:j THE PHYSICAL CHEMISTRY OF THE METALS BY RUDOLPH SCHENCK J Professor of Physical Chemistry in the Technischen Hochschule in Aachen TRANSLATED AND ANNOTATED BY REGINALD SCOTT DEAN Research Metallurgist, American Zinc, Lead and Smelling Co., St. Louis, Mo. FIRST EDITION NEW YORK JOHN WILEY & SONS, INC. LONDON: CHAPMAN & HALL, LIMITED 1919 COPYRIGHT, 1919, BY REGINALD S. DEAN PRESS OF BRAUNWORTH & CO. BOOK MANUFACTURERS BROOKLYN, N. Y, PREFACE THERE is perhaps no field where physical chemistry is of more value than in the field of metallurgy and metallography. In the latter, field its usefulness has been recognized almost from the first but in the field of metallurgy its application has not been so general. It is hoped that this translation may aid in making the value of chemical dynamics and equilibrium clear to the metallurgist and metallurgical student. Such additions as have seemed necessary have been incor- porated in the text and the numerical data have been revised to agree with the accepted values. I have deemed it advisable not to go into the recent investigations concerning the electron theory since the scope of the work did not seem to warrant an extended treatise on this subject. The book has been changed from lecture to text-book form and the references shifted from the appendix to the body of the book. My thanks are due to Mr. A. T. McPherson of the U. S. Bureau of Standards for reading the manuscript as well as for many valuable suggestions. I wish also to thank Dr. Edward Schramm, director of this laboratory, for his .encouragement and cooperation in the work of translation. REGINALD S, DEAN. ST. Louis, Mo. March, 1919. 111 PREFACE TO THE GERMAN EDITION THIS little book is the outcome of a series of lectures which I delivered in 1907, in the " Technischen Hochschule," at Aachen. Their purpose was to show the engineers of the Rhenish indus- trial district, before whom they were delivered, the use of chem- ical statics and to deepen their understanding of smelting opera- tions and metallurgical processes. I have endeavored, especially, to develop the principles of equilibrium clearly and so far as possible by the use of per- tinent examples. In the systematic survey of such a field new problems nat- urally arise and there is found in these lectures some heretofore unpublished data bearing on these problems, among which may be mentioned: the equilibrium between the various components of steel, the quantitative determination of amorphous carbon and graphite, and the investigation of the sulfatizing roast. It has not been possible to consider here, all of the experi- mental matter relating to the physical chemistry of metals, but all fundamental questions have been treated rather thoroughly. I wish to thank Dr. P. Goerens, for the preparation of the metallographs and photographs; Dr. Hemplemann, for his assistance in the preparation of the index, and Dr. Ratzbach, for the preparation of the diagrams. THE AUTHOR. AACHEN. July, 1908. CONTENTS PREFACE iii PREFACE TO GERMAN EDITION v CHAPTER I INTRODUCTION, PROPERTIES OF METALS 1-38 Introduction, i. Form Changes, 3. Vapor Pressure and Volatility, 3. Monatomic State of Metal Vapors, 5. Color of Metal Vapors, 6. Vapor Pressure and the Definition of Melting Point, 7. Melting Points of Metals, 8. Density Change on Melting, 9. Heat of Fusion, 9. Cooling Curves, 10. Polymorphy, n. Analogy of Polymorphic Transition to Melting, 14. Determination of the Transition Point, 14. Explosive Antimony, 17. Enantiotropy and Monotropy, 18. Crystal Growth, 20. Electrical and Optical Properties of Metals, Conductivity, 21. Faraday's Law, 22. Metallic and Gaseous Conduction, 22. Electron Theory, 23. Law of Wiedemann and Franz, 23. Optical Properties of Metals; Light Reflection and Absorption, 27. Metallic Luster, 28. Electron Concentration of Metals, 30. Temperature Coefficient of Conductivity, 31. Thermo-Electric Force, 32. Passivity in Metals, 34. CHAPTER II METALLIC SOLUTIONS AND ALLOYS 39-81 Colloidal Metal Solutions, 39. Dilute Metallic Solutions, 41. Vapor Pressure of Metallic Solutions, 42. Freezing Point of Alloys, 44. Metals only Partially Miscible in the Liquid State, 46. The Parkes .Process, 47. Solidification Curve of Binary Alloys, the System Cadmium- Zinc, 49. The Pattison Process, 52. Segregation in Alloys, 52. Investi- gation of Alloy Structure, 53. Solid Solutions, 55. The Heraeus Process of Plating Platinum with Gold, 57. Crystallization Diagrams for Metals Forming Solid Solutions, 57. Inter-metallic Compounds, 60. Ternary and Quaternary Alloys, 64. Relations of Mechanical and Physical Properties to Alloy Structure, 68. Bearing Metals, 69. Den- sity of Alloys, 70. Potential of Alloys, 72. Electrical Resistance of Alloys, 73. Thermo-electric Force of Alloys, 80. vii viii CONTENTS CHAPTER III PAGE ALLOYS OF METALS WITH CARBIDES, OXIDES AND SULFIDES, IRON AND STEEL, MATTES, PHASE RULE ' 82-138 Compounds with Metallic Properties, 82. The Iron Carbon Alloys, 83. The Crystallization of the System Iron Cementite, 85. Perlite and its Transition Products, 93. The Precipitation of Carbon in Iron Carbon Alloys, 98. The Uses of Additions to Iron and Steel, 104. Alloys of Metals and Oxides; Copper Copper Oxide, no. Silver Silver Oxide, 113. Alloys of Metals and Sulfides, 114. Lead Lead Sulfide, 115. Antimony Antimony Sulfide, 115. Copper Copper Sulfide, 116. Silver Silver Sulfide, 118. Iron Iron Sulfide, 120. Nickel Nickel Sulfide, 122. Alloys between Sulfides, 123. Phosphorus and Arsenic Containing Alloys, 125. Silicides of Metals, 131. The Phase Rule, 133. CHAPTER IV THE METALLURGICAL REACTIONS, OXIDATION AND REDUCTION 139-169 Equilibrium between Metal, Oxide and Oxygen, 140. The Applica- tion of the Phase Rule, 140. Oxygen Tension of Oxides, 141. Oxygen Tension in Atmospheres, 143. The Equilibrium between two Oxides and Oxygen, 145. Direct Decomposition of Oxides by Heat, 148. Reduction by Metals, 149. Reduction by Gaseous Reducing Agents, 149. Reduction by Hydrogen, 150. The Phase Rule, 150. LeChatelier's Principle, 151. The Mass Law, 152. Van't Hoff's Equation, 156. Reduction by Carbon Monoxide, 159. CHAPTER V DECOMPOSITION OF CARBON MONOXIDE, BLAST FURNACE PROCESS 170-204 Catalytic Decomposition of Carbon Monoxide, 1 70. Equilibrium between Carbon Monoxide and Iron, 173. The Analysis of Mixtures of Graphite and Amorphous Carbon, 178. The Blast Furnace Process, 180. The Mond Nickel Process, 202. CHAPTER VI THE REACTIONS OF SULFIDES 205-227 THE PHYSICAL CHEMISTRY OF THE METALS CHAPTER I INTRODUCTION, PROPERTIES OF METALS IT is a well-known fact that the metal industry is based largely on chemical knowledge. The metals were the first chemically defined substances to be recognized, and the first to be obtained by chemical means; it was thus that man acquired the beginning of his hard-earned chemical knowledge. What a heritage of facts has come to the present day chemist simply through the efforts of the alchemist to change base metals into noble. Interest in the chemistry of metals has never lagged; the mines and the pits of the furnaces were to Berzelius and many other chemists the inspiration to notable new investi- gations; the application of new agents, particularly the voltaic battery, in the electric decomposition of salts, have brought to light new metals with remarkable properties, for example, sodium, potassium, aluminium, magnesium and many more. The names of Davy, Wohler, and Winkler recall to our minds what great advances the chemistry of metals has made in the past century. These include the discovery not only of new metals but also of the methods by which mineral substances are recognized and their amounts determined with certainty. Technical ch.emists, especially metallurgists, are already convinced of the value of chemical analysis, not only in the in- vestigation of ores, but in the control of processes. Experience has shown that the success or failure of a smelter, may, under 2 THE PHYSICAL CHEMISTRY OF THE METALS certain conditions, depend on the reliability of its analyst. The introduction of chemical analysis into smelters was the first great advance toward rational operation. In the course of the last century the machine industry and the electrical industry have made unheard-of progress; the metal- lurgical industries, especially that of iron and steel, are now growing into a sphere of new problems. The purchaser demands not only large amounts of material at the lowest possible price, but he also places entirely new and definite demands on the quality of the same. Bridge and machine builders need steel of definite mechanical properties, the electrical industry needs steel of definite magnetic properties, the tool manufacturer, a steel that retains its hardness when hot, ships and machines, needles and pins, wire and sheet, cook stoves and furnaces all are made of iron and each one requires an entirely different mate- rial. Iron must play the part of a veritable Proteus to possess all these properties which are often diametrically opposite. It is astounding that one metal is endowed with so great a varia- tion in its properties and it is remarkable that the industries have found a use for this material in each of its notable properties. The growth of our metal industries is largely due to machine and bridge building, since the carrying out of a large number of these projects required appropriate material and the smelter- man was thereby directed to experiments and tests for their production. Many problems in metallurgy cannot be solved by means of analytical chemistry, since it is often found that material of basically different properties gives identically the same analysis and that extraordinary relations in physical properties are afforded no explanation by the analysis. Structural chemistry, as developed for organic compounds, also fails to be of value here, and accordingly the only resort was the purely empirical. Every empirical process that we know works uneconomically; a success follows ten failures; every suc- cessful advance is costly in experience and money. All the phases of metallurgy have not yet been placed on a scientific basis, but some of the obscurities have been disposed INTRODUCTION PROPERTIES OF METALS 3 of and progress made in the solution of some of the problems presented by the smelting art. This has been accomplished by means of physical chemistry; especially through chemical mechanics which for metallurgical engineers and technical chem- ists possesses the same importance that physical mechanics does for bridge and machine builders. We will now attempt with the examples at hand to present a picture of the manner in which the problems of metallurgy are solved with the help of physical chemistry. We will not consider the electrochemical relations of the metals as electro-metallurgy has long since left the scientific nursery, gone out into life and proved its worth. We may not, however, overlook the youngest branch of electrical science, the electron theory of metals, since it explains many of the remarkable physical peculiarities of the metallic state. Form Changes. We must first consider certain physical phenomena which the metals show in common with non-metallic substances but which play an especially important role in the working of the metals. They are the form changes which the metals undergo at high temperatures, the phenomena of vaporization, of melting, and the polymorphic transitions in the solid state. For the present we shall limit ourselves to the pure metals, later taking up solu- tions and mixtures. Vapor Pressure and Volatility. A number of metals go over easily to the gaseous state. The volatility of mercury is detectable at atmospheric temperatures and cadmium and zinc are obtained by the use of a distillation process. The boiling points of these metals are relatively low. At high temperatures such as can be obtained in the oxy-hydrogen flame, the less volatile metals are vaporized. It is known that silver distils at such temperatures since Stas in his classical investigations on atomic weights purified silver in this way. At the highest temperature of the electric arc there is no metal that does not vaporize. 4 THE PHYSICAL CHEMISTRY OF THE METALS Moissan * succeeded with the help of a strong current in boiling and distilling most metals; copper and gold as well as platinum, palladium, iridium, rhodium, ruthenium, and osmium can be vaporized and precipitated again on a cooled metal sur- face. Of the iron group, manganese is known to be the most volatile, then nickel and then chromium. Iron itself boils with relative difficulty and the boiling points of uranium, molybdenum and tungsten are still higher but they are undoubtedly volatile in an electric arc. Many metals that are difficultly vaporized at ordinary pres- sures are easily brought to vaporization and distillation under strongly reduced pressure. Krafft and Kahlbaum f and Roth and Siedler J have made investigations of this kind. The latter have purified gold, silver, copper, lead, bismuth and antimony by distillation in the highest obtainable vacuum and investi- gated a series of properties of the metals so purified. The following table gives the approximate values for the boil- ing points of metals so far as they have been determined. The values are those given by J. Johnston, J. Ind. Eng. Chem. 9, 873 (1917), unless otherwise noted. Metal. Boiling Point.* Degrees C Metal. Boiling Point.* Degrees C Cadmium ?8o Chromium 22OO Zinc Q2O Tin 226o Magnesium 1 1 2O Copper . 2? CO Thallium . 1 3OO Nickel 2J.OO Bismuth I4.4.O Iron 24.C.O Antimony I44O Arsenic 610 Lead 1640 Platinum f . . . 3QO7 Aluminum Manganese I800 IOOO Molybdenum f Tungsten t 3617 4827 Silver 2OQO Sodium J 742 Mercurv 1. . 3C7. 2C * i.e., the temperature at which the vapor pressure is 760 mm. t Langmuir, Phys. Rev. 2, 329 (1913). j From Schenck's table. * Compt. rend., 141, 853: 977 (1905). t Ber. 36, 1690 (1903). JZ. anorg. Chem., 29, 177 (1902). 142, 189; 425; 673 (1906). INTRODUCTION PROPERTIES OF METALS For volatile substances every temperature corresponds to a definite vapor pressure which increases with rising temperature. The relation between the two has been investigated by Barus and others. The approximate boiling temperatures of metals at various pressures, i.e., the temperature at which their vapor pressure is io~ 3 mm., etc., is given in the following table taken from Johnston. Metal. p IN MM. MERCURY. io- 3 io- 2 10 ' I 10 So 100 760 Cadmium . . C 220 2QO 380 500 540 540 62O 730 7QO Q20 980 IOIO 1080 IIOO 1130 220 C 270 35 440 570 620 620 710 830 890 1030 1090 1130 1200 I22O 1250 260 C 330 420 520 660 720 720 820 950 IO2O 1160 1230 1270 1340 1370 1400 310 c 410 500 620 770 840 840 960 1090 1170 1320 1400 1440 1520 1550 1590 360 C 500 610 750 910 990 990 1130 1280 1360 3520 1610 1660 1740 1780 1820 430 C 590 700 860 1030 1130 1130 1290 1440 1530 1700 1800 1850 1930 1970 2OIO 490 C 630 750 920 1090 1 200 1200 1360 1520 1610 1780 1890 1940 2030 2070 2IIO 5!0 C 7 80 920 1120 1300 1440 1440 1640 I800 1900 2O9O 2200 2260 2350 2400 2450 610 Zinc Magnesium . . . Thallium Bismuth Antimony . Lead Aluminum Manganese Silver Chromium Tin CoDoer Nickel Iron Arsenic If the relation of the vapor pressure to the temperature for any metal be represented graphically, a curve is obtained which corresponds to the vapor pressure curve of non-metallic sub- stances, for example, water. (Fig. i.) Monatomic State of Metal Vapors. In calculating the molecular weight from the vapor density of metals it is seen that the molecules consist of only one atom, differing in this respect from most other substances. The monatomic state of the vapor has been established not only for the easily vaporized metals, mercury, cadmium and zinc (Victor Meyer) * but also for the difficultly vaporizable metals * Ber. 12, 1426 (1879). 6 THE PHYSICAL CHEMISTRY OF THE METALS bismuth, antimony, lead, and thallium (von Wartenberg.*) The monatomic state is also revealed by another circumstance. Gases have recognizedly two different specific heats. The amount of heat which we must bring into a gas in a closed constant volume to raise it to a higher temperature is smaller than that required to accomplish the heating at constant pres- sure. In the latter case the gas increases in volume and thereby performs work. It can be deduced from the kinetic theory 800m m. 700 700 600 600' 500 500 400 400 300' 300 200 200 100 100 200 300 500 700 900 1100 1300 C Vapor Pressure of the Metals Hg , Cd > Zn and Bi FIG. i. 1500 c that for a monatomic gas the ratio of the specific heat at con- stant pressure to that at constant volume must have a definite value, namely 1.667. Kundt and Warburg f found the value 1.666 fpr mercury vapor. Besides the metals only the rare gases of the atmosphere, argon, xenon, helium, krypton, and neon are monatomic. The molecules of all other substances are polyatomic. Color of Metal Vapors. The vapors of the metals often show an intense coloring which is especially easy to observe with potassium vaporized * Z. anorg. Chem., 56, 320 (1907). f Pogg. Ann., 157, 353 (1878). INTRODUCTION PROPERTIES OF METALS in an atmosphere of hydrogen. The color of the vapor is in- tensely green. Sodium vapor is blue. It is the powerful absorption of these metal vapors which gives rise to the Frauen- hofer lines in the solar spectrum. Vapor Pressure and the Definition of Melting Point. We must ascribe to all substances, the capacity of sending off vapor molecules from their upper surfaces. That we cannot observe them is due to their smallness and the relative insensi- bility of our methods. Vaporization is not confined to liquids, for solids also vaporize and have a definite vapor tension. For the volatile solid substances as iodine and camphor its mag- nitude can be readily deter- mined. Substances like benzene and naphthalene whose vapor pressures can be readily meas- ured in both the solid and the liquid state have been investi- gated and the relation of vapor pressure to temperature estab- lished for both states. From this it has been found that each state has a special vapor pressure curve. It is seen from Fig. 2 that these curves must intersect once and the temperature of this intersec- tion is no other than the melt- ing point. The melting point of a substance may accordingly be defined as the temperature at which the vapor pressure of the solid substance is equal to that of the liquid. The melting point is an equilibrium point at which solid and liquid can exist together. If heat be added to a substance at its melting point the solid will disappear without change of temperature after which the temperature will rise. If heat be subtracted the liquid will solidify and the temperature will only fall after the entire mass is solid. It is sometimes Temp. M.P. FlG. 2. 8 THE PHYSICAL CHEMISTRY OF THE METALS possible, however, to cool a liquid below its melting point with- out the separation of solid taking place. Such liquids are said to be supercooled. The supercooled liquid has a vapor pressure curve which is a direct continuation of that part of the liquid curve which lies above the melting point. Its vapor pressure is always higher than that of the solid at the same temperature, crystallization is therefore accompanied by a decrease in vapor pressure. Above the melting point the liquid has a lower vapor pressure than the solid would possess if it could be observed there. The transition of liquid into solid, or conversely of solid into liquid always takes place in the direction which results in a decrease of vapor pressure. The state with the lower vapor pressure is always stable, that with the higher always metastable. At the point of equal vapor pressure, that is, the melting point, both states are equally stable and there is no tendency to change. Melting Points of Metals. The melting points vary greatly for different metals. The most recent values are given in the table below. [Mercury 'C ^8 8? Barium C 850 Caesium 26 Germanium 958 Gallium OQ I Silver 060 S Rubidium . . 38 Gold 1063 . o Potassium 62 3 Copper 1083.0 Sodium 07 "? Manganese 1230 Indium ICC Nickel 1452 1 86 Cobalt 14.80 Tin 231 Chromium 1615 Bismuth 271 Iron IC.3O Thallium 3O2 Palladium 1549 720 O Vanadium . . . 1720 Lead J27 4. Platinum 1755 Zinc 4.IQ 4. Titanium 1800 Antimony 6^O Uranium Caw \ \ \ ^ > C >=* rs \ N \ > ) 5 10 15 20, 25 30 35' 40 45 50 55 G( Seconds. FIG. 3. In all cases, however where a change of state is involved, the solidification point is marked in the cooling curve,* by means of which we can determine its position. This point has a special meaning in complex systems of mixtures and alloys as we shall see later. Polymorphy. Transitions may take place in the solid state ; for example, we know tin as a malleable silver white metal, but a remarkable * For the methods of determining cooling curves, see Burgess, Bulletin Bureau of Standards, 1908, 5, 199; also an excellent chapter in Deschs' "Metallography," p. 123. 12 THE PHYSICAL CHEMISTRY OF THE METALS change is sometimes observed, especially in cold regions. In the mass there appear here and there spots of a gray color; the metal falls to a brittle product which occupies a much greater volume than the material out of which it was formed. Due to this volume change there forms at the transition places " Pustules " that fall to pieces by touching. This transformation begins at one spot and travels out from it until it involves the entire vicinity and the metal breaks to a brittle powder. Where this Time FIG. 4. phenomenon once shows itself all tin is in danger. If a grain of the transition product be placed on the intact metal the transi- tion is brought on there. The metal is said to be " sick;" it becomes the tin " pest." This phenomenon was first observed in Russia, where, in a military magazine, a block of tin was found fallen entirely to a powder. The tin pest is also known in Germany; for example, the eaves on the Post building at Rothenberg are infected with it. INTRODUCTION PROPERTIES OF METALS 13 A number of chemists have investigated the tin pest; its com- plete explanation is due to Schaum* and Cohen, f The chemical analysis of the brittle powder, to which the white tin falls, shows nothing but metallic tin. The tin pest has thus caused no chemical change. It cannot accordingly be con- sidered otherwise than a new allotropic modification of tin, which, because of its gray color is designated as gray tin. The above-named investigators have now established the reciprocal relations of the two forms of tin. If white tin be pow- dered by allowing the molten metal to crystallize with strong FIG. 5. Tin Infected with the Tin "Pest." rubbing in a mortar and this powder mixed with some gray tin and placed in a Dewar vessel for several days at the tempera- ture of solid carbon dioxide and ether, the entire mass changes to gray tin. The reverse change of gray tin into white tin can be accomplished by warming the mass on a water bath. One may accordingly prepare at will, by simple temperature changes, either the white or the gray modification. The process, as we have seen, is entirely reversible. At low temperatures the gray is the stable form, at high temperatures the white. Between these temperatures there must be a point at which both * Ann. 308, 30 (1899). fZ. Physik. Chem., 30, 601 (1899). 14 THE PHYSICAL CHEMISTRY OF THE METALS forms are equally stable and are in equilibrium with each other. By experiments on the electro-motive force at various tempera- tures, of a galvanic cell, formed from the two modifications of tin as electrodes, in a stannous chloride solution as electrolyte, it is possible to determine the position of this equilibrium. At equilibrium the potential difference of the two modifications is nil. At high temperatures the gray modification is the pos- itive pole and the metastable modification, at low temperatures the white. The equilibrium temperature is about 20. Below this point a transformation occurs to the gray, above it, to the white. The two differ also in their crystalline form, the gray having a columnar structure. Analogy of Polymorphic Transition to Melting. Such a multiplicity of forms or polymorphy in the solid state is also found in other elements, e.g., sulfur, of which there are rhombic and monoclinic forms. For this element the poly- morphic transition and its peculiarities have been carefully studied. It has been shown that an extraordinary similarity exists between the process of polymorphic transition and the process of melting and solidification. The equilibrium point which is also designated as a transition point, can be compared in many ways to the melting point. The transition of one modification into the other is accompanied by a change of den- sity that may, in some cases, be very considerable. In the transi- tion of gray to white tin it increases from 5.85 to 7.30. As with melting the transition is accompanied by an absorption of heat. The so-called transition heat is however ordinarily less than the heat of fusion. The two processes differ in only one respect, namely, the low temperature form can frequently be observed above the transition point, while it is not possible to heat a solid substance above its melting point without its going over to a liquid. Determination of the Transition Point. The phenomena of volume change and transition heat can both be used to ascertain the position of the transition point. INTRODUCTION PROPERTIES OF METALS 15 It is of special importance, for example, with iron which has at least three polymorphic modifications sharply separated from each other by two transition points. We have the three forms a iron or ferrite, and |8 and 7 iron. The transition point a. > /? lies at 768 C., the second 0-> 7 at 898-909 C* The cooling curve of iron shows slight arrests at these tem- peratures which are brought about in precisely the same way as Temperature C. \ 1530 Solidi ficatk n Poii It \ S V Ir on i i \ / / / / \ /' / / AC> / j/905 898 / ^ 3 800 / / s \r on / Ac r\ 768 i \ Vr 2 // \ Oi Iror \ / / X X 10 20 30 40 50 60 7Q Minutes FIG. 6. 90 100 the points of arrest due to the solidification of a molten liquid. Fig. 6 shows a cooling curve for pure iron. The volume change of the transition process is the cause of the re-expansion of iron strips on cooling which has been observed by Kinder. The length change in a strip 90 meters long is about 200 mm. The change of 7 into /3 iron is accompanied by a volume increase, the 7 iron being denser than the /3, but the * Burgess and Crowe, Bur. Standards Bull., 10, 317 (1913). This paper also contains an excellent bibliography on the allotropy of iron. 16 THE PHYSICAL CHEMISTRY OF THE METALS change of the latter into the a modification gives again a denser product and a contraction of the piece.* The a iron possesses the property of magnetism which the other forms do not. This can be easily shown by a simple experiment (Fig. 7). If a small piece of wrought iron be sus- pended on a wire before a magnet it will naturally be drawn to the magnet, but if it be heated to a white heat, the action of the FIG. 7. magnet ceases. The attraction occurs again on cooling when the temperature has sunk below 768 C. the transition point for the reaction j8 > a. Accordingly if the /3 or the 7 were the stable form at ordinary temperatures our entire electrical industry in *The volume changes of the iron transitions have been studied by Charpy and Grenet, Bull. Soc. d'Encouragement, 104, 464 (1892); Broniewski, Compt. rend., p. 1983 (1913); Rosenhain and Humphrey, Proc. Roy. Soc., 83, 200 (1909); Le Chatelier, Compt. rend., 129, 279 (1899). INTRODUCTION PROPERTIES OF METALS 17 the form in which we now possess it would be impossible. Dyn- amo machines and electric motors could not be built if we did not possess a material so strongly magnetic as a iron. We also encounter polymorphy in other metals. The fol- lowing table from Guertler* gives the transition points of the ele- ments so far as they have been determined.! Metal. Transition Points, C. Iron . 808 768 Cobalt I I^O Nickel . . . 7 2O Zinc l6 iron. If this is exceeded the large a crys- tals fall to pieces to small crystals so that the structure again becomes finely crystalline. Electrical and Optical Properties of Metals : Conductivity. The previously considered phenomena are of an entirely general nature and are in no way peculiar to the metallic state. There is however a whole series of properties which are charac- teristic of the metals. The lustrous appearance and the ability to conduct heat and electricity distinguish the metals from all other substances. These properties of the elementary metals appear in a weak degree, in certain metallic compounds espe- cially the sulfides. For an insight into the nature of metals we are indebted to the views of physicists on the nature of elec- tricity, the so-called electron theory. Electrical conductivity is not characteristic of the metals alone but solutions of acids, bases and salts, as well as molten salts and certain solid oxides, such as constitute the Nernst glower, conduct the electric current. A theory has been devel- oped by which we are able to account for each of these things. The conduction of electricity by acids, bases and salts dis- solved in water has led to the theory that these substances are dissociated in solution into electrically charged split molecules. The products of the dissociation are ions, that is, electrically charged atoms or atom complexes. It follows from the dis- sociation process that there are as many positive as negative charges, the number of charges which an ion possesses expressing its valence. The positively charged ions migrate to the cathode and are called cathions, while the negative anions go to the anode. If a solution of an electrolyte be brought between two elec- 22 THE PHYSICAL CHEMISTRY OF THE METALS trodes and the circuit closed, the electrodes become charged, one with positive and the other with negative electricity. Under the influence of the charges, motion of the ions to the oppositely charged electrode results. On its arrival the electrically charged particle gives up its charge and becomes electrically neutral. From the silver ion metallic silver results, from the chlorine ion, chlorine gas, etc. The passage of the current, therefore, involves chemical changes, the electrolyte being decomposed at the elec- trodes. The transport of electricity through the liquid is accomplished by ions. The displacement of the electrical charges bound to the material particles is greater, the stronger the current. The con- ductivity is greater the greater the concentration of ions and the smaller the frictional resistance which their motion inside the liquid encounters. The movement of the cathion and the anion in the same electrolyte is generally unequal, and as a result of such migration differences, concentration displacements occur in the vicinity of the electrodes. There is an increase in the concentration at the electrode, of the ion which moves in that direction and a decrease in the concentration of the other ion. Faraday's Law. The amount of the electrode discharges and of substance pre- cipitated depend on the current strength. Faraday has shown that by passing the same current through different cells, the amounts of different substances precipitated are in the ratio of their equivalent weights. Equivalent amounts of ions carry equal amounts of electricity. The gram equivalent carries 96,540 coulombs, a current of i ampere must flow 96,540 seconds or 26.8 hours to precipitate 108 gm. of silver, i.oi gm. of hydrogen or 31.8 gm. of copper. Metallic and Gaseous Conduction. In the conduction of electricity through metals there is no displacement by the current. The conductor remains abso- lutely unaltered. In this circumstance the metallic conductors do not stand alone, a similar relation existing in the conduction INTRODUCTION PROPERTIES OF METALS 23 of current by dilute gases under the action of cathode rays and /8 Bequerel rays sent out by radioactive substances. These rays consist of negatively charged electrical particles streaming with great velocity, which can be deflected from their path by electro- static and electro-magnetic influences. From the magnitude of the electro-static and electro-magnetic deflection both the velocity of the charged particles and their mass, which is the unit of electrical transportation, can be calculated. Electron Theory. For the mass we get a very small value. One gram of sub- stance will transport i .88 X io 8 coulombs. One gram of hydrogen, the lightest of all known substances, carries in the ionic state, 96,540 coulombs. It is calculated therefrom that the equivalent weight of the ray substance is 0.000513. In grams this is the mass of cathode and /3 Bequerel rays which will be associated with 96,540 coulombs. Other phenomena have also been used to determine this magnitude and give about the same value. These carriers of negative charges which are in round numbers 2000 times as light as hydrogen particles, are now generally desig- nated as electrons. It remains to be pointed out that such small carriers of positive charges are not known. The electrons also cause current conduction inside the metals according to the new view. The electron theory of metals that has been evolved by J. J. Thomson and P. Drude * supposes that within the metal a dissociation is undergone into positive metal ions and negative electrons, and that the negative par- ticles, under the influence of the potential difference attached to the metal, fall to streaming in the same way as the electrolytic ions in an electrolytic cell. Law of Wiedemann and Franz. Wiedemann and Franz have noted a remarkable parallelism between the electrical and heat conductivity of metals. The good conductors of electricity being also good conductors of * Ann. Physik., 1, 566 (1900); 8,369(1900); 7, 687 (1902); 14,936 (1904); Physik. Z., 1, 161 (1900). 24 THE PHYSICAL CHEMISTRY OF THE METALS heat and if the ratio of the heat conductivity K to the elec- trical conductivity a is calculated it is shown that for the same temperature this ratio possesses a large value which varies within narrow limits. This fact is shown in the following table from the experiments of Jaeger and Diesselhorst:* Metal. - for 18. ff * for 100. _K_ . K al8 ' trioo* Aluminum 636 844 I. 12 Copper. . . 665 862 . 7Q Silver 686 881 28 Gold 727 Q2v s x X s x X x x Au 1023 FIG. 21. curves for Bi-Sb alloys. Curve Bi-A-Sb is the line for the beginning of melting, curve Bi-C-B-Sb is the line for the begin- ning of solidification. The two curves enclose a field outside of which solutions in either the liquid and the solid state can exist. In the gap only heterogenous mixtures of solid and liquid solu- Au 1060 i ==; 1 * 1040 v^ X 1020 \ 1000 V \ 980 ! \K 10 20 30 40 50 60 70 80 90 100 FlG. 22. 960 tions are possible. As the temperature is lowered the first crystals to precipitate are rich in antimony but as the tempera- ture is lowered the bismuth content grows steadily. The com- pletely solidified alloy is accordingly not homogenous, but con- sists of a stratification of crystals of different composition. In METALLIC SOLUTIONS AND ALLOYS 59 some cases it is possible to make the alloys homogenous, by sup- plementary heating, at a temperature near the melting point, when an adjustment of the concentration takes place by dif- fusion. This diffusion is very slow with Bi-Sb alloys but with Au-Ag and Au-Pt alloys the thermal treatment is quite effective. All alloys whose components are miscible in all proportions do not show the kind of diagram described above, that is, the steady decrease of melting and crystallization points from the melting point of the higher-melting component to that of the lower. Exactly as with solutions of liquids, whose boiling point UU 600 500 400 300 268 9nn Sb i B, X x / X, i AI x C .k x / / ,/-' .^' | 1 1 1 Bi 622 100 FIG. 23. is higher or lower than the boiling point of either of the com- ponents, there is a mixture which has a maximum or minimum boiling point, so with a solid solution whose melting point is higher or lower than that of the components, there is a mixture of maximum or minimum melting point. The liquid mixture with a maximum or minimum boiling point, boils at a constant temperature, the composition of the vapor being the same as that of the liquid. Similarly, there is no concentration change in the melting of a solid solution of maximum or minimum melt- ing point, the alloy has the same composition in the solid and liquid state. Alloys of this kind are designated as coincident melting. Schematic diagrams of solid solutions with a maximum 60 THE PHYSICAL CHEMISTRY OF THE METALS and minimum melting point are shown in Fig. 240 and b. Out- side of the pure components only the mixture of maximum or minimum boiling point is coincident melting; all the rest melt over a temperature range. With solid solutions of the type Sb-Bi there is in general no coincident melting alloy. The miscibility of the metals, in the solid state, is frequently only partial, for example, in the metal pairs Cu-Fe, Cu-Pt, and others. Tammann has formulated a rule concerning these cases, it is " the metal with the higher melting point, always dis- Temp. Melt Solid Solution Melt Solid Solution Temp. FlG. 24. \ solves more of the metal with the lower melting point than vice versa." The crystallization diagrams of alloys, with partial mis- cibility of the components, show many different types of phe- nomena and are best illustrated by examples. The gold-nickel alloys (Fig. 25), which have been studied by Levin, and the cadmium-mercury alloys, which have been studied by Bijl, are such examples. In the first case the occurrence of a eutectic mixture of the two kinds of solid solutions is interesting. Inter-metallic Compounds. A whole series of phenomena indicate that chemical reaction takes place when certain metals are melted together. Among METALLIC SOLUTIONS AND ALLOYS 61 J.UUV 1500 1400 1300 1200 1100 1064 1000 900 NU1484 / ' / / / Au , / \ \/A _j B c 20 40 60 80 100 FIG. 25. -50 350 f i j t ) C A 10 20 30 40 50 60 70 80 90 10o|cd 100 90 80 70 60 50 40 30 20 10 -Hg FIG. 26. 62 THE PHYSICAL CHEMISTRY OF THE METALS these may be mentioned the strong heat evolution when sodium and hot mercury are mixed and also when copper and zinc are mixed. The large volume difference between alloys and a physi- cal mixture of their components, is best explained by such a compound formation. Alloys frequently show properties which the components do not possess. The alloys of antimony and manganese, for example, are extraordinary and are well suited to lecture demonstration of the change of components by melting. If a mixture of equal molecular proportions of antimony and manganese are fused, in FIG. 27. Powdered Mn-Sb Alloy in a Magnetic Field. a hard glass tube, a reaction takes place, and after cooling, the powdered alloy behaves like iron filings, as regards a magnet, it being possible to show the magnetic lines of force by means of it. (See Fig. 27.) Neither manganese nor antimony as elements are magnetic, but manganese compounds frequently are, and in all the magnetic alloys discovered by Heusler * the magnetic character is connected with the presence of manganese com- pounds; for example, manganese and aluminum. * Heusler, verb. Deutsch. Physikalisch. Gesselschaft, 1903, 219. Stark & Haupt, Ibid., 1903, 222. Heusler, Z. Angew. Chem., 1904, 260. METALLIC SOLUTIONS AND ALLOYS 63 Compounds do not ordinarily occur between .metals that stand near each other in the periodic system. (Tammann).* They are, however, no rare occurrence and many metal pairs form several compounds. The existence of the following com- pounds has been demonstrated: Cu 2 Cd 3 , Cu 3 Cd, Cu 3 Al, CuAl, Cu 4 Sn, Cu 3 Sn, CuSn, Cu 3 Sb, Cu 2 Sb, Ag 3 2n 2 , AgZn, Ag 2 Zn 3 , Ag 2 Zn 5 , AgsAl, Ag 2 Al, Ag 3 Sb, Au 3 Zn 5 , AuZn 8 , Au 2 Zn, Au4Cd 3 , AuCd 3 , Au4Al, Au 5 Al 2 , Au 2 Al, AuAl, AuAl 2 , AuSn, AuSn 2 , AuSn4, AuSb2, Au 2 Pb, AuPb 2 , Na4Sn, Na 2 Sn, Na4Sn 3 , NaSn, A+ Eatectic A, AB | A B + Eutectic A, AB AB j-f- Eutectic I B, A B B -4- Eutectic B, AB Per cent B FIG. 28. NaSn 2 , Na 4 Pb, Na 2 Pb, NaPb, Na 2 Pb 5 , Mg 2 Sn, Mg 2 Pb, SbAl, Sb 2 Zn 3 , SbZn, SbNa 3 , SbNa, Mg 3 Sb, BiNa 3 , BiNa, ,BiMg 3 , Ni 3 Sn 2 . These compounds can occur as separate structure constit- uents. They can also form solid solutions with the components, and if several compounds are capable of existence, dissociation and recombination come into consideration. The equilibrium diagram may be very complex, if all these phenomena occur simultaneously with a single metal pair. It is not possible to discuss all known cases here, so we will limit ourselves to the simplest and most typical. * Z. anorg. Chem., 49, 113 (1906). 64 THE PHYSICAL CHEMISTRY OF THE METALS If a stable compound of the two components is formed a dia- gram like Fig. 28 is obtained. A maximum is seen in the curve which represents the melting point of the compound A-B. This melting point is lowered by an excess of A or B and curves are obtained which represent the equilibrium of differently com- posed melts with the solid compound; these solidification curves intersect on the one hand the crystallization curve of component A, and on the other hand, that of component B. Two eutectic points and two eutectic mixtures are accordingly possible, for Per cent B ^ FIG. 29. an alloy of two components, which form one compound. If the compound A-B decomposes, before it melts, an equilibrium diagram is obtained as shown in Fig. 29. Miscroscopically the existence of two eutectics can be recognized, but only one eutectic point can be observed by the cooling curve method. The diagram for the occurrence of many compounds is still more complex, but the principles involved are the same as with the cases discussed. Ternary and Quaternary Alloys. In practice, alloys are frequently met which are composed of more than two metals. We shall consider here only the bronzes METALLIC SOLUTIONS AND ALLOYS 65 and the bearing metals of which a whole series of different com- position have been placed on the market. The easiest to treat are the three component or ternary alloys containing copper, zinc, antimony, lead and tin. The great number of possible structure components make the study of these complex systems quite difficult. But to give an idea of FIG. 30 Ternary Alloy of Sn, Bi, Pb. White Portion Bismuth crystals, Lighter Portion of the Ground Mass, Binary (Bi-Sn) Eutectic, Dark Portion Ter- nary Eutectic. the method of attack on ternary systems, the combination of the three metals Zn, Bi, Pb, which is not complicated by the forma- tion of solid solutions or compounds, will be considered rather closely. It has been investigated by Charpy.* We will first consider a metallograph of this kind of alloy (See Fig. 30), the picture shows three different parts, the bright * Contribution a 1'etude des alliages, 121. 66 THE PHYSICAL CHEMISTRY OF THE METALS large surfaces are doubtless the first component to be precipitated and can be considered as Bi. The crystals are embedded in a mass which has the appearance of a eutectic, but this ground mass is not uniform, and light and dark fields can be distinguished from each other in it. The light places are binary eutectic, the dark places ternary eutectic. The binary (Bi, Zn) pre- cipitates before the ternary. A graphical representation of the solidification relations can- not be made in the simple way we have used for binary systems. Representation in a plane is only possible, when the number of variables is two. With more variables a space model must be constructed. We may, however, represent a three-component system at constant temperature. The points no longer form a line but a plane. Of the different possible representations, that of triangular coordinate axes is ordinarily used. The side of an equilateral triangle is divided in the relation a : b : c of the three components, a parallel to the side AC is then drawn through the end point of a and this made equal to 5. In this way a point P in the interior of the triangle is obtained, which has notable properties. It can easily be proven, that a perpendic- ular which is dropped from it to the three sides AB, BC, AC are in the same proportion as the coordinates a : b : c, and that the sides of the three small equilateral triangles of which P is the mutual vertex are equal to a : b : c respectively. This point adapts itself excellently for the graphic representation of the composition of a mixture of the proportions a : b : c. It can be easily seen that the components can be represented by the vertices of the triangle and systems of two components along the triangle sides. (Fig. 31.) The temperature axis is now erected in a vertical plane, and a three-sided prism is obtained, in which the equilibrium diagram in space is placed. The equilibrium diagrams of the two component systems are now represented on the sides of the prism. In the case of the system Zn, Bi, Pb, these are simple pairs of curves with a eutectic point. Experiment shows now that by the addition of a third metal to a binary eutectic, the solidification temperature is lowered. This is true of all the metal pairs and there are three space METALLIC SOLUTIONS AND ALLOYS 67 curves, which represent the equilibrium between the melt and the two precipitated metals. The equilibrium of the melt with any one solid metal is represented by a surface. The three space curves intersect in a point, a ternary eutectic point, at which there exists a simultaneous equilibrium between the melt and the three different precipitated metals. The crystalliza- tion is complete at the ternary eutectic point. FIG. 32. The relations between the binary and ternary eutectic points and the composition of the different eutectics are shown in the following table: Components. Eutectic Temp. Atomic Proportions. Binary eutectics (Sn Pb \ Sn Bi 183 175 77 : 23 58 : 42 Ternary eutectics I Pb Bi Sn Pb Bi 125 06 56.2 : 4 3-8 24 I 4.6 I 28 Ternary eutectic with addition of a fourth metal Sn Pb BiCd 65 c Woods metal 68 THE PHYSICAL CHEMISTRY OF THE METALS A quick insight into the equilibrium changes in a ternary alloy can be obtained from a projection of the space diagram on the basal plane and connecting the points representing combinations which crystallize at the same temperature. In this way a mul- titude of isotherms is obtained which show the height above the basal plane, just as the contour lines on topographical maps. Such a projection is shown in Fig. 33. The addition of a fourth metal sometimes causes a de- pression of the melting point below that of the eutectic point FIG. 33. of the ternary alloy as, for ex- ample, shown for Woods metal in the preceding table. The different low melting alloys are of this type. Some of these alloys melt in hot water as, for example, the metals of Newton, Lichtenberg and Wood and have found practical use in many ways. Relations of Mechanical and Physical Properties to Alloy Structure. The mechanical and physical properties of alloys are depend- ent on the structure and this fact has given practical importance to metallography for the testing of materials. It is clear, that an alloy with the most homogenous possible structure, which has the same components as another, con- taining interstratified large crystals, will excel in tensile and compressive strength. Where the size is important the use of eutectic alloys or alloys that consist of a homogenous solid solu- tion is to be recommended. The elastic limit seems always to be a maximum in the eutectic alloys so far as the few investigations allow a conclusion. (Ssaposhnikow.) It is still premature to make any general statements due to the great difference which the several metals show in their properties. The question of the relation of hardness to structure has been METALLIC SOLUTIONS AND ALLOYS 69 recently treated by Russian investigators. Ssaposhnikow * and his co-workers have established that with normal alloys of two components the eutectic possesses the maximum hardness. The results with the zinc-cadmium alloys are given in the following table: Zinc, Per Cent. Hardness. Zinc, Per Cent. Hardness. Zinc, Per Cent. Hardness. 15-9 24.4 32.2 79-9 43-0 IO 31-5 30 27.2 80 52.0 12.5 33-o 40 31.2 81 46 17.2 36-8 SO 35-o 82.2 39-0 17-6 38.8 60 34-o 90 39-o 20 34-5 70 34-5 100 35-0 The second hardness maximum in the alloy of 80 per cent Zn is not clearly understood. (The pressure in Kg/sq.cm. neces- sary to press a steel ball of 10 mm. diameter into the alloy serves as a measure of the hardness). Hardness maximums are also shown by alloys of the com- ponents Pb-Sn, Al-Zn. Solid solutions are usually somewhat harder than their com- ponents and alloy systems, which are made up of a continuous series of solid solutions usually have a hardness maximum. Bearing Metals. The hardness relations of the structure components play an important role in the so-called bearing metals. The theoretical treatment of this subject is due to Charpy.f A good bearing metal must consist of a plastic ground mass in which hard crystals are embedded in a uniform way. Such a structure offers the best security that the bearing itself will be continually adjusted to the form of the axle without the material becoming attached to the axle. Further there occurs a continual yield of the bearing material, and the abrasion coefficient of it must, therefore, be high so that the bearing is not easily heated. An entirely hard * Chem. Zentr., 1908, I, in. t Contribution a 1'etude des alliages, 121. 70 THE PHYSICAL CHEMISTRY OF THE METALS material will not cling to the axle, but it does not give sufficiently to the pressure. All these conditions are met by an alloy of the above-described structure, since by a change of pressure the hard crystals in the yielding mass are immediately pressed out. The structure of a bearing metal is shown in Fig. 34. It repre- sents an alloy of the components Sn, Sb, Pb, the bright crystals FIG. 34. Bearing Metal Alloy of Tin, Antimony, and Lead. are of the compound SbSn and the dark ground mass is ternary eutectic. Density of Alloys. The relations of certain other constants of alloys to the structure is known with considerable accuracy, viz., the density, the electrical potential and especially the electrical resistance. In many cases the density of an alloy can be calculated from the density of its components, by the rule of mixtures. The reciprocal value of the density, the specific volume, can be fre- quently represented, as Maey * has shown, by a linear equation of the form V L = V A -b.p. B * Z. Phys. Chem. 38, 299 (1901). METALLIC SOLUTIONS AND ALLOYS 71 Where VL is the specific volume of the alloy, V A that of the com- ponent A , pB the percentage of B and b a constant. In the fol- lowing table are given the values for V A and b which are neces- sary to calculate VL. COMPONENTS. V A . &. Kind of Alloy. A. B. Lead Antimony 0.08791 0.08791 0.08791 o. 0881 i 0-IIS54 0.05191 0.05191 0.05191 0.04461 0.07368 0.07366 o . 0948 0-0955 0.10181 o. 10181 o. 10181 0.13710 0.13710 0.0006106 0.0002763 0.000076 0.00049 o. (3002156 o . 000605 o . 000605 0.000852 0.0000191 0.0001422 0.0006345 0.000169 0.000063 0.0004715 0.0001373 0.000353 0.0001187 0.00004 K K K K K M(p) M(l) K K M(/) K VM K Lead Cadmium Lead Silver Lead Tin Cadmium Tin Gold CoDoer Gold Gold .... Silver Tin Iridium Platinum Mercury Lead Mercury . . . Tin Silver . Copper Silver Bismuth Bismuth Antimony Cadmium Bismuth Bismuth, Zinc Tin Antimony Zinc Tin K means simultaneous deposition of the two metals, M forma- tion of solid solutions, / total miscibility, p partial miscibility, V formation of compounds. No volume change occurs in the formation of solid solutions in these alloys. There is, however, a contraction, if the com- ponents combine with each other, hence a minimum of specific volume occurs in alloys of Sb and Fe and Cu and Sn when their composition expresses that of a compound, as FeSb or CusSn (Fig. 35). A large contraction cannot always be explained by the formation of a compound, for example, with tin and silver the minimum volume corresponds to a composition of 29.2 per cent tin and with lead and bismuth the maximum contraction occurs with 40-50 per cent bismuth while, according to the equilibrium 72 THE PHYSICAL CHEMISTRY OF THE METALS diagrams the alloys of this composition must consist of normal conglomerates of their pure components (Roozeboom). Potential of Alloys. Of the electrical magnitudes which are capable of giving an insight into the nature of structure components we shall first consider potential, which has been recently investigated by the Russian investigator Puschin.* He measured the potential difference between the binary alloys and their least noble com- 8.5 8.0 7.5 7.0 100 20 30 70 40 GO 50 50 70 30 20 00 10 FIG. 35. 9.0 8.5 8.0 7.5 7.0 100 ^Sn o "Cu ponents using a salt of the latter as an electrolyte. The follow- ing rules have been formulated for the relation of potential to composition. (1) If the alloy is a conglomerate of components the poten- tial difference is independent of the composition. (2) With the formation of a continuous series of solid solu- tions the potential difference rises with increasing content of the noble metal. (3) With the formation of compounds sudden changes occur in the potential difference. * Chem. Zentr., 1907, 1, 1724; II, 1315, 2026; 1908, 1, 108. METALLIC SOLUTIONS AND ALLOYS 73 Electrical Resistance of Alloys. The electrical resistance of an alloy frequently shows import- ant imprints of its character and from the relation of the con- 10 20 30 40 50 60 70 80 90 100 ^Cd 100 90 80 70 60 50 40 30 20 10 FIG. 36. Ag- >Au 10 20 30 40 90 60 70 80 90 100$ Ag 100 90 80 70 60 50 40 30 20 10 "All FlG. 37. ductivity to the composition, conclusions can frequently be drawn, regarding the structure. 74 THE PHYSICAL CHEMISTRY OF THE METALS There is a great deal of experimental material concerning the conductivity of alloys. The investigations of Mathiessen * are of especial value. By comparison of his results with the metallographic investigations of Guertler f and the earlier inves- tigations of LeChatelier { the following general rules have developed. (i) Alloys, which are purely mixtures of their components, have a conductivity which is additively made up of the conduc- tivity of their components. Copper Antimony 14 13- 12 11- 10- 9- 8- 7- 6- 5- Cu 2 Sb -13 12 20 40 60 FIG. 38. 80 fcSb (2) Alloys which consist of solid solutions show an important lowering of conductivity. (3) The occurrence of compounds can frequently be recog- nized by the occurrence of peaks in the composition-conductivity diagram. The alloy of zinc and cadmium is an example of a mixture, (see Fig. 36), while the pair silver, gold is an example of the solid solutions (see Fig. 37). The alloys of copper-antimony show the effect of compound formation (see Fig. 38). * Pogg. Ann., 110, 222 (1860). t J. Inst. Metals, No. 2, 1911, VI, p. 135. t Rev. gen. Sci., 30, June, 1895. METALLIC SOLUTIONS AND ALLOYS 75 The temperature coefficient of the conductivity shows a similar regularity which has also been noted by Guertler.* With conglomerates, the temperature coefficient is equal to that of the pure metals, and the resistance of all this kind of alloys van- ishes as the metals reach absolute zero. The alloys which con- sist of solid solutions, show in comparison to their components, a very small temperature coefficient, and their resistance does not 26 AU FIG. 39. vanish at absolute zero. In general there is a proportionality be- tween the conductivity of alloys and their temperature coefficient. The form of the curve expressing the temperature coefficient as a function of the volume composition, has, without exception, the same form as the curve for conductivity. These facts can be recognized by a comparison of Fig. 39 with the earlier figures. *Z. anorg. Chem., 51, 397 (1906); 54, 58 (1907); Z. Elektrochem., 13, 441 (1907). 76 THE PHYSICAL CHEMISTRY OF THE METALS The electrical relations of solid solutions are not only of interest for their bearing on metallography, but also for their practical importance. The valuable resistance metals con- stantan and manganin are solid solutions. We will, therefore, look a little farther into the theoretical problems which they present, and see if we can discover what factors condition the strong influence of small additions of a second metal on the con- ductivity. As the following table shows, considerable decrease in the conductivity of metallic copper is caused by the addition of 2 per cent of other material. Thermo-electric force of the dis- Copper +2% Metal Atomic Weight of addition Electrical Cond. Xio* solved metal against copper. Temp. diff. 100. Millivolt Cu. Dure 62.1 WtftJ ^/HAV* Ag 108 55-i o.oi Au 197.2 47-5 +0.02 Zn 65.4 45 +0.03 Sn 118.5 26.5 0.30 Ni 58.5 26.0 -2.15 Al 27.1 17 0.32 Mn 55 12 Fe 56 13 + I.IQ Especially large lowering is obtained with alloys of copper and nickel, whose components are miscible in all proportions, this is shown in diagram 40 and the next table. Metal Conductivity Xio Ratio Copper. . . C7.A IOO Nickel . . 13*30 23. 3 Constantan, 6oCu, 4oNi. 2.04 2. er The smallness of the temperature coefficient, is brought out by reference to the next table. It is even possible to obtain an alloy whose resistance is independent of the temperature. This kind of a resistance metal, is of great practical importance, for METALLIC SOLUTIONS AND ALLOYS 77 the preparation of precision resistances, as they are uninfluenced by the temperature. Mets il Temp. Coeff. of resistance at 18 Metal Temp. Coeff. of resistance at 18 Cu 1 00% -{-O 00428 Au 1 00% -j-o 00368 Cu 80%, Ni 20%.. -(-0.000262 Au 90%, Ag 10% +o 00124 Cu u% Ni 46% -|-o oooo Au66%, Ag^% ~f~o 00067 Ni 1 00% -4-Q OO^CK Ag 100% -j-o 00400 Constantan and manganin which show these valuable prop- erties are solid homogenous solutions. It is of interest in this connection, that there is also an elementary metal, which possesses an extraordinarily low tem- perature coefficient, of the same order of magnitude, as that of many alloys, namely, 0.0008. It is mercury in the liquid state. Liebenow has drawn the conclusion from this, that mercury is a solution, which contains in it a second molecular species. It is assumed for example, that in it outside of the mercury molecules Hg, a second molecule Hg2 is present in equilibrium with the first. In the solid state mercury behaves normally. The question of the remarkable conductivity lowering of alloys is often raised but it has not been answered in every way. Lord Rayleigh and later Liebenow have attempted to explain the phenomena by the Peltier effect, that is, the thermo-current causes an increase in resistance. However, at the time of the presenting of this theory, we were not so well-informed on the structure relations of alloys, and in the light of our present knowledge it is quite inadequate. The conductivity lowering can be explained, however, in terms of the electron theory. It is first necessary to find whether the heat conductivity shows a similar lowering for these alloys, and to ascertain whether the law of Wiedemann and Franz holds with certainty. With the Bi-Pb and Bi-Sn alloys in which the ability to form solid solutions is confined to narrow limits a parallelism has been established between the electrical and heat conductivity (F. A. Schulze).* Compare the following table with Fig. 41. * Habilitationschrift, Marburg, 1902. 78 THE PHYSICAL CHEMISTRY OF THE METALS From these measurements it follows that the ratio of the two conductivities of the alloys is somewhat greater than for pure metals. The same results have been obtained by differently carried out investigations on the Ni-Cu series. Alloy. Volume, Per Cent. Heat Con- ductivity. X Electrical Conduct- ivity. I0 5 X a Bi Pb IOO 0.019 0.830 230 Bismuth-Lead < 99-57 98.44 97-31 0-43 1.56 2.69 0.0188 0.0119 O.OII7 0.766 0.468 0.444 245 256 263 96.47 IOO 99-54 3-53 0.0129 0.0190 0.0140 0.514 0.830 0-595 251 2 3 235 0.46 Bismuth-Tin 99-05 0-95 0.0126 0.506 248 97-13 2.87 O.OIIO 0.448 245 90.26 9-74 0.0126 0.488 258 Metal. X a Observer. Cu, pure 665 Jaeger and Diesselhorst * Ni, pure 669 Jaeger and Diesselhorst 60 Cu, 40 Ni 1106 Jaeger and Diesselhorst <4 Cu 4.6 Ni 90 1 Gruneisen f 84 Cu 4 Ni 12 Mn 014 Jaeger and Diesselhorst * Loc. cit. t Ann. Phys., 3, 71 (1900). Copper and nickel are miscible in all proportions and accord- ingly the curve which represents the relations of the conductivity quotient to the composition possesses a maximum. The increase of the conductivity quotient shows that the dissolved metal lowers the heat conductivity relatively less than the electrical. How is this fact to be explained? If we take as our basis the Drude electron theory we see that the heat con- ductivity is concerned with heavier particles" than the con- duction of electricity. As diffusible particles we have now in these solid solutions, beside the electrons, the dissolved mole- METALLIC SOLUTIONS AND ALLOYS 79 cules of the second metal. These will fulfill the requirements for heat conductivity but will not take part in the electrical con- ductivity. They are to be considered as electrically neutral and undissociated. Now, however, the electrical conductivity decreases. Such a decrease can, according to our previous considerations, result in two ways; in decreasing the number of electrons or in the ,20 10" 10 Copper Nickel V 3 2 1 1 10 1.00 0.90 .0.80 4 1 , I" K0.60 0.50 0.40 i-Sn Z. 19 15 11 20 40 60 80 100 Volume per cent Nickel FlG. 40. 95 96 97 98 99 100 Volume per cent Bi FlG. 41. resistance which they encounter in their motion under the influ- ence of the e.m.f. becoming greater due to the increase of internal friction. Electrons and molecules both diffuse in the solid metal- lic solutions; we have accordingly to do with the diffusion of a mixture. The kinetic theory allows us to draw certain con- clusions concerning the friction in a gas mixture, which have been found to agree with the experimental facts. The vis- cosity of a gas is increased by the addition of a second gas even though the second gas possesses a smaller coefficient than the 80 THE PHYSICAL CHEMISTRY OF THE METALS first. (This is true only in certain cases, and is not true in case of oxygen-hydrogen, nitrogen-hydrogen, hydrogen-helium or oxygen-nitrogen, but is true for carbon dioxide-hydrogen and helium-argon). Graham, and later Puluj, both obtained the remarkable, and on first thought improbable, result, that the internal friction of C02 is not lowered but raised by the addition of a small amount of hydrogen. This increase reaches a maxi- mum with 27 per cent H2 and 73 per cent C02 at room tem- perature. Now, if we carry over the kinetic theory to a mixture of electrons and molecules, it follows that the internal friction of the electrons increases with the addition of molecules and, therefore, the electrical conductivity decreases. The lowering of the conductivity can be explained then even if the number of electrons does not decrease. A decrease of the electron number in the alloys could be ascer- tained as we have previously seen by a change in the optical properties (compare p. 31), especially by a lowering of the reflection constant. There is, however, no experimental data on the optical constants of alloys as compared with their com- ponents. Thermo-electric Force of Alloys. The comparison of the thermo-electric forces of the alloys and their components gives us a second means, but the experimental material is also scarce in this field. However, there is enough for a preliminary consideration. Jaeger and Diesselhorst have determined the thermo-electric force of constantan and man- ganin against copper. If one junction is held at o and the other at 100 there results the following electromotive-force. Constan tan-copper 3440 micro volts Manganin copper -f- 5 70 micro volts According to Drude the thermo-electric force is connected with the electron number N of the two metals by the equation (v. p. 32). N cu (micro- volts) = i .71 (Tz T\ ) log e AT A alloy METALLIC SOLUTIONS AND ALLOYS 81 The positive pole at the warm junction is the electron richer metal. It follows, therefore, that manganin has a higher and constantan a lower concentration than copper. From the given results the ratio of the electron numbers is calculated. Ncu *r- - = 1-223, i V constantan Ncu = manganin N O.97O If the electron number of copper is now placed arbitrarily at 100 it follows that N cu ' N Constantan = TOO : 81, Ncu - N Manganin = 100 : 103. We see, therefore, that the electron concentration of an alloy is not necessarily smaller than that of the pure metals. The ratio of the electrical conductivity between copper and its two alloys is acu ' o- Constantan = 100 : 3.55, - A continuation of the heating causes an enlargement of the structure and we obtain sorbite and, if the lamellar structure is resolvable with the help of the microscope, perlite. Figs. 57 and 58 show pictures of these two intermediate states between the solid solution and perlite. The steps of the rearrangement are shown in the following equation: Austenitev ^jTroostite >Sorbite ^lamellar Perlite Granular Martensite' Perlite * Z .Chem. Ind. Kolloide, 1910 p. 290. ALLOYS OF METALS WITH CARBIDES, ETC. 97 FIG. 57. Troostite (Black); Austemite (Light), Shot Through with Cementite Needles (Goerens). Xioo. FIG. 58. Sorbite, Over Light Ferrite (Goerens). Xioo. 98 THE PHYSICAL CHEMISTRY OF THE METALS If perlite be heated above the eutectic point, 710, a solid solution of iron and carbide, martensite forms. This structural element may be fixed by quenching and is recognized by its hard- ness, being the constituent of hardened steels. By heating and slow cooling it separates again into its components and softens. Steel hardened by heating and quenching is accordingly in the metastable state. The Precipitation of Carbon in Iron Carbon Alloys. Under certain conditions, namely slow cooling of the melt or continued heating at high temperature the carbon does not precipitate out of the concentrated melt as cementite but as the stable form, graphite, or, at relatively low temperatures as amorphous carbon; the so-called temper carbon. Due to its small specific gravity it collects at the surface of liquid pig iron, which is held somewhat above its melting point, as in casting practice, and is called " kish." In the slowly solidified gray pig iron it is possible to recognize the graphite leaves with the naked eye. The equilibrium between carbide, iron and graphite and also that between carbide, iron and temper carbon are unknown. It is certain, however, that amorphous carbon, for example, sugar carbon as well as finely divided graphite dissolves in iron with the formation of carbide. It is the general conception that the total dissolved carbon in the melt is contained as carbide; we may further suppose that an equilibrium is established between this solution and solid carbon, which can be formulated. Fe+C (solid) = Fe 3 C (dissolved). It is even possible that the equilibrium relations would also hold for the solid iron-carbide solutions. That this equilibrium has not previously been established is due to the fact that the dissolved carbide especially in not very concentrated solutions is very slowly decomposed. It has, therefore, been impossible up till now to establish the eutectic point for the precipitation of a mixture of graphite and solid solution by cooling experiments. These led always to the eutec- tic point of the metastable system solid solution cementite. ALLOYS OF METALS WITH CARBIDES, ETC. 99 The equilibrium relations set forth above do not hold for solid crystallized cementite, this is always labile and its decomposi- tion velocity in the neighborhood of the solidification point of pig iron (1130 C.) is very considerable. This fact has led a num- ber of authors to assume that the precipitation of graphite must always be preceded by a precipitation of cementite crystals in the melt followed by the irreversible reaction. Fe 3 C (solid) - Graphite-f-melt. FIG. 59. Graphite Druse, Surrounded by Light Eutectic (Goerens). Xso. With pig iron of small carbon content this indirect way of decom- position is favored, however, with melts very rich in carbon in which, according to the law of mass action, there must be a very great decomposition velocity the possibility of the direct crys- tallization of graphite is not rejected. Since the graphite is stable compared to the cementite, the carbon content of the melt with which the graphite is in equili- brium is smaller than that of the melt out of which the solid cementite crystallizes. These relations must also hold for the solid solutions which are in equilibrium with the graphite. This 100 THE PHYSICAL CHEMISTRY OF THE METALS deduction is confirmed by the quenching test of graphite-con- taining iron-carbon alloys which do not contain foreign sub- stances, especially silicon. The part of the solid solution that is directly in contact with the graphite is always carbon poorer than that at a distance from which the carbon has not sep- arated. In the microphotograph this fact can be at once recog- nized. In Fig. 59 is seen a light solid solution-cementite eutectic; in the center of this is a dark field consisting of carbon poor iron FIG. 60. Graphite in Eutectic, Strongly Enlarged (Goerens). ' inside of which are easily recognized graphite grains. These relations are shown still better and under stronger magnification in Fig. 60. The necessary experimental material is not extant for the setting up of a complete equilibrium diagram of iron-graphite. A schematic representation must suffice. From our earlier consid- erations it can be deduced that the curves which express the solu- bility of graphite in the melt and in the solid solution are de- flected to the left as compared to the cementite curve. Also an ALLOYS OF METALS WITH CARBIDES, ETC 101 analogous equilibrium line which lies in the space between the graphite and cementite curves is to be given for the amorphous carbon. (Fig. 61.) In practice, cases are not met in which the graphite pre- cipitation goes on till the equilibrium is established. Micro- scopic investigation of all pig iron shows greater or smaller portions with the structure constituents of the metastable Solid Solution -F ' Eutectic Eutectic + Cementite Crystals Perlite+ Cementite Per cent Carbon 6 Cementite Fe 3 C FIG. 61. systems. The necessary time for equilibrium is even greater than the duration of cooling of the preparation. If the gray pig iron is heated for a long time at 1100 and quenched in ice water there are obtained, as would be expected, graphite and the needle-like martensite as the structural com- ponents; unchanged cementite crystals are also often met in such specimens. In single cases troostite and austentite occur. If the cooling is allowed to proceed slowly the specimen shows graphite and cementite embedded in lamellar perlite. These 102 THE PHYSICAL CHEMISTRY OF THE METALS FIG. 62. Graphite +Martensite (Goerens). FIG. 63. Gray Pig Iron; Black Particles of Graphite, Light Cementite Em- bedded in Lamellar Perlite. A Needle Scratch is on the Surface, which does not Scratch the Hard Cementite. (Goerens). Xsoo. ALLOYS OF METALS WTIH CARBIDES, ETC. 103 are the components that one finds in normal gray pig iron. A micrograph of such a specimen is shown in Fig. 63 ; to indicate the hardness of the different components a needle scratch is drawn over the surface, showing in the soft perlite as a deep furrow while the hard cementite is entirely uninjured. A pic- ture of gray pig iron with strong graphite precipitation is shown in Fig. 64. By long-continued heating of white pig iron, the carbon is not obtained as crystal leaves, but in the finely divided state, as the so-called temper carbon, which is probably amorphous. FIG. 64. Gray Pig Iron ; Black Particles of Graphite, Light of Cementite. Ground Mass Lamellar Perlite (Goerens). This process has been followed metallographically by Wiist. The pictures of his preparations are given in Figs. 65, 66 and 67. Fig. 65 shows the original material, a white pig iron which has been rather quickly cooled. It consists as can be seen of an intimate mixture of cementite and solid solution which already shows a tendency to decomposition; it is "sorbitic." By heat- ing 50 hours in vacuum at 980 and slow cooling the structure is totally changed. This can be recognized from Figs. 66 and 67. One obtains black excretions which are surrounded with light halos, embedded in lamellar perlite. The excretions are temper carbon which have surrounded themselves with a circuit of fer- rite. It would be interesting to see -how a specimen which had 104 THE PHYSICAL CHEMISTRY OF THE METALS been heated for a long time and then quenched would be con- stituted, it would then be shown whether the carbon excretions result equally well from isothermal decomposition of cementite into carbon and solid solution, poor in carbon, or whether these flocks are a secondary condition and a result of the precipitation and the crystallization of ferrite as we have seen in the formation of granular perlite. The practical use of this process in the malleabilizing of cast iron is known to all. At temperatures FIG. 65. White Pig Iron; Cementite (Light) +Sorbite (Dark) (Goerens). Xsoo. below 700 the decomposition of cementite into its elements is still unobservable and would require a very long time. The Use of Additions to Iron and Steel. The various kinds of iron and steel which are met in practice are seldom pure iron carbon alloys. Generally they contain still other elements. We shall not consider here such undesirable elements as sulfur and phosphorus which are excluded as far as possible, but only those which are intentionally added. Of these manganese and silicon are especially important and for the preparation of special steels nickel, chromium, tungsten, molybde- ALLOYS OF METALS WITH CARBIDES, ETC. 105 FIG. 66. Material of Fig. 65 Heated to 900 and Cooled Slowly. Black Nuclei of Temper Carbon, Surrounded by Light Ferrite; Ground Mass Perlite (Goerens). FIG. 67. Material of Fig. 66 with Great Magnification (Goerens). 106 THE PHYSICAL CHEMISTRY OF THE METALS num and vanadium. They all modify the properties of the pure iron carbon alloys. We will not develop here the space diagrams of these ternary and quaternary alloys but will confine ourselves to a general dis- cussion. The changes which these substances produce in the properties of iron-carbon alloys can be different in nature. They depend on the fact that either the stability of the carbide is lessened or increased or that the transition temperature and the eutectic point and with them the boundaries of the solid solution fields are displaced. It has been known for a long time that the presence of silicon in iron favored the decomposition of the carbide into its elements; it has even been supposed that the presence of sil- icon was a necessary preliminary condition for the occurrence of graphite; this, however, has been disproved by the investi- gations of Wiist on pure iron-carbon alloys which, by sufficiently long heating near the melting point leave graphite. In all the cases considered the presence of this substance in iron decreases the stability of cementite and the solubility of the carbide in liquid iron. Investigations of the lowering of the solubility of carbon in iron by the addition of silicon have been carried out by Petersen.* In carbon saturated silicon containing alloys, which have been heated for a long time at the melting point the con- tent of carbide (cementite+ dissolved carbide) decreases with rising temperature as the following table shows: Silicon, Per Cent. Carbon, Per Cent. F. P. Silicon, Per Cent. Carbon, Per Cent. F. P. 4-3 1130 3-25 3-41 1187 0.13 4.29 H38 3-69 3-32 "97 O.2I 4-23 H3I 3- 9 6 3-24 1205 0.41 4.11 1152 4.86 3-08 I2IO 0.66 4-05 "55 5.06 2.86 1215 1.14 3.96 1160 13-54 1.94 1233 2.07 3-79 1185 26.93 0.87 1255 2.68 3.56 1185 Metallurgie, 3, 8n (1906). ALLOYS OF METALS WITH CARBIDES, ETC. 107 In contrast to silicon stands manganese which makes the labile carbide stable. This action is easily understood when we think that the manganese cementite MnsC, which corresponds to iron cementite FesC is a stable compound which is not split into its elements and that both cementites are isomorphous and are miscible in all proportions in the solid state. As a general rule, the properties of isomorphous mixtures are made up additively from the properties of their components and hence the small tendency of manganese carbide to decompose into its elements is shared by the iron-carbon alloys containing manganese. We meet silicon as well as manganese in pig irons. The white pig iron always contains greater or less amounts of man- ganese while the gray in which the black graphite leaves are easily recognized on a broken surface shows a content of silicon. These two elements are also not without importance in the iron-carbon alloys poor in carbon. The silicon occurs in the iron, not in the elementary form but in the combined state. Guertler and Tammann, who have established the equilibrium diagram for the binary system iron silicon, found the silicide FeSi and con- sider the existence of a second silicide Fe2Si probable. The latter forms with a ferrite an unbroken series of solid solutions, promotes the occurrence of this structure component and in- hibits the polymorphic change as well as the formation of car- bide solid solutions. There is also a displacement of the eutectic point toward the side of the higher temperature so that the solid solution field of silicon containing steel is smaller than that of pure carbon steel. With manganese we have the complete miscibility of the metal with iron on the one hand and the complete miscibility of the corresponding cementite on the other. The presence of man- ganese causes a lowering of the eutectic point that goes hand in hand with a widening of the solid solution field. As for the rest of the substances which are added to the iron- carbon alloys in the manufacture of special steels, nickel, chro- mium, tungsten and molybdenum, their influence on the equilib- rium diagram, at least so long as only small percentages of the third component are considered, is entirely similar to that of 108 THE PHYSICAL CHEMISTRY OF THE METALS silicon and manganese. At higher concentrations new structure components appear. Nickel which with a iron and with 7 iron forms solid solutions in all proportions, enlarges the martensite field and not only lowers the eutectic point but also the transi- tion point of the iron modifications. The higher of the two points is lowered in a larger measure than the lower so that finally a direct transition of 7 solid solution into a solid solu- tion takes place. Chromium and tungsten which favor the formation of f errite, form with it solid solutions. The possibility of the occurrence of martensite is not completely destroyed but the perlite eutectic point is raised and the stable 7 solid solution field lessened. The entrance of these different metals into the solid solution acts similarly to an increased content of dissolved carbide, that is, the addition increases the hardness. We chance here again on the previously mentioned fact that solid solutions are harder than their components. These special steels, especially nickel steel have the great advantage over the pure carbon steels that in spite of their hardness they are not brittle. The strong " hysteresis " of the transition phenomena which is characteristic of all ternary and quaternary steels is of practical and theoretical interest. This phenomenon is not entirely missing in the binary system iron-carbon but it is not so strongly marked as in the special steels. In the determination of the transition point and the eutectic point by the survey of the cooling and heat- ing curves it is frequently observed that the transition phenomena do not take place at the same point with falling temperature as with rising temperature (Fig. 6). The alloy can accordingly be under-cooled or overheated before the transition is released ; a retardation or hysteresis takes place which, in general, is greater, the greater the amount of the third component. This phe- nomenon is especially marked with nickel steel where the transi- tion is from 7 solid solution into a solid solution. The occur- rence of this change is especially easy to observe as the a solid solution is magnetic and the 7 solid solution is non-magnetic. The change can be recognized by the appearance or the disap- pearance of the magnetism. ALLOYS OF METALS WITH CARBIDES, ETC. 109 The position of the magnetic transition point on cooling and heating and its relation to the nickel content is given in the fol- lowing table which is taken from the work of Osmond * who has done the most toward the explanation of these phenomena. COMPOSITION OF THE NICKEL STEEL. MAGNETIC TRANSITION POINT. C. Ni. Si. Mn. On Cooling. On Heating. 0.19 . 0.27 0.31 0.79 715 735 0.19 3-82 O.2O 0.65 628 710 0.17 7.65 0.21 0.68 530 710 0.23 I5-48 o. 24 o-93 145 623 o. 16 24-75 o. 24 0.32 27 540 0.61 26. 20 0.24 0.46 O 540 The position of the point is determined not only by the com- position, but also by the thermal treatment, which the material has undergone. Osmond's experiments on chrome steel show that the point of arrest on cooling varies considerably with the temperature to which it has been heated. The higher the pre- vious heating the lower is the transition point of solid solution into perlite. Heating Temperature, Degrees C. Transition Point on Cooling, Degrees C. 835 713-716 1030 682-692 I22O 635-643 1320 640-600 These hysteresis phenomena of steel are not without analogy in other fields. It is a known and often observed fact that a labile modification of a polymorphic substance which is scarcely obtainable with the pure substance may be relatively stable if the substance is impure or if a foreign substance is intentionally added. Heating to a high temperature acts in a like sense. It has been deduced from these facts that at high temperatures, * Compt. rend., 128, 304 (1898). 110 THE PHYSICAL CHEMISTRY OF THE METALS the crystal neuclei which start the occurrence of the stable crystals at low temperatures are destroyed and the more com- pletely the higher the temperature. The solubility of the solid solution is especially great in quaternary steels, in which two metals are added simultaneously to the carbon-iron alloy. In the chromium tungsten steels, espe- cially if they be exposed before cooling to a very high tempera- ture it is not necessary to quench in order to obtain them hard since there is practically no change to the stable perlite form, the precipitation of cementite out of the solid mass being extraor- dinarily hindered. This steel also does not lose its hardness by heating to 600 as does ordinary steel. In order to obtain the perlite structure in a chrome-tungsten steel it is necessary to heat it for a long time due to its very slow transformation which requires at least one hour at a temperature of 700. This kind of steel is used as a tool steel since it possesses the great advan- tage that it may become hot during use without losing its hard- ness. The quaternary steels which remain hard on heating are called " high-speed steels. " There are other metal, metal-compound systems which are of importance in metallurgy and we will study here a few of the systems. Alloys of Metals and Oxides. Copper Copper oxide. Holborn and Day f always observed lower values for the solidification point of copper in the presence of air than in an indifferent gas atmosphere. The air must, accordingly, form with the liquid metal a soluble product which causes a depression of the copper melting point. This product can be no other than cuprous oxide. Concerning the solubility of this substance in liquid copper and the solidification phenomena as well as the structure of the solidified melt Heyn * has made a thorough study and has arrived at the important result that the alloys of metals and sub- oxides differ in no respect from the alloys of metal pairs. The *E. Heyn, Mittelungen, aus den Konigl. Techn. Versuchsanstalt,, 18, 320 (1900). t Ann. Phys. 4, 99 (1901). ALLOYS OF METALS WITH CARBIDES, ETC. Ill mutual solubility of the two components exists only in the liquid state, a solid conglomerate of the two components exists in the solid state. The typical solidification diagram is obtained with two limbs which cut at the eutectic point (see Fig. 68). The de- 1180 1160 1140 1120 1100 1080 Cu-Cu 2 5 6 FIG. 68. 8 9 10$Cu 2 pendence of the solidification point on the oxide content is shown in the following table: (The temperatures are probably high). Content of Cu2O, Per Cent. SOLIDIFICATION. Begins, Degrees C. Ends, Degrees C. 0.08 IIO2 1.16 IOQS 1085 i-75 1089 1084 3-5 1084 1084 3-4 1084 1084 4-7 1116 1084 6-3 1149 1084 9.0 1186 1084 The structure of the crystallized copper is shown in the accompanying pictures (Figs. 69-71). Fig. 69 shows the typical eutectic structure (compare to this Fig. 20) . Fig. 70 is a picture of an alloy which is poorer in oxygen; Fig. 71, one which is richer 112 THE PHYSICAL CHEMISTRY OF THE METALS in oxygen than the eutectic. In the first are shown particles of the eutectic embedded between large flakes of metal, in the latter, grainy deposits of oxide in the eutectic. From the amount FIG. 69. Cu-Cu 2 Eutectic (Heyn). FIG/;. Cu-Cu 2 O Eutectic Scattered Through Metal (Heyn). of these grains the oxygen content can be approximately ascer- tained. The metallographic method which is quite simple to use can, in this case, displace the analytical. This method is of ALLOYS OF METALS WITH CARBIDES, ETC. 113 practical importance since conclusions in regard to the mechanical properties and usability of the copper metal can be drawn from the oxygen content. A high oxide content lowers the mallea- bility of the material very essentially. Silver Silver-oxide. Liquid silver also dissolves its oxide. It at least takes oxygen from the air which is given up again on solidifying. Since the oxide cannot exist in the free state at the melting point FIG. 71- Granules of Cu 2 in Cu 2 O-Cu Eutectic (Heyn). of silver it breaks down into its elements as soon as the concen- tration in the melt is large, thereby showing the phenomenon of sprouting. The content of dissolved oxide is likewise recog- nized, in that the melting point of silver determined in the absence of air is higher than that ascertained in the presence of air, the temperatures are 961.5 and 955 respectively. (Hoi- born and Day.) There are also cases in which the solubility of the oxide in the metal occurs in the solid state, for example, palladium oxide forms a solid solution with palladium metal. 114 THE PHYSICAL CHEMISTRY OF THE METALS Alloys of Metals and Sulfides. The number of alloys of sulfides with metals is much greater than that of oxides with metals. It also happens that a large number of sulfides show metallic properties and, accordingly, these alloys are in many respects similar to those of elementary metals. Partial and complete miscibility in the solid and liquid state and formation of compounds are known in these alloys as in those of metals. These systems are also of practical importance. They play, for example, a role in the so-called precipitation process, that is, the smelting process by which the metals are precipitated out of their sulfides by decomposition with metallic iron. Lead Lead Sulfide. We meet in case of lead a slight solubility of the sulfide in the metal. The solidification relations of this solution have been investigated by K. Friedrich and A Leroux.* The crystalliza- tion diagram would be entirely normal if the solubility of the sulfide in the neighborhood of the melting point of lead were not so very small. The result is that the eutectic point is very close to the melting point of lead. The dependence of the solidifica- tion point on the composition of the melt is given in the following table: COMPOSITION. F. P. Eutectic Point. COMPOSITION. F. P. Eutectic Point. PbS. Pb. PbS. Pb. 100 94.6 1103 1085 43-0 38-0 57-0 62.0 1036 1030 328 329 5-4 327 89-3 10.7 1073 327 33-2 66.8 1033 329 83-9 16.1 1060 327 28.3 71.7 1015 331 78-7 21.3 1054 326 23-5 76-5 998 325 73-5 26.5 1057 326 18.7 8i-3 956 325 68.3 31-7 1050 325 14.0 86.0 931 326 63-2 36-8 1047 329 9-3 90.7 881 325 58.i 53-o 48.0 41.9 47.0 52.0 1041 1049 1040 329 327 327 4-7 95-4 IOO 806 326 327 * Metallurgie, 2, 536 (1905). ALLOYS OF METALS WITH CARBIDES, ETC. 115 From this and from microphotographs it appears that both components, whether forming solid solutions with one another or compounds, go into sulfides somewhat poorer in sulfur. Fig. 72 FIG. 72. Pb-PbS Alloy, 78.7 Per Cent PbS; PbS Light, Metal Dark (Friedrich and Leroux). shows the metallograph of a sulfide rich (78.7 per cent), Fig. 73 that of a sulfide poor (18.7 per cent) alloy. The same structure components can be recognized in both. FIG. 73 Pb-PbS Alloy, 18.7 Per Cent PbS; PbS Light, Metal Dark (Friedrich and Leroux). Antimony Antimony Sulfide. In the system antimony antimony sulfide we meet partial miscibility in the molten state. These relations have been 116 THE PHYSICAL CHEMISTRY OF THE METALS somewhat thoroughly investigated by the French investigators Grunchant and Chretien * on the one hand and by Pelabon on the other. They show that the molten sulfide dissolves small amounts of the metal and that thereby the melting point of the solvent is depressed. From the magnitude of this depression the molecular weight of the metal can be ascertained; it is as in most other cases monatomic, the observed value being 113 as compared to the theoretical 120. The depression constant has the value 790. At larger concentrations a separation occurs into two layers of different density,which can be observed up to the boiling point of the metal. DENSITY OF MIXTURES OF Sb AND Sb 2 S 3 Temperature (in degrees) . . Sb layer 13 6 7<; 643 698 6 1116 1156 6 4."\ Density Sb2Ss layer 6 4.3 * 80 Eute :tic Line 1067 C oE c )0 90 80 70 60 50 40 30 20 10 r 10 20 30 40 50 60 70 80 '90 100$Cu 2 S FIG. 74. tion curve for the metallic copper. Besides these curves there is between 1102 and the melting point of the sulfide 1127 still a further curve which represents the precipitation of sulfide from melts rich in sulfur (line BG in Fig. 74). Fig. 74 shows the equil- ibrium diagram. The eutectic point lies at 1067. The com- position of the eutectic alloy is 3.8 per cent sulfide and 96.2 per cent metal. 118 THE PHYSICAL CHEMISTRY OF THE METALS The microscopic structure of the solidified melt is again very characteristic and it is possible exactly as with the presence of oxide to ascertain very small amounts of sulfur in copper in the microscopic way. Heyn and Bauer found amounts that were overlooked by the analytical methods, viz., o.oi per cent sulfide corresponding to 0.002 per cent sulfur as strings of sulfide eutectic. A series of section pictures which represent different parts of the equilibrium diagram are given in Figs. 75-78. Fig. 75 shows the eutectic with 3.8 per cent sulfide content (magnification 117). Fig. 76 represents an alloy with 0.49 per FIG. 75. Cu-Cu 2 S Alloy, Eutectic 3.8 Per Cent Cu 2 S (Heyn and Bauer). Xn?. cent sulfide (magnification 117.) Fig. 77 represents an alloy with 8 per cent sulfide, the same shows noticeable sulfide crystals (magnification 117.) Fig. 78 shows an alloy of 95 per cent with metal inclusions (magnification 117). Silver Silver Sulfide. Nearly the same equilibrium diagram as described above is found for the system silver silver sulfide, which has been inves- tigated by Friedrich and Leroux * the miscibility of the com- * Metallurgie, 3, 361 (1906). ALLOYS OF METALS WITH CARBIDES, ETC. 119 FIG. 76. Cu-Cu 2 S; 0.49 Per Cent Cu 2 S; Strings of Eutectic in Ground Mass of Metal (Heyn and Bauer). Xn?. FIG. 77 Cu-Cu 2 S; 8 Per Cent Cu 2 S; Noticeable Sulfide Crystallites (Heyn and Bauer). Xn7. 120 THE PHYSICAL CHEMISTRY OF THE METALS ponents in the liquid state is also limited here. There exists a mixing interval (at 906 for solutions with a content between 1.3 and 97 per cent sulfide). . For all mixtures whose composition lies between these limits there exists a constant temperature for the precipitation of metal (906). Solutions with a content less than 13 per cent show higher crystallization temperatures; the corresponding curve limb has its origin in the melting point of the pure metal. The remaining curve limb, the limb for the pre- cipitation of the sulfide out of the rich melt which starts at the melting point of that compound, 835 and that for the crystalliza- FIG. 78.-^Cu-Cu 2 S; Metal Inclusion K (Heyn and Bauer). Xn?. tion of metal from high percentage melts, which turns off hori- zontally at 906, are yet to be observed, only the eutectic tem- perature at which the two cut is known, namely, 807. In the cooling curves of the alloys there is shown a still further point of arrest at 175 at which -an allo tropic transition of the sulfide occurs. Iron Iron Sulfide. The partial miscibility between metals and sulfides appears to be the rule; at least we find still others like the systems con- ALLOYS OF METALS WITH CARBIDES, ETC. 121 sidered. Treitschke and Tammann* found these phenomena also for the mixtures of iron with iron sulfide, since the two liquids are very viscous, layer formation does not take place, but an emulsion of the two melts forms. The mixing gap at 1400 lies between 20 and 92 per cent metal. The miscibility of the com- ponents is also continued in the solid state but is only partial. A solid solution of little sulfide is known in 7 iron (c) and also of 1600 1 1500 1400 FeSisool 1200 p. 1 I 300 1 200 1 100 Solid Solution a Melt 1100 H^ Solid 1000 Solution! [[) 900 Melt 800 10 .700 >HI Solid 2 Liquids /3 Solid Solution D+ So l -9 "^iT Solution D+ OL a Fe S +ac.Fe Iron id Solution^ n V ron 1540 Fe 970 850 780 130 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Per cent by Weight Fe FIG. 79. iron in sulfide (D). The first separate on cooling with the precipitation of iron (along ik). Thereby the solid solution is enriched in sulfide up to a content of 4 per cent at 780, below this temperature it decomposes to a solid mixture of a ferrite and solid solution D which at 128 undergo an allotropic modi- fication into a sulfide. The complicated equilibrium diagram is shown in Fig. 79. Z. anorg. Chem., 49, 320 (1906). 122 THE PHYSICAL CHEMISTRY OF THE METALS Nickel Nickel Sulfide. The diagram for alloys of nickel and nickel sulfide studied by Bornemann* which is given in Fig. 80 is still more complicated. The only sulfide capable of existence in the molten state has the Temp. 1500 hNi 1400 1300 1200 1100 1000 FIG. 80. composition Ni 3 S 2 ; at low temperatures still further compounds exist namely NiS, NisS^ NiS2 and apparently also NieSs. As in all nickel containing systems, the tendency to form solid solu- * Metallurgie, 5, 13 (1908). ALLOYS OF METALS WITH CARBIDES, ETC. 123 tions is very great; there precipitates from the melt principally solid solutions which separate from the melt in a very complex way. Alloys Between Sulfides. In the metallurgy of copper, nickel and lead, there occur as important intermediate products, the so-called mattes, copper matte, lead matte and nickel matte, alloys of the corresponding sulfides with iron sulfide. It is of importance to obtain a picture of their important structural constituents. In many cases compounds occur between the components, in part these are such as occur in nature as ores, we have accordingly before us a field which must arouse the interest of the smelterman to a high degree. Here also metallographic methods give the explana- tion of the nature and formation of these products. In most cases the equilibrium diagrams are comparatively simple. Out of the melts, miscible in all proportions the com- ponents crystallize on solidification in the pure state. The diagram consists of two curves which cut in a eutectic point. This form we find in all alloys in which lead and zinc sulfide occur as components. The melting points of the components, the position of the eutectic temperature as well as the composi- tion of the eutectic alloy are given in the following table: COMPONENTS. MELTING POINT. Eutectic Tempera- ture. Degrees C. Composition of Eutectic. Observer. A. * ! A. Degrees C. B. Degrees C. PbS FeS 1114 1187 863 70% A PbS AgS | III4 835 630 77% A PbS CuS i III4 II2I 535 49% A ZnS ZnS PbS AgS : 1600 1600 1114 835 1044 807 6% A 3% A K. Friedrich* ZnS FeS 1600 1171 1162 5% A ZnS CuS 1600 II2I { near } I H2I / ? * Metallurgie, 4, 479 (1907); 4, 672 (1907); 5. 114 (1908). There are also cases in which solid solutions occur between the components; in the system copper sulfide, silver sulfide 124 THE PHYSICAL CHEMISTRY OF THE METALS which has been investigated by Friedrich * there appears to be present complete miscibility in the solid state. The solidifica- tion curve has a minimum at 677. The following table shows the dependence of the solidification point on the composition: COMPOSITION. COMPOSITION. AgjS. CuS. AgjS. CuzS. p.p. IOO 835 40 60 701 90 10 749 30 70 876 80 20 698 20 80 945 70 30 677 / IO 90 1050 60 40 688 IOO 1 121 50 50 719 According to the investigation of P. Rontgen there are a number of compounds which occur between the components of copper matte; those recognizable by a maximum in the solidi- fication diagram are (Cu2S)s.FeS, Cu2S-FeS and probably a third (Cu2S)2(FeS)5; also in the system silver sulfide antimony sulfide we find a maximum at the places on the diagram which represent the composition of the compounds Sb 2 S3Ag 2 S and Sb2Sa-3Ag2S. Both occur naturally, the first as miargyrite, and the last as pyrargyrite or ruby silver ore. The position of the melting point of the components and com- pounds (maxima) as well as the co-ordinates of the eutectic point are given for the two systems in the following table : Components and Compounds. Melting Points and Maxima., Degrees. Eutectic Temp., Degrees. Composition of Eutectic Alloy. Observer. Cu 2 S io8< QQC Cu 2 S, 21.8%; FeS, 78.2% P. Rontgen (Cu S) 3 FeS . . 106$ Cu 2 S FeS (CuiSMFeS),.... FeS Sb 2 S 3 1030 980 1133 562 1000 895 Cu 2 S, 67%; FeS, 33% Cu 2 S, 33; FeS, 67% Sb 2 8 3 ,Ag 2 S Sb 2 S 3 3 Ag 2 S 503 438 44-O Sb 2 S 3 ,82. S %,Ag 2 8,i7.5% SboSs 46 *% AffoS (- H 2 and further that the equilibrium mixture is richer in water vapor the higher the concentration of the oxygen. This conclusion is of importance if we think of our homogeneous system set in equilibrium with metal and metal oxide. By this contact the oxygen is brought to a constant value corresponding to the oxygen tension of the oxide. 154 THE PHYSICAL CHEMISTRY OF THE METALS In case of the complete equilibrium with all the substances together, the system must be in equilibrium as regards oxygen and the two other gases and with metal and oxide; we bring now all the constants on one side and obtain Co 2 C H2 o \K_ ^ We see therefrom that for the equilibrium between metal, metal oxide, water vapor and hydrogen, a definite ratio between the two gases is required. This is only true if the metal is always present as a solid or liquid phase; with volatile metals the con- centration of the metal vapor also enters into the constant. Since the absolute value of these concentrations and therewith also the pressure under which the gas stands is without meaning, then the ratio 77 can exist under different pressures. We, accord- ingly, obtain the same result which we had previously deduced with the help of the principle of LeChatelier. Every deviation of the gas composition from the ratio 77 requires reaction, if the ratio is greater than 77 a reduction of the oxide takes place, if it is smaller an oxidation of metal to oxide. If we represent graphically the conditions for oxidation and reduction in our system, as depending on the pressure and composition of the gas atmosphere at constant temperature, we obtain (see Fig. 93), as limits for the oxidation and reduction fields, the geometrical locus of all gas systems which are in equilibrium with metal and oxide. For all these the abscissa x which gives the number of hydrogen molecules (the sum of the hydrogen and water vapor molecules is equal to i) is defined by the equation x = 77. I X If x is constant, the curve is parallel to the p axis. The equilibrium constant 77 is, as we have above deduced, THE METALLURGICAL REACTIONS 155 and is dependent on the dissociation constant of water K, and also on the oxygen tension of the oxide. If we compare the value of rj for different metals, we see that rj grows with decrease of the decomposition pressure of the oxide, that the amount of hydrogen in the equilibrium ratio for a noble metal need be very small, for a base metal it must be under some conditions ex- traordinarily large. This agrees with the experimental fact that noble metals are not attacked by water vapor, their oxides Oxydation Reduction 10$ 20$ 30$ 40 70$ 80$ 90$ 100X, FIG. 93. are reduced by the smallest addition of hydrogen, and that with base metals the reduction of oxides with hydrogen is difficult. If more than one oxide is known for a metal, we obtain a number of values for rj and in the diagram a number of parallels. Concerning the other magnitude on which the constant 77 depends, the dissociation constant of water vapor, exact experi- ments have recently been carried out by Nernst and v. Warten- berg.* The dissociation of water vapor is first observable at very high temperature. It is determined by passing the water * Z. Physik. Chem., 56, 534 (1906). 156 THE PHYSICAL CHEMISTRY OF THE METALS vapor through a porcelain pipette heated in an electric furnace and measuring the resulting oxy-hydrogen in a fine eudiometer tube. For temperatures above 1800 the method worked out by Lowenstein * is used. The water is led through an electrically heated iridium tube, and is partially decomposed into oxy- hydrogen gas. The hydrogen partial pressure in the same is measured in by means of an evacuated iridium bulb connected with a manometer and placed in the reaction space. This metal is permeable to hydrogen but not to the other gases, the dif- fusion into the inside proceeds till the pressure of the hydrogen gas inside the bulb is equal to the partial pressure of the hydrogen from the dissociation of the water vapor. The manometer determines, accordingly, the hydrogen pressure directly. From the experiments the following values have been found. Water vapor at i atmosphere pressure undergoes by heating to the tem- perature T a decomposition of X per cent. r c. T Abs. X Per Cent. K. 1124 1397 0.0073 2.312.10-13 1207 1480 0.0189 3.794.10-12 1288 1561 0.034 2.094. lo-u 1882 2155 1.18 6.418.10-7 1984 . 2257 1.77 2.080.10-6 The dissociation is dependent on the temperature, and increases with increasing temperature. For low temperatures, the degree of decomposition is so small that one can no longer measure the resulting oxy-hydrogen gas. Van't Hoff's Equation. Thermodynamic methods are used to reckon the equilibrium which cannot be experimentally determined. Van't Hoff has deduced an equation which connects the dissociation constant with the reaction heat and the temperature. It is * Z. Physik. Chem., 54, 715 (1906). THE METALLURGICAL REACTIONS 157 if Q is the heat of combustion of the hydrogen. Now, however, the reaction heat is in no way a constant but changes with the temperature. The change is according to thermodynamics equal to the difference between the specific heats of the factors and the products. These latter magnitudes are dependent on the temperature and they have recently been exactly deter- mined by Holborn and Henning who have given the following relation for the molecular heats of the gases concerned (molecular heat = product of the molecular weight and the specific heat at constant volume C) for water vapor. C P (H 2 O) = 5.62-0.00077T The molecular heats of the permanent gases oxygen and hydrogen are equal and C, = 4.68-o.ooo267\ From this it follows that the temperature coefficient of the reaction 2H 2 -fO 2 = 2H 2 O is - 2C,(H 2 0) = 2.8o-o.ooo 7 6r. The value of the heat of combustion at 100 and constant volume to form two molecules of water is Qioo= 11,500 cal. The tem- perature function is If we substitute this value in the above differential equation and integrate it follows if we replace natural logarithms with Briggsian log tf' = log -25,030--1.40 log -o. If one value of K has been determined experimentally we can, with the help of this equation, arrive at the value of K for every other temperature. From this the per cent of decomposed water vapor (X) can be calculated at every other gas pressure, 158 THE PHYSICAL CHEMISTRY OF THE METALS from the mass law; the values given in the following table have been arrived at in this way : T Abs. X Per Cent. K. T Abs. X Per Cent K. 1000 3.00.10-5 2. 245. IO-2O 1800 0.199 3.628. 10-9 IIOO I . 82 . 10-4 4.5l6.IO-l8 1900 0-354 1.937.10-8 1 200 8.13.10-4 3.693.10-16 2000 0.588 8.461.10-8 1300 2.91.10-3 1.564.10-14 2IOO 0-935 3.251.10-7 1400 8.61. 10-3 3.763.10-13 22OO 1.42 i .092. 10-6 1500 2. 21. 10-2 5.944.IO-I2 2300 2.08 3.311.10-6 1600 5,07.10-2 6.733.IO-II 2400 2.92 8.847.10-6 1700 0.105 5.633.IO-IO 2500 3.98 2.225.10-5 TAbs. P = io Atm. X Per Cent. P=i Atm. X Per Cent. P=o.i Atm. X Per Cent. P =0.0 1 Atm. X Per Cent. 1000 I.4Q.IO-5 3.000.10-5 6.46. 10-5 1.39.10-4 1500 I.03.IO-2 2. 21. 10-2 4.76.10-2 0.103 2000 0.273 . 0.588 1.26 2.70 2500 I. 9 8 3.98 8.16 16.6 From the differential equation it can be at once deduced whether K increases or decreases with rising temperature. If Q is positive, that is, during the dissocia- tion heat is supplied, K increases with increasing temperature. If the process goes on with heat evolution, if heat is given to the surroundings from the system, the differential quotient is nega- tive and K is smaller with rising temperature. " The equilibrium of a chemical system is displaced by raising the temperature in the sense that the product resulting from heat absorption is favored." This law can be derived from LeChatelier's principle, since also in thermal relations chemical equilibrium systems are elastic, and by cooling the equilibrium swings back and liberates the amount of heat which has been added. We are now oriented concerning the constant K, for the reaction between hydrogen and oxygen. If we combine it in THE METALLURGICAL REACTIONS 159 the above-named way with the oxygen dissociation pressure of oxides, we are, able to obtain mathematically the equilibrium ratio rj for all these oxides. In practice there comes into consideration only the reaction with which we started and to which we now return. Fe 3 4 +4H 2 <=* 3 The equilibrium has been investigated by Deville * and more recently by G. Preuner.f We give in the following table the values which appear to be the nearest correct. Temperature in Degrees. CH 2 O 77 ~ CH 2 Per Cent H 2 Observer. 2OO 20.41 95-32 265 14.49 93-56 1 360 440 8.405 5-682 89.39 85-06 Deville, Liebig's Ann., 157, 71 (1872). 770 -852 64.94 Q2O 515 60.23 900 449 59-27 ' 1025 .282 56.18 Preuner, Physik. Chem., 47, 416 (1904) 1150 163 53.76 J The value of t\ decreases with rising temperature; we may draw the conclusion therefrom that the reduction of FesC^ goes on with heat absorption. Reduction by Cat bon Monoxide. The second gaseous reducing agent, carbon-monoxide, plays in practice .a still greater role than hydrogen. In general there exists between the two reducing gases close agreement in proper- ties. In the treatment of oxides with carbon monoxide we meet a reversible action precisely as in the treatment with hydrogen. This can be represented by the general scheme. MeO+CO <= Me+C0 2 . This decomposition also takes place without change of volume; the reaction products occupy the same space as the factors. * Ann., 157, 71 (1872). f Z. Physik. Chem., 47, 416 (1904). 160 THE PHYSICAL CHEMISTRY OF THE METALS The reaction is independent of the pressure of the gas and comes to rest when a definite ratio between carbon monoxide and diox- ide exists; which ratio is dependent on the temperature.* This equilibrium ratio = Cco 2 * \'Co 2 ' is dependent, as the corresponding magnitude for hydrogen, on the oxygen pressure of the oxide to be reduced and the dissoci- ation constant of C0 2 , K', which is defined by the equation ~, C 2 co-Co 2 "-- Then also the combustion of CO into C0 2 is a reversible reaction. Carbon dioxide decomposes at high temperature into CO and O 2 . This can be represented by the equation 2CO+O 2 <=> 2C0 2 . The equilibrium ratio of this system has been studied by Nernst and v. Wartenberg, by the same method as was described for the dissociation of water vapor. They observed the following * This again is only true for non-volatile metals. If we consider, for example, the system ZnO+CO=Zn+CO 2 , we find that since at temperatures where the reaction goes on, zinc only exists as a vapor, the ratio of CO to COz is therefore not dependent only on the tem- perature but the equilibrium constant for the system is given by ZnXCO 2 _ ir ~CO~ and hence the concentration of zinc vapor must be taken into account. In the various unsatisfactory investigations which have been made of the zinc equilibrium this has not been taken into account. See, for example, the recent paper by Fulton, Bull. Am. Inst. Min. Eng., 140, 1375 (1918); also the earlier paper of Lencauchez, Mem. Soc. Ing. Civils, 1877, 568; Eng. Min. J., 26, n (1878). THE METALLURGICAL REACTIONS 161 decomposition values, x-per cent decomposed CO 2 at the given temperatures T abs. x per cent 1300 0.00414 1400 O.OI O.O2 1478 0.029-0.035 From these values and the thermal data the heat of combus- tion of CO, 6800 cal. (at constant pressure) the molecular heat of C0 2 . and for the permanent gases CO and O2 the dissociation constant K and the per cent of decomposed gas x for different temperatures and pressures, is calculated in a similar way as for the H 2 O dissociation. Tabs. Degrees. X Per Cent. K'. Tabs. Degrees. X Per Cent. K Pressure I Atm. 1000 1.58.10-5 3. 280. IO-2I 1800 0.507 6.016.10-8 IIOO 2.00.10-4 5.99I.IO-I8 1900 0.978 4.III. 10-7 1200 8.94. 10-4 4.908.10-16 2000 1.77 2.335.10-6 1300 3.89.10-3 3.735.IO-I4 2IOO 3-03 1.130.10-5 I4OO 1.38.10-2 I.550.IO-I2 2200 4.88 4.595.10-5 1500 4.06. IO-2 3.684. io-ii 2300 7-55 1.675.10-4 I600 0.104 5.813. 10-10 24OO 11.3 5.631.10-1 I7OO o. 242 6.905.10-9 2500 15-8 1-552 Tabs. Degrees 10 At. i At. o.i At. o.oi At. =Pressu:e. IOOO 7.31.10-6 1.58 3-40 7-31=*% 1500 1.88 4.06 8.72 0.188 2OOO 0.818 1.77 3-73 7.88 25OO 7.08 15-8 30-7 53-o The equilibrium ratio has been determined for the system iron oxide, iron, carbon monoxide, carbon dioxide. The reaction FeO+CO<=Fe+CO 2 , 162 THE PHYSICAL CHEMISTRY OF THE METALS gives the following values : Temperature- Degrees CO, Per Cent. CO2, Per Cent. CO C0" 2 =77 - Observer. 552 53-7 46.3 .160 556 53-4 46.6 . 146 Schenck, 561 ' 53-6 46.4 155 * 596 55-5 44- S .247 Semiller and 6lp 56.8 43-2 315 6 5 I 57-9 42.1 375 Falcke 662 58.4 41.6 .404 The equilibrium constant increases with increasing tempera- ture, that is, the left side of the reaction is favored as the reduc- tion takes place with heat evolution. The reaction Fe 3 4 +CO => 3FeO+C0 2 , also has a measureable equilibrium which has been determined by Baur and Glasner.* They found at 581, 41.9 per cent CO and 57.4 per cent CO2 corresponding to 77 = 0.721. The reversibility of the action of CO on iron oxide, the oxida- tion of metal by C02 is of the greatest importance in the blast furnace process which we shall consider in the next chapter. It gives us the explanation of the remarkable fact that the carbon monoxide used as a reducing agent does not go completely to CO2 but occurs in rich amounts in the exit gases of the blast furnace and becomes a valuable by-product. We have become familiar with the characteristics of the two gaseous reducing agents CO and H2. We will find which is the strongest reducing agent. The question is now easily answered since we are familiar with the dissociation ratios of the oxidation products of H2 and CO. The strongest reducing agent is the substance whose oxidation product is the least dissociated, that is, whose dissociation constant possesses the smallest value. In order to make clear the comparative ratios of the two reducing gases we plot the value of X or K for water vapor and carbon dioxide or better their logarithms against temperature, we * Z. Physik. Chem., 43, 354 (1903). THE METALLURGICAL REACTIONS 163 obtain two curves (Fig. 94). These cut at a temperature of 1140 abs. or 867 C. Here the H 2 and C02 are equally strongly dissociated, hydrogen and CO are equally strong reduc- ing agents, beneath this temperature the CO reduces more ener- getically; above it the reduction is more active with hydrogen. This, however, does not exhaust what we have to say con- cerning CO. The gas undergoes, under definite conditions, a charateristic decomposition giving as the products C0 2 and solid C. This reaction which was first observed by St. Clair Deville * Dissociation of C 2 and H 2 Pressure 1 Atm. x per cent decomposed Gas 1000 1200 1400 1600 1800 2000 2200 2400 2600 Temp.abs FlG. 94. does not proceed alone but requires the presence of certain con- tact materials. The metals of the iron group Ni, Co, Pe and Mn, are especially active in starting the decomposition process and, indeed, they do this without being thereby changed. Their action is the greater, the greater the surface. We are familiar with the reverse of the above process. The method by which CO is obtained by leading CO 2 over glowing C is a very useful one and plays an important role in the formation of producer gas. The reaction Compt. rend., 59, 873 (1854). 164 THE PHYSICAL CHEMISTRY OF THE METALS is accordingly reversible, and runs to completion in neither direction but comes to equilibrium between the concerned substances. Concerning this equilibrium of two components C and divided between two phases the solid C and the gas mixture, the phase rule says that it possesses a two-degree free- dom, that is, the two phases can exist together under different temperatures and pressures. Concerning the direction in which the composition of the gas is changed by change of pressure, LeChatelier's rule says, through raising the pressure that reaction is favored which goes on with a decrease of volume. It is the decomposition of CO since two volumes of this gas give one volume of CCb, the simul- taneously formed solid substance is neglected in comparison to the gas. The CO can accordingly be decomposed in the pres- ence of C at a constant temperature by compressing the gaseous atmosphere.* By lowering the pressure the C02 acts on the carbon and forms the substance with the greater volume, that is the CO again (naturally these reactions take place more quickly in the presence of a contact substance). At constant tempera- ture every gas pressure corresponds to an entirely definite ratio of the two gaseous oxides. The dependence of this ratio on pressure can be determined from the law of mass action. We consider again the reaction in the gas phase and assume that a small immeasurable quantity of carbon vapor is in it. In a similar way as above for the dis- sociation process, the relations for equilibrium can be devel- oped. C 2 co Cc-Cco 2 = Const. The concentration of carbon vapor that is in equilibrium with the solid phase is, at constant temperature, a constant magnitude, * For the effect of pressure on the reaction 2 CO *=> C+CO 2 , see Rhead and Wheeler, Trans. Chem. Soc., 97, 2181 (1910). These investigators find a very close agreement between the effect of pressure as experimentally de- termined and as calculated from the mass law. THE METALLURGICAL REACTIONS 165 depending on the vapor pressure of the carbon. If we combine this constant with the constants on the right side and designate the magnitude obtained by we obtain as the equilibrium rela- tion for the two oxides of carbon, if they are in contact with solid C the equation C 2 co r Cco 2 * We will now make the assumption which we have previously employed that the sum of the partial pressures of the two oxides is constant; we set this equal to i. If the number of monoxide molecules present is x the dioxide molecules are i x. If we set the total pressure of the two gases equal to P the partial pressures are x-P and (i x) P respectively. The concentrations of the two gases are proportional to these magnitudes. If one substi- tutes them in the equation it follows: or - IX X 2 These equations give us the relation between the composi- tion of the gas atmosphere (x.ioo is the CO per cent) and the pressure under which it stands in case there is equilibrium between the substances present. In the discussion of the equations it is brought out that x increases with rising values of P. The requirements of Le- Chatelier's principle which we have previously spoken of are accordingly fulfilled. To get a conception of the existing rela- tions we will represent them graphically. The given equation represents a curve of the third degree. Its position is shown in Figs. 95 and 950. Since in our case the abscissa x cannot be greater than i, only that part of the curve with the abscissa between O and i has a real meaning. For later consideration it is well to give here its further extent. In Fig. 950 the real part is drawn. For x 2 the curve reaches a minimum, with increas- ing x it passes through a point of inflection and approaches the abscissa axis asymptotically intersecting it at x = o. We have, accordingly, a cubic hyperbola. This form of curve is met with in 166 THE PHYSICAL CHEMISTRY OF THE METALS many homogeneous equilibrium systems, for example, in all binary dissociation processes it has considerable importance. %co FIG. 95. FIG. 950. The curve which represents the geometrical locus of all pos- sible mixtures of monoxide and dioxide that are in equilibrium with solid carbon, is accordingly an isotherm dividing the field THE METALLURGICAL REACTIONS 167 (Fig. 95) in two halves. In these, reactions take place, in the right decomposition and in the left formation of CO. We find in the drawing still another curve (dotted) which begins at the point x,= i and P = Po and terminates at x = o and P = it is the reaction path of the decomposing of CO, 2 i.e., the series of points through which the decomposing CO of original pressure PQ passes during its decomposition. If we carry the reaction entirely to the end the complete transforma- tion into carbon dioxide is reached when the gas pressure has become half of the original pressure. The form of this reaction path is indicated; the partial pressure of the still present monox- ide in the decomposing gas p stands in the following relation to the original pressure PQ and the observed total pressure P: then the observed difference from the original pressure repre- sents the double difference of the decomposing CO. Now = x-P and thence follows This equation is represented geometrically by a hyperbola. If we consider P as variable we obtain a series of hyperbolas for which P is infinite for the case x = 2. The reaction path is also a section of a hyperbola. This usually reaches its end as the equilibrium between the finely divided carbon and its oxides is reached, i.e., in the intersection of the equilibrium curve and the reaction path. The equilibrium between carbon and its oxides is now, as practical experience in the preparation of generator gas has also shown, strongly dependent on the temperature and at high temperatures very much more monoxide is observed than at low temperatures. From this we draw the conclusion that the forma- tion of carbon monoxide is a process which goes on with heat absorption. Correspondingly heat is evolved in the decomposi- tion of the monoxide, the thermochemistry of the decomposition reaction is expressed by the following equation. 2CO=C+C0 2 -39,ooo cal. 168 THE PHYSICAL CHEMISTRY OF THE METALS there is accordingly a very considerable amount of energy freed in the form of heat. With increasing temperature the equilibrium constant also increases. Its change with the temperature can be gotten from measurements of Boudouard * on the composition of monoxide dioxide mixtures, that are in equilibrium with carbon at atmos- pheric pressure and different temperatures. The following table gives the values found from the observations and thermo- 1100 1000 900 800 700 600 500 C 2 CO^ Pressure I Atmosphere 90 100$ CO 10 20 30 40 50 60 70 I FIG. 96. dynamic calculations. The dependence on temperature is shown still better by the graphical representation of the results. (Fig. 96) Temperature, Degrees. Co, Per Cent. C0, Per Cent. Temperature, Degrees. CO, Per Cent. COj, Per Cent. 450 2 98 800 90 IO 500 5 95 850 94 6 550 ii 89 900 96.5 3-5 600 23 77 925 97 3-0 650 39 61 950 98.5 i-5 700 68 32 IOOO 99-3 0.7 750 76 24 1050 99.6 0.4 * Ann. Chim. Phys., (7), 24, 5 (1901). More recent and accurate measurements of this equilibrium have been made by Rhead and Wheeler, Trans. Chem. Soc., 97, 2181 (1910). For a discussion of the results of these investigators from the standpoint of thermodynamics see Lewis and Randall, J. Am. Chem. Soc., 37, 458(1915). THE METALLURGICAL REACTIONS 169 The values of f vary still further with the different modifica- tions of carbon. Graphite gives different results from amorphous, wood or sugar carbon, or the modification resulting from the decomposition of the monoxide. The ratio between these equilibrium constants can be determined in a way which we shall consider later. Schenck and Heller * found the following numbers : Temperature, Degrees. Graphite. CO-carbon. Sugar Carbon. 600 660 I I 5 6 5-5 6.6 It is accordingly not immaterial in the preparation of gener- ator gas whether a graphite like coal is used or an amorphous form, the latter is to be desired since it gives gas richer in a carbon monoxide under the same conditions. These relations can be at once seen if we recall the develop- ment of with the help of the mass law. We see that the con- centration of carbon vapor which is in equilibrium with the car- bon enters into these magnitudes. This carbon vapor pressure is different for the different carbon modifications. We can get from this as Smits f has done that the relations between the values of f are the direct relations between the vapor pressures of the different carbon modifications. We have considered rather fully the relations of carbon monoxide. We must, however, study its peculiarities, its rela- tion to carbon dioxide on the one hand, and to carbon on the other since in all reduction with the help of carbon this oxide occurs as a reaction product and frequently influences the course of the reaction. Ber., 38, 2139 (1905). t Ber., 38, 4027 (1905). CHAPTER V DECOMPOSITION OF CARBON MONOXIDE, BLAST- FURNACE PROCESSES We have already, in a previous chapter, learned of the remarkable decomposition which carbon monoxide suffers under certain conditions, namely, its decomposition into carbon dioxide and elementary carbon. We have seen that this process can only take place in the presence of contact material and that the finely divided metals of the iron group, nickel, cobalt iron and manganese serve as such. We have also determined the reac- tion path of the decomposing monoxide and the place at which th'e decomposition halts. Catalytic Decomposition of Carbon Monoxide. All these theoretical conclusions are confirmed if the decom- position is carried through at a constant temperature, near the boiling point of sulfur, using as a catalyst finely divided nickel or cobalt. Fig. 97 shows the change of pressure with time, of a given amount of carbon monoxide at constant volume, during the decomposition reaction in the presence of nickel and iron. This has been observed by Schenck and Zimmermann in their investigation on the decomposition of carbon monoxide. As is seen, there is a considerable difference in the action of the two metals. With nickel the reaction comes to a standstill, when the equilibrium between carbon and its two gaseous oxides is reached as we should expect from our earlier experiences. The final pressure reached should, therefore, not be under half of the original pressure of the pure carbon monoxide. To our great astonishment, we meet with iron an entirely different condition, a very large pressure decrease, that we can 170 DECOMPOSITION OF CARBON MONOXIDE, ETC. 171 only explain by a transition of the gas into a solid product. Sometimes the final pressure is only a small per cent of the orig- inal pressure. If we investigate the contact mass more care- fully it is found that along with the precipitation of carbon an oxidation of metallic iron has taken place. We encounter the paradox, that iron is oxidized by the strong reducing agent, pure carbon monoxide. At first sight this fact appears contrary to all we have learned in the previous lecture. i Pressure m m. Nickel 445' 50 100 150 .200 250 300 350 400 450 500 .Minutes FIG. 97. The entire problem is apparently thereby complicated since with other original pressures the decomposition goes as in the pres- ence of cobalt and nickel. The solution of the problem is reached when one thinks that from the decomposition of carbon monoxide, the dioxide results, and that its concentration can, under some conditions, grow so great that the equilibrium ratio of the two gases with iron and iron oxide is exceeded. It then follows that the metal is oxidized with the formation of carbon monoxide, and, on the other hand, this gas decomposes again with the precipita- tion of carbon. By the alternation of the two processes, oxi- 172 THE PHYSICAL CHEMISTRY OF THE METALS dation of the metal and the decomposition of the carbon monox- ide, all the carbon as well as all the oxygen can go over to the solid state. The reaction path (see page 133 and Fig. 98), is under these conditions complicated further than with the simple decomposition in the presence of nickel. If we assume the same original pressure and temperature for the reaction in the presence of iron and of nickel then the two hyperbola limbs fall together until the ratio of the two gaseous oxides of carbon has reached a value at which oxidation of the iron results. In the presence of nickel, which is not oxidized, the reaction proceeds along the %CO FIG. 98. hyperbola till the curve is reached which represents the equilib- rium between carbon and its oxides. In the presence of iron there is a break in the reaction path, the pressure decreases strongly while the composition of the gases approaches a value at which it is in equilibrium with carbon and with iron and iron oxide. We will see further on, that it depends on the position of this total equilibrium and on the original pressure of the carbon monoxide whether an oxidation of the metal by the gas takes place or whether the metal remains unoxidized and acts as con- tact material for the decomposition of carbon monoxide into carbon dioxide and solid carbon as nickel does. DECOMPOSITION OF CARBON MONOXIDE, ETC. 173 Equilibrium Between Carbon Monoxide and Iron. The total equilibrium between iron, iron oxide, carbon and the two gaseous oxides, carbon monoxide and carbon dioxide is rep- resented as an equilibrium of a three-component system (com- ponents, carbon, iron and oxygen) with the simultaneous pres- ence of four phases (Fe, FeO, C, gas). According to the phase rule such a system must possess one degree of freedom, the equilibrium is univariant precisely as a vapor pressure. Every temperature corresponds to an entirely definite equilibrium pressure. We arrive at the same results by the use of the mass law. We have for our case, the equilibrium conditions for two bivari- ant systems 1. FeO+CO<=>Fe+C0 2 , 2. 2CO<=C+C0 2 . that must be simultaneously fulfilled. The first may, as we have seen earlier, be expressed by the equation and second by x 2 2. i-x If we eliminate x from these two simultaneous equations we obtain an expression for the total pressure of the two gases P in which this magnitude depends only on the equilibrium con- stants t] and It follows from this so long as f and TJ are constants for a given temperature, that every temperature corresponds to an entirely definite gas pressure, a definite sum for the partial pressures. From Eq. (i) it follows further that also the composition of the gas phase is absolutely fixed. The facts can be made very plain if we show the equations graphically. 174 THE PHYSICAL CHEMISTRY OF THE METALS We have already recognized a parallel to the ordinate axis as a geometrical locus of the equilibrium between metal, oxide and the two gases, and a cubic hyperbola as the geometrical ex- pression of the equilibrium between carbon and its oxides. The two intersect, and at the intersection, the two equilibria are simultaneous, it represents the total equilibrium between metal, metal oxide, carbon and the two gaseous oxides (see Fig. 98). The two equilibrium lines, straight and cubic hyperbola divide the plane of our picture into four fields, in these the following reactions take place: Field I. 2CO-+C+C0 2 FeO+CO->Fe+C0 2 Field II. 2CO-C+C0 2 Fe+C0 2 -Fe+CO Field III. C+C0 2 -2CO Fe+C0 2 ->Fe+CO Field IV. C+CO 2 -+2CO FeO+CO-*Fe+C0 2 The reduction of iron oxide or generally of oxide to metal accordingly goes on only in fields I and IV. Carbon monoxide can act as a reducing agent inside of both. Solid carbon can only act as a direct reducing agent in field IV. In field II the two sub- stances carbon and oxide, remain together, in field III the carbon gradually disappears without thereby attacking the metal oxide. Of all points of the PX diagram the solid phases, metal oxide and carbon, are only present at the point of total equilib- rium, the intersection of the curve with the line. In which way this occurs we have already seen in the consider- ation of the decomposition of carbon monoxide in the presence of iron. We have there confined ourselves to a special case ; we have already mentioned, however, that the decomposition of the pure carbon monoxide in the presence of iron can take place so that no oxidation of metal results. Now we will determine the conditions under which the two forms of carbon monoxide decom- position take place. DECOMPOSITION OF CARBON MONOXIDE, ETC. 175 The original pressure of the pure carbon monoxide is decisive for this, on it depends whether the hyperbolic reaction path (see above) lies above or below the intersection of the cubic curve, with the line. We have learned of the first case above, the reac- tion path meets first the line and suffers a deflection going into field I and field II where oxidation of the metal results. If the hyperbola runs beneath the equilibrium point, so it strikes the decomposition curve first, and with the attainment of the simple carbon equilibrium the reaction comes to rest, since inside of fields I and IV the oxidation of the metal is impossible. The limiting case between the two is given if the reaction path cuts directly at the point. It can be easily calculated what the original pressure of CO was when this condition is satisfied. The equation of this hyperbolic reaction path has been derived in the fourth chapter. It is P = (2-X)P. If we substitute in this equation the ordinates of the total equilibrium, namely, and " *=. it follows for the limiting case: M a -^) f ?~ r If the original pressure of the carbon monoxide is greater than this value, the metal is itself finally oxidized, if it is smaller the metal acts only as contact substance. The position of the equilibrium point and the magnitude of the equilibrium pressure are dependent on the nature of the metal, the carbon modification present and the temperature, since these factors are determinative for the magnitudes f and rj. If we consider now the relations at constant temperature we see that with the same modification of carbon (constant ) P becomes smaller with increasing 77. 176 THE PHYSICAL CHEMISTRY OF THE METALS The less noble the metal is, the smaller will be the equilibrium pressure for the reduction of its oxide with carbon. It is the higher the more noble the metal. Now we understand why the nickel acts only as a contact material toward the carbon monoxide ; why it is not, as under certain conditions with iron, oxidized by the decomposing gas. Nickel is somewhat more noble than iron and we may assume that its equilibrium pressure is many atmospheres, so that under ordinary working conditions the decomposition hyperbola of the carbon monoxide runs far beneath this pressure. Iron is the only metal with which this pressure can be conveniently observed, with the more noble it is so high that it is difficult of measure- ment and with the less noble it is exceedingly small. If we consider only one metal (constant TJ) but different modi- fications of carbon, the equilibrium pressure is the greatest, with the form with greatest f. Accordingly a greater equilibrium pressure corresponds to the equilibrium with amorphous carbon than with graphite. The magnitudes of the equilibrium pressures, for different forms of carbon are, as can be seen from the equation directly proportional to the values of . Both 77 and are dependent on the temperature and, indeed, they both increase, as we have seen earlier, with rising tempera- ture. Whether P also increases with the temperature, cannot be derived from the equation directly. It depends on the ratio of the temperature coefficients of the two equilibrium constants. Experiment has shown that the pressure increases with the tem- perature. For the determination of the equilibrium pressure, several ways are available. We can reach it as we have seen above, if we allow carbon monoxide at a sufficiently high original pressure to react with metallic iron at a given temperature. It is only necessary to follow the pressure decrease and the reaction comes to rest when the equilibrium pressure is reached. We arrive at the same value by gradual heating of the solid phases iron, carbon DECOMPOSITION OF CARBON MONOXIDE, ETC. 177 and ferrous oxide, in a previously evacuated tube connected with a manometer. Investigations concerning these reduction equilibria have been carried out by Schenck and Heller * as well as Schenck, Semiller and Falcke.f Their results are collected in the fol- lowing table. Besides the pressure values the table gives the composition of the gas phase. Temperature in Degrees. Pressure, Mm. COMPOSITION OF THE GAS PHASE. Observers. CO, Per Cent. CO 2 , Per Cent. 455 27 510 43 538 81 552 130 53-7 46.3 556 137 53-4 46.6 561 142 53-6 46.4 Fe, FeO, C am- orphous from CO 562 586 596 177 266 296 55-5 44-5 Schenck and Semiller 616 401 619 411 56.8 43-2 629 469 643 561 | 651 57i 57-9 42.1 662 662 58.4 41.6 670 858 408 5-6 465 10.4 Fe, FeO, sugar 560 161. 7 Schenck and carbon 590 3I4-5 Heller 627 546.8 649 750-1 500 12.3 536 27-3 Fe, FeO, graph- < 568 36.8 49-2 Schenck and ' Falcke ite 582 69-3 609 77-5 660 129.0 59-6 40.4 Ber., 38, 2131 (1905). tBer., 40, 1708(1907). 178 THE PHYSICAL CHEMISTRY OF THE METALS As carbon modifications, the amorphous carbon resulting from the decomposition of carbon monoxide, sugar carbon, and graph- ite are used. It can be seen from the equation, that at the same tempera- ture, differences of pressure due to variation of the carbon modi- fication do not also require differences in gas compositions. These changes influence only the values of f but not that of 77. Also the reduction of FesC^ to FeO by means of amorphous carbon has measurable equilibrium pressures which are collected in the following table: Solid Phases. Temperature in Degrees. Pressure in Millimeters. Observers. 528 121 535 129 Fe 3 O 4 , FeO, C (amorphous) 55i 560 about 175 303 Schenck and Falcke 568 about 430 58i 699 The Analysis of Mixtures of Graphite and Amorphous Carbon. We will now put the results of our theoretical and experi- mental investigations concerning the reduction of iron oxide by carbon to a practical test. We will first consider an analytical problem. The chemist has often been given the problem of determining quantitatively graphite in the presence of amorphous carbon. This problem has been impossible of solution with the previously available methods. Our intimate knowledge of reduction by carbon puts us in a position to arrive at a method which will reach the stated go'al. It can be directly deduced from the diagram, which represents the equilibrium between iron and ferrous oxide, as well as the different carbon modifications with the gaseous oxides of carbon at constant temperature. We show on Fig. 99 beside the vertical line for the iron equilibrium, also the equilibrium isotherm for amorphous and graphitic carbon. We select a suitable temperature of about 750. From this diagram we arrive at the conditions under which amorphous carbon will be oxidized by ferrous oxide while the graphite will remain unattacked. DECOMPOSITION OF CARBON MONOXIDE, ETC. 179 Our earlier considerations (compare page 140 and Fig. 98) show that a transformation of carbon into its gaseous oxides by ferrous oxide is only possible inside field IV as in Fig. 99, field IV for amorphous carbon is considerably greater than the corre- sponding one of graphite (curve G = graphite, A = amorphous carbon.) The remaining fields are not suitable for our purpose since both forms of carbon react with ferrous oxide. The shaded Pmm. Isotherms for 700 III 100 20 80 30 70 50 50 60 40 900 800 700 600 500 400 300 200 100 80 90 100$ CO 20 10 FIG. 99. field (IV) in which only the amorphous carbon is burnt to car- bon monoxide is however suitable. In the practical performance, we mix the sample of mixed carbon forms with ferrous oxide or more conveniently dry ferrous oxalate, which, by heating to 300 in an indifferent atmosphere, leaves ferrous oxide. If we now heat this mixture in a stream of mixed carbon monoxide, and dioxide, the gas pressure and composition represents a point in field IV, so we have the con- ditions under which the amorphous carbon is burnt. If we 180 THE PHYSICAL CHEMISTRY OF THE METALS choose for the performance a temperature between 700 and 720, which can be held sufficiently constant by means of a Hergeus resistance furnace, we can work under atmospheric pressure. A convenient composition of the gas mixture is 65 per cent monoxide and 35 per cent dioxide. The Blast-furnace Process. Further, our discussion concerning reduction forms the basis for the theory of our most important technical process, the iron- blast furnace. As we all know the blast furnace consists of a shaft of double conical form in which the iron ore to be reduced is interlayed with carbon and slag-forming material in the upper cone. The under bowl sets with its narrow part in the " frame," in which the supply for the hot blast, the so-called " tuyeres," are built in. The pit still beneath forms the hearth in which the blast furnace products, liquid pig iron and above this the slag settle, the latter flows continuously through a side opening, while the metal is drawn from time to time. The process with which we are concerned takes place between the tuyeres and the throat. In front of the tuyeres the carbon is burnt by the oxygen of the air and as a result of the pre- heating of the air there is a very high temperature of combustion, which is above 1100, the combustion product is accordingly nearly pure carbon monoxide since the gas mixture, which is in equilibrium with carbon at this high temperature and the lowest possible pressure, of the combustion product contains only traces of dioxide. As at the ordinary pressure of 750 mm. the partial pressure of the oxygen is 150 mm. and since the two volumes result from one volume of oxygen the pressure maximum is ^ atmosphere (two volumes CO 4 volumes nitrogen) i.e., 250 mm. The sum of the partial pressures of CO2 and CO cannot exceed this pres- sure and, as a rule, it remains under this value. The gases containing CO come up from below and encounter the solid charge which melts in the hot zone and is displaced by fresh material from above. On their way the hot gases give up DECOMPOSITION OF CARBON MONOXIDE, ETC. 181 their temperature and finally enter into reaction with the heated ore and carbon. We obtain thus a temperature gradient inside of the furnace from below to above. At the throat of the furnace the gases still have a temperature of 400-500. In this upper zone chemical reactions do not take place but the moisture which the charge contains is vaporized and so a drying of the solid material results. With sinking into the hot sphere the reduction begins, first, that of the higher oxides to ferrous oxide, then that of ferrous oxide .to metal. This is cemented by a further action of the carbon monoxide and carbon. Somewhat below the tuyeres the temperature finally becomes so high that both the iron and the slag melt. This liquid mass collects in the hearth and the task of blast fur- nace, namely the preparation of pig iron, is finished. The problem is to conduct the process so that the single processes which go on from oxidized ore to the iron carbon alloys, take place smoothly and without disturbance. The right tem- perature and the right composition of the gas stream are of the greatest importance to reach this goal. Before we go into the possible disturbances we will turn to an important by-product of the blast furnace, the throat gases. Still at the beginning of the ipth century the gases were simply allowed to escape from the throat, it was, however noticed that they contained a large amount of burnable substance which, if allowed to go to waste, meant the loss of large heat values. In 1836, Fabre du Faure sought to remedy this disadvantage and by the use of the throat gas, for regenerators, brought the blast fur- nace process to a rational form. At this time the demands on iron due to the introduction of railroads and steam machines and interest in cheap prices for the raw materials was growing steadily, and caused everything necessary for economic production of the pig iron to be energetically considered, and occasioned also in 1836 the Kurfurstliche Hessische Bergdirektion zu Kassel, to put to the professor of chemistry at Marburg, Robert Bunsen, the problem of making a close investigation of the blast-furnace process. The title of his communication on the results of his research 182 THE PHYSICAL CHEMISTRY OF THE METALS in Poggendorffs Annalen, " Concerning the gaseous products of the blast furnace and its use as burning material," betrayed at once from which side he attacked the problem. It was on this occasion that he worked out his well-known methods of gas analysis which he used for the study of the composition of the blast-furnace gases. He confirmed their high CO content and calculated that under certain conditions up to three-fourths of the carbon used was lost if the gases were allowed to escape unused, into the atmosphere. The amount of combustible constituents is in fact large and the following small table gives the average composition of the blast-furnace gases: Nitrogen 54~66 per cent Carbon Dioxide 7~iQ Carbon Monoxide 21-31 Hydrogen i- 6 Hydrocarbons 0-6 The technic has made use of this knowledge and the gases are led out and used in the rational way. In part, they are burnt in the Cowper apparatus and so warm the blast which the furnace receives through the tuyeres and in part they are used to heat steam boilers. Later experience has taught that their energy content is best used, if they are purified and mixed with air in gas engines. The useful effect is here much greater than by the use in connection with boiler and steam engines. Blast-furnace gas motors of very large dimensions are now used. They fur- nish the mechanical energy for the operation of steel and rolling mills, as well as drive dynamos and furnish light and power for a wide circle of our industrial operations. The exploitation of this original by-product has been of the greatest technical im- portance. Yet before we had progressed so far, the question was fre- quently asked whether it was possible to use the strongly reducing gas for further reduction of ore. Large means have been used to lengthen the layer of ore through which the gas passed. Blast furnaces have even been built 30 meters high; however, no matter how high they were the CO content of the throat gases was not DECOMPOSITION OF CARBON MONOXIDE, ETC. 183 appreciably reduced. This negative result of costly experi- ments cannot be wondered at by one conversant with chemical equilibrium, he knows that the reduction of iron is a reversible reaction, he knows that by the action of FeO on CO a certain amount of the gas must remain over and that the remainder does not come into consideration in the least for the preparation of iron. However, this scientific knowledge is unfortunately younger than the troubles concerning the utilization of blast- furnace gases. The cooler is the zone in which the gases arrive the higher rises the content of carbon dioxide. This is not alone due to the reduction; the same phenomena would be observed if the ore in the furnace were entirely lacking. The influence of the temper- ature on the composition of the gases is due to its effect on the equilibrium of carbon monoxide with carbon. With decreasing temperature the carbon monoxide formed at the tuyeres, breaks down more and more into dioxide and finely divided carbon. The presence of metallic iron greatly favors this reaction. The attainment of the theoretical equilibrium is, however, only par- tial, due to the great velocity of the gas stream. Therefore, the blast-furnace gases are always richer in carbon monoxide than corresponds to the equilibrium ratio and the zone temperature. Since the temperature at which reduced iron is in contact with the gas, is quite high, and the pressure of the total equilibrium between metal, oxide, carbon and the gases corresponding to these temperatures is over | atmosphere, we may, accordingly, designate the field of the equilibrium diagram that represents the pressure and the composition of the blast-furnace gases. According to the above laid-down rules only field I comes into consideration and, indeed, only that part in which the pressure is smaller than the pressure of the total equilibrium. We recog- nize therefrom that the reduction in the blast furnace is essen- tially by means of CO. Under certain conditions, for example, as a result of stop- pages in the blast supply, the region in which the ore is reduced to metal may be cooled to a temperature of 500 or still lower. At this temperature the pressure of the total equilibrium is less 184 THE PHYSICAL CHEMISTRY OF THE METALS than the sum of the partial pressure of monoxide and dioxide, the composition of the gases then represents a point in field II. Oxidation of the metal takes place and simultaneous precipitation of finely divided carbon, which stops the furnace. The approach of danger may be learned from the increase of C02 in the throat gases. The lowest temperature at which reduced metal may be in contact with the furnace gases, without reoxidation of iron taking place, is that at which the pressure of the total equilibrium is equal to the sum of the partial pressure of the carbon monoxide and dioxide in the gases. If we take this sum as 200 mm., the temperature above which no difficulty will ensue is 570. Be- side the reduction, there takes place in the furnace, the cement- ing of the iron, which is a very important process for the smooth operation of the blast furnace. As we have seen in an earlier lecture the melting point of the iron-cementite is very much lower than that of pure iron. This fact has the important practical consequence, that one can operate a furnace, in which pig iron is to be melted, at a much lower temperature than one in which pure iron is to be melted. This temperature difference is nearly 400, a condition which is very essential for the permanence of the furnace material. Concerning the conditions for the cementing of iron by car- bon monoxide we will supplement the experiments by means of theoretical calculation. We are limited for experimental mate- rial to an investigation of the question at temperatures up to 700, below this the formation of solid solutions of iron and car- bide does not occur, the results of these experiments give us only a general schematic picture of the run of cementation at higher temperatures and with solid solution formation. In laboratory investigations concerning the action of CO on metallic iron, it has frequently been observed that if the amount of carbon monoxide present is large compared to the amount of iron, that there are entirely different and much lower equilib- rium pressures than under the same conditions of temperature and original pressure with relatively smaller amounts of gas and larger amounts of metal. The former pressure value is near that which represents the equilibrium between iron, FeO, graphite and DECOMPOSITION OF CARBON MONOXIDE, ETC. 185 the gases and its origination is explained by the transition of labile amorphous carbon into the stable graphite. On the basis of our earlier experiences we have a means of testing this explanation for its correctness. If the difference of the twQ pressures is conditional on the carbon modifications, there must be at constant temperature, the same ratio of carbon monoxide and carbon dioxide, the monoxide content of gas must vary accordingly between 53 per cent and 60 per cent (see table p. 143). To our astonishment the content proves to be much higher, between 85 and 90 per cent. The difference of the two equilibrium systems can not accordingly be conditioned by the difference in the carbon modifications. Another possible explanation is that, in the place of FeO, another more difficultly reducible oxide has been formed, but there is no other evidence of this. The only logical assumption is that a change takes place in the iron phase. Since by the ac- tion of carbon monoxide on metallic iron, there takes place first, a transformation of the upper surface into cementite (FesC), according to the equation and this carbide is oxidized by further decomposition of the CO, as we have earlier seen with the metal, till finally there results a total equilibrium between cementite, ferrous oxide, amorphous carbon and the gas mixture. The observation data follow: COMPOSITION OF GAS. Tempera- ture, Degrees. Pressure, Millimeters. Observers. CO, CO 2 , Per Cent. Per Cent. 468 10.3 O.O O.O 540 30.0 600 6 Fe 3 C+4C0 2 . S 8 " b S -i 1 Fe,C, FeO, 00,002 Fe, Fe 0, CO, 8 FIG. 101. m m, The conditions for the first We already know, for the second mass law gives the constant v5 rP. -rH" The graphic representation for the first equation is the cubic hyperbola with which we are familiar, the second equation is represented by a similar curve of a higher order, which, beside the starting point (P = 0,x = i) cuts the carbon curve in still a second point. 188 THE PHYSICAL CHEMISTRY OF THE METALS The coordinates of this intersection can be deduced, in a very simple way, from the two equations which must be simulta- neously fulfilled. The division of the second by the first gives From this and from the first it follows: From the isotherm, in which we represent, beside the two above treated curves, still the vertical straight lines for the equilibrium: FeO+CO^Fe+C0 2 , with the equation, we see that the new curve (see Fig. 102) in its lower part is strongly curved and above it is a straight line nearly normal to the x axis. Its tendency toward the left is only small, never- less it causes an intersection of this curve of higher order with the equilibrium line for the system (Fe, FeO, CO, C02) at high values of P. (Compare here Fig. 103). At the intersection the two equations = r? and x v 4 -P = fl, I ~~~X \L X) are simultaneously fulfilled; from these the coordinates of the intersection follow: x , and P = DECOMPOSITION OF CARBON MONOXIDE, ETC. 189 This point represents a new univariant equilibrium not pre- viously treated and not directly observable, between the four phases, Fe, FeO, FesC, and gas. We may represent it by the symbol FeO+Fe 3 C <=> 4Fe+CO. The corresponding equilibrium pressure cannot be directly measured, but we are able to calculate it from the equilibrium 800 700 600 500 400 300 200 100 Pressure m.m. Isothermes for 650" 10 20 30 40 5,0 60 70 80 90 100 $CO FIG. 102. pressure of the systems we can observe. The calculation gives for the temperatures 650 and 700, the values P<>5Q = 5 I -9 2 atm.; P 700 = 166.3 atm. The composition of the gases is naturally the same as in the reduction of FeO, by carbon monoxide, forming metal and car- bon dioxide as the calculated points already belong to the equilibrium lines for this reaction system. Accordingly X 65Q o = 0.58 and X 70QO =0.60. 190 THE PHYSICAL CHEMISTRY OF THE METALS The calculation of this not directly observable equilibrium is not as it might appear idle play. The coordinates found in CO i.e. FIG. 103. this way give us a knowledge of a magnitude that is of great practical importance. It makes it possible to give the condi- DECOMPOSITION OF CARBON MONOXIDE, ETC. 191 tions under which metallic iron is cemented by carbon monoxide. The above-determined equilibrium point is not only the inter- section for the curves of the bi variant system (Fe, FeO, CO, C02) and (FeaC, FeO, CO, CQ2). Since in it three solid phases are in equilibrium with the gas, the equilibrium curve for the third possible combination (FesC, Fe, CO, CO2) must also pass through it. The equilibrium 3Fe+2CO <= Fe 3 C + C0 2 also exists at the calculated point. This equilibrium is also bivariant since it consists of three components (C, Fe, O) simultaneously present as three phases. . The equilibrium curve has, exactly as for the coexistence of carbon in contact with its gaseous oxide, the equation it is accordingly a cubic hyperbola. The constant ju is connected with the two constants 77 and # through a simple relation. For the two intersection points, we have Pi~\"n j r> O i~l~*? = )u , ana *?.**& . From this follows: For 650 #=83,240 and 17= 1.381, for 700 5 = 384,000 and 77 = 1.500 hence M650 = 13600; M700 = 113,800. From these values and the equation x" IX P, the composition of the gases can be calculated which are in equilibrium with iron and cementite at a determined pressure; they are = -96, and #700 = 0.99. 192 THE PHYSICAL CHEMISTRY OF THE METALS At the temperatures 650 and 700 accordingly a transforma- tion of iron into cementite is only possible with such mixtures of CO and C0 2 which contain more than 96 and 99 per cent CO, respectively. The position of the cementation curve is shown in the isotherm for 650 (see Fig. 102) and one can get from this the necessary CO concentration for carbide formation at other pressures. From the equilibrium diagram for 650 (Fig. 102) we know further that ferrous oxide goes into cementite easier than the metal; 86 per cent being sufficient for that reaction. In the previously considered cases pure cementite always results. One dare not draw the conclusion from these investi- gations that the conditions for cementation always require so high a content of monoxide. If the temperature allows of the formation of solid solutions between iron and carbon the relations are somewhat different. The influence of the solid solution for- mation on the CO concentration necessary for the cementation can also be seen from the schematic diagram (see Fig. 103). In this diagram we recognize again our four curves for the bivariant equilibria: (A) 2CO ^C + C0 2 (B) FeO+CO <=Fe+C02 (C) 3 FeO + 5 CO <=> Fe 3 C+ 4 C0 2 CD) 3Fe+2CO <=Fe 3 C+C0 2 . including their intersections. If some of the iron dissolves car- bide so must the lines representing the equilibrium with metallic phases, the lines b and the curve d be displaced. The sense of this displacement is at once clear. Through the presence of the easily oxidized carbide, the solid solution, compared with the pure iron, is also increased in oxidizability ; it requires for trans- formation into FeO even smaller C0 2 concentrations than the pure metal; that is the line b is displaced to the right. The result of this changed position is a sinking of the pressure value for the four-phase system. The pressure of the system (solid solu- tion FeO, Fe 3 C, CO, CO 2 ) is smaller than that for the system (ferrite, FeO, Fe 3 C, CO, C0 2 ). Since the cementation curve DECOMPOSITION OF CARBON MONOXIDE, ETC. 193 for the solid solution must go through this point, it can be seen directly from the figure, that the cementation of solid solutions can take place with lower monoxide concentrations than that of pure iron. The cementation curve is pushed toward the left (the curve for the solid solution is dotted). Since the curve c is very steep relatively small displacements of x, respectively to t\ represent large depressions of the inter- section -4. The ordinate equation also holds for this point. & retains its value since, in the equi- librium c no solid solution phase appears. How large this influence is, a rough calculation will show. We will assume that at a temperature of 700 solid solution for- mation already takes place and that through this condition the equilibrium value X for the oxidation of this solid solution is raised to 0.62 (compared to 0.60 for pure iron). Thereby t\ in- creased from 1.50-1.63. The value of # is at this temperature 384,000. From the above given equation we calculate for the equilibrium pressure of the solid solution with the simultaneous presence of cementite and FeO, the value 115.4 as compared to 166.3 atm. if we replace the solid solution with pure iron. The depression is accordingly 51 atm. for the small concentration increase of CO (around 2 per cent). That the constant n of the cementation equilibrium (FeaC, Fe, gas) changes with a dis- placement of 77 is shown by the equation: Accordingly this change is relatively large for small deviations of the variable 77, since M is approximately proportional to the third power of 77. It decreases with the above given change of 77 from 113,800 to 88,650. Beside the intersection A, the intersection of curves a and b is depressed by solid solution formation. Since the curve a is flatter than the curve d, the influence is not so great as with a, 194 THE PHYSICAL CHEMISTRY OF THE METALS but it is sufficiently large, that the pressure value, which is above atmospheric pressure for pure iron, is depressed below this value so that the equilibrium of the solid solution becomes conveniently G CO i.e.- x FIG. 104. measurable. And this is not without practical importance, since we may orient ourselves experimentally concerning the depend- ence of the constant rj on the carbide content of the solid solution and from the determined relations follow the influence of the solid DECOMPOSITION OF CARBON MONOXIDE, ETC. 195 solution formation on the cementation process mathematically. Nothing further being necessary, than, after the occurrence of equilibrium between solid solution, ferrous oxide, carbon and the gas atmosphere, to determine the composition and pressure of the latter and to analyze the solid solution. All these experimental data still fail to place us in a position to predict the direction of the displacement of the equilibrium and the reaction fields under the complex conditions for the formation of solid solutions such as can occur at high tempera- tures. After this digression we return to the simple relation such as we have seen holds under 700, we draw again a schematic diagram of the equilibrium curves and their intersections at con- stant temperature, considering, however, also the ferrous oxide (see Fig. 104). We have, accordingly besides the representations of the equilibria FeO+CO^Fe+C0 2 3 FeO+5CO <=> Fe 3 C+ 4 C0 2 also that for Fe 3 4 +CO <= 3FeO+CO 2 These five curves divide our diagram, the X, P plane into twelve fields, in which the following reactions go on: Field A 2. 3 3. FeO+CO = Fe+C0 2 4. Fe 3 4 +CO = 3FeO+C 5. 2CO =C+C0 2 196 THE PHYSICAL CHEMISTRY OF THE METALS Field B l Field B 2 2. 3. 4. Fe+CO = Fe+CO 2 = C+C0 2 Field FeO+CO = Fe+C0 2 3. = C+C0 2 5. C+C0 2 Field 2. 3. 4. 5. 2. Fe 3 C+4C0 2 =3FeO+5CO 3. Fe+CO 2 = FeO+CO 4. 5. Field 2. e 3. Fe+C0 2 = FeO+CO 4. 5. 2. 2. 3- 4- 5- 2. as n Field C 2 as in Ci Field D 2 i. 2. 3- 4- 5. i. 2. 3- 4- as n Field E 2 as n 5. C+C0 2 = Field F 3. Fe+C0 2 = FeO+CO 4. Fe 3 4 +CO-3FeO+C0 2 5. DECOMPOSITION OF CARBON MONOXIDE, ETC. 197 Field G 2. 3. Fe+C0 2 = FeO+CO 4. Fe 3 04+CO = 3FeO+C0 2 5. 2CO = C+C0 2 Field H 1. 3Fe+2CO=Fe 3 C+C0 2 2. Fe 3 C+5C0 2 = 3FeO+4C0 2 3. Fe+C0 2 = FeO+CO 4. 3FeO+C0 2 = Fe 3 4 +CO 5. 2 CO = C+C0 2 As end products of the reaction also occur: Cementite, together with carbon in fields A and F, metallic iron with carbon in fields BI and Ci, without simultaneous precipitation of carbon in I$2 and C 2 . Ferrous oxide with carbon in fields D\ and G. Ferrous oxide without carbon in field D 2 . Ferrosoferric oxide with carbon in fields E\ and H. Ferrosoferric oxide without carbon in field 2 . We can accordingly, at the same temperature obtain any of the four solid substances as the stable phase by simple variations of pressure and composition of the gas phase. By raising the temperature all fields suffer deformation, all curve constants increase, as a result, the curvature changes and displacements occur toward the right. If we think now of all isotherms as vertical to a temperature axis of a space model with the coordinate axes, T, P, and X, so we have a geometrical representation of the whole theory of the blast furnace. We will now study one important conclusion which our diagram brings out. It so happens that the cementite as a phase in opposition to the other solid products is only stable in such fields as the CO is labile and is subject to decomposition 198 THE PHYSICAL CHEMISTRY OF THE METALS into C02+C. From this it arises that the cementite is not formed from carbon and metal, that the only cementing agent by which one can reach a direct formation of carbide is CO. It can, however, be formed by the action of carbon on ferrous oxide (in B). It is, however, labile here. However, the solid carbon can also act as cementing agent if it is not in the formation of pure cementite as a separate phase, but in the formation of a solid solution of this substance with metallic iron. Suppose that the prevailing temperature for the formation of solid solution falls inside of field 2, the field in which at temperature less than 770, metallic iron is the only stable solid phase. At this temperature reaction i in the above given table is no longer unidirectional, but is reversible for a large interval of X. We must replace it with the equation in which the carbide belongs to the same phase as the metal. If we remember the Mass Law, we see that the composition of the solid solution is dependent on the pressure and com- position of the gas atmosphere. That is also the case for the reaction. FeO+CO <= Fe (Fe 3 C containing) +C0 2 . The fields F and G which on account of the high pressure values are not observable with iron deserve a few further words. The mutual relations of metal, carbide and oxide in the presence of gas can be made clear with other metals, especially with manganese. Practical experience has taught that by the reduction of manganese oxide with carbon and carbon monoxide a carbon- free metal never results but always a carbide. We might draw the conclusion from this that the oxide is more easily reduced to the carbide than to the metal, which would be the case at point A if the CO concentration necessary for cementation were smaller than that required for reduction. DECOMPOSITION OF CARBON MONOXIDE, ETC. 199 The equilibrium pressure occurring with manganese is very small. Carbon monoxide is, up to over 1000, easily and com- pletely absorbed by metallic manganese with strong heat evo- lution. At a temperature of 1200 the equilibrium pressure reaches a measurable value. This relation can be easily under- stood from our earlier theoretical considerations, since with manganese we have a metal which is much less noble than iron and as a consequence possesses a larger value for the reduc- tion constant rj. What holds for rj holds also for $, the constant which gives the conditions for the reduction of the oxide to the carbide. The result of increase of ?? and $ is the strong decrease 800- 700- 600- 500- 400- 300- 200- 100- m m. Temp. 600 700 800 900 1000 1100 FIG. 105. of the equilibrium pressure for the systems. Metal oxide, carbon (amorphous) gas, and carbide oxide, carbon (amor- phous) gas. This influence makes itself felt in the action of CO on solid solutions of iron and manganese as well as those of iron and manganese cementite. Small manganese contents cause, as the observations show, great decrease in the equilibrium pressure. In the following table the results of one experiment are given, which has been taken from the observations of Schenck and Semiller on the action of CO on manganese containing pig iron. 200 THE PHYSICAL CHEMISTRY OF THE METALS From the numerical values and still better from the graphical representation (Fig. 105) the influence of the manganese content on the equilibrium pressure for the presence of the solid phases, cementite, oxide, amorphous carbide can be seen. For com- parison the values for manganese free iron cementite are given. COMPOSITION OF MANGANESE CONTAINING PIG IRON Preparation. Mn, Per Cent. Fe, Per Cent. C, Per Cent. I o-9S 96.6 2.IS II 4.01 93-o 2.99 III 6.38 93-2 1.32 Cementite. Manganese Free PREPARATION I. PREPARATION II. PREPARATION III. Temp, in Degrees. Pressure, Mm. Temp, in Deg. Pressure. Mm. Temp., Deg. Pressure, Mm. Temp., Deg., Pressure, Mm. 634 83 629 28 6 7 2 131 679 41 6 9 I 195 * 722 2 9 8 734 341 730 86 774 562 751 IOI 779 657 820 180 849 217 853 17 880 349 900 396 911 28 943 IS 0^8 617 0^6 20 VO U 959 802 965 46 vo 981 20 IOIO 74 1031 34 1088 162 1086 62 1093 68 1103 170 IIIO 96 If we also distort the results by the formation of ferrite- cementite solid solution, we obtain a picture, at least qualitative, of the influence on the equilibrium of the presence of man- ganese. According to our previous consideration we would expect that the reduction of the metallic oxide by carbon and CO would be easier the more noble the metal and coordinately the DECOMPOSITION OF CARBON MONOXIDE, ETC. 201 greater the field B 2 of our diagram. We should encounter accordingly very little danger of falling into a field where the C02 concentration is so high that the metal present is oxidized. The experience of practice and the laboratory show us now that one is bitterly deceived in these expectations. Nickel oxide, as well as tin oxide, are vainly subjected to conditions, which with iron oxide would produce the metal without dif- ficulty. To obtain an energetic reaction much higher temper- atures must be used than are necessary according to the posi- tion of the equilibrium pressure, which with noble metals is already many atmospheres at relatively low temperatures. We have heretofore left entirely out of consideration a factor which is of essential meaning for an energetic reaction, namely : the reaction velocity, the velocity with which a chemical process approaches the equilibrium. Even with the oxides of noble metals, the reduction with carbon is only feasible if the reaction velocity is not too small. The resistance, which is opposed to an energetic completion of the reaction is in many cases due to the physical condition of the oxide. With one and the same substance, we can meet different degrees of reaction energy, if it has received different previous treatment. Strongly ignited sintered masses are much more inert than finely pulverized porous preparations. The surface is the essential factor on which the reaction velocity depends. A second factor is the temperature and indeed the velocity increases with the temperature under all conditions. Of this fact use has long been made in practice. The reduction of NiO, which at 600 goes only very slowly, is conveniently carried on 100 higher provided the oxide and the reducing agent are intimately mixed. There must be considered for the reduction of oxides, beside the equilibrium ratio also suf- ficiently great reaction velocity. In many cases it is easier to obtain a base metal from its oxide than a noble one, if the latter oxide shows inertness toward the reducing agent. 202 THE PHYSICAL CHEMISTRY OF THE METALS The Mond Nickel Process. We dare not close our considerations of the action of CO on the oxides and metals without mentioning a remarkable reaction, which is of interest to the metallurgist and has been made by Mond to play an important role in the production of nickel. Iron and nickel, have, at temperatures slightly removed from room temperature, the peculiarity of adding CO and form- ing with it the remarkable compounds Fe (CO)?, Fe(CO)5, Ni(CO) 4 . The most important of these is nickel carbonyl Ni(CO)4. Concerning its formation and existence conditions, Mittach * has made thorough investigations. Nickel-carbonyl is a water clear, very volatile liquid which boils at 40 and at higher temperatures decomposes easily into the components from which it was formed. The ease of formation on the one hand and the decomposition on the other can be demonstrated by a simple experiment. If we fill a glass tube, drawn to a point on both ends, with the vapor of nickel-carbonyl and place it, after the ends have been sealed, in a boiling water-bath, after a time we observe the precipitation of a nickel mirror. If then the tube be taken out and allowed to lie a few days at room temperature, the mirror disappears again, with re-for- mation of nickel-carbonyl vapor. We are accordingly dealing with a reversible reaction, which can be expressed by the symbol Ni-h4CO<=Ni(CO) 4 . As previously, we can deduce the equilibrium conditions from the mass law. They are C co ' K. We have, so long as the nickel-carbonyl is gaseous, a two- component system with two phases; it is accordingly bivariant * Z. Physik. Chem., 40, i (1902). DECOMPOSITION OF CARBON MONOXIDE, ETC. 203 and dependent on the pressure and temperature. The equation of the isotherm is as one can easily derive. i x which gives us the relation between the pressure and the com- position of the gas phase. The higher the pressure the greater the amount of the nickel-carbonyl vapor in it. Temperature in Degrees. logK. Temperature in Degrees. logK. "3 2.719 62.0 5-575 16.1 3-04 70.4 5-948 29.8 3.812 8o.O 6-443 35-9 4.199 90.0 6.859 50.2 4.925 99-3 7.178 It is sufficient to have the values for K or its logarithm for a series of temperatures between 10 and 100. From 10 20 30 40 50 60 70' FIG. 106. 90 100 these we can calculate the composition of the gas for every other temperature. For an illustration of the displacement of the equilibrium with the temperature, see Fig. 106. It shows that as the amount of CO in the gas mixture increases with rising temperature, that of nickel-carbonyl decreases. 204 THE PHYSICAL CHEMISTRY OF THE METALS The constants vary somewhat with the condition of the metallic nickel present, and the value for the finely divided metal is higher than for the compact, an analogy to the fact that small crystals show a greater solubility than large. The velocity of the formation is influenced by a large num- ber of factors. That increasing pressure, that is, raising the CO concentration, is favorable to the reaction is a result of the mass law and that fine division of the nickel also aids the reaction can be understood from our earlier experience. It is remarkable, however, that traces of foreign substances, for example of oxygen, cause very great variations from the nornal run. They work directly as " poison " for the nickel and destroy its combining ability. It would be going too far here to follow this influence further. Our problem, the action of carbon and carbon-monoxide on the ores and metals, and the study of the occurring equilib- brium phenomena is now disposed of, and we now understand the factors on which the oxidation reduction and cementation depend. We will now turn to another field, application of theory, to the chemical processes which are important in the smelting of sulfide ores. CHAPTER VI THE REACTIONS OF SULFIDES WHILE for the preparation of iron only oxide ores come into consideration, the principal amount of the other prac- tically important metals are obtained from the sulfides. This holds especially for copper, lead, zinc, and mercury. A thorough knowledge of the relations of the metallic sul- fides to the metals on the one hand, and to the oxides on the other, is therefore essential if we would specify the conditions under which the reactions, which are used in practice for the transformation of sulfides to the metal, run in the desired direction. It is a well-known fact that at high temperatures the sulfides undergo a dissociation into their elements or into sulfur and sulfur-poorer sulfides. From some, for example, the iron sulfide FeSs, the sulfur can be directly distilled. What now are the kinds of dissociation phenomena concerned, and are the relations exactly as with the dissociation of the oxides. The reaction MeS < Me+S, is characterized by a definite dissociation tension, dependent on the temperature which with the above-mentioned iron sulfide amounts to over one atmosphere at the distillation temperature. Naturally, the occurrence of the definite dissociation pres- sure is connected with the condition, that two solid phases must be simultaneously present. Due to the frequently observed miscibility between metals and sulfides, the mixtures consist in many cases of a single phase, a solid or liquid solution and 205 206 THE PHYSICAL CHEMISTRY OF THE METALS it depends entirely on the concentration of these solutions whether at the same temperature, the sulfur tension is great or small. The experimental material concerning the sulfur tension of sulfides is extraordinarily sparse. There is still difference of opinion as to the series which gives the sulfides, according to the magnitude of their tension. The reason for this is the experimental difficulty encountered in handling the problem, and the great influence of mutual solubility which the various authors have not taken sufficiently into consideration. That the difference of the dissociation magnitude for dif- ferent sulfides in some cases must be very large, may be con- cluded from the possibility of the so-called precipitation process. This metallurgical process consists in heating the sulfide of a valuable metal with iron, whereby the metal is set free and the sulfur combined with the iron. Many uses are made of this reaction; for example, the mercury out of cinnabar and the antimony out of stibnite are precipitated. In some cases lead has even been obtained from galena by this method. Iron sulfide may also give up sulfur even in the liquid state by the addition of manganese or ferromanganese, to the melt. This desulfurization process which is used in practice would not be possible if the affinity of manganese for sulfur were not considerably greater than that of iron. This " precipitation process " has its exact analogy in the reduction of oxides of noble metals by less noble, which has been treated in the fourth chapter. Also in this case, reaction takes place if the dissociation tension of the desulfurized sulfide, is greater than that corresponding to the sulfide of the metal used. However, the relations are not so simple as with the oxides. A series which gives the ease of desulfurization of the single sulfides is extraordinarily difficult to arrange. That this dif- ficulty is principally due to the solubility of the metals in the sulfides we have already mentioned. It is further increased since the mutual solubility of the 'sulfides themselves is very THE REACTIONS OF SULFIDES 207 considerable even in the solid state. I need recall only that in the complex diagram of the nickel matte, the solid solutions between nickel sulfide and iron sulfide play an important role. Further it is no rarity for the sulfides to go into chemical com- pounds with each other as we also recall from the nickel matte. The degree of the desulfurization reached will under conditions be dependent on the relative amounts of the sulfide and the desulfurizing metals that enter into the reaction. If we would become clear concerning the yield of the metal to be expected, we must investigate first the equilibrium diagram of the concerned ternary systems. Only in rare cases does the reaction take place without matte formation, the desul- furization being generally only partial. We have now, in desulfurization by iron, at least one method by which the dissociation tension of the original sulfide can be considerably depressed and thereby the energy of desulfuriza- tion essentially raised. This means is an addition of Na 2 S or CaS to the liquid melt. These substances form with iron sulfide complex sulfides, so-called sulfurets, whose sulfur tension is considerably smaller than even that of the iron sulfide. This has the added advantage that the sulfide dissolved in the metal bath is increased, not to mention the change of physical prop- erties which is very favorable. The sulfuret melt is considerably lighter and more liquid .than that of the matte, it separates much better from the regulus and can be separated very readily from it after cooling. The limitation of the usefulness of this modification of the precipitation process is naturally that the sulfide to be desulfurized may form stable complex compounds with alkali sulfides. The latter appears to be the case with lead sulfide, where the experience of practice shows that ad- dition of alkali sulfide influences the precipitation process adversely. We recognize from this that it is difficult and not suf- ficiently reliable to arrive at an orientation of the relative affinities of the metals for sulfur, by means of decomposition of the sulfides with metals. We will therefore experiment in another way to get consistency, at least in some cases. 208 THE PHYSICAL CHEMISTRY OF THE METALS We saw with the oxides that a conclusion concerning the affinity of metals for oxygen could be drawn, if the equilibrium between metal oxide, water vapor and hydrogen is known. In an entirely similar way a conclusion concerning the affinity of the metals for sulfur may be drawn from measurements of the equilibrium. MeS+H 2 <=Me+H 2 S. Ag 2 S+H 8 ^2 Ag-l-H 2 S 800 700 600 506 400 300 10 20 30 40 50 60 FIG. 107. 70 80 90 H 2 S Such measurements have been made for several sulfides by Pelabon.* His results are given in the following table and the graphic representation in Fig. 107: Ag 2 S +H 2 <= Ag 2 +H 2 S. HgS+H 2 4=Hg+H 2 S. Sb 2 S 3 +3H 2 ^ Sb 2 +3H 2 S. Temp, in Degrees. H 2 S. Per Cent. Temp, in Degrees. H 2 S, Per Cent. Temp, in Degrees. H 2 S, Per Cent. 360 440 520 6lS 710 21. 02 19.85 18.60 17.00 16.03 360 440 520 78.67 85-26 Q2.IO 440 510 555 610 625 44.3 ! two solid 48 . 6 / phases * " two liquod 56-01 > . phases 56.9 J We have also here to deal with the measurements of in- complete equilibria and reversible reactions. We find a definite * Ann. Chim. Phys. (7), 25, 365 (1902). THE REACTIONS OF SULFIDES 209 equilibrium relation of the two gases depending only on the temperature, so long as sulfide and metal are present as two solid or liquid phases. The latter possibility is met, as we have seen in chapter 3, rather frequently in systems of sulfides and metals. Such a case of limited miscibility in the liquid state is the antimony-antimony sulfide equilibrium in the above table. The percentage of hydrogen sulfide in the equilibrium mix- ture of gases, as we see from our examples, decreases with tem- perature (e.g., with silver) or as with mercury and antimony increases. The direction of the temperature change of the equilibrium will according to the rules of chemical thermo- dynamics, be conditioned on the thermo-chemistry of the de- composition. That side of the reaction will be favored by rising temperature which is formed with heat absorption. The stability of silver sulfide accordingly increases with rising temperature, the affinity for sulfur increases. With mercury and antimony the reverse is true, an increase in temperature favoring a precipitation of the elementary metals. To be sure the equilibrium constants may only be used as a correct measure of the affinity of sulfur for metal, if the metal and sulfide exist as phases in the pure state. If one or the other, or both of the " Bodenkorper " consist of solutions, the hydrogen sulfide content of the gas phase shows us only the tendency of the sulfide-richer phase to go over to a sulfide- poorer. If the sulfide and metal dissolve completely in one another the hydrogen sulfide content of the gas atmosphere depends entirely on the mixture relations of the two substances in the solution. If we investigate solutions of increasing sulfide content at constant temperature, the hydrogen sulfide con- centration in the gas increases in the same direction; it is first constant when the solution becomes saturated with sulfide. These relations may be readily followed with the system bis- muth bismuth sulfide, as has been shown by Pelabon. It lends itself particularly to graphic representation (see Fig. 108). At 600 we have complete miscibility, at 440 the curve goes upward with increasing sulfide content and at a determined 210 THE PHYSICAL CHEMISTRY OF THE METALS place, namely, the limit of solubility, it becomes a straight line. From this point on the hydrogen sulfide content is constant. From all these facts we may conclude, that the possibility of determining the affinity of metals for sulfur from the equilib- rium atio of a hydrogen-hydrogen sulfide atmosphere, which is in contact with a metal and its sulfide, is very limited. The precipitation process for obtaining metals from their sulfides is only used in practice in special cases. The process of reduction after previous roasting is much more general. 440 610 90 80 70 GO 50 40 30 -20 -10 10 20 30 40 50 60 70 80 90 H 2 S FlG. 108. By the roasting of sulfides is understood, their oxidation by atmospheric oxygen, a reaction which can be generally represented by the equation (Me indicating any bivalent metal) . The reverse of this reaction has never been observed. There occurs, however, sometimes in place of the oxide, other oxida- tion products, under some conditions sulfate is formed. The sulfatizing roast plays a role, for example, in the Ziervogel THE REACTIONS OF SULFIDES 211 process which consists in roasting the argentiferous copper matte so that the silver, but not the accompanying iron and copper, is changed to sulfate, so that by leaching the silver goes into solution, which is then precipitated by copper, washed, pressed and melted in a crucible. The copper and iron remain as oxides in the residue. There is also no difficulty in changing copper sulfide to sulfate and one observes in the Ziervogel process always the formation of copper sulfate as well as iron and silver sulfate. By heating to higher temperatures it de- composes as does the iron sulfate, while the silver sulfate remains unchanged. We now come to the question : " Under what conditions does the sulfide go on roasting to oxide, and under what con- ditions to sulfate?" Frequently the result of the roasting depends on the temperature and the composition of the gas phase which is over the material to be roasted. From analytical practice we know that by strong heating the sulfates of the heavy metals are decomposed and that thereby sulfur trioxide as well as sulfur dioxide and oxygen are formed. This reaction has also technical importance, or rather it has had, since on it depends the old method for the prepara- tion of fuming sulfuric acid and sulfur trioxide, in which iron sulfate was heated in retorts and the evolved vapor condensed in prolongs which contained a small amount of water. The decomposition of the sulfate is a dissociation process which can be repiesented by the equation MeS0 4 <=MeO+S0 3 . We may conclude on the basis of the phase rule, that a definite sulfur trioxide pressure exists for every temperature. If a sulfate is heated in a previously evacuated vessel, connected with a manometer, at a constant temperature, a definite pressure results. Since a number of metal oxides favor the decom- position of sulfur trioxide into sulfur dioxide and oxygen, these gaseous substances are found in the dissociation products of the sulfate and raise the dissociation pressure. This does not interfere, however, if we heat them with platinum gauze so that 212 THE PHYSICAL CHEMISTRY OF THE METALS the equilibrium between the two oxides of sulfur and oxygen can take place. The resulting gas pressure allows us to cal- culate the partial pressure of SOs and from that the SOa tension of the sulfate. The observed pressure is equal to the sum of the partial pressures of the separate gases present. and since the number of molecules of oxygen that result from the decomposition of the trioxide is half as great as that of the trioxide, so it follows Recently Keppeler,* and L. Wohler, Plliddemann and P. Wohler f have made observations concerning the tension of sulfates and measured the values of P for a series of metallic sulfates. The equilibrium between SOs, 862, and 0, has been rather completely investigated by Knietsch,J Bodlander, and further by Bodenstein and Pohl.|| The equilibrium conditions for the reversible reaction are 2S0 3 =* 2S0 2 +0 2 , K C 2 S0 3 The magnitude for the constant K as measured for different temperatures by Bodenstein and Pohl are given in the following table: Temperature in Degrees. K. Temperature in Degrees. K. 528 1.55.10-5 727 3.45.10-3 579 7-55-10-5 789 1.26.10-2 627 3.16.10-4 832 2.8o. IO-2 680 1.12.10-3 8 97 8. 16. 10-2 * Z. angew. Chem., 21, 532 (1908). t Ber., 41, 703 (1908). t Ber., 34 4059 (1901). Z. Elektrochem., 9, 787 (1903). || Z. Elektrochem., 11, 373 (1905). THE REACTIONS OF SULFIDES 213 TENSION P ACCORDING TO L. WOHLER, PLUDDEMANN AND P. WOHLER Temp, in Degrees. Pressure in Mm. Temp, in Degrees. Pressure in Mm. Fe 2 (S04) => FejOa+3S03(S02,0 2 ). Al2(S0 4 )s. = A1 2 3 +3803(802,02). 553 23 572 28 570 33 621 51 592 35 681 120 614 70 702 180 634 H3 720 261 650 149 731 356 660 182 742 480 680 286 748 692 690 401 699 560 , 707 715 2CuSO 4 . <= 2CuO-SO3+SO3(SO2O 2 ). 2CuO.SOs. <= 2CuO+SO3(SO 2 ,Oj). 546 43 600 62 588 55 653 98 615 700 686 123 642 98 705 139 665 , 130 728 173 700 233 745 209 714 324 775 298 725 460 805 542 Temp, in Degrees. Pressure in Mm. ZnSOi. <= ZnO -r-SCWSCfc.Oz). 675 690 5 6 720 750 24 61 775 800 112 l8 9 214 THE PHYSICAL CHEMISTRY OF THE METALS From the plotting of log K the values for the intermediate temperatures can be obtained. (Fig. 109.) If we rearrange the equilibrium conditions somewhat we obtain C 2 S0 3 K i.e., Cso T' If we designate the S0 3 concentration corresponding to the dissociation tension by Csoa, the equilibrium between the solid -l -2 -4 500 C 60.0 700 800 c FIG. 109. 900 products, metal sulfate, and oxide and the gases SOs, 862, and O2 requires, C so ,= I c , VT- If the expression on the right is greater the oxide changes to sulfate, if it is smaller the sulfate decomposes. The same conditions also hold for the roasting process. By roasting an atmosphere is obtained, which contains SC>2 and the unused oxygen. If we designate the concentrations with Cs02 and Co2, the conditions for the formation of oxide are SO,- C/ ^02 so 2 "V/^T- THE REACTIONS OF SULFIDES The conditions for the sulfatizing roast are: C 215 It is frequently more convenient to use the partial pressure of the gases, in place of the concentrations, these are obtained, expressed in atmospheres (which, however, are easily converted into millimeters, by multiplying by 760) by the general gas equation: = c.R.T. The roast reactions are of the form (oxidizing) 2Pb+2SOj. II. PbS+sPbO. 2. Out of this sulfate containing melt we may precipitate by cooling different kinds of products, there may, as a single possibility, be solid solution or compounds between oxide and sulfate present in the solid phase and the variation of the gas evolu- tion may be due to these. These we can only decide by investigation of the crystal- lization diagram of the system lead oxide-lead sulfate. The 700 500 400 300 200 100 S0 3 Pressure m m. 600 650 700 750 800 850 900 FIG. no. equilibrium diagram of this pair of components is shown in Fig. in. From this the remarkable fact is apparent that there are a number of compounds between lead sulfate and lead oxide. Two of them are recognized by maxima in the crystallization curve. They are the two basic sulfates PbO PbS0 4 and 2PbO PbS0 4 . Beside these there is still a third compound, 3PbO-PbSO 4 , whose maximum is not, however, to be observed, since it decomposes before it is reached. The position of the melting point for the components and their compounds are as follows: PbSO 4 1 100 PbS0 4 PbO 966 PbO 882 PbS0 4 -2PbO 95 220 THE PHYSICAL CHEMISTRY OF THE METALS The compound PbSO^PbO decomposes above 880 into the compound PbSO^PbO and melt. Eutectic points are present in three cases: PbO PbS0 4 -3PbO at 820 and 87 per cent PbO. PbS0 4 -2PbO PbSO 4 -PbO at 940 and 553 per cent PbO. PbSCVPbO PbSO 4 at 950 and 30 per cent PbO. Besides these there occur two horizontals in the diagram which correspond to transitions in the solid state. The one THE REACTIONS OF SULFIDES 221 at 845 corresponds to a transition of the pure lead sulfate while the second at 450 corresponds to a rearrangement of the compound PbSO 4 2PbO. This rearrangement is reversible. It shows a halt-point with both rising and falling temperature. Whether this is a change of modification or the chemical reaction 2 PbS0 4 2PbO <=> PbS0 4 PbO + PbSO 4 cannot be decided easily and for our special purpose it makes no difference. The temperature field of 550-900 is the only one of impor- tance to us, inside of this we have to do with three solid phases due to the fact that three basic sulfates of lead exist. We may conclude from this test that the raised sulfur dioxide tension which occurs with mixtures of oxide, sulfide and metal, previ- ously heated to melting is due to the presence of these basic sulfates and indeed it turns out that the deviation is conditioned by the formation of PbS0 4 -PbO and that the curve represents the equilibrium. 2 (PbSO 4 -PbO)+3PbS <= 7Pb+5SO 2 . The formation of this basic sulfate can be easily understood. The action of 862 on molten lead oxide is under all conditions to form sulfate. If we allow such a melt to crystallize we obtain a eutectic which consists of basic sulfates. For the greatest sulfur dioxide tension, that substance is necessary which with sulfide will develop the highest pressure and that is naturally the sulfate-richest substance. The pressure value and curves for the other basic salts must be obtained by special experiments. A glance at the diagram shows us that to procure the melts which would allow us to measure the 862 tension of the equilibrium system, basic salt, sulfide, metal, gas, would require special .preparations. The measurements must always be carried out at temperatures not exceeding 830. 222 THE PHYSICAL CHEMISTRY OF THE METALS The Structure Constituents Present in the Solid Mass. 1. PbS0 4 -3PbOwithPbO 2. PbS0 4 - 2PbO with PbS0 4 3PbO 3. PbS0 4 PbO with PbSO 4 2?bO 4. PbS0 4 with PbSO 4 PbO The Composition of the Melts, PbSO 4 , Per cent. 25 2 5-39 39-60 60 For every basic salt we would expect a curve which repre- sents the equilibrium between each of these salts, sulfide metal and gas. Masses, formed by melting oxide and sulfate, in the ratio, which No. 3 of the given table represents, are rich in the basic sulfate PbS0 4 -PbO and quickly reach equilibrium with sul- fide, metal and 862 and a definite 862 tension is obtained which increases with rising temperature. With the oxide rich melts in which the other basic sulphates are contained equilibrium is reached only with difficulty and the position of their cor- responding equilibrium curves is uncertain, we are therefore limited in our considerations to the normal basic sulfate. The dependence of the S02 tension on the temperature is given in the following table: Temperature in Degrees. Pressure in Mm. Temperature in Degrees. Pressure in Mm. 681 16 7 8o 217 7 l6 42 795 306 741 81 810 440 759 130 821 548 770 184 830 710 If we compare these values with those of the system ILz Fig. no, we see that these fall completely together with the curve for the basic sulphate. By the action of S02 on lead oxide during the cooling basic sulfate has formed. We obtain accordingly three equilibrium curves (if those for the di- and tri-basic sulfates were known there would be, in all, five) their comparative position is shown in Fig. 112. THE REACTIONS OF SULFIDES 223 These curves divide the coordinate plane into four different fields, inside of these, different possible reactions take place and different end products remain. The following table gives 800 700 600 500 400 300 200 100 Temp. 600 650 700 750 800 850 900 950 FlG. 112. us a general view of the reaction system, the stable solid pro- ducts are in heavy type: Field I. (a) 2 Pb+2S0 2 (>) M =PbSO 4 +PbS, = 2(PbS0 4 -PbO)+3PbS, = 2PbO+PbS. Field II. (a) PbS0 4 +PbS (c) 3 Pb+S0 2 Field III. (a) PbSO^PbS (b) (2PbSO (c) 3 Pb+S0 2 = 2 (PbSO 4 PbO) +3?bS, = 2PbO+PbS. = 2PbO+PbS, = 7Pb+5S0 2 , = 2PbO+PbS. 224 THE PHYSICAL CHEMISTRY OF THE METALS Field IV. (a) PbS0 4 +PbS = 2Pb+2SO 2 , (b) 2(PbSO 4 -PbO)+3PbS = 7Pb+5SO 2 , (c) 2PbO+PbS = 3 Pb+S0 2 . In Field I sulfate and sulfide do not react with each other, in Field II the basic sulfate does not act on the lead sulfide and in Field III oxide and sulfide are compatible. It can now be easily shown that the action of sulfur dioxide on lead oxide always gives these stable products. Out of Eqs. a and b of Field I it follows by elimination of the metal 4S0 2 +4?bO = 3PbSO 4 +PbS. The same end products are left by the basic sulfate with S0 2 . In Field II one obtains the basic sulfate by the action of sulfur dioxide on sulfate, as well as on oxide, and in Field III, the reaction takes place to which one ascribes the principal role in the Huntington and Heberlein process: The stable products always result. It is easiest to reach the most stable state by the action of sulfur dioxide on the metal. The progress of the reaction with time can be followed very conveniently by means of a manometer and further there is the possibility by variation of the gas pressure and temperature to start from every point in the field. The experiments on the reaction of metallic lead with S0 2 in Field I are easily understood. Their results are shown in Fig. 113. Metallic lead absorbs S0 2 ; if we observe a closed volume of this gas in contact with lead at constant temperature, we notice a decrease of the gas pressure to a constant value which represents the equilibrium of sulfate with sulfide and metal. The values 653 : 101 mm., 690 : 240 mm., fall exactly with the sulfate curve, observed from the other direction. If the experiments be continued still further so would a further decrease in pressure occur after a time, especially if the sulfate undergoes rearrangement to basic sulfate by tran- THE REACTIONS OF SULFIDES 225 sitions in the solid phase. When the next curve is reached, the pressure would remain again constant for a time. Also this arrest is finally passed and if the amount of lead reacting is very large the pressure sinks to very small values. Such a complete experiment would require considerable time; it would, 226 THE PHYSICAL CHEMISTRY OF THE METALS however, be excellent to determine the position of the resulting equilibrium at the considered temperature. Reactions which are connected with the evolution and absorption of gases are best investigated at different temperatures by the surveying of the reaction isotherms and from the position of the pressure halt-points, the equilibrium diagram is constructed in an en- tirely similar way as we put one together for solidificatio*- and transition processes from the temperature halt-points. Since we have not a complete absorption diagram at our disposal we will have to be content with single sections of such. A closed volume of 862 shows in the presence of lead, which has previously been treated with lead sulfide at 800, the follow- ing change of gas pressure with time: Time in Minutes. Pressure in Mm. Time in Minutes. Pressure in Mm. Time in Minutes. Pressure in Mm. O 745 96 135 132 125 22 559 IO2 134 192 no 24 403 108 132 228 IOO 42 251 114 131 72 156 1 20 I2Q 8 4 142 126 130 465 40 The halt-point lies inside of the Field III, somewhat above the value which the pure oxide shows at 800, it belongs accordingly to the oxide rich basic salt, probably, however, the lead is still not sufficiently saturated with sulfide. In all the above treated equilibria the metal phase consists not of pure but of sulfide saturated lead. If the sulfide amount present does not reach saturation equilibrium may be obtained; this is, however, not a univariant but a divariant system, and the SO2 pressure depends on the sulfide concentration of the lead bath. In Field IV equilibrium curves do not exist and it is more remarkable that we find a halt-point there, as the experiment represented in Fig. 114, shows. At 880 is shown a retardation in the decrease at a pressure of 90 mm., at 940 in the interval 470 mm.-430 mm. It is certainly not to be treated as a univari- THE REACTIONS OF SULFIDES 227 ant equilibrium. There is formed on the upper surface of the metal, first a relatively concentrated solution of sulfide in metal which, due to its lower specific gravity, floats on the surface. This, in contact with lead oxide, gives a sulfur dioxide tension which remains till the concentration of the solution is decreased by diffusion into the lower layers and since the diffusion equaliza- tion requires some time, the retardation is easily understood. Also in Field 4, the reaction = 2PbO+PbS 60 120 180 240 300 Minutes FIG. 114. can take place, this is true, however, only so long as the solution is not saturated. At the high temperature, it is further to be considered that one no longer has a solid oxide or sulfate phase, but a liquid solution of sulfate in oxide whose concentration depends on the external conditions. In any case we see that to get an understanding of so great a multiplicity of phenomena, as the reactions between the components of the ternary system, lead-sulfur-oxygen show, the consideration of the solubility of the sulfide in the metal and the miscibility of oxide and sulfate in the molten state is absolutely necessary. NAME INDEX Allen, H. S., 35 Arrhenius, S., 42 Bancroft, W. D., 134 Barus, C., 5 Bauer, 116, 118, 119, 120, 126 Baur, 162 Benedicks, 84, 85, 96 Berzelius, i Bijl, 60 Bodenstein, 212 Bodlander, 212 Bornemann, 122, 125 Bornstein, 9, 10, 41 Boudouard, 168 Bredig, 40 Broniewski, 16 Burgess, n, 15 Byers, 37 Bunsen, R., 181 Carpenter, 84 Chamberlain, 9 Charpy, 16, 65, 69 Chretien, 116 Cohen, 13, 17, 18 Crow, 15 Darrin, 37 Davy, i Day, no, 113 Dean, R. S., 35 Desch, ii Dewar, 13 Diesselhorst, 24, 78, 80 Drude, 23, 24, 25, 26, 28, 30, 32, 33, 34, 78, 80, 81 Ewen, 9 Fabre du Faure, 181 Falcke, 162, 177 185 Faraday, 22 Findlay, 134 Foerster, 84 Franklin, 37 Franz, 23, 26, 77 Friedrich, 114, 115, 118, 123, 124, 128 Getman, 17 Gibbs, 134, 135 Glaesner, 162 Goerens, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 99, ioo, 102, 103, 104, 105 Goldschmidt, 149 Goubau, 8 Graham, 80 Grenet, 16 Grunchant, 116 Griineisen, 78 Guertler, 9, 17, 74, 75, 107, 133 Guldberg, 152 Hagen, 29 Hanke, 37 Heberlein, 216, 224 Heller, 169, 177 Henning, 157 Heraeus, 57, 180 Heusler, 62 229 230 NAME INDEX Heycock, 45, 53 Heyn, 54, 84, no, 112, 113, 116, 118, 119, 120, 126 van't Hoff, 45, 156 Holborn, no, 113, 157 Humphrey, 16, 84 Huntington, 216, 224 Houghton, 9 Jager, 24, 78, 80 Jenkins, 216 Johnston, 4, 5 v. Juptner, 143 Kahlbaum, 4 Keppler, 212 Kinder, 15, 20 Knietsch, 212 Koessler, 37 Krafft, 6, 28 Kurbatoff, 95 Landolt, 9, 10, 41 Langmuir, I., 8, 35 Lea, Carey, 39 LeChatelier, 16, 54, 74, 84, 140, 145, 147, 151, 152, 154, 158, 164 Lehman, 18 Leroux, 114, 115, 118 Levin, 60 Lewis, 141, 143, 168 Lichtenberg, 68 Liebenow, 77 Lorenz, 40 Lowenstein, 156 Maey, 70 Martens, 54 Matthiesen, 74 Meyer, G., 41 Meyer, V., 5 Meyerhoffer, 9, 10, 41 Mittasch, 202 Moissan, 4 Murray, 9 Mond, 202 Nernst, 21, 143, 155, 158, 169 Neville, 45, 53 Newton, 68 Nichols, 37 Osmond, 84, 109 Parkes, 47 Pattison, 52 Pelabon, 116, 141, 208, 209 Peltier, 77 Petersen, 106 Pluddemann, 212 Poggendorff, 181 Pohl, 212 Preuner, 159 Puluj, 80 Puschin, 72 Ramsay, 43, 44 Randall, 168 Rathke, 84 Rayleigh, 77 Reinganum, 22 Rhead, 164 Richards, T. W., 135 Roberts- Austen, 41, 84 Romanoff, 47 Rontgen, 124 Roozeboom, 51, 84 Roschdestwensky, 31 Rosenhain, 16, 84 Roth, 4 Rubens, 29 Rudolphi, 131 Schaum, 13 Scheele, 140 Schenck, 169, 170, 177 Schule, F. A., 77 Schuster, 30 Semiller, 177 Siedler, 4 Smith, 216 Smits, 169 Sorby, 54 Spring, 47 Ssaposhnikow, 68, 69 Stahl, 143, 149 Stansfield, 84 NAME INDEX 231 Stas, 3 Stead, 127, 128 Tammann, 45, 54, 57, 63, 107, 121, 131, i33 Thomson, J. J., 23 Thomson, W., 20 Treitschke, 121 Turner, 9 Waage, 152 Warburg, 6 v. Wartenberg, 155, 160, 6 Wheeler, 164 Wiedemann, G., 23, 26, 77 Winkler, i Wohler, i Wohler, L., 212, 148, 141 Wohler, P., 21 Wood, 68 Wiist, 84, 85, 103, 106 Zeemann, 31 Ziervogel, 211 Zimmermann, 170 SUBJECT INDEX Absorption of dyes, 28 metals, 28 Alloys between sul fides, 123 binary, 49 coincident melting, 60 density of, 70 electrical resistance of, 73 freezing point of, 44 hardness maxima in, 69 heat conductivity of, 77 low melting, 68 magnetic, 62 . maximum melting, 60 microscopic examination of, 55 minimum melting, 60 of metals with metallic compounds, 82 sul fides, 114 potential of, 72 relation of physical properties to structure of, 68 segregation in, 52 structure of, 53 temperature coefficient of electrical conductivity of, 75 ternary, 64 Aluminum sulfate, sulfur trioxide tension of, 213 Analysis of blast furnace gases, 182 graphite and amorphous carbon, 178 Aniline dyes, absorption of, 28 luster of, 27 Antimony antimony sulfide, 115 explosive, 17 manganese, 62 sulfide-hydrogen, 208 Arsenic alloys, 125 copper, 129 iron, 129 lead, 129 nickel, 131 233 234 SUBJECT INDEX Arsenic silver, 129 zinc, 129 Austenite, 95, 96, 97, 101 Barium peroxide barium oxide, 146 Blast furnace processes, 180 gas, 182 Bearing metals, 69 Binary alloys, 49 Bismuth antimony, 58 bismuth sulfide, 209 lead-tin, 65 zinc, 47 Boiling points of metals, 4 Cadmium mercury, 61 zinc, 49, 55, 56 Calcium plumbate-plumbite, 147 Carbon iron, 83 precipitation of, .in iron-carbon alloys, 98 Carbon monoxide, catalysis of 37, 164, 166, 270 iron, 173 iron oxide, 162 reduction by, 159 zinc oxide, 160 Catalysis and passivity, 37 of carbon monoxide, 37, 164, 166, 170 Cementite, 85, 87, 88, 89 iron, 85 Charge on one electron, 25 Chromium tungsten steel, 108 influence of, on iron carbon alloys, 104 steel, 1 08, 109 Coefficient of absorption, 28 temperature, of conductivity, 31 Coincident melting alloys, 60 Colloidal solutions, aqueous, 39 solid or molten, 40 Colored metals, 34 Color of metal vapors, 6 Cooling curves, 10 Component, 135 Compounds, intermetallic, 60, 61, 63 with metallic properties, 82 Conductivity, 21 relation between heat and electrical, 23 Constantan, 76, 77, 81 SUBJECT INDEX 235 Copper antimony, 74 arsenic, 129 copper oxide, no copper-sulfide, 116 matte, 123 silicon, 132 silver, 51 Critical solution temperature, 47 Crystal growth, 20 Crystallization diagrams, of solid solutions, 57, 58, 59 compounds, 61 Decomposition of oxides by heat, 148 Degrees of freedom, 135 Density change in transition of tin 14 change on melting, 9 of alloys, 70 of explosive antimony, 17 Desulfurization with iron, 207 Diffusion of metals, 41 Distillation of metals, 3- Elastic limit, 68 Electrical conductivity of solutions, 21 gases, 22 metals, 22 properties of metals, 21 Electromagnetic theory of light, 29 Electrons, concentration of, in metals, 30 normality of, in metals, 31 Electron theory, 22 Enantiotropy, 18 Equivalent weight of electrons, 24 Faraday's law, 22 Ferrite, 91, 92, 93 Form changes, 32 Frauenhoffer lines, 7 Freezing point depression, 45 of alloys, 44 Fusion, heat of, 9 Gases, blast furnace, 182 Gold nickel, 60 platinum, 57, 59 silver, 58 sodium, 54 Goldschmidt process, 149 236 SUBJECT INDEX Graphical representation of three component systems, 66 Graphite and amorphous carbon, analysis of, 178 iron, 101 Hardness of alloys, 69 Heat conductivity, 23 conductivity of alloys, 77 of transition, 18 specific, ratio of at constant pressure and at const, vol., 6 of fusion, 9, 10 . decomposition of oxides by, 148 Heraeus process, 57 Heusler's alloys, 62 van't Hoff's equation, 156 Hysteresis, 108 Index of refraction of metals, 28 Iron arsenic, 129 carbide, 151 carbon, 151 carbon monoxide, 173 iron sulfide, 1 20 magnetism of, 16 modifications of, 15, 83 oxide-carbon, 177 oxide-hydrogen, 159 passivity of, 34 phosphorus, 127 silicon, 142 sulfate, sulfur trioxide tension of, 213 transition points of, 15 volume change of, transition, 15 Kinetic theory, 24 Kish, 98 Kurbatoff reagent, 95 Law of mass action, 152 Wiedemann and Franz, 28 Faraday's, 22 LeChatelier's principle, 151 Light absorption, 27 reflection, 27 Lead arsenic, 129 bismuth tin, 65 bismuth zinc, 47 lead sulfide, 114 SUBJECT INDEX 237 Lead sulfate basic, 222 sulfur oxygen, 218 Luster metallic, 21, 28 Magnetism and passivity, 35, 36, 37 of iron, 16 Magnetic alloys, 62 Manganese antimony, 62 effect on cementation of iron, 198 effect on steel, 106 Manganin, 76, 77, 81 Martensite, 90, 91, 96 Mass Law, 152 Mattes, 123 Melting points definition of, 7 of metals, 8 Mercury oxygen, 141 cadmium, 61 sulfide-hydrogen, 208 Metallic luster, 28 solutions, 39 Metal oxide oxygen, 140 Metals, reduction by, 149 Miscibility of metals, 46, 57 Molecular weights of dissolved metals, 44 Monatomic state of metal vapors, 5, 6 Mond nickel process, 203 Monotropy, 18 Nickel arsenic, 130 gold, 6 1 nickel sulfide, 122 silicon, 133 steel, 108, 109 Normality of electrons in metals, 31 Newton's metal, 68 Oxygen tension, 141 of metal oxides, 143 Erin's methods for, 146 Optical constants of metals, 21, 27, 28 Osmotic pressure, 42 Oxides, alloys of metals with, no Palladium oxygen, 144 Parkes process, 47 Passivity, 37, 36, 35, 34 Pattison process, 52 238 SUBJECT INDEX Perlite, 93, 94, 95 Permeability and passivity, 35 Pest, the tin, 12 Phase Rule, 133-138, 140, 150 Phosphorus alloys, 125 copper, 126 iron, 126 Potential of alloys, 72 Polymorphy, n Precipitation process, 207 Radiographs of gold sodium alloys, 54 Ratio of heat conductivity to electrical conductivity, 24 Rates of reaction, 201 Reactions of sulfides, 205 Reagent, Kurbatoff's, 95 Reduction by gases, 149 CO, 159 H,i 5 o metals, 149 Reflection, light, 27 constant, 28 Resistance, electrical, of alloys, 73 Roast, sulfatizing, 210, 211 Roasting process, 210 reactions, '216 Rontgen rays in the investigation of alloy structure, 58 Segregation in alloys, 52 Silicides, 131 Silicon copper, 142 effect on steel, 106 iron, 132 nickel, 133 Silver copper, 51 oxygen, 113, 141 silver sulfide, 118 lead, 51 sulfide-hydrogen, 208 Speiss, 129 Solidification curve of binary alloys, 149 Solutions, metallic, 39 colloidal, 39, 40 solid, crystallization diagrams of, 57, 58, 59 electrical conductivity of, 21 vapor pressure of metallic, 42 Sorbite, 95 Steel, chromium, 108, 109 SUBJECT INDEX 239 Steel, high speed, no tungsten, 108 chromium tungsten, 108 nickel. 108, 109 Sulfatizing roast, 210, 211 Sulfides, alloys between, 1 23 of metals and, 114 reactions of, 205 Sulfates, sulfur trioxide tension of, 213, 214 Sulfurets, 207 Sulfur tensions, 206 trioxide, dissociation of, 212 Supercooling, 7, 8, n Temperature coefficient of conductivity, 31, 75 eutectic, 51 Ternary alloys, 64 Tension, oxygen, 141 sulfur trioxide, Tin pest, 12 bismuth lead, 65 gray, 12 Transition points, 20 of iron, 15 of tin, 14 of metals, 17 Transition, volume change of, 15 heat of, 18 density change of, 14 Troostite, 95, 96 Tungsten steel, 108 Vapor density of metals, 6 pressure of metals, 5 metallic solutions, 42 Vapors, color of metal, 6 monatomic state of metal, 5, 6 Viscosity of gas, 31 Water vapor, dissociation of, 155 Woods metal, 67 X-rays in the investigation of alloy structure, 53 Zinc bismuth lead, 47 cadmium, 49, 55, 56 arsenic, 129 oxide carbon monoxide, 160 Ziervogel process, 210 THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW AN INITIAL FINE OF 25 CENTS WILL BE ASSESSED FOR FAILURE TO RETURN THIS BOOK ON THE DATE DUE. THE PENALTY WILL INCREASE TO 5O CENTS ON THE FOURTH DAY AND TO $1.OO ON THE SEVENTH DAY OVERDUE. OCT311SM NOV 1 t984 aiW 14 1938 JUL 27 19* JUL 27 W2 LD. ;YC 21726 O c UNIVERSITY OF CALIFORNIA LIBRARY mi