CHEMICAL COMBINATION AMONG METALS CHEMICAL COMBINATION AMONG METALS BY DR. MICHELE GIUA PROFESSOR OF GENF.RAL CHEMISTRY IN THE ROYAL UNIVERSITY OF SASSARI AND DR. CLARA GIUA-LOLLINI AWARDED THE PRIZE OF THE CAGNOLA FOUNDATION BY THE ROYAL LOMBARDY INSTITUTE OF SCIENCE AND LITERATURE TRANSLATED BY GILBERT WOODING ROBINSON ADVISER IN AGRICULTURAL CHEMISTRY, UNIVERSITY COLLEGE, BANGOR PHILADELPHIA P. BLAKISTON'S SON & CO. 1012, WALNUT STREET 1918 Printed in Great Britain. PREFACE. THE subject of chemical combination among metals is of considerable interest and importance in general chemistry. Viewing the recent developments of the chemistry of metals, we cannot fail to remark the immense strides made in little less than a score of years. This progress has been made possible by the rapid and incessant development of modern metallography. To give a historical picture would involve the detailed discussion of almost all the problems around which metallography is grouped. The chemistry of metals has been largely studied by means of thermal analysis, and it may be fairly said that the subject has only acquired clearness since the introduction of Tammann's thermal method, dating from the beginning of the present century and still growing steadily in value and importance. In our treatment of the subject of the capacity of metals for combination with each other this method of investigation will be outlined. The groundwork of any treatise which aims at explaining the nature of chemical combination among metals must be that part in which are described the various states of equilibrium which can be examined by quantitative methods and are, in consequence, susceptible of scientific interpreta- tion. The foundation of the various methods used for defining the conditions of thermal equilibrium is Gibbs' Phase Eule. In the development of our subject we shall proceed, after a brief account of the various types of equilibrium diagram for binary systems, to define the nature of intermetallic combina- tion. This is a matter of considerable difficulty but of great vi PKEFACE. interest. With it is bound up the question of chemical compounds of variable composition, recently thrown into new light by the fundamental work of the Kussian chemist Kurnakoff. This part of the subject forms, indeed, a fine memorial to the memory of the great Berthollet, whose figure ever retains the character of modernity. It is the distinctive mark of a genius that he anticipates by many years the developments of philosophic or scientific thought and leaves, thereby, a deep mark on the hard rocks of science a safe guide to succeeding generations. There has been a lack, up to the present, of a complete collection of all the binary systems in which intermetallic chemical compounds appear. The chapters on " Homo- polar intermetallic compounds " and " Heteropolar inter- metallic compounds " now fill this gap. All the studies made up to 1915 have been collected and every system has been discussed, above all, with reference to the various phases of equilibrium which are observed in the fusion of metals. With this latter account one may say that the develop- ment of the work is complete. A short chapter at the end on ternary combinations gives an idea of this as yet little- known field of research. So far the investigations on ternary combinations are few in number and not even completely worked out, but they are enough to claim the attention of the student of this branch of the chemistry of the metals. An experimental contribution attempts at an investiga- tion of the degree of dissociation of intermetallic compounds, a subject hitherto untouched, but which serves to give a clearer interpretation of the nature of these combinations. MendelejefPs and Meyer's classification of the elements, together with the thermal method, have afforded a safe guide in the development of the subject. The subject of chemical combination among metals has been considerably elucidated by the application of physical methods to the study of metallic alloys. A complete, though summary, description is therefore necessary of the PEEFACE. vii physical properties of those alloys in which chemical com- pounds are formed. This task has also been attempted. An extended historical account of the subject has not been traced for various reasons. Above all, this branch of chemistry is still rather in fieri than in esse, and is not com- plete in its various branches. In addition, an account of the development of metallography would be needed, which has already been ably traced elsewhere. 1 We have, therefore, in the chapter on the nature of intermetallic compounds, preferred to confine ourselves to the purely chemical side of the history of the subject. M. AND C. GIUA. 1 Guertler, Metallographie, Vol. I., Part la, 1912. Desch., Metallography, 1910. ABBREVIATIONS FOR PERIODICALS, ETC. CITED. Amer. Journ. of Sci, Ann. Chem. Ann. Chim. Phys. Ann. d. Phys. . Arch. Pharm. . Ber Bull. Soc. Chim. Bull. Soc. d'Encour. . C relies Journ. . Chem. Zentr. . Chem. Neivs C.R Dingl. Polyt. Journ. . D.R.P D. A. or Drud. Ann. . Ferrum .... Gazz. Chim. Hal. Genie Civil Gilb. Ann. Int. Z. f. Metalt. Jahr. her. J. Am. C.S.. Journ. Ch. Phys. J.C.S J. de Phys. Journ. of Phys. Chem. J. Russ. Phys. Chem. Soc. . Journ. of Science Journ. praJct. Ch. Journ. Soc. Ch. Ind. Mem. 1st. Lombardo . Metall. . Mon. f. Ch. . Mon. Sclent. . Nature .... Nuovo Cimento . Phil. Mag. Phil. Trans. . Phys. Zeitschr. Pogg. Ann. Proc. Inst. Mech. Engin. . Proc. Roy. Soc. of London . R. Ace. Lincei . Rec. P.-B. Rend. Ace. Scienze Fis. e Nat. Napoli. Rend. Soc. Chim. Ital. American Journal of Science. Liebigs Annalen der Chemie u. Pharmacie Annales de Chimie et de Physique, Paris. Annalen der Physik. Archiv der Pharmacie. Berichte der deutschen chemischen Gesellschaft. Bulletin Societe Chimique de Paris. Bulletin Societe d 1 Encouragement, Paris. Crelles Journal, Journ. fur die reine und angewandte mathematik. Chemisches Zentralblatt, Berlin. Chemical News, London. Compts rendus de VAcademie des Sciences, Paris. Dinglers polytechnisches Journal, Berlin. Deutsche-Re iche-Patent. See Annalen der Physik. Ferrum. See Metallurgie. Gazzetta Chimica Italiana, Rome. Le Genie Civil, Paris. Annalen der physik u. der physikalischen Chemie. International Zeitschrift fur Metallographie. Jahresberichte der Chemie. Journal of American Chemical Society. Journal de Chimie Physique, Geneva. Journal of the Chemical Society, London. Journal de Physique, Paris. Journal of Physical Chemistry. Journal of the Russian Physico-chemical Society, Petrograd. See Philosophical Magazine. Journal fur praktische Chemie. Journal of the Society of Chemical Industrie, London. Memorie del R. Istituto Lombardo di Scienze e Letter e, Milan. Metallurgie. Monatshefte fur Chemie. Moniteur Scientifique, Paris. Nature, London. II Nuovo Cimento, Pisa. Philosophical Magazine, London. Philosophical Transactions of the R. Society of London. Physikalische Zeitschrift, Leipzig. Annalen der Physik u. Chemie. Proceedings Institute Mechanical Engineering. Proceedings of the Royal Society of London. Rendiconti della R. Accademia dei Lincei, Rome. Recueil des Travaux Chimique des Pays-Bas y Amsterdam. Rendiconti Accademia Scienze Fisiche e Naturali, Naples. Rendiconti Societd Chimica Italiana, Rome. x ABBREVIATIONS FOE PERIODICALS, ETC. Rep. Brit. Ass Rev. de Me'tallurgie . Rev. Ge'ne'r. des Sciences Trans, Royal Soc. of London Verh. K. Ak. Wetensch. Amsterd. Wied. Ann Zeit. anorg. Ch. ,, Elektroch.. ., /. angew. Ch. f.Kryst ,, phys. Ch. . Report of the British Association for the Advance- ment of Science, London. Revue de Me'tallurgie, Paris. Revue Ge'nerale des Sciences, Paris. Transactions of the Royal Society of London. Werhandelingen der Koning. Akademie van Weien- schappen, Amsterdam. Annalen der Physik (Wiedemann). Zeitschrift f. anorganische Chemie. Zeitschrift f. Ekktrochemie. Zeitschrift f. angewandte Chemie. Zeitschrift f. Krystallographie. Zeitschrift f. physikalische Chemie CONTENTS. PAGE PKEFACE v CHAPTER I. EQUILIBRIUM DIAGRAMS 1 Binary Equilibria ........ 1 Class I. (a) The two components form neither chemical com- pounds nor solid solutions, 2. (b) The two components are par- tially soluble in the liquid state and do not form compounds, 4. Class II. The two components form definite compounds, 5. (a) The compound melts at a definite temperature to a homo- geneous liquid, 6. Abnormal maxima, 7. (b) The compound has no definite melting point and decomposes on melting, 9. (c) The compound is strongly dissociated in the fused state, 11. Class III. The two components form solid solutions, 14. Type I. of Roozeboom, 14. Type II., 16. Type III., 17. Type IV., 17. Type V., 18. CHAPTER II. THERMAL ANALYSIS 20 CHAPTER III. THE NATURE OF INTERMETALLIC COMPOUNDS . . 23 The Concept of Chemical Combination and the Phase Rule . 23 Intermetallic Combinations of Variable Composition . . 26 Intermetallic Compounds and the Theory of Valency (Tammann's Rules) .34 The Degree of Dissociation of Intermetallic Compounds . . 39 Existence of Intermetallic Compounds in the Vapour State . 48 CHAPTER IV. PHYSICAL PROPERTIES 51 Influence exerted by the presence of Intermetallic Com- pounds on the physical properties of alloys, 5 1 . Specific Volume, 52. Specific Heat, 57. Electrical Conductivity, 64. Matthies- sen's Rule, 66. Barus' Rule, 67. Magnetic Properties, 70. Electrolytic Potential, 73. Thermo-electric Power, 81. Thermal Conductivity, 84. Thermal Dilatation, 86. Hardness, 89. Variation of Hardness with the Composition of Alloys, 90. Compressibility, 97. Crystalline Form of Binary Compounds, 99. Natural Intermetallic Compounds, 101. xii CONTENTS. CHAPTER V. PAGE HOMOPOLAR INTERMETALLIC COMPOUNDS . . 105 Compounds of Elements of Group I. with each other . . 105 1st sub-group, 105, Na-K, 105. 2nd sub-group, 106. Na- Au, 106. Compounds of Elements of Group II. with each other . . 108 1st sub-group, 108. 2nd sub-group, 108. Mg-Zn, 108. Mg-Cd, 109. Hg-Mg, 110. Hg-Ca, 112. Hg-Sr, 112. Hg- Ba, 112. Compounds of Calcium with members of 2nd sub- group, 113. Ca-Mg, 113. Ca-Zn, 114. Ca-Cd, 115. Compounds of Elements of Group III. with each other . .117 Al-La, 117. Compounds of Elements of Group IV. with each other . .117 Ce-Sn, 117. Ce-Pb, 119. Compounds of Elements of Group V. with each other . .119 Compounds of Elements of Group VI. with each other . . 120 Compounds of Elements of Group VII. with each other . .120 Compounds of Elements of Group VIII. with each other . .120 Compounds of Iron, 121. Fe-Co, 121. Fe-Ni, 122. Compounds of Metals of Group I. with Metals of other Groups. 123 Lithium Compounds, 123. Li-Cd, 123. Li-Hg, 125. Li-Sn, 125. Sodium Compounds, 127. Na-Mg, 127. Na-Zn, 127. Na- Cd, 129. Na-Hg, 129. Na-Al, 131. Na-Tl, 131. Na-Sn, 132. Na-Pb, 134. Na-Sb, 135. Na-Bi, 136. Potassium Compounds, 137. K-Zn, 137. K-Cd, 138. K-Hg, 139. K-T1, 140. K-Sn, 141. K-Pb, 142. K-Sb, 143. K-Bi, 143. Rubidium Compounds, 144. Rb-Hg, 144. Caesium Compounds, 145. Cs-Hg, 145. Copper Compounds, 145. Cu-Be, 145. Cu-Mg, 147. Cu-Zn, 148. Cu-Ca, 151. Cu-Cd, 153. Cu-Hg, 154. Cu-Al, 154. Cu-Sn, 156. Cu-Sb, 158. Silver Compounds, 159. Ag-Mg, 159. Ag-Ca, 161. Ag-Zn, 162. Ag-Cd, 163. Ag-Hg, 164. Ag-Al, 165. Ag-Sn, 167. Ag-Sb, 168. Ag-Mn, 168. Ag-Pt, 169. Gold Compounds, 170. Au-Mg, 170. Au-Zn. 171. Au-Cd, 173. Au-Hg, 174. Au-Al, 175. Au-Sn, 175. Au-Pb, 177. Au-Sb, 179. Au-Mn, 179. Compounds of Metals of Group II. with Metals of other Groups . 180 Beryllium Compounds, 180. Be-Fe, 180. Magnesium Compounds, 181. Mg-Al, 181. Mg-Tl, 182. Mg- Sn, 184. Mg-Ce, 184. Mg-Pb, 185. Mg-Sb, 186. Mg-Bi, 187 Mg-Ni, 188. Calcium Compounds, 190. Ca-Al, 190. Ca-Tl, 191. Ca-Sn, 192. Ca-Pb, 193. Ca-Sb, 194. Ca-Bi, 194. Zinc Compounds, 194. Zn-Al, 194. Zn-Sb, 194. Zn-Mn, 195. Zn-Fe, 196. Zn-Co, 197. Zn-Ni, 198. Cadmium Compounds, 200. Cd-Sn, 200. Cd-Sb, 200. Cd-Cr 201. Cd-Fe, 202. Cd-Co, 202. Cd-Ni, 202. Mercury Compounds, 203. Hg-Al, 203. Hg-Ga, 204. Hg-In, 204. Hg-Tl, 204. Hg-Sn, 205. Hg-Ce, 206. Hg-Sb, 206. Hg-U,206. Hg-Cr,206. Hg-Mo,207. Hg-Mn,207. Hg-Fe, 208. Hg-Co,208. Hg-Ni,208. Hg-Pt, 208. CONTENTS. xiii PAGE Compounds of Metals of Group III. with Metals of other Groups . 208 Aluminium Compounds, 208. Al-Ce, 208. Al-Sb, 210. Al-Mn, 211. Al-Pe, 213. Al-Co. 213. Al-Ni, 216. Al-Cr, 217. Thallium Compounds, 218. Tl-Pb, 218. Tl-Sb, 219. Tl-Bi, 220. Tl-Pt, 222. Compounds of Metals of Group IV. with Metals of other Groups . 223 Tin Compounds, 223. Sn-Sb, 223. Sn-Bi, 223. Sn-Mn, 224. Sn-Fe, 226. Sn-Co, 226. Sn-Ni, 228. Sn-Pt, 230. Cerium Compounds, 231. Ce-Bi, 231. Ce-Fe, 233. Lead Compounds, 233. Pb-Bi, 233. Pb-Pd, 233. Pb-Pt, 235. Compounds of Metals of Group V. with Metals of other Groups . 236 Antimony Compounds, 236. Sb-Mn, 236. Sb-Fe, 237. Sb-Co, 238. Sb-Ni, 239. Sb-Pd, 240. Sb-Pt, 242. Sb-Cr, 243. Bismuth Compounds, 244. Bi-Mn, 244. Bi-Ni, 246. Compounds of Metals of Group VI. with Metals of other Groups . 246 Chromium Compounds, 246. Cr-Fe, 246. Cr-Co, 247. Cr-Ni, 247. Molybdenum Compounds, 248. Mo-Fe, 248. Mo-Co, 248. Mo-Ni, 249. CHAPTER VI. HETEROPOLAR INTERMETALLIC COMPOUNDS . .251 General Remarks. ........ 251 Boron Compounds, 252. B-Fe, 252. B-Ni, 252. Carbon Compounds, 254. C-B, 254. C-A1, 254. C-Ti, 255. C-U, 255. C-Cr, 255. C-Mo, 255. C-W, 255. C-V, 255. C-Mn, 255. C-Fe, 256. C-Ni, 256. Silicon Compounds, 256. Si-Li, 256. Si-Cu, 256. Si-Mg, 258. Si-Ba, 258. Si-Sr, 258. Si-Ca, 258. Si-Ce, 259. Si-Ti, 260. Si-Zr, 260. Si-Th, 260. Si-V, 261. Si-Ta, 262. Si-Cr, 262. Si-Mo, 262. Si-W,262. Si-U,262. Si-Mn,262. Si-Fe,263. Si- Co, 264. Si-Ni, 265. Si-Ru, 267. Si-Pt, 267. Si-Pd, 267. Phosphorus Compounds, 267. P-Cu, 267, P-Ag, 268. P-Au, 268. P-Mg,269. P-Zn,269. P-Cd,269. P-Hg,269. P-Sn,269. P-Bi, 269. P-Cr, 269. P-W, 270. P-Mn, 270. P-Fe, 271. P-Co, 272. P-Ni, 273. P-Pd, 275. P-Ir, 275. P-Pt, 275. Arsenic Compounds, 275. As-Cu, 275. As-Ag, 276. As-Au, 276. As-Mg, 276. As-Zn, 276. As-Cd, 277. As-Hg, 278. As-Tl, 278. As-Pb,278. As-Sn, 278. As-Mn, 279. As-Fe, 280. As-Co, 281. As-Ni, 282. As-Pt, 283. Sulphur Compounds, 284. S-Rb, 284. S-Cs, 286. S-Cu, 286. S-Ag,287. S-Au,287. S-Bi,287. S-In, 288. S-T1,288. S-Sn, 289. S-Pb, 290. S-As, 290. S-Se, 292. S-Mo, 292. S-Mn, 292. S-Fe, 293. S-Co, 295. S-Ni, 296. S-Pd, 296. Selenium Compounds, 296. Se-Cu, 296. Se-Ag, 297. Se-Zn, 298. Se-Cd, 298. Se-Hg, 298. Se-In, 298. Se-Tl, 298. Se-Sn, 299. Se-Pb, 300. Se-Sb, 301. Se-Bi, 302. Se-Cr, 302. Se-Mn, 303. Se-Fe, 303. Se-Ni, 303. Se-Co, 303. Se-Pd, 303. Se-Pt, 303. Tellurium Compounds, 303. Te-Cu, 303. Te-Ag, 304. Te-Au, 305. Te-Zn, 306. Te-Cd, 306. Te-Hg, 307. Te-In, 308. Te-Tl, 308. Te-Sn, 309. Te-Pb, 310. Te-As, 310. Te-Sb, 311. Te-Bi, 312. Te-Fe, 313. Te-Co, 313. Te-Ni, 313. Te-Pt, 313. xiv CONTENTS. CHAPTER VII. PAGE TERNARY INTERMETALLIC COMPOUNDS . . 314 General Remarks 314 Fundamental Types . . . . . . .316 Class I. The three components crystallise in the pure state without formation of mixed crystals or chemical compounds . 316 Class II. The three components form mixed crystals but do not combine chemically 318 Class III. The three components combine chemically to form either binary or ternary compounds ..... 318 Type I. Of the three components, completely miscible in the liquid state, two, A and B, form a binary compound D, 318. Type II. Of the three components, completely miscible in the liquid state, A and B form a compound D, and A and C form a compound E, 319. Type III. The three components are completely miscible in the liquid state, A and B form a compound D, A and C form a compound E, and B and C form a compound F, 320. Type IV. The three com- ponents form a ternary compound, 320. Na-K-Hg, 321. Na-Cd-Hg, 321. Mg-Al-Zn, 322. Ag-Au-Te, 324. TABLES 326 Melting Points and Atomic Weights of the more important Metals and Metalloids, 326. Periodic System, 327. Binary Systems studied thermally . . . . . . 328 Binary Systems in which Chemical Combinations do not occur 329 INDEX OF AUTHORS 333 INDEX OF BINARY SYSTEMS 339 CHEMICAL COMBINATION AMONG METALS. CHAPTEK I. EQUILIBRIUM DIAGRAMS. Binary Equilibria. THE application of the principles of the Phase Rule to binary systems may now be regarded as almost completely elaborated. All the forms of diagram theoretically possible have been discussed, and the examples experimentally studied have not only corroborated theoretical deductions, but have enlarged our view of the subject. The knowledge which we possess of the Phase Rule, as applied to the study of two-component systems, serves as a guide and help in the interpretation of phenomena not only of theoretical but also of practical importance. All the phenomena encountered in the theoretical and experimental study of binary systems may be grouped, according to Roberts-Austen, into three general classes. CLASS I. The two components form neither chemical com- pounds nor solid solutions. CLASS II. The two components form chemical compounds. CLASS III. The two components form solid solutions. This classification has a purely systematic value, because in reality there are cases in which the components form solid solutions to a limited extent and in which, while chemical combinations occur, polymorphic transformations exist with more or less extended series of mixed crystals and partial lack of miscibility. C.M. 2* : CHEMCA^L :C6MBINATION AMONG METALS. We may allude to other cases encountered practically : for instance/the two components may be completely immiscible in the fused state or only miscible to a limited degree. The schematic reduction to a few types is necessary and suffi- cient for the interpretation of the different equilibrium diagrams. For our purpose the second and third classes alone have a direct importance ; but to make the treatment of the subject complete we shall describe all the cases out- lined above. CLASS I. (a) The two components form neither chemical compounds nor solid solutions. This case is quite simple. The possible form of equilibrium is shown in Fig. 1. Each of the two components A and B lowers the melting point of EQUILIBRIUM DIAGRAMS. 3 the other so that the curve has two branches meeting in the point c, the so-called eutectic point. This point always lies below the melting points of the components. The liquid phase exists in the region above the line a c b, the liquidus curve as it is commonly called. The solid phase exists below the line. In the region ace the pure component A separates out ; in the region b c e 1 the pure component B separates out. In the region e c A K the pure component A exists together with the eutectic alloy, while in the region e 1 c B K the pure component B exists together w r ith the eutectic alloy. The phenomena which occur on the solidification of fused mixtures of A and B are clearly deduced from the diagram. Let us image a fused mixture of composition represented by m. On cooling, the pure component A will separate out at m and a thermometer immersed in the mixture will show slackening in the cooling curve at the temperature repre- sented by that point. 1 When the mixture cools eventually to the point e', the eutectic alloy separates out. This separation is indicated by a further thermometric arrest proportional to the content in eutectic alloy of the mixture and represented by e' /'. A fused mixture of composition K, corresponding to that of the eutectic alloy, solidifies at the eutectic point c. Here the thermometric arrest is greater than that for all the other possible mixtures of A and B, and is represented by the line c /. The region e j e l represents the eutectic arrests for all the mixtures comprised in the diagram. It should be noted that at the point c the eutectic mixture does not solidify homogeneously but is made up of two com- ponents. If a homogeneous phase solidified at c, it would be a compound, not a mixture. The composition of the eutectic alloy does not correspond to any simple molecular proportion of the components, or if it does, it is quite 1 The withdrawal of A from the melt will alter the composition in the direction of .B and thus the temperature will fall until the eutectic point is reached, when the remain- ing mixture of eutectic composition will solidify. 12 4 CHEMICAL COMBINATION AMONG METALS. fortuitous. The position of the eutectic point depends almost always on the melting points of the two components, as has been noted by A. Miolati. 1 CLASS I. (b) The two components are partially soluble in the liquid state and do not form compounds. This case, which occurs in certain pairs of metals, is fairly common among organic and inorganic compounds. The typical case is that of ether and water. The first investigation on organic substances partially soluble in water is that of Alexejeff, 2 who showed that salicylic acid warmed in a closed tube with water, in which in the cold it is partially soluble, mixes 1 Zeit. phys. Chem., 9, 649 (1892). * Journ. prakt. Chem., 133, 518 (1882) ; Bull. soc. chim., 38, 145 (1882) ; Wied. Ann., 28,305(1886). EQUILIBRIUM DIAGRAMS. 5 with it in all proportions. He has been able to establish the fact that between liquids which only mix in limited proportions, two solutions are always formed, one of com- ponent A in component B, and the others of component B in component A, according as one or the other is present in greater proportion at the saturation point. 1 The equilibrium diagram when the two components A and B are completely insoluble in the solid state and partially soluble in the liquid state is of the form indicated in Pig. 2, in which the liquidus curve is represented by the curve a ef g b. The horizontal line/ g indicates the extent of the gap in miscibility. All the liquid mixtures of com- position between /' and g' consist of two liquid strata, namely, the component B and the alloy of composition/'. The process of solidification of mixtures of compositions lying between/' and g' is clearly shown from the diagram. Along the line b g, the component B is first deposited, while along the line a e, A is first deposited. Along e /, B is deposited together with the eutectic alloy. W. Spring and Romanoff ' 2 have met this case in the study of the equilibrium diagrams of lead-zinc and bismuth-zinc. Other alloys which show this peculiarity are those of lead- nickel, sodium-aluminium, sodium-magnesium, bismuth- aluminium, etc'. This case has been particularly studied by Tammann. 3 CLASS II. The two components form definite compounds. When chemical combination takes place in the fusion of two components the diagram is more complicated ; for our pur- pose these curves have the greatest importance. The com- pounds formed may have variable conditions of existence and yet be more or less stable. We shall reduce to three the types of diagrams in which definite compounds occur without the formation 1 H. C. Jones : Elements of Physical Chemistry, p. 179. 2 Zeit. anorg. Chem., 13, 29 (1896). 3 Ibid., 47, 293 (1905). 6 CHEMICAL COMBINATION AMONG METALS. of solid solutions between the compounds and the pure components. (a) The compound melts at a definite temperature to a homo- geneous liquid. (b) The compound has no definite melting point and partially decomposes on melting. FIG. 3. (c) 1. The compound is strongly dissociated in the liquid state. 2. The compound on fusion forms two liquid strata. (a) Let us imagine a case in which, in addition to the occurrence of a compound with a definite melting point, the three kinds of crystals which exist in the system, i.e., the two components A and B, and the compound A m B n , are completely soluble in the liquid state and do not form solid solutions. The equilibrium diagram then assumes the form indicated in Fig. 3. This shows three branches, of which EQUILIBEIUM DIAGRAMS. 7 the middle branch e c e' indicates the conditions of existence of the compound, whose melting point corresponds to the maximum c of the curve. The points e and e f are the two eutectics, e being that between the compound and the pure component A, and e' that between the compound and the pure component B. In the case indicated in Fig. 3, a fused mixture of the components corresponding in composition to A m B n which is allowed to cool, solidifies homogeneously and completely at the temperature corresponding to c. The fused mixtures of composition e and e r solidify at the tem- peratures corresponding to those points on the diagram and show maximal thermometric arrests. The thermal pheno- mena which accompany the solidification of mixtures of intermediate composition can easily be deduced from what has been said above. Abnormal maxima. In the study of the antimony- aluminium alloys (see Chapter V.) the occurrence of two maxima has been observed, without a corresponding crystalline form in one of the cases. This second maximum is formed, as Tammann has noted 1 by an abnormal process which can be understood from the following considerations. The two components A and B (see Fig. 4) are miscible in all proportions, but the compound, A m B n , forms slowly and separates, after sufficiently long heating, along the line e df, which is its equilibrium curve. If, however, the melt is not heated long enough, then the formation of the compound A m B n is not complete, so that, compared to a melt which has been subjected to longer heating, it has a greater content of the two components A and B. The temperature at which crystals of A or B separate out is lower. With an equal degree of heating of the various melts the dotted line ae^f^ gives the temperatures at which A,A m B n and B separate on the respective branches of the curve. If the concentration of the melt has become by the separation of crystals of A or J5, 1 Zeit. anorg. Chem., 48, 53 (1906). 8 CHEMICAL COMBINATION AMONG METALS. equal to that indicated by the point e, there will be three kinds of molecules present, namely, A,A m B n and B. Since, compared to the velocity with which crystallisation takes place, the velocity of formation of A m B n is inconsiderable, then, even with the simultaneous separation of A and A m B n , the concentration of the melt e will alter until at a lower FIG. 4. temperature than t 3 , or about / 3 ' crystals of B will begin to separate and the remainder will crystallise at a constant temperature with the formation of three kinds of crystals. Melts richer in B will behave in a similar manner on cooling. Let us consider now the curve e 1 d 1 f l : if with increase of component A the velocity of formation of the compound increases, then the maximum is displaced to d or even a EQUILIBRIUM DIAGRAMS. 9 second maximum is formed. The form that the curve assumes depends always on the duration of the heating of the melt. As has been said, no conclusions as to the com- position of a compound can be drawn from the concentration at such a maximum. If, before the compound is formed in appreciable quantity, the two liquids are immiscible, the phenomena of crystallisa- tion are greatly complicated and do not admit of any quan- titative deductions being made. (b) The compound has no definite melting point, and decom- poses on melting. In this case the diagram assumes the form shown in Fig. 5. As was said above, the compound does not melt to a homogeneous liquid, but at a given temperature decomposes 10 CHEMICAL COMBINATION AMONG METALS. into a liquid of definite composition and a crystalline portion with a higher melting point, which in the case indicated in the figure is the component B. This process is expressed by the following equation : A m B n ^lCB + [(n - c) B + mA\ (crystals) (liquid) in which C indicates the number of molecules of B produced per molecule of A m B n . The process of solidification corresponds to the case in which the compound can only exist in the presence of an excess of one of the components. The dotted line which is prolonged from c indicates a region of metastable equilibria and the maximum g corresponds to the composition of the compound. From the diagram we may deduce the thermal phenomena which accompany the solidification of mixtures of varying composition. It is worth while noticing the changes which accompany the cooling of a melt of com- position A m B n . The solidification begins at /, at which temperature the pure component B separates. The sepa- ration of B continues until the temperature has fallen to that of the horizontal through c, below which the compound can exist without decomposition. Here the separation of B ceases and the separation of the compound begins with a thermometric arrest corresponding to the distance between h and the horizontal. This arrest does not indicate the formation of an eutectic, but of the compound. The ther- mometric arrest is proportional to the quantity of the com- pound which is formed for a given quantity of mixture. Tammann, 1 who has discussed particularly the phenomena of crystallisation which occur, in the type of system just described, states that thermal analysis can safely be applied only when the following conditions are satisfied. 1. The reactions which occur in the melt before and after crystallisation must occur fast enough to keep pace with the process of crystallisation. 1 Zeit. anorg. Chem., 45, 24 (1905). EQUILIBRIUM DIAGRAMS. 11 2. The quantity which crystallises in unit time should depend solely on the quantity of heat yielded up by the crystallising mass. The phenomena of crystallisation are still more compli- cated if the compound splits up into a definite melt and another compound of different composition. A B 11. FIG. 6. (c) 1. The compound is strongly dissociated in the fused state. A particular case which in many respects resembles the one just examined is that indicated in Fig. 6. The maximum c' which indicates the compound does not exist in reality, being flattened out to c. But this temperature corresponds simply to that of the two points e and e', which are the two eutectics one between the component A and the compound, and the other between the component B and the 12 CHEMICAL COMBINATION AMONG METALS. compound. The solidification of a mixture corresponding to the composition of A m B n is accompanied by a trans- formation at c where the compound separates. The solidifi- cation of the melts corresponding to the two eutectics takes place at the same temperature, but the arrests in these cases are indicated by e e l and e' e\ respectively. The equilibrium diagram of the system magnesium- aluminium, in which Mg 4 Al 3 occurs as a compound, approxi- mates to this, but a perfect example has not yet been encountered in binary intermetallic systems. Such types have, however, been noted in certain organic binary equilibria. 1 2. A very interesting case is that of a compound which forms on melting two layers miscible with each other at a higher temperature. The diagram in this case is referable to the simplest type shown in Fig. 7. Applying to this diagram the reasoning developed in the preceding cases, its structure is perfectly clear. The eutectics between the compound and the components are e and e'. Along the line a e the pure component B separates. The horizontal tract M N deserves special notice. During the fusion of the mixtures of com- positions between M and N, two liquid layers are formed, so that the compound under certain conditions can exist in equilibrium with the two components. Within the curve M k N the two liquids exist in a form similar to the well- known emulsions of the organic chemist ; but above this curve they mix, forming a homogeneous liquid. The mixture of composition /c, which corresponds to the compound A m B,, in composition, show r s a maximal arrest which, according to the considerations already put forward in preceding cases, is not to be attributed to the formation of an eutectic. The liquidus curve is a e M N e' b. The quantities of liquid which crystallise at the temperatures t l9 t 2 , and t 3 1 Cf. R. Kreemann : Mon.f. Chem., 25, 1215 (1904) ; 29, 877 (1908) ; J. P. Wibaut Rec. P.-B., 32, 244 (1913). M. Giua, Ber., 47, 1718 (1914) ; Gazz. chim. ital., 45, I., 339, II., 348 (1915). EQUILIBEIUM DIAGRAMS. 13 respectively from equal quantities of homogeneous melts are represented by ordinates in Tammann's diagram. 1 The greatest quantity is given by the liquid which crystallises at ti at the composition of the compound, while the quantities FIG. 7. which crystallise at t 2 and t 3 become nil for the composition of the compound. The duration of crystallisation also depends on the composition of the melt. There are thus three modes of determining composition. In this type the horizontal line M N may be supposed to culminate in /c'. Zeit. anorg. Chem., 47, 296 (1905). 14 CHEMICAL COMBINATION AMONG METALS. A sodium-zinc compound of the formula NaZn n is of the type just described. CLASS III. The two components form solid solutions. The diagrams which represent the curves of crystallisation of binary mixtures where solid solutions occur have been described and assembled into five types by Bakhuis Kooze- boom. 1 The five types of solidification have been divided into two groups according as they show partial or total miscibility. Although Eoozeboom's types taken singly do not include the case of the formation of definite compounds, yet they are of great interest because one can often verify the fact that compounds resulting from two components form mixed crystals between more or less wide limits. In the first group are comprised the types I, II and III, in which complete solubility exists between the two com- ponents in both the liquid and the solid states ; in other words the fused components solidify forming a continuous series of mixed crystals. Types IV and V form the second group ; in these, the components are only partially miscible in the solid state. TYPE I. The two components form a continuous series of mixed crystals. They are completely miscible in the liquid and solid state, and form crystalline masses in which the components are not separate but homogeneously built up. The equilibrium curve assumes the form of Fig. 8. In this, the component A lowers the melting point of component B, while the latter raises the melting point of the former. No maximum, therefore, occurs in this diagram. Three regions may be distinguished in the diagram : (1) that of homogeneous liquid, (2) that lying between the liquidus curve and the solidus curve, and (3) that below the solidus curve in which homogeneous mixed crystals occur. A melt of composition a, if cooled, deposits mixed crystals of composition b, and thereby becomes richer in the com- 1 Zeit. phys. Chem., 30, 385 (1899). EQUILIBKIUM DIAGEAMS. 15 ponent A. Consequently, from the same melt there will separate along the curve b c a series of mixed crystals increasingly richer in A . The curve a c gives the composition of the corresponding melts. If it be desired to know the relative amount of mixed FIG. 8. crystals formed from a mixture of composition s at a temperature t v it may be readily found from the diagram. If a line ef be drawn horizontally through t l9 then e will indicate the composition of the fused liquid and / will indicate the composition of the mixed crystals. But the average composition will be represented by g, and therefore 16 CHEMICAL COMBINATION AMONG METALS. x = weight of mixed crystals g e weight of fused mixture " gf, a relation which holds for melts and mixed crystals following the lines a h and b i respectively. In the vicinity of the line h i the residual liquid is almost nil and the alloy consists almost entirely of mixed crystals. Several alloys of gold, copper and iron belong to this type. TYPE II. The fused mixtures deposit a continuous series of M FIG. !). mixed crystals : the curve passes through a maximum. Each component raises the melting point of the other, as is shown in Fig. 9. In this case the curves are quite simple. The liquidus curve is indicated by af c b and the solidus curve by a k c k' b. A mixture of composition M between a and c (or between b and c) solidifies in the temperature interval/ k. The composition of the mixed crystals, as in the preceding type, will differ from that of the liquid. The mixture of com- position c has, however, a definite solidification point, and to that extent resembles a simple substance. It is constituted of EQUILIBRIUM DIAGRAMS. 17 homogeneous crystals just as a chemical individual, and appears thus under microscopical examination. This type of crystallisation has been recognised in the binary system lead-thallium (see Chapter V). TYPE III. The fused mixtures deposit a continuous series of mixed crystals ; the curve passes through a minimum. Each of the components lowers the melting point of the other. The diagram is indicated in Fig. 10. The considerations put forward in the last case hold in an inverse sense' for mixtures of this type. The alloys copper-manganese, copper-gold, nickel-manganese and various others belong to this type. TYPE IV. The solidification curve exhibits a transition point. -This type is shown in Fig. 11. As is seen, component A lowers the melting point of component B, while the latter raises the melting point of the former. It has already been stated that in this type and the next mixed crystals are not formed in a continuous series. The two liquidus curves are O.M. 2 18 CHEMICAL COMBINATION AMONG METALS. a c and be; the solidus curves a d and b e intersect the horizontal through c in d and e. At the temperature of the line c d e the melt af composition represented by c is in equilibrium with two Separate mixed crystals of composition represented by d and e respectively. At this point there is, of course, a discontinuity. In the solidification of a fused mixture n, intermediate in composition between c and d, mixed crystals of composition n 1 will separate at the point n. FIG. 11. Below the point c the mixed crystals which separate will lie along the curve d n 3 and be in equilibrium with melts lying along the line c n 2 . TYPE V. The solidification curve exhibits an eutectic point. -Each of the two metals lowers the melting point of the other (see Fig. 12). This type is similar to that described on p. 2. The liquidus curve is shown by a e b, the solidus curve by af g b. In the solidification of a mixture of com- position e, the liquid will be at that point in equilibrium with mixed crystals/ and g. This solidification takes place EQUILIBRIUM DIAGRAMS. 19 at a constant temperature and causes a maximal thermo- metric arrest. It will, of course, be quite easy to trace the course of solidification of a mixture n : its behaviour will be exactly FIG. 12. similar to that of the mixture considered in Type I (see p. 14). This type of solidification is fairly common in binary mixtures of metals, as in the alloys copper-gold, nickel-gold, aluminium-zinc, bismuth-lead, and many others. CHAPTEK II. THERMAL ANALYSIS. IF, in the study of equilibrium curves a maximum is found, it may be inferred with considerable probability that a compound has been formed whose composition is indicated by the position of the highest point of that branch of the curve in which the said compound occurs. It is frequently the case, however, that this criterion does not suffice to estab- lish the formation of such a compound, and it is only with the aid of thermal analysis, recently introduced by Tammann, that safe conclusions can be drawn. Tammann's method is based on the interpretation, not only of the first arrests noted in cooling curves (which serve to define the liquidus curve), but also of other arrests due to the solidification of eutectics or to polymorphic transformations within single phases. It is necessary to call attention to the importance of the magnitude of the arrest, which is of course propor- tional to the thermal change, in thermal analysis. Generally speaking, a maximum thermometric arrest corresponds to the formation of an eutectic alloy ; but this cannot always be inferred. If, for example, at the temperature of eutectic solidification, the separation of a phase occurs, the ordinary thermal changes may not take place and the eutectic arrest may thereby be masked. The results obtained by thermal analysis are not, then, of absolute value in determining the capacity of two sub- stances to form definite compounds with each other. In the first place, such capacity can only be determined by the thermal method over a range of temperature limited by the conditions of the experiment. There is also a second limitation : thermal analysis studies processes which take THEEMAL ANALYSIS. 21 place in heterogeneous systems and cannot take account of processes occurring in homogeneous liquids. In liquid mix- tures phenomena are noted concerned with the molecules, which act as crystallisation nuclei. Given the limitation of our knowledge and the theoretical and experimental diffi- culties which hinder a complete elucidation of the subject, the thermal method is not always sufficient for the solution of a given problem. Kecently, however, by the application of physico-chemical methods to the problems of equilibrium new cases have been brought to light and our knowledge has been extended. A classical example is given by the bismuth- thallium alloys. The investigations of this system by M. Chikashige 1 had given evidence of the existence of a compound of the formula Bi 5 Tl 3 ; but recent researches of Kurnakoff and his collaborators 2 on the fusion curves and electrical conductivity have demonstrated the existence of a y phase which occurs between 55 per cent, and 64 per cent, (atomic) of bismuth. This phase exists perfectly inde- pendently, but cannot be fitted into any of Eoozeboom's types for solid solutions. Further, microscopic examination showed that it undoubtedly possessed the characteristics of a chemical compound. According to Kurnakoff, Chika- shige's contention that there exists a compound Bi 5 Tl 3 capable of forming solid solutions with excess of thallium or bismuth is not admissible, because the singular points noted do not correspond to the limits of the compound's existence. We shall have occasion to return later to this case, which constitutes a clear case of a chemical individual which does not obey the law of definite proportions. The phenomena observed in the cooling of a liquid sub- stance are well known, as also the course of the cooling curve of two or three substances fused together ; the pheno- mena differ according to the degree of miscibility in the solid and liquid states. All these cases have been sufficiently 1 Zeit. anorg. Chem,, 51, 328 (1906). 2 Ibid., 83, 200 (1913). 22 CHEMICAL COMBINATION AMONG METALS. discussed. We prefer to confine ourselves particularly to the questions relating to the formation of chemical com- pounds. The forms of the principal curves have been already described (see Chapter I). As has been said, and as we shall repeat, the mutual behaviour of two substances is comprehensively and completely described by their equili- brium diagram. With changes in the concentration of two substances there are always definite thermal changes from which general ideas can be obtained as to the structure of the solid phase (see p. 20). As we have seen in the foregoing discussion on equilibria, the interpretation of diagrams where compounds occur is often rendered difficult by the appearance of new phases. Many compounds which have definite maxima are able to form solid solutions with one or both components and between varied limits. Generally, in the diagrams for metals, various phases appear which may render it impossible to interpret rationally the variations of thermal equilibrium. The diagrams representing systems in which compounds occur capable of forming solid solutions with the components are of great importance from the point of view of chemical theory. We shall speak of these in the section on chemical compounds of variable composition. CHAPTER III. THE NATURE OF INTERMETALLIC COMPOUNDS. The Concept of Chemical Combination and the Phase Rule. THE fundamental principles of the Phase Rule and its varied applications in physical chemistry have already been expounded in numerous publications treating of metallo- graphy. This is not the place for even an outline of the subject and we shall therefore assume as well known the fundamentals of Gibbs' theory of heterogeneous equilibria. 1 In dealing with intermetallic compounds, however, whose nature is not yet completely understood and about which considerable differences of opinion exist, it will be well to begin by elucidating a concept which has an important historical significance. In the development of the natural sciences, since the first realistic efforts of the Renaissance, certain general principles or fundamental concepts have formed the bases on which experimental science has been built up. To confine our- selves only to the matter which concerns us in the develop- ment of our subject, the concepts of element, simple sub- stance and chemical compound formed for some centuries the principal subject of chemical investigation, so that all the theories which have held and still hold in science seem intimately bound up with these concepts. Even in ancient times logical and scientific methods developed securely on the bases of fundamental concepts, although these may have been causes of error. Among these concepts, that of " element " has dominated the greatest philosophic minds from the most remote antiquity. " The influence exerted by this doctrine o/ the elements," 1 Roozeboora : Helerogenen Gkichgewichte, Vols. I. and II. 24 CHEMICAL COMBINATION AMONG METALS. says E. von Meyer, 1 writing of Aristotle and Empedocles, " together with its philosophic offshoots on all trends of thought was remarkable. All the scholasticism of the Middle Ages was grounded on this assumption, which became the fount of innumerable errors." Nevertheless the logical basis of science has remained, as also the idea that matter in all its varied forms is derived from the elements. 2 For modern chemistry the most important property of chemical compounds is constancy of composition. But now, according to Kurnakoff, 3 the terms compound and individual have almost the same significance. It has fallen to W. J. Gibbs 4 to introduce a new funda- mental concept, namely, that of phases, i.e., a system of equilibrium formed of homogeneous bodies separated from each other by surfaces. We must attribute an almost abso- lute individuality to this concept. By means, of it, inter- metallic compounds, as also many other complexes, organic and inorganic, which remained completely obscure on the basis of older principles, have received a rational explanation. It should also be mentioned that the concept of phase is much wider than that of chemical individual in the Daltonian or Berzelian sense, because by it is indicated not only bodies of constant composition but also the important class of solutions and homogeneous bodies of variable composition. According to the idea first expressed in 1897 by the technical chemist, Fr. Wald, 5 a chemical individual repre- 1 E. von Meyer : History of Chemistry. 2 Meyer also writes of the ancients : " In the concept of chemical combination, also, we occasionally meet opinions diametrically opposed to those of to-day. In those days the formation of a substance from the interaction of other substances was con- sidered as the creation of a new substance and it was considered that an annihilation took plare of the substances from which it was derived. In all directions the mind was the slave of theoretical speculations, without rigorous experimental proof. This lack appears clearly evident from the way in which the ancients viewed the numerous chemical phenomena with which they became acquainted either casually or empirically. 3 Zeit. anorg. Chem., 88, 109 (1914). 4 Tfarnwdynamic Studies, 187678. 5 Zeit. yhys. Chtm., 18, 337 (1895) ; 19, 607 (1886) ; 22, 253 ; 23, 78 : 24, 315 (1897) 25, 525 ; 26, 77 (1898) ; 28, 13 (1899). NATURE OF INTEBMETALLIC COMPOUNDS. 25 sents a phase which in a series of equilibrium changes possesses a notably constant composition. Kurnakoff, 1 holds that this definition gives a new aid to the recognition of a chemical compound. '* The pure naturalistic and logical conception of a phase here coincides with the mathematical conception of a determinate com- pound. The phase which exists independently would appear to have individual characters and show substantially the manifestations of the ideal complex of atoms which we call a compound. Many definite compounds have been discovered by means of their reactions or from the diagram of their properties ; but up to the present they have not been considered as chemical individuals. In order to be able to demonstrate their existence, it has been considered necessary to isolate them in the form of single and independent phases." The first task, therefore, in the investigation of compound systems is that of establishing the genetic relation among existing phases and of classifying the individuals. From this point of view the domain of chemistry comprehends the classification and study not only of compounds of constant composition, governed by the well-known principles of Proust, but also of solutions, till now considered as homo- geneous physical mixtures. Under the phase rule they come to be treated by the same logical method. As far back as 1907, B. Nasini 2 had placed in clear relief the importance of indefinite combinations in physico- chemical research. He says, 3 " Another important and fertile region of unlooked-for results is that of the study of physical mixtures, of complex combinations, of combina- 1 Zeit. anorg. Chem., 88, 109 (1914). 2 Pv. Nasini : La chimica-fisica, il suo passato, quello che e e quetto che si propone. Padua, 1907. 3 Op. cit., p. 38. The same concept has been developed by Le Chatelier, who writes : " On account of the clearness which was given to chemistry by the concept of definite chemical compounds investigators have long occupied themselves especially with the study of these bodies. Compounds with variable composition, solid solutions and homogeneous liquids were on the contrary held of small account, although the importance and significance of these bodies in the study of natural phenomena are considerable." Lecons sur le charbpn, p. 385. Paris, 1908. 26 CHEMICAL COMBINATION AMONG METALS. tions of indefinite composition, as Guldberg calls them and as Mendelejeff called them before him. There is no doubt JJiat in the past chemistry has been too exclusively occupied although to a large extent necessarily with the simplest chemical species and has taken little notice of solutions or homogeneous mixtures, including solid solutions and the complex combinations which can exist in and be formed from them in various ways. Wherefore it may very reasonably be said that chemistry even physical chemistry until a few years ago was the science only of the most simple and stable chemical species. Having regard to the isolated occurrence in nature of such species and their laboratory origin, these are almost rarities and singular cases obtained and isolated ivith considerable trouble and artifice. Wald said, indeed, that chemistry only studied and made collection of rarities. Physical chemistry, in fact, finally turned its attention to homogeneous and heterogeneous physical mixtures, to the more complex combinations, those of indefinite com- position ; it has in other words turned to the study of materials ivhich directly and actually present themselves and with which we find ourselves in contact without the preoccupation of attempting to isolate from them simple chemical species." Without wishing to discuss, for the moment, the value of the new concept of phases introduced into modern physico- chemical investigation, we will only say that it has been and is of indispensable value in the treatment of indeter- minate complexes as stated in the judgment of Nasini quoted above. In the following pages the problem of intermetallic compounds of variable composition will be more particularly discussed. Intermetallic Combinations of Variable Composition. Before discussing this important aspect of equilibrium diagrams for binary systems and making any observations of a general character, it may be well to mention that in our NATURE OF INTEKMETALLIC COMPOUNDS. 27 treatment of the subject we include all cases in which the compound or compounds formed can give solid solutions with one or both components over more or less wide ranges of concentration. The types described in Chapter I relating to the existence, as shown by the equilibrium diagram, of definite compounds, fall according to the extent of the series of solid solutions into the different types in Eoozeboom's classification as outlined on p. 14. As in the preceding pages, we shall here give some FIG. 13. indications on the possible interpretation of equilibrium diagrams. The following types will be discussed : I. The compound is completely soluble in its components even in the solid state. II. The compound, which has a definite melting point, forms solid solutions with excess of the components. III. The compound, though forming solid solutions with its components, dissociates on fusion. TYPE I. In this type may be grouped the systems whose diagrams are indicated by Figs. 13, 14, and 15 respectively, according as the melting point of the compound lies between, 28 CHEMICAL COMBINATION AMONG METALS. below or above the melting points of its'components. In the three diagrams the alloys of composition A m B n correspond- FIG. 14. FIG. 15. ing to definite compounds solidify homogeneously at the point c. For alloys lying between a and c or b and c the solidification proceeds as in Boozeboom's first type[described on p. 14. .NATURE OF INTERMETALLIC COMPOUNDS. 29 TYPE II. This is indicated in Fig. 16 and is equivalent to the fusion in juxtaposition of two systems, each belonging to Roozeboom's fifth type. The considerations developed concerning this type are sufficient to explain all the processes of solidification of FIG. 16. fused mixtures whose compositions vary from A to B. The liquidus curve is represented by the line a e x e' b and the solidus curve by the line a f g x f g r b. This type is very common in metallic alloys which form compounds with well-marked melting points. The thermal results have often been confirmed by micrographic analysis. The alloys corresponding to the two eutectics e and e f are distinguished 30 CHEMICAL COMBINATION AMONG METALS. * by a very characteristic stratified structure ; in the region where solid solutions occur, the structure is quite homo- geneous. It may be seen, then, from the diagram, how the composition of the solid phase of a compound can vary between limits. As has already been said, solid solutions frequently occur in the formation of compounds, so that generally, the concentration of the liquid or solid phase of a compound is variable, and in such cases the concentration FIG. 17. corresponding to the maximum of the diagram cannot always serve as a criterion to establish the true composition of the compound. In the later part of the work we shall frequently have occasion to return to this problem, which is becoming of increasing importance. The study of the physical properties of alloys (see Chapter IV) is of great service in the interpretation of those problems for which the thermal method has proved inadequate. TYPE III. The diagram is shown in Fig. 17. The liquidus curve is given by the line a e c b while the solidus curve is NATURE OF INTEEMETALLIC COMPOUNDS. 31 given by a f g h h' b. From the diagram the processes of solidification along a e, c e and c b become quite clear. The alloys from c to h' in which the compound A m B n is formed have very complicated crystallisation processes, because the compound does not actually exist as such even in the solid state, being split up into conjugate solid solutions. The systems silver-antimony, silver-tin and various others belong to this type. The recent developments of the thermal method and of physical methods in general, which are of such service in the investigations of modern metallography have led various investigators to give a scientific value to the idea of the existence of compounds of variable composition or indefi- nite compounds, an idea championed by Berthollet in his classical dispute with Proust on the limits of chemical combination. Berthollet foresaw clearly this unknown realm of science 1 and asserted that "compounds which are formed with slight contraction can be found in all pro- portions and their composition is only limited by their capacity for saturation. Thus, alloys, glasses and minerals are formed in diverse proportions in which occasional gaps occur." Avogadro also, who was an admirer of Berthollet, quickly recognised the importance of such indeterminate compounds in chemical theory. Guareschi in his work on Avogadro 2 has given evidence of the breadth of view of the great 1 C. L. Berthollet : Essai de Statique Chimique, L, 373 (1803). 2 I. Guareschi : Opere scelte di Amedeo Avogadro. Turin, 1911. In the critical and historical discourse introductory to the literature of Avogadro' s works, Guareschi writes (p. cix.), " Avogadro was disposed to admit Berthollet's ideas on chemical com- bination in indefinite proportions. Naturally, he believed in the definite proportions in which gases combined ; yet he believed that in certain cases the ideas of Proust and Dalton could be reconciled with those of Berthollet. At the end of his classical memoir (J. de Phys., 1811, LXXIIL, pp. 58 76) he writes, ' If we consider the approach of molecules in solid or liquid bodies, when the spaces between the integrating molecules are not of the same order as those between the elementary molecules, combinations may occur in complicated proportions, but these combinations are of a different nature ; this idea may serve to reconcile the ideas of Berthollet with the theory of fixed proportions.' These ideas are entirely modern and the work on colloids has rendered this kind of combination very probable." 32 CHEMICAL COMBINATION AMONG METALS. Piedmontese physicist and chemist. Guldberg, 1 in 1870, recognised the importance of indeterminate compounds. They have recently received fresh light from the labours of Fr. Wald, already noted. The treatment of the matter by this ardent investigator is certainly not the most adapted to throw into relief the importance of the subject of indeter- minate compounds because he has attempted to extend to all branches of chemistry particular concepts which have only a limited application, seeking, in fact, to diminish the importance of the atomic and molecular theory a theory which we may say was born in Italy, first of Avogadro and then of Canizzaro. E. Nasini 2 has amply dealt with this argument in a notable contribution to which we shall return later, as it has no direct bearing on the present theme. It is of importance to note that the fundamental principles of chemistry are always essential in the study of compounds of constant composition, but that, naturally, they cannot serve as a guide to the investigation of the indeterminate compounds, whose existence and importance we have indi- cated. In order to obtain knowledge of this still obscure subject, the principles of the Phase Eule, together with some concepts developed by Fr. Wald, may guide us and give us fresh light. We may record in this connection the contribution of Kurnakoff 3 on the interpretation of equi- libria among different phases. In an equilibrium diagram in which a compound capable of forming solid solutions with its components appears, there is a characteristic which should be noted at once. When the electrical conductivity of these solid solutions is studied, the isotherm shows two branches which intersect in a point corresponding to the maximum on the thermal diagram (see Chapter IV). 1 Guldberg : Beitrag zur Theorie der unbestimmten chemischen Verbindungen Ostwaltfs Klassiker, n. 139. 2 Gazz. Chim. Ital, 36, 1., 540 (1906) ; ibid., 37, II., 137 (1907). 3 Zeit. anorg. Chem., 83, 500 (1913). NATUKE OF INTERMETALLIC COMPOUNDS. 33 The meeting point of two single branches on curves of physical properties is called by Kurnakoff a singular point and characterises the composition of a definite com- pound. This conception, taken from the theory of algebraic curves, is very happy on account of the light it sheds on the question of equilibria among diverse phases. These singular points, which are necessary criteria for the determination of compounds in solid or liquid homogeneous media, have been called by Kurnakoff Daltonian points, since they illus- trate Dalton's law of multiple proportions. By means of this conception the composition of a definite compound may be indicated. Kurnakoff says, 1 " It is not the composition of the phase, which is generally variable, but the composition at the singular or Daltonian point which is characteristic on the diagrams showing the properties of a determinate compound. . . . A chemical individual repre- sents a phase which shows singular or Daltonian points on the curves of its properties. The composition which corre- sponds to these points is the same in all changes of the factors of equilibrium of the system. 1 " This definition is not valid for a compound of variable composition, because singular points are lacking on the diagrams of its properties. We shall, therefore, define an indeterminate compound 2 in the following manner : " A chemical individual of variable composition represents a phase which does not show singular or Daltonian points on the curves of its properties." In the description of equilibrium diagrams we shall frequently note the existence of intermetallic compounds of variable composition. In the systems thallium-bismuth and mercury-thallium, such compounds appear ; they are more common than is generally believed, for among them may be placed those independent solid phases, indicated in the 1 Zeit. anorg. Chem., 88, 109 (1914). 2 Kurnakoff proposed to term such compounds Berthollidcs ; this denomination, however much an act of remembrance of the great chemist, should not be introduced into use, strictly scientific terms being preferable. C.M. 3 34 CHEMICAL COMBINATION AMONG METALS. description of binary systems by Greek letters. Such are the phases ft, y and S of the ferro-silicon alloys and those of copper and silver with tin, cadmium and zinc (see Chapter V) . Intermetallic Compounds and the Theory of Valency (TAMMANN'S KULES). The description of the behaviour of the elements, which forms the main theme of general chemistry, presents to-day three gaps which will in time be filled as a result of the efforts made to extend our knowledge. These are in- organic complexes, intermetallic compounds, and organic additive molecular compounds. The co-ordination numbers of Werner, illustrated by numerous and careful experi- mental investigations, have shed considerable light on the inorganic compounds, so that it may be safely affirmed that they constitute the most enterprising attempt at systematisation since the periodic law of Mendelejeff and Lothar Meyer. The same cannot be said of intermetallic compounds and the molecular compounds of organic chemistry. The ordinary ideas of valency are insufficient in most cases. It is known also that some intermetallic com- pounds do not conform to the law of definite proportions, thus bringing to life the celebrated dispute between Proust and Berthollet on the limits of chemical combination (compare p. 31). As Tammann remarks, 1 the theory of valency has given abundant fruit in two great chemical groups, namely, the carbon compounds and the inorganic salts. The latter are, above all, governed by very simple rules in which the character of the element has a relative importance. Kegard- ing, in place of these, the intermetallic compounds, it is seen at once that only a very small minority can be explained on the basis of the known saline valencies of their constituent elements. Tammann may well conclude, therefore, that 1 Cf. Lehrbuch der Metallographie, Chemie und PhysiJc der Metatte und ihre Legierungen, p. 229. Leipzig, 1914. NATUKE OP INTEEMETALLIC COMPOUNDS. 35 " Judging from the material at our disposition relating to the binary compounds, it can be maintained that their formulas, where these compounds are not saline in character, are not generally determined by saline valencies." 1 Confining ourselves only to a few examples, we will con- sider two groups of metallic compounds the amalgams of the alkali metals and the compounds of magnesium and thallium. Lithium, potassium sodium and caesium form with mercury a complete series of compounds of the type E Hg 2 (where E == alkali metal). 2 Eubidium forms with mercury the compound EbHg 6 alone. But besides the series mentioned, which is the most stable, we rind the following compounds : Li 3 Hg KHg- Na 3 Hg Cs 2 Hg Li 2 Hg KHg 2 Na 5 Hg a CsHg LiHg KHg 3 Na 3 Hg 2 CsHg a LiHg 2 K 2 Hg 9 NaHg CsHg 4 LiHg 3 KH g , where Q is the heat of formation at constant pressure and ^A> Xjj, X 7 , the latent heats of evaporation of A, B and D respectively. Then '" PA, PR and P D being the partial pressures of the gases T and the absolute temperature. In this equation, the stability of the compound D increases with increase of the quotient on the left of the equation. The quantity Q may be determined by means of the heat of solution. The latent heats of evaporation may be obtained from Trouton's rule, according to which the molecular heat of evaporation is proportional to the boiling point on the absolute scale. H. v. Wartenberg found that the gaseous compound MgZn 2 is fairly stable at low temperatures but unstable at higher temperatures. The existence of the compound Na 3 Hg in the vapour state has been demonstrated. The fact that some intermetallic compounds can exist in the vapour state, although it does not solve the question of the nature of the linkage which binds together the constituent atoms of intermetallic compounds, offers some prospect of the possibility of explaining such combinations by the principles of the theory of valency. Between saline metallic C.M. 50 CHEMICAL COMBINATION AMONG METALS. compounds and the true mixtures which make up inanimate nature there is no sharp break but rather a gradual transition which modern research has explored in the study of solid solutions and labile combinations (intermetallic compounds, additive compounds, etc.). CHAPTEK IV. PHYSICAL PROPERTIES. Influence exercised by the presence of Intermetallic Compounds on the physical properties of Alloys. IN the preceding pages we have called attention to certain limitations of the thermal method in the complete study of equilibrium diagrams, above all, when it is a question of inferring the formation of one or more compounds. But, in addition to the thermal method, there is opened up a wide field of investigation in the accurate study of certain physical properties of alloys, such as specific volume, thermal and electrical conductivities, hardness, magnetic properties, thermo-electric potential, electrolytic potential, heat of formation, specific heat, microscopic characters and crystal- line form. The presence of compounds always has some effect on the physical properties of alloys. A greater or less degree of discontinuity is often sufficient, not only to corroborate the results obtained by thermal analysis, but also to fill in any gaps which could not be explored by the thermal method alone. For example, the fact, placed in evidence by Kurnakoff, of the existence of intermetallic compounds with irrational maxima (i.e. compounds which do not obey Dalton's law of constant proportions) has only been capable of development and interpretation by means of the study of physical properties. According to Kurnakoff, there should be, for every determinate compound, a corresponding dis- continuity or singular point in the curve of physical pro- perties and composition. When this discontinuity is lacking, a maximum noted on the curve obtained by the thermal 42 52 CHEMICAL COMBINATION AMONG METALS. method would correspond to a compound of variable com- position. It is the absence of this discontinuity in the electrical conductivity and compressibility curves in the bismuth-thallium and mercury-thallium alloys which led Kurnakoff (see Chapter V) and, more recently, his pupil P. Paulo vitch x to postulate irrational maxima in these systems. In alloys where the formation of compounds does not take place, the physical properties are almost always linear functions of the composition, as in the case of specific volumes, or continuous functions showing maxima or minima, as in the case of the electrical conductivity and hardness of solid solutions. Specific Volume. The specific volume of a mixture formed from two com- ponents is a linear function of the composition and can generally be calculated from the specific volumes of the single components ; it is, in short, an additive property. If x and y are the masses of the two components and v and v 2 their specific volumes, the specific volume of a mixture is deduced easily from the mixture rule : xvi + yvo if V= - 2-* = V - , 2 - j - x + y v x + r y or, putting - - = r B x + y This equation means that in the mixing of the com- ponents there is no volume change. This rule is, however, not always strictly valid. Small deviations have been noted in the case of the formation of solid solutions and eutectic mixtures. It must also be mentioned in this connection 1 Bull. Soc. Chim., (4), 20, 2 (1916). PHYSICAL PROPERTIES. 53 that a linear relation can subsist in cases where the com- ponents form a continuous series of solid solutions. The formation of a chemical individual introduces a discontinuity into the curve of specific volumes. If for example the specific volumes of two components which do not combine are plotted for varying concentrations, they will form a straight line ; if, on the other hand, a compound appears, the curve will consist of two branches which intersect in a point which, if secondary phenomena do not interfere, indicates the composition of the compound. The case is different where the compound forms solid solutions with one or both of the components ; the two lines of the preceding case are then intersected by a third line so that two points are obtained, neither of which correspond to the maximum of the compound. As a rough approximation this maximum can be obtained by producing the two lines till they meet. The point of intersection should give the maximum of the compound. The following arithmetical rule, deduced by Maey 1 may be applied to a mixture of t\vo compounds resulting from the union of two elements A and B. If x an y are the masses of the two elements in a mixture of the two com- pounds (1) and (2) formed from the union of the said elements, x l9 y l9 and x 2 . y 2 , the masses of the elements in the two compounds, then . x = x + x 2 and 2/ = J/i + y% so that i x + y" 2 + 2/2 i + 2/i) + ( X 2 + 2/2) Vl ( *2+2/2 -\-JL- = r BI +yi\ fai + 2/1) + 0*2 + yj x + y /ji is the relative amount of the element y, - ^ X 1"T ^r X 2 + 2/2 and ^r = r 2 are the relative amounts of y in the Zeit. phys. Cltcm., 29, 122 (1899) 54 CHEMICAL COMBINATION AMONG METALS. i 'i ^2 ~> v2 two unknown compounds, while ~ ( ~ + ^ + ^ + ^ - 9 is the relative amount of the compound (2) in the mixture. The relation can be stated as r==r 2 q + r l (1 - q) == r 1 + (r 2 - r x ) q v or in other words, the relative content of the mixture in one of two elements is a linear function of its content in one of the two compounds. As a deduction from this simple relation it may be recorded that specific volume can be substituted in the calculation of atomic volume. We may consider, for example, the mixtures of mercury with the alkali metals. If V 1 be the atomic volume of mercury, F 2 that of the alkali metal in the mixture, % the number of atoms of mer- cury and n 2 the number of atoms of the alkali metal ; and if F be the mean atomic volume of the mixture and E the relative atomic content in alkali metal, then R= -> These two equations can also be obtained from the atomic weights of the elements comprising the mixture. For if M l and M 2 be the atomic weights of the elements and m l and w 2 their relative amounts in the mixture, and hence M, H "M 2 _ / MA M t ~ I X 71/T / ~T 7\/f I ( MA B '- = V 'Mi or TT = m, V M X 7 r M! M 2 M 2 fm l + rt PHYSICAL PKOPEETIES. 55 For F, 7 = - V n m< n. Putting v = a + b r where a and b are two constants, by substitution V==aM l + [M l (a + b) - M l a] R. From this expression it is seen that F is a linear function of E. Maey has applied these relations to many alloys, even in some cases where compounds are formed. In the following table we reproduce some values from Maey for alloys which do not show sufficiently great deviations in their specific volumes for existence of compounds to be inferred. ALLOYS SHOWING CONTRACTION IN SPECIFIC VOLUME. Alloys. v = a -f b.p. AV. p.AV. Av v Bi -Cd 10181 + -0001373 p -00015 51-8 -001 Ag-Bi 0955 + -000063 p - -0004 49-0 - -004 Hg-Sb 07368 + -0001 422 p - -0008 50-8 - -010 Hg-Sn 07366 + -0006345 p - -00091 53-7 -009 ALLOYS SHOWING INCREASE IN SPECIFIC VOLUME. Alloys. v = a + b.p. Av. P.AV. Av V ' Sn -Sb 13710 + -0001187 p + -0011 51-4 + 007 Sn -Zn 13710 + -000010 p + -0005 75-0 + -005 Pb -Cd 08791 + -0002763 p -f -00035 8-3 + 001 Pb -Sb 08791 + -0006106 p -f -00100 54-1 + -009 56 CHEMICAL COMBINATION AMONG METALS. ALLOYS SHOWING BOTH POSITIVE AND NEGATIVE DEVIATIONS. Alloys. v = a + b.p. AV. p.Av. Av V ' Bi -Sb 10181 + -0004715 p - -00013 37-1 -001 + -00003 22-7 + -0003 Bi Sn 10181 + -0003530 p + -00105 3-3 + 010 - -00058 62-5 -005 Cd -Sn 11544+ -0002156 p + -00063 80-5 + -005 - -00012 14-7 -001 Pb - Sn 08811 + -0004900 p + -00085 16-0 + -009 - -00073 69-5 006 Pb-Ag 08791 + .0000761 p + -00064 11-5 + -007 - -00043 67-6 - -005 Ag -Cu 01591 + -000176 p + -0007 5-6 + -0074 - -00003 50-35 -003 Au -Cu 05191 + -000605 p + -0004 6-8 + 007 - -00003 1-99 -0006 Au Ag 05191 + -0004309 p + -00016 76-5 + 002 - -00007 12-0 -001 Au - Sn 05191 + -000852 p + -00108 37-0 + -013 - -00104 22-7 -015 Ir -Pt 04461 + -0000190 p + -00011 66-7 + -002 - -00007 95-0 -002 From the study of specific volume concentration curves, Maey 1 has affirmed the existence of the following com- pounds : SnAg 3 Au 2 Bi 3 AuPb 3 O D/jT\ Sb 2 Cd 3 SbAg 3 SbCu, FeSb SuCu, CuZn 2 AgZn 4 AgCd 2 or AgCd 3 CuCd 3 AgHg The specific volume method has been applied by Maey to the study of the potassium, sodium and lithium amalgams, but the numerous compounds formed in these series have led to the results being partly untrustworthy. In the cadmium-arsenic alloys an example has been found of the existence of two compounds formed with expansion. 1 Zcit. phys. CJiem., 38, 292 (1901) ; 50, 200 (1905). PHYSICAL PKOPEETIES. 57 In this case solid solutions do not occur, at least within wide limits : it was found that the curve consisted of two straight lines intersected by a third. Here, the points of intersection give the maxima of the compounds. As a conclusion of these investigations we can affirm that the specific volumes of alloys which are simply mechanical mixtures can be calculated to within 1 per cent, by the mixture rule ; deviations, when they occur are generally to be attributed to the formation of compounds. C. Hoitsema l has made a contribution to the subject in a study of the density of the copper-gold and gold silver alloys. As is known, the density is the inverse of the specific volume. In the following table are shown some values for the copper-gold alloys ; the calculated specific volumes have been obtained by the mixture rule and are concordant with those experimentally found. COPPER-GOLD ALLOYS. Gold, Specific Specific volumes. per cent. 15 C Observed. Calculated. (100-0) (19-26) (-05192) 91-7 17-35 05764 05715 _ .907 83-3 15-86 06305 06244 - i-o, 75-0 14-74 06784 06768 - -3, 58-3 12-69 07880 07820 - '8, 25-0 10-035 09965 09919 - '5, 0-0 (8-7) (11494) Specific Heat of Intermetallic Compounds. The energy content of solid bodies is a subject of great importance and is intimately bound up with that of crystal- Ze.it. anon). Chcm., 41, 63 (1904). 58 CHEMICAL COMBINATION AMONG METALS. line form. Domenico Guglielmini, in 1688 and 1705, 1 recog- nised that every substance has its own crystalline form, governed by definite rules. Among the many theories on the energy content of solid bodies is that of the uniform distribution of this energy. This has, however, recently been thrown in doubt by the investigations of Nernst and his co-workers on the relation between specific heat and temperature. The kinetic theory of the energy content of a monatomic body, which- is the simplest case, postulates that the function r f c v d T o is indicated by the oscillations of the atoms, which are supposed to be at rest at absolute zero. At a given tempera- ture the atoms are in motion and, in an isotropic body, this motion occurs in three planes perpendicular to each other. With rise in temperature the energy content of bodies varies. For a solid body such energy is given by the displacement of the atoms from their equilibrium positions, or the localised potential energy. According to Nernst, 2 considering the case in which atoms revolve in a circular path about their mean positions, and assuming that the force attracting them to such posi- tions is proportional to the displacement, 3 the equation for the centrifugal force for an atom of mass m revolving with velocity u in a circle of radius r is m 9 u* = A r where A is the force per unit displacement attracting the atom to its position of rest. 1 On the works of this founder of crystallography see Guareschi : Domenico Gugliclmini e la sua opera scientifica. Turin, 1914. ' 2 The Theory of tJte Solid Slate. London, 1914, p. 11. 3 Boltzinann : Vorks u. GastJieorie, II., 126. PHYSICAL PROPEKTIES. 5 ( J The kinetic energy of the atom is m r r 2 from which u 2 = -^ iii 2 . which establishes the equality between kinetic and potential energy. Consequently, if it is known that the kinetic energy of a monatomic gas is ~ ET 2-98Tper gram atom, z from the above equation the atomic heat of a monatomic solid body is Cv = m = 5-955, for E== 1-985. This simply expresses the law of Dulong and Petit. For the content in heat energy of metallic compounds, as appears from two recent works of H. Schimpff 1 and F. Schubel, 2 the Neumann-Kopp law holds, according to which the molecular heat of a solid compound is equal to the sum of the molecular heats of the elements contained in it. The specific heat, c, of a compound is calculated from the equation nM l c l nM^ where c lt c 2 are the specific heats, M lf M 2 the atomic weights and n, m, the number of atoms of the components in the compound. In metallic compounds, specific heat is an additive pro- perty. Before the study of metallic compounds had reached its present importance, Regnault 3 had shown that the specific heats of fusible metals followed the mixture rule at tempera- tures below the melting point, diverging therefrom at higher 1 Zeit. phys. Chem., 71, 288 (1910). 2 Zeit. anorg. Chcm., 87, 101 (1914). 3 Ann. Ch. Phys., (3), 1, 129 (1841). 60 CHEMICAL COMBINATION AMONG METALS. temperatures. The recent investigations of Schiibel show the additive character of the specific heats of metallic compounds. The following table shows the results obtained by Schiibel for a number of definite compounds at temperatures below their melting points. Com- pounds. Specific heats at Melting Points. 100 200 300 400 500 600 Cu 2 Mg 1184 1230 1283 1365 . . _ 797 Cu 3 Al 1093 1135 1167 1205 1260 1050 CuAc 1310 1363 1413 1468 625 CuAlo 1526 1585 1630 1665 1690 . . 590 CllgSb 0815 0837 0860 0890 687 Cu 2 Sb 0760 0784 0806 AgMg Ag 3 Al 0910 0695 0942 0724 0974 0745 1004 0762 1042 0872 1070 0798 820 771 Ag 2 Al 0763 0785 0806 0831 0868 I -0913 721 Ag 3 Sb 0560 0574 0630 . - 559 MgZn 3 1180 1234 1292 1450 595 Ni 2 Mg 1305 1385 1424 1460 1480 1508 1145 Co 2 Sn 0824 0876 0900 0926 0944 0962 Ni 3 Sn 0836 0872 0907 0940 0972 ! -1002 1162 FeSi 1465 1540 1600 1645 1690 1720 NiSi . . 1469 1523 1572 1615 Ni 2 Si 1190 1250 1290 1320 1355 ! -1386 Mg 2 Si 2250 2455 2640 1102 As a result of the preceding data it appears that the com- pounds Cu 2 Mg, Cu 3 Al, CuAl 2 , AgMg, Ag 3 Al, MgZn 2 , Co 3 Sn, follow the Neumann-Kopp law to within 2 per cent. Other compounds show greater deviations ; the compound Ni 3 Sn shows a deviation of about 7-3 per cent. From the data given on p. 61 it appears that the Neumann- Kopp law is, especially for high temperatures, an approxi- mate statement. The divergences between the observed and calculated values are independent of the temperature for nearly half the compounds shown. We shall refer to a PHYSICAL PEOPEETIES. 61 MOLECULAR HEATS OF CERTAIN METALLIC COMPOUNDS (SCHUBEL). Deviations of observed from calculated values. Mol. heats -150 -100 100 200 300 400 500 Cu 2 Mg obs. 12-81 14-73 17-13 17-95 18-65 19-45 20-69 calc. 12-88 14-89 17-37 18-11 18-68 19-29 19-78 diff. -0-07 -0-16 -0-24 0-16 -0-03 + 0-16 + 0-91 . . % -0.06 -1-1 -1-4 0-9 -0-2 + 0-8 + 4-6 . CUaAl obs. 16-28 19-08 22-60 23-82 24-73 25-43 26-26 27-46 calc. 16-58 19-48 23-00 23-98 24-76 25-52 26-07 26-71 diff. 0-30 0-40 0-40 -0-16 0-03 -0-9 + 0-19 + 0-75 -1-8 -2-0 -1-7 -0-7 0-1 -0-4 + 0-7 + 2-8 CuAl obs. 7-58 9-10 11-10 11-88 12-36 12-82 13-32 calc. 8.00 9-52 11-42 12-02 12.48 12-88 13-15 . diff. 0-42 -0-42 0-32 -0-14 0-12 0-06 + 0-17 - - % 5-2 4-4 -2-8 - 1-1 1-0 -0-5 + 1-3 CuAl 2 obs. 11-70 14-07 17-04 17-98 18-67 19-20 19-61 19-22 calc. 11-71 14-07 17-05 18-06 18-82 19-44 19-84 20-27 diff. -0-01 0-01 0-08 0-15 0-2 0-23 -0.35 % -0-1 0-1 0-4 -0-8 -1-2 1-1 j.y Cu,Sb obs. 19-44 21-60 24-28 25-33 26-01 26-73 27-66 calc. 17-82 20-33 23-31 25-06 24-71 25-42 26-03 - diff. + 1-62 + 1-27 + 0-97 + 1-27 + 1-30 + 1-31 + 1-63 . % + 9-0 + 6-2 + 4-1 + 5-2 + 5-2 + 5-1 + 6-2 . - Cu Sb obs. 14-40 15-99 17-97 18-79 19-32 19-94 calc. 13-53 15-35 17-52 18-08 18-57 19-10 . . diff. + 0-87 + 0-64 + 0-45 + 0-71 + 0-75 + 0-84 + 6-4 + 4-1 + 2-6 + 3-9 + 4-0 + 4-4 - - AgMg obs. 8-76 9-90 11-36 12-02 12-46 12-88 13-26 13-79 calc. 9-27 10-39 11-80 12-23 12-54 12-91 13-26 13-81 diff. -0-51 -0-49 0-44 0-21 0-08 0-03 0-02 % 5-5 -4-7 -3-7 -1-7 0-6 0-2 0-2 Ag.Al obs. 19-08 20-92 23-36 24-56 25-40 26-14 26-73 27-43 calc. 18-62 20-92 23-66 24-28 24-76 25-31 25-89 27-16 diff. + 0-46 -0-30 + 0-28 + 0-64 + 0-83 + 0-84 + 0-27 % + 2-5 -1-3 + 1-2 + 2-6 + 3-0 + 3-2 + 1- AgaAl obs. 13-77 15-45 17-70 18-53 19-07 19-58 20-18 21-08 calc. 13-65 15-46 17-65 18-20 18-62 19-06 19-49 20-38 diff. + 0-12 0-01 + 0-05 + 0-33 + 0-45 + 0-52 + 0-69 + 0-70 % + 0-9 + 0-3 + 1-8 + 2-4 + 2-7 + 3-5 + 3-5 Ag 8 Sb obs. 20-48 22-44 24-52 24-85 25-47 27-95 . calc. 19-86 21-67 23-97 24-36 24-71 25-21 . - diff. + 0-62 + 0-77 + 0-55 + 0-49 + 0-76 + 2-74 % + 3-1 + 3-5 + 2-3 + 2-0 + 3-1 + 10-8 . - MgZn 2 obs. 13-71 15-27 17-37 18-31 19-15 20-05 calc. 13-98 15-57 17-71 18-69 19-44 20-05 diff. -0-27 0-30 -0-34 0-38 -0-29 - % 1-9 1-9 -1-9 -2-0 -1-5 . Ni 2 Mg obs. 14-18 16-80 18-50 19-64 20-19 20-70 20-99 calc. 14-51 17-91 19-45 20-92 22-05 21-46 21-77 diff. -0-33 -1-11 0-95 1-28 -1-86 0-76 0-78 2 6-3 -4-9 -6-1 -8-4 3-5 -3-6 62 CHEMICAL COMBINATION AMONG METALS. rational interpretation of the divergences shown with rise in temperature in the following pages. ATOMIC HEATS OF CERTAIN METALLIC COMPOUNDS (FROM SCHUBEL.) Compounds 100 200 300 400 500 600 Cu 2 Mg 5-98 6-22 648 6-89 __ _ Cu 3 Al 5-95 6-18 6-36 6-56 6-86 CuAl 5-94 6-18 641 6-66 CuAl a Cu 3 Sb 5-99 6-33 6-22 6-50 640 6-68 6-54 6-91 6-64 I Cu 2 Sb 6-26 644 6-65 . . ., . AgMg Ag 3 Al 5-89 6-23 6-14 6-35 644 6-54 6-63 6-68 6-89 6-86 7-07 7-00 Ag 2 Al 6-17 6-36 6-53 6-73 7-03 740 Ag 3 Sb 6-21 6-37 6-99 MgZn 2 6-10 6-38 6-68 748 Ni 2 Mg 6-17 6-55 6-73 6-90 6-98 7-12 Co 2 Sn 6-51 6-92 7-10 7-33 746 7-60 Ni 3 Sn 6-17 648 6-70 6-95 7-18 7-39 FeSi 6-16 648 6-74 6-93 7-11 7-25 NiSi 5-73 6-39 6-64 6-85 7-04 Ni 2 Si 5-77 6-07 6-27 640 6-59 7-67 Mg,Si 5-78 6-31 6-79 The relation of atomic heat to temperature above 100 is generally almost linear. If the above values are plotted along with the known values for the pure metals it will be seen that the behaviour of intermetallic compounds is perfectly similar to that of pure metals. Inasmuch as the figures quoted do not show a perfect concordance between observed and calculated values, the Neumann-Kopp law must be considered as an approxima- tion, particularly at high temperatures. Planck's l quantum theory, proposed in his researches on the phenomena of radiation has been generalised by 1 Theorie d. Warmestrahlung, p. 157. Leipzig, 1906. PHYSICAL PEOPEETIES. 63 Einstein * for the movements of atoms ; it gives a rational explanation of the deviations from the law of Dulong and Petit, which forms the foundation for that of Neumann- Kopp. Einstein's formula for the total energy content W of a gram atom is given by the equation, deduced from the ordi- nary equations for the distribution of energy, eT-1, and for atomic heat, differentiating with respect to T : (elf- e T 1 According to the quantum theory, definite limits are assigned to the oscillations of an atom about its equilibrium position. In the case of a solid monatomic body it may be assumed that its molecules are surrounded by a gas which may be considered monatomic for simplicity. The molecules of this gas oppose the oscillations of the atoms of the solid body, producing thereby a condition of equilibrium. For a monatomic gas the number of oscillations is nil ; it can move freely by its kinetic energy according to Max\vell's conception. But for a solid body or any other union of molecules the quantum theory leads to other results. Einstein's formula, which gives the specific heat as a function of the temperature, leads to a simple qualitative result. Quantitatively there are divergences between theory and observation which increase at low temperatures. Nernst and Lindemann 2 have amended the formula of Einstein, introducing in addition to the number of oscilla- tions r a second number 5 which gives values more in 1 Ann. d. Phys., 22, 180 (1907). 2 Ber., 43, 26 (1910). 64 CHEMICAL COMBINATION AMONG METALS. accordance with experiment. The Nernst-Lindemann formula is the following : dT __ eT\ T f ftv y ( f*JL V \e T -- V \e 2 T - V Although it has been shown necessary to introduce into Einstein's formula a greater number of oscillations, the value ^-in the above formula has no theoretical basis. 2 Other theoretical possibilities with regard to the relation of specific heat and temperature have been discussed by Debye * and Duclaux. 2 We must, however, limit our treat- ment of the subject to the short outline given above. Electrical Conductivity. The problem of the electrical conductivity of alloys has become of considerable importance with the development of modern metallography an importance both theoretical and practical because alloys with constant and small tem- perature coefficients are of great use in electrotechny. The first exhaustive study of this subject was made by Matthies- sen. 3 In his earlier researches Matthiessen found that binary mixtures of metals can be divided into two groups ; in the first the conductivity, o-, is an additive property as in the case of alloys of lead zinc, tin and cadmium ; in the second group the conductivity is less than that calcu- lated by the mixture rule, as in the case of alloys of copper, silver, gold, aluminium, bismuth, platinum, antimony, iron and others. These two types of conductivity curves are shown in Fig. 19. The abscissae represent the percentage of the second component and the ordinates the con- 1 Ann. d. Phys., 39, 789 (1912). 2 C. R., 155, 1015, 1509 (1912). 3 Fogg. Ann., 103, 428 (1858) ; 110, 190 (1860) ; 116, 369 (1862) ; 122, 19 and 68 1864). PHYSICAL PROPERTIES. 65 ductivity. The straight line I. indicates that in a mixture of two metals the conductivity can be calculated by the mixture rule, i.e., the conductivity of the alloy is the sum of the single conductivities of its constituents obtained from their relative proportions. Curve II. shows the course of conductivity in alloys of the second group. If a quantity of an element of the second group be added to a metal of the first group the conductivity curve will be lowered appreciably and the curve will be of the type III. Curve IV. represents the case of the formation of a compound. Among the investigations carried out on the conductivity of metallic alloys must also be mentioned those of Le Chatelier l and Liebenoff. 2 Lord Rayleigh 3 has also dis- cussed the matter on the theoretical side. 1 C. E., 112, 40 (1891) ; 126, 1709 (1898). 2 Zeit. Elektr., 4, 201 (189798). 3 Nature, 54, 154 (1896). C.M. 5 66 CHEMICAL COMBINATION AMONG METALS. Le Chatelier attempted to give an explanation of the diverse behaviour of the metals of Matthiessen's first and second groups. If an alloy is a " mixture of crystals " of two metals, the conductivity may be calculated by the mix- ture rule ; but if it forms " crystals of mixture " the values of the conductivity are then different. 1 Le Chatelier further noted that the addition of small quantities of non- metals, such as phosphorus, carbon and arsenic, exercised a great influence on the conductivity of certain metals. The conductivity of iron, for example, is greatly altered by the addition of small quantities of carbon. 2 The electrical conductivity method has been used of late years to define the existence of various intermetallic com- pounds. By the investigations of Guertler, 3 of Kurnakoff and Zemczuzny 4 and of Stepanoff, 5 it has become one of the most valued methods of investigation for the solution of problems left unsolved by the thermal method. The experimental methods followed in the investigations men- tioned are based on measurements of conductivity or of specific resistance. MATTHIESSEN'S EULE. If the relative increase of conductivity between and 100 be expressed by the equation P (a) = 100 - == 100 " " "o 2 10 where %, the specific resistances, then for pure metals P (a) == 29. If n i-F Pi where P l is the electrolytic solution pressure, p 1 the osmotic T) pressure of the cations M 1? n the valency of M 1? the electro- lytic constant of gases and T the absolute temperature. When part of the ions of M x are substituted by ions of M 2 , p l becomes smaller and hence x 1 becomes greater. The same argument applies to # 2 ; the difference of potential between the metal M 2 and a solution of M 2 Z is influenced by the presence of ions of Mj as in the first case. For equilibrium X-^ XQ. Representing graphically the values x v x 2 as a function of the composition of an electrolyte in which the total concen- tration of the ions is constant, the curve shown in Fig. 21 is obtained. A point in the line A D gives the difference of potential and the concentration of the electrolyte in equili- 1 Zeit. phy*, Chem., 22, 539 (1898). 76 CHEMICAL COMBINATION AMONG METALS. brium with the metal M x ; similarly a point in B D gives the difference of potential and the concentration of the electrolyte in equilibrium with the metal M 2 . At the point D the electro- lyte is in equilibrium with both metals. For this point x 1 =- x 29 or or P. "" Pi and putting % = n 2 B M-L FIG. 21. The ratio oj the ionic concentrations in the equilibrium con- dition is equal to the ratio of the electrolytic solution pressures. II. The two metals form mixed crystals with each other. In this case the electrolytic solution pressure of M 3 is lowered by presence of M 2 ; when in the metallic phase there are x atoms of M 2 and 1 x atoms of M x (x being very small), the lowering of the solution pressure is proportional to the number of molecules of the second metal which are in the solution. The solution pressure of the first metal is PHYSICAL PKOPERTIES. 77 P' (1 x) and of the second Kx, in which K is an unknown factor. For equilibrium IP' (lx) IKx_ V V, V v, ' or Vp 2 V Kx and if n 1 = n. VP f (lx) p 2 K x ' A/, FIG. 22. which means that the ratio of the ions in the electrolyte is, in atomic proportions of the metals, as K : P'. If the two metals form a continuous series of mixed crystals the potential curve assumes the form of Fig. 22, in which the continuous line represents the metallic phase and the dotted line the corresponding electrolyte. When there is a gap in the series of mixed crystals the two cases represented in Figs. 23 and 24 are exemplified. The two components form 78 CHEMICAL COMBINATION AMONG METALS. two series of solid solutions ; between the limits C and D the two metallic phases are in equilibrium. At E the electrolyte co-exists. The type represented by Fig. 23 has been found by Herschkovitch 1 in the cadmium-tin and cadmium-lead alloys. The type shown in Fig. 24 has been met with by the same author in the zinc-tin and zinc-lead alloys and also by Jaeger 2 and Bijl 3 in the cadmium amalgams. III. The iivo metals form a compound. If the compound forms ions of its own composition the curve shows a maxi- FIG. 23. FIG. 24. mum as in Fig. 25. The solution tension of the compound, as in the case of a pure metal, is a specific constant. Indicating it by P! . 2 we obtain by Nernst's formula ET n,,oF . log P V The ions of the compound are dissociated into the ions of the components, and if in the component there are a atoms of Mj and b atoms of M 2 , equilibrium will exist when If the total concentration of the ions p l + p 2 is constant, 1 Loc. cit. 2 Wied. Ann., 65, 106 (1898). Zeit. phys. Chem., 41, 641 (1902). PHYSICAL PKOPEKTIES. 79 p V2 reaches a maximum when p : p 2 = a : b. In other words x reaches a maximum 1 when the ratio of the ions M 1 and M 2 in the electrolyte is equal to the ratio of the metals in the compound. This case is indicated in Fig. 25. Along AG the potential is that of the electrolyte in equilibrium with pure M! ; G is a non- variant point in which the electrolyte of that composition is in equilibrium with M t and the compound MiMa. The solid phase in equilibrium with electrolytes lying between G and K is the compound M, FIG. 25. M^Mg, and the potential follows the dotted curve. K is another non-variant point in which the compound M^ and the pure component M 2 co-exist with the electrolyte. Along KB the electrolyte is in equilibrium with component M 2 . If n compounds are formed between the two metals M t and M 2 , for the point corresponding to each compound there occurs in the curve a fall of potential, so that there are n falls of potential as Herschkovitch stated. Reinders discusses also the cases in which the compound forms series of mixed crystals with the components M x and 1 Thus in Reinders' paper. It would appear, however, that when p 1>2 is a maximum, x should be a minimum. TRANSLATOR'S NOTE. 80 CHEMICAL COMBINATION AMONG METALS. M 2 . From what has been said already, however, the form of the potential curve will be quite clear in this as in the other cases, since the diagram is a combination of the types comprised in it. The complications which may occur in potential curves when, in addition to a compound, other variable phases appear with more or less wide limits, render the method somewhat unreliable, so that it is not always possible from observations of potential to deduce the true composition of the compound. A. Sucheni, 1 in a study of the potential of thallium amal- gams, noted that the compound Hg 2 Tl formed solid solutions with mercury at an atomic concentration of 33-3 per cent, of thallium. The electromotive force increases up to the concentration of the compound, after which it remains con- stant. L. Cambi 2 has recently used the measurements of electrolytic potential to corroborate the equilibrium diagrams of the calcium and magnesium amalgams made from thermal data (see Chapter V). In the following table are shown some results obtained by Pushin in the study of the electrolytic potential of various binary alloys in which definite compounds exist. These results are not always in accord with those obtained by the thermal method. System. Compounds found by Pushin. Compounds now admitted. Ag-Se Ag 2 Se Ag 2 Se Ag-Te Ag./Te Ag 2 Te, AgTe Cu-Te CuTe, Cu 2 Te Cu 4 Te, Cu 2 Te Pb-Te PbTe PbTe Sn-Te SnTe SnTe Sn-Cu SnCu 2 , SnCu 3 SnCu 2 Sn-Ag SnAg 3 , Ag 6 Sn or Ag 5 Sn SnA g3 Sn-Au Sn 2 Au, SnAu SnAu, Sn 2 Au, Sn 4 ^ Ui Zn-Cu Zn 6 Cu, Zn. 2 Cu, ZnCu, ZnCu 2 ZnCu, Zn 3 Cu 2 Zn-Ag ZiisAg, Zn 4 Ag, Zn 2 Ag, ZnAg Zn 2 Ag 8 , ZnAg, Zn H Ags, Zn 5 Ag 2 Zn-Au Zn 5 Au ?, Zn 2 Au, ZnAu ZnAu, Zn 5 Au s , Ziv An Cd-Cu Cd 2 Cu CdCu 2 , Cd 3 Cu 2 1 Zeit. f. Elektroch, 12, 726 (1900). 2 R. Ace. Lincei, 23, II. (1914) ; 24, I. (1915). PHYSICAL PEOPEETIES. 81 Thermo-electric Power. The study of the variation of thermo-electric power with the composition of an alloy has only recently been made in a systematic way by E. Budolfi l and W. Haken. 2 The researches of E. Becquerel 3 had led to the belief that the thermo-electric power of an alloy reaches its maximum when the components are present in equivalent proportions. The recent studies have shown this generalisation to be incomplete, but the fact noted by Becquerel served to call attention to this important field of study. Cj J. FIG. 26. It is known that in a " couple " of two metals the magni- tude of the electromotive force E depends on the temperature as well as the nature of the metals. In addition to the researches of Becquerel on the measure of E in alloys, the earlier work of Siebeck 4 and C. L. Weber 5 should be mentioned. 1 Zeit. anorg. Chem., 67, 65 (1910). 2 Ann. d. Phys., (4), 32, 291 (1910). 3 Ann. Chim. Phys., (4), 8, 408 (1866). 4 Gilb. Ann., 73, 115, 480 (1823) ; Ann. Chim. Phys., 199 (1823). 5 Wied. Ann., 23, 447 (1884). C.M. 6 82 CHEMICAL COMBINATION AMONG METALS. Rudolfi studied the thermo-electric power of various alloys, as 1. Tin-cadmium. 2. Tin-zinc. 3. Cadmium-zinc. 4. Tin-lead. 5. Bismuth-cadmium. 6. Lead-antimony. 7. Gold-silver. 8. Gold-copper. 9. Nickel-copper. 10. Platinum-palladium. Coi-tx.<7 /tCto^vc-o-i-i^ FIG. 27. Fir.. 28 PHYSICAL PROPERTIES. 83 He gives a general summary of the variations of thermo- electric power with the composition of alloys, laying down the following four types with accompanying diagrams. TYPE 1. The two components mix in the liquid state but solidify separately. The curve of thermo-electric power assumes the form indicated in Fig. 26. It is a continuous line which can be calculated by the mixture rule. TYPE 2. The tivo components form a series of mixed crystals. The curve here assumes the form of Fig. 27, showing a continuous variation passing through a minimum. FIG. 29. TYPE 3. The two components form a limited series of mixed crystals. The curve (Fig. 28) shows that along B G mixed crystals separate while along A C conglomerates of mixed crystals together with the pure component separate. TYPE 4. The two components form a compound. The point C of Fig. 29 shows the composition of the compound. If solid solutions are formed in addition to the compound, the curve is modified according to cases 2 and 3, but this aspect of the problem is not yet well worked out. Haken (loc. cit.}, who has studied the alloys of tellurium with antimony, tin, bismuth and lead, has succeeded in determining the limits of 62 84 CHEMICAL COMBINATION AMONG METALS. chemical combination in these alloys. He has further noted that for an alloy which shows a smaller electrical con- ductivity the thermo-electric force is greater. Thermal Conductivity. From the intimate relations existing between electrical and thermal conductivity recently elucidated to a con- siderable extent by J. J. Thomson's corpuscular theory the laws regulating electrical conductivity may be expected to hold also for thermal conductivity. One well-defined relation is that the thermal conductivities of metals are nearly proportional to their electrical conductivities. This rule was demonstrated by the work of G. Wiedemann. 1 In the following table we give some values obtained by Wiede- mann for the copper-zinc and tin-bismuth alloys. The coefficient of thermal conductivity is given under K, that of electrical conductivity under X. Alloys. K A Cu-Zn ratio 8:1 27-3 25-5 6-5: 1 29-9 30-9 4-7 : 1 31-1 29-2 2-1 : 1 25-8 25-7 Sn-Bi ' 3:1 10-1 9 ? 1 : 1 5-6 4-3 1 :3 2-7 2 The thermal conductivity K of certain binary alloys can be calculated according to F. A. Schulze 2 by the mixture rule ; in the case of the bismuth-lead and bismuth- tin alloys, on the other hand, the value of K is less than that calculated by this rule. According to W. Voigt, 3 the determination of the ratio 1 Pogg. Ann., 108, 393 (1859) ; Ann. Chim. Phys., (3), 58, 126 (1860) ; Phil. Mag., (4) 19, 243 (1860). 2 F. A. Schulze, D.-A., 9, 555 (1902). 3 Wied. Ann., 64, 95 (1898). PHYSICAL PKOPEKTIES. 85 K : K L of the thermal conductivities of two substances in the case of heat flowing across the junction of two laminae may be given by K : /q = tan (f> 1 : tan <, where , <^ 1 are the angles formed by the directions of heat flow in the two substances with the normal to their line of separation. The number of determinations made hitherto on the thermal conductivity of alloys is small. Among the most important researches bearing on the question of the occur- rence of intermetallic compounds are those recently made by A. Eucken and G. Gehlhoff, 1 who have determined the conductivity of the cadmium-antimony alloys in which the compound CdSb (see Chapter V) occurs. They found that while for the metals the temperature coefficient of conduc- tivity has a value approximately 1-3, for the compound the value is about 2-8. It must, however, be borne in mind that the antimony-cadmium compound approximates in character to a metalloid, which renders it impossible to draw general conclusions from these results. At present this branch of the subject is comparatively little understood, and what knowledge we possess does not permit us to make any generalisations. The methods employed for the determina- tion of thermal conductivity are as various as the theoretical principles underlying them. 2 A physical constant of considerable importance is the ratio between the electrical conductivity X and the thermal con- ductivity K ; it has been used, for example, in the study of the compound SbCd by Eucken and Gehlhoff (loc. cil.). This ratio, it can be argued, is influenced by temperature and increases proportionately to it. Some determinations of the ratio - between 600 and have shown that it has an K almost constant value for metals and is approximately 1-367. The value becomes greater for alloys in which solid solutions or compounds occur. 1 Ber., 14, 169(1912). 2 Of. Chwolson, Traite de Physique, Vol. III., pp. 348 et seq. Paris, 1909. 86 CHEMICAL COMBINATION AMONG METALS. The temperature coefficient of thermal conductivity has been found to be positive for aluminium, gold and platinum and for certain binary and ternary alloys, such as manganin and constantin ; negative values are given by bismuth, tin, lead, cadmium, copper, silver, zinc, iron and nickel. Thermal Dilatation. The study of thermal dilatation has only been applied to a limited degree to metallic alloys. The investigations of FIG. 30. Matthiessen 1 and Le Chatelier 2 have led to certain gen rali- sations of value. For a solid isotropic body, length is in general a function of temperature and may be expressed by the formula : where / and 1 Q indicate the length of the body at t and 1 Pogg. Ann., 130, 50 (1867) ; Phil. Trans., 156, 861 (1866) ; Phil. Mag., (4), 31, 149 (1866): 32,472(1866). 2 C. R., 108, 1096 (1889) ; 128, 1444 (1899) ; 129, 331 (1899). Cf. also by the same author : Sur les Proprietes des Alliages in the volume Contribution d VEhide des A lli ages , p. 387. Paris, 1901. PHYSICAL PROPERTIES. 87 respectively. A, B, C, etc., are constants. The coefficient of dilatation between and t is expressed by oc : = A + 2 Bt + 3 Ct 2 + .... This coefficient has been principally used for the study of metallic alloys. Matthiessen, in the investigations mentioned, found that the dilatation of an alloy can be calculated by the mixture rule, for it is equal to the sum of the dilatations of its components. This conclusion is, however, only valid in the cases in which the alloys are uniform conglomerate mix- tures. Polymorphic transformations, for example, are always accompanied by discontinuities in the dilatation curves. A variation in the dilatation curve is to be attributed to the formation of solid solutions or of compounds. The presence of a compound, as Le Chatelier found, is indicated by the occurrence of a maximum in the dilatation curve. Le Chatelier studied the dilatation of the copper-antimony and copper-aluminium alloys, in which the compounds SbCu 2 , SbCu 4 , AlCu 3 and Al 2 Cu appear. Comparing the fusion curve with the dilatation curve of the system CuAl, it is seen that at the first maximum in the one curve there corresponds a discontinuity in the other curve. The methods used for the determination of dilatometric data are various, 1 but will not be set out in the present volume. It may be mentioned that Tammann in col- laboration with Sahmen 2 has described a very accurate dilatometer. The investigations of Svedelius 3 may also be mentioned. He has noted that in the cooling as well as in the heating of steel there occurs at about 660 and 730 an abnormal contraction which is connected with the carbon content of the metal. Guillaume 4 has also worked on the alloys of nickel. 1 See Chwolson, Traile de Physique, Vol. III., pp. 90 et seq., and Cuillet, Etude Theorique des Alliages Me'talliques f , p. 147. Paris, 1904. Le Chatelier describes an optical method. Of. Contrib. a V Etude des A lliages, p. 387. 2 Ann. d. Phys., (4), 10, 879 (1903). 3 Phil Mag., (5), 46, 173 (1898). 4 C. R., 124, 170, 752 (1897) ; 125, 235 (1897) : 136, 303, 350 (1903). Jour, de Phy*., (3), 7, 264 (1898). 88 CHEMICAL COMBINATION AMONG METALS. Hardness. Although it is difficult to give an exact scientific definition of the hardness of a body, it has come to mean the resistance which it opposes to a force acting on its surface. It has been found that the hardness of alloys is propor- tional to the composition when the constituents are present in the form of a conglomerate. The presence of solid solu- tions or definite compounds has, however, a considerable influence on the form of the hardness -composition curve. The most important investigations on this property in rela- tion to the composition of binary metallic alloys are those of Kurnakoff and his co-workers. 1 Their studies have resulted in a rapid development of this method of investigation, which often furnishes reliable evidence as to the composition of alloys. Before describing the applications of the study of hard- ness to metallic alloys in which compounds occur, we may mention that this physical property has often claimed the attention of investigators. Thus, J. K. Eyd- berg 2 found that the hardness of elements is a periodic function of their atomic weights. He has given the values for all the elements in the order of Mendelejeff's classification. The values have been expressed by means of Moh's scale of hardness, in which the following are the degrees : 1. Talc. 6. Orthoclase. 2. Gypsum. 7. Quartz. 3. Calc-spar. 8. Topaz. 4. Fluor-spar. 9. Corundum. 5. Apatite. 10. Diamond. 1 Kurnakcff and Zemczuzny, Ze.it. anorg. CJiem., 60, 1 (1908) ; 64, 149 (1909) Kurnakoff and Pushin, ibid., 68, 123 (1910). Kurnakoff and Smirnoff, ibid., 72, 31 (1911). 2 Ze.it. phys. Cliem., 33, 353 (1900). PHYSICAL PEOPEETIES. 89 HARDNESS OF THE ELEMENTS (KYDBERG). I. II. III. IV. V. VI. VII. VIII. Li Be B C N F 6 Na Mg 9-5 Al 10 Si (2) P (-5) S (2) Cl 4 2 2-9 7 5 2 (4) K Ca Si Ti V Cr Mn Fe Co Ni 5 1-5 (3) (4) (6) 9 (6) 4-5 (5) (5) Cu Zn Ga Gy As Se Br 3 2-5 1-5 (3) 3-5 2 (-6) Rb Sr Y Zr Nb Mo Ru Rh Pd 3 1-8 (3) (4) (6) (8-5) 6-5 (6) 4-8 Ag CJ In Sn Sb Te I 2-7 2 1-2 1-8 3 2-3 (8) Cs Ba La Ce Di 2 2 (3) (3) (5) Ta W Os Ir Pt (7) (9) 7 6-5 4-3 An Hg Tl Pb Bi 2-5 1-5 1-2 1-5 2-5 The scale of hardness devised by F. Auerbach l is more rational and expresses the property in absolute degrees. Auerbach's method is based on the following argument which had previously been elaborated by H. Hertz. 2 If a convex surface of a body is compressed against a plane surface of the same body or another body a circular surface of contact will be produced, and the pressure upon this circular area eventually produces a rupture about its centre. As a measure of hardness, Auerbach takes the pressure P on the circle of contact (whose magnitude is determined microscopically) at the moment of rupture. If q be the area of the circle of contact and D its diameter, 3 P 6 P P l varies inversely as the cube root of the radius of the com- 1 Wied. Ann., 43, 60 (1891); 58,380(1896). 2 C relies Journal, 92, 156 (1882). Gesamte Werke, Vol. I., p. 155. 90 CHEMICAL COMBINATION AMONG METALS. pressed sphere. The values obtained by Auerbach for the fundamental substances on Moh's scale are the following expressed as kilograms per square millimetre : 1. Talc 5. 6. Orthoclase 237. 2. Gypsum 14. 7. Quartz 308. 3. Calc-spar 92. 8. Topaz 525. 4. Fluor-spar 110. 9. Corundum 1,150. 5. Apatite 170. 10. Diamond ? According to Bottone * the hardness of metals is propor- tional to where d is the density and m the atomic weight itv of the metal. E. Benedicks, 2 in a later paper, in which the hardness of alloys and metals is studied from a general and theoretical point of view, examines Bottone's results, from which, as shown, he deduces the formula 0=4 The ratio is called by Benedicks the atomic concentra- 7ft' tion. In this property of solid bodies it is easy to trace an analogy with gases 3 for which according to Avogadro's law the density is proportional to the atomic concentration at a given temperature. VARIATION OF HARDNESS WITH THE COMPOSITION OF ALLOYS. It has already been said that in binary alloys formed of simple conglomerates, the hardness varies almost always in a linear manner with the composition ; but, in practice, small deviations may be noted. Considering, however, that the observations are not always made under comparable con- 1 Chem. News, 27, 215 (1873). 2 Zcit. phys. Chem., 36, 529 (1901). 8 Benedicks calls attention to this analogy, which is indeed not the only analogy existing between different states of matter. PHYSICAL PEOPEETIES. 91 ditions and that the mode of preparation of an alloy has an influence on its hardness, such deviations do not greatly invalidate the general conclusion. Fig. 31 shows the form of the hardness curve where the alloys are composed of simple conglomerates. The hardness varies in a linear fashion between the pure components A and B, whose hardness is represented by A' and B f . The case is different where alloys are constituted of isomorphous mixtures (solid solutions) or contain one or FIG. 31. more definite compounds. The investigations of Kurnakoff already noted have not only embraced such cases but also cases in which a compound forms solid solutions with the components. Kurnakoff's first rule for the hardness of isomorphous mixtures is that the property increases so as to reach a maximum for the median composition of the series of mixed crystals. This is shown in Fig. 32. Tammann 1 has given a rational explanation of this 1 Cf. Tammann, Ueber die fip.ziehungm zwixchen den inneren Kmflen und Eiyensch. derLosunyen, p. 35 (1907) ; also Lehrb. d. Metallographie, p. 332. 92 CHEMICAL COMBINATION AMONG METALS. behaviour. He supposes that in a given mixture the forces of attraction between dissimilar molecules are greater than C O FIG. 32. those between similar molecules. From this it follows that for liquids, the internal pressure, resulting from the sum of the forces of attraction per unit surface, is increased by the Cone FIG. 33 PHYSICAL PKOPEBTIES. 93 addition of another substance, and this must be the case with alloys. According to Tammann the attraction of two dissimilar molecules is greater both in the isotropic and in the anisotropic state. When the formation of a compound takes place between two metals, the hardness increases and the maximum corre- sponds to the composition of the compound. The curve representing this behaviour consists therefore of two straight lines starting from points representing the hard- ness of the components and intersecting in a maximum point (Fig. 33). The following table, giving the hardness of various inter- metallic compounds, illustrates this point. The values for the pure components are also given. Hardness is given in degrees on Moh's scale. Compounds. Hardness of Compound. Components. Mg 2 Sn 3-5 Mg- 2 Sn - 1-8 Mg 2 Pb 3-5 3 J Pb - 1-5 MgCu 2 4-5 33 Cu 3 Mg 2 Cu 4-5 53 Zn 2 Cu 4-5 Zn - 2 Cu 3 Sn 4-5 Sn 1-8 Cu 3 P 6 P -5 Cu 3 Sb 4-5 Sb -3 CdSb 3-5 55 Cd -2 CoSb 5-5 33 Co -4 + NiSb 5 - 5-5 3 3 Ni -4 + CoSn 5-5 Sn - 1-8 Co -4 + CoSn 2 5-5 33 53 PbSb 2 5-5 Sb -3 Pb - 1-5 Ni-As 2 6-5 Ni -4 As -3-5 Ni 5 P 2 5-5 33 P -5 PtSn 5-5 Pt - 4-3 Sn - 1-8 NaCdj, >.3 Na 4 Cd 2 Na 2 Pb 2-5 Pb - 1-5 NaPb 3 + 3 3 3 3 NaSn 3 53 Sn - 1-8 94 CHEMICAL COMBINATION AMONG METALS. When a compound forms solid solutions with the com- ponents, the hardness curve differs from the preceding cases described. Here two maxima appear and a point of inter- section. The maxima correspond to the formation of mixed crystals and the point of intersection of the two branches to the compound (see Fig. 34). Kurnakoff and Smirnoff 1 have classified the hardness curves for alloys in which, in addition to the compound, solid solutions are formed. We FIG. 34. have discussed the question of compounds of variable composition in Chapter III. Kurnakoff and Smirnoff divide the cases in question into two groups: (1) com- pounds which are not dissociated either in the liquid or the solid state ; and (2) compounds which dissociate on fusion. We will examine xnore closely the significance of this classification. Group 1. These compounds belong to the category of hylotropic phases mentioned by Ostwald. 2 Two types are 1 Zeit. anorg. Chem., 72, 31 (1911). 2 Zeit. /. Elektrochemie, 10, 572 (1904). PHYSICAL PBOPEBT1ES. 95 distinguished ; in the first, (a), the compound forms a con- tinuous series of solid solutions with the components ; in the Conceit tx(x$>'o 114 A"iDn FIG. 35. S 96 CHEMICAL COMBINATION AMONG METALS. second, (b), the compound forms solid solutions only to a limited extent. TYPE (a). This case is shown in Fig. 35, where c repre- sents the melting point of the binary compound A m B n . Composition FIG. 37. JJ0 This may be between that of the two components, as in the diagram, or even above or below them. The two systems of isomorphous mixtures A + A m B n and A m B n + B are represented by the two curves % df and/ e b 1 with maximum points in d and e. The minimum/ corresponds to the com- PHYSICAL PROPEKTIES. 97 pound A m B n ; this maybe found also above a x and b l9 i.e., the hardness of the compound A m B n may exceed that of the components. TYPE (b). As was said, this type is characterised by a gap in the series of solid solutions. The compound (see Fig. 36) shows a maximum and forms solid solutions with excess of A and B in the region limited by the lines e l c 1 g d^f v The curve I f n represents the variations of hardness in the interval of concentration e 1 / x . The presence of solid solutions is accompanied by an increase in hardness and the two branches intersect in /, which represents the com- pound. Group 2. The behaviour of substances of this group is represented by Fig. 37. The compound melts at the point D and gives rise to the component B. D l which is the point of intersection of the horizontal D D l with the curve C l M (which indicates the concentration of the solid solutions) represents the highest temperature at which the solid phase of the compound is stable. The branches D M K and D l M K v which pass through the obscured maximum M, belong to the labile state of the compound. Discontinuities occur at P and Q corresponding to changes in the hardness of the solid solutions with A and B. The measurements of hardness made up to the present have been carried out either by means of the sclerometer or by Brinell's method. Compressibility. In direct relation with the hardness of a substance is the property of compressibility under pressure. Kurnakoff has used this property also in his investigations on the variation of the physical properties of alloys with their composition (see Bibliography, p. 88). Since this property is intimately connected with hardness there is no purpose in giving any particular detailed treatment of the subject. It is sufficient to mention that in isomorphous mixtures the compressibility C-M. 98 CHEMICAL COMBINATION AMONG METALS. as well as the hardness, exhibits a maximum on the curve showing its relation to composition. The apparatus used by Kurnakoff is described in the first memoir, 1 and has been used for conglomerates of mixed crystals, metals, inorganic salts and organic substances. Among the substances thus examined are the thallium-lead and thallium-bismuth alloys, binary mixtures of silver chloride and bromide, potassium iodide and bromide, stearic and palmitic acids, p-dichloro- benzene and p-dibromobenzene. The researches of Kurnakoff on the compressibility of the thallium-bismuth alloys are above all worthy of note since they have served in the determination of the irrational maximum of the compound of variable composition whose formula is approximately Bi 5 Tl 3 (see Chapter V). Crystalline Form of Binary Compounds. Comparatively few intermetallic compounds have been studied crystallographically, so that we possess very little definite information on this subject. Studies are also want- ing on the relations between crystalline form and degree of saturation of compounds. The crystalline forms of metals are generally known ; they crystallise most often in the regular or the hexagonal- rhombohedral systems. Silver, cadmium, iron, iridium, mercury, nickel, gold, osmium, lead, platinum, palladium and others of the platinum group belong to the regular system. The latter are, however, dimorphous and crystallise also in the hexagonal system. Arsenic, beryllium, antimony, bismuth, magnesium, tellurium and zinc crystallise in the hexagonal system. Potassium and sodium are tetragonal, while tin crystallises in the tetragonal and also in the rhombohedral system. Zirconium is monoclinic. The inter- metallic compounds given in the following table have been studied crystallographically : Zeit. anorg. Chem., 60, 22 (1908). PHYSICAL PKOPEETIES. 99 Compound. Crystalline system. Compound. Crystalline system. NaCd 2 Cubic FeZn 7 Hexagonal Mg 2 Sn NiAs (Niccolite) Ag- 2 Te ! NiSb Ag 4 Zn Bj 2 Te 3 (Tetradimite) Ag 2 Zn Ni 2 Te s (Melanite) AgZn Cd s Sb 2 Orthor lombic Cu 6 Ni FeSb 2 Cu 2 Zn AgaSb (Dyscrasite) CuZn ZnSb CuZn 8 FeAs 2 (Lollingite) PbTe (Altaite) CoAs 2 (Smaltite) NiAs 2 (Ranimelsbergite) CoAs 2 (Safflorite) NiAs 2 (Cloantite) CeAl 4 CoAs 3 LaAl 4 PbAs 2 (Sperrilite) 1 ThAl 4 CuSn 2 Hex a gonal FeAl 3 Monoclinic CuSn Cu 4 Sn ; AsSn 2 i Tetragonal Our information on other intermetallic compounds is very limited. Groth, in his Chemische Krystallographie, 1 gives certain data about artificially prepared compounds, whose crystallographic characters are not yet well under- stood. (a) Bivalent Metals. The compounds of beryllium and magnesium with monovalent metals have not yet been studied crystallographically. The zinc compounds with copper and silver described by Behrens 2 have been given in the above table. The amal- gams of lithium, sodium and potassium are very numerous ; they crystallise partly in needles and partly in hexahedral forms. Some of the copper, silver and gold amalgams crystallise in the cubic system. The compound CoHg 5 was obtained in prisms which, however, were not examined very exactly. Crystalline amalgams exist of barium and strontium. 1 Vol. I., pp. 43 et seq., 1906, 2 Gefiige d. Metallc u. Legierungen, pp. 46 and 100. 1894. Hamburg and Leipzig, 7 2 100 CHEMICAL COMBINATION AMONG METALS. (b) Trivalent Metals Behrens l and Guillet 2 obtained the following crystalline compounds of copper and aluminium : Cu 3 Al, CuAl and Cu 2 Al 3 ; Petrenko 3 obtained two crystalline compounds of aluminium and silver, namely, Ag 3 Al and Ag 2 Al ; Heycock and Neville 4 prepared from aluminium and gold, Au 4 Al, Au 2 Al and AuAl 2 ; but the crystalline form of all these compounds was not determined. Thallium forms compounds with the monovalent metals in which it behaves as a heavy metal. KurnakofT, 5 from the thallium-sodium and thallium-potassium alloys, obtained the compounds NaTl and KT1, the former in the form of three-rayed aggregates and the latter in the form of cubes. The thallides of the bivalent metals, of which some of magnesium are known, are very little recognised. (c) Tetravalent Metals. In addition to the stannides shown in the preceding table, Groth 6 refers to the com- pound Cu 3 Sn which crystallises in small prisms. The compound CuSn 2 was described by Miller 7 and Eam- melsberg 8 as occurring in hexagonal prisms. The stan- nides of silver have not yet been isolated nor have the compounds with gold, AuSn, AuSn 2 and AuSn 4 described by Vogel. 9 Copper and silver form double plumbides : Curlt 10 describes the compound CuAgPb 2 as forming regular octahedra. The gold plumbides described by Vogel, 11 Au 2 Pb and AuP 2 crystallise respectively as rhombs and long needles. Among the stannides of iron, Fe 4 Sn, Fe 3 Sn, FeSn and Op. cit., p. 107. C. R., 133, 684 (1901). Zeit. anorg. Chem., 46, 49 (1905). Trans. Hoy. Soc., 194, 201 (1900). Zeit. anorg. Chem., 30, 86 (1902). Op. cit., p. 46. Phil. Mag., 1835, p. 107, and Pogg. Ann., 36 478. 8 Zeit. anorg. Chem., 46, 60 (1905). 9 Pogg. Ann., 120, 34 (1863). 10 Uebers. d. pyrogen. Kunstl. Mineralien, Freiberg (1857), p. 17. 11 Zeit. anorg. Chem., 45 (1905). PHYSICAL PEOPERTIES. 101 FeSn 2 , described by Headden and Stevanovitch, 1 are known. The bistannide crystallises in long needles. Few stannides of the trivalent metals are known : the stannides of aluminium, AlSn 4 and AlSn, prepared by Guillet 2 have not been studied crystallographically. It is worth while recording here the observations of Barlow and Pope 3 on a regularity in the crystalline form of binary compounds formed by the union of elements of equal valency ; such compounds crystallise in the cubic or hexagonal systems and generally in classes with a lesser degree of symmetry. According to these authors, the mole- cules which form the homogeneous structure which is the crystal are constituted of similar atoms. But beyond this similarity of the atoms it is possible to think of other factors which influence crystalline form. The homogeneity or uniform structure of the crystal is conditioned by two oppos- ing forces : (a) a repulsive force due to the kinetic energy of the atom, and (b) an attractive force varying as the square of the distance between the atoms. Concerning this question of the nature of the forces which maintain the component atoms of crystals in equilibrium, Nernst and Lindemann 4 have recently stated that the attractive force is identical with that of chemical affinity. This, of course, throws additional light on the regularity noted by Barlow and Pope. Natural Intermetallic Compounds. Among the intermetallic compounds studied, the majority are homopolar, resulting from the action of unitary forces. Compounds of this group are not ionisable and exhibit in their properties the properties of their constituent elements. All intermetallic compounds, however, are not formed by 1 Zeit. f. Krystall, 40, 327 (1905). 2 C. R,, 133, 935 (1901). 8 Trans. Chem. Soc., 89, 1675 (1900). 4 Cf. Nernst, The Theory of the Solid State, p. 4. London, 1914. 102 CHEMICAL COMBINATION AMONG METALS. the union of homopolar elements. Indeed, certain metal- loidal elements are able to form with metals true alloys having markedly metallic characters. Among such metalloidal elements are carbon, silicon, boron, tellurium, selenium, phosphorus and arsenic. The natural intermetallic compounds are generally hetero- polar. Native metals are few in number and chiefly alloys of the elements of the iron group (iron, cobalt and nickel) and of platinum, mercury, copper, silver and gold. Among these metals those with the highest melting points tend to form solid solutions. Thus gold forms solid solutions with silver (electrum), palladium and rhodium, osmium and iridium from iridosmin ; and lastly, in meteorites, nickel and iron are found united in solid solutions. The following tellurides occur naturally : AgTe Silvanite and em- pi essite. Ag 2 Te Hessite. Ag 4 Te Stutzite. AuTe 9 Colaverite. AgAuTe 2 - - Krennerite. AgAuTe -- Muthmannite. HgAu 2 Ag Te Kalgoartite, Bi 2 Te 3 Tetradimite. HgTe - - Coloradoite. PbTe - Altaite. Melanite. (AgAu) 2 Petzite. (AgAu) 2 Te 6 Goldschmite. Silvanite, colaverite and krennerite, though recognised as definite minerals have not been encountered in the study of fusion diagrams. Pellini 1 has found recently in the study of the system silver-gold-tellurium by thermal analysis the compound (AgAu) 3 Te 2 , which has not been discovered among the gold minerals hitherto investigated. Empressite was found recently by M. W. Bradley 2 ; Pellini and Quercigh 3 have also noted the existence of the compound AgTe by thermal methods. Stutzite is probably a mixture of Ag 2 Te and silver, while melanite probably contains the compound NiTe. 1 Gazz. Chim. Ital, 45, I., 47 (1915). 2 Journ. of Science, 38, 163 (1914). 3 It. Ace. Lincei, 19, II., 415, 445 (1910). PHYSICAL PROPERTIES. 103 A bismuth telluride must also be mentioned whose com- position is not exactly known and which has been called Joseite. The following native antimonides and arsenides occur : Ag 3 Sb Dyscrasite. Ag 6 Sb (variety of dyscra- site). Cu 6 Sb Horsfordite. p A j^^u^. NiSb Breithaufite. LS2 |Safflorite. Ag 3 As Arsenoargentite. CoAs 3 Skutterodite. NiAs Niccolite. Cu 3 As Domeichite. Cu 6 As Algodonite. Cu 9 As Whitnegite. (Lollingite. ^{Arsenoferrite. Fe 3 As 4 Lemopyrite. p A (Smaltite. LS2 |Safflorite. CoAs 3 Skutterodite. NiAs Niccolite. NiAs { Cnloantite - 2 (Rammelsbergite. PbAs 2 Sperrylite. Co(AsBi) 3 Bismuthosmal- tite. A selenide of bismuth, guanajnatite, exists of the com- position Bi 2 Se 3 ; silvanite, to which at first the formula Bi 3 Se was assigned, 1 has been since 2 recognised to be a mixture of Bi 2 Se 3 and bismuth. N. Parravano, 3 in a study of the system Bi Se, has demonstrated the existence of two selenides, Bi 2 Se 3 and BiSe. The mineral naumannite, Ag 2 Se, found in nature, is isomorphous with argentite and hessite. Naumannite and hessite are found in isomorphous mixtures in the mineral aguilarite. The system silver-selenium has recently been studied by Pellini. 4 Some of the minerals here mentioned and at present regarded as definite compounds are probably isomorphous mixtures. Although a mineral may have a uniform structure and a constant composition it is not always a true chemical individual. In this connection it is well to bear in mind that several native compounds have not been 1 Zeit.f. Kryst., 1, 499 (1877). 2 Ibid., 6, 96 (1888). 3 Gazz. Chim. ltd., 43, L, 201 (1913). 4 Ibid., 45, I., 533 (1915). 104 CHEMICAL COMBINATION AMONG METALS. recorded in the thermal study of the mixtures of their constituent elements. The recent developments of the study of heterogeneous equilibria and the introduction of Tammann's method into physico-chemical research, although they have not led to a complete elucidation of this important aspect of mineralogy, have nevertheless furnished powerful means of investigation which promise much for the future. CHAPTER V. HOMOPOLAR INTERMETALLIC COMPOUNDS. Compounds of the Elements of Group I. among themselves. 1st sub-group. The only compound known in this sub- group is that formed by sodium with potassium. The sodium-potassium compound has the formula Na 2 K. The system has been studied by Kurnakoff and Pushin l and the diagram is shown in Fig. 38. According to Kurnakoff the curve shows a break corresponding to 40 per cent, of potassium at 638 ; this may demonstrate the existence of a compound of the formula Na 2 K or Na 3 K 2 . Bornemann 2 on the other hand believes that there are three species of mixed crystals between the two pure components. The branches of the curve Na a and Ky are convex to the axis representing concentration, showing that the two metals instead of separating in the pure state form solid solutions. According to Bornemann the presence of a strongly dis- sociated compound of the type Na n K must be admitted. If we hold with Kurnakoff that the break on the fusion curve corresponds to the compound, so that the crystals {3 saturated with sodium are identical with the compound, we may have Na 2 K or Na 3 K. The composition represented by Na 3 K is near a point of arrest. Bornemann holds that it should be characterised by a more decided point of arrest and that the compound is probably Na 2 K. 3 Lithium is miscible with sodium and potassium with 1 Zeit, anorg. Chem., 30, 109 (1902). 2 Die binaren Metallegierungen, Halle (1905), p. 7. 3 Of. also Van Bloiswyk, Zeit. anorg. Chem., 74, 152 (1912). 106 CHEMICAL COMBINATION AMONG METALS. difficulty. 1 Nothing is known of the behaviour of lithium and sodium with caesium and rubidium. 2nd sub-group. Copper, silver, and gold give alloys with to" so" io" 20 AO 20 y/ ///x v/ XPS. TO 60 -10O FIG. 38. each other with complete or partial formation of solid solu- tions. Sodium alone among the elements of the first sub- group forms the compound NaAu 2 . The system sodium- 1 Zeit. anorg. Chem., 67, 183 (1910). HOMOPOLAR INTEBMETALLIC COMPOUNDS. 107 gold has been studied by Mathewson, 1 who has noted the formation of a compound. Silver and copper on the con- trary do not combine with sodium, which, of course, is contrary to Tammann's rule. The diagram (see Fig. 39) is fairly simple. At a concen- l*Q FO 60 7-0 90 010 lo FIG. 39. tration of 3-6 per cent, of gold there is an eutectic point ; at 66-6 per cent, of gold and at 989 there is a maximum corre- sponding to the compound NaAu 2 . A second eutectic between the compound and gold occurs at 876. There is no indication of the formation of solid solutions in the diagram. The compound NaAu 2 is chemically very resistant and exhibits a considerable degree of hardness. 1 Intern. Zeit. /. Metall , 1, 85 (1911). 108 CHEMICAL COMBINATION AMONG METALS. Compounds of Elements of Group II. with each other. 1st sub-group. The behaviour of the metals of this sub- group has not been studied as yet. 2nd sub-group. Magnesium forms with zinc the compound MgZn 2 ; with cadmium it forms the compound MgCd, and with mercury a compound whose composition is not yet determined. 700 C 300- 10 -100 FIG. 40. Magnesium-zinc. The system magnesium-zinc has been investigated by Boudouard l and Grube. 2 From the diagram constructed by the ] atter, who has traced cooling curves from 650 to 250 (see Fig. 40), Grube inferred the existence of a compound MgZn 2 , which is formed at 590, and at 33 per cent, magnesium. A small deviation noted 1 C. R., 139, 424 (1904). Bull Soc. Chim., (3), 31, 1201. 2 Zeit. anorg. Chem,, 49, 77 (1906). HOMOPOLAR INTERMETALLIC COMPOUNDS. 109 near the eutectic point between magnesium and MgZn 2 is due to the presence of magnesium or the compound, which must be attributed to super-cooling. Grube denies the formation of solid solutions on the one hand between the compound and magnesium, and on the other hand between zinc and the compound. Magnesium tends on first separation, if present in small quantity, to crystallise in dendritic forms. Accord- ing to Boudouard the compound Mg 4 Zn is also formed. This, however, is denied by Grube, arguing from the theory of heterogeneous systems, since the eutectic horizontals of 700' 600- 300" 100 FIG. 41. the compound MgZn 2 extend over the whole diagram. Microscopic analysis also excludes the formation of Bou- douard's supposed compound. The compound MgZn 2 is unattacked by water and air ; it has a brilliant white colour and is a little harder than its constituents. It is very brittle. Magnesium-cadmium. The system magnesium- cadmium has been investigated by Boudouard l and Grube. 2 Accord- ing to Boudouard the compounds MgCd, Mg 4 Cd and Mg 3 Cd occur, while Grube only reports one compound, namely, MgCd. From the diagram (Fig. 41) he argues complete miscibility both in the liquid and solid states. 1 C. R., 134, 1431 (1902). Bull. Soc. Chim., (3), 27, 854 (1902). 2 Zeit. anorg. Chem., 49, 72 (1906). 110 CHEMICAL COMBINATION AMONG METALS. Near 50 per cent, of cadmium the crystallisation interval between liquid and solid practically vanishes so that the liquidus and solidus curves touch. Between 35-9 per cent, and 66-4 per cent, of magnesium the compound MgCd and the mixed crystals of the compound and its components undergo a transformation into another series of mixed crystals. By crystallising slowly, homo- geneous alloys are obtained. -foo FIG. 42. The compound MgCd is greyish-white in colour, a little harder than its constituents, and very resistant to water and moist air. MAGNESIUM AND CALCIUM AMALGAMS. Mercury -magnesium. This system has recently been studied by L. Cambi and G. Speroni. 1 The two elements combine to form the compounds MgHg 2 and MgHg. The system has been investigated up to a concentration of 50 per 1 R. Ace. Lincei, 24, I., 734, 932 (1915). HOMOPOLAK INTEEMETALLIC COMPOUNDS. Ill cent, magnesium. Preliminary researches were made by Wanklin and Chapman 1 and by Kerp and Bottger. 2 By addition of a very small quantity of magnesium the melting point of mercury is lowered (see Fig. 42). The diagram then rises in a curve to 168, where there is a dis- continuity corresponding to the compound MgHg 2 . The existence of the compound MgHg has not been proved directly by the thermal method, for this series of observations \oo FIG. 43. terminates at 5-08 per cent, by weight or 30 per cent, atomic of mercury. The amalgam with 32 34 per cent, atomic magnesium boils at 415. Cambi has also studied the elec- tromotive force of magnesium amalgams. In 1904 Evans and Fetsch, 3 working on magnesium amal- gam as a reducing agent, prepared a homogeneous amalgam corresponding to the formula MgHg 2 from one part mag- nesium and eighteen parts of mercury. The formation of 1 J. C. S., (2), 4, 141 (1866). 2 Zeit. anorg. Chem., 25, 33 (1000). 3 /. Am, (7, S., 26, 1158 (1904). 112 CHEMICAL COMBINATION AMONG METALS. this compound is shown clearly by the thermal study of the system. Mercury-calcium. The affinity of calcium for mercury is very slight ; at ordinary temperatures calcium dissolves slowly in mercury. The capacity for combination between the two metals has long been known, but the system has only been studied recently by Cambi and Speroni 1 who have shown the existence of the compound CaHg 4 . The diagram is shown in Fig. 43 and only extends to 33 per cent, atomic of calcium. The compound melts with decom- position at 266 C. Thermal analyses do not indicate the existence of any other compounds. Moissan and Chavanne 2 had previously described a compound of the formula CaHg 2 crystallising in hexagonal prisms. J. Schiirger 3 recorded the existence of a crystalline compound CaHg 5 while J. Feree 4 described a compound Ca 3 Hg 4 easily altered on exposure to air. Thermal analysis, however, renders the existence of the last three compounds very improbable. Strontium-mercury. Strontium like the other metals of the alkaline earths combines chemically with mercury. From the researches of Kerp and Bottger 5 the existence of the compound SrHg 12 , in equilibrium with the liquid phase at 30, appears established. It decomposes at 60 to 70 and has a silvery appearance. An amalgam of strontium containing 5-37 per cent, of this metal separates a crystalline phase at 81. The system has not yet been investigated by thermal methods. Barium-mercury. Barium combines with mercury. Kerp 6 and Kerp and Bottger 5 isolated two amalgams of the com- position BaHg 12 and BaHg 13 by electrolysis of a saturated solution of barium chloride, using a mercury cathode. 1 R. Ace. Lincei, 23, II., 599 (1914). 2 C. R., 140, 125 (1905). 3 Zeit. anorg. Chem., 25, 426 (1900). 4 C. R., 127, 619 (1898). 5 Zeit. anorg. Chem., 25, 1 (1900). Ibid., 17, 284 (1898). HOMOPOLAE 1NTEEMETALLIC COMPOUNDS. 113 BaHg 12 crystallises in silvery cubes which become oxidised in the air. The amalgam containing 5-34 per cent, of barium separates a crystalline phase at 99. The system has not yet been studied thermally. COMPOUNDS OP CALCIUM WITH METALS OF THE SECOND SUB-GROUP. Calcium-magnesium. Calcium combines with magnesium, forming the compound Ca 3 Mg 4 . The system has been loo (To J o |o / ""- /> f ' J& ^-- FIG. 41. worked out by Baar. 1 The diagram traced by him (Fig. 44) shows that in the liquid state magnesium mixes with calcium in all proportions. The liquidus curve shows a maximum at 55 per cent, (by weight) of calcium and 715. Here the compound Ca 3 Mg 4 separates. Microscopic examination confirms the deductions made from the diagram, and shows that the compound exists in polyhedric crystals. The magnesium-calcium alloys, particularly those found in the vicinity of the composition of the compound, are C.M. Zeit. anonj Chem., 70, 362 (1911). 114 CHEMICAL COMBINATION AMONG METALS. very brittle ; the alloys richer in calcium are more ductile. Exposed to air these alloys crumble to a grey powder in which are seen shining particles of the compound which is stable in air. Calcium-zinc. This system was studied by Donski, 1 who FIG. 45. noted the formation of four compounds, namely, CaZn 10 , CaZn 4 , CaZn 3 , and Ca 4 Zn. On the fusion curve (see Fig. 45) between 9 and 20 per cent. Ca, there are distinct breaks at 677 and 717. The compound CaZn io separates from its melt and from melts containing large proportions of it only 1 Zeit. anorg. Ckem., 57, 185 (1908). HOMOPOLAB INTEKMETALLIC COMPOUNDS. 115 after a super-cooling of about 16. The second compound crystallises at 680 from melts containing 9-23 per cent, of calcium. This compound either melts without decomposi- tion or else in melting splits off a small quantity of the com- pound CaZn 10 . The first case occurs if the composition coincides with a point of transition, the second case when the composition is represented by points lying to the left of the said point. The curve rises again up to 40 per cent, calcium at a temperature of 688, where the compound Ca 2 Zn 3 separates. In the further course of the curve small arrests are noted at 431, which probably correspond to a compound CaZn. Other arrests occur at 385, between 52 and 84 per cent, of calcium with a maximal arrest at 80 calcium. This means that another species of crystal is formed having the formula Ca 4 Zn. The alloys from 6 per cent, calcium have the colour of pure zinc and are a little harder ; they are fairly stable in air, while those containing from 6 19 per cent, of calcium soon become grey. From 20 29 per cent, they decompose water, giving a black powder, and with greater energy the richer they are in calcium. The brittleness of these alloys increases from pure zinc up to 30 per cent, of calcium, diminishing thence as the proportion of the latter metal increases. Calcium-cadmium. Calcium combines with cadmium, forming the three compounds CaCd 3 , CaCd and Ca 3 Cd 2 , according to Donski. 1 The curve (Fig. 46) shows a break at 615 and 24 per cent, of calcium and an eutectic point at 316. The compound CaCd 3 separates only after a super-cooling of about 8. At 685 from 27 to 84 per cent, of calcium, a lack of miscibility in the liquid state is noted. The compound CaCd forms at 50 per cent, of calcium, and arrests are observed corresponding to it. At about 515, a new species of crystal separates which may be the compound of the formula Ca 3 Cd 2 , but this is not quite certain since the eutectic horizontal extends to 50 per cent, of calcium, whereas it 1 Zeit. anorg. Chem., 57, 193 (1908). 82 116 CHEMICAL COMBINATION AMONG METALS. should extend only to 40 per cent. The last arrests are at 415 from 50 95 per cent, calcium, and correspond probably to mixed crystals. The alloys containing up to 10 per cent, of calcium are fairly stable in the air and a little harder than calcium ; from 750' 700 100 300* 250 o u Fm. 46. 10 26 per cent, calcium they are less stable and decompose cold water ; alloys containing higher proportions of calcium are still more easily oxidised. Alloys become increasingly brittle from 10 up to 40 per cent, calcium, and then there is a decrease in this property. Zinc does not form compounds with cadmium or mercury, nor mercury with cadmium. HOMOPOLAK INTEBMETALLIC COMPOUNDS. 117 Compounds of Elements of Group III. with each other. The reciprocal behaviour of the elements of this group has not yet been closely studied. Tammann's first rule holds for this group. Gallium, indium and thallium form a com- plete series of solid solutions. Thallium and aluminium are not miscible in the liquid state according to Doerinckel. 1 It is probable that aluminium can combine with lanthanum. Aluminium-lanthanum. This system has not been studied thermally. Muthmann and Beck 2 by fusing together the two elements have obtained rhombic or monoclinic crystals, isomorphous with the compound of cerium and aluminium ; their composition corresponds to the formula LaAl 4 . The two metals react energetically on fusion. The compound has a white colour ; its specific gravity is 3-923, which is approximately equal to the value calculated by the mixture rule. Compounds of Elements of Group IV. with each other. Two elements, namely, carbon and silicon, occur in this group, which are characterised by great capacity for com- bination. We shall have occasion later to deal at length with the carbides and silicides ; they are of great importance on account of their relationship with compounds of a true metallic character. Silicon and carbon form a compound SiC called carborundum, which apart from its physical pro- perties is of interest because it occupies a position inter- mediate between intermetallic compounds and the class of compounds to which sulphides and oxides belong. Little is known at present of the alloys of titanium, zirconium, germanium and cerium with each other and with other elements of this group. Cerium combines with tin and lead. Cerium-tin. Cerium forms with tin the compounds Ce 2 Sn, Ce 2 Sn 3 and CeSn 2 . The system has been investi- 1 Zeit. anorg. Chem., 48, 188 (1906). 2 Ann. Chem,, 331, 51 (1904). 118 CHEMICAL COMBINATION AMONG METALS. gated by Vogel. 1 As the diagram (Fig. 47) shows, 2 the three compounds melt at a temperature much above the melting points of their components. Thus CeSn 2 melts at 1135, Ce 2 Sn 3 at 1165 and Ce 2 Sn at 1400. At low temperatures the solubility of the first and the third compounds in their FIG. 47. components is very small. Ce 2 Sn separates out at about 30 per cent, of Sn, Ce 2 Sn 3 at 56 per cent. Sn, and CeSn 2 at 64 per cent. Sn. The formation of the first two is accom- panied by a small degree of super-cooling (up to 15) which however is not observed in the case of CeSn 2 . 1 Zeit. anorg. Chem., 72, 319 (1911). 2 In this diagram the proportions are by weight and not atomic proportions. HOMOPOLAK INTEKMETALLIC COMPOUNDS. 119 It is characteristic of these compounds that they are formed with great evolution of heat, with increasing intensity in the order CeSn 2 , Ce 2 Sn 3 , Ce 2 Sn. All the alloys of cerium and tin, with the exception of those containing more than 80 per cent, of tin, are pyrophoric ; above all, those containing CeSn 2 . It should be added that these alloys are very hard and brittle. Cerium-lead. It appears from the researches of Muth- mann and Beck l that cerium has the power of reacting with other metals with ease. They isolated by chemical means a compound CeAl 4 ; cerium and zinc, also, combine with great heat evolution or even explosion. Cerium and magnesium mix with absorption of heat. Investigations with tin and lead did not give good results. Vogel 2 has studied the cerium-lead as well as the cerium- tin alloys. Adding cerium to molten lead a solid homo- geneous vitreous alloy is obtained. According to Vogel, cerium behaves with lead as with tin, so that the two equili- brium diagrams have points of similarity. Further, both the cerium-lead and the cerium-tin alloys are formed with energetic evolution of heat, giving rise to several compounds whose melting points are higher than those of their compo- nents, and which are decomposed by water with the evolu- tion of gas. The fusion diagram for the cerium-lead alloys has, however, not yet been published. Lead and tin do not form compounds ; the thermal investi- gation of the system made by Kosenhain and Tukes 3 shows the existence of solid solutions between limits. Compounds of the Elements of Group V. with each other. Nitrogen, phosphorus, arsenic, antimony form a natural group ; metallic characters are, however, only displayed to a marked extent by antimony and bismuth. By reason of 1 Ann. Ckem., 331, 46 (1904). 2 Zeit. anorg. Chem., 72, 320 (1911), 3 Phil. Trans., 209 (1908). 120 CHEMICAL COMBINATION AMONG METALS. their affinities, phosphides and arsenides are grouped apart, together with selenides, tellurides and sulphides. Antimony and bismuth do not combine chemically, but form an almost complete series of solid solutions. 1 The reciprocal behaviour of vanadium, niobium and tungsten is not as yet known. Compounds of the Elements of Group VI. with each other. Nothing is kno\vn of the reciprocal behaviour of chromium, tungsten and uranium. The selenides, tellurides and sul- phides are dealt with together with the phosphides and arsenides. The compounds formed by selenium, tellurium and sulphur with the metals of this group do not display true metallic characters. Compounds of the Elements of Group VII. with each other. The only metallic element known in this group is man- ganese ; the other elements are, of course, the halogens. Among these, notable exceptions occur to Tammann's first rule. Chlorine and iodine, indeed, form two compounds, IC1 and IC1 3 2 , while bromine and iodine form the compound BrI. 3 As these compounds do not show metallic characters it will not be necessary to give any data as to their respective systems. Compounds of the Elements of Group VIII, with each other. This group comprises three sub-groups : (1) iron, cobalt, nickel ; (2) ruthenium, rhodium, palladium ; (3) osmium, iridium, platinum. 1 Cf. Gautier, Contrib. a VEtude des Attiages, p. 114 (1901), Paris. Charpy, ibid., p. 138. Huttnerand Tammann, Ze.it. anor-j. Chem., 44, 131 (1905). Parravano and Viviani, Gazz. Chim. Ital, 40, II., 446 (1910). 2 The equilibrium diagram has been described by Stortenbecker. Zeit. phys. Chem., 3, 11 (1889). 3 C7.Meerum-Terwogt, Zeit. anory. Chem., 47, 203 (1905). HOMOPOLAE INTEEMETALLIC COMPOUNDS. 121 Iron, cobalt, and nickel mix in the liquid state with other metals in all proportions, with the exception of tin. If a gap in miscibility occurs in one case it occurs also in the other two. Such lack of miscibility decreases from iron to nickel. In the solid state, although these metals generally behave as in the liquid state, there are exceptions. For cobalt, accord- ing to Levkonja, 1 Tammann's rule holds, namely, that a metal with a higher melting point can dissolve a greater quantity of a metal with a lower melting point than can the metal with lower melting point of the metal with higher melting point. Tammann's rule for the elements of a natural group also holds for the iron group. If iron does not form a compound with other elements, neither do cobalt and nickel. This rule has been verified in forty-five cases ; the single exception is given by the compound which nickel forms with bismuth. Of the forty-five compounds known, the greatest number, nineteen, is given by nickel ; cobalt gives thirteen and iron eight. The most common formula is AB 3 then AB 2 and A 2 B 3 . According to the Periodic System iron, cobalt and nickel should each form a group, (1) with ruthenium and osmium ; (2) with rhodium and iridium ; and (3) with palladium and platinum. Levkonja (loc. cit.) maintains that his researches would appear to give support to the proposal made by Biltz 2 to reunite iron, cobalt and nickel in a single natural group. COMPOUNDS OF IKON. Iron-cobalt. This system was examined by Euer and Kaneko. 3 Two series of mixed crystals occur (see Fig. 48), one from to 83 per cent, of iron cobalt ft, and the other from 83 to 100 per cent, iron iron y. The transformation of cobalt ft into cobalt a, though occurring along a continuous curve from 30 to 83 per cent, shows a maximum correspond- 1 Zeit. anorg. Chem., 59, 339 (1908). 2 Ber., 35, 562 (1902). 3 Ferrum, 11, 33 (1913). 122 CHEMICAL COMBINATION AMONG METALS. ing to the compound Fe 4 Co 3 . Alloys belonging to this curve exhibit magnetic properties. In the mixed crystals between 30 and 83 per cent, of iron the molecules of cobalt and iron unite with the compound, which forms mixed crystals with excess of the constituents. Below 30 per cent, the crystals of Co fi change into crystals of Co a (ferro- magnetic). Above 83 per cent., crystals of Fe /3 are formed from Fe y and crystals of Fe a (strongly magnetic) from Fe /3. Iron-nickel. This system was studied by Guertler and Tammann l and by Kuer and Schiiz. 2 At about 1465 and 66 per cent, of nickel the compound 1550' Co 50 Fe 100 1V 0.0 FIG. 48. FeNi 2 occurs. The mixed crystals in equilibrium with the melt on the first branch of the curve should be considered as mixtures of the compound FeNi 2 with nickel and iron in excess, while the crystals on the second branch are to be considered as iron y in which nickel is dissolved. These two series of mixed crystals are sharply distinguished from the crystallographic point of view ; those from to 35 per cent, nickel are isomorphous with iron y while those from 35 to 100 per cent, nickel are isomorphous with the form of nickel stable at high temperatures and with the compound FeNi 2 . On cooling the two series of crystals stable at high tempera- tures, transformations occur into other species of crystals. It appears from the study of these alloys that one of the pro- 1 Zeit. anorg. Chem., 45, 211 (1905). 2 MdalL, 7, 415 (1910). HOMOPOLAR INTERMETALLIC COMPOUNDS. 123 perties which changes with greatest intensity in these trans- formations is the magnetic permeability. Compounds of the Metals of Group I. with Metals of other Groups. COMPOUNDS OF LITHIUM. The behaviour of lithium with beryllium, calcium, stron- tium and barium is not known ; with magnesium there is no combination, nor is it known whether lithium reacts with zinc. 150 O 10 Li ZO JO 50 &0 70 80 30 100 Cd FIG. 49. Lithium-cadmium. Lithium forms two compounds with cadmium, LiCd and LiCd 2 . The system was studied by Masing and Tammann, 1 and is shown in Fig. 49. Lithium and cadmium form an uninterrupted series of mixed crystals with a maximum at 50 per cent, of cadmium. The tempera- tures at which crystallisation begins can be taken with exactness in all cases, but similar precision is not obtainable for the completion of crystallisation. As Kuez 2 has 1 Ze.it. anorg. Chem., 67, 183 (1910). 2 Zeit. phys. Chem., 59, 16 (1907). 124 CHEMICAL COMBINATION AMONG METALS. observed, there is at 67-7 per cent, cadmium, in place of an interval of crystallisation a marked arrest, and the alloy crystallises at constant temperature. A small arrest at 356 is noted on the cooling curve of this alloy. The alloy with 67-7 per cent, of cadmium, in addition to showing an arrest, has also a sharp transformation point ; it can be considered to be the compound LiCd 2 - At 50 per cent, of cadmium, FIG. 50. corresponding to the maximum of the curve, another arrest point is observed which probably indicates the compound LiCd. The alloys obtained present a homogeneous appearance ; only the alloy with 92 per cent, of cadmium shows a poly- hedric structure when treated with dilute hydrochloric acid. Alloys containing from 100 to 70 per cent, cadmium are stable in the air ; with increase in the lithium content their oxidisability increases ; thus the alloys with more than 50 HOMOPOLAK INTEEMETALLIC COMPOUNDS. 125 per cent, lithium become red, those with 13 per cent, of cadmium dark brown, and those with still less of this metal black. Lithium-mercury. Lithium forms the following com- pounds with mercury : Li 3 Hg, Li 2 Hg, LiHg, LiHg 2 and LiHg 3 . The equilibrium diagram of the lithium amalgams is shown in Fig. 50 and has been traced by Gr. Zukovski. 1 Kerp, Bottger, Winter and Iggena, 2 studying the amalgams of the alkali and alkaline earth metals had already obtained a compound LiHg 5 . Guntz and Feree 3 admit the existence of this latter compound. Maey, 4 from a study of the varia- tions of specific volume with composition, established the existence of the compounds LiHg 5 , LiHg 3 , LiHg and Li 3 Hg. In the diagram traced by Zukovski the maximum point of the system is at 50 per cent, mercury, corresponding to a compound LiHg. The cooling curve shows other arrests from 2-4 to 24-8 per cent, mercury, corresponding to the com- pound Li 5 Hg. G represents a new compound which decom- poses on melting. The transformation point G indicates a new solid phase of the composition LiHg 2 . (The existence of this compound was confirmed by calorimetric measure- ments.) The point H represents the compound LiHg 3 . In the rest of the curve no other arrest points are noted, which contradicts the contention of other workers as to the occur- rence of a compound LiHg 5 . The compound LiHg separates in needle-shaped crystals. Lithium-tin. Lithium combines with tin forming the following compounds : Li 2 Sn 5 , Li 3 Sn 2 and Li 4 Sn. The diagram has been worked out by Masing and Tam- mann 5 and is shown in Fig. 51. Adding lithium to tin, the 1 Zeit. anorg. Chem., 11, 403 (1911). * Ibid., 25, 16(1900). 3 Bull. Soc. Chim., 15, 834 (1896). 4 Zeit. phys. Chem., 29, 119 (1898). 5 Zeit. anorg. Chem., 67, 183 (1910). 126 CHEMICAL COMBINATION AMONG METALS. melting point of th3 latter is lowered until the eutectic point b is reached, corresponding to about 95 per cent, of tin. The cooling curve of the alloy of this composition shows an arrest at 214, and the structuie of the solid alloy is eutectic. Adding more lithium, the liquidus curve rises along the branch b c. The compound Li 2 Sn 5 first separates at 320 and 72 per cent, of tin in the form of white crystals, sur- 700 ffO* 100 O 10 FIG. 51. rounded by the eutectic, which slowly become yellow when exposed to the air. At 465 and 40 per cent, tin a slackening is shown which corresponds to the compound Li 3 Sn 2 . The reaction between Li 3 Sn 2 and the melt c does not take place completely, for the eutectic horizontal b o extends beyond the point p, which corresponds to the composition of Li 2 Sn 5 . In consequence, alloys containing less than 77 per cent, of tin are composed of three species of crystals ; the crystals HOMOPOLAK INTERMETALLIC COMPOUNDS. 127 Li 3 Sn 2 are surrounded by Li 2 Sn 5 , and among the latter is found the eutectic b, containing tin. The crystals of Li 3 Sn 2 are long, and on exposure to air become first yellow and then brown. From e the curve rises rapidly to the maxi- mum / at 20 per cent, of tin, which probably corresponds to the compound Li 4 Sn. The crystals at this point have a 9oo 800 ZOO 00 SOO 4 00 300 ~ .,- ^o Co & ^-- OLto \-\*Ji cYci- FIG. 52. micaceous structure and are very brittle ; they acquire a dark blue colour on exposure. COMPOUNDS OF SODIUM. Sodium-magnesium. Sodium forms no compound with magnesium. The system was investigated by Mathewson. 1 Sodium-zinc. With zinc, sodium forms the compound 1 Zeit. anorg. Chem., 48, 193 (1906). 128 CHEMICAL COMBINATION AMONG METALS. NaZn n or NaZn 12 . The equilibrium diagram, due to Mathewson (loc. cit.), is shown in Fig. 52. The two metals are only slightly soluble in each other even at high tempera- tures. At 557, zinc dissolves about 6 per cent, of sodium, while sodium scarcely dissolves zinc at all. However, at SO 10 FIG. 53. about this temperature and at a concentration of 8-36 per cent, of sodium a compound is formed of uncertain formula, for while the diagram suggests NaZn n as most probable, analyses of the lower strata of certain melts give NaZn 12 . On melting, the compound gives a liquid richer in zinc together with almost pure sodium. The compound is harder and more brittle than zinc, and HOMOPOLAR INTEEMETALLIC COMPOUNDS. 129 on exposure to air is slowly covered with a white layer of zinc hydroxide. Sodium-cadmium. With cadmium, sodium forms the two compounds NaCd 5 and NaCd 2 . Fig. 53 gives the equilibrium diagram of this system, studied first by Kurnakoff, 1 then by Mathewson, 2 and finally by Kurnakoff and Kusnetzoff. 3 The diagram shown is due to Mathewson, who has estab- lished, both by means of thermal analysis and also directly by removing the liquid in a pipette, the gap in solubility which occurs from 30 to 40 per cent, of cadmium. Two maxima are observed corresponding to the two compounds, for NaCd 5 at 360 and 16-4 per cent, sodium, and for NaCd 2 at 382 and 33'06 per cent, of sodium. According to Kurna- koff, in the place of NaCd 5 , NaCd 6 is formed. Mathewson does not report the occurrence of mixed crystals ; Kurna- koff, on the other hand, states that NaCd 6 can dissolve a certain quantity of NaCd 2 in solid solution. The compound NaCd 2 is shining and brittle. The alloys of the compound NaCd 5 and the eutectic are harder and show a finer structure than the compound NaCd 2 . The alloys of the compound NaCd 5 and the eutectic with 16 per cent, of sodium can be cut with a knife. The compounds are harder than cadmium ; NaCd 5 is as hard as calc spar, and NaCd 2 is a little harder. In moist air both compounds are oxidised, NaCd 2 more rapidly than NaCd 5 . Sodium-mercury. The sodium amalgams are very diverse and include compounds of the following formulae : Na 3 Hg, Na 5 Hg 2 , Na 3 Hg 2 , NaHg, Na 7 Hg 8 , NaHg 2 and NaHg 4 . The system was studied thermally by Schuller, 4 who reports the existence of a compound Na 12 Hg 13 , while E. Vanstone 5 from a study of the specific volumes of the 1 Zeit. anorg. Chem., 23, 439 (1900). 2 Ibid., 50, 180 (1906). 3 Ibid., 52, 173 (1907). 4 Ibid., 40,385(1904). 5 Chem. Neivs, 103, 181, 198, 207 (1911). C.M. 130 CHEMICAL COMBINATION AMONG METALS. system sodium-mercury believes the existence of a com- pound Na : Hg s more probable. Fig. 54 shows the equilibrium diagram for sodium-mercury. The compound NaHg is formed at 50 per cent, sodium and 219, with such strong super-cooling, however, that the corre- 30 40 5-0 6O ^o tio ^ ^o -100 fc *^j "'^ CX^Cyvxvx C-VcX. FIG. 54. sponding arrest points are not found on the horizontal as would be expected. An imperfect equilibrium is established, and hence the arrest points at 123 show a quantity of sodium less than 50 per cent. At 61 per cent, of sodium and 125 the compound Na 3 Hg 2 is probably formed ; in this region, also, super-cooling is observed. Na 12 Hg 13 is shown by a discontinuity at 227 and 48 per cent, of sodium. But, as HOMOPOLAR INTERMETALLIC COMPOUNDS. 131 has already been said, the formula of this compound is uncertain. The formulae attributed to the other compounds may be considered reliable. Mixed crystals are formed among the compounds. Sodium-aluminium. Sodium forms no compounds with 400 Afa. 77 / \ \ too SO AOO A. I/ 7 i FIG. 55. aluminium. The system has been studied by Mathewson l and Smith. 2 Sodium-thallium. This system was studied by Kurnakoff and Pushin, 3 who have established the existence of the com- pound NaTl, which agrees with the monovalency of thallium. The diagram shown in Fig. 55 indicates the existence of two other compounds which separate respectively at 158 and 33 per cent, thallium and 77*9 and 17-2 per cent, thallium ; 1 Zeit. anorg. Chem., 48, 192 (1906). 2 Ibid., 56, 112(1907). 3 Ibid., 30, 87 (1902). 92 132 CHEMICAL COMBINATION AMONG METALS. it is doubtful, however, if the formulae Na 2 Tl and Na 5 Tl attributed to them are reliable. NaTl crystallises in three- rayed arboriform growths, and is formed with considerable development of heat. The melting point of this compound is 305-8, or somewhat higher than that of pure thallium. The compound is harder and more brittle than its compo- nents. Sodium-tin. Sodium forms the following compounds with 700 FIG. 56. tin : Na 4 Sn, Na 2 Sn, Na 4 Sn 3 , NaSn and NaSn 2 . The system has been worked out by Mathewson, 1 and its diagram is shown in Fig. 56. The fusion curve shows two distinct maxima. The compound in equilibrium with the melt at 405 and at a concentration of 80 per cent, sodium has the formula Na 4 Sn. At 405 it decomposes to a melt of com- position indicated by B and the compound Na 2 Sn. The latter compound corresponds to the maximum at 66-9 per cent, sodium and 477, while Na 4 Sn 3 occurs at 57 per cent. 1 Zeit. anorg. Chem. r 46, 94 (1905). HOMOPOLAR INTERMETALLIC COMPOUNDS. 133 of sodium. At 478 the latter decomposes, forming crystals of the compound NaSn, which is found at the maximum point at a concentration of 50 per cent, of sodium. At 483 a polymorphic transformation takes place. Finally NaSn 2 is found at 33-5 per cent, sodium. When warmed it decom- ioo so FIG. 57. poses at 305 into crystals of the compound NaSn and a melt containing about 20 per cent, sodium. Freshly cut, NaSn 2 has a steel blue colour ; Na 4 Sn 3 has a pale blue colour ; the other compounds are similar in colour either to sodium or to tin. The compound Na 4 Sn 3 is fairly hard and brittle ; the other compounds are more brittle. The latent heats of fusion have been calculated approximately from the 134 CHEMICAL COMBINATION AMONG METALS. thermometric arrests by Tammann's method ; the values obtained are as follows : Na 4 Sn = 11 ; Na 2 Sn = 12 ; Na 4 Sn 3 = 11 ; Na 4 Sn 3 (transformation) = 4 ; NaSn 14 ; NaSn (transformation) == 7 ; NaSn 2 = 9 ; NaSn 2 (trans- formation) = 4. Sodium-lead. This system was studied by Kurnakoff, 1 9oo 300* 200 10 *io 1,0 to (, o 70 v o 9o IDO FIG. 58. and later by Mathewson, 2 who have noted the occurrence of the following compounds : Na 4 Pb, Na 2 Pb, NaPb and Na 2 Pb 5 . The curve of fusion is shown in Fig. 57 taken from Mathewson's paper and exhibits four maxima, the first at 386 and 80 per cent, sodium, the second at 405 and 67-2 per cent, sodium, the 1 Zeit. anorg. Chem., 23, 439 (1900). 2 Ibid. f 50, 172 (1906). HOMOPOLAK 1NTEKMETALLIC COMPOUNDS. 135 third at 367 and 49-6 per cent, sodium, and the last at 319 and 28-2 per cent, sodium. Some doubt, however, exists about the compound Na 2 Pb 5 , which might be replaced by the compound NaPb 3 . Mathewson, however, decides for Na 2 Pb 5 , seeing that while its melt shows a sharp Arrest point, that of the second only gives a discontinuity. Two of the compounds, Na 2 Pb and Na 4 Pb, form mixed crystals with each other. The alloys of this series oxidise easily those with high sodium content more easily than those with low content of this metal. As to hardness, NaPb and Na 2 Pb 5 are almost as hard as calc spar, while Na 4 Pb and Na 2 Pb are rather less hard. Na 2 Pb has a light blue colour, while NaPb is light grey. Sodium-antimony. Sodium combines with antimony, forming the two compounds Na 3 Sb and NaSb. Fig. 58, taken from Mathewson, 1 is the fusion diagram. Two maxima are shown, one at 856 and about 75 per cent, sodium, corresponding to the first compound, and the other at 465 and 50 per cent, sodium, corresponding to the second compound. NaSb is almost as hard as gypsum and is of the same colour as antimony, while Na 3 Sb is deep blue and some- what harder. The sodium-antimony alloys ignite spon- taneously in the air. Sodium-bismuth. This system has been investigated by Kurnakoff 2 and Mathewson. 3 The diagram (Fig. 59) shows the existence of the two compounds Na 3 Bi and NaBi. The first is indicated by a distinct maximum at 775 and 75-15 per cent, sodium, the second is represented by a break and is formed at 445 and about 49 per cent, of sodium. Here also, as in the case of NaSb, the two metals develop heat strongly on being melted together. Na 3 Bi has a violet blue colour when freshly cut ; pieces larger than 10 grams inflame spontaneously in the air on slight heating. The hardness 1 Zeit. anorg. Chem., 56, 192 (1906), 2 Ibid., 23, 439 (1900). 3 Ibid., 50, 187 (1906). 136 CHEMICAL COMBINATION AMONG METALS. of the two compounds is almost equal to that of bismuth and of calc spar. POTASSIUM COMPOUNDS. Potassium-zinc. Potassium forms a compound with zinc whose probable formula is KZn n or KZn 12 . The system was boo Zoo 400 FIG. 59. studied by Smith. 1 As Fig. 60 shows, the two metals have a very small reciprocal solubility at 600, since at that tem- perature most melts consist of two liquid strata. At 585 a metastable form of the compound separates, passing into the stable form when the solidification is scarcely complete. Between 405 and 510, from 9 to 40 per cent, potassium, thermal effects are indicated. The horizontal at 510 marks 1 Zeit, anorg. Chem., 56, 113 (1907). HOMOPOLAK INTEKMETALLIC COMPOUNDS. 137 a transformation of the compound KZn 12 . The structure of the alloys is finely granular. They are easily altered in the air. Potassium-cadmium. Potassium combines with cadmium, forming the compounds KCd 12 and KCd 7 . Fig. 61, taken from Smith 1 is the fusion diagram. The mutual solubility roe*' 600 FIG. 60. of the metals is small. At 468 there is a gap of solubility between 17 and 99 per cent, potassium. At 473 and 12-5 per cent, potassium, KCd 7 separates and at about 485 and 7 per cent, potassium, KCd 12 . The formula given to the latter is, however, not perfectly reliable ; KCd n might be substituted for it. These alloys oxidise easily in air. Potassium-mercury. The potassium amalgams, like those Zeit. anorg. Chem.,56, 113 (1907). 138 CHEMICAL COMBINATION AMONG METALS. of sodium, are numerous. Kuinakoff 1 first studied the system and his data were later confirmed by Janecke. 2 The compounds probably formed are KHg, KHg 2 , KHg 3 , K 2 Hg 9 , and KHg 9 , all of which Kurnakoff admits with the exception of K 2 Hg 9 whose formula he believes to be KHg 5 . Fig. 62 gives the fusion diagram. The compound KHg, which melts at 178, is shown by a discontinuity in the curve ; the compound KHg 2 alone shows a maximum, which occurs 40 10 30 iO 60 -- > % ^ FIG. 61. 10 ffO 90 /OO c C d, at 279 and about 65 per cent, of mercury. The other com- pounds are formed in a Kmitel concentration interval. KHg 3 occurs at 204 and 78 per cent, mercury, K 2 Hg 9 at 173 and about 83 per cent, mercury, and KHg 9 at 70 and 90 per cent, of mercury. The homogeneity of the last three, on which some doubt might be cast, is shown by microscopic analysis, for the first crystallises in rods about one centi- 1 Zeit. anorg. Chem., 23, 439 (1900). 2 Zeit. pht/s. Cfam., 58, 245 (1907). HOMOPOLAR INTERMETALLIC COMPOUNDS. 139 metre long, the second in hexagonal plates, and the third in regular, mainly cubic form. Potassium-thallium. As noted by Kurnakoff andPushin, 1 potassium forms with thallium the compounds KT1 and K 2 T1. ioo "boc Zoo c FIG. 62. The equilibrium curve is shown in Fig. 63. The existence of the second compound has been established by analogy with the behaviour of sodium with thallium, since the dis- continuity which is found at 242 and 32-9 per cent, is not very pronounced. At 335, the curve shows a maximum, corresponding to KT1. The melting point of this compound 1 Zeit. anorg. Ckem., 30, 87 (1902). 140 CHEMICAL COMBINATION AMONG METALS. is, as may be seen from the diagram, somewhat higher than that of pure thallium. The compound KT1 crystallises in compact brittle cubes, which react with moist air. It is formed with considerable evolution of heat. Potassium-tin. Potassium forms with tin numerous com- pounds, from which, however, K 4 Sn, corresponding to the saline valency of tin, is lacking, although a corresponding 40CF 50 SCO 200 ACO FIG. 63. sodium compound occurs, Na 4 Sn. The following are the potassium-tin compounds : KSn 4 , KSn 2 , KSn and K 2 Sn. Fig. 64 indicates the diagram of the system which was studied by Smith. 1 The melting points of the alloys are, over a certain range, above the boiling point of pure potassium (757). KSn 4 separates at 600 and 20 per cent, potassium from the melt and from a compound of uncertain formula, possibly KSn 2 , formed at a higher temperature. At above 1 Zeit. anorg. Chcm., 56, 129 (1907). HOMOPOLAR INTERMETALLIC COMPOUNDS. 141 600 and 50 per cent., KSn is formed which, reacting with the melt, gives the compound K 2 Sn (?) with a maximal arrest at 535. Potassium-lead. Smith x has also studied this system in which the compounds KPb 4 , KPb 2 and K 2 Pb are formed as indicated in Fig. 65. In this series the type K 4 Pb, which occurs in the sodium-lead system and corresponds to a saline Qco FIG. 64. valency of lead, is missing. At 568, between 35 and 75 per cent, of potassium, a compound crystallises out from the two liquid strata to which the formula K 2 Pb is given. At 380 a transformation occurs, probably into another form of the compound. At 337 and 33-33 per cent, potassium, KPb 2 is formed, while at 295 there separates from the crystals of this compound, and from the melt, KPb 4 , whose formula is not, however, well established. It is, further, uncertain 1 Loc. cit. 142 CHEMICAL COMBINATION AMONG METALS. whether the eutectic horizontal at 376 consists of two hori- zontals, of which one corresponds to a compound X between K 2 Pb and KPb 2 . Smith did not observe an eutectic point between K and K 2 Pb ; yet the alloys between 65 and 98 per cent, potassium show a point of arrest 4 to 6 lower than the melting point of potassium. Potassium-antimony. Although we have no very exten- <*oc FIG. 65. sive knowledge of the compounds between potassium and antimony, it is well established that the two elements do combine. The system potassium-antimony has recently been studied by Parravano. 1 The diagram is shown in Fig. 66. It will be seen that the curve is quite simple and shows the existence of the two compounds K 3 Sb and KSb, having melting points at 812 and 605 respectively. The formation of these compounds is attended with a consider- i Gazz. Chim. Ital, I., 485 (1915). HOMOPOLAR INTEEMETALLIC COMPOUNDS. 148 able evolution of heat. The compound K 3 Sb, in which the trivalency of antimony is shown, has a yellowish-green colour and alters rapidly in the air ; the compound KSb crystallises in long slender prisms of a colour similar to antimony and is less rapidly attacked by air than the former compound. FIG. 66. Potassium-bismuth. Potassium combines with bismuth, according to Smith x to form a rather numerous series of com- pounds. They comprise KBi 2 , K 9 Bi 7 (?), K 3 Bi 2 and K 3 Bi. Fig. 67 is the diagram for this system. K 3 Bi gives a maximum at 75 per cent, of potassium and 671. At 286, arrest points occur between 60 and 83 per cent, of potassium, 1 Zeit. anorg. Chem., 56, 125 (1907). 144 CHEMICAL COMBINATION AMONG METALS. with a maximum at 75 per cent. These probably indicate a transformation of a K 3 Bi into ft K 3 Bi a companied by an increase in volume. Further, at 420 with a maximal arrest at 60 per cent, potassium, a compound, probably K 3 Bi 2 , separates. From these crystals and the melt at 54 per cent., another compound separates at 373, also of uncertain formula, K 9 Bi 7 . Another maximum occurs on the curve at &CO ZOO FIG. 67. about 550 and 33-33 per cent, of potassium, due to a com- pound whose formula can safely be taken as KBi 2 . The two metals develop a considerable amount of heat on being melted together, due, it is maintained, to the formation of K 3 Bi. RUBIDIUM COMPOUNDS. Rubidium-mercury. Little is at present known of the compounds of rubidium with other metals. The system HOMOPOLAR INTERMETALLIC COMPOUNDS. 145 rubidium-mercury has been studied by Kurnakoff and Zukovsky, 1 but only over a limited range of concentration. The existence of a compound RbHg 6 was noted, which melted at 136-5 with decomposition. At 70-2 there is a transformation point, and here a compound richer in mer- cury begins to separate, to which, by analogy with the corre- sponding caesium compound, the formula RbHg 10 may be given. CESIUM COMPOUNDS. Ccesium-mercury. Our information on the metallic com- pounds of caesium is also very scanty. The compounds with mercury are well known. The system cesium-mercury studied by Kurnakoff and Zukovsky (loc. cit.) is shown in Fig. 68. The following compounds are formed : CsHg 10 , CsHg,, CsHg 4 , CsHg 2 , CsHg and Cs 2 Hg. Three maxima are seen in the diagram representing the three compounds CsHg 2 , CsHg 4 and CsHg 6 respectively. They occur at 208-2 and 67 per cent, mercury, 163-5 and 80 per cent, mercury and 157-7 and 86 percent, mercury, and melt without decomposition. At 188 weak arrests are noted between 62-4 and 65-8 per cent, mercury, which probably imply a polymorphic transformation of the com- pound CsHg 2 . The formation of solid solutions is observed on the curve. At 13-1 and 91 per cent, mercury, the com- pound CsHg 6 is transformed into CsHg 10 ; this formula is, however, not well established. Other arrest points occur at 50 per cent, and at 37 per cent, of mercury (CsHg and CsHg 2 ), but as to these compounds there is considerable uncertainty. COPPER COMPOUNDS. Copper-beryllium. Little is known of the beryllium alloys. The alloys with copper have only been studied by Lebeau 2 and, more recently, by G. Oesterheld, 3 who studied the system 1 Zeit. anorg. Che.m., 52, 416 (1907). 2 C. P., 125, 1172 (1897). Bull. Soc. Chem., (3), 19, 64 (1898). Ann. Chim. Phys., (7), 16.498(1899). 3 Zeit. anorg. Chem., 97, 6 (1916). C.M. 10 146 CHEMICAL COMBINATION AMONG METALS. over a limited range. The compound CuBe 3 occurs. The melting point of copper is lowered until a concentration of 10 per cent, beryllium is reached, and the lowering is accom- panied by the formation of solid solutions. Beyond this So FIG. 68. limit for a short interval the solidus and liquidus curves have an unusual form and subsequently pass through a minimum and a point of inflection. At 575 there is an eutectic point for 31 per cent, beryllium, and thence the curve rises to a maximum corresponding to the compound mentioned. The copper-beryllium alloys can dissolve in nitric acid. HOMOPOLAK INTERMETALLIC COMPOUNDS. 147 Copper-Magnesium. Copper forms with magnesium two compounds, namely, CuMg 2 and Cu 2 Mg. The system has HAOO Qoo 10 10 <30 10 fO 60 70 80 90 100 been studied by several workers, including Boudouard, 1 Urasoff 2 and Sahmen. 3 1 C. R., 135, 794 (1902). 2 Chcm.Centr., 1908, I., 1038. 3 Zeit. anonj. Ghent., 57, 20 (1908). 10-2 148 CHEMICAL COMBINATION AMONG METALS. The data obtained by the two latter are quite concordant. As Fig. 69 shows, there are two maxima, one for Cu 2 Mg at 33-3 per cent, magnesium and 797, and the other for CuMg 2 at 66-7 per cent, magnesium and 570. Mixed crystals are lacking in this system. Boudouard also reports a compound CuMg. Sahmen states that the eutectic point between Cu 2 Mg and CuMg is found between 55 and 57 per cent, of ^ C M FIG. 70. magnesium, while according to Urasoff it is found at 58-5 per cent, of magnesium. Both compounds are very brittle and have the colour of pure magnesium. Only those alloys rich in copper which contain a large quantity of that metal in the free state exhibit a red colour. Copper-zinc. Copper combines with zinc, forming the two compounds Cu 2 Zn 3 and CuZn. The equilibrium diagram of these alloys has been investigated by Eoberts-Austen, 1 1 Proc. Inst. Mech. Eng., 1897, p. 31. HOMOPOLAK INTERMETALLIC COMPOUNDS. 149 Shepherd, 1 Tafel, 2 Carpenter and Edwards, 3 and lastly by Parravano. 4 As is seen from the diagram, which epitomises our knowledge of the brass alloys (see Fig. 70), Cu 2 Zn 3 separates at 833 and 60 per cent, zinc, and CuZn at 1005 and about 50 per cent. It was noted that the region of existence of this latter compound became more restricted with fall of temperature. Shepherd maintains that the curve does not necessitate . the unconditional admission that compounds are formed. Bornemann 5 refutes this for various reasons. Above all be- cause the field of existence of homogeneous crystals enlarges with fall of temperature towards the Zn axis. The solubility of the substance richer in zinc in the solid solution indicated by increases with fall of temperature. The solution process must be exothermic and an exothermic compound must be formed. To support his contentions, Bornemann cites the work of Backer 6 and Herschkovitch 7 on the heats of formation of the copper-zinc alloys ; for all concentrations the values obtained were positive. Further, allowing the existence of the compound, and this Tafel holds to be certain, as well defined and practically undissociated in the pure state, the micrographic study and the diagram should be in perfect agreement. To resolve the question, Bornemann examined the methods used to determine the constitution of compounds by observ- ing the relation between concentration and (a) electrical conductivity and temperature coefficient of electrical resis- tance, (b) specific gravity and specific volume, (c) electro- lytic potential, and (d) chemical and electrochemical reactivity. In all cases the presence of compounds is dis- tinguished from the presence of mixed crystals, since the 1 J. Phys. Chem., 3, 421 (1904). 2 Metatturgie, 5, 349, 375 (1908). 3 Int. Zelt. Metall, 2, 129 (1912). 4 Gazz. Chim. Ital, 44, II., 475 (1914). 5 Die binaren Metalkgierungen, p. 19, Halle, 1909. 6 Kelt, phys. Chem., 38, f>30 (1901). 7 Ibid., 27, 164(1898). 150 CHEMICAL COMBINATION AMONG METALS. formation of the former produces much more marked changes than the latter. (a) Le Chatelier 1 and Guertler 2 with regard to the first method have laid down the following rules : (1) the curve of concentration-electrical conductivity is practically a straight line in all cases where an alloy consists of a hetero- geneous mixture of two constituents ; (2) wherever mixed crystals occur the curve shows a marked lowering. The measurements carried out by Matthiessen, 3 Haas, 4 and Weber, 5 show that while the existence of the compound CuZn is probable, other compounds do not exist between this and pure copper. The compound Cu 2 Zn 3 may exist although apparently it shows no maximal conductivity. (b) Specific volume should change in a linear manner with concentration in the case of heterogeneous mixtures. In the case of mixed crystals a change from one series to another should be marked by a discontinuity in the curve. If at the point of discontinuity mixed crystals do not exist the new phase is to be taken as a compound. The diagram showing Maey's 6 observations shows no marked discontinuities with the exception of one at 40 per cent, copper, corresponding to the compound Cu 2 Zn 3 . (c) With regard to electrolytic potential two principles must be noted : (1) where mixed crystals occur the potential should fall in a continuous curve ; (2) in heterogeneous systems of saturated mixed crystals the potential should remain constant. In the case where compounds are present without the formation of mixed crystals, each compound must be reckoned as a single new substance, and the diagram is divided up accordingly into corresponding parts. A compound should be marked by a decided fall of potential. In practice the distinction between mixed crystals and com- Rev. Gener. des Sciences, 6, 531 (1895). Contrib. a V Etude des Alii ages, Paris, 1901, p. 446. Rep. Brit. Ass., 1863, 127. Wied. Ann., 52, 673 (1894). Ibid., 68, 705 (1899). Zeit. phrjs. Chem., 38, 291 and 299 (1901). HOMOPOLAK INTEEMETALLiO COMPOUNDS. 151 pounds is not so simple, since most usually the fall of potential extends over a certain range of concentration and is graphically represented by a curve. Consequently it is necessary to examine also the fusion diagram. Pushin's 1 curve, though showing some errors in measurement, indi- cates a sharp fall of potential corresponding to 60 per cent, zinc, i.e., to the compound Cu 2 Zn 3 . Another, though small, fall takes place at about 50 per cent, and may correspond to strongly dissociated CuZn. The other falls of potential given in Pushin's curve are very probably not due to com- pounds. Sackur, 2 by chemical means, has demonstrated two considerable falls of potential. (d) Sackur, 3 and Lincoln, Klein and Howe 4 have studied the chemical and electro-chemical reactivity of the copper- zinc alloys. The former has determined the solubility of alloys in dilute acids in the presence of air. He observed that there were distinct changes for the crystals ft and y, and these may consequently be compounds. The other authors studied electrolytic reactivity in neutral saline solutions, obtaining hydrates or basic salts which were mechanically removed from the anodes. From their curve it appears that at about 40 per cent, the quantity of copper oxidised is reduced almost to nothing, which is another proof of the existence of the compound Cu 2 Zn 3 . Summarising the evidence it may be said that the com- pound Cu 2 Zn 3 certainly exists and can melt without decom- position. A compound CuSn also probably exists which is strongly dissociated at high temperatures and to a marked degree at lower temperatures, giving rise to copper and the compound Cu 2 Zn 3 , accompanied by the formation of mixed crystals with the components. Copper-calcium. Copper is said to combine with calcium, forming the compounds Cu 4 Ca and CuCa 4 . The existence 1 Ze.it. anorg. Chem., 56, 28 (1907). 2 Ber., 38, 2186 (1905). 3 Ibid., 38, 2190 (1905). 4 J. Phys. Chem., ii, 501 (1907). 152 CHEMICAL COMBINATION AMONG METALS. of the latter compound is somewhat doubtful, as N. Baar l has shown in his study of the system. The diagram is given in Fig. 71. From the eutectic point the curve rises up to 13*7 per cent, by weight of calcium, a maximum correspond- ing to the compound Cu 4 Ca, which melts at 933. From this point to the next, eutectic crystals of Cu 4 Ca separate. From melts containing more than 38 per cent, of calcium, a series of mixed crystals separate whose last member contains lioo FIG. 71. 56 per cent. On the cooling curve of alloys between 23-5 and 90 per cent, calcium, points of arrest are noted at 480 with a maximum for 70 per cent, of the metal. This alloy, which may be a mixed crystal or the compound CuCa 4 , has a homo- geneous appearance and passes without change of composi- tion from an a to a ft form. Alloys containing about -8 per cent, calcium are not acted upon by water and are copper-coloured, those with 1-25 per cent, of calcium are a little harder than the pure metal. The others are more brittle, decompose water, and decompose in 1 Zeit. anorg. Chem., 7C, 532 (1911). HOMOPOLAR INTERMETALLIC COMPOUNDS. 153 air to a powder, which up to 35 per cent, calcium has a brassy colour. From 35 per cent, upwards of calcium the alloys are silvery white when freshly prepared, but decompose when exposed to the air. Copper-cadmium. Copper forms two compounds with 4000 9oo &oo a 700 cadmium, namely : Cu 2 Cd and Cu 2 Cd 3 . The diagram (Fig. 72) has been drawn from thermal data by Sahmen. 1 From melts containing up to 42 per cent, cadmium, pure copper separates out at temperatures between 1084 and 552. At 552, copper, reacting with the melt, forms a compound crystallising in long needles. The eutectic horizontal is prolonged to alloys richer in copper, and ends at 33 per cent. 1 Zeit. anorg. Chem,, 49, 301 (1900). 154 CHEMICAL COMBINATION AMONG METALS. cadmium, rendering almost certain the existence of the com- pound Cu 2 Cd. This compound was obtained in crystals by Mylius and Fromm 1 by precipitating a 1 per cent, solution of copper sulphate with cadmium. A slightly denned maximum is observed on the fusion curve, corresponding to the second compound. On both sides of the maximum mixed crystals are formed, on the one side between Cu 2 Cd 3 and Cu 2 Cd and, on the other, between Cu 2 Cd 3 and cadmium. Sahmen believes this formula to be more probably correct than Cu 5 Cd 7 . The alloys richer in cadmium are soft ; with decreasing content of this metal they become more hard and brittle, but the brittleness decreases on further increase of the copper content. Alloys containing up to 40 per cent, copper are grey, but with increase of copper they become more reddish until the copper colour is reached. Copper-mercury. The copper amalgams have not as yet been studied thermally. Chemical and physico-chemical researches on these alloys are, however, numerous. 2 The researches of J. Joule 3 and E. De Souza 4 on the capacity for chemical combination of the two metals should be recorded. The former obtained from liquid amalgams well- defined crystals corresponding to the formula HgCu. The existence of Hg 2 Cu 3 is doubtful, a solution of copper in HgCu having probably been mistaken for it. According to De Souza the compounds HgCu 14 and HgCu 10 are formed. Copper amalgams on account of their plasticity are frequently used for technical purposes. Copper-aluminium. Copper forms with aluminium the three compounds Cu 3 Al, CuAl, and CuAl 2 . The copper- aluminium alloys have been frequently studied by various 1 Ber., 27, L, 630 (1894). 2 Regnault, C. R., 52, 533 (1861). Becquerel, ibid., 75, 1729 (1872). Merz and Weith, Ber., 14, 1438 (1881). Battelli, Rend. Ace. Lincei, (4), 3, II., 37 ; 4, 206 (1887). Bachmetjeff, Jahrber., 109 (1893). Gouy, Jour, de Phys., (3), 4, 320 (1895). Humphrey, /. C. 8., 69, 343 (1896). Coehn, Zeit. phys. Chem., 38, 609 (1901). Haber, ibid., 41, 399 (1902). 3 ./. 4(J FIG. 80. ceding investigations of Gautier, 2 Guillet, 3 and Heycock and Neville 4 may also be mentioned. The diagram shows that the two compounds mentioned above are formed. The cooling curves of the compounds are similar in character to those for other simple substances, showing only single arrest 1 Zeit. anorg. Chem., 46, 49 (1905). 2 Butt. Soc. d'Encour., 1312 (1896). 3 Genie Civil, 1902. 4 Phil Trans., A, 69 (1897). 166 CHEMICAL COMBINATION AMONG METALS. points. The interval of eutectic crystallisation at 567 is for the first compound reduced to zero at the point {3, while for the second compound at 75 per cent, silver and 770, the interval is greater than for all other melts. Transformation points are shown at 718 and 610 respectively. These alloys consist in section of a single species of crystals. 1000 900 100 10D , O^ FIG. 81, The tw r o alloys form mixed crystals between 610 and 718. It is uncertain whether the transformation of mixed crystals takes place with or without separation of their components, AlAg 2 and AlAg 3 . On the curve above 718 ft crystals of AlAg 2 are in equilibrium w r ith the melt, while below 718, y crystals are similarly in equilibrium. The presence of this branch indicating the equilibria of the y crystals cannot, HOMOPOLAB INTEEMETALLIC COMPOUNDS. 167 however, be demonstrated directly. A horizontal indicates the transformation of <3 into a crystals. Microscopic examination confirms the foregoing deduc- tions, since it reveals the presence of the five groups of alloys which are presumed to exist from thermal considerations. Silver-tin. Silver forms the compound Ag 3 Sn with tin as may be gathered from Fig. 81, which reproduces the diagram constructed by Petrenko 1 from thermal data. Heycock and Neville 2 had previously studied the system. There are two branches on Petrenko 's diagram, the one straight and the other sinuous; at 480, 232 and 220 respectively, horizontals occur. At 480 and 27 per cent, of tin, the compound Ag 3 Sn separates. Microscopic examination shows that the alloy of this composition consists of polyhedra. The compound is dimorphic. A maximum interval of transformation occurs at 232 at a concentration corresponding to the compound. Silver-antimony. Silver combines with antimony to form the compound Ag 3 Sb, as appears from the investigations of Gautier, 3 Heycock and Neville, 4 and, finally, of Petrenko. 5 Maey 6 from volumetric observations argues that silver and antimony form a single compound Ag 3 Sb. Pushin, 7 how- ever, from his measurements of electro-motive force, argues that there are two compounds, Ag 2 Sb and Ag 3 Sb. Petrenko's diagram, shown in Fig. 82, demonstrates that by addition of antimony the melting point of silver is lowered sharply. Mixed crystals containing antimony separate, which are recognisable by microscopic examina- tion. At 560 and 27-07 per cent, of antimony, there is a sharp arrest point corresponding to the compound Ag 3 Sb. From 27-07 to 45 per cent, antimony the compound separates and at 485, the eutectic alloy. Microscopic Zeit. anorg. Chem., 53, 200 (1907). Phil Trans., 189, A, 140 (1897) Bull Soc. d'Encour., 1896, pp. 1309, 1310. Phil. Trans., 189, A, 25 (1897). Zeit. anorg. Chem., 50, 139 (1906). Zeit. phys. Chem., 50 (1905). Journ. Buss. phys. Chem, Soc., May, 1905. 168 CHEMICAL COMBINATION AMONG METALS. examination confirms these data. The compound appears in polygonal crystals separated by fine lines. Silver-manganese. It is not certain whether silver forms compounds with manganese. The system was studied by thermal and miscroscopic methods by Hindrichs * and Arrivaut 2 ; the latter also used chemical methods and determined electrolytic potentials. Hindrichs does not 400 300 FIG. 82. report the formation of compounds, but states that two strata exist with limited mutual solubility. Arrivaut, while admitting the limited miscibility of the two metals, maintains that at 978 and at 20 per cent, of man- ganese, the compound Ag 2 Mn is formed. The compound forms a continuous series of mixed crystals with silver. Guertler, 3 however, is of opinion that the supposed com- pound is, rather, a saturated solution of manganese in silver. 1 Zeit. anorg. Chem., 59, 437 (1908). 2 Ibid., 83, 193 (1913). 3 Melallographie, Berlin, 1912, Vol. I., Part I., p. 98. HOMOPOLAR INTEEMETALLIC COMPOUNDS. 169 By analogy with gold, the existence of a combination between silver and manganese is probable. Silver-platinum. There is some doubt as to the occur- rence of chemical combination between silver and platinum. According to Doerinckel's 1 investigations, the compound Ag 2 Pt is formed between the two metals. Thompson and 1800' 1700 1600' 1500 1400 1000 900 i oo Fm. 83. Miller 2 stated that such a compound was formed and noted that platinum, when alloyed with silver, was soluble in nitric acid. Fig. 83 is Doerinckel's diagram. It cannot be drawn above 80 per cent, platinum on account of the high melting point of such alloys. At 1184 and 20 per cent, platinum there is a distinct discontinuity. Up to 35 per cent, there is a continuous series of mixed crystals, the last member of 1 Zeit. anorg. Chem., 54, 338 (1907). 2 J. Amer. C. S., 28, 1115 (190fi). 170 CHEMICAL COMBINATION AMONG METALS. which contains 48 per cent, by weight of platinum. This corresponds fairly nearly to the formula Ag 2 Pt, which would require 47-5 per cent, of platinum. It is doubtful, however, whether this is a compound or mixed crystal. The alloys from 10 to 30 per cent, are a little harder than their components. From 40 per cent, of platinum the hard- ness increases slowly and at 70 per cent, exceeds that of calc spar. COMPOUNDS OF GOLD. Gold-magnesium. Magnesium forms four compounds with (0 20 30 40 Sa 60 10 90 9O IOO FIG. 84. gold, in none of which saline valencies are shown. They are AuMg, AuMg 2 , AuMg 3 and Au 2 Mg 5 . The system has been studied by Vogel, 1 by Urasoff 2 and by these two authors in collaboration. 3 The results obtained are in general agree- ment (see Fig. 84). Vogel admits the first three compounds, 1 Zeit. anorg. Chem., 63, 169 (1909). 2 Ibid,, 64,375(1909). 3 Ibid., 67, 442 (1910). HOMOPOLAB INTEBMETALLIC COMPOUNDS. 171 while Urasoff adds Au 2 Mg 5 which would appear to separate at 796 and 72 per cent, of magnesium. For the first three compounds, whose concentrations are respectively 50, 67 and 75 per cent, of magnesium, Vogel gives the melting points as 1160, 796 and 830, while Urasoff gives 11 50, 788 and 818. Between 30 and 34 per cent, of magnesium Vogel has observed that although the alloys at 830 are homogeneous, when cooled slowly to 818, they separate out crystals, which would show a diminution in the solubility of the alloys from to 30 per cent, in the foregoing alloys. At 720 and 72 per cent, of magnesium, Urasoff observed a thermal effect due to a transformation of the compound Au 2 Mg 5 . Gold forms solid solutions with magnesium from 17 to 27 per cent, magnesium ; magnesium in the crystalline state does not dissolve notable quantities of gold. Gold-zinc. Gold and zinc form the following compounds : AuZn, Au 3 Zn 5 and AuZn 8 . The system was studied by Vogel x and the diagram is shown in Fig. 85. At 50 per cent, zinc and 744 there is a maximum corresponding to the com- pound AuZn which, by reaction with the melt, gives mixed crystals with gold and zinc. Between 50 and 63 per cent, of zinc, mixed crystals separate with a higher content of zinc than is demanded by the formula AuZn, and at about 650, a homogeneous substance crystallises which is the compound of the formula Au 3 Zn 5 . At 486 saturated mixed crystals y are changed into a third compound. The maximal trans- formation is at 88 per cent, of zinc corresponding to the formula AuZn 8 . Here a ne\v series of mixed crystals occurs. Alloys rich in gold have the same hardness as that metal, are less tenacious, and not at all brittle. At above 31 percent, of zinc the alloys show considerable hardness and brittleness. After 61 per cent, they become gradually less hard and brittle. Gold-cadmium. Gold forms two compounds with cad- 1 Zeit. anorg. Chem., 48, 319 (1906). 172 CHEMICAL COMBINATION AMONG METALS. mium, Au 4 Cd 3 and AuCd 3 as shown by Vogel, 1 who has constructed the diagram shown in Fig. 86 from thermal data. The crystallisation of the two compounds is marked by two distinct discontinuities, one at 43 per cent, cadmium and 623, and the other at 75 per cent, cadmium and 493. From to 28 per cent, cadmium, a series of mixed crystals is formed. Au 4 Cd 3 forms with the saturated mixed crystal ft, occurring at 51 per cent, cadmium, a series of mixed crystals noo 8 1000 1QO FIG. 85. rich in cadmium, while AuCd 3 crystallises in eutectic alloy from melts rich in cadmium. The compound AuCd, whose formation was reported by Heycock and Neville, 2 might be, according to Vogel, a mixed crystal containing Au 4 Cd 3 of the ft series, richer in cadmium, and only having the composition corresponding to AuCd fortuitously. The more recent studies on the gold-cadmium alloys made by Saldau 3 have regard to thermal phenomena, hardness, the special properties of 1 Zdt. anorg. Chem., 48, 333 (1906). 2 J. C. S. ,61, 888(1892). 3 Int. Zeit. f. Metalkgr., VIL, 3 (1914). HOMOPOLAR INTERMETALLIC COMPOUNDS. 173 eutectic alloys, electrolytic potential and microscopic struc- ture. Saldau, from his investigations, maintains that these alloys include the compounds AuCd and AuCd 3 . He derives the former not only from the fusion diagram, but also from the diagrams for hardness and electrical conductivity. The first compound separates at 50 per cent, and about 625 and 1100 1000" the second at 75 per cent, of cadmium and 490. According to Saldau the compound stated by Vogel to be Au 4 Cd 3 has no real existence, for this concentration simply represents the limit of saturation of the compound AuCd with gold in mixed crystals. The two compounds form mixed crystals with the components, AuCd between 46 and 59 per cent, cadmium, and AuCd 3 between 74 and 79 per cent, cadmium. Two other series of mixed crystals exist, one between gold 174 CHEMICAL COMBINATION AMONG METALS. and cadmium up to 35 per cent, of cadmium, and the other between cadmium and gold up to 2 per cent. gold. The alloys show maximum hardness at 18 to 30 per cent, of cadmium and 51 to 63 per cent, cadmium ; the maximum brittleness occurs at 51 to 63 per cent, cadmium. Gold-mercury. Gold amalgams occur naturally as solids 1200 1100 1000 _ 300 FIG. 87. or semi-solids according to their mercury content ; the solid amalgams are well crystallised. The two metals combine chemically, but the system has not yet been studied thermally. Bottger 1 prepared arti- ficial amalgams by direct union of the two elements. Croockewit 2 isolated a compound of the formula AuHg 2 . 1 Jour. pr. Chem., 3, 278 (1834). 2 Ibid., 45, 87 (1847). HOMOPOLAE INTEKMETALLIC COMPOUNDS. 175 The numerous investigations l on these amalgams lead us to suppose the existence of other combinations richer in gold. The existence of the compounds Au 2 Hg and Au 4 Hg seems probable. The gold amalgams have always been of great importance in the extraction of gold from its minerals. Gold- aluminium. The system gold-aluminium was studied by Heycock and Neville 2 thermally and micrographically. The two metals form the compounds Au 4 Al, Au 5 Al 2 , Au 2 Al, AuAl, AuAl 2 . The curve shows two distinct maxima, one at about 1060 and 33-3 per cent, gold and the other at about 625 and 66 per cent, of gold, corresponding to the com- pounds AuAl 2 and Au 2 Al respectively. There are three dis- continuities at which the other compounds separate. The first occurs at about 56 per cent, of gold and 625 ; probably it corresponds to the compound AuAl. Heycock and Neville noticed, however, that there is not a perfect equili- brium between AuAl 2 and the eutectic at 569, for on microscopical examination they detected the presence of AuAl 2 characterised by its purple colour. The compound AuAl, indeed, does not separate in a pure state. The eutectic horizontal at 569, which should only extend to 50 per cent, of gold, reaches 40 per cent., which is explained by the fact that at that point the equilibrium is complete. The other two compounds Au 5 Al 2 and Au 4 Al separate respec- tively at 575 and 72 per cent, of gold and 545 and 78-5 per cent. gold. Gold-tin. The following compounds are formed between gold and tin : AuSn, AuSn 2 and AuSn 4 . They have been recognised by Vogel 3 by means of thermal analysis. Fig. 88 represents the equilibrium diagram. Preceding Vogel's work, the system had been studied from the point of view of 1 Henry, Phil Mag., (4), 9, 458 (] 855) ; Knafel, Dingl. Poly. Journ., 168, 282 (1863) ; Rammelsberg, Pogg. Ann., 120, 54 (1863) ; De Souza, Her., 9, 1050 (1876) ; Chester Ann. Jour. Sci., (3), 16, 29 (1878) ; Kasauzeff, Bull. Soc. Chim., (2), 30, 20 (1878) ; Merz and Weith, Ber., 14, 1438 (1881) ; Wilm., Zeit. anorg. Chem., 4, 325 (1893) ; Gouy, Jour, de Phys., (3), 4, 320 (1895). 2 Phil. Trans., A, 194, 201 (1900). 3 Zeit. anorg. Chem., 43, 60 (1905). 176 CHEMICAL COMBINATION AMONG METALS. electrical conductivity by Matthiessen, 1 while Maey 2 and Heycock and Neville 3 studied the specific volume, and Lawrie 4 the electrolytic potential. The fusion curve falls from the melting point of pure gold to an eutectic point, and 1000 _ 900 200 100' FIG. 88. subsequently rises to a distinct maximum. In the rest of the curve two discontinuities occur which are met by thermal horizontals from 10 to 33 per cent, and from 30 to 50 per cent, respectively. From melts rich in gold, mixed crystals separate containing up to 8 per cent, of tin. The 1 Pogg. Ann., 110, 190 (1860) ; Phil Trans., 150, 161 (1860). 2 Zeit. phys. Chem., 38, 292 (1901). 3 ,/. C. S. t 59,936(1891). Phil Mag., (5), 33, 94 (1892). HOMOPOLAK IKTTEEMETALLIC COMPOUNDS. 177 compound corresponding to the maximum has the formula AuSn and contains 37-63 per cent, by weight of tin ; it melts at 418. At the temperature of the eutectic horizontal, crystals of AuSn react with the melt and pass into a com- pound richer in tin of the formula AuSn 2 , containing 54-68 1100 1000 500 400 300 200 100 l_ per cent, by weight of tin and melting at 308. This com- pound passes in turn into a third compound, AuSn 4 , which melts at 252. All these alloys are very brittle and acid resistant, particularly the alloy AuSn, which is as brittle as glass. Gold-lead. Lead and gold form the compounds Au 2 Pb 12 178 CHEMICAL COMBINATION AMONG METALS. and AuPb 2 . Vogel 1 has made a thermal study of this system. Maey 2 concluded from a study of specific volumes that the compound Au 2 Pb 3 should also occur. The diagram is reproduced in Fig. 89. Two very distinct discontinuities are seen ; the first is at 418 and corresponds 1100 1000 900 800 700 600 600" 400 300' 200 Au Sb 2 St Au 5b 5b. 10 10 1.0 50 FlG. 90. 90 -100 to the compound Au 2 Pb with a maximal arrest for 35 per cent. lead. At 211 and the same concentration, a poly- morphic transformation takes place. The second discon- tinuity is at 254 and the corresponding thermal arrest, with a maximum at 67-8 per cent, of lead, indicates the compound AuPb 2 . This compound also undergoes a polymorphic transformation at 211. The compound Au 2 Pb crystallises 1 Zeit. anorg. Chem., 45, 11 (1905). 2 Zeit. phy*. Chem., 38, 292 (1901). HOMOPOLAE INTERMETALLIC COMPOUNDS. 179 in large white crystals, AuPb 2 in long needles. The two compounds form mixed crystals with each other and with the components. Maey's supposed compound Au 2 Pb 3 is explained by Vogel, who argues that the specific volume method can be used for alloys with two structural elements, but not for alloys such as those of this system between 10 to 72 per cent, lead having three such structural elements. Both compounds are brittle Au 2 Pb more so than AuPb 2 . Gold-antimony. These metals according to Vogel 1 form the compound AuSb 2 . It is seen from the diagram (Fig. 90) that each metal lowers the melting point of the other. Primary separation of gold and antimony occurs at 34 per cent, antimony and 73 100 per cent, antimony respec- tively. Melts containing 35 per cent, of gold separate the compound AuSb 2 at 460. The primary separation of this compound continues till 34 per cent, of antimony is reached, when eutectic solidification occurs at a temperature of 360. The compound AuSb 2 is harder than its constituents ; it is very brittle and more resistant than antimony to the action of acids. In section it has a shining shell-like crystalline appearance. Gold-manganese. Manganese combines with gold to form the compound AuMn. This system was studied by Parra- vano and Perret. 2 The capacity of gold for combination with manganese differentiates it from the first member of its group, namely, copper, which does not combine with man- ganese, but forms a continuous series of mixed crystals. In the case of silver, as mentioned above, its capacity for com- bination with manganese is not yet well established. In the fusion diagram (Fig. 91) described by Parravano and Perret, there is in addition to the formation of the compound AuMn, a lack of miscibility between 50 and 57-5 per cent, by weight of manganese. Further, transformations have 1 Zeit. anorg. Chfm., 50, 151 (1906). 2 Gazz. Chim. ltd., 45, I., 293 (1915). 122 180 CHEMICAL COMBINATION AMONG METALS. been observed in the solid state which, however, are not yet w r ell explained. Manganese lowers the melting point of gold to 990 for a concentration of 10-5 per cent, man- ganese. In this interval mixed crystals separate. There is a maximum at 1225 corresponding to the compound (21-8 per cent, by weight manganese) ; thence the curve falls to 1080 and 46 per cent, by weight of manganese. Prom this point the curve rises to the melting point of manganese, but between 50 and 57 per cent., of manganese a gap in as* /ill 199 jo 2o > % in peso. liquid miscibility occurs. The formation of the compound has been confirmed by measurements of the' hardness of alloys ; the hardness curve of alloys up to 35 per cent, manganese is similar to the type represented in Fig. 34. Compounds of Metals of Group II. with Metals of other Groups. COMPOUNDS OF BERYLLIUM. Beryllium-iron. According to G. Oersheld, 1 the two metals combine chemically, and in all probability the compound 1 Zeit. anorg. Chem., 97, 6 (1916). HOMOPOLAE INTEEMETALLIC COMPOUNDS. 181 FeBe 2 is produced. The system has been studied up to 21 per cent, of beryllium. At 1155 and a concentration of 38-4 per cent, (atomic) of beryllium, there is an eutectic point ; up to 29 per cent., solid solutions are formed. By addition of beryllium the transformation temperature of iron ft into iron a is lowered till it becomes constant. The compound is blackened by alkalies. COMPOUNDS OF MAGNESIUM. Magnesium- aluminium. - - Aluminium and magnesium form the compound Mg 4 Al 3 . The system has been studied by Boudouard 1 and Grube. 2 The diagram (Fig. 92), taken 700' 50 'S 500 50 500 50 AI 3 Mg x>^ FIG. 92. 10 ^ 30 OLtom.1 <=^6a -100 from Grube, is of interest because, although the compound has a range of existence of about 30 per cent., it displays no decided maximum ; this system approaches, therefore, to the type c described on p. 11. The maximum occurs at 1 Butt. Soc. Chim.. (3), 27, 5, 45 (1902). 2 Zeit. anorg. Chem., 45, 225 (1905). 182 CHEMICAL COMBINATION AMONG METALS. 462-7 and 54-9 per cent, of magnesium. All the alloys from 35 to 55 per cent, of magnesium crystallise as chemically homogeneous substances ; they are in fact mixtures of Mg 4 Al 3 and excess of aluminium. The compound Mg 4 Al 3 has a silvery white colour and is 650- 600 go toe FIG. 93, very brittle. The brittleness decreases from the composition of the compound to the two pure components. Alloys between 35 and 50 per cent, magnesium have, after polishing, a very bright mirror-like surface. The alloys of magnesium and aluminium are well known as " magnalium." Magnesium-thallium. Magnesium forms the following HOMOPOLAB INTEBMETALLIC COMPOUNDS. 183 compounds with thallium : Mg 3 Tl 2 , Mg 2 Tl and Mg 8 Tl 3 . The study of the system is due to Giube. 1 The diagram (Fig. 93) shows a distinct maximum corresponding to 27-4 per cent, thallium and two obscured maxima. For the latter, the FIG 94. concentrations are 33-33 per cent, and 40 per cent, thallium respectively, corresponding to the compounds Mg 2 Tl and Mg 3 Tl 2 . The maximum, of course, represents the com- pound Mg 8 Tl 3 , which melts to a homogeneous liquid, while the other two break up on fusion into crystals of different composition and into melts. In the region on the 1 Zeit. anorg. Chem., 46, 84 (1905). 184 CHEMICAL COMBINATION AMONG METALS. diagram between the fusion curve and the eutectic hori- zontals there is equilibrium between one species of crystal and the melt. These alloys oxidise in air, particularly in a moist atmo- sphere ; the compound Mg 2 Tl though not very stable is slightly more resistant than the others. Magnesium-tin. Magnesium and tin only form one com- pound, Mg 2 Sn. The system has been investigated by Grube x and Kurnakoff and Stepanoff. 2 The diagram (Fig. 94) shows a maximum at 783-4 and a concentration of 66-5 per cent, of magnesium ; this corresponds to the com- pound Mg 2 Sn. Below the fusion curve and above the eutectic horizontals there is in every region of the curve one species of crystals in equilibrium with the melt. The alloy Mg 2 Sn is formed with great development of heat ; it is brittle and has a steel blue colour which is tarnished on exposure by a stratum of black oxide. The diagram shows three groups of alloys : (1) alloys containing magnesium of primary separation ; (2) alloys containing Mg 2 Sn of primary separation ; and (3) alloys rich in tin containing tin of primary separation. Magnesium-cerium. Vogel, 3 who studied this system, has recorded the formation of four compounds, CeMg, CeMg 3 , CeMg 9 and Ce 4 Mg. His diagram is shown in Fig. 95. The two branches descending from A and C, along which cerium and the compound CeMg respectively separate, should apparently intersect in the point B. On the cooling curves of these alloys are noted two arrest points, one at about 632 and the other at 497 at a concentration of 20 per cent, magnesium. Vogel explains the presence of a compound here by assuming a decomposition to take place in the solid state. A small branch with a slight maximum at 20 per cent, should fill the gap between A and C. The compound CeMg separates at a concentration of 50 per cent, and 738. 1 Zeil. anorg. Chem., 46, 76 (1905). r 2 Ibid., 46, 177 (1905). 8 Ibid., 91, 277 (1915). HOMOPOLAR INTEEMBTALLIC COMPOUNDS. 185 The arrest point at E, which practically occurs a little lower, corresponds with the separation of saturated mixed crystals containing magnesium. From melts with 60 to 75 per cent, the latter being a maximum on the curve the compound CeMg 3 separates at about 780. Finally, at 90 per cent, of magnesium, the compound CeMg 9 separates. The compound Ce 4 Mg absorbs heat in its formation, i.e., it is endothermic ; it has a very small field of existence and 1ioo 10 00 100 FIG. 95. a low melting point. This compound, like all the alloys from 20 to about 62 per cent, of magnesium, ignites spon- taneously. The compound CeMg is very hard ; a freshly broken surface shows a reddish grey colour. CeMg 3 is less hard than CeMg and is not so easily oxidised. CeMg 9 is brittle and has a silvery lustre when freshly broken. It is more resistant than the other compounds to oxidation and the action of acids. Magnesium-lead. Lead forms with magnesium the com- 186 CHEMICAL COMBINATION AMONG METALS. pound Mg 2 Pb, as was found by Grube 1 and Kurnakoff and Stepanoff. 2 The latter authors consider compounds of the type Mg 2 E (where R = Sn or Pb) as belonging to the hypo- thetical type KH 4 by replacement of four atoms of hydrogen by two atoms of bivalent magnesium. The curve (see Fig. 96) is similar to the curve for magnesium-tin. The 60 70 90 90 -<00 FIG. 96. diagram consists of three parts ; in the first, lead separates ; in the third, magnesium ; while in the second, with a maxi- mum at 550 and 66-66 per cent, magnesium, the compound Mg 2 Pb separates. This compound is oxidised in moist air and has a steel grey colour. Magnesium- antimony. Antimony combines with mag- nesium forming the compound Mg 3 Sb 2 . The system has Zeit. anorg. Chem., 44, 117~(1905). 3 'Jbid., 46, 177 (1905). HOMOPOLAE INTEKMETALLIC COMPOUNDS. 187 been studied by Grube, 1 and the diagram constructed by him is shown in Fig. 97. A maximum occurs at 961 and 40 per cent, of antimony, corresponding to the compound Mg 3 Sb 2 . The eutectic horizontals are prolonged to the point indicating the concentration of the compound : there is, consequently, 1000 50 900 FIG. 97. no formation of mixed crystals. The compound crystallises in steel grey needles which are oxidised in the air. Alloys increase in brittleness up to 49-5 per cent, of antimony, and between this percentage and 95 per cent, of antimony are exceedingly brittle. Magnesium-bismuth. Magnesium forms the compound Mg 3 Bi 2 with bismuth, as noted by Grube. 2 The diagram 1 Zeit. anorg. Chem., 49, 87 (1906). 2 Ibid., 49, 183 (1906). 188 CHEMICAL COMBINATION AMONG METALS. (Fig. 98) shows two branches intersecting in an eutectic point at about 18 per cent, of bismuth. The one branch has a maximum at 710 and about 40 per cent, of bismuth, which corresponds to the compound mentioned. The eutectic horizontal extends from pure magnesium to the composition of the compound, thus excluding the possibility of solid solutions being formed. The compound is very slightly 900 800 700 600 i 500 400 300 200 \ \\ -fo 2.0 60 -YO SO . .90 FIG. 98. soluble in bismuth, and the eutectic alloy with bismuth consists almost entirely of bismuth. The compound is strongly exothermic. Freshly prepared it has a dark grey colour. It is very brittle ; in dry air it is stable, but in moist air it is oxidised in time to a black powder. Magnesium-nickel. Nickel and magnesium form, accord- ing to Voss, 1 two compounds. Mg 2 Ni and MgNi 2 . The diagram (Fig. 99) shows that at the point D the compound MgNi 2 separates from the melt. On the liquidus curve, Zeit. anorg. Chem., 57, 61 (1908). HOMOPOLAK INTEEMETALLIC COMPOUNDS. 189 however, instead of a maximum, a flat portion is observed. Voss maintains that along this portion of the curve crystal- lisation takes place from two liquid strata. He has not succeeded in demonstrating this, as hot magnesium attacks 1500 40 50 60 70 SO go -lOO 400 FIG. 99. porcelain strongly with resultant perforation and loss of liquid. At 768, NiMg 2 separates. On the corresponding horizontal a maximal arrest is shown for 66 per cent. magnesium. The compound MgNi 2 crystallises in thin platelets. Freshly broken it has a red colour but alters quickly in the 190 CHEMICAL COMBINATION AMONG METALS. air. The compound Mg 2 Ni has not been prepared in a pure state ; alloys containing it are composed of the compound MgNi 2 surrounded by Mg 2 Ni, together with a certain quantity of the eutectic alloy. CALCIUM COMPOUNDS. Calcium-aluminium. Donski, 1 working in Tammann's laboratory, has demonstrated the formation of the compound 800 e 600 50 500 10O FIG. 100. CaAl 3 . The curve is shown in Fig. 100. Arrest points are noted at about 692 between 12 and 34 per cent, of calcium. Since two liquid strata are observed above 692, it is supposed that they react, forming the compound CaAl 3 at the concen- tration, 25-5 per cent, of calcium, corresponding to the maximal arrest. Alloys up to 8 per cent, calcium have the colour of pure aluminium and are a little harder than this metal. They are fairly stable in air and cold water ; they react with hot water giving hydrogen. Alloys with a medium calcium content are brittle, porous and show exfoliations of 1 Zeit. anorg. Chem., 57, 185 (1908). HOMOPOLAK INTEEMETALLIC COMPOUNDS. 191 coarse silvery-white crystals ; they decompose cold water with evolution of hydrogen. The alloys richest in calcium are less brittle and are unstable in air. Calcium-thallium. Thallium and calcium combine to form the compounds CaTl 3 , Ca 3 Tl 4 , and CaTl. These compounds have recently been examined thermally by Baar. 1 The pre- ceding researches of Donski 2 are somewhat incomplete. FIG. 101. Up to 3 per cent, of calcium (see Fig. 101) a small series of mixed crystals separates ; then the compound CaTl 3 is formed as is deduced from the prolongation of the eutectic horizontal at 310 up to the composition indicated by this formula. By addition of calcium the curve rises slowly to D, after which the compound Ca 3 Tl 4 separates, its composi- tion being shown by a maximal arrest at 43-7 per cent, calcium and 556. A maximum occurs at 969 and 50-5 per cent, of calcium, corresponding to the compound CaTl. The 1 Zeit. anorg. Chem.. 70, 366 (1911). 2 Ibid., 57,206(1908). 192 CHEMICAL COMBINATION AMONG METALS. remainder of the curve indicates two series of mixed crystals between the latter compound and pure calcium. All these alloys are unstable in air, in fact, so quickly do alloys with 55 to 85 per cent, calcium oxidise that micro - FIG. 102. scopical examination is impossible. Alloys with 30 to 55 per cent, calcium are harder than thallium ; those richest in calcium are very brittle. Calcium-tin. This system has been studied partially by Donski. 1 The formation of the compound CaSn 3 melting at 624 has been noted. The curve (Fig. 102) rises directly 1 Zeit. anorg. Chem., 57, 206 (1908). HOMOPOLAR INTERMETALLIC COMPOUNDS. 193 from the melting point of tin to the melting point of the compound. A maximum occurs at 23-7 to 26-8 per cent, of calcium and 624. Mixed crystals are formed between 25 per cent, and 28-5 per cent, calcium containing high propor- tions of the compound. The alloys so far studied are oxydised in the air ; those containing even as little as 2 per cent, of calcium decompose cold water with evolution of hydrogen. Calcium-lead. Calcium forms with lead the compounds FIG. 103. CaPb 3 , CaPb, and Ca 2 Pb. The system has been studied by Donski * and Baar. 2 The curve (Fig. 103) is very similar to that for calcium-tin. At 25 per cent, and 649 there is a maximum due to the compound CaPb 3 . From the eutectic point D the curve rises quickly up to 982 and 51-8 per cent, calcium with separation of the compound PbCa. The curve then reaches a maximum at 1105 corresponding to the compound Ca 2 Pb. From 66-6 to 89 per cent, of calcium 1 Zeit. anorg. Chem., 57, 208 (1908). 8 Ibid., 70, 372(1911). c.M. 13 194 CHEMICAL COMBINATION AMONG METALS. mixed crystals separate. Maximum eutectic crystallisation occurs at 89 per cent, where the resulting alloy is a con- glomerate of calcium and saturated mixed crystals of calcium and the compound. These alloys are quickly oxydised in air to a black powder ; only those with 35 to 50 per cent, of calcium have been examined microscopically. Calcium-antimony. This system also has been studied by Donski l up to a concentration of 9 per cent, of calcium. The existence of a compound seems probable, but it has not been possible to define it. The investigation had to be abandoned on account of experimental difficulties, namely, the high melting point of calcium and the extreme instability of the alloys in air. Calcium-bismuth. Donski has also studied this system over a limited range. The formula? of the compounds which probably occur are unknown. Arrest points have been observed between 20 and 37 per cent, of calcium at about 500. COMPOUNDS OF ZINC. Zinc-aluminium. According to Tammann, 2 zinc forms with aluminium the compound Al 2 Zn 3 . This compound is, however, not admitted by Hey cock and Neville, 3 Shepherd, 4 and Eger, 5 who have only encountered solid solutions in their investigations of the system. Zinc-antimony. Antimony and zinc form the compounds Sb 2 Zn 3 and SbZn. The system has been investigated by Gosselin, 6 Monkmeyer, 7 and Zemczuzny. 8 The curve con- structed by the latter is given in Fig. 104. There is a maxi- mum at 566 and 60 per cent, of zinc corresponding to the 1 Loc. c.it. 2 Lehrb. d. MetaJL, p. 222. J. C. S., 71, 389 (1897). J. Phys. Chem., 9, 504 (1905). Int. Z. f. Metatt., 4, 35 (1913). Bull. Soc. d'Encour., (5), 1, 1312 (1896). Zeit. anorg. Chem., 43, 182 (1905). 8 Ibid., 49. 384 (1906). HOMOPOLAR INTEKMETALLIC COMPOUNDS. 195 compound Sb 2 Zn 3 . Zemczuzny denies the existence of ZnSb, believed by Gosselin to exist. In the region corre- sponding to this compound, he observed marked super- cooling due to the reaction Zn 3 Sb 2 + Sb = 3ZnSb. The curve shows irregularities which, together with the formation of metastable crystals, are not observed if the liquid is seeded with crystals of ZnSb. This compound melts with decom- position, so that the curve only shows a transformation point. 700* 600" 50 50IT 50 Zn Zn -10 jo 90 FIG. 104. 60 1-Q > lw OLto-ntL 9C 100 Zn 3 Sb 2 shows a modification which was only observed on one side of the maximum. Zinc-manganese. This system has been investigated by Parravano * up to 29-7 per cent, of manganese. He has noted the formation of the compounds MnZn 7 and MnZn 3 . The diagram is shown in Fig. 105. The melting point of zinc is lowered a few degrees by the addition of manganese, so that the first eutectic is very close to pure zinc on the diagram and only 2 below it. The two compounds are not indicated by well-defined arrests but rather by slackening in 1 Gazz. Chim.Ital,, 45, L, 1 (1915). 132 196 CHEMICAL COMBINATION AMONG METALS. the cooling curves, often accompanied by super-cooling. The alloys of this system are hard and brittle. Zinc-iron. Iron forms with zinc the compounds Zn 7 Fe and Zn 3 Fe. There are also three series of mixed crystals. The system has been studied by Wologdine, 1 Guertler, 2 von Vegesack 3 and finally by Kaydt and Tammann. 4 The diagram (Fig. 106) is that of von Vegesack, completed by /In?* FIG. 105. Kaydt and Tammann. At 777 and 25 per cent, of iron the melt separates the compound Zn 3 Fe which undergoes a transformation to the compound Zn 7 Fe at 662 and 15 per cent. iron. As above mentioned there are three species of mixed crystals, one up to -7 per cent, of iron, the second from 7*3 to 11 per cent, of iron, and the third from 80 to 100 per cent, of iron. From 25 to 86 per cent., the alloys consist of 1 Rev. d. Metall., 3, 701 (1906). 2 Int. Z. /. Metatt., L, 355. 8 Zeit. anorg. Chem., 52, 36 (1907). * Ibid., 83, 257 (1913). HOMOPOLAR INTERMETALLIC COMPOUNDS. 197 saturated mixed crystals and the compound Zn 3 Fe ; from 25 to 15 per cent., of the two compounds Zn 3 Fe and Zn 7 Fe and from 7-3 to ! per cent., of the two saturated mixed crystals. Magnetic properties are observed in alloys from 26-2 to 96 per cent, of iron ; such properties diminish in intensity with decrease of iron content. The alloys with 11 to 22 per 100 FIG. 106. cent, of iron containing the two compounds are exceptional. Probably their properties are due to the presence of small quantities of iron. On heating, such alloys lose their magnetic properties. The iron-zinc alloys are porous, hard, and very brittle. Zinc-cobalt. This system has been studied by Levkonja l 1 Zeit. anorg. Chem., 59, 319 (1908). 198 CHEMICAL COMBINATION AMONG METALS. to a limited degree, and, up to the present, the only com- pound recognised is one having the formula CoZn 4 . The maximum corresponding to this compound occurs at about 880 and 18-5 per cent, by weight of cobalt (see Fig. 107). From 5 to 13-4 per cent, of cobalt mixed crystals are formed. 95" 100 % u. (xc<,,vu 2V,. FIG. 107. From the melts of alloys containing 5 to 13-4 per cent, cobalt, there are separated, primarily, unsaturated mixed crystals which by addition of zinc pass, on cooling to 419, to the saturated mixed crystals a. Levkonja has not observed magnetic properties in the alloys containing up to 18-4 per cent, of cobalt. Zinc-nickel. Nickel forms with zinc the compound HOMOPOLAR 1NTERMETALLIC COMPOUNDS. 199 Zn 3 Ni and ZnNi, according to Tammann 1 and Voss. 2 The diagram constructed by the latter is given in Fig. 108. The melting point of zinc is not lowered by addition of nickel, and the curve rises directly from the zinc axis. Between 14-5 and 23 per cent, of nickel a distinct interval occurs, WOT 100 FIG. 108. together with a point of discontinuity. This range corre- sponds to the mixed crystals formed between Zn 3 Ni and Zn, the saturated mixed crystal occurring at 14-5 per cent, nickel. Microscopical examination confirms the results of thermal analysis. 1 MelalJurgie, 4 ; 781 (1007). 2 Zeit, anorg. Chem., 57, 67 (1908). 200 CHEMICAL COMBINATION AMONG METALS. CADMIUM COMPOUNDS. Cadmium-tin. Information on this system is as yet incomplete ; the compound CdSn 4 is probably formed. The system has been investigated by Kapp 1 and Stoffel, 2 and the diagram is shown in Fig. 109. Stoffel studied the alloys not only by thermal but also by microscopical and dilato- metric methods ; he also made determinations of electro- motive force. The last method, for some unknown reason, gave negative results, while the microscopical results were uncertain. He found a thermal arrest at 122 for all alloy -tco FIG. 109. from 2-5 to 50 per cent, of cadmium and accompanying small changes of volume, of which, however, he was unable to note the maximum. Cadmium-antimony. Antimony and cadmium form well- defined compounds, namely, CdSb and Cd 3 Sb 2 , the latter of which is in accordance with the known valencies of the two metals. The results obtained by Treitschke 3 and Kurna- koff and Konstantinoff 4 are in fair agreement. The diagram (Fig. 110) shows a maximum at 455 and 50 per cent. An eutectic horizontal at 290 reaches to a point below the 1 Ann. d. Phys., 6, 754 (1901). 2 Zcit. anorg. Chem., 53, 140 (1907), 3 Ibid., 50/217 (1906). 4 Ibid., 58, 12 (1908). HOMOPOLAR INTERMETALLIC COMPOUNDS, 201 maximum of the liquidus curve. Between 36 and 51-6 per cent, of antimony the compound CdSb separates as a solid phase of constant composition. By very slow cooling, another line (shown in dots on the diagram) is obtained, which corresponds to Cd 3 Sb 2 , with a maximum at 423. The cooling curves in this region show the formation of solid solutions. At 260 to 290 a reaction takes place in the solid FIG. 110. state with a rise of temperature of 20 to 30 degrees. This evolution of heat reaches a maximum for alloys containing 50 per cent, of antimony. The transformation is, of course, that of Cd 3 Sb 2 , which is unstable, into CdSb, which is stable, and has a higher melting point (Cd 3 Sb 2 + Sb = 3CdSb). The transformation point corresponding to stable Cd 3 Sb 2 is at 409 and 66-5 per cent, of cadmium. Cadmium-chromium. Hindrichs 1 in a study of this 1 Ze.il. anorg. Chem., 59, 427 (1908). 202 CHEMICAL COMBINATION AMONG METALS. system did not succeed in obtaining positive results and our information is incomplete. Cadmium-iron. The degree of chemical combination between cadmium and iron is not as 3^et completely known. The system has been investigated by Isaac and Tammann, 1 who found that on adding iron dust to molten cadmium and heating for a long time at 650 only one arrest point was obtained, which was in fact identical with the melting point of cadmium (321). Microscopical analysis revealed the presence of conglomerates, which Isaac and Tammann believe to contain a compound of the two metals. It is doubtful whether any compound rich in cadmium exists (as in the case of iron and zinc) or whether such compound is practically insoluble at its melting point. If such w r ere the case a single arrest would be noted at the melting point of cadmium, which actually occurs. Cadmium-cobalt. In this case also it is doubtful whether compounds are formed. The system has been studied by Levkonja, 2 who found an eutectic crystallisation in alloys with 2-5 to 10 per cent, cobalt at 316 to 6 below the melting point of pure cadmium. It was not possible to ascertain whether chemical combination took place. The alloys obtained did not exhibit magnetic properties at ordinary temperatures. Cadmium-nickel. Nickel and cadmium combine to form the compound Cd 4 Ni. The system has been studied, though to a limited extent, by Voss, 3 whose observations do not extend below 15 per cent, on account of the volatility of cadmium. An eutectic horizontal at 321 extends from 20 per cent, of nickel, the composition of the compound, to pure cadmium (see Fig. 111). Another horizontal is shown at 405 whose nature is not exactly understood. It evidently indicates a new species of mixed crystals. The nature of the solid phase above 502 is not known. 1 Zeit. anorg. Chem., 55, 61 (1907). 2 Ibid., 59, 322 (1908). 3 Ibid., 57, 69 (1908). HOMOPOLAR INTERMETALLIC COMPOUNDS. 203 COMPOUNDS OF MERCURY (AMALGAMS). The amalgams are of considerable historical importance in the study of the solubility of metals. It may be men- tioned that Ramsay and Tammann used mercury as a solvent in their studies on the atomicity of metallic molecules by the cryoscopic method. Our knowledge of the chemical com- pounds of mercury with other metals, although abundant, is somewhat inexact, as only in a few instances has the thermal method been used to ascertain formulae. We shall, how- 700 60(T 500* FIG. 111. i ever, not confine our description to the amalgams studied by the thermal methods which have so much enlarged the field of scientific investigation. A reliable summary of our knowledge of the various amalgams studied by the older methods of chemistry has been made by Ley. 1 We shall now allude briefly to the information which we actually possess of the crystalline amalgams. Mercury-aluminium. - - The aluminium amalgams have received considerable attention, but at present little is known of the capacity for chemical combination between the two metals. 1 Cf. Handhuch d. anorg. Chem,, R. Abegg, Vol. II., pp. 509 et seq. 204 CHEMICAL COMBINATION AMONG METALS. Cossa 1 prepared an amalgam by fusion in an atmosphere of inert gas ; Tarugi 2 allowed mercury vapour to act on aluminium. The compound Hg 3 Al 2 is supposed to exist, but this is not certain. Aluminium amalgam is chemically very reactive. 3 Mercury-gallium. This system has not yet received any attention. Eamsay 4 states that gallium dissolves easily in mercury, in which it agrees with thallium. It is, therefore, not improbable that the two metals combine chemically. Mercury-indium. T. W. Eichards and Wilson 5 studied the electro-chemical potential of indium amalgams, and obtained values in excess of those calculated by the mixture rule. According to Hildebrand 6 the compound Hgjn can probably occur. Mercury -thallium. - - This system has been studied by Kurnakoff and Pushing who noted the existence of the compound Hg 2 Tl. Studies have also been made by Regnault 8 and Carstanjen, 9 the former of whom affirmed the existence of the compounds Hg 5 Tl 2 and Hg^lg. The most recent investigations are those of Paulovitch, 10 who has con- firmed the work of Kurnakoff and Pushin and added some considerations on irrational dystectics. The behaviour of thallium with mercury resembles to an extent that of the alkali metals which give the type EHg 2 . The diagram of Kurnakoff and Pushin (Fig. 112) gives a maximum at 15 and 33-3 per cent, of thallium. The branch 1 Nuovo Cimento, (2), 3, 75 (1870). 2 Gazz. Chim. ltd., 34, II., 486 (1904). 3 Cf. Tissier, C. R., 49, 54. (1859) ; Casamajor, Chem, News, 34, 34 (1876) ; Jehu. Ber., 11, 360 (1878) ; Kronchkoll, J. de Phys., (2), 3, 139 (1884) ; Ramsay, J. C. S., 55, 521 (1889) ; Schumann, Wied. Ann., 43, 101 (1891) ; Coehn and Ormandy, Ber., 28, 1505 (1895) ; Biernacki, Wied. Ann., 59, 664 (1896) ; Humphreys, J. C. S., 69, 1679 (1896); Konovalotf, Chem. Centr., 1896, II., p. 338; Richards, Chem. News, 74, 30 (1896). 4 J. C. S., 55, 553 (1889). 5 Zeit. phi/8. Chem., 72, 141, 157, 164 (1910). J. Amer. C. 8., 35, 513 (1913). 7 Zeit. anorg. Chem., 30, 104 (1902). 8 C. E., 64, 111 (1867). 9 J. prakt. Chem., 102, 84 (1867). 10 Journ. Russ. phys.-chem. Soc., 47, 2946 (1915) ; Bull. Soc. Chim., (4), 20, 2 (191 6). HOMOPOLAR INTEEMETALLIC COMPOUNDS. 205 of the curve on which this maximum occurs belongs therefore to the compound TlHg 2 which melts without decomposition. Along the branch a, liquid amalgams separate. At low tem- peratures thallium is very soluble in mercury and lowers the freezing point of the latter to 60, which is the temperature 100 50 200* 50 100" 50 /\ l 9 J 50 100" 1o 30 l.o $0 60 10 SO, , 90 -too l^. CiL&m*. r/a FIG. 112. of eutectic solidification. It may be added that this is the lowest freezing point hitherto observed for any alloy. Mercury-tin. Tin dissolves in mercury, even at ordinary temperatures, with evolution of heat. According to Kooze- boom l and Van Heteren, 2 tin and mercury form solid solu- tions. Pushin 3 has also noticed the formation of solid 1 Verh. K. Ak. Wetensch., Amsterdam, 420 (1902). 2 Zeit. anorg. Chem., 42, 129 (1904). 3 Ibid., 36, 207 (1903). 206 CHEMICAL COMBINATION AMONG METALS. solutions. Kupffer l and Matthiessen 2 have determined the specific gravity of these amalgams. Tammann 3 studied the effect of the addition of small quantities of tin on the freezing point of mercury. He found that the freezing point was raised as follows : 063 gram. Sn dissolvel in 100 gm. Hg raised the f.p. '06 148 1-1 219 2-1 281 2-4- According to Tammann the existence of the compound Hg 3 Sn is probable, but the actual data render more probable the existence simply of isomorphous mixtures of the two metals. Mercury-cerium. The solubility of cerium in mercury was observed by Winkler, 4 and by Muthmann and Beck. 5 Cerium is very soluble in boiling mercury. Nothing is known of the capacity for combination of these metals. Mercury-antimony. Mercury is without action on anti- mony in the cold ; an amalgam is formed at higher tempera- tures. Amalgams of mercury have been prepared by means of the usual indirect methods by Bottger 6 and Vortmann. 7 According to Partheil and Mannheim, 8 the compound Hg 3 Sb 2 is probably formed. Mercury -uranium. The amalgams of uranium, prepared electrolytically by Feree, 9 lose mercury on heating to 242, leaving a residue of uranium which ignites spontaneously. Mercury-chromium. Chromium amalgams have been prepared by the direct fusion of the two elements. Schon- bein 10 and Vincent n obtained them by acting upon sodium 1 Ann. Chem., 40, 293. 2 Pogg. Ann., 112, 445. 3 Zeit. yhys. Chem., 3, 441 (1889). 1 Ber., 24, 1883 (1891). 5 Ann. Chem., 331, 56 (1904). J. prakt. Chem., 12, 350 (1837). 7 Ber., 24, 2762 (1891). 8 Arch. Pharm., 238, 160 (1900). 9 Bull. Soc. Chim., (3), 25, 622 (1901). 10 Jahresber., 1861, p. 95. " Phil. Mag., (4), 24, 328 (1862). HOMOPOLAK INTEEMETALLIC COMPOUNDS. 207 and potassium amalgams with concentrated chromium chloride solution. They have also been prepared and studied by Moissan. I Myers 2 and Feree 3 obtained chromium amalgams by electrolysis of chromium sulphate solution with a mercury cathode. The compounds Hg 3 Cr and HgCr have been described in chemical literature. The compound Hg 3 Cr loses mercury easily under pressure giving the other hypothetical com- pound HgCr. This in its turn loses mercury on heating to 300, leaving a residue of chromium (Feree). It is probable that these metals form solid solutions rather than definite compounds. Mercury-molybdenum. The amalgams of molybdenum have been prepared by electrolysis of a solution of molybdic anhydride in hydrochloric 4 or sulphuric 5 acid. Feree 6 obtained a semi-solid amalgam by electrolysis : from the crystalline phase, the compounds Hg 9 Mo, Hg 2 Mo, and Hg 3 Mo 2 separated under pressure. These compounds, however, lose their mercury completely on heating, leaving behind a residue of pyrophoric molybdenum. Mercury -manganese. Manganese amalgams have been prepared by Bottger, 7 Giles, 8 Moissan, 9 and J. Schumann, 10 by the action of sodium amalgam on manganous chloride. Ramsay u obtained an amalgam by electrolysis of manganese chloride solution using a mercury cathode. The researches of 0. Prelinger 12 must also be mentioned ; from these the existence of the compound Mn 2 Hg 5 is argued. C. R., 88, 180 (1879) ; Ann. Chim. Phus., (5), 21, 250 (1880). 2 J. Am. C. S., 26, 1126 (1904). 3 C. B., 121, 823 (1895). 4 Zeit. Elektroch., 12, 146, 154 (1906). 5 J. Am. C., S., 26, 1124 (1904). 6 C. R., 122, 733 (1896). 7 J. prakt. Chem., 12, 350 (1837) Phil. Mag., (4), 24, 328 (1862). 9 C. R., 86, 180 (1879). 10 Wied. Ann., 43, 110 (1891). " J. C. S., 55, 532 (1889). 12 Ber. Wien. Akad., 102, 346 (1893). 208 CHEMICAL COMBINATION AMONG METALS. Mercury-iron. These amalgams have been prepared exclusively by indirect methods, either electrolytically, or else by the action of alkali amalgams on solutions of iron salts. Cailletet 1 showed that iron may be superficially amalgamated. From amalgams containing about 15 per cent, of iron Joule 2 obtained, by the application of pressure, a crystalline phase of composition FeHg. Zamboni 3 pre- pared an iron amalgam by electrolysis of ferrous ammonium sulphate, using a mercury cathode, Kammann, 4 by the same method as Joule obtained a solid phase of composition Fe 2 Hg 3 . It is not, however, certain whether mercury and iron combine chemically. Mercury-cobalt and mercury-nickel. These amalgams have been little studied with reference to the capacity of their components for chemical combination. They have been prepared by Darmour, 5 Moissan 6 and Schumann. 7 Mercury -platinum, etc. Little is known of the amalgams of platinum, palladium, osmium, iridium, rhodium and ruthenium. Joule 8 obtained a crystalline amalgam of platinum of composition PtHg 12 . It is most probable that these metals form simple isomorphous mixtures with mercury. Compounds of Metals of Group III. with Metals of other Groups. COMPOUNDS OF ALUMINIUM. Aluminium- cerium. The system has been investigated by Vogel, 9 who claims that five compounds are formed, namely, AlCe 3 , AlCe 2 , AlCe, Al 2 Ce, and Al 4 Ce. Two of these com- C. R., 44, 1250 (1857). J. C. S., (2), 1, 378 (1863). Nuovo Cimento, (4), 2, 26 (1895). Ber., 14, 1433 (1881), J. prakf. Chem., 17, 346 (1839). C. R., 88, 180 (1870). Wied. Ann., 43, 111 (1891). Jahresber., 1850, p. 333 ; 1863, p. 280. Zeit. anorg. Chem., 75. 41 (1912). HOMOPOLAE INTERMETALLIC COMPOUNDS. 209 pounds, Al 2 Ce and Al 4 Ce, have melting points much above those of their components. The curve (Fig. 113) shows two maxima, one very pronounced at about 1460 and 33 per cent, of cerium due to the compound CeAl 2 ,and the other, not so pronounced, at about 614, corresponding to Ce 3 Al. The asymmetry of the latter maximum is due in all probability to the compound Ce 3 Al being strongly dissociated, while the FIG. 113. compound Ce 2 Al is less dissociated. The latter compound, formed at 33 per cent, of aluminium, melts at 595 with decomposition into CeAl and a melt ; CeAl separates primarily in beautiful prisms which at 780 split up into CeAl 2 and a fused mixture. Above 700, between 35 and 50 per cent, of aluminium, weak thermal effects are noted which may be due to the occurrence of lanthanium and didymium in the cerium used for experiment. After CeAL 2 , which separates at about 67 per cent, of aluminium, comes the compound Al 4 Ce which C.M. 14 210 CHEMICAL COMBINATION AMONG METALS. separates at 1245 and 80 per cent, of aluminium. This compound undergoes a polymorphic transformation at 1005 which is shown in the diagram as a weak thermal effect. The change from the form ft, stable at higher temperatures, to the form a, stable at lower temperatures, is characterised by a contraction. The cerium-aluminium alloys are in general stable in the air, unacted upon by water, and more resistant to the action of acids than the pure metals. They exhibit distinct hard- ness between 35 and 80 per cent, of aluminium with a maxi- mum for CeAl 2 . With this hardness is associated extreme brittleness. Aluminium-antimony. The various chemical combina- tions of these metals, if indeed they exist, have not been well elucidated up to the present. The existence of a com- pound AlSb seems certain. Wright 1 noted that antimony and aluminium form an alloy with a high melting point, and containing 81-6 per cent, by weight of antimony (correspond- ing to AlSb). Eoche 2 and Gautier 3 have confirmed this result. The study of the equilibrium system is due to Campbell and Matthews 4 and, later, to Tammann. 5 The fusion diagram (Fig. 114) shows two maxima, one at 50 to 53 per cent, of aluminium and the other at 90 per cent, aluminium. Of these, the former corresponds to the com- pound AlSb ; no crystalline species has been found to corre- spond to the latter maximum. Tammann has attempted to solve this problem by showing that the form of the curve is different when this single compound of aluminium and anti- mony is formed slowly from its fused components. If the compound is formed more quickly there may occur a dis- placement of the maximum, an increase of the quantity of one of the components, or even the appearance of a second 1 J. C. 8. /., 1892, p. 493. * Mon. Scient., (4), 7, 269 (1893). 3 Butt. Soc. d'Encour., (5), 1, (1896). 4 J. Am. C. 8., 241, 259 (1902). 6 Zeit. anorg. Chem., 48, 53 (1906). HOMOPOLAR INTEEMETALLIC COMPOUNDS. 211 maximum. The compound AlSb is formed slowly from its components at about 700 ; at 1100 the reaction proceeds very rapidly. (Cf. p. 7). Aluminium-manganese. Two compounds of these metals with each other occur, but their composition has not been ascertained with certainty up to the present. According to Gwyer, 1 their formulae are probably AlMn 3 and Al 3 Mn, respectively. Guillet 2 prepared alloys of aluminium and FIG. 1H. manganese and found the compounds Al 3 Mn, Al 3 Mn 2 and Al 4 Mn. Hindrichs 3 has studied the system by means of thermal methods The diagram (Fig. 115) shows that the melting point of manganese is raised by addition of aluminium. At about 90 per cent, of manganese there occurs, at 1279, the first separation. Between 10 and 35 per cent, of aluminium the temperature only changes by about 2. Further, on the cooling curves of melts from 5 to 40 per cent, of aluminium, 1 Zeit. anorg. Chem.. 57, 150 (1908). 2 C. R. t 134, 236 (1908) ; Le Genie Civil, 41, 139, 156, 169, 188, 363, 377, 393 (1902). 3 Zeit. anorg. Chem., 59, 44 (1908). 142 212 CHEMICAL COMBINATION AMONG METALS. distinct crystallisation intervals occur with a minimum at 14 per cent, of aluminium. The temperature change of 2 on the fusion curve may well be due to ordinary experimental error, for the temperature along that portion of the curve should either remain constant or else show r a maximum. In 600 10 10 90 FIG. 115. the former case a compound would be formed from the melts in question, in the latter case there would occur a crystallisa- tion of a series of homogeneous melts into a continuous series of mixed crystals, of which that mixed crystal with the highest melting point would be a definite compound. Thermal analysis does not enable us to decide which is actually the case that occurs. At 1040 to 1050 small arrests are noted in the alloys up to 45 per cent, of aluminium with a HOMOPOLAE INTEKMETALLIC COMPOUNDS. 213 maximal arrest at about 27 per cent, which should be due to the decomposition of mixed crystals with 50 per cent, of aluminium into manganese and the crystals a. If the mixed crystals formed from the melt consist of the compound AlMn 3 there ought here to be a decomposition of AlMn 3 into manganese and the mixed crystals a. From 35 to 73 per cent, mixed crystals separate. Between 73 and 95 per cent, there exists a gap in miscibility. The formation of mixed crystals is accompanied by super-cooling. At 670 the melt a reacts with the saturated mixed crystals forming a com- pound which may be detected by microscopical examination. Its formula is probably Al 3 Mn. Aluminium-iron. Gwyer 1 states that these metals form a compound Al 3 Fe. Between 100 and 43 per cent, of iron (see Fig. 116) a series of mixed crystals is formed. From 30 to 50 per cent, of iron there occur on the cooling curve either breaks due to primary separation or else eutectic arrest points ; these occur at 1087 for alloys with 65, 62-5 and 60 per cent, of aluminium, and at higher temperatures for alloys between 57-5 and 50 per cent, of aluminium. From 67-5 to 76-8 per cent, aluminium, crystallisation intervals are noted which correspond to a series of mixed crystals of which the last member, at about 75 per cent., probably represents the compound Fe A1 3 . In the alloys from 41 to per cent, of iron the proportion of FeAl 3 diminishes, while that of aluminium increases. In the cooling of these alloys a very small thermal effect is observed at 550, believed by Gwyer to be due to the formation of a compound from Al 3 Fe and Al ; this compound is, however, not detected by microscopical analysis. The reaction between the two metals is accompanied by such an evolution of heat that Gwyer was unable to make use of porcelain fusion tubes and was obliged to substitute tubes of magnesia. Aluminium-cobalt. These metals combine, forming the 1 Zeit. anorg. Chem., 57, 129 (1908). 214 CHEMICAL COMBINATION AMONG METALS. compounds Co 3 Al 13 , Co 2 Al 5 and CoAl. This system also was studied by Gwyer. 1 The curve (Fig. 117) is somewhat similar to that for FeAl. It shows two discontinuities and a maximum. Between 100 and 80 per cent, cobalt mixed crystals of the two metals are formed. From 80 to 50 per 1500" 600 10 T.O 30 60 vo ^0 9o ~ 100 j;> 7 ixv*. Cttc;7-u Al /a FIG. 116. cent, all cooling curves show crystallisation intervals, but microscopical analysis shows that only those alloys with 50 to 65 per cent, cobalt have a homogeneous structure, while from 80 to 65 per cent, two structural elements are observed. Since these alloys even on long annealing do not acquire homogeneity, Gwyer concludes that there is here an 1 Zeit. anorg. Chem., 57, 136 (1908). HOMOPOLAE INTBBMETALLIG COMPOUNDS. 215 interrupted series of mixed crystals (with a maximum) or else a gap in miscibility between two phases. At 1628 and 50 per cent, the compound CoAl crystalHses. The latter, reacting with the melt, gives at 1165 and 22-5 per cent, of 1700' 1600 1500 60 >10 SO 90 100 10 FIG. 117. cobalt, the compound Co 2 Al 5 . At 1110 small breaks are noted which probably indicate a polymorphic transforma- tion. From Co 2 Al 5 and the melt, crystals of Co 3 Al 13 separate at 940 and 18 per cent, cobalt. The formulae of the last two compounds are indicated by the positions of th,e maximararrests, 216 CHEMICAL COMBINATION AMONG METALS. Gwyer observed very small thermal effects also at 550 ; he believes it possible that there are two such effects, one for a reaction between -Co 3 Al ]3 and aluminium, and the other in alloys richer in Co 3 Al 13 due to a polymorphic transformation in this compound. 6C 10 ?..U~ OLton-U. A/ SJaJUU 80 90 -1oo FIG. 118. Aluminium-nickel. These metals form the following com- pounds : NiAl 3 , NiAl 2 and NiAl. The system was studied by Gwyer (loc. cit.). The curve (Fig. 118) shows a maximum, two discontinuities and an eutectic point. Between 100 HOMOPOLAB INTERMETALLIC COMPOUNDS. 217 and 73-15 per cent, of nickel there separates a series of mixed crystals of the two metals. From 63 to 73-5 per cent, small arrests are noted at 1370 ; these alloys consist of two species of crystals. At lower temperatures, however, this lack of miscibility changes ; in fact the alloy with 63 per cent, of nickel after long annealing has a homogeneous structure. From 73 to 50 per cent, of nickel there is a series of mixed crystals of NiAl and nickel. The maximum shows that at about 1640 and 50 per cent., the compound NiAl is formed. At 1130, by action of the melt containing 25 per cent, nickel on the crystals of NiAl, the compound NiAl 2 is formed at a concentration of 34 per cent. These crystals of primary separation react at 825 with a melt containing 15 per cent, nickel, forming at 25 per cent, a compound which is probably NiAl 3 . Aluminium- chromium. The compound AlCr 2 is probably formed from these two metals. The system was studied by Hindrichs 1 to a limited extent. Working with alloys con- taining more than 70 per cent, of aluminium, the temperature of fusion was so high that the aluminium attacked and destroyed the magnesia tubes which were used. It was, therefore, necessary to study these alloys by micrographical methods alone. At 644 (see Fig. 119) arrests were noted as far as 60 per cent, of chromium. By extrapolation it would appear that they die out at 85 per cent, chromium. From 5 to 70 per cent, there are also arrests at 975. The separation which occurs here is often accompanied by slight super- cooling. On the higher branch of the curve there probably occurs the separation of compound with a very high melting point. Microscopical examination shows that alloys from 85 to 96 per cent, of aluminium have a homogeneous structure ; they represent mixed crystals of chromium and a compound of the two metals. The formula is probably AlCr 3 , which would require 85-12 per cent, by weight of chromium. 1 Ze.it. anorg. C/iem., 59, 433 (1908). 218 CHEMICAL COMBINATION AMONG METALS. THALLIUM COMPOUNDS. Thallium-lead. With thallium, lead probably forms the compound PbTP. The system has been studied by Lev- konja l and by Kurnakoff and Pushin. 2 It is of interest by reason of certain considerations which may be made with 15-00 FIG. 119. reference to the application of thermal analysis to the interpretation of equilibrium diagrams. Each metal raises the melting point of the other, and the fusion curve passes through a point representing a much higher temperature than either of the melting points of the pure metals. 1 Zeit. anorg. Chem., 52, 454 (1907). 2 Ibid., p. 435. L. Rolla studied the thermo-electric power and specific volume of these alloys. Oazz. Chim. Ital., 45, L, 185 (1915). HOMOPOLAR INTERMETALLIC COMPOUNDS. 219 Levkonja holds that this maximum (see Fig. 120) which occurs at about 275 and 34 per cent, of lead indicates the compound PbTl 2 which forms solid solutions with both components. Kurnakoff and Pushin, however, contend that the curve merely indicates the formation of solid solu- tions and belongs to Roozeboom's second type (see p. 16). This question can only be resolved by microscopical exami- nation. Kurnakoif has, however, given powerful support to his contention by observing that the addition of tin displaces 380- 300' 280 10 1 OO FIG. 120. the maximum to a point where Tl : Pb = 1 : 2-5 instead of 1-7 --1-8:1. Thallium- antimony. Antimony and thallium combine to form the compound SbTl 3 . The system has been investi- gated by Williams 1 (see Fig. 121). Pure antimony crystal- lises from melts rich in that metal. Thallium occurs in two polymorphic forms which transform at 225. From melts between 29-8 and 22 per cent, of antimony, pure thallium does not separate, but a series of mixed crystals. At 195 conglomerates are formed which consist of antimony and mixed crystals of a Tl with 22 per cent, of antimony ; from 22 to per cent, they consist of a series of mixed crystals of Zeit. anorg. Chem., 50, 129 (1906). 220 CHEMICAL COMBINATION AMONG METALS. composition corresponding to that of the melt. At 187 or 8 below the eutectic temperature and 25 per cent, of anti- mony a new species of crystals is formed, namely, the com- pound SbTl 3 , with a super-cooling of about 2. Alloys with less than 50 per cent, of thallium are hard, brittle and capable of receiving a polish. With increase of thallium they become softer and can only be polished with difficulty. Microscopical analysis has confirmed these results, FIG. 121. although at 187 Williams failed to obtain a homogeneous alloy, for there occurred an admixture of rod-like crystals of antimony. On subjecting this heterogeneous alloy to pro- longed annealing, however, the free antimony diminished in amount and finally almost disappeared. At 22 per cent, the quantity of the compound decreased while that of the mixed crystals a TISb increased. Thallium-bismuth. This system is of particular interest. The two metals combine but, as was mentioned in a preceding chapter, their combination cannot be reconciled with the law of definite proportions. Chikashige l first studied the system 1 Zeit. anorg. Chem., 51, 330 (1906). HOMOPOLAK INTEEMETALLIC COMPOUNDS. 221 which was subsequently examined by KurnakofT, Zemczuzny and Tararin. 1 The fusion diagram shows three maxima, two of which are distinct and one less pronounced. The first of these maxima occurs at -9 per cent, of bismuth at 301-5 ; it corresponds to the separation of a solid solution of bismuth in thallium (phase ft). At 5-8 per cent, these solid solutions 25 100 FIG. 122. cease to exist. The second maximum is at 12-05 per cent, of bismuth and 303' 7 : between 5-9 and 33 per cent, of bismuth more solid solutions are formed. The third maximum is at 62-8 per cent, and 211-7, and a further series of solid solu- tions crystallises from 55 to 64 per cent, of bismuth. The last branch of the curve corresponds to the separation of pure bismuth without any notable formation of solid solutions. 1 Zeit. anorg. Chem., 83, 200 (1913). 222 CHEMICAL COMBINATION AMONG METALS. According to Kurnakoff and his co-workers, the three maxima do not correspond to any rational atomic propor- tions. Chikashige, though admitting the irrationality of two of the maxima, maintains that the third corresponds to the formula Bi 5 Tl 3 . Kurnakoff and his collaborators from a study of the electrical conductivity, compressibility, hardness and micro- scopic structure of these alloys conclude that no definite formula can be assigned to correspond with any of the three maxima. 2000* 1800* 1600' 1200' 1000* '800* 60(f 200' 0' ys Pr -f/ 7 . Tl Tl =Ft- W 10 10 50 60 ?0 go 90 FIG. 123. Thallium-iron. Nothing is known of the alloys of .these metals. Isaac and Tammann, 1 who made some preliminary investigations, obtained no results, the boiling point of thallium being lower than the melting point of iron. They did in fact observe a certain lowering of the melting point of iron by thallium, but the thallium very quickly distilled off unaltered. Thallium-platinum. According to Hackspill, 2 thallium 1 Zeit. anorg. Chem., 55, 61 (1907). 2 C. ., 146, 820 (1908). HOMOPOLAE INTEKMETALLIC COMPOUNDS. 223 and platinum form a compound TIPt. Thallium dissolves easily in platinum. The curve (Fig. 123) shows that the addition of platinum to thallium causes lowering of the melt- ing point by a few degrees. With increase of the content in platinum there occurs, at 50 per cent, and 685, the sepa- ration of crystals of the formula PtTl. The curve then falls slightly and subsequently rises so that at 70 per cent, of platinum it reaches about 1000. Platinum forms solid solutions with thallium. Between the compound and platinum no solid solutions are formed according to Hack- spill's observations. Compounds of Metals of Group IV. with Metals of other Groups. COMPOUNDS OF TIN. Antimony-tin. It is probable, though not certain, that these metals combine to form a compound SbSn. The system has been studied by Eeinders, 1 Gallacher 2 and Williams. 3 The diagram constructed by the last named is shown in Fig. 124. On cooling melts rich in antimony, a series of mixed crystals of the two metals separates. At 420, from 50 to 90 per cent, of antimony there forms, after separa- tion of mixed crystals, a new crystalline species surrounding the mixed crystals rich in antimony. There are no corre- sponding arrests on the cooling curve but only slackening. At 243, however, up to 50 per cent, of antimony, distinct arrests are noted. The mixed crystals in all probability are transformed into SnSb crystals at 50 per cent, by addition of tin. The existence of this compound is, however, not well established. Tin-bismuth. Chemical combination between these metals, although stated to take place by Tammann, 4 is 1 Zeit. anorg. Chem., 25, 113 (1900). 2 J. Phys. Chem., 10, 93 (1906). 3 Zeit. anorg. Chem., 55, 14 (1907). Lehrb. d. MetalL, p. 222 (1914), Leipzig. 224 CHEMICAL COMBINATION AMONG METALS. excluded by the complete investigations of Mazzotto. 1 Ihis is also shown by the experimental work recorded on p. 45. The system has also been studied by Stoffel 2 and Lepkovsky. 3 650' 600 200 150 QO 100 Tm-manganese. Manganese forms with tin the com- pound SnMn 4 , SnMn 2 and possibly SnMn. The diagram (Fig. 125) has been traced by Williams. 4 The three com- pounds are represented by three breaks in the curve, the first at 988 and 80-1 per cent, of magnanese, the second at 1 Mem. Inst. Lomb., 16, (1886), and Nuovo Cimento, 18 (1900). 2 Zeit. anorg. Chem., 53, 148 (1907). 3 Ibid., 59, 287 (1908). 4 Ibid., 55, 26 (1907). HOMOPOLAR INTERMETALLIC COMPOUNDS. 225 898 and 64-8 per cent, of manganese, and the third at 541 and 15 per cent, of manganese. The reaction which gives rise to the latter does not, however, take place to completion, but part of the crystals of SnMn 2 surrounded by SnMn are deposited in the presence of the melt. 1300" 1200 1100 200 \oo FIG. 125. There is a series of mixed crystals rich in manganese, the last number of which is the saturated mixed crystal with about 4 per cent, of tin. SnMn 4 is less brittle than man- ganese and, when polished, has a surface like that of steel. Hardness, 4-5. SnMn 2 is similar in colour to the preceding C.M. 15 226 CHEMICAL COMBINATION AMONG, METALS. compound and its hardness is 3 4. SnMn has a silvery white colour. These compounds are weakly attacked by acids. SnMn 4 is magnetic, SnMn 2 to a less degree, and SnMn least of the three. Tin-iron. There is doubt as to the occurrence of com- pounds of tin and iron. Levin and Tammann * and Isaac and Tammann 2 have studied the system and constructed a diagram (Fig. 126). Iron and tin do not mix in all propor- tions in the liquid state. At 1140, between 32 and 79 per cent, of tin, two strata are found, one rich in iron and the other rich in tin. The greatest arrest due to crystallisation occurs at 32 per cent, of tin. At 893 between 10 to 93 per cent, tin, arrest points occur. The mixed crystal y reacts with the melt containing 79 per cent, of tin to form a compound. The thermal effect is ? however, small and sho\vs no marked maximum. From 10 to 95 per cent, tin there occur at 780 further arrests which may be due to poly- morphic transformation in the compound formed at 893. There is a distinct maximal arrest at 24 per cent, of tin which should correspond to a compound SnFe 3 . At 496, between 35 and 97-5 per cent, of tin, arrests occur which, how r ever, have not been explained. These alloys display notable magnetic properties ; the tin content has a marked influence on the temperature at which iron loses its magnetic permeability. Tin-cobalt. Two compounds are formed between these metals, namely, SnCo and SnCo 2 . The system has been studied by Levkonja 3 and Zemczuzny and Belynski. 4 The results obtained by these workers are in accordance. Levkonja's diagram is shown in Fig. 127. The curve shows a maximum and a break, the maximum at 1150 and 33 per cent, of tin, corresponding to SnCo 2 , and the break at 948 and 75 per cent, of tin, corresponding to the compound 1 Zeit. anorg. Chem., 47, 141 (1905). 2 Ibid., 53, 281 (1907). 3 Ibid., 59, 298 (1908). 4 Ibid., p. 368. HOMOPOLAE INTEEMETALLIC COMPOUNDS. 227 CoSn, 1 which undergoes a polymorphic transformation at 520. At 229 there are arrest points which are due to the eutectic between CoSn and pure tin. The compounds CoSn and Co 2 Sn are harder than their 1500 1300 500' too FIG. 126. component elements. With regard to magnetic properties the cobalt-tin alloys may be divided into two groups ; up to 50 per cent, of cobalt, magnetic properties are wanting ; the alloys containing more than 50 per cent, of cobalt are increasingly magnetic with increase of cobalt. 1 The composition of this compound is indicated by the position of the maximal arrests at 948 and 520. Translator's Note. 152 228 CHEMICAL COMBINATION AMONG METALS. Tin-nickel. Nickel combines with tin, and the following compounds have been shown to exist by thermal analysis : 1500 100 300 200 FIG. 127. Ni 3 Sn 2 , Ni 3 Sn, and Ni 4 Sn. The diagiam (Fig. 128) was traced by Voss. 1 At the beginning of the curve mixed crystals separate. At 1135 and 82-5 per cent., eutectic 1 Zeit. anorg. Chem., 57, 38 (1908). HOMOPOLAE INTEEMETALLIC COMPOUNDS. 229 crystallisation occurs. At 1162 and 60 per cent, of nickel, the compound Ni 3 Sn 2 is formed. From 64 to 42 per cent, of nickel a lack of miscibility is noted and two liquid phases FIG. 128. occur. Another gap in miscibility occurs also between 30 and 7 per cent, of nickel ; these melts are in equilibrium with the compound Ni 3 Sn 2 . At 885, arrests are noted between 85 and 60 per cent, of nickel with a maximum for 230 CHEMICAL COMBINATION AMONG METALS. 67 per cent, (by weight) of nickel, corresponding probably to the compound Ni 4 Sn. Microscopic analysis confirms the thermal data. Tin-platinum. Platinum forms the following compounds with tin : SnPt 3 , SnPt, Sn 3 Pt 2 , and Sn 8 Pt 3 . The system was studied by Doerinckel l and is given in Fig. 129. SnPt 3 2000 FIG. 129. >%inatomi is formed at about 1375 and 75 per cent, of platinum by reaction between crystals of platinum and the melt contain- ing 70 per cent, of platinum. Crystals of SnPt then begin to separate ; the maximum occurs at 50 per cent, and 1281. At 846 crystals of SnPt react with the melt of composition fi. 1 Zeit. anorg. Chem., 54,35(1907). HOMOPOLAE INTEBMETALLIC COMPOUNDS. 231 The arrests at this temperature show a maximum for about 40 per cent, platinum. According to Doerinckel this maximal arrest is due to the compound Sn 3 Pt 2 . A small difference between the position of the observed maximal arrest and that required for the compound may be explained by the formation of a layer around the crystals which hinders the completion of the reaction, as appears evident on micro- scopical examination. A further series of arrests occur at 738 with a maximum at 40 per cent, platinum ; these indicate, therefore, a polymorphic transformation of the compound Sn 3 Pt 2 . At 537 this modification of the com- pound Sn 3 Pt 2 reacts with the melt having 5 per cent, of platinum, giving a compound at 28 per cent, of platinum which is probably Pt 3 Sn 8 . The hardness of these alloys grows with increase of platinum content ; at 40 per cent, it is equal to that of calc spar, at 60 per cent, to that of fluor spar. Maximum hard- ness is found at 80 per cent., at which it is slightly greater than that of apatite. Mineral acids attack these alloys in inverse proportion to their platinum content. CERIUM COMPOUNDS. Cerium-bismuth. - - These alloys have been studied by Vogel, 1 who noted the formation of the compounds BiCe 3 , Bi 3 Ce 4 , BiCe, and Bi 2 Ce. The maximum (see Fig. 130), which is found at the high temperature of 1630, corresponds to the compound Bi 3 Ce 4 , which requires a concentration of about 53 per cent, (by weight) of bismuth. At 1400, as shown by microscopical analysis, Bi 3 Ce 4 reacts with the melt forming the compound BiCe 3 which has an obscured maxi- mum at 32 per cent, of bismuth. This compound has its primary separation along B C and at 757 crystallises eutectically with cerium. Small thermal effects are noted between 830 and 860, due to reactions between bismuth and small quantities of lanthanum and didymium present as ] Zeit. anorg. Chem., 84, 327 (1914). 232 CHEMICAL COMBINATION AMONG METALS. impurities in the cerium used in the experiment. At 1525 and about 60 per cent, of bismuth, the compound BiCe is formed which separates primarily between 61 and 82 per cent, of bismuth. At 882 it reacts with the melt forming a (100 A goo a. ion 306 <\\ to 30 30 to so 60 fo 80 FIG. 130. <)o compound richer in bismuth (Bi 2 Ce, requiring 74-8 per cent, bismuth). These alloys are very easily oxidisable ; exposed to air, they are quickly altered to a black powder and are energetic- ally attacked by acids. The hardness is greatest for medium concentrations. HOMOPOLAE 1NTERMETALLIC COMPOUNDS. 233 Cerium-iron. These alloys, which are noteworthy for their pyrophoric properties, are known through the work of Auer von Welsbach. land2 COMPOUNDS OF LEAD. Lead-bismuth. It is doubtful whether chemical combina- tion occurs between lead and bismuth. The system has been studied by Mazzotto, 3 Kapp, 4 Stoffel, 5 and Barlow. 6 The two branches of which the equilibrium curve consists intersect at 125 and 55 per cent, of bismuth. According to Kapp, the thermal effects corresponding to eutectic crystallisation occur from 36 to 98 per cent, of bismuth, which would show that up to 36 per cent, of bismuth solid solutions are formed. Lead-palladium. Palladium forms a number of com- pounds with lead, namely, Pb 2 Pd, PbPd, Pb 3 Pd 4 , PbPd 2 and PbPd 3 . Some doubt exists as to the occurrence of the last ; the first compound has been admitted to exist by Pushin and Paschsky, 7 as a result of their measurements of the electro- lytic potential of these alloys. The system has been studied by Euer, 8 whose diagram is reproduced in Fig. 131. Pb 2 Pd separates at 454 and 30-49 per cent, of palladium. This compound melts without decomposition. From 40 to 75 per cent, of palladium, three series of arrests indicate three 1 D.R. P., 154,807(1903). 2 Vogel has recently published a paper on the alloys of cerium and iron (Zeit. anorg. Chem., 99, 25 (1917) ). He found that the metals are miscible in all proportions and form two compounds, namely, CeFe 2 and Ce 2 Fe 3 . The first is changed into the second at 773 ; Ge^Fe 3 decomposes at 1085 into a liquid and a solid solution rich in iron ; the solution contains 15 per cent, of cerium and becomes poorer in this element on cooling. The compound CeFe 2 is magnetic, but loses its magnetic properties at 116. It is not certain whether the second compound is magnetic. Pyrophoric properties are exhibited by these alloys, those with 70 percent, of cerium displaying this property to a most marked degree so that a slight scratch is sufficient to cause ignition. The compounds are hard and brittle ; they are not oxidised at ordinary temperatures. Translator's Note. 3 Memt. 1st. Lomb., (3), 7 (1886), and Nuovo Cimento, 18 (1909). 4 Drud. Ann., 6, 754 (1901). 5 Zeit. anorg. Chem., 53, 150 (1907). 6 J. Am. C. 8., 32, 1394 (1910); Zeit. anorg. Chem., 70, 183 (1911). 7 Zeit. anorg. Chem., 62, 360 (1909). 8 Ibid., 52,347(1907). 234 CHEMICAL COMBINATION AMONG METALS. compounds, each of which on melting decomposes into a melt, and a new species of crystal. The first is formed at 500 and 50 per cent, of palladium and is the compound PbPd. The second compound is indicated by a weak maximal arrest at 595 and 56 per cent, of palladium ; the most probable formula is Pb 3 Pd 4 . The formulae Pb 6 Pd 7 , Pb 5 Pd 6 , and Pb 4 Pd 5 are less probable. The third compound . FIG. 131. gives a maximal arrest for 67 per cent, of palladium here again the maximum is not well defined. The of the last compound should be PbPd 2 . The fifth compound, PbPd 3 , gives a maximum on the fusion 1230 and 75 per cent, of palladium. The hardness of these alloys is greater than that and increases up to about 78 per cent, of palladium, reaches a maximum (5), and afterwards decreases. at 830 ; formula and last curve at of lead, where it Alloys HOMOPOLAE INTEEMETALLIC COMPOUNDS. 235 with 27 to 74 per cent, of palladium are very brittle ; the remainder are not so markedly brittle. Lead-platinum. These metals combine to form a com- pound PbPt. The existence of two other compounds is doubtful. The system, of which the diagram is reproduced in Fig. 132, was studied by Doerinckel. At 40 per cent, of 1800 \ So 10 to 30 (Jo fo JO OP 4/100 0 "I / fo FIG. 181. also been made by Pellini. 2 Fig. 182 is the diagram con- structed by the last named. From the melting point of i Dp / o oo JOO Jfoo 10 /, S*. / fi J~ ^ /J FIG. 182. selenium (217) the curve rises rapidly to 696, and remains horizontal from 5 to 52 per cent, of silver, rising then 1 Metall, 5, 355 (1908). 2 Qazz. CMm. Ital, 45 L, 533 (1915). 298 CHEMICAL COMBINATION AMONG METALS. to 897 and 66*6 per cent, of silver, the maximum cor- responding to the compound Ag 2 Se. The curve falls and then rises to 890 and remains horizontal from 68 to 89 per cent, of silver, rising thence to the melting point of the metal. The two horizontal tracts represent gaps in miscibility. At 122 at all concentrations, arrest points are noted which are due to a transformation of the compound Ag 2 Se. Pellini performed his experiments in an atmosphere of nitrogen. Selenium-zinc. Eio (1828) found a compound ZnSe 4 combined with HgS in a mineral from Mexico. Fonces- Diacon (1900) by heating zinc chloride in hydrogen selenide obtained the compound ZnSe in crystalline form. Selenium- cadmium. Margottet (1877) obtained from cad- mium and hydrogen selenide the compound CdSe, prepared also by Fonces-Diacon (1901). Selenium-mercury. This system has been studied by Pellini. 1 The compounds Hg 2 Se and HgSe are formed ; the latter can be distilled without decomposition. The highest temperature at which the distillation can be performed is from 600 to 650. From 500 to 650 the mixture is semi- liquid. From 132 to 139 an arrest is observed due to the solidification of selenium after super-cooling. Selenium-indium. According to Thiel and Koelsch (1910) these elements form a compound In 2 Se 3 . The existence of another compound poorer in selenium is. also probable. Indium and selenium mix with a brisk reaction, giving a dark liquid. Selenium- thallium. The system has been studied by Pelabon. 2 The compounds Tl 2 Se, TISe and T1 2 S 5 are formed. The fusion curve, leaving the melting point of thallium (302), runs at first horizontally at 400. Corre- sponding to this horizontal tract two liquid layers exist, the lower of pure thallium and the upper a mixture of Tl 2 Se and selenium. The curve then falls to an eutectic at 23 per 1 Rend. Ace. Lincei, 18, II., 211 (1910) ; Oazz. Chim. Ital., 40, 44 (1910). 2 C. It. 145, 118 (1907). HETEEOPOLAE INTEBMETALLIC COMPOUNDS. 299 cent, of selenium, from which it rises again to 338, where a maximum corresponding to the compound TISe occurs. From this maximum the curve falls to the composition of the compound T1 2 S 5 , after which there is another horizontal tract at 195. Here again two liquid layers occur, the upper being a solution of a selenide in selenium, and the lower being the selenide Tl 2 Se 5 . SVO" 800 JO 20 30 40 50 60 10 SO .90 100% FIG. 183. Selenium-tin. The selenides SnSe, Sn 2 Se 3 and SnSe 2 are known ; they were found by Biltz and Mecklenburg 1 in a study of the equilibrium between these elements. The compound SnSe shows a maximum at about 40 per cent, of selenium and 861 (Fig. 183). The second compound occurs at 50 per cent, and 645, and the third, not well authenticated, at the same temperature and about 57 per cent, of selenium. The curve falls to the left from the maximum at first quickly, 1 Zeit. anorg. Chem., 64, 232 (1909). 300 CHEMICAL COMBINATION AMONG METALS. then slowly, and finally almost vertically to the eutectic, which practically coincides with the melting point of pure tin. On the other side of the maximum the curve, after falling steeply, shows a discontinuity corresponding to the second compound. A more or less horizontal tract then suggests the occurrence of a lack in miscibility. The eutectic, which crystallises at 217, is almost pure selenium. \\ to- FIG. 184. Selenium-lead. The system lead-lead selenide has been studied by Friedrich and Leroux l and by Pelabori. 2 The melting point of the selenide is very high. According to Friedrich and Leroux it is at 1088, according to Pelabori at 1065. From a concentration of 50 per cent., at which the compound is formed, the curve falls continuously to the melting point of pure lead (Fig. 184). Corresponding with the melting point of lead (326), arrests are noted for all mixtures from to 50 per cent. 1 MetalL, 5,355(1908). 2 C. R., 144, 1159 (1897). HETEEOPOLAE INTEEMETALLIC COMPOUNDS. 301 Pelabon has investigated the further course of the curve in the region between the selenide and pure selenium. The curve falls to 70 per cent, of selenium, and remains level at 673, indicating the formation of two strata, the upper of which contains free selenium and the lower the compound. The existence of a compound PbSe 2 is not indicated. Selenium- antimony. Selenium forms with antimony the compound Sb 2 Se 3 reported by Pelabon, 1 by Chretien, 2 who mentioned in addition the compounds SbSe, Sb 3 Se 4 and Sb 4 Se 5 , and also by Parravano, 3 who studied the system thermally. The compound (see Fig. 185) separates at 50 per cent, and 630. It does not mix with antimony in all proportions in the liquid state ; two liquid strata occur with 11 and 35 per cent, of selenium respectively. Between 60 and 70 per cent, of selenium a change in direction of the curve indicates, according to Parravano, a lack of miscibility in the liquid state with an upper or lower critical point. 1 J. Ch. Phys., 2, 437 (1904) ; C. R., 142, 207 (1906). 2 C. R., 142, 1339 (1906). 3 (Jazz. Chim. ltd., 43, L, 210 (1913). 302 CHEMICAL COMBINATION AMONG METALS. From 50 to 100 per cent, of selenium marked supercooling is observed. Selenium-bismuth. Selenium forms two compounds with bismuth, BiSe and Bi 2 Se 3 , observed by the thermal method through the work of Parravano. 1 The curve (Fig. 186) rises rapidly from the melting point of bismuth up to about 27 per cent, of selenium, where, between 600 and 610, thermal effects are observed due to the formation of the compound FIG. 186. BiSe ; this is confirmed by microscopical evidence. The curve again rises to 706 and 37 per cent, of selenium and shows a maximum for the compound Bi 2 Se. It then des.cends to about 625, remains horizontal along a tract and subsequently falls almost vertically to the melting point of pure selenium. The compound Bi 2 Se 3 is not completely miscible with selenium ; for some mixtures, two liquid layers occur in the liquid state, the lower of which consists mainly of the compound and the upper mainly of selenium. Selenium- chromium. Moissan (1880) obtained the com- 1 Gazz. Chim. Hal, 43, I., 210 (1913). HETBEOPOLAE INTERMETALLIC COMPOUNDS. 303 pounds Cr 2 Se 3 and CrSe by heating chromium chloride in an atmosphere of hydrogen selenide. The existence of these compounds is, however, not well authenticated. Selenium-manganese. Fonces-Diacon (1900) obtained the compound MnSe, which, he states, does not decompose on heating. Selenium-iron. Fonces-Diacon also prepared the follow- ing selenides of iron : Fe 2 Se 3 , Fe 3 Se 4 , or Fe 7 Se 8 and FeSe 2 . These on heating are changed into a compound whose com- position approximates to the formula FeSe. The homo- geneity of this product has not been clearly demonstrated. Selenium-nickel and Selenium- Cobalt. - - According to Fonces-Diacon (1900) selenium forms with nickel and cobalt compounds with the following formulae : M 2 Se, MSe, M 3 Se 4 , M 2 Se 3 and MSe 2 , where M = Co or Ni. His researches, however, were not founded on the most trustworthy methods. Selenium- palladium. Rossler (1895) isolated a selenide, PdSe. An alloy with 6 per cent, of selenium, on treatment with nitric acid, gave a compound with the formula Pd 4 Se. Selenium-platinum. Rossler also obtained a selenide with the formula PtSe, while Minozzi claimed to have obtained the compounds PtSe 3 and PfcSe 2 , the first by precipitating a solution of the double cyanide of platinum and selenium with formaldehyde, and the second by reduction of the first. COMPOUNDS OF TELLUKIUM (TELLURIDES). Tellurium- copper. Two compounds are formed, namely, Cu 2 Te and Cu 4 Te 3 . The system has been investigated by Chikashige 1 ; the diagram is given in Fig. 187. The first compound is formed at 855 at a concentration of about 66- 6 per cent, of copper. It separates secondarily in the form of a series of mixed crystals with tellurium from melts con- taining 100 to 66 per cent, of copper, and primarily from melts with 66 to 50 per cent, of copper. At 623 the last 1 Zeit. anorg. C hem., 51, 50 (1907). 304 CHEMICAL COMBINATION AMONG METALS. species of crystal reacts with the melt of composition Z), forming the compound Cu 4 Te 3 . At 365 there is a thermal effect, Cu 4 Te 3 being transformed into another crystalline species. Since the maximal arrest occurs at the composition of the compound Cu 4 Te 3 it would appear that this is a poly- morphic transformation of the compound. Tellurium- silver. The compounds Ag 2 Te and AgTe are known. Pushin l has investigated the system and also ' 3600 ca. Cerium .... Ce 140-25 > 800 Caesium .... Cs 132-9 26 Cobalt .... Co 59-0 1480 Chromium Cr 52-1 1520 Copper . Cu 63-6 1084-1 Gallium Ga 70-0 30 Gold .... Au 197-2 1063-5 Indium .... In 115-0 155 Iridium .... Ir 193-0 2350 (?) Iron .... Fe 55-9 1530 Lanthanum La 138-9 810 (?) Lead .... Pb 206-9 327-4 Lithium. Li 7-03 186 Magnesium Mg 24-36 635 Manganese Mn 55-0 1260 Mercury. Hg 200-0 38-9 Molybdenum . Mo 96-0 2500 (?) Nickel .... Ni 58-7 1451 Osmium. Os 191-0 2700 (?) Palladium Pd 106-5 1549 Phosphorus P 31-0 I. 4411. 930 Platinum Pt 194-8 1780 Potassium K 39-15 62-5 Rubidium Rb 85-5 38 Ruthenium Ru 101-7 2450 (?) Selenium Se 79-2 217 Silicon .... Si 28-4 1420 Silver .... Ag 107-93 961-5 Sodium .... Na 23-5 97-5 Sulphur S 32-6 | I. 112-8 II. 119-2 III. 106-2 Strontium Sr 87-6 > Ca, < Ba (?) Thallium Tl 204-1 302 Tellurium Te 127-6 450 Tin .... Sn 119-0 232 Titanium Ti 48-1 1800 Tungsten W 184-0 > 3000 Uranium Ur 238-5 < 1850 Vanadium V 51-2 1720 Zinc .... Zn 65-4 419 TABLES. 327 i 1 - ft . 6 II II II ^j co cb 222 II H II I 9 o oo ^ co 4f OS Oi OS II n n II II II ii ^ II II II | 80' ^ 43 o co OS ft t- O 10 05 CO J3 i I a O pj OS r ( CO II || w II 11 d _! II P=H O PH HH CO CO 3 ,_; O jj PH 9 CO II o CO 10 II os CO OS (M GO r 1 10 00 CO o pO fH O 02 H O S3 02 o^ PH H ft Q 9 r 1 i i t- O II ?-( fJJ w ^ w XI sauog 1 I-H cq CO * 1O 1 CO t- 00 os o 328 CHEMICAL COMBINATIONS AMONG METALS. BINARY SYSTEMS STUDIED THERMALLY. (Homopolar Combinations.) |Na| K jRb C8JCu[Af|AujBc CaJMgjZn Cd|ligJAl In|TI SB p Pb|Ce Sb|Bi[Cr|MojMn Fe|Ni|Co|Pdl Pt y |O|O| 1 1 1 1 1 IIH C c ^ 1 Na| |C| _|0|C| 1 |OjC|C c of- C|C|C|- C c L 1 1 K| -l-l- II 1 1 O c C|C|0 C|C c c 1 1 1 u Rb i i r* 1 ii ^ 1 Cs| 1 1 1 s~* 1 1 II 1 ^ 1 1 1 Cu 0|0|C C|C|C|C C O|C|O |C|OJO|-|0|O|O|0|O|O Ag |O| C|C|C C| |C| O|C|O| C|O O| c 0|0|O[O|C Au| 1 ^l c c c |C| jOiCiCj c |_|-|C|0|0 0|0|0 Be| 1 1 i 1 I r Ca| 1 1 1 c c c C|C - C|C|C |C|C| | | | | | | | Mil 1 |C|C|C c C|C|C|C|C|C 1 f~** 1 1 II 1 Zo| 1 |0|0|0||0|0|0 C|0|C|-|C|C|C Q 1 1 Cd| | |O|0| O|O|O| |C|O|C C|C|C|- Hf| 1 1 1 C* 1 C* o O| |O| -l-l- 1 All | 1 1 1 ! |O|0|OiC|C|O|C -JC|C|C c la I | *l 1 1 1 1 |0 o ii 1 1 1 T.| 1 1 1 1 1 1 1 1 c |C|C| -|0 0|0|0|- 1C Sn| C C |0 - c o c o Crj c C|0| MoJ - A c c c Ma! |0|0|0 - Fe| 1 l c ! c l i' Nil 1 1 |o|l Co| 1 1 1 1 1 1 . l-l- Pdl 1 1 1 1 1 1 1 1 1 1 1 1 1 1 !0 BINARY SYSTEMS. 329 BINARY SYSTEMS IN WHICH CHEMICAL COMBINATIONS DO NOT OCCUR. System. Bibliography. Ag-As Ag-Au Ag-Bi Ag-Co Ag-Cr Ag-Cu Ag-Mn Ag-Na Ag-Ni Ag-Pb Ag-Pd Ag-Si Ag-Tl Al-Be Al-Bi Al-Cd Al-K Al-Na Al-Pb Al-Si Al-Sn A1-T1 Al-Zn i As-Au As-Bi As-Pb As-Zn Au-Bi Au-Co Au-Cu Au-Fe Au-Ni Au-Pd Au-Pt ! Au-Tl K. FRIEDRICH AND A. LEROUX, Metall., 3, 194 (1906). ROBERTS-AUSTEN AND KIRKE ROSE, Chem. News, 87, 2 (1904) ; HEYCOCK AND NEVILLE, Phil. Trans., 189, A 69 (1897). G. J. PETRENKO, Z. anorg. Cli., 50, 138 (1906). 53,215(1907). G. HINDRICHS, Z. anorg. Ch., 59, 425 (1908). K. FRIEDRICH AND A. LEROUX, Metall., 2, 298 (1907); W. v. LEPKOVSKI, Z. anorg. Ch., 59, 290 (1908) ; HEYCOCK AND NEVILLE, Phil. Trans., 189, A 25 (1897); N. KURNAKOFF, N. PUSHIN AND ZUKOVSKI, Z. anorg. Ch., 68, 123 (1910). G. HINDRICHS, Z. anorg. Ch., 59, 440 (1908). E. QUERCIGH, Z. anorg. Ch., 68, 303 (1910) ; C. H. MATIIEWSON, Int. Z. Metall., 1, 57 (1911). G. J. PETRENKO, Z. anorg. Ch., 53, 213 (1907). K. FRIEDRICH, Metall., 3, 398 (1906) ; PETRENKO, Z. anorg. Ch., 53, 202 (1907) ; HEYCOCK AND NEVILLE, Phil. Trans., 189, A 37 (1897). R. RUER, Z. anorg. Ch., 51, 316 (1906). G. ARRIVAUT, Z. anorg. Ch., 60, 439 (1908). PETRENKO, Z. anorg. Ch., 50, 135 (1906). G. OESTERHELD, Z. anorg. Ch., 97, 6 40 (1916). A. G. C. GWYER, Z. anorg. Ch., 49, 318 (1906). 57, 150 (1908). D. P. SMITH, Z. anorg. Ch., 56, 113 (1908). C. H. MATHEWSON, Z. anorg. Ch.. 48, 193 (1906). A. G. C. GWYER, Z. anorg. Ch., 57, 149 (1908). W. FRAENKEL, Z. anorg. Ch., 58, 157 (1908). A. G. C. GWYER, Z. anorg. Ch., 49, 315 (1906) ; HEYCOCK AND NEVILLE, Journ. Chem. Soc., 57, 376 (1890). FR. DOERINCKEL, Z. anorg. Ch., 48, 189 (1906). PIEYCOCK AND NEVILLE, Journ. Chem. Soc., 71, 383 (1897) : SHEPHERD, Journ. of Phys. Chem., 9, 504 (1905). K. FRIEDRICH, Metall, 5, 360 (1908). K. FRIEDRICH AND P. LEROUX, Metall, 5, 148 (1908). K. FRIEDRICH, Metall, 3, 46 (1906). K. FRIEDRICH AND A. LEROUX, Metall, 3, 477 (1906). R, VOGEL, Z. anorg. Ch., 50, 147 (1906). W. WAHL, Z. anorg. Ch., 66, 65 (1910). N. S. KURNAKOFF AND ZEMCZUZNY, Z. anorg. Ch., 54, 164 (1907) ; ROBERTS AUSTEN, Proc. Roy. Soc., 67, 105 (1901). E. ISAAK AND G. TAMMANN, Z. anorg. Ch., 53, 294 (1907). M. LEVIN, Z. anorg. Ch., 45, 239 (1905). R. RUER, Z. anorg. Ch., 51, 393 (1906). FR. DOERINCKEL. Z. anorg. Ch., 54, 347 (1907). M. LEVIN, Z. anorg. Ch., 45, 34 (1905). 330 BINARY SYSTEMS. BINARY SYSTEMS IN WHICH CHEMICAL COMBINATIONS DO NOT OCCUR continued. System Bibliography. Bi-Ca Bi-Cd Bi-Co Bi-Cr Bi-Cu Bi-Fe Bi-Hg Bi-Pb Bi-Sb Bi-Si Bi-Sn Bi-Zn C-Ni Ga-Fe Ca-Sb Cd-Co Cd-Cr Cd-Fe Cd-Hg Cd-Pb Cd-Sn Cd-Tl Cd-Zn Co-Cu Co-Ni Co-Pb Co-Tl Cr-Cu Cr-Pb Cr-Sn Cr-Zn Cu-Mn L. DONSKI, Z. anorg. Ch., 57, 215 (1908). A. STOFFEL, Z. anorg. Ch., 53, 149 (1907). K. LEVKONJA, Z. anorg. Oh., 59, 317 (1908). R. S. WILLIAMS, Z. anorg. Ch., 55, 24 (1907). K. JERIOMIN, Z. anorg. Ch., 55, 413 (1907) ; GAUTIER, Contr. a V etude des alliages, 1901, p. 1 10 ; HETCOCK AND NEVILLE, Phil. Trans., 189, A 25 (1897) ; ROLAND -GossELiN, Bull. Soc. tfEncour. (5), 1, 1310 (1896). E. ISAAK AND G. TAMMANN, Z. anorg. Ch., 55, 60 (1907). N. A. PUSHIN, Z. anorg. Ch., 36, 214 (1903). A. STOFFEL, Z. anorg. Ch., 53, 150 (1907). K. HUTTNER AND G. TAMMANN, Z. anorg. Ch., 44, 138 (1905). R. S. WILLIAMS, Z. anorg. Ch., 55, 22 (1907). W. v. LEPKOVSKI, Z. anorg. Ch., 59, 287 (1908) ; A. STOFFEL, ibid., 53, 148 (1907). ARNEMANN, Metall, 7, 201 (1901) ; HEYCOCK AND NEVILLE, Journ. Chem. Soc., 71, 394 (1897) ; SPRING AND ROMANOFF, Z. anorg. Ch., 13, 29 (1897). K. FRIEDRICH AND P. LEROUX, Metall., 7, 10 (1910). C. QUASEBART, Metall., 3, 28 (1906) ; L. STOCKEM, ibid., 3, 147 (1906). L. DONSKI, Z. anorg. Ch., 57, 217 (1908). K. LEVKONJA, Z. anorg. Ch., 59, 322 (1908). G. HINDRICHS, Z. anorg. Ch., 59, 427 (1908). E. ISAAK AND G. TAMMANN, Z. anorg. Ch., 55, 61 (1907). BIJL, Z. phys. Ch., 41, 641 (1902). A. STOFFEL, Z. anorg. Ch., 53, 152 (1907). 53,146(1907) KURNAKOFF AND PUSHIN, Z. anorg. Ch., 30, 106 (1902). G. HINDRICHS, Z. anorg. Ch., 55, 417 (1907) ; HEYCOCK AND NEVILLE, Journ. Chem. Soc., 71, 383 (1897) ; GAUTIER, Butt. Soe. dEncour. (5), 1, 1293 (1896). R. SAHMEN, Z. anorg. Ch., 57, 3 (1908). W. GUERTLER AND G. TAMMANN, Z. anorg. Ch., 42, 361 (1904). K. LEVKONJA, Z. anorg. Ch., 59, 314 (1908). " 59,318(1908). G. HINDRICHS, Z. anorg. Ch., 59, 422 (1908). 59,429(1908). 59,418(1908). 59, 427 (1908) ; H. LE CHATELIER, Bull. Soc. d'Encour. (4), 10, 388 (1895). R. SAHMEN, Z. anorg. Ch., 57, 23 (1908) ; ZEMCZUZNY, URASOFF AND RYKOVSKOFF, ibid., 57, 256 (1908). BINARY SYSTEMS. 331 BINARY SYSTEMS IN WHICH CHEMICAL COMBINATIONS DO NOT OCCUR continued. System Bibliography. Ou-Ni Cu-Pb Cu-Pd Cu-Pt Cu-Tl Fe-Mn Fe-Pb Fe-Pt Fe-Tl Fe-V Fe-W Hg-Pb Hg-Sn Hg-Zn In-Pb In-Tl K-Li K-Mg Li-Mg LiNa Mg-Na Mn-Ni Mn-Pb Mn-Tl Ni-Pb Ni-Tl Pb-Pt Pb-Sb Pb-Si Pb-Sn Pb-Zn GUERTLER AND TAMMANN, Z. anorg. Ch.. 52, 27 (1907) ; KURNA- KOFF AND ZEMCZUZNY, ibid., 54, 153 (1907) ; GAUTIER, Bull. Soc. d'Encour. (5), 1, 1310 (1896). K. FRIEDRICH AND A. LEROUX, Metall, 4, 300 (1907). R. RUER, Z. anorg. Ch., 51, 225 (1906). FR. DOERINCKEL, Z. anorg. Ch., 54, 337 (1907). R. SAHMEN, Z. anorg. Ch., 57, 13 (1908). M. LEVIN AND G. TAMMANN, Z. anorg. Ch., 47, 141 (1905). E. ISAAK AND G. TAMMANN, Z. anorg. Ch. 55, 59 (1907). 55,66(1907). 55, 61 (1907). R. VOGEL AND G. TAMMANN, Z. anorg. Ch., 58, 77 (1908). H. HARKORT, Metall, 4, 617, 673 (1907). N. A. PUSHIN, Z. anorg. Ch., 36, 213 (1903) ; JANECKE, Z. phys. Ch., 60, 399 (1907). VAN HETEREN, Z. anorg. Ch., 42, 129 (1904). Cf. p. 205. N. A. PUSHIN, Z. anorg. Ch., 36, 214 (1903). N. S. KURNAKOFF AND N. A. PusiiiN, Z. anorg. Ch., 52, 444 (1907). 52,445(1907). G. MASING AND G. TAMMANN, Z. anorg. Ch., 67, 189 (1910). D. P. SMITH, Z. anorg. Ch., 56, 114 (1908). G. MASING AND G. TAMMANN, Z. anorg. Ch. 67, 197 (1910), 67, 189 (1910). C. H. MATHEWSON, Z. anorg. Ch., 48, 194 (1906). ZEMCZUZNY, URASOFF AND RYKOVSKOFF, Z. anorg. Ch., 57, 263 (1908). R. S. WILLIAMS, ibid., 55, 32 (1907). N. BAAR, ibid., 70, 360 (1911). G. Voss,. anorg. Ch., 57, 47(1908). ,,57,50(1908). FR. DOERINCKEL, Z. anorg. Ch., 54, 361 (1907). W. GOUTERMANN, Z. anorg. Ch., 55, 421 (1907); ROLAND- GOSSELIN, Bull. Soc. tfEncour. (5), 1, 1301 (1896). S. TAMARU, Z. anorg. Ch., 61, 43 (1909). P. N. DEGENS, Z. anorg. Gh., 63, 212 (1909); A. STOFFEL, ibid., 53, 139 (1907) ; D. MAZZOTTO, Int. Z. Metall., 1, 289 (1911). HEYCOCK AND NEVILLE, Journ. Chem. Soc., 71, 304 (1897) ; SPRING AND ROMANOFF, Z. anorg. Gh., 13, 29 (1897); ARNE- MANN, Metall., 7, 201 (1910). 332 BINAEY SYSTEMS. BINARY SYSTEMS IN WHICH CHEMICAL COMBINATIONS DO NOT OCCUR continued. System. Bibliography. S-Se S-Te Se-Te Sb-Si Sb-Sn Si-Sn Si-Tl Sn-Tl Sn-Zn Tl-Zn W. E. RINGER, Z. anorg. Ch., 32, 202 (1902). M. CHIKASIIIGE, Z. anorg. Ch., 72, 112 (1911). G. PELLINI AND G. Rio, E. Ace. Line., 15, 46 (1906). R. S. WILLIAMS, Z. anorg. Ch., 55, 20 (1907). ' 55,14(1907); REINDERS, ibid., 25, 113 (1900). S. TAMARU, Z. anorg. Ch. } 61, 42 (1909). 61,45(1909). N. S. KURNAKOFF AND N. A. PuSHiN, Journ. Chem. Soc., 30, 106 (1902). HEYCOCK AND NEVILLE, Journ. Chem. Soc., 71, 383 (1897) ; ARNEMANN, Metall, 7, 201 (1910). A. v. VEGESACK, Z. anorg. Ch., 52, 32 (1907). INDEX OF AUTHORS. A. ABEGG, R.,38, 203. Ahrens, 251. Alex6jeff, 4. Amort, 278. Aristotle, 24. Arnemann, 277, 330, 331, 332. Arrivaut, 168, 329. Ashley, 311. Aten, A. H. W., 288. Auerbach, F., 89. Avogadro, 31, 32. B. BAAR, N., 113, 152, 161, 191, 193, 249, 331. Bauer, 267, 286. Bachmetjeff, 154. Backer, 149. Baikoff, 158, 160. Barlow, 101, 233. Bartoli, 41. Barus, 67. Battelli, A., 154. Beck, 117, 119, 206. Becquerel, 81, 154, 164. Behrens, 99, 100. Bekier, 245. Belynsky, 226. Benedicks, C., 66, 90. Bengougli, 275. Berry, 48. Berthelot, 164. Berth ollet, 31, 34. Berzelius, 38. Bessler, 288. Biernacki, 204. Bijl, 78, 330. Biltz, W., 121, 285, 286, 289, 299, 309. B lough, 156. Blunt, 269. Bottger, 112, 125, 164, 174, 206, 207. Boltzmann, 58. Bornemann, 105, 149, 296. Borodovsky, 291. Bottone, 90. Boudouard, 108, 109, 147, 160, 181. Bradley, M. W., 102. Brinell, 97. Bruni, G., 39, 163. C. CAILLETET, 207. Cambi, L., 80, 110, 112. Campbell, 155, 210, 269. Cannizzaro, 32. Carnot, 262. Carpenter, 149, 155. Carrara, G-., 74. Carstanjen, 204, 278. Casamajor, 204. Cavazzi, 269. Chapman, 111. Charpy, 120, 316. Chavanne, 112. Chester, A., 175. Chikashige, M., 21, 45, 220, 303, 308, 332 Chretien, 301. Chwolson, 66, 70, 85, 87. Clarke, 275. Clausius, 41. Coehn, 154, 164, 204. Cossa, 204. Croockewit, 164, 174. Curlt, 100. Curry, 155. D. D ALTON, 33, 51. D arm our, 208. Debye, P., 64. De Cesaris, 278. De-Chalmet, 257. Defacqz, 262, 270. Degens, 331. Delepine, 251. Deseamps, 276. De Souza, 154, 164, 175. Dieckmann, 245, 279. Doerinckel, 117, 169, 230, 235, 262, 329 Donski, 114, 115, 190, 191, 192, 193, 194, 330. 334 INDEX OP AUTHOES. Ducelliez, 281. Duclaux, F., 64. Dumas, 164. Dumesnil, 278. Dupuis, 278. E. EDWARDS, 149, 155. Efremoff, 270. Eger, 194, 322. Einstein, 63. Emmerling, 268, 269. Empedocles, 24. Eucken, A., 85. Evans, 111. F. FABRE, 313. Faraday, M., 36, 70. Fay, 310,311. Feodoroff, 163. F6ree, J., 112, 125,206, 207. Fetsch, 111. Fonces-Diacon, 298, 303. Fraenkel, 329. Friedrich, H., 71, 242, 275, 276, 277, 278, 279, 280, 281, 282, 283, 286, 290, 293, 295, 296, 297, 300, 329, 330, 331. Fromm, 154. GALLAGHER, 223. Gastaldi, G., 325. Gattermann, 258. Gautier, 120, 162, 163, 165, 168, 210, 269, 330, 331. Gehlhoff, G., 85. Gercke, 271. Gibbs, W. J., 24, 74, 314. Giebelhausen, 252, 261. Giles, W. B., 207. Gillson, 310. Gin, 255. Giolitti, 158. Giua, M., 12. Goldsclimidt, 258. Gosselin, 194, 330, 331. Goutal, 262. Goutermann, 331. Gouy, 154, 164, 175. Granger, 268, 269, 275. Groth, 99, 100. Grotthus, 41. Grube, 108, 109, 181, 183, 184, 186, 187. Guareschi, 31, 74. Guertler, 66, 67, 122, 150, 159, 168, 196, 256, 263, 265, 275, 296, 330. Gugliehnini, D., 58. Guichard, 292. Guillaume, 87. Guillet, 87, 100, 155, 165, 210. Guldberg, 26, 32. Guntz, 125. Gwyer, 155, 210, 213, 216, 329. H. HAAS, 150. Haber, F., 154. Hackspill, 222. Haken, W., 81,83, 312. Hannesen, 252. Harkort, 331. Hautefeuille, 252, 255. Headden, 101. Henry, T., 175. Herschkovitch, M., 74, 78, 79, 149. Hertz, H., 89. Heteren, van, 205, 331. Heusler, 72. Heycock, 46, 100, 156, 162, 165, 167, 168, 173, 175, 176, 194, 329, 330, 331, 332. Heyn, E., 267, 286. Hildebrand, J. H., 204. Hill, 275. Hilpert, 245. Hindrichs, 168, 201, 210, 217, 329, 330. Hiorns, 158, 275. Honigschmid, 251, 260, 262. Hoff, van't, 42. Hogg, 72. Hoitsema, C., 57. Holt, 261. Honda, K., 71. Hovsleff, 269. Howe, 151. Hiittner, 120, 330. Humphreys, W., 154, 204. I. IGGENA, 125. Isaac, 202, 222, 226, 329, 330, 331. J. JACOBS, 258. Jaeger, 78. Janecke, 139, 318, 321, 331. Jeriomin, 330. Jolibois, 267, 269, 278. Joly, 254. INDEX OF AUTHORS. 335 Jones, II. C., 5, 36. Jonker, 291. Joslin, 275. Joule, J., 154, 164, 208. K. KAIILBAUM, G. W. A., 48. Kaneko, 121. Kapp, 200, 233. Kasauzeff, 175. Kerp, 112, 125. Kirke-Rose, 163, 164, 269, 329. Klein, 151. Knaffl, L., 175. Kobayaski, 306, 309. Koelsch, 288, 298, 308. Konovaloff, 204. KonstantinofE, 200, 237, 271, 273. Kremann, R., 12, 42. Kronchkoll, 204. Kupffer, 206. Kurnakoff.N. S.. 21, 24, 25, 32, 33, 35, 37, 45, 47, 51, 52, 66, 88, 91, 94, 97, 100, 105, 129, 131, 134, 136, 139, 140, 145, 184, 186, 200, 204, 218, 221, 237, 238, 329, 330, 331, 332. Kusnetzoff, 129. L. LAMB, 275. Laschtschenko, 236. Laurin, 288. Lautsch, 248. Lawrie, 176. Lebeau, 145, 257, 258, 262, 267. Le Chatelier, 25, 65, 66, 86, 87. 150, 155, 272, 293. Leleux, 255. Lepkovski, 224, 329, 330. Leroux, A., 242, 255, 277, 283, 290, 296, 297, 300, 329, 330, 331. Levin, 226, 329, 331. Levkonja, 121, 197, 202, 218, 226, 238, 247, 264, 330. Levy, 260. Ley, 203. Liebenoff. 65. Lincoln, 151. Lindemann, 63, 101. Littleton, F., 164. Losseff, K., 239. M. MAEY, 53, 56, 125, 150, 164, 168, 176, 178. Manchot, 267. Mannheim, 206. Margottet, 298, 313. Maronneau. 269. Martin, 256. Martins, 269. Masing, 123, 125, 331. Mathews, 155, 210. Mathewson. 107, 127, 129, 131, 132, 134, 136, 329, 331. Matthiessen, 64, 86, 87, 150, 176,206. Maxwell, 63. Mazzotto, D., 224, 233, 331. Mecklenburg, 289, 299, 309. Meerum-Teiwogt, 120. Mendelejeff, D., 26, 34, 38. Merz, 154, 164, 175. Meyer, E. v., 24. Meyer, L., 34. Miller, 100, 169. Minozzi, 303. Miolati, A., 4. Moh, 88, 90, 93. Moissan, H., 71, 112, 207, 208, 251, 254, 255, 256, 258, 261, 262, 267, 302. Monkmeyer, 194. Mourlot, 287. Muthmann, 117, 119, 206. Myers, 207. Mylius, 154. N. NAGAOKA, 71. Nasini, K, 25, 32. Nernst, W., 38, 49, 58, 63, 73, 75, 101. Neville, 46, 100, 162, 165, 167, 168, 173, 175, 176, 194, 329, 330, 331, 332. 0. OESTERHELD, G., 145, 329. Ogg, 164. Ormaridy, 204. Ostwald, W., 94. P. PARKINSON, 269, 276. Parravano, 103, 120, 143, 149, 158, 179, 195, 245, 278, 301, 302, 318. Partheil, 206, 278, Paschsky, 233. Paulo vitch, P., 52, 204. P61abon, 287, 288, 298, 300, 301, 305, 310, 311, 312. Pellini, 102, 103, 297, 298, 304, 305, 307, 324, 332. Ferret, 179, 245. 336 INDEX OF AUTHOES. Petrenko, 100, 162, 163, 165, 167, 168, 329. Planck, 62. Plato, W., 47. Podkapajeff, 238. Pope, 101. Portevin, 246. Prelinger, 0., 207. Proust, 25, 31, 34. Pushin, 36, 73, 80, 88, 105, 131, 140, 151, 168, 204, 205, 218, 233, 236, 304, 329, 330, 331, 332. Q. QUASEBART, 330. Quercigh, 102, 163, 304, 305, 329. K. KAMANN, 208. Rammelsberg, 100, 164, 175. Ramsay, 203, 204, 207, 278. Rausch, 36. Raydt, 196, 248. Rayleigli, Lord, 65. Regnault, 59, 154. Reinders, W., 73, 75, 79, 223, 332. Renault, 204, 269. Richards, 204. Richarz, 72. Ringer, 292, 332. Rio, 298. Rio, O., 332. Roberts-Austen, 1, 148, 156, 329. Roche, 210. Rdssler, 287, 296, 303, 313. Rolla, L., 218. Romanoff, 5, 330, 331. Roozeboom, Bakhuis, 14, 23, 205, 256, 314. Rosenhain, 119. Rudolfi, E., 81, 82, 257. Ruer, 121, 122, 233, 329. Ruez, 123. Ruff, 256. Rydberg, J. R., 88. Rykovskoff, 330, 331. S. SACKUR, 151. Sahmen, 87, 147, 153, 330, 331. Saklatvalla, 271. Saldau, 173. Sander, W., 241. Schepeleff, 272. Schimpff, H., 59. Schneider, 296. Schoen, 279. Schonbein, 206. Schreinemakers, 318. Schrotter, 268, 269, 275. Schubel, F., 59. Schiiller, 129. Schiirger, J., 112. Schiiz, 122. Schulze, F. A., 84. Schumann, 164, 204, 207, 208. Shepherd, 149, 156, 194, 329. Siebeck, 81. Sirovich, 318. Smirnoff, 88, 94. Smith, 131, 137, 138, 141, 142, 143, 329, 331. Speroni, G., 110, 112. Spring, W., 5, 330, 331. Stadeler, 255. Staigmiiller, 39. Stanfield, 156. Stead, 269. Stepanoff, 66, 184, 186. Stevanovitch, 101. Stockem, 330. Stoffel, 200, 224, 233, 330. Stokes, 314. Stortenbecker, 120. Suchein, A., 80. Svedelius, 87. T. TAFEL, 149. Tamaru, 331, 332. Tammann, G., 5, 7, 10, 13, 34, 36, 37, 46, 47. 70, 87, 91, 93, 117, 120, 121, 122, 123, 125, 196, 199, 202, 203, 206, 210, 222, 223, 226, 246, 248, 258, 263, 265, 293, 329, 330, 331. Tararin, 221. Tarugi, 204. Tavanti, 156. Thiel, 288, 298. Thompson, 169. Thomson, J. J., 72, 84. Tibbals, 313. Tissier, 204. Tivoli, 276. Trabacchi, 313. Traubenberg, 36. Treitschke, 200, 246, 293. Troost, 252, 255. Trouton, 49. Tukes, 119. U. URASOFF, 147, 170, 330, 331. INDEX OF AUTHOKS. 337 v. VANSTONE, E., 129. Vegesack, v., 196, 332. Vigier, 269. Vigouroux, 257, 258, 262, 267. Vincent, 206. Viviani, 120, 158. Vogel, 100, 118, 119, 170, 173, 175, 178, 179, 184, 208, 231, 233, 259, 329, 331. Voigt, W., 84. Vortmann, G., 206, 278. Voss, 188, 199, 202, 228, 246, 331. W. WAHL, 329. Wald, F., 24, 32. Wanklin, 111. Wartenberg, v. H.,'48, 49. Weber, C. L., 81, 150. Wedekind, E., 72, 244, 260. Weith, 154, 164, 175, 244. Welsbach, Auer v., 233. Werner, A., 34. Wibaut, J. P., 12. Wiedemann, G., 84. Wilke-Dorfurt, L., 285, 286. Williams, 219, 223, 224, 243, 255, 330, 332. Wilm, T., 175. Wilson, 204. Winkler, C., 206, 258. Winter, 125. Wohler, 258. Wologdine, 196, 272. Wright, 210. Wiinsche, H., 71. ZAMBONI, 208. Zemczuzny, 66, 88, 160, 194, 221 226, 270, 272. 277, 329, 330, 331, Zettel, 262. Ziegler, 293. Zukowski, G., 35, 125, 145, 329. CM. 22 INDEX OF BINARY SYSTEMS. Ag-AI, 165. Au-As, 276. Ca-Cd, 115. Cr-Co, 247. Ag-As, 276. Au-Cd, 173. Ca-Cu, 151. Cr-Fe, 246. Ag-Ca, 161. Au-Hg, 174. Ca-Hg, 112. Cr-Hg, 206. Ag-Cd, 163. Au-Mg, 170. Ca-Mg, 113. Cr-Ni, 247. Ag-Hg, 164. Au-Mn, 179. Ca-Pb, 193. Cr-P, 269. Ag-Mg, 159. Au-Na, 106. Ca-Sb, 194, Cr-Se, 302. Ag-Mn, 168. Au-P, 268. Ca-Si, 258 Cs-Hg, 145. Ag-P, 268. Au-Pb, 177. Ca-Sn, 1G2. Cs-S, 286. Ag-Pt. 169. Au-S, 287. Ca-Tl, 191. Cu-Al, 154. Ag-S, 287. Au-Sb, 179. Ca-Zn, 114. CU-As, 275. Ag-Sb, 168. Au-Sn, 175. Cd-Ag, 163. Cu-Ca, 151. Ag-Se, 297. Au-Te, 305. Cd-As, 277. Cu-Cd, 153. Ag-Sn, 167. Au-Zn, 171. Cd-Au, 173. Cu-Hg, 154. Ag-Te, 304. Cd-Ca, 115. Cu-Mg, 147. Ag-Zn, 162 B-C, 254. Cd-Co, 202. Cu-P, 267. Al-Ag, 165. B-Fe, 252. Cd-Cr, 201. Cu-S, 286. Al-Au, 175. B-Ni, 252. Cd-Cu, 153. Cu-Sb, 158. Al-C, 254. Ba-Hg, 112. Cd-Fe, 202. Cu-Se, 296. Al-Ca, 190. Ba-Si, 258. Cd-Li, 123. Cu-Si, 256. Al-Ce, 208. Be-Cu, 145. Cd-Mg, 109. Cu-Sn, 156. Al-Co, 213. Be-Fe, 180. Cd-Na, 129. Cu-Te, 303. Al-Cr, 217. Bi-Ca, 194. Cd-Ni, 202. Cu-Zn, 156. Al-Cu, 154. Bi-Ce, 231. Cd-P, 269. Al-Hg, 203. Bi-K, 143. Cd-Sb, 200. Fe-As, 280. Al-La, 117. Bi-Mg, 187. Cd-Se, 298. Fe-B, 252. AI-Mg, 181. Bi-Mn, 244. Cd-Sn, 200. Fe-Be, 180. Al-Mn, 211. Bi-Na, 136. Cd-Te, 306. Fe-C, 256. Al-Ni, 216. Bi-Ni, 246. Ce-Al, 208. Fe-Cd, 202. Al-Sb, 210. Bi-P, 269. Ce-Bi, 231. Fe-Co, 121. Al-Zn, 194. Bi-S, 287. Ce-Mg, 184. Fe-Cr, 246. As-Ag, 276. Bi-Se, 302. Ce-Pb, 119. Fe-Hg, 208. As-Au, 276. Bi-Te, 312. Ce-Si, 259. Fe-Mo, 207. As-Cd, 277. Bi-Tl, 220. Ce-Sn, 117. Fe-Ni, 122. As-Co, 281. Co-Al, 213. Fa-P, 271. As-Cu, 275. C-A1, 254. Co- As, 281. Fe-S, 293. As-Fe, 280. C-B, 254. Co-Cd, 202. Fe-Sb, 237. As-Hg, 278. C-Cr, 255. Co-Cr, 247. Fe-Se, 303. As-Mg, 276. C-Fe, 256. Co-Fe, 121. Fe-Si, 263. As-Mn, 279. C-Mn, 255 Co-Mo, 248. Fe-Sn, 226. As-Ni, 282. C-Mo, 255. Co-P, 272. Fe-Tc, 313. As-Pb, 278. C-Ni, 256. Co-S, 295. Fe-Zn, 196. As-Pt, 283. C-Ti, 255. Co-Sb, 238. As-S, 290. C-U, 255. Co-Se, 303. As-Sn, 278. C-V, 255. Co-Si, 284. Ga-Hg, 204. As-Te, 310. C-W, 255. Co-Sn, 226. As-Tl, 278. Ca-Ag, 161. Co-Te, 313. Hg-Ag, 164. As-Zn, 276. Ca-Al, 190. Co-Zn, 197. Hg-Al, 203. Au-Al, 175. Ca-Bi, 194. Cr-A!, 217. Hg-As, 278. 840 INDEX OF BINARY SYSTEMS. Hg-Au, 174. Mg-P, 269. P-Ag, 286. S-Bi,i287. Hg-Ba, 112. Mg-Pb, 185. P-Au, 268. S-Co, 295. Hg-Ca, 112. Mg-Sb, 186. P-Bi, 269. S-Cs, 286. Hg-Cr, 206. Mg-Si, 258. P-Cd, 269. S-Cu, 286. Hg-Cs, 145. Mg-Sn, 184. P-Co, 272. S-Fe, 293. Hg-Cu, 154. Mg-Tl, 182. P-Cr, 269. S-In, 288. Hg-Fe, 208. Mg-Zn, 108. P-Cu, 267. S-Mn, 292. Hg-Ga, 204. Mn-Ag, 168. P-Fe, 271. S-Mo, 292. Hg-In, 204. Mn-Al, 211. P-Hg, 269. S-Ni, 296. Hg-K, 139. Mn-As, 279. P-Ir, 275. S-Pb, 290. Hg-Li, 125. Mn-Au, 179. P-Mg, 269. S-Pd, 296. Hg-Mg, 110. Mn-Bi,244. P-Mn, 270. S-Rb, 284. Hg-Mn, 207. Mn-C, 255. P-Ni, 273. S-Se, 292. Hg-Mo, 207. Mn-Hg, 207. P-Pd, 275. S-Sn, 289. Hg-Na, 129. Mn-P, 270. P-Pt, 275. S-T1, 288. Hg-P, 269. Mn-S, 292. P-Sn, 269. Sb-Ag, 168. Hg-Pt, 208. Mn-Sb, 236. P-W, 270. Sb-Al, 210. Hg-Rb, 144. Mn-Se, 303. P-Zn, 269. Sb-Au, 179. Hg-Sb, 206. Mn-Si, 262. Pb-As, 278. Sb-Ca, 194. Hg-Se, 298. Mn-Sn, 224. Pb-Au, 177. Sb-Cd, 200. Hg-Sn, 205. Mn-Zn, 195. Pb-Ca, 193. Sb-Co, 238. Hg-Sr, 112. Mo-C, 255. Pb-Ce, 119. Sb-Cr, 243. Hg-Te,!307. Mo-Co, 248. Pb-K, 142. Sb-Cu, 158. Hg-Tl, 204. Mo-Fe, 248. Pb-Mg, 185. Sb-Fe, 237. Mo-Hg, 207. Pb-Na, 134. Sb-Hg, 206. In-Hg, 204. Mo-Ni, 249. Pb-Pd, 233. Sb-K, 143. In-S, 288. Mo-S, 292. Pb-Pt, 235. Sb-Mn, 236. In-Se, 298. Mo-Si, 282. Pb-S, 290. Sb-Na, 135. In-Te, 308. Pb-Se, 300. Sb-Ni, 239. Ir-P, 275. Na-Au, 106. Pb-Te, 310. Sb-Pd, 240. Na-Bi, 136. Pb-Tl, 218. Sb-Pt, 242. K-Bi, 143. Na-Cd, 129. Pd-P, 275. Sb-Se, 301. K-Cd, 138. Na-Hg, 129. Pd-Pb, 233. Sb-Sn, 223. K-Hg, 139. Na-K, 105. Pd-S, 296. Sb-Te, 311. K-Na, 105. Na-Pb, 134. Pd-Sb, 240. Sb-Tl, 219. K-Pb, 142. Na-Sb, 135. Pd-Se, 303. Sb-Zn, 194. K-Sb, 143. Na-Sn, 132. Pd-Si, 267. Se-Ag, 297. K-Sn, 141. Na-Tl, 131. Pt-Ag, 169. Se-Bi, 302. K-T1, 140. Na-Zn, 127. Pt-As, 283. Se-Cd, 298. K-Zn, 137. Ni-Al, 216. Pt-Hg, 208. Se-Co, 303. Ni-As, 282. Pt-P, 275. Se-Cr, 302. La-AI, 117. Ni-B, 252. Pt-Pb, 235. Se-Cu, 296. Li-Cd, 123. Ni-Bi, 246. Pt-Se, 303. Se-Fe, 303. Li-Si, 256. Ni-C, 256. Pt-Sb, 242. Se-Hg, 298. Li-3n, 125. Ni-Cd, 202. Pt-Si, 267. Se-In, 298. Ni-Cr, 247. Pt-Sn, 230. Se-Mn, 303. Mg-Ag, 159. Ni-Fe, 122. Pt-Te, 313. Se-Ni, 303. Mg-Al, 181. Ni-Mg, 188. Pt-Tl, 222. Se-Pb, 300. Mg-As, 276. Ni-Mo, 249. Se-Pd, 303. Mg-Au, 170. Mg-Bi, 187. Ni-P, 273. Ni-S, 296. Rb-Hg, 144. tth Q 9QJ. Se-Pt, 303. Se-S, 292. Mg-Ca, 113. Mg-Cd, 109. Ni-Sb, 239. Ni-Se, 303. JriU-o, Aot. Ru-Si, 267. Se-Sb, 301. Se-Sn, 299. Mg-Ce, 184. Ni-Si, 265. Se-Tl, 298. Mg-Cu, 147. Ni-Sn, 228. S-Ag, 287. Se-Zn, 298. Mg-Hg, 110. Ni-Te, 313. S-As, 290. Si-Ba, 258. Mg-Ni, 188. Ni-Zn, 198. S-Au, 287. Si-Ca, 258. INDEX OF BINARY SYSTEMS. 341 Si-Ce, 259. Si-Co, 264. Si-Cr, 262. Si-Cu, 256. Si-Fe, 263. Si-Li, 256. Si-Mg, 258. Si-Mn, 262. Si-Mo, 262. Si-Ni, 265. Si-Pd, 267. Si-Pt, 267. Si-Ru, 267. Si-Sr, 258. Si-Ta, 262. Si-Th, 260. Si-Ti, 260. Si-U, 262. Si-V, 261. Si-W, 262. Si-Zr, 260. Sn-Ag, 167. Sn-As, 278. Sn-Au, 175. Sn-Ca, 192. Sn-Cd, 200. Sn-Ce, 117. Sn-Co, 226. Sn-Cu, 156. Sn-Fe, 226. Sn-Hg, 205. Sn-K, 141. Sn-Li, 125. Sn-Mg, 184. Sn-Mn, 224. Sn-Na, 132. Sn-Ni, 228. Sn-P, 269. Sn-Pt, 230. Sn-S, 289. Sn-Sb, 223. Sn-Se, 299. Sn-Te, 309. Sr-Hg, 112. Sr-Si, 258. Ta-Si, 262. Te-Ag, 304. Te-As, 310. Te-Au, 305. Te-Bi, 312. Te-Cd, 306. Te-Co, 313. Te-Cti, 303. Te-Fe, 313. Te-Hg, 307. Te-In, 308. Te-Ni, 313. Te-Pb, 310. Te-Pt, 313. Te-Sb, 311. Te-Sn, 309. Te-Tl, 308. Te-Zn, 306. Th-Si, 260. Ti-C, 255. Tl-As, 278. Tl-Bi, 220. Tl-Ca, 191. Tl-Hg, 204. Tl-K, 140. Tl-Mg, 182. Tl-Na, 131. Tl-Pb, 218. Tl-Pt, 222. Tl-S, 288. Tl-Sb, 219. Tl-Se, 298. Tl-Te, 308. U-C, 255. U-Si, 262. V-C, 255. V-Si, 261. W-C, 255. W-Si, 262. Zn-Ag, 162. Zn-Al, 194. Zn-As, 276. Zn-Au, 171. Zn-Ca, 114. Zn-Co, 197. Zn-Cu, 148. Zn-Fe, 196. Zn-K, 137. Zn-Mg, 108. Zn-Mn, 195. Zn-Na, 127. Zn-Ni, 198. Zn-P, 269. Zn-Sb, 194. Zn-Se, 298. Zn-Te, 306. Zr-Si, 260. C.M THK WHITEFRIARS PRESS, LTD., LONDON AND TONBR1PGK. 23 -,1 *- _ -ss^ 33 YC 2 1 "724 UNIVERSITY OF CALIFORNIA LIBRARY %&$&fm'- lifli