- imum melting-point ( 22.5 C.). Reactions. 1. Heated in closed vessels at 100 C., it forms a black mass, which when afterward heated in an open tube at 50 C. gives off NH 4 CN and (CN) 2 leaving a carbonaceous residue. A current of HCN passed through a hot tube gives (CN) 2 , H, N and C.; this reaction is reversible (see Cyanogen: Reaction No. 3). 2. When the vapor is decomposed by electric sparks it yields acetylene and nitrogen 2HCN r; C 2 H 2 + N 2 (reversible). An electric current passed between platinum electrodes, through liquid HCN, gives H at the negative pole, but no gas is evolved at the positive pole, as the cyanogen attacks the plati- num. 1 3. Burns (in air or oxygen) with a white flame, bordered with purple. Explodes in O with great violence, yielding CO 2 , H 2 0, N, and a little HNO 3 . 4. Action of non-metals: Sulphur vapor gives the same pro- ducts as H 2 S on CN (see Cyanogen: Reaction No. 7). Phosphorus has no action. Dry Cl and anhydrous HCN give HCN + C1 2 = HC1 + CNC1. Cl and aqueous HCN give only a little CNC1, the chief products being CO 2 , CO, HC1 and NH 3 . Nascent H gives methylamine: HCN + 2H 2 = CH 3 NH 2 . 5. Action of strong acids: Strong HC1 and moderately con- centrated H 2 SO 4 convert it into ammonium formate, which is decomposed on heating, yielding formic acid: HCN + 2H 2 O = H -C0 2 NH 4 . HCO 2 NH 4 + HC1 = NH 4 C1 + H CO 2 H. The gaseous acids (HC1, HBr, HI) form addition-compounds with anhydrous HCN, usually of the type HX-HCN. Concentrated H 2 SO 4 , heated in a closed vessel with HCN to 30-35 C., turns brown; SO 2 , CO 2 , and N are given off. 6. Action of alkalis: At ordinary temperatures, cyanides are produced: HCN + KOH = RCN 1 Davy, "Journ. Science and Arts," I, 288. CYANOGEN AND ITS COMPOUNDS WITH NON-METALS 55 On heating, a formate of the metal is produced and ammonia is evolved: KCN + ^O = HCO 2 K + NH 3 . 7. Action of metallic oxides : Most metallic oxides yield cyan- ides, in some cases with evolution of cyanogen. 8. Action of metallic salts: Hydrocyanic acid is a very weak acid; when absolutely pure it does not redden litmus: it is set free from its compounds by almost every acid, even by CO 2 . Alkaline carbonates and most acids are, however, decomposed by HCN, the reaction being to some extent reversible. Salts of metals which give insoluble cyanides are decomposed by HCN with precipitation of the cyanide (e.g., salts of Ag,Cu, Pb). Alkaline polysulphides give thiocyanates (MCNS). 9. Spontaneous decomposition: Anhydrous HCN may be preserved in sealed vessels without decomposition. If, however, a trace of ammonia or alkali be present, it decomposes, espe- cially if exposed to light, turning brown and yielding ammonium cyanide and azulmic acid, probably first polymerizing to (HCN) 3 . Poisonous Action of Hydrocyanic Acid. The poisonous action seems to be chiefly paralysis of the nervous centers. Anything which will act as a stimulus to the nerves, such as inhaling chlo- rine or ammonia, will act as an antidote, although the products of the reaction in these cases are also strong poisons. The symp- toms produced by inhaling small quantities of hydrocyanic acid gas are headache, giddiness, a peculiar sensation at the back of the throat, and sometimes nausea. Long continual respiration of small quantities of hydrocyanic acid, as in some operations in the cyanide plant, particularly in cleaning up, may give rise to chronic affections of the heart and throat (palpitations and hoarse- ness). The poisonous action of soluble cyanides is due to the liberation of HCN by the acids of the stomach. Various antidotes have been proposed, such as salts of cobalt, hypodermic injection of peroxide of hydrogen, etc., but the simplest and most effective seems to be freshly prepared ferrous carbonate. H. C. Jenkens * recommends the following solutions to be prepared and kept handy as an antidote for cases of cyanide poisoning: (a) A hermetically sealed bottle containing 7 grams ferrous sulphate in 30 c.c. of distilled water, previously well boiled, (b) A hermetically sealed bottle containing 1 grams i Trans. I. M. M.. XIII. 484. 56 THE CYANIDE HANDBOOK NaOH in 300 c.c. water, (c) A tube containing 2 grams magnesia (MgO). When used, the sealed ends of the two bottles are to be broken off and the solutions mixed, the magnesia added, and the whole administered as rapidly as possible. Anything which will induce vomiting also has a beneficial effect. Soluble cyanides introduced into cuts in the skin act as blood poisons, producing sores and eruptions. (C) COMPOUNDS OF CYANOGEN WITH NON-METALS The compounds with sulphur have been already mentioned. The following compounds with the haloid elements are known: CNC1 CNBr CNI (CNC1) 3 (CNBr) 3 The chloride CNC1 is gaseous at the ordinary temperature; the others are solid. CHLORIDES OF CYANOGEN GASEOUS CHLORIDE (CNC1) Prepared: 1. By action of Cl on anhydrous HCN; the excess of Cl is removed by agitation with mercury. 2. By action of Cl on mercuric cyanide (solid or in saturated solution) : Hg(CN) 2 + 2C1 2 = HgCl 2 + 2CNC1. 3. By passing Cl through a solution of KCN surrounded by a freezing-mixture: the product is then distilled and the gaseous chloride collected and liquefied in a cooled condenser: KCN + C1 2 = KC1 + CNC1. Physical Properties. Very volatile liquid; boils at 15.5 C. Density of vapor, 20.2 (H - 1). Solidifies at - 5 C. Very poisonous; has extremely irritating odor; produces coughing and tears. Soluble in alcohol, ether, and water. Reactions. 1. Polymerizes, forming the solid chloride (CNC1) 3 in presence of a slight excess of Cl. 2. Decomposed when heated with certain metals, yielding a cyanide or cyanogen: CyCl + 2K = KC1 + KCy. 3CyCl + Sb = SbCl 3 + 3Cy. CYANOGEN AND ITS COMPOUNDS WITH NON-METALS 57 3. Decomposed by water in two stages: CNC1 + H 2 O = HC1 + HCNO. HCNO + HC1 + H 2 O = NH 4 C1 + CO 2 . 4. Alkalis give an analogous reaction: CNC1 + 2KOH = KC1 + KCNO + H 2 O KCNO + 2HC1 + H 2 O = KC1 + NH 4 C1 + CO 2 . 5. Dry ammonia gas forms ammonium chloride and cyan- amide: CNC1 + 2NH3 = NH4C1 + CN . NH2 6. Does not precipitate AgCl or AgCN from silver nitrate. SOLID CHLORIDE (CNC1) 3 Preparation. 1 . By exposing to sunlight a mixture of chlorine with excess of Hg(CN) 2 in sealed tubes: 3Hg(CN) 2 + 6C1 2 = 3HgCl 2 + 2(CNC1) 3 . 2. By action of phosphorus pentachloride on cyanuric acid: (CNOH) 3 + 3PC1 5 = 3POC1 3 + 3HC1 + (CNC1) 8 . 3. By slightly heating dry KCNS in a current of dry chlorine: 3KCNS + 6C1 2 = 3KC1 + 3SC1 2 + (CNC1) 8 . The (CNC1) 3 crystallizes from the mixture. Physical Properties. Small needles (density 1.32), melting at 140 C. to a transparent liquid which boils at 190 C. Vapor density, 6.35 (air = 1). Very poisonous. Easily pro- vokes tears. Smells of chlorine, but also has an odor suggestive of mice. Reactions. 1. Decomposed by water (more rapidly by alkalis), forming a cyanurate: (CNC1). + 3H 2 = 3HC1 + (CNOH) 3 . 2. Ammonia gives chlorocyanuramide: (CN) 3 (NH 2 ) 2 -C1. BROMIDES OF CYANOGEN MONOBROMIDE (CNBr) Preparation. 1. By adding a strong solution of KCN grad- ually to an excess of bromine until the latter is discolored, then distilling and condensing the product KCN + Br 2 = KBr + CNBr. 58 THE CYANIDE HANDBOOK 2. By acting upon sodium bromide and bromate with sodium cyanide and sulphuric acid at 70 C. (Gopner.) 1 2NaBr + NaBrO 3 + 3NaCN + 3H 2 SO 4 = 3Na 2 SO 4 + 3H 2 O + SCNBr. 3. By adding bromine drop by drop to a solution of HCN: HCN + Br 2 = HBr + CNBr. When Br is added to an excess of cyanide (reversing the order of procedure in the first method), there is a tendency to formation of the polymer (CNBr) 3 . 4. One part of Br is added gradually to two parts mercuric cyanide: Hg(CN) 2 + 2Br 2 = HgBr 2 + 2CNBr. The same reaction may be brought about by adding a mixture of mercuric cyanide and hydrochloric acid to bromine in a cooled vessel, and then distilling off the CNBr by placing the vessel in hot water. Physical Properties. A transparent crystalline solid, said to melt at 4 12 C.; at ordinary pressures, however, it volatizes without melting and is gaseous above 61 C. It is volatile above 15 C. It has a very penetrating odor, attacking the eyes and causing a flow of tears. Very poisonous, but less so than HCN. Very soluble in water and alcohol. With water it forms a hydrate which is less volatile than the pure bromide of cyanogen. Solution bleaches litmus without reddening it. Crystallizes in long hexagonal needles, which soon change to cubic or tabular crystals. 2 Reactions. 1. Gradually decomposed by water with forma- tion of cyanic acid and hydrobromic acid: CNBr + H 2 O = HCNO + HBr. When an aqueous solution of CNBr is evaporated to dryness, ammonium bromide is obtained and carbon dioxide evolved: CNBr + 2H 2 O = NH 4 Br + CO 2 . 2. Decomposed rapidly by alkalis, with formation of a bro- mide and cyanate: CNBr + 2KOH = KBr + KCNO + H 2 O. 3. Acids (H 2 SO 4 , HC1, HNO 3 ) dissolve it without decomposi- tion, but S0 2 is oxidized by CNBr in presence of water: CNBr + SO 2 + 2H 2 O = HBr + H 2 SO 4 + HCN. 1 See "Journ. Soc. Chem. Ind.," XXV, 1130. 2 H. L. Sulman, "Proc. Chem. Met. and Min. Soc. of South Africa," I, 114. CYANOGEN AND ITS COMPOUNDS WITH NON-METALS 59 4. Soluble cyanides act rapidly, probably evolving cyanogen as follows: KCN + CNBr = KBr + (CN)a> It is possible, however, that hydrocyanic acid is evolved, as ollows: CNBr + 2KCN + H20 = KCNO + KBr + 2HCN The liberation of cyanogen in an active form appears to be the cause of the rapid solvent power on gold of the mixture of KCN and CNBr. 5. Action on Metals. Certain metals are attacked, giving a metallic bromide with evolution of cyanogen: Hg + 2CNBr = HgBr 2 + (CN) 2 (in solution). A similar reaction occurs with the vapor of antimony. Gold and silver are rapidly dissolved by a mixture of CNBr and an alkali cyanide, forming the soluble double cyanides: 3KCN + CNBr + 2Au = 2KAu(CN) 2 + KBr. 3KCN + CNBr + 2Ag = 2KAg(CN) 2 + KBr. This represents the final effect of the reaction, but it may take place in several stages. Bromide of cyanogen alone does not attack gold or silver. 6. Ammonia decomposes cyanogen bromide, giving bromide and cyanate of ammonium: CNBr + 2NH 3 + H 2 O = NH 4 Br + NH 4 CNO. POLYMER OF CYANOGEN BROMIDE (CNBr) 3 This substance is sometimes obtained during the preparation of ordinary cyanogen bromide. It is formed by heating CNBr for 5 to 6 hours at 130-140 in sealed tubes, but is partially decom- posed with liberation of bromine. It is obtained in a purer con- dition by heating CNBr with anhydrous ether. This is a solid body, melting at 300 C., and boiling, with decomposition, at a higher temperature. It is insoluble in alco- hol and nearly insoluble in ether. Exposed to moist air, or heated with water to 100 in sealed tubes, it forms cyanuric acid: (CNBr) 3 + 3H 2 O = (CNOH) 3 + 3HBr. IODIDE OF CYANOGEN (CNI) Preparation. 1. Not formed by direct union of the ele- ments. 60 THE CYANIDE HANDBOOK 2. Formed by action of free iodine on the cyanides of mercury, silver, or potassium: Hg(CN) 2 + 2I 2 = 2CNI + HgI 2 . KCN + I 2 = CNI + KI. The iodide of cyanogen is separated from the mixture by dis- tilling at a temperature below 135 C. Physical Properties. Forms long needles, soluble in water, alcohol or ether. Crystallizes from these solutions in four-sided tables. Volatile at ordinary temperatures and boils below 100 C. Has a penetrating odor resembling CN and I. Poisonous. Does not redden litmus. Reactions. 1. Decomposed by heat, on passing through a red-hot tube: 2CNI = (CN)z + Iz 2. Gradually decomposed by water: CNI + H 2 O = HCNO + HI. 3. Dissolved by alkalis with formation of a cyanide and an iodate, a little cyanate being formed at the same time: 6KOH + 3CNI = 3KCN + KIO 3 + 2KI + 3H 2 O. 2KOH + CNI = KCNO + KI + H 2 O. (Note difference from CNBr.) 4. Action of Acids. HC1 and H 2 SO 4 slowly decompose it with liberation of HCN and I at ordinary temperatures. Sulphurous acid acts as follows: CNI + SO 2 + 2H 2 O = HI + H 2 SO 4 + HCN. 5. Soluble cyanides react as with CNBr, but less vigorously. The action on metals is also similar. 6. Action of Non-metals. Chlorine does not decompose dry CNI. Phosphorus has a violent action, often accompanied by light; iodide of phosphorus is formed and cyanogen probably liberated. (Phosphorus has only a slight action on CNBr.) 7. Dry H 2 S gives: 2CNI + H2S = SI 2 + 2HCN. In presence of water, the reaction is: CNI + H 2 S = S + HI + HCN 8. Ammonia is slowly absorbed, giving ammonium iodide and cyanamide: CNI CYANOGEN AND ITS COMPOUNDS WITH NON-METALS 61 9. Solution of CNI in water gives no precipitate with solutions of silver salts. 10. On mixing, in the following order, solutions of CNI, KOH, FeSO 4 and HC1, a green precipitate is obtained. This reaction does not occur with CNC1 or CNBr, unless the FeSO 4 be added before the KOH. SECTION II SIMPLE METALLIC CYANIDES (INCLUDING THE EASILY-DECOMPOSED DOUBLE CYANIDES) WE shall now consider those metallic compounds in which the group CN occurs as the negative radical. This includes all the substances which can correctly be described as metallic cyan- ides. Compounds such as ferrocyanides, nitroprussides, thio- cyanates, etc., have little in common with the true cyanides, and are more conveniently considered as compounds of entirely dis- tinct radicals, Fe(CN) 6 , Fe(CN) 5 NO, (SCN), etc. The true cyanides have the general formula /CN m' C = N, m" <5!?l m'"r-CN, etc. \CN \ CN where m', m", etc., are positive monad, dyad, etc., elements or radicals. They show, however, a great tendency to combine with each other, forming compounds of the type miCN m 2 CN, possibly mi C = N. m a C = N. where n\ l and m 2 are different positive radicals. It is also possible that isomeric modifications exist in which the positive radical is united directly with carbon or nitrogen: m C = N'" orm Nv = C. corresponding with the organic groups known as nitriles and car- bamides. General Modes of Formation. 1. By direct action of cyan- ogen on metals: (CN)JJ + 2R = 2RCN 2. By action of cyanogen on dissolved alkalis with simulta- neous formation of a cyanate: (CN) 2 + 2KOH = KCN + KCNO + H 2 O. 62 SIMPLE METALLIC CYANIDES 63 3. By action of hydrocyanic acid on metals, oxides, or metallic salts. 4. By double decomposition, with metallic salt and other cyanides: MjCN + MjiX = M]X where M 1? M 2 are two different metallic radicals and X a negative (acid) radical. 5. By calcination of nitrogenous organic matter at a high temperature with alkalis or alkaline carbonates. 6. By passing free nitrogen over a mixture of carbon and an alkaline carbonate heated to redness. 7. By calcining certain organic salts such as acetates and tartrates, with the nitrates or nitrites of the alkalis. General Properties. Cyanides of the alkalis and alkaline earths are soluble in water; the simple cyanides of mercury and cadmium are also soluble; other simple cyanides are insoluble. Many of the insoluble cyanides unite with alkaline or other soluble cyanides to form soluble double salts. Soluble compounds are also formed with ammonium salts, and with certain other metallic salts. Most soluble cyanides have an alkaline reaction and are very poisonous. The soluble cyanides are colorless. The insoluble cyanides often have characteristic colors. General Reactions. 1. Action of heat: The alkali cyanides, when anhydrous, are not decomposed by ordinary heat. KCN and NaCN may be fused without decomposition. Hydrated NaCN is partially decomposed with evolution of NH 3 . Of the cyanides of the alkaline earths, BaCy 2 is the most stable and MgCy 2 the least. The former may be heated to redness without decomposi- tion. Some cyanides (Zn, Cu) are decomposed by heat into a mixture of C with the metal or a metallic carbide, N being given off; others (Hg, Ag) are decomposed into metal and CN. 2. Action of Oxygen. Alkaline cyanides are converted into cyanates, and by stronger heating into carbonates with evolution of N. The reaction takes place more readily in presence of an oxidizing agent. With other cyanides, CO 2 and N are given off, and an oxide of the metal remains. Cyanides explode in contact with certain oxidizing bodies (nitrates, chlorates). 3. Action of Chlorine. Varies under different conditions. In 64 THE CYANIDE HANDBOOK general a chloride is formed, with free CN and more or less CNC1. Bromine and iodine act similarly. 4. Action of Water. Decomposes all cyanides at a higher or lower temperature, generally at 100 C. Boiling solutions of alkaline cyanides are converted into alkaline formates with evolu- tion of ammonia: mCN + 2H 2 O = HCO 2 m + NH 3 . Cyanides of the heavy metals, heated in presence of water, decompose, yielding CO 2 , CO, HCN, NH 3 , the metal remaining as oxide, or in the free state mixed with a little carbon. Cyanides of Hg and Ag, heated with water in sealed tubes to 280 C., give ammonium carbonate and the free metal. 1 5. Action of Acids. Dilute mineral acids decompose most metallic cyanides with evolution of HCN. Many organic acids also decompose cyanides. Weak acids, such as CO 2 , decompose the cyanides of the alkalis and alkaline earths, and some double cyanides. The cyanides of Hg, Ag, and Pd are not decomposed by dilute HC1, but on the contrary HCN displaces HC1 in the chlorides of these metals. Nitric acid decomposes cyanides with liberation of N and C0 2 and formation of oxalic acid. Concen- trated H 2 SO 4 gives a sulphate of the metal, and of ammonium, and forms CO. 6. Mercuric oxide decomposes most cyanides, when boiled with their solution, giving mercuric cyanide (HgCy 2 ) and the corresponding oxide. This reaction is owing to the great stability of mercuric cyanide. 1 DETAILS RESPECTING CERTAIN METALLIC CYANIDES Ammonium Cyanide (NH 4 CN) : Preparation. 1. By heat- ing ammonium ferrocyanide and condensing the product in a cooled receiver: (NH 4 ) 4 FeCy 6 = FeCy 2 + 4NH 4 Cy. 2. By heating ammonium salts with certain cyanides or ferro- cyanides, e.g.: K 4 FeCy 6 + 4NH 4 C1 = 4NH 4 Cy + 4KC1 + FeCy 2 . HgCy 2 + 2NH 4 C1 = HgCl 2 + 2NH 4 Cy. 3. By direct union of NH 3 gas and HCN. i It is probable that most cyanides are ionized in dilute solution. With mer- curic cyanide, however, this is not the case. SIMPLE METALLIC CYANIDES 65 4. By passing NH 3 over red-hot carbon: 2NH 3 + C = NH 4 CN + H 2 . also 2NH 3 + CO = NH 4 CN + H 2 O. It is also one of the products of combustion of illuminating gas containing ammoniacal vapors, and is formed by action of HNO 3 on some organic substances. Physical Properties. Crystallizes in cubes. Very volatile and poisonous. Boils about 36 C., and is easily dissociated into HCN and NH 3 . Vapor density at 100 C. 0.79 (air = 1). Reactions. 1. Burns in air or oxygen, giving a yellowish flame, depositing ammonium carbonate. 2. Very unstable, particularly at high temperatures. Turns brown and passes into azulmic acid. 3. Decomposed by chlorine or bromine: NH 4 Cy + C1 2 = CyCl + NH 4 C1. CYANIDES OF ALKALI METALS Potassium Cyanide (KCN) : Preparation (methods mostly applicable also to sodium cyanide). 1. By direct action of hydrocyanic acid on potash (best in alcoholic solution) : HCN + KOH = H 2 O + KCN. 2. By decomposition of ferrocy anide : (a) By heating in closed vessels; (b) by heating with dry potassium carbonate; (c) by action of metallic sodium (giving a mixture of NaCy and KCy). These methods will be discussed in detail under the title "Manufacture of Cyanide." (a) The effect of heating K 4 FeCy 6 by itself in a closed vessel is to partially break it up into N and a carbide of iron, mixed with KCN. (b) When heated with K 2 CO 3 : K 4 FeCy 6 + KzCOa = 5KCy + KCyO + CO 2 + Fe. Similar reaction with Prussian bhie and K 2 CO 3 . (c) With metallic sodium: K 4 FeCy e + Na 2 = 2NaCy + 4KCy + Fe. 3. By calcining nitrogenous organic matter with K 2 CO 3 . A part of the carbon liberated reduces K 2 CO 3 to K, which then reacts with N and more C to form KCN. 66 THE CYANIDE HANDBOOK 4. By passing atmospheric nitrogen over a mixture of carbon and potash heated to redness: Nz + 4C+ K 2 CO 3 = 2KCN + 3CO. or N 2 + 3C + 2KOH = 2KCN + CO + H 2 O. Probably K is first liberated in this case also. 5. By precipitating mercuric cyanide with potassium sul- P hide: HgCy 2 + K 2 S = HgS + 2KCy. Physical Properties. Anhydrous. Crystallizes in cubic sys- .\ tern by fusion and in cubic octahedra from aqueous solutions. \ Deliquescent; readily soluble in water; insoluble in absolute alcohol. Strongly alkaline taste and reaction. Very poisonous. Gives off HCN when exposed to air. Melts at red heat and volatil- izes at a white heat without decomposition. Reactions. 1. When heated in contact with air a little KCNO is always formed. When heated with metallic oxides, such as those of Pb, Sb, Fe, As, Sn, or Mn (MnO 2 ), larger quanti- ties of cyanate are produced, and the oxide is reduced to metal, or, in the case of MnO 2 , to MnO. When heated with oxidizing agents (nitrates or chlorates) it explodes violently. KMnO 4 gives as oxidation products CO 2 , HCO 2 H,(CO 2 H) 2 , CO(NH 2 ) 2 , HNO 2 and NH 3 . Chloride of lime gives calcium cyanate. Potassium sulphate gives cyanate and sulphide: K 2 SO 4 + 4KCN = 4KCNO + KaS. 2. The aqueous solution decomposes even at ordinary temper- ature, giving potassium formate and ammonia, but much more rapidly on boiling: KCN + 2H 2 O = HCO 2 K + NH 3 . 3. On electrolysis it gives CO 2 , NH 3 and KOH. 4. When melted with sulphur it forms a thiocyanate: KCN + S = KSCN. Aqueous solution of KCN does not dissolve sulphur, but poly- sulphides react with KCN in solution as follows: 2KCN + K^Ss = 2KCNS + KaS. Selenium reacts in a similar way, but also dissolves in KCN solution, even at ordinary temperature, to form seleno cyanide SIMPLE METALLIC CYANIDES 67 KSeCN (soluble in water to a colorless solution). When fused with tellurium it forms potassium telluride (giving a pink solution) : 2KCN + Te = K 2 Te + (ON),. 5. Sulphureted hydrogen passed into a solution of KCN gives small yellow needles of " chrysean " and potassium and ammonium sulphides in solution: 4KCN + 5H2S = 2K2S + NH 4 HS + CJIsNaSa. 6. On heating with alkalis, it is transformed first into HCO 2 K and NH 3 ; but on heating to redness H is evolved and an alkaline carbonate remains. 7. Thiosulphates convert it into thiocyanate: KaSaO, + KCN = K.S03 + KCNS. 8. The reactions with Cl, Br and I' are given under Section I (c). 9. Certain metals are dissolved in aqueous solution of potas- sium cyanide: (a) In presence of oxygen: 2Au + 4KCy + O + H 2 O = 2KAuCy 2 + 2KOH. 2Ag + 4KCy + O + H 2 O = 2KAgCy 2 + 2KOH. (b) With evolution of hydrogen: Zn + 4KCy + 2H 2 O = K 2 ZnCy 4 + 2KOH + H 2 . (also with Cu, Fe and Ni). (c) The following are insoluble: Pt, Hg, Sn. (Hg, however, dissolves to some extent, under the conditions of cyanide treat- ment, from amalgamated plates in the battery.) 10. Many metallic oxides dissolve to form the corresponding cyanide, or double cyanide of the metal and potassium. Sodium Cyanide (NaCN) : Formation (see under " Potassium Cyanide," above). 1. By. passing a current of anhydrous HCN into alcoholic solution of caustic soda. Sodium cyanide is pre- cipitated in the anhydrous form, and must be rapidly washed with alcohol and dried. 2. When ammonia is passed over a heated mixture of metallic sodium, or sodium carbonate and carbon, it forms sodium cyanide, with sodium cyanamide as intermediate product: Na 2 + C + 2NH 3 - N a2 CN 2 + 3H 2 . N a2 CN 2 + C - 2NaCN. 68 THE CYANIDE HANDBOOK 3. By fusing barium cyanide with sodium carbonate: BaCy 2 + Na 2 CO 3 = BaCO 3 + 2NaCy. The sodium cyanide is separated from the insoluble BaCO 3 by extracting with water. 4. By fixation of atmospheric nitrogen, using the following series of reactions: CaO + 2C"+ 2N = CaCN 2 + CO (in electric furnace). 2CaCN 2 + 4H 2 O = (CN NH 2 ) 2 + 2Ca(OH) 2 . (CNNH 2 ) 2 + Na,CO 3 + 2C = 2NaCN + NH 3 + SCO + H + N (These processes will be more fully discussed under "Manu- facture of Cyanide. ") Properties. Sodium cyanide is a white crystalline solid, similar in general appearance and properties to potassium cyanide. It is very deliquescent, and forms two hydrates, of the composi- tion, NaCN 2H 2 O and 2NaCN H 2 O, respectively. The first is white, crystallizing in thin plates, the second yellowish. Both these hydrates are unstable, and give off ammonia, even when protected from the air, at the ordinary temperature. Reactions. Analogous to those of potassium cyanide. Cer- tain alleged differences in the behavior of the two cyanides in the treatment of ores will be discussed later. CYANIDES OF THE ALKALINE EARTH METALS Barium Cyanide (Ba(CN) 2 ): Preparation. 1. By calcination of Ba 2 Fe(CN) 6 . 2. By passing anhydrous HCN over Ba(OH) 2 . 3. By passing a current of air over a mixture of BaO and carbon at a high temperature, or, better, by passing air previously freed from oxygen and water vapor over a heated mixture of BaCO 3 and charcoal, tar, or other carbonaceous matter. Barium carbide (BaC 2 ) is first formed, which combines with nitrogen, giving Ba(CN) 2 . Properties. White, anhydrous, or crystallizing as BaCy 2 - 2H 2 O. Loses water in vacuo at ordinary temperature, forming BaCy 2 H 2 O. Very soluble and deliquescent. Completely dehy- drated by heating in a current of dry air at 75 and then at 100 C. Reactions. 1. Decomposes on heating: Ba(CN) 2 + 3H 2 O = BaCO 3 + NH 3 + HCN + H 2 . Ba(CN) 2 + 2H 2 = BaO + CO + NH 3 + HCN. SIMPLE METALLIC CYANIDES 69 2. Decomposed by carbon dioxide: Ba(CN) 2 + CO 2 + H 2 O = 2HCN + BaCO 3 . 3. Fused with alkaline carbonates it yields BaCO 3 and the corresponding alkaline cyanide. Strontium Cyanide (Sr(CN) 2 ). Analogous in preparation and properties to the above. Crystallizes as SrCy 2 . 4H 2 O. Less stable than BaCy 2 ; more so than CaCy 2 . Calcium Cyanide (Ca(CN) 2 ): Preparation. 1. By action of HCN on lime. 2. By passing nitrogen over calcium carbide, the chief product is calcium cyanamide: CaC 2 + N 2 = CaCN 2 + C (at 800 C.). When this is heated to a higher temperature (preferably with addition of salt to facilitate the reaction) it is decomposed as follows: CaCN 2 + C = Ca(CN) 2 Physical properties. Said to have been obtained in anhy- drous cubes, but very unstable. Reactions. 1. Dilute solutions are tolerably stable, but a concentrated solution placed in vacuo gradually decomposes, blackens, and gives the ordinary products of decomposition of HCN. In presence of dehydrating agents and KOH, small crys- talline needles of the oxycyanide, 6CaOCaCy 2 -15H 2 O, are formed. 2. Concentrated alcohol gives a precipitate of calcium hydrate. Magnesium Cyanide (Mg(CN) 2 ) [existence doubtful] : Prepara- tion. 1. By dissolving freshly precipitated Mg(OH) 2 in HCN (Scheele). 2. By passing CO or CO 2 over heated magnesium nitride: M g3 N 2 . Properties. Very unstable. Solution decomposes on evap- oration, liberating HCN and leaving Mg(OH) 2 . Decomposed by CO 2 with precipitation of MgCO 3 . Magnesium salts decompose alkaline cyanides thus: MgSO 4 + 2KCN + 2H 2 O = KaSC^ + Mg(OH) 2 + 2HCN. CYANIDES AND DOUBLE CYANIDES OF ZINC Zinc Cyanide (Zn(CN) 2 ): Preparation. 1. By precipitating a salt of zinc with an alkaline cyanide in suitable proportions: 2KCy + ZnSO 4 = KO 4 + ZnCy 2 . 70 THE CYANIDE HANDBOOK 2. By the action of HCN on the oxide or acetate of zinc. These reactions are incomplete and appear to be reversible: Zn(CH 3 C0 2 ) 2 + 2HCN ^ 2CHsCO 2 H + Zn(CN) 2 . ZnO + 2HCN ^ H 2 O + Zn(CN) 2 . Properties. White, amorphous or crystallized in orthorhom- bic prisms. Insoluble in water. Unalterable when dry. Reactions. 1. Heated in a closed vessel, gives a black car- bonaceous residue. 2. Dissolves in alkalis, forming a double cyanide and zincate: 2ZhCy 2 + 4KOH = K 2 ZnCy 4 + Zn(OK) 2 + 2H 2 O. probably in two stages: '(a) ZnCy 2 + 2KOH = 2KCy + Zn(OH) 2 (6) ZnCy 2 + Zn(OH) 2 + 2KCy + 2KOH = K 2 ZnCy 4 + Zn(OK) 2 + 2H 2 O. 3. Dissolves in alkaline cyanides, forming double cyanides (ZnCy 2 2mCy) : ZnCy 2 + 2mCy = m 2 ZnCy 4 [m = K, Na, NH 4 ]. Zinc-Potassium Cyanide (K 2 Zn(CN) 4 ): Preparation. 1. By dissolving ZnCy 2 in excess of KCy (see above). 2. By acting on metallic zinc with KCy in excess: Zn + 4KCy + 2H 2 O - K 2 ZnCy 4 + 2KOH + H 2 . 3. By dissolving ZnO or ZnCO 3 in a mixture of KCy and HCy: ZnCO 3 + 2KCy + 2HCy = K 2 ZnCy 4 + H 2 O + CO 2 . Properties. Regular transparent colorless octahedra. Can be fused without decomposition. Very soluble in water at all tem- peratures. Reactions. 1. Decomposed by dilute acids, first with pre- cipitation of ZnCy 2 , which is further decomposed by excess of acicl : K 2 ZnCy 4 + 2HC1 - ZnCy 2 + 2HCy + 2KC1. ZnCy 2 + 2HC1 = ZnCl 2 + 2HCy. 2. Gives precipitates of insoluble double cyanides of zinc with various solutions of metallic salts (e.g., those of Ba ; Ca, Ni, Co, Al, Fe, Hg, Cu, Pb). 3. With silver nitrate it gives first a precipitate of ZnCy 2 , and on adding excess of AgNO 3 this is decomposed, giving AgCy: K 2 ZnCy 4 + AgNO 3 - ZnCy 2 + KAgCy 2 + KNO 3 . ZnCy 2 + 2AgNO 3 = 2AgCy + Zn(NO 3 ) 2 1 This reaction is discussed in detail in Section V. SIMPLE METALLIC CYANIDES 71 Other Double Cyanides of Zinc. The double cyanides of zinc, with NH 4 , Na, K, and Ca, are soluble in water; the remainder are more of less insoluble. The double cyanide of Zn and ammonium is unstable, and decomposes in solution: (NH 4 ) 2 ZnCy 4 ^ ZnCy 2 + 2NH 4 Cy. CYANIDES OF THE IRON GROUP The simple cyanides of iron, manganese, nickel, and cobalt are unstable; the chromic cyanide is tolerably stable. No simple cyanide of aluminium is known. All these compounds show a great tendency to form double salts by reacting with alkaline cyanides. The double cyanides of nickel are easily decomposable, and are more or less analogous to those of zinc; they are of the general type NiCy 2 2mCy, and are decomposed by dilute acids in the same way. Thus: 2HC1 + NiCy 2 2KCy = NiCy 2 + 2KC1 + 2HCy. NiCy 2 + 2HC1 . = NiCl 2 + 2HCy. The remainder form characteristic compounds of complex cyanogen radicals, which are not decomposed by dilute acids, and will be described later. The characteristic colors of the simple cyanides of this class are as follows: Iron (ferrous) Reddish yellow. " (ferric) Deep brown ( solution only). Manganese (manganous) White, turning brown. Nickel Apple green. Cobalt Pink. Chromium (chromous) White. (chromic) Grayish blue. CYANIDES OF THE PRECIOUS METALS Cyanides of Gold. Gold forms two simple cyanides: Aurous cyanide (AuCN) and Auric cyanide (Au(CN) 3 ). It also forms two corresponding classes of double cyanides: Aurous (mAu(CN) 2 ) and Auric (mAu(CN) 4 ). 72 THE CYANIDE HANDBOOK Aurous cyanide (AuCN) : Preparation. 1. By decomposing auro-potassic cyanide with acids: KAuCy 2 + HC1 = AuCy + HCy + KC1. The mixture remains clear at ordinary temperatures, but becomes turbid and deposits AuCy at 50 C. 2. By evaporating to dryness a mixture of HgCy 2 and AuCl 3 . Only part of the Au is obtained as AuCy, the remainder forming a double cyanide of Au and Hg. Properties. A citron-yellow crystalline powder, inodorous, tasteless, unalterable in air; not affected by light if dry, but in presence of moisture takes a greenish tint. Insoluble in water. Reactions. 1. Decomposed by heat into Au and CN. 2. Burns, leaving residue of metallic gold. 3. Most acids do not decompose it, even on boiling. When recently precipitated it is soluble in some acids without change. H 2 SO 4 and HN0 3 + HC1 dissolve it slowly on boiling. H 2 S does riot decompose it. 4. Ammonia and ammonium sulphide dissolve it; sulphide of gold is precipitated from the latter solution on adding acid. 5. KOH, on boiling, gives Au and KAuCy 2 . 6. Dissolves in alkaline cyanides: KCy + AuCy = KAuCy 2 . 7. Dissolves in sodium thiosulphate. Auric Cyanide (Au(CN) 3 ): Preparation. 1. By dissolving KAuCy 2 in an acid and evaporating. 2. When potassium auricyanide (KAuCy 4 ) is treated with excess of AgNO 3 , a precipitate is formed, consisting of AgAuCy 4 . This, on treatment with HC1, decomposes as follows: AgAuCy 4 + HC1 = AgCl + HCy + AuCy 3 . The filtrate from AgCl is evaporated in vacuo. Properties. Small colorless crystals, melting at 50 C. Easily soluble in water, alcohol, or ether. Reactions. 1. Decomposes on gentle heating, first giving the aurous cyanide AuCy, then Au and CN. 2. Mercurous nitrate forms AuCy and HgCy 2 . 3. Oxalic acid does not precipitate Au from a solution of AuCy 3 in water. Double Aurous Cyanides. The following are known : SIMPLE METALLIC CYANIDES 73 Auropotassic cyanide (KAuCy 2 ). Aurosodic cyanide (NaAuCy 2 ). Auroammonic cyanide (NH 4 AuCy 2 ), Aurobarytic cyanide (Ba(AuCy 2 ) 2 ). Aurozincic cyanide (Zn(AuCy 2 ) 2 ). Auroargentic cyanide (AgAuCy 2 ). Aurppotassic Cyanide (KAu(CN) 2 ) : Preparation. - 1. By dis- solving metallic gold in potassium cyanide, in presence of air or oxygen: 4 KCy + 2 Au + O + H 2 O = 2KOH + 2KAuCy 2 . [This, and the analogous reaction with NaCy, are probably the chief means by which gold is dissolved in cyanide treatment]. 2. By dissolving AuCy in KCy (see above). 3. By dissolving oxide of gold or fulminating gold (AuCNO) 2 in KCy. Properties. Crystallizes in colorless plates composed of orthorhombic prisms with octahedral faces. Dissolves in 4 parts of cold water and 0.8 parts of boiling water. Very slightly soluble in alcohol. Unalterable in air. Reactions. 1. Decomposed by heat into Au, Cy, and KCy. 2. Slowly decomposed by acids, giving off HCy and deposit- ing AuCy. 3. Not decomposed by H 2 S. 4. Decomposed by mercuric chloride, forming a chlorocyanide of mercury and potassium: HgCl 2 + KAuCy 2 = KC1 HgCy + AuCy + Cl. 5. Forms addition-products with the haloid elements, having the general formula KCyAuCyX 2 , nH 2 O where X = Cl , Br or I. . Iodine decomposes it into AuCy, Cy, and KI. Aurosodic Cyanide (NaAu(CN) 2 ). Its preparation and reac- tions are analogous to those of KAuCy 2 . Also prepared by double decomposition, thus: Na2SO 4 + Ba(AuCy 2 ) 2 = BaSO 4 + 2NaAuCy 2 . Less soluble in water than KAuCy 2 . Very slightly soluble in alcohol. Crystallizes in anhydrous scales which decompose at about 200. Also forms addition compounds with Cl, Br, and I. Auripotassic Cyanide (KAu(CN) 4 ) : Preparation. 1. By act- 74 THE CYANIDE HANDBOOK ing on a mixture of Au and AuCl 3 with concentrated solution of KCy, and crystallizing. Properties. Colorless crystals. Containing H 2 O, which is only completely given off at 200 C. Reactions. 1. Melts to a brown liquid, evolving Cy and leaving KAuCy 2 . 2. Slowly decomposed by acids. 3. Chlorine decomposes it on heating, giving CNC1. Iodine gives KAuCy 2 I 2 . 4. Mercuric chloride precipitates AuCy 3 . HgCl 2 + KAuCy 4 = AuCy 3 + KC1 + HgClCy 2 . Analogous compounds: Auriammonic cyanide (NH 4 Au(CN) 4 ) and auriargentic cyanide (AgAu(CN) 4 ). Cyanides of Silver. Silver forms a cyanide (AgCN) and double cyanides of the type mAg(CN) 2 . Cyanide of Silver (AgCN): Preparation. 1. By precipitat- ing a silver salt with HCN or a metallic cyanide in suitable propor- tions. The best method of obtaining it pure is by precipitating the double cyanide of silver with silver nitrate: KAgCy 2 + AgNO 3 = KNO 3 + 2AgCy. (It is thus obtained free from carbonates, cyanates, chlorides and ferrocyanides, which might be present in commercial KCy or NaCy). 2. Many cyanides, when treated, especially by the aid of heat, with AgNO 3 , are decomposed, yielding AgCN. Thus, ZnCy 2 + 2AgNO 3 = Zn(NO 3 ) 2 + 2AgCy. 3. By adding an acid to a solution of the double cyanide: KAgCy 2 + HN0 3 = KNO 3 + HCy + AgCy. Properties. White, curdy substance resembling AgCl. In- soluble in water and in cold, moderately concentrated acids. Turns brown on exposure to light. (According to Fresenius, it is not affected by light.) Reactions, 1. When heated without access of air, it first melts, then decomposes, swelling considerably, giving up CN and leaving a compound, Ag 6 (CN) 3 , which, on further heating, yields N and carbide of silver (Liebig). 2. Heated in contact with air or oxygen, it leaves only metallic silver. SIMPLE METALLIC CYANIDES 75 3. Chlorine forms AgCl and CN. When the cyanide is com- pletely decomposed, the CN unites with excess of Cl, forming CNC1. 4. Sulphur heated with AgCN gives AgSCN. 5. Hot acids decompose it, giving HCN. It is completely dissolved and decomposed by hot 50 per cent. H 2 SO 4 . [AgCl is not decomposed under these conditions]. 6. Heated with water in a sealed tube to 280 C., it is decom- posed into Ag and (NH 4 ) 2 CO 3 . 7. Boiling solution of KC1 decomposes it: AgCy + KC1 = AgCl + KCy. 8. Soluble in ammonia, ammoniacal salts, ferrocyanides, and thiosulphates. 9. Soluble in alkaline cyanides, forming double salts: AgCy + KCy = KAg(Cy) 2 . 10. Decomposed by a solution of sulphur in CS 2 , forming com- pounds which appear to be sulphides of cyanogen. Argento Potassic Cyanide (KAg(CN) 2 ) : Preparation. 1. By dissolving metallic silver in a solution of potassium cyanide in presence of air or oxygen: 4KCy + 2Ag + O + H 2 O = 2KOH + 2KAgCy 2 . 2. By dissolving certain insoluble salts of silver (e.g., AgCN, AgCl, Agl, Ag 4 FeCy 6 , Ag 3 FeCy c , and AgCNO) in an excess of KCN. Properties. Crystallizes in colorless, regular octahedra, with faces often depressed to the center. Inodorous, with metallic taste. Soluble in 8 parts of cold water, in 4 parts water at 20 C., and in 1 part boiling water. Also soluble in alcohol. Reactions. 1. Decomposed by acids with formation of AgCN (see above). 2. Decomposed by H 2 S or alkaline sulphides, with precipita- tion of Ag 2 S: 2KAgCy 2 + H2S = A g2 S + 2KCy + 2HCy. 2KAgCy 2 + K 2 S = A g2 S + 4KCy. 3. Many metallic solutions precipitate insoluble double cyan- ides of silver; e.g., 2KAgCy 2 + ZnSO 4 = K,SO 4 + Zn(AgCy 2 ) 2 . 4. Certain metals, such as zinc, decompose the double cyanide, 76 THE CYANIDE HANDBOOK with precipitation of silver, probably as the result of secondary reactions : KAgCy 2 + Zn + H 2 O = KOH + H + ZnCy 2 + Ag. Other Double Cyanides of Silver. Among others, the follow- ing are known: Argentocyanic acid (HAgCy 2 ). Sodic argentocyanide (NaAgCy 2 ). Barium argentocyanide (Ba(AgCy 2 ) 2 H 2 O). Sodio-potassic argentocyanide (NaK 3 (AgCy 2 ) 4 ) = NaAgCy 2 + 3KAgCy 2 . Zinc argentocyanide (Zn(AgCy 2 ) 2 ). Similar compounds are formed with Cd, Ni, Co, Mn, Fe, Cu, Pb, and Hg. Ammonia forms the compound NH 3 AgCy. The compound AgNO 3 2AgCy is formed by dissolving AgCy in a concentrated solution of AgNO 3 . Cyanides of Other Precious Metals. Cyanides and various classes of double cyanides are formed with the platinum metals and other less common elements, which need not be considered here. CYANIDES OF COPPER Three simple cyanides have been described, viz., cuprous cyanide (Cu 2 (CN) 2 ), cupric cyanide (Cu(CN) 2 ), cuproso-cupric cyanide (Cu 3 (CN) 4 ). There are also numerous double cyanides of various types. Cuprous Cyanide (Cu 2 (CN) 2 ) : Preparation. 1. By the action of HCN on recently precipitated cuprous hydrate, on cuprous chloride dissolved in HC1, or on cupric chloride in presence of S0 2 . 2. By decomposition of cupric cyanide in presence of water. 3. By precipitation of cupropotassic cyanide with acids: 2KCuCy 2 + H 2 SO 4 = K 2 SO 4 + Cu 2 Cy 2 + 2HCy. Properties. A white, curdy precipitate, insoluble in water and dilute acids. This is the most stable of the cyanides of copper. Reactions. 1. On heating it melts, giving off much water and leaving a swollen brown mass. 2. Decomposed by nitric acid with evolution of N 2 O 2 . SIMPLE METALLIC CYANIDES 77 3. Dissolves without change in concentrated HC1, and is pre- cipitated from this solution by H 2 O or KOH. 4. Soluble in ammonia and in many ammonium salts, forming complex cyanides. 5. Soluble in alkaline cyanides: Cu 2 Cy 2 + 2KCy = 2KCuCy 2 . With excess of KCy, other compounds are formed (see below). 6. Decomposed by ferric chloride in the cold, with evolution of cyanogen: Fe 2 Cl 6 + Cu 2 Cy 2 = 2FeCl 2 + Cu\ 2 + Cy 2 . 7. Decomposed by acetic acid in presence of an oxidizing agent, giving cyanogen. Cupric Cyanide (CuCy 2 ): Preparation. 1. By treating car- bonate of copper with HCy in solution : CuCO 3 + 2HCy = CuCy 2 + H 2 O + CO 2 . 2. By acting on acetate of copper with HCy. 3. By adding KCy to an excess of a cupric salt: 2KCy + CuS0 4 = CuCy 2 + K^CX. Properties. A yellowish-brown, unstable substance, only known in hydrated condition. Reactions. 1. Changes rapidly and spontaneously at the ordinary temperature, forming greenish or yellowish-gray cuproso- cupric cyanide (?Cu 3 Cy 4 H 2 O), which, on heating, gives Cu 2 Cy 2 + Cy. Cuproso-Cupric Cyanides. There appear to be two of these, viz - : Cu 2 Cy 2 CuCy 2 = Cu 3 Cy 4 . 2Cu 2 Cy 2 CuCy 2 = Cu 5 Cy 6 . These are, perhaps, only mixtures of Cu 2 Cy 2 and CuCy 2 . Preparation. 1. By treating cuprous oxide with HCy. 2. By adding HCy to CuSO 4 . 3. By spontaneous decomposition of cupric cyanide: 3CuCy 2 = Cu 3 Cy 4 + Cy 2 . Properties. According to mode of formation, it is yellowish- gray, or crystallizes in transparent prisms of a brilliant canary- green color. Reactions. 1. Heated to 100 it loses Cy and water, and is converted into Cu 2 Cy 2 without change of form. 78 THE CYANIDE HANDBOOK 2. Strong acids liberate HCy, leaving a mixture of cuprous and cupric salts. 3. Alkalis form double salts of copper and the alkali metal, probably 3Cu 2 Cy 2 4mCy, and precipitating Cu(OH) 2 . 4. Alkaline cyanides give (when in excess) the compound Cu 2 Cy 2 6mCy. (m = K, Na, etc.) 5. Ammonia dissolves it to a blue solution. The compound is only slightly soluble, and sometimes deposits green crystals (see below). 6. Ammonium carbonate dissolves it in the cold, other ammo- nium salts only on heating. Double Cyanides of Copper. These are very numerous, the best known being those of ammonium and potassium. The ammo- nio-cyanides of copper are of two distinct kinds, represented re- spectively by the formulae. 1 (a) x(NH 4 Cy)y(CuCy 2 )z(Cu 2 Cy 2 ); (6) x(NH 3 )y(CuCy 2 )z(Cu 2 Cy 2 ). The potassium compounds have the general formula, x(KCy) y(Cu 2 Cy 2 ). Double Cyanides of Copper and Ammonium. The compounds of class (a), which are true double cyanides of copper and ammo- nium, are colorless, highly soluble, and stable in absence of other metallic salts. They act feebly as solvents of gold. They are decomposed by excess of a copper salt, yielding a mixture of cuprous and cupric cyanides. The compounds of the second class (6), having the formula x(NH 3 )y(CuCy 2 )z(Cu 2 Cy 2 ), are formed by dissolving mCuCy 2 -f nCu 2 Cy 2 in ammonia. They are generally regarded as cupram- monium cyanides, i.e., as compounds in which the ammonia is intimately united with copper, forming a complex radical, to which Rose 2 gives the formula Cu(NH 3 ) 4 (cuprammonium), and which enters into unstable combinations with cyanogen. These substances are generally blue or green crystalline bodies, only slightly soluble in water. They yield deep blue solutions and appear to be "ionized" in dilute solutions, with liberation of + + CN and the cuprammonium ion Cu(NH 3 ) 4 , to which the blue color of the liquid is due. i Sulman, Trans. I. M. M., XIV, 363. a Trans. I. M. M., XIV, 369. SIMPLE METALLIC CYANIDES 79 Preparation and Properties. 1. When a salt of copper is precipitated by ammonium cyanide, cyanogen is abundantly evolved and a greenish-blue substance formed having the formula 2NH 3 CuCy 2 Cu 2 Cy 2 H 2 O. This is an amorphous powder, slightly soluble in cold water and decomposed at 100 C. 2. By acting on the above with further quantities of ammonia, a blue liquid is obtained which, on evaporation, yields 4NH 3 . CuCy 2 . 2Cu 2 Cy 2 , in small, bright-green, needle-like crystals, pris- matic, with metallic luster. The same substance is obtained by passing a current of HCN into oxide of copper suspended in ammonia. 3. When the preceding compound is heated with ammonia and a current of NH 3 gas is passed through the liquid, the latter deposits, on cooling, the substance 6NH 3 CuCy 2 Cu 2 Cy 2 . This is a crystalline precipitate, in prismatic needles of a fine blue color. Exposed to the air, it loses NH 3 and becomes green. Reactions. Solutions of these compounds readily give up cyanogen to any substance capable of combining with it. Hence they act as powerful solvents of gold. This reaction is accom- panied by deposition of the slightly soluble dicuproso-cupric cyanide 4NH 3 CuCy 2 Cu 2 Cy 2 , and takes place without the intervention of oxygen. Double Cyanides of Copper and Potassium. The following are known: l (a) 4KCy 3Cu 2 Cy 2 . (6) 2KCy-Cu 2 Cy 2 . (c) 4KCyCu 2 Cy 2 . (d) 6KCyCu 2 Cy 2 . Preparation and Properties. (a) The first of these is ob- tained by the action of fused cuprous cyanide on potash and is little known. (6) This compound is formed by the action of KCy on Cu 2 Cy 2 in the proportions indicated by the formula of the compound. It is also formed by acting on cupric hydrate with KCN solution and slowly evaporating, 2 or by acting on cuproso-cupric cyanide with KCN. It forms transparent, colorless clinorhombic prisms, fusible Sulman, Trans. I. M. M., X, 125. Virgoe, Ibid., X, 107. 80 THE CYANIDE HANDBOOK without decomposition, but liberating metallic copper on con- tinued fusion. The compound 2KCy Cu 2 Cy 2 is sparingly soluble in water, and partially decomposes in solution into Cu 2 Cy 2 and 6KCy Cu 2 Cy 2 . It is not a solvent of gold or silver. (c) This compound, 4KCy Cu 2 Cy 2 , is the form in which copper, according to Virgoe (loc. cit., p. 141), normally exists in cyanide solutions, though Sulman (loc. cit., p. 127) considers that it is more frequently present as sulphocyanide (Cu 2 (SCy) 2 ) dissolved in KCy. It is prepared by agitating cupric hydrate with excess of KCy: i 2Cu(OH) 2 + 8KCy = 4KCy Cu 2 Cy 2 + 4KOH + Cy 2 . Virgoe considers that an intermediate product exists, to which he gives the formula 2KCy Cu 2 Cy 2 4KCy Cu 2 Cy 2 , which might be written 3KCy Cu 2 Cy 2 . This latter is said to be a good solvent of gold and silver (loc. cit., p. 142). (d) 6KCy Cu 2 Cy 2 . This is obtained by dissolving metallic copper in KCN: Cu 2 + 8KCy + H 2 O + O = 6KCy Cu 2 Cy 2 + 2KOH. Colorless transparent prisms, with 6 sides, terminated by 6-sided pyramids, almost unalterable in air, but gradually becoming blue. May be fused with separation of a little copper. [The salts de- scribed by Virgoe as existing in solution may possibly be only mix- tures of 2KCy Cu 2 Cy 2 and 6KCy Cu 2 CyJ. Reactions. 1. The double cyanides of K and Cu are all decomposed by acids, forming Cu 2 Cy 2 and HCy. 2. H 2 S does not precipitate copper sulphide from their solu- tions, or only to a very slight extent from dilute solutions. Alka- line sulphides give no precipitate. 3. Ferric salts precipitate Cu 2 Cy 2 , Fe(OH) 6 and liberate HCy. 4. Mercuric salts give Cu 2 Cy 2 , HgCy 2 , and the potassium salt of the radical originally combined with the mercury. 5. Excess of copper salts give a pale yellow precipitate which may be a cuproso-cupric cyanide xCuCy 2 yCu 2 Cy 2 . 6. Silver nitrate converts a part of the KCy in these compounds 1 The reaction (in presence of an excess of alkali) which, according to the writer's experiments, seems most in accordance with the facts, is as follows: 2Cu(OH) 2 + 7KCy = 4KCy Cu 2 Cy 2 + 2KOH + KCyO + H 2 O This indicates a consumption of 3.6 parts KCy for every part of Cu dissolved. SIMPLE METALLIC CYANIDES 81 into the soluble KAgCy 2 before any precipitation of AgCy takes place. The usual reaction is (Virgoe, loc. cit., p 139): AgN0 3 + 4KCy Cu 2 Cy 2 = 2KCy -Cu 2 Cy 2 + KAgCy. + KNO 3 . On further addition of AgNO 3 , there is a precipitation of Cu 2 Cy 2 as follows: AgNO 3 + 2KCy -Cu 2 Cy 2 = Cu 2 Cy 2 + KAgCy 2 + KNO 3 . When KI is added, however, there is an immediate precipitation of Agl, which is not soluble in the cupropotassic cyanides. Other Double Cyanides of Copper. Other compounds are de- scribed with Na, Ba, Zn, Mn, Co, Ni, Fe, Pb, and other metals. CYANIDES OF MERCURY Only one well-defined simple cyanide exists, namely, mercuric cyanide (Hg(CN) 2 ). There are two basic cyanides (oxycy anides) , namely, HgO- HgCy 2 , and HgO 3HgCy 2 , besides numerous double cyanides with the alkali metals and ammonia. Many chlorocyanides are also known, of the general formula HgCy 2 RC1 + nH 2 O. Mercuric Cyanide (Hg(CN) 2 ) : Preparation. 1. By action of HCy on mercuric oxide suspended in water: 2HCy + HgO = HgCy 2 + H 2 O. 2. By double decomposition of mercuric salt and hydrocyanic acid: Hg(NO 3 ) 2 + 2HCy = HgCy 2 + 2HNO 3 . [With mercurous salts a mixture of HgCy 2 and metallic mer- cury is obtained]. 3. By the action of mercuric oxide on cyanides or ferrocyan- ides, or (with boiling) on Prussian blue. 4. By heating mercuric sulphate with ferrocyanide. 5. By acting on the ferrocyanides of lead or barium with mercuric sulphate, and filtering from the insoluble PbSO 4 or BaSO 4 . Properties. Crystallizes in colorless, rectangular prisms with square base, but occurs in other forms. Anhydrous. Soluble in 8 parts cold water; more soluble in hot water. Insoluble in abso- lute alcohol. Reactions. 1. On heating, gives Hg and (CN) 2 (see above). 82 THE CYANIDE HANDBOOK 2. Decomposed by concentrated acids, giving HCN and the corresponding mercury salt. 3. Chlorine in the presence of moisture or on exposure to light gives HgCl 2 and CNC1. Bromine and iodine act similarly. 4. Water at 280 C. forms Hg and (NH 4 ) 2 CO 3 . 5. Decomposed by HC1, HBr, HI, and H 2 S, giving the corre- sponding Hg salt and HCy. Gradually decomposed by alkaline sulphides, forming HgS and a thiocyanate of the alkali. Basic Cyanides of Mercury. These are white or yellowish crystalline explosive bodies, formed by acting on HgCy 2 with HgO in various proportions. Double Cyanide of Mercury and Potassium (K 2 Hg(Cy) 4 ): Preparation. 1. By dissolving HgCy 2 in KCy and evaporating till the salt crystallizes: HgCy 2 + 2KCy = K 2 HgCy 4 . 2. By treating HgCy 2 with HCy and K 2 CO 3 : HgCy 2 + 2HCy + K 2 CO 3 - K 2 HgCy 4 + H 2 O + CO 2 . 3. By dissolving HgO in KCy: HgO + 4KCy + H 2 O = K 2 HgCy 4 + 2KOH. 4. By dissolving HgCl 2 in KCy: HgCl 2 + 4KCy = K 2 HgCy 4 + 2KC1. Properties. Crystallizes in regular octahedra. Transparent. Unalterable in air. Dissolves in 4 parts cold water. Solution smells slightly of HCy. Reactions. 1 . Decrepitates violently on heating, and melts to a brown liquid which gives off Cy 2 and Hg. 2. Decomposed by acids, either completely or with formation of HgCy 2 and HCy. 4HC1 + K 2 HgCy 4 = 2KC1 + HgCl 2 + 4HCy. 2HN0 3 + K 2 HgCy 4 - 2KNO 3 + HgCy 2 + 2HCy. 3. Decomposed by alkaline sulphides or by H 2 S with precipi- tation of HgS : KjS + K2HgCy4 = Hgg + 4RCy 4. Acts as a solvent of gold and silver, with deposition of mercury and without intervention of oxygen: K 2 HgCy 4 + 2Au = Hg + 2KAuCy 2 . K 2 HgCy 4 + 2Ag = Hg + 2KAgCy 2 . SIMPLE METALLIC CYANIDES 83 5. Decomposed by other metals (zinc, for example) with deposition of mercury: K 2 HgCy 4 + Zn = K 2 ZnCy 4 + Hg. [A similar double cyanide is formed with sodium, namely, Na 2 HgCy 4 . Double cyanides of mercury with Zn and Pb are also known. Many double salts are formed by combination of HgCy 2 with other metallic salts, e.g., nitrates, thiosulphates, thiocyanates, ferrocyanides, etc. CYANIDES OF LEAD No simple cyanide of lead of the formula Pb(CN) 2 is known. The precipitate obtained by acting on lead salts with an alkaline cyanide appears to have the composition 2PbO Pb(CN) 2 H 2 O. It is best prepared by acting on neutral lead acetate with KCN in presence of ammonia. It is a white precipitate, insoluble in water, but decomposed by dilute acids with evolution of HCy. Double Cyanide of Zinc and Lead. When a salt of lead is precipitated by K 2 ZnCy 4 , a white precipitate is obtained having the composition 2ZnCy 2 PbCy 2 . SECTION III METALLIC COMPOUNDS OF COMPLEX CYANOGEN RADICALS (A) METALLIC COMPOUNDS OF THE RADICAL RC 6 N 6 THE compounds we now have to consider cannot legitimately be described as cyanides. Although under certain conditions they may be obtained from, or may yield, cyanogen or cyanides of the type previously discussed, they differ widely from the cyanides in their general properties and reactions, and behave as though containing a negative radical into the composition of which a metal (usually of the iron type) enters as a constituent part. This radical goes in and out of combination unchanged, and may be replaced by different non-metallic elements or rad- icals. The general formula for this radical is RC 6 N 6 , where R is an atom of a metal such as Fe, Mn, Cr, Co, or of the platinum group, such as Pt, Ir, Pd, Os, Rh, Ru. A closely related class of substances, known as the nitroprus- sides, contain the radical RC 5 N 6 O, probably RC 5 N 5 (NO). The radical RC 6 N 6 is capable of acting as a dyad, triad, or tetrad, forming compounds of the types M 2 RC 6 N 6 where M represents an M 3 RC 6 N 6 atom of a monad element or M 4 RC 6 N 6 group. Compounds of the first type are very unstable; those of the second may perhaps have the constitution M 3 EE(RC 6 N 6 ) I M 3 =(RC N 6 ) on the assumption that the group RC N 6 is normally tetrad. 84 COMPLEX METALLIC CYANIDES 85 COMPOUNDS IN WHICH IRON ENTERS INTO THE NEGATIVE RADICAL These compounds are as follows: (a) Ferrocyanides, M 4 FeC 6 N 6 . (b) Ferricyanides, M 3 FeC 6 N 6 . (c) Perferricyanides, M 2 FeC 6 N 6 . (?) (d) Nitroprussides, M 2 Fe(CN) 5 (NO). For simplicity we may use the symbol FeCy 6 to indicate the radical FeC 6 N 6 in the first three of these groups of compounds. (a) FERROCYANIDES (M 4 FeCy 6 ). General Characteristics. 1 . Most f errocyanides are colored ; the characteristic colors of these compounds are of value in analysis. . 2. The soluble alkali ferrocyanides are non-poisonous. 3. They do not evolve HCN when treated with very dilute acids, but generally yield hydroferrocyanic acid (E^FeCyg) . 4. Those ferrocyanides which are completely dehydrated by heat without decomposition yield on further heating N and car- bide of iron. In addition to these there may be formed a cyanide or carbide of the positive metal, or the metal may be liberated with evolution of (CN) 2 . 5. Those ferrocyanides which cannot be completely dehydrated without decomposition yield HCN, NH 3 , CO 2 and a mixture or compound of each of the metals with carbon. 6. When aqueous solutions of alkali ferrocyanides are electro- lyzed, the alkali separates at the pole, and HCN and Prussian blue at the + pole; in some cases a cyanide of the metal forming the electrode is produced. 7. When heated with concentrated H 2 SO 4 , sulphates of Fe, NH 4 , and the metal forming the + radical are produced, and SO 2 , CO, CO 2 , and N are evolved. 8. H 2 S has no action on alkaline ferrocyanides. Some others are decomposed, giving a sulphide of the + radical, and H 4 FeCy 6 . 9. Ferrocyanides of heavy metals are generally decomposed by aqueous alkalis, giving a ferrocyanide of the alkali and hydrate of the heavy metal. 10. Decomposed by boiling with HgO and water, forming HgCy 2 and oxides or hydrates of the metals. 86 THE CYANIDE HANDBOOK 11. Alkaline ferrocyanides are converted into the correspond- ing ferricyanides by oxidizing agents. Potassium Ferrocyanide (K 4 FeCy 6 3H 2 0) : Preparation. 1. By boiling Prussian blue with potash: Fe4(FeCy 6 )3 + 12KOH = 3K 4 FeCy 6 + 4Fe(OH) 3 . 2. By fusing nitrogenous animal matter with K 2 CO 3 and scrap iron, lixiviating with water, filtering, and crystallizing. KCy is first formed, which reacts on the iron compounds during the extrac- tion with water, yielding K 4 FeCy 6 . [This process will be dis- cussed more fully under the heading "Manufacture of Cyanide."] 3. By heating ammonium thiocyanate (NH 4 SCN) with scrap iron to dull redness and extracting with water. 4. By action of KCN on ferrous compounds: 2KCN + FeSO 4 = K 2 SO 4 + Fe(CN) 2 . 4KCN + Fe(CN) 2 = K 4 Fe(CN) 6 . Similarly with FeS, FeCO 3 , Fe(OH) 2 , etc. 5. By acting upon various metallic ferrocyanides with KOH. Properties. Crystallizes with 3 molecules H 2 O as yellow quadratic pyramids. Non-poisonous. Loses water at 60 to 80 C. Unchanged at ordinary temperatures. Soluble in 4 parts cold and 2 parts boiling water. Insoluble in alcohol. Reactions. 1. Heated in a closed vessel, it melts a little above red heat, evolves N, and leaves a mixture of Fe carbide and KCy. When the hydrated salt is used, C0 2 , NH 3 , HCN, and N are evolved. 2. Aqueous solution is gradually decomposed on exposure to light, forming Prussian blue and evolving HCN. 3. When heated to redness in air, or with reducible metallic oxides, it gives KCyO. Ozone changes it slowly into K 3 FeCy 6 . Oxidizers such as KMnO 4 , MnO 2 , PbO 2 , etc., also form K 3 FeCy 6 : 2K 4 FeCy 6 + RO 2 = RO + K 2 O + 2K3FeCy 6 . Chlorine and bromine act as follows: 2K 4 FeCy 6 + C1 2 = 2KC1 + 2K 3 FeCy 6 . Iodine gives a double compound of KI and K 3 FeCy 6 . 4. 1 On electrolysis it gives K 3 FeCy 6 at the + pole and KOH and H at the pole (Joannis). 1 According to Morley and Muir (Watts, " Diet, of Chemistry," II, 333), when aqueous solutions of alkali ferrocyanides are electrolyzed, alkali separates at the - pole and HCy and Prussian blue at the + pole. COMPLEX METALLIC CYANIDES 87 5. Dilute cold sulphuric acid gives H 4 FeCy 6 : K 4 FeCy 6 + 2H2SO 4 = H 4 FeCy 6 + 2KO<. Dilute warm sulphuric acid gives 2K 4 FeCy 6 + 3H 2 SO 4 = 3K 2 SO 4 + K 2 Fe FeCy 6 + 6HCy. Concentrated sulphuric acid gives K 4 FeCy 6 + 6H 2 SO 4 + 6H 2 O = 2K 2 SO 4 + 3(NH 4 )2SO4 + FeSO 4 4- 6CO. 6. Fairly concentrated HNO 3 (say 50 per cent. HNO 3 by volume) forms nitroprussic acid (H 2 FeCy 5 NO H 2 O). Very concentrated HNO 3 decomposes K 4 FeCy 6 into N, CN, NO, CO 2 , KNO 3 and Fe 2 O 3 . 7. Boiled with mercuric oxide: K 4 FeCy 6 + 3HgO + 3H 2 O = 3HgCy 2 + 4KOH + Fe(OH) 2 . 8. Boiled with NH 4 C1, it gives NH 4 Cy, which volatilizes: K 4 FeCy + 6NH 4 C1 = 6NH 4 Cy + 4KC1 + FeCl 2 . 9. Heated with ammoniacal silver nitrate, it forms Fe(OH) 3 and KAgCy 2 . 10. Most metallic salts give precipitates of insoluble ferro- cyanides. CHARACTERISTIC INSOLUBLE FERROCYANIDES The more important of these are as follows: Barium BaFeCy 6 - 6H 2 O White, slightly sol- uble. Cobalt Co 2 FeCy 6 7H 2 O Blue, changing to red. Copper Cu 2 FeCy 6 xH 2 O (cupric) Reddish brown. Iron Fe 2 (FeCy 6 ) (ferrous) White. Fe 4 (FeCy 6 ) 3 (ferric) Blue. Lead Pb 2 FeCy 6 - 3H 2 O White. Manganese Mn 2 FeCy 6 7H 2 O White. Mercury composition doubtful White. Nickel Ni 2 FeCy 6 - (7 to 1 1) H 2 O Greenish white. Tin Sn 2 FeCy 6 4H 2 O (stannous) White. " SnFeCy 6 4H 2 O (stannic) Brownish. Zinc Zn 2 FeCy e 3H 2 O White. Silver Ag 2 FeCy 6 White. 88 THE CYANIDE HANDBOOK The more important soluble ferrocyanides are those of NH 4 , Ca, H, Mg, Na, K. There are also numerous compounds in which more than one metal forms part of the positive radical, for example, K 2 ZnFeCy 8 . Hydrogen Ferrocyanide : Hydroferrocyanic Acid (H 4 FeCy 6 ). This is the acid corre- sponding to the metallic ferrocyanides. It may be prepared: 1. By decomposing Ba 2 FeCy 6 with an equivalent of H 2 S0 4 , and filtering from BaSO 4 . 2. By the action of cold concentrated HC1 on an air-free solu- tion of K 4 FeCy 6 , and washing the precipitate with alcohol and ether. Properties. A white, crystalline powder, becoming blue in moist air with evolution of HCy. Soluble in water. Strongly acid to litmus. Reactions. Solution decomposes carbonates, acetates, oxal- ates, and tartrates. When boiled with water it decomposes: 2H 4 FeCy 6 = FeH 2 (FeCy 6 ) + 6HCy. Ferrocyanides of Copper: Cuprous Ferrocyanide (Cu 4 FeCy 6 ) is said to be formed by add- ing K 4 FeCy 6 to Cu 2 Cl 2 dissolved in HC1. Cupric Ferrocyanide (Cu 2 (FeCy 6 ) xH 2 0) is formed: 1. By adding K 4 FeCy 6 to a cupric salt. By this means the precipitate formed is mixed with double ferrocyanidesCuK 2 (FeCy 6 ), and Cu 3 K 2 (FeCy 6 ) 2 . 2. By adding H 4 FeCy 6 to a cupric salt. A reddish-brown pre- cipitate, insoluble in dilute acids; soluble in ammonia and alkalis. Ferrocyanides of Iron: Ferrous Ferrocyanide Fe 2 (FeCy 6 ). Obtained by adding fer- rous salts to H 4 FeCy 6 , both compounds being freed from air and ferric compounds. A white amorphous precipitate, soon oxidiz- ing, with formation of blue compounds: 3Fe 2 (FeCy 6 ) + 3O + 3H 2 O = Fe 2 (OH) 6 + Fe 4 (FeCy 6 ) 3 . Potassium Ferrous Ferrocyanide (K 2 Fe FeCy 6 ) (Everitt's salt). Obtained by decomposing K 4 FeCy 6 with hot dilute sul- phuric acid. (See under " Potassium Ferrocyanide. ") White, microscopic quadratic crystals, becoming blue in air. Converted by oxidizers into potassium-ferrous ferricyanide (FeK-FeCy 6 ). COMPLEX METALLIC CYANIDES 89 Ferric Ferrocyanide (Prussian Blue: Fe 4 (FeCy 6 ) 3 ) : Prepara- tion. 1. By acting on a ferric salt with potassium ferrocyanide, keeping the ferric salt in excess: 3K 4 FeCy 6 + 4FeCl 3 = 12KC1 + Fe4(FeCy 6 ) 3 . 2. By acting on a soluble cyanide with a ferrous and ferric salt, adding excess of KOH or NaOH, then excess of acid: (a) 6KCy + FeSO 4 = K 2 SO 4 + K 4 FeCy 6 . (6) 3K 4 FeCy 6 + 4FeCl 3 = 12KCI + Fe 4 (FeCy 6 ) 3 . 3. Prussian blue is obtained commercially by mixing K 4 FeCy 6 with partially oxidized ferrous sulphate, and heating the light- blue precipitate so obtained with various oxidizing agents. In addition to ferric ferrocyanide (Fe 4 (FeCy 6 ) 3 ) it usually contains also Ferrous ferrocyanide (Fe 2 FeCy 6 ) , Ferrous ferricyanide (Fe 3 (FeCy 6 ) 2 ,) and sometimes Everitt's salt (K 2 Fe FeCy 6 ) . Properties. A dark-blue amorphous solid with luster re- sembling that of copper. Retains water up to 250 C., at which temperature it begins to decompose. Insoluble in water, alco- hol, ether, and dilute acids. Reactions. 1. When strongly heated it ignites, leaving Fe 2 3 . Heated in a partially closed vessel, it first gives off NH 4 Cy and (NH 4 ) 2 CO 3 , and finally HCN, CO, and CO 2 . 2. Alkalis (including MgO) decompose it, yielding ferric hydrate and a ferrocyanide of the alkali. Ammonia gives Fe(OH) 3 and (NH 4 ) 4 FeCy 6 . Alkaline carbonates give similar reactions. 3. Boiled with HgO and water, it gives Fe 2 O 3 and HgCy 2 . 4. Boiled with lead oxide, it gives lead ferrocyanide, a ferri- cyanide, and Fe 2 3 . 5. Reduced by H 2 S or by Fe or Zn, it gives white ferrous ferro- cyanide (Fe 2 -FeCy e ). Sodium Ferrocyanide (Na 4 FeCy - 12H 2 0). Obtained by boil- ing Prussian blue with NaOH, filtering, and cooling. Pale yellow monoclinic crystals, which effloresce in air. Reactions similar to those of K 4 FeCy 6 . Zinc Ferrocyanide (Zn 2 FeCy 6 -3H 2 O). Obtained pure by add- ing excess of ZnSO 4 to H 4 FeCy 6 . The precipitates formed by ZnS0 4 , or other zinc -salts, with ferrocyanides contain more or less of double ferrocyanides of zinc and the alkali metal. White precipitate, insoluble in dilute acids, soluble in alkalis 90 THE CYANIDE HANDBOOK and ammonia. Double salts are also formed with ammonia, one of which is Zn 2 FeCy 6 3NH 3 - H 2 O. The white precipitate found in the zinc-boxes used in precipi- tation of gold-cyanide solutions, and formed when insufficient alkali is present, appears to be a mixture of ZnCy 2 , Zn(OH) 2 and double ferrocyanides, to which the formulae K 2 Zn(FeCy 6 ) and K 2 Zn 3 (FeCy e ) 2 have been assigned. 1 (See Section V: " Chemistry of Precipitation. ") (b) FERRICYANIDES (M 3 FeCy 6 ) These substances are formed generally by the action of oxidiz- ing agent on ferrocyanides: 2M 4 FeCy 6 + RO = M 2 O + R + 2M 3 FeCy 6 ; where RO represents any substance containing " available" oxygen. The alkali ferricyanides are soluble in water: most others are insoluble. Many have characteristic colors. They act as oxidiz- ing agents by reaction the converse of that given above. CHARACTERISTIC FERRICYANIDES Metal Color Solubility, etc. NH 4 Ruby-red Soluble Ca Red Co Red-brown fCu' . Brown-red Insoluble (cuprous} Cu W Fe" Yellowish Blue Insoluble (cupric) Fe'Fe"' Green Insoluble (forrosoferric) H Pb Dark red Mg . Red-brown Snlnhlp Mn . Brown Ni K Red q l n ui a Ae.. Orange-yellow Na Ruby-red Sn White The insoluble ferricyanides are soluble in alkalis and ammonia. Potassium Ferricyanide (K 3 FeCy 6 ) : Preparation. 1. By act- 1 "Proc. Chem., Met. and Min. Soc. of South Africa/' V. 56, 75. COMPLEX METALLIC CYANIDES 91 ing upon K 4 FeCy 6 with Cl, Br, ozonized oxygen, PbO 2 , KMnO 4 , or bleaching powder. In some cases the action requires the presence of an acid, thus: 5K 4 FeCy 6 + 4H 2 SO 4 + KMnO 4 = SKsFeCve + 3K 2 SO 4 + MnSO 4 + 4H 2 O. When chlorine is used, it is passed into cold solution of K 4 FeCy 6 , with constant agitation, till a few drops of the liquid give a brown- red color with FeCl 3 , instead of a blue color. K 3 FeCy 6 is also one of the products of electrolysis of K 4 FeCy 6 . Properties. Large red prismatic monoclinic crystals. More soluble in hot than in cold water. Nearly insoluble in alcohol. Reactions. 1. Heated in a closed vessel, it decrepitates, evolves (CN) 2 and a little N, leaving a residue of KCy, K 4 FeCy 6 , Fe 4 (FeCy 6 ) 3 , C, Fe, and probably paracyanogen (CN)x. 2. Heated in air (CN) 2 is evolved, and Fe 2 O 3 and KCy remain. 3. Reduced to K 4 FeCy 6 by many reagents, such as H 2 S, Na 2 S, Na 2 S 2 O 3 ; also by finely divided metals (Ag, Fe, Zn, etc.), by SO 2 , by hot ferrous salts, phosphites, etc., especially in presence of alkalis. The general type of these reactions may be represented thus: KsFeCye + KOH + R - KJFeCye + R - OH. R being an oxidizable substance. 4. Certain metallic oxides (e.g., PbO, Cr 2 O 3 , MnO, SnO, etc.), when boiled with K 3 FeCy 6 in presence of KOH, pass to a higher state of oxidation and form K 4 FeCy 6 , thus: 2K 3 FeCy 6 + 2KOH -f RO = 2K 4 FeCy 6 -I- RO 2 + H 2 O. 5. Alkaline peroxides are decomposed with evolution of oxygen, thus: KsFeCye + K 2 O 2 + H 2 O = K 4 FeCy b + KOH + O 2 . 6. When alkalis are boiled with concentrated ferricyanide solutions, Fe 2 O 3 is precipitated and ferrocyanide, cyanide, and (CN) 2 produced. Ammonia forms ferrocyanides of K and NH 4 , and evolves N. Mercuric oxide gives HgCy 2 and Fe 2 3 : K 3 FeCy 6 + 3HgO + 3H 2 O = 3HgCy 2 + 3KOH + Fe(OH) 3 . 7. A mixture of KCy and K 3 FeCy 6 acts as a solvent of gold or silver without intervention of oxygen: K 3 FeCy 6 + 2KCy + Au = K 4 FeCy + KAuCy 2 . 8. Decomposed by chlorine or bromine, forming first the haloid 92 THE CYANIDE HANDBOOK cyanogen compounds (CNC1 and CNBr) and HCy; on continued boiling various ferrosoferric ferricy anides are formed. 9. Potassium iodide forms an unstable compound which readily breaks up into I and K 4 FeCy e . 10. Boiling HC1 forms KC1, FeCl 3 and "Turnbull's blue/' Fe 3 (FeCy 6 ) 2 . 11. Evaporation with nitric acid yields potassium nitroprusside (K 2 FeCy 5 NO) and KNO 3 . Hydrogen Ferri cyanide (H 3 FeCy 6 ) : Hydroferricyanic Acid. Prepared by action of H 2 S0 4 on lead ferricyanide, or by adding cold concentrated K 3 FeCy 6 to 2 to 3 times its volume of concentrated HC1 and filtering. Lustrous brownish-green needles. Very soluble in water or alcohol. Decom- poses in air, giving HCy and a blue residue. Ferricyanides of Copper: Cuprous Ferricyanide (Cu 3 (FeCy 6 ) or Cu 6 (FeCy 6 ) 2 ). brown- ish-red precipitate, obtained by adding a solution of Cu 2 Cl 2 in HC1 to K 3 FeCy 6 . Soluble in NH 3 but not in NH 4 salts. Cupric Ferricyanide (Cu 3 (FeCy 6 ) 2 ). Yellowish precipitate, obtained by adding a cupric salt to K 3 FeCy 6 . Probably contains K. Soluble in NH 3 and in NH 4 salts. Ferricyanides of Iron: Ferrous Ferricyanide (Turnbull's Blue: Fe 3 (FeCy 6 ) 2 ). Pre- pared by adding K 3 FeCy 6 or, better, H 3 FeCy 6 to a ferrous salt, and, after digesting for some time, washing the precipitate with hot water. Also prepared by partial oxidation of ferrous ferro- cyanide, Fe 2 FeCy 6 : 2K 3 FeCy f) + 3FeSO 4 = Fe 3 (FeCy 6 ) 2 + 3K 2 SO 4 . 2Fe 2 FeCy 6 + O - FeO + Fe 3 (FeCy 6 ) 2 . A deep blue powder, with a tinge of copper-red. Insoluble in water, alcohol, and dilute acids. Soluble in oxalic acid. Cannot be dehydrated without decomposition. Exposed to moist air, it yields Prussian blue and ferric oxide : 6Fe 3 (FeCy 6 ) 2 + O 3 = 4Fe 4 (FeCy 6 ) 3 + Fe 2 O 3 . Decomposed by alkalis, alkaline carbonates, and ammonia, giving a ferricyanide and Fe 3 O 4 . Ferrosoferric Ferricyanide (Prussian Green ; Pelouze's Green : Fe 7 (FeCy e ) 6 ). -- A green precipitate, obtained by passing Cl into K 8 FeCy, and boiling, then washing the residue with con- COMPLEX METALLIC CYANIDES 93 centrated boiling HC1. Changes to Prussian blue on contact with air. A black ferrosoferric ferricyanide Fe 5 (FeCy 6 ) 4 ? is formed by acting on K 3 FeCy 6 with Br. Potassium Ferrous Ferricyanide (Soluble Prussian Blue: FeK FeCy 6 ). Obtained by mixing solutions of FeCl 3 and K 4 FeCy 6 in equivalent proportions, stirring, and at once washing with cold water. Also by action of pure FeSO 4 on K 3 FeCy 6 in absence of air. When dried in vacuo it has the composition 4 (FeK FeCy 6 ) 7H 2 O. A blue solid, soluble in cold water. Solution decomposed by boiling, giving a yellowish precipitate; on adding acids or metallic salts, a blue precipitate. Alkalis and ammonia give Fe(OH) 3 and a ferrocyanide. Ferric salts give Prussian blue. Ferrous salts give Turnbull's blue. Ferricyanide of Silver (Ag 3 FeCy 6 ). A reddish-yellow precipi- tate obtained by adding K 3 FeCy 6 to AgNO 3 . Acted on by ammo- nia, forming a reddish double salt, which is soluble in excess of ammonia. (c) PERFERRICYANIDES (M 2 FeCy 6 ) When K 4 FeCy 6 solution is heated with iodine a greenish-brown liquid is formed, from which alcohol precipitates a crystalline salt, which appears to be K 2 FeCy 6 . It can be better prepared by acting on K 4 FeCy 6 with HC1 and KC1O 3 , neutralizing, and precipitating with alcohol. Nearly black. Dissolves to a deep violet solution. Gives green precipitates with many metallic salts. Very unstable. Acts as a powerful oxidizer. (d) NITROPRUSSIDES (M 2 FeCy 5 (NO) ) Formation. These compounds are formed by acting on alkali, ferrocyanides, or ferricyanides with HNO 3 . Two parts powdered K 4 FeCy 6 , 5 parts HNO 3 , and 5 parts water are warmed on a water- bath. A coffee-colored liquid is formed which gives off HCy, Cy, N, and C0 2 . When the liquid gives a dark-green or slate-colored precipitate with a ferrous salt (instead of blue), the reaction is complete. After cooling and filtering, the liquid is neutralized with an alkali, again filtered, and evaporated till the alkaline nitroprusside formed crystallizes. They are formed also by the action of nitrites on ferrocyanides, or on a boiling mixture of FeCl 3 and KCy. 94 THE CYANIDE HANDBOOK General Properties. The alkaline nitroprussides are soluble; also those of the alkaline earths. Most others are insoluble. Generally colored. CHARACTERISTIC NITROPRUSSIDES Metal Color of Salt Remarks NH 4 .... Very Unstable Ba Dark red Very Soluble Ca Very Soluble H Dark red Deliquescent K .... Dark red Deliquescent Na Ruby-red Non-deliquescent Cu Greenish Insoluble Fe" Yellowish pink Insoluble Ag Flesh-colored Insoluble Zn Yellow-rose Insoluble General Reactions. 1. Solutions of alkali nitroprussides give a deep, brilliant purple color with alkali sulphides. The color soon fades/ but is often used as a characteristic test for sulphides in solution. 2. Aqueous solutions decompose on heating or exposure to light, giving a blue precipitate. 3. When boiled with alkalis, Fe(OH) 3 is precipitated, N evolved, and a nitrite and ferrocyanide formed. 4. H 2 S precipitates Prussian blue, leaving a ferrocyanide in solution. 5. Soluble nitroprussides act as powerful oxidizing agents. 6. KMnO 4 forms NaN0 3 and a ferricyanide. COMPOUNDS CONTAINING A METAL ANALOGOUS TO IRON, FORMING A COMPLEX RADICAL WITH CYANOGEN The chief classes of these compounds are: Chromocyanides (M 4 CrCy 6 ) Chromicyanides (M 3 CrCy 6 ) Cobaltocyanides (M 4 CoCy 6 : unstable) Cobalticyanides (M 3 CoCy 6 ) Manganocyanides(M 4 MnCy 6 ) Manganicyanides (M 3 MnCy 6 ) . J The products of decomposition are ferrocyanide, thiocyanate, nitrite, sulphur, and ferric oxide. COMPLEX METALLIC CYANIDES 95 The chromo-, cobalto-, and mangano-cyanides (corresponding to ferrocyanides) are obtained by acting on the salts of the lower oxide of the respective metals with strong solutions of KCy. The chromi-, cobalti-, and mahgani-cyanides are obtained by oxidizing the above compounds, or by acting on the higher oxides or salt of the respective metals with KCy. Some of the acids belonging to this series, namely, H 4 CrCy 6 , H 4 CoCy 6 , H 3 CoCy 6 , and H 4 MnCy 6 , have been prepared. It has been suggested that mixtures of KCy with cobalti- cyanides or manganicyanides might be used as gold solvents. (B) METALLIC COMPOUNDS OF RADICALS IN WHICH CYANOGEN is COMBINED WITH OXYGEN OR A SIMILAR ELEMENT. A number of substances are known in which cyanogen appears to form part of a negative radical, in which it is intimately united with oxygen or another element (S, Se) of the oxygen group. These compounds have the general formula (MR"CN)x, where M is a + metal or group, and R" an element (O, S, or Se) forming part of a negative radical containing C and N in the pro- portions to form cyanogen. Since we may suppose the element R" to be directly united either with the C or with the N of the cyanogen, we have two classes of isomeric compounds, corresponding with the formulas X M - R" - C = N'" and M - R" - N v = C, respectively. The latter are distinguished by the prefix "iso." Compounds of the type (MRCN)x show a great tendency to polymerize, and different classes of bodies are known in which x = 1, 2, and 3, respectively. Different theories have been advanced to explain their con- stitution, which need not concern us here, and innumerable organic derivatives have been described, often of great complexity, in which M is a positive radical of the "alky!" type (e.g., CH 3 , C 2 H 5 ,C 6 H 5 , etc). The most important substances for our purpose fall under the following groups: 1 Other arrangements, e.g., M - C = N = R" and M - N = C = R", are also possible, and may occur in certain compounds. 96 THE CYANIDE HANDBOOK Cyanates, MOCN. Isocyanates, MONO. Fulminates, (MOCN) 2 . Isodicyanates, (MONC) 2 . Cyanurates, (MOCN) 3 . Isocyanurates, (MONC) 3 . Thiocyanates, MSCN. Dithiocyanates, (MSCN) 2 . Thiocyanurates, (MSCN) 3 . Selenocyanates, MSeCN. CYANATES (MOCN) Cyanic Acid (HOCN). This acid is only formed in small quantity by decomposing cyanates by stronger acids, as it im- mediately undergoes further decomposition. It may be made by the action of heat on its polymer, cyanuric acid (HOCN) 3 , or by dehydrating urea (with phosphorus pentoxide), or by passing dry HC1 over dry AgCNO. It is a thin, colorless liquid which reddens litmus and has a pungent odor resembling acetic acid. It excites tears, and raises blisters on the skin. Soluble without decomposition in ice-cold water. Readily polymerizes. It is decomposed by water as follows: HOCN + H 2 O - NH 3 + CO 2 . Metallic Cyanates: General Methods of Formation. 1. By fusing alkaline cyanides or ferrocyanides with easily-reduced oxides, such as Pb0 2 , or with oxidizing agents (nitrates, bichro- mates, etc.) : MCN + Q = MQCN 2. By passing cyanogen into a solution of an alkali or alkaline earth; e.g.. (CN)a + 2KOH = KCN + KOCN 3. By heating alkaline carbonates to low redness in an atmos- phere of cyanogen, or with mercuric cyanide: M 2 CO 3 + (CN) 2 = 2MOCN + CO. 4. By electrolysis of soluble cyanides. General Properties and Reactions. 1. Most metallic cyanates are soluble in water; those of Ag, Cu, Pb, and Hg are only slightly soluble. 2. Alkali cyanates are not decomposed by heating to dull redness in dry air. In moist air they give carbonates of am- monium and the alkali: 2MOCN + 4H 2 O = M 2 CO 3 + (NH 4 ) 2 CO 3 . COMPLEX METALLIC CYANIDES 97 Cyanates of the heavy metals are decomposed by heat into CO 2 and a cyanide of the metal: 2M'(OCN), = M*(CN)+M'+2CQ+Ni 3. Easily decomposed in solution by water into carbonates and ammonium salt (see above); or by dilute acids into CO 2 , an ammonium salt of the acid, and a metallic salt of the acid; e '9'> 1 MOCN + 2HC1 + H 2 O = MCI + NH 4 C1 + CO 2 . 4. Reduced to cyanides by heating with certain reducing agents H, C, K, or Fe. 5. Melted with sulphur, they give sulphides, sulphates, and thiocyanates. 6. Sodium amalgam reacts with neutral cyanates to produce formamide (HCONH 2 ). '7. Certain metallic cyanates are obtained by acting on salts of the required metal with alkaline cyanates, and crystallizing from alcohol. 8. Silver Cyanate: Silver salts give a white precipitate (perhaps AgONC), insoluble in cold water, soluble in boiling water; dissolved and decomposed by dilute nitric acid: AgOCN + 2HNO 3 + H 2 O = AgNO 3 + CO 2 + NH 4 NO 3 . Silver cyanate is decomposed by heat into Ag mixed with C and some N and CO 2 . It is soluble in NH 3 , giving a double com- pound which loses NH 3 in the air. Potassium Cyanate (KOCN) : Preparation. By the general methods given above; best as follows: 4 parts K 4 FeCy 6 and 3 parts K 2 Cr 2 7 , both dry and pulverized, are mixed, and the mixture heated, a little at a time, till it blackens. After cooling it is extracted with boiling alcohol, and KCNO crystallized out and dried in vacuo. Properties. Small, colorless, odorless laminae resembling KC10 3 . Soluble in water; fairly soluble in boiling hydrated alcohol. Insoluble in absolute alcohol. 1. According to some writers, 2 the salt present in commercial cyanide is an isocyanate (presumably KONG or KNCO), which is gradually transformed in solution into the normal cyanate 1 Concentrated HCl or dilute H 2 SO 4 give a certain amount of HOCN. 2 Bettel and Feldtmann, " Proc. Chem., Met. and Min. Soc. of South Africa," I, 273. 98 THE CYANIDE HANDBOOK (KOCN), which latter has no action on silver nitrate. This, however, is denied by other authorities, 1 who maintain that the decomposition taking place is merely the gradual formation of ammonium and potassium carbonate, referred to above: 2KOCN + 4H 2 = K 2 C0 3 + (NH 4 ) 2 CO 3 . 2. When BaCl 2 is added to a solution of KOCN, it gives at first no precipitate; but on boiling, barium carbonate is thrown down: (a) 2KCNO + BaCl 2 - Ba(CNO) 2 + 2KC1. (6) Ba(CNO) 2 + 4H 2 O = BaCO 3 + (NH 4 ) 2 CO 3 . 3. A solution of cobalt acetate gives, with potassium cyanate, dark blue quadratic crystals of a double cyanate (CoK 2 (CNO) 4 ). This reaction has been proposed as a test for cyanates. FULMINATES (MOCN) 2 These are unstable, often explosive substances, formed by the action of alcohol on metallic nitrates. The acid itself has not been isolated. The most important are those of silver, mercury, and gold. Mercuric fulminate (Hg(OCN) 2 ). Used in the manufacture of percussion caps. It is prepared by dissolving mercury in cold nitric acid, and pouring the solution into 90 to 92 per cent, alcohol. The salt deposits and is recrystallized from water. Crystallizes from alcohol in octahedra; from water in needles. Very slightly soluble in cold water, more soluble in hot water. Explodes by heat, friction, percussion, or treatment with H 2 SO 4 , yielding Hg, N, and CO. When heated with water it polymerizes to fulminurate. Forms double salts with cyanides. CYANURATES (MOCN) 3 Cyanuric acid is a product of the decomposition of urea by heat: 3CO(NH 2 ) 2 = (HOCN) 3 + 3NH 3 . It is precipitated, on dissolving the residue and diluting with water, in colorless, oblique, rhombic prisms (with 2H 2 0). Ef- floresces. Converted into anhydrous octahedra at 100-120 C. Slightly soluble in cold water, more soluble in hot water. Soluble in hot acids without decomposition. Heated in a closed tube, it gives cyanic acid: (HCNQ)3 = 3HOCN 1 Victor, in "Zeit. fur anal. Chem./' XL, 462-465. COMPLEX METALLIC CYANIDES 99 When dissolved in NH 3 it gives a fine pink precipitate with ammonio-sulphate of copper. PC1 5 forms (CNC1) 3 . Cyanurates are precipitated by Ba salts but not by Ca salts. They may occur sometimes in working cyanide solutions. The following formula has been suggested as representing their con- stitution: OH I C Only organic derivatives of the isodicyanates and isocyanurates are known. THIOCYANATES (MSCN) General Methods of Formation. 1. By fusion of cyanides or ferrocyanides with sulphur or with alkaline sulphides or poly- sulphides, or thiosulphates. 2. By the action of alkaline polysulphides on cyanides in aqueous solution. Properties. Soluble, often deliquescent salts. The thio- cyanates of copper (cuprous), lead, mercury, and silver are white or yellowish insoluble salts. General Reactions. 1. Solutions give a deep-red color with ferric salts: 3 KCNS + FeCl 3 = 3KC1 + Fe(SCN) 3 . 2. Alkali thiocyanates are decomposed by heat in presence of air, giving a sulphate and cyanate of the metal, and evolving SO 2 . Thiocyanates of the heavy metals give a sulphide, S, CS 2 , and finally (CN) 2 and N. 3. Oxidizing agents form sulphates: e.g., 5KSCN + 6KMnO 4 + 12H2SO 4 = 11KHSO 4 + 6MnSO 4 + 5HCN + 4H 2 O. 4. Copper salts in presence of a reducing agent precipitate white cuprous thiocyanate, insoluble in water and dilute acids: 2KCNS + 2CuSO 4 4- N a2 SO 3 + H 2 O = Cu 2 (SCN) 2 -I- KzSO* + 2NaHSO 4 , 5. Silver salts give a white curdy precipitate of silver thio- cyanate: KCNS + AgNO3 = AgSCN + KNOa. 100 THE CYANIDE HANDBOOK AgSCN is insoluble in water and dilute acids, soluble in am- monia, alkaline cyanides, and thiocyanates, and also in mercu- rous nitrate. Forms soluble double thiocyanates with KSCN. (NH 4 )SCN, etc. 6. Salts of mercury give white precipitates; the mercurous salt is unstable and readily decomposes with separation of Hg. Mercuric thiocyanate swells greatly on heating, giving off Hg, N, and CS 2 . Ammonium Thiocyanate (NH 4 SCN) : Preparation. 1. By digesting HCN with ammonium polysulphide (S dissolved in (NH 4 ) 2 S). On boiling, S separates, and the filtrate contains NH 4 SCN. 2. By decomposing cupric thiocyanate with NH 4 HS, filter- ing and evaporating. 3. By evaporating a mixture of ammonia with alcoholic solution of CS 2 . 4. By heating ammonium sulphate with carbon and sulphur: (NH 4 ) 2 SO 4 + C + S = NH 4 SCN + SO 2 + 2H 2 O. Properties. Large, white, deliquescent plates, very soluble in water and alcohol. Melts at 159 C. Dissolves in water with great absorption of heat. Reactions. 1. Heated to 180-190 C., it gives off CS 2 , H 2 S, and NH 3 . 2. Heated for some time nearly to melting-point, it is con- verted into its isomer, thiourea (CS(NH 2 ) 2 ). 3. Forms double salts with HgCy 2 and with various thio- cyanates. Potassium Thiocyanate (KSCN) : Formation. 1. By fusing potassium sulphide with ferrocyanide, K 2 S 2 3 is first formed; at a higher temperature this reacts to form KSCN. 2. By acting on ammonium thiocyanate with potassium sul- phide or potash at high temperature: 2NH 4 SCN + K 2 S = 2KSCN + (NH 4 ) 2 S. NH 4 SCN + KOH = KSCN + NH 3 + H 2 O. Properties. Long, white, striated prisms resembling niter. Very soluble in water, with great absorption of heat. MeJte at 161.2 C. Non-poisonous. Reactions. 1. On melting, the salt turns green, then indigo, COMPLEX METALLIC CYANIDES 101 but becomes white again on cooling. Heated in moist air, it gives CO 2 , NH 3 , and K 2 S. 2. Aqueous solution slowly decomposes (rapidly on boiling), evolving ammonia,: 2KSCN + 4H 2 = K 2 S + (NH 4 ) 2 S + 2CO 2 . 3. Heated with iron, it forms FeS, K 2 S, and K 4 FeCy e . 4. On electrolysis, it gives H 2 SO 4 , SO 2 , HCN, and S. 5. HC1 gas reacts violently with molten KSCN, forming HCN, CS 2 , and NH 4 C1. 6. Concentrated nitric acid forms " pseudo-sulphocyanogen " (C 3 N 3 HS 3 ). The same compound is formed by action of Cl on concentrated solution of KSCN. With excess of Cl the products are NH 4 C1, (NH 4 ) 2 SO 4 , and CO 2 . Cl, passed into molten KSCN, gives S 2 C1 2 and (CNC1) 3 . 7. Heated gently with PC1 5 , it gives KC1, CNC1, and PSC1 3 . 8. Forms double salts with HgCy 2 and HgI 2 . Thiocyanates of Gold and Potassium: Auro potassic Thiocyanate (KAu(SCN) 2 ). When auric chlo- ride (AuCl 3 ) is added drop by drop to KSCN at 80 as long as the precipitate dissolves, and the liquid is evaporated and crystallized, straw-yellow prisms are obtained, melting at 100 C. Decom- posed by heat into Au, S, CS 2 , and KSCN. Solution blackens in light. Gives precipitates with various salt solutions. NH 3 precipitates NH 3 AuSCN. Auripotassic Thiocyanate (KAu(SCN) 4 ). When AuCl 3 is added to excess of KSCN in the cold, this substance is formed, which crystallizes from warm water in orange-red needles. Soluble in water, alcohol, and ether. Forms double compounds. An analogous aurisodic thiocyanate, NaAu(SCN) 4 , is also known. SECTION IV CHEMISTRY OF THE DISSOLVING PROCESS Necessity of Oxygen in Ordinary Cyanide Process. The re- searches of Bagration, Eisner, Faraday, and latterly of J. S. Maclaurin, 1 have demonstrated that under ordinary circumstances the presence of oxygen is necessary for the solution of metallic gold and silver in cyanide solutions. The reaction may in fact be expressed in the form frequently quoted as " Eisner's equation," though not originally due to him : 2Au + 4KCy + O + H 2 O = 2KAuCy 2 + 2KOH. Intermediate Products of Reaction. The correctness of this equation as representing the final result of the reaction has been abundantly verified by Maclaurin and other investigators, but it is quite possible that intermediate products are formed, so that the actual mechanism of the reaction is not so simple as would appear from the above. Bodlander 2 states that hydrogen per- oxide is first produced, and that a portion of this is further de- composed, an additional amount of gold being dissolved. The reactions he gives are as follows: (a) 2Au + 4KCy + 2H 2 O + O 2 = 2KAuCy 2 + H 2 2 + 2KOH. (6) 2Au + 4KCy + H 2 O 2 = 2KAuCy 2 + 2KOH. It is stated that H 2 2 , or some substance having similar re- actions, can be detected in the solution. Bettel, however, states that the intermediate product is potassium auricyanide, and proposes the following: (a) 2Au + 6KCy + 2H 2 O + O 2 = KAuCy 2 + KAuCy 4 + 4KOH. (6) 2Au + 2KCy + KAuCy 4 = 3KAuCy 2 . The reactions in the case of silver appear to be precisely analogous. 1 Maclaurin, "Proc. Chem. Soc.," CXII, p. 81; CXLVI, pp. 7-8; CLXVIII, p. 149. 2 Bodlander, " Zeit. fur angew. Chem., " 1896, pp. 583-587. 102 CHEMISTRY OF THE DISSOLVING PROCESS 103 Use of Artificial Oxidizers. The necessity for oxygen being thus recognized, it seemed natural to suppose that the reaction would be assisted and hastened by the addition of substances which would readily yield oxygen under the conditions of cyanide treatment. The advantage, however, was only apparent in a few cases, namely, in those of alkaline peroxides, ferricyanides, and permanganates. In the case of peroxides the decomposition takes place rapidly in presence of water: Na 2 O 2 + H 2 O = 2NaOH + O; and consequently they are of little use unless added to dry ore previous to cyanide treatment. In the case of ferricyanides, the reaction is very probably quite independent of oxygen, as already pointed out. The reaction with permanganate is some- what complicated; formates, oxalates, nitrites, urea, and am- monia being among the products (in alkaline solutions), and it does not appear to be known with any certainty how permanga- nates act in aiding the solution of gold. If manganicyanides are formed, the reaction is probably analogous to that of ferricyanides. Solutions completely deprived of oxygen were found to have no solvent effect whatever; and the addition to such solutions of chlorates, perchlorates, chromates, bichromates, nitrates, nitrites, or bleaching powder had no effect, and did not bring about solu- tion of the gold. 1 Solution of Metals in Cyanide Independently of Oxygen. We have already noted several types of reaction in which gold is dissolved (usually by the interaction of two different cyanogen compounds, one of them being a simple alkali cyanide), in which oxygen plays no part. We may here briefly enumerate the fol- lowing cases, some of which have been already referred to: (1) The haloid compounds of cyanogen; e.g., 3KCy + CyBr + 2Au = 2KAuCy 2 + KBr. (2) Double cyanides of copper and alkali metal; e.g., C U2 Cy 2 - 4KCy. (3) Cuprammonium cyanides of the type x(NH 3 ) -y(CuCy 2 ) -z(Cu 2 Cy 2 ). These readily decompose with formation of the insoluble salt 1 Bettel and Marais; quoted by T. K. Rose, "Metallurgy of Gold," 4th ed., p. 380. 104 THE CYANIDE HANDBOOK 4NH 3 CuCy 2 2Cu 2 Cy 2 and liberation of cyanogen, which at the moment of formation attacks gold in presence of KCy: KCy + Cy + Au = KAuCy 2 . (4) Double cyanides of mercury and alkalis: K 2 HgCy 4 + 2Au = 2KAuCy 2 + Hg. (5) Double cyanides of zinc and alkalis. [These are stated to be solvents of gold by themselves, but it is evident that no reaction analogous to No. 4, above, can take place. It seems more probable that the molecule breaks up in solution: K 2 ZnCy 4 = 2KCy + ZnCy 2 . and that the dissolving of gold is due to the ordinary reaction of KCy in presence of O and H 2 O.] (6) Ferricyanides and analogous compounds: 2KCy + K 3 FeCy 6 + Au =K 4 FeCy G + KAuCy 2 . 2KCy + KsMnCye + Au = K 4 MnCy 6 + KAuCy 2 . In practice, these aids to solution have had only a limited application, but arrangements for bringing air into intimate con- tact with the material undergoing treatment, and with the cyanide solution, have sometimes proved of great value. In such cases it may be doubted whether the oxygen directly aids the solution of gold, or whether the beneficial action is not rather due to the destruction of cyanicides, or of precipitants of gold, present in the material to be treated. According to H. F. Julian, 1 free oxygen plays no primary part in the dissolution of gold by cyanide, but, on the contrary, exerts a retarding influence. He ascribes the solvent action of cyanide (as did Faraday) to the formation of local voltaic circuits. "These, in the first instance, deposit hydrogen and oxygen, which it may be assumed become occluded at their respective electrodes until the systems are in equilib- rium." When this condition is reached, less expenditure of energy is required to remove Cy from the solution than to occlude a further quantity of O, and accordingly AuCy is formed and is deposited in films. This compound has a high potential, and acts as an electrode, forming a couple with the uncombined gold. The AuCy then dissolves in the excess of KCy, " as one salt dissolves 1 Julian, "Brit. Assoc. Reports," South Africa, 1905, p. 369. CHEMISTRY OF THE DISSOLVING PROCESS 105 in the solution of another." When the solution contains dis- solved oxygen, this aids in a secondary manner by oxidizing the occluded hydrogen produced by the local voltaic currents, and thus upsetting the equilibrium. Factors that Influence Solubility of Gold and Silver. What- ever may be the correct theoretical explanation of the reaction, it has been observed in practice that certain factors exert an influence on the rate of solution of the precious metals in cyanide. The most important of these are: (1) Degree of concentration of the solution. (2) Temperature. (3) Presence or absence of "Cyanicides," i.e., of substances other than the precious metals, capable of decomposing or com- bining with cyanides. (4) Amount of "available oxygen," either free or in some easily decomposed compound. (5) Physical condition of the surfaces of metallic gold and silver in contact with the solution. (6) Area of metal exposed in proportion to weight. (7) Relative masses of solvent, and matter undergoing treat- ment in contact with it. Influence of Strength of Solution. It has been found by Maclaurin * that, other things being equal, both gold and silver dissolve most rapidly in solutions of a strength corresponding to 0.25 per cent. KCy (i.e., 0.10 per cent. Cy). The rate of solution falls off gradually above or below this strength, but is nearly constant between -0.1 per cent., and 0.25 per cent. KCy. His experiments show that the rate of solution of silver is about two- thirds that of gold, but otherwise the phenomena are similar. L. Janin's experiments, 2 however, show, for cement silver, a maximum solubility at 1 per cent. KCy. The reduced efficiency of stronger solutions has been ascribed to the diminished solu- bility of oxygen in strong KCy. It would seem, however, that + + the dissociation of cyanide into its ions, K (or Na) and CN, is an essential condition to the solution of gold or any metal therein. Since this dissociation takes place the more completely, the more dilute the solution, there must be some point of maximum effi- 1 J. S. Maclaurin, "Journ. Chem. Soc.," LXIII, 731; LXVII, 199. 2 L. Janin, " Eng. and Min. Journ.," Dec. 29, 1888. 106 THE CYANIDE HANDBOOK ciency at which the solution contains the greatest possible num- 4 ^ V ' ber of CN ions per unit volume. \^ V Influence of Temperature. It has long been known that U* \,* hot solutions dissolve the metals more rapidly than cold ones, and the greater efficiency of hot cyanide solution for dissolving gold was pointed out by Bagration in 1843. Experimental trials with solutions heated to about 130 F. (55 C.) have generally shown a somewhat higher extraction than at the ordinary tem- / (. perature. In the practical application of this observation, how- f\ ever, there are economic difficulties, depending on the cost of raising large volumes of liquid to a high temperature; and more- over, there is in hot solutions a somewhat rapid decomposition of cyanide, as described below. Influence of other Soluble Substances: Electrochemical Order of Metals in Cyanide Solution: The relative solubilities of metals in cyanide (or other solvent) depend upon their electrochemical order in that solvent. Determinations of the electrochemical order of metals in potassium cyanide solution, under various conditions of dilution and temperature, have been published by Gore, Skey, Christy, and others. Generally speaking, the order, from positive to negative, of the ordinary metals, is as follows: + Mg Ag Al Hg Zn Pb Cu Fe Au - Pt Carbon is to Pt. From this table it may be inferred that any metal in the list would tend to dissolve in cyanide more readily than the metal below, and would displace that metal from solution and precipitate it; for example, zinc will precipitate Cu, Au, or Ag; copper will precipitate Au or Ag; gold or silver will precipitate mercury; magnesium or aluminium will precipitate gold or silver more readily than will zinc. The solubility, and in some cases the order, is affected by differences of concentration and temperature. Experiments on the electromotive force of minerals in cyanide CHEMISTRY OF THE DISSOLVING PROCESS 107 solutions * show that most minerals are electronegative to metallic gold, so that gold tends to dissolve in preference to such minerals. In ore treatment, it is generally found that the percentage of gold extracted is higher than that of silver, but this may easily be accounted for when it is considered that the actual weight of silver present in an ore is generally much greater than that of gold, and that the silver is largely present, not as metal, but as a compound such as Ag 2 S, only slightly soluble in KCy. Selective Action of Dilute Solutions. It is claimed that dilute cyanide solutions have a "selective action" for gold and silver in preference to base metals. In the case of the majority of ores, the actual experimental facts may be stated as follows: (1) It is not usually the case that the actual amount of base metal dissolved is less than that of precious metal; but whether the solution be strong or weak, the amount of base metal dissolved relatively to the total quantity of base metal present is less than the quantity of precious metal dissolved relatively to the total quantity of precious metal present. (2) The weight of precious metal dissolved per unit of base metal or compound dissolved in a given case is the greater the more dilute the solution. In all actual cases, the precious and base metals are dissolved simultaneously, but the rate of solu- tion of the precious metals is less diminished by dilution than that of the base metals. These phenomena are nx>t by any means confined to cyanide solutions: the same laws are observed in the action of thiosul- phates on silver chloride, in ores which have undergone a chloridiz- ing roast, and in the action of acids on powdered rock-mixtures containing different percentages of almost any two or more in- gredients soluble in the acid. In general terms, it may be said that the ingredient which is present in the smaller percentage is dissolved in the greater relative amount; the difference becoming more marked the more dilute the solution. Since the formation of base-metal compounds is the main source of consumption of cyanide in the treatment of most ores, it follows that the consumption of cyanide may be taken, roughly, as an index of the amount of base metal dissolved in any given case. It is nearly always found that the consumption of cyanide increases regularly as the strength of the solution is increased, > Christy, Trans. A. I. M. E. (Sept., 1899), XXX, 1-83. 108 THE CYANIDE HANDBOOK but that the amount of gold or silver dissolved only increases slightly, if at all, with increasing strength of cyanide, provided a reasonably sufficient time be given for the action. The experi- ments by Mac Arthur, quoted by Rose, 1 illustrate this point, as do also the following tests by Caldecott; 2 in which charges of 34 tons each, assaying originally 100 grains per ton, were treated for three days with various strengths of solution, and subsequently water-washed. Strength of Cyanide Used. KCy, Per Cent. Strength after Treatment. KCy, Per Cent. KCy Consumed. Per Cent. Gold Extracted. Per Cent. .041 .010 .017 83 .110 .068 .024 84 .373 .277 .055 83 1.023 .860 .095 85 3.333 3.020 .180 83 Many other experiments might be quoted from which it would appear that while the extraction is practically constant, the con- sumption is higher the greater the cyanide strength. Decomposition of Cyanide by External Influences. In its early days, objections were frequently raised against the practical application of the cyanide process on account of the instability of potassium cyanide when exposed to atmospheric influences. It was urged that this substance undergoes spontaneous decom- position, and that it is rapidly affected by atmospheric oxygen and carbonic acid. A dilute solution of KCN, as already pointed + out, is probably to a large extent dissociated into the ions K and CN. Hydrolysis. In addition to this, but to a smaller extent, a decomposition known as "hydrolysis" occurs, as follows: KCy + H 2 O = KOH + HCy. The HCy thus formed is continually volatilized, and the smell of it is always noticeable over vessels containing a solution of any alkali cyanide. The alkali cyanides in the solid state are not stable unless 1 "Metallurgy of Gold," 4th edition, p. 392. 2 "Proc. Chem., Met. and Min. Soc. of South Africa," I, 294. CHEMISTRY OF THE DISSOLVING PROCESS 109 perfectly dry; sodium cyanide is especially liable to decomposition when existing as a hydrate, and reacts with water evolving ammonia; thus: NaCN + 2H2O = HCO2Na + NH3- A similar decomposition occurs with all soluble cyanides of the alkalis and alkaline earths, but much more rapidly in hot solutions, producing a formate of the metal and evolving am- monia gas. The hydrolytic decomposition into formates and ammonia is also promoted by the presence of large amounts of alkali. It has been stated that sodium cyanide is less stable in solu- tion than potassium cyanide of corresponding cyanogen strength, or than the mixed cyanides frequently obtained as a commercial product. Oxidation. The cyanides also, both in the solid state and in solution, gradually absorb oxygen and carbonic acid from the air; the solid cyanides are deliquescent and are gradually acted upon by the moisture they absorb from the air. They are ulti- mately converted into carbonates, possibly in some cases with formation of cyanates as intermediate products, the ultimate reaction being: 2KCN + O 2 + 4H 2 O = K 2 CO 3 + (NH^Oa. The oxidation in solution is, however, very slow under ordinary circumstances, and air or oxygen may be injected under pressure into a cyanide solution for a long time without any appreciable oxidation of cyanide taking place. Action of Cafbonic Acid. In the absence of free alkali, carbonic acid decomposes cyanide; thus: KCN + CO 2 + H 2 O = HCN + KHCO 3 . This does not take place, however, if sufficient caustic alkali (NaOH, KOH, Ca(OH) 2 , Ba(OH) 2 ) or ammonia, or alkaline car- bonate (K 2 CO 3 , Na 2 CO 3 ), be present to neutralize the CO 2 accord- ing to the reactions: KOH + CO 2 = KHCO 3 . -f CO 2 + H 2 O = 2KHCO 3 . Alkali bicarbonates afford no protection against CO 2 ; as soon as all hydrates and carbonates have been converted into bicarbonates, any further quantities of CO 2 entering the solution at once de- compose the cyanide. These decompositions are, however, of 110 THE CYANIDE HANDBOOK less importance in dilute solutions than might be supposed, and the addition of a slight excess of alkali prevents the rapid evolu- tion of HCy, even when solutions are exposed to the. air in open vessels. Action of Base Metal Compounds: Iron Compounds. In the treatment of ores a certain amount of cyanide is always con- sumed in attacking the compounds of base metals contained therein. Of these metals the commonest is iron, which occurs in a variety of conditions, some of which are unaffected by cyanide, while others are rapidly attacked. Iron in the form of ferric oxide is quite unaffected by cyanides; the same is also probably true of anhydrous ferrous oxide and magnetic oxide. Ferrous hydrate, however, is rapidly acted upon with formation of f errocyanide : Fe(OH) 2 + 2KCy = FeCy 2 + 2KOH. FeCy 2 -f 4KCy = K 4 FeCy 6 . Similarly with ferrous carbonate: FeC0 3 + 6KCy - K 4 FeCy 6 + K 2 CO 3 . which reaction explains the efficiency of ferrous carbonate as an antidote for cyanide poisoning. Iron pyrites exist in ores in two different crystalline conditions, known as "pyrite" and "marcasite." Pyrite, which occurs in dense yellowish cubical crystals, is only slowly acted on by cyanide, whereas marcasite, which is softer, is much more readily attacked, perhaps directly as follows: FeS 2 + KCy = FeS + KSCy. In any case, under ordinary conditions of treatment, pyrites rapidly undergoes oxidation, forming sulphates and basic sul- phates of iron, which act readily as cyanicides. Caldecott * gives the following as the main stages in the decomposition of pyrite or marcasite: (1) FeS 2 Iron pyrites. (2) FeS + S Ferrous sulphide, sulphur. (3) FeSO 4 + H 2 SO 4 .... Ferrous sulphate, sulphuric acid. (4) Fe 2 (SO 4 ) 3 Ferric sulphate. (5) 2Fe 2 O 3 SO 3 Insoluble basic ferric sulphate. (6) Fe 2 O 3 Ferric oxide. Ferrous sulphide is injurious owing to its power of absorbing i " Proc. Chem., Met. and Min. Soc. of South Africa," II, 98. CHEMISTRY OF THE DISSOLVING PROCESS 111 oxygen, and may perhaps actually decompose KAuCy 2 so as to precipitate gold that has been previously dissolved. Thio- cyanates are almost invariably present in cyanide solutions after use in extraction of ores. As sulphur does not dissolve directly in cyanide, they may be due to the action on pyrites, as stated above, or to action on ferrous sulphide: FeS + 7KCy + H 2 O + O = KCyS + K 4 FeCy 6 + 2KOH. All the iron in working cyanide solutions probably ultimately takes the form of soluble alkali ferrocyanides (K 4 FeCy e or Na 4 FeCy 6 ) , since ferrous thiocyanate, if formed at all, is unstable and would react with KCy: Fe(CyS) 2 + 6KCy = K 4 FeCy 6 + 2KCyS. Ferrous and ferric sulphates act directly as cyanicides, ulti- mately giving rise to Prussian blue and similar compounds, or to hydrocyanic acid. In the case of ferrous sulphate, some such series of reactions as the following may be supposed to take place: FeS0 4 + 2KCy = FeCy 2 FeCy 2 + 4KCy = K 4 FeCy 6 . 2FeSO 4 + K 4 FeCy 6 = Fe 2 FeCy 6 + 2K 2 SO 4 . 3Fe 2 FeCy 6 + 3O + 3H 2 O = Fe 4 (FeCy 6 ) 3 + 2Fe(OH) 3 . In the case of ferric sulphate: Fe 2 (SO 4 ) 3 + 6KCy = 2FeCy 3 + KzSO*. FeCy 3 + 3H 2 = Fe(OH) 3 + 3HCy. In presence of sufficient alkali the soluble ferric salts are pre- cipitated as ferric hydrate, which is unacted upon by cyanide; l ferrous salts give ferrous hydrate, which readily forms ferro- cyanide, as above described. The basic sulphates, although insoluble in water, are attacked by cyanide, giving a similar series of reactions. Action of Copper Compounds on Cyanide. Copper exists in ores most frequently as sulphide, often combined with sulphides of other metals, copper pyrites being a compound, or intimate mixture, of the sulphides of iron and copper. It also occurs as native copper, and in oxidized ores as carbonate, oxide, silicate, and other forms. i According to Julian and Smart (" Cyaniding Gold and Silver Ores," 2d edition, p. 112), "ferric hydrate" is an indefinite compound, Fe 2 O 3 .xH 2 O, which, under certain conditions, may dissolve in cyanide, forming a ferrocyanide. 112 THE CYANIDE HANDBOOK Action of Cuprous Sulphide. The natural sulphides of copper are much less readily acted on by cyanide than are the carbonates, or than artificially prepared sulphide, but they grad- ually dissolve, probably with formation, finally, of cupric thio- cyanate, by some such series of reactions as the following: (a) 2Cu 2 S + 4KCy + 2H 2 O + O 2 = Cu 2 (CyS) 2 + Cu 2 Cy 2 + 4KOH. The insoluble cuprous thiocyanate and cyanide readily dis- solve in excess of KCy: (6) 2Cu 2 (CyS) 2 + 8KCy + H 2 O + O = 2Cu(CyS) 2 + Cu 2 Cy 2 6KCy + 2KOH. (c) Cu 2 Cy 2 + 6KCy - Cu 2 Cy 2 6KCy. Since the solutions almost invariably contain an excess of alkali thiocyanate, the further reaction is probably as follows: (d) Cu 2 Cy 2 -6KCy + 4KCyS + H 2 O + O = 2Cu(CyS) 2 + 8KCy + 2KOH. Action of Carbonate of Copper. Carbonate of copper, both native and artificial, is attacked by cyanide with extreme rapidity, and is one of the most troublesome cyanicides. The action, when KCy is in excess, appears to be as follows: CuCO 3 4- 2KCy = CuCy 2 + K 2 CO 3 . 2CuCy 2 + 4KCy = Cu 2 Cy 2 4KCy + Cy 2 . When sufficient alkali is also present: Cy 2 + 2KOH = KCy + KCyO + H 2 O; so that the entire reaction becomes: 2CuCO 3 + 7KCy + 2KOH = Cu 2 Cy 2 4KCy + KCyO + 2K 2 CO 3 + H 2 O. Soluble cupric salts (CuSO 4 , etc.) act in an analogous way, evolving cyanogen in unprotected solutions, and forming double cyanides probably of the type Cu 2 Cy 2 nKCy. Removal of Copper Before or During Treatment. Copper, when present in the form of carbonate, oxide, hydrate, etc., may be largely removed from the ore: (1) By preliminary acid treatment, with H 2 SO 4 or H 2 S0 3 ; any acid remaining in the charge after this operation may be neutralized with alkali previous to cyanide treatment. (2) By preliminary treatment with ammonia, which dissolves many copper compounds that are insoluble in water. (3) In the course of cyanide treatment, by the use of a mix- ture of cyanide with ammonia or ammonium salts. Such mixtures CHEMISTRY OF THE DISSOLVING PROCESS 113 + + + are largely dissociated in dilute solutions into the ions NH 4 , K, Na, and CN, Cl, OH, etc.; they, therefore, act as powerful solvents, and appear to attack gold under certain conditions in preference to copper. The copper forms " cuprammonium " compounds of the type xNH 3 yCuCy 2 zCu 2 Cy 2 , which are only slightly soluble in water, forming blue solutions from which green, insolu- ble, needle-like crystals of ammonium dicuprosocupric cyanide (4NH 3 CuCy 2 2Cu 2 Cy 2 ) are deposited. At the same time cyano- gen is liberated, which at the moment of formation readily attacks gold in presence of excess of KCy: KCy + Cy + Au = KAuCy 2 . Gold is thus dissolved and copper precipitated at the same time. Copper may also be precipitated from solutions by the addition of acids; thus: -nKCy + nH 2 SO 4 = Cu 2 Cy 2 + nHCy + nKHSO 4 ; 2Cu(SCy) 2 + H 2 SO 4 = Cu 2 (SCy) 2 + HSCy + HCy + 2SO 2 forming the insoluble cuprous cyanide and thiocyanate. General Action of Metallic Sulphides. Many metallic sul- phides are more or less soluble in alkalis, with formation of alka- line sulphides: 2KOH + R ,, s = R * (O H) 2 + K,S. Since the cyanide solutions almost invariably contain an excess of caustic alkali, this reaction takes place to some extent in the treatment of ores containing such sulphides. The soluble alka- line sulphide locally formed may sometimes act as an injurious factor in the process, owing to its powerful reducing action, which removes the oxygen necessary for the solution of the gold. It is, however, probably *soon converted into thiocyanate by the reaction KzS + KCy + H 2 O + O = KCyS + 2KOH. and also thrown out of solution by the reaction KaS + K 2 ZnCy 4 = ZnS + 4KCy. It is partly owing to this latter reaction that the zinc, dis- solved as double cyanide in the precipitation process, is removed from the solutions before it has accumulated to an injurious extent. As a general rule, clean unoxidized minerals have only a slight 114 THE CYANIDE HANDBOOK action on cyanide solutions. Christy's experiments 1 show that nearly all such minerals are electronegative to gold in cyanide solutions, so that the latter would dissolve in preference. Never- theless, a large consumption of cyanide is sometimes observed in the treatment of ores containing, for example, mispickel, realgar, stibnite, cinnabar, galena, blende, fahl ore, sulphides of nickel and cobalt, etc., especially when the sulphides have under- gone partial oxidation. In some cases, as with the sulphides of arsenic, antimony, and tin, it is possible to extract the cyanicide by a preliminary opera- tion, in which the ore is treated with a (preferably hot) solution of an alkali or alkaline sulphide. Zinc-blende is not much attacked by cyanide, but calamine is a rapid cyanicide; thus: ZnCO 3 + 4KCy = K 2 ZnCy 4 + K 2 CO 3 . Action of Selenium and Tellurium. Metallic selenides, especially when finely crushed, are dissolved in cyanide without much difficulty, the element passing into solution as a seleno- cyanide; e.g., KSeCy. Tellurium dissolves with more difficulty, perhaps first forming an alkaline telluride, K 2 Te, which is rapidly converted into a tellurite, K 2 TeO 3 . Ores containing considerable quantities of refractory minerals, especially those of arsenic and tellurium, are perhaps best treated by preliminary roasting, to remove or oxidize the greater part of these elements; but in some cases they have been successfully treated by very fine crushing, followed by bromocyanide. When an ore has been partially oxidized, either by natural processes or by roasting, it is very necessary to extract the soluble matter by water-washing previous to cyanide treatment. Such partially oxidized material may contain large amounts of soluble sulphates, which may cause a very much larger consumption of cyanide than the raw unoxidized ore. Addition of alkalis, with- out previous water-washing, is not always a satisfactory remedy, as metallic hydrates, such as Fe(OH) 2 , may be precipitated, which are themselves cyanicides. Oxides of manganese sometimes appear to form very easily decomposable double cyanides, which readily deposit the brown hydrated oxide. i Trans. A. I. M. E. (Sept., 1899), XXX, 33. CHEMISTRY OF THE DISSOLVING PROCESS 115 Regeneration of Solutions. Since the main cause of cyanide consumption is the formation of soluble double cyanides or com- plex cyanogen compounds of the base metals, and since solutions highly charged with such compounds are more or less inefficient as solvents of gold and silver, it has been suggested that the cyanogen in such liquors might be recovered in the form of simple alkali cyanides, by treatment with suitable chemicals. This has been carried out in practice in some cases, but generally speaking the cost of chemicals, power, labor, etc., required for such treatment outweighs the advantage gained by it. The following methods have been suggested : (1) The solution is acidulated, and the cyanogen converted wholly or partially into hydrocyanic acid, in some cases with precipitation of a portion as insoluble cyanides; thus: K 2 ZnCy 4 + 2H 2 SO 4 = ZnCy 2 + 2HCy + 2KHSO 4 . ZnCy 2 + H 2 SO 4 = ZnSO 4 + 2HCy. KAgCy 2 + H2SO 4 = AgCy + HCy + KHSO 4 . 4KCy -Cu 2 Cy 2 + 4H2SO 4 = Cu 2 Cy 2 + 4HCy + 4KHSO*. (2) Ferric salts may be added, to precipitate ferrocyanides as Prussian blue (for reactions see above). After drying, the Prussian blue is converted into an alkali cyanide by fusion in a closed vessel with alkali or alkaline carbonate. (3) Alkaline sulphides decompose the double cyanides of silver, mercury, zinc, cadmium and nickel, precipitating these metals more or less completely as sulphides and forming a soluble alkali cyanide; e.g. : 2KAgCy 2 + K.S = A g2 S + 4KCy, etc. The reaction, however, is never complete, as the sulphides formed are not absolutely insoluble in the excess of cyanide. Moreover, any excess of sulphide must be removed and the solu- tion thoroughly aerated before it can be successfully used in treat- ing fresh charges of ore. Reprecipitation of Gold and Silver during Dissolving Process. Under certain conditions, a premature precipitation of gold and silver may occur in the leaching tanks or other parts of the dissolv- ing plant, and may lead to losses. It is well known that charcoal is a precipitant of gold from cyanide solutions; an admixture of carbonaceous matter may therefore cause local precipitation of values previously dissolved. This was a frequent cause of trouble 116 THE CYANIDE HANDBOOK with the black cyanide formerly in use (containing carbon and carbides of iron). When this cyanide was dissolved, it left a layer of black sediment, which sometimes contained considerable values after being some time in contact with the auriferous solu- tions. Decaying organic matter, such as grass roots, etc., may likewise cause losses in this way. Local acidity of the ore may cause precipitation of insoluble aurous cyanide: KAuCy 2 + H 2 SO 4 = AuCy + HCy + KHSO 4 . Under some circumstances gold and silver may be thrown down as insoluble salts, by the action of soluble base metal salts; thus: 2KAgCy 2 + ZnSO 4 = K 2 SO 4 + ZnAg 2 Cy 4 . The action of alkaline sulphides on silver double cyanide has already been noted; gold is not precipitated by these reagents. Some metallic sulphides, such as Cu 2 S, seem to have the power of throwing down gold from a cyanide solution. SECTION V CHEMISTRY OF PRECIPITATION AND SMELTING PROCESSES IT has already been pointed out that gold and silver are electro- negative to zinc in cyanide solutions; hence we should expect that zinc would act as a precipitant of the precious metals. This effect, however, is obviously complicated by various secondary re- actions. The following points are always noticed in this connec- tion: (a) Precipitation only takes place in presence of sufficient free cyanide. (6) It is always accompanied by increase in the alkalinity of the solution. (c) It is always accompanied by evolution of hydrogen. The entire effect, as regards precipitation of gold, is probably expressed by the equation: KAuCy 2 + 2KCy + Zn + H 2 O = K 2 ZnCy 4 + Au + H + KOH; but the solution of zinc goes on to a large extent quite inde- pendently of the precipitation of gold, and bears no necessary proportion to it. The probable reaction is Zn + 4KCy + 2H 2 O = K 2 ZnCy 4 + 2KOH + H 2 , though some writers have supposed that an alkaline zincate, such as Zn(OK) 2 , may be formed. It would appear that the electric couples Zn: Pb and Zn: Au, etc., give rise to local currents which electrolyze water, the oxygen at the moment of formation attacking zinc to form hydroxide: Zn + O + H 2 = Zn(OH) 2 . Under the usual working conditions this redissolves in the excess of cyanide; thus: Zn(OH) 2 + 4KCy = K 2 ZnCy 4 + 2KOH; but an insufficiency of free cyanide may lead to its deposition in the form of a white precipitate on the surface of the shavings. 117 118 THE CYANIDE HANDBOOK As before pointed out, the double cyanides of zinc and alkali metals may be assumed partially to dissociate in dilute solutions, so that we have KCy, NaCy, and ZnCy 2 actually in solution, and these are, perhaps, still further split up. In the absence of suffi- cient free alkali or alkaline cyanide, this ZnCy 2 may also be de- posited. Sharwood 1 has investigated the conditions under which zinc oxide dissolves in cyanide, and zinc cyanide in alkaline hydrox- ides. He finds that ZnO is soluble in .65 per cent. KCy with formation of a double cyanide, and of a compound of zincate with excess of alkaline hydrate: 20KCy + 6ZnO + 4H 2 O = 5K 2 ZnCy 4 + K 2 ZnO 2 -8KOH; the latter being the result of a secondary reaction; thus: 4KCy + ZnO + H 2 O = K 2 ZnC yi + 2KOH; 10KOH + ZnO = K 2 ZnO 2 8KOH + H 2 O. The action of KOH on ZnCy 2 is as follows: (a) When the ZnCy 2 is in less proportion than ZnCy 2 : 2KOH it dissolves completely and permanently: 4KOH + 2ZnCy 2 = K 2 ZnCy 4 + K 2 ZnO 2 + 2H 2 O. On heating, the zincate decomposes: K 2 ZnO 2 + 2H 2 O = Zn(OH) 2 + 2KOH. (b) When ZnCy 2 is in the proportion ZnCy 2 :KOH, it is tem- porarily dissolved, but soon begins to precipitate Zn(OH) 2 , as follows: 2ZnCy 2 + 2KOH = K 2 ZnCy 4 + Zn(OH) 2 . (c) With larger proportions of ZnCy 2 any excess remains un- affected by KOH. NaOH acts in a similar way. It seems probable that any excess of KCy would convert the zincate into the double cyanide, but as increasing the alkalinity undoubtedly increases the solvent efficiency of the solution, it may be that we have here a reversible reaction: K 2 ZnCy 4 + 4KOH ;=! Zn(OK) 2 + 4KCy + 2H 2 O. Various investigators have made more or less complete analyses of the white precipitate above referred to; somewhat conflicting results have been obtained, and the substance no doubt varies according to the conditions of formation. Experience has shown i "Bull. Univ. Cal.," April 27, 1904. CHEMISTRY OF PRECIPITATION 119 that the formation of the precipitate is greatly promoted by exposure of the zinc shavings to the air. It occurs chiefly as a flocculent, whitish deposit in the upper compartments of the boxes, and often interferes seriously with effective precipitation of the gold and silver. The general conclusion is that it consists principally of zinc hydrate, with smaller quantities of cyanide and ferrocyanide of zinc, with small amounts -of the sulphates and carbonates of calcium, lead, etc.; also alumina, silica, and other ingredients of less importance. The following analyses give some idea of its composition: (1) By A. Whitby 1 (the metals being calculated as their equivalent oxides) : ZnO 54.00 per cent. SiO 2 and insoluble 2.10 Fe 2 O 3 + A1 2 O 3 75 PbO 73 CaO . .50 CN 2.50 K 2 O 70 Loss on ignition 36.80 (2) By Bay and Prister 2 : Zinc hydrate, Zn(OH) 2 54.79 per cent. Zinc cyanide, ZnCy 2 22.73 Zinc-potassium ferrocyanide, K 2 Zn 3 (FeCy 6 ) 2 10.45 Silica, SiO 2 1.03 Ferric oxide, Fe 2 O 3 1.00 Cupric oxide, CuO 40 The zinc hydrate and cyanide are readily soluble in excess of KCy or of caustic alkali (also in dilute acids), but the ferro- cyanide is not easily soluble, and if once formed generally re- mains in the boxes until the clean-up. There are grounds for supposing that gold and silver are pre- cipitated on zinc, either in the metallic condition or as an alloy, according to the concentration of the liquors undergoing pre- cipitation in metal and cyanide. Argall states that when the solutions pass 1.5 ounces gold per ton, the gold on the shavings assumes a yellow color, and passing 2 ounces and upward is usually golden yellow in the first compartment, shading off in 1 " Proc. Chem., Met. and Min. Soc. of South Africa," V, 55. *IUd., V, 77. 120 THE CYANIDE HANDBOOK the following compartments to the usual black-colored deposit obtained from poor solutions. The alloys so formed are more energetically attacked than pure zinc by the cyanide solutions, and hence act more efficiently as precipitants. Silver also presents a totally different appearance, according to the conditions of precipitation. Action of Copper in Precipitation with Zinc. Solutions con- taining copper deposit this metal in the upper compartments of the zinc-box; the copper forms a closely-adherent metallic film, which coats the zinc shavings in such a way that all further action ceases and the precipitation of gold and silver is prevented. In strong solutions, copper is precipitated in preference to gold and silver. When copper has accumulated to a certain extent in the solution, it becomes practically impossible to use the liquor either for extraction of gold and silver or for precipitation, and the copper must be removed by some means or the liquor run to waste. The zinc-lead couple has been sometimes used with advantage for cupriferous solutions; certain compartments filled with zinc shavings previously dipped in lead acetate are found to collect most of the copper. Barker, 1 however, says that although " with a good spongy coating of lead on the zinc the precipitation of the copper is at first good, the copper coming down in a loose spongy form," the action is not continuous. He therefore sug- gests precipitating as cuprous cyanide by the addition of sulphuric acid. When the copper exists in the solutions as a double cy- anide (e.g., Cu 2 Cy 2 4KCy, or Cu 2 Cy 2 6KCy), the precipitation of copper is accompanied by regeneration of cyanide, as shown by the following reactions : Cu 2 Cy 2 4KCy + Zn = Cu 2 + K 2 ZnCy 4 + 2KCy. 6KCy + Zn = Cu 2 + K 2 ZnCy 4 + 4KCy. Small quantities of mercury are almost invariably present in the solution before precipitation. Mercury is carried off mechani- cally with the tailings from the battery and gradually dissolves to furm K 2 HgCy 4 . In some cases it may also be partly derived from mercury minerals contained in the ore itself. The com- pound is directly decomposed by zinc, as follows: K 2 HgCy 4 + Zn = K 2 ZnCy 4 + Hg. * Trans. I. M. M., XII, 399. CHEMISTRY OF PRECIPITATION 121 As already pointed out, the double cyanides of mercury are gold solvents, and also act by decomposing soluble sulphides, so that the presence of this metal in small quantities is a distinct advantage. The precipitate formed in the zinc-boxes occasionally contains other elements; such as arsenic, antimony, selenium, tellurium, etc., derived from the material treated. The removal of these in the' subsequent treatment may involve special operations, as noted below. Whenever the zinc shavings have been treated with lead acetate, the precipitate will contain a certain amount of metallic lead. Electrolytic Precipitation. Many methods have been tried or suggested for precipitating gold and silver electrolytically from cyanide solutions. The best known is that of Siemens and Halske, which, in the form originally adopted in South Africa, used anodes of iron and cathodes of lead. The cathodes were not attacked under ordinary circumstances, but an adherent deposit of gold was formed upon them. Under the influence of the electric current the aurocyanide is ionized into + +. + K(or Na), Au and Cy Cy. The K, of course, goes to form KOH and liberates H at the cathode. The Cy attacks the iron anode, ultimately forming ferrocyanide and a certain amount of Prussian blue, ferric hydrate, and per- haps other by-products. Afterward, the cathodes and anodes in use were both of lead, but the lead anodes were found to become brittle after a time and were coated superficially with a white deposit, probably oxycyanide of lead. Tinned iron cathodes with lead anodes were also used at one time; the current density was so regulated that the gold was precipitated, not in an adherent form, but as a loose powder which continually fell off the cathodes and settled at the bottom of the box, whence it could be easily collected in cleaning up. In the latest form of the process, adopted at the San Sebastian Mine, Salvador, Central America, 1 the ' cathodes are rolled lead plates, and the anodes are similar lead plates which have been previously coated with a layer of peroxide of lead by immersing i C. P. Richmond, " Eng. and Min. Journ., " LXXX, 512. 122 THE CYANIDE HANDBOOK in a solution of permanganate. The difficulty with regard to copper is got over in the plant referred to by using electric pre- cipitation for recovering part of the copper and gold and then passing the solution, comparatively free from copper, through ordinary zinc-boxes to recover the residual gold. Other electrolytic methods, using amalgamated zinc or copper as electrodes, have also been tried, but with little practical suc- cess. Aluminium has been suggested as a material for the elec- trodes, but although effective as a precipitant, it causes much difficulty by forming infusible compounds in the subsequent smelt- ing operations. Other Precipitation Processes. Aluminium in the form of shavings has also been suggested, and occasionally tried on a working scale, as a substitute for zinc, without the aid of elec- tricity. It is even more electro-positive than zinc, but a deposit of alumina is formed on the shavings and causes much trouble in the treatment of the precipitate. Sodium amalgam has been used, either directly or in conjunc- tion with an electric current. Theoretically, the action is very Simple: KAuCy 2 + Na = Au + KCy + NaCy; but in practice it is found that the large surface required precludes the economic application of the process. In the Gilmour- Young process the ore is subjected to com- bined pan-amalgamation, and cyanide treatment. After grinding in pans with mercury and cyanide solution, copper and zinc amalgam are added, which precipitate the gold from the solution. The gold amalgamates and can be recovered by retorting, like ordinary battery gold. Other methods involving the use of amalgams are the Pelatan- Clerici and Riecken processes. Charcoal was at one time extensively used as a precipitant in Victoria, Australia, The chemical action by which it pre- cipitates gold is somewhat obscure, but is supposed to depend on the occlusion of certain gases in the pores of the charcoal which act on and decompose the aurocyanide in the percolating solutions. An enormous volume is required for effective pre- cipitation. Other suggested methods of precipitation involve the use of chemicals which destroy, or at least enter into combination with, CHEMISTRY OF PRECIPITATION 123 the whole of the cyanogen present in the solution as simple cyanides. These cannot be considered to have passed the experi- mental stage. S. B. Christy and P. de Wilde have observed independently that soluble cuprous salts, added to a cyanide solution, produce a precipitate of white cuprous cyanide, which carries down with it practically the whole of the gold and silver; but this reaction only takes place in acid solutions, and would involve the neutralization of enormous quantities of alkali. Soluble silver salts give a precipitate of silver cyanide when added in sufficient excess, or to an acidulated solution; practically the whole of the gold is carried down; but this, of course, is only a laboratory method. Final Operations. In the clean-up, apart from mechanical operations, we have to consider three processes involving chemical reactions: (1) Acid treatment; (2) roasting; (3) smelting. Acid Treatment. The precipitate as obtained from the zinc-boxes invariably contains a considerable amount of metallic zinc in fine shreds or particles which pass through the sieves. This can be largely removed by treatment with sulphuric or hydrochloric acid, the zinc dissolving as sulphate or chloride. In some cases sodium bisulphate (NaHSO 4 ), which is a by- product in the manufacture of nitric acid and may be obtained at a low cost, has been used with advantage as a substitute for H 2 SO 4 , the reaction being: 2NaHSO 4 + Zn = NaaSO* + ZnSO 4 + H 2 . In the case of precipitates containing As, Sb, Se, and Te, these elements are partially evolved as the hydrides AsH 3 , SbH 3 , SeH 2 , and TeH 2 , on treatment with sulphuric or hydrochloric acids; these gases are very offensive and poisonous, and in the case of AsH 3 have several times been the cause of fatal accidents. In all cases, considerable amounts of HCy are evolved from the decomposition of the cyanogen compounds contained in the precipitate. Preliminary treatment with nitric acid has there- fore been recommended; this converts cyanides into cyanates, arsenic into arsenic acid, selenium and tellurium into selenious and tellurous acids. These compounds, being non-volatile, give rise to no noxious fumes, and are carried off in solution after the treatment. There may, however, be losses of silver and possibly also of gold. 124 THE CYANIDE HANDBOOK When zinc dust is used as a precipitant, some difficulty has been experienced (as at Mercur, Utah) , in treating the precipitate obtained, when sulphuric acid alone has been used. It packs very hard, and the particles of zinc appear in some cases to become coated with a layer of gold which protects them from further action. A mixture of sulphuric and nitric acids was found to be a more efficient solvent. Objection has been raised to the use of hydrochloric acid on the ground that it may liberate chlorine under certain circum- stances, thus involving loss of gold. It is, moreover, considerably more costly than sulphuric acid in most places. Whatever treatment be adopted, the solution of the zinc is always incomplete. It is possible that among the various alloys which may be formed there are some that are not readily attacked by acids; also in some cases protective coatings of PbSO 4 or CaS0 4 may be formed. Roasting. The main object of this operation is to oxidize the metallic zinc to ZnO, which in the subsequent smelting passes into the slag as silicate of zinc, whereas metallic zinc would pass into the bullion. Any other oxidizable metals are also oxidized, and cyanogen compounds either completely destroyed or con- verted into cyanates. Arsenic, antimony, selenium and tellu- rium are partially volatilized as oxides, but it is not possible to completely remove them in this way, and the fumes may mechanically carry off considerable amounts of gold and silver. A partial oxidation of the zinc is sometimes brought about in a preliminary operation by mixing the precipitate with niter and drying in shallow pans at a low heat. The roasting may be done in pans over an open fire, in a muffle, or in a reverberatory furnace. Smelting. This operation aims at producing bullion as free as possible from base metals and other impurities. As the pre- cipitate, after acid treatment and roasting, still contains silica, insoluble calcium salts, and considerable amounts of zinc, lead, and other base metals, the flux required must be varied according to its composition. The chief ingredients used as fluxes are: 1. Borax, to produce fusible borates of iron, aluminium, and other metals. 2. Silica (if not already present in sufficient quantity), to form silicate of zinc. CHEMISTRY OF PRECIPITATION 125 3. Carbonate of soda, to form a more fusible silicate, to flux any excess of silica, and in some cases to form fusible double sulphides. 4. Fluorspar, to flux infusible calcium sulphate, and generally to increase the fluidity of the slag. 5. Oxidizing agents (niter, MnO 2 , etc.), to oxidize any base- metal sulphides, selenides, tellurides, carbonaceous matter, etc. Metallic iron is also sometimes added, to reduce base-metal sulphides, etc. In some cases a matte is formed, which requires special treatment for reduction to bullion. The slags are some- times remelted with addition of fresh flux, to which litharge is added, whereby a large part of the value is collected as lead bullion, and the resulting low-grade slag may safely be rejected. Lead Smelting of Zinc Precipitate. In the Tavener process the zinc precipitate is smelted direct with litharge, siliceous material, and a reducing agent, the object being to form a lead alloy of the gold and silver and to flux off the zinc, etc., as fusible silicates in one operation. The litharge is reduced as follows: 2PbO + C = CO 2 + 2Pb. The lead bullion is then cupeled, and the litharge formed in this operation recovered for use in future smelting charges. Bullion is often further refined by remelting with fresh flux, and in some cases by oxidizing by means of a blast of air directed on the surface of the molten metal. Attempts have also been made to refine by injecting air or oxygen into the crucible con- taining the molten bullion. According to T. K. Rose, 1 the following is the order in which metals are oxidized under such conditions; the table given also Metal Oxide Heat of Combination with O Zinc ZnO 827 Iron Fe 2 O 3 637 Lead PbO 503 Nickel Ni 2 O 3 401 Copper Cu 2 O 372 Platinum PtO 179 Silver Ag 2 O 59 Gold AuO 44 Trans. I. M. M., XIV, p. 384. 126 THE CYANIDE HANDBOOK shows the heat of combination, the number given being the num- ber of grams of water which would be raised from to 100 C., by the combination of 16 grams of oxygen with the metal in question. SECTION VI MANUFACTURE OF CYANIDE THE various processes for the manufacture of cyanide may be classified according to the source from which the nitrogen is derived. The principal methods in use are: (a) Those in which refuse animal matter is used as the nitrogenous raw material, ferrocyanide being generally produced as an intermediate product. (6) Those in which atmospheric nitrogen is employed to form cyanide compounds, directly or indirectly. (c) Those in which ammonia or ammonium compounds form the nitrogenous raw material, including methods which utilize residues from gas-works. (a) PRODUCTION OF CYANIDES FROM REFUSE ANIMAL MATTER Until about the year 1890, this was the method almost uni- versally used. The raw materials required are: (1) Nitrogenous animal matter, such as horns, hoofs, dried blood, wool, woollen rags, hair, feathers, leather-clippings, etc. (2) An alkaline car- bonate, such as pearl-ash, soda-ash, etc. (3) Iron filings or borings. The alkali and the iron are first fused together at a moderate heat in an iron pan, or other suitable vessel, contained in a reverberatory furnace. The well-dried animal matter is then introduced in small quantities at a time and stirred in. The heat is then raised and the furnace closed so as to maintain a reducing atmosphere. The hard black mass which forms is then taken out and lixiviated with nearly boiling water. The crude ferro- cyanide containing sulphides, sulphates, carbonates and thio- cyanates, is crystallized out and purified by recrystallization. The ferrocyanide was formerly converted into cyanide by first dehydrating and then fusing, either alone or with alkaline carbonate : K4Fe(CN) 6 + K 2 CO 3 = 5KCN + KCNO + Fc + CO a 127 128 THE CYANIDE HANDBOOK The cyanide so formed is always contaminated with cyanates and carbonates, and generally with small amounts of other salts (sulphides, chlorides, thiocyanates, etc.). The procedure frequently adopted at present is to fuse with metallic sodium: K 4 FeCy 6 + 2Na = 4KCy + 2NaCy + Fe; thus yielding a mixture of potassium and sodium cyanides free from cyanates, etc. The presence of sulphides and thiocyanates in the product is due chiefly to the sulphur contained in the organic matter. These compounds are partially decomposed and removed by metallic iron during the fusion. When the cy- anide is made by direct fusion of ferrocyanide, the product con- tains carbide of iron, some nitrogen being given off in the process. Most of the volatile organic nitrogen is lost in the form of ammonia or nitrogen gas during the fusion for ferrocyanides, in the first stage of the process. (6) PRODUCTION OF CYANIDES FROM ATMOSPHERIC NITROGEN It was observed by Scheele that when nitrogen is passed over a mixture of K 2 CO 3 and charcoal heated to redness, a cyanide of potassium is formed. Many attempts were made throughout the nineteenth century to utilize this reaction for industrial purposes. One of the earliest was that of Possoz and Boissiere, who used a mixture of charcoal with 30 per cent, of potassium carbonate. This was kept at a red heat in fire-clay cylinders, through which a mixture of N and CO, produced by passing air over red-hot alkalized carbon, was allowed to pass for about 10 hours. The product was then lixiviated with water in presence of ferrous carbonate (spathic iron ore) to give a ferrocyanide. It was also observed at an early date (by Clark, in 1837), that cyanides are formed as an efflorescence in blast-furnaces, and that the gases of these furnaces contain cyanogen. Bunsen proposed a special blast-furnace for the production of cyanide, in which coke and potash in alternate layers were to be heated by a strong blast, the fused cyanide running off at the bottom of the furnace. It was found, however, that a very high tempera- ture was necessary, as the potassium compound must be reduced to metallic potassium before combination with atmospheric nitro- gen takes place. MANUFACTURE OF CYANIDE 129 Better results were obtained by substituting barium carbonate for K 2 CO 3 , as in the process of Margueritte and DeSourdeval (1861). Air (deoxygenated by hot carbon) was passed over a previously ignited mixture of BaCO 3 , iron-filings, coal tar, and sawdust, whereby barium cyanide is produced. This is converted into sodium cyanide by fusion with sodium carbonate. The BaC0 3 is first reduced to barium carbide (BaC 2 ) : BaCO 3 + 4C = BaC 2 + SCO. This then combines with nitrogen to form Ba(CN) 2 . It has been found, however, that only about 30 per cent, of the barium is converted to cyanide, the remainder forming barium cyanamide by a secondary reaction: Ba(CN) 2 = C + BaCN 2 . When calcium is substituted for barium in this process, prac- tically the whole is converted into calcium cyanamide: CaC 2 + N 2 = C + CaCN 2 . Calcium cyanamide is also formed in the electric resistance furnace by passing nitrogen over a mixture of lime and charcoal: CaO + 2C + N 2 = CaCN 2 + CO. By heating the product at a high temperature with a further quantity of carbon, with the addition of salt to prevent frothing and facilitate the reaction, the cyanamide is converted into cyanide as follows: CaCN2 + c = Ca(CN)2 . The crude mixture so formed has been used as a substitute for potassium cyanide under the name of "Cyankalium surrogat," and is equivalent in cyanogen contents to about 30 per cent. KCN. [See Erlwein and Frank, U. S. patent, 708,333.] An improved method more recently introduced is to convert the calcium cyanamide into sodium cyanide by the following series of reactions: (1) By leaching with water, a crystallizable, easily purified salt is obtained, known as dicy an-diamide : 2CaCN 2 + 4H 2 = (CN NH 2 ) 2 + 2Ca(OH) 2 . (2) This, when fused with sodium carbonate and carbon, is largely converted into sodium cyanide: (CN -NH 2 ) 2 + N a2 C0 3 + 2C = 2NaCN +NH 3 + N + H + 3CO !3Q THE CYANIDE HANDBOOK A portion of the dicyan-diamide sublimes and polymerizes ; this is recovered and re-treated with Na 2 CO 3 in a subsequent operation. The cyanamide and di- cyan-diamide are also utilized as sources of products valuable as manures, as they can be readily converted into ammonia, ammonium carbonate, urea, etc. Cyanides may also be formed by the action of metallic sodium and carbon on atmospheric nitrogen (Castner) ; but it is prefer- able to use ammonia as the source of nitrogen in this reaction (see below). (c) PRODUCTION OF CYANIDES FROM AMMONIA OR AMMONIUM COMPOUNDS When ammonia is passed over mixtures of heated alkali and carbon, only small quantities are converted into cyanide; better results are obtained by passing CO and NH 3 through a molten mixture of KOH and carbon, but even by this means much of the ammonia is dissociated into N and H, owing to the great heat which is necessary. It is supposed that potassamide is an intermediate product: KOH + NH 3 !=; K - NH 2 + H 2 O (reversible). K NH 2 + CO = KCN + H 2 O. In Castner's process (Brit, patents, 12,218, 12,219, of 1894), molten sodium is allowed to flow through a mass of heated coke while ammonia gas is passed upward, the reaction being 2NH 3 + 2C + 2Na = 2NaCN + 3H 2 . The reaction takes place at a much lower temperature than in the previous process with KOH, and the losses by dissociation of NH 3 and volatilization of the cyanide are consequently smaller. It probably takes place in two stages, forming sodamide as an intermediate product: (1) NH 3 + Na = NaNH 2 + H (at 300 to 400 C). (2) NaNH 2 + C = NaCN + H 2 (at dull red heat). In a modification of this method used by the Deutsche Gold and Silber Scheide-Anstalt, metallic sodium is melted with car- bonaceous matter in a crucible, and then ammonia is led in at 400 C. 600 C.; this forms disodium cyanamide: Na 2 + C + 2NH 3 = Na 2 CN 2 + 3H 2 . By then raising the temperature to 700-800 C., the excess of C. interacts, forming sodium cyanide : Na 2 CN 2 + C =2NaCN; the whole operation being conducted in the same crucible MANUFACTURE OF CYANIDE 131 Methods have also been proposed by J. Mactear, H. C. Wol- tereck, and others, in which cyanogen compounds are produced by the action of ammonia gas on gaseous carbon compounds at a high temperature. In Mactear's method (Brit, patent, No. 5037, of 1899) the reaction 2NH;} + co = NH 4 CN + H 2 O is supposed to take place in a closed chamber at 1800 to 2000 F., the products being condensed and absorbed in alkali hydrate and the ammonia liberated for reuse. Instead of CO, a mixture of CO with N and H (producer gas) may be used. In Woltereck's method (Brit, patent, No. 19,804, of 1902) "perfectly dry ammonia and a volatilized or gaseous carbon compound, also perfectly dry, are passed together with hydro- gen, in equal volumes, over a strongly heated catalytic agent, such as platinized pumice. One volume of NH 3 and two volumes of 'water-gas' (CO -f H 2 ) make a convenient mixture. The HCN produced is absorbed in an alkaline solution." Cyanide has also been made from the trimethylamine (CH 3 ) 3 N obtained by the distillation of beet-root molasses at a high tem- perature. This, at a red heat, decomposes, giving NH 3 , HCN, and H. Another source of cyanogen compounds is the crude illumi- nating gas from the distillation of coal. In Knublauch's method (Brit, patent, No. 15,164, of 1887) the gas is passed through a solution of an alkali or alkaline earth containing ferrous hydrate in suspension. The gas carries with it ammonium cyanide and thiocyanate, which are absorbed by the mixture and converted into ferrocyanide. In Rowland's process (U. S. patent, No. 465,600, of 1891) the gas is passed through a solution of an iron salt, thus forming ammonium ferrocyanide. This is converted into the calcium salt by boiling with lime. The calcium ferrocyanide may then be con- verted into the required alkali ferrocyanide by decomposing with an alkaline carbonate. Bueb's process (Brit, patent, No. 9075, of 1898) is a modi- fication of the above, in which the cyanogen is separated in the form of an insoluble double compound by using a concentrated iron solution (FeSO 4 ) ; the reactions said to take place are : FeSO 4 + HaS + 2NH 3 = FeS 2FeS + 6NH 4 CN = (NH 4 ) 2 Fe 2 (CN) 6 132 THE CYANIDE HANDBOOK The insoluble product, known as "cyanide mud," is then treated to obtain marketable cyanogen compounds. Many other modifications have been suggested. Cyanides may also be obtained by desulphurizing thiocyanates by means of iron, or by zinc and carbon. PART III PREPARATORY TREATMENT OF ORE FOR CYANIDING ONLY a brief outline can be given here of the mechanical principles and processes which form a necessary preliminary to cyanide treatment or are involved in the carrying out of the process itself. Details of the construction and mode of action of the machines used will be found in any text-book of metallurgy, and much information on these points is usually given in the cata- logues supplied by the makers. Whatever metallurgical process is to be adopted, it is evident that the ore as taken from the mine will require to undergo a certain amount of crushing before its valuable content can be extracted, so that much of what follows does not specifically concern the cyanide process. We shall endeavor, however, to point out more particularly those considerations in the preparatory treatment which have a direct bearing on subsequent cyanide work. SECTION I CRUSHING MACHINERY (A) ROCK-BREAKERS THE reduction of an ore to particles of a size suitable for any metallurgical process is almost always carried out in at least two stages, for which different types of machines are required. These may be distinguished, somewhat arbitrarily, as Crushing and Grinding. In the first, or crushing, stage, the ore is broken by a succession of blows delivered at intervals by a hard moving body, such as the jaw of a rock-breaker or the shoe of a stamp; the rock being crushed against another (generally stationary) hard body, as the die of the stamp battery. In the grinding process the comparatively coarse lumps produced by crushing are pressed more or less continuously between two hard surfaces to reduce them to a smaller size, although in many machines (e.g., Gates crusher, Ball mill, Tube mill, etc.), both actions go on to some extent simultaneously. For economical work, it may be stated as a nearly universal rule that crushing must precede grinding, and further, that separate machines should be used for different stages of crushing. The first, or coarse, crushing is carried out in a machine known as a rock-breaker, the object of which is to break the ore as it comes from the mine, which 'may be in pieces the size of a man's head or larger, to lumps of J in. to J in. diameter, suitable for further crushing in stamps or other fine crushers. The rock-breakers in general use are of two types: (1) those with a fixed and movable jaw, the opening between which is regulated as required, such as the Blake, Blake-Marsden, and Dodge crushers; (2) those with a gyrating vertical shaft, work- ing against plates on a stationary circular shell surrounding the shaft, as in the Gates and Comet crushers. In the first type (jaw crushers) the fixed and movable jaws are inclined at an angle to one another, the smaller opening being 135 13 6 THE CYANIDE. HANDBOOK at the bottom. The ore is fed in at the top, between the jaws. When the machine is at work, the movable jaw rapidly advances and recedes about half an inch, so that the rock gradually slips down and is crushed between the jaws, which are faced with removable plates of hard steel. The Dodge crusher is said to be suitable for somewhat finer crushing than the Blake. (Fig. 1.) In the second type (gyratory crushers), the top end of the shaft is usually pivoted, while the bottom describes a circle. A conical piece of chilled iron or steel (breaking head) is fitted to the shaft, the surface of which is inclined at an angle to the crush- ing faces fixed to the shell or hopper surrounding the shaft. As the shaft gyrates, the surface of the breaking head alternately advances toward and recedes from the fixed crushing faces, thus performing the same functions as the movable jaw in the jaw crushers. The crushing surfaces are generally corrugated, and are composed of chilled iron, chrome steel, manganese steel, or some similar material combining hardness and toughness. (Fig. 2.) In a typical case, a Blake rock-breaker will crush 300 tons of rock in 24 hours to 1 J-in. size, running at 250 r.p.m. and requiring 14 h.p. 1 The accompanying illustration, Fig. 1, shows a slightly modified form of Blake crusher manufactured by Hadfield's Steel Foundry Co., of Sheffield, England. The rock is broken between a fixed jaw plate and a movable jaw plate actuated by a powerful toggle movement communicated to it from the driving pulley through an eccentric shaft and pitman. Fig. 2 illustrates the Gates rock and ore breaker (Style K), one of the gyratory type manufactured by the Allis Chalmers Company, of Milwaukee. The main shaft (25) is suspended in the upper part from spider arms (44) that span the top of the opening, and gyrates without revolving, in such a way that the ore is crushed between the concave shell (19) and a removable block or "head" (18). This head has a circular and rolling movement, and approaches successively every point in the interior of the throat, owing to the movement of the lower end of the axis of the shaft around a small circle. The rock is broken rather than crushed. It falls on an inclined diaphragm which protects the eccentric and gear, and thence passes to the discharge spout (32). The main shaft is adjustable, 1 Rose, " Metallurgy of Gold," 4th edition, p. 103. CRUSHING MACHINERY 137 and can be raised or lowered to regulate the size of product required. FIG. 1. Modified Blake Crusher. (B) STAMPS The finer crushing of gold ores is almost always performed by means of stamps, except in cases (1) where dry crushing is desirable, or (2) where the ore is unusually soft. In modern prac- tice, crushing with stamps is frequently followed by grinding in pans or tube mills, as described below. The stamps consist of sets (usually five in each set) of heavy iron hammers, 700 to 1250 Ib. in weight. Each stamp has a vertical iron rod (stem), provided at the lower end with a cylindrical block of cast iron (head), into which the stem fits. The lower end of the head has 138 THE CYANIDE HANDBOOK a socket to receive a replaceable block of hard iron (shoe) which forms the crushing surface. The stamps are alternately raised and dropped by means of a horizontal cam-shaft. As the shaft FIG. 2. Gates Rock and Ore Breaker. revolves, the cams (set at different angles on the shaft) strike against a block (tappet) fixed to the upper part of the stem, so that the stamp is raised a few inches as the cam revolves, and CRUSHING MACHINERY 139 drops by its own weight when the point of the cam passes its highest position. The cams are set in such a way on the shaft that each stamp is lifted and dropped in succession, in a par- FIG. 3. Elevation of 10-stamp Battery, showing Mortar Box, etc. [Eraser and Chalmers.] ticular order. As the stamp falls, the lower surface of the shoe strikes the layer of ore resting on a hard iron or steel block (die), which thus acts as an anvil. The dies corresponding to each set 140 THE CYANIDE HANDBOOK of five stamps are set in an oblong cast-iron box (mortar-box). (See Fig. 7.) The ore is fed in on one side of the row of stamps, and discharged on the other side through a screen fitted to an opening in the mortar-box, the lower edge of which is a few inches above the top of the die. (Figs. 3, 4, 5, 6.) CRUSHING MACHINERY 141 After leaving the rock-breakers and before passing to the stamps or other mill, the ore is generally sifted, through a coarse screen consisting of parallel steel bars, known as a ''grizzly" (see Fig. 16), the oversize returning to rock-breakers. When stamps are used, as is generally the case, for wet crushing, a stream of water is fed into the mortar-box along with the ore, and a further supply of water is usually added, outside the battery, FIG. 5. New Blanton Cam. [Fraser and Chalmers.] FIG. 6. Cam Shaft, showing 5 Cams in Position. [Fras3r and Chalmers.] to wash away the ore as it passes through the screens. The fineness to which the ore is crushed, which is the principal point of interest with regard to subsequent cyanide treatment, depends chiefly on the (1) size of openings in screen used in mortar-box; (2) relative amount of ore and water fed in; (3) hight of discharge above dies; (4) hight of drop of stamps; (5) rate at which ore is fed into mortar-box. Where the object is to treat as much of the product as possible by percolation, it is of course desirable to crush the ore only so fine as may be necessary to render a sufficient percentage of the 142 THE CYANIDE HANDBOOK values accessible to the solvent; this will be attained by (1) using a somewhat coarse screen; (2) using much water in proportion to ore, so as to wash out the crushed ore as soon as it is reduced to the required size; (3) having the discharge as close as possible to the top of the dies, to avoid crushed ore being repeatedly liu. /. Section of Mortar-box. [Fraser and Chalmers.] thrown back under the stamps; (4) using heavy stamps running at a high speed, with the least hight of drop possible in practice; this results in a maximum quantity of material being crushed for a given expenditure of power; (5) keeping the rate of feed low, so as to avoid banking of the ore against the screen, which would lead to unnecessary recrushing of the same material. In many cases, however, it is more advantageous to crush finer, even although the output of ore per stamp may be reduced, CRUSHING MACHINERY 143 as the extraction, both by cyanide and by amalgamation, is usually greater the finer the state of division to which the ore particles are reduced. Frequently the ore contains two constituents of different degrees of hardness, quartz and pyrites, one of which carries practically the whole of the value. If, as is the case on the Rand, the quartz is barren, while the whole of the gold is associated with or imbedded in the pyrites, it would seem that the best method would be to crush coarsely at first, then separate the pyrites by some form of concentration or other mechanical means, and regrind the pyritic portion alone as finely as pos- sible. In general, it may be said that when the ore is to be sub- sequently treated by cyanide, less trouble and loss will be caused by crushing too fine than by crushing too coarse, since in the coarse particles the gold or silver is absolutely inaccessible to the solvent; whereas the difficulties in the way of the mechanical handling of slimes in order to extract their dissolved values have been greatly reduced if not entirely overcome by modern methods. Another point of great importance for cyanide work is uniformity of -the mill product, both as regards fineness of the ore particles, and as to proportion of water. The whole system of treatment must be regulated in accordance with these factors, and any unforeseen variation may give rise to bad extractions and heavy losses. Where " spitzlutten " (see below) are used for separation of sands and slimes, they are arranged to give the required product by regulating the size of outlet, etc., and any variation in the ratio of ore to water will result either in slime passing into the sand tanks or in fine sand being carried away with the slime. It is obvious that the whole metallurgical treatment should be re- garded as one problem, and the preliminary operations of crush- ing and grinding should be made subservient to the extraction processes which are to follow, and not carried out merely with a view to passing through the maximum quantity of ore in a given time. As an example of typical stamp battery practice under modern conditions, the following is given, on the authority of T. Roskelley, 1 as having been used in the crushing of banket ore at the Robin- son Deep G. M. Co.'s battery: i " Proc. Chem., Met. and Min. Soc. of South Africa," III, 338. 144 THE CYANIDE HANDBOOK Standard Screen Used Hight of Drop: Inches Speed: Drops per Minute Water Used per Ton Crushed Tons Crushed per Stamp per 24 Hours Slimes Produced Per Cent. 600 8i 96 6 5.45 25.4 700 81 96 6 5.28 29.0 900 81 96 61 5.07 31.8 1000 71 100 7 5.06 35.3 1200 74 100 74 4.95 35.0 Fall of Plates = 1 fa in. per foot. Hight of discharge = 2 1 in. Other authorities (e.g., F. Alexander) recommend a greater hight of discharge, such as 6 or 8 in., which, with other conditions remaining the same, would result in a finer product and dimin- ished output. A full account of this subject will be found in the following papers, to which reference should be made for further details as to mill practice and its effect on cyanide treatment: "The Stamp Milling of Gold Ores in its relation to Cyaniding," by E. H. Johnson, in " Proc. Chem., Met. and Min. Soc. of South Africa," II, 176 (1897). " Notes on the Common Practice of Quartz Milling on the Rand," by F. Alexander. Ibid., Ill, 298 (1903). Much useful information as to the effectiveness of the crush- ing with a view to cyanide treatment may be obtained by sifting a sample of the product through a number of screens of varying mesh, and assaying the portions passing through successive screens, similar tests being also made on the residues after cyanide treatment. When the battery pulp, before cyanide treatment, is examined in this way, it is generally found that one portion of the crushed ore is richer than another, and frequently that the finer products are the richer. With the residues after cyanide, however, it is almost invariably the case that the coarser portions are the richer, showing that finer crushing of this part of the ore would result in higher extraction. 1 Finer crushing usually results also in improved amalgamation, not only because more gold is brought into actual contact with the mercury, but because me- 1 Tests on these lines made weekly for a number of years at the Redjang Lebong mine, Sumatra, showed scarcely an exception to this rule, both as regards gold and silver values. CRUSHING MACHINERY 145 chanical losses of amalgam, due to the passage of coarse material over the surface of the plates, are much lessened. Any amalgam carried off in this way is practically lost, as but little dissolves in cyanide during the time of treatment usually given. This matter will be further discussed under amalgamation. In this connection the following remarks by P. S. Tavener l are very much to the point: " Would it not be better to take the pulp leaving the mortar- boxes and pass it direct to a grinding machine, or to separate, by means of classifiers, the particles requiring regrinding before amalgamation or afterwards? It would appear to be better to do so before, for if done afterward the reground portion would have to pass over the plates for a second time. Is it not likely that it is the coarser particles that help to scour the plates?" As the stamp mill is not adapted for crushing effectively and economically to a very high degree of fineness, it is desirable to pass a part at least of the product from the battery through some grinding machine; hence 'the modern practice is to use the stamp battery for comparatively coarse crushing, supplementing it by one or other of the grinding machines described in the following section. i " Proc. Chem., Met. and Min. Soc. of South Africa," III, 329. SECTION II GRINDING MACHINERY THE type of machine to be used for grinding will depend largely on the hardness and general character of the ore and on the treat- ment which it has previously undergone. In some cases the ore may pass direct from rock-breakers to the grinding machines, but this can in general only be done with comparatively soft material, or in cases where the ore is to be crushed dry. The grinding machines principally used in connection with cyaniding are rolls, grinding pans, ball mills, and tube mills. (A) ROLLS These consist of two parallel cylinders with their axes placed horizontally. (See Fig. 8.) The tires, or crushing faces, are of forged steel or chilled iron, usually smooth, and are replaceable when worn. The rolls are mounted so that they revolve at a dis- tance of f to -gV in. apart, according to the fineness of the product required. The usual sizes are 12 to 16 in. across the face, and 22 to 40 in. diameter. The product obtained varies from -| in. size to about 70-mesh, depending on the distance between the crush- ing faces of the two rolls. They are used almost exclusively for dry crushing, the ore being fed between the revolving rolls by means of a hopper. After leaving the rock-breakers and before reaching the rolls, the ore is passed over screens, and the over- size, too large to be fed direct to the rolls, is returned to the rock- breakers. In some cases a succession of rolls is used to reduce the ore to varying degrees of fineness, with suitable screens between each pair. The speed may be from 60 to 150 r.p.m., the velocity at the circumference being (say) 750 to 1000 feet per minute. In some cases the two rolls are made to revolve at different velocities. A detailed account of modern practice with rolls will be found in a paper by P. Argall, "Sampling and Dry Crushing in Colorado." 1 (See Fig. 9.) 1 Trans. I. M. M. X, 234-273. 146 GRINDING MACHINERY 147 In general, H may be stated that rolls are the most suitable machines for dry crushing where the object is to produce as little fine dust as possible; when set to a particular opening, they pro- FIG. 8. Smooth Crushing Rolls, manufactured by Hadfield's Steel Foundry Co., Sheffield. The frames are of toughened cast steel and the shells (crushing surfaces of the rolls) are of "Era" manganese steel or, when used for fine crushing, of best quality toughened cast steel. The rolls are driven by means of belts running on two heavy fly-wheel pulleys. duce a minimum of material crushed to a greater degree of fine- ness. They are not usually suitable for very coarse or very fine crushing; the material fed to the rolls should have been pre- viously reduced at least to 2-in. size by rock-breakers, and if very fine grinding is necessary, the product from the last set of rolls should be passed to grinding pans or tube mills. Screen analyses of the product obtained by rolls on Cripple Creek ore are given by Argall (loc. cit., pp. 269, 270), as follows: I. Size of opening = .0212 in. (andesitic and granite ores). Remaining on Ore n 40-mesh ' 40 to 60-mesh . (a) Per cent. 21 17 (ft) Per cent. 13 20 W Per cent. 31 16 ' 60 to 120-mesh . 41 54 44 ' 120 to 200-mesh 15 12 6 Passinsr 200-mesh 6 1 3 148 THE CYANIDE HANDBOOK II. Size of opening = .02 in. (base ore). Remaining on n 50-mesh (<*) Per cent. 31.5 () Per cent. 32.5 ' 50 to 100-mesh . ' 100 to 200-mesh 26 16.5 26 16.5 Passing 200-mesh . 26 25 It is seldom advantageous to use rolls for crushing finer than about 30-mesh. When ore requires to be roasted previous to FIG. 9. Argall Rolls, furnished by Fraser and Chalmers. These are belt- driven, and one roll is fixed while the other is movable. The bearings of the movable roll are connected by a heavy cast-iron yoke, insuring that the face of this roll will always remain parallel to the face of the fixed roll. cyanide treatment, it is very generally crushed by means of rolls, as it is desirable to avoid producing an excessive quantity of fine material. In some cases the whole or part of the roasted product is ground finer before undergoing treatment. (B) GRINDING MILLS These are of many types. Those used in the process of pan- amalgamation consist essentially of a heavy shoe or muller, revolving about a vertical axis, the ore being ground between the under surface of this muller and the bottom of the pan. GRINDING MACHINERY 149 Descriptions of these machines will be found in any work on the metallurgy of silver. 1 The bottom of the pan is lined with replace- able cast-iron " dies." The muller revolves at about 70 r.p.m., and the ore, previously crushed to a sufficient degree of fineness, FIG. 10. Evans- Waddell Chilian Mill, manufactured by Fraser and Chal- mers. This machine consists of three crushing rollers mounted on hori- zontal arms, connected with a central vertical driving-shaft, the latter being driven by means of a horizontal pinion-shaft and bevel-gear at- tached to the lower part of the vertical shaft. The mortar or pan is lined inside with hard iron wearing-plates, and the die-ring and roller shells are of the best quality forged steel. There are six screen openings for discharging the crushed material. is fed in with sufficient water to make a thick pulp; after the ore has been ground for some hours, mercury is added and the grind- ing is continued, sometimes with the addition of other chemicals. Other forms of grinding machines which are in use, either with or without the addition of mercury, may here be mentioned. i See also Rose, loc. cit., pp. 169-173. 150 THE CYANIDE HANDBOOK Among the best known are the Arrastra, Chilian mill, Griffin mill, Bryan mill, and Huntington mill. These machines, although of. widely different construction in detail, have certain features in common. In all of them the grinding part is made to revolve about a central vertical shaft, and acts against a circular or annular surface forming the bottom or side of the machine. The Arrastra, probably the oldest and most primitive machine used in ore treatment, is still employed to a large extent in some parts of America. It consists of a bed or pavement of hard stone, with a wooden shaft set in the center, bearing a horizontal revol- ving arm or arms, to which heavy stones are attached; the ore, previously broken in small pieces, is ground between the revolving stone and the pavement. It is capable of very fine grinding, but the output is of course very small. The Chilian mill consists of large wheels or rollers, which may be as much as 8 ft. in diameter and 6 in. wide, revolving on horizontal axes in an annular trough or mortar. (Fig. 10.) The Bryan mill is of similar construction, consisting of three rollers mounted on horizontal axles, which revolve about the central axis of the machine. The rollers run in an annular mortar lined with segmental steel dies; the central shaft makes about 30 r.p.m., the pulp at the periphery running round at over 300 feet per minute. The capacity of a 4-foot mill is about 15 to 20 tons of quartz ore per day through a 40-mesh screen. 1 These mills are suitable for both hard and soft ores. The Huntington mill is somewhat similar, but the rollers are mounted on vertical axles suspended from a circular horizontal carrier revolving about a central vertical axis. The grinding is due to the pressure of these rollers against the sides of the mill, which are lined with an annular replaceable die of steel or cast iron. Above this is an annular screen for the discharge of the crushed ore. A hopper is also placed at the side of the mill for feeding in the ore, which should be previously crushed to nut size in a rock-breaker. The rollers are suspended about one inch from the bottom of the pan, and being free to swing radially as well as to revolve on their own axes, they are driven by cen- trifugal force against the ring-die when the carrier revolves. The speed is 45 to 90 r.p.m., usual diameter 5 ft. This machine is particularly suitable for soft ores, which it crushes much more 1 Rose, loc. cit., p. 165. GRINDING MACHINERY 151 effectively than stamps, and owing to the freer discharge it pro- duces less slime than the latter. Enough water is added in the feed hopper to make a thick pulp, and in cases where the mill is FIG. 11. Improved Huntington Mill manufactured by Fraser and Chalmers. It is claimed that this mill can be run with less expenditure of power per ton of ore crushed than any other similar mill. The ore is discharged through the screen as soon as it is crushed fine enough, thus avoiding the production of slime and giving a product very suitable for concen- tration. to be used as an amalgamating machine, from 17 to 25 Ibs. of mercury are added, forming a layer at the bottom just beneath the rollers. The cost is about two-thirds that of stamps for the same crushing capacity; less power is required per ton crushed, 1 and the cost of repairs and renewals is less. There is also less floured mercury than is given by stamps with inside amalgama- tion. 2 (Figs. 11 and 12.) i Tho power required is : For mill 3^ ft. diam., speed 90 r.p.m., 4 h.p. 5 " " " 70 " 6 " 6 " " " 55 " 8 " 2/feid., p. 161. 152 THE CYANIDE HANDBOOK (C) BALL MILLS These consist essentially of cylindrical revolving drums con- taining a number of steel balls of various sizes. The interior of the drum is lined with steel plates, arranged so that a portion of each projects inward, forming a shelf which has the effect of FIG. 12. Sectional view of a Huntington Mill as made by Fraser and Chalmers. raising the balls and dropping them during the rotation of the drum. The other portion of the lining plates is perforated. The ore is fed in through a hopper at one end of the drum, and is crushed by the impact of the balls against the grinding surfaces of the lining and against each other. The action is in fact a combination of crushing and grinding. The crushed ore passes through the perforations to a set of inner and outer sieves con- GRINDING MACHINERY 153 centric with the circumference of the drum. The first or inner screens consist of perforated steel plates which serve to protect the outer wire screens. The oversize returns automatically to the interior of the drum through openings between the lining plates. The material passing the sieves goes through a hopper forming part of an outer casing of sheet iron, surrounding the revolving drum, which discharges the crushed ore at the bottom. (Fig. 13.) FIG. 13. Ball Mill, No. 8 size. [From photograph furnished by Cyanide Plant Supply Co.] These machines may be used either for wet or dry crushing, but are more commonly used for the latter purpose. Where the object is to crush a large percentage of the product to a tolerably high degree of fineness, ball mills are more effective than stamps or rolls. Julian and Smart 1 state that ball mills of the Krupp- Grusonwerk type are generally preferable to rolls or stamps for dry crushing, and recommend crushing coarsely in the first in- stance, say 12 to 20 mesh, and passing the product through a second machine for the finer grinding. The following screen- 1 " Cyaniding Gold and Silver Ores," 2d edition., p 47. 154 THE CYANIDE HANDBOOK analysis given by the same authorities (loc. cit., p. 199) will serve to illustrate the nature of the product obtained by crushing in ball mills. An ore from the Kalgurli mines, West Australia, fed in 2^ in. size to a No. 5 Krupp ball mill with 35-mesh screens, gave a product : 35 to 80-mesh 30 per cent. 80 to 120-mesh 30 " Passing 120-mesh 40 " Another example is given by Dr. Simon ("Trans. I. M. M." X, 290), from No. 4 Krupp ball mills, crushing dry with 30-mesh screens, on ore from Norseman Mine, West Australia: Remaining on 30-mesh 0.6 per cent 30to40-mesh .'. . 16.1 40 to 60-mesh 17.1 60 to 80-mesh 25.2 Passing 80-mesh 41.0 " These may be compared with the figures given for the product obtained by crushing with rolls, given by P. Argall for Cripple Creek ore (see above), where the percentage passing 120-mesh was from 9 to 21 per cent. Where ball mills are used for wet crushing, the water is led into the interior of the drum and di- rected in a spray against the sieves, the discharge hopper being arranged as a spitzkasten, for separating sands and slimes. The power required is 8 or 9 h.p. for an output of 200-300 Ib. per h.p.-hour, crushing to 40-mesh (Argall, loc. cit., p. 296). Ball mills are also extensively used for crushing lime, cement, slag, and other materials, and sometimes for mixing the fluxes and precipitate from the zinc-boxes previous to smelting. (D) TUBE MILLS These machines, which have been very extensively introduced of late years for the fine grinding of ore before cyanide treatment, consist of sheet steel drums 13 to 22 ft. in length, having an in- ternal diameter of 3 to 5 ft. They are lined internally with re- placeable blocks of some hard material, such as hardened steel or "silex," and the remaining space is generally rather more than half filled with rounded flint pebbles. The cylinder is mounted on trunnions, with its axis slightly inclined downwards from the feed to the discharge end. The ore or battery pulp, GRINDING MACHINERY 155 reduced to a suitable consistency by removal of superfluous water by means of spitz kasten or otherwise, is fed in by means of a hopper through an opening in one of the ends of the drum. Dur- ing the rotation of the machine the pebbles cling to the lining FIG. 14. Davidsen Wet Crushing Tube Mill. Feed end. for some distance up the ascending side, and then drop to the bottom of the drum, crushing the ore by their impact. The crushed ore is discharged through an opening at the opposite end of the drum, and usually passes to spitzlutten, from which the coarser product is returned to the feed end of the mill and reground. (Figs. 14 atid 15.) To obtain the maximum efficiency from tube mills, great care is necessary in regulating the speed, the rate of feeding, consistency c-f pulp, and charge of pebbles. According to W. R. Dowling, 1 the tube mills of the Robinson Deep G. M. Co. are run at a speed of 29 r.p.m.; about 230 tons of sand are fed to each mill per 24 hours; the pulp should be as thick as possible (in the case cited it contains about 40 per cent, of water to 60 per cent, of solids) ; the load of pebbles is rather more than half the inter- nal volume of the mill. The nature of the liners is also a matter of considerable im- portance. It appears to have been generally found that "silex" linings give more satisfactory and economical results than man- ganese steel. The silex consists of hard flint cut in blocks about 2J in. thick, and secured to the inside of the shell by a setting of Portland cement. Blocks of 4 in. have been used and are con- siderably more economical, as they last longer. The pebbles are 1 " Tube Mill Practice," in Journ. Chem., Met. and Min. Soc. of South Africa, VI, 308 (April, 1906). 156 THE CYANIDE HANDBOOK 2 to 3 in. in diameter and are mostly imported from the Danish coast, but in South Africa the practice has been introduced of using pieces of the banket ore itself as the crushing medium; lumps 3 to 6 in. in diameter are quickly reduced by the action of the mill to the form of a round pebble. The size of pebbles required seems to depend on the size of the product to be crushed; the finer the particles entering the mill the smaller the pebbles needed. FIG. 15. Davidsen Wet Crushing Tube Mill. Discharge end, nished by the Cyanide Plant Supply Co.] [Fur- The feed is regulated by reducing or enlarging the discharge opening of the spitz kasten used for dewateririg the pulp before it enters the tube mill. According to Bowling (toe. ciL, p. 310), "the best indication that a tube mill is being given a sufficient feed is the continual presence of a few inches of sand at the bottom of the dewatering spitzkasten." The action of the mill is thus described: "The thicker the pulp, the better the grinding and the less the wear on both liners and pebbles, since the sluggish pulp clings to both pebbles and liners and forms a layer to take the blow of the falling and the abrasion of the rolling pebbles. The clinging action of the thick pulp can be seen at the discharge end of the mill, and this is a ready means of telling whether the feed is of approximately the right consistency." GRINDING MACHINERY 157 The crushing efficiency of a tube mill is generally measured by comparing the percentages of coarse and fine sand in the pulp fed to the mill and in the product delivered by it. A somewhat arbitrary rule is to take the sum of the decrease in material remaining on 60-mesh and the increase of material passing 90- mesh as representing the work done by the mill. The nature of the products entering and leaving the tube mill, in ordinary working conditions, is shown by the following table given by Bowling (loc. cit., p. 313) at the Robinson Deep. Grade PULP LEAVING BATTERY (Entering Tube Mill) PULP ENTERING CYANIDE WORKS (From Tube Mill) WORK DONE Per Cent, of Total Wt. Tons Per Cent, of Total Wt. Tons Per Cent. Tons Remaining on 60-mesh 60 to 90-mesh Passing 90-mesh 28.66 17.43 53.91 8,599 5,230 16,175 5.28 16.50 78.22 1,584 4,951 23,469 23.38 7,015 24.31 7,294 100.00 30,004 100.40 30,004 47.69 14,309 30-mesh 30 to 40-mesh 40 to 60-mesh ' 60 to 100-mesh 100 to 150-mesh Before Tube Milling Per cent. 5.32 9.77 15.94 13.96 12.29 After Tube Milling Per cent. 0.03 0.12 1.13 7.43 18.42 150-mesh 42.72 72.87 A more detailed analysis of the gradings at the Waihi mine, New Zealand, is given by E. G. Banks: x Remaining on Passing A comparative test between crushing in a tube mill with gold quartz (Rand banket ore) and Danish pebbles is given by K. L. Graham, 2 showing results slightly in favor of the quartz. In both cases the mills were lined with silex blocks, 6X6X4 in., laid on edge. The wear on the linings was certainly not greater when quartz was used. The results of a month's work (January, 1907) are given as follows: i " Trans. A. I. M. E," XIII, p. 63 (Jan., 1907). a " Journ. Chem., Met. and Min. Soc. of South Africa," VII, 318 (April, 1907). 158 THE CYANIDE HANDBOOK No. 1 MILL No. 2 MILL Quartz fed 171.5 tons. Pebbles 10 tons. Entering: Leaving: Entering: Leaving: Product: per cent. per cent. per cent. per cent. Remaining on 60-mesh 68.68 9.73 68.68 11.32 60 to 90-mesh 19.57 26.28 19.57 27.24 Passing 90-mesh 11.75 63.99 11.75 61.44 The lumps of ore, up to 4^ in. diameter, are introduced by means of a feeder consisting of a hopper leading to a bent cast-iron tube, 5 in. in diameter at the narrowest part, and passing into the hollow trunnion at the feed end of the mill. The pulp from the spitzkasten is fed into the same tube near the bottom, oppo- site the outlet of the bend (where it enters the trunnion). At Waihi, New Zealand, liners have been used consisting of cast-iron frames divided into square compartments about 3^ in. deep; these are filled in with hard quartz or quartzite, embedded in Portland cement mixed with ground flint fragments. These "honeycomb liners" appear to be as efficient as silex or quartzite blocks, and cost much less. At the Waihi, 4 mills, 22 ft. in length, are run at 27.5 r.p.m. Each mill has a load of 5.5 tons of flints and requires 50 h.p. to operate it. SECTION III MECHANICAL HANDLING OF MATERIAL FOR CYANIDE TREATMENT DURING the early years of the cyanide process it was applied almost exclusively to the treatment of accumulations of already crushed ore, which had previously undergone treatment by amal- gamation or other metallurgical process. Such material was usually collected in heaps or had been allowed to settle in pits or natural hollows; in either case it had been exposed to atmos- pheric influences frequently for long periods of time. Later, when these old accumulations had been largely worked out, the process was adapted for the immediate treatment of material as delivered from the crushing plant, generally after plate amal- gamation. Further developments of the process are its appli- cation in certain cases to the direct treatment of ore from the mine, without other extraction process, and in the treatment of special classes of material, such as concentrates, and by-products of other processes. The methods adopted in handling the material naturally differ considerably in these various applications of the process. We shall consider them in the above order. (A) HANDLING OF OLD ACCUMULATIONS Effect of Natural Settlement. In all cases the crushed ore consists of particles differing in size and in specific gravity. When these are carried along in a stream of water of diminishing velocity, the particles tend to settle, the larger and heavier particles coming to rest first, while the smaller and lighter particles are carried farther. As a consequence of this, the material is roughly sorted into a layer of grains diminishing in size as they are deposited farther from the crushing plant. As, however, the point of inflow and the direction of the stream is varied from time to time, the final effect is the production of layers of coarse and fine material, superimposed and intermixed in a most irregular way. When 159 160 THE CYANIDE HANDBOOK any attempt is made to treat such material by the ordinary percolation method, difficulties at once arise. If the material be dug out of the pits and dumped indiscriminately into the leaching vats, the charge will consist of masses of particles of varying degrees of fineness; the solution, instead of flowing uni- formly through the whole charge, will form channels wherever there is least resistance, that is, in those parts consisting of coarse particles. The very fine material (slimes) forms layers which are practically impervious to liquids; when these slime layers are broken up and dumped into the vats together with the sand, the lumps of slime remain practically untreated. If, however, the slime lumps have become partly dry before charging, they may absorb solution, which remains in them when the tank is dis- charged ; if the solution so absorbed contains gold extracted from other portions, this may lead to considerable losses. Difficulties in Treatment of Unsized Tailings. Moreover, the percolation of a charge consisting of particles of various sizes is always more difficult than when the particles are uniform, even if small, because in the first case the interstices of the larger particles are partially filled with the smaller particles; whereas in the latter case the interstices are chiefly filled with air or liquid. It therefore becomes desirable to separate, at least roughly, those portions which can readily be treated by percolation from those which cannot be so treated. It is generally impossible, in digging the material from the dumps or pits, to separate the sandy and slimy portions sufficiently, as the slime occurs usually in thin layers intermixed with layers of sand, but one of the first suggestions was to separate the slime-lumps by passing the whole of the tailings over a coarse screen or grizzly. (Fig. 16.) When the quantity of slime is not too large, it can in many cases be treated by partially drying the material sifted out as above de- scribed, and mixing thoroughly with a large proportion of sand, sifting again to break up any lumps. Experiments, however, show that the extraction from the sands diminishes in proportion to the amount of slime mixed with them. 1 Attempts have sometimes been made to mix the sand and slime in situ, by plowing and harrowing the surface of the tail- ings in the dumps or pits, and allowing them to dry as much as possible by exposure before charging into the vats. The material 1 Julian and Smart, loc. cit., p. 60. MECHANICAL HANDLING OF MATERIAL 161 FIG. 16. is then passed through a screen or disintegrator to obtain a uni- form mixture for treatment. (Figs. 17 and 18.) The Callow Screen (Fig. 18) consists of a perforated endless traveling belt of wire cloth, mounted on revolving rollers. The ore and water are fed on from above. The belt travels continu- ously at 25 to 125 feet per minute. The coarsest and largest par- ticles strike the screen ahead of the smaller ones, thus leaving a free space for the passage of particles small enough to go through the meshes of the belt. The undersize is discharged into a hopper beneath the upper surface of the traveling belt. In treating old accumulations much trouble is sometimes caused by the presence of masses of vegetable matter, such as grass-roots, etc., derived from the surfaces on which the tailings were originally deposited; such foreign matter, especially if par- tially decomposed, causes serious difficulty in the treatment, 162 THE CYANIDE HANDBOOK MECHANICAL HANDLING OF MATERIAL 163 not only by preventing uniform percolation, but by introducing organic compounds into the solution, which act as powerful deoxidizing agents, and may even cause local precipitation of gold in the tanks. Whenever possible, such impurities should be carefully screened out. Another source of trouble is the partial oxidation by the influence of air and moisture of iron pyrites and other sulphides contained in the tailings, giving rise FIG. 18. Callow Screen, supplied by Fraser & Chalmers. to compounds, soluble in water or cyanide solution, which attack and destroy considerable quantities of cyanide. This matter will be more fully considered in a later section of this work; it may here be mentioned that this difficulty is partially overcome by preliminary water-washing and alkali treatment, applied to the tailings before running on the cyanide solution. Methods of Transfer. In general, the tailings are transferred from dumps or pits to the treatment tanks by means of metal trucks or cars running on rails, which are adjusted from time to time to bring them into the most convenient position for loading. The cars are drawn by mules or pushed by native laborers, accord- ing to local conditions. Where circumstances allow, this work may be carried out much more economically and conveniently by mechanical haulage, using conveyor belts, bucket elevators, or similar devices. (Fig. 19.) The rails are generally carried over the top of the treatment tanks, and the loaded cars are 164 THE CYANIDE HANDBOOK FIG. 19. Bucket Conveyor, furnished by Hadfield' Steel Foundry Co., Sheffield, Eng. MECHANICAL HANDLING OF MATERIAL 165 tipped so as to discharge their contents on to a wooden framework or grizzly, to assist in breaking up any lumps and to secure a uniform product for treatment. When the tank is filled, the surface is leveled off and raked over before applying the first wash. (B) HANDLING OF CURRENT MILL-PRODUCT If the tailings leaving a battery be allowed to run direct into filter tanks, it is found in most cases, after the superfluous water has been allowed to drain off or has been decanted from the settled sand, that the material so collected is quite untreatable by cyanide, owing to its mechanical condition : a partial separation of grains, according to size, takes place in the tank as in the tail- ings dams, and the result is a dense, almost impervious, mass. If an attempt be made to treat this, the solution will either not percolate at all, or will only pass through the coarser material deposited near the inlet. Settling Pits. The first suggestion for overcoming this diffi- culty was to construct comparatively small dams or pits which would collect sufficient clean sand for one or more tank charges, and allow the water and the bulk of the slime to run to waste. The sand so collected was then dug out and transferred to the treatment tanks, the stream of battery pulp being meanwhile diverted to another pit. The outlet of the collecting pit was provided with a "slat gate" consisting of two upright grooved posts, boards being placed in the grooves so as to raise the hight of the overflow -from time to time during the filling of the pit. These pits were frequently lined with masonry and the sides inclined toward the center. The objections to this method, which is still used in many places, are that much fine sand escapes along with the slime and water at the exit end (or, on the other hand, if the pit is made very large, the tailings which settle near the exit end are too. slimy to be successfully treated by percola- tion), and the transfer of the collected sand from the pits to the treatment tanks is an operation which would be eliminated by a successful method of direct filling. Direct Filling by Means of Hose. As the chief difficulty in the percolation of a charge of tailings which is allowed to run direct into the leaching tank arises from the uneven distribution of particles of different sizes, it was suggested that this trouble 166 THE CYANIDE HANDBOOK might be overcome by passing the stream of battery pulp through a hose, which could be moved by hand to different parts of the tank as the filling proceeded. This system was introduced on the Rand by Hennen Jennings about 1894. The tanks were provided with adjustable slat gates, over which the slime and water passed to the exit pipe or launder. This method is still in extensive use, but it requires a good deal of supervision, as, if the hose be not constantly moved from place to place, there will be local deposits of coarse sand and slime and consequent bad percolation. In some more recent applications of this system, the slat gates are replaced by canvas blinds, which are unrolled upward as the filling proceeds. Intermediate Collecting Tanks. A third method, introduced about the same time as the above described hose-filling system, was applied by Charles Butters, and consists in the use of special tanks for collecting the material to be treated. These tanks are filled with water before use, and when the battery pulp is turned into them, the slime and water overflow at the edges into a cir- cular peripheral launder, whence they are carried away for storage or further treatment. When a sufficient quantity of leachable sand has collected, this is transferred, through discharge doors in the bottom of the collecting tank, either direct into leaching tanks placed beneath, or into trucks by which it is conveyed to the leaching tanks. Belt conveyors may be advantageously used for this purpose. The collecting tanks are provided with filter bottoms to assist in draining off superfluous water. In order to secure uniform mixture and distribution of the material, the bat- tery pulp entering the collecting tank usually passes through an automatic distributor, designed by Charles Butters and Captain Mein. This consists of a conical bowl mounted on a pillar placed at the center of the tank, and provided with radial arms, consist- ing of tubes of different lengths bent at the exit end; the flow of liquid and the reaction of water leaving the bent pipes causes the apparatus to revolve, and so produces a uniform distribution of the pulp. (Fig. 20.) Double Treatment. In some cases the product collected in this way in the collecting tanks is sufficiently leachable for direct cyanide treatment, but in most cases it is found advantageous, after adding sufficient cyanide solution in the collecting tank to saturate the charge, to transfer the latter for further treatment MECHANICAL HANDLING OF MATERIAL 167 to another tank. This method of double treatment has the ad- vantage of exposing the material while saturated with cyanide to the action of atmospheric oxygen, under conditions favorable for solution of the gold. Occasionally charges are transferred FIG. 20. Revolving Sand Distributor for feeding tailings pulp from battery into collecting tanks. With conical hopper, revolving arms and mouth- pieces, ball bearings, spindle and flanged pillar. [From photograph furnished by the Cyanide Plant Supply Co.] a second or third time from one tank to another during the course of treatment. (See Part IV., Section I.) (Figs. 21 and 28.) (C) TRANSFER OF MATERIAL TO HIGHER LEVELS As the ore and tailings undergoing treatment have to pass through a number of machines and appliances, it is obviously an advantage, from the point of view of economy, that this trans- fer should take place as much as possible automatically. Gravity is utilized by placing each successive apparatus at a lower level than the preceding one. It frequently happens, however, that the conformation of the site on which the works are erected does not allow of sufficient fall for every transfer to take place by gravity, and in this case means must be found to raise the material mechanically to a higher level. Conveyors, both of the belt and bucket type, are often used to convey ore from the bins to the crushers (see Fig. 22); and also for transferring tailings from dams or settling pits to treat- ment tanks, for transferring from one tank to another, for trans- porting residues to the dump, etc. Tailings wheels (see Fig. 23) are very commonly adopted where wet material, such as battery pulp or the coarse discharge from hydraulic separators, has to be raised to a higher level. These 168 THE CYANIDE HANDBOOK MECHANICAL HANDLING OF MATERIAL 169 are steel structures, similar in general construction to a bicycle wheel, of a diameter depending on the vertical night to which the material must be raised say 20 to 70 ft. The rim is fur- nished with a frame divided into compartments or buckets, gen- erally on the inside, the walls of which are inclined all in one FIG. 22. Ore Conveyor (belt system) as supplied by Fraser and Chalmers, transporting rock from ore bins to crushers. direction, so that liquid or pulp may be fed into them at the bottom from a wooden trough or launder and discharged into another launder situated just below the top of the wheel. The speed is generally from 3 to 5 r.p.m. 1 Sand and Tailings Pumps. In some cases tailings are ele- vated by means of ordinary plunger pumps, running at a high speed, and with special contrivances to prevent the settling of sand in any part of the pump or delivery pipe. The Frenier 1 For further details, see Julian and Smart, loc. cit., p. 352. 170 THE CYANIDE HANDBOOK Spiral Sand Pump 1 is a well-known apparatus for this purpose. It consists of a spiral tube which is caused to revolve so that at each turn the open end scoops up some of the tailings from the launder, in which it dips. As the spiral is only partially immersed, it takes up air during part of its revolution. On reaching the center of the spiral, the pulp is discharged into a delivery pipe, to which the spiral is connected by means of a detachable tube and stuffing- FIG. 23. Tailings Wheel, Sons ofGwalia Mine. [Furnished by the Cyanide Plant Supply Co.] box; through this delivery tube air and pulp are forced by the hydrostatic pressure of the liquid in the spiral. The speed is about 20 r.p.m., and as the construction is very simple and the cost for power comparatively small, the apparatus presents considerable advantages over the centrifugal pump. The lift is from 14 to 24 ft. On the other hand, it requires considerable care, both in erection and operation, particularly in starting and stopping. (D) PREPARATION OF ORE FOR DIRECT CYANIDE TREATMENT In some cases it has been found advantageous to treat the ore by cyanide without previous amalgamation. When this system 1 Julian and Smart, loc. cit., p. 362. MECHANICAL HANDLING OF MATERIAL 171 is adopted the ore is usually crushed dry by means of rock-breakers, rolls, ball mills, etc., as previously described (Sections I and II). With exceptionally porous ores, such as that of the Mercur mine, Utah, the material may be successfully treated after merely reducing to about \ in. size by rock-breakers, and charging direct into the treatment tanks. A similar method was adopted at the George and May mine, near Johannesburg. Frequently it is necessary partially to dry the ore, so as to reduce the percentage of moisture below 2 per cent., before it can be successfully handled by dry-crushing machinery. In the case of refractory ores, containing large amounts of base metal sulphides, arsenic, tellurium, etc., it is generally neces- sary to roast before cyanide treatment, this operation being carried out either by hand, in reverberatory furnaces, or mechani- cally in one or other of the numerous types of revolving furnace, or in furnaces provided with mechanical stirrers. A description of some of the principal types of furnace used for this purpose will be found in Rose's "Metallurgy of Gold" (4th ed., pp. 246- 262). Belt conveyors or similar devices are frequently used for trans- ferring the crushed ore from the rolls or ball mills to the roasting furnaces, and from the latter to the treatment tanks. Special appliances are also used for cooling the roasted ore; thus it may be passed through a revolving drum containing tubes of cold water, or the conveyor itself may be water-jacketed. In roasting for cyanide treatment it may be observed that a partial roast is -sometimes worse than none at all, as it may con- vert comparatively harmless sulphides, etc., into soluble oxi- dized salts, such as base metal sulphates, which act as rapid and powerful cyanicides. In general it may be said that where roast- ing is adopted at all, a dead roast is essential. Many ores formerly treated by roasting have been found to be amenable to direct cyanide treatment in the raw condition, if finely enough ground. In some cases, however, the effect of roasting is to render the ore more porous, so that heavily mineralized ore, which could only be treated raw after very fine grinding, may be roasted after comparatively coarse crushing and treated direct. The subject of roasting for cyanide treatment will be more fully dis- cussed in a later section. (Part VI, Section III.) SECTION IV HYDRAULIC CLASSIFICATION, SEPARATION, AND SETTLEMENT OF SLIMES (A) HYDRAULIC CLASSIFICATION BY MEANS OF POINTED BOXES IT has already been pointed out in Section III that it is desir- able to separate the finer from the coarser portions of the crushed ore prior to cyanide treatment, as a different procedure is neces- sary for each in order to get the best results. This mechanical separation can be more effectively and cheaply carried out in water than by any other means. Various devices have been used from time to time with this object, such as flat shaking screens or jigs, revolving screens or trommels, etc.; but the simplest and most effective for the present purpose is the pyramidal box with pointed bottom, introduced in its original form by Rittinger, about the middle of the last century. These boxes are of two principal types: those which merely have an outlet for the coarse material at the bottom being known as Spitzkasten, while those which have, in addition to this outlet, a supply of water introduced near the bottom so as to cause an ascending current are distinguished as Spitzlutten. When a stream of water carrying particles of ore of different sizes and specific gravities is allowed to flow in at one side of such a box and out at the opposite side, the question whether a given par- ticle will sink to the bottom and be discharged through the out- let, or whether it will be carried away by the overflow at the top, depends, (1) on the area exposed at right angles to the direction of flow; (2) on the specific gravity of the particle; (3) on the velocity of the stream carrying the particle through the box; (4) on the velocities of any opposing currents which it may meet with. In very shallow water, the velocity of a falling particle varies / ft \ approximately as g M J, where g = acceleration due to gravity; 5 = density of water; D = density of particle. 172 HYDRAULIC CLASSIFICATION 173 In sufficiently deep still water, the velocity is uniform and varies as a (D 8), where a 2 = area of section of particle at right angles to line of fall. When the particle falls in opposition to a vertical ascending current, the pressure opposing its fall varies as / v 2 , where / = area of surface opposed at right angles to current ; v = velocity of ascending current. It is evident, therefore, that these machines are not concen- trators in the strict sense of the term, since a large particle of a light mineral and a small particle of a heavy mineral might fall at the same rate and be discharged together. The mathematical questions involved in the construction of these boxes cannot be discussed here. A full discussion of the theory of bodies falling in water will be found in Rittinger's "Lehrbuch der Auf bereitungskunde " ; an abstract of his results is given by Julian and Smart, "Cyaniding Gold and Silver Ores" (2d ed., pp., 369-375). See also Rose, "Metallurgy of Gold" (4th ed., p. 176). It should be noted that the shape of the particles influences their rate of fall, bodies presenting a pointed (wedge-shaped or conical) surface to the opposing current falling faster than bodies of the same weight and volume presenting a flat surface. In a case where the material treated consists entirely of fairly large particles of the same specific gravity, the hydraulic separa- tion will take place between fine and coarse particles, that is, the material will be classified according to size. In a case where the particles are approximately of the same size but of different specific gravities, the separation takes place according to density, and the apparatus becomes a true concentrator. A large spitz- kasten or spitzlutte with a relatively small orifice will yield, with ordinary material consisting of particles varying both in size and specific gravity, a product consisting of approximately equal falling grains, that is, grains which fall at an equal rate in still water. The use of hydraulic separators in cyanide practice is twofold. In the first place, it is desirable to collect for separate treatment the very coarse portion of the crushed* ore together with a large proportion of the heavy mineral; this is done usually by passing the pulp through a small spitzlutte having a fairly strong up- 174 THE CYANIDE HANDBOOK current. In the second place, the very fine (unleachable) part of the ore, generally described as slime, has to be separated from the leachable sandy portion. In this operation the conditions governing the separation are totally different from those concerned in the separation of coarse particles. Gravity has little influence on the result, and other factors the viscosity of the pulp, surface tension and electrostatic repulsion come into play. For the separation of slime from fine sand a much larger spitzlutte with a relatively slow ascending current must be used. Details of the construction of pointed boxes are given in many works on metallurgy. 1 We shall here draw attention only to a few points of special interest in cyanide work. In order to prevent the settling of sand near the top of the apparatus, where the ascending current is least effective, it is customary to make the upper part of the box with vertical sides, the lower part being made either pyramidal, with four triangular faces meeting at the outlet, or in the form of a triangular prism, with inclined rectangular sides in the direction of flow and tri- angular vertical sides joining these. In order to prevent surface currents from carrying the pulp direct to the overflow, a board is placed across the stream near the inlet end, either vertical or inclined, so as to force the stream of pulp downward toward the outlet and bring it in contact with the ascending current of clear water. According to Julian and Smart (loc. cit., p. 375), "for a good separation it is necessary that the rising stream should have, as nearly as possible, the same velocity at every part of its cross-section. This desideratum is attained by passing it through a rectangular passage, one of whose dimensions is very small relatively to the other." This is attained in some forms of spitzlutte by making the whole apparatus in the form of a triangular prism, with a smaller prism of similar shape forming the partition, so as to leave a V-shaped passage, rectangular in cross-section, between the two. An ad- justable form of spitzlutte has also been introduced, in which the inner V can be raised or lowered to vary the width of the passage and so regulate the nature of the product obtained. The clear water is introduced by means of a two-way pipe at the bot- tom of the apparatus; the continuation of the pipe forming the 1 See diagrams given by Julian and Smart (loc, cit., Figs. 172 to 176, pp. 375-379). HYDRAULIC CLASSIFICATION 175 outlet is bent upward and provided with a flexible hose, so that the level of outflow can be varied as required. The inclination of the pointed bottoms should be not less than 55 with the hori- zontal, otherwise sand may collect at the angles, eventually falling and blocking the outlets. The diameter of the discharge opening required is propor- tional to the square root of the depth; the minimum size neces- sary to prevent choking is about f in. diameter. The pressure of the ascending clear water current in spitzlutten must be rather greater than that of the descending pulp stream in order to ensure a clear outflow. By raising the level of outflow as above described, a back pressure is set up, which enables a product free from slime to be obtained without excessive waste of clear water. When a number of pointed boxes are used in series so as to obtain a set of products of varying fineness, each successive box must be made wider than the last, the width of each box being from 1J to 2 times greater than the preceding one. For the first box, from 4 to 6 ft. of width (according to different authorities) should be allowed for each cubic foot of pulp flowing per second through the box. The length of the successive boxes should also be increased, though this is a matter of less consequence. Where large quantities are handled, the stream of pulp is sometimes divided between a number of boxes placed in parallel, so as to avoid the necessity of constructing very deep boxes. Cone Classifiers. A form of hydraulic separator which offers some advantages over the ordinary spitzkasten and spitzlutten consists of a hollow inverted cone of wood or sheet iron having an outlet at the apex. The pulp is fed in either by means of a distributor with revolving arms which dip just below the sur- face of the water with which the cone is filled, or through a cen- trally placed vertical pipe which descends some distance beneath the surface. The slimes overflow into a peripheral launder at the circumference of the cone. The underflow may be made to pass to a second cone, or a number of cones arranged in series for more complete removal of slimes. The apparatus may be arranged either as spitzkasten, without fresh water supply at outlet, or as spitzlutte with ascending current introduced at the bottom (apex) of the cone. It is essential to have the apparatus quite level, so that a uniform overflow takes place in all direc- tions. The cones have an angle of about 55. Owing to the 176 THE CYANIDE HANDBOOK even and regular flow, a very perfect separation can be attained by proper adjustments. (See Fig. 24.) FIG. 24. Callow Settling and Pulp-thickening Tank, furnished by the Utah Mining and Machinery Co., Salt Lake City, Utah. This is a form of cone classifier which may be adapted to serve either as spitzkasten or spitzlutte. (B) SEPARATION AND SETTLEMENT OF SLIMES When the ore is very finely divided, in the condition generally described as slime, frequently it will riot settle in any ordinary time, and artificial means must be employed to remove the super- fluous water. Filtration, except under pressure, is not usually successful, and may often be too costly for practical use. Since the gold values in such suspended material are generally very easily dissolved in cyanide, it might be supposed that settlement and separation of the mineral matter could be dispensed with. Usually, however, it is very difficult to recover the dissolved gold in a concentrated form suitable for smelting, unless the suspended ore particles are removed, although from time to time attempts have been made to dissolve and precipitate in one vessel, as for example in the Pelatan-Clerici process. It is generally simpler HYDRAULIC CLASSIFICATION 177 to collect the slimy material, after causing it to settle in a more or less concentrated form, by the addition of some coagulating agent. The substance universally used for this purpose is lime, one part of which is said to be capable of clarifying between 6000 and 7000 parts of water containing clay in suspension. There are other substances which are equally or even more effective in causing settlement, but their use is precluded either by their cost or on account of some chemical property which interferes with the proper working of the process by precipitating the precious metals or by setting up other injurious reactions with the ore or solution. Among substances having a coagulating effect on suspended mineral matter may be mentioned many salts of calcium, magnesium, iron, and aluminium, notably the alums, and, in a lesser degree, neutral salts of the alkali metals, such as sodium chloride and various mineral acids. The follow- ing figures are taken from a table given by Julian and Smart (loc. cit., p. 220), representing the results of their experiments on the relative coagulating effect of certain salts: Weights Required to Produce Equal Co- agulating Effect. Aluminium sulphate 100 Potash alum 143 Lime 654 Magnesia 748 Calcium chloride 1095 " carbonate 1215 sulphate 2870 Magnesium sulphate 3460 Sodium chloride 45900 " sulphate 61700 It is thus seen that, weight for weight, alum has about five times the coagulating power of lime. Application of Lime in Preparatory Treatment of Slimes. In modern practice a certain amount of lime is usually added in the battery, sufficient to dissolve any grease from the crushing machinery which might find its way onto the plates. This has the further effect of coagulating the slime and thus bringing it into more intimate contact with the amalgamated surface. Hence it has been observed that where lime is used in the battery, the extraction by amalgamation is higher, and the assay value of the slimy portion of the tailings is lower, than when the same ore is crushed without addition of lime. As a rule, also, a further 178 THE CYANIDE HANDBOOK quantity of lime is added to the pulp in the launder carrying away the overflow from the collecting tanks or spitzlutten used in the separation of the sand. (See Sections III and IV, A.) An ingenious device has been introduced by P. S. Tavener, 1 by which lime is fed automatically at a regular rate into this launder. The overflow carrying the slimy portion of the pulp is led into a large spitzkasten, in which the coagulated slime settles in practically still water. If sufficient lime has been added, the overflow from this spitzkasten is clear enough to be used at once in the mill, and is pumped back for that purpose, while the underflow, that is, the discharge from the outlet at the bottom, consists of slime carrying perhaps fifteen times its weight of water. This material is led into a collecting-tank having a conical bottom, where further settlement takes place. By using a sufficiently deep tank for this purpose, Ihe settled slime finally obtained may contain not more than 40 per cent, of moisture. The overflow from the collecting tanks also consists of clear water which can be returned to the mill. 1 For details see Julian and Smart, loc. cit., p. 222. SECTION V AMALGAMATION AND CONCENTRATION (A) AMALGAMATION IN RELATION TO CYANIDE TREATMENT A CONSIDERATION of concentration, properly so called, might naturally follow the discussion of hydraulic classification, but as in practice some process of amalgamation almost invariably precedes concentration, these operations will be considered in the same .order. In its earlier applications, the cyanide process has been looked upon merely as an adjunct to amalgamation. It has been applied simply as a means of recovering some part of the gold which is lost in the older process, or which could only be obtained by amalgamation with much trouble and expense. This view still very largely prevails, and probably governs the design of metallurgical plants at the majority of mines in most parts of the world. As a consequence, while much care is taken to secure the maximum recovery on the plates, little consideration is given to the question whether the product leaving the plates is in a suitable condition for yielding the best extraction by cyanide treatment. The two processes are looked upon as more or less antagonistic, and are usually under separate managements, between which there is often considerable rivalry. In a few localities, notably in the Rand, it has been recognized that the metallurgy of the ore should be regarded as a single problem, and that method or combination of methods adopted which is found by experience to yield the maximum profit. Thus, in some instances it has proved more advantageous to aim at a lower extraction by amalgamation, as this may allow of an in- creased tonnage crushed per month, the subsequent cyanide treatment being relied on to recover values which otherwise might have been saved in the mill. Generally speaking, how- ever, it may be said that it is cheaper to recover gold by mercury than by cyanide, especially when the gold is in fairly coarse 179 180 THE CYANIDE HANDBOOK particles, and it is therefore good policy to make the amalgama- tion as efficient as possible, consistently with leaving the tailings in such a condition that the values not recoverable by amalgama- tion may be readily extracted by cyanide. Fortunately, the conditions which promote good amalgamation are in general also beneficial to cyanide treatment, as, for example: (1) sufficiently fine crushing of the ore (already referred to in describing the action of stamps: Section I, B); (2) a slight alkalinity of the water in which the pulp is suspended, obtained by the addition of lime in the battery; (3) absence of soluble and easily decomposed compounds of the base metals, which stain or " sicken " the mer- cury and also act as cyanicides; (4) absence of oil, grease, and similar matter. Amalgamation is carried out in one of the two following ways: (1) by causing the crushed ore, suspended in water, to come in contact with surfaces coated with a layer of mercury (plate amalgamation); (2) by grinding ore and mercury together in an iron pan, with sufficient water to form a paste, other chemicals being added if necessary (pan amalgamation). Other modi- fications of less importance need not be considered here. In the case of ores in which the value consists principally of gold, plate amalgamation is almost universally used; when the material to be treated is essentially a silver ore, pan amalgama- tion is more usual. In plate amalgamation sheets of copper coated with mercury are placed inside the mortar-box, generally both on the feed and discharge sides, attached either to the mortar-box itself or to wooden blocks. Amalgamated plates are also placed on a sloping table immediately in front of the screens, so that the pulp flows over them as it leaves the mill. The practice of amalgamation inside the mortar-box is open to much criticism, on the ground of difficulties in regulating the work, losses of mercury involved, etc., and has been entirely abandoned in many places. It is, however, defended by some metallurgists (see Rose, " Metallurgy of Gold," pp. 149, 150). It is probable also that the practice of placing the outside plates close to the screens may be largely abandoned in the near future, as it seems desirable for various reasons that the amalgamation should be conducted in a room or building entirely separate from the crushing machinery. The plates used for amalgamation consist of pure copper, of copper AMALGAMATION AND CONCENTRATION 181 previously coated by electroplating with a thin film of silver, or of Muntz metal, an alloy of copper and zinc. They are from T V to f in. in thickness, thicker plates being required for inside than for outside amalgamation. For the latter, the plates are laid on an inclined wooden table, usually 4 or 5 ft. wide and 12 to 14 ft. long, having raised edges to retain the pulp. It is found that plates previously amalgamated with gold or silver. are more effective in catching the precious metals than copper plates amalgamated only with pure mercury; hence the use of silvered plates alluded to above. Effective amalgamation depends chiefly on keeping the sur- face of the plate as clean as possible, and in so regulating the amount of mercury that it forms a pasty amalgam with the gold and silver which it extracts from the ore; when too little mercury is added, the amalgam becomes hard and ceases to take up fresh quantities of the precious metals; when too much is used, it runs off in drops and is carried away with the tailings, together with any gold and silver which it may have absorbed. Other impor- tant points are: (1) the due regulation of the water-supply so that the pulp passing over the plates may be neither too thick nor too thin; (2) proper inclination of the table, so that the pulp comes into effective contact with the amalgamated surface without allowing a deposit of sand or mineral to settle permanently on it. This inclination is from 1 to 2J in. per foot, according to the nature of the ore, a steeper grade being naturally required when . much heavy material is present. At intervals during the day each battery of five stamps is stopped and the plates connected with it cleaned. A part of the amalgam is removed at the same time by means of a rubber scraper and fresh mer- cury applied. In many mills a certain quantity of mercury is also fed at intervals into the mortar-boxes. The total amount of mercury used varies, of course, according to the nature and richness of the ore treated, but may be put at about 1 to 2 oz. per ounce of gold amalgamated. Most of this is recovered in re- torting the amalgam, but there are some losses which must now be considered, as they are of importance in connection with cyanide treatment. The practice of feeding mercury into the mortar-boxes causes some of it to be broken up, by the action of the stamps and the agitation to which it is subjected, into very small globules known 182 THE CYANIDE HANDBOOK as " floured mercury." This passes through the screens and may be carried off the tables by the stream of pulp, ultimately finding its way into the cyanide vats. If ordinary base metals, such as copper, zinc, lead, or tin, be present as impurities in the mercury used for amalgamation, or occur in the ore in the form of easily decomposable compounds, these metals give rise to films which coat the surface of the mercury and entirely prevent amalgama- tion. When this occurs the mercury is said to be "sickened." These films consist of the oxides, sulphides, sulphates, carbonates, or other compounds of the base metal. In some cases actual alloys of mercury with another metal may be formed, and occa- sionally sulphates or other salts of mercury. 1 Lead forms an amalgam which separates as a frothy scum, carrying with it any gold which may be present; this easily becomes detached from the surface of the mercury and is then carried away by the pulp. Arsenic and antimony are particularly harmful. Arsenic, whether in the metallic state, as sulphide (As 2 S 3 ), or as mis- pickel (FeAsS), produces a black coating which is a mixture of metallic arsenic and finely divided mercury, no amalgam being formed. Antimony separates from its compounds rapidly in a similar manner, but the metal forms an actual amalgam, and another part of the mercury is converted into sulphide. Bismuth has a similar but less rapid action. All these compounds readily break up into small particles, so that sickened mercury is very liable to be carried away by the stream of ore and water and hence to find its way into the cyanide plant, giving rise to various undesirable reactions. All metallic sulphides except clean, coarse, undecomposed iron and copper pyrites have some action on mercury. Gold in iron pyrites largely escapes amalgamation, unless the mineral be very finely crushed and treated in pans, in which case much sickening and loss of mercury takes place. The difficulty of amalgamating gold in pyrites appears to be due to the fact that the metal occurs in thin layers or plates on the surfaces of the crystals of pyrites, or occupies fissures in it, the gold itself being possibly coated by thin films of the mineral or of sulphur, so that contact with mercury is prevented. Certain gangue minerals may also cause losses of mercury; among these may be mentioned heavy spar (barium sulphate), talc, steatite, and similar greasy 1 Rose, loc. cit., pp. 133, 143, 144. AMALGAMATION AND CONCENTRATION 183 hydrated silicates of magnesia and alumina. The latter cause frothing and form a coating on the gold which prevents amalgama- tion. Occasionally, the gold itself, although apparently free, is coated with a thin film, consisting of sulphur, oxide of iron, silica, arsenic, metallic sulphides, etc., which prevents contact with the mercury surface. Such gold is said to be " rusty," and is not recoverable by amalgamation unless means be taken to remove the film.' When the ore or the water used in the battery contain any base metal compound or any acid capable of acting on copper, yellow, brown or green stains are observed on the surface of the plates, which interfere with amalgamation, and if not removed, will increase until all contact between the gold and mercury is prevented. These consist usually of carbonates, oxides, and sulphides of copper, and perhaps in some cases of mercurial compounds. From the present point of view, the matters of chief interest are those practices which affect the subsequent cyanide treat- ment of the tailings, and some reference must therefore be made to the chemicals used to promote amalgamation. It has already been mentioned that lime is frequently added to the ore in the mortar-boxes, or to the battery water. This neutralizes any acid substances, such as soluble salts of iron, which may be pres- ent, and which might tarnish the plates, and also helps to dis- solve any oily matter introduced into the ore or feed water from the grease used in lubricating the crushing machinery. Any kind of grease ha*s a very deleterious effect on amalgamation. The addition of lime, moreover, causes flocculation and settle- ment of the fine suspended ore particles, thus bringing them in contact with the amalgamated surfaces. This treatment renders the material more suitable for cyaniding, both from a mechanical and from a chemical point of view, but in certain cases it may render the amalgamation so effective that the values remaining in the tailings are reduced below the point at which cyanide treatment is profitable. (See Jourri. Chem., Met. and Min. Soc. of South Africa, II, 87.) (Proc. 2, 657.) In order to remove the stains on the plates to which reference is made above, and to produce a bright clean surface for amalgama- tion, certain chemicals are used, the chief being sal ammoniac and cyanide. Both of these are solvents of the copper com- 184 THE CYANIDE HANDBOOK pounds which cause the discoloration. When they are to be applied, the battery is stopped and the spots scrubbed with the required solution. The chemical is then washed off, and gen- erally a fresh quantity of mercury is added before allowing the pulp again to flow over the plate. It is obvious that the use of cyanide for this purpose is open to very serious objections. It is applied in the form of a fairly strong solution, perhaps 0.5 per cent. KCy, and even though the mill is stopped during the dressing of the plates, some part of the solution used must almost inevitably find its way into the launder which conveys the tailings from the mill, where it will dissolve some gold. Unless the whole of the battery water is returned to the mill, a part of these dissolved values will be lost. If used at all, the plates should be carefully washed with fresh water immediately after the cyanide has been applied, and such water should be run into a separate receiver. 1 The practice of adding cyanide in the mortar- boxes is, of course, still more objectionable, and could only be defended where complete arrangements are made for crushing with cyanide solution, which involves passing all the liquid used in the mill through the precipitation boxes. Where no lime is used in the battery and the water contains acid iron salts, these would probably destroy in most cases any excess of cyanide used in cleaning the plates before any appreciable amount of gold had been dissolved from the tailings. Cyanide is also to some extent a solvent of oily and fatty matters, but probably only by virtue of any free alkali it may contain. For this purpose a solution of caustic soda or of sodium carbonate is much more effective; and as already mentioned, lime acts beneficially in this way. Clean floured mercury may be collected and recovered by agitation with a large mass of fresh mercury, especially if a little alkali be added to dissolve any film of grease which may be present. The sickening of mercury may often be remedied by the addition of sodium amalgam, which contains 2 to 3 per cent, of metallic sodium. The action of this substance is to reduce the metallic oxides, forming caustic soda, which at once dissolves in the water and is incidentally beneficial as a solvent of grease and some other impurities. The base metal is liberated in the metallic form and 1 See paper by A. von Gernet, " Losses of Gold in Mill Water," in " Proc. Chem., Met. and Min. Soc. of South Africa," II, 529. AMALGAMATION AND CONCENTRATION 185 usually amalgamates with the mercury. In the case of antimony, however, the sulphide of mercury present in the film is attacked, forming sodium sulphide, which acts upon a further quantity of the antimonial mineral, forming more antimony amalgam and liberating sulphureted hydrogen, so that in this case the use of sodium amalgam probably does more harm than good. Hydrochloric, or still better, nitric acid, will dissolve some of the impurities in sickened mercury; the latter readily dissolves some of the mercury itself, exposing a bright metallic surface, so that the particles will coalesce on agitation. Nitrate of mer- cury also removes the stains on the plates, forming metallic mercury and copper nitrate. The foreign substances introduced during the crushing and amalgamating processes, and which may influence the results of subsequent cyanide treatment, may be here enumerated: (1) Metallic mercury, due to the use of excessive quantities in dressing the plates, or to flouring or sickening, as described above. (2) Gold and silver amalgam, scoured off the plates by coarse particles of ore. (3) Base metals, originally present in the mercury and ulti- mately converted into oxides or alloys, coating the sickened mercury. (4) Lime, in solution in the water carrying the pulp from the battery. (5) Oil and grease from the bearings of machinery, from feed water, or from rmine candles. (6) Fragments of metallic iron from the shoes, dies, and other parts of the crushing machinery. (7) Particles of brass and copper from plates, screens, and the percussion caps used in blasting. (8) Chips of wood, rubber, leather, and other miscellaneous materials which may accidentally fall into the stream of battery pulp. (B) CONCENTRATION IN RELATION TO CYANIDE TREATMENT General Principles of Concentration Concentration may be defined as any process by which a mass of material is separated into two portions, one containing 186 THE CYANIDE HANDBOOK a greater proportion by weight of the valuable constituents than the other. This separation is generally carried out by utilizing some difference in the physical nature of the particles which constitute the mass, with a consequent difference in behavior when subjected to the same forces. The methods of hydraulic separation by means of spitzkasten and spitzlutten have already been discussed. In these methods the forces involved are prin- cipally gravity, inertia, resistance to hydrostatic pressure, and (in the case of the finer particles) viscosity and probably elec- trostatic repulsion. Concentration by machines of this class must always be imperfect from a metallurgical point of view, since, as pointed out above (Section IV, A), the product obtained consists of particles differing in size and specific gravity, small particles of heavy material being collected together with large particles of light material; whereas an ideally perfect concentrator would separate the whole of whatever mineral or substance it was designed to save, and the product yielded by it would there- fore consist entirely of particles having the same specific gravity. No machine actually fulfils this condition. What is aimed at in practice is the separation of the material to be dealt with into two products: (a) Concentrates, consisting of a relatively small total mass, containing a relatively large part of the total valuable constitu- ents originally present, and a relatively small part of the worth- less ingredients. (b) Tailings, having a relatively large total mass containing as little as possible of the original valuable contents. The forces which are utilized in the ordinary systems of con- centration, and which effect their purpose owing to the difference of their action on different kinds of particles, are: (1) gravity; (2) inertia and centrifugal force; (3) adhesion between surfaces of concentrating machine and the class of particles to be separated; (4) Adhesion between the fluid used in the apparatus and the particles to be separated. When applied to gold and silver ores, concentration as a rule follows amalgamation. The pulp, or crushed ore suspended in water, as it leaves the plates, passes to the concentrating ma- chines, which are of three or four principal types. AMALGAMATION AND CONCENTRATION 187 Old Types of Concentrator The oldest and most primitive types of concentrator depend for their action solely on the fact that in a flowing stream the heavier particles carried by the water tend to settle and collect against any obstacle or irregularity in the bed of the stream. This is the principle of the "Long Tom" and other forms of riffled 'sluices, in which the material is washed down an inclined trough or launder, either stationary or subjected to a rocking motion, the coarser particles of gold and the heavier minerals being caught against strips of wood (riffles) nailed at intervals across the bed of the launder. Another primitive device is to allow the pulp to flow over a sloping table covered with blankets or some kind of rough cloth, the heavy particles which settle to the bottom of the stream being caught and retained by the hairs and interstices of the cloth. These cloths are removed at inter- vals and the heavy material, known as "blanketings," which has collected on them is brushed or washed off, by hand or mechani- cally, into any suitable receptacle. In many concentrators some device is used for keeping the pulp in a state of agitation, so that the lighter particles may be thrown into suspension and carried off by the flowing stream, while the heavier particles are moved only a short distance along the bed of the machine. In the various types of "buddies" the pulp flows over a stationary inclined bed, on which it is stirred by means of revolving arms carrying brushes or similar con- trivances; or the same effect is produced by revolving the inclined bed and using stationary brushes. These machines generally have a conical form; in the convex buddle the pulp is fed at the apex and the tailings discharged into a launder at the circum- ference; in the concave buddle the feed is at the circumference and the tailings flow away through an opening in the center. The heavy mineral collects near the feed and the lighter particles are carried off as tailings; but in all such devices a large quantity of ore is spread out over the surface between the feed and dis- charge, consisting of a mixture of different kinds of particles, gradually diminishing in richness from the feed to the discharge. To obtain a satisfactory concentration, this "middle product" must be collected from time to time and again fed onto the ma- chine, necessitating repeated handling of the same material, so 188 THE CYANIDE HANDBOOK that the capacity of the concentrator is very small in compari- son with the labor and power required to work it. Modern Types of Concentrator Modern concentrating machinery, as applied to gold ores, apart from the pointed boxes already described, is almost entirely of two types: percussion tables, and endless-belt tables. (a) Percussion Tables. In these machines the material to be concentrated is fed onto a sloping table having a smooth sur- face of wood, iron, sheet copper, linoleum, etc., together with sufficient water to make a thin pulp, some device being often used for securing a uniform distribution and regular feed of ore and water onto the table. The ore, thus spread out in a thin layer, is kept in a constant state of agitation by a series of blows delivered against one end of the table by a set of cams or some similar contrivance. These blows cause the heavier particles which have settled out of the flowing pulp to travel gradually toward one end of the table, where they are collected in a box or other receptacle; the lighter particles are thrown into suspen- sion and carried off by the stream of water, either to the oppo- site end of the table or over the side into a launder, by which they are discharged. In some types of percussion table, the ore spreads itself out into a fan-shaped layer in which the different minerals are sorted according to their specific gravity; the heaviest particles form a line at the extreme edge and may be collected separately; then follow in succession bands of (say) galena, lead carbonate, pyrites, blende, black oxide of iron, coarse and fine sand, and finally slime. The middle products, if necessary, may be collected in a separate launder leading to an elevator, which returns them to the table for reconcentration. Among the older types of machine based on the percussion principle may be mentioned the Gilt Edge concentrator; among modern improved types are the Wilfley, Woodbury, James, and many other concentrating tables of similar design. Through the courtesy of the Wilfley Mining Machinery Co., the following brief description of this well-known machine is given. The latest form (No. 5, 1906) of the Wilfley Concentrator consists of a flat deck or table mounted on rockers and supported by cast-iron plates bolted rigidly to a timber frame. The move- ment is imparted at the upper end of the table by means of a toggle AMALGAMATION AND CONCENTRATION 189 operated by an eccentric shaft, the length of stroke being ad- justed by turning a hand wheel, which raises and lowers a wedge block at one end of the toggle, varying the stroke from to 1 in. The shaft makes about 240 r.p.m. When in use the table is set so as to be slightly inclined from the feed to the discharge side. It is generally constructed of redwood, in narrow strips laid diagonally to avoid warping, and strengthened by steel stringers extending the full length of the table, to obviate any bending of the table surface. It is covered with linoleum, over a part of which are tacked parallel strips of wood, forming riffles. This kind of surface readily holds the ore particles, and is also durable, impervious to water, and easily replaced. (See Figs. 25 and 26.) FIG. 25. Wilfley Concentrator. (Showing driving gear. From cut furnished by the Wilfley Mining Machinery Co.) The pulp is fed onto the table through holes in the upper side of the pulp-box, placed near the upper corner on the feed side of the table. Water is delivered by an open, perforated box extend- ing the whole length of the feed side, below the pulp-box. The tailings are discharged over the edge of the table on the discharge side (opposite the feed) into a launder. The concentrates pass off at the lower end of the table (opposite the driving gear) . Any required portion of the tailings may be returned as middlings, by a separate launder, to an elevating wheel connected with the same eccentric shaft which moves the table, and passed back 190 THE CYANIDE HANDBOOK to the feed box. This arrangement prevents losses of mineral which might otherwise occur through irregular feeding of pulp or wash-water. The bank of mineral formed by the return ele- vator tends to prevent a rush of material toward the tailings launder. The finest slimes are discharged near the head of the table; the remaining minerals arrange themselves in the order FIG. 26. Wilfley Concentrator. General view showing distributing box and riffles. [From cut furnished by the Wilfley Mining Machinery Co.] of their specific gravity, the heaviest passing off at the foot of the table. It is sometimes desirable to reconcentrate the middlings after further crushing, on a separate table, instead of using an elevator as above; in this case the feed to the table should be regular. (6) Belt Tables. Belt concentrators usually consist of an endless, revolving, rubber belt, flanged at the edges, which travels slowly over a couple of rollers, placed in such a position that the upper part of the belt forms a slightly inclined surface, having an area of about 12X4 feet. At the same time a slight but rapid shaking motion is imparted to the machine, either laterally or longitudinally. Ore and water are distributed as uniformly as possible across the width of the traveling surface, near its upper end, and the belt is caused to revolve in the opposite direction to the flow of the pulp; that is, it travels up the incline while passing above the rollers. By means of another pair of rollers placed beneath the machine, the belt on its return journey is made to pass through a tank or trough containing water, which serves to collect the concentrates. The heavy particles of mineral AMALGAMATION AND CONCENTRATION 191 fall to the bottom of the pulp stream and cling to the surface of the rubber; as the belt ascends, it carries them past the pulp distributor, behind which a perforated pipe delivers a number of small jets of water. These serve to clean the concentrates still further by washing away the lighter particles, while the heavier material is carried past them and delivered to the collecting box underneath. The light portions are carried by the stream of water down 'the slope of the belt and discharged into a launder at the lower end. To this type of machine belong the Frue Vanner (Fig. 27), the Embrey Concentrator, the Liihrig Vanner, etc. The lat- FIG. 27. Frue Vanner, supplied by Fraser and Chalmers. ter delivers a middle product at the side of the belt, and receives a series of blows from a cam-shaft, in these respects resembling the Wilfley table. (See above.) Other Systems of Concentration Other types of concentrator are used in special circumstances such as jigs, shaking screens, dry blowers, centrifugal separators, etc. ; but these need not be described here. Some reference, how- ever, may be made to the recently introduced methods of "oil concentration" and "flotation," in which gravity plays no part whatever. In the Elmore oil concentration process, the mixture of crushed ore and water is fed into a slightly inclined revolving cylinder together with a considerable quantity of crude mineral oil. Owing 192 THE CYANIDE HANDBOOK to some physical action at present imperfectly understood, cer- tain minerals, particularly those having a bright "metallic" surface, adhere to the oil. On discharging the contents of the cylinder into tanks containing water, the oil floats to the top, carrying with it the adhering minerals, which are then separated from it by means of a rapidly revolving centrifugal extractor. By this means gold, galena, copper, and iron pyrites and other lustrous minerals may be separated from earthy carbonates, sulphates, oxides, etc., and from sand and siliceous minerals generally. The "flotation" process utilizes the surface tension of water as a means of separating the same class of minerals, namely, those having a lustrous surface. The pulp is distributed by means of a revolving cylinder, rifled internally, so that it falls at a certain angle on the surface of water. Those mineral particles which under the circumstances are not "wetted," float on the surface and are carried to a receptacle where they are caused to sink by agitation. The remainder of the pulp sinks at once in the collect- ing tanks and is discharged from the bottom. Concentration Previous to Cyanide Treatment In the early days of the cyanide process it was thought that the solvent could not be successfully applied to material contain- ing coarse gold or refractory minerals having considerable chemical action on cyanide; hence the system was generally adopted only for tailings after concentration. The metallurgical scheme com- monly carried out in dealing with gold ores was as follows : Crush- ing by rock-breakers and stamps; amalgamation of coarse gold in the battery; concentration of battery tailings, generally by means of Frue Vanners (the object being to obtain the maximum gold value with the minimum of other substances) ; cyanide treat- ment of the tailings from the concentrators in cases where the values still remaining in them were sufficient to pay the cost. The concentrates were generally roasted in reverberatory or other furnaces and treated by chlorination, yielding a high percentage of extraction, but at such a heavy cost that the operation could only be conducted in large establishments treating the accumu- lated product from a number of mines. This system has gradually been replaced by the method of AMALGAMATION AND CONCENTRATION 193 hydraulic classification described above, in which the tailings after amalgamation undergo only a rough concentration in spitz- lutten, the object being not to obtain a close saving of the values by concentration, but to obtain a product carrying the bulk of the coarser and heavier mineral particles, so that this product may receive special treatment, leaving the main portion of the pulp as fine sand suitable for ordinary treatment. Roughly speaking, the pulp leaving the battery is finally obtained in three products: (1) a comparatively small amount of coarse heavy sand, requiring a long time of treatment with rather strong solu- tions; (2) a much larger amount of fine sand, amenable to cyanide treatment with weak solutions, not over 0.25 per cent. KCN, and (say), 3 to 4 days' contact; (3) slime, capable of treatment with very weak solution by some system of agitation, followed by settlement and decantation, or by filter-pressing. Under ordinary circumstances this more recent system is found to yield a greater profit, though probably a lower per- centage of extraction than the earlier method. It also has the advantage that the whole of the gold recovery can be made in a metallurgical plant under direct control of the management at each mine, none of the products being necessarily sold to smelters or reduction works. 1 A still more modern system, largely adopted in Western Australia and coming into general use in other fields, consists in reducing the whole of the ore to such a fine state of division that it can be treated (with or without amalgamation and hydraulic classification) by agitation with cyanide solution, the gold-bearing liquor being finally separated by filter-pressing. This system has only been rendered possible by the introduction of cheap devices for fine grinding, such as the tube mill already described. (See Section II, D.) Cyanide Treatment of Concentrates Where the heavy mineral consists almost entirely of clean unoxidized iron pyrites, the concentrates, even when carrying a large percentage of sulphides, can generally be treated success- fully by cyanide, but a long period of treatment, sometimes amounting to three or four weeks, is necessary, and some arti- 1 See also Rose, loc. tit., pp. 221-222. 194 THE CYANIDE HANDBOOK ficial means must generally be adopted for securing the amount of oxygen needed for the reaction. The concentrates must be turned over or discharged from one tank to another from time to time during the treatment, and it is sometimes advantageous to aerate the solutions. In special cases oxidizing agents or other chemicals may be added as an aid to solution. In cases where the concentrates have become partially oxidized by exposure, difficulties arise owing to the action of soluble sulphates of iron and of insoluble basic sulphates on the cyanide. These diffi- culties may be largely overcome by giving a preliminary treat- ment with a mineral acid (H 2 SO 4 , or HC1). 1 The reactions involved are discussed in the section of this work dealing with the chemistry of the process. The treatment of spitzlutte concentrates presents less diffi- culty, both because the percentage of refractory material is smaller and because the coarse sand renders the mass readily leachable; moreover, since the material is collected by settlement from a large bulk of water, ail soluble salts will have been removed before the cyanide is applied, and partial oxidation of the pyrites to basic sulphates is not likely to occur unless the treatment of the product be unduly delayed. 1 See " Proc. Chem., Met. and Min. Soc. of South Africa," I, 98-103, 133-134. PART IV THE DISSOLVING PROCESS THE extraction of gold and silver from a mass of pulverized material by means of cyanide or other solvent may be brought about in one of two ways: (1) By causing the solution to pass, either by gravity or by hydraulic or atmospheric pressure, through a stationary body of the material to be treated. (2) By agitating solvent and material together by suitable appliances, so as to bring them into intimate contact, and then separating them, either by settlement and decantation or by filtration. The term " percolation " is used simply to signify the passage of a liquid through porous material, whether anything is dis- solved by the liquid or not. The term "leaching" means the extraction of soluble matter from a mass of material by the pas- sage of a liquid through it. The two terms, however, are often used in technical literature as though they were interchangeable. The words "vat" and "tank" also are commonly used as inter- changeable, although some attempts have been made to restrict the use of the word "tank" to vessels holding liquids. SECTION I PERCOLATION (A) LEACHING VATS IN modern practice, percolation is almost invariably carried out in circular vats of wood or iron, having a flat bottom on which rests a filter-frame, composed of thin parallel strips of wood covered with canvas or other suitable cloth to form a filter. In the early days of the process, square wooden vats were com- monly used, but these presented many drawbacks, both in con- struction and use, and usually caused heavy losses by leakage. Cement-lined masonry tanks have also been used occasionally. The vats should rest on a solid foundation, supported above the ground on masonry piers, so as to allow a free passage under- neath. Wooden vats are constructed of a number of upright staves, beveled at the edges, so that when fitted together and tightened by the hoops they form a water-tight structure. Each of the staves has a notch or "check," a few inches from the bottom, to receive the floor or bottom of the tank. The latter is constructed of planks with tongues and grooves, so that they may be tightly joined, and is cut in the form of a circle of the required diameter, allowing for the check in the staves. The hoops are made from lengths of round iron, the ends being tightly screwed together. 1 Steel vats are constructed of plates t\ in- to f in. in thickness, riveted together. Those used for the sides of the tanks are bent to the required curve by passing them through rolls. The vats are strengthened by angle-iron rings at top and bottom. 1 When carefully constructed there is little tendency to leak. Small leaks may be stopped with fine slime, paraffin wax, or vaseline. Either type of vat has its advantages and disadvantages. 1 For details of construction see Julian and Smart, " Cyaniding Gold and Silver Ores," 2d edition., Ch. XXXIV and XXXV. 197 198 THE CYANIDE HANDBOOK Wooden vats are more liable to leakage, and a certain quantity of solution containing gold and silver values may be absorbed by the wood. Iron vats, on the other hand, are more expensive, both in first cost and maintenance; leaks, when they do occur, are more difficult to stop, especially at the bottom of the tank, after the filter has been laid. They are also liable to distortion PERCOLATION 199 under internal pressure when the ore in the vat is irregularly distributed. This causes cracks through the sand charge and consequently bad percolation, as the liquid forms channels. The latter defect can be remedied by using sufficiently heavy material and bracing with angle-iron. The- accompanying illustration (Fig. 28) shows a large South African Plant with double treat- ment tanks constructed of steel. With regard to the absorption of solution by wooden vats, experiments by F. L. Bosqui 1 and others would seem to show that this is' not a very serious matter. Somewhat greater absorp- tion appears to take place when the immersion in solution is intermittent than when continuous. Thus a piece of redwood having an area of 180 sq. in., immersed continuously for 3 weeks in a gold cyanide solution, absorbed gold to the value of $1.07; whereas a similar piece of wood of the same area, treated for the same length of time, but immersed for 16 hours and dried in the sun for 8 hours each day, absorbed gold to the value of $2.45 per ton of wood. Tests on a screen frame of rough pine, which had been in contact with solution for 12 months, gave, however, the following result: Bottom of frame, continuously immersed: Gold absorbed, $5.32 per ton of wood. Top of frame, alternately exposed and immersed, about half time under solution: Gold absorbed, $2.87 per ton of wood. S. H. Williams 2 states that a launder through which 63,000 tons of auriferous solution carrying 0.07 per cent, cyanide had passed was burned, and that the ashes were found to contain 2.766 oz. gold and 32.464 oz. silver. Filter-frames. These are formed by a number of parallel wooden slats 6 to 9 in., apart, above which, and at right angles to them, are nailed wooden strips of, say, 1 sq. in. section, laid 1 in. apart. Large frames are made in sections, and round holes are cut over the discharge doors. Filter-cloths. These are generally of cocoanut matting, sometimes covered with an upper cloth of duck or canvas or, better, hessian. The edges of the cloth are tucked tightly be- tween the sides of the vat and the filter-frame, and the joint is sometimes made water-tight by wedging a rope into this space. i"Min. Sc. Press," April 14, 1906. 2 "Trans. I. M. M.," Bui. No. 11, Aug. 17, 1905. 200 THE CYANIDE HANDBOOK Parallel strips of wood are sometimes laid above the cloth and at right angles to the strips below, to protect the cloth during the discharging of the vat. These are fastened lightly, so as to be easily removed. Discharge Doors. Some description of these is given below. Holes for the doors are cut in symmetrical positions in the bottom of the tank and through the filter-frame and cloth, and are so arranged that the discharge of the tank can be carried out with the minimum of labor. 1 Solution Outlet. This is placed in any convenient position below the filter-frame, and is furnished with a cock and a pipe leading to the precipitation department. Dimensions of Vats. Tanks have been constructed to hold any quantity up to 600 tons of tailings, and may be as much as 60 ft. in diameter. There is no difficulty in constructing larger ones if required. The depth is generally considerably less than the diameter. Ordinary sizes are 20 to 30 ft. in diameter, and 8 to 14 ft. inside depth. Tanks for storing solution are commonly made deeper in proportion to diameter than leaching tanks, as in the latter the efficiency of percolation has to be considered. (B) THE LEACHING PROCESS Conditions for Effective Percolation. As already pointed out, it is very desirable to separate the slime as much as possible from the material which is to be treated by percolation; and better results are obtained when the grains of sand are more or less of the same size than when all sizes are mixed together. (See Part III.) To ensure good results it is necessary to lay the material uni- formly throughout the vat, without undue compression, and finally to level the surface by raking it over before introducing cyanide solution. By this means the solution percolates evenly through all parts without forming channels. The tank is com- monly filled to within an inch or so of the top, but the material settles considerably when moistened with solution. The valves at the outlet pipes are closed until all bubbling (due to displace- ment of air) has ceased; this ensures a better aeration of the charge than if the valves are left open in order that the solution sinking into the mass may drive the air out through the exit pipe. 1 Julian and Smart, loc. tit,, p. 297. PERCOLATION 201 In certain cases, as in the treatment of accumulated sand which has not undergone hydraulic separation, or where sufficient lime has not been added in the battery or elsewhere to neutralize the latent acidity of the material, it is necessary to add lime to the charge as the tank is filled. This is sometimes spread in a layer over the surface of the tank, and raked over before adding the first wash of solution. Preliminary Water-wash. This is occasionally given for the purpose of removing soluble cyanicides, that is, substances which would consume cyanide during the treatment. It is generally unnecessary when the sands have been obtained by hydraulic separation or have settled from water in collecting tanks. When used for the purpose of removing soluble salts of iron, copper, etc., it is best not to add lime or other alkali to the charge or to the water used in washing, as these salts would be wholly or partially precipitated by the alkali. Alkali Wash. It is a common practice to give a preliminary treatment with an alkaline solution, commonly of lime, the chief object being to secure the neutralization of insoluble cyanicides, such as basic ferric sulphate, before adding the strong cyanide solution, and thus avoid any undue consumption of cyanide in the latter. In practice this amounts to a treatment with very weak cyanide solution, as a certain amount of cyanide from previous charges finds its way into the water used for the alkali wash, for which a special storage tank is commonly reserved. The alkali wash is pumped on and allowed to leach through con- tinuously until the effluent is distinctly alkaline. The first por- tions of liquid passing through are sometimes run to waste, but as they commonly contain gold, it is better to allow the whole of the effluent to return to the alkali storage tank until it shows sufficient strength in cyanide for effective precipitation; then it is diverted to the precipitation boxes, the small quantity of cyanide thus introduced into the alkali wash being no serious disadvantage. The tank is allowed to drain as much as possible, so as to avoid undue dilution of the strong solution which is after- ward applied. Occasionally the strong solution is preceded by a weak cyanide wash, also with the object of diminishing the dilution which would otherwise take place in the strength of the strong solution; but the more usual practice is to add the latter as soon as the alkali wash has sufficiently drained off, it being then 202 THE CYANIDE HANDBOOK assumed that latent acidity in every part of the charge has been properly neutralized. Strong Solution. Before adding the strong solution the out- let cock is closed. The solution is then pumped on to the vat from the strong solution storage tank, in which it has previously been made up to the required strength in cyanide, usually by adding the calculated amount of liquor from the "dissolving tank/' a small vat placed above the storage tanks and used for dissolving the solid cyanide from the cases. In some plants the cyanide is dissolved directly over the leaching tanks, by placing the solid lumps in a perforated box into which weak solution from the storage tanks is pumped. This method is not to be recom- mended, as it must lead to the treatment of some parts of the charge with unnecessarily strong solution and consequent waste of cyanide. The solution is pumped on until the charge remains well covered with liquid to a depth of two or three inches. It is then allowed to stand, frequently for twenty-four hours or longer, so as to allow the charge to soak thoroughly in strong solution. It is a better plan, however, to draw off the solution as soon as all bubbling has ceased, say after one or two hours, and allow it to run off until the effluent shows a sufficient strength in cyanide for effective dissolution of the values; the outlet valve is then closed and more strong solution is pumped on, until the charge is again covered to the required depth. The charge is now left standing under strong solution as long as may be considered necessary. The object of this procedure is to ensure the satura- tion of the whole charge with solution of sufficient strength, since but little diffusion takes place in the leaching tank, one solution dis- placing another without mingling with it to any considerable extent. The strength of solution to be used depends on the nature of the material to be treated. Roughly speaking, the richer it is in gold, the stronger will be the cyanide solution required, but this rule is by no means universal. Where silver is to be ex- tracted, much stronger solutions are commonly required than for gold alone, partly owing to the greater weight of metal which has to be dissolved and partly to the presence of accompanying minerals which consume much cyanide. This matter is discussed in the sections dealing with the chemistry of the process. Under normal conditions a strength of 0.25 per cent. KCN * (=0.1 per 1 5 Ib. potassium cyanide or 3.8 Ib. sodium cyanide per ton of solution. PERCOLATION 203 cent. CN) is a good standard to work with, though in the case of clean tailings free from coarse gold, a strength of 0.1 per cent. KCN (= 0.04 per cent. CN) is often sufficient to give a good extraction of the values. (See pp. 105-107.) When the charge has stood long enough under strong solution, the outlet cock is opened and the liquid allowed to drain off to the precipitation boxes; the effluent is tested for cyanide and also for alkali to ensure that it is in good condition for precipitation. This will generally be the case if the preliminary alkali wash has been properly carried out. For the treatment of silver ores and concentrates, strengths of 0.3 per cent, to 0.5 per cent. KCN (6 to 10 Ib. per ton) are commonly used, and in some cases as high as 1 per cent. Weak Solution. When the strong solution has sufficiently drained off, weak solution is pumped on from the storage tanks. It is often desirable to allow the charge to remain exposed to the air for some time before pumping on weak solution, and occasionally the surface is raked over, so as to allow access of oxygen to the material, to enable the weak solution to dissolve any gold not extracted by the strong solution. The liquor used for weak solution is commonly that which passes from the pre- cipitation boxes, used again without further addition of cyanide; but it may in some cases be necessary to bring up its strength, say to 0.15 per cent. KCN (= 0.06 per cent. CN, i.e., 3 Ib. KCN or 2J Ib. NaCN per ton of solution), by adding the required amount of stock solution from the dissolving tank. The weak solution is generally added in a number of succes- sive washes, gradually diminishing in strength to say 0.05 per cent. KCN ( 0.02 per cent. CN) ; these are either drawn off im- mediately or left in contact only a short time, a fresh wash being added as soon as the previous one has sunk well below the sur- face of the charge. More effective extraction, however, is gen- erally obtained by allowing each wash to drain as completely as possible before adding the next, though this is not always prac- ticable, owing to considerations of time. Unless the sand treated is absolutely free from slime, the rate of percolation gradually diminishes, in consequence of the settling of slime particles in the interstices of the charge and on the filter-cloth. The main object of the weak solution treatment is to displace the strong solution and thus remove from the charge the values which have already been dissolved. 204 THE CYANIDE HANDBOOK The amount of weak solution necessary varies greatly accord- ing to circumstances; 'the total volume used is generally con- siderably greater than that of the strong solution, and the time occupied is also greater. In some plants a medium solution is used after the strong solution, a special storage tank being em- ployed for this. Final Water-wash. As the liquid with which the charge is saturated after the last weak wash has been drawn off still con- tains gold in solution, it is the custom where practicable to dis- place this by addition of water. It will be obvious, however, since the liquid so displaced is added to the general stock of solu- tion in the plant, that the amount of water which can be used in this way is very limited. The amount of solution in stock is a fixed quantity, depending upon the . storage accommodation. The amount of final water-wash, therefore, cannot exceed the difference between the moisture originally present in the charge and that in the tailings as discharged, except in cases where a part of the solution is run to waste or where losses occur by evapo- ration. In practice, the final wash is generally given with the weakest cyanide solution available. Quantities of Solution Used in Percolation. No definite rule can be given as to the quantities of each class of solution that should be used, nor as to the time occupied by each in contact and leaching. These considerations depend partly on the arrange- ment of the plant and partly on the nature of the material treated. In an ordinary case, about 1J tons of liquid will pass through the vat for every ton of sand (dry weight) which it contains; say J ton of strong solution, 1 ton of weak solution, and | ton of alkali wash and final water-wash. It is convenient to have the size of the tanks such that one tank will contain the whole amount of sand produced per day. Thus, with a plant consisting of seven leaching tanks, one might be filled and one discharged every day, allowing five days for the actual cyanide treatment two days for strong solution and three for weak. Some such arrangement as this greatly simplifies the routine of operations in the plant. Upward Percolation. In the case of material containing slime, which is apt to settle on the filter-cloth and impede filtra- tion, one or more of the solutions may be introduced from below and forced upward through the charge by hydraulic pressure. When this is done slowly and carefully, the formation of channels PERCOLATION 205 is avoided and the filter-cloth kept free from slime. After the charge has stood covered with solution for a sufficient time, the liquid is again drawn off from below, and subsequent washes generally added from above and allowed to percolate downward. (C) DOUBLE TREATMENT The system of intermediate filling has already been described (Part III, Section III), and it has been noted that the collecting vats are sometimes used as preliminary treatment vats. Usually, the final treatment vats are placed immediately beneath the collecting vats, so that the material may be readily transferred from one to the other by bottom discharge doors. The general method is to give the alkali wash, if any, and at least a part of the strong solution, treatment in the upper vat, so that when the transfer takes place the material saturated with strong solution is freely exposed to the air. The remainder of the strong solution treatment and the final washing with weak solution and water takes place in the lower vat. In some cases, especially in the treatment of concentrates, it is found advantageous to transfer the charge a second or third time. The following scheme of treatment, based on a method pro- posed by W. R. Feldtmann for treatment of tailings at the Luipaards Vlei Estate/ will illustrate the method: FIRST TREATMENT: CHARGE, 165-170 TONS TAILINGS Tons Solution put on Strength: KCy Time Leaching: Hours Remarks Strong . . . Medium . . 27 27 0.25 per cent 0,20 " " 66 to 70 Drained off without Weak .... 27 0.15 " " standing Extraction by first treatment, 67 per cent. SECOND TREATMENT Tons Solution put on Strength: KCy Time Leaching: Hours Remarks Medium . . Weak .... Water 20 to 25 75 20 to 30 0.20 percent. 0.15 to 0.10 " " 179 In successive washes F. White, " Trans. I. M. M.," VII, 124 (1899). 206 THE CYANIDE HANDBOOK It is stated that in some cases a difference of over 20 per cent, in the extraction has been obtained through the introduction of double treatment. The method is extensively adopted in the treatment of pyritic ores, and in cases where refractory minerals are present, as in most silver ores. (See below.) (D) DISCHARGING OF TREATED MATERIAL (RESIDUES) Where facilities exist for the purpose, the treated sand may be rapidly and economically discharged by sluicing, using a strong jet of water. This method is frequently adopted for discharging treated slimes. (See below.) When the first cyanide plants were erected in South Africa, the only method of discharging the vats was by shoveling over the side into trucks, this procedure invol- ving great waste of labor, even with the shallow square tanks then in use. In 1891 the system of "bottom discharge" was introduced by Charles Butters, whereby the cost of discharging was reduced to about a quarter of that for the previous system. Tanks have also been constructed with doors in the side for dis- charging, but since it is evident that the material to be discharged must be thrown a greater average distance when the door is at the side than when at the bottom, the bottom doors have been almost universally adopted. The type of discharge door first designed by Butters and still in general use consists essentially of a circular steel or cast-iron FIG. 29. James Patent Discharge Door as furnished by the Cyanide Plant Supply Co. disk, fitting into a flanged cast-iron tube which passes through the filter-cloth, frame, and bottom of the tank. These doors are 9 to 16 in. in diameter, and have a bolt passing through the center, carrying a butterfly-nut underneath the tank, by which the door can be fastened or unfastened as required. (See Fig. 29.) PERCOLATION 207 A more recent type of door is hinged on one side and secured by a bolt and nut on the opposite side. In either case certain precautions are necessary to obtain a water-tight joint. When a tank is to be filled, the door is closed and the discharge tube (passing through the filter) is filled with sand, with a covering of clay or slime. In discharging, which is generally done by hand, the door is first unscrewed, a rod is pushed upward into the tank, and a hole is dug down to the opening. The material is then shoveled through the latter into a truck standing on rails beneath the vat. When the tanks are very deep an additional steel tube is sometimes attached to the upper edge of the discharge tube before the tank is filled, and when needed other sections can be added above this. It is then only necessary to dig down to the uppermost section before beginning to discharge. The extra tubes are removed in succession as the discharging proceeds. In places where labor is costly, various mechanical appli- ances have been introduced for discharging tanks. One of these, the Blaisdell vat excavator, consists of a number of revolving radial arms carrying steel disks inclined at an angle with the arms. The machine is placed over the top of the tank to be dis- charged, and as it revolves the material is pushed by the disks toward the center and passes through a discharge door to trucks or to a belt-conveyor beneath. A shorter pair of radial arms, also carrying disks, is used to counteract the tendency of the material to pile up toward the center. 1 1 For details, see Julian and Smart, loc. cit., p. 192. SECTION II AGITATION (A) AGITATION BY MECHANICAL STIRRERS IN the first plant erected in South Africa the experimental plant near the Old Salisbury battery, which was in operation in the early part of 1890 a system of agitation with paddles was used, combined with subsequent filtration of the pulp by suction. It was considered that a better contact of solution and ore, and consequently more rapid dissolution of the gold, would be ob- tained by the agitation system. This view was no doubt correct, but the method was very soon abandoned in favor of percolation, which required so much less expenditure of power and was far simpler in execution. Modern practice, however, shows a ten- dency to revert to agitation methods, and consequently percola- tion is a far less important factor in cyanide work than it was a few years ago. Many ores have been shown to yield a much higher percentage of their values when every portion is crushed at least fine enough to pass a 150-mesh sieve than when crushed moderately fine, say to 30 or 40 mesh, and a separation made of sands and slimes for different treatment. Ore crushed to 150- mesh cannot, strictly speaking, be regarded as slime, as it is generally possible to separate from it a considerable amount of fine sand which is perfectly leachable. It may, however, be con- veniently treated by agitation, and in many cases this system would be more profitable than hydraulic separation of the sands for percolation. The overflow from the spitzliitten or collecting tanks in some plants the entire mill-product reduced to a sufficient fine- ness generally goes first to a large spitzkasten, the overflow from which, consisting of practically clear water, is returned to the battery. The underflow from this, consisting of thickened slime, passes either direct to the agitation tanks or into special collecting tanks with conical bottoms, whence a further quantity 208 AGITATION . 209 of water is withdrawn. The moisture in the pulp withdrawn from the bottom of these tanks may in some cases be reduced by this means to 50 per cent. The agitation tanks are provided with paddles revolving about a central vertical shaft. The latter is supported by a platform running across the centers of the vats; at the upper end of the vertical shaft is a bevel-wheel, operated by a horizontal countershaft. When a sufficient charge has been collected, the agitator is set in motion, arid a sample taken of the well-mixed pulp. Lime is added, if necessary, and then sufficient cyanide solution to give a pulp of about 4 to 5 tons liquid for every ton of dry slime present. When the material is to be subsequently treated by filter-pressing, a much thicker pulp than this is com- monly used, say 1| to 1| tons liquid per ton of slime. The strength of solution depends on the material to be treated, but it is commonly much less than that required for percolation treatment of sands. In many cases a satisfactory percentage of the gold may be dissolved with solutions of 0.01 per cent. KCN (= 0.004 per cent. CN, i.e., 0.2 Ib. KCN or 0.15 Ib. NaCN per ton of solution, or 4 to 5 times these amounts per ton of dry slime). Slimes from pyritic or refractory ores, especially if silver or copper minerals be present, may require much stronger solutions. The agitation is continued until all soluble gold or silver may be assumed to be dissolved. The time required for this also varies greatly, from 3 to 36 or even 48 hours. In an ordinary case, 6 hours will be sufficient. In a thick pulp the values gen- erally dissolve more slowly than in a thin one, and the strength of solution also affects the result. An ordinary strength for the treatment of clean slimes is 0.05 per cent. KCN ( = 0.02 per cent. CN), but twice or three times this strength may be needed for refractory material. The agitator is then stopped; in some cases the stirring gear is so arranged that the paddles can be lifted above the level to which the settled slimes will reach. The tank is allowed to stand without further agitation till a sufficiently complete settlement has taken place, which may require from 9 to as much as 60 hours. As soon as a clear liquid shows above the surface of the settling pulp, decantation is begun. This is generally carried out by means of a pipe, jointed near the bottom of the tank and with the upper end supported by means of a float just below the surface of the liquid. The exit pipe passes out near the bottom 210 THE CYANIDE HANDBOOK of the tank. The pipe is lowered as the settlement proceeds, so that clear solution is continuously drawn off and passes on to the precipitation plant. When filter-presses or some form of suction filter are to be used, the pulp is transferred direct from the agitation tanks, without settlement, by means of hydraulic or atmospheric pressure, or by suitable pumps, to the filtering apparatus. The apparatus known as "montejus" is frequently used for this purpose. (B) AGITATION BY CIRCULATING THE PULP AND INJECTION OF AIR A method often used in place of, or in conjunction with, mechanical stirrers is to withdraw the pulp continuously from the bottom of the tank and return it to the top by means of cen- trifugal or other pumps. This method has the advantage of secur- ing a more complete mixture of the particles, especially when a little coarse material is present, than could possibly be obtained by any form of paddle agitator, and the further advantage that during the transfer the pulp is thoroughly exposed to the air and the action of the cyanide thereby accelerated. In cases where further oxidation is considered necessary, arrangements have been made for causing the pulp as it returns to the agitation tank to fall from a hight and to be scattered in the form of spray. Another device, due to H. T. Durant, which seems to be still more effective in securing oxidation, is to place a small suction valve on the intake pipe of the centrifugal pump; this draws in air, in the form of minute bubbles intimately mixed with the pulp, though of course with some loss in the lifting power of the pump. Sometimes agitation by injection of compressed air through a perforated pipe is also used. In this case the pipe is bent around the bottom of the vat, so as to allow air to pass through the charge at as many points as possible. In carrying out this method of agitation, the best results are obtained by using tanks that are very deep in comparison with their diameter. Tanks for this purpose erected at Pachuca, Mexico, are 15 ft. in diameter and 45 ft. high; 1 they are constructed with a conical bottom, the air being delivered through the apex of the cone under suffi- cient pressure to prevent settling of the pulp on the conical sur- 1 M. R. Lamb, " Notes on Air-agitation," " Eng. and Min. Journ.," LXXXVI, 901 (Nov. 7, 1908). AGITATION 211 face. The pressure required is from 22 to 35 Ib. per sq. in.; 15 to 20 cu. ft. of air are used per charge, which it is stated gives a vigor- ous agitation and excellent extraction. Slime, fine sand, and con- centrates are treated by this method. The apparatus used, known as the "Brown Agitator," 1 was first introduced at Komata Reefs, New Zealand. It acts on the principle of the air lift. The vat used is a long, narrow, vertical, cylindrical vessel, with conical bottom, from 40. to 55 ft. high and 10 to 15 ft. diameter, with a central column 1 in. in diameter for every foot of vat diameter, both ends of which are open and immersed in the pulp. Air is intro- duced through a narrower pipe opening at the bottom of the center column, at just such a pressure as will overcome the weight of a column of slime-pulp at the point of introduction, thus estab- lishing a circulation of the slime. The power required is 2^ h.p. per charge of 50 tons, which would be about the slime content of a 40 X 10 ft. vat. Apparatus on this principle has been intro- duced at the Waihi and Waihi Grand Junction, New Zealand; also, as above mentioned, at Pachuca, Mexico. (C) ANDREW F. CROSSE'S SLIME TREATMENT PROCESS The slime, after collection by any suitable method, with addi- tion of the necessary amount of lime and removal of excess of water, is washed by weak cyanide solution into a vessel consisting of two concentric steel cones, each having an angle of 45. The inner cone has a comparatively large opening at the bottom; its sides are parallel with those of the outside cone, and its diameter about half that of the outside cone. This inside cone acts as a baffle between the pulp, which is being agitated within it, and the outside portion, which is quiescent on the surface; it is sus- pended on two girders resting on the top of the outside cone. An air-lift in the center of the apparatus reaches nearly to the bottom of the outside cone; the air pipe for working this passes through the bottom of the cone, at which point there is also a suitable discharge pipe and cock. The upper end of the air-lift pump is several feet above the surface of the pulp, and the overflow runs into a cylinder having two or more horizontal discharge pipes, so arranged that the discharged pulp causes the pulp in the inside cone to revolve i " Min. Sc. Press," XCV, 689. 212 THE CYANIDE HANDBOOK and mix with the clear solution returning from the zinc-box. The slime-pulp, mixed with cyanide solution, is pumped or run into the inside cone, and as soon as this pulp reaches half-way up the air-lift pump, the air valve is opened, to prevent any settlement of slime. As the slime-pulp rises in the vat, the outside portion tends to settle and leaves several inches of quite clear liquid on the surface. The pulp is allowed to flow in until the cone is nearly full, then the two decanting arms are lowered, and the clear liquid drawn off and run through a zinc- box adjoining the vat at a slightly lower level. The effluent from the zinc-box is returned by a small pump to the inside cone, the whole process being thus very simple. A by-pass is provided to allow the clear decanted solution to circulate with- out going through the zinc-box, as the first liquid drawn off would contain no gold or silver. The time required for solu- tion and extraction depends on the nature of the material treated. The whole process is practically automatic and one shiftsman can attend to a number of vats. After treatment, the pulp is discharged into a settling tank and the clear liquid decanted from the settled pulp. If necessary, the strong cya- nide may be displaced by a weak solution by allowing the liquid leaving the zinc-box to run into a sump and running in a very weak solution at the same rate, so as to displace the stronger solution. This would be necessary in treating silver ores. Successful experiments with this method have been made at Frankfort (near Pilgrim's Rest, Transvaal) and at Johnson & Sons' smelting works, Finsbury, London. The advantages claimed are: high percentage extraction of gold and silver; simplicity and cheapness in working; very great saving in capital outlay. (D) TREATMENT OF SLIMES BY SETTLEMENT AND DEC ANT ATI ON This system was for several years the only one adopted on the South African gold-fields, and even to-day it is, in many parts of the world, the only method economically possible for handling very low-grade material that cannot be treated by percolation. It involves the use of large volumes of liquid, and consequently is not well adapted for countries like Western Australia, where water is very scarce. The method consists of successive agitation with cyanide solution, settlement and decantation, the cycle of operations being AGITATION 213 repeated twice or three times, and in exceptional cases oftener. The extraction obtained by each cycle depends (1) on the pro- portion of the total value which goes into solution under the given conditions; (2) on the relative volumes of the clear-settled liquid and the moisture retained in the residual pulp after de- cantation, it being assumed that the dissolved values are dis- tributed uniformly throughout the whole mass of liquid; (3) on the values originally present in the solution added; as the pre- cipitation is not absolutely perfect, some gold and silver will always be present in the liquid used. Suppose, for example, the slime to carry originally 5 dwt. per ton, of which 4.5 dwt. are actually dissolved during the agitation and settlement stages of treatment, and that the cyanide solution used in the proportion of 4 tons solution to 1 of dry slime carries originally 0.6 dwt. per ton. The settled slime after decantation retains, say, 50 per cent, of moisture. The total value dissolved consists of 4.5 dwt. per ton of dry slime, distributed over 4 tons of solution, or 1.125 dwt. per ton of solution, which in addition to the 0.6 dwt. originally present gives 1.725 dwt. The total gold drawn off is therefore 3 X 1.725 = 5.175 dwt. per ton of dry slime treated. The value left in the residue consists of 1 ton dry slime carrying 0.5 dwt., with 1 ton solution carrying 1.725 dwt., or 2.225 dwt. per ton of dry slime. The extraction, based on original contents of dry slime, is therefore 5 2.225 = 2.775 dwt., or 55.5 per cent. Suppose now that a second treatment is given with 3 tons of solution carrying 0.4 dwt. per ton. This, with the 1 ton of solu- tion carrying 1.725 dwt. remaining after the first decantation, gives after agitation and settlement 4 tons of solution carrying 0.731 dwt. per ton, assuming that no further dissolution of gold takes place. Then if, as before, the settled residue retains 50 per cent, of moisture, the result is 3 tons of solution drawn off, carrying 3 X 0.731 = 2.193 dwt., and a residue consisting of 1 ton dry slime at 0.5 dwt. and 1 ton solution at 0.731 dwt., or a total of 1.231 dwt. per ton of dry slime, bringing the total extrac- tion by two treatments up to 75.4 per cent. This example will serve to illustrate some of the weak points in the decantation system. It is obvious that a high extraction can only be obtained by a number of successive decantations, which involves the handling of a very large volume of liquid per 214 THE CYANIDE HANDBOOK ton of material treated. Ample storage room must be provided for the solutions, so that the first cost of the plant is very con- siderable. Good extractions, also, can only be obtained by using solutions from which the values have been very perfectly precipitated before use in the slime tanks. The percentage of moisture in the residues may be materially reduced and the extraction correspondingly improved by the use of very large, deep tanks, generally with a pointed bottom. After one or two treatments by agitation and decantation, as already described, the pulp is transferred to the large tank. When several charges have been thus introduced, the settled material at the bottom is under very great hydrostatic pressure and may be withdrawn as a very thick pulp carrying 40 per cent, or less of moisture. There is also an additional economy in the use of these deep tanks; as the final settlement takes place in them instead of in the agitation tanks, the latter are set free for the treatment of fresh charges sooner than would otherwise be the case. This system was first suggested by Charles Butters, and has been carried out on a large scale, both on the Rand and in other districts. By making the final settlement tank sufficiently large and introducing the pulp by means of a pipe passing some distance below the surface of the liquid contained in the tank, the process may be made continuous, the apparatus thus acting as a spitz- kasten and allowing clear liquor to overflow at the periphery, while the thickened slime is constantly drawn off at the bottom. Several systems of treating slime by continuous settlement have been devised, in which this principle is utilized. One has been described by E. T. Rand. 1 After removal of superfluous water by means of spitzkasten, the thickened pulp is first agitated with cyanide solution, then pumped to a settling-vat of suitable size and form, the thickened underflow from which, mixed with fresh solution, goes to a second settling-vat. The clear solution overflow from the first settler goes direct to the precipitation box; that from the second settler goes to the launder that feeds the agitation vat, and, with addition of fresh cyanide, forms the dissolving solution. The precipitated liquor is pumped to the launder by which the pulp passes from the first to the second settler, and forms the final wash. 1 " Proc. Chem., Met. and Min. Soc. of South Africa," II, 686 (1899). AGITATION 215 The Clancy system * is very similar in general design, but the agitation tank is dispensed with and more settlers are used. None of these methods appears to have come into general use, probably owing to the difficulty of obtaining constantly a residue sufficiently low in solution contents to be economically discharged; it is not likely that the final residue would average as little as 50 per cent, moisture, a condition easily attained with the ordinary settlement and decantation process. i " Proc. Chem., Met. and Min. Soc. of South Africa," II, 741. SECTION III FILTRATION BY PRESSURE AND SUCTION (A) ADVANTAGES OF THE SYSTEM As already pointed out, the methods of treating slime or finely divided ore by settlement and decantation involve the use of tanks of large capacity for the settlement of the pulp and for the storage of solutions. It has been shown also that a high extraction could be obtained only by repeating the cycle of operations several times. By using some method of filtration by pressure or suction, the greater part of the dissolved values may be extracted in one operation, in far less time and with the aid of a much smaller quantity of solution than is required for the decantation system. In the latter system, much more time is consumed in separating the solution carrying the values from the insoluble residue than in actually dissolving these values; whereas by the system now to be described the separation of so- lution from pulp may be made with little delay as soon as a suffi- cient percentage of the values has been dissolved. On the other hand, the expense of maintenance and repairs and the cost of labor and supervision are probably very much higher for any system of pressure or suction filtering than for the decantation process. Consequently the latter is almost invariably adopted where finely divided low-grade material has to be dealt with. In the filter processes the quantities of material dealt with in one operation are necessarily very much smaller than in the decantation process. (B) FILTER-PRESSES General Principles. Filter-presses consist essentially of a number of cells or elements into which the material to be filtered is forced by hydraulic or atmospheric pressure. These chambers are lined internally with filter-cloths, through which the filtered 216 FILTRATION BY PRESSURE AND SUCTION 217 liquid is forced as a result of this pressure, passing out through channels in the walls of the cells, either to separate outlets or to a common solution-outlet channel. Details of Filter-press Construction. The following account is condensed from a description given by Clement Dixon. 1 "A filter-press consists of a number of hollow frames (from 30 to 50 in a 6-ton press) placed alternately between solid flanged plates, with filter-cloths of strong duck material between each. The hollow frames in a 6-ton press would be 3 ft. 6 in. or 4 ft. square inside, and 2, 3, or 4 in. in depth or thickness, as required for the particular ore treated. It will thus be seen that we have in reality a number of little vats standing on end, the hollow frames forming the sides, and the solid plates and filter-cloths the bottoms. Each of the hollow frames (which receive the slimes) is connected by a slot with the slimes inlet passage, and every alternate solid plate is connected by a slot with the wash-solu- tion inlet passage; these solid flanged plates act as filter-bottoms for admitting wash solutions to the slime cakes. The remaining solid plates are connected with the wash-solution outlet passage and serve to carry away the wash solution after it has passed through the slime cakes. On each of the solid flanged plates is a cock, used for draining the press when filling same with slimes." This description applies to the Dehne and other similar types of press, which, however, show various modifications in the less important details of construction. Illustrations are here given of the plates and frames used in the " Cyanippus " press, furnished by the Cyanide Plant Supply Co. ; in this type the use of drainage cocks for the separate chambers is dispensed with. In these figures a is the channel through which the slimes pulp is intro- duced and which is connected only with the frames (Fig. 32); b is the channel for introducing wash solution, and is con- nected only with each alternate plate (those known as high- pressure plates (Fig. 31); c is the channel which receives the solution that has been forced out through the filter-cloths in the case of the high-pressure plates; d is a separate channel serving as an outlet for the filtered solution on the low-pressure plates only (Fig. 30) . A general view of filter-presses in use for treatment of slimes at Lake View Consols, West Australia, is also given (Fig. 33). 1 "Proc. Chem., Met. and Min. Soc. of South Africa," III, 13. 218 THE CYANIDE HANDBOOK FIG. 30. Cyanippus Filter Press, Low-pressure Plate. FIG. 31. Cyanippus Filter Press, High-pressure Plate. FIG. 32. Cyanippus Filter Press. Frame. FILTRATION BY PRESSURE AND SUCTiON 219 To facilitate the passage of solution through the cloths, the surfaces of the plates are grooved or corrugated. Some mechanical arrangement is also required for tightening the plates and frames while the press is in use. This is generally done by means of screws and levers at one end of the press. FIG. 33. Filter Press Slimes Plant at Lake View, Consols, W. Australia. Showing type of press furnished by the Cyanide Plant Supply Co. Methods of Filling Filter-presses. Although some solution of gold and silver ntay take place during the process of pressing, it is generally better practice to make sure that a sufficient percentage has been dissolved before transferring to the presses. Hence some system of agitation is commonly employed, after which the pulp, thickened to the proportion of about one part dry slime to one part of solution, is forced by compressed air or otherwise into the filter-presses. For this purpose a montejus, an air- tight cylindrical vessel, is often used. This is provided with inlets at the top for the pulp and compressed air, and an outlet pipe, reaching nearly to the bottom, for conveying the pulp to the slime-pulp feed channels of the presses. The pressure re- quired may be from 30 to 75 or even 100 Ib. per square inch. In other cases the presses are filled by means of a pump; the pumping is continued until the cells of the press are filled with 220 THE CYANIDE HANDBOOK slime under a sufficient pressure to give an efficient filtration say 75 to 80 Ib. per square inch. It is important to increase the pressure steadily and uniformly; hence some form of air chamber is commonly necessary. Treatment in the Press. After the solution originally present in the pulp has been forced out as much as possible, the pulp inlet channel is closed and fresh solution is forced in under pres- sure by means of a pump. This passes by the solution inlet channel (b, Fig. 31) into the chambers of the presses and through the slime cakes, passing out through the outlet channels c and d (Figs. 30 and 31). Wash-water may subsequently be introduced in the same way, and continued until the effluent is sufficiently low in values. In some presses a special inlet channel for wash- water is provided. Arrangements are also made by which com- pressed air may be forced into the press, in order to remove as much liquid as possible at the end of the operation. It is stated that by this means the residual moisture may be reduced in some cases from 25 to about 14 per cent. Discharging Filter-presses. When the operation is complete, the closing-screw of the press is loosened and the plates and frames drawn apart. The slime-pulp remains usually in the frames in the form of a square cake, which is discharged by tilting or tapping the frame, dropping the pulp through a discharge hopper into trucks running beneath the presses. After each set of operations, the plates, frames, and filter-cloths are washed and cleaned, and the press again closed ready for a fresh charge. Capacity of Filter-presses. The size varies greatly according to the work required. There may be 20 to 50 frames, with a total capacity equal to 1 to 5 tons of dry slime per press charge. In West Australia an average of 6 charges are treated per day in each press, one cycle of operations occupying 4 hours. 1 Where conditions allow, the discharge of the treated slimes may be made by sluicing. In The Merrill Press, 2 designed by C. W. Merrill and introduced at various plants in the Black Hills, South Dakota, arrangements are made for introducing water under pressure, whereby the treated slime may be sluiced out of the frames without opening the press. The following particulars relating to this type of press are summarized from 1 Julian and Smart, loc. cit., p. 258. 2 " Eng. and Min. Journ.," LXXXI, 76 (1906). FILTRATION BY PRESSURE AND SUCTION 221 an account given by F. L. Bosqui. 1 The press is of a common flush plate and distance-frame pattern, but consists of larger units, the dimensions being: Number of frames 92 Size of frame 4 X 6 ft. Length of press 45 ft. Capacity 26 tons Weight of press 65 tons Thickness of cake 4 in. In addition to the ordinary channels for introducing slime- pulp and solution, there is provided at the bottom of the frames a continuous channel, within which lies a sluicing pipe with nozzles projecting into each compartment. This pipe can be revolved through an arc of any magnitude, so as to play a stream into any part of the cake, washing it down into the outlet channel. When the press is being filled, and during the cyanide treatment, the discharge ends of the pipe are sealed. The mode of operating is as follows: The slime-pulp, 3 parts water to 1 of solids, is charged by gravity to the presses under about 30 Ib. pressure. The cyanide treatment is carried out in the press itself, the effluent solution going to four precipitating vats, where the gold is recovered by zinc-dust. There is no power cost for agitating or elevating slime-pulp, but only for elevating solution to the press. The quantity of solution required is 0.6 ton per ton of dry slime, of which only 0.3 ton is precipitated. The power required is T V h.p. per ton of dry slime treated. Four tons of water per ton of slime are required for sluicing. All filtering is done by gravity, at a cost of 2 cents per ton. In ex- perimental tests on this system a recovery of 91 per cent, of the original value of the slime was obtained. (C) FILTRATION BY SUCTION Early Applications. As already mentioned in Section II, A, the principle of suction as an aid to filtration was employed at an early stage in the development of the cyanide process. In the first plant where the treatment of slimes was attempted on a working scale, namely at the Robinson slime plant, erected by Charles Butters, 2 the slime pulp, after being agitated sufficiently " Min. Sc. Press," Dec. 15, 1906. 2 " Proc. Chem., Met. and Min. Soc. of South Africa," II, 241 (Feb. 1898). 222 THE CYANIDE HANDBOOK to dissolve the gold, was run on to a filter-vat, provided with stirring gear that could be raised or lowered; the material was allowed to settle, the clear liquid was decanted from above, while at the same time suction was applied below the filter-cloth. By this means about 10 tons of slimes could be settled, decanted and sucked dry by a vacuum pump to about 28 per cent, of moisture, and would remain on the filter-cloth as a tough leathery layer about 6 to 8 in. thick. Fresh solution was applied and the paddle gradually lowered while revolving at the rate of about 16 r.p.m., so that in half an hour the slimes were washed into a pulp again. After further dilution, the process of simultaneous settlement, decantation, and suction by vacuum pump was re- peated. The method was abandoned owing to the wear and tear on the filter-cloths and the cost of keeping the suction pumps in order, together with the power and attention required for running them. Filtration aided by suction is used also in some cases in the treatment of fine sands by percolation and in the direct cyaniding of dry-crushed ore. For this purpose the material is generally treated in comparatively shallow vats and suction applied by the use of a steel cylinder connected with the bottom of the tank, below the filter-cloth, by a pipe, furnished with a cock. This cylinder is also connected with a vacuum pump by a separate pipe. The communication with the tank is first shut off and the steel chamber exhausted by the vacuum pump. The valve connecting with the latter is then closed and the cock below the filter-tank is opened, the solution from the latter being sucked into the exhausted cylinder. The advantage of this arrange- ment is that the solution from the tank does not pass direct through the vacuum pump, so that the valves of the pump are not liable to injury from grit passing through the filter. The Moore Filter. Among the more modern appliances for filtering by means of suction may be mentioned the Moore filter, the salient feature of which is that the filtering apparatus is movable and is transferred from one tank to another as required. The following description is quoted from an article by R. Oilman Brown, on " Cyanide Practice with the Moore Filter/' 1 The filter consists of a frame or "basket" carrying a number of "plates." Each plate is composed of a double thickness of 1 " Min. Sc. Press, Sept. 1, 1906. FILTRATION BY PRESSURE AND SUCTION 223 canvas, of medium quality, 5 X 16 ft., sewed around three edges. The fourth (long) edge, forming the top of the filter, is bolted between strips of wood 1^ X 6 in. section. In the bottom edge of each filter-plate is a J-in. channel-iron which serves as a launder for collecting the filtered solution. The filters are stitched ver- tically through both sides at 4 in.-distances, and in the compart- ments so formed are placed J X 1 in. strips to allow circulation. Each plate is provided with a 1-in. vertical suction pipe, dipping into the launder at the bottom, the end of the pipe being flattened for this purpose. The upper end of this pipe is connected by means of suction hose with a 3-in. manifold and vacuum pump. There are 49 of these plates in each basket. The mode of operation is as follows. The basket is lowered into a vat full of pulp until the upper edge of the plates is sub- merged. The suction pump is then started. If the solution drawn through is at first muddy, it is returned to the vat. In order readily to locate any leaks in the filter-plates, a glass nipple is placed on the pipe connecting each plate with the manifold; any plate showing leakage may thus be at once cut off. Suction is maintained with intermittent agitation till a sufficient coat of slime, averaging f in., is obtained, the time required being about one hour. The basket, with the suction pump still running, is then raised by a traveling crane, transferred to a wash tank and lowered therein. Water is sucked through the slime layer ad- hering to the plates until the effluent is sufficiently low in cyanide (approximately 0.0075 per cent.), about 0.7 tons of wash-water being needed per ton of dry slime. The cake then contains 40 per cent, of moisture. The basket is now raised again and trans- ferred to the discharge hopper; the suction is continued until excess of moisture is removed; the action is then reversed and an air pressure of 35 Ib. per square inch given in successive blasts for a few seconds, causing the cakes to drop off. The residual moisture is about 32 per cent. When it is necessary to clean the filters, water pressure instead of air is applied until the pores of the cloth are freed from slime. This is done about every alter- nate day. In some cases a second cyanide treatment is given before water-washing. The apparatus is charged with thick slime- pulp in the first tank, transferred to a second containing weak solution, where the suction is continued for a time, say 20 to 40 224 THE CYANIDE HANDBOOK minutes, then to the water tank, and finally to the discharge hopper. In general, it is found that water pressure is preferable to air for discharging, as it brings the slime cake off in one mass instead of in patches. One of the difficulties encountered in using this system was the imperfect adhesion of the slime during transfer, owing to jarring and partial drying. This caused un- even suction and consequent poor extraction. Among the plants where the Moore system was first adopted may be mentioned the Lundberg and Dorr Mill, Terry, South Dakota; the Liberty Bell, Colorado, and the Standard, Bodie, California. It was originally introduced by G. Moore at the Con- solidated Mercur mines, Utah, in 1902 or 1903. The Butters Filter. This apparatus is in many respects similar to the Moore filter. The two systems have the following points in common: (1) The filtering apparatus is immersed in the p'ulp that is to be filtered; (2) the solution is withdrawn by suction, by the use of a vacuum pump; (3) the cakes of slime formed adhere to the outside of the filter-plates; (4) a large filtering area is obtained in a cheap and compact form. The essential difference is that in the Butters system all the operations are conducted in one vat. The filter itself is station- ary, but so arranged that the separate plates or "leaves" may be lifted out if necessary. The pulp is first brought to the proper consistency by settlement and decantation; the thickened pulp is then run into a special rectangular filter-box, in which the apparatus is suspended. Suction is applied until a sufficient coating of slime is obtained ; the surplus pulp is then withdrawn from the filter-box by suitable pumps arid returned to the settling tank or slime reservoir. The box is now filled with weak cyanide solution, which is sucked through as long as may be necessary for extraction of the values. The surplus solution is then pumped back to the storage tank, water substituted, and suction con- tinued till soluble values are extracted. Finally, the suction is cut off, water pressure is applied to force off the adhering cakes, which fall into the filter-box and thence through hoppers with sides inclined at a sharp angle, whence they are discharged into a launder by means of a gate-valve at the apex of each hopper. On closing the valves, the filter-box may be again charged with slime-pulp. The construction of the " leaves" in this system differs some- FILTRATION BY PRESSURE AND SUCTION 225 what from that of the " plates " in the Moore system. A detailed description by E. M. Hamilton is given by Julian and Smart; l it need only be said here that each leaf consists essentially of an oblong frame suspended from a wooden bar forming the upper side of the leaf. Its remaining three sides are formed by a pipe bent at right angles and perforated along its upper edge in the part forming the bottom of the frame. The inner space is filled with cocoanut matting. The whole is enclosed in canvas, stitched vertically as in the Moore system, and forming the filtering sur- face to which the slime adheres. The perforated pipe is closed at one end and communicates at the open end with the vacuum apparatus, and also with the water tank used when water pres- sure is required for discharging. The Butters filter, which is the result of experiments carried out by Charles Butters and staff, has been installed at the Butters plant, Virginia City, Nevada; at the Butters Copala Syndicate Mill. Mexico; at the Combination Mill, Goldfield, Nevada, and elsewhere. At the first-mentioned plant it is stated that 150 tons of slime are treated per day at a cost of 11} cents per ton. The Cassel filter is of very similar design to the Butters filter. The chief distinction is that the filter-cloth is separately supported on a movable frame. At the end of the operation the slime-cake is detached by oscillating this frame. The following particulars of the costs and power required for working the Butters system are given by F. L. Bosqui : 2 At the 200- ton plant, Virginia City, Nevada, 10 cents per ton; At the 40-ton plant, Combination Mill, Nevada, 45 cents per ton. The latter could be reduced to 31 cents per ton by working the plant at its full capacity of 56 tons. Former cost of filter-pressing was $1 per ton. The power required at the Combination Mill is: For filtering 9 h.p. For agitating slime-pulp 3 " For pumping and other purposes 9 " The pumps required are: one 4-in. Butters centrifugal pump; one 12 X 10-in. Gould's vacuum pump; one 2-in. centrifugal pump for raising filtered solution to clarifying filter-press; and Loc. tit., p. 246. 2 " Min. Sc. Press," Dec. 15, 1906. 226 THE CYANIDE HANDBOOK one 2-in. centrifugal pump for returning slime overflow from leaching vats to slime settlers. 1 Traveling Belt Filters. As early as 1893 it was suggested by W. Brunton to use an endless traveling belt of filter-cloth, passing over vacuum chambers, for filtering slime-pulp, the slime being spread over the cloth at the head of the machine. Several other filters on the same or somewhat similar lines have been suggested. 2 THE CRUSH FILTERS The courtesy of the Cyanide Plant Supply Co. permits the following description of these appliances, of which that company controls the patents. They are of two types, distinguished as the "Traversing Filter" and the "Fixed Immersed Filter." The Crush Traversing Filter This system was elaborated by Barry and Banks at the Waihi mine, New Zealand, and in its general features somewhat resembles the Moore filter, with the important exception that no reverse pressure of air or other fluid is necessary for discharging the cakes, and that the material to be filtered is agitated during filtration. Each basket carries a number (generally ten) of filter-plates or leaves hung between two joists, and is suspended from a traver- sing crane, by which it may be raised or lowered and transferred from the pulp tank to the wash tank or to the discharge hopper, as required. Each filter-plate consists of a sheet of thin gal- vanized iron, carried by a wooden bar and enclosed in a frame of iron piping. The whole is enveloped in cotton twill or duck, between which and the iron may be placed cocoanut matting or other filtering medium if required. The plates are kept apart by the supporting wooden bars or by wires arranged at an angle. Each frame communicates by flexible rubber tubes with a main, this being further connected by flexible hose with a vacuum reservoir or pump. The frames are also connected to another main having a valve opening to the atmosphere. (See Figs. 33a and 33 b.) The mode of operation is similar to that of the Moore filter. 1 As the filter-cloths in time become choked with lime and other salts, it is necessary at intervals to immerse the leaves in dilute hydrochloric acid, which dissolves the deposits and thus restores the efficiency of the filter. 2 See Julian and Smart, loc. cit., p. 251. FILTRATION BY PRESSURE AND SUCTION 227 (See above.) When the basket is lowered into the pulp tank, the valve on the atmospheric main is left open to allow the displaced air to escape. As soon as the filter is immersed, the valve on the atmospheric main is shut off and that on the vacuum main turned FIG. 33a. Crush Traversing Filter immersed in Pulp and Solution Tanks. on. The pulp is agitated with compressed air or otherwise to prevent settlement during the filtering operation. The time required for formation of the cake is 30 to 45 minutes, and for extraction in the wash tank about 30 minutes. The whole cycle of operations, including charging, discharging, and transferring, takes about 2 hours. The cake varies in thickness from J to 1J in. In discharging, the vacuum is maintained until 228 THE CYANIDE HANDBOOK a crack begins to form at the junction of the supporting bar with the corrugated plates; the vacuum is then cut off and air ad- mitted. The cake immediately discharges itself with a rolling motion, and after a little cleaning the basket is ready for a fresh charge. FIG. 336. Crush Traversing Filter, raised for transfer- ring or discharging. During the drying operation before discharging, the inrush of air has the effect of lowering the vacuum and so interfering with the suction on the other baskets; it is therefore advisable either to have a separate vacuum pump for each basket or a separate connection with high- and low-pressure pumps, the latter, giving a vacuum of 5 to 10 in., to be applied as soon as the filter is raised out of the wash tank for discharging. The frames are made in two sizes 16 X 4 ft. and 10 X 5 ft., with 10 frames in each basket. Each basket of the first type treats 40 to 60 tons of slime per day; the filter-cloth requires renewal about once in 6 months, 1 sq. yard of cloth treating about 4 tons of slime. The power required for the crane is 10 h.p.; the vacuum pump FILTRATION BY PRESSURE AND SUCTION 229 requires 6 h.p. per basket. When air agitation is used, additional power is needed. (See Figs. 33a and 336.) The Crush Fixed Immersed Filter In this type the filter is stationary and the crane is dispensed with. In principle it resembles the Butters and Cassel filters. The pulp is agitated in the filter-box by means of centrifugal pumps, and after a sufficient thickness of cake has accumulated on the plates, the surplus pulp is returned to a pulp storage-tank (a, Figs. 34a, and 346.) This type of apparatus is claimed to have a Pulp Tank b Pulp Filter c Gold Solution Tank d Centrifugal Pump e Vacuum Pump FIG. 34a. Plan of Crush Filter Plant. (Fixed immersed type.) the drawback that there is a danger of untreated material being discharged, as the filtration and discharge of the slime both take place in the same vessel. This can only be obviated by carefully attending to the washing out of the pulp-box (b, Fig. 34 a), after removing the surplus pulp when the cake has formed. These washings should of course be returned to the general stock in the supply tank, which, however, may thus become unduly diluted. The construction of the filter-plates is identical with that of the traversing type of filter, but no hessian or cocoanut matting 230 THE CYANIDE HANDBOOK is used in the leaves, and it is claimed that the latter are not liable to be choked with lime or fine particles. It is advisable in this case also to have high- and low-pressure vacuum pumps in con- nection with the frames, the former to be used during the actual filtering operations and the latter while filling or emptying the pulp-box. Too high a vacuum while the plates are exposed to air would cause cracks to form in the slime cakes, rendering the subsequent washing inefficient. Another drawback pointed out by the Cyanide Plant Supply Co. is the liability of the cakes to drop off during the rather long period occupied in emptying the tank of surplus pulp and refill- ing with solution or wash-water. This renders the method un- a Pulp Tank b Pulp Filter c Gold Solution Tank d Centrifugal Pump c Vacuum Pump FIG. 346. Elevation of Crush Filter Plant. (Fixed immersed type.) suitable for sandy or granular matter. The difficulty may be remedied by the use of very large pumps for transferring pulp and solution, but this adds greatly to the cost of the installation. The capacity of a standard plant on this system is about six to eight charges per day, say 100 tons of dry slime, or about 1 ton per day for every 30 sq. ft. of filtering area. The power required is about 9 h.p. when the slime-pulp is fed by gravity to the filter- tank, and about 30 h.p. when pump circulation is used. The Ridgway Filter A further extension of the idea of a traveling suction-filter, first introduced in the case of the Moore apparatus, already de- scribed, is seen in the Ridgway filter, of which illustrations in plan and section are here given (Figs. 35 and 36). The carriers bearing the filter-plates revolve about a central hollow axis and travel along a circular track in such a way that during part of the circuit the filter-plates are immersed in slime-pulp. The latter, containing the necessary amount of cyanide for dissolution of the values, is fed continuously into a section of the annular space between the tracks, and a portion of the pulp adheres to each FILTRATION BY PRESSURE AND SUCTION 231 plate as it passes through, hi consequence of the suction. Every plate is connected by a separate pipe with the central hollow shaft about which the apparatus revolves; this communicates with FIG. 35. Plan of Ridgway Filter. [From illustration furnished by the Cyanide Plant Supply Co.] the receiver of a vacuum pump, by which the filtered solution is continuously withdrawn. At a certain point in the circuit, the outer edge of the track ascends so that the traveler rises out of Compressed Air Strong Solution ounm Pump FIG. 36. Sectional Elevation of Ridgway Filter. the slime-pulp; beyond this point it descends again into a separate section of the annular trough, which is kept filled with water. As the filter-plates travel through this water, the latter is sucked continuously through the adhering slime cake, displacing the 232 THE CYANIDE HANDBOOK solution. Finally, the outer track again rises so that the carrier emerges from the liquid and at the same time the suction is automatically cut off, and compressed air applied by separate pipes connected with each filter-plate. As a result of this pres- sure, the slime cakes drop off and are received in a discharge hopper at this point. The carrier then descends again into the section of the apparatus containing the slime-pulp and the cycle of operations is repeated. This apparatus is also furnished by the Cyanide Plant Supply Co. The following additional particulars with regard to the Ridg- way filter may be of interest. 1 The first trials with the machine were made at the Great Boulder mine, Kalgoorlie, West Australia, January, 1906. There are 12 to 14 radial arms or levers, each bearing a flat cast-iron filtering frame, corrugated on the under surface; to these corrugated surfaces screens are attached over which ordinary filter-cloth is fixed, giving 4 sq. ft. of filtering area per frame. The central column, in its lower part, has two internal compartments, each connected separately with a vacuum pump and serving as channels for withdrawing the effluent gold solution and wash solution respectively. The upper portion of the column is either connected with an air compressor or is con- structed to act as an air compressor, driven by the machine itself. Each filtering frame is connected by 3 radial pipes and rubber hose with the two compartments of the central column and with the compressed-air section. The outer ends of the levers are provided with rollers running along a circular track at the edge of the annular trough which holds the slime-pulp and wash-water. The frames are suspended horizontally from the levers in such a way that the lower surface only is immersed; at the points between the sections of the trough, where the track is raised, the filtering frames emerge from the liquid, and at the same time one of the three valves connecting with the central column is closed; on descending again, another valve is opened. These valves are automatically operated while the machine is in motion by small rollers passing over elevations on one or other of three concentric tracks. The series of opera- tions is as follows: After the frame has passed the discharge chute, it descends *" Monthly Journal Ch. of Mines," West Australia, Nov., 1906. See also "Min. Sc. Press," P'eb. 9, 1907. FILTRATION BY PRESSURE AND SUCTION 233 into the slime-pulp; at the same time the gold-solution valve is opened by a roller running up an incline on its track, opening direct connection between the filter-frame and the vacuum which is maintained in the center column. When the frame ascends from the pulp, the gold solution valve is closed; as it descends into the wash solution, the wash valve connected with the other section of the center column is opened, causing the vacuum to draw wash solution through the adhering pulp. The frame con- tinues in the wash trough for half a circle; when it again rises, the wash valve is not closed till the discharge chute is reached, so as to allow the slime cake to be extracted by suction as thor- oughly as possible. When the discharge chute is reached, the wash valve closes and the air valve opens, admitting a puff of compressed air, which enters the filter-frame and causes the resi- due to drop off into the chute and thence into a truck. The machine makes about one revolution per minute. Air-tight con- nections between the rotating and stationary parts are secured by means of glands and stuffing-boxes. The supply of pulp, wash solution, and water to the annular troughs is regulated by float valves in the receiving compart- ments, which are placed outside the peripheral track. The pulp is kept from settling by the revolution of small agitators driven independently of the other parts of the machine. From 0.8 to 1 ton of wash solution is used per ton of dry slime : the residues carry 28 to 33 per cent, of moisture, showing only a trace of dissolved gold and cyanide. An extraction of 89 per cent, is claimed in the treatment of old battery slime. The vacuum pump requires 3 h.p., the agitators 1 h.p., and the filter machine J h.p., making a total of 5 h.p. The capacity is 25 to 66 tons per day, according to the nature of material treated. The Bertram Hunt Filter The following account of this apparatus is summarized from description given by the inventor. 1 The filter-bed is composed of gravel and sand, resting on wooden slats and supported on an annular concrete structure, raised at the edges 2 in. above the filter-bed. The slats are i to f in. apart. The top layer of the filter is composed of sand of 8 to 12-mesh size, and is one " Min. Sc. Press," XCVII, 430 (Sept. 26, 1908). 234 THE CYANIDE HANDBOOK inch in thickness. A vacuum filter is applied beneath, but no filter-cloth is used. A carriage bearing scrapers revolves about the central hollow pillar, and travels over the annular filter- bed, being supported on the raised edges of the concrete bottom. By means of this carriage, the material to be treated is con- tinuously fed onto arid removed from the filter-bed. The appa- ratus carries a scraper in front, which removes the upper layer of slime and transfers it to a screw conveyor, by which it is dis- charged or conveyed to the central hollow pillar, whence it is discharged by sluicing. Behind the scraper is a distributor for the sandy portion of the pulp, followed by another for the slimy portion. Revolving pipes to spray solution and wash-water fol- low the carriage at suitable distances. The outside diameter of the machine is 15 ft., the annular filter-bed being 3 ft. wide, thus giving a filtering surface of 113 sq. ft. The speed is 1 r.p.m. Taking the quantity of material discharged during each revolution as represented by a layer J-in. thick, say 2.26 cu. ft., carrying 50 per cent, moisture, or 123 Ib. of dry residue, the capacity of the apparatus would be 80 tons per day. The power required to drive the machine is about 1 h.p., and for the vacuum filter 4 h.p., or a total of 5 h.p. It is stated that a clear filtrate is always obtained. Whenever it is desirable to clean the permanent filter-bed, the adjustable scraper is lowered and the sand above the slats removed. Clean sand is then spread to the required depth. The apparatus can be arranged with a separate vacuum for strong and weak solutions. SECTION IV HANDLING OF SOLUTIONS (A) SUMPS AND STORAGE TANKS IN addition to the vats or other vessels used for holding the material under treatment, storage room must be provided for the solutions, before and after use. The volume occupied by the solution may be taken as about 31 to 32 cu. ft. per ton. The volume occupied by ordinary tailings varies from 20 to about 26 cu. ft. per ton, according to the density with which it is packed into the leaching vats. It is impossible to give any fixed rule as to the capacity required for storing solutions, but it may be taken as about equal to the vat capacity required for the material actually undergoing treatment at any one time. As already mentioned, there are in general three classes of solution, for which separate tanks must be provided, namely, strong cyanide, weak cyanide, and alkali. The storage tanks are generally similar in construction to the leaching tanks, but without filter-frames, and may be of wood or iron. It is usual to make them deeper in proportion to diameter than the leaching tanks, as this form is more economical in con- struction, and there is in this case no limitation of depth on account of difficulties in percolation. They are generally placed either above the level of the leaching vats, so that the solution can pass from them to the latter by gravity, or below the outlet of the precipitation boxes, so that the precipitated solution flows into them by gravity. In the latter arrangement the storage tanks are commonly described as "sumps." In the first case, arrange- ments must be provided for continuously pumping the precipi- tated solution as it leaves the boxes up to the storage tanks. The solution flowing from the boxes runs into a small cement- lined depression or sump in the floor of the extractor house, from which it is pumped at the same rate. In the second case, with the storage tanks below the precipitation boxes, every solution 235 236 THE CYANIDE HANDBOOK run on to the leaching tanks must be pumped up, as required, but the pumping will not be continuous. It is obvious that the same amount of liquid must be raised in both cases; but in the first case the average hight of lift will be greater, as all the solution must be raised from the level of the boxes to the top of the storage tanks, whereas in the second case the average lift will be from half the depth of the storage tanks to the top of the leaching vats. Dissolving Tank. A special small tank is generally provided for dissolving the solid cyanide and preparing a concentrated solution, from which the necessary amount is drawn off as re- quired for making up the solutions in the storage tanks. (B) SETTLING AND CLARIFYING TANKS It frequently happens, for various reasons, that the solutions leaving the leaching vats, or decanted from slime-settling tanks, and also those drawn off from filter-presses or suction filters, are not perfectly clear. Consequently it is the usual practice to run such solutions into a separate tank, instead of allowing them to flow direct to the head of the precipitation box. When the liquid has settled clear, it is drawn off by syphon or otherwise, and the introduction of sand or other suspended matter into the boxes is avoided. The arrangement has the additional advantage that a perfectly regular flow of liquid through the boxes can be main- tained, as the rate of flow need not depend on the rate of leaching, decanting, etc. In some cases it is advisable to allow the solu- tions to percolate through a sufficient thickness of clean sand, in what is described as a clarifying tank. Other materials, such as cotton waste, coir matting, etc., are likewise used for clarifying. Sometimes the solution is introduced from below and caused to percolate upwards through the clarifying medium. Another arrangement sometimes adopted is to pass the liquid to be clarified through a special small filter-press adapted for this class of work. An account of this method is given by Truscott and Yates. 1 Previous to the introduction of the clarifying press, at Redjang Lebong, Sumatra, slime had been carried into the zinc-boxes, notwithstanding the use of several successive settle- ment tanks and the addition of lime in the launder from the filter- 1 " Journ. Chem., Met and Min. Soc. of South Africa," VII, 3 (July, 1906). HANDLING OF SOLUTIONS 237 presses, and caused much trouble and expense in the smelting of the precipitate. It was found that the solution could be effectively clarified by passing it through two small Johnson presses; these were opened twice a week and the cloths well scrubbed to free the pores from fine slime. The beneficial effect of the filter-press clarification upon the solution was at once felt in the smelting; the precipitate became clean and was easily reduced, the smelting costs being reduced from 3.07 pence to 0.87 pence per ounce of fine metal recovered. (C) PUMPS The pumps commonly used for transferring solutions or for circulating slime-pulp are centrifugal pumps, the size varying with the work required. A 4-in. pump is a convenient size for many purposes. When the pumping required is for the con- tinuous transfer of small quantities of solution, several small pumps, one at least for each class of solution, will be needed; this will be the case when the storage tanks are above the leach- ing tanks. Where pumping is only necessary at intervals and the object is to transfer large quantities of solution rapidly, one or more large pumps are necessary. This is the case when the storage tanks are below the leaching vats. Pumps with brass or copper fittings should be avoided; all parts liable to come in contact with solution should be of iron or steel. For special purposes other types of pump are used. For transferring sandy or gritty material the " air lift " is com- monly employed. Some account of the Frenier sand pump has already been given. (See Part III.) The power required for trans- ferring slime-pulp is considerably greater than that needed for solution. (D) PIPING FOR CONVEYING SOLUTIONS Iron pipes are used for transferring solution (1) from the storage tanks to the leaching tanks, agitation tanks, or filter- presses; (2) from the treatment tanks to the precipitation boxes or clarifying tanks; (3) from the precipitation boxes to the storage tanks. When, as is usually the case, several different kinds of solution are used, arrangements must be made to convey each class of solution as required to each treatment tank. This may be done either by having a separate pipe for each class of solution, 238 THE CYANIDE HANDBOOK with a cock over each treatment tank, or by having a common "solution main" for all classes of solution, with separate pipes and cocks between it and the storage tanks, and also separate branch pipes and cocks for each treatment tank. The latter system is more economical as regards piping and the number of cocks required, but is perhaps somewhat more complicated in actual working, as there is more likelihood of mistaking the solu- tion required; and it also involves the necessity of opening two cocks for each transfer. When a pump has to be used for filling a tank or transferring a solution to a treatment tank, it is necessary to open the cock before starting the pump; otherwise the pipe may burst or the driving belt of the pump may be thrown off. The pipes used for running solution to the tanks should be fairly large, so that this operation may be carried out as quickly as possible. For conveying solution from the treatment tanks to the settling or clarifying tanks, or direct to the precipitation boxes, a separate pipe from each treatment tank is desirable, so that the nature of the effluent and the rate of flow from each tank may be readily observed. Some arrangement is also neces- sary to divert the effluent as required into any precipitation box. This may easily be done by means of a length of rubber hose at the exit end of the pipe conveying solution to the boxes. In some plants separate receivers or launders are provided for each class of solution, with which the pipes from the tanks may be connected as required; these receivers are connected with one or more of the boxes set apart for each class of solution. For conveying solution from the boxes to the storage tanks, it is often necessary to arrange for putting any box in communica- tion with any of the storage tanks; the system will therefore be similar to that for connecting the storage tanks with the treat- ment tanks. In cases where separate boxes are reserved exclu- sively for one class of solution, however, it is only necessary to have a single pipe connecting each box with its appropriate storage tank. In general, it may be stated that it is desirable to keep the different classes of solution as distinct as possible throughout the whole series of operations; where large volumes of slime- pulp have to be rapidly transferred, open wooden launders may often be substituted for pipes with great advantage, both owing HANDLING OF SOLUTIONS 239 to reduced friction and owing to the opportunity for aeration which they afford. The inside diameters of solution pipes vary from 1 to 4 in., according to the rate of flow required in the liquid passing through them. (E) ARRANGEMENTS FOR HEATING SOLUTIONS In very cold climates it is necessary to enclose the whole plant in a suitable shed or building, warmed sufficiently to prevent the solutions from freezing. In some cases it has been found that higher extractions are obtained by the use of hot solutions, at a temperature of, say, 100 to 130 F., and provision is sometimes made for heating solution or slime-pulp by the injection of steam. According to Julian and Smart's experiments, 1 the maximum solubility of gold in cyanide solutions of 0.25 per cent. KCy is observed at a temperature of 85 C. (185 F.). But as the tem- perature of the solution also affects the action of cyanide on other constituents of the ore, heating the solution may in some cases be a disadvantage. In most cases it is doubtful whether the improved extraction outweighs the increased cost. Experiments made by the writer in conjunction with Mr. J. J. Johnston at Minas Prietas, Mexico, showed no distinct advantage; other factors in the treatment seemed to be of greater importance, as the higher extractions were obtained sometimes with cold and sometimes with hot solutions. 1 Loc. tit., p. 91. PART V THE PRECIPITATION AND SMELTING PROCESSES THE processes of precipitation and smelting have for their object the recovery of the gold and silver extracted by the cyanide solution in a marketable form. In the precipitation process, the precious metals are thrown out of solution with as little admixture of foreign impurities as possible and obtained in the solid state, but usually in a very finely divided condition. In the smelting process, the finely divided precipitate is brought by the action of heat into a coherent metallic mass, and impurities are removed as far as practicable by the use of suitable fluxes and reagents. Many different pro- cesses of precipitation have been proposed, some of which will be described in a subsequent part of this book; the present discussion will be confined to the usual method, namely, the deposition of the gold and silver on zinc shavings or filaments. The finely-divided precipitate obtained in this process is frequently referred to as "slimes"; here, however, this term will be restricted exclusively to its more usual sense, namely, the finely divided and generally unleachable material obtained in the crushing or grinding of ores. SECTION I ZINC-BOX CONSTRUCTION AND PRACTICE (A) ZiNc-Box CONSTRUCTION THE precipitation by zinc shavings is carried out in rectangular boxes or troughs, divided by transverse partitions into a number of compartments. The partitions are so arranged that the solu- tion flows alternately downward through a narrow compartment, and upward through a wide one, the latter alone containing the zinc shavings. The hight of every alternate partition is one or two inches lower than that of the succeeding one, so that the liquid, after ascending through a wide compartment, overflows the lower partition into the narrow space between it and the next (higher) partition; the lower edge of the latter is raised several inches above the floor of the box, so that the solution flows under it into the next wide compartment, and so on. (See Fig. 37.) The dimensions of the boxes vary, according to the work required of them; the following may be taken as representing ordinary conditions: Length, 12 to 24 ft.; width, 1J to 3 ft.; hight, 2i to 3 ft. The sides, ends, and bottoms are usually constructed of IJ-in. boards, strongly held together by bolts passing horizontally from side to side of the box, generally through the narrow compartments. Vertical bolts are also used for securing the floor of the box to the sides. The partitions, which may be of thinner boards, fit into vertical grooves. The box is preferably raised somewhat above the floor of the extractor house, and should have a slight fall from the feed to the discharge end. Sometimes the bottom is inclined to one side, and plugs are provided on this side just above the bottom, so that the precipitate may be easily collected during the clean-up. The advantage of this latter arrangement is somewhat doubtful, as it may lead to losses through leakage, etc. In each of the wide compartments is a tray, resting loosely 243 244 THE CYANIDE HANDBOOK li I! II I 8^ ij = ZnCa(CO 3 ) 2 + Ca(CN) 2 4 2KCN + 2H 2 O. (I) TESTS FOR REGULATING THE WORKING OF THE BOXES The methods used in making these tests will be fully described in a later section of this book. Samples of the solutions entering and leaving the boxes should be taken at frequent intervals. This is generally done by means of glass tubes provided with i " Proc. Chem., Met. and Min. Soc. of South Africa," IV, 636. ZINC-BOX CONSTRUCTION AND PRACTICE 257 cocks and passing through the top and bottom ends of the boxes. The cock is turned so as to allow a small uniform drip of solution, which is caught in a bottle beneath. These samples are tested at regular intervals, say every 12 or 24 hours, for cyanide, alkali, and gold (for silver also where this metal is of importance). These represent the average results for equal periods of time. Special samples for cyanide and alkali are generally required at more frequent intervals for regulating the leaching operations, the supply of cyanide for the storage tanks, etc., and a test is commonly made of each solution run on and drawn off from each charge under treatment, unless the material is so uniform and the conditions so regular as to render this unnecessary. The strength of cyanide to be maintained depends on such a variety of circumstances that no general rule can be given. Some information on this head will be found in Part IV, Section I. The alkalinity needed also depends to a certain extent on the nature of the material under treatment. A slight but constant alkalinity should be maintained in the entering solution, this being secured by the addition of suitable amounts of lime to the ore or pulp to be treated, or, occasionally, by the addition of caustic soda to the solutions themselves. For slimes plant solu- tions, W. A. Caldecott recommends an alkali strength equivalent to 0.006 to 0.01 per cent. NaOH, 1 but frequently as much as 0.05 per cent, or more is admissible. i " Proc. Chem., Met. and Min. Soc. of South Africa," II, 101. SECTION II CLEAN-UP OF ZINC-BOXES (A) PRELIMINARY OPERATIONS THE frequency of the clean-up will depend on the nature and amount of the precipitate which is formed. In large plants it may take place once a week, in small ones once a month, the com- mon practice being twice or three times a month. When much white precipitate has formed, it may be necessary to clean up oftener than would be the case with a clean gold deposit contain- ing the same values. Before cleaning up it is usual to displace the solution in the box with water, to which alum may be added to promote settlement of the precipitate during the process. In some cases the strength of the entering solution is raised con- siderably for an hour or so before it is displaced with water. This is for the purpose of loosening the deposit which adheres to the zinc. After all cyanide has been displaced, a little acid may be added, which also assists settlement. (B) SEPARATION OF PRECIPITATE FROM COARSE ZINC The trays are first lifted in the compartments, and shaken up and down to allow as much of the precipitate as possible to pass through the screen. The operator, wearing rubber gloves, gen- erally also shakes the shavings under water in the compartment, and then holds them over an adjoining (lower) compartment where they are sprayed with a jet of water. In some cases the tray and contents are lifted bodily from the box and placed over the clean- up tank, into which the fine adhering precipitate is washed. The washed shavings are then either set aside in any suitable water- tight vessels, such as enameled-iron pans, buckets or trays, or are transferred to the lower compartments of another zinc-box. Each compartment in succession is treated in this manner, be- ginning at the head and proceeding 'to the foot of the box. 258 CLEAN-UP OF ZINC-BOXES 259 (C) REMOVAL OF PRECIPITATE FROM Box After the whole of the coarse zinc has been lifted out, the trays are turned over, washed and scrubbed to detach any adher- ing precipitate. The liquid and precipitate remaining in the box are then trans- ferred to a special settling tank, usually about 4 ft. diameter by 3 ft. deep, known as the clean-up tank. This is generally of wood, sometimes lined with lead. As this tank is often used for the subsequent acid treatment of the precipitate, it is sometimes provided with a wooden paddle mounted on a central vertical shaft, or with other means of stirring. When there are no out- lets to the compartments, the transfer is made by means of enameled-iron buckets. In some plants, the contents of the box are allowed to stand and settle for some time, and the clear liquid syphoned off. The residue is then scooped out and carried to the clean-up tank. When plugs are provided at the bottom of the compartments, these are pulled out and the liquid allowed to run into a launder beneath, which leads to the clean-up tank. The sides of the compartments are then washed down with a hose until the whole of the precipitate has been cleaned out and run into the tank. When the box is perfectly clean, the coarse zinc is replaced on the trays in the same order, unless it has been already transferred to another box. The box is at once filled up with solution, so as to avoid oxidation of the zinc, and is ready for use after any fresh zinc that may be necessary has been added. In some modern plants, the whole of the liquid drawn off during the clean-up is passed through a small filter-press. The precipitate, after passing through a fine screen (60- to 80-mesh), is also filter-pressed, and the material retained on the screen treated separately. Formerly the fine precipitate was squeezed by hand through filter-bags. (D) SETTLEMENT AND SIFTING OF PRECIPITATE Where filter-presses are not used, the liquid is allowed to stand for 24 to 48 hours in the clean-up tank, sometimes with addition of a little alum, until the precipitate has completely settled. The clear liquid is then carefully syphoned off and run to waste. The precipitate during transfer to the settling tank is frequently run through a fine sieve, 40- to 80-mesh, and any 260 THE CYANIDE HANDBOOK material retained on the sieve (generally known as "shorts" or "metallics") is set aside for separate treatment. When suffi- ciently coarse, it may be returned to the top compartment of the box and placed on a layer of clean shavings; if there is too much of this product, however, or if it is too fine, it would be liable to choke the screens of the boxes, and it is usually treated with acid to dissolve as much as possible of the zinc, as described below. If acid treatment is not to be used, the settled precipitate, if free from shorts, may be at once pumped to a filter-press. Lead-lined tanks should not be used when the precipitate, as is often the case, contains mercury, as this metal amalgamates with the lead and rapidly corrodes the lining. In some plants vacuum filters are used. As the precipitate is very finely divided, the filtering medium must be very closely woven, such as flannel or very thick canvas. The liquid is drawn off by a suction pump and a wash of fresh water added, and this is again sucked through to extract as much as possible of any soluble matter which may be present. Any wash-waters drawn off in the above processes should be assayed before finally running to waste, as they occasionally carry sufficient values to be worth recovering. In that case they should be returned to the boxes, after addition of alkali, if necessary. SECTION III ACID. TREATMENT AND ROASTING OF PRECIPITATE AFTER the precipitate has been separated as far as possible from the coarse zinc, there is considerable divergence in the sub- sequent procedure. The methods adopted may be briefly classified as follows: (1) Direct fusion. (2) Roasting and fusion. (3) Mixing with niter, roasting and fusion. (4) Acid treatment and fusion. (5) Acid treatment, roasting, and fusion. (6) Roasting, acid treatment, and fusion. (7) Smelting with litharge and cupellation. Since acid treatment generally precedes roasting, the operations will be considered in that order. (A) ACID TREATMENT The precipitate, having been transferred to the acid tank with or without filtering, is allowed to settle somewhat, and as much as possible of the clear liquid drawn off. Sulphuric acid is then added, in quantity more than sufficient to dissolve the zinc supposed to be present, and diluted with enough water to form a 10 per cent, solution. The heat produced by the mixture of acid and water, with that generated by the chemical action of the acid on the zinc, is generally sufficient to maintain the re- action with sufficient vigor till the greater part of the zinc is dis- solved. In some cases it may be necessary to apply artificial heat, by injection of steam or otherwise. The mixture is stirred from time to time, mechanically or by hand. At first there is a very violent disengagement of gases, chiefly hydrogen, but mixed also with hydrocyanic acid and, where arsenic is present, with arseniureted hydrogen. As the latter are ex- tremely poisonous, it is necessary to provide a hood with a good draft, for carrying off the fumes. The tank must be sufficiently large to avoid the danger of frothing over. When all action has 261 262 THE CYANIDE HANDBOOK ceased, the tank is filled up with water, preferably hot, as the zinc sulphate formed is more soluble in hot water. The mixture is stirred, allowed to settle, and the clear liquid drawn off, this operation being repeated several times. The acid tank may be of wood, with or without a lead lining, or of enameled iron. Sodium bisulphate (a by-product in the manufacture of nitric acid) is sometimes used as a substitute for sulphuric acid. An 8 per cent, solution of the salt is an effective strength. The tem- perature should not be above 33 C., as the solubility of the salt decreases with rise of temperature. 1 This material is cheaper and more easily transported than sulphuric acid. In cases where much arsenic is present, it is advisable to add nitric acid in the proportion of about 1 part HNO 3 to 2 parts H 2 SO 4 ; this has the effect of oxidizing the arsenic to arsenic or arsenious acid, and at the same time converts hydrocyanic into cyanic acid, thus avoiding the evolution of large quantities of poisonous fumes. It must be mentioned, however, that under certain circumstances nitric acid may lead to a loss of gold. Experiments have shown that this loss is greater with nitric than with hydrochloric acid, and greater with hydrochloric than with sulphuric acid. Nitric acid has sometimes been added for the purpose of dissolving the lead introduced by the use of the lead zinc couple in precipitating, but it seems preferable to remove this metal by suitable fluxes in the subsequent fusion. At the Metallic Extraction Co.'s mill near Florence, Colorado, a treatment with hydrochloric acid was given after roasting, the roasted precipitate being treated in a lead-lined sheet-steel pan with a mixture of one part hydrochloric acid to two of water, the residue being afterward filtered, washed, and dried. After acid treatment, the precipitate is frequently pumped to a filter-press, through which hot water is forced to extract as much as possible of the zinc sulphate still retained, and finally dried by forcing air through the cakes. The product may also be extracted and dried by vacuum pans. The following is a description of the method employed at the Myalls United Mines, Australia, given by C. J. Morris. 2 The zinc is scrubbed, and the precipitate is placed, together with the 1 J. E. Thomas and G. Williams, " Journ. Chem., Met. and Min. Soc. of South Africa," V, 334. 2 " Trans. I. M. M.," XV, 543 (1906). ACID TREATMENT AND ROASTING OF PRECIPITATE 263 shorts, in a wooden acid tank 5 ft. deep by 4 ft. diameter, pro- vided with a 4-armed revolving paddle, plugged holes for draining off the wash, and a locked cover. Twenty pounds of acid are added to the charge at one time, and a further quantity when the violent action has ceased, until the required amount has been added. The charge remains under treatment all night, with occasional stirring, and exhaust steam from the solution pump is led in through the cover by means of an iron pipe. In the morning all action has ceased, and twelve times the bulk of water is added, with agitation. After settling, the wash-water is drawn off. This contains only 1 to 2 dwt. of gold per ton, which might be recovered by long settling. The washed precipitate is then dried with addition of 5 per cent, of niter, roasted and fluxed. It is claimed that acid treatment, as described, gives far better results than addition of niter, roasting, and fusion without acid treatment. In many cases it is found advisable to keep the shorts and fine precipitate separate, as they may be advantageously treated by different fluxes. At Redjang Lebong, a considerable difference was noticed in the composition of the bullion obtained from these two products. As the shorts contain very much more metallic zinc than the fines, they will naturally require a larger amount of sulphuric acid. In some cases it is found that the solution of the zinc by sulphuric acid is imperfect. This may be due to the metal being present as an alloy insoluble in that acid, or to the presence of protective coatings of insoluble sulphates (PbSO 4 or CaSOJ. Composition of Acid-treated Precipitate. The zinc-gold pre- cipitate as generally obtained is a very complex mixture, some- times containing, in addition to gold, silver and zinc, nearly all the common metals, with silica, carbonates, sulphates, sulphides, and various cyanogen compounds ; of these ingredients the pro- portions vary enormously in different samples, so that no fixed rules can be given as to the best method for its reduction. An analysis of acid treated and roasted precipitate by A. Whitby 1 gave the following percentages: Gold 34.5 Ferric oxide 3.65 Silver 4.75 Zinc oxide 7.0 Lead 12.5 Sulphuric acid 6.95 (SO 3 ) Copper 2.55 Nickel oxide 1.0 Silica 21.0 i " Proc. Chem., and Met. and Min. Soc. of South Africa," III, 46 (1902). 264 THE CYANIDE HANDBOOK (B) ROASTING OF ZINC-GOLD PRECIPITATE The precipitate, whether acid-treated or not, is best treated by filter-pressing or vacuum filtration to remove as much of the moisture as possible. By the use of hot air, the drying may be practically completed in the filter-press, but more usually the cakes of precipitate are broken up and placed on sheet-iron trays about H m - deep, on which the material is first carefully dried over a slow fire and finally roasted, either over an open grate or in a muffle or reverberatory furnace. It is a common practice to mix about 5 per cent, of niter with the dried and powdered precipitate before roasting, especially in cases where acid treat- ment is not used. The mixture, when perfectly dry, may be ignited and burns of itself, all carbonaceous matter being thus removed and a complete oxidation of the zinc obtained. Caldecott recommends mixing 30 parts of silica and 30 parts of niter with every 100 parts of dried precipitate; this yields a product which does not undergo losses by dusting, and the sand serves as a flux for the zinc oxide in the smelting. Other operators saturate the precipitate, before drying, with a strong solution of niter. The roasting should be done slowly, at a dull red heat, with careful regulation of the draft, since heavy losses of gold may occur when copious fumes of zinc are given off. The exact amount of this loss has never been ascertained, but it is the general ex- perience that the best recovery of the contained values is obtained when the zinc is eliminated as completely as possible before roasting. Stirring may be occasionally necessary during the roasting, but should be avoided as much as possible, as it may lead to mechanical losses. In some works the precipitate is placed in shallow iron pans, which fit inside the muffle, and allowed to remain over night at a moderate heat, with little or no stirring. When cool, the trays are withdrawn and the lumps ground and mixed with the necessary fluxes for smelting. When the acid treatment is very carefully and thoroughly performed, the roasting operation is of less im- portance, and may in some cases be advantageously omitted. As a precaution against " dusting " during the early stages of the roasting, the precipitate is generally charged into the furnace in a slightly moist condition, and is often compressed or briquetted. Probably more loss of gold and silver occurs in the roasting ACID TREATMENT AND ROASTING OF PRECIPITATE 265 than in any other stage of the process. When metallic zinc is volatilized, it carries with it considerable quantities of the pre- cious metals, probably mechanically. In contact with the air, the zinc vapor burns with a greenish flame, producing dense fumes of white oxide of zinc. These are deposited in the flues and cooler parts of the furnace, but large quantities escape into the air, carrying with them gold and silver in a state of minute subdivision. Zinc oxide itself is non-volatile, and if the zinc can be oxidized by chemical means before roasting, this loss does not occur. This is the chief object of adding niter in the manner described above. SECTION IV FLUXING, SMELTING, AND REFINING OF THE PRECIPITATE (A) FLUXES FOR ZINC-GOLD PRECIPITATE THE fluxes to be used will depend to a large extent on the nature of the materials contained in the precipitate, and on the preliminary treatment which it has undergone. The following are given as examples, but in any case the most suitable flux should be ascertained by trial. They are tabulated according to the preliminary treatment of the precipitate. 1. For direct fusion without acid treatment or roasting: Parts by weight (a) (b) Precipitate 100 100 Borax 35 60 Bicarbonate of soda 50 7 Sand 15 11.5 Niter 2 to 4 19 (a) Julian and Smart, "Cyaniding Gold and Silver Ores," 2d ed., p. 168. (ft) T. L. Carter, "Trans. Chem., Met. and Min. Soc. of South Africa," II, 444. The method generally yields a very- base bullion containing much zinc, and of irregular composition; (6) was tried for treat- ment of precipitate from zinc-lead couple. 2. Roasting with niter, without acid treatment: Parts by weight (a) Precipitate 100 Borax 40 Bicarbonate of soda 40 Sand 15 (a) Julian and Smart, loc. cit., p. 168. When Caldecott's system of roasting with sand and niter is used, the silica is omitted from the flux and the following is used: 266 REFINING OF THE PRECIPITATE 267 Parts by weight <*) Precipitate 100 Borax 40 Carbonate of soda 10 (6) "Journ. Soc. Chem. Ind.," XVII, 3 (1897). 3. Acid treatment without roasting: Parts by weight (a) (&) (c) . Precipitate .100 100 100 Borax 25-30 66 10 Carbonate of soda 50 9 40 Fluor-spar 9 2 Sand 15 (a) T. H. Leggett, "Trans. I. M. M.," V, 147. (6) E. H. Johnson, "Eng. and Min. Journ.," p. 220 (1899). (c) J. Gross, "Trans, A. I. M. E.," XXXV, 616. 4. After acid treatment and roasting: Clay liners used when much lead is present : Parts by weight (a) (6) (c) (d) Precipitate 100 100 100 100 Fused borax 20-35 60 70 50 Carbonate of soda 25 50 Manganese dioxide 20-40 15-20 Fluor-spar 5 10 0-10 10 Sand 15-40 5 10 44 (a) E. H. Johnson and W. A. Caldecott, " Trans. Chem., Met. and Min. Soc. of South Africa," III, 51 (1902). (6) B. W. Begeer, "Metallurgy of Gold on the Rand," 1898. (c) Used at Abosso G. M. Co., Gold Coast, West Africa, 1907-08, yielding bullion about 940 fine. Previously roasted with 5 per cent, niter. (d) The present writer has found this flux satisfactory for a base precipi- tate containing much copper and lead, together with sulphur and a little arsenic. (B) MIXING OF PRECIPITATE AND FLUXES The various materials used for the flux should be thoroughly dried, especially the borax and soda, which are liable to contain considerable amounts of water. In small plants, the necessary quantities for each charge are weighed out, mixed on an iron tray or other suitable receptacle with a scoop, and fed into the previ- ously heated crucible. Where large quantities are dealt with at a time the mixture is put through a small ball mill specially reserved for this purpose, so that the different ingredients are thoroughly incorporated and all lumps broken up. The flux must be well mixed with the precipitate before transferring to 268 THE CYANIDE HANDBOOK the crucible. The mixture is generally fed in by means of a long iron scoop, the lower end of which is formed into a tube to avoid loss by scattering of the charge while feeding into the pot. A wide-necked iron hopper or funnel is also used for this purpose. (C) THE FUSION PROCESS The fusion is almost invariably made in graphite crucibles (actually made from a mixture of graphite and fire-clay), the kind known as "salamander" being frequently used. The size of crucible will depend on the amount of fluxed precipitate which can be conveniently smelted at one time in the furnace. Size No. 60 is largely used. From 10 to 20 Ib. of dried precipitate can be smelted in one crucible charge. In cases where lead is present, or when for any reason the graphite of the crucible might act injuriously in the fusion, a "liner" is used, consisting of a fire-clay crucible that fits inside a large graphite pot. Both pots and liners should be carefully dried and annealed by gradual heating outside the furnace, and raised to a dull redness in the furnace before charging in the mixture to be smelted. The furnaces used are square, and generally similar to the wind fusion-furnaces used in assaying, but somewhat larger and built so that the mouth of the furnace is flush with the floor. It is a good plan to have the building so arranged that fuel can be supplied to the furnace and the ashes removed by openings out- side the smelting room proper. The floor of the latter may thus be kept clean throughout the operation, and no unskilled work- men need be admitted while the smelting is going on, thus di- minishing the chances of loss or theft. The ash-pit should, however, be accessible from the smelting room in case an acci- dent should occur, such as the breaking or upsetting of a pot in the fire. The fuel generally used is coke; where coal or charcoal is employed, the furnace must be somewhat larger to allow of a sufficient quantity of fuel being packed round the crucibles. Furnaces have also been built for gasoline or other liquid fuel. Where large quantities of precipitate have to be smelted, a re- verberatory furnace, as introduced by Charles Butters, is some- times used. Another appliance very convenient for this purpose is a tilting furnace, similar in design to that proposed by H. L. REFINING OF THE PRECIPITATE 269 Sulman, 1 consisting of a rectangular furnace cased with iron and lined internally with brick. It is mounted on trunnions and contains an inner (replaceable) retort of graphite or fire-clay, which serves the place of the crucible and holds the charge to be smelted. Arrangements are provided, so that when the fusion is complete the entire furnace may be tilted and the contents of the crucible poured into a mold. A furnace of this type was used by the Redjang Lebong Mining Co., for running down acid-treated and roasted-zinc precipitate containing a large percentage of silver and some selenium. The latter was partially expelled as oxide by introducing into the furnace a blast of air from an air-com- pressor. There is considerable saving of labor and probably less risk of loss in using some such arrangement, but the cost of in- stallation and repairs is greater than with the ordinary pot furnace. The crucible, in the usual square furnace, should be supported firmly by a fire-brick resting on the furnace bars. It should be well packed all around with coke, and covered with a graphite or fire-clay lid. At the beginning it should not be more than three- quarters filled with the mixture of flux and precipitate, for the charge swells considerably through the disengagement of carbonic acid, etc., as it becomes heated. After it is thoroughly melted it subsides, and a further quantity of the mixture may now be added. Toward the end of the operation, when the fusion is nearly complete, the contents of the pot are stirred with an iron rod and any infusible dross which remains on the surface removed by skimming. When the zinc has not been perfectly removed by acid treatment, dense fumes will be given off during the fusion, and some of the zinc will be volatilized and burn at the mouth of the crucible with a greenish flame. Under these circumstances some loss of gold and silver by volatilization undoubtedly takes place, but opinions differ as to the extent of this loss. Deposits of zinc dust and zinc oxide carrying high values are sometimes recovered from the flues, and dust chambers are occasionally used for col- lecting this material. When the fusion is complete, the pot is lifted from the furnace by means of basket tongs; when large pots are used, a block and tackle arrangement should be provided for lifting the pots. The contents are then poured into a conical mold which should be previously oiled and heated. As soon as the bullion has set, the i " Journ. Soc. Chem. Ind.," December, 1897. 270 THE CYANIDE HANDBOOK mold is inverted onto an iron tray and the slag detached from the button of metal by one or two blows of a hammer. The slag should be clear and uniform; it is generally of a light greenish- gray color, and is set aside to be crushed and panned, as it usually contains sufficient value in the form of small shots of metal to be worth treating. When the whole quantity of fluxed precipitate which is to be smelted has been run down in this way, the buttons from the conical molds are remelted in a clean crucible with a little borax and cast into an oblong ingot-mold. 1 It is usual to take a sample of this bullion just before casting, by means of a dipper formed from a piece of an old graphite crucible. The molten metal should be well stirred and the dipper heated before sampling. (See Part VII.) (D) USE OF SPECIAL FLUXES Niter is used to oxidize any metallic zinc which may still be present in the precipitate after the preliminary treatment. It is generally unnecessary when acid treatment has been given, as in that case any zinc remaining is probably present as sulphate. It is ineffective for oxidizing lead, as it decomposes and gives off its oxygen below the temperature required for the formation of litharge. Manganese dioxide is used as a substitute for niter when lead is present; it is not advisable to use it when much silver is present, as it has a tendency to carry the silver into the slag. Fluor-spar is used chiefly to give fluidity to the charge. It is also of some use in fluxing calcium sulphate. Sand is added for the purpose of forming a fusible silicate with the zinc and other bases present. It is omitted when practically all the zinc has been removed by preliminary operations. Carbonate of soda is used only in small quantity and is some- times omitted. When silica is present in the precipitate or has been added to the charge, it helps by forming a fusible double silicate of zinc and sodium. Borax is the principal flux for all bases present. It also 1 The following flux answers very well for remelting the buttons into bars. Weight of metal remelted 9 to 10 kilos. Fused Borax 1.7 " Fluor-spar 0.2 " Niter (added at moment of fusion) 0.15 to 0,2 " REFINING OF THE PRECIPITATE 271 increases the fluidity of the charge, apart from any chemical action. (E) TREATMENT OF MATTE It has already been noted that sulphur, from various sources, is liable to be present in the deposit collected from the zinc-boxes. Unless the precipitate after acid treatment is thoroughly roasted, this sulphur, in the subsequent fusion, will form a "matte," gen- erally carrying gold, silver, iron, zinc, and lead, also copper if the latter metal is present in the material treated. This matte forms a brittle but strongly adherent layer above the button of metal, between it and the slag, and as it frequently carries high values, some means must be found to reduce it to a marketable bullion. Selenium, tellurium, arsenic, and antimony form similar products when they occur in the material smelted. A method of recovering fine bullion from matte is described by A. E. Drucker 1 as follows. The fluxes used for the reduction are borax and cyanide. The matte, borax, and cyanide are put separately through a rock-breaker and crushed fine. Alternate layers of borax, matte, and cyanide are charged into a No. 60 graphite crucible until the pot is nearly full, when a layer of borax is added as a cover. The crucible is now put into the furnace and maintained at a white heat for two or three hours, until the charge subsides and bubbling ceases. Large volumes of sulphur are given off at the end of the first hour, and burn at the mouth of the pot. When the action is complete the slag becomes quite thick and must be removed by a skimmer. The remaining contents are then poured into a conical mold. If excess of cyanide has been used it forms a crust just above the gold button and can be broken off with a blow from a hammer. The matte is com- pletely decomposed, and 85 to 94 per cent, of the total values contained in it are recovered as fine metal. Only a light porous slag remains, which may be remelted with a subsequent charge. The usual method of treating this matte is to remelt with additional flux and scrap-iron, niter being sometimes added to assist in oxidizing the sulphur. This method, as stated by Drucker, is slow and incomplete, and when graphite crucibles are used they are rapidly corroded by the niter. i " Min. Sc. Press," May 18, 1907. 272 THE CYANIDE HANDBOOK (F) SMELTING OF ZINC-GOLD PRECIPITATE WITH LITHARGE In the early days of the cyanide process, the suggestion of melting the zinc-precipitate with lead, and afterward cupeling, seems to have been carried out. G. H. Clevenger 1 states that at the Balbach Smelting and Refining Co.'s plant, Newark, New Jersey, the precipitates, tied up in paper sacks in parcels of 1 to 5 lb., were charged from time to time on a bath of molten lead in a cupeling furnace. The gold and silver were quickly ab- sorbed by the lead, and until the mass was well melted, the pre- caution was taken of keeping all drafts closed. The slag was then removed by skimming and the bullion cupeled and refined. For this method it would seem that there is considerable proba- bility of metallic zinc being volatilized with consequent loss of gold and silver. This is avoided in P. S. Tavener's process, first introduced at the Bonanza mine, Johannesburg, in August, 1899, in which the precipitate, after filter-pressing (but without acid- treatment or roasting), is mixed with litharge, slag from previous operations, slag from assay fusions, sand and sawdust, and smelted in a reverberatory furnace. The zinc is. fluxed off with- out being volatilized to any appreciable extent, and lead bullion is obtained which is afterward refined on a bone-ash "test," in an ordinary cupellation furnace. The process may be likened to a scorification assay on a large scale. The following account of the process is summarized from a detailed description given by Tavener. 2 The fine precipitate from the filter-press and the zinc shorts are dried separately on trays in an oven for 15 minutes. The dry precipitate is then rubbed through a sieve 'of four holes to the linear inch, roughly weighed, and mixed with the fluxes. These are as follows : Parts by weight. Precipitate 100 Litharge 60 Assay slag 10 to 15 Slag previously used 10 to 15 Sand 10 to 20 After adding the precipitate, the mixture is again sifted to ensure i " Trans. A. I. M. E.," XXXIV, 891 (October, 1903). 2 "Proc. Chem., Met. and Min. Soc. of South Africa" Vol. III., p. 112 (Oc- tober, 1902). REFINING OF THE PRECIPITATE 273 thorough mixing, and shoveled into the furnace, which is as yet unlighted. The fine zinc is then mixed as follows: Parts by weight. Fine zinc (shorts) 100 Litharge 150 Slag 20 This is put into the furnace on the top of the precipitate to pre- vent 'loss by dusting and also to ensure the greater part of the litharge being on the top of the charge. A slow fire is now lighted and the charge allowed to dry for a couple of hours. The temper- ature is then raised, and in four or five hours the charge is reduced; any sweepings or by-products which have to be worked up are now added, and are quickly absorbed. When the slag has again become fluid, it is well stirred with a rabble, and sawdust thrown in to reduce the excess of litharge. The slag is now tapped off through the slag-door above the level of the lead, and the latter is skimmed; a shovelful of lime is thrown in and a final skimming is given. The clean lead-bath is then stirred and sampled, the tap-hole is opened, and the lead bullion run into molds. The lead bullion is now cupeled on a test, consisting of an oval cast-iron frame filled with bone-ash ground to pass a 20-mesh screen, and mixed with 3 per cent, caustic potash and 10 to 11 per cent, water. The mixture is again sifted to break up lumps. The test-frame is placed on a cast-iron plate and filled with the bone-ash, which is then tamped in. The center is then hollowed out, leaving a rim round the sides. About 300 Ib. of bone-ash are used for one test, which may be used about four times, cupeling altogether over 1\ tons of lead bullion. The test should be dried slowly for some weeks before use. When a new test is put in the furnace, a slow fire is kept up for 3 or 4 hours. An iron blast-pipe, 3 in. diameter, flattened and turned down at one end to allow the blast to strike on the molten lead, is now fitted to the back end of the cupel. Temperature is raised, and the lead bars fed in one at a time through the working- door of the furnace onto the test. A channel is cut in the rim of the latter } to \ in. deep and 1 J in. wide, and communicating with a hole by which the molten litharge runs down to a suitable pot placed beneath. When the test is filled with molten lead nearly to the level of this channel, the temperature is increased to the melting-point of litharge, and when the lead is covered with molten 274 THE CYANIDE HANDBOOK litharge the blast is turned on, and the litharge allowed to flow away along the channel, which is deepened as the operation proceeds. Fresh lead bars are fed in until all are melted. The temperature is raised as the proportion of gold in the alloy re- maining on the cupel increases. At the finish it is necessary to add a little assay-slag, which is melted and run off. When the operation is complete, the gold freezes or solidifies. It is then broken in pieces, remelted in crucibles, and cast into bars. This method possesses many advantages over the ordinary systems of reducing zinc-gold precipitate. The cost of material is much lower; a large part of the litharge is recovered for re-use, and the slag used as flux in this process is a waste product from which the contained values could otherwise be obtained only at a high cost. It also affords a means of easily disposing of a num- ber of troublesome by-products of the mill and cyanide works which cannot be economically treated by other methods. The cost of installation, however, is so high that the method could only be economically carried out on a large scale; moreover, skilled labor is required, especially in the cupellation process. The method would seem in general to be more suitable for customs works than for individual mines. In spite of the fact that com- parative tests made with great care showed a much higher recovery by the litharge-smelting process as compared with the ordinary acid-treatment, roasting, and smelting the system has not been generally adopted. In some plants, where the ordinary method of smelting is used, it is customary to remelt all the slags from the first fusion of precipitate, with litharge and a little additional flux, adding a reducing agent if necessary or adding a small quantity of metallic lead at the finish. The resulting lead bullion is then cupeled. (G) NATURE AND PROPERTIES OF CYANIDE BULLION The bullion bars, after the final smelting, are frequently turned out of the ingot molds into a vessel of water as soon as they are set. The adhering slag is detached by hammering, and in some cases the bar is " pickled " by immersing in dilute nitric acid and cleaned by scrubbing with a hard brush. If dip samples have not been taken, it is commonly sampled by machine drill as described below, in Part VII, and then immediately packed for shipment. REFINING OF THE PRECIPITATE 275 The nature and composition of the bullion varies according t.> the proportion of gold and silver, and to the method by which it has been obtained. When much zinc is present, it is of a pale yellowish-green color, and shows much variation in composition of different parts of the bar, the general tendency being for the gold value to concentrate toward the center of the bar. 1 This refers chiefly to bars of 650 to 800 fineness; with richer bars the liquation is less marked. It is quite possible to obtain cyanide gold which is little, if at all, inferior to battery gold in fineness. This is mainly a question of care in the clean-up and smelting, more particularly in the thorough extraction of metallic zinc by acid treatment and in the selection of the proper fluxes in the fusion process. The following analyses, given by Dr. T. Kirke Rose, 2 will illustrate the character of bullion produced when direct smelting is used without preliminary treatment to remove the zinc: (1) (2) (3) Per cent. Per cent. Per cent. Gold 60.3 61.7 72.6 Silver 7.3 8.1 9.2 Zinc 15.0 9.5 7.1 Lead 7.0 16.4 4.9 Copper 6.5 4.0 4.8 Iron 2.2 0.3 1.4 Nickel 2.0 By acid treatment and roasting the amount of impurity may be reduced to 5 or 6 per cent., particularly if manganese dioxide be used for fluxing off the lead. (H) REFINING OF BULLION On account of the deductions made by bullion buyers, con- siderable efforts have been made to devise a means of economically purifying the low-grade bullion produced by the cyanide process. It may be remarked, however, that it is generally easier and more satisfactory to remove the impurities before conversion into bullion than after. The method of remelting with borax is very ancient, as is also the practice of sprinkling borax on the metal at the moment of fusion. 1 See paper by F Stockhausen, " Proc. Chem. Met. and Min. Soc. of South Africa," II, 46. 2 "Metallurgy of Gold," 4th edition, p. 331. 276 THE CYANIDE HANDBOOK Refining by leading a current of chlorine gas through the molten bullion was practised for many years in dealing with battery gold in Australia, where it was introduced by Miller in 1867 and used at the Sydney mint. The process was in use at the Pretoria mint at one time for refining cyanide bullion, but has the drawback that it involves a separate operation for the recovery of the silver, as the scoria formed contains this metal as chloride. Dr. T. Kirke Rose l has experimented on a method of refining by injecting oxygen gas into the molten metal. Air may also be used, and is equally effective in removing base metals, but the action is slower. The method is also applicable to zinc-gold precipitate without previous smelting; sand, borax, and charcoal are added, the mixture fused in a graphite pot with clay liner, and air or oxygen injected into the molten mixture through a J-in. clay pipe, by means of a Root's blower. i " Trans. I. M. M.," XIV, 378-441 (April, 1905). PART VI SPECIAL MODIFICATIONS OF THE CYANIDE PROCESS THIS part of the book will be devoted to the discussion of certain departures from the normal course of cyanide treatment adopted in particular cases and under special circumstances. We shall refer briefly to certain new developments of the process which have not yet stood the test of experience long enough to demonstrate their practical value. Some account will likewise be given of obsolete or nearly obsolete methods which possess a historical interest for cyanide workers. As the admission of such descriptions may be criticized, it may be pointed out that all improvements in the cyanide process are merely applications of well-known principles employed in other branches of industry; that much may be learned by studying the causes of past failures; and that the experience of the past often points out the directest road to advance in the future. These special processes will be considered under the following heads: (1) Direct treatment after dry or wet crushing. (2) Crushing with cyanide solution. (3) Roasting before cyanide treatment. (4) Use of auxiliary dissolving agents. (5) Elec- trolytic precipitation processes. (6) Other precipitation processes (7) Special treatment of cupriferous ores. SECTION I DIRECT TREATMENT AFTER DRY OR WET CRUSHING DIRECT treatment may be defined as any process in which the ore is crushed dry or with water, and the cyanide solution applied at once to the crushed product, the latter undergoing no amalga- mation, hydraulic separation, concentration, roasting, or other intermediate process previous to cyanide treatment. When this system is adopted, dry crushing is nearly always employed in preference to wet, chiefly because the wet-crushed product is seldom in a condition suitable for cyanide treatment as a whole, being generally unleachable without some form of hydraulic separation; also because the 'cost of plate amalgamation is gen- erally so trifling in comparison with that of crushing, that there is no sufficient advantage in omitting it in the case of wet-crushed ore. For diy crushing, Chilian, Griffin or Ball mills and rolls are preferable to stamps. Some particulars with regard to these machines are given in Part III. The system is particularly suitable for cases in which the ore is exceptionally friable or porous; in such cases good, extractions can sometimes be obtained even with very coarse crushing, as at the Mercur mine, Utah, where the oxidized ore was treated direct after dry crushing to J-in. size. At the George and May mine, Johannesburg, a porous oxidized ore was crushed coarsely in a Gates crusher and treated direct by cyanide, giving an extraction of 70 per cent, of the gold. 1 At the Lisbon Berlyn mine, Lydenburg, Transvaal, a 10 dwt. ore was put through a Blake rock-breaker, Marsden* fine crusher, and Gates rolls; the product gave a 68 per cent, extraction by direct cyanide treatment. It is generally necessary to dry the ore sufficiently to reduce the percentage of moisture below 2 per cent, previous to dry crushing, care being taken in this operation to avoid partial roast- i Julian and Smart, loc. tit., p. 202. 279 280 THE CYANIDE HANDBOOK ing, which might lead to the formation of cyanicides. To avoid the formation of large amounts of fine dust, the crushing is best done in several stages, using screens and returning the oversize for recrushing. As a rule, shallow tanks must be used for the leach- ing of the dry-crushed product, and in many cases nitration is aided by suction. Since no moisture has to be displaced by solu- tion, it is possible to give a more thorough final water-wash than is the case when wet crushing is used, without increasing the stock of solution in the plant. SECTION II CRUSHING WITH CYANIDE SOLUTION GENERAL CONSIDERATIONS THE crushing of ore with cyanide solution instead of water was attempted at least as early as 1892, when experiments in this direction were made at the May Consolidated Battery near Johannesburg. 1 The system was introduced at the Crown mine, Karangahake, New Zealand, in 1897, and at Central City, South Dakota, in 1899. Five mills on this principle were established in South Dakota in 1904, since when the method has been ex- tensively adopted in the United States and Mexico. The early attempts at crushing with cyanide seem to have been abandoned on account of difficulties in handling slimes. The proportion of ore to liquid in the material crushed is neces- sarily much greater than when water is used, and it is of course impossible to run the slimes to waste without losing the greater part of the dissolved values. Owing to the thicker pulp, the ore is also crushed finer than when water is used under ordinary conditions, and the product is therefore less adapted for direct leaching. In modern practice, however, these difficulties are overcome by a system of hydraulic separation, in which the clear overflow is returned to the mill and the sand and slime collected for separate treatment. To avoid the necessity of handling enormous volumes of solution, the amount of liquid used in the battery is generally from 1^ to 1 ton of solution per ton of ore crushed, but as much as 7 tons is sometimes used. The advantages of the system are briefly as follows: (1) The solution of gold begins from the moment the ore enters the battery. (2) There is no necessity for special appliances for agitation and aeration, these being sufficiently secured by the transfer of the pulp from the battery to the classifiers and treatment tanks. l" E. and M. J.," Oct. 8, 1892. 281 282 THE CYANIDE HANDBOOK (3) Less water is required than when the crushing is done in water. (4) There is no necessity for running any solution to waste, as must be generally done when final water-washes are given in the ordinary system of treatment. The objections are: (1) There is danger of loss, owing to the fact that gold in solution is transferred for a considerable dis- tance and through a number of different appliances. (2) Owing to the thickness of the battery pulp, the capacity of the mill is reduced, and, as already pointed out, a larger percentage of fines is produced. (3) The action of the cyanide hardens the amalgam on the plates, and corrodes the plates themselves, owing to the solvent action of cyanide on copper. (4) It is impossible to give preliminary water or alkali washes. This is a serious objection in the case of acid ores, or of ores containing soluble cyanicides, which might be removed by preliminary treatment in the ordinary process. The first objection is not of much consequence in a well-con- structed plant, where proper means are taken to prevent leakage; the second is of little importance in the case of ores which must be crushed fine in any case in order to obtain a good extraction. When very dilute solutions (.03 to .07 per cent. KCN) are used, as is generally the case, the action on the plates is not serious, and it may possibly be avoided by using Muntz metal or some other alloy instead of copper. Illustrations from Practical Working On the Rand, the system of crushing with cyanide solution forms part of the scheme of treatment introduced by G. A. and H. S. Denny. 1 In this system the ore is crushed with a solution containing .03 per cent, cyanide (as KCy) and .004 per cent, alkali (calculated as NaOH), in the proportion of 6.6 tons of solution to 1 ton of ore. After passing over amalgamated plates, the pulp, mixed with the returned coarse product from the spitz kasten following the tube mills, goes to hydraulic classi- fiers, where coarse sand and concentrates are separated from fine sand and slime. The coarse sand and concentrates are passed through tube mills, and thence over shaking amalgamated plates to spitzkasten, the overflow from which goes to the conical slime settlers, while the underflow (coarse product) is elevated and i " E. and M. J.," LXXXII, 1217 (Dec. 29, 1906> CRUSHING WITH CYANIDE SOLUTION 283 returned to the pulp leaving the battery plates. The fine sand and slimes go through another set of spitzkasten, from which the sandy product goes to percolation tanks, while the slimy portion goes to a large conical tank, whence the clear solution overflows and is returned to a solution tank supplying the battery. The thickened slime-pulp, together with the slime overflow from the tube-mill product, goes to conical slime settlers and is then pumped to filter-presses. The precipitated solutions are also returned to the mill-supply tank. This system could only be successfully applied in cases where the gold in the finely crushed ore is very rapidly dissolved; on the Rand it is claimed that 98 per cent, of the gold recovered from the slime is dissolved before the slime is settled, and that 70 per cent, of the gold recovered from sands is in solution before per- colation begins. In the pulp leaving the mortar-boxes 12.65 per cent, of the total gold is already in solution. The results of a month's treatment at the Meyer and Charlton G. M. Company's plant, working on this system, are given as follows: Ore treated 10,740 tons Average assay before treatment $10.90 per ton Average assay of residues $ 0.51 " " Extracted 011 battery plates 43.85 per cent. " on shaking plates 3.01 " " " in sand treatment 9.76 " " from ore in transit 38.67 " " Total extraction 95.29 Of the total ore milled, 52.3 per cent, is reground in tube mills; approximately 70 per cent, is treated by percolation as fine sand, and 30 per cent, as slime. It will be noticed that comparatively little of the gold is recovered from the sand in the percolation process, most of it having been already dissolved and carried off with the solution in the slime overflow. This suggests a further modification of the process, in which percolation is eliminated altogether, and the whole of the ore ground fine enough to be treated by settlement, decantation of clear solution, and filter- pressing of the settled pulp. It is doubtful, however, whether the reground sand would in all cases yield its gold values to the solu- tion as rapidly as is the case with true slime. In ores where the values are less rapidly dissolved, some form of mechanical agitation and aeration will generally be necessary. 284 THE CYANIDE HANDBOOK In the United States, similar methods have been adopted at the Liberty Bell mine, Colorado, and at various plants in the Black Hills, South Dakota, and in Nevada. In cases where the finely ground pulp is treated by suction filters on the Moore or Butters principle, it is found advisable to pass the overflow from the first set of hydraulic cone classifiers through a second set, returning the coarse products from each set for regrinding, or delivering them to the leaching tanks, and allowing only the fine product from the second classifiers to go to the slime filter tank. It is found that these filters are best adapted for treating slime which is as free as possible from fine sand. 1 The presence of a little slime in the sand treated by percolation is of less consequence, as the material is very thoroughly mixed by the use of the Blais- dell sand distributor and by double treatment. In the Black Hills, 2 on the other hand, where the Merrill filter-press is in use, the practice is to make as clean a sand as possible. The stream of pulp is passed through a succession of cones, the final ones being arranged as spitzlutten that is, they have an upward stream of solution introduced at the bottom. The fine product treated by filter-pressing contains 15 to 20 per cent, of sand passing a 150-mesh screen. The battery pulp is delivered to the cones by means of sand pumps, and solution is added in the launder which carries the sands to the distributor, so as to dilute the pulp in the ratio of 5 to 1. The battery solu- tion has a strength of .06 to .065 per cent. KCN and .04 per cent. NaOH; the solution after precipitation carries .075 to .08 per cent. KCN and .05 to .06 per cent. NaOH. (These figures refer to the practice at the Maitland properties, South Dakota, in 1904.) At the Desert mill (Tonopah Mining Company), Millers, Nevada. 3 the ore is crushed with a .15 per cent, solution, using 7 tons solution per ton of ore. The pulp is concentrated on Wilfley tables, the tailings from which pass to cone hydraulic classifiers, and clean slime is separated from them as already described. The sand goes to collecting vats, in which it is drained. It is then excavated by the Blaisdell apparatus and carried by conveyors 1 This is contrary to the experience in other localities, where a small proportion of fine sand in the slime is found to be essential to the successful working of suc- tion filters. 2 J. Gross, "Trans. A. I. M. E.," XXXV, 616. s A. R. Parsons, " Min. Sci. Press," XCV, 494. CRUSHING WITH CYANIDE SOLUTION 285 to leaching vats; lime is added in the collecting vats and lead acetate during the transfer. The solutions used in leaching con- tain .25 to .15 per cent. KCN. After five days' treatment, includ- ing transfer, the sand is passed to a second set of vats and similarly treated for another five days, then to a third set for three to five days, the total time of treatment being twelve to fifteen days. The residue carries 0.6 dwt. gold and 3.1 oz. silver. The slime is collected in vats of 36 ft. diameter by 20 ft. deep, with rim overflow, the clear solution being returned to the battery storage tank. The thickened pulp goes to agitation vats previously filled with precipitated solution, to which the requisite quantities of lime and cyanide have been added. Agitation is continued for thirty hours, assisted by injection of compressed air and cir- culation with centrifugal pump, the dilution of pulp being about 4 to 1. After settling for six hours, the clear solution is drawn off for precipitation and the slime is agitated with fresh solution for a further twenty-four hours, after which it goes to the Butters filter plant. In Mexico, the process is in use at various plants treating gold and silver ores. At Butters' Copala mines, Sinaloa, the ore treated carries gold $1.96 and silver 15.8 oz. It is crushed by stamps with 12-mesh screens, with a solution containing .07 per cent. NaCN and .135 per cent, alkali calculated as NaOH, using 16 tons of solution per ton of ore. The battery product goes to cone classifiers, from which the coarse product goes to tube mills and the fine product to a second classifier. The coarse product from the latter also goes to the tube mills, the effluent from which returns to the second classifier. The fine product from the second classifier goes to a third, which makes the final separation of 42 per cent, sand 58 per cent, slime. The sand is leached with 0.3 per cent. NaCN, and the slime is treated in agitation vats and Butters filters. SECTION III ROASTING BEFORE CYANIDE TREATMENT THE conditions under which roasting is necessary before cyanide treatment have been briefly referred to in Part III of this book. In general, it may be stated that roasting is only necessary or advantageous in cases where the gold is combined, or closely associated, with some element which prevents or re- tards its solution in cyanide. The chief applications of roasting in practice are: (1) to telluride ores, such as those of Cripple Creek and Western Australia, where some part of the gold is probably in chemical combination with tellurium, the compound being unattacked by ordinary cyanide solution, even in presence of excess of dissolved oxygen; (2) to arsenical sulphide ores, such as those of the Mercur district, Utah. Two types of furnace are in general use, viz.: (1) furnaces with single or superimposed hearths and revolving rabbles, such as the Edwards and Merton furnaces; (2) furnaces consisting of an inclined revolving cylinder, such as the Argall roaster. Roast- ing by hand in reverberatory furnaces was formerly employed as a preliminary operation in the cyaniding of concentrates on the Rand and elsewhere, but is now superseded in most parts of the world by mechanical contrivances. FURNACES WITH REVOLVING RABBLES The Edwards furnace consists of a single hearth enclosed in an iron framework, and supported above the ground on central pivots, which enable the hearth to be tilted to any required angle to aid the passage of the ore along the bed of the furnace. The furnace is kept at the desired inclination by means of screw- jacks. The rabbles are supported on vertical shafts passing through the roof of the furnace, and are driven by means of a horizontal shaft and gear-wheels above the furnace. To these vertical shafts are attached a number of removable plows or rakes, 286 ROASTING BEFORE CYANIDE TREATMENT 287 so arranged that each revolves in the opposite direction to the succeeding one. The ore is thus made to travel across the fur- nace in a zigzag direction. In some cases the five rabbles nearest the fire-box are water-cooled; the shafts are made hollow and a stream of water is allowed to descend through the central pipe, pass through a cavity in the rabble itself, and escape into a cir- cular channel surrounding the vertical shaft above the furnace arch. The ore is fed in at the receiving end of the furnace by means of fluted rolls, consisting of cast-iron cylinders having eight V-- shaped channels on the circumference; these rolls are driven by toothed gearing. The rabbles then gradually work the ore toward the discharge end (nearest the fire-box), where the heat is greatest. The roasted ore is finally discharged through an opening on to a " push conveyor." This consists of a semicircular trough provided with transverse blades free to move only in one direction. The conveyor moves horizontally to and fro on rollers, and at each forward movement the ore is pushed by the blades a distance of 20 in.; on the return movement the blades are free to swing and pass back over the heaps of ore formed by the forward movement. The ore is thus gradually pushed to the end of the trough, and at the same time turned over by the blades, thus exposing fresh surfaces and allowing an opportunity for improving the final roast. The roasted ore is then ready for cyanide treatment, and may be discharged directly from the con- veyor into a stream of solution. 1 The Merlon furnace has three superimposed horizontal hearths connected by vertical discharge holes. The rabbles, as in the Edwards furnace, are carried by vertical shafts, of which there are four or five, each shaft passing through the whole series of hearths, and being mounted at the lower end on a footstep below the bottom hearth. Attached to each shaft are three rabbles, one on each hearth, arranged so that the radii of the circles they describe in revolving are a little less than the distance between the shafts. By this means the ore is passed successively from one rabble to the next, and eventually falls through the discharge door on to the hearth below. A sliding bar at the discharge doors regulates the rate of discharge. There is also a finishing hearth next the fire-box, with rabbles which are sometimes water- i E. W. Simpson in " Trans. I. M. M.," XIII, 27-33. 288 THE CYANIDE HANDBOOK jacketed. The vertical shafts are driven by bevel or worm- gearing at 1 to 2 r.p.m., and are provided with special arrange- ments for counteracting the effect of expansion. The Merton furnace is more economical, both in first cost and in working, than the Edwards, but has the disadvantage that the inclination of the hearths cannot be varied. There is also more difficulty in withdrawing the shafts of the rabbles for repair. Doors are provided at the ends and sides for admitting air and for removing and renewing the rabbles. The capacity of this furnace varies from 5 to 25 tons per day, for ores containing from 25 to 6 per cent, of sulphur. The roasting is very perfect, and the percentage of sulphur may be reduced to .05 per cent, or less. 1 In Australia the fuel used is generally eucalyptus wood, but oil or gas may be used with advantage, and give a more regu- lar temperature. The accompanying drawing (Fig. 38) shows a recent type of this furnace in plan and sectional elevation. REVOLVING CYLINDRICAL FURNACES Several furnaces of this type have been in use for many years for roasting ores for the chlorination process, as, for example, the Bruckner cylinder. Most of these, however, are costly, both in initial expense and in repairs. We shall here describe only the Argall roaster, which has been successfully applied in pre- paring for cyanide treatment the ores of the Cripple Creek district, Colorado. A similar apparatus is used as a drier for ore previous to crushing in rolls. The furnace consists of four or more parallel steel tubes lined with brick or tile, fire-brick being used in the parts exposed to the greatest heat. These tubes are mounted together, so as to revolve as a single cylinder supported on two tires. The latter revolves on friction wheels, operated by differential gear. (See Figs. 39 and 40.) The apparatus is inclined from the feed to the discharge end at a slight angle. The ore to be roasted is fed in mechanically by a shoot, and travels in a thin layer through each of the revolving tubes; the discharge end is provided with a hood with suitable openings, through which the ore drops con- tinuously into a hopper beneath. This hood opens into a station- ary cylinder of steel plate which conveys the heated gases directly from the fire-box to the hood and thence to the tubes. The fire- box is mounted on wheels, and can thus be easily withdrawn from 1 Julian and Smart, loc. cit., p. 439. ROASTING BEFORE CYANIDE TREATMENT 289 FIG. 38. Merton Furnace. Plan and sectional elevation. [From drawing furnished by Fraser & Chalmers.] 290 THE CYANIDE HANDBOOK ROASTING BEFORE CYANIDE TREATMENT 291 292 THE CYANIDE HANDBOOK the furnace if necessary. The layer of ore in the tubes gradually diminishes from the feed to the discharge end, so that it is thinnest where exposed to the greatest heat. It is claimed that this furnace gives a practically perfect roast, with less dust than other furnaces and small expense for repairs. The accompanying illustrations are furnished by the Cyanide Plant Supply Company. In all processes of treating roasted ore with cyanide, it is advisable to cool the ore before charging into the treatment tanks. This is done sometimes by spreading on special cooling floors. In other cases the ore is sufficiently cool as it is discharged from the conveyors to be fed direct into a vessel or launder con- taining the cyanide solution. If the ore is sufficiently hot to raise the solution to boiling-point, however, a very large con- sumption of cyanide takes place, and the extraction is ineffective, because the oxygen is thereby expelled from the solution. SECTION IV USE OF AUXILIARY DISSOLVING AGENTS OXIDIZERS IN CONJUNCTION WITH CYANIDE As soon as the function of oxygen in the reaction taking place between gold and the alkaline cyanides was recognized, the sug- gestion of adding a reagent to supply the necessary oxygen in an active and concentrated form naturally occurred to many in- vestigators. A very large number of substances have been tried experimentally, and many processes involving their use have been suggested and patented; but the only oxidizers which appear to have had even a limited application in practice are: (1) At- mospheric air, injected under pressure into the pulp or solution, or beneath the filter-cloth; (2) sodium peroxide; (3) potassium permanganate; (4) manganese dioxide. Other substances, such as potassium ferricyanide, have been used, which, although not, strictly speaking, oxidizers, may aid the reaction by liberating cyanogen in an active form. Oxidizers act in two ways: by supplying oxygen in a nascent or active condition, so accelerating the solution of gold; by oxidiz- ing deleterious impurities that may be present in the ore or solu- tion. The general conclusion arrived at by many experimenters in this direction is that, with a given ore crushed to a given degree of fineness, the ultimate maximum extraction of gold which can be attained is the same, whether an oxidizer be employed or not. The only advantage secured by the addition of the oxidizer is greater rapidity of extraction. In practice, this advantage is frequently nullified by the fact that much time is necessarily consumed in washing out the dissolved gold, and as this washing is done, as a rule, with cyanide solution, the dissolving process is carried to its limit in either case. Sodium peroxide has been used to some extent in America in connection with the so-called Kendall process. The reagent is 293 294 THE CYANIDE HANDBOOK generally mixed with dry-crushed ore previous to cyanide treat- ment. It gives up its oxygen very rapidly in contact with water; hence, if mixed with the moist ore or tailings previous to or during the process of charging into the vat, the oxygen will be liberated and the greater part of its efficiency lost before the cyanide solu- tion has come in contact with the ore. Potassium permanganate has been occasionally used in South Africa in the form of a preliminary wash, especially in the treat- ment of concentrates. It should, however, be removed by water- washing previous to cyanide treatment, as any excess remaining reacts with and destroys the cyanide. In some instances a mix- ture of permanganate and sulphuric acid has been used for the preliminary oxidation and removal of cyanicides. Manganese dioxide may be added with advantage in some cases. It gives up oxygen gradually, and thus the beneficial effect is continued throughout the treatment. 1 In this respect manganese dioxide is superior to sodium peroxide and similar reagents, but if used in excessive amounts it causes large loss of cyanide, and may also give rise to objectionable deposits in the zinc-boxes. The use of potassium ferricyanide in conjunction with cyanide was suggested by C. Moldenhauer in 1892. It is not in reality an oxidizing agent, but reacts by liberating a part of cyanogen from the cyanide and forming potassium ferrocyanide KsFeCyo + KCy = K 4 FeCy 6 + Cy Au + Cy + KCy = KAuCy 2 . It thus, theoretically at least, enables the cyanide to dissolve gold in the absence of oxygen. The reaction is very rapid and effective, but other secondary reactions occur which give rise to an accumulation of salts in the solution and deposits in the zinc- boxes, both of which are detrimental. Quantities of Oxidizers to be Added. No definite rule can be given as to the amounts required, as this will depend on the gold or silver value of the material to be treated and on the nature and amount of the oxidizable materials present. In all cases small- scale preliminary experiments with varying quantities should be made, and the effect on precipitation tested as well as on extrac- tion. 1 As is well known, compounds of manganese exhibit a tendency to act as " car- riers " of oxygen, by alternately combining with this element and giving it up to any substance capable of absorbing it. USE OF AUXILIARY DISSOLVING AGENTS 295 THE BROMOCYANIDE PROCESS An auxiliary agent which for several years was in extensive and successful use was introduced in 1894 by Sulman and Teed. This consists of the bromide of cyanogen, BrCy. Other haloid compounds of cyanogen have been proposed, and their use was claimed in the Sulman-Teed patents, but they have never been successfully applied in practice. The reagent is a volatile crystal- line solid, soluble in water, and when mixed with an alkaline cyanide forms a very powerful and rapid solvent for gold. Its action is discussed in Part II, but it may be repeated here that the reaction of bromide of cyanogen on alkaline cyanides is accom- panied by the evolution of cyanogen or hydrocyanic acid, and that in presence of an excess of alkali the bromide of cyanogen is immediately destroyed with formation of bromide, cyanate, and (probably) bromate of the alkali metal. It must therefore be used in solution containing little or no free alkali, and is most effective if added in small quantities at a time. The bromide of cyanogen is not per se a solvent of gold. The simplest theory of the reaction is represented thus: KCy + BrCy = KBr + Cy 2 . Au + KCy + Cy = KAuCy 2 BROMOCYANIDE PRACTICE IN WESTERN AUSTRALIA The chief centers in which the bromocyanide process has been applied are the Deloro mine, Ontario, Canada, where a mispickel ore was treated without roasting, and at various mines in the Kalgurli district, Western Australia, where it was introduced in conjunction with tube milling, or fine grinding in pans, under the name of the Diehl process. Details of the process will be found in the following papers: "The Sulman-Teed Gold Extraction Process," Sulman and Teed, in "Journ. Soc. Chem. Ind.," XVI, 961 (1897); "The Diehl Cyanide Process," by H. Knutsen, in "Trans. I. M. M.," XII, 2 (1902); "Metallurgy of the Kalgoorlie Goldfield," by Gerard W. Williams, in "Eng. and Min. Journ.," LXXXV, 345 (Feb. 15, 1908). The following particulars are summarized from these publications. The ores to which the process is applied in Western Australia are sulpho-tellurides, the gold in which cannot be extracted by treating raw with ordinary cyanide solutions. In the Diehl 296 THE CYANIDE HANDBOOK process the operations are: (1) Crushing and grinding to a high degree of fineness. (2) Treating the finely ground ore by agita- tion with a cyanide solution, to which bromide of cyanogen is added at intervals. (3) Separating solution from sludge by means of filter-presses. (4) Precipitation on zinc shavings. In some plants amalgamation and concentration are used, the concentrates being generally roasted and treated by ordinary cyanide. The grinding is done in Krupp tube mills, 18 ft. long by 4ft. diameter, the fine sands being reground until the whole product passes a 200-mesh sieve, less than 3 per cent, remaining on 220. The slimes, after removal of excess of water by spitzkasten, go to agitators, where they are treated first with ordinary cyanide. The dilution of pulp is about 2 parts ore to 3 of water. Cyanide is added in quantity sufficient to form a solution of 0.10 per cent. KCy, equivalent to 0.15 parts KCy per 100 parts of dry slime. After from one and a half to two hours' agitation with cyanide alone, the bro'mocyanide is added in the proportion of 0.04 parts per 100 parts of dry slime, equivalent to a strength of about 0.027 per cent. BrCy in the solution. This mixture is agitated for twenty-four hours, lime being added two hours before the finish in the proportion of 3 to 4 Ib. per ton of dry ore. The sludge is then run into a receiver, from which it is forced by com- pressed air into the filter-presses. The above details refer to the treatment at Hannan's Star mill, 1902. The system varies slightly at different plants, but in general the amount of bromocyanide used is from J to J that of the cyanide, corresponding approximately with the proportions indicated by the equations given above. According to G. W. Williams (loc. ciL), the treatment at the mines of the Ivanhoe Gold Corporation is as follows: After con- centration on Wilfley tables, the sands are ground in pans, recon- centrated, the fines separated by spitzlutten and conveyed to settlers. The slimes after settlement are pumped to agitators, where they are diluted with weak solution to form a pulp with the proportions 1:1, dry ore to solution. After two hours' agi- tation, 0.6 Ib. bromocyanide is added per ton of dry slimes and the agitation continued for twelve hours. Lime is then added in the proportion of 1 Ib. per ton of dry slimes. The pulp is then filter-pressed. The agitators are 20 ft. in diameter by 8 ft. deep, closed in and fitted with mechanical stirring gear. The method USE OF AUXILIARY DISSOLVING AGENTS 297 involves a separate treatment of the concentrates, which amount to 18 per cent, of the total tonnage. These are roasted in Edwards furnaces, mixed with spent cyanide solution, ground fine in pans, agitated with 0.1 per cent, cyanide and filter-pressed. The entire scheme of treatment comprises Recovery: per cent of total gold. Amalgamation 28 Treatment of concentrates by roasting and ordinary cyanide .... 13 Treatment of sands: 5 days percolation 17.5 Treatment of slimes by bromocyanide 28 Total recovery 86.5 The total treatment cost is given as $2.20 per ton. The recovery by similar systems of treatment in other plants varies from 85 to 95 per cent. Other plants treating a portion of their product by bromocyanide are the Golden Horseshoe, Oroya-Brownhill, and Lake View Consols. Owing to the instability of the reagent and the unpleasantness of handling it in quantity, mixtures of various salts are prepared and shipped to the mines, which, on addition to the cyanide solu- tion, produce bromo-cyanide in the required proportion, for instance, a mixture of cyanide, bromide, bromate, and bisulphate of sodium. USE OF MERCURIC CHLORIDE Keith and Hood have proposed the addition of mercuric chloride or of a double cyanide of mercury to the solution. It is found that the action of the cyanide on gold, and still more on silver, is greatly accelerated. This reagent is occasionally em- ployed as an auxiliary to cyanide in the treatment of ores con- taining silver as sulphide. The presence of small amounts of mercury in the solution is also beneficial in assisting the pre^ cipitation of the precious metals in the zinc-boxes. The solvent effect on sulphide ores is apparently due to the great affinity of mercury for sulphur, and may be represented thus: A g2 S = HgS + 2KAgCy 2 SECTION V ELECTROLYTIC PRECIPITATION PROCESSES THE idea of using an electric current for precipitating the pre- cious metals from their cyanide solutions dates at least as far back as 1840, when Elkington's patent for electroplating and electro- gilding was taken out (E. P. No. 8447; Sept. 25, 1840). In fact, the supposition that an electric current was necessary, at least for the precipitation if not for the solution of metals in cyanide, appears to have been general among investigators up to the time of the discovery of zinc-thread precipitation. THE SIEMENS-HALSKE PROCESS About the year 1887 the firm of Siemens and Halske, of Berlin, introduced an electrolytic process for depositing copper, zinc, gold, silver, etc., from cyanide solutions. (See E. P. No. -3533; Feb. 27, 1889.) This process was subsequently applied to the treatment of solutions resulting from the cyanide extrac- tion of gold ores and tailings, and for several years found an extensive application in South Africa, where it was introduced by A. von Gernet in 1893. In its original form, this process consisted in precipitating the metals from cyanide solutions by means of an electric current passed between anodes of iron and cathodes of sheet lead, the precious metals being deposited on the latter, while the iron was gradually converted into oxide, and ultimately into soluble and insoluble cyanogen compounds, such as ferrocyanides. Some account of the introduction and working of the process at the plant of the Worcester mine, Johannesburg, where it was first applied in South Africa on a working scale, is given by von Gernet. 1 The method of applying this process usually adopted in South Africa was to place the anode and cathode plates in the compartments of a box similar to the ordinary zinc-box, but of larger dimensions, so as to allow from 15 to 25 cu. ft. per ton of solution per 24 hours. i " Proc. Chem., Met. and Min. Soc. of S. A." I, 28 (Aug. 18, 1894). 298 ELECTROLYTIC PRECIPITATION PROCESSES 299 The plates were commonly arranged parallel to the sides of the box; in each compartment the anodes and cathodes were placed alternately side by side, the cathodes of one compartment being connected electrically with the anodes of the next lower com- partment. The anodes of the top compartment were connected with the positive pole of a battery, dynamo, or other source of current, the cathodes of the bottom compartment being connected with the negative pole of the same. The anodes were rectangular plates of sheet iron, J to J in. thick, with an effective area of, say, 5 sq. ft. They were usually held in grooves or by means of slats at the sides of the box, and were enclosed in canvas bags to insulate them from the cathodes and to collect the products of the reaction (ferric oxide, etc.) referred to above. The cathodes were of thin sheet lead suspended on horizontal wires between each pair of anodes, at about 3-in. distances. As it was found advantageous to increase the cathode area as much as possible, in later practice the lead sheets were cut into long strips and the threads separated as much as possible so as to give the maximum surface for precipitation. A current of .04 ampere per sq. ft. of anode surface was usually sufficient, and was found to deposit the gold on the cathodes in a coherent form. When sufficiently coated with precious metal, the cathodes were removed and replaced by fresh ones, without interrupting the regular working of the box; they were then melted in a reverberatory furnace and cast into lead-bullion bars. These were then cupeled in an ordinary cupeling furnace. A part of the lead was recovered for re-use by reduction of the litharge resulting from the latter operation. A further quantity of gold was obtained by treating the sludge which formed inside the sacks containing the anodes. The chief advantages claimed for the Siemens-Halske process were: (1) The possibility of precipitating gold. from solutions very weak in cyanide, which did not admit of treatment by the ordinary zinc process. It was thus practicable to employ only the minimum quantity of cyanide needed for dissolving the gold. This was of particular importance in the treatment of slimes by the decantation process, where the volume of liquid used per ton of material treated was necessarily large. (2) The purity of the bullion obtained by cupellation of the lead bars as compared with that produced by the ordinary smelting of zinc precipitate. 300 THE CYANIDE HANDBOOK (3) There was no necessity for interrupting the regular work of the plant during the clean-up. (4) Copper and other obnoxious base metals were removed from the solution very effectively. (5) There was no accumulation of double cyanide of zinc, etc., in the solution, which remained always in good working condition and did not become charged with foreign salts. On the other hand there were certain evident disadvantages: (1) The necessity, when dealing with solutions low in gold value, of a large surface for deposition, involving the use of boxes of great size, and hence considerable initial outlay. (2) More skill and attention was needed to secure satisfactory working than with zinc precipitation. With the introduction of the zinc-lead couple in 1898, the first and principal advantage of the Siemens-Halske process dis- appeared, as it was then found possible to precipitate effectively by means of zinc from solutions as weak in cyanide as those dealt with by the electrolytic process. By the use of the Tavener process (see Part V), or by various methods of refining zinc pre- cipitate, as for example by fluxing with dioxide of manganese in clay-lined pots, it is also possible to obtain bullion from the zinc process having the same fineness as that furnished by cupella- tion in the Siemens-Halske process. LATER DEVELOPMENTS OF ELECTROLYTIC PROCESS As the action of the solution on the anodes gave rise to a troublesome by-product, many attempts were made to find a material for the anodes which would be insoluble in the cyanide solution. Thick sheet lead was used for some time by Charles Butters, but this had a tendency to become hard and brittle, and under certain conditions became rapidly coated with cyanide of lead, so that after use for a short time the plates crumbled to pieces. E. Andreoli and others used anodes of lead coated with a layer of peroxide of lead; this was obtained by immersing the sheets in a bath of potassium permanganate or of plumbate of soda. They were then immersed in strong cyanide solution and an electric current passed through, which caused the coating of peroxide to become firm and hard. These plates were practically insoluble in the solution, and hence no by-products were formed. Many attempts were also made to find a cathode from which the deposit could be removed as required without destruction ELECTROLYTIC PRECIPITATION PROCESSES 301 of the cathode itself. Sheet-iron carefully freed from oxide has been used; the gold deposit is removed by immersing the cathode in a bath of molten lead. A more satisfactory method, however, is to use a cathode of tinned iron. This was applied successfully at Charles Butters and Company's works at Minas Prietas, Mexico, and at Virginia City, Nevada. The gold and silver are deposited in a loose form, so that they may be rubbed off from time to time, to collect as a sludge at the bottom of the box. This is cleaned up at intervals and smelted direct, yielding high-grade bullion, while the cathode remains intact. When much base metal, e.g., copper, is present in the solution, however, there is a dense hard deposit on the plates, which is scraped off with difficulty; also a considerable amount of gold adheres to the plate before any of the deposit begins to fall. off. ELECTROLYTIC PROCESS FOR GOLD-COPPER ORES The following system of combined electrolytic and zinc pre- cipitation was used by C. P. Richmond, at the San Sebastian mine, Salvador, Central America, 1 for the treatment of a complex ore containing copper, arsenic, antimony, and tellurium. The ore is crushed in ball mills and roasted. The sand and slime are then separated in cyanide solution, the sand being treated by percolation and the slime by agitation and filter-pressing. The greater part of the gold, together with some copper, is precipi- tated by electrolysis, after which the solutions are passed through a zinc-box to recover the residual gold. The accumulation of copper in the solution is thus prevented. The cyanide solution, averaging 16 dwt. per ton, flows through the electric box at the rate of 150 tons per 24 hours; the box is 30 ft. long by 10 ft. wide, and 4 ft. 8 in. deep, and is inclined 1 in. per foot. There are twelve compartments, but the last two are used for settlement only, so that the effective capacity is 1 166 cu. ft., or about 8 cu. ft. per ton per 24 hours. Each compartment has 25 anodes and 24 cathodes. The anodes are rolled lead plates 22 X 48 X J in. The cathodes are lead plates of similar size, but only ^ in. thick. The anodes are peroxidized by immersing them in a solution of potassium permanganate, with or without sulphuric acid. One method is to dip them in a solu- tion of 1 per cent. KMnO 4 and 2 per cent. H 2 SO 4 , keeping them i Chas. P. Richmond, " E. and M. J.," LXXXIII, 512. 302 THE CYANIDE HANDBOOK for six hours under a current of 2.5 amperes per square foot. By this means a coating of lead peroxide is obtained which lasts 8 to 12 months. In the boxes, the anodes and cathodes of successive compart- ments are connected in series, with a current strength of 1 ampere per square foot. The gold and copper form a hard dense coating on the cathodes, with no tendency to fall off as sludge, but there is a gradual accumulation of low-grade precipitate on the anodes and at the bottom of the box. After passing the electric box, the solution flows through two zinc-boxes 22 ft. long, with fourteen compartments, each 2 X 2 X 1 ft., which reduces the gold con- tents to about 0.06 to 0.10 dwt. per ton. Hardly any copper is deposited in the zinc-box. The electrical box precipitates 80 to 90 per cent, of the gold, together with much copper, which is then separated as follows: After remaining in the cyanide electric box for 20 to 30 days ; the cathodes are transferred to another box, where they are used as anodes. This box contains sulphuric acid of 2 to 3 per cent. H 2 SO 4 , the cathodes used being lead plates T V in. thick. There are four compartments, each containing 5 anodes and 6 cathodes, 4 in. apart and connected in series. The current used is 5 amperes per square foot of anode surface. The anodes are hung in a wooden frame with closed bottom and open sides, over which is stretched a sack of cotton cloth. The copper dissolves and precipitates on the cathodes, where it forms a slime, which falls to the bottom. The gold remains in a loose form in the anode sacks. The anode plates are then washed and returned, as cathodes, to the cyanide electric box. Special arrangements are employed to secure circulation of the acid liquid and to pre- vent short-circuiting. To recover the gold, the acid liquid is syphoned off, and the anode frames are raised and allowed to drain over the box. They are then suspended over a filter-tank, the sacking at the sides is cut away, and the gold precipitate washed into the tank. The frames and old sacks are washed in the same tank. When thor- oughly drained, the precipitate is removed from the tank, dried, and smelted in graphite pots. The copper slime from the cathodes is flushed through holes near the bottom of the box, filtered, and dried. The bullion recovered is high-grade, and owing to the constant precipitation of copper there is a regeneration of cyanide amounting to 30,000 Ib. in a year's run. Sufficient copper is ELECTROLYTIC PRECIPITATION PROCESSES 303 extracted to cover the cost of refining by the acid box and smelt- ing of gold precipitate. OTHER ELECTROLYTIC PROCESSES Numerous processes have been suggested and tried experi- mentally, but very few have achieved even a temporary com- mercial success. In the so-called " electro-zinc " process, zinc plates were used as anodes, and amalgamated zinc or copper plates as cathodes. The anode being soluble, a smaller current could be used than with insoluble or slightly soluble anodes, but the method was unsatisfactory owing to the rapid accumulation of zinc ferrocyanide. The effect of the current is also to produce a hard non-adherent amalgam. Many attempts have been made to use carbon as anode, but it is difficult to obtain it in a form which will not disintegrate under the current. Aluminium has been used as cathode by S. Cowper Coles, but there appears to be much difficulty in removing the precipitated gold. Mercury is not suitable as a cathode, for the reason pointed out by von Gernet, that it is not practicable to obtain the enormous area required for effective precipitation of solutions low in gold. Amalgamated copper plates have, however, been frequently tried, and an interesting series of experiments with them has been made by T. K. Rose. 1 These experiments show that with solutions rich in gold, a current of 0.03 amperes per square foot precipitates the gold as a black powder which does not amal- gamate. Other objections arise from the cost of the process, the fact that the plates absorb a portion of the gold, and that they are corroded by the action of the current. Douglas Lay 2 enumerates several drawbacks inherent to electrocyanide processes in general. A very weak current must necessarily be used, to avoid excessive action on the anodes; this in turn involves, as already pointed out, a very large cathode area when solutions low in gold are to be treated. He also states that deposition of carbonate of lime on the cathodes takes place, involving a decomposition of cyanide and so coating the cathodes as to prevent precipitation of the precious metals. !" Trans. I. M. and M.," VIII, 369. *"E. and M. J.," April 11, 1908, p. 765. 304 THE CYANIDE HANDBOOK SOLUTION AND ELECTROLYTIC PRECIPITATION IN THE SAME VESSEL The idea of dissolving and precipitating the gold simultaneously in one and the same tank or receiver is an old and favorite one with inventors; it figures in J. H. Rae's apparatus, patented in 1867, and reappears in various forms in the Pelatan-Clerici, Riecken, Gilmour- Young and other processes, in which an electric current also is generally introduced, as an aid both to solution and precipitation. There appears, however, to be very little evidence that the electric current aids the solution of the precious metals in any way, and with regard to precipitation it is for various reasons more effective in solutions from which the sus- pended solid matter has been removed. In the Pelatan-Clerici process, used at Delamar, Idaho, in 1897, the pulp was agitated with cyanide solution in vats, the bottoms of which were composed of sheet-copper, covered with mercury, and constituting the cathodes. The anodes were in the form of sheet-iron agitators, so that the gold could be dis- solved and precipitated in the same vessel. The gold was re- covered in the form of amalgam. The Riecken process was introduced at South Kalgurli; Western Australia, in 1900. The ore is roasted, mixed to a thick pulp with cyanide solution, and fed to agitation vats, of about 13 tons capacity, having vertical ends and sloping sides. The latter are provided with amalgamated copper plates, over which a thin stream of mercury constantly flows. The coarse gold amalgamates directly. An electric current is passed through the vat by means of iron rods, which constitute the anodes, the amalgamated plates forming the cathodes. The current used is 150 amperes at 1.5 volts, with a density of 0.4 ampere per square foot. A charge of 13 tons can be treated in 18 hours with an extraction of 93 per cent, from ore assaying over an ounce per ton. The cost, exclusive of crushing and roasting, is given as 9 s. 4.8 d. per ton. 1 In the Molloy process the cathode consisted of mercury, and the anode of peroxidized lead immersed in sodium carbonate, contained in a compartment which separated it from the cyanide * T. K. Rose, " Metallurgy of Gold," 4th edition, p. 346. ELECTROLYTIC PRECIPITATION PROCESSES 305 solution. The result of the electrolysis was to form sodium amalgam which permeated the mercury in the outer vessel. The gold was precipitated and amalgamated, with regeneration of cyanide, thus: Na + KAuCy 2 = Au + KCy + NaCy SECTION VI OTHER PRECIPITATION PROCESSES IN this section we propose to give a brief description of various processes not involving the use of an electric current, which have been proposed as alternatives to the ordinary method of zinc pre- cipitation. These* may be classified as follows: (1) Methods in which the values are dissolved and amalgamated in the same vessel. (2) Precipitation with finely divided zinc (dust or fume). (3) Precipitation with charcoal. (4) Precipitation with cuprous salt in acid solution. The Gilmour-Young process, employed at the Santa Francisca mine, Nicaragua, consists in agitating the gold-bearing material, which contains much clay, as a thick pulp, in pans, with addition of mercury and cyanide. After two hours of this treatment, copper and zinc amalgam are added and the pan run for four hours longer. Thus the same vessel is used for solution and precipitation. The gold and silver are obtained as amalgam, which is separated and retorted in the usual way. Extraction of 80 to 90 per cent, of the gold is said to have been obtained. For details see "Min. Ind.," VII, 334. 1 Zinc-dust Precipitation. This method was originally applied by H. L. Sulman 2 and H. K. Picard, at the Deloro mine, Ontario, Canada. It has been extensively used in the United States, notably at Mercur, Utah, and in the Black Hills, South Dakota. The zinc-dust consists of the finely divided metal obtained by the distillation and condensation of zinc, and generally contains about 90 per cent, of metallic zinc. A small amount of lead is usually present, which is advantageous for precipitation. The zinc is very finely divided, over 95 per cent, of the powder passing a 200-mesh sieve. In this condition it is very easily oxidized, and Sulman has proposed the use of ammonia and ammonium salts for dissolving the oxide and cleansing the surface previous to use. lAlso, "Trans., I. M. and M." Nov. 16, 1898. 2Eng. Patents Nos. 18003 and 18146, 1894. 306 OTHER PRECIPITATION PROCESSES 307 The following description of the use of this reagent is given by W. J. Sharwood. 1 The solution to be precipitated is run into a collecting tank and agitated for a few minutes. A suitable quantity of the dust is then introduced, which may be sprayed into the tank as an emulsion. The mixture is then forced by a pump through a filter-press, from which the barren liquid flows to the storage tanks, the precipitated metals and any excess of zinc remaining in the press. These form a porous mass which offers slight resistance to the passage of solution. The extra power required, as compared with zinc-thread precipitation, is slight, but some trouble is experienced, owing to the clogging of the filter-cloths when the solutions are turbid with suspended siliceous matter. From J to J Ib. of zinc-dust is used per ton of solution treated. About 6 tons of liquid per hour can be passed through every 100 sq. ft. of filter-cloth in the press. When a clean-up is to be made, the contents of the press are washed by pumping a little water, then partially dried by forcing air through it. When the press is opened and the frames shaken, the cakes fall out readily. At this stage, the product usually carries about one-third of its weight of water; its composition does not differ materially from that of the precipitate obtained with zinc sha- vings, and it may be treated in the same way. It is, however, more uniform, owing to the absence of threads of "short zinc." The blowing operation oxidizes some of the zinc and sometimes causes the presses to become warm. The evolution of hydrogen which accompanies precipitation continues slowly while the moist precipitate remains in the press. Care must therefore be taken to bring no naked lights near while the clean-up is proceeding. One filter-press, with 16 frames 2 ft. square, working 15 to 16 hours per day, requires 2 h.p. (neglecting the additional fric- tion in pumps) in excess of the power required for pumping when zinc shavings are used. In other cases, however, difficulties have been experienced in treating the precipitate after removal from the presses. It was found to pack so hard that it was difficult to disintegrate it for the acid treatment, so that it was practically impossible to dissolve the zinc in dilute sulphuric acid. The particles are said to become coated with a thin layer of metallic i " E. and M. J.," LXXIX, 752 (from 13th annual convention California Miners' Assoc., December, 1904). 308 THE CYANIDE HANDBOOK gold which protects them from the acid, but a mixture of nitric and sulphuric acids has been found to be effective. 1 The precipitation by zinc-dust or fume appears to be rapid and complete, five minutes' agitation being sometimes sufficient for the reaction; but the difficulties in collecting and refining the precipitate have led to the abandonment of the process in several cases where it had been originally adopted, e.g., at various plants in the Black Hills. Much ingenuity has been exercised in devising apparatus for mixing the zinc effectively with the solution, and for settling the precipitate after mixture. Picard used a conical vessel, in which the fume was fed in through a central funnel and met an ascending current of solution which was passed through a perforated conical distributor near the bottom of the cone. The settler contained a number of transverse plates of glass or smooth wood, set at an angle of 45, on which any particles not collected in the cone are deposited. 2 . Charcoal Precipitation. The fact that gold is precipitated by charcoal from cyanide solutions has long been known, though no satisfactory explanation of the reaction has yet been given. It has been alleged to be due to occluded hydrogen or hydro- carbons contained in the pores of the charcoal. The method has been used in several small plants in Victoria, Australia, on a working scale. 3 The charcoal, crushed to a suitable size, is placed in tubs about 2 ft. in diameter, having a central vertical cylinder also filled with charcoal. These are used much in the same way as Caldecott's vats for zinc precipitation; the solution passes downward through the central column and upward in the outer portion, the overflow passing in this way through a succession of tubs each at a slightly lower level than the last. As about 24 Ib. of charcoal are required for the precipitation of an ounce of gold, the quantity needed in a plant of any size would be very large. To recover the gold, the charcoal is burnt to ashes in a special furnace so arranged that the products of combustion pass through water, to avoid mechanical loss. The ash is then smelted with a mixture of sand and borax. According to experiments made by Prof. S. B. Christy, the efficiency of the charcoal rapidly falls off with continued passage of the solution. Precipitation with Cuprous Salts. This method seems to have i " Min. Sci. Press./* XCIIT, 607 (Nov. 17, 1906). 2 "Trans. Fed. Inst. Min. Eng.," XV, 417. J. T. Lowles, " Trans. I. M. and M.," VII, 192. OTHER PRECIPITATION PROCESSES 309 been suggested independently, about 1895, by Prof. P. de Wilde, of Brussels, and Prof. S. B. Christy, of Berkeley, California. It depends on the fact that gold is precipitated completely as aurous cyanide on acidifying the cyanide solution and adding a sufficient excess of a cuprous salt. Several modifications have been pro- posed, in which an attempt is made to regenerate the cyanide decomposed in the reaction, or to precipitate the cyanogen in some form from which a soluble cyanide may subsequently be recovered before precipitating the gold. It is obvious, however, that a method involving the acidification, or at least the neutraliza- tion, of 100 tons or more of alkaline liquor per day is hardly likely to find a practical application. The precipitants suggested are: (1) a mixture of copper sulphate and sulphurous acid (De Wilde); (2) freshly precipitated cuprous sulphide, to be agitated with the solution (Christy); (3) cuprous chloride dissolved in common salt. After calcining the precipitate, the oxide of copper is fco be removed by dissolving in sulphuric acid, leaving nearly pure gold. The method forms the basis of the analytical methods for determining gold in cyanide solutions proposed by Christy and A. Whitby. (See Part VIII.) SECTION VII TREATMENT OF CUPRIFEROUS ORES THE causes which render the treatment of cupriferous ores by the cyanide process particularly difficult have already been discussed in the earlier parts of this book; they may be sum- marized thus: (1) From certain minerals, notably the carbonates and oxides, copper is readily dissolved, forming a double cyanide with the alkaline cyanide. (2) The metal coats the zinc shavings in the boxes with a thin adherent metallic film, which prevents the deposition of gold and silver. (3) The presence of much copper in the precipitate gives rise to a low-grade bullion, as the metal is only imperfectly removed in the ordinary smelting and refining operations: The remedies which have been applied or proposed are, for the most part, as follows: (1) The copper is removed from the ore by a preliminary operation, previous to cyanide treatment. (2) The double cyanide is decomposed in such a manner that the copper is separated and the free alkaline cyanide regenerated. (3) A modification of the precipitation process is used, whereby the gold, silver, and copper may be precipitated simultaneously or successively. (4) The precipitate or the bullion undergoes special refining processes for separation of the copper and other base metals. PRELIMINARY TREATMENT FOR REMOVAL OP COPPER Sulphuric Acid. In cases where the copper exists as car- bonate, satisfactory extractions have sometimess been obtained by giving a preliminary wash with dilute sulphuric acid. This system was at one time applied at the Butters Plant, Virginia City, and has been used at Cobar, New South Wales, by Messrs. Nicholas and Nicols. 1 The solution used at the latter plant averaged 0.65 per cent. H 2 SO 4 . The liquor drawn off was passed iW. S. Brown, "Trans. I. M. and M.," XV, 445. 310 TREATMENT OF CUPRIFEROUS ORES 311 through boxes containing scrap-iron for precipitation and recovery of the copper. The charges in the vats were then water-washed, mixed with a sufficient quantity of lime to neutralize any remain- ing acidity, and transferred to other vats, where they were treated with cyanide, up to 0.3 per cent. KCy, for extraction of the gold. The first cyanide solution used ran off with little or no cyanide and generally with low gold values; this was precipitated in a special zinc-box and again used for the final wash before dis- charging. Some interesting features in connection with this plant were the use of charcoal after the zinc-box for the recovery of gold not precipitated by zinc from weak solutions, and the use of nitric acid in the clean-up. The acid vats had a capacity of 25 tons and the cyanide vats of 75 tons, so that three charges of the acid vats were treated together in each cyanide vat. The acid treatment consisted of 10 to 12 tons of dilute sulphuric acid (0.65 per cent.), followed by two water-washes. The acid was first partially drained off, then allowed to digest in contact with the ore for an hour or so, and finally passed slowly through the scrap-iron boxes. In the cyanide treatment, 15 tons of weak solution were used to displace moisture; this solution carried little gold, and was used again as a final wash before discharging; after passing slowly through a zinc-box it was allowed to flow through about 20 cu. ft. of packed charcoal. The precipitate from the zinc-boxes, in which much lead acetate was used, was extremely base, and was first treated with sulphuric acid for removal of the zinc, then washed with distilled water, and finally treated with nitric acid to dissolve copper and lead. In order to ,prevent gold going into solution, any chlorides present were precipitated with silver. Bullion was produced over 900 fine. Sulphurous acid has been tried as a preliminary solvent of copper by A. von Gernet and others, and is claimed to have some action not only on carbonates, but also on sulphides of copper. Ammonia. The 'use of ammonia as a preliminary solvent of copper has been suggested by H. Hirsching. 1 The ammoniacal solution is distilled, and the ammonia recovered for re-use, while the copper is precipitated as oxide. Double Cyanides of Copper. It has been shown by Scrymgeour 2 that cupriferous cyanide solutions are capable of extracting a 1 H. Hirsching, " The Ammonia Process." 2" E. and M. J.," Dec. 20, 1902, p. 816. 312 THE CYANIDE HANDBOOK further quantity of copper from certain ores. The copper is finally recovered from this solution by electrolysis, with regenera- tion of a part of the cyanides. After this preliminary wash, the ore is cyanided as usual and the solutions precipitated electrically. This process appears to depend on the instability of the cupri- cyanides of the alkalis, which are readily transformed into cupro- cyanides by taking up a further quantity of copper. PRECIPITATION OF COPPER FROM CYANIDE SOLUTIONS Several methods have been suggested for precipitating copper from cyanide solutions. We have already described the elec- trolytic method practised at Butters' Salvador mines (see Section V, above), and the effect of copper in the zinc-boxes, with and without the lead-zinc couple, has been discussed. It is well known that certain acids precipitate copper from cyanide solutions as white insoluble cuprous cyanide. It has been proposed by H. A. Barker * to use sulphuric acid for precipitating copper and gold from the solutions, which are then to be filtered and made alkaline with caustic soda; it is stated that cyanide is regenerated in the process, but the portion of cyanogen precipitated with the copper is obviously lost. The reactions are probably as follows: K 2 Cu.Cy 4 + H 2 SO 4 = Cu,Cy 2 + 2HCy + K 2 SO 4 HCy + NaOH = NaCy + H 2 O. It is also suggested that this method of precipitation might be used as an adjunct to Scrymgeour's process, described above, the weak solution used as a preliminary wash in that process being precipitated with sulphuric acid when sufficiently charged with copper. USE OF AMMONIA AS AUXILIARY SOLVENT It has been found that a mixture of ammonia or ammonium salt with cyanide forms a much more effective solvent for gold in cupriferous ores than cyanide alone. When an ammonium salt is mixed with an equivalent of an alkaline cyanide in solution, decomposition takes place, with formation of ammonium cyanide, thus: 2KCN + (NHOiSCX = K S0 4 + 2NH 4 CN. "Trans. Inst. Min. and Met." XII, p. 399 (May 21, 1903). TREATMENT OF CUPRIFEROUS ORES 313 A series of experiments made by Jarman and Brereton * appears to show that a mixture of ammonia and potassium cy- anide dissolves less copper than does potassium cyanide alone. A process based on the use of this solvent has been applied by Bertram Hunt in treating tailings from the Comstock Lode, and cupriferous ore at Dale, California. The observations on which this method is based are summarized as follows by J. S. MacArthur, in a discussion on Jarman and Brereton's results: 2 (1) A weak solution of ammonia generally does more work as a solvent of copper than a corresponding quantity of a stronger solution. (2) A solution of ammonia along with a cyanide dissolves less copper than the sum of the amounts dissolved by the solvents separately. (3) For ordinary strengths of cyanide and ordinary grades of ore, the addition of ammonia equal to 0.11 per cent, of NH 3 produces a solution having a maximum solvent effect on gold and a minimum solvent effect on copper. " Trans. Inst. Min. and Met." XIV, p. 289 (Feb. 16, 1905). it., p. 331. PART VII ASSAYING SECTION I SAMPLING (A) SAMPLING OF ORES AND SIMILAR MATERIAL IN BULK Samples of Hard, Coarse Material. The sampling of ore " in place" does not generally come within the limits of the work carried out by the staff of a cyanide plant. It may be necessary, however, in certain cases, to deal with large masses of coarsely broken rock, such as ore dumps, ore as delivered from the mine to the mill, slag heaps, and other similar accumulations of hard material, so that a few words on the methods of sampling them may not be out of place. Definition of Sample. The essential point in all sampling is to select from a relatively large mass of material, containing various ingredients unevenly distributed, a relatively small quantity which shall contain each of these ingredients in the same proportions as those which occur in the entire mass. Sampling Large Ore Heaps. A large pile of ore is generally sampled in the first instance by making cuts into it at various points round the base; when circumstances allow, a channel is dug right across the heap through the center, or, better, two or more intersecting channels. A plan sometimes adopted is to shovel alternately to right and left, and to throw every third shovelful into a barrow, the portion going into the barrow con- stituting the sample. The assumptions involved in this process are: (1) That a cut made through the heap from side to side, passing through the center, will consist of material of the same average composition as the entire heap; and (2) that every third shovelful is, on the average, of the same composition as the re- maining two. The first of these assumptions appears less justifi- able than the second; hence it is safer to make more than one channel. Coning and Quartering. The further division of the sample must generally be preceded by breaking up the larger lumps, which may be done either mechanically or by hand. For this 317 FIG. 41. Vezin Sampler, as furnished by Hadfield's Steel Foundry Co. Sheffield. SAMPLING 319 purpose it may be spread evenly over a smooth surface, such as a cement floor, and any pieces broken up that are obviously larger than the rest. The entire sample is then mixed by turning over with a shovel; to ensure thorough mixing the process of "coning" is often resorted to. A space is swept clean, and the ore thrown into this, a shovelful at a time, each shovelful being thrown on the top of the conical heap so formed, so that the material rolls down as evenly as possible on all sides. If the heap is still very large, it may be again sampled by trenching; if not too large to be conveniently handled, it is spread out in a circular flat-topped heap and " quartered " by means of boards held edgewise across the center so as to divide it into four segments as nearly equal as possible. Another method is to form the material into a ring on the floor, and make a heap by shoveling toward the center. The operations of coning and quartering are repeated succes- sively a number of times; each time two opposite quarters are rejected and removed from the floor, and the remainder broken smaller before proceeding to quarter again. Care must be taken to sweep away all dust belonging to the rejected portions before mixing the other quarters. Sizes to which Samples must be Crushed. The degree of crush- ing necessary at each stage will depend on the nature of the material, and must be varied in special cases. The following figures are given by A. Harvey l as representing the sizes found to be safely allowable in actual work with a Vezin automatic sampling machine, cutting out J of the entire mass put through at each operation. (This machine is so constructed that it deflects the entire Stream falling through a chute at regular intervals of time ; say every two or three seconds.) (See Fig. 41.) TABLE I. ALLOWABLE SIZES OF ORE PIECES IN SAMPLING Diameter of Largest Pieces in Sample: Inches Minimum Weight of Sample: Pounds Diameter of Largest Pieces in Sample: Inches Minimum Weight of Sample: Pounds 5J 79,300 f 256 4 69,109 i 32 3i 44,958 i 4 2i 16,384 A | H 2,048 * A Min. and Sci. Press," LXXXVIII, p. 78 (Jan. 30, 1904). 320 THE CYANIDE HANDBOOK W. Glenn * gives the following rules, applicable to such ma- terial as 10 per cent, copper ore, originally in a 10-ton pile con- sisting of large lumps; 1 ton, obtained as a sample by "trenching" as above described, is quartered to 500 lb., breaking the larger lumps with a hammer. This is treated as follows: Diameter of Weight o' largest pieces: sample: Inches Pounds 1 500 250 "fine gravel" 125 "coarse sand" 60 Proceed to mix and quarter on a sheet of paper. "At this point the sample would weigh about 15 lb.; its larger grains would be in size like coarse sand. It would be safe now, without further breaking, to mix and quarter it twice, or until its weight did not exceed 4 lb. Run this through the mortar and then mix and quarter it twice, or down to 1 lb. weight. Grind this to something approaching powder, and for the last time mix and quarter it." Relation of Size of Sample to Grade of Ore. D. W. Brunton 2 discusses the whole subject of mixing and quartering from a mathematical standpoint, and shows that the size to which ore must be crushed, so that the error in sampling may be within allowable limits, depends: (1) On the weight or bulk of sample (the smaller the sample the finer it must be crushed) ; (2) on the relative proportion between the value of the richest minerals and the average value of the ore (with high-grade ores we may crush more coarsely than with low-grade, for the same percentage of error); (3) on the specific gravity of the richest mineral present. The. higher the specific gravity of the richest mineral, the greater the value of a particle of given size and grade, and hence the greater the influence of such particle on the sample. Finer crush- ing is therefore required when the rich material is also of high specific gravity. Formula for Dividing Samples. S. A. Reed 3 gives the fol- lowing formula for determining the maximum diameter allowable for the particles in any stage of the sampling of a given ore: i Trans. A. I. M. E., XX, p. 155 (June, 1891). a Trans. A. I. M. E., XXV, p. 826 (October, 1895). 3" Sch. Mines Quart.," VI, p. 351 (1885). SAMPLING 321 D -= diameter of largest pieces in inches; p = quantity of the lot in Troy ounces; / = number of parts into which p is to be divided (one part to be chosen as sample) ; k = percentage of metal sought (gold or silver) in the richest specimens of the lot: s = sp. gr. of richest minerals; m = average grade of ore (ounces per ton); a number of pieces of D size and k value that can be in excess or deficit in the portion chosen as sample; I = largest allowable percentage of error. Before cutting to -, we must crush or grind the lot so that D = .053 m p I k (f - 1) a The following table is given for ores of different grades, assum- ing in all cases that s = 7, I = 1, and that with samples: Class A, m = 50 k = 1 Medium, tt B, m = 75 k = 10 high-grade, n c, m = 500 k = 30 very rich. TABLE II VALUE OF D, IN INCHES aampie JK.eauceu irom Class A Class B Class C 100 to 10 tons .... 5.28 2 96 2 58 10 to 1 ton - 2 46 1 38 i o 2000 to 200 Ibs 1.14 06 56 200 to 5 " 03 18 01 A 5 Ibs. to 10 assay tons. . . . 0.034 0.02 0.018 These figures correspond roughly with those given by Harvey. (See above.) Sampling Ore in Mine. As the assayer may occasionally be called upon to take samples of ore from the mine itself, the following description of the method employed on the Rand is quoted from a paper by G. A. Denny " Observations on Sam- pling, Computation of Assay Averages," etc. 1 "The sampling of the mine is entrusted to men of experience who take as their equipment a water squirt, with which to wash i Tr^ns. A. I. M. E., XIX, p. 294 (June 15, 1900). 322 THE CYANIDE HANDBOOK down the rock at the point to be sampled, hammer, chisels (both diamond and chisel-pointed), a receptacle in which to catch the samplings, and sacks about 15 in. deep and 9 in. wide wherein to place the completed sample. . . . The sampler first proceeds to measure equal distances along the level, rise, or other work- ing which he is to sample. The -distances taken are either 5-ft. or 10-ft. intervals; in the case of thin rich leaders, always 5 ft. Having chalked or otherwise marked the intervals, he proceeds to the work of sampling. In cases where the reef is thin, say from 1 in. to 3 in. in width, the sample is taken over a width of 6 in., that width being carefully channeled out to a certain breadth and depth, so that as nearly as possible a true average may be obtained. Samples should be taken of approximately a minimum weight of 5 to 6 lb.; in general the bigger the sample the nearer it is to the true average." When the reef is "solid" (i.e., contains no waste), and less than 18 in. wide, it is sampled as one section; if of more than 18 in. width, it is sampled in so many sections of 18 in., each section being separately sacked and recorded. This is done to avoid getting undue quantities of ore from one section, as may happen when the hanging and foot-wall sides are of different character. " In cases where the reef is divided into bands of ore and quartzite, only the ore is sampled, the bands of waste being measured and calculated on the total sampled width in the following manner": TABLE III Band Width in Inches Nature Assay Value: Dwts. Assay Inches A 6 Reef 15 90 B c 12 10 Waste Reef 20 200 D E 30 2 Waste Reef 60 120 60 6.83 410 Where samples have been taken at regular intervals, the aver- age assay is found by dividing the total assay-inches by the sum of the widths in inches of all the samples taken. Where samples have been taken at irregular intervals, the length of reef represented by each sample also enters into the calculation. SAMPLING 323 It may be added that all mine samples should be taken as nearly as possible in a plane at right angles both to the dip and strike of the vein or deposit to be sampled; otherwise an erroneous impression will be given of the quantity of ore represented. Machinery for Crushing Samples. The work of breaking up the larger lumps, before mixing and dividing ore samples, is carried out either by hand, with a large hammer and anvil, or by machines, of which the two principal types are represented by the Blake and Gates crushers respectively. The finer crushing is performed by smaller machines of the same types, or by the pestle and mortar, or by machines imitating the action of the latter. Crushers of the Ball mill type are unsuitable for sampling work on account of the labor and difficulty of cleaning them effectively. The finest crushing is best done on a flat steel or cast-iron plate with raised edges on three sides (known as a "buckboard"), the ore being ground on this by means of a heavy iron block ("muller"), worked backwards and forwards by means of a wooden handle. Precautions in Preparing Samples. Attention may here be drawn to certain points of the utmost importance, neglect of which is probably the chief calise of inaccuracy in the preparation of ore samples for assay: (a) When any portion has been selected as a sample and is to be passed through a sieve of particular mesh, the entire por- tion so selected must be made to pass the sieve. If the harder portions, which naturally resist crushing the longest, be rejected, the sample is worthless. (6) After the operations of crushing, sifting, etc., have been completed on one sample, every appliance used in the process should be carefully cleaned before proceeding with another sample. (c) After crushing and sifting, and before dividing, every sample must be thoroughly mixed. The operation of sifting in itself causes a certain amount of separation in the different in- gredients, so that the sifted sample is never homogeneous. When, as is frequently the case in mine or cyanide works' assay offices, the preparation of the sample is entrusted to natives not under close supervision, one or all of these precautions will probably be shirked. Cleaning Sampling Implements. Great care is necessary in 324 THE CYANIDE HANDBOOK order to prevent the mixing of one sample with material from others, or with foreign impurities. All vessels and instruments used should be so constructed that they may be easily cleaned. The absence of this condition is a fatal drawback to various other- wise effective devices which have been designed for mechanically crushing, mixing, and dividing samples. Sieves for Samples. The sieves can generally be cleaned effectively with a hard brush; but sometimes it may be necessary to pass through the crushers, sieves, etc., a quantity of barren rock, sand, charcoal, or other material for the purpose of remov- ing adhering particles of a previous sample. The sieves used for this purpose are of brass-wire screening, set in a metal frame; as an additional precaution they may be soldered all round on both sides at the joint of the screen with the frame. All sheets of iron, cloth, and rubber, and all sampling floors and tables, should be carefully swept and dusted after each sample is finished; they should also be washed occasionally. Necessity for Separate Sampling Room. Whenever possible, the room in which the samples are crushed and sifted should be entirely distinct from the assay room. These preliminary opera- tions inevitably cause a certain amount of dust which should be carefully kept away from the supplies of flux, and from the cruci- bles and other utensils used in the assay itself. Importance of Sampling. The importance of correct sampling can hardly be overestimated; it is obvious that the subsequent assay is valueless, however carefully conducted, unless the sample assayed really represents the average composition of the material to be examined. (B) SAMPLING OF SAND AND SIMILAR MATERIAL In cyanide plants we have to deal, as a rule, with material which has already been more or less finely crushed, and in most cases with ore from which the coarser metallic particles have already been removed by amalgamation. Difficulties of Sand Sampling. To obtain a correct average sample, representing, say, the material delivered from the mill to the cyanide tanks during a run of 24 hours, or the charge run into some particular tank, is by no means so simple a matter as might appear at first sight. The crushed ore is usually far from uniform in composition, and the methods adopted for filling the SAMPLING 325 tanks tend to produce layers or rings of material differing in com- position and value from the remainder. Points where Samples may be Taken. Samples for regulating the work in a sands plant may be taken: (1) Before entering the tanks; (2) in place, after the tanks are filled; (3) while the tanks are being discharged. Where the sand is collected in one tank and treated in another, the sampling is often done during the transfer from the collecting to the treatment tank. Samples Taken before Charging Tanks. It is customary in many reduction works to take two samples daily in the battery, one representing the material passing through the screens of the mortar-boxes immediately before it comes in contact with the amalgamated plates, and the other after it leaves the plates and enters the launder leading to the hydraulic separators. These samples are taken by placing a metal scoop or other suitable vessel across the whole width of the discharge, and intercepting a portion of it at regular intervals (say every hour). The por- tions so collected are transferred to a bucket, and at the end of the day's run the combined sample is drained, mixed, quartered down, and assayed. Automatic Samplers. An automatic sampler is sometimes used on the launder leading to the separators. These samplers are of various designs, the general principle being, however, that the entire stream of pulp is diverted into a special vessel for a few seconds at regular intervals of time. D. Simpson 1 states that at the New Goch G. M. Company, the method of catching the outflow from the delivery hose at regular intervals, and also that of the automatic sampler, were discarded at an early date as unre- liable. In general, it may be remarked that when intercepting a stream by means of a scoop it is extremely difficult to avoid loss by splashing, so that the sample taken in this way is liable to contain an undue proportion of the coarser and heavier par- ticles. Truck Samples. In cases where comparatively dry sands, such as accumulated tailings, are to be treated, and these are de- livered into the tanks in trucks (cars), it is usually found that a satisfactory sample of the charge can be obtained by taking small equal portions from each truckload, provided the number of loads required to fill the tank is fairly large. i " Sand Sampling in Cyanide Works," I. M. M., Bull. No. 25, Oct. 11, 1906, p. 6. 326 THE CYANIDE HANDBOOK Sand Samples taken in Place by Sampling-Rod. After the tanks have been filled it is customary to take a sample of the con- tents by means of a sampling-rod. This consists of a metal tube, long enough to penetrate the whole vertical depth of the material to be sampled say from 4 to 10 ft. length and 1J to 2 in. diameter. This tube is slit from the bottom to within a few inches of the upper end and is provided with a crosspiece at right angles, serving as a handle. The edges of the groove and the lower end of the instrument are sharpened to allow it to cut into the mass of sand. When the rod is pushed into the charge with a slight twisting motion, the hollow fills with sand. The rod is then withdrawn and the surface scraped smooth. The core of sand is then cut out and set aside to form part of the sample, using a suitably shaped piece of hard wood or metal for cleaning the groove. A number of cores are taken in this way from various parts of the tank, and the whole mixed and quartered. The same instru- ment is commonly used for sampling tailings in dumps or pits, or for any similar accumulations of moist sandy material. Objections to Rod-Sampling. It is a frequent complaint that the sampling-rod does not really take out equal quantities of sand at different depths in the tank. In passing through the upper layers, the lower part of the tube becomes choked, especially if the material be somewhat slimy, so that the lower layers are not represented at all, although the rod has been pushed to the bottom of the tank. This difficulty may be partially obviated by taking samples to half the depth of the tank, cleaning the rod, and .then pushing it down to the bottom of the tank through the hole already made, the second portion taken thus representing only the bottom layers. As it frequently happens that the bottom layer (6 in. or 12 in.) is of very different composition from the remainder, it cannot be considered that this method may be safely relied on to give accurate results, unless special precautions are taken. Location of Samples in Tank. The number of rod-samples taken per charge, and the points at which they are taken, are also matters of considerable importance. If the charge were abso- lutely uniform, a single sample taken at any point would correctly represent its average composition. This, however, is seldom, if ever, the case. When the charge is not uniform, a better result will be obtained by taking samples in such places that each one SAMPLING 327 represents an equal mass of sand, rather than by haphazard sampling in any part of the charge. The tanks used at the present day are practically always cylindrical in form, with a flat bottom at right angles to the sides. Hence, by describing on the upper surface a number of circles of varying diameters, the charge may be divided into any required number of sections, each representing an equal, mass of sand. Thus, a circle 40 feet in diameter may be divided into 4 equal areas by describing from its center circles having radii of 10 ft., 14.14 ft., and 17.32 ft. The general formula is as follows: Let n = number of equal areas required; T\, r-2, rs, = radii of circles described from center of upper surface; r n = radius of the tank itself; Then a = area of each section. Then, a = HT, = ir(r 2 2 - n 2 ) = 7r(r 3 2 - r* 2 ) = ir(r 2 -i - r 2 w ). Hence, n = \/ , r 2 = \/2r,, r 3 = v/3n, r = \/n n, a = ir^L. ir n These areas may again be subdivided by diameters of the 40-ft. circle inclined at equal angles. Thus, by taking 2 diameters at right angles we obtain 16 equal areas, and by taking a rod- sample as nearly as possible in the center of each we should obtain a tolerably good average of the whole charge. Causes of Irregularity in Distribution of Sand. From the above it will be evident that the method sometimes adopted of taking samples at equal distances along a diameter gives an undue proportion of material from the central parts of the tank, where the coarser and heavier particles generally tend to collect. The lumps of slime generally accumulate at the bottom and sides of the tank. When material is transferred from an upper to a lower tank, the lumps of slime and any lighter or finer material tend to roll down the sides of the cones formed below each dis- charge door, and form layers and walls on the filter-mat at the junctions of the cones with one another and the sides of the tank. Section Sampling. Where the conditions permit, according to Simpson, 1 an accurate sample may be obtained by discharging sand from a tank, leaving a vertical face in the plane of a diameter of the tank. This is divided into a number of horizontal zones of equal width, and samples taken by scraping the face of each d., p. 3. 328 THE CYANIDE HANDBOOK zone into a separate bucket. These samples are then assayed separately, or a combined sample is made by mixing equal quan- tities of each. Samples taken while Discharging Tanks. When the tanks are discharged by means of trucks, a good sample of the residues may generally be obtained by taking with a scoop or, better, with a sampling-rod an approximately equal portion from each truckload, the sample thus obtained being afterward mixed and divided. In cases where the material is entirely composed of comparatively small grains and is fairly uniform, only a small quantity need be taken from each truck. Where the material contains large lumps, as may be the case with slimy tailings, it is necessary to take larger amounts. All such lumps included in the sample must be thoroughly broken up before quartering. When the tanks are discharged by a belt-conveyor, or similar mechanical device, samples may be taken, at equal intervals of time, as the material passes a certain point; or an equal portion may be taken once or twice during each revolution of the belt. Precautions as to Residue Samples. When residue samples have been taken after cyanide treatment, it is essential to place them in a vessel which will not absorb liquid, and which is not affected by cyanide in any way. An ordinary metal bucket is not suitable for this purpose, but an enameled iron bucket in good condition may generally be used. The samples should be kept covered as much as possible, especially if a moisture determination is to be made. Mixing, Dividing, and Quartering Samples. For assay pur- poses, all samples taken as above described require to be reduced in bulk. The general principles of this process are the same as those underlying the reduction of ore samples, already discussed, the essential point being that the average composition of the sample must not be altered. In all cases a thorough mixing is necessary. The sample is spread out on a smooth surface of suitable material, and sufficiently large to allow of the sample being turned over freely with a shovel. When the material is not very wet, a cement floor answers the purpose. All lumps must be carefully broken up, and in many cases it is very desir- able to put the whole sample through a coarse sieve, after which it is reduced by coning and quartering. (See above.) It has already been remarked that any lumps of slimes and such-like SAMPLING 329 material tend to roll to the bottom of the cone, so 'that it is as well not to rely exclusively on this process for mixing the sample. When the mixing is considered to be sufficient, the conical or hemispherical heap is divided crosswise into four quarters. Two opposite quarters are dug away and rejected. The remaining quarters are mixed together, any lumps further broken up, and again quartered, this operation being repeated until the required amount is obtained for the determinations needed. After the weight of the sample has been sufficiently reduced, the mixing may conveniently be done by rolling on a sheet of oil-cloth or rubber, taking care to alter the direction of rolling as often as possible, so as to avoid the formation of bands of concentrates. In the case of ordinary sand samples a very satisfactory mixture may be obtained by rolling, provided the surface be sufficient and the operation be repeated several times. Samples containing a moderate amount of moisture and fairly uniform in composition may be divided by spreading out in a uniform layer a few inches thick, and selecting with a horn scoop, spatula, or other instrument, approximately equal portions at regular intervals all over the layer. It is a good plan to rule two sets of parallel lines at right angles, so as to divide the layer into a considerable number of small equal squares, and to take equal small portions from each square. Before doing this, how- ever, the sample should have been well mixed by coning, rolling, or otherwise. Moisture Samples. It is frequently necessary to make a determination of the amount of moisture in charge or residue samples in order to calculate the dry weight of material under treatment. Thus, the dry contents of a tank are known if we know the average weight of the truckloads of wet sand required to fill it, the number of such loads, and the percentage of moisture. All such "moisture samples" are taken with as little delay as possible. Probably 'the most satisfactory method is to spread out a large quantity of the roughly mixed sample, take out por- tions at regular intervals in the way described above, and weigh on a good balance 500 or 1000 grams of the sample. This is placed in a porcelain or enameled-iron dish and dried at the same temperature that is used for drying the ordinary assay samples. When cool, the dried sample is again weighed, the difference indicating the amount of moisture. In the case of 330 THE CYANIDE HANDBOOK very wet material (slimes and similar samples) it is best to weigh dish and contents together, and' determine the weight of the empty dish after cleaning and drying. Drying Assay Samples. After the sample has been reduced by successive quarterings or otherwise to about 5 Ibs. (or, say, 2 kg.), it may conveniently be dried by heating in a suitable vessel over a moderately hot fire. In this operation the nature of the material to be treated must be carefully considered. Where pyrites is present, the heat must not be sufficient to cause any roasting. Drying at a temperature of 100 or 110C. would in general be far too slow for ordinary assay work, though some such standard is usually adopted for samples required for analysis. An effective arrangement for some time used by the writer con- sisted of a wide chamber about a foot high, with wooden and can- vas sides, the whole being heated above and below by means of a coiled steam-pipe, maintaining a temperature of about 90 C. About a dozen ordinary samples, containing, say, 5 per cent, moisture, could be dried if left over night in this apparatus. Samples dried in this and similar ways probably always retain a small percentage of hygroscopic moisture; this, however, does not seriously affect the assay, nor does it affect the accuracy of any calculations of the values per ton of material, provided the moisture sample has been dried at the same temperature. Drying samples over an open fire in iron pans, as is frequently done, is not to be recommended. Unless constantly watched, there is danger of the sample being overheated; also, in removing the sample from the pan, pieces of metal or rust from the latter are apt to scale off. Samples containing dissolved gold and silver, such as the residues from cyanide-treated tailings, should on no account be dried in unprotected metal vessels. Porcelain or enameled-iron vessels are the most suitable, but when these are used the temperature must be carefully regulated. Reduction of Dried Sample. When quite dry, the entire sample must be passed through a sieve of suitable mesh, depend- ing on the nature of the material. For this first sifting a screen of 4 holes to the linear inch may be used, to break up any lumps of slime, etc. The sample is then turned out on an oilcloth or rubber sheet, and thoroughly mixed by rolling. All lumps not passing the sieve are ground up in a mortar or on the buckboard. The softer lumps may be broken by simply rubbing on the sieve SAMPLING 331 with a smooth block of wood or cork. The use of metal weights and similar objects for this purpose is objectionable, as they tend to wear the wires and enlarge the holes of the sieve. It is essential that the whole sample be passed through, as the portions remaining on the sieve are almost invariably of different average composition from the remainder. After sift- ing and mixing, the sample is quartered, either by hand or by some mechanical device. Many different types of apparatus have been brought out for this purpose, mostly based on the principle of cutting out a given fraction (say J or ) from a fall- ing stream of sand. The fraction retained as the sample is then passed through a finer sieve, all particles which do not at once pass the sieve being crushed to the requisite degree of fineness. The sample is then again mixed and further divided. This process is repeated, using successively finer sieves, until the assay sample is obtained. This should be crushed at least to 60-mesh; in some cases it is desirable to use a much finer mesh, say 120 or 200. The finer crushing is generally done on a buckboard, as described above. Size of Grains at Different Stages. The maximum size of lumps allowable at any stage of the process depends on the weight of the sample at that stage. This subject is discussed above in reference to rock samples, and some examples are given from different writers, but no absolute rule can be given. The point should be determined, in the case of any particular class of ma- terial, by making separate assays of each pair of opposite quarters in one or more lots of the material. The following, frequently used by the writer, may be taken as an example in dealing with ordinary sand charges and residues: Entire sample passed (wet) through J-inch screen. From this 2000 grams are roughly weighed out after mixing, dried, crushed to J in., and halved. 1000 grams crushed to 10-mesh and halved 500 " " 30- " " 250 " " 60- " " 125 " " 80- " for sample. The last step is taken only in the case of fairly rich material. The final sample allows of 4 assays of 1 assay ton each; or if the division is stopped at 250 grams of 4 assays of 2 assay tons each. 332 THE CYANIDE HANDBOOK Weight of the Final Sample. Sufficient material should be retained to allow of the assay being repeated three or four times if necessary. Generally speaking, the poorer the sample, the larger must be the quantity taken for each assay. The assay charge commonly taken varies from 30 to 60 grams, but in special cases it may be as low as 15 grams or as high as 250 grams; the assay sample may therefore vary from 60 to 1000 grams. Sampling Slimes. L. Ehrmann * gives the following method for sampling "battery slimes." The slimes are put through a small filter-press after the pulp has been thoroughly mixed; then two, three, or more cakes are made, according to the volume of the original sample, the cakes weighing about 1J kg. wet or 1 kg. dry. Now from each fresh cake two opposite triangular parts are cut and the portions so selected are made into a pulp with water and again put through the press, until only one cake is obtained. This cake will be an average of the lot, and will weigh, say, 35 assay tons. From this, triangular opposite parts are cut so as to get 16 to 17 assay tons. This may then be dried for the assay sample. 'Samples Containing Dissolved Values. When samples con- tain soluble gold and silver it is best to make a thorough extrac- tion of the soluble matter, and assay the extract and the washed residue separately, then calculating the results on the weight of the original sample. It has been supposed that losses of precious metal take place in drying finely divided samples containing cyanide solution, but the evidence on this point is not very con- clusive. 2 1 " Proc. Chem., Met. and Min. Soc. of South Africa," II, p. 697. ., II, 372. SECTION II ASSAY OF A TYPICAL SILICEOUS ORE (A) CRUCIBLE FIRE ASSAY Object of Assaying. The process of assaying aims at extract- ing the whole of the gold and silver in a pure and weighable form from a weighed portion of the ore or other material to be assayed, which portion is assumed to be of the same average composition as the entire mass which it represents. Quantity of Ore to be Used for Assay. The amount to be weighed out for the fusion will be determined by the probable richness of the sample. Where results have to be reported in ounces or pennyweights per ton, it is convenient to use what is known as the " assay ton " system. An assay ton is a weight so chosen that it contains as many milligrams as there are ounces (troy) in a ton. For the ton of 2000 Ib. this weight is 29J grams (29166 mg.) For the ton of 2240 Ib. the assay ton is 32f grams (32666f mg.). If this weight be taken for an assay, every milligram of metal found will represent 1 oz. (or 20 dwt.) per ton of the material assayed; if 2 assay tons be taken for an assay, every milligram found will represent J oz. (or 10 dwt.); or every 0.1 mg. will represent 1 dwt. per ton of the material assayed. The abbreviations " 1 a. t.," "2 a. t.," etc., are commonly used to represent "one assay ton," "two assay ton," etc. When not otherwise stated, the "ton" referred to in this work is the ton of 2000 Ibs. Assays are often made, however, on an even number of grams (say 30, 60, or 100) ; in this case, if the result is to be reported on the " ounce per ton " or " pennyweight per ton " system, a table for converting the results must be used. Where the metric system is used throughout, 1 mg. per kilo- gram represents 1 gram per ton of 1000 kilograms. In this case a convenient quantity for assay is 25 or 50 grams: 1 mg. on 25 grams = 40 grams per ton. 1 mg. on 50 " = 20 " per ton. 333 334 THE CYANIDE HANDBOOK Fluxes. The " crucible " or " fusion " assay of an ore is carried out in several stages, the first of which consists in forming an alloy of lead with the precious metals; other substances are added at the same time to form fusible compounds with the re- maining constituents of the ore, so that every particle of precious metal may readily sink through the mass and alloy with the lead. In order to effectively secure every particle of gold and silver, it is found necessary to use a mixture which will generate metallic lead in a state of very fine division ; for this purpose an oxide of lead mixed with some reducing agent is necessary; when this is sufficiently heated, the latter acts upon the oxide, producing minute globules of metallic lead in all parts of the mass; these gradually sink, at the same time combining with and carrying down with them any particles of gold and silver which they encounter. The remaining constituents of the ore fall into one of two groups, generally described as "acid" or "basic." In order to effectively separate these from the metallic particles, substances must be added to form a fusible compound with each of these classes. Acid substances will require a basic flux, and basic sub- stances an acid flux. We require, therefore: 1. An oxide of lead (litharge, red lead); 2. A reducing agent (charcoal, flour, argol) ; 3. A basic flux (carbonate of soda, bicarbonate of soda, car- bonate of potash); 4. An acid flux (borax, silica). Other substances, such as niter, fluor-spar, sulphur, metallic iron, etc., are occasionally added for special purposes (see below). Proportions of Ingredients in Flux. The nature of the flux necessary and the proportions and total quantity of each ingredi- ent should be ascertained by experiment for each particular class of sample which has to be dealt with. The information given on this subject in different text-books is contradictory and often misleading, but by bearing in mind the object of the different classes of fluxing materials, the proper proportions may generally be discovered after a few trials, and where many samples of similar material have to be assayed, a large stock of flux may be pre- pared for use. Standard Flux for Siliceous Ores. The following may be taken as applicable to a large number of samples of what may fairly ASSAY OF A TYPICAL SILICEOUS ORE 335 be called siliceous or quartzose ores. They were adopted by the writer after an extensive series of experiments on an ore consist- ing mainly of silica, but containing a fair amount of oxide of iron, and smaller quantities of the silicates of aluminium, magnesium, etc.: Charge of ore 1 assay ton 2 assay tons Carbonate of soda 32 grams 57 grams 1 Fused borax 5 " 8 " Litharge 57 " 120 " Charcoal 1.1 " 2 " Size of crucible required G H Weight of lead produced 25 grams 40 grams Quality of Fluxes. When large quantities of flux are pre- pared, the ingredients should be carefully mixed, all lumps broken up, and the whole sifted through a coarse screen kept for this special purpose. The flux should be preserved in a suitable box or tin, kept in a dry place, and not exposed to dust from ore samples, etc. The proportions of borax and soda to be used will vary con- siderably with the quality of these reagents, which sometimes contain a considerable percentage of moisture and other ingredi- ents. It is not by any means unusual to find commercial " dry " borax with over 30 per cent, of moisture. If used in the flux, about double the quantities given above must be taken, but such material should be avoided as it swells enormously on heating, and causes violent bubbling in the crucible during the early stages of the fusion. There seems to be no object in using bicarbonate of soda in place of the carbonate; the Na 2 O being the effective constituent, the use of bicarbonate merely implies loss of fuel and time in expelling the extra quantity of carbonic acid. Proportions of Litharge and Reducer. The litharge and char- coal should be so regulated that the button of lead obtained at the end of the fusion is of convenient size for cupellation, and may be reasonably supposed to have taken up practically the whole of the values in the material fused. Some assayers insist on reducing the whole of the lead from the litharge used in the flux; others are equally emphatic as to the necessity of using a large excess of litharge; in the writer's experience it is best to leave a small excess of litharge in the slag. The weight of lead button found to be most generally suitable is about 18 to 23 grams for an assay on 1 assay ton of ore, and 25 to 30 grams for 2 assay tons of ore. 336 THE CYANIDE HANDBOOK In some cases, better results are obtained by increasing both litharge and reducing agent so as to bring down a large button of lead, which is afterwards reduced, as described below, to a con- venient size for cupellation. The special modifications of the process for different classes of ores containing base metals, etc., will be discussed in a later section. Crucibles. The crucibles generally used are fire-clay pots of different shapes and sizes. The kinds marked "G" and "H" are those most generally useful for assays on 1 assay ton and 2 assay tons respectively. Previous to use, they should be stored in a warm dry place and carefully cleaned before charging. Covers of the same material are used, which serve to prevent the entrance of coal or other foreign matter into the crucible during the fusion. The same crucible may, with care, be used for five or six succes- sive fusions, but must be carefully examined for cracks before recharging. Weighing the Assay Charges. It is advisable to have the portions of flux needed for each assay ready before weighing out the ore charges. The flux need not be weighed with great accu- racy; in fact, for all practical purposes the necessary amount may be measured out by means of a small metal cup or other con- venient instrument which allows of approximately equal quanti- ties being selected. The ore is best spread out on a sheet of smooth rubber in front of the pulp balance or ore balance used for weighing. It is as- sumed that the sample has already been thoroughly mixed, but as an additional precaution the assay charge may be chosen by removing small amounts with a spatula at regular intervals all over the layer of ore. The charge is placed on the left-hand pan of the balance and the assay-ton weight, or other weight required, on the right. The weight for assay may then be adjusted by adding or taking off small portions of ore until it is correct to about 0.1 gram. In an ordinary assay it is a waste of time to adjust more closely than this, as it is obvious that a difference of 0.1 gram on a charge of, say, 30 grams would introduce an error of only y 1 ^ dwt. in an assay of a 20-dwt. ore; whereas a balance sensitive to T ^ mg. can only detect a difference equivalent to oz., or $ dwt., in an assay of 20 dwt. Mixing with Flux. The weighed portions of an ore are ASSAY OF A TYPICAL SILICEOUS ORE 337 commonly mixed with the flux on a sheet of glazed paper, using a spatula to obtain a uniform mixture. The charge may then be readily transferred, to a clay crucible of suitable size, and the paper swept clean by a small flat brush. Another, perhaps preferable, method is to mix flux and ore in a smooth porcelain mortar, large enough to enable the mixture to be stirred freely. In many assay offices the ore and flux, after mixing, are wrapped in thin tissue paper (Japanese copying-paper answers the purpose), arid the little bundle placed in a crucible which has been previously heated. The fusion in this way is probably more rapid than when the charge is added to a cold crucible; but there are objections to the method, as will be shown later. The Fusion Process. The assay fusions are generally carried out in small square furnaces (8 in. to 12 in. square, inside measure- ment), capable of holding from 2 to 6 crucibles at a time, and provided with horizontal or inclined sliding covers. The fur- naces and such parts of the flue as are exposed to strong heat are lined with fire-brick. The crucibles should be firmly supported on fire-bricks resting on the bars at the bottom, but arranged in such a way that they may be uniformly heated all round; and the mouths of the crucibles should not be at a higher level than the opening of the flue. When charcoal is used, the crucibles, after putting on the fire-clay lids, may be entirely covered by a layer of fuel. With coke it is hardly necessary to use a support beneath the crucibles. The bars of the furnace must be clear of ash before setting the crucibles in their places. The operation of fusion usually takes afrout 40 minutes. For the first 10 minutes or so the temperature should be kept down to avoid loss through the violent expulsion of carbonic acid, water vapor, etc. At this stage, signs of fusion generally begin to appear at the sides, and the lids had better be removed, provided the fuel is burning quietly and there is no danger of lumps of charcoal, etc., being projected into the pots. A semi-fused crust sometimes forms on the surface, which may occasionally be raised nearly to the mouth of the crucible by the action of gases beneath, but if the proper conditions be observed this crust soon sinks down and no projection of the contents from the crucible takes place. As soon as the action begins to moderate, the temperature may be raised (by opening the damper more fully and closing the covers of the furnace) , and the heat is continued until all signs 338 THE CYANIDE HANDBOOK of ebullition have ceased. The lids are then replaced for the last five to ten minutes, after which the contents should be ready to pour. A fairly bright red heat is needed for the fusion, but very intense heat is neither necessary nor desirable. Fusions in the Muffle. The fusions are sometimes made by placing the crucibles in a large muffle, and with care good results can be obtained; but the method has the disadvantage that the actions going on cannot be easily observed and controlled; also, with poor ores, either a very large muffle must be used to accom- modate the large crucibles which are necessary, or a number of assays on each sample must be combined, which is laborious and troublesome. Pouring. When ready for pouring, the contents of the cru- cible should be quite tranquil; the crucibles are then lifted out of the furnace by means of iron tongs, bent at the points so as to grip the edge of the crucible firmly, and the molten contents poured into iron molds, previously cleaned and slightly warmed. The fused mass should pour freely, like oil, and should leave no metal or lumps of any kind in the crucible, which, when cool, should show internally only a thin layer of slag. Hammering Lead Buttons. The molds are left about five minutes to cool, then inverted, and the slag removed from the buttons of lead by one or two blows with a hammer. No shots or detached particles of lead should appear in the slag, but in all cases it should be carefully examined, and any detached lead added to the main button. The color and appearance of the slag will depend largely on the nature of the ore; also to some extent on the time and temperature of fusion. When the ore consists chiefly of silica, with small quantites of oxide of iron, as in the case we are considering, the color will be light green; with large amounts of iron it will be dark brown, almost black. When copper is present a characteristic reddish-brown color is produced (due to cuprous oxide). The lead should be easily detached from the slag, and should not leave a scale or film behind on breaking off; this takes place sometimes when the slag is exces- sively hard and stony, as may be the case if too much borax or acid flux has been used. Often the slag is very brittle and flies to pieces with some violence on merely touching with a hammer. The hammering should be done on a steel block, mounted on a firm wooden support and surrounded on three sides by a frame ASSAY OF A TYPICAL SILICEOUS ORE 339 to retain flying particles. Hammer and anvil faces should be kept carefully cleaned. The detached lead buttons are held in steel forceps and struck on the edges before hammering the upper and lower surfaces, otherwise particles of slag may be hammered into the block of lead, causing spitting and loss in cupellation. By hammering at the edges and brushing the lower surface (where the button was in contact with the slag) with a wire brush, the lead may be effectually cleaned. The block of lead is then beaten into a roughly cubical shape, and placed in a compart- ment of a cupel tray. Arrangement of Assays. In working a batch of assays it is very necessary to arrange them systematically and to adhere to the same order of arrangement at every stage of the process. Each sample is first marked with a number, and the same number entered in the assay book, with sufficient description for identi- fying the sample with certainty. In placing the crucibles in the fusion furnace, the lead buttons on the cupel trays, or in the muffle for cupellation, etc., these numbers are taken in the order of the words on a page of a book. Thus, 12 assays in 2 fusion furnaces might be arranged: Furnace No. 1 1 2 3 456 Furnace No. 2 789 10 11 12 When subsequently transferred to the cupel tray they might be placed thus: 1234 5678 9 10 11 12 If this principle be strictly adhered to, no confusion or am- biguity can occur, even if the crucibles, cupels, etc., be not actually marked. It is a good plan to arrange the samples so that the lower- grade assays (e.g., cyanide residues) are made first. The same crucibles may then be used for the richer samples without danger of " salting " should any lead remain in the crucible from a previous imperfect fusion, as this could hardly contain a weigh- able quantity of gold. Fusion in Hot Crucibles. The method of adding fresh charges 340 THE CYANIDE HANDBOOK immediately to the hot crucibles, whether these charges are wrapped in paper or not, is not to be recommended, as it results in very violent and rapid action in the early stages, giving rise very probably to loss of material through ejection of particles from the crucible. It is also possible that the lead particles will form and sink to the bottom of the crucible before the siliceous matter is properly fused, and that particles of precious metal will therefore remain in suspension in the slag without ever com- ing in contact with lead. When the temperature is kept low in the early stages the action is more moderate; the lead formed sinks gradually through the mass and has a better chance of secur- ing all the particles of precious metal. (B) CUPELLATION AND PARTING Cupellation. This process depends upon the fact that when an alloy of gold and silver with lead is heated, with free access of air, in contact with certain porous materials, such as finely pow- dered bone-ash, the lead oxidizes, this oxide being partly volatil- ized and partly absorbed by the substance on which it rests; while the gold and silver undergo little or no oxidation, but, after all the lead has disappeared, remain behind in the form of a nearly spherical bead of fine metal. Muffle- Furnaces. The operation is generally carried out in a small fire-clay oven, called a " muffle," consisting of an oblong floor surmounted by an arched roof and closed at one end. This is supported within an outer furnace ("muffle-furnace"), so arranged that fuel may be placed below, above, and on both sides of the muffle. The latter is at least partially open in front, and the semicircular end is usually provided with a slit to secure a draft. Slits are also sometimes made in the sides of the arch. Coal, charcoal, gas, or oil are used as fuel. Cupels. The " cupels " in general use are small cylinders of porous material with a hollow at one end for the reception of the lead button. Formerly, they were always made of bone- ash, but at the present day various refractory materials, such as magnesia and magnesium borate, have been applied with some success in the manufacture of cupels. The efficiency of bone-ash cupels depends largely on the amount of water added, the method of molding and pressing, the fineness and purity of the bone-ash, and the time of drying the cupels ASSAY OF A TYPICAL SILICEOUS ORE 341 before use. Some assayers add small quantities of alkaline car- bonates and other ingredients, but the general opinion seems to be that with pure bone-ash the best results are obtained by moderately fine crushing, using only distilled water. The amount of water should be only just sufficient to cause the material to cohere, and not sufficient to make it pasty. The cupels are then made by means of a mold consisting of a hollow metal cylinder, which is filled with the moist bone-ash; the plunger is then pushed in and struck once or twice with a mallet. The mold is then in- verted and the cupel turned out. Better results can in general be obtained by the use of manufactured cupels, which are much more uniform in quality than can be the case with those made by hand on the spot. Previous to use, the cupels should be dried for a long time in a moderately warm place. Charging into Muffle. The cupels are placed, empty, in the muffle, in rows of 3 or 4, as may be convenient, with the hollow uppermost. After about 10 minutes, with a good fire, they will begin to show a red heat, and the lead buttons may then be added by means of suitably shaped cupel tongs, each button being care- fully placed in the hollow of its corresponding cupel. It is a good plan to begin by adding the front row (i.e., the row nearest the mouth of the mu ffle) , as these are generally cooler than the rest, and require more time to complete the cupellation, The other rows are then added in succession from front to back. The num- ber of assays to be cupeled simultaneously will depend on the size of the furnace and of the cupels, but it is not a good plan to crowd too many into the muffle at one time; if the furnace be heated sufficiently to cupel the front row in a full muffle, some of the back rows may be much overheated, leading to heavy losses by volatilization. Size and Form of Cupels. The size of cupels to be used will depend on the weight of the lead buttons. It is generally assumed that a bone-ash cupel will absorb its own weight of lead. For buttons of 20-25 grams, the size known as No. 7 is convenient; for 15 grams, No. 5 may be used. The depth of the hollow is also of some importance. If too shallow, the molten lead may overflow, although there may be sufficient absorbent material; if too deep, the oxidation proceeds very slowly and the losses of gold and silver by volatilization may be heavy. A layer of bone- ash should be placed on the floor of the muffle, as a precaution 342 THE CYANIDE HANDBOOK against accidents and to keep the litharge from attacking the fire-clay if it should soak through the bottoms of the cupels. Process of Cupellation. When the lead buttons are placed on the cupels, in a few moments the lead melts, and the surface shortly afterward brightens and begins to exhibit a rotary move- ment. This bright appearance continues until the operation is complete, provided a sufficient temperature be maintained. When the operation is finished the movement ceases, but the bead of gold and silver may remain molten for some time. When drawn forward to a cooler part of the muffle, it suddenly solidifies, emitting a bright glow of light; if cooled too rapidly particles of metal may be projected onto the edges of the cupel or entirely lost. This action is generally Ascribed to the expulsion of absorbed oxygen as the silver solidifies. Large buttons exhibit a play of iridescent interference colors a few moments before the finish of the operation. At a suitable temperature, a button of lead weighing 25 grams will be cupeled in about half an hour. Effects of Varying Temperature. If the temperature of the muffle during cupellation sinks below the melting-point of litharge, a crust forms on the buttons and the movement ceases. This is spoken of as " freezing." When this has taken place it requires a much higher temperature to re-start the cupellation, and there is always a possibility of loss in doing so. The temperature required to start cupellation and also to complete the final stages is considerably higher than that required to maintain the operation during the earlier part, after it has once started. Losses both of gold and silver occur in all cases, by volatilization and by absorption in the cupels, and as these losses are increased by rise of temperature, an accurate assay, especially where the determination of silver is of importance, is always conducted at the lowest temperature which can be used without danger of freezing. When the cupellation has been carried out at a fairly low temperature, the bead at the finish is surrounded by a ring of feathery yellowish crystals of litharge. Losses of Gold and Silver in Cupellation. According to Hillebrand and Allen, 1 "when cupellation takes place at a low temperature, with formation of considerable feather litharge, the loss by volatilization is practically negligible, or at any rate i Bull. No. 253, U. S. G. Survey (1905). ASSAY OF A TYPICAL SILICEOUS ORE 343 is perhaps compensated by retention of lead; but when the cupella- tion is made at a higher temperature, the loss is considerable. " The following average losses were determined by them in samples of telluride ore, by assaying the slags from the original fusions and the cupels respectively. The results give the amounts of gold and silver together per assay. In Slags: In Cupels: Sample Milligram Milligram 1 055 .085 2 08 .10 3 145 .16 4 05 .18 This cupellation loss is the amount absorbed, and does not in- clude the loss by volatilization. The latter could only be deter- mined by making a check assay with known quantities of gold and silver under conditions of temperature, etc., as nearly as possible identical with those of the actual assay. Experiments made by Hillebrand and Allen 1 showed that, contrary to the usual opinion, the losses of gold in cupellation are not negligible, especially with rich ores. Tests made by cupeling various known weights of gold and silver with 25 grams of lead in each case, to correspond with an ordinary assay button, led to the following conclusions: (1) That the losses of gold and silver increase progressively from front to back of muffle. (2) That, for equal amounts of gold, the loss is greater, with 25 grams of lead than with 5 grams. (3) That the percentage of loss of gold is greater with small than with large amounts of gold, especially in the hotter part of the muffle. (4) That the loss of gold by absorption is much greater than the loss by volatilization; when no silver is present the loss is chiefly due to absorption. (5) That the loss of gold is about the same whether silver is present or not. They insist on the necessity for corrected assays where accu- rate results are required, and state that the most exact results are obtained when feather litharge is still abundant at the time of brightening. 344 THE CYANIDE HANDBOOK Miller and Fulton 1 state that the absorption of silver by the cupel increases with the size of button, but the amount absorbed per gram of lead button cupeled diminishes regularly as the lead increases. They find also that the loss in diminishing a lead button by scorification is irregular, but usually greater than by cupeling direct. The corrections never account for the whole of the loss of silver. F. P. Dewey 2 remarks: "While the correction for slag and cupel loss is easily made, and ought always to be made when accurate statistics are kept, there is yet the volatilization loss to correct, and some means of doing this is very desirable. While a check assay answers very well for bullion, it would hardly be possible to construct check material for the varying characters of ores and products ordinarily met with." W. P. Mason and J. W. Bowman 3 give a table showing a number of results of losses of gold and silver on weighed quan- tities, when scorified and cupeled in a Battersea F muffle, the conditions under which the assays were made, such as heat of muffle, draft, and manipulations in general, being such as would obtain in careful, practical work. The average of their results show: TABLE IV. SCORIFICATION AND CUPELLATION LOSSES IN A BATTERSEA F MUFFLE PERCENTAGE Loss Silver Gold In Scorification. 0.55 1.99 0.574 0.296 In Cupellation For entire process 2.54 0.87 The methods of assaying slags and cupels will be discussed in a later section. Cleaning and Weighing Fine Metal. When the cupels are sufficiently cool, the beads of fine metal (gold and silver) are removed by means of a small pair of pliers, pressed tightly, and if large enough brushed with a wire scratch brush, then hammered i"Sch. Mines Quart.," XVII, p. 160 (1896). 2" Journ. Am. Chem. Soc.," XVI, p. 505. a" Journ. Am. Chem. Soc.," XVI, p. 313 (Oct. 6, 1893). ASSAY OF A TYPICAL SILICEOUS ORE 345 on a smooth clean anvil with a clean smooth-faced hammer. The flattened beads may conveniently be transferred to small porcelain crucibles, arranged on a cupel tray or wooden frame, in the same order as the cupels. They are then weighed on the assay balance. For ordinary work it is quite sufficient to weigh the fine metal within .05 or even .1 mg., as the variations in the losses of silver in different parts of the muffle probably exceed this amount. Inquartation. In cases where the amount of silver present is less than 2^ times that of the gold, the latter protects the silver to some extent from the action of the nitric acid in the subsequent operation of parting. When this is the case it will be necessary to add a small quantity of silver foil, a milligram or two in excess of the amount theoretically required, wrap the bead and extra silver in a small piece of lead-foil, and cupel in a small cupel (say No. 3 size). A better plan, however, is to make two assays of the material one without adding silver, and the other with the necessary silver for parting added to the flux before transferring to the crucible. It is advisable to examine the silver- foil used and make sure that it is absolutely free from gold. Parting. This operation may be carried out either in porce- lain crucibles or in small long-necked glass flasks. The former method is quite satisfactory, and if carefully carried out requires less manipulation and trouble than the method with parting-flasks. The acid required for parting the beads from ordinary ore assays is nitric acid, 20 per cent, by volume -i.e., a mixture of 20 parts pure acid (say 1.42 sp. gr.) with J80 parts distilled water. From 10 to 15 c.c. of acid are required for each bead. The cru- cibles are 4 to 4.5 cm. in diameter, and when this quantity of acid is added are filled about two- thirds full. The crucibles may be conveniently arranged on a perforated metal plate sup- ported over a uniform and moderate source of heat; an oil-stove with large wicks can be used, but the heat is not really uniform in all parts. The acid must be heated to boiling, but not allowed to boil too violently, as in some cases the beads may break up and pieces of gold may be projected from the crucibles and lost. The heating must be continued until no signs of red fumes can be observed, and until no further evolution of small bubbles takes place. Care must be taken not to allow the liquid to evaporate to dryness. 346 THE CYANIDE HANDBOOK With ordinary assay beads a second parting is hardly neces- sary. With large beads, however, it is advisable to pour off the first (weak) acid and add an equal amount of a stronger acid (50 per cent, by volume). This is heated to boiling and again poured off when the action appears to be finished. The crucible is then filled up with distilled water, which is carefully decanted; if the gold has broken up, as will be the case when a large excess of silver was present, this operation requires close attention, and it is best to pour the water off along a glass rod into a porcelain basin. After washing once or twice in this way, the crucibles are dried slowly to avoid spurting: generally this may be safely done by setting them on the perforated frame in a slightly inclined position and turning the lamp down. They are finally heated until the gold acquires its natural yellow color. When cool it is ready for weighing. When parting-flasks are used they are generally supported on a specially constructed frame so that they rest in a slightly inclined position; this is heated with the necessary precautions against bumping and spurting. After pouring off the acid the flask is filled to the top with distilled water and inverted into a small crucible of unglazed porous material (annealing-cup). This may easily be done without spilling any liquid by placing the crucible over the mouth of the flask before inverting the latter. The gold then sinks to the bottom of the annealing cup; the in- verted flask is gradually raised, allowing the cup to fill with water, the flask is then withdrawn by a quick lateral movement and the surplus water poured off. For small beads, test-tubes may preferably be used instead of parting-flasks. Generally speak- ing, this method is more troublesome than parting in crucibles. A glazed porcelain crucible may of course be used instead of the porous annealing-cup, and is perhaps preferable, as small particles are liable to chip off the edges of the latter and might occasion errors in weighing broken gold, but greater care is necessary in drying the gold in a glazed crucible. Purity of Parting Acid. Care must be taken that the nitric acid used is free from chlorides, free chlorine, hydrocyanic acid, or any impurity which could precipitate silver in acid solution or dissolve gold, either by itself or in conjunction with nitric acid. Silver nitrate added to the diluted parting acid should give no turbidity. If any milkiness be observed, it is best to add a ASSAY OF A TYPICAL SILICEOUS ORE 347 certain quantity of silver nitrate to all the parting acid made up from that lot of strong acid, and to allow the turbidity to settle before using. Nitrous acid has generally been considered to be a solvent for gold; 1 but the researches of Hillebrand and Allen 2 would appear to show that this acid has no such action. Nitrous acid is shown by a reddish brown color and by the dilute acid liberating iodine from potassic iodide (test with starch solution). The same writers also found no evidence of the solution of gold by prolonged boiling, in pure nitric acid. There is, however, some reason for believing that, under conditions not clearly under- stood, a quite perceptible quantity of gold may be dissolved, so that the assayer will do well to use the purest acid obtainable and avoid prolonging the process of boiling after the action on the silver has entirely ceased. The distilled water used must of course also be free from any impurities which might precipitate silver, act on the gold, or otherwise interfere with the process. Adjustment of Assay Balance. Before weighing a batch of gold assays it is essential to verify the adjustment of the assay balance. The case should be free from dust, inside and out, and the plate on which the balance rests should be quite level, usually shown by two small spirit-levels placed at right angles inside the balance case. If necessary, it must be leveled by raising or lowering the supporting screws. If the balance is much out of adjustment, it is rectified by moving the vane or adjusting screw in the required direction. Small variations in adjustment are best corrected by means of a rider. If the right arm of the balance be used for the weights, as is generally the case, the adjusting rider may conveniently be placed on the left arm, which is kept permanently one- or two-tenths of a mg. lighter than the right arm. Any small variation of adjustment occurring during the course of weighing (due to differences of temperature, etc.) can thus be corrected by moving the left- hand rider, without any disturbance of the balance. Weighing Parted Gold. The crucibles or annealing-cups containing the parted gold are arranged in their proper order on a cupel tray, or on any suitable frame which allows them to be moved simultaneously. The gold is then transferred from iG. H. Making, in " Journ Chem. Soc.," XIII, p. 97 (1861). 2 Bull. No. 253, U. S. G. Survey (1905). 348 THE CYANIDE HANDBOOK the crucible to the balance pan by means of a needle or very fine camel's-hair brush; sometimes a slight tap may be needed to move it. The pan should be placed on a sheet of glass or a smooth card, so that any loose particles which might be lost in transfer may be at once seen. A good balance will determine the weight within .01 mg., which is sufficient for most purposes, this amount corresponding to 4.8 grains per ton (20 cents gold value) on an as- say of 1 assay ton, or 2.4 grains (10 cents) on an assay of 2 assay tons. Most balances are so constructed that weights less than 0.05 mg. and weights lying between 0.9 and 1.0 mg. cannot be con- veniently weighed with the rider alone. In such cases a counter- poise of known weight (say 0.5 mg.) may be placed in the pan, or the left-hand rider may be moved one or more divisions, so that the zero-point in weighing is no longer at the center of the beam, care being taken to return the rider to its normal position after each such weighing. After weighing, the gold should be immediately transferred to a small cup or other vessel placed in the balance case, so that the accumulated beads may be recovered, and also to avoid any risk of their finding their way into other samples. Reporting Results. A convenient method of recording the assay results is to use the headings here shown: No. of Sample Description Date WEIGHT FOUND ASSAY OF SAMPLE Gold and Silver: mg. Gold: mg. Gold : Dwt. Silver: Dwt. The weights of fine metal in milligrams are noted in the column headed " Gold and Silver," and the weights of gold are subse- quently entered in the next column, headed "Gold: mg." The difference of these entries gives the weight of silver in milligrams. If the assay be on 1 assay ton, the weights of gold and silver in milligrams must be each multiplied by twenty to obtain the assay in pennyweights. If the assay be on 2 assay tons, they must be multiplied by ten. SECTION III SPECIAL METHODS OF ASSAY FOR PARTICULAR ORES AND PRODUCTS (A) THE SCORIFICATION ASSAY Variable Nature of Material for Assay. The ores treated, and the materials obtained as by-products in the treatment of gold and silver ores by the cyanide process, are of such varied nature that no single method of assaying is applicable in all cases. When these ores or materials are of high assay value, or present diffi- culties in the ordinary method of assaying as described in the previous section, special care and experience are required in dealing with them. In doubtful cases it is always well to try two or more distinct methods on the same sample. The agree- ment of duplicate assays by the same method cannot be accepted as a proof of correctness, as experience often shows that a different method may yield a totally different result. Scarification. The most generally applicable method for many different classes of material is scorification, and this will be first considered. As a rule the method can only be conveniently applied to comparatively rich ores or products, as the amount taken for each assay of the substance is necessarily small, an ordinary charge being 5 to 8 grams. In exceptional cases as much as 15-20 grams may be taken, but this involves the use of large scorifiers. The scorifiers in general use are small fire-clay dishes varying in size from 1^ in. up to 3 or 4 in. in diameter. The substance to be assayed is mixed with pure lead in the form of small grains, and placed in the scorifiers. The general custom is to mix half the lead with the material to be scorified and add the remainder as a cover, together with a very small amount of- borax or other flux. In some cases, litharge may be added with advantage instead of, or as well as, metallic lead. The following five charges are given as examples for ordinary 349 350 THE CYANIDE HANDBOOK cases; the quantities given are in grams, unless otherwise desig- nated : Charge 1 Charge 2 Charge 3 Charge 4 Charge 5 Weight of material Grain lead Litharge Fused borax Silica Diam. of scorifier a. t. 50 0.2 5 60 0.25 2 Jin. 5 30 35 0.25 2 1 in. 70 0.25 Sin ia.t. 27 22 0.5 0.4 3 in. The last (No. 5) is suitable for pyritic ore or similar material. The amount of silica may be varied according to the quantity of pyrites present in the ore. In this case the litharge and silica are to be mixed with the ore and the lead and borax added as a cover. Charges for Material Containing Copper. The following is suitable for material containing small amounts of copper and low-grade in precious metals: Charge 6 Material for assay 14^ grams = \ a.t. Grain lead 60 Fused borax 0.05 " Scorifier 3^ in. For highly cupriferous material: Charge 7 Substance for assay 7.23 grams = \ a.t. Grain lead 75 Borax 0.2 to 0.3 " Scorifier 3^ in. or 4 in. Charges for Rich Silver Ores. For rich silver-bearing material one of the following may be used: Charge 8 Charge 9 Substance for assay ^ a.t. ^V a.t. Grain lead 30-70 grams 40-60 grams Borax 0.3-3 " 0.5 " Scorifier 2^ in. to 2f in. 2^ in. Zinc-Box Precipitate. The following is given by R. W. Lodge l for the assay of zinc-box precipitate : i Trans. A. I. M. E., XXXIV, p. 432 (October, 1903). SPECIAL METHODS OF ASSAY 351 Charge 10 Precipitate 0.05 a.t. Grain lead 65 grams Borax glass 10 " Scorifier 3 in. or 4 in. 35 grams of the lead to be mixed with the material and 30 grams to be used as cover Lead Required for Scorifying Different Kinds of Material. As will be seen by the above figures, the amounts of grain lead required to scorify various kinds of material are very different, according to the nature of the material. The following table gives some of the usually accepted figures on this subject: TABLE V. GRAIN LEAD REQUIRED WITH VARIOUS MATERIALS Nature of Material Taken To BE ADDED PER GRAM OF MATERIAL ASSAYED Lead : Grams Borax: Grams Galena 5- 8 8-10 8-10 10-15 12 12-20 16 16-18 0.1-0.15 nil 0.25-1 0.1-0.2 0.05 0.1-0.15 0.1-1 0.1 (with litharge cover) Quartzose ores Basic ores Pyritic ore 1 Blende J Graphite Fahl ore Arsenical ore ] Antimonial ore . . I Telluride ore Mode of Procedure. In an ordinary scorification assay, the materials, mixed as directed, are placed in the hollow of the scorifier, which is then introduced into the muffle at a low red heat. In many cases it is preferable to add the borax little by little, as the action proceeds. The temperature is then raised gradually to bright redness, keeping the muffle door closed, say, for 15 minutes, by which time the mass should be completely fused. The door is then opened sufficiently to give a good current of air, when litharge is rapidly formed, and may be seen forming a ring round a bright central spot or "eye." This spot gradually diminishes until it is completely covered by the layer of molten slag. The temperature is now again raised, by closing the door or increasing the draft. At this stage it is sometimes a good 352 THE CYANIDE HANDBOOK plan to add a small quantity of a reducing agent, say 0.5 to 0.8 part of anthracite powder to every 100 parts of lead originally taken. This is wrapped in tissue paper and placed quickly on top of the slag; it serves to reduce any silicates of silver that may have been formed. When all action appears to have ceased and the contents of the scorifier are quite liquid, the latter is removed from the muffle by means of " scorifier " tongs, consisting of two arms, the upper one of which rests across the top of the scorifier while the lower one is branched in U-shape, so as to hold the bottom of the scorifier firmly. While the scorifier is held in this way its contents are poured into a dry, clean mold. When cool, the slag is detached and the lead button cleaned for cupellation. The process must not be continued too long after the "eye" has closed, i.e., after the lead is completely covered by the layer of slag, as the size of the lead button gradually diminishes, and when too small much of the gold and silver may pass into the slag. Remelting Slag. In any case it is advisable, with all samples that are at all rich, to remelt the slag with fresh flux. This may be done by grinding the slag in a mortar and fusing in a clay crucible with 20 grams litharge, 2 grams sodium carbonate, and 1 gram charcoal. The resulting lead button is cupeled and the result obtained added to that from the scorification assay; or the two buttons may be cupeled together. Theory of Scorification Assay. The theory of the process is described by J. Daniell 1 as follows: " The oxidation of lead produces litharge, or litharge is added with the charge. With the silica (of the scorifier or of the ma- terial assayed) this forms the extremely fusible silicate Pb 2 SiO 4 . By sulphides, the oxide is reduced back to metallic lead with formation of sulphur dioxide in presence of a current of air, and a metallic oxide which is dissolved by the excess of litharge. In fact, fused litharge is an extremely powerful solvent, and there are very few substances which are not attacked and held in 1 igneous solution ; by it. The noble metals are thus concentrated in the unoxidized metallic lead and are subsequently obtained by cupellation." Inconveniences of Scorifying in Muffle. With ordinary muffle furnaces there is a tendency, already noted, for the back of the muffle to become much hotter than the front; hence the i " Proc. Chem., Met. and Min. Soc. of South Africa," II, 277. SPECIAL METHODS OF ASSAY 353 scorifications at the back are apt to finish first, and where large numbers are assayed simultaneously it may be troublesome to remove these at the proper time without upsetting or interfering with those in front. Scarification Assay in Fusion Furnace. A modified form of the process, described by Daniell, 1 will meet this difficulty. The operation is carried out in a "Cornish fire." " Duplicate trials of about 3 grams of ore were taken, mixed with 25 grams of granulated assay lead and a small quantity of a flux composed of equal weights of carbonate of soda and borax. Another quantity of 25 grams gran- ulated lead was then put in and covered with a further quantity of flux. The crucibles were placed in a full fire at a low heat [small wide-mouthed crucibles were used instead of scorifiers], with damper closed, 4, 6 or more at a time, depending on the size of the furnace. When melted, the bricks were opened, and damper raised, causing a current of air to play upon the crucibles, ensuring rapid oxidation of the lead and a thorough scorification of the ore. . . . The use of the Cornish fire in this way has been thoroughly tested and gives per- fectly satisfactory results." Cautions. T. Kirke Rose 2 remarks, in reference to scorifica- tion: "Effervescence and spurting may occur, especially if the scorifier has not been well dried by warming before it is used." Daniell 3 says: "There are, of course, a few precautions which have to be taken. For instance, pyritic material, especially arsenical pyrites, has a tendency to decrepitate, throwing out minute particles which burn with a characteristic sparkle and entail loss. This can be obviated by careful attention to the heat at the commencement of the operation. Again, carbonaceous matter in presence of litharge generates gas, which, rising through the melted lead, may occasion loss by projection of particles." Reduction of Large or Impure Buttons. It sometimes happens on pouring a scorification assay that the button of lead obtained is too large to be conveniently cupeled, and in other cases it may be hard, brittle, or coppery in appearance. In such cases the slag should be removed and the button replaced in the same scorifier, fresh lead being added in the case of hard or impure buttons. The scorification is continued until the button is reduced to the proper size for cupellation and appears quite clean and malleable. According to Rose 4 (loc. cit.) less loss of the precious metals is i ibid. a" Metallurgy of Gold," p. 383. 3 " Proc. Chem., Met. and Min. Soc. of South Africa," II, 277. 4 Compare Miller & Fulton, " Sch. Mines Quart.," XVII, p. 160. 354 THE CYANIDE HANDBOOK incurred by scorifying lead than by cupeling it, and consequently it is better to reduce any lead button weighing more than 20 grams by rescorification before cupellation. Assay of Graphite Crucibles. The following method is given by T. L. Carter 1 for assaying graphite crucibles (i.e., old pots that have been used for smelting operations) : " A quantity not exceeding 6-7 grains (say .4 to .5 grams) of the finely powdered material is taken; if poor, several buttons are combined. The bottom of a large scorifier is rubbed with silica, then 40 grains (say 3 grams) of pure litharge is introduced, followed by the plumbago arid 4 grains (say .25 gram) of niter. These materials are mixed together, 30 grains (i.e., 2 grams) litharge added, and finally a covering of borax. The temperature is kept very low for a few minutes, then gradually raised to a white heat till the mass fuses completely. As a check, 3 assays are made by scorification and 3 by pot fusion." Acid-Treated Residues. Scorifioation is also used sometimes for assaying the residues from ores and other products which are largely soluble in acids, as certain cupriferous materials, etc. This will be referred to later. Limitations of Scorification Assay. Dr. Loevy 2 observes: " For Rand ores and for all kinds of tailings and slimes, scorifica- tion is entirely out of the question; but for a certain class of ma- terial, such as graphite crucibles, certain slags, antimonial ores, lead ores, etc., the scorification assay is the only reliable one." This statement appears a little too sweeping, especially as regards antimonial and lead ores. (See below.) (B) CRUCIBLE FUSIONS FOR VARIOUS CLASSES OF ORE Classes of Ore for Assay. As already remarked, the flux to be used must always be adjusted to the nature of the ore, and no fixed rules can be given. The following examples may, however, be of some use as a guide. The classes which will be here con- sidered are: Class I. Siliceous (Quartzose Ores). Class II. Basic (Oxidized) Ores. Class III. Pyritic Ores. " Eng. and Min. Journ.," Aug. 5, 1899 (p. 155). a " Proc. Chem., Met. and Min. Soc. of South Africa," II, 206. SPECIAL METHODS OF ASSAY 355 Class IV. Lead-Zinc Ores. Class V. Arsenical Ores. Class VI. Antimonial Ores. Class VII. Cupriferous Ores. Class VIII. Telluride Ores. Methods of Assay. The methods of assay employed may be grouped under the following heads: (a) Direct fusion with ordinary fluxes (i.e., soda, borax, litharge and charcoal, or substitutes for these). (b) Direct fusion with special oxidizing fluxes (niter, manganese dioxide, etc.). (c) Direct fusion with a desulphurizing agent generally metallic iron. (d) Fusion after preliminary roast or oxidizing process. (e) Fusion after preliminary wet chemical treatment. Class I. Siliceous Ores Consisting mainly of quartz, with very small quantities of oxide of iron, iron pyrites, calcium and magnesium silicates and carbonates, etc. The quantity taken for assay will vary from J assay ton up to 4 assay tons, according to richness, but on a cyanide plant it is seldom desirable to use less than 1 assay ton. Fluxes for Free-Milling Ores. We give here a number of fluxes adapted from those recommended by different writers, so as to be applicable to charges of 1 assay ton. The quantities are given in grams: Flux No. 1 No. 2 No. 3 No. 4 No. 5 No. 6 Bicarbonate of soda 45-68 50 60 60 50 48 Carbonate of soda 30-46 33 40 40 33 32 Borax glass 7-15 10 6 cover nil 5 Litharge 15-30 40 45 98 49.55 57 Charcoal 1-1.5 as required 1 0.45 1.1 Flour 2.4 Nearly all writers on assaying give bicarbonate of soda in the fluxes recommended. If the monocarbonate (Na 2 CO 3 ) is to be used, the amount Of bicarbonate should be reduced in the pro- portion of 2 to 3; thus 40 grams Na.C0 3 are equivalent to 60 grams NaHCO 3 . Of the above fluxes, No. 1 is based on the recommendations of T. Kirke Rose. 1 The quantities there given are "From 1 to 1J " Metallurgy of Gold," 4th edition, p. 470. 356 THE CYANIDE HANDBOOK assay tons of soda carbonate and from J to J assay ton of borax to 1 assay ton of ore." And with regard to litharge he remarks: "Litharge or red lead is added in the proportion of one or two parts to two of ore; if too much litharge is used the slags are not clean, as a slag containing lead may mean a loss of silver and gold." [The amount of litharge here recommended appears to be much less than that which the general experience of assayers shows to be desirable. J. E. C.] No. 2 is given as applicable to ordinary quartz ores on the Rand, and is recommended by B. W. Begeer. 1 No. 3 is adapted from the flux given by C. and J. Beringer. 2 As- suming the reducing power of charcoal to be twice that of flour, we might substitute 1.2 grams charcoal for the 2.4 grams flour above. No. 4 is a flux calculated on the lines of that given by C. H. Fulton 3 for the production of a " sesquisilicate " slag, with a charge of 0.5 assay ton. The amount of NaHCO 3 is doubled, and the litharge is calculated on the assumption that 76 grams are required for the slag and 22 grams to give a 20-gram lead button. No. 5 is a flux recommended by G. B. Hogenraad 4 for the ore at Redjang Lebong, Sumatra, which is practically a free- milling ore, but contains small amounts of pyrites, manganese peroxide, copper, and selenium. Borax is omitted, as it was thought to carry part of the silver into the slag. No. 6 is the flux adopted by the writer for free-milling ores of ordinary types. (See Section II, A.) The following fluxes are adopted in a similar way for charges of 2 assay tons: Flux No. 1 No. 2 No. 3 No. 4 No. 5 Bicarbonate* of soda . . . . 100 100 117 110 85 Carbonate of soda 07 67 78 73 57 Fused borax 30 20 12 8 Litharge 45 80 88 89.7 120 Charcoal - as required 0.3 2 Flour - 4.7 Argol 2.5 Glass - 20 The quantities are given in grams. No. 1 is a flux formerly used in the assay office of the Robinson G. M. Company of Johannesburg; it was found to give lead buttons i " Metallurgy of Gold on the Rand," 1898. 2 "Text Book of assaying," 9th edition, p. 138. 3" A Manual of Fire Assaying," 1907, p. 66. 4" Journ. Chem., Met. and Min. Soc. of South Africa," VIII, 73 (1907). SPECIAL METHODS OF ASSAY 357 of about 20 grams, with a clear green slag. The borax seems excessive. No. 2 is given by B. W. Begeer * for tailings on the Rand. No. 3 is adapted from a flux given by C. and J. Beringer, 2 the quantities there given for 50 grams being increased in the pro- portion 58J : 50. We might substitute 2.35 grams of charcoal for the. 4. 7 grams of flour above given. No. 4 is a flux recommended by G. B. Hogenraad for the 2 assay ton charges at Redjang Lebong. It is said to give lead buttons of 30 grams, with a clear green slag. With pure siliceous ores more charcoal would be necessary. No. 5 adopted by the writer. (See Section II, A.) The following flux, suitable for cyanide tailings, is given by C. and J. Beringer, 3 for a charge of 3 assay tons. Bicarbonate of soda 135 grams Carbonate of soda 90 " Borax 22 " Litharge 135 " Charcoal as required. The writer uses the following for charges of 4 assay tons, the fusion being made in a Battersea No. 12 fluxing pot. Bicarbonate of soda 161 grams Carbonate of soda 108 " Fused borax 15.7 " Litharge 141 " Charcoal . 2.3 " Total, 428 grams Probably the total quantity of flux per charge might be increased with advantage, but in that case a larger crucible would be neces- sary. General Remarks With certain classes of ore the amount of litharge in the flux may be reduced to some extent, if the amount of soda be corre- spondingly increased, and vice versa. As already pointed out, carbonate of soda is for various reasons preferable to bicarbo- nate. (Section II, A.) An excessive quantity of borax should be avoided, as this gives rise to a very hard stony slag to which the lead adheres very tenaciously. On hammering such a slag it is liable to fly to pieces with explosive violence, carrying with it portions of the lead. There is also a possibility that some gold i " Metallurgy of Gold on the Rand," 1898. 2" Text Book of Assaying," 9th edition, p. 138. 3 Ibid., p. 140. 358 THE CYANIDE HANDBOOK and silver may be retained chemically in slags high in borax. With little or no borax the slag is usually thick and not very fusible. Excess of soda gives a very deliquescent slag, which is liable to attack the crucibles. Excess of charcoal gives a very infusible black slag, and may cause heavy losses in gold and silver. The use of salt, sometimes recommended as a cover, is chiefly to moderate effervescence during the expulsion of C0 2 in the early stages of the fusion, and to prevent oxidation of lead by exposure to air. It is held by many assayers that the addition of salt causes losses, especially of silver, hence where such a cover is needed it is better to use borax. The writer sometimes uses a cover of soda. Class II. Basic Ores. The general principle in assaying material of this class is to increase the borax and reducing agent, and to somewhat diminish the alkaline flux (soda, litharge). Ferric iron should be con- verted as much as possible into ferrous. For very basic ores it is advantageous to add silica or other acid flux in addition to borax. The following fluxes are suggested for assays on 1 assay ton of ore: Flux. No. 1 No. 2 Carbonate of soda 18 30 Borax 12 30 Litharge 40 45 Charcoal 2.5 2 No. 1 is adapted from a flux given by E. A. Smith, 1 the litharge and charcoal being unchanged, but the soda and borax reduced proportionally to the amount of ore taken. No. 2 is given by C. and J. Beringer (loc. cit., p. 139). For assays on 2 assay tons of ore: No. 1 No. 2 No. 3 Carbonate of soda 90 36 12-18 Borax 30 24 36 Litharge 45 40 50 Charcoal - 2.5 4-6 Sand or glass - 12-24 Argol 3 i "Assaying of Complex Gold Ores," in Trans. I. M. M., IX, p. 320 (1901). SPECIAL METHODS OF ASSAY 359 Flux No. 1 was used on the Rand for chlorination tailings and similar material that was mainly siliceous, but contained some ferric oxide. No. 2 is adapted from a flux given by E. A. Smith (loc. cit.), the soda and borax being increased proportionally to the amount of ore, leaving the litharge and charcoal unchanged. No. 3 is similarly adapted from a flux given by Smith for very basic ore. C. and J. Beringer (loc. cit., p. 139) also give the following: No. 1 No. 2 No. 2a Charge of ore 2.5 assay tons 50 grams 2 assay tons Soda (bicarbonate) .... 75 grams 30-50 grams 36-60 grams Soda (carbonate) 50 " 20-33 " 24-40 " Borax 75 " 30 " 36 " Litharge 120 " 50 " 50 " Silica - " 10-25 " 12-30 " Charcoal 4 " Flour - " 7 " 7 No. 1 is for low-grade material such as chlorination tailings; to be assayed in a large crucible (size I) . No 2 is for ore consist- ing chiefly of hematite. It is very similar to the flux given by Smith for very basic ore, except as regards the amount of soda. Some assayers use metallic iron with this class of ore, but it would seem to be unnecessary if sufficient reducing agent be used. Finely powdered gas-carbon has been recommended as a very efficient reducing agent; it has a higher reducing power, weight for weight, than -charcoal, and being less rapidly oxidized acts for a longer time and at a higher temperature during the course of the fusion. Class III. Pyritic Ores Methods of Assaying Pyritic Ores. The principal methods used in assaying ores containing larger or smaller amounts of iron pyrites are: (a) Roasting. (b) Fusion with excess of litharge and little or no reducing agent. (c) Fusion with metallic iron. (d) Fusion with an oxidizing agent, generally niter. These will be considered in the order given. (a) Assay by roasting. According to the richness of the ore, 1 or 2 assay tons are weighed out and placed in a shallow clay dish 360 THE CYANIDE HANDBOOK 4 to 5 in. in diameter and spread out in a thin layer. Rectangular dishes made to fit inside the muffle are sometimes used. For highly pyritic ores it is better to place a layer of silica (15-20 grams) on the dish before adding the ore. This serves to prevent the roasted material from adhering to the dish and also aids as a flux in the subsequent fusion. The roasting-dish is placed at the mouth of a large muffle, which is kept for some time at a very low heat. To prevent " caking/' or the formation of hard lumps during this process, the contents of the dish are stirred occasionally, but not too much, with a thin iron rod, bent and flattened at the end. After some time, if much pyrites be present, a blue flame of burning sulphur will be observed over the dish. The temperature is raised gradually to dull redness and the dish, pushed toward the back of the muffle. After all sulphur fumes have ceased and there are no perceptible sparks on stirring, the temperature may be further raised to complete the roast. The material when roasted " dead " should have a uniform appearance, and should emit no smell whatever of sulphur, etc. Arsenical and antimonial ores require special treatment in roasting, which will be described later. When the roasting is complete, the dish and contents are allowed to cool (under a cover, if left for some time), and the roasted ore swept off the dish into a mixture of suitable flux by means of a small flat brush. It is then well mixed with the flux and assayed in the ordinary way. When roasted without silica, the ordinary pyritic ores, con- sisting mainly of siliceous materials and containing not more than 25-30 per cent, iron pyrites, may be assayed by one of the first four fluxes given for basic ores; if very pyritic (over 30 per cent. FeS 2 ) the flux for very basic ores or for hematite should be used for the roasted material, since in roasting the sulphides of iron are converted into oxides. The process takes place in several stages, but the ultimate result may be represented thus: 2FeS 2 + 11O = Fe 2 O 3 + 4SO 2 When the roasting process is properly carried out, the heat kept low enough in the early stages and not hurried, and when the proper flux is used for the fusion of the roasted material, the results are generally at least as good as those obtained by any direct fusion method. However, the time occupied in the roast- ing, often an hour or more, is a disadvantage; hence the method SPECIAL METHODS OF ASSAY 361 of fusion with excess of litharge is to be preferred for slightly pyritic material. Assay of Pyritic Concentrates by Roasting. The method usually adopted by the writer for the assay of pyritic concentrates on the Rand was as follows: About i assay ton (say 15 grams) of siliceous sand, free from gold or silver, was spread out on a 5-in. roasting-dish, and 1 assay ton of the concentrates laid in a thin layer on the sand. This was roasted as above described, and when cool mixed thoroughly with Carbonate of soda 90 grams Borax 30 " Litharge 50 " Gas carbon 1.3 " and fused. The carbon was varied, if necessary, so as to give a button of 18 to 25 grams. (b) Fusion with Excess of Litharge. The principle of this method is that lead oxide and pyrites act on one another when the former is in excess, with formation of oxides of iron and sul- phur, metallic lead being liberated. The pyrites thus supplies the place of the whole or part of the reducing agent. In the flux used, rather more than the ordinary amount of litharge is taken; the charcoal is reduced as the amount of pyrites increases, and in some cases may be omitted altogether. The following fluxes maybe used for an ordinary type of slightly pyritic ore: No. 1 No. 2 Ore taken 1 assay ton 2 assay tons Carbonate of soda 45 grams 90 grams Borax 15 " 15 " Litharge 50 " 70 " Charcoal 0.8 to nil " 0.8 to nil " Mitchell gives the following, which could probably be used for most ores of this class: Ore 1 assay ton Carbonate of soda 30 grams Borax 30 " Litharge 150 " Salt cover 30 " Argol, sufficient to give a button of 13 grams. [A larger button would probably give better results. J. E. C.] 362 THE CYANIDE HANDBOOK Limitations of the Method. The method is not convenient for highly pyritic ores, as in such cases a large amount of litharge must be added to ensure having an excess, and consequently very large lead buttons are produced, which must be reduced subse- quently by scorification. If the litharge be not in excess, the lead button will be brittle, consisting partly of lead sulphide, and more or less of the values will be lost. (c) Fusion with Metallic Iron. The principle of this method is to convert the pyrites into ferrous sulphide in accordance with the reaction: FeSa + Fe = 2 FeS The ferrous sulphide passes into the slag without causing any considerable loss of precious metal, and a malleable lead button is formed. Large nails are sometimes used, but a better plan is to bend a piece of hoop-iron in U-shape; this is pushed into the charge in the crucible; when the fusion appears to be complete and the slag is quite fluid, the strip of iron is removed with the tongs, tapped once or twice to detach any adhering slag or lead, removed from the crucible, and the fusion continued about 5 minutes longer before pouring. E. A. Smith 1 gives the following flux as suitable with this method of assay (adapted for assay of 2 assay tons) : Ore ................................ 2 assay tons Carbonate of soda .................... 48-60 grams Borax .............................. 12-18 " Red lead ............................ 48-60 " Charcoal ............................ 2 to nil Hoop-iron (thick) .......... 2 or 3 pieces to be added Aaron 2 gives the following: Ore .............. 1 assay ton Carbonate of soda . . 90 grams Borax ............. 15 " Litharge ......... 30 " Sulphur... ......... 3 " Flour ............. 3 " Iron nails ........ 3 to be added (with glass, if a more acid flux is needed, and a cover of salt). [The object of the sulphur is not particularly obvious. J. E. C.] This method requires considerable care and experience for i " Assaying of Complex Gold, Ores," in Trans. I. M. M., IX, p. 320. a Rose, " Metallurgy of Gold," 4th edition, p. 469. SPECIAL METHODS OF ASSAY 363 satisfactory results. The strips of iron when withdrawn from the crucibles are apt to carry small adhering shots of lead, and should be carefully examined. (d) Assay with Niter. In this method it is presupposed that the quantity of pyrites in the ore is more than sufficient to bring down the required amount of lead. The surplus lead is oxidized by the addition to the flux of a carefully adjusted quantity of niter. [It is more probable that the niter attacks the iron pyrites rather than the lead, as the slag in this process invariably con- tains potassium sulphate.] An excess of niter causes heavy losses of precious metals, and sometimes causes corrosion of the crucible itself; hence when this method is tried on an unknown ore, a trial fusion is always made to determine the " reducing power " of the ore. In such a case, some such mixture as the following is made: Ore i assay ton (7.292 grams) Litharge 60 grams Carbonate of soda 4 " Miller, Hall, and Falk 1 give the following: Ore 3 grams Soda bicarbonate 10 " Litharge 50 " Salt cover The mixture is fused at a bright red heat for from 10 to 15 minutes, then poured, and the resulting lead button weighed. The weight of this button is a measure of the reducing power of the ore. It is generally assumed that 1 gram of niter will oxidize 4 grams of lead. E. A. Smith (loc. cit., p. 335) says 4 to 5 grams, but the amount varies with the nature of the ore, and may range from 3.2 to 5.3. Formula for Niter Required. The following is a general formula which can be applied in all cases for calculating the amount of niter required: Let a = weight of ore taken for trial fusion. ' A = weight of ore taken for final assay. < w = weight of lead button from trial fusion. ' W= weight of lead button required. ' n = number of grams lead oxidized by one gram of niter. ' x = weight of niter required. Then Aw Wa x = na i Trans. A. I. M. E., XXXIV, p. 387. 364 THE CYANIDE HANDBOOK In the special case where 1 gram niter oxidizes 4 grams- lead, and J assay ton is taken for trial fusion, 1 assay ton for final fusion, and a button of 20 grams is required, we have x = w 5 Oxidizing Power of Niter on Various Minerals. Smith (loc. cit., p. 327) gives the following table to determine the oxidizing effect of niter, when the nature and amount of the oxidizable material is known: TABLE VI. OXIDIZING EFFECT OF NITER Parts Niter Required to 1 Part of " Sulphuret " Iron pyrites 2-2 parts Copper pyrite, fahlerz, or zinc-blende Antiinonite . li-2 " U " Galena * " No. 2 | assay ton 15 grams 15 " _ 70 " Final Fusion. For the final fusion, as a rule, a flux is used containing much litharge and little soda, together with the indi- cated amount of niter and a cover of salt to modify the effer- vescence. The violent action is a frequent cause of loss in assays with niter. The following may be given as examples of fluxes for the final fusion : No. 1 Ore 2 assay tons Bicarbonate of soda ... 45 grams Carbonate of soda 30 " Borax 12 " Litharge Red lead 120 " Niter as indicated. Salt cover. No. 1 is given by E. A. Smith (loc. cit.). No. 2 is given by Miller, Hall, and Falk (loc. cit.). These authors found that the use of borax was objectionable, giving an excessively hard and stony slag, with hard brittle lead buttons. [This would probably apply only to certain classes of ore. J. E. C.] They summarized their experience of the process as follows: " It is necessary to determine the oxidizing power of niter with that substance and charge with which it is subsequently used; the niter method, as modified, gives accurate results and is neither as long nor as tedious as the roasting method; it also gives a SPECIAL METHODS OF ASSAY 365 charge which does not boil over, and yields a lead button agree- ing closely with the calculated weight, and a clean slag." Another method of applying niter in assaying will be described under "Antimonial Ores" (Class VI). Class IV. Lead-Zinc Ores Ores containing galena and zinc-blende are best assayed by fusion with metallic iron, as already described under "Pyritic Ores" (Class III, c). Where galena predominates, the amount of litharge may be reduced; when the zinc mineral predominates, the amount of soda and borax should be increased. One of the two fluxes given above will probably meet most cases. Hall and Popper 1 give the following as suitable to certain kinds of zinc ore (sphalerite) : Ore i assay ton, say 10 grams Carbonate of soda 1J " "40 " Borax glass i " "15 " Litharge | " "24 " Argol as required. Add iron if more than 15 per cent, pyrites is present. [The amount of soda here seems excessive, and in fact the authors remark that the silica of the crucible appeared to be attacked, and the charge was inclined to boil excessively. J. E. C.] Roasting may be employed when the ore is chiefly or exclusively zinc-blende, but is unsatisfactory with any but very small quan- tities of galena. Class V. Arsenical Ores Mode of Occurrence. The arsenic in gold ores usually occurs as mispickel (FeAsS), occasionally as arsenide of nickel or as realgar and orpiment (sulphides of arsenic). Roasting with Charcoal. The best method is to first roast at a low temperature, eliminating part of the arsenic, then allow to cool somewhat; add 5 to 10 per cent, of powdered charcoal, stirring well, to mix with the charge on the roasting-dish, and continue heating at a somewhat higher temperature until a dead roast is obtained. The roasted material is then fluxed as described under basic ores, or pyritic ores after roasting. i " Sch. Mines Quart.," XXV, p. 355 (1904). 366 THE CYANIDE HANDBOOK Theory of Method. The object of adding charcoal is to de- compose the arseniates formed in the first stage of roasting. These compounds, if not destroyed, would carry gold and silver into the slag. It has been stated 1 that the addition of a small quantity of potassium cyanide to the flux facilitates the decomposition of any arseniates remaining in the roasted product. It is not advisable to assay ores of this class by means of niter, as this tends to form arseniates. Class VI. Antimonial Ores Various Methods for Antimonial Ores. The methods usually adopted for assaying ores of this class may be divided as follows: (a) Fusion with niter. (6) Preliminary oxidation with niter or other oxidizer, fol- lowed by fusion with a reducing flux. (c) Preliminary acid treatment, to eliminate as much as pos- sible of the antimony, followed by scorification or fusion of the residue. (a) Fusion with Niter. The details of this method have already been given under "Pyritic Ores" (Class III, d). E. A. Smith (toe. cit., p. 334) gives the following flux for an ore containing 75 per cent, of stibnite (here adapted for assay on 2 assay tons) : Ore 2 assay tons Carbonate of soda 50 grams Borax 12 " Red lead 120 " Niter 36-48 " Salt cover. A large crucible must be used, raising the temperature very gradually at first. The slag is very fluid. " When effervescence has ceased and the contents of the crucible have become thoroughly liquid and tranquil, it is advisable to lift the crucible out of the fire and subject it to a rotary motion, in order to thoroughly mix the liquid contents. If this causes effervescence it indicates that the action of the niter is not complete, and the crucible must be returned to the fire and the fusion continued " (Smith) . (6) Preliminary Oxidation with Niter. This method requires some previous knowledge of the nature of the ore. The charge of ore is placed in a crucible with niter equal to twice the weight * E. A. Smith, Ibid., p. 332, SPECIAL METHODS OF ASSAY 367 of the pyritic material contained in the ore, together with soda and borax, say for 2 assay tons of ore: soda 50 grams, borax 10 to 15 grams, niter (as required) up to 80 or 90 grams. When the action appears to be complete and the fusion is tranquil, the crucible is withdrawn from the fire and allowed to cool until the slag begins to thicken. The following mixture is then added, and fused for 10 to 15 minutes: Borax 10 grams Red lead 40 " Charcoal 2 " Several variations of this method have been used, and a some- what similar method has been proposed by K. Sander for assaying carbonaceous matter containing lead, in which the preliminary oxidation is made with a mixture of niter and sodium peroxide. (See below.) (c) Preliminary Acid Treatment. The following combined wet and dry method depends on the fact that sulphide of antimony is soluble in strong hydrochloric acid. The native oxides are, however, only partially soluble. "The weighed quantity of ore (30 to 50 grams) is treated with concentrated hydrochloric acid and heated until decomposition is complete; the solution is diluted with a small quantity of water containing a little tartaric acid to prevent precipitation of antimony as oxychloride. The insoluble residue is allowed to settle, and when the solution is clear as much of the liquid as possible is poured off through a filter, without disturbing the residue. The latter is finally transferred to the filter, allowed to drain, and dried. The filter-paper is burnt, and the ash, with the insoluble residue, is scorified with ten times its weight of grain-lead and a borax cover, or fused with the follow- ing flux:" E. A. Smith (loc. cit.). Soda carbonate and borax together 30 grams Red lead 30 " Charcoal 1.5 " Class VII. Cupriferous Ores Methods of Assaying. The methods usually applied for this class of ore are: (a) Preliminary acid treatment to remove the bulk of the copper, followed by fusion or scorification of the residue. 368 THE CYANIDE HANDBOOK (b) Fusion with a large excess of lead oxide, using charcoal or niter according to the nature of the ore. (a) Assay with Preliminary Acid Treatment. Dr. J. Percy 1 gives the following method (grains are here expressed in the nearest equivalent in grams) : "Six or seven grams of ore are heated with nitric acid, or a mixture of nitric and sulphuric acids. The liquid is then diluted, a little hydrochloric acid or salt added to precipitate silver, after which the solution is filtered. The filtrate is clear; contains most of the copper, and may be rejected. The residue is dried and fused with Soda and borax 20 grams Red lead 20 " Charcoal 1 " for 10 to 15 minutes." In a modification of this method given by E. A. Smith (loc. cit.) , 30 to 50 grams of the ore are treated with fuming nitric acid till decomposed, heated to expel nitrous acid, then diluted and mixed with a little HC1 or salt to precipitate silver. The ore is then stirred, settled, and decanted through a filter. The dried residue, with the filter-paper, is fluxed with Soda and borax (together) 30 grams Red lead 30 " Charcoal 1.5 " or it may be scorified with ten times its weight of lead. The writer's experience of this method is that the filtration is very tedious unless a long time be allowed for settlement, as with most ores a considerable portion is crushed fine enough to produce slimes, which soon choke the filter. Moreover, the method for precipitating the silver is certainly faulty, for if insufficient HC1 or salt be added, the silver will not be completely precipi- tated, whereas an excess of either, together with the nitric acid present, would dissolve gold. A better plan would be to filter without addition of chlorides and precipitate or titrate the silver separately in the filtrate; or make an entirely independent assay of it by scorification or some other method. A similar method, due to C. Whitehead, 2 in which the silver 1 " Metallurgy: Silver and Gold," Part I, p. 115. 2 See R. W. Lodge in Trans. A. I. M. E., XXXIV, p. 432. SPECIAL METHODS OF ASSAY 369 in the filtrate is precipitated by potassium bromide, was found by Lodge to give low results. (b) Direct Fusion with Excess of Litharge. In all cases where cupriferous ores are fused without preliminary acid treatment, a large quantity of lead oxide must be used, and the resulting large lead button reduced by scorification before cupeling. When the buttons are hard and have a coppery appearance, they should be still further scorified, with addition of more lead. According to Percy (loc. cit.) one part by weight of copper requires about 16 parts of lead for cupellation, and although a smaller quantity of lead will pass the copper into the cupel, yet in that case some considerable amount of gold and silver is also absorbed by the cupel. The three following fluxes (adapted from Percy) are given by E. A. Smith (loc. cit.) for ores containing various amounts of copper (the weights in grams) : No. 1 No. 2 No. 3 Ore 10-50 10-50 1 assay ton Soda Carbonate .... 20-30 120 Borax 15-20 Silica 10-25 Red lead 200-300 100 150 Charcoal 1 5-3.5 3.5 Niter 18 No. 1 is for ores containing only moderate amounts of copper; No. 2 for ores after roasting, No. 3 for gray antimonial-copper ore containing say 30 per cent. Cu and 150 oz. Ag per ton. The quantities of charcoal and niter would of course be varied according to circumstances. Class VIII. Telluride Ores. Points to be Observed in Assaying Telluride Ores. An excel- lent article on the assaying of gold telluride ores, by W. F. Hille- brand and E. T. Allen, will be found in Bulletin No. 253 of the United States Geological Survey (1905). The essential points in assaying these ores appear to be: (a) Very fine grinding of the sample, say to 150- or 200-mesh; (b) considerable excess of lead 370 THE CYANIDE HANDBOOK oxide; (c) avoidance of roasting, scorification, or any oxidation process, either for the ore or the lead buttons; hence the button produced should not be larger than can be cupeled direct. Flux for Cripple Creek Tellurides. Hillebrand and Allen found the following flux to give very satisfactory results on Cripple Creek telluride ores, using a Battersea " F " crucible for the fusion: Ore 1 assay ton Bicarbonate of soda 30 grams Carbonate of soda 20 Fused borax 10 " Litharge 120-180 grams Salt cover. This gave lead buttons of about 25 grams weight, without addition of iron or niter. The losses in the slag were very slight; those by absorption in the cupels were more considerable and were corrected by a subsequent assay of the cupels. Precautions. With ores of this class great care is necessary in moderating the heat in the early stages of the fusion, and with an ore not previously tested it would always be as well to remelt the slag with fresh flux. In the case of rich tellurides it is better to reduce the charge of ore rather than increase the litharge, so as to avoid producing an unduly large lead button. The losses in oxidizing telluride ores or lead buttons containing tellurium appear to be due to the volatilization of tellurium trioxide, which carries off gold mechanically. Smith (loc. cit., p. 344) states that in presence of lead oxide tellurium is readily converted into telluric acid, which combines with the lead oxide when the latter is in excess: Te + 4pbo = 3pb + PbTeO4 but if the contrary be the case, the excess of telluric acid is volatil- ized and telluride of lead is produced: 2Te + 3PbO = TePb 3 + TeO 3 . Fluxes for Telluride Ores. Smith gives the following fluxes (adapted) : u/. Ore . No. 1 (for ordinary assay) 2 assay tons No. 2 (for very rich ores) 4- assay ton Carbonate of soda . Borax . 60 grams 24 " 50 grams u Siliceous sand Red lead Charcoal . 12-36 " .50-200 " 1-2 " 5-10 " 100-150 " 1 " Salt cover. SPECIAL METHODS OF ASSAY 371 The red lead is varied according to the tellurium present, and charcoal according to the reducing power of the ore. The char- coal may be omitted in some cases. Combined Wet and Dry Assay. The following method is given by Hillebrand and Allen (loc. cit.). "Fifty-gram portions were mixed with water in large porcelain basins, and strong nitric acid was added by degrees, with constant stirring, till the action had nearly ceased; then bromine water was added to oxidize the gold and sulphur and precipitate the silver, no arti- ficial heat being used." After thorough settling, the residue was filtered and washed repeatedly with water containing Br and HC1, and then washed two or three times with pure water. The combined filtrates were evaporated in the original basins to near dryness, then treated with HC1 and covered for a time. When action ceased the covers were removed and evaporation continued, HC1 being added two or three times more at intervals. The residues were now digested with dilute HC1 and a little bromine water to insure solution of the gold. Considerable calcium sulphate remained undissolved, which was filtered off and found to be practically free from gold. The filtrate was precipitated with ferrous sulphate, allowed to stand for 24 hours, and the precipitated gold collected on a filter, washed, dried, weighed, cupeled, and again weighed (using a check of proof gold for cor- rection). The gold thus thrown down was found to be free from tellurium. The residue, containing the silver and a small quantity of gold not extracted by the washing, was dried and assayed with the following flux, yielding a lead button of 20 grams: Carbonate of soda 30 grams Borax 10 " Litharge 60 " Argol 2 " Salt cover. (C) ASSAY OF CERTAIN REFRACTORY MATERIALS Among the products which the cyanide chemist may occa- sionally be called upon to assay may be mentioned zinc-gold precipitate (with or without roasting or acid treatment), slags from fusion of zinc precipitate, etc., slags and cupels from the assay of different classes of ore (when exact results are required 372 THE CYANIDE HANDBOOK on rich material), old crucibles, cupel and furnace bottoms, bat- tery screening, scalings from plates, prussian blue, ferric hydrate and other deposits from tanks or precipitation boxes, residues from smelting retorts, sweep from smelting room, old sacking and filter-cloths, and other matters of very varied description. Of course, no general rule can be given for the treatment of such material, but a few remarks may be of use: Where the material does not admit of fine subdivision no reli- able assay can be made, unless a very large sample be taken and the whole of it weighed and subjected to some process for redu- cing the bulk without loss of values. Thus, samples consisting largely of metal may often be treated by dissolving in some suitable acid, depending on the nature of the material; an assay is then made by scorification or otherwise on the dried residue insoluble in that acid; the residue, after acid treatment, is of course weighed again, unless the whole of it can be dealt with in one operation for extraction of its values. In other cases, where a large part of the sample is inflammable, the whole may be weighed, and then roasted or burnt with suitable precautions and an assay made on the ash. Samples Containing Metallics. Material of this class fre- quently contains metallic particles, such as shots or flakes of bullion, of totally different and usually much higher assay value than the remainder. In all such cases the entire sample should be first weighed. It is then crushed, and as much as possible passed through a sufficiently fine sieve. The portion which passes the sieve, and the metallic and other particles remaining on the sieve, are then weighed and assayed separately. Formula for Assays with Metallics. The following formula may be used in the ordinary case, where the whole of the metallic portion is treated to recover its entire gold or silver contents, and an assay made in the usual way on the fine portion. Let M = total weight of metallics in grams. F = total weight of fines in grams. m = weight in mg. of the metal sought (gold or silver) contained inM. a = assay of metallics (in dwt. per ton of 2000 Ib.) b = assay of fines (in dwt. per ton of 2000 Ib.) x = assay of entire sample (in dwt. per ton of 2000 Ib.) 12 Fb + 7000m 12 (M + F) M + F SPECIAL METHODS OF ASSAY 373 In the case where an assay is made, on a part only of each product, ^ Ma + Fb Some of the substances above mentioned are similar in com- position to one or other of the classes of ore previously discussed, and may be assayed by the methods of crucible fusion already given, but scorification is more generally applicable. Assay of Slags. Slags (from melting of zinc-gold precipitate, etc.) usually require a considerable amount of litharge and re- ducer, so as to give a large lead button. Occasionally they are mixed with so much carbonaceous matter that niter must be used instead of reducing agent. Roasting cannot be used for this class of material, as it would fuse and attack the silica of the dish, forming a crust which would not afterwards be separated without loss. Slags from Assay Fusions frequently have to be remelted with fresh flux to recover the values passing into them in the first fusion. It is usually sufficient to powder the slag coarsely and return it to the original crucible, together with about 50 grams of the ordinary assay flux. This is fused for 10 to 15 minutes, and the small resulting lead button cupeled with the original button, or preferably on a separate cupel, so that the amount recovered from the slag may be noted. The following fluxes may be mentioned, given by different authorities as suitable for re- melting the assay slags from various kinds of material: No. 1. E. A. Smith (loc. cit., p. 322): Soda carbonate 10 grams Litharge 30 " Charcoal 1.5 " Borax added if mixture does not fuse quickly. No. 2. Adapted from Percy, loc. cit. (fuse for 15 to 20 minutes) Soda carbonate 2-3.5 grams Red lead 20-33 Charcoal 1-2 " No. 3. For rich silver-lead ore cerussite (Miller and Ful- ton 1 ). Slag is first remelted with litharge, 1 assay ton, argol, 2 grams, and then the large lead button resulting is scorified. The slag from this scorification is further run down with i" Sch. Mines Quart.," XVII, p. 160. 374 THE CYANIDE HANDBOOK Carbonate of soda 1 assay ton Silica " Litharge 1 " " Argol 2 grams No. 4. For remelting slags from assay of telluride ores (Hille- brand and Allen, loc. cit.) : Litharge 1 assay ton Argol 2 grams Salt cover. Assay of Bone-ash Cupels. It is usually only necessary to test the portion of the cupel which has been stained by absorbed litharge. The remainder is scraped off and rejected, and the portion for assay finely powdered (at least to 80-mesh). T. Kirke Rose 1 gives the following as a suitable flux: Ground cupels 100 parts Carbonate of soda 100 " Borax 50 " Sand 75 " Fluor-spar 75 " Litharge 50 " Charcoal 4 " Percy (loc. cit.) gives: Carbonate of soda 7-17 grams Borax 7-17 " Fluor-spar 7-17 " Charcoal, 4 to 5 parts for every part of litharge in the cupel ; if the latter is insufficient, add red lead, 20 grams, charcoal, 1 gram. Instead of borax, glass free from lead may be used. The two following fluxes also are given by (1) Hillebrand and Allen, and (2) Miller and Fulton: No. 1 No. 2 Carbonate of soda 30 grams 30 grams Borax 45 " 30 " Litharge 60 " 60 " Argol 2 " 2 " Cover of salt. For cupels which have been used in the assay of zinc-gold precipitate the following are given by (1) R. W. Lodge (Trans. A. I. M. E., XXXIV, p. 432), and (2) W. Magenau ("Min. and Sci. Press," LXXX, p. 464). The quantities are given in grams: i ".Metallurgy of Gold," 4th edition, p. 481. SPECIAL METHODS OF ASSAY 375 No. 1 No. 2 Cupels 30-40 Carbonate of soda 15 40 Borax 10 10 Silica - 10 Glass 15 Litharge 60 20 Argol 2 Flour - 1 Cover of - borax Graphite Crucibles. A method for assaying graphite cru- cibles by scorification has already been given in Section III. The following methods of crucible fusion may here be noted. R. Dures, 1 after sifting out all shots, etc., from the finely- ground material, fluxes the remainder thus: Finely ground crucible 5 grams Carbonate of soda 25 " Borax 15 " Silica 10 " Litharge 85 " Niter 7 " A. F. Crosse 2 gives a method of assaying this material by preliminary oxidation with manganese peroxide, taking Powdered crucibles 10 grams Manganese dioxide 35 " After heating this mixture to bright redness in an " H " cru- cible, the temperature is reduced, and the following flux added : Potassium carbonate 50 grams Borax 25 " Salt 25 " Litharge 50 " Silica 20 " Flour 2 " In some cases the manganese dioxide may be omitted and the material given a preliminary roast with silica, using Powdered crucibles : 1 assay ton Quartz 2 assay tons Other assayers, however, obtained very low results by Crosse's method, and Dr. J. Loevy 3 recommends scorifying 1 " Journ. Chem., Met. and Min. Soc. of South Africa," II, p. 577. " Proc. Chem., Met. and Min. Soc. of South Africa," II, p. 124. " Proc. Chem., Met. and Min. Soc. of South Africa," II, p. 205. 376 THE CYANIDE HANDBOOK Powdered crucibles 5 grams (Lead 60 " I or Litharge 70 " with a small quantity of borax or glass. Battery Scrapings (and other coppery material of similar nature). R. Dures (loc. cit.} prefers scorification to nitric acid treatment for this class of material, on the ground that the nitrous acid always present in the commercial acid causes loss of gold. (See Section II, b.) Williams and Crosse, however, advocate acid treatment, and the latter adds lead acetate and sulphuric acid to form a precipitate of lead sulphate, which helps to carry down the gold. Zinc-Box Precipitate. No. 1. The following scorification charge is given by R. W. Lodge l : Precipitate 0.05 assay ton Grain lead 65 grams Borax glass 10 " Of the 65 grams of lead, 35 grams are to be mixed with the charge and the remaining 30 grams are to be used as a cover. Use a 3-in. to 4-in. scorifier. The slag from this to be remelted with Soda carbonate 15 grams Glass 10 " Litharge 30 " Argol 2 " and the cupels assayed as described above. No. 2. C. H. Fulton and C. H. Crawford 2 give the following: Precipitate 0.05 assay ton Grain lead 70 grams Litharge 15 " ) , , , -n t it t added as cover Borax 1 ) But these writers prefer a "wet and dry" method, as follows: No. 3. One-tenth assay ton of the precipitate is boiled with a mixture of 20 c.c. cone. H 2 SO 4 and 60 c.c. water; then 75 cc. normal NaCl and 20 c.c. lead acetate solution (strength not given) are added. The precipitate is settled for an hour, filtered, washed, and dried, care being taken to wash into the point of the filter. The paper is then burnt at a low heat and the whole residue i Trans. A. I. M. E , XXXIV, p. 432 (October, 1903). a" Sch. Mines Quart.," XXII, p. 153 (1901). SPECIAL METHODS OF ASSAY 377 scorified with 30 to 40 grams of lead. This gives lower gold results but more silver than the following: No. 4. W. Magenau's l method for crucible fusion: Zinc precipitate ^ ff assay ton Dry soda carbonate 5 grams Borax glass 2 " Silica 5 " Litharge 70 " Flour 1 " Salt cover. Carbonaceous Residues (left in retorts after distillation of Zinc). K. Sander 2 gives a method for assaying the residue remaining \n the muffles after zinc distillation. This residue contains lead and a large amount of carbon, and offers difficulties in the ordi- nary scorification and crucible assays. He gives a preliminary oxidation as follows: The mixture Powdered residue 20 grams Potassium nitrate 40 " Sodium peroxide 10 " is mixed with an iron spatula and gradually introduced into a red-hot iron crucible. After the violent action has ceased, the following flux is added: Soda 14 parts Calcined borax 8 " Litharge 10 " Argol 2 " Pour when tranquil. The results quoted, however, are in general lower than those obtained by scorification. i"Min. and Sci. Press," LXXX, p. 464 (1900). a" Zeit. fur Angew. Chem.," XV, p. 32 (1902). SECTION IV ASSAY OF GOLD AND SILVER BULLION (A) FIRE ASSAY METHOD Sampling: Dip-samples. The generally accepted opinion is that the most accurate method of sampling bullion is by means of what is known as a "dip-sample." This is taken from the crucible or smelting-furnace while the metal is in a molten con- dition, immediately before pouring. The bath of metal is first well stirred with an iron bar (rabbling-iron) somewhat flattened at the end, and a small portion of the metal is taken out by means of a plumbago spoon, held by a pair of tongs. The dipper used is often a section cut from an old graphite crucible. When the dip-sample is poured there should be a thin layer of slag over the molten metal to prevent oxidation. The dip-sample may be either poured into a small mold and subsequently flattened and cut into pieces, or it may be granulated by pouring into water. Drill-Samples. Samples are also very frequently taken by drilling one or more holes in the bars after casting. These " drill- samples " are usually taken by means of a small vertical machine drill, which is caused to revolve with considerable speed, cutting out fine turnings of the metal. A common plan is to take one drilling near a corner of the bar and the other at the middle of the opposite side, each hole passing half-way through the bar. In bars containing much base metal, there is often a marked difference in the composition of the central and superficial por- tions, owing to the phenomenon of liquation which occurs during the cooling of the metal. Cut-Samples. Samples are also sometimes taken by chipping fragments from the corners or edges of bars, but this procedure can only give correct results when no liquation has occurred. In general it is not to be recommended. The results obtained by Stockhausen 1 showed that drill-samples gave a higher, and cut- i " Proc. Chem., Met. and Min. Soc. of South Africa," II, pp. 46-48. 378 ASSAY OF GOLD AND SILVER BULLION 379 samples from the edges and corners a lower, result than dip- samples from the same bars, and also that the cut-samples ex- hibit much greater fluctuations than either. Preparation of Samples. The samples must be carefully examined before weighing, and any particles of slag or other extraneous matter which may have been accidentally mixed with them are removed. Occasionally it may be necessary to cut them up finer; this may be done by hammering the fragments and cutting with a clean pair of scissors; when the material is brittle, as with some qualities of cyanide bullion, it may be further reduced by grinding in an agate mortar. Weighing-in. The sample is next spread out on a clean sheet of paper and the charge weighed. This charge is usually 500 mg., and the bullion taken may be either adjusted to this exact amount, by clipping or filing the fragments, or an approximate amount may be taken and its exact weight determined and recorded. With a good assay balance the correct weight may be determined within 0.01 mg., but this degree of accuracy is commonly not required; for ordinary purposes it is sufficient to weigh to 0.05 mg. Addition of Lead. When the portion of bullion for assay has been weighed it is at once transferred by means of a fine brush to a small capsule of lead-foil. The weight of lead to be used depends largely on the nature of the impurities contained in the bullion, as does also the question whether scorification should be employed or not. In an ordinary case about four times the weight of the bullion taken (i.e., about 2 grams of lead-foil) will suffice. With certain kinds of base bullion, particularly such as contain arsenic, antimony, selenium or tellurium, a much larger amount may be necessary, say 15 grams. Additions of Copper and Silver. When the bullion contains no copper, it is advisable to add a small amount, say about 20 to 30 mg. in each assay, which prevents " spitting " after the cupella- tion. When the amount of silver is insufficient for parting the gold (i.e., less than about two and a quarter times the weight of gold present), it is necessary to add a sufficient quantity of silver to the assay to make up this difference. When the bullion is to be assayed for both silver and gold, the most satisfactory method is to make these determinations on separate portions. In many cases, however, sufficiently exact results are obtained by re- cupeling the bead of mixed metal, with the addition of the requisite 380 THE CYANIDE HANDBOOK amount of fine silver and sufficient lead-foil. It is necessary to ascertain that the lead-foil used is free from gold, and its silver contents, if any, must be determined and allowed for. Checks. With every batch of bullion assays one or more checks are added. These are made by weighing out exact quanti- ties of pure gold and silver corresponding approximately to the amounts expected to occur in the actual bullion assays. When these are unknown, the approximate composition of the bullion is determined by a preliminary trial. The gold is first adjusted by cutting pieces from a thin sheet of proof gold, to within 0.5 mg. of the theoretical amount; the weight of this quantity is then determined as accurately as possible, say to 0.01 mg. It is then transferred to a small lead capsule (a sheet of lead-foil coiled into a small conical pocket answers the purpose). The amount of silver judged to be present in 500 mg. of the bullion is then weighed out, together with an additional quantity to allow for volatiliza- tion and absorption in the muffle; this extra amount depends upon the position the assays are to occupy in cupellation and the tem- perature at which this operation is to be carried out. The silver is adjusted to within 2 mg. of the amount theoretically required, the exact weight being then determined to within 0.02 mg. It is then added to the capsule containing the gold. It is usual to add also a small quantity of copper, and it is desirable, at least occasionally, to ascertain the effect of adding other ingredients corresponding to the amounts present in the bars to be assayed. In an ordinary case 15 to 20 mg. of copper may be added. Losses of Silver in Cupellation. A series of tests were made by the writer on known weights of gold and silver, approximately 400 rng. Ag to 60 mg. Au, with about 15 grams lead-foil in each case, cupeling in an ordinary fire-clay muffle, size " J," with char- coal as fuel, and using 4 cross-rows of 3 cupels each. The extra silver required was shown by these tests to be approximately as follows : Cupels used Size Morganite. No. 5 Loss of Silver Mg. Bone-ash. No. 5 Loss of Silver Mg. First row (at back) 6-8 11-14 Second row 5-7 Q 10 Third row 4-6 7 10 Fourth row (in front) 3-5 5-S ASSAY OF GOLD AND SILVER BULLION 381 (See remarks on this subject in Section II, a.) Position of Checks. The lead packets containing the weighed bullion or check pieces are rolled up and placed in their proper positions in a suitable tray. Generally, two checks with each muffle charge will be sufficient, but where, as in the experiment quoted above, the differences of temperature in different parts of the muffle are considerable, it is as well to use a check with each row, the two remaining places in the row being occupied by duplicates of the same assay sample. In order to equalize the effects of any possible differences of temperature in various parts of the muffle, the position of the checks may be altered with each batch of assays, as shown in the following arrangements, where the numbers 1, 2, etc., represent the bar assays, and the letters a, b, c, d the checks: No. 1 No. 2 No. 3 No. 4 all lal lla lal 262 622 262 226 33c 3 c 3 c33 3c3 4 d 4 4 4 d 4 d 4 d 4 4 In an ordinary muffle-furnace it will be found that the tem- perature is lower and the losses of silver therefore less in the cen- tral line of the muffle than at the sides, and the back rows are also considerably hotter than the front. Cupellation. The cupels are heated for some time in the muffle before introducing the assay pieces. Claudet 1 recommends baking in the furnace for about an hour, but in most cases a shorter time would seem to be sufficient. The safest plan is to introduce a small piece of pure lead into each, when the cupels appear to be hot enough. When this has melted and begins to cupel freely the assay pieces and checks may be transferred by means of cupel-tongs from the tray to the corresponding cupels in the muffle, each packet being dropped carefully into its proper cupel, where it is at once absorbed in the bath of molten lead. The furnace, where coke or charcoal is used, should be well filled up with fuel, and a moderate red heat maintained throughout the operation. Towards the end the door should be closed as a somewhat higher temperature is required. To secure greater uniformity of temperature, a block of red-hot coke or charcoal i Inst. Min. and Met. Bull No. 27, p. 16 (Dec. 13, 1906). 382 THE CYANIDE HANDBOOK is sometimes placed at the mouth of the muffle. Another plan is to surround the assay pieces by rows of empty cupels, so that the actual cupellation is carried on only in the central part of the muffle. The time occupied in cupellation will vary from ten minutes to half an hour or more, according to the weight of lead and the temperature at which the operation is conducted. When approaching the finish, a play of iridescent colors is noticed on the surface of the beads; these are interference colors produced by the very thin transparent film of litharge surrounding the beads. The cupels are then withdrawn in their proper order and placed for a few minutes at the mouth of the muffle until the beads have solidified. This is accompanied by a sudden flash of light. If the cooling has been too sudden, this process may be accom- panied by violent projection of particles of silver, by " sprouting" or "vegetating" (i.e., the formation of irregular processes pro- jecting from the surface of the bead). These effects are generally ascribed to the sudden ejection of occluded oxygen absorbed by the molten silver during the cupellation. When the proper con- ditions are carefully attended to, this will not occur, and the beads when withdrawn from the muffle will show a perfectly smooth, bright surface. The sizes of cupels from No. 3 to No. 5 are most commonly used for bullion work. Various materials are used, several sub- stitutes for bone-ash having been extensively adopted of late years. Scarification. Where the bullion is very base, and especially when it contains ingredients which would be liable to crack or corrode the cupels, a preliminary operation is resorted to. In- stead of transferring the lead packets direct to the cupels, they are placed in small previously heated scorifiers (l\ in. to 1J in. diameter), with the addition of a very small quantity (say 0.1 gram) of borax. Oxidation of the base metals then takes place, and the molten oxides form fusible silicates with the material of the scorifier itself, producing a slag which floats on the surface of the molten lead. The details of the process have been already described in Section III (a). When the operation is complete, the scorifiers are taken out and their contents poured into small molds previously warmed. The clean lead button is detached when sufficiently cool, and cupeled in the ordinary way. The slag should be uniform; when chiefly consisting of lead silicate, it ASSAY OF GOLD AND SILVER BULLION 383 has a yellowish-brown color, but may be stained dark brown by iron, or green by copper, manganese, etc. Cleaning and Flattening the Beads. The beads are detached from the cupels by means of pliers and squeezed with some force to loosen any adhering matter. A convenient method of clean- ing the bead is to hold it edgewise with a small pair of steel forceps, striking it once or twice on the edge with a smooth-faced hammer on a smooth, clean anvil. The rough under surface, where the bead was in contact with the cupel, may then be easily cleaned todth a wire scratch-brush. The bead is then flattened by one or two blows of the hammer. Small beads containing much silver may quite well be hammered out to about 1 c.m. diameter. The hammer and anvil for this purpose should be kept perfectly clean and polished, and used exclusively for hammering assay beads. Weighing Fine Metal. The beads are now weighed, this weight being recorded as "fine metal"; if the operation has been properly conducted, this is the sum of the gold and silver, less whatever losses have taken place by volatilization and absorption in the cupels during cupellation. It is generally sufficient to weigh to 0.02 mg. The variation in the adjustment of the balance may be corrected as described in Section II (6), or the method of " substitution " may be used (in which the bead is first placed in the right-hand pan of the balance, and then counterpoised by pieces of metal or rough weights in the left-hand pan) . The bullion bead is then removed and accurate weights added on the right until equilibrium' is obtained, the final adjustment, of course, being made with the rider. Annealing and Rolling. The beads are generally annealed by heating to dull redness on fire-clay tiles, or in a small spirit or gas flame, and rolled into strips about 3 in. in length (fillets) by passing between smooth steel rolls. The most convenient form of rolls have a single adjusting screw, working in such a way that the rolls always remain exactly parallel; when there are two screws, this is difficult to secure. The rolls are first loosened so as to allow the metal to pass through without much force; then they are gradually tightened, so that the strip becomes some- what longer and thinner each time it is passed through. The edges should be smooth and show no signs of frilling or cracks. After rolling, the fillets are again annealed and loosely coiled, taking care to leave the rougher side, corresponding to the bottom 384 THE CYANIDE HANDBOOK of the bead on the cupel, on the outer side of the roll, so as to be more easily attacked by the acid. Beads containing a large per- centage of silver (say 5 to 7 times the weight of the gold) may quite well be parted without rolling out, provided they are well flattened by hammering to at least 1 cm. diameter. Parting. This operation is usually carried out in special glass vessels (parting-flasks) having a somewhat lengthened bulb and a long narrow neck. The method of using these is described in Section II (6). Equally good results may, with care, be ob- tained by using porcelain crucibles of somewhat large size (say 4J cm. diameter by 3 cm. deep). Where large numbers of bullion assays are made at a time, as in mints, etc., it is customary to use a platinum parting apparatus, consisting of a tray with a number of perforated recesses ("thimbles"), in which the rolled assay beads (" cornets ") are placed. The whole tray is immersed in a platinum vessel containing nitric acid, so that all the cornets receive uniform treatment. Arthur C. Claudet l gives the following description of parting with the platinum apparatus: " The cornets are placed in a platinum tray, capable of holding sixty comets, in their respective thimbles. Twenty-five ounces of nitric acid, sp. gr. 1 160, are put into No. 1 platinum boiler, and the tray of cornets immersed in it. The boiling is continued till no more nitrous fumes are observed. The tray is then taken out and drained from the acid liquor, then washed in distilled water and immersed in No. 2 boiler, containing 25 oz. nitric acid, sp. gr. 1250; boiling is continued for 20 minutes; the tray is then re- moved from the boiler and the acid liquor drained off; the tray is immersed, without washing, in No. 3 boiler, containing 25 oz. nitric acid, sp. gr. 1250, and boiling continued for 20 minutes. The tray is then removed, drained, washed twice with distilled water, drained again, and dried quietly over a gas-burner covered with wire gauze. The tray is placed in the muffle-furnace on a suitable revolving platform to anneal, and when the cornets are thoroughly red-hot, is withdrawn. The cornets are then of a bright gold color." In the office of a mine or cyanide plant, where such expensive apparatus is commonly not available, the following method will generally be found to work satisfactorily. The flattened beads or >I. M. M., Bull No. 27, Dec. 13, 1906, p. 17. ASSAY OF GOLD AND SILVER BULLION 385 cornets are placed in porcelain crucibles of convenient size, sup- ported on a perforated metal plate over a uniform source of heat, such as a large oil-stove, and 15 c.c. of "weak acid" poured over each. The weak acid is prepared by diluting 1 part of pure nitric acid (sp. gr. 1.42) with 7 parts of water (i.e., 12.5 per cent, by volume of concentrated acid). After the cornets have been well boiled, the lamp is turned down, the acid being kept hot, but not actually boiling, for say 15 minutes, after which the liquid is again boiled till quite colorless, and until the bubbles given off no longer show the slightest appearance of brown fumes. If on cooling for a moment a rapid evolution of small bubbles still takes place, the boiling is further continued, but usually about 20 minutes' treat- ment in the weak acid is sufficient. This acid is then drained off and poured away into the silver residue bottle. The cornets are washed once with distilled water, and 15 c.c. of " strong" acid poured on. The strong acid is prepared by mixing 3 volumes of pure nitric acid (sp. gr. 1.42) with 1 volume of water, and thus contains 75 per cent, by volume of concentrated acid. The liquid is raised cautiously to the boiling-point; the lamp is then turned down for say 25 minutes, thus keeping the crucibles hot but not actually boiling, and adding fresh acid if necessary. At the finish the temperature is again raised to the boiling-point. The time allowed for strong acid treatment is usually about 30 minutes. For good results it is necessary to boil thoroughly in both acids. Precautions. Evaporation to dryness must be carefully avoided, especially during the weak-acid treatment, for if the cornets are heated sufficiently at that stage to cause them to acquire a natural gold color, it is afterwards impossible to extract the residual silver by the strong-acid treatment, and the result will be too high. By the treatment above described, beads containing originally 60 to 80 mg. of gold and 350 to 420 mg. of silver usually remain unbroken. When the proportion of silver is much larger they break to pieces and some care is necessary to avoid loss. Too violent boiling must also be carefully guarded against, as this is liable to cause spurting, especially during the strong acid treat- ment, and may result in the cornet suddenly breaking in pieces, in which case fragments are likely to be projected out of the crucible. Solvent Action of Nitric Acid on Gold. Contradictory opinions 386 THE CYANIDE HANDBOOK are expressed by different writers on this subject. (See Section II, b.) Some maintain that neither nitric nor nitrous acids by themselves are capable of dissolving gold. It has been observed, in parting cornets containing a rather high percentage of silver, with nitric acid perfectly free_from chlorides, that on diluting the clear liquid containing the dissolved silver, either with cold acid or water, a dark purplish turbidity was formed. On boiling, the liquid became clear again, but a blackish scum was formed which seems to consist, at least partially, of finely divided gold. Igniting the Cornets. After pouring off the strong acid, the cornets are washed and the inside surface of the crucible carefully rinsed out with distilled water. This operation is repeated at least twice. The crucibles are then replaced on the perforated frame over the lamp and allowed to dry for a few minutes at a moderate heat. They are then ignited by being placed for a minute or so in a muffle at a dull red heat, or by holding the cru- cibles in the flame of a blast lamp until they acquire a bright golden color. Care must be taken not to heat so intensely as to fuse the gold. Weighing Gold Cornets. When cool, the gold cornets are weighed on the fine assay balance, recording the correct weight to 0.01 mg. A good assay balance will weigh to 0.005 mg., but this degree of accuracy is not usually necessary. The cornets must be a good color; if pale, it is a sign that they contain undis- solved silver. The adjustment of the balance must be carefully attended to, and corrected, if necessary, as described under ore assaying. It is advisable to verify the correctness of the adjust- ment after every two or three weighings by allowing the balance to swing with the pans empty, a special rider on the left arm of the beam being used to correct any inequality in the swings. Errors from this source may also be corrected by the method of substitution above noted. Accuracy of Assay Weights. It is necessary to carefully examine the weights used, and to determine their true relative value with reference to the 500 mg. weight used for weighing the bullion-charges, as the sets supplied, even by firms of high stand- ing, are by no means absolutely correct. Milliemes. Some assayers use the " millieme " system, in which the 500 mg. weight is marked 1000, and its subdivisions numbered in milliemes, or half-milligrams. This has the advan- ASSAY OF GOLD AND SILVER BULLION 387 tage of giving the "fineness" on a charge of 0.5 gram of bullion directly, without calculation, but is very confusing if the weights are used for other purposes, as the numbers are apt to be mis- taken for milligrams. The fineness of the bullion is generally reported in parts per 1000. As 500 mg. are taken for each assay, it is merely necessary to multiply the result by two to obtain the fineness. Correction of Results by Checks. A correction is required to be made for losses of gold and silver in cupellation, and for silver retained in the cornets after parting. In the gold assay, the weight actually obtained is the net result of the losses in the first operation and the gains in the second. It is assumed that the fine gold used in the check has undergone the same losses and gains as that from the bar. This is only strictly true when the check is very closely of the same composition as the bullion to be assayed, and has been subjected to .precisely similar treatment. If the check, as is usually the case, shows an increase of weight, the amount of this increase is deducted from the weight of gold obtained from the bars adjoining that particular check in the muffle; if the check has diminished, the amount of such decrease is added to the weight of gold from each bar assay. Amount of Surcharge. According to A. C. Claudet 1 the amount of surcharge in gold bullion assays may vary from 0.7 to 1.0 per millieme (i.e., 0.35 to 0.5 mg. on an assay of 500 mg.) and is always an increase on the amount taken for the check; this amount must therefore be deducted from the weight of the cornet in each bar assay. Professor W. Gowland 2 stated, in the discussion on Claudet's paper, that the surcharge at the Mint of Japan varied during 6 months from 0.5 to 0.7, averaging 0.55. He also remarks that a small surcharge indicates that the cupellation has been carried out at too high a temperature, and that in such a case the assay should be repeated, still more so if there should be no sur- charge or a loss on the checks. These figures, however, all prob- ably refer to bullion consisting chiefly of gold. According to T. W. Wood, 3 the surcharge varies according to the relative proportion of gold and silver, and with bullion consisting chiefly of silver the gold checks usually show a loss. The present writer's experience il. M. M., Bull. No. 27, p. 17. 2 Ibid., Bull. No. 28, p. 29. 3" Proc. Chem.,-Met. and Min. Soc. of South Africa," II, p. 3. 388 THE CYANIDE HANDBOOK with bullion averaging about 800 in silver and 60 to 70 in gold was that the surcharge rarely exceeded 0.4 per 1000 (i.e., 0.2 mg. on 500 mg.). Under normal conditions the silver always shows a loss, so that the correction indicated by the check has to be added to the silver found. It occasionally happens that the front row in the muffle shows an apparent increase in silver. This is due to reten- tion of lead ; in such cases the correction cannot be safely applied and the assays must always be repeated, using a somewhat higher temperature. Examples of Corrections for Bullion Assays. (a) Gold Bullion (e.g., battery gold); 500 mg. taken for assay, in duplicate. Gold: Silver: mg. mg. Taken for check 284.52 207.85 Found.. 284.86 206.42 Correction .34 + 1.43 Gold: Bar assay: Found Corrected 284.93 284.59 284.79 284.45 Silver: Bar assay: Found Corrected 206.37 207.80 206.25 207.68 Gold Silver Mean Fineness 569.04 415.48 (b) Silver Bullion (Cyanide) . Gold: Silver: mg. mg. Taken for check 76.79 397.59 Found 76.88 391.97 Correction -.09 +5.62 Gold: Bar assay: Found Corrected 76.70 76.61 76.68 76.59 ASSAY OF GOLD AND SILVER BULLION Silver: Found Corrected 391.80 397.42 392.76 398.38 Gold Silver Mean fineness 153.2 795.8 (B) VOLUMETRIC METHODS There are two methods in general use for determining the silver in bullion; several others have been proposed, but these two only need be considered here. They are: (1) Gay-Lussac's method; (2) Volhard's method. (1) Gay-Lussac's Method. This method depends on the fact that silver is precipitated by solutions of sodium chloride as a white chloride, insoluble in dilute acid solutions; the only other metals precipitated under the same conditions being lead and mercury. Since the chloride coagulates and settles rapidly on agitating the solution, the finishing point of the reaction may be found by noting the point when fresh additions of salt solution fail to produce any further precipitate. It has been found, however, that silver chloride is slightly soluble in the sodium nitrate produced in the reaction: NaCl + AgNO 3 = NaNO 3 + AgCl. Hence the true finishing point occurs slightly before the point at which fresh NaCl ceases to give a precipitate. In fact, at or near the point of exact correspondence the solution will give a pre- cipitate of AgCl if either salt or silver solution be added; the true end-point may be found by determining (1) the point at which NaCl ceases to give a precipitate, (2) the point at which, after a slight excess of NaCl has been added, a corresponding solution of AgNO 3 ceases to give a precipitate, and taking the mean of the two results. The amount of AgCl dissolved in this way de- pends on the temperature and degree of dilution of the solutions. A simpler method is to determine the finishing point, by means of an accurately prepared salt solution, for different measured quantities of a silver solution of known strength; after a number of such observations have been made, a factor can be deduced which will give the proper correction for any reading. The solu- tions commonly used for assays by this method are: 390 THE CYANIDE HANDBOOK 1. A "standard" salt solution, made by dissolving 5.4202 grams of the purest obtainable dry sodium chloride in distilled water, and making up to a liter; 1 cc. of this solution = .01 gram Ag. The pure salt is best prepared by neutralizing pure sodium carbonate or bicarbonate with hydrochloric acid, evaporating to dryness, and igniting. 2. A salt solution, prepared from the above of exactly one- tenth the strength, known as the "decimal" salt solution; 1 cc. of this solution = .001 gram Ag. 3. A solution of silver nitrate corresponding exactly to the deci- mal salt solution, that is, containing 1 gram of Ag dissolved in pure nitric acid, heated to expel nitrous fumes and diluted to a liter. In making a test, the quantity of bullion taken must be such that it contains very slightly more than 1 gram of silver, this being ascertained by a previous approximate assay. This is introduced into a stoppered bottle and dissolved in nitric acid, heating on the water-bath till completely dissolved, after which the stopper is removed and the nitrous fumes dispelled by blowing through a bent tube. Then 100 cc. of the standard salt solution are added by means of an accurately graduated pipette, so that exactly 1 gram of silver is precipitated. The stopper is replaced and the bottle shaken, being kept in the dark as much as possible, until the silver chloride settles clear. The decimal salt solution is then added, a little at a time, as long as a further drop, after agi- tation and settling, continues to produce a precipitate. A me- chanical device is often used for the agitation of the bottles, so arranged that they are kept dark, since silver chloride is rapidly acted upon and decomposed by exposure to light. With care, this method yields very accurate results, but the necessity of waiting after each addition of the test solution for the settlement of the precipitate renders it very tedious. (2) Volhard's Method. This depends on the fact that silver in nitric acid solution is precipitated by solutions of a soluble thiocyanate (sulphocyanide), and that if a ferric salt be present the red ferric thiocyanate is decomposed and the color destroyed as long as silver is in excess. The reactions are: KCNS + AgNO 3 = KNO 3 + AgCNS. 3KCNS + Fe(NO 3 ) 3 = 3KNO 3 + Fe(CNS) 3 . As soon as all the silver is precipitated, the red color remains permanent. ASSAY OF GOLD AND SILVER BULLION 391 Standard and decimal solutions of potassium or ammonium thiocyanate may be prepared, such that 1 cc. = .01 gram Ag, and 1 cc. = .001 gram Ag, respectively. 1 In this case the solutions cannot conveniently be prepared by direct weighing of the thio- cyanate, on account of the deliquescent nature of the salt, but may be adjusted by comparison with a silver solution of exactly known strength. With this process also there is a difficulty as to the exact finish- ing point, owing to the fact that the red ferric thiocyanate color disappears very slowly and only after continued agitation when only a small excess of silver is present. Various methods of determining the end-point have been suggested, as follows : 1. A somewhat sharper finish is obtained by adding a slight excess of thiocyanate, shaking thoroughly, then cautiously add- ing a dilute silver nitrate solution (say 1 cc. = .0005 gram Ag), drop by drop, until the red color just disappears, deducting the equivalent of AgNO 3 used from the thiocyanate taken. 2. Precipitating the bulk of the silver as chloride by means of a decimal salt solution (as in Gay-Lussac's method), filtering, and titrating the nitrate only with dilute thiocyanate. 3. Adding a slight excess of thiocyanate, estimating the amount of this excess (or rather its equivalent in terms of silver) by colorimetric comparison with measured quantities of solution of known strength necessary to give a tint of equal intensity, and deducting this from the silver equivalent of the total thiocyanate used. The solution must be filtered before making the compari- son. (See E. A. Smith, in I. M. M., Bull. No. 28.) The method may be used in presence of most of the ordinary metals, such as lead, iron, zinc, copper, etc., which may be present in considerable amount without affecting the result. The presence of copper alters the tint of the solution at the finish, and hence interferes with the colorimetric method of determining the end point. The writer finds it advisable to avoid filtration. Where the quantity of silver in the solution is considerable (more, say, than 0.1 gram), it is better to add the thiocyanate in slight excess, shake thoroughly, allow to settle for some time, decant the almost clear liquid (tinged red by the ferric indicator and excess of thio- i It is more convenient to work with weaker solutions, say 1 cc. = .005 gram and 1 cc. - .0005 gram. 392 THE CYANIDE HANDBOOK cyanate) into a separate vessel, adding water, agitating, settling and decanting several times, and titrate the whole liquid poured off by means of the dilute standard silver solution until the red color disappears. PART VIII ANALYTICAL OPERATIONS IN this part of the present volume, some description will be given of the analytical methods required for the examination of such substances as may occasionally demand analysis in connec- tion with cyanide work. The subject is divided into the follow- ing sections: I. Analysis of ores and similar siliceous material, including II. Analysis of bullion and other mainly metallic products. III. Analysis of cyanide solutions after use in ore treatment. IV. Analysis of commercial cyanide (solid). V. Analysis of sundry materials (lime, coal, water) used in the process. SECTION I ANALYSIS OF ORES AND SIMILAR SILICEOUS MATERIAL No attempt can be made in a work of this kind to give minute details of analytical methods, but a general outline will be here presented, with references to standard works, where the reader may obtain fuller information. In this section the analysis will be described of such material as might commonly be expected to contain 50 per cent, or more of silica. The constituents whose determination is followed are those which may be of interest and importance in cyanide work, and for convenience of refer- ence they are arranged alphabetically, though of course the actual order of analysis would be determined by the nature of the sample. PRELIMINARY OPERATIONS The operations of sampling, crushing, and drying are of at least as much importance in analysis as in assaying; in fact, certain operations which are quite allowable in ordinary assay work would not be admissible in an analysis. Crushing. The use of an ordinary iron pestle and mortar might introduce sufficient iron, in the form of metal or oxide, to entirely vitiate the result. The metallic portion may, it is true, be removed with a magnet, but many minerals themselves con- tain magnetic particles, and the magnet does not remove foreign particles of oxide of iron. The crushing can only be carried out safely by means of a well-polished hard-steel hammer and anvil, provided with a guard to retain flying particles, followed by fine grinding of a well-mixed average portion with an agate pestle and mortar. Sifting. Metal sieves are liable to induce contamination, and even sifting through cloth may give rise to organic impu- rities. In some cases, sifting is entirely unnecessary; in others, judgment must be used as to whether the possible contamina- 395 396 THE CYANIDE HANDBOOK tion of the sample is sufficient to necessitate a special determina- tion in an unsifted sample of the elements (iron, copper, zinc) which might arise from the wearing of the screen. Drying. The presence of moisture in samples for analysis is sometimes overlooked. Many siliceous substances, when finely divided, readily absorb moisture from the air, which adds to their weight without imparting any visible dampness to the sample. This so-called "hygroscopic water," if neglected, will cause unaccountable variations in the determination of the other constituents. On this account it is customary to dry the por- tion of the sample to be taken for analysis for some hours at a fixed temperature (generally 100 or 110 C.), weighing at inter- vals in a glass-stoppered bottle, which is of course left open during the drying, until two consecutive weighings agree within sufficiently narrow limits. Another method sometimes employed is to take the crushed mineral in the undried condition, and weigh out, as nearly as possible at the same time, all the portions required for the different determinations, including one for hygro- scopic water. The percentages of the various constituents in the dry sample are then obtained by calculation. The drying is carried out in an air-bath or steam oven at constant tempera- ture. The determination of combined water will be described below. GENERAL OUTLINE OP ANALYSIS After the preliminary operations above described, the next step, in the treatment of siliceous material, is to "open up" the substance by some method which will render the bases soluble, and the whole of the silica insoluble, in dilute acids. Where conditions allow, this is done by fusion in platinum vessels with alkalis or alkaline carbonates; in other cases the sample is evap- orated or boiled with nitric, hydrochloric, or sulphuric acid, according to circumstances. In the filtrate from the insoluble matter, iron and alumina are generally precipitated together by adding excess of ammonia; then, successively, the group con- taining manganese, zinc, nickel, and cobalt by ammonium sul- phide, calcium by ammonium oxalate, and magnesium by an alkaline phosphate. The alkali metals are estimated in a sep- arate portion of the sample. In cases where lead, bismuth, arsenic, antimony, or other heavy metals capable of attacking ANALYSIS OF ORES AND SIMILAR MATERIAL 397 platinum are present, these must be eliminated before making any fusions in platinum vessels. ESTIMATION OF VARIOUS INGREDIENTS ALKALI METALS 1. These are generally estimated by Lawrence Smith's method, as follows: A weighed portion, say 1 gram, of the finely pow- dered ore is mixed in a mortar with an equal weight of ammo- nium chloride, then with eight times its weight of pure calcium carbonate, and transferred to a platinum crucible provided with a funnel-shaped prolongation to condense any alkaline chloride which may be volatilized during ignition. The crucible is heated, gently at first, until ammonium salts are decomposed, then kept at redness for, say, three quarters of an hour. When cool the contents are turned into a dish and the crucible washed into the same dish with water (60 to 80 cc.) and boiled. The residue is then filtered, the filtrate mixed with 1.5 gram ammonium carbonate and concentrated somewhat; then more ammonium carbonate and ammonia are added. The liquid is now again filtered into a weighed platinum dish, evaporated to dryness, and ignited cautiously to below redness. When cool the dish is weighed, the increase in weight giving the combined chlorides of the alkali metals. The residue is then dissolved in water, and the chlorine titrated by standard silver nitrate. Deducting the weight of chlorine so found from that of the mixed chlorides, we have the total weight of alkali metals present. (For separa- tion of sodium and potassium see Potasskfm.) 2. In cases where Lawrence Smith's method is inapplicable, as when the sample contains large quantities of the heavy metals, evaporate to dryness with aqua regia, ignite to render silica insoluble, take up with dilute hydrochloric acid, and precipitate with hydrogen sulphide. Treat the filtrate with ammonia and ammonium sulphide; boil, filter, and wash with dilute ammonium sulphide. Evaporate the filtrate to dryness and ignite to expel all ammonium chloride, etc. Re-dissolve in a little water, add ammonium carbonate and evaporate to dryness on a water- bath. Re-dissolve in warm water and filter into platinum dish. Evaporate on water-bath, ignite gently, and weigh mixed chlo- rides as above. 398 THE CYANIDE HANDBOOK ALUMINIUM This metal is generally precipitated as alumina, together with Fe, Cr, Ti, Zr, and in some cases phosphates and small amounts of Si0 2 , in the filtrate from the silica. After determining the total weight of the ignited precipitate, and the separate amounts of the other constituents, the alumina (A1 2 O 3 ) is found by differ- ence. The precipitation is made either by ammonia or by sodium acetate. 1. The nitrate from silica is heated nearly to boiling, and ammonia added until the liquid becomes slightly alkaline. Suffi- cient ammonium chloride must be present to prevent the precip- itation of magnesia, etc.; this is best secured by acidulating the liquid, if not already strongly acid, with additional hydrochloric acid before adding ammonia. After stirring and allowing to settle, wash by decantation with hot water containing a little ammonia; finally transfer to a filter, allow to drain, and wash back the precipitate into the original dish. Re-dissolve in hot dilute HC1, re-precipitate with ammonia from the nearly boiling liquid, again filter, and add the second filtrate to the first. The precipitate is dried carefully in an air-bath, removed from the papers, which should be burnt separately, ignited over an oxidiz- ing blast flame in a platinum crucible, and weighed as alumina, etc. 2. When much manganese is present, it is better to precip- itate with sodium acetate, as follows: The acid filtrate from the silica is nearly neutralized with caustic soda, taking care, how- ever, not to produce a permanent precipitate. From 2 to 3 grams of sodium acetate are then added, and the liquid boiled, allowed to settle somewhat, filtered and washed slightly. The precipitate is re-dissolved in nitric acid, transferred to the orig- inal vessel, and heated nearly to boiling. It is then precipitated with ammonia, filtered, and washed with a 2 per cent, solution of ammonium nitrate, care being taken to keep the two filtrates separate. By concentrating these filtrates separately, an addi- tional quantity of alumina, etc., is recovered. The first filtrate is then re-filtered, and the second poured on to the same filter, giving a final wash with hot dilute ammonia. The united fil- trate thus obtained serves for the estimation of Mn, etc. The additional precipitate recovered is added to the main bulk, dried ANALYSIS OF ORES AND SIMILAR MATERIAL 399 in the air-bath, and treated as already described in the first method. 3. The nitrate from the silica is oxidized to bring all iron into the ferric state. It is diluted to 100 cc. and 15 cc. of a saturated solution of microcosmic salt (Na NH 4 HPO 4 8H 2 O) is added, followed by ammonia, till a slight permanent precipitate remains. Heat to boiling; acidulate very slightly with HC1, add 25 cc. saturated sodium thiosulphate and 5 cc. glacial acetic acid; boil for 10 minutes. A white precipitate is formed consisting of aluminium phosphate mixed with sulphur. Filter, dry, and ig- nite with the paper; finally ignite at bright-red heat and weigh as A1PCV A1 2 O 3 X 0.5303 - Al. A1PO 4 X 0.4185 = A1 2 O 3 . A1PO 4 X 0.2219 = Al. ANTIMONY In ores containing antimony, this metal may be brought into solution: (1) By treatment with hydrochloric acid and potas- sium chlorate. (2) By fusion with an alkali and sulphur. The mass extracted with water contains the antimony as a soluble sulphantimonite. (3) By evaporation with sulphuric and tar- taric acids, the latter being added to ensure the conversion of antimony into an antimonious compound; (4) In certain ores, such as antimonite, the metal may be dissolved by simply boil- ing with hydrochloric acid. The following precautions are to be noted in manipulating solutions containing antimony: (1) The chloride is very liable to precipitate on dilution, as SbOCL,; this may be prevented by the addition of a little tartaric acid before precipitating the anti- mony, for instance, with H 2 S. (2) Treatment of antimony ores with nitric acid forms antimonic acid, which is soluble with great difficulty. It is therefore preferable to use a mixture of HC1 and KC10 3 . (3) Antimonious chloride, SbCl 3 , is volatile on evaporation; solutions, should therefore be made alkaline before evaporating. Separation. 1. Having obtained the metals in HC1 solu- tion, the liquid is mixed with excess of caustic soda and a little sulphur; a current of H 2 S gas is passed through for some min- i H. T. Waller, Bull. No. 49, I. M. M., Oct. 8, 1908. 400 THE CYANIDE HANDBOOK utes, and the mixture allowed to stand in a warm place for an hour; it is then filtered and washed, the Sb being in the filtrate as sulphantimonite (together with As and Sn if present). Acid- ulate filtrate with HC1, which leaves As 2 S 3 undissolved on boil- ing. Filter, dilute, add a little tartaric acid; re-precipitate with H 2 S. Filter and wash free from chlorides. Determine Sb as below. In case tin also is present, transfer the precipitate to a weighed dish and treat cautiously with fuming nitric acid. Evaporate, ignite, and weigh as xSb 2 O 4 + ySnO 2 . The residue in the dish is then transferred to a flask and digested for an hour with con- centrated tartaric acid solution, on a water-bath, avoiding undue evaporation. Residue is filtered, washed, ignited, and weighed as ySn0 2 . The precipitation of tin by H 2 S may be prevented by adding a sufficient amount of concentrated oxalic -acid solu- tion to the mixed chlorides. 2. The ore is heated in a flask with a mixture of tartaric acid, sulphuric acid, and potassium bisulphate until completely de- composed; all carbon separated is burnt off and the mixture heated till most of the H 2 SO 4 is expelled. Allow to cool and dissolve in dilute HC1 with addition of tartaric acid. Heat nearly to boiling and pass in H 2 S. Filter and wash with H 2 S water. Transfer the precipitate to a beaker and warm with potassium sulphide; pass again through same filter, washing with warm dilute K 2 S. The Sb (with perhaps As and Sn) is in solution; convert into chlorides by repeating the original treatment with sulphuric acid, etc., finally dissolving in strong HC1. Now pass in H 2 S, which, in strongly acid solution, precipitates only the As. Filter, wash with HC1, dilute filtrate with warm water, add oxalic acid if tin is present, and precipitate antimony as Sb 2 S 3 by passing H 2 S. (For details see A. H. Low, "Technical Methods of Ore Analysis," 3d edition, pp. 32-35.) Estimation. 1. The antimony is assumed to have been separated as antimonious chloride, SbCl 3 , free from other metals; add excess of HC1, boil slightly if necessary to expel SO 2 , dilute considerably, and titrate with standard permanganate or bi- chromate : 5 SbCl 3 + 2KMnO 4 + 16HC1 = 5SbCl s + 2KC1 + 2MnCl 2 + 8H 2 O V 1 gram KMnO 4 =1.9 gram Sb; Ice. ^ KMnO 4 = 0.006 gram Sb. ANALYSIS OF ORES AND SIMILAR MATERIAL 401 2. When the antimony has been obtained as antimonic chloride (SbCl 5 ), as in the case where the ore has been treated with HC1 and KC10 3 , the excess of chlorine is boiled off, and the solution cooled, diluted, and mixed with potassium iodide. SbCl 5 + 2KI = SbCl 3 + 2KC1 + I 2 . Ice. TJJJ iodine <= 0.006 gram Sb. After standing a few minutes, the liberated iodine is titrated with standard thiosulphate. The thiosulphate must be standard- ized against a known quantity of antimony, either the pure metal dissolved in HC1 + KC1O 3 or tartar emetic, also heated to con- vert the antimony into the antimonic condition. 3. Antimonious compounds may also be titrated with a standard solution of iodine in potassium iodide, as described under arsenic. 1 Detection. The presence of antimony in an ore may often be detected by heating the extract obtained with hydrochloric acid or aqua regia, in a dish containing a strip of platinum in contact with metallic zinc. A black deposit on the platinum, insoluble in dilute HC1 but soluble in HNO 3 , indicates antimony. ARSENIC Ores containing arsenic, such as mispickel, realgar, orpiment, and some other minerals, are occasionally treated by cyanide; since the presence of arsenic introduces some difficulties in the process, attention may be drawn to the methods of detecting and estimating this element. Dissolving. 1. The methods already described for anti- mony may be applied in general for arsenic. 2. Arsenic may be obtained in the arsenious condition by evaporating the powdered ore with nitric acid, and adding the mixture gradually to a strong solution of alkaline sulphide, warm- ing and filtering, the arsenic being in the filtrate as sulpharsenite. 3. It may also be obtained as sulpharsenite by fusing with sodium carbonate and sulphur. 4. It may be obtained as an alkaline arseniate by digesting the ore with nitric acid, evaporating to dryness, and taking up with an alkali. 1 Sutton, "Volumetric Analysis," 8th edition, p. 160. 402 THE CYANIDE HANDBOOK 5. As arseniate, it may be obtained by fusing with sodium carbonate and niter, and extracting with water. 6. By suspending the finely powdered ore in a caustic alkali solution and passing in chlorine till saturated. 7. By distilling with zinc and sulphuric acid, which produce hydrogen arsenide; this gas when led through fuming nitric acid forms arsenic acid. Separation. 1. Having obtained the arsenic (together with antimony and tin) as a sulpharsenite, the insoluble residue, if any, is collected on a filter and washed with dilute potassium sulphide. The filtrate is acidulated with sulphuric acid and potassium bisulphate, adding tartaric acid if antimony is present. Heat till the free sulphur and most of the acid is expelled, then cool and add strong hydrochloric acid. Warm gently and satu- rate with H 2 S. The arsenic is precipitated as As 2 S 3 , leaving anti- mony and tin in solution. Filter and wash with HC1 (2: I). 1 2. In cases where the arsenic has been converted by prelim- inary treatment into an arsenious compound, it may be separated from other (non-volatile) ingredients by distillation as follows: The substance is placed in a flask connected with a bulbed U- tube or other suitable condenser, containing a little water. Add to the flask, for each gram of material treated, 30 grams calcium chloride, 15 grams ferric chloride, and 30 cc. hydrochloric acid. Distil for 20 to 30 minutes without evaporating to dryness. The arsenic is present in the distillate as arsenious chloride. In a modification of this method, ferrous sulphate is used instead of ferric chloride, and serves to reduce any arsenic compound that may be present. Estimation. 1. Having obtained the arsenic as an arsen- ious compound by distillation as described above, or by dissol- ving the sulphide in ammonium sulphide and evaporating with sulphuric acid and potassium bisulphate, the solution is made slightly alkaline with ammonia, then slightly acid with hydro- chloric acid. When cool, a sufficient excess of sodium bicar- bonate is given, and the mixture titrated with standard iodine and starch indicator. The iodine solution is standardized against known quantities of arsenious acid (As 2 O 3 ) dissolved in hydro- chloric acid. 2 1 For details see A. H. Low, "Technical Methods of Ore Analysis," pp. 33-41. 2 C. and J. Beringer, "Text Book of Assaying," pp. 383-384. ANALYSIS OF ORES AND SIMILAR MATERIAL 403 N . 1 cc. r^ iodine = 0.00375 gram As. = 0.00495 gram As 2 O 3 . 2. Gravimetric Method. When the arsenic has been ob- tained as an arseniate, the mixture is made alkaline with am- monia, and a mixture of magnesium sulphate and ammonium chloride added (1 part of the former to 2 parts of the latter). The precipitate is allowed to settle, filtered, and washed with a minimum quantity of dilute ammonia, then dried very slowly and carefully to expel ammonium salts, ignited, and weighed as magnesium pyroarseniate, Mg 2 As 2 7 1 . Mg 2 As 2 O7 X 0.4828 = As. 3. Volumetric Method, for Arseniates. The solution is made slightly alkaline with ammonia, then mixed with sodium acetate and acetic acid till distinctly acid, heated to boiling, and at once titrated with standard uranium acetate, using ferro- cyanide as external indicator; the appearance of a brown color in the spots of ferrocyanide marks the end of the titration. 1 4. After fusing with sodium carbonate and niter (5 parts of each to 1 part of the ore), dissolve the residue in warm water, filter, and wash with cold water. Acidulate slightly with nitric acid, add silver nitrate, then successively ammonia and nitric acid, till the precipitate just dissolves and the liquid is faintly acid. Add sodium acetate in excess; the arsenic is precipitated as silver arsenate, Ag 3 AsO 4 ; heat to boiling, filter, and wash with cold water till "free from soluble silver salts, re-dissolve precip- itate with dilute nitric acid, and estimate the silver by Volhard's method. (See p. 390.) [Pearce's method.] 2 N 1 cc. thiocyanate = 0.0025 gram As. Detection. 1. The substance is dissolved in hydrochloric acid, or the arsenic obtained as AsCL, as described above. A strip of copper is introduced into the liquid, which is then warmed. Small quantities of arsenic give a gray deposit, and larger amounts a black deposit, on the copper. 2. Marsh's Test. The substance to be examined is intro- duced into a vessel containing zinc and sulphuric acid, from 1 C. and J. Beringer, loc tit., p. 389. 2 For details, see A. H. Low, "Technical Methods of Ore Analysis," 3d edi- tion, p. 42. 404 THE CYANIDE HANDBOOK which a current of hydrogen is being generated. The evolved gases are led through a horizontal tube drawn to a fine point, forming a jet, where the escaping hydrogen may be ignited. By holding a cold porcelain plate in the flame, a black deposit is obtained consisting of reduced metallic arsenic or antimony; the arsenic deposit may be distinguished from that of antimony by its greater solubility in a solution of a hypochlorite. A mirror of arsenic may also be obtained by heating the tube at a certain point; the AsH 3 is decomposed and metallic arsenic depos- ited in the cooler part of the tube. 3. The substance is mixed with dry charcoal and heated in a bulb tube. A sublimate of arsenic is formed, and when much is present an odor resembling garlic is observed. BARIUM When the ore contains the metal as carbonate, it may be extracted by simply boiling with hydrochloric acid. In other cases the ore is treated successively with nitric and sulphuric acids as described under LEAD; barium and strontium remain as sulphates, together with silica, in the portion insoluble in ammo- nium acetate. They may be separated from silica by treating the residue in a platinum crucible with H 2 SO 4 and HF. In the absence of lead, however, it is better to determine barium in a separate portion of the sample; this is treated with hydrofluoric and sulphuric acids, and evaporated repeatedly with sulphuric acid until every trace of fluorine has been expelled. The evap- orated residue is digested for some time with 5 per cent. H 2 S0 4 , boiled if necessary, and filtered. The filtrate contains any other non-volatile metals which may have been present, and may be used for their estimation. The residue, consisting of BaSO 4 and SrS0 4 , with perhaps traces of CaS0 4 , is ignited with sodium carbonate and extracted with water. The residual carbonates are then dissolved in HC1. If strontium be present, the filtrate is made alkaline with ammonia, a slight excess of acetic acid is added, and potassium chromate, which precipitates the barium as BaCrO 4 ; this is filtered, washed with ammonium acetate, ignited gently and weighed. BaCrO 4 X 0.542 = Ba. Strontium remains in the filtrate, and may be precipitated by ANALYSIS OF ORES AND SIMILAR MATERIAL 405 H 2 SO 4 . When no strontium is present, the BaCO 3 formed by fusion with sodium carbonate is at once dissolved in dilute HC1, diluted, heated to boiling, and precipitated with H 2 SO 4 . After settling, it is filtered, washed with hot water, ignited, and weighed as BaSO 4 . BaSO4 x 5885 = Ba If any calcium should be present in this precipitate of BaSO 4 , it may be separated by dissolving the whole in concentrated H 2 SO 4 , and re-precipitating the BaSO 4 by dilution, the CaSO 4 remaining dissolved. BISMUTH This metal is obtained together with copper in precipitating the acid extract of the ore with H 2 S, after removal of lead, etc., with H 2 SO 4 , as described under copper. In order to separate it, the sulphide precipitate is dissolved in nitric acid and evaporated to dryness, taken up with a little sulphuric acid, warmed and filtered. The filtrate is nearly neutralized with ammonia, am- monium carbonate added in excess, boiled and filtered. The precipitate is then re-dissolved in HNO 3 , and re-precipitated from boiling solution by ammonium carbonate, filtered, washed, dried, ignited at low-red heat, and weighed as Bi 2 3 . Bi 2 O 3 X 0.8966 = Bi. CADMIUM This metal is precipitated by H 2 S with copper and bismuth. It remains in solution together with copper when the bismuth is precipitated as carbonate in ammoniacal solution. Cadmium may be detected and estimated by adding cyanide to the blue ammoniacal liquor till it is colorless, then passing in H 2 S, which gives a yellow precipitate, leaving the copper in solution. The precipitate is collected on a weighed filter, washed with sulphu- reted hydrogen water, then extracted with carbon bisulphide to remove free sulphur, dried at 100 C. and weighed as CdS. CdS x 0.7781 = Cd. CALCIUM Separation. 1 . Calcium is generally determined in the filtrate from the group containing zinc and manganese; this filtrate is boiled to expel excess of ammonium sulphide, filtered if not quite 406 THE CYANIDE HANDBOOK clear, then mixed with a boiling solution of ammonium oxalate. After settling clear, the precipitate is filtered off and washed with hot water. In cases where much calcium is present the precipitate is re-dissolved in a minimum quantity of HC1 and re-precipitated from boiling solution with ammonia and ammo- nium oxalate. This double precipitation is necessary because calcium oxalate retains both magnesium and sodium salts very tenaciously. 2. A direct separation in the filtrate from silica may be made as follows. The liquid is heated nearly to boiling, excess of am- monia added, then solid oxalic acid till iron is dissolved. Neu- tralize exactly with ammonia and add oxalic acid until the ferric hydrate just dissolves. On boiling, the calcium is precipitated as oxalate, free from iron. In presence of manganese it may be necessary to add bromine water after the first addition of ammonia; boil, filter, and wash with hot water before proceeding to add oxalic acid. Estimation. 1 . When calcium has been precipitated as oxa- late it may be conveniently and accurately estimated by titra- tion with standard permanganate. The precipitate is washed thoroughly with hot water, dissolved in dilute sulphuric acid, heated to 70 C., and titrated with permanganate till the pink tint remains permanent. The permanganate should be standard- ized on pure crystallized oxalic acid. 1 part oxalic acid (C 2 H 2 O 4 2H 2 O) = 0.4451 parts CaO. Special methods for determining the percentage of caustic lime in commercial samples will be described in a later section. Ice. ^ KMnO 4 = 0.0028 gram CaO. = 0.0020 gram Ca. A convenient standard solution is one containing 1.58 gram KMn0 4 per liter, so that 1 cc. = 0.001 gram Ca. The reactions are as follows: Ca(C 2 O 4 ) + H 2 SO 4 = CaSO 4 + H 2 C 2 O 4 . 2KMnO 4 + 5H 2 C 2 O 4 + SHzSO* = K 2 SO 4 + 2MnSO 4 + 10CO 2 + 8H 2 O 2. Calcium may also be determined gravimetrically by strong ignition of the precipitate, and weighing as lime, CaO, or by gentle ignition, afterwards moistening the residue with ammo- ANALYSIS OF ORES AND SIMILAR MATERIAL 407 nium carbonate, evaporating to dryness, heating just to redness, and weighing as calcium carbonate, CaCO 3 . CaCO 3 X 0.4006 = Ca. CaCO 3 X 0.5604 = CaO. CARBONIC ACID The carbonic acid in ores exists in the form of carbonates, generally insoluble in water but soluble in dilute acids. It may be estimated in one or other of the numerous types of carbonic acid apparatus, of which the general principle is as follows: The apparatus is weighed, first without and afterwards with the por- tion of ore to be analyzed. The dilute acid (HNO 3 or HC1) which serves to decompose the carbonate is then allowed to come in contact with the ore by turning the tap of the bulb in which it is contained and momentarily opening the stopper. The escaping gas has to pass through a bulb containing sulphuric acid, which retains any moisture. Gentle heat is necessary in some cases for complete expulsion of carbonic acid. When the action is complete, a slow current of air is aspirated through the appa- ratus, which is then again closed and weighed, the loss of weight giving the CO 2 expelled. 3. For more exact determinations the gas is passed through a series of tubes containing desiccating agents, such as strong sulphuric acid, anhydrous copper sulphate, calcium chloride, etc., and the C0 2 is absorbed by soda-lime contained in a weighed U-tube. The inlet of the flask in which the decomposition of the carbonate takes place is protected by a U-tube of soda lime to absorb any atmospheric CO 2 which might otherwise enter the apparatus, and the exit end of the weighed tube is protected by a U-tube containing a desiccating agent and absorbent of CO 2 . Air is then aspirated through the whole apparatus, and the weighed U-tube detached, closed, and weighed. The increase of weight gives the CO 2 . CHLORINE Chlorine occurs as chlorides in small quantities in certain ores. Silver occurs occasionally as "horn-silver" (AgCl), in which form it is readily extracted by cyanide. When insoluble chlorides are present, the crushed ore is fused with sodium carbonate and chlorine determined in the aqueous extract. Any heavy metals 408 THE CYANIDE HANDBOOK present in the extract are removed by H 2 S, the excess of which must be boiled off. The liquid is then made acid with HNO 3 and the chlorine determined. Estimation. 1 . If present in sufficient quantity, it may be titrated in neutral solution by AgNO 3 , with potassium chromate (K 2 CrO 4 ) as indicator. The silver solution is added till a per- manent reddish tint remains after agitation. 1 2. In cases where it is inconvenient to neutralize the solu- tion, a measured amount of standard silver nitrate is added, more than sufficient to precipitate the chlorine, and the liquid acidified with HNO 3 . The AgCl is then filtered off and the resid- ual AgNO 3 determined by titrating with standard potassium or ammonium thiocyanate, using ferric nitrate or sulphate as in- dicator. 2 3. Very small quantities are best estimated gravimetrically as AgCl. The liquid, which must be slightly acid with HNO 3 , is mixed with a few drops of AgNO 3 , boiled thoroughly, and allowed to stand for some time. The AgCl should settle in a dense form. It is then washed by decantation, without filtering, transferred to a weighed porcelain crucible, dried and heated to incipient fusion, cooled, and weighed as AgCl. AgCl X 0.2472 = Cl. CHROMIUM This metal occurs as chrome iron ore; it is precipitated along with Al and Fe in treating the filtrate from silica with ammonia or with sodium acetate. It is preferable, however, to determine it in a separate portion of the ore. This must be ground extremely fine and fused, either with sodium carbonate and niter, or with sodium peroxide, to convert the chromium into a soluble yellow chromate. The metal is dissolved in water, giving a yel- low solution. Determination. 1. When only small quantities are present, the chromium may be determined colorimetrically by compari- son with a standard solution of K 2 CrO 4 . 2. If present in larger amount, the solution, after boiling off any excess of sodium peroxide, is mixed with ammonium carbo- nate to partially neutralize the caustic alkali, filtered, and acidified 1 For details, see Sutton, "Volumetric Analysis," p.. 152 (8th edition). 2 Sutton, loc. tit., pp. 155, 184. ANALYSIS OF ORES AND SIMILAR MATERIAL 409 with dilute sulphuric acid. An excess of ferrous ammonium sulphate is then given, by adding a weighed quantity of the salt, and the residual ferrous iron titrated with permanganate. 1 cart Fe = 0.3107 parts Cr. COBALT The sample is evaporated to dryness with HNO 3 and KC1O 3 . The residue is dissolved in dilute HC1, heated to boiling, and pre- cipitated with H 2 S. The nitrate from H 2 S precipitate is boiled to expel excess of H 2 S, a few drops of HNO 3 added, and the iron precipitated by ammonia. The filtrate is then made strongly acid with acetic acid, adding ammonium acetate and heating to about 70 C. On again passing in H 2 S, cobalt, nickel, and zinc are precipitated, leaving manganese in solution. Filter, wash with warm water, and re-dissolve the precipitate in HC1 + HNO 3 . Add a considerable excess of ammonium chloride, evaporate to dryness, and ignite to volatilize zinc chloride; re-dissolve the residue, which now contains only Co and Ni, in HC1 + HNO 3 . Determination of Cobalt. The extract containing Co and Ni is made alkaline with caustic alkali, acidified with acetic acid, and mixed with a concentrated solution of potassium ni- trite. On standing in a warm place (best for 24 hours) the cobalt separates as a yellow precipitate (dauble nitrite of cobalt and potassium). This is filtered off, washed with potassium acetate solution and finally with alcohol, dried, heated in a porcelain crucible with sulphuric acid, ignited at a very low red heat and weighed as 2 CoSO 4 + 3 K 2 SO 4 . Weight of residue X 0.1416 = Co. COPPER The presence of copper in ores to be treated by cyanide intro- duces such difficulties that the detection and estimation of the metal may become a matter of considerable importance. Detection. In most cases the presence of copper is readily detected by boiling a small quantity of the ore in strong nitric acid, diluting and neutralizing the acid with ammonia. On allowing to settle, the presence of copper is shown by a blue color in the supernatant liquid. It may be confirmed by filtering a little of the fluid, rendering just acid with hydrochloric or acetic 410 THE CYANIDE HANDBOOK acid and adding a drop of ferrocyanide solution, which gives a reddish-brown coloration. In order to determine whether the copper in an ore is in a form readily attacked by cyanide, a little of the crushed ore may be shaken with cyanide solution and filtered; if the filtrate be acidulated and a drop of ferrocyanide added, the characteristic reddish-brown color is obtained if any copper has been dissolved. Separation. 1 . In cases where lead has also to be determined, the copper may be estimated in the filtrate from the lead sulphate precipitate, after the alcohol has been boiled off. It may be precipitated as sulphide, best from a hot solution, by sulphureted hydrogen or by sodium thiosulphate, or it may be precipitated on aluminium as described below, in the form of finely divided metallic copper. 2. A portion of the crushed ore (0.5 gram to 2 or 3 grams, according to richness in copper) is evaporated nearly to dryness with HN0 3 . It is then boiled with HC1 and finally with H 2 SO 4 , heating till white fumes are freely evolved. Cool, dilute, and filter. Wash with hot water. The copper in the filtrate is now precipitated as sulphide by passing a current of H 2 S gas through the nearly boiling liquid; when saturated, allow to settle for a few minutes, filter while still hot, and wash with H 2 S water. In presence of much iron, the sulphide precipitate must be re-dis- solved in HNO 3 , again boiled with H 2 S0 4 , diluted and re-precip- itated with H 2 S. Dissolve the precipitate in concentrated HNO 3 , adding a little potassium chlorate if arsenic or antimony is pres- ent; boil thoroughly, filter off any undissolved sulphur. If any undissolved precipitate remains on the filter-paper, dry and ignite the latter, dissolve the ash in boiling HNO 3 , and add to the bulk of the solution. The copper is in solution as nitrate, with perhaps Bi, Cd, or some other metals. In presence of Bi, add excess of ammonia, boil and filter. 3. Proceed as in method No. 2, until white fumes are given off on heating with H 2 SO 4 . Then cool, dilute and heat to boiling. Filter off any precipitate of PbSO 4 , etc. Wash with hot water, dilute to (say) 75 cc., add one or more sheets of aluminium, bent so as not to lie flat in the beaker. Boil about ten minutes, till the copper is precipitated in a spongy form. Decant the liquid through a filter, and wash the copper on to the filter with H 2 S water till all iron, etc., is removed. The copper adhering to the ANALYSIS OF ORES AND SIMILAR MATERIAL 411 aluminium is then dissolved in hot 50 per cent, nitric acid, which is afterwards poured over the precipitate in the funnel, together with a little bromine or KC1O 3 ; collect the filtrate in a flask, wash with hot water, boil thoroughly to expel Br or Cl, and add am- monia. 4. Certain slags contain all or nearly all the silica in a form soluble in HC1, the copper being present as sulphide, insoluble in that acid. In such cases the sample is simply treated with hot dilute HC1 and filtered. Dry, ignite, and treat the residue with HNO 3 and a little HC1. Boil off red fumes, dilute, add am- monia, boil, and filter. Estimation. 1. If the copper has been obtained in nitric acid solution, neutralize with sodium carbonate, and add ammonia till a clear-blue solution is obtained. Filter if necessary (e.g., if Bi is present). If the copper contents are approximately known, prepare a similar solution containing about the same amounts of copper, nitric acid, and ammonia, using pure copper foil. This solution is used as a standard. Titrate the blue solu- tions as nearly as possible at the same time with a solution of potassium cyanide 0.5 to 2 per cent. KCN, according to circum- stances, running in the solution until a faint violet color of equal intensity remains in each. When approaching the finish, the cyanide solution should be run in a little at a time, in quantities of say 0.5 cc., waiting some time before making a fresh addition, as the reaction takes place slowly. 2. The nitric acid solution of the copper is neutralized with sodium carbonate, and acetic acid added until the precipitate dissolves. (Ammonia may be used instead of sodium carbon- ate.) Cool, and add about 10 parts of potassium iodide for every part of copper supposed to be present. The following reaction takes place: 2Cu(CH 3 CO 2 ) 2 + 4KI = Cu 2 I 2 + I 2 + 4 (KCH 3 CO 2 ). The liberated iodine is then titrated at once with standard sodium thiosulphate, adding starch solution towards the finish. 1 Ice. thiosulphate = 0.0063 gram Cu. In this process care must be taken that no iron or nitrous acid are present, and the liquid must not be too dilute; if these 1 Sutton, loc. tit., p. 201. Beringer, loc. cit., p. 199. 412 THE CYANIDE HANDBOOK conditions are neglected, the blue color of the iodine returns on standing for a few moments and no definite end-point can be obtained. Bismuth interferes with the process, as it gives a color with iodine; if present, it may be separated by boiling the solution with carbonate of ammonium and filtering, before add- ing the acetic acid. 3. When only small quantities of copper are present, they may be determined colorimetrically, by comparing the tint of the ammoniacal solution with that of a liquid in a similar vessel containing amounts of nitric acid and ammonia approximately corresponding to those in the assay, adding standard copper nitrate until the colors are alike. Still smaller quantities may be determined colorimetrically by using a standard ferrocyanide solution. 4. The copper is obtained in slightly acid HC1 solution, and is reduced by sodium sulphite and precipitated with thiocyanate. The cuprous thiocyanate is then boiled, filtered, and washed thoroughly with hot water. It is next dissolved off the filter with boiling caustic soda, and the resulting sodium thiocyanate titrated with permanganate after acidulating with H 2 SO 4 . The permanganate solution should be standardized on pure copper under the conditions of the assay, as cuprous thiocyanate is slightly soluble in the reagents used. 5. The writer has found that small amounts of copper may be readily and accurately estimated by a method nearly iden- tical in principle with that of T. Moore for the estimation of nickel. 1 (See below.) The copper is obtained in a nitrate or sulphate solution free from other metals. Sufficient caustic soda is added to precipitate the whole of the copper, avoiding a large excess of alkali; then standard cyanide solution, in measured amount, until a perfectly colorless liquid is obtained. Potassium iodide is then added and the excess of cyanide titrated with silver nitrate. The presence of ammonia or ammonium salts vitiates the test. A constant excess of cyanide should be used in all cases, as the amount of copper indicated increases slightly with increased excess of cyanide. Satisfactory results were obtained with solutions containing from 1 to 50 mg. Cu. The results were on a par with the iodide method in point of accuracy, and superior to the ordinary cyanide method. i "Chem. News," LXXII, 92. ANALYSIS OF ORES AND SIMILAR MATERIAL 413 FLUORINE Fluorine occurs in ores chiefly as fluor-spar (CaF 2 ) and as cryolite (Na 3 AlF 6 ). It may be determined l by mixing the finely-ground ore with an .excess of powdered silica and dis- tilling with concentrated sulphuric acid for two hours at a gentle heat, whereby the fluorine is expelled as gaseous silicon fluoride : 2CaF 2 + 2H2SO 4 + SiO 2 = 2CaSO 4 + 2H 2 O + SiF 4 . During the distillation a current of air is made to pass slowly through the apparatus, and the silicon fluoride, which should be cooled by leading through a bulb-tube immersed in water, is passed into an alcoholic solution of potassium chloride (30 parts KC1 in 100 parts alcohol and 100 water), by which it is decom- posed as follows: 3SiF 4 + 2H 2 O + 4KC1 = SiO 2 + 2K2SiF 6 + 4HC1. The potassium silicofluoride remains undissolved in the alcoholic solution. By titrating the HC1 with standard alkali and lac- moid indicator, the amount of fluorine may be determined. 1 equivalent HC1 = 3 equivalents F. Ice. ~ HC1 = 0.0057 gram F. IRON Separation. The estimation of iron is generally made in the filtrate after separation of silica. The methods of separation with ammonia and with sodium acetate are described under aluminium. In a few cases the precipitate of alumina, etc., formed under these circumstances, can be ignited and weighed, and afterwards completely dissolved by boiling with concen- trated sulphuric acid. In general, however, the ferric oxide, after ignition, resists the action of acids very strongly and can only be got completely into solution by long-continued fusion (2 to 4 hours) with potassium bisulphate. The fusion should be conducted at as low a temperature as possible, and the bisul- phate should have been previously fused by itself, to expel water- vapor and excess of acid. The molten mass is allowed to cool, dissolved in water to which a little sulphuric acid is added, boiled and filtered. Any residual silica is here collected and estimated. i Penfield, "Chemical Engineer," III, 65. 414 THE CYANIDE HANDBOOK Estimation. 1. Trie liquid obtained as above is treated with a reducing agent, cooled and titrated with permanganate. The reduction may be carried out in many ways. The writer prefers boiling with clean aluminium turnings until a drop of the liquid no longer reacts with a drop of thiocyanate solution. Other methods of reduction commonly employed are with zinc, with sulphureted hydrogen, and with sodium sulphite. 1 Care must be taken to avoid oxidation of the iron in the solution dur- ing the cooling process. When much iron is present, it may be necessary to dilute with boiled, cooled, distilled water before titrating with permanganate. The permanganate solution is standardized on pure iron wire dissolved in pure 20 per cent, sulphuric acid, or on a known weight of pure dry ferrous ammo- nium sulphate, FeSO 4 (NH 4 ) 2 SO 4 6H 2 O. 2. The liquid, reduced in the manner described, or (more rapidly) by addition of a small excess of stannous chloride which excess is afterwards removed by cautious addition of mercuric chloride is titrated with a standard solution of potas- sium bichromate, using ferricyanide of potassium, in spots on a white porcelain plate, as external indicator. For the estimation of iron in slags, 0.5 grams of the sample may be dissolved in 25 cc. boiling water, with addition of 20 cc. HC1. The mixture is stirred vigorously and boiled to expel H 2 S. Then add SnCl 2 in slight excess, cool, and add 20 cc. of a saturated solution of mer- curic chloride. Titrate with bichromate solution, 4.392 grams K 2 Cr 2 O 7 per liter. 2 3. Very small amounts of iron are best estimated by the colorimetric method. The ferric hydrate or basic acetate pre- cipitate is dissolved in HC1, a sufficient excess of thiocyanate solution added, and the tint compared with that produced under similar conditions by measured quantities of standard ferric chloride solution, using the same amount of thiocyanate in each case. The standard solutions may be conveniently adjusted so that 1 cc. = 0.01 or 0.001 gram Fe, according to the amount to be estimated. LEAD Separation. In some ores, such as pure galena, the lead may be completely extracted by boiling with strong hydrochloric 1 C. and J. Beringer, Text Book of Assaying, 9th edition, p. 235. 2 H. T. Waller, Bull. I. M. M., No. 49, Oct. 8, 1908. ANALYSIS OF ORES AND SIMILAR MATERIAL 415 acid; in other cases nitric acid is used, with addition of a few drops of HC1. After boiling for some time, the flask is allowed to cool, and sufficient sulphuric acid added to precipitate the whole of the lead as PbSO 4 , together with barium and perhaps some other metals. Evaporate until dense white fumes of sul- phur trioxide are given off. Cool, dilute carefully, and boil. Again cool, add alcohol, stir and settle thoroughly (over night if possible). Decant the clear liquid and wash by decantation with dilute sulphuric acid containing about 10 per cent, of alco- hol, finally with pure alcohol, leaving the residue as much as possible in the beaker. The filtrate may serve for the estima- tion of copper, etc. To the residue add a strong solution of ammonium acetate, with a slight excess of acetic acid, to dis- solve PbSO 4 , leaving Si0 2 , BaS0 4 , CaSO 4 , etc., undissolved. Moisten the filter with ammonia and pass the liquid through into a clean flask or beaker. Wash first with dilute ammonium acetate, then with hot water. Estimation. 1. Gravimetric Method. Acidulate the fil- trate, obtained as above, with sulphuric acid. Allow to settle for some time and filter. The precipitate is separated as much as possible from the paper, ignited carefully, and the ash of the paper, separately ignited and treated successively with HNO 3 and H 2 SO 4 , is added. Weigh as PbSO 4 . . It is preferable to collect the final precipitate of PbSO 4 on a weighed perforated porcelain crucible with asbestos filter. This is washed with absolute alco- hol, dried, ignited, and again weighed. PbSO 4 X 0.6829 = Pb. 2. Several volumetric methods are in use, 1 which are appli- cable where the quantity of lead is considerable, (a) The lead sulphate is dissolved in ammonium acetate, and titrated in hot solution with standard ammonium molybdate, using tannin as indicator, which gives a brown tint with excess of molybdate. A correction is required for the amount of molybdate solution necessary to develop this color. (6) Chromate of potassium is added in excess; the excess is then determined either colorimet- rically or by means of a ferrous iron solution, (c) The lead dis- solved as acetate is titrated with standard ferrocyanide at about 60 C., using uranium acetate as indicator. 1 See Beringer, loc. tit.; also Irving C. Bull. "Sch. of Mines Qly.," XXIII, No. 4; "Chem. News," LXXXVII, 40. 416 THE CYANIDE HANDBOOK MAGNESIUM 1. After precipitation of calcium as oxalate, the nitrate is made strongly alkaline with ammonia, and mixed with an excess of sodium or ammonium phosphate, well stirred without touch- ing the sides of the vessel with the stirring-rod, and allowed to settle for a considerable time (2 or 3 hours). The precipitate is then filtered off and washed with 2.5 per cent, ammonia. In exact analysis it should be re-dissolved in HC1 and re-precip- itated with ammonia and phosphate. In washing, the mini- mum necessary amount of wash-water should be used, and the first portions of the precipitate are best washed on to the filter, by means of a small quantity of the filtrate. When free from chlorides, the precipitate is dried, ignited, moistened with nitric acid, cautiously evaporated to dryness, then again ignited and weighed as magnesium pyro phosphate (Mg 2 P 2 O 7 ). Mg 2 P 2 O 7 X 0.2188 = Mg. 2. Magnesium may also be estimated volumetrically by dis- solving the precipitate of magnesium ammonium phosphate in decinormal sulphuric acid and titrating the excess of acid with standard alkali and methyl orange. 1 Ice. H 2 SO 4 = 0.001218 gram Mg. = 0.002018 gram MgO. MANGANESE Separation. 1. Manganese is contained in the filtrate obtained after precipitating iron, alumina, etc., with ammonia or with sodium acetate. Its presence is generally indicated at an earlier stage by the greenish, sometimes purplish, color of the melt obtained with the ore and sodium carbonate; this color disappears on treatment with hydrochloric acid. Manganese may be precipitated, along with some other metals, by treating the alkaline filtrate from Al, etc., with H 2 S or ammonium sulphide solution, and allowing to stand, covered, for a considerable time. A more convenient method, which at once separates the manganese from accompanying metals, is to make the filtrate just acid with sulphuric acid. It is then warmed to about 70 C. and precip- itated by a brisk current of H 2 S gas. This throws down zinc, i Handy, " Journ. Am. Chem. Soc.," XXII, 31. ANALYSIS OF ORES AND SIMILAR MATERIAL 417 nickel, cobalt, and also copper and any other of the heavy metals which may be present, together with any traces of platinum which have passed through from previous operations. The precipitate is filtered off and washed with dilute H 2 SO 4 containing H 2 S. The manganese is contained in the filtrate. 2. The ore is decomposed by treatment with hydrochloric and nitric acids in the usual way. It is then heated to white fumes with sulphuric acid, cooled somewhat, and diluted with water, warmed till iron has dissolved, and then mixed with an emulsion of zinc oxide or pure precipitated zinc hydrate. This precipitates iron, leaving manganese in solution. Filter, add a saturated solution of bromine in water, and 2 to 3 grams of sodium acetate. 1 On boiling, the manganese is precipitated as hydrated manganese dioxide. Boil and wash with hot water. Estimation. 1. When manganese has been separated in H 2 SO 4 solution, as in the: first method, the filtrate is made alka- line with ammonia, which, with the excess of H 2 S, causes the manganese to precipitate as sulphide. The flask is stoppered and set aside over night, or preferably for 24 hours. The clear liquid is then decanted through a filter, the precipitate washed with dilute ammonium sulphide, then dissolved in HC1 and H 2 SO 4 , boiled to expel excess of HC1, and the manganese deter- mined by any suitable gravimetric or volumetric method. Per- haps the simplest in ordinary circumstances is to add carbonate of soda in slight excess to the manganese solution, which is pre- viously boiled in a porcelain or platinum vessel. The manga- nese, precipitated as carbonate, is filtered off, washed with hot water till free from alkali, dried, ignited cautiously in an open crucible, and weighed as Mn 3 O 4 . 2. Where manganese has been precipitated by bromine as MnO 2 , it may be conveniently estimated by adding an excess of oxalic acid, together with some dilute sulphuric acid, heating nearly to boiling, diluting with water, and titrating the warm solution with standard permanganate, to determine the excess of oxalic acid. 126 parts crystallized oxalic acid = 55 parts Mn or 87 parts MnO2. Reduction by oxalic acid in presence of dilute sulphuric acid affords an easy means of determining the amount of actual MnO 2 1 For details see A. H. Low, "Technical Methods of Ore Analysis," 3d edition, p. 146. 418 THE CYANIDE HANDBOOK in a manganese ore when other oxides are present, as the latter do not oxidize oxalic acid under the conditions of the test. 3. Very small quantities of manganese are best determined by the colorimetric method; l the ore is oxidized by boiling with nitric acid and lead peroxide, and the tint of the resulting solu- tion compared with that of a standard permanganate solution containing a known quantity of manganese. For further details on the estimation of manganese see Berin- ger, "Text Book of Assaying," pp. 299-306, and Sutton, "Volu- metric Analysis, 8th edition, pp. 255-266." MERCURY Gold and silver ores occasionally contain cinnabar, and metal- lic mercury is frequently present in such products as tailings and concentrates, so that its determination may be of importance in connection with the cyanide process. Whenever possible, sev- eral pounds of the sample should be carefully panned, in cases where the presence of metallic mercury is suspected. If floured, the metal may generally be obtained in a single globule by the addition of cyanide or ammonium chloride. Estimation. 1. The following rough method frequently suffices. The powdered sample, mixed with soda-lime, is placed in a porcelain crucible covered with a disk of gold or silver shaped like a watch-glass, the hollow side uppermost and filled with water. The mixture is cautiously heated; the cinnabar or other ore is decomposed, and the volatilized mercury condenses on the under side of the disk. When the action is complete, the disk is removed, washed with' alcohol, and dried in a desiccator. It is then weighed, the increase of weight giving the amount of mercury deposited. 2. The mineral may also be mixed with lime and heated in a combustion tube closed at one end. The closed end is previously filled with powdered magnesium carbonate. The open end is bent and drawn to a point, dipping under water. The carbonic acid evolved from the magnesium carbonate drives out the last portions of mercury vapor; the sublimed mercury is condensed by the water, collected, dried, and weighed. 2 1 Beringer, loc. tit., p. 306. 2 Beringer, loc. tit., p. 172. ANALYSIS OF ORES AND SIMILAR MATERIAL 419 NICKEL Separation. Proceed as described under cobalt, until a solution is obtained containing only these metals. In cases where a separation of the two metals is necessary, precipitate the cobalt as described, with potassium nitrite, filter, precipitate the nickel from a boiling solution, by caustic alkali, filter, and wash with hot water. Estimation. 1 Dissolve the precipitate of nickel hydroxide in dilute ammonia. Add a measured quantity of a solution of potassium cyanide which has been accurately standardized on a silver nitrate solution of known strength, using potassium iodide indicator. If the nickel precipitate does not dissolve readily in ammonia, add a little ammonium chloride. When sufficient cyanide has been added to give a perfectly clear solution, a mod- erate excess is added beyond this point, then potassium iodide is added, and the excess of cyanide titrated by means of the standard silver solution. Cobalt and nickel may be estimated together by this method, but the presence of cobalt causes a darkening of the solution. The cyanide solution should also be standardized on a nickel solution of known strength. OXYGEN In most cases the amount of oxygen in an ore can only be determined by difference, after all other ingredients have been estimated. Sulphates and carbonates are separately deter- mined and their oxygen contents calculated. The nature and degree of oxidation of the metallic oxides present may sometimes be ascertained by examining uncrushed specimens of the ore. In a few cases oxygen may be directly determined (e.g., in cas- siterite) by igniting the ore in a stream of hydrogen and collect- ing the water formed by means of a suitable absorbent in a weighed tube. The hydrogen must of course be previously dried and freed from any admixed oxygen, and any combined water present in the mineral must be previously expelled, or allowed for in the determination. In cases where oxides of a given metal in various degrees of oxidation are present in an ore, methods may be employed for determining one or other of these separately and deducing the amount of the other from the total quantity ' T. Moore, "Chem. News," LXXII, 92. 420 THE CYANIDE HANDBOOK of the given metal. Thus in an ore containing ferrous and ferric oxides, we may determine the total iron, and the iron present as ferrous oxide, by separate tests. A simple calculation then gives the amount of each oxide present. PHOSPHORIC ACID Separation. Ores frequently contain small quantities of phosphoric acid in the form of metallic phosphates, and calcium phosphate is found pure or mixed with other minerals. Most phosphates are soluble in hydrochloric acid. 1. Boil with hydrochloric acid, filter, ignite the insoluble residue in a platinum crucible and add it to the extract. Evap- orate the whole to dryness. Take up the residue with nitric acid, concentrate by boiling, dilute, filter, and wash with water. Precipitate the warm filtrate with a warm clear solution of am- monium nitro-molybdate, which gives a yellow precipitate on standing. 2. In cases where arsenic is present, take up the evaporated residue with HC1 instead of HNO 3 , precipitate the As with H 2 S, filter, boil off excess of H 2 S, then add HNO 3 and proceed as above, adding ammonium nitro-molybdate. 3. The substance is treated with HC1 as already described, evaporated to dryness, taken up with HC1, diluted and precip- itated with H 2 S. The precipitate is filtered off and the filtrate boiled to expel excess of H 2 S and treated with HNO 3 . Ferric chloride is then added, if sufficient iron is not already present, to convert the whole of the phosphoric acid into ferric phosphate ; the solution is then just neutralized with ammonia, sodium ace- tate and acetic acid added, boiled slightly, and filtered. The precipitate, after washing, is transferred to a flask and treated with ammonia and H 2 S to remove the iron. It is then filtered, the filtrate containing the phosphoric acid as ammonium phosphate. Estimation. 1. When the phosphate has been separated with nitro-molybdate, the precipitate is collected on a weighed filter-paper, and washed, first with dilute nitric acid, then with alcohol, dried at 110 and weighed as ammonium phospho- molybdate (NH 4 ) 3 12MoO 3 P X 0.00044 = gram CO 2 as carbonate; (/A - 20) X 0.00044 = gram CO 2 as bicarbonate. (c) Hydrates and Carbonates. The water is alkaline to both indicators. N (i) Titrate with acid and phenol-phthalein. 100 Reactions: NaOH + HC1 = NaCl + H 2 O. Na 2 CO 3 + HC1 = NaHCOa + NaCl. Ice. ^ acid = 0.0004 gram NaOH. J-LfU = 0.00044 gram CO 2 as carbonate. N (ii) Titrate with acid and methyl orange. 100 Reaction : NaOH + HC1 = NaCl + H 2 O. Na 2 CO 3 + 2HC1 = 2NaCl + H 2 O +CO 2 . (2 - /A) X 0.0004 = grams hydrate, as NaOH. ( /A - 0) X 0.00044 = grams CO 2 as carbonate. (d) Carbonates Alone Present. The water is alkaline to both indicators, ^ = 2 . X 0.00044 = grams CO 2 as carbonate. A* X Q.00022 = grams CO 2 as carbonate. (e) Bicarbonates Alone Present. The water is neutral to phenol-phthalein and alkaline to methyl orange. At X 0.00044 = grams CO 2 as bicarbonate. ANALYSIS OF SUNDRY MATERIALS 471 These formulae are not applicable where other alkalis, such as ammonia, sulphides, or cyanides, are present in the water. The carbonates occurring in ordinary waters are those of the alkali metals (Na and K), and also those of the metals Ca, Mg, Fe, Mn, held in solution by the excess of carbonic acid, and perhaps forming unstable compounds analogous to the alkaline bicar- bonates. Carbonic Acid. The following method for determining total carbonic acid (free and combined) may be used as a check on the titration given above. A measured volume (100 to 200 cc.) of the water is mixed with a slight excess of ammonia and suffi- cient calcium chloride solution to precipitate all the C0 2 present as CaCO 3 . The flask is immersed for two hours in a vessel of boiling water. The precipitate is washed by decantation with hot water containing a little ammonia. The liquid poured off is filtered, and the washing continued till the residue is free from chlorides. Any material collected on the filter is then washed back into the flask, standard acid is added in excess, and the N residual acid titrated with alkali and methyl orange. 100 CaCO 3 + 2HC1 = CaCl 2 + CO 2 + H 2 O. Ice. ^ acid = 0.00022 gram CO 2 . Solids in Suspension and Solution. These are determined as described in Section III (for cyanide solutions), except that the solids in solution may generally be evaporated at a higher tem- perature, first on a water-bath and finally in an air-bath at 150 C. Care must be taken not to heat sufficiently to expel CO 2 from carbonates or to decompose and char any organic matter. After cooling and weighing it is sometimes desirable to determine the weight of residue on ignition at a low temperature. The / dish is heated until organic matter is burnt off, without however, decomposing carbonates. The remaining residue may be used for determination of SiO 2 , Fe 2 3 , A1 2 O 3 , CaO, MgO, etc. Hardness. The " hardness " of water is estimated by the amount of calcium carbonate contained. One degree of hard- ness = 1 grain CaCO 3 per gallon. It may, therefore, be calcu- lated from the amount of calcium carbonate found in the analysis. For Clark's test with soap solution, see Sutton, " Volu- 472 THE CYANIDE HANDBOOK metric Analysis," 8th edition, p. 488. " Permanent hardness " is generally measured by the amounts of calcium and magnesium sulphates present. Sulphates. Except in the case of water containing consider- able quantities of organic matter, silicates or nitrates, the sul- phuric acid may be estimated directly by acidulating a measured volume with HC1, heating to boiling, and precipitating with BaCl 2 in the usual way. Where interfering substances are pres- ent, evaporate to dryness with HC1, ignite and re-dissolve residue in dilute HC1 before precipitating with BaCl 2 . Chlorides. If the water is approximately neutral these may be estimated by titration with standard AgNO 3 and neutral chromate indicator. If acid or alkaline, neutralize with standard NaOH or HNO 3 as required before adding AgNO 3 . If there be any doubt as to the end-point, pour off half the liquid into an- other vessel, add another drop of AgNO 3 to one portion, and compare the tints; if any difference is observed the titration may be regarded as having been finished before adding the extra drop. 1 When the water is strongly alkaline, it is perhaps better to add excess of HNO 3 and AgNO 3 , and titrate excess AgNO 3 with thio- cyanate and ferric indicator. Silica is determined by evaporating and igniting, and filter- ing off the insoluble residue. The latter is then taken up with hot dilute HC1, filtered, and the residue on the filter again strongly ignited and weighed as SiO 2 . Silica may exist in suspension as free SiO 2 , or in solution as a silicate of an alkali metal. Calcium and magnesium may be determined in the filtrate from the silica by heating to boiling, adding ammonia in slight excess, filtering and proceeding as described in previous sec- tions. (See above.) Sodium and Potassium. As a check on the direct estimation of these elements, it is desirable to convert the total evaporated residue into chlorides or sulphates, and weigh as such. Evap- orate to dryness and ignite sufficiently to render SiO 2 , Fe 2 O 3 , and A1 2 O 3 insoluble. Re-dissolve residue in water, filter, add HC1 or H 2 SO 4 , and re-evaporate to dryness. Weigh residue. In cases where the water contains no sulphates, hydrochloric acid may advantageously be used. In this case the total metals may be found by simply titrating the chlorine in an aliquot part of 1 See A. H. Low, "Technical Methods of Ore Analysis," p. 71. ANALYSIS OF SUNDRY MATERIALS 473 the dissolved residue. If the amounts of Ca and Mg are known, these are deducted, together with the total chlorine in the evap- orated residue; the balance may generally be taken as xNa + yK. In cases where sulphates are originally present, it is better to use sulphuric acid. The Ca and Mg are calculated to CaSO 4 and M'gSO 4 ; the difference gives xNa 2 SO 4 + yK 2 SO 4 ; from which Na and K can be obtained by calculation. The sulphuric acid is determined in an aliquot part of the dissolved residue, and calculated to SO 4 . Deducting this, we obtain the total metals, and deducting Ca and Mg, we obtain the combined weight of Na and K. It is rarely necessary to make separate determina- tions of Na and K. If required, this can be done as described in Sections I, III, and IV. Iron and Alumina. A sufficient quantity of the water is evaporated with HC1, adding a little HNO 3 and igniting residue to destroy organic matter. The residue is then boiled with HC1, filtered, and the filtrate mixed with ammonia in slight excess. After again boiling it is filtered, and the precipitate ignited and weighed as xFe 2 O 3 + yA! 2 O 3 . If iron is to be determined sep- arately, instead of igniting the precipitate, dissolve in HC1 and estimate calorimetrically with thiocyanate. (See Section I.) Organic Matter. Evaporate at 150 C. and weigh residue. Ignite gently to burn off carbonaceous matter without decom- posing metallic carbonates, and weigh again. The difference of the two weights- gives a rough idea of the amount of organic matter present. A simple test may also be made as follows : * A flask is rinsed out, first with sulphuric acid, then with pure distilled water. A measured volume (100 to 250 cc.) of the water to be tested is run into the flask, which is then stoppered and heated to 80 C. Add 10 cc. sulphuric acid (25 per cent, by vol.) to which enough KMnO 4 has previously been added to give a faint permanent pink tint. Add also 10 cc. potassium permanganate (0.395 gram KMnO 4 per liter). Allow to stand, stoppered, for four hours, then add a little KI and titrate iodine liberated, repre- senting residual KMnO 4 , with a corresponding thiosulphate solu- tion, using a little starch indicator at the finish. Ice. KMnO< consumed = 0.0001 gram available oxygen. Sutton, "Volumetric Analysis," 8th edition, p. 518. 474 THE CYANIDE HANDBOOK A rule 1 based on average analyses of organic carbon in waters gives the following factors: Organic carbon = O absorbed X 2.38 (for river water). = O absorbed X 5.8 (for deep well water). i Ibid., p. 508. PART IX METALLURGICAL TESTS WE will now describe certain operations which are frequently necessary in controlling the work of a cyanide plant, and in determining the proper conditions of treatment. These consist of various tests arid measurements based on well-known mathe- matical, physical, and chemical principles, but distinct from ordinary analytical processes, and may be roughly described as metallurgical tests. These will be considered under the follow- ing heads: I. Density determinations. II. Acidity and alkalinity tests. III. Grading and screening tests. IV. Concentration and amalgamation tests. V. Cyanide extraction and consumption tests. VI. Tables and formulae. SECTION I DENSITY DETERMINATIONS IN cyanide treatment it is very essential to have as exact a knowledge as practicable of the quantities of material treated per day or per charge, in order to regulate operations and to check the actual recovery of values. Differences between "theo- retical " and " actual " extraction may often be traced to inaccu- rate estimates of the quantities treated. In the case of crushed ore, moderately dry tailings and similar material, this may be done by direct weighing, but in many cases, such as slime pulp, it is practically impossible to weigh the charge in bulk, and the weight can only be calculated indirectly after ascertaining the densities of the pulp and of the dry material. (a) To find weight of a given volume of sand or similar material. In cases where it is not convenient to weigh the entire charge direct, the wet and dry weight of a charge is often ascertained by allowing a" vessel of known capacity, say a large cubical or cylin- drical box, to be filled at the same time and in the same manner as the tank or other container in which the material is to be treated. When the charge is completed, the surface is leveled off without pressing down and the box and contents weighed on a platform scale or suitable balance. The entire charge is also leveled and its volume computed from the known dimensions of the containing vessel. In the case of a cylindrical tank, V (volume) T r*h. where r = radius of tank. h = average depth charged with material. (22 \ = -=- approximately J In the case of a conical vessel, 3 where r = radius of circle forming upper surface of mass of material filling cone. h = depth from surface to apex of cone. 3.14159 f = y approximately V 477 478 THE CYANIDE HANDBOOK After weighing the sample, a sufficient quantity, say 1 or 2 kg. (4 or 5 Ib.) is weighed separately, spread out in a thin layer on a metal tray, and dried at a moderate heat. After drying the tray and contents are re-weighed and the percentage of moisture calculated. The wet weight of the charge is then easily calculated as fol- lows: Let W = wet weight of charge. v = volume of sample. w = wet weight of sample. h = depth charged in tank (average). r = radius of tank. Then v:irr*h: :w: W. w " r * hw or r The value of h is deduced from the average of a number of measurements taken at different parts of the surface to deter- mine the depth left empty between the top of the tank and the charge. The dry weight of the charge, D, is calculated from p } the per- centage of moisture, as follows: D:TT::100 - p: 100 or D (b) To find the weight of a given volume of wet pulp (e.g., a charge of slime). In cases where the material is of such a nature that all the interstices between the particles are filled with liquid, as in a charge of slime mixed with water or cyanide solution, the total dry weight may be calculated from the following data: V = total volume of charge. P = density (specific gravity) of pulp. S = density (specific gravity) of the dry material. The total volume (V) is determined as already described, from the known dimensions of the containing vessels. We may define the density or specific gravity of a substance as the ratio between its weight and the weight of an equal volume of water at a tem- perature of 4 C. At that temperature the weight of 1 cc. of water is 1 gram; hence the density of any liquid or solid is the DENSITY DETERMINATIONS 479 weight in grams of a measured quantity, divided by the number of cubic centimeters which it occupies. The density of the pulp is determined by filling a vessel of exactly known volume, for instance a liter flask, with an aver- age sample of the pulp and weighing its contents. The volume of the containing vessel need not be known, if the weight of water filling it to the mark be ascertained. In this case a correction is required for the expansion of water if the determination be made at any other temperature than 4 C., though this is negli- gible for most practical purposes. The density of the dry substance of which the pulp is formed is generally found by weighing a certain amount of a carefully dried average sample, placing it in a vessel of known capacity and filling up to the mark with distilled water. The weight of the vessel filled with water alone is also determined. Before weighing the vessel containing the material and water, it is neces- sary in some cases to immerse it for one or two hours in boiling water up to the neck, leaving the stopper out, to allow the expul- sion of any air contained in the dry powder. It is then cooled and weighed. The density of the dry material may then be cal- culated as follows: Let m - weight of dry substance taken. a = weight of water filling vessel to mark when no substance is added. 6 = weight of substance together with amount of water required to fill vessel to mark after substance has been added. c, = weight of residual water. d weight of water displaced by dry material s = density of dry material. Then b = m + c d = a c mm m d a c m + a b Or let W\ = weight of vessel when filled with water to mark. TF 2 = weight of vessel with dry material, filled up with water to mark. Wo = weight of empty vessel Then a = W l - W b=W 2 -W m and s TIT ifr m + rr i rrj The weight of dry material in the total charge may then be calculated from the known values of P and S. 480 THE CYANIDE HANDBOOK Let P density of pulp. S = density of dry material V = total volume of pulp m = total weight of dry material in pulp c = weight of water in pulp w = weight of a unit volume of water From (1) c = VwP - m From (2) c = - ^ Hence m = This formula is subject to the same correction for expansion of water which is involved in calculating the value of P. (c) Useful data deducible from the above formula}. 1 (1) Percentage of dry material in a charge of pulp. Let p = percentage of dry material Then p: 100: : m: m + c (100 - p)m Hence c = - P ,, m + c m + cS From (1) and (2) above: Vw = -p = ^ - and p m (S - P) . 100 - p S-P S (P-l}' henCG F- 100 S(P- 1) _ (2) Percentage of solution in a charge of pulp. q = percentage of solution 100 QS -P) * = ~ p= (3) Ratio of solution to dry material (R). r P ~ A + B (3) To find weight of one solution which must be added to another to make it up to required strength. Using same symbols as above, orB= ^n4 A 498 THE CYANIDE HANDBOOK (4) Formulae for converting percentages into results per ton. P = percentage 20P = Ib. av. per ton of 2000 Ib. 22.4 P = Ib. av. per ton of 2240 Ib. 10P = kg. per ton of 1000 kg. = grams per kg. L^. p = iip p = oz. Troy per ton of 2000 lb. oz. Troy per ton of 2240 lb- dwt. per ton of 2000 lb. -P = dwt. per ton of 2240 lb. CONVERSION OF METRIC WEIGHTS AND MEASURES APPROXIMATELY 1 meter = 39.370432 inches = 39.4 = 3.280869 feet = 3.3 " = 1.093623 yard = 1.1 1 inch = 2.539977 centimeters = 2.54 1 foot = 0.304797 meter = 0.3 1 yard = 0.914392 meter = 0.9 1 cubic meter = 35.3156 cubic feet = 35.3 1 liter = 61.0254 cubic inches = 61 " = 0.035316 cubic feet = ^ 1 cubic foot = 0.0283161 cubic meters = ^ 1 gram = 15.43234874 grains = 15.4 " = 0.643015 dwt. = | = 0.032151 oz. Troy = & 1 kilogram = 2.20462125 lb. av. =2.2 1 grain = 0.06479895 gram = 0.065 1 dwt. = 1.555176 gram = 1 1 oz. Troy = 31.10352 grams = 31 1 lb. Av. = 453.592652 grams = 454 1 ton of 1000 kg. = 2204.621 lb. Av. = 2205 1 ton of 2000 lb. = 907.185 kilograms = 907 1 ton of 2240 lb. = 1016.048 kilograms = 1016 1 gram per ton of 1000 kg. =14 grains per ton of 2000 lb. 1 " = _^ dwt. tt it < t 1 " " " . " __ O7 tt tt a tt 1 dwt. per ton of 2000 lb. = ty gram per ton of 1000 kg. 1 oz. per ton of 2000 lb. = -^ grams per ton of 1000 kg. MISCELLANEOUS DATA 1 ton of 2000 lb. = 291 66f oz. Troy 1 lb. Av. = 14.58333 oz. Troy 1 " " = W oz. Troy = 1^ dwt. = 29 If dwt. USEFUL FORMULAE 499 1 Ib. Av. 1 oz. Troy 1 oz. Av. 1 Ib. per ton of 2000 Ib. 1 kg. per ton of 1000 kg. 1 Ib. Av. of water 1 gallon of water 1 gallon (U. S) of water 1 cubic foot of water ' 7000 grains 480 grains : 437.5 grains 0.05 per cent. 0.5 kilo per ton of 1000 kg. 0.1 per cent. 2 Ib. per ton of 2000 Ib. 27.68122 cubic inches 10 Ib. = 276.8122 cubic inches 8.345 Ib. = 231 cubic inches 62.425 Ib. = 62^ approx. ATOMIC WEIGHTS (1909) Aluminum Al Antimony Sb Argon A Arsenic As Barium Ba Bismuth Bi Boron B Bromine Br Cadmium Cd Caesium Cs Calcium Ca Carbon C Cerium Ce Chlorine Cl Chromium Cr Cobalt "... .Co Columbium Cb Copper Cu Dysprosium Dy Erbium Er Europium Eu Fluorine . . .S. F Gadolinum Gd Gallium . . Ga Germanium Ge Glucinum Gl Gold Au Helium He Hydrogen H Indium In Iodine I Iridium . . .Ir 27.1 Iron Fe 55.85 120.2 Krypton Kr 81.8 39.9 Lanthanum La 139 75.0 Lead . Pb 207.1 137.37 Lithium Li 7 208 Lutecium Lu 174 11 Magnesium Mg 24.32 79.92 Manganese Mn 54.93 112.40 Mercury Hg 200 132.9 Molybdenum Mo 96 40.09 Neodymium Nd 144.3 12 Neon Ne 20 140.25 Nickel. Ni 58.68 35.46 Nitrogen N 14.01 52.1 Osmium Os 190.9 58.97 Oxygen O 16 93.5 Palladium Pd 106.7 63.57 Phosphorus P 31 162.5 Platinum Pt 195 167.4 Potassium K 39.1 152 Praseodymium Pr 140.6 19 Radium Ra 226.4 157.3 Rhodium Rh 102.9 69.9 Rubidium , Rb 85.45 72.5 Ruthenium Ru 101.7 9.1 Samarium Sa 150.4 197.2 Scandium Sc 44.1 4 Selenium Se 79.2 1.008 Silicon Si 28.3 114.8 Silver Ag 107.88 126.92 Sodium Na 23 193.1 Strontium . .Sr 87.62 500 THE CYANIDE HANDBOOK Sulphur S 32.09 Tantalum Ta 181 Tellurium Te 127.5 Terbium Tb 159.2 Thallium Tl 204 Thorium Th 232.42 Thulium Tm 168.5 Tin Sn 119 Titanium . . .Ti 48.1 Tungsten W 184 Uranium U 238.5 Vanadium V 51.2 Xenon Xe 128 Ytterbium Yb 172 Yttrium Y 89 Zinc Zn 65.37 Zirconium Zr 90.6 INDEX PAGE Aaron, C. H., assay of pyritic ores 362 Abosso Gold Mining Co., flux for zinc precipitate 267 Absorption of solution by wooden vats 199 Accumulation of zinc in solutions 39, 256 Accumulations of tailings, handling of 159 Acid radicals, estimation of, in cyanide 456 in solutions 440 Acid-treated precipitate, composition of 263 Acid treatment of concentrates 194 of cupriferous ores 310 of zinc-gold precipitate 41, 123, 261 use of sodium bisulphate in 123 Acidity of ores, estimation of 483 of water, estimation of 468 Aeration of slime pulp 210 tests 494 Agitation by air 210 by mechanical stirrers 208 by pumps 210 process of cyanide treatment. .33, 36, 208 Alexander, F., stamp milling practice . 144 Alkali consumption tests 483 estimation of, in solutions 451 in water 468 hydrates, estimation of 451 influence of, in zinc precipitation . . 249 on extraction 109 Alkali-metal cyanides, properties of . . 65 Alkali metals, estimation of 397 Alkali, strength required in slime treat- ment 257 washes in sand treatment 201 use of, in cyanide process 27, 201 All-sliming process 283 Alloys, analysis of 430 Aluminium, estimatio/f'of, in ores, etc. . 398 in water 473 precipitation with 122 use of, as anode 303 Amalgam, method of obtaining from plates 181 Amalgamation, conditions of good . 180, 181 cyanide as an aid to 21 cyanide in, of concentrates 16 effect of fine crushing on 144 metals on 182 minerals on 182 films on plates in 182 PAGE Amalgamation, foreign matter intro- duced during 185 in relation to cyaniding 179 inside 180 pan 180 plate 180 previous to cyanide treatment 34 use of chemicals in 183 use of cyanide in 14, 21, 184 Ammonia, estimation of, in cyanide solutions 451 manufacture of cyanide from "130 use of, as preliminary solvent for cupriferous ores 112, 311 Ammonium argentocyanide 76 compounds, estimation of 451 manufacture of cyanide from ... 130 Ammonium cyanide, preparation of . 64 properties and reactions 64 Ammonium, cyanides of copper and . 78 estimation of, in cyanides 460 in solutions 451 thiocyanate 100 Analytical operations 393 Andreoli, E., peroxidized* lead anodes. 300 Annealing cornets in bullion assay . . 383 Anodes in electric precipitation 121 in Siemens-Halske process 299 iron 121, 299, 300 peroxidized lead 121, 300 Anticipation of cyanide process 10-24 Antimonial ores, assay of 366 Antimony, effect of, in amalgamation 182 in zinc precipitation 254 estimation of 399, 400 separation of 399 Argall, P., conditions of zinc precipi- tation 119 modern practice with rolls 146 nature of zinc-box deposits 254 power required for Ball mills 154 roasting furnace 288 rolls H8 Argentic cyanide 74 Argentocyanic acid 76 Argento-potassic cyanide 75 Arrastra 150 Arsenic detection of 403 estimation of 401 501 502 INDEX Arsenic in amalgamation 182 in zinc precipitate 123, 254 separation of 402 Arsenical ores, assay of 365 cyanide treatment of 295 Artificial oxidizers, use of, in cyaniding 103 Ash, estimation of, in coal 466 Assay, arrangement of 339 balance, adjustment of 347 crucible fire 333, 354 of antimonial ores 366 of arsenical ores 365 of basic ores 358 of battery scrapings 376 of cupels 374 of cupriferous ores 367 of gold and silver bullion 378 of graphite crucibles 354 of lead ores 365 of pyritic concentrates 361 of pyritic ores 360 of refractory materials 371 of siliceous ores 333, 355 of slags 373 of telluride ores 369 of zinc ores 365 quantities of flux in 335 quantities of ore in 333 scorification 349 silver, by volumetric methods 389 special methods of 349 weights, accuracy of 386 Assaying 315 with metallic iron 362 with niter 363 Atomic weights (1909) 499 Auriammonic cyanide 74 Auriargentic cyanide 74 Auric cyanide 72 Auripotassic cyanide 73 thiocyanate 101 Aurisodic thiocyanate 101 Auropotassic cyanide 73 thiocyanate 101 Aurosodic cyanide 73 Aurous cyanide 72 double 72 Automatic lime feeder 178 pulp distributor 166 sampler 319 Auxiliary dissolving agents 293 Bagration, Prince P.; solubility of gold in potassium cyanide and ferro- cyanide 10 Balbach Smelting and Refining Co., treatment of zinc-gold precip- itate 272 Ball mills . . 152 PAGE Ball mills, power required for 154 Banket ore, use of in Tube mills . . 156, 157 Banks, E. G., grading test of tube mill product 157 Barium argentocyanide 76 cyanide 68 estimation of 404 Barker, H. A., precipitation in pres- ence of copper 120 precipitation of copper from cyanide solution by means of acids .... 312 Base metal compounds, action of, on cyanide 110 Base metals, estimation of, inbullion . . 433 Basic ores, assay of 358 Battery practice 143, 180 scrapings, assay of 376 Bay and Prister, analysis of white precipitate 119 Begeer, B. W., assay of low grade samples 356 quartz ores 355 fluxing of zinc precipitate 267 Belt concentrating tables 190 conveyors 167 Bergman and Guyton, researches on prussic acid 4 Beringer, C. and J. (text book of assay- ing), assay of basic ores 358 cyanide tailings 357 siliceous ores 355 estimation of arsenic 402 copper 411 iron 414 lead 415 manganese 418 mercury 418 tin 426 separation of gold and silver in bullion by means of cadmium 431 Berthelot, cyanide chemistry 9 Berthollet, chloride of cyanogen 5 composition of prussic acid 5 Berzelius, cyanide chemistry 9 Bertram Hunt filter 233 use of ammonia in treatment of cupriferous ore 313 Bettel and Marais, effect of oxidizers on cyanide extraction 103 Bettel, W., reaction of oxygen and cyanide on gold 102 Bicarbonate of soda, use as flux 335 Bicarbonates, estimation in cyanide . 456 in solutions 448 in water 469 Bismuth, effect of, in amalgamation . . 182 estimation of 405 Black Hills (South Dakota), Merrill filter press at 220 INDEX 503 PAGE Black Hills, system of treatment at . . 284 zinc-dust precipitation at 306, 308 Blake crusher 135 Blake-Marsden crusher 135 Blaisdell excavator 207 Blanketings 187 Bodlander, reaction of gold and cyanide. 102 Bonanza mine, Tavener's process at. . 272 Bosqui, F. L., absorption of solution by wooden vats 199 Butters filter press costs 225 Merrill filter press 221 Bromide of cyanogen, action of, on gold and silver 59 action of, on metals 59 on cyanide 59 mono- , 57 physical properties of 58 polymer of 59 preparation of 57 reactions of 58 Bromocyanide practice in W. Aus- tralia 295 process 295 Brown, F. C., air-lift agitator 211 R. Oilman, cyanide practice with the Moore filter 222 W. S. f treatment of cupriferous ores 310 Brunton, W., traveling belt filter . . . 226 D. W., theory of sampling 320 Bryan mill 150 Buckboard 323 Buddies 187 Bueb , J . , process of cyanide manufacture 131 Bull, Irving C., estimation of lead . . . 415 Bullion, analysis of 430 assay of gold and silver 378 corrections in assay . , 387 from cyanide process 42, 274 smelting of 270 volumetric assay of silver in 389 Bunsen, R. W., manufacture of cyanide from atmospheric nitrogen .... 128 Butters and Mein, automatic distributor 166 Butters, Charles, bottom discharge door 206 deep tanks fot^slime settlement 214 filter 224 intermediate collecting tanks 166 reverberatory furnace for smelting zinc precipitate 268 Robinson slime plant 221 sheet lead anodes \ . 300 Butters Copala Syndicate, slime fil- tration 225 Cadmium, estimation of 405 use of, in separating gold and silver . 431 Calcium cyanide, preparation and properties 69 PAGE Calcium, estimation of, in metallic products 434 in ores 406 in solutions 446 in water 472 separation of 405 Calculation of cyanide analyses 461 of tank contents 477 of sand charges 477 of slime pulp densities 479 of specific gravities 479 of strength of solutions 496 Caldecott, W. A., alkali strength of slimes solutions 257 circular vats for zinc precipitation 245 influence of strength of solution ... 108 roasting zinc precipitate with niter. 264 slime pulp formulae 480 Caldecott, W. A., and Johnson, E. H. f precipitating surface of zinc shavings 250 use of lead-zinc shavings 253 Callow cone settling tank 176 screen -. 161 Calorific power of coal, estimation of 466 Cams 138 Cam-shaft 138 Carbon, estimation of, in coal 466 organic, estimation of, in water . . . 473 Carbonaceous residues, assay of 377 Carbonate of ammonia, use of, in con- junction with cyanide 22 of copper, action of, on cyanide . . 112 of soda, use as assay flux 335, 357 use as flux for zinc precipitate . . 125 Carbonates, estimation in cyanide . . 456 solutions 448 water 469 Carbonic acid, action on cyanide .... 109 estimation in ores 407 water 469, 471 Carter, T. L., assay of graphite crucibles 354 fluxing of zinc precipitate 266 lead acetate drip 249 zinc-lead couple 247 Cassel filter 225 gold extracting process 24 H. R., colorimetric estimation of gold 445 Castner, manufacture of cyanide .... 130 Cathodes, amalgamated copper 303 in electric precipitation 121 in Siemens-Halske process 121, 299 lead 121,300 tinned iron 121, 301 Caustic lime, estimation of 464 Caustic soda, use of, in amalgamation . 184 Central City (S. Dakota), crushing with cyanide solution 281 Charcoal precipitation 122, 308 504 INDEX PAGE Charcoal, use as reducing agent in assaying 334 Checks, for bullion assays 380 Chemical losses of zinc 255 Chemistry 45 of dissolving process 102 of precipitation process 117 of smelting process 124 Chilian mill 149 Chloride of cyanogen 5, 56 gaseous 56 solid 57 Chloride of sodium, use of, in con- junction with cyanide 21 Chlorides, estimation of, in solutions 449 Chlorine, estimation of, in ores 407 in solutions 449 in water 472 Christy, S. B., charcoal precipita- tion 308 electrochemical order of metals in cyanide solutions 106 electromotive force of minerals in cyanide solutions 106 estimation of gold in solutions .... 443 precipitation of cyanide solutions by cuprous salts 123, 309 Chromicyanides 94 Chromium, cyanides of 71 estimation of 408 Chromocyanides 94 Clancy, continuous system of slime settlement 215 Clarification of solutions before pre- cipitation 38-236 Clarifying tanks 236 Clark, Thos. C., patent 21 Classifiers, hydraulic 34, 172 Claudet, A. C., cupellation of bullion . . 381 platinum parting apparatus 384 surcharge of bullion assays 387 Clean-up of zinc-boxes 258 Clean-up tanks 259 Clevenger, G. H., treatment of zinc- gold precipitate with litharge . . 272 Clouet, artificial production of cyanides from ammonia 6 Coagulating agents, for settling slimes . . 177 Coal, analysis of 465 estimation of ash in 466 of calorific power of 466 of carbon in 466 of iron and phosphorus in 467 of moisture in 465 of sulphur in 466 of volatile matter in 465 Cobalt, cyanide of 71 estimation of 409 Cobalticyanides and cobaltocyanides . 94 PAGE Cobar (N. S. W.), treatment of cuprif- erous ores at 310 Collecting tanks 166 Collection of sands by direct filling . 165 by intermediate filling 166 slimes 35, 208 Colorimetric estimation of chromium 408 of copper 412 of gold 445 of iron 414 of selenium 423, 450 of sulphides 449 of thiocyanates 442 of titanium 427 Combination mill (Nevada), slime treat- ment 225 Commercial cyanide, analysis of .... 455 sampling of 455 Complex cyanogen radicals, metallic compounds of 84 Comet crusher 135 Concentrates, amalgamation of, by aid of cyanide 16 cyanide treatment of 193 Concentrating tables 187 Concentration by oil 191 definition of 186 general principles of 185 in relation to cyanide treatment . . 185 previous to cyanide treatment .... 192 tests 487 Concentrators, belt 190 Buddie 187: modern types of 188 old types of 187 percussion 188 Wilfley 188 Conditions, affecting size of product in Stamp mills 141 determining separation of particles in a flowing stream 172 for coarse crushing 141 for effective percolation 200 for good precipitation 38, 248 Cone classifiers 175 calculating volume of 477 Conical slime treatment vats 211, 284 Coning and quartering 317 Consumption of cyanide, causes of . . 108 tests to determine 490 Consumption of zinc 41 Conveyors 163, 164, 167, 171 Copper, action of, in zinc precipitation 120 and ammonium, double cyanides of 78 and potassium, double cyanides of . . 79 anodes, amalgamated 303 compounds, action of, on cyanide . Ill cyanides of 76 detection of . . . 409 INDEX 505 PAGE Copper, double cyanides of 78 use as solvents 311 estimation of, in ores, etc 411 in solutions 446 ferricyanides 92 ferrocyanides 88 ores, treatment of 301 removal of, before cyanide treat- ment 112, 310 separation of, in analysis..- 410 Cornets, gold, annealing of - 383 treatment of, in assaying 384 Corrections for bullion assays 387 Costs of treatment with Butters' filter 225 Cowper-Coles, S., aluminium cathodes . 303 Cripple Creek (Colo.), crushing with rolls 154 assay of telluride ores from 370 Croghan, E. H., analysis of lime 464 composition of commercial lime . . . 463 Crosse, Andrew F., assay of battery scrapings 376 assay of graphite crucibles 375 slime-treatment process 211 Crown mine (Karangahake, New Zea- land), crushing with cyanide solu- tion 281 early cyanide trials at 30 Crucible fusions 337, 354 Crucibles for assays 336 Crush filters, fixed immersed type . . . 220 traversing type 226 Crushing fine, effect on amalgamation. 144 machinery 135 of ores as affecting cyanidation. . . . 143 samples for analysis 395 for assay ~ 323 Crushing with cyanide solution 281 advantages of 281 objections to 282 Black Hills practice 284 Mexico practice 285 Nevada practice 284 Rand practice 282 United States practice 284 Gumming, A. C., and Masson, O., estimation of cyanates 458 Cupellation assays 340 of bullion 381 details of 342 effect of temperature in 342 losses of gold and silver in 342, 380 of lead from treatment of zinc pre- cipitate 272 Cupels 340 assay of 374 making bone ash 340 size and form of 341 PAGE Cupric cyanide, preparation and prop- erties 77 ferricyanide 92 Cupriferous ores, assay of 367 difficulties in treatment of 310 preliminary acid, treatment of 310 use of ammonia in treating 313 Cuproso-cupric cyanides 77 Cuprous cyanide, preparation and prop- erties 76 ferricyanide 92 salts, precipitation by 123, 308 Cut samples of bullion 378 Cyanamides as intermediate products in cyanide manufacture 129 Cyanates, action of acids on 97 estimation of 457 general properties and reactions of . . 96 metallic 96 of potassium 97 of silver 97 Cyanic acid, discovery of 8 preparation and properties 96 Cyanicides 108 Cyanide, as aid to amalgamation ... 14, 21 as solvent with other chemicals ... 22 as solvent of metallic sulphides .... 16 extraction tests 490 in conjunction with electric current.10, 15 manufacture of 9, 127 relative solubility of metals in 17 solubility of gold in 6 solubility of silver in 20 solutions, analysis of 437 use of, in amalgamation 14, 183 Cyanide process, early history of .... 3 importance of, in metallurgy 31 introduction of 30 later developments of 31 outline of operations in 33 stages of 33 Cyanides, artificial formation of 6 Proust's researches on 6 metallic 62 Cyanippus filter press 217 Cyanogen, action of acids on ... 50 action of alkalis on 50 action of metals on 49 bromide of 57 chemical characteristics of 48 Cyanogen compounds, early use of . . 3 estimation of 440 with non-metals 47, 56 Cyanogen, Davy's researches on 8 estimation of 434 in cyanide 456 formation of 47 Gay-Lussac's researches on 7 physical properties of . 48 506 INDEX PAGE Cyanogen, Proust's researches on .... '7 reactions of 49 use of, in gold extraction 26 Cyanurates 98 Cyanuric acid 98 Daniell, J. , scorification in fusion furnace 353 theory of scorification 352 Davy, Sir H., discovery of iodide of cyanogen 8 electrolysis of hydrocyanic acid ... 54 researches on cyanogen 8 Decantation of solution from slime . . 209 process 36, 209 defects of 213 treatment of slimes 212 Decomposition of cyanide by car- bonic acid 109 by copper salts Ill by external influences 108 by iron salts 110 by oxygen 109 by sulphides 113 determination of 490 Deep tanks for slime settlement '214 Dehne filler press 217 Deloro mine (Ontario, Canada), bro- mocyanide treatment at 295 zinc-dust precipitation at 306 Deniges, G., estimation of silver 433 Denny, G. A., sampling of ore in mine. 321 G. A. and H. S., crushing with cyan- ide solution 282 Density determinations of sand 477 of slime 478 Deposits in zinc-boxes 118, 252 in electric precipitation 121, 299 Desert mill (Nevada), cyanide prac- tice at 284 Deutsche Gold-Silber Scheide-Anstalt, process of cyanide manufacture 130 Dewey, F. P., losses in cupellation. . . 344 Diehl process 295 Diesbach, discovery of Prussian blue . . 3 Dilute solutions, claim for use of 26 selective action of 107 Dip samples of bullion 378 Direct cyanide treatment 170, 279 Direct filling of sand tanks 165 Discharge doors for cyanide tanks . . 200 Butters 206 James 206 Discharging residues 206 from bottom of tank 206 from side of tank 206 Disposal of residues 38, 167, 206 solutions 39, 235 Dissolved solids, estimation of, in solu- tions . 453 PAGE Dissolving process 33, 102, 195 tanks 202, 236 Distributor for charging pulp into tanks 166 Dittmar and MacArthur, estimation of potassium 422 Dixon, Clement, details of filler press construction and practice 217 Dixon, W. A., action of oxidizers on metallic gold in cyanide solution . 18 precipitation of gold from cyanide solutions 19 treatment of pyritic ores with cyano- gen compounds and oxidizers. . 18 Dodge crushers 135 Donath, E. and Margoshes, B. M., es- timation of ferrocyanides 435 Double treatment 166, 205 Dowling, W. R., tube mill practice 155 Drill samples of bullion 378 Drip of cyanide at head of zinc-boxes . . 249 of lead acetate in zinc-boxes 249 Drucker, A. E., treatment of matte. . . 271 Dry crushing 279 advantages of 280 Duflos, Dr., early cyanide extraction experiment 12 Durant, H. T., estimation of gold in solutions 444 suction valve on centrifugal pump. . 210 Dures, R., assay of battery scrapings. . 376 assay of graphite crucibles 375 Early cyanide, extraction test 12 history of the cyanide process 3 use of cyanogen compounds 3 use of suction filters 208 Edwards' roasting furnace 286 Ehrmann, L., sampling slimes 332 Electric current, influence of, on extrac- tion 10,15 precipitation, advantages of 40 Siemens -Halske process 40, 121, 289 Electrochemical order of metals in cyanide solutions, 18, 106 Electrolysis of cyanide solutionslO,40,121, 298 Elkington's process of 10 Rae's process of 15 Electrolytic processes 10, 40, 121, 298 recent developments of 300 Electromotive order of metals in cyanide 18 of minerals in cyanide 106 Electro-zinc process 303 Elkington, J. R., and H., patent 10, 298 Ellis, C. J., estimation of cyanurates. . 441 oxalates 452 sulphides 450 urea 452 Elmore oil concentration process 191 INDEX 507 PAGE Eisner, L., classification of metals by their reactions with cyanide .... 12 influence of oxygen in action of met- als on cyanide 11 Eisner's equation 102 Endlich and Muhlenberger, alleged an- ticipations of cyanide process ... 23 Erlwein and Frank, manufacture of cyanide 129 Eschka,' determination of sulphur in coal 466 Evans-Waddell Chilian mill 149 Everitt, action of acids on ferrocyanides 9 Excavator, Blaisdall 207 Extraction by cyanide process 38 by successive decantations 213 effect of oxidizers on 103 influence of solution strength on . . 108 tests to determine 490 Faraday, M., researches on solubility of gold-leaf in cyanide 13 Faucett, H. J., patent 21 Feldtmann, W. R., double treatment of tailings 205 Feldtmann, W. R., and Bettel, W., con- stitution of cyanates 97 Ferricyanide of copper 92 of iron 92 of potassium 90 properties and reactions 91 of silver 93 Ferricyanides, characteristic colors of 90 estimation of 442 use as auxiliary solvents 294 gold solvents 104 Ferrocyanide, ferric 89 of hydrogen v 88 of potassium 86 of sodium 89 of zinc 89 Ferrocyanides, action of acids on .... 9, 85 action of HsS on 85 characteristic insoluble 87 cuprous and cupric 88 discovery of /('. 4 electrolysis of 7, 85 estimation of 435 in cyanide 459 estimation of, in solutions 441 general characteristics 85 of copper 88 Porrett's researches on 7 Proust's researches on 6 solvent action of, on gold 10 Ferroso-ferric ferricyanide 92 Ferrous ferricyanide 92 ferrocyanide 88 potassium ferrocyanide 88 PAGE Ferrous salts, action of, on cyanide . Ill Films, on amalgamated plates 182 Filter cloths 199 frames 199 Filtering slimes 216, 221 Filter plates, construction of 222 Filter presses 37, 216 capacity of 220 clarifying solution by 236 cyanippus 217 Dehne 217 details of construction 217 general principles of 216 Merrill 220 method of discharging 220 method of filling 219 treatment of slime in 220 use of compressed air with 220 washing cakes in 220 Filter pressing of ore slimes 37, 216 of zinc-dust precipitate 307 of zinc-gold precipitate 260, 262 Filters, Bertram Hunt 233 Brunton 226 Butters 224 Crush 226 Moore 222 Ridgway 230 Filtration by pressure 216 by suction 221 of slimes 37, 216 vacuum 221 Fine crushing, for cyanide treatment. 37, 154 grinding, advantages of 208 metal, in assaying, cleaning and weighing 344 Finkener, method for estimation of potassium 422 Flint mill 37 Florence (Colorado), acid treatment of zinc precipitate at 262 Flotation process of concentrating ore . 192 Fluorine, estimation of 413 Fluor-spar, use of, as flux 125 Fluxes, acid and basic 334 for crucible assays 334, 355 for free-milling ores 355 for smelting cine precipitate 124, 266 quality of assay 335 standard (for siliceous ores) 334 Formulae (useful) 496 Frankland, growth of cyanide indus- try 31 Free cyanide, estimation of 438 influence of, on zinc precipitation . . 248 Frenier spiral sand pump 169 Fresenius, C. R. (quantitative analysis), estimation of cyanogen 435 Frue vanner l&l 508 INDEX PAGE Fulminates 98 Fulton, C. H., assay of siliceous ores. . 355 and Crawford, C. H., assay of zinc- box precipitate 376 Furnace, Argall 288 Edwards 286 Merton 287 Furnaces for roasting ores 286 for roasting zinc precipitate 268 revolving cylindrical 288 with revolving rabbles 286 Fusion in hot crucibles 339 in muffle 338 of zinc precipitate 42, 268 process for assays 337 with excess of litharge 361 with niter, for zinc precipitate 270 Fusions, pouring, in assaying 338 Gates, rock-breaker 136 Gay-Lussac, analysis of cyanogen ... 7 prussic acid 7 assay of silver 389 Gaze, W. H., proportion of zinc required per ton of solution 250 George and May mine (Johannes- burg) , direct treatment of coarse- crushed ore at 171, 279 Gernet, A. von, electrolytic precipita- tion 40,298 electrolytic precipitation on mercury 303 introduction of the electric process. . 40 losses of gold in mill water 184 treatment of cupriferous ores 311 Gilmour- Young process 122, 304, 306 Glassford and Napier, researches on cyanides and double cyanides. . 11 Glenn, W., sizes allowable with samples of different weights . . . 320 Gold, action of cyanide on 73, 102 analysis of bullion 431 assay of, in ores, tailings, etc 315 bullion assay 378 coated, in amalgamation 183 cyanides 71 cyanides of, and ammonium 74 and potassium 73 and silver, researches on 4, 11 and sodium 73 estimation of, in solutions 443 leaf, experiments on action of cyanide with 13 loss of, in mill water 184 ; thiocyanates of, and potassium 101 Gopner, production of bromocyanide . . 58 .Gowland, W., surcharge in bullion assays 387 .Grading tests 485 of Ball mill products 154 PAGE Grading tests of Rolls products 147 of Tube mill products 157 reporting 486 Graham, K. L., comparative test of banket and pebbles in Tube mill 157 Graphite crucibles, crucible assay of ... 375 scorification assay of 354 use of, in smelting 268 Great Boulder mine (W. Australia), Ridgway filter at 232 Green, L. M., estimation of protective alkali 439 estimation of alkaline hydrates .... 451 Grinding machinery 146 mills 148 Grizzly 141 Gross, J., fluxing of zinc precipitate. . . 267 Growth of cyanide industry 31 Gyratory crushers 136 Hagen, solubility of gold in cyanide ... 6 Hahn, H. C., cyanide as a solvent of metallic sulphides 16 Haloid cyanogen compounds, discovery of 5, 9 estimation of 441 preparation and properties 56 use of, in cyanide process 59, 295 Hall and Popper, assay of zinc ores . . . 365 Hamilton, E. M., practice with Butters filter 224 Hammering lead buttons 338 Handling of material for cyanide treat- ment 159 of old accumulations of tailings 159 of residues 206 of solutions 235 Handy, J. O., volumetric estimation of magnesium 416 Hannan's Star Mill (W. Australia), bromocyanide practice at 297 Hardness of water, estimation of 471 Harvey, A., sampling of ore 319 Heat of combination of metals with oxygen 125 Heating solutions 239 Herting, O., estimation of cyanates . 457 Hillebrand and Allen, action of nitrous acid on gold 347 assay of cupels 374 slags from telluride ore 374 telluride ores 369 losses in cupellation 342 Hirsching, H., use of ammonia in treatment of cupriferous ores. . 311 History of cyanide process 3 of Mac Arthur-Forrest discovery. ... 24 Hogenraad, G. B., fluxes for siliceous ores . . . 356 INDEX 509 PAGE Hood process 297 Hose filling of sand tanks 165 Hot solutions, advantages of 106 tests made with 239 Hunt filter 233 Huntington mill 150 improved 151 Hydrates, estimation of, in cyanide . . 456 in water 470 Hydraulic classification 34, 172 classifiers 172 separation tests 485 Hydrochloric acid, use of, in treating zinc precipitate 262 Hydrocyanic acid 51 action of acids on 54 action of alkalis on 54 action of metallic salts on . 55 action of non-metals on 54 action of oxides on 55 antidote for 55 electrolysis of 54 estimation of 440 occurrence of, in nature 53 physical properties of 53 poisonous nature of 55 polymer of 55 preparation of 51 reactions of 54 synthesis of 52 Hydroferricyanic acid 92 Hydroferrocyanic acid 88 Hydrogen, ferricyanide of 92 evolution of, in zinc precipitation . . . 252 evolution of, in zinc precipitation . . . 252 ferrocyanide of 88 Hydrogen peroxide, occurrence of, as intermediate product in reaction of cyanide on gold 102 Hydrolysis of cyanide 108 Immersed type of vacuum filter .... 229 Indicators, action of acids and alkalis on 468 Intermediate filling of sand tanks 166 Inquartation . .^ 345 Insoluble matter in cyanide, estimation of . 461 Iodide method of copper assay 411 Iodide of cyanogen, discovery of 8, 9 preparation and properties of 59 reactions of 60 Iodine method of zinc estimation . . . 447 Iron anodes 121, 299 cathodes (tinned) 121, 301 compounds, action of cyanides on . 110 cyanogen radicals 85 estimation of, in ores, etc 414 in bullion . . .433 PAGE Iron, estimation of, in coal 467 in solutions 447 in water 473 ferrocyanides 88 group, cyanides of 71 pyrites, action on cyanide 110 salts, action on cyanide 110 separation of 413 tanks 197 use of, in assaying 362 Ittner, researches on prussic acid ... 7 Ivanhoe Gold Corporation, bromocy- anide practice 296 James, Alfred, discharge door 206 influence of foreign salts on precipi- tation with zinc 249 Janin, Louis, Jr., alleged anticipitation of cyanide process 23 criticism of cyanide process 23 solubility of silver in cyanide ..... 105 Jarman and Brereton, use of ammonia as auxiliary solvent in cyanide treatment 313 Jaw crushers 135 Jenkens, H. C., antidote for cyanide poisoning 55 Jennings, Hennen, hose filling of tanks 166 Joannis, chemistry of cyanide com- pounds 47 electrolysis of ferrocyanide 86 Johnson, E. H., stamp milling in rela- tion to cyaniding 144 Johnson, E. H., and Caldecott, W. A., fluxing of zinc-gold precipitate 267 Julian, H. F., reaction of cyanide on gold 104 Julian, H. F., and Smart, E., "Cyanid- ing Gold and Silver Ores" au- tomatic lime feeder 178 Ball mill practice 153 Butters filter-beams 225 coagulating agents for slime 177 composition of ferric hydrate Ill discharge doors for leaching tanks . 200 excavators for discharging tanks . . 207 experiments on hot solutions 239 falling particles in water, laws of . 173 filter press practice 220 fluxing of zinc precipitate 266 leaching vats 197 Merton furnace, details of 288 sand pumps 170 slimy sand, treatment of 160 zinc-box construction 245 Kalgurli (Kalgoorlie, W. Australia), Ball mill practice at 154 bromocyanide practice at 295 510 INDEX PAGE Karangahake (New Zealand), establish- ment of cyanide process at, in 1889 30 Keith and Hood, use of mercuric chloride as auxiliary solvent. . . 297 Kendall process 293 Knublauch, manufacture of cyanides from coal-gas 131 Krupp-Grusonwerk Ball mill 153 Lake View Consols, filter presses at . 219 Lamb, M. R., agitation with air 210 Latent acidity 483 Lay, Douglas, electrolytic processes . . . 303 Leaching, definition of 195 process 200 (see also Percolation, tank) Lead acetate, use of in precipitation . . . 249 buttons, scorification of 353 treatment of, in assay 338 cathodes 121, 300 cyanides of 83 double cyanide of, and zinc 83 effect of, in amalgamation 182 estimation of 415, 433 in solutions 448 ores, assay of 365 peroxidized, use of, for anodes .... 121 separation of 414 smelting of zinc precipitate . . 42, 1'25, 272 zinc couple 38, 246 Leaves, vacuum filter 225 Leggett, T. H., fluxing of zinc precipitate 267 Lenssen and Mohr, estimation of ferrocyanides 442 Lime, analysis of 463 composition of commercial 463 effect of, in amalgamation 177, 183 feeder, automatic 178 use of, in preparatory treatment of slimes 177 use of, in settling slimes 177 Liners for graphite crucibles 267, 268 for Tube mills 154, 158 Liquation of gold in bullion bars . . . 275 Lisbon-Berlyn mine, direct treatment of coarsely crushed ore at 279 zinc-lead couple at 247 Litharge, smelting zinc precipitate with 42,125,272 use of, in assaying 335 Lodge, R. W., assay of cupels 374 assay of cupriferous ores 368 zinc-box precipitate 350 Loevy, Dr. J., assay of graphite cru- cibles 375 estimation of sulphides 450 limits of scoriScation assay 354 Long Tom 187 PAGE Losses in cupellation 342, 380 in roasting zinc precipitate 264 of gold in mill water 184 of zinc in cyanide process 255 Low, A. H., "Technical Methods of Ore Analysis " estimation of antimony 400 arsenic 402 chlorides 472 coal 466 manganese 417 potassium 421 Lowles, J. T., charcoal precipitations . . . 308 Lxihrig Vanner 191 Luipaard's Vlei Estate, details of double treatment at 205 MacArthur-Forrest process 24 amended claim for 27 first patent 25 history of discovery 24 introduction of 30 second patent 27 use of alkalis in 27 MacArthur, J. S., influence of ammonia in cyaniding cupriferous ores . 313 lead-zinc couple 246 selective action of dilute solutions . 26 researches on gold solvents 24 and Dittmar, W., estimation of potassium 422 Machinery for classifying 172 concentrating 187 crushing 135 for grinding 146 Maclaurin, J. S., influence of oxygen on solution of gold and silver in cyanide 102 influence of strength of solution 105 Macquer, researches on Prussian blue . 4 Mactear, J., method for manufacture of cyanide 131 Magenau, V., assay of rich cupels 374 zinc-box precipitate 377 Magnesium cyanide, properties of. ... 69 Magnesium, estimation of 416 in water 472 volumetric estimation of 416 Makins, G. H., action of nitrous acid on gold 347 Manganese, cyanide of 71 Manganese dioxide, as auxiliary solvent 294 as flux 125, 270 estimation of, available 418 Manganese, estimation of 417 in solutions 448 separation of 416 Manganicyanides 94 use of, as gold solvents 104 Mauganocyanides , 94 INDEX 511 PAGE Manufacture of cyanide 127 from ammonia 130 ammonium compounds 130 animal matter 127 atmospheric nitrogen 128 illuminating gas 131 thiocyanates 132 trimethylamine 131 raw materials for 127 Marais (see Bettel) Margueritte and DeSourdeval, manu- facture of cyanide 129 Margoshes, B. M., estimation of ferro- cyanide 435 Marsden crusher 135 Marsh's test for arsenic 403 Mason, W. P., and Bowman, J. W., losses in cupellation 344 Materials used in cyanide manufac- ture 127 used in cyanide process, analysis of . . 463 Matte, analysis of 430 treatment of, in fusions of zinc-gold precipitate 271 May Consolidated mine (Johannes- burg), crushing with cyanide solution 281 Mechanical difficulties in treatment of crushed ore 33 handling of material for cyanide treatment 159 haulage of tailings 163 losses of zinc in precipitation 254 Mein, Capt. G. A., and Butters, C., automatic pulp-distributor . . . 166 Mercur mine (Utah), coarse crushing for cyanide treatment 171 Mercur (Utah), zinc-dust precipitation at 306 Mercuric chloride, use of, as auxiliary. 297 Mercuric cyanide, analysis of, by Gay- Lussac 7 preparation and properties of 81 solvent effect of 17 Mercury, basic cyanides of 82 cyanides of . 81 double cyanide of, and potassium ... 82 effect of, on zinc precipitation 120 estimation of 418 feeding into mortar-boxes 181 fulminate of 98 sickening of 182 use of, as cathode 303 use of, in amalgamation 179 Merrill, C. W., filter press 221 Merton furnace 287 Metallic coatings in zinc precipitation . . 254 compounds of complex cyanogen radicals 84 PAGE Metallic compounds of cyanogen with oxygen or similar elements .... 95 cyanates 96 Metallic cyanides, action of acids on ... 64 action of chlorine on 63 heat on 63 oxygen on 63 water on 64 complex 84 constitution of 63 general methods of formation 62 general properties and reactions .... 63 simple 62 Metallic Extraction Co. (Colorado), acid treatment of roasted pre- cipitate 262 Metallic radicals in cyanide, estimation of 459 Metallics in assay samples 372 Metallic sulphides, action of, on cyanide 1 13 Metallurgical tests 475 Metals, dissolved in cyanide without oxygen 103 electrochemical order of, in cyanide 18, 106 estimation of, in solid cyanide 459 in solutions 447 in bullion 433 Metric weights and measures, conver- sion of 498 Mexico, crushing with cyanide in 285 electrolytic precipitation in 301 slime treatment by aeration in 210 Meyer and Charlton Gold Mining Co., details of cyanide treatment .... 283 Mill-product, handling of, for cyanide treatment 165 Mill, Ball 152 Bryan 150 Chilian 149 Flint 37 Huntington 150 Stamp 137 Tube 37,154 Miller and Fulton, assay of cupels 374 assay of rich slags 373 in cupellation 344 in scorification 353 Miller, Hall and Falk, assay of pyritic ore with niter 363 Millieme system of reporting assays . . . 386 Milling practice with rolls 146 with stamps 144 with tube mills 155 Minas Prietas, Mexico, electrolytic precipitation at 301 experiments on hot solutions at 239 Mine sampling 321 Minerals, electrochemical order of, in cyanide solutions 106 512 INDEX PAGE Mitchell, assay of pyritic ore 361 Mixing of precipitate and fluxes .... 267 of assay samples 328 Modern systems of concentration 188 of metallurgy 193 Modifications of cyanide process 278 Moldenhauer, C., use of ferricyanide as auxiliary solvent 294 Moir, Dr. J., colorimetric estimation of gold 445 Moisture in assay samples 329 in coal, estimation of 465 in cyanide, estimation of 461 in ores, estimation of 427 in samples for analysis 396 Molloy process 304 Montejus 210, 219 Moore filter 222 details of working 223 Moore, T., estimation of nickel 419 Morris, C. J., details of zinc clean- up 262 Mortar-box, construction of 142 Muffle furnaces . . : 340 charging cupels into 341 Muhlenberger, alleged anticipation of cyanide process 23 Muntz metal, use of, for amalgamated plates in battery 181 Myall's United mines (Australia), treatment of zinc precipitate at . 262 Natural settlement of sands and slimes . 159 Nevada, crushing with cyanide solution in 284 cyanide practice in 225, 284 slime treatment in 225 New Goch Gold Mining Co., sand sampling 325 Nicholas and Nicols, treatment of cupriferous ores 310 Nickel, cyanide of 71 separation and estimation of 419 Niter, assaying with 363 use of, in assaying ores of antimony . . 366 use of, in fluxing zinc-gold precipi- tate 125, 266 use of, in roasting zinc precipitate . . 264 Nitrates, estimation of, in solutions. . . 449 Nitric acid, action of, on gold 385 use of, in assaying 345, 384 use of, in treatment of zinc-gold pre- cipitate 123, 262 Nitrites, estimation of 449 Nitrogen, atmospheric, manufacture of cyanide from 128 Nitrogenous matter (organic), manu- acture of cyanide from 127 Nitroprussides 93 PAGE Nitroprussides, characteristic colors of 94 general properties and reactions .... 94 use of, in detecting and estimating sulphides in solution 450 Nitrous acid, action of, on gold in part- ing 347 Non-metals, action of hydrocyanic acid on 54 cyanogen compounds of 47, 56 estimation of, in cyanide 456 solutions 448 Norseman mine (W. Australia), Ball mill, products at 154 Oil, concentration by 191 effect of, in amalgamation 183 Old tailings, preparation of, for cyanid- ing 159 Ore, assay of 333-371 analysis of 395-429 concentration of 185, 192 crushing of 135, 143, 281 grinding of 146 sampling in mine 321 treatment of cupriferous 112, 301 telluride 286, 297 Organic' carbon, estimation of 473 Organic matter, estimation of, in solu- tions 452 in water 473 Outline of ore analysis 396 of operations in cyanide process .... 33 Oxalates, estimation of, in solutions. . . 452 Oxidation, effect of, in cyanide treat- ment 11, 18, 102, 109, 293 of metals in molten bullion 125 Oxides, action of, on cyanide 66 on hydrocyanic acid 55 Oxidizers, action of, on gold in cyanide solution 18 in conjunction with cyanide. . . 18, 103, 293 Oxygen, Dixon's experiments on, in con- junction with cyanide 18 Eisner's experiments on 11 estimation of, in ores 419 Faraday's experiments on 13 heat of combustion of metals with 125 influence of, on cyanide 63, 109 on solubility of metals in cyanide 11, 13, 18, 102,293 Maclaurin's experiments on 102 necessity of, in cyanide process.... 102 refining zinc-precipitate by 43, 276 Pachuca agitation tanks 210 Pan amalgamation 180 test to imitate 489 Paracyanogen, preparation and prop- erties .. 51 INDEX 513 PAGE Parting assay beads 345, 384 bullion assays 384 in crucibles 345 in flasks 346, 384 platinum apparatus for 384 precautions in 385 testing acid for use in 346 Pearce, Dr. R., estimation of arsenic . . . 403 Pelatan-Clerici process 304 Pelouze, J.-, production of formates from cyanides 9 Pelouze's green 92 Penfield, estimation of fluorine 413 Percolation, conditions for effective . . 200 definition of 195 process 35, 195 tests 493 upward 204- Percussion tables 188 Percy, J., assay of cupels 374 assay of cupriferous ores 368 assay of slags 373 proportions of lead in cupellation of copper ores 369 solubility of silver in cyanide solu- tions 20 Perferricyanides 93 Permanganate, use of, as auxiliary dis- solving agent 294 Peroxide of hydrogen, formation of, in reaction of cyanide on gold 102 Peroxide of lead, coating anodes with . . 300 Phosphoric acid, in ores, separation and estimation of 420 Phosphorus in coal, estimation of 467 Picard, H. K., mixer for zinc-dust 308 zinc-dust in connection with bromo- cyanide process 306 Piping for conveying solutions 237 Platinum parting apparatus 384 Plates, amalgamated copper, in battery 180 filter 222 Muntz metal 181 Poisoning with arseniureted hydrogen . 261 with cyanide and hydrocyanic acid . . 55 with cyanide, antidotes for 55 Polymer of cyanifc acid 98 of cyanogen 5-1 of cyanogen bromide 59 of cyanogen chloride 57 of hydrocyanic acid 55 Porrett, R., discovery of sulphocyanides 7 electrolysis of ferrocyanides 7 Possoz and Boissiere, manufacture of cyanide 128 Potassium aurocyanide 73 copper cyanides 79 cyanate 97 Potassium cyanide, action of alkalis on 67 PAGE Potassium cyanide, action of metals on 67 analysis of commercial 455 manufacture of 9, 127 preparation of 65 properties and reactions of 66 synthesis of 65 use of in amalgamation 14, 21, 183 Potassium, estimation of, in cyanide. . . 460 in ores 421 in water 472 ferricyanide 90 use of, as auxiliary dissolving agent 294 ferrocyanide preparation, properties and reactions 86 -ferrous ferricyanide 93 -silver cyanide 75 thiocyanate 100 -zinc cyanide 70 -zinc ferrocyanide 90 Power required for Ball mills 154 for Huntington mills 151 for Rock-breakers 136 for Tube mills 158 Precipitate, white, composition of.... 119 Precipitating surface, influence of 251 Precipitation, alternative methods of . . 39 by aluminium 122 by charcoal 122, 308 by cuprous salts 123, 309 by zinc-dust 40, 124, 306 by zinc shavings 28, 38, 117, 241 chemistry of 117 conditions of 38, 117, 248 effect of copper in 120 in circular vats 245 influence of various factors on .... 38 of copper by acids in cyanide solu- tion 312 on silver and copper ,. . . 19 reactions in 117 Precious metals, estimation of, in bullion 431 solubility of, in cyanide 4, 20, 105, 454 Preparation of charges for leaching .... 200 of ore for direct cyaniding 170 Preparatory treatment of ore for cyanid- ing 133 Presses (see Filter-Presses) Pressure filters 216 Protective alkali, estimation of 439 Proust, investigations on prussiates. . . 6 obtains cyanogen 7 Prussian blue, discovery of 3 manufacture of 3 preparation, properties, and reac- tions 89 researches on 4, 5 soluble . 93 514 INDEX PAGE Prussian green 92 Prussiates, early investigations on ... 6 Prussic acid (see also Hydrocyanic Acid) composition 5 discovery of 4 early use of 3 in natural products 6 Ittner's researches on 7 Scheele's researches on 4 Pulp distributor 166 Pumps, agitation by centrigufal 210 centrifugal 237 for transferring solution 237 Frenier sand 169 Push conveyor 287 Pyrites, action of, on cyanide 110 effect of, in amalgamation 182 Pyritic concentrates, assay of 361 Pyritic ore, assay of 359 Quantities and strength of solution in sand treatment 35, 203 Quartering assay samples 317 Quartz ores, assay of 333, 355 Rae, J. H., patent 15 Rand, E. T., continuous slime settle- ment 214 Rand, battery practice 144 crushing with cyanide solution on.. 282 fluxing of zinc-gold precipitate 286 flux for chlorination tailings 358 pyritic concentrates 361 siliceous ore and tailings 356 mine sampling on 321 sand collecting on 166 scorification assay on 254 slime treatment on 212, 221 Rate of flow in zinc-boxes 250 Redjang Lebong mine (Sumatra) clarifying solutions by filter-press at. 236 composition of bullion at 263 flux for siliceous ore at 356 grading tests on sands before and after cyaniding at 144 selenium in zinc-precipitate at 253 tilting furnace at 269 Reducing agents in assaying 334 power of solutions, estimation of ... 453 Reed, S. A., theory of sampling 320 Refining of bullion 275 with oxygen 43, 276 Refractory materials, assay of certain. . 371 ores, cyanide treatment of 286 Regeneration of cyanide solutions by acids 115 Remelting of cyanide bullion 270 Reporting assay results 348 Reprecipitation of gold and silver during the dissolving process 115 PAGE Residues, discharging of sand 206 discharging slime .' .214, 220 from zinc retorts, assay of 377 Reverberatory furnace for roasting zinc- gold precipitate 268 Revolving cylindrical furnaces 288 Richmond, C. P., electrolytic treat- ment of gold-copper ores 121, 301 Ridgway filter 230 practice with, at Great Boulder, W. Australia 232 Riecken process 304 Rittinger, pointed boxes for classifying 172 Roasting, assay by 359 for cyanide treatment 114, 171, 286 furnaces 286 Argall 288 Edwards 286 Merton 287 zinc precipitate 42, 124, 264, 268 Robinson Deep Gold Mining Co., stamp battery practice at 143 Tube mill, products at 157 Robinson Gold Mining Co. (Transvaal), early slime treatment plant at .. 221 establishment of cyanide process at . . 30 flux for low-grade ore at 356 Rock-breakers 135 jaw 135 gyratory 136 Rodgers, F. and E., manufacture of cyanide 9 Rolls, crushing with, at Cripple Creek . . 154 Argall 148 Hadfield 147 modern practice with 146 Rose, Dr. T. Kirke, amalgamated copper plates as cathodes .... 303 analysis of bullion from zinc process . 275 assay of cupels 374 early practice of amalgamation 14 efficiency of rock-breakers 136 falling bodies in water 173 flux for siliceous ores 355 inside amalgamation 180 refining zinc precipitate and bullion with air or oxygen 43, 125 remarks on scorification 353 roasting of ores 171 Roskelley, T., stamp battery practice. 143 Rowland process of cyanide manufac- ture 131 Sal-ammoniac, use of, in amalgamation . 183 Salisbury battery (Johannesburg), in- troduction of cyanide process at . 30 Sample, definition of 317 Samplers, automatic 325 Vezin . . .318 INDEX 515 Samples, bullion 378, 430 containing dissolved values 332 drying 330 location of, in tanks 326 mixing, dividing and quartering . . . 328 moisture 329 precautions in preparing 323 reduction of 330 residue 328 taken in discharging sand tanks . . 328 Sampling 317 bullion 430 coal 465 conditions for accurate 324 cyanide 455 implements 323 large heaps 317 lime 463 mine ore 321 ores 317 rod 326 objections to 326 sand and similar material 324 difficulties in 324 slime 332 theory of 320 trucks 325 water 468 Sand charges, calculating 477 determining density of 477 collection of 34 definition of 33 natural settlement of, in pits 159 pumps 169 sampling 325 treatment by percolation 35, 195 Sander,K.,assay of carbonaceous matter 377 Sanders, J. F., patent .r 21 Sands and slimes 33 San Sebastian mine (Salvador), elec- trolytic process at 121, 301 Santa Francisca mine (Nicaragua), Gilmour- Young process at . . . . 306 Scheele, observations on cyanides of gold and silver 4 researches on prussic acid 4 Scorification assay 349, 351 theory of 352 difficulties in 352 in fusion furnaces 353 lead required in 351 of bullion assays 382 of large lead buttons 353 of material containing copper 350 of rich silver ore 350 of zinc box precipitate 351 precautions in 353 theory of 352 Scorifiers 349 PAGE Scorifier tongs 352 Screen analysis 435 Callow iQi Screening tests 435 Scrymgeour, use of cupriferous solu- tions as preliminary solvents. . . 311 Selective action of dilute solutions ... 26, 107 Selenium, action of cyanide on 66 detection of 423 estimation of, in bullion 434 in ores 422 in solutions 450 in zinc precipitate 123, 253 removal of, in smelting 269 Separation of coarse zinc (shorts) 258 of dissolved values in slime treatment 36 of gold and silver by cadmium 431 of sands and slimes. 172 tests (hydraulic) 485 Serullas, discovery of cyanogen bromide 9 Settlement, agents used for 177 of slimes 176, 212 of zinc precipitate 259 tanks, conical 176 for slime 176, 214 Settling pits 165 Shaking amalgamated plates 282 Sharwood, W. J., reaction of cyanide on zinc oxide 118 zinc-dust practice 307 "Shorts" 260 treatment of 263 Siemens, Dr. W., invention of electroly- tic process 40 Siemens-Halske process 40, 121, 298 advantages and disadvantages of .40,299 Sifting samples for assay 324 samples for analysis 395 Silica, estimation of, in ores 423 in water 472 use of as flux for zinc precipitate .... 266 in assaying 334 Silicates, estimation of, in cyanide 459 Siliceous material, analysis of 395 ores, assay of 333, 355 Silver cyanate 97 Silver cyanide 74 action of acids and sulphur on 75 preparation, properties, and reactions 74 Silver, Deniges method of estimating 433 double cyanide of potassium and . . 75 double cyanides of 75, 76 estimation in bullion 432 solutions 446 ferrocyanide of 87 ferricyanide of 93 Gay-Lussac's method of estimating . 389 Silver, separation from gold in alloys 43 J. solubility in cyanide 20 516 INDEX PAGE Silver, Volhard's method of estimat- ing 390, 433 volumetric assays of 389 Simon, Dr. F., Ball mill practice 154 Simpson, D., automatic sand sampler. 325 sampling sand in tanks 327 Simpson, E. W., details of furnace practice 287 Simpson, J. W., patent 22 Size of stamp-mill product, condition regulating 141 Sizes of samples allowable at different stages of crushing 319, 331 in relation to grade of ore 320 Sizing tests, method of making in laboratory 485 on Ball mill product 154 on Roll product 147 on Tube mill product 157 Skey, W., conditions of amalgamation. 17 electromotive order of metals in cyanide 18 relative solubility of metals in cyanide 17 Slag, assay of, by scorification 352 assay of, from zinc precipitate 373 from fusions of zinc precipitate .... 270 treatment of, by litharge smelt- ing 42,125,272 Slime, aeration of 210 agitation treatment of 36, 209 coagulation of 177 collection of 35 decantation treatment of 36, 212 deep tanks for settling 214 definition of 34 deposition of, by natural settlement 160 determining density of 478 effect of, in sand treatment 160 first plant in S. Africa for treating. . 221 pulp densities, calculation of 479 sampling 332 separation of 172 settlement of 176, 212 solutions, requisite alkalinity of ... 257 use of lime in settling 177 Smelting, chemistry of 124 of bullion 270 of zinc precipitate 41, 124, 241 of zinc precipitate with litharge 42, 125, 272 Smith, E. A., assay of antimonial ores. . 366 arsenical ores 366 basic ores 358 cupriferous ores 368 pyritic ores 362 slags 373 telluride ores 370 assay with niter 363 oxidizing effect of niter 364 PAGE Smith, E. A., volumetric assay of silver 391 Smith, Lawrence, estimation of alkali metals in ores 397 Smith, R., solubility of silver in cyanide 20 Soda, caustic, use of in amalgamation. . 184 carbonate (see Sodium Carbonate) Sodium amalgam, precipitation by. ... 122 use of in amalgamation 184 argentocyanide 76 bisulphate, use of in clean-up . . . 123, 262 carbonate, use of, for precipitating foul solutions 250 chloride, use of, in conjunction with cyanide 21 cyanide, formation and preparation of 67 properties and reactions 68 estimation of, in cyanide 459 in ores 421 in water 472 ferrocyanide 89 gold cyanides 73 metallic, use of, in manufacture of cyanide 128, 130 nitroprusside 94 use of as indicator 450 peroxide, use of, as auxiliary solvent 293 potassium argentocyanide 76 Solids, dissolved, estimation of, in cy- anide solutions 453 in water 471 suspended, estimation of, in solutions 453 in water 471 Solubility of gold and silver in cyanide 4, 6, 20, 105 Elkington's observations on 10 effect of alkali on 109 effect of base metal compounds on 110 effect of carbonic acid on 109 copper on Ill external influences on 108 other substances on 106 oxygen on 109 strength of solution on 105, 107 temperature on 106 Hagen's observations on 6 MacArthur's observations on 24 Maclaurin's observations on 102 Scheele's observations on 4 Wright's observations on 10 Solution, absorption of, by wooden vats 199 analysis of 437 and electrolytic precipitation in same vessel 304 crushing with cyanide 382 Solvent activity, estimation of 454 South Africa (see Rand, etc.) South Dakota (see Black Hills) Specific gravity determinations 478 INDEX 517 PAGE Specific gravity of slime pulp, calcula- tion of 479 Spitzkasten 36, 172 Spitzliitten 34, 172 conditions of working 175 conical 175 details of construction 174 Stamp battery practice 143 Stamps, crushing with 137 mode of action 138 Standard cyanide test 490 flux for siliceous ore 334 Stines, N. S., estimation of gold in solutions 444 Stockhausen, F., composition of bul- lion bars 275 sampling of bullion 378 Storage tanks, dimensions of 235 for solutions 235 Strength of solution, influence of, in dissolving gold and silver 105 in sand treatment 202 in slime treatment 209 Strong solution treatment 202 Strontium cyanide 69 estimation of 404 Suction filters, early use of 221 Sulman, H. L., bromocyanide reactions 58 double cyanides of copper 79 tilting furnace for zinc precipitate . 269 Sulman and Picard, zinc-dust precipi- tation 40,306 Sulman-Teed, bromocyanide process . . 295 Sulphates, effect of, in acid treatment. 263 estimation of, in solutions 449 water 472 insoluble, deposited in zinc-boxes . . 253 Sulphides, colorimetric estimation of 450 . detection of ". 450 estimation of, in cyanide 459 in solutions 449 metallic, action of, on cyanide 16, 113 Sulphocyanides (see Thiocyanates) Sulphotelluride ores, treatment of .... 294 Sulphur, estimation of, in coal 466 in metals S- 434 in ores 425 in solution 450 occurrence of, in zinc precipitate. . . . 253 separation of, in ores 425 Sulphuric acid, use of, in treatment of cupriferous ores 310 use of, in treating zinc precipi- tate 41,123,261 Sulphurous acid, use of, in preliminary treatment of cupriferous ores.. 311 Sumps 235 Surcharge of bullion assays 387 Surrogat, cyankalium 129 PAGE Suspended matter, estimation of ..453,471 influence of, in zinc precipitation .... 248 Sutton, F. (volumetric analysis) estimation of antimony 401 chlorine 408 copper 411 ferricyanides 442 hardness of water 471 lime 464 nitrates 449 organic matter in water 473 urea 452 Synthesis of cyanogen 48 of hydrocyanic acid 52, 54 of potassium cyanide 62, 65 of urea 8 System of crushing with cyanide at Black Hills 284 on Rand 282 System of continuous slime settlement 214 System of treatment by concentration followed by cyanide 192 Tables, concentrating 187 endless belt 191 percussion 188 Wilfley 188 Tables for converting metric weights and measures 498 of atomic weights 499 Tailings, accumulated, treatment of. . 159 current, handling of 165 double treatment of 205 pumps 169 transfer of- 163 treatment of unsized 161 wheels 167 Tanks, absorption of, solution by wooden 199 calculating contents of 477 clarifying 236 deep, for slime settlement 214 filling, by hose 165 intermediate filling 166 iron 197 leaching 195 remedies for leaks in 198 square 197 steel 197 wooden 197 Tatlock, estimation of potassium 421 Tavener, P. S., automatic lime feeder ... 178 conditions of crushing for amalgama- tion 145 process for treating zinc precipi- tate 42,125,272 smelting precipitate with litharge 42, 125, 272 vats for zinc precipitation 245 518 INDEX PAGE Telluride ores, assay of 369 treatment by bromocyanide 297 treatment by roasting 286 Tellurium, action of, on cyanide 114 assay of ores containing 369 detection of 426 estimation of, in ores 423 in solution 450 in zinc precipitate 123 Temperature, effect of, in cupellation 342 on extraction 239 "Test" for cupellation of lead bullion 273 Tests, amalgamation 488 concentration 487 cyanide extraction 490 for acidity of ores 483 for regulating zinc-box work 256 grading 485 hydraulic separation 485 used in routine of plant 438 Thames Gold Field (New Zealand), use of cyanide in amalgamation .... 17 Theft of zinc precipitate, precautions against 256 Theory of sampling 320 of scorification 352 of slime treatment by decantation 212 Thiocyanates, discovery of 7 estimation of 442 general modes of formation 99 manufacture of cyanide from 132 properties and reactions 99 use as indicator for iron 414 Thiosulphates, estimation of 450 occurrence in cyanide 459 Thomas, J. E., and Williams, G., use of sodium bisulphate in treat- ment of zinc-gold precipitate . . . 262 Tin, estimation of 426 Titanium, estimation of 427 Tonopah Mining Co. (Nevada), cyanide practice 284 Total alkali, estimation of 451 Total cyanide, estimation of 440 cyanogen, estimation of 440 dissolved solids, estimation of 453 sulphur, estimation of 450 Transfer of material 163, 167 by conveyors 167 by tailings wheels 167 Traveling belt niters 226 belts 167 Traversing type of vacuum filter .... 226 Truscott, S. H. and Yates, A., clarify- ing solutions by filter press .... 236 Tube mills 37, 154 conditions for maximum efficiency of 155 Davidsen 155 liners for . . . . 154, 158 Tube mills, power required for. use of banket in Turnbull's blue Types of roasting furnaces . . . vacuum filters . . PAGE 158 156 92 286 221 Union mine (Amador Co., California), use of cyanide in battery at. ... 16 United States, cyanide practice in 284 Upward percolation 204 Urea, estimation of, in solutions 452 synthesis of 8 Useful formula 496 Vacuum filters 37, 221 filtration of sands 222 slimes 222 Valves, arrangement of, on solution pipes 237 suction, for aerating pulp 210 Vats (see also Tanks) conical 211,284 steel 197 wooden 197 Vauquelin, discovery of cyanic acid ... 8 occurrence of prussic acid in nature. 6 Vegetable matter in tailings 161 Vezin sampler 319 Victor, E., reactions of cyanates 98 Victoria (Australia), charcoal precipita- tion in 308 Virginia City (Nevada), slime filtration at 225 treatment of cupriferous ore at 310 Virgoe, W. H., double cyanides of copper 79 losses of zinc in precipitation 254 mercury in zinc precipitate 251 precipitation of zinc and lime in the extractor boxes 256 Volatile matter, in coal, estimation of . 465 Volhard, estimation of silver 390, 433 Volumes of conical vessels, calculation of 477 tanks 477 Volumetric estimation of cyanide 438 earliest notice of 11 of silver 389 Waihi Gold Mining Co. (New Zealand), air-lift slime agitator at 211 crush filter at 226 gradings of Tube mill, products at. . 157 Tube mill liners at 158 Waller, H. T., estimation of aluminium 399 iron in slags 414 zinc in slags 429 Washing, alkali 201 cakes in filter press 220 final water . . . 204 INDEX 519 PAGE Washing, preliminary water 201 Water, analysis of 467 estimation of, in ores 427 hardness of 471 sampling . 467 use of, in cyanide treatment 201 Weak solution in treating sands 203 Weighing assays, charge of ore 336 bullion assays 379, 383 fine metal 344 gold cornets 386 parted gold 347 weights and measures 498 assay 386 West Africa, treatment of zinc precipi- tate in 267 West Australia, Ball mill practice in. . 154 bromocyanide practice in 297 Ridgway filter in 232 roasting for cyanide treatment in ... 286 Whitby, A., analysis of "white precipi- tate" 119 analysis of acid-treated and roasted zinc-precipitate 363 estimation of gold in solutions 443 White, F., treatment of tailings 205 White precipitate in zinc-boxes 118, 252 analysis of 430 composition of 119 conditions and formation 118 Whitehead, C., assay of cupriferous ores 368 Wilde, P. de, precipitation of gold from " cyanide solutions by cuprous salts 123, 309 Wilfley concentrating table 188 Williams, Gerard, W., bromocyanide practice 296 estimation of carbonates 448 lime 464 Williams, S. H., absorption of solution by wood 1" Witwatersrand (see Rand) Woltereck, H. C., manufacture of cyanide 131 Wood, T. W., surcharge of bullion assays 387 Wooden vats, absorption of solution by 199 construction of 197 Woodward, manufacture of Prussian blue 3 Wohler, F., composition of azulmic acid 5( > synthesis of urea 8 Worcester mine (Transvaal), Siemens- Halske, process introduced at . 298 Wright, Dr. A., solubility of gold and silver in cyanide 10 Wurtz, Prof. H., use of cyanide in amal- - gamation 14 PAGE Yates, A. and Truscott, S. H., clarify- ing solutions by filter press . . . 236 Zinc, accumulation of, in solutions 39 ammonium cyanide 71 argentocyanide 76 assay of ores containing 365 Zinc-boxes 38, 243 charging of 247 clean-up of 258 construction of 243 deposits in 118, 252 dimensions of 243 empty compartments in 247 practice 247 rate of flow in 250 trays for 243 Zinc, consumption of 254 -copper couple 251 Zinc cyanide 69 action of heat on 70 preparation of 69 properties and reactions 70 Zinc, double cyanides of 70 double ferrocyanides of 90 Zinc-dust 40, 124, 306 advantages of precipitating with . . 40 at Black Hills 221 at Deloro, Canada 306 clean-up 124 details of working with 306 difficulties in precipitating with 308 mixer for 308 recovery of, in smelting* 269 Zinc, estimation of 428 in solutions 447 ferrocyanide 89 -gold precipitate (see Zinc Precipitate) -lead couple 38, 246 -mercury couple 251 ores, assay on 365 oxide, solubility in cyanide 118 -potassium cyanide 70 Zinc precipitate acid treatment of 41, 123, 262 arsenic in 123 assay of 376 composition of 41 fluxes for 124, 266 fluxing with niter 125, 266 ingredients occurring in 121 roasting 264 roasting with niter 264 treatment of, with bisulphate 123 with hydrochloric acid 262 Zinc precipitation 28, 117, 241, 306 circular vats for 245 conditions which influence 248 details of . ..117,241 520 INDEX PAGE *AOP, Zinc precipitation, difficulties in 252 Zinc, Simpson's process of 22 discovery of 29 separation of 428 hydrogen evolved in 252 shavings, cutting and preparing ... 246 influence of iron screens in 255 first use of for precipitating MacArthur-Forrest, process of 28 cyanide solutions 29 outline of 38 slimes 241 VD 07556' M127G43 cofo 3 THE UNIVERSITY OF CALIFORNIA LI