THE CRYSTALLIZATION OF IRON AND STEEL BY THE SAME AUTHOR HIGHER MATHEMATICS FOR STUDENTS OF CHEMISTRY AND PHYSICS, With Special Reference to Practical Work. With 142 Diagrams. 8vo. I2s. 6d. net. CHEMICAL STATICS AND DYNAMICS. Including the Theories of Chemical Change, Catalysis, and Explosions. (Text-Books of Physical Chemistry.} Crown 8vo. *js. 6d. LONGMANS, GREEN, AND CO. LONDON, NEW YORK, AND BOMBAY. THE CRYSTALLIZATION OF IRON AND STEEL AN INTRODUCTION TO THE STUDY OF METALLOGRAPHY BY J. W. MELLOR, D.Sc. LONGMANS, GREEN, AND CO. 39 PATERNOSTER ROW, LONDON NEW YORK AND BOMBAY 1905 All rights reserved DEDICATED TO MY FRIEND A. E. BONE, ESQ. DUNEDIN, NEW ZEALAND 14064-3 PREFACE THIS course of six lectures delivered to the Engineer- ing Students of the Staffordshire County Technical Classes at the Newcastle High School, in November and; December, 1904 was intended to summarize the results of the more important researches which have been made during the last ten years upon the constitution of alloys of iron and steel. The claims of the new science metallography which has revealed the internal structure of the metallic alloys, have been too long overlooked. The far-reaching importance of this knowledge is begin- ning to dawn upon the most conservative minds. Metallurgists are now rending the veil which has so long concealed the internal structure of the metals and their alloys; and many phenomena connected with the industrial treatment of the metals, so long inexplicable, are now yielding up their secrets to the indefatigable methods of modern scientific research. Unfortunately, much of this important work is disseminated throughout various scientific journals, vii vlii PREFACE and the facts are so frequently obscured beneath a mass of controversial matter that it is difficult for the uninitiated to get into touch with the work. In consequence, I have been lead to reproduce my lectures in a form suitable for publication, in the hope that they may help those who have not specially studied the subject to appreciate the remarkable chapter which metallurgists have recently added to physical chemistry. I desire to thank Profs. Arnold, Beilby, Cohen, Ewing, Heyn, Newth, Osmond, Popplewell, and Stead for permission to use their micro-photographs ; Dr. G. T. Beilby, the Secretary of the Iron and Steel Institute, the proprietors of The Philosophical Transactions and of The Iron Age for the loan of blocks; and Mr. H. Fowler for help with the diagrams. J. W. MELLOR. March 4, 1905. CONTENTS (The bracketed numbers refer to pages.) PAGE THE SOLIDIFICATION AND COOLING OF ALLOYS ... 1 1, Atoms and molecules (1) ; 2, The degradation of energy (3) ; 3, Passive resistance (4) ; 4, Allotropy (5) ; 5, Transi- tion temperatures (6) ; 6, Velocity of transformation (7) ; 7, Cooling curves (9) ; 8, Surfusion and recalescence (11) ; 9, The cooling curve of pure iron (12) ; 10, The freezing of salt water (16) ; 11, The solidification of copper-silver alloys (18) ; 12, The solidification of copper-antimony alloys (20) ; 13, The cooling of iron-carbon alloys (22) ; 14, The colour names for high temperatures (24). THE CONSTITUENTS OF IRON AND STEEL 26 15, Eutexia (26) ; 16, The relative proportions of ferrite, cementite, and pearlite (27) ; 17, Graphitic, hardening, and cement carbon (30) ; 18, Compounds, mixtures, and solutions (31) ; 19, The solidification of molten iron (33) ; 20, Martensite, hardenite, austenite (39) ; 21, Sorbite, troostite (42) ; 22, The phase rule (43). THE HARDENING, ANNEALING, AND TEMPERING OF STEEL 49 23, General properties of hypo- and hyper- eutectic steels (49) ; 24, The influence of rate of cooling (51) ; 25, The allotropic modifications of iron (54) ; 26, Annealing, tempering, and hardening of steel (58) ; 27, The law of mass action (60) ; 28, Theories of annealing and hardening (62). ii CONTENTS PAGE THE CRYSTALLIZATION OF IRON AND STEEL .... 67 29, The crystallization of iron (67) ; 30, The development of crystalline grains (68) ; 31, Grain size and fracture (70) ; 32, Influence of mechanical work (77) ; 33, Influence of other elements (79). THE INFLUENCE OF STRESS AND STRAIN 82 34, Intercrystalline or intergranular weakness (82) ; 35, Intracrystalline or cleavage weakness (85) ; 36, Birth, growth, and structure of crystals (88) ; 37, Effects of progressively augmented strain (90) ; 38, Effects of repeated alterations of stress (92). HOW TO PREPARE A SPECIMEN FOR THE MICROSCOPE 95 39, The cutting of a sample (95) ; 40, Filing and rough polishing (96) ; 41, Fine polishing (96) ; 42, Polishing in relief (97) ; 43, Etching (98) ; 44, Osmond's polish attack (102) ; 45, Heat tinting (104) ; 46, Mounting (104) ; 47, Preservation of polished specimens (105) ; 48, The microscope and its accessories (105) ; 49, Photography (107) ; 50, Miscellaneous (107). APPENDIX GLOSSARY Ill INDEX 135 THE '. f .-.\< CRYSTALLIZATION OF IRON AND STEEL THE SOLIDIFICATION AND COOLING OF ALLOYS i. Atoms and Molecules CHEMISTS have invested matter with an imaginary constitution, which explains very well the various transformations which matter undergoes. Matter is supposed to be made up of extremely small particles, called molecules. No successful attempt has been made to describe how the molecules associate together except in the case of crystalline substances. Here all the evidence points to a symmetrical and fixed mode of arrangement, which finally produces regular geometrical figures called crystals. More or less approximate attempts to calculate the size of the molecules show that, if a drop of water were magnified to the size of the earth, the size of the molecules of water would be between that of small shot and of cricket balls. Molecules lie quite outside the range of observation, and we must accept here, in i B 2 CRYSTALLIZA TION OF IRON AND STEEL i good faith, the large mass of circumstantial evidence accumulated by the chemist. By analysis it. has been found that the infinite variety of substances known to man can all be reduced to. about .eighty simple forms, called elements. No analyst ha -ever .separated from an element anything but itself. Pure iron will yield nothing but pure iron. Iron is, therefore, an element. Pure copperas, on the other hand, will furnish iron, sulphur, and oxygen. Copperas is not an element. Facts like these seem to indicate that molecules are made up of still smaller particles. These are called atoms. Atoms of the same kind make up the molecules of elements; atoms of different kinds make up the molecules of compounds. All substances known to man are supposed to be made up from different combinations of some eighty different kinds of atoms. Eecent investigations seem to show that the different atoms are different com- binations of still smaller particles, called corpuscles, or, if they be charged with electricity, electrons. All corpuscles are the same. Matter, whether it be a mummy, a piano stool, or a toothpick, is essentially one universal substance. Variety enters when the corpuscles arrange themselves in groups of atoms, when the atoms unite into molecules, and when the molecules aggregate into masses of matter. The stages are Corpuscles > atoms -> molecules > matter en masse. The molecules of a gas lead a more or less inde- pendent existence. This is illustrated by the rapidity with which the molecules of, say, ammonia gas travel from one end of a room to the other and affect the 2 THE COOLING OF ALLOYS 3 sense of smell. In liquids, however, the molecules are much less mobile. This can easily be proved by dropping a small grain of aniline dye into a tumbler of clear still water. The water will be uniformly coloured in a few days. The molecules of a solid substance have practically lost their mobility. But not all. Carbon laid in contact with pure, hot, solid iron will diffuse into the mass of the metal ; and gold in contact with lead will, in a few years, diffuse into the lead in appreciable quantities. 2. The Degradation of Energy The infinite variety of changes continually taking place in the properties of bodies around us is often said to be due to the action of an external agent, called energy, upon matter. Just as water will always run down from a high to the lowest level that circumstances will permit, so will energy at a high potential always run down to energy at a low potential. And one of the most interesting phenomena in connection with all natural changes is this constant running down or degradation of energy. Still keeping the same analogy, just as water may descend from the top of a hill in many ways rivers or rain, underground channels, glaciers, or avalanches so may energy give rise to electrical, thermal, or chemical phenomena in its descent from a high to a low potential. But I need say little to engineers on this subject. The electric light, steam engine, electric tramcars, gas engines, water-wheels, watches, and clocks all bear testimony to the ubiquity of the law. An ancient philosopher has said that all things are in motion, and we might add 4 CRYSTALLIZA TION OF IRON AND STEEL 2 that the motion always involves & degradation of energy. Motion only ceases when energy has run down to the level of its surroundings. The system is then said to be in a state of equilibrium. 3. Passive Resistance There is also another remarkable law the law of passive resistance. Equilibrium may be apparent. The running down of energy may be resisted in some way. It is a common thing to find energy at a higher potential than we should expect. Energy does not always, of itself, run down to its lowest level. Just as the throttle-valve of a steam-engine must be moved before the degradation of high-pressure energy com- mences, and the engine can start on its journey, so may a preliminary impulse be required to set the process of degradation in motion. We therefore distinguish between two states of equilibrium. The one is stable, the other unstable. The one is a real state of equilibrium, the other is only apparent. When you see water in a liquid state at a temperature below its normal freezing-point, C., you know that some agent must be at work which prevents the freezing of the water. This unknown agent is called passive resistance. Sodium thiosulphate is a convenient substance to illustrate these facts. At ordinary temperatures this salt is a white crystalline solid. On heating to 56 it melts to a clear liquid. This is also the freezing-point of the liquid. But it is possible to cool the molten salt down to the temperature of the room without solidification. The sodium thiosulphate is then said to be in a state of apparent or false equilibrium, FIG. 1. Octahedral Sulphur. (G-. S. Newth.) FIG. 2. Prismatic or " Needle-shaped " Sulphur Crystals, (fl. S. Xewth.) [To face p. 5. 4 THE COOLING OF ALLOYS 5 and it can be kept in this state an indefinite time. Now put a crystal of sodium thiosulphate into the liquid mass. The passive resistance is overcome in some way, for now the liquid assumes the stable crystalline condition, and during the transition from the liquid to the solid states energy is degraded. 4. Allotropy Sulphur, at the temperature of this room, is a pale yellow crystalline solid. The crystals are shaped like octahedrons (Fig. 1). If sulphur be heated above 96, these pass into needle-shaped crystals (Fig. 2). Still further, if sulphur be heated to near its boiling-point, and suddenly quenched by pouring into cold water, an amorphous, non- crystalline, plastic, and elastic mass is produced. Here, then, you have the element sulphur existing in three different forms plastic, octahedral, and needle-shaped crystals. Each form is said to be an allotropic modification of sulphur. This word allotropic. How can the same substance exist in different forms ? Just as the builder can with the same kind of bricks build up various structures, so can Nature with the same kind of atoms build up molecules with very different properties. The atoms of sulphur, for instance, may form molecules which crystallize as octahedral or as needle-shaped crystals ; atoms of carbon form three allotropic modifications diamond, graphite, and amorphous carbon. Allotropy occurs when a substance exists in two or more forms which differ in some of tlieir properties. The term is not usually applied to the different states of aggrega- tion of a substance solid, liquid, or gaseous. Allo- tropic transformations are usually accompanied by 6 CRYSTALLIZA TION OF IRON AND STEEL 4 changes in the internal energy of the substance concerned. Energy is at a higher potential in one allotropic form than in another. The one form which has energy at the higher potential must be unstable. 5. Transition Temperatures The octahedral crystals are alone stable at ordinary temperatures. Both plastic sulphur and the needle- shaped crystals pass spontaneously into octahedral crystals at the temperature of the room. But above 96 the needle-shaped crystals are stable, while the plastic and the octahedral crystals slowly assume the needle-like form. You might just as well try to prevent water running down a hill as to prevent these changes taking place at these temperatures. Plastic sulphur will crystallize in octahedrons at the tem- perature of the room because the process involves the running down of energy. Sulphur, therefore, has two crystalline forms, one of which is stable above 96, and the other below 96. The critical temperature, 96, is called the transition temperature. Mercury iodide also exists in two allotropic forms, one of which is red and the other yellow. The red form is stable at ordinary temperatures, and passes into the yellow modification when heated above 126. The yellow form is stable above 126, but passes into the red form when cooled below the transition temperature, 126. After a particularly cold winter, 1867-68, some blocks of tin stored in the Customs House, and some tin buttons in the Military Stores at St. Petersburg had mysteriously crumbled to a grey powder. It has since been proved that tin exists in two allotropic FIG. 3. Surface of Diseased Tin. (E. Cohen.) [To face p. 7. 6 THE COOLING OF ALLOYS ^ modifications white malleable metal, and grey powder. The transition temperature is 20, just a little above the average temperature of the air. The grey powder is the stable form below 20. Hence it follows that all the malleable tin in the world, except on the hottest summer days, is in an unstable condition. It is only passive resistance of some kind which prevents all the tin vessels in the world slowly crumbling to powdered grey tin. 6. Velocity of Transformation We have just seen that a crystal of sodium thio- sulphate will make the unstable liquid thiosulphate pass into the stable form, so will the presence of a little grey tin facilitate the transformation of white into grey tin. This crumbling of tin to a grey powder is known as the tin pest. The disease is, therefore, infectious. The surface of a piece of diseased tin is shown in Fig. 3. The change is slow at ordinary temperatures. But articles of tin which have been buried a few hundred years are in almost every case in a more or less advanced state of disintegration. The comparative rigidity or immobility of the molecules of a solid offers a kind of frictional resistance to change, analogous to the action of a brake upon the wheels of a car. If it were not for passive resistance the speed of transformation from one allotropic form to another would be faster -the more distant the temperature away from the transition-point. Experi- ment shows that the rate of transformation of white into grey tin increases as the temperature is reduced below the transition-point, 20. At 50, for instance, the transformation is very rapid. As a general rule, 8 CRYSTALUZA TION OF IRON AND STEEL 6 passive resistance increases as the temperature falls below the transition-point. The one effect works against the other. If 6 be the prevailing temperature, we may write transition temperature less & Velocity of change at e = pa88ive resi8tance at g o Or, if V denotes the velocity of transformation, R the magnitude of the passive resistance at 0, and E the difference of the temperature between the transition point and 0, we have " R a result which bears a close formal analogy with Ohm's well-known law. It may be assumed that E also represents the amount of energy to be degraded in the process. This formula states in symbols the observed facts that the greater the value of E, the greater the velocity of transformation ; and the greater the value of R, the less the velocity. These experiments teach us four important facts which must be clearly understood : (1) A substance may exist in two or more forms having different properties. (2) Only one of these forms is, in general, stable at any given temperature. (3) The transformation of a substance from its unstable to its stable form occupies time. (4) The transformation from the unstable to the stable form may be hindered or even arrested by passive resistance for an indefinite time. The phenomena are not always so obtrusive as the changes which take place with sulphur, tin, and 7 THE COOLING OF ALLOYS mercury iodide. We naturally ask, how can we tell whether a substance is capable of existing in different allotropic forms ? As a matter of fact, we select some physical property of the substance and measure it at different temperatures : if there is a sudden change in the physical property of the substance at any particular temperature, we infer that there is some drastic change going on in the internal structure of the substance. 7. Cooling Curves Let the temperature of a cooling copper bar at 200 be measured every ten minutes. Let distances at right angles to the line 0-200 (Fig. 4) represent 200* ISO" 50 w 20 40 60 Time FIG. 4. Cooling Curve of Solid Copper. time, and vertical distances from the line - 60, the corresponding temperatures of the bar. We thus obtain the series of points shown in Fig. 4. Draw a line so as to lie most evenly among the points. The result is a so-called cooling curve. The simple form of the cooling curve in Fig. 4 gives no evidence of any sudden change in the nature of the cooling copper. If a curve is drawn for water cooling down from io CRYSTALLIZATION OF IRON AND STEEL 7 20 to 20 C., we get a terrace in the cooling curve, as shown in Fig. 5. This tells us that some change has taken place in the nature of the substance at 0. We cv 10' -vr -4 v \ Freezit W \ ^ X ^ } 20 40 60 Time FIG. 5. Water. see directly that this change corresponds with the passage of water from the liquid to the solid state of aggregation. Now draw the cooling curve of molten sodium thiosulphate. We know that the molten liquid " ought" to freeze at 56 (Fig. 6). But the cooling curve goes 90 70 56 50' 30 10 20 40 60 Time FIG. 6. Sodium Thiosulphate. on quite normally below that temperature until, at length, there is a great evolution of heat, and the liquid 8 THE COOLING OF ALLOYS u solidifies. The temperature may even rise above 56. The cooling curve of the solid is quite normal. The amount of heat evolved as the molten liquid solidifies corresponds with the " latent " heat absorbed as the solid melts. 8. Surfusion and Recalescence The molten sodium thiosulphate as it cools down the " surfusion " curve is at a lower temperature than its normal point of solidification or freezing. The liquid is then said to be in a state of superfusion or surfusion. The system is in unstable equilibrium. We may get a similar state of things when a saturated solution of a substance is slowly cooled. More salt may be in solution than the true solubility of the salt. The result is a supersaturated solution. Agitation, or the addition of a trace of something which will serve as a nucleus for crystallization, will generally suffice to start the system on its passage to a state of stable equilibrium. But when the trans- formation does set in, it usually takes place very rapidly, and is accompanied with a rise of temperature. The cooling curve is distorted in a corresponding manner (Fig. 6). It is interesting to put a little ether in a small bulb blown at the end of a piece of glass tubing, placed in supercooled sodium thiosulphate (Fig. 7). Drop in a crystal of sodium thiosulphate. The evolution of heat as the liquid solidifies raises the temperature high enough to vaporize the ether. The vapour of ether will burn at the mouth of the tube with a steady flame when ignited. When a steel bar is cooling, an evolution of heat 12 CRYSTALLIZATION OF IRON AND STEEL 8 occurs at about 690. The amount of heat evolved is so great that the metal visibly brightens in colour. IUUU 800" 600* W \ ^v ^ A "" Ar, \ ^ ^ FIG. 7. 20 40 60 80 100 TIME FIG. 8. Kecalescence. The phenomenon is called recalescence. The cooling curve is shown in Fig. 8. 9. The Cooling Curve of Pure Iron The cooling curve of iron from the molten condition is shown in Fig. 9. The iron was practically pure. It M ^ fccxtn \ 7 WMT IGOO' 600' ?firt \ An 8&F ft *9 ^ i ^ 20 40 60 Time. FIG. 9. Iron. only contained O'Ol per cent, of carbon. F. Osmond, or 9 THE COOLING OF a celebrated French metallurgist, maintains that the existence of the transition-points, or discontinuities, Ar a and Ar2, in the cooling curve of the solidified metal, points to the existence of three allotropic modifications of solid iron : i. Alpha Iron. Below Ar2, that is 750, we have what he calls a-iron, or alpha iron. ii. Seta Iron. Between Ar 2 and Ar 3 , that is between 750 and 860, we have what he calls /3-iron, or beta iron. Beta iron is non-magnetic. Heat is evolved when iron passes from the /3- to the a-state, and mag- netic properties are developed at the same time. iii. Gamma Iron. Above the Ar 3 critical point, namely 860, we are supposed to have y-iron, or gamma iron. This variety is non-magnetic. Each critical point is found to be associated with a change in the mechanical properties, the microscopic appearance, the electrical conductivity, the magnetic properties, and the specific gravity of the metal. 1 The changes which occur during the cooling of a substance are reversed when the substance is heated. The cooling curve of steel, with 1-2 per cent, of carbon, shown in Fig. 10, is reversed on heating, as shown by the heating curve in the same diagram. There is only one critical point at about 690, called the Ar! critical point. The critical points Aci, Aca, Acs on the heating curve of mild steel are generally a few degrees higher 1 O. Boudouard, Jaurn. Iron and Sted Intt., 63. i. 229, 1903 ; H. le Chatelier, Compt. Rend., 128. 1444, 1899; 129. 299, 331, 497, 1899; Metallographist, 2. 334, 1899 ; 3. 38, 152, 1900; G. E. Syedelius, Phil. Mag., [5], 46. 173, 1898 ; G. Charpy and L. Grenet, Compt. Bend., 124. 540, 598, 1902; MetaUographist, 6. 240, 1903; S. Curie, ibid., 1. 107, 229, 1898. 14 CRYSTALLIZA TION OF IRON AND STEEL 9 than the corresponding points Ari, Ar 2 , and Ar 3 respectively. There seems to be a kind of molecular inertia, or lag, which prevents the y to )3, the |3 to a, or the reverse changes taking place sharply. The critical 900ir soo* 700 600* 50(T Ac. 20 40 60 Time FIG. 10. Steel. points on the cooling curve are, in consequence of this lag, a few degrees below the true critical point. The lag induces a state somewhat analogous to surfusion in molten sodium sulphate. The critical points on the heating curve are a little too high, and for a similar reason. 1 The critical points of iron really represent ranges of temperature, although, for the sake of inconvenience, we call them points. The Ar 3 with soft steel commences at 845, and finishes at 800 ; it is most marked at 820. The Ar 2 extends from 755 to 710 ; and the Ari from 680 to 645. The "r" of "Ar" comes from the French word refroidissant, for cooling; the "c" of "Ac" from chauffant, heating. This notation is due to D. Tscher- noff, 2 the Russian metallographist. 1 F. Osmond, Metallographist, I. 270, 1893; 2. 169, 1899; H. M. Howe, ibid.,2. 257, 1899; M. Aliament, La ledricien, 49, 1903. 2 Otherwise spelt " D. Chernoff." 9 THE COOLING OF ALLOYS 15 Now y-iron is said to be hard, a-iron soft. If, therefore, y-iron be quickly cooled past the Ar 2 critical point, the passage of the hard y-iron to the soft a-iron is retarded ; the iron is then in an unstable hardened condition, ready on the least provocation to pass into the stable soft ,a-form. We naturally ask is there any method of helping the passive resistance so that the iron will not readily change from, say, the y to the a modifications at ordinary atmospheric temperatures? It is supposed that the presence of many foreign substances, like carbon, nickel, and manganese, augment the passive resistance so as to render the hard y-iron more stable and permanent at low temperatures. On the other hand, the presence of chromium, tungsten, aluminium, silicon, phosphorus, arsenic, and sulphur facilitate the passage of hard beta 1 iron to the soft alpha iron. The influence of minute traces of foreign substances upon the properties of the metals is a most important subject. The effects seem inexplicable. The presence of 0*05 per cent, of tellurium alters the properties of bismuth so much that we seem to be dealing with a totally different substance; a few hundredths of one per cent, of sulphur will determine the success or failure of iron; and the presence of O'l per cent, of bismuth in copper lowers its conductivity so much that if copper so contaminated had been alone avail- able, it would have been fatal to the success of the Atlantic cable. 1 Either "gamma" or "beta." We are not sure which. Both are supposed to be hard. 16 CRYSTALLIZATION OF IRON AND STEEL 10 10. The Freezing of Salt Water I have been speaking of pure or almost pure iron, and now we naturally turn to alloys of iron with carbon. Cast iron and steel are, as you well know, alloys or solidified solutions of carbon in iron. These alloys are so complex that it will be profitable for us to examine some other solutions which do not present such complications as occur in the case of the iron- carbon alloys. It is a most interesting fact to find that the same general laws hold good for the cooling of metallic alloys, for the separation of ice when sea water is frozen, the separation of crystals in the glazes of the potter, the devitrification of old glass, for alloys of carbon and iron, and the formation of rocks when the world was a-building. True enough, with the iron- carbon alloys other phenomena are superposed upon, and hence modify the course of the simple phenomenon as it occurs during the freezing of sea water. The freezing-point of a 5 per cent, solution of sodium chloride is below that of pure water. If more salt be added, the freezing-point is reduced still more ; and this goes on until the solution contains 23^ per cent, of sodium chloride, when further additions of salt raise the freezing-point. The experimental results are shown in Fig. II. 1 But the experiment reveals something more in- teresting than this. If the solution contains, say, 5 per cent, of salt, pure ice separates out as the solution freezes, and, in consequence, the solution which remains unfrozen has more than 5 per cent, of salt dissolved in it. The freezing-point of the 1 F. Guthrie, Phil. Mag., [5], 1. 354, 1876. 10 THE COOLING OF ALLOYS mother liquid is therefore lower than that of the original solution. This separation of pure ice, and the lowering of the freezing-point, goes on along the curve AP (Fig. 11), until the solution contains 23 J- per cent, of salt. After that, the residual mixture freezes en masse at 22. If the solution contains more than 23J per cent, of salt, then pure salt separates from the solution, and 10 20 23i 30 %SaU FIG. 11. Freezing Curves of Aqueous Salt Solutions. the separation of salt, and the lowering of the freezing- point of the solution goes on along the curve BP (Fig. 11), until the solution contains 23J per cent, of salt, and the whole residue then solidifies at - 22. If the solution contains just 23J per cent, of sodium chloride, it freezes en masse at 22. No other mixture of salt and water freezes at a lower temperature than this. Hence this mixture is called a eutectic mixture. Guthrie used to think that the mixture which separated at this temperature was a definite chemical compound of water with salt, which he called a cryohydrate. Ponsot calls the mixture a "cryosel." The term eutectic mixture is to be preferred. We know now that Guthrie's cryohydrate is nothing but a mechanical i8 CRYSTALLIZATION OF IRON AND STEEL 10 mixture of ice and salt. The one is entangled with the other. Under the microscope the crystals of ice can be seen lying in a matrix of salt. 1 ii. The Solidification of Copper-Silver Alloys A like phenomenon occurs when molten mixtures of silver and copper are allowed to cool. At 770, when the alloy has the composition 28 per cent, of copper and 72 per cent, of silver, the whole solidifies en masse. If the mixture contains less than this per- centage of copper, pure silver separates at temperatures along the " silver " line (Fig. 12) ; while, if the molten 1100" 1000' 000 700 A ^ Lute 25 50 75%COPPER FIG. 12. Fusibility Curves of Copper-Silver Alloys. mixture contains more than 28 per cent, of copper, pure copper separates, and continues separating along the "copper" line until the mixture has the above composition, when the whole solidifies as a eutectic mixture at 770. No other alloy of silver and copper 1 A. Ponsot, Ann. Chim. Pliys., [7], 10. 79, 1897 ; T. Andrews, Proc. Boy. Soc., 40. 544, 1890; 48. 106, 1890; J. Y. Buchanan, Proc. Roy. Soc. Edin., 14. 129, 1888. FIG. 13. Polished Surface of Cu-Ag Alloy. (F. Osmond.) 'FiG. 14. Polished Surface of Cu-Ag Alloy. (F. Osmond.) [To face p. 19. ii THE COOLING OF ALLOYS 19 melts at so low a temperature. The resulting alloy is a network of the two metals, pure silver and pure copper, as shown in Fig. 13, where the heterogeneous nature of the alloy is clearly seen. In Fig. 14 we have an alloy of 15 per cent, of copper and 85 per cent, of silver. The alloy has a greater percentage of silver than the eutectic alloy, and in consequence silver separates out until the residue has the eutectic com- position. This is in harmony with the microscopic appearance of the alloy, which shows large masses of silver embedded in a network of the eutectic alloy. An alloy of these metals appears to possess two freezing-points: (i.) The temperature at which the mass begins to solidify ; and (ii.) the temperature at which the whole is solidified. The pasty condition of solder tin with 66 per cent, of lead is due to the fact that there are two freezing-points. Solid lead separates first, and on this fact depends the facility with which a joint can be wiped with plumber's solder. Wa get similar results with binary alloys of antimony and lead, tin and lead, tin and bismuth, tin and zinc, lead and silver, zinc and aluminium, and with copper and gold. 1 The fusibility curve is very much simpler if the one constituent is mutually soluble in the other in all proportions. The fusibility curve is then approxi- mately a straight line (AB, Fig. 15). This is the case 1 W. Campbell, Jonrn. Franklin Inst, 154. 1, 131, 201, 1902; Metallographist, 5. 286, 1902; J. E. Stead, {bid., 5. 110, 1902; H. M. Howe, {bid., 5. 166, 1902 ; A. W. Kapp, Drude's Ann., 6. 754. 1901 ; W. C. Roberts- Austen, Proc. Roy. Soc., 23. 481, 1884 ; 63. 452, 1898 ; G. T. Heycock and F. H. Neville, Phil. Trans., 189. 25, 1897 ; A. Dahms, Wied. Ann., 54. 486, 1895 ; H. le Chatelier, Compt. Rend., 118. 350, 415, 800, 1894. 20 CRYSTALLIZATION OF IRON AND STEEL n with alloys of silver and gold. The same thing occurs with alloys of antimony and bismuth. 1200 1100 1000' 900 ^0 25 50 75%SILVER FIG. 15. Fusibility Curve of Gold-Silver Alloys. 12. The Solidification of Copper-Antimony Alloys Silver and copper do not form a chemical compound. Many metals, however, do form compounds. Copper and antimony, for example, form a compound having the chemical formula SbCu 2 , or, according to H. le Chatelier, Sb 2 Cu 8 . This behaves as if it were a single and distinct element. The addition of either copper or antimony lowers the melting-point of the compound SbCu 2 in the ordinary way. In Fig 16, G represents the melting-point of the pure compound SbCu 2 . The line CPi represents the effect of the addition of antimony to the compound ; while the line GP 2 repre- sents the effect of additions of copper. There are two eutectic points, the one, P 2 , corresponding with the eutectic mixture of the compound SbCu 2 with copper, 1 A. Gautier, Bull. Soc. d' Encouragement, [5], 1. 1293, 1896 ; W. C. Roberts-Austen, Proc. Roy. Soc., 67. 105, 1900; with T. K. Roso, 71. 161, 1903. 12 THE COOLING OF ALLOYS 21 and the other, PI, corresponding with a eutectic mixture of SbCu 2 with antimony. The results depicted in Figs. 11, 12, and 16 can be represented graphically in another way. 1 Take the more complicated case (Fig. 16). Let the horizontal nor 800 TOO* A 500 F> 25 50 75% COPPER FIG. 16. Fusibility Curve of Copper-Antimony Alloys. line (Fig. 17) represent the ultimate composition of the alloy in terms of copper; the vertical lines, the structural composition in terms of the various con- stituents copper, SbCu a , antimony, or the two 10 20 30 40 50 60 70 FIG. 17. (After A. Sanveur.) 80 90%COPPE( eutectics. As an example, an alloy with 10 per cent, of copper will have 40 per cent, of the first eutectic A. Sauveur, MetallograpMut, 1. 103, 1898. 22 CRYSTALLIZATION OF IRON AND STEEL 12 and 60 per cent, of antimony ; an alloy with 33 per cent, of copper will have 30 per cent, of the first eutectic, and 70 per cent, of SbCu 2 . Alloys of gold and antimony present similar phenomena. 1 13. The Cooling of Iron-Carbon Alloys But we can go a step further. Let us consider what takes place when an iron bar containing, say, 0*6 per cent, of carbon and 99*4 per cent, of iron cools from 900. The cooling curve shows nothing very remarkable until a temperature of about 720 is attained. There is then a sudden evolution of heat. The critical points, Ar 3 and Ar 2 , of pure iron coalesce into one. At this point pure iron, or ferrite, as 1200 0-5 1-0 I-5%CARBON Ifrj. 18. Cooling of Solid Iron." Howe calls it, separates from the solid solution. The separation of ferrite goes on along the curve AP (Fig. 18) until the temperature reaches about 660, 1 C. T. Heycock and F. H. Neville, Proc. Boy. Soc., 68. 171, 1901 ; J. E. Stead, Journ. Soc. Cham. Ind., 17. 1111, 1898; Metallographist, 1. 179, 1898; 2. 314, 1899; G. Charpy, ibid., 1. 87, 192, 1898. PIG. 19. Surface of Lamellar Pearlite. (F. Osmond.) FIG. 20. Surface of Granular Pearlite. (E. Heyn.) [To face p. 23. 13 THE COOLING OF ALLOYS 23 when another recalescence point occurs (Ari). No other noteworthy change occurs as the system cools down to the normal temperature of the atmosphere. Other alloys containing different amounts of carbon furnish a set of curves quite analogous to the freezing curves of salt water and of silver-copper alloys. But with iron these changes take place in the solid cooling metal. If the alloy contains less than 0'89 per cent, of carbon there is a separation, or, better, segregation of ferrite ; if the alloy contains more than 0*89 per cent, of carbon, there is a separation, not of carbon, but of a chemical compound of carbon with iron, called normal iron carbide, or cementite, and represented in chemical symbols by Fe 3 C. Cementite contains 6*9 per cent, of carbon. The separation of cementite occurs along the curve BP (Fig. 18). We are there- fore dealing with a mixture of ferrite and of cementite. The eutectic alloy contains 13 per cent, of cementite (i.e. 0-89 per cent, of carbon) and 87 per cent, of ferrite roughly, six of ferrite to one of cementite. The microscopic appearance of the eutectic reminds one forcibly of other eutectic mixtures. The eutectic mixture of cementite and ferrite is called pearlite, owing to the fact that it generally shows the rainbow tints of mother-of-pearl under microscopic treatment. Sorby called it the " pearly constituent " of steel. Fig. 19 shows a fine specimen of lamellar pearlite obtained from a steel containing 1/0 per cent, of carbon. The black streaks are ferrite. Fig. 20 shows a specimen of what is sometimes called granular pearlite, from a 1 According to H. le Chatelier, Journ. Iron and Steel Inst., 61. i. 40, 1902, there are two or three allotropic forms of cementite. 24 CRYSTALLIZATION OF IRON AND STEEL 13 forged bar of crucible steel containing 0'92 per cent, of carbon. It is magnified 1240 times. 14. Colour Names for High Temperatures I always speak of temperatures on the centigrade scale. For reference purposes I will give a scale of corresponding centigrade and Fahrenheit tempera- tures, and also the colour names as determined by White and Taylor with the Le Chatelier pyrometer. 1 These results are said to be more accurate than Pouillet's old numbers. The colour names do not really correspond with any particular temperature, but rather with a certain range of temperature. The results, too, depend upon so many external factors physical and physiological that different numbers might be obtained by different observers, and by the same observers at different times. Colour names. 532 566 635 682 746 835 941 QQfi /yo 1079 1205 990 1050 1175 1250 1375 1550 1650 1725 1825 1975 2200 Dark blood red, black red, incipient red, rouge naissant. Dark red, blood red, low red, rouge sombre. Dark cherry red, incipient cherry red, cerise naissant. Medium cherry red. Cherry, full red, cerise. Light cherry, bright cherry, scaling heat, 2 light red, cerise clair. Orange, salmon ; free scaling heat, orange foncee. Light orange, light salmon, orange clair. Yellow. Light yellow. White, blanc. 1 M. White and F. W. Taylor, Metallographist, 3. 41, 1900 ; H. M. Howe, ibid., 3. 43, 1900; C. S. M. Pouillet, Compt. Eend., 3. 784,1836. 2 Scale forms and adheres, i.e. does not fall away from the piece when cooled in air. THE COOLING OF ALLOYS There is also a colour scale of temperatures used in the tempering and annealing of iron. The colours are due to the formation of a thin film of oxide on the surface of the metal. c. F. Colour names. 220 408 Faint yellow. 230 446 Straw yellow. 255 491 Brownish yellow. 265 509 Purple brown. 277 280 531 536 Purple. Violet. 288 550 Light blue. 293 559 Dark blue. 316 601 Blackish blue. 400 732 Black THE CONSTITUENTS OF IRON AND STEEL 15. Eutexia THE phenomenon of eutexia, as we have seen, plays a most important part in the structural changes which take place during the cooling of iron. Figs. 11, 12, and 16 abbreviate in one diagram the cooling curves of every possible mixture of the two components con- cerned. Eleven cooling curves of silver-copper alloys, containing various proportions of the two metals, are shown in Fig. 21 ; and when arranged in this manner Time, FIG. 21. Set of Cooling Curves of Ag-Cu Alloys. you can see the genesis of the eutectic diagrams with- out further study. 26 1 6 CONSTITUENTS OF IRON AND STEEL 27 I hope that you can trace the analogy between the freezing of salt water, the solidification of silver-copper alloys, and the cooling of steel bars. Let me attempt a comparison of the cooling of steel with the freezing of a solution of salt and water, leaving you to extend the comparison to the solidification of the copper-silver alloy. Solution of salt and water. Iron-carbon alloy. 1. Larger per cent, of water than the eutectic mixture. Crys- tals of ice in cryohydrate. 2 % Larger per cent, of salt than the eutectic mixture. Crys- tals of salt in cryohydrate. 3. Same composition as eutectic mixture (23J per cent, of ealt). Solidifies at -22 as network of ice and salt, i.e. of the cryohydrate. 1. Larger per cent, of iron than the eutectic mixture. Crys- tals of ferrite in pearlite. 2. Larger per cent, of carbide than eutectic. Grains or bands of cementite in pear- lite. 3. Same composition as eutectic mixture (0'9 per cent, of carbon). System consists en- tirely pearlite. 16. The Relative Proportions of Ferrite, Cementite, and Pearlite The remarkable analogy which obtains between the constitution of steel and of a crystalline igneous rock was first pointed out by H. C. Sorby, of Sheffield, in 1864. x In granite, for instance, we have certain specific constituents of definite chemical composition mica, feldspar, and quartz; while in the iron-carbon 1 H. C. Sorby, B. A. Reports, ii. 189, 1864; i. 139, 1865; Journ. Iran and Steel Inst., 29. i. 140, 1886 ; 33. i. 255, 1888. Steel," said Sorby, " must be regarded as an artificial crystallized rock, and to get a complete knowledge of it, it must be regarded as such." 28 CRYSTALLIZA TION OF IRON AND STEEL 16 alloys there are also two well-defined constituents ferrite or pure iron, cementite or iron carbide and the eutectic mixture, pearlite. All three constituents were discovered by Sorby, and their present names were suggested by Howe. They segregate from the cooling solid according to the laws just outlined, and they can be readily seen when a polished surface of the alloy is examined under the microscope. If large tracts of ferrite be present, the alloy contains less than 0'89 per cent, of carbon ; while if large areas of cementite be present, the alloy contains more than 0*89 per cent, of carbon. Fig. 22 shows '0 0-5 1-0 1-5 2-0 2-5 3-0 3'5 407OFCARBON FIG. 22. (After A. Sauveur.) diagrammatically the relation between the percentage of combined carbon, and the percentage amounts of the constituents ferrite, cementite, and pearlite in various steels and cast irons. The diagram shows, for example, that a steel containing O5 per cent, of carbon has nearly 58 per cent, of pearlite, and 42 per cent, of ferrite ; cast iron containing 3*5 per cent, of carbon has 62 per cent, of pearlite, and 38 per cent, of cementite. Eutectic, or saturated, steel, composed wholly of pearlite, has 0'9 per cent, of carbon ; a steel containing less than this amount of carbon is said to be a hypo- eutectic, or unsaturated, steel ; while if the steel 16 CONSTITUENTS OF IRON AND STEEL 29 has more than 0'9 per cent, of carbon, it is said to be a hyper-eutectic, or supersaturated, steel. NOTE. To calculate the amount of pearlite and excess ferrite or excess cementite in steel containing a given percentage of carbon, assum- ing that the iron has been cooled slowly, so that everything is in equilibrium. The formulae are deduced by the ordinary methods used in chemical computations. I. Hypoeutectic or unsaturated steels (ferrite in excess). Per cent, of cementite in pearlite = 15 x per cent, of carbon . . (1) fenite in pearlite = 6'4 X per cent, of cementite . . (2) ,. pearlite = per cent, ferrite + per cent, cementite . . (3) excess ferrite = 100 per cent, pearlite .... (4) total ferrite = per cent, excess ferrite + per cent. ferrite in pearlite (5) ll.Butectic steels (0'9 per cent, of carbon), pearlite alone present. Only (1) and (2) is required for the calculation. III. Hypereutectic or supersaturated steels (cementite in excess). Per cent, of total cementite = 15 x per cent, of carbon .... (6) ferrite in pearlite = |100 per cent, of total cementite (7) cementite in pearlite = per cent, of ferrite -f- 6'4 . . (8) pearlite = per cent, cementite + per cent, ferrite . . (9) cementite in excess = per cent, total cementite per cent, in pearlite (10) EXAMPLES. (1) A low carbon steel has 0-1 per cent, of carbon. What are the percentage amounts of the constituents ? Obviously the ferrite will be in excess, and there will be 15 x O'l = 1*5 per cent, of cementite in pearlite. 6'4 x 1*5 = 9-6 per cent, of ferrite in pearlite. 1-5 -f 9-6 = 11 -1 per cent, of pearlite. 100 - 11-1 = 88-9 per cent, of ferrite in excess. 88-9 + 9-6 = 98-5 per cent, of total ferrite. (2) Calculate the percentage amounts of the various constituents of white cast iron containing 2 per cent, of carbon. The cementite will be in excess, and there will be 15 X 2 = 30 per cent, of cementite. 100 - 30 = 70 per cent, of ferrite. 70 -*- 6'4 = 10'9 per cent, of cementite in pearlite. 10-9 4- 70 = 80-9 per cent, of pearlite. 30 10 f 9 = 19-1 per cent, of cementite in excess. 30 CRYSTALLIZATION OF IRON AND STEEL 17 17. Graphitic, Hardening, and Cement Carbon Molten iron can dissolve as much as 7 per cent, of carbon. The amount dissolved depends upon the temperature. At the eutectic temperature, between 1100 and 1200, the molten solution contains 4*3 per cent, of carbon ; and the eutectic mixture when solidified has 2*3 per cent, of carbon in the form of graphite. The carbon which is rejected as the molten solution cools is called graphitic carbon. The separation of graphitic carbon goes along the line BP (Figs. 24 and 25). Graphitic carbon is found almost exclusively in cast iron. It is not acted upon by boiling hydro- chloric acid, nor by nitric acid (sp. gr. T2). The graphite which is formed during the solidification of the alloy is generally in the form of thin plates or flakes, often one-eighth of an inch or more in thickness, as shown in Fig. 23, which represents a specimen of grey pig iron with 0*72 per cent, of silicon, magnified 90 diameters. The rounded areas consist of cementite surrounded by depressed pearlite. About 1822, Faraday and Caron noticed that the carbon in quenched and annealed steels, containing no graphitic carbon, exists in two distinct forms. One kind, called hardening carbon, escapes as hydro- carbon gas when steel is digested with dilute hydro- chloric or sulphuric acid ; the other form of carbon cement carbon, or cementite carbon is left as an insoluble residue when steel is digested in the acids just named. Cement carbon dissolves in dilute nitric acid to form a brown solution, which is utilized in the well-known Eggertz's colour test for combined carbon. 1 1 H. Juptncr von Jonstorff, Journ. Iron and Steel lust., 51. i. 248, FIG. 23. Graphite in Grey Fig Iron. (E. Heyn.) [To face p. 30. 18 CONSTITUENTS OF IRON AND STEEL 31 It is interesting to notice that the presence of chromium in steel seems to stimulate the formation of hardening carbon, and to prevent the latter passing into cement carbon. The special property of chrome steels is extreme hardness. Chromium appears not to confer hardness upon steels in the absence of carbon. We do not know in what form hardening carbon exists in steel. It may be simply a solid solution of carbon in iron. What is a solid solution ? 18. Compounds, Mixtures, and Solutions There are two special features about chemical com- pounds which it is well to notice. The elements which make up a chemical compound are so completely merged one in the other as to form a new substance having properties quite distinct from any of its com- ponents. Neither by the microscope nor by any other known means is it possible to detect the components of the compound so long as the compound retains its individuality. In copper oxide, for example, the com- ponents copper and oxygen are absolutely indis- tinguishable. The compound is quite homogeneous. The second feature is that the elements are combined together in certain fixed and definite proportions. In black copper oxide there are 63*5 parts of copper for every 16 parts of oxygen; in red copper oxide, 127 parts of copper are united with 16 parts of oxygen. Copper and oxygen unite in no other proportions. On the other hand, in glass, the lime, soda, and silica are united together so as to form a perfectly 1897 ; Metallographist, 1. 154, 1898 ; Report with A. A. Blair, G. Dillucr, and J. E. SStead, Joitrn, Iron and Steel Inst., 66. ii. 221, 1904. 32 CRYSTALLIZATION OF IRON AND STEEL 18 homogeneous mass. The constituents are merged to- gether as if the glass were a true chemical compound. But glass does not satisfy the second criterion. The constituents can be mixed in many different propor- tions. The amounts of silica or lime can be varied between wide limits and still produce glass. Glass may be called a solid solution in contradistinction to ordinary liquid solutions, like whiskey and water, or salt and water. Solutions satisfy the ^first but not the second mark of chemical compounds. The components of an ordinary mixture are not so completely merged as in chemical compounds. The constituents of a mixture can generally be separated by mechanical means, and they may be mixed together in any proportions we choose. In a eutectic mixture, however, the constituents are mechanically mixed to- gether in fixed and constant proportions. To summarize, a given substance may be TT.. /Indefinite proportions ...... Ordinary mixture. 3rogcn> \Definite proportions ...... Eutectic mixture. /One component ........ Element. The terms " isomorphous mixture " and " mixed crystals " are not to be used in place of " solid solu- tion." Isomorphous salts are those which furnish crystals having the same shape or form ; isomorphous substances will frequently crystallize together to form single crystals, called mixed crystals. A solidified solution, although it be a solid, may yet retain the essential characteristics of a liquid solution. 1 Of course, a solidified solution is not 1 V. Rothmund, Zeit. phys. Chem.,2G. 433, 1897. 19 CONSTITUENTS OF IRON AND STEEL 33 necessarily a solid solution. As a general rule a solvent can dissolve more of any constituent the higher the temperature. At any given temperature there is usually a limit to the amount the solvent can dissolve. If a solution is saturated at any given temperature, then, on cooling, the solution will reject all in excess of the maximum amount it can dissolve at the lower temperature. The solution theory of carbon-iron alloys affirms that 1. Molten carburized iron is a solution of carbon in iron. 2. The solidified mass is a solid solution. 3. The molten and solid solutions obey the laws of ordinary fluid solutions. 1 19. The Solidification of Molten Iron Let us examine in more detail a freezing solution of carbon and iron. The freezing curves are shown in Fig. 24. If molten iron containing less than 4J per cent, of carbon be cooled, a solid solution of carbon in iron begins to separate along the line AP' (Fig. 24). This solid solution of carbon in iron is called martensite, in honour of the German metallurgist, A. Martens. 2 There is a complication. The solid solution of carbon in iron which separates is not a definite chemical compound. The iron will not retain, in solid solution, more than 2 per cent, of carbon, whereas the molten mass may have as much as 4J per cent, of carbon. We have a new curve (AQ, Fig. 24). 1 A. S. Stansfield, MetallograpJiist, 3. 24, 300, 1900 ; Journ. Iron and Steel In**., 56. ii. 169, 1899; 58. ii. 317, 1900. 2 W. C. Roberts- Austen, Metallographist, 2. 186, 1899. D 34 CRYSTALLIZATION OF IRON AND STEEL 19 The abscissa, at any point on the curve AQ, represents the compositions of the solid, which separates when the solution has the composition represented by the abscissae of the curve AP corresponding with the same ordiuate. For example, the solid which separates at the temperature OT has the composition TS, while the composition of the freezing liquid is represented by TE. The eutectic mixture which separates at P' con- 2000 1600 1000 %CARBOM FIG. 24. Fusibility Curves of Iron-Carbon Alloys. sists of graphite associated with martensite having the composition represented by MQ. As the iron cools down from the eutectic temperature OM, the solid martensite rejects more carbon. The curve QN is the solubility curve of carbon in iron. When the tem- perature falls to 1000, the solid solution, martensite, only contains ON that is, 1'8 per cent, of carbon. The carbon which separates out from the solid metal at temperatures below the point of solidification is usually in the form of a very fine powder, and it is called temper carbon, temper graphite, or anneal- ing carbon. Graphite, you will remember, is an allotropic modification of carbon. 19 CONSTITUENTS OF IRON AND STEEL 35 Eoozeboom 1 has collected the results depicted in diagrams Figs. 18, 22, and 24 into one diagram, shown, slightly modified, in Fig. 25. Here P is the 1500 1400 1300 1300 1100 1000 900 800 700 $| .^ L n in rr \ ^ X, Mo lien Ire n/ \ V ^z /W2 ^ X j forte) s isile \ ^x V, B em * \ 9 ^^ x^ c y-fe rrite JU vten* tie a ^f^ l^p/w fe 4 /H \ 1 f Mart msib ant J ^W7 ,*/*# e -sj\ 7 600' snn a April le %arl It a. td ( Tenter ft'^ \CtiK Ffm te FIG. 25. Diagrammatic. (After H. W. B. Roozeboom.) eutectic point of Fig. 18, and P ; of Fig. 24. Fig. 25 may be taken provisionally to represent the relation 1 H. W. B. Roozeboom, Zeit. Phys. Chem., 34. 437, 1900 ; improved in Zeit. EleUrochem., 10. 489, 1904; Metallographwt, 3. 293, 1900; H. le Chatelier, ibid., 3, 290, 1900 ; 4. 161, 1901 ; F. Osmond, 4. 150, 1901 ; H. Juptner von Jonstorff, ibid*, 5. 210, 1902. r \ 36 CRYSTALLIZATION OF IRON AND STEEL 19 between the various constituents and the temperature, when the system is in stable equilibrium. 1 Owing to the lack of suitable measurements, some of the curves are only approximately known. After what has been said, I think one example will be sufficient to illustrate the use of the diagram. Take a steel with 014 per cent, of carbon. Solidification commences at a and finishes at 6 ; from 6 to c we have a solid* solution of carbon in y-iron ; at c ferrite begins to segregate from the solid solution in the form of /3-iron ; at d the iron passes into the a-form, while the solid solution still segregates a-iron, and at the same time becomes richer in carbon. The separation of a-ferrite continues until the point e is reached. The system then contains a-iron or a-ferrite, and a solid solution of carbon in iron, or hardenite, with 0'89 per cent, of carbon. The hardenite at e passes directly into pearlite. Let us now follow the changes which take place as a molten solution of iron with 4 per cent, of carbon cools down to ordinary temperatures, forming cast iron with 4 per cent, of carbon. Region I. In region J. we have a mixture of molten solution and solid martensite with 2 per cent, of carbon. Region II. This solidifies into martensite with 2 per cent, of carbon, and graphite on passing into region II. There is no cementite formed in this region, because cementite decomposes into iron and graphite at about 1000. It is interesting to notice here that the presence of sulphur or manganese retards, while that of silicon 1 H, C. H. Carpenter and B. F. E. Keeling have examined this experimentally (Journ, Iron and Steel Inst., 67, i, 224, 1904), 19 CONSTITUENTS OF IRON AND STEEL 3? stimulates, the separation of graphite. 1 The silicon might act as indicated in the chemical equation 2 Fe 3 C + Si = Fe 3 Si + C(graphite) An iron relatively free from silicon, and rich in sulphur or manganese, will have little free graphite; while iron rich in silicon and comparatively poor in sulphur and manganese will be suffused with free graphite. Hence the manufacturer can play silicon and sulphur against one another in order to preserve a uniform percentage of graphite. If the amount of sulphur is great, the amount of silicon should be augmented, and vice versa. Region III. As the temperature cools down into region JZT., the graphite which was produced in regions /. and II. combines with iron to form cementite, and at the same time the martensite is resolved into cementite and martensite with 0*9 per cent, of carbon, or Arnold's iron subcarbide (Fe 24 C). Region IV. As the temperature cools down into region IV. the iron subcarbide decomposes into pearlite. The cementite in the final product thus comes both from the decomposition of the martensite and the carbon which is rejected as the molten solution of carbon in iron cools in region 7. But the phenomenon of surfusion plays a part. Passive resistance may interfere with the above sequence of changes, and we may get grey, ultra grey, white, or malleable cast iron from the same solution. i. Grey Cast Iron. Here the sequence of changes 1 G. Charpy and L. Grenet, Metallographist, 5. 202, 1902 ; W. J. Keep, Cast Iron, New York, 1902. 2 W. Campbell, Journ. Iron and Steel Intt., 59. i. 211, 1901. 38 CRYSTALLIZATION OF IRON AND STEEL 19 is interrupted in region ///. The transformation of the iron and graphite of region //. into cementite is suppressed. The consequence is that the 2 per cent, of graphite which was present in region //. is still found in the final product. The cementite of grey cast iron is therefore wholly derived from the decom- position of martensite with 2 per cent, of carbon. ii. "Ultra Grey Cast Iron. Here, owing to the presence of foreign substances, like silicon, the sepa- ration of graphite is greatly stimulated ; so much so that the martensite passes directly into ferrite and graphite. The final result is a mixture of ferrite and graphite. iii. White Cast Iron. The formation of the eutectic mixture, martensite and graphite, in region /. is sup- pressed. We have instead a supersaturated solution containing 4 per cent, of carbon dissolved in iron. The martensite so formed passes unchanged into region //. ; but it is decomposed into cementite and ferrite on passing into region ///. The final product is, therefore, a mixture of pearlite with cementite in excess. There is practically no free graphite. iv. Malleable Cast Iron. In malleable castings, white cast iron is heated in a bed of iron oxide, " mill cinder," for five or six days in region //. (Fig. 25), so that the graphite which would have separated if the castings had been slowly cooled has time to segregate. The iron oxide removes the graphite near the surface of the casting, thus Fe 3 4 + C = 3FeO + CO. Part of the carbon from within diffuses outwards, and is removed as before. The net result is a casting of grey cast iron with about O'l per cent, of combined carbon. 20 CONSTITUENTS OF IRON AND STEEL 39 The remaining carbon exists in the form of fine particles of free graphite Ledebur's temper carbon, in fact. The particles are so small that there is little danger of weakening the casting, as must occur when coarse particles separate, in the ordinary process of manufacture of cast iron. The advantage of malleable castings is that the cost of conversion into steel is avoided, and they are neither so weak nor so brittle as white or grey cast iron. When the cooling metal is suddenly quenched, various transitional forms are produced, which have received special names hardenite, martensite, sorbite, troostite, and austenite. Metallurgists are by no means all agreed as to the identification of the different forms ; some are denied the right to the name " constituent." 20. Martensite, Hardenite, and Austenite We have already met with martensite as the solid solution of carbon in iron which separates during the solidification of molten iron. It may contain as much as 2 per cent, of carbon above 1130; but, as the solution cools, cementite gradually separates out, and the remaining solid solution of carbon becomes poorer and poorer in carbon. The separation of cementite continues until the solid solution has O9 per cent, of carbon. This is the eutectic mixture which segregates into pearlite. Some reserve the term martensite for the unsegregated eutectic mixture containing O9 per cent, of carbon. 1 1 J. O. Arnold and A. M'WiUiam, Journ. Iron and Steel Inst., 55. i. 85, 1899; MetaUographist, 2. 278, 1899; A. Sauveur, ibid., 2. 305, 1899. 40 CRYSTALLIZATION OF IRON AND STEEL 20 Osmond gives the following relation between the amounts of ferrite and martensite in a steel containing O14 per cent, of carbon, and 0'19 per cent, of man- ganese when quenched at different temperatures. Quenched at C. Martensite. Per cent. Ferrite. Per cent. 1340 90 10 1000 61 39 820 46 54 770 24 76 670 14 86 Is the unsegregated eutectic a chemical compound ? Arnold thinks that it is a definite compound, and calls it iron subcarbide (Fe 24 C), or hardenite. Others think that we are dealing with a compound of cemen- tite with iron in the form of " iron of crystallization " (Fe 3 C . 21Fe), analogous to the water of crystallization in crystallized sodium carbonate (Na 2 C0 3 . 10H 2 0), or the " alcohol of crystallization " in Ca01 2 . 4CH 3 OH. It is supposed that, below the Ari critical point, the martensite with 0'9 per cent, of carbon dissociates into ferrite and cementite. The term martensite thus in- cludes both the unsegregated eutectic containing 0*9 per cent, of carbon (Arnold) and the solid solution of hardening carbon in iron (Osmond). Martensite is best formed when a steel containing O2 to 0'8 per cent, of carbon is cooled from above the Ar 3 critical point (830) slowly to the Ar 2 point (730), and then suddenly quenched in a freezing mixture at - 20. Martensite has the appearance of interlacing needles, shown in Fig. 26, which is a specimen of soft rail steel. Fig. 27 is a specimen of hardened tool FIG. 26. Martensite. (E. Heyn.) FIG. 27. Hardenite. (E. Heyn.) [To face p. 40. 20 CONSTITUENTS OF IRON AND STEEL 41 steel. It is frequently very difficult, if not impossible, to detect the crystalline structure of martensite in properly hardened steel. When the steel has just the eutectic proportion of carbon, namely, 0'9 per cent., the martensite is called hardenite, and, by Arnold, iron subcarbide (Fe^C) ; while if there is a greater amount of carbon than this, say 1 to 2 per cent., the result, on quenching, is a mixture of hardenite with a softer constituent, called, by Osmond, austenite, in honour of W. C. Koberts- Austen. Austenite is said to be so soft that it can be scratched with a knitting needle. Some deny the existence of austenite altogether. Fig. 28 shows a specimen of steel containing 1*57 per cent, of carbon, in which barbed plates of hardenite are embedded in a matrix of the alleged austenite. NOTE. To trace the structured changes in the composition of an alloy containing 4'5 per cent, of carbon as it cools down from above 1130. (Tiemann. 1 ) Above 1130 the carbon in excess of 4'3 per cent, separates as graphite. There is 95-5 per cent, of pure iron, and if this has 4'3 per cent, of carbon it will have 4-3 x 95'5 -f- 100 = 41 per cent. /. 4-5 - 41 = 0-4 per cent, graphite separates at 1130. At 1130, the eutectic of graphite with martensite containing 2 per cent, of carbon separates. There is 95*5 per cent, of iron, and this has 2 per cent, of carbon ; and 2 x 95 '5 * 100 = 1-9, /. 41 - 1-9 = 2-2 per cent, graphite separates at 1130. Between 1130 and 1000 the percentage of carbon in the martensite diminishes from 2 to T8 per cent. Hence, since 1-8 X 95'5 * 100 = 1-7, we have 1-9 - 1*7 = 0'2 per cent, of carbon. /. Total graphite = 0-2 + 2-2 + 0-4 = 2'8 per cent. Martensite = 100 - 2-8 = 97'2 per cent. 1 H. P. Tiemann, Metattographist, 4. 313, 1901. 42 CRYSTALLIZA TION OF IRON AND STEEL 20 Below 1000 martensite combines with graphite to form cemen- tite : Martensite (with T8 per cent, of carbon) + graphite = Fe 3 C. Cementite has 6*7 per cent, of carbon, and, therefore, to convert all the ferrite to cementite, we require 67 X 95*5 -j- 100 = 6'4 per cent, of carbon. This would make .a total of over 100 per cent. The martensite already has 1*8 per cent, of carbon, and hence, 6'4 1*8 = 4'6 per cent, more carbon is needed to convert the 97'2 per cent, of martensite into cementite; but only 2 -8 per cent, of graphite are available, hence the 2'8 per cent, of graphite will form 2'8 X 97'2 -j- 4-6 = 59-2 per cent, of cementite; and 100 - 59*2 = 40'8 per cent, of martensite will remain. Below 690 the martensite forms pearlite and cementite. Pearlite has 0'9 per cent, of carbon, and cementite 6'7 per cent. Let x denote the percentage of pearlite, and y that of cementite, .". x + y = 40-8 ; 0'9 + G'ly = 1-8 X 40'8 ; hence, x = 34'5, y = 5-3. Hence, the total cementile is 59'2 + 6*3 = 65-5 per cent. The percentage amounts of cementite and pearlite can also be calculated by the rule given on p. 29. 21. Sorbite and Troostite If the temperature at which the pearlite segregates be hastened by quenching the hot metal in lead, the pearlite loses its well-defined lamellar appearance, and we get what is called sorbitic pearlite, or sorbite, after the pioneer worker on the microstructure of steel, H. 0. Sorby. Sorbite is unsegregated pearlite, a transitional form between martensite and pearlite. A specimen of sorbitic pearlite magnified 1500 diameters is shown in Fig. 29. The specimen contained 1 per cent, of carbon. Some hold that what we have called granular pearlite, Fig. 20, is really sorbite. Whatever sorbite may be, it is not homogeneous. There is another transitional form between martensite and pearlite rather difficult to prepare. It is produced when a metal containing about 045 t 28. Hardenite in Austenite. (F. Osmond.) FIG. 29. Lamellar Sorbite. (F. Osmond.) [To face p. 42. FIG. 30. Troostito. (F. Osmond.) [To face p. 43. 22 CONSTITUENTS OF IRON AND STEEL 43 per cent, of carbon is cooled down to the Ari point (690) and quenched in water at atmospheric tem- perature. 1 This constituent is called troostite, after a celebrated French chemist, L. Troost. A specimen of troostite embedded in martensite is shown magnified about 1000 times in Fig. 30. The steel contained - 45 per cent, of carbon. There are, therefore, two transitional forms between the martensite and pearlite stages of cooling steel. Martensite -> troostite -> sorbite - pearlite. Many metallurgists maintain that troostite and sorbite only refer to particular patterns which the con- stituents of the cooling alloy assume when the metal is quenched under special conditions. I dare say that it would be possible to get an infinite number of gradations between martensite proper and pearlite. All we can say is that two predominating patterns are called troostite and sorbite. To prepare troostite, Le Chatelier recommends the heating of a steel bar containing O9 per cent, of carbon in a furnace, and leaving the other end free, so that the variation of temperature is uniform throughout the whole length of the bar. After quenching, find the points of medium hardness by means of a file, and, on suitably etching them, the untransformed pearlite will be found separated from the completely transformed pearlite by a zone of troostite. 22. The Phase Rule The relation between the pressure and volume of a gas confined in a vessel at any fixed temperature is given by the well-known law of Boyle: pressure X volume is constant. If suitable units are chosen, 1 H. C. Boynton, Journ, Iron and Steel Inst., 67. i. 262, 1904. 44 CRYSTALLIZATION OF IRON AND STEEL 22 we may write p x v = 1. If the pressure be 0*5 units, the volume must be 2 units. If not, the gas will either expand or contract until the product p X v is unity. The equation p x v = 1 is called the con- dition of equilibrium for the gas. There is another relation of a similar nature which enables us to see whether the components of a mixture are those neces- sary for equilibrium. This condition of equilibrium is called Gibbs' phase rule. Before describing the phase rule it will be well to fix special meanings to three terms component, phase, and degree of freedom. The components of a mixture are those entities which are undecomposable under the conditions of the experiment, and which take part in the reaction. The components of a mixture may either be elements, or " undecomposable " compounds which behave in the given system as if they were elements. The com- ponents of an aqueous solution of salt are sodium chloride (NaCl) and water (H 2 0); copper and anti- mony are the components of copper-antimony alloys ; iron and carbon are the components of steel. The components of a system may group themselves in various ways. They may pass from one physical state to another, as when water boils or freezes ; they may combine with one another in various ways, as when antimony and copper form copper antimonide (Cu 2 Sb) ; they may form simple solutions, as when salt dissolves in water. Every homogeneous state solid, liquid, or gaseous ; element, compound, or solution which the components may produce is called a phase. The phases of a system are the physical states in which the components exist. A eutectic is not a phase. If a gas be confined in a vessel, the volume v, 22 CONSTITUENTS OF IRON AND STEEL 45 temperature T, and pressure p, are related by the con- dition of equilibrium (p x v = 2T). If only one of these variables is fixed, say the volume, the state of the system will remain undefined, because the vapour may yet retain the fixed volume, while the temperature and pressure have very different values. Two of the three variables must be known before the state of the system can be defined unequivocally. If, say, both the volume and the temperature are fixed, the remain- ing variable, p, can only assume one definite value. The two fixed variables are said to be arbitrary or independent variables; the third, which can be calculated from the equation of equilibrium when the other two are known, is called a dependent variable. Another term for independent variable is degree of freedom. The above-described system has two degrees of freedom. A system consisting of liquid water and water vapour can be defined by the two variables pressure and temperature. So long as liquid water is present the pressure of the water vapour is determined solely by its temperature. The state of the system is defined by two variables ; one is the dependent vari- able, the other the independent variable. In other words, the system has one degree of freedom. If ice, liquid water, and water vapour all exist together none of the three variables can be changed without destroying one of the phases. In consequence the system has no degrees of freedom. The variability, variance, or degree of freedom of a system is the number of independent variables which must be fixed before the state of the system can be unequivocally defined. The degree of freedom also shows whether a system can survive an arbitrary 46 CRYSTALLIZATION OF IRON AND STEEL 22 variation of any one of these three variables without passing into some other condition. If a system has no degrees of freedom it is said to be invariant ; if it has one degree of freedom, univariant ; if two degrees of freedom, bivariant ; and if the variance of the system be greater than this, multivariant. According to Gibbs' phase rule, a system will be in equilibrium when it has G P + 2 degrees of freedom, where C denotes the number of components, and P the number of phases. If F denotes the number of degrees of freedom of the system, then we may write the phase rule F = C - P + 2 . . . . (1) If, for instance, a system has two components and four phases, it will have 2 4 + 2 = degrees of freedom, and the system will not be able to survive any variation of temperature, pressure, or in the concen- tration of its components. A one-component system cannot be in equilibrium if four phases are present, because the system cannot have less than no degrees of freedom. In the application of the phase rule to alloys it is usual to neglect the vapour pressure, and to consider only two variables, volume or concentration, and temperature. Hence the phase rule assumes the simpler form F = C - P + 1 . . . . (2) We may now apply the phase rule (2) to a number of examples. Molten hypoeutectic steel is divariant; the com- ponents are iron and carbon; the phase is a molten solution. The same steel solidified at 1300 is also 22 CONSTITUENTS OF IRON divariant, with the same components, and the phase martensite. At about 720 the phases are martensite and ferrite ; hence the system is univariant below 720. At 690 the system has the phases martensite, ferrite, and cementite ; the system is invariant. The pearlite present is not a phase. If the same steel be suddenly cooled from above 690 the phases are also martensite, cementite, ferrite ; the system is therefore invariant, neither the temperature nor the concen- tration of the system can be altered without breaking up the system. As a matter of fact, the system is in an unstable condition. The test for the stability of a system is to find whether the system can survive a small change of temperature. Invariant systems can only be in equilibrium stable equilibrium when each variable has one fixed and definite value. There is only one particular temperature, for example, at which the two phases ice and water can exist together in the same system. System. Components. Phases. Degrees of freedom. Freezing water Water Liquid, solid Invariant. Freezing sodium Sod. thiosul- Liquid, solid Invariant. thiosuphate phate Aqueous sol. of salt . Water, salt Solution Bivariant. Copper-silver alloy : Molten Copper, silver ' Solution Bivariant. Solid . Copper, silver '> Two solids : copper Univariant. and silver Freezing eutectic selective Copper, silver ( Two solids, one liq. Copper, silver ( One solid, one liq. Invariant. Univariant. White cast iron Iron, carbon Martensite, cemen- Invariant. tite, graphite Grey cast iron Iron, carbon Ferrite, cementite, Invariant. graphite 48 CRYSTALLIZATION OF IRON AND STEEL 22 Granite is composed of the minerals muscovite (K 2 . 3A1 2 3 . 6Si0 2 ), quartz (8iO a ), and orthoclase (K 2 . A1 2 3 . 6Si0 2 ). The components are silica, alumina, and potash; the phases are the three solid minerals ; hence the system is univariant, and it is in equilibrium. A. Findlay's The Phase Rule and its Applications, London, 1903, is a text-book devoted to this branch of chemistry. EXEBCISES. (1) Find the percentage of pearlite and excess ferrite in medium carbon steels containing 0'3 and 07 per cent, of carbon respectively. An*. 33 and 67 ; and 78 and 22. (2) An alloy containing a high percentage of graphite is required. If the castings, owing to their thinness, or from the nature of the mould used, are expected to cool rapidly, why should the percentage of silicon be increased? Am. To stimulate the formation of graphite, since rapid cooling retards the separation of graphite. (3) Trace the changes which steels containing (K, 0'8, and 1-6 per cent, of carbon undergo as they cool slowly from the molten condition. (4) Find the total ferrite, and total cementite in a low carbon steel containing 0-2 per cent, of carbon. Ans. 97 and 3. THE HARDENING, ANNEALING, AND TEMPERING OF STEEL 23. General Properties of Hypo- and Hyper- eutectic Steels THE mechanical properties of iron-carbon alloys are closely connected with the relative amounts of the two elements. The relation between the percentage of carbon in an alloy and the tenacity in tons per square inch is indicated l in the following table : Per cent, of carbon = 0-05 01 0'2 0'4 0'6 0'8 I'O 1-3 Tenacity . . = 25 "00 26'0 31 "0 36'0 43'0 58'0 60'0 44'0 The gradual increase in the tenacity of the metal as the amount of carbon approaches the eutectic pro- portions is brought out very clearly. The results are shown graphically in Fig. 31. Cementite is a very brittle substance, harder than glass, while ferrite is as soft and as ductile as copper. The relative proportions and the distribution of these two constituents in any alloy must affect its mechanical properties. In hypoeutectic steels the presence of an excess of ferrite renders the metal ductile and tenacious. On the other hand, in hypereutectic 1 H. M. Howe, Eng. and Mining Journ., 241, 1887. 49 E So CRYSTALLIZATION OF IRON AND STEEL 23 steels, the presence of an excess of cementite diminishes the ductility and tenacity of the metal. In a chisel which has to be subjected to blows, there should be no structurally free cementite. Chisel steels, in consequence, contain approximately 0*9 per cent, of carbon. A cutting file which has to withstand DU 50 40 30 ?n 5 sr / N I 7 / / / 04 0-8 1-2% CARBON FIG. 31. Tenacity of Iron-Carbon Alloys. but little shock may have a little free cementite say, 5 per cent, excess cementite or 1/2 per cent, of carbon ; while a razor which has no shock to meet, and wants essentially a keen cutting edge, may have as much as 10 per cent, of excess cementite, or 1/5 per cent, of carbon. The brittleness of such a steel is not a serious objection. When the metal has 2 or more per cent, of carbon, the resulting alloys are called cast irons. 2 White cast iron, for example, has so much cementite that it cannot be filed or drilled with ordinary tools. In white cast iron the carbon is principally in the form of cementite. By heating white cast iron to a suitable 1 H. M. Howe, Metallographist, 4, 177, 1901 ; 6. 203, 1903; Encyc. Brit. 29. 570, 1902. FIG. 32. White Cast Iron. (F. Popplewell.) [To face p. 51 24 HARDENING AND TEMPERING OF STEEL 51 temperature (region //., Fig. 23), the cementite is decomposed into free graphitic carbon. The result is grey cast iron, in which the carbon is not combined with the iron, but is present in the form of graphitic carbon. The effect of free graphite on the properties of the alloy is Only mechanical. It destroys the con- tinuity of the metal, and so renders it liable to fracture under the influence of mechanical stresses. Grey cast iron is soft enough to be filed, drilled, and subjected to other mechanical operations. Figs. 32 and 23 represent the microscopic appearance of white and grey cast irons respectively. The dark spots on Fig. 32 consist of pearlite; the lighter portions are cementite. 24. The Influence of Rate of Cooling The structure and properties of steel and cast iron may be profoundly modified by the rate at which the metal is cooled from a high temperature. Two pieces of steel having exactly the same properties, chemical composition may have entirely different physical properties. The hardness, tenacity, and other proper- ties of the metal depend upon the relative proportions of the allotropic modifications of iron, and of mar- tensite, ferrite, and cementite present in the metal; and these proportions, in turn, depend upon the rate of cooling. There is a particular temperature at which the speed of the transformation of hard martensite into soft pearlite goes on most rapidly ; and if the metal be cooled down past this temperature before the hard martensite has time to pass into soft pearlite, then the passive resistance to further change becomes 52 CRYSTALLIZATION OF IRON AND STEEL 24 so great that any further change is arrested, and the properties of the hard martensite will predominate. On the other hand, if the metal be cooled down past this critical temperature very slowly, the martensite will pass into peaiiite, and the properties of the metal will be altered accordingly. The physical properties of steel thus depend upon (1) Its chemical composition ; and (2) The heat treatment to which it has been subjected. The sole object of hardening, tempering, and annealing steel is to make the metal pass through the various critical temperatures with the proper velocity ; in other words, the different transformations which take place at the critical points are arrested when the constituents are distributed in the proportions neces- sary to confer upon the metal the required degree of hardness. When it is remembered how dependent the rate of cooling of a mass of metal is upon external conditions specific heat, thermal conductivity, etc., of the quenching fluid it is easy to see how so many empirical directions for the tempering of steel for special purposes have crept into metallurgical practice. The belief in the efficacy of special nostrums and solutions persists even at this day. But I think that it would be difficult to match the two following receipts for quenching liquids taken from a book of trade secrets published about 1530. 1 In the first you are directed to boil snails in rain water collected in the first two months of the harvest. Bed-hot iron quenched in this resulting liquid is said to be as hard as steel. The author suggests the alternative recipe 1 W. C. Roberts- Austen, An Introduction to the Study of Metallurgy, 147, London, 1902. 24 HARDENING AND TEMPERING OF STEEL 53 which I give in his own words : " Ye may do the like with the blood of a man of XXX years of age, and of a sanguine complexion, being of a merry nature and pleasant . . . ., distilled in the middst of May." In passing, another point may be noticed. For the uniform tempering of a mass of metal it is neces- sary for the whole mass of metal to undergo the same variations of temperature. This is hardly possible in practice, because the cooling must go on through the outer surface of the metal, but the rate of cooling increases as the square of the surface, while the total quantity of heat to be removed increases as the cube of the mass of the metal. Hence the rate of cooling diminishes as the size of the specimen increases. A specimen may even be so large that real hardening becomes impossible. The mass cannot be cooled fast enough. Although the properties of an alloy are closely connected with its chemical composition, yet there is much useful information to be obtained from a micro- scopic study of the metal which can be obtained in no other way. Knowledge gained by the microscope is not expected to supplant, but rather to supplement the results of the chemical analysis and of the mechanical and physical t tests. It is indeed possible that in the near future specifications will be made out for steel with stated amounts of the above-named con- stituents, sorbite, pearlite, ferrite, etc., when certain specified properties are required. A metal with a large amount of sorbite, for example, is particularly tenacious, and metallurgists have therefore investigated the best heat treatment for retaining a maximum amount of sorbite in the specimen. Hence we have " sorbitic steel rails ; " and the so-called " patented 54 CRYSTALLIZATION OF IRON AND STEEL 24 wire" alloys, which, by a special treatment, 1 have a large percentage amount of sorbite. The valued wear- ing qualities of gun-barrels, tires, and armour-plates are enhanced when the sorbite content is high. As an example, Sauveur and Boynton 2 state that a steel with 0*55 per cent, of carbon, cooled from 1150, has these properties How cooled. Prevailing constituent. Elastic limit. Ibs. per sq. inch. Tenacity. Ibs. per sq. inch. In furnace . In air . . . Pearlite Sorbite 39,901 55,000 81,162 99,979 25. The Allotropic Modifications of Iron Besides all this, the relative proportions of the three allotropic modifications of iron (p. 14) play an important part. Gamma iron is said to be as hard as chilled steel, while alpha iron is soft and ductile. The properties of beta iron are not very well known, but they are supposed to be intermediate between the properties of y- and of a-iron. Let me summarize the probable properties of the three allotropic forms of iron. Allotropic form. Normal limits of stability. Magnetic properties. Hardness. Ductility. Alpha, or a . Beta, or ft . Gamma, or 7 680 - 645 755 - 710 845 - 800 Magnetic Non-magnetic Non-magnetic Soft Hard Hard Ductile. Brittle. Ductile. 1 J. E. Stead, Jonrn. Iron and Steel Imt,, 64. ii. 141, 1903. 2 A. Sauveur and H. 0. Boynton, Metallographist, 6. 148, 1903. 25 HARDENING AND TEMPERING OF STEEL 55 The critical points are very much affected by the presence of foreign substances. The influence of carbon is shown roughly in Fig. 33. The three critical points, Ar x , Ar 2 , Ar 3 , gradually converge into one critical point, at about 690, as the percentage of carbon increases. Tims, Tim* Mild Steel Medium Steel Pure Iron Mild Steel Medium Steel Hard Steel FIG. 33.The Influence of Carbon on the Critical Points. The effect of adding carbon is to increase the stability of hard y-iron. Pure y-iron is very unstable at ordinary temperatures, and it passes rapidly back to normal a-iron at ordinary temperatures. 1 But if a little carbon be present the rate of transformation is reduced ; and if the y- or a-iron be suddenly cooled the transformation of the hard y-iron to the soft a-iron goes on very slowly. The rate of transformation is then fastest at 690, and gradually slows down to zero as the metal cools down to the temperature of the surrounding air. Still more remarkable effects are produced by 1 H. M. Howe, Mineral Industry, 8. 380, 1900; A. Sauveur, Metallographist, 1. 27, 1898 ; H. le Chfttelier, ibid., 1. 52, 1898 ; T. M. Lowry, Technics, 1. 450, 1904. 56 CRYSTALLIZATION OF IRON AND STEEL 25 alloying the metal with manganese, nickel, chromium, and tungsten. These elements may lower the Ari point from 690 down to atmospheric temperatures or less. In air- or self-hardening steels discovered by E. Mushet the transition temperature is brought down to ordinary atmospheric temperatures by alloying the metal with carbon, tungsten, and manganese. Again, the presence of 35 per cent, of nickel brings down the transition temperature of y-iron to a-iron as low as C. the freezing-point of water. A steel of this kind maintains its cutting edge at a much higher temperature than ordinary carbon steel. In the high- speed or rapid-cutting steels 1 these qualities are much enhanced by variations in the composition of the alloy and suitable heat treatment. There are about fifty brands on the market, and all are com- pounds of iron and carbon with (i.) tungsten and chromium ; or (ii.) tungsten, chromium, and molybde- num ; or (iii.) molybdenum and chromium. In the non-expansive alloys, or "invar" alloys of iron, containing about 36 per cent, of nickel, dis- covered by A. E. Guillaume, 2 the transformation of y-iron to a-iron is just beginning at ordinary tem- peratures. Now a-iron occupies a greater volume than 1 J. M. Gledhill, Technics, 1. 591, 1904 ; 2. 17, 1904 ; J. T. Nichol- son, ibid., 1. 84, 1904. 2 E. H. Saniter, Metallograpliist, 1. 251, 1898 ; G. Gore, Proc. Roy. Soc., 17. 260, 1869; W. F. Barrett, Phil. Mag. [4], 46. 472, 1873; F. Osmond, Compt. Rend., 128. 304, 1899; 118. 532, 1894; Hopkinson, Proc. Roy. Soc., 48. 1, 442, 1890 ; H. le Chatelier, Compt. Rend., 110. 283, 1890 ; 111. 454, 1890 ; A. E. Guillaume, ibid,, 124. 176, 1515, 1897 ; 125. 235, 1897 ; 126. 738, 1898 ; Metallographist, 6. 162, 1903 ; Nature, 71. 134, 1904; Les Applications des Aciers au Nickel avec un Appendice sur la Thtforie des Aciers au Nickel,. Paris, 1904; E. Dumont, Compt. Rend., 126. 741, 1898; E. A. Hadfield, Journ. Iron and Steel Inst., 64. ii., 14, 1903. 25 HARDENING AND TEMPERING OF STEEL 57 y-iron. The transformation from one allotropic form to the other is spread over a certain range of tem- perature. The normal expansion or contraction of iron, as the temperature rises or falls, is just counter- balanced by the decrease or increase in volume as iron passes from the a- to the y- condition, or in the reverse direction. Hence it is not necessary to com- pensate for the effects of temperature in clock pendulums and in geodesic measuring instruments. A similar state of things obtains with the elastic properties of nickel steels, and we are promised nickel steel hair springs for watches, which render it unneces- sary to employ compensated balance-wheels. If ordinary steel be melted with 8 per cent, of manganese and 2*5 per cent, of nickel, the result is a non-magnetic steel. 1 There is no break in the cooling curve. The explanation is that the non- magnetic y-iron has not passed through its critical points at the ordinary temperatures. A steel with 25 per cent, of nickel does not begin to show mag- netization at C., but if cooled in liquid air the transformation of y-iron to a-iron takes place very quickly, and the iron becomes magnetic. Magnetization is not a property peculiar to the atoms of the so-called magnetic elements; rather does it seem to be due to a peculiarity in the arrangement of the atoms in the molecule, or of the molecules themselves. If it were possible to make the molecules or atoms of the other elements assume the peculiar arrangement which is the cause of their magnetic properties, then we might make copper, brass, etc., magnetic. This has been done by Heusler. 2 Alloys 1 L. Dumas, Compt. Rend., 129. 42, 1899 ; Metallographist, 3. 48, 1900: F. Osmond, ibid., 1. 266, 1898; 2. 136, 1899; W. F. Barrett and W. Brown, Technics, 1. 123, 1904. 2 F. Heusler, Ueber die ferromagnetischen Eigenschaften von Legier- vngen unmagnetischer Metcdle : Marburg, 1904. $8 CRYSTALLIZATION OF IRON AND STEEL 25 containing the non-magnetic elements, copper, 60 per cent.; manganese, 26 per cent. ; aluminium, 14 per cent., are magnetic the so-called Heusler's magnetic alloys. What is more, if the magnetic iron be com- bined with Heusler's alloy, the alloy loses its magnetic properties ! 26. Annealing, Tempering, and Hardening of Steel But let us return to the iron- carbon alloys. When steel is suddenly cooled from above the Ar 3 to the Ari it becomes very hard. A lot depends upon the per- centage of carbon. If 1 per cent, of carbon be present, the steel is nearly as hard and as brittle as glass. Low carbon steel is, however, not much affected. The hardness and brittleness increases with the rapidity of cooling. For example, the cooling of a mild steel containing Ol per cent, of carbon furnished the following numbers : * Cooling agent. Hardness. Soda solution (20) . . Brine (20) .... Cold water (20) . . Wood tar (80) . . . Boiling water (100) . Lead (350) .... 202 156 149 121 118 112 Hardness before quenching = 99. Still further, it does not matter very much from what temperatures the cooling begins provided it is above the critical range, 720. This follows from the subjoined tests, due to Arnold, 2 with steel containing 1 I. A. Brinell, Journ. Iron and Steel Inst., 59. i. 269, 1901 ; H. M. Howe, Iron, Steel and other Alloys, Boston, 225, 1903. 2 J. O. Arnold, Engineering, 64. 49, 1897 ; F. Osmond, Metallo- graphist, 2. 80, 1899. 26 HARDENING AND TEMPERING OF STEEL 59 0'07 per cent, of carbon. The metal was quenched in brine at 5 C. Quenched at Tenacity tons per sq. inch. 918 31 -4 \ 887 32-6 ( 820 29-5 ( 780 29-87 650 25-6 600 23-0 \ 525 22-5 400 22-0 ( 15 21-4; The explanation is based upon the fact that the hard martensite has not time to change into the rela- tively soft pearlite when the steel is suddenly cooled. The rate of change from martensite to pearlite at ordinary temperatures is extremely slow. If steel, quenched at ordinary temperatures, be heated, the changes which were arrested are resumed; and the nearer the steel be heated to 600 the more rapid the change. Steel is annealed or softened when heated for some time at as low a temperature as 230 or 300. Steel is tempered at 230 when great hardness is desired 1 and the metal has no shock to meet, as in the case of razors. If the brittleness of this tempering is objectionable, as in the case of chisels and cutting tools which have to withstand certain shocks, such as blows from a hammer, the steel is tempered at a higher temperature, say 300. But this tempering is at the expense of hardness ; the loss of hardness is the price paid to get rid of brittleness. 1 P. Longmuir, Technics, 1. 333, 1904; 2. 568, 1904; 3. 32, 1905. 60 CRYSTALLIZATION OF IRON AND STEEL 26 The object of annealing is to render the metal tough, soft, or ductile. This is done by diminishing the abnormal qualities the metal acquired through hardening and mechanical working. The metal is annealed by heating it to the Aci critical point, and keeping it at that temperature a sufficient length of time to change the "hardening carbon" to "cement carbon," and the hard into soft allotropic iron. When the effects of hardening have been removed by heating the steel to the Aci point it is quite immaterial whether the steel be slowly or quickly cooled the metal will be annealed. In fine (1) Each critical point is accompanied ly structural changes, which begin and end with it. (2) No change in the structural composition takes place in a range of temperature where there is no critical point. (3) Sudden cooling serves to fix the structure possessed "by steel immediately before cooling. 27. The Law of Mass Action Let A denote the state of true equilibrium of steel below the Aci critical point ; B the apparent state of equilibrium which is prevented by passive resistance from passing into A. Obviously, A is annealed steel the state of true equilibrium below the Aci point ; B is hardened steel the state of true equilibrium at a high temperature above Aci. If steel be slowly cooled, the successive stages in the transformation from B to A are effected without hindrance; but if steel be quickly cooled, the brake of passive resistance is in full activity before the passage from B to A is 27 HARDENING AND TEMPERING OF STEEL 61 completed. Suppose that 10 per cent, of B has passed into A before the brake of passive resistance is suffi- ciently powerful to prevent any further change, then, if the temperature be raised to 230, the brake is relaxed, and, say, 10 per cent, more of B passes into A. The remaining 80 per cent, is prevented from passing into A by the passive resistance. Again, by raising the temperature to 300, the brake will be sufficiently relaxed to allow a little more of B to pass into A ; and generally, the higher the temperature, the less the passive resistance. At Aci the brake appears to be sufficiently relaxed to allow the whole of B to pass into A. This analogy must not be pushed too far. For this reason. If the passage from B into A below the ATI obeys the laws of all chemical changes, then at any fixed temperature the velocity of the transfor- mation of B into A, at any instant, will be proportional to the amount of B remaining to be transformed into A. This is the law of mass action. By "amount" is meant the number of grams of B per unit volume. Let a denote the amount of B originally present, and x the amount which has already been transformed at any given instant, t ; then a x will denote the amount of B yet remaining to be transformed into A. The law of mass action may now be expressed in symbols, the velocity of the reaction V = k(a - x) 9 where k is a numerical constant, whose value depends on the conditions of the experiment, the magnitude of the passive resistance, the temperature, etc. It is easy to see that the velocity of the trans- formation must slacken down as time goes on. To fix 62 CRYSTALLIZATION OF IRON AND STEEL 27 the idea, let k be unity, and the amount of a be 100 grams. At the beginning of the action the velocity will be V = 100 grams per hour ; but when, say, 10 per cent, has been transformed, V will be 90 grams per hour; when 10 per cent, of B remains untransformed, the velocity of the reaction will be at the rate of 10 grams per hour. When we say that " at 230 the passive resistance is relaxed so that 10 per cent, of B passes into A," the meaning is that when 10 per cent, of B has passed into A the rate of transformation of the remainder of B is too slow to affect the temper very materially when the metal is heated for a short time at the given temperature. It is reasonable to suppose that a more or less prolonged exposure at 100 would anneal hardened steel just as effectually as a shorter exposure at a higher tem- perature, and that, if the heating at 100 were con- tinued long enough, the whole of B would pass into A. Those who are familiar with the calculus will see that the Telocity of the reaction should be written When all B has passed into A, we have a = x ; but this can only happen after the elapse of an indefinite length of time. See J. W. Mellor, Cliemical Statics and Dynamics, London, 1904. 28. Theories of Annealing and Hardening The constitution of steel may thus be viewed from two important aspects I. The Allotropic Changes of the Iron itself. The explanation which emphasizes the allotropic changes in the iron is known as the allotropic theory. 1 1 H. M. Howe, Metallographist, 1. 150, 1898. 28 HARDENING AND TEMPERING OF STEEL 63 Osmond l has dealt particularly with this phase of the work. The hardening of suddenly cooled steel is supposed to be due to the presence of hard 7- or /3-iron. This state of things is favoured by the presence of foreign substances, like carbon, nickel, etc. But we do not know if other elements, in the absence of carbon, will effect similar changes in iron. Carbon seems to play an essential part in the action. But other explanations have been suggested. II. The Relations between Iron and Carbon. Ac- cording to the carbon theory, 2 the whole of the facts observed during the hardening of steel can be ex- plained on the assumption that carbon exists in the two states hardening carbon and cement carbon already described. The cause of hardening by sudden cooling is due to the retention of carbon in the hardening state. This view does not explain the critical points in the cooling curve of pure iron, and the accompanying changes in, say, the magnetic properties of the metal. It has also been suggested that the hardening of suddenly cooled steel is due to the presence of hard carbides of y- or /3-iron, which are decomposed at the critical points if the steel be cooled slowly ; but, if cooled quickly, passive resistance sets in before the carbides have time to decompose. This is the so- called carbo-allotropic theory of Howe. 3 1 F. Osmond and J. Werth, Compt. Rend., 100. 450, 1885 ; Annales des Mines, [8], 8. 5, 1885 ; F. Osmond, ibid., [8], 14. 1, 1888 ; F. Osmond, Transformations du Fer et du Carbone dans les Fers, les Aciers et Fontes Blanches : Paris, 1888. 2 A. Ledebur, Journ. Iron and Steel In**., 44. ii. 53, 1893; Stahl und Eisen, 14. 523, 1894; 17. 302, 436, 1897. 3 H. le Chatelier, Metattographist, 1. 52, 1898. 64 CRYSTALLIZATION OF IRON AND STEEL 28 J. 0. Arnold 1 has developed an interesting ex- planation, which is known as the subcarbide theory. The points of this theory are as follows : In eutectic or saturated steel there is only one critical point, Ari, which marks the passage of pearlite into an homo- geneous mass corresponding with the empirical formula Fe24C. There is no evidence to show that this sub- stance changes at higher temperatures, or that carbon separates from combination with the iron and passes into a solid solution of elementary carbon in iron. On dissolving this material in acid, practically the whole of the carbon is evolved as hydrocarbon gas. In hypereutectic or supersaturated steel, with, say, 1*4 per cent, of carbon, the pearlite changes into hardenite at about 700, but the cementite only dissolves in the hardenite above 900. With unsaturated steel containing, say, 0*2 per cent, of carbon at the Ari, 700, " the pearlite areas pass into hardenite ; " at the Ar 2 , 750, point " the hardenite areas dissolve in the beta ferrite ; " at the Ar 3 point " there is a dilation of iron like that of water at 4 C." The terms "beta" and "gamma" are used in the sense of a range of temperature, and not of allotropic modi- fications of iron. Arnold lays no stress on the Ar 2 and the Ar 3 points. Tensile tests made on bars of iron containing 0*2 per cent, of carbon, and quenched in iced brine at temperatures ranging from atmo- spheric up to 1000 in an atmosphere of nitrogen, showed that the tenacity increased from about 500 to 950 proportionally with the quenching temperatures. There was no marked increase in the tenacity at the 1 J. 0. Arnold, /own. Iron and Steel Imt., 45. i. 314, 1894. 28 HARDENING AND TEMPERING OF STEEL 65 Ar 2 or the Ar 3 critical points. The hardening of steel by sudden cooling is supposed to be due to the retention of hard subcarbide. Let us, then, compare the explanation offered by the allotropic solution theory with Arnold's subcarbide theory for the condition of carbon in cooling iron. Temperature. Allotropic theory. Snbcarbide theory. Above Ar 3 . . AtAr 3 . . . . Bet. Ar 3 and Ar 2 At Ar 2 . . . . Bet. Ar, and Ar x AtA^" . . . Below Ar, . . Solid sol. carbon in ?-iron y-iron > /3-iron Solid sol. carbon in /3-iron /3-iron -> a-iron Solid sol. carbon in a-iron 3Fe + C -> Fe 3 C Ferrite and cementite Sol. of hardenite in iron. Maximum density of sol. SoL of hardenite in iron. Segregation of Fe 24 C. Segregated hardenite. Fe 24 C -> Fe 3 C + 21Fe. Ferrite and cementite. I have frequently laid stress upon the fact that we can apply the ordinary laws of liquid solutions to the solidified solutions of carbon in iron. This has led to the use of the term "solid solution," as previously mentioned. We are indebted to Koberts- Austen for developing the subject on this side. I have treated the carbo-allotropic theory from the point of view of the theory of solutions, and summarized the results in Fig. 25. It would be an easy matter to reset the diagram so as to summarize Arnold's interpretation of the facts. I do not suppose for one moment that any of these hypotheses is dressed up in its final form. Each one has its weak and its strong points. All are, or ought to be, agreed as to the facts. But metal- lurgists are yet only groping for the true explanation. Many attempts have been made, to calculate the molecular weight of carbon, or of the carbides dissolved in iron. It is shown in Elements of Physical Cliemistry that the lowering of the freezing-point of any P 66 CRYSTALLIZATION OF IRON AND STEEL 28 solution below that of the pure solvent is directly related to the molecular weight of the dissolved substance by formulae resembling weight of substance Molecular weight = constant X d ep re88 L of freezing-point On this view, for instance, it is possible to decide whether carbon disulphide dissolved in methyl alcohol is present as OS 2 , C 2 S 4 , C 3 S 6 , . . . Chemical analysis only shows! that the relative proportions of C ! S = 12 : 64. So it has been sought whether the carbon is dissolved as C, C 2 , C 3) . . ., and whether the carbide is dissolved as Fe 3 C, Fe 6 C 2 , Fe 9 3 , ... in solutions of carbon in iron. The results cannot be accepted without reservations. 1 1 H. Juptner von Jonstorff, Journ. Iron and Steel Tnst., 55. i. 204, 1899 ; 57. i. 219, 1900. FIG. 34. Crystal of Ferrite. (D. Tschernoff.) [To face p. 67. THE CRYSTALLIZATION OF IRON AND STEEL 29. The Crystallization of Iron WE can all give a more or less crude guess of what is meant by a crystal. The word is generally associated with a definite and regular external form which has been produced spontaneously, and not artificially cut. 1 The octahedral and needle-shaped crystals shown in Figs. 1 and 2 are examples. Then again, in Fig. 34, we have the beautiful crystal of ferrite photographed from a crystal 15 cm. (or 15 inches) long, which Tschernoff found in a cavity of a cast steel ingot. But it is not always so easy to decide whether a sub- stance is, or is not, built up of crystals. In Fig. 35, for example, you see the microscopic appearance of a section from a bar of pure Swedish iron. It is gene- rally supposed that the lines in the diagram represent the surfaces of contact of one crystal with another, and that the crystals have not had sufficient space in which 1 J. E. Stead, Journ. Iron and Steel lnst. t 53. i. 145, 1898 ; Metallo- graphist, 1. 289, 1898 ; F. Osmond, Annales des Mines, [9], 17. 110, 1900 ; Metalhgraphist, 3. 181, 275, 1900 ; with G. Chartaud, ibid., 4. 119, 236, 1901 ; Annales des Mines, [9], 18. 113, 1900 ; D. Tschernoff, Metallographist, 2. 74, 1899. 67 68 CRYSTALLIZATION OF IRON AND STEEL 29 to develop their regular crystalline form. The result is a compact mass of irregular-shaped crystals, called crystalline grains, or simply grains. The metal appears as if it had been built up in the form of a mosaic with irregular-shaped stones. We know that iron is a crystalline substance. Pure iron has been prepared in the form of cubic crystals represented in Fig. 36. But in studying the structure of alloys, the crystals are usually so ill-defined and imperfect that it is impossible to decide from their external shape whether they be true crystals or simply amorphous grains. But we can leave this question with those more particularly interested. 30. The Development of Crystalline Grains The junctions of the crystalline grains of pure iron, shown in Fig. 35, and of pure copper, shown in Fig. 37, are typical of pure metals ; but when impurities are present the crystals of the pure metal, in the act of crystallizing, reject the impurities which collect at the crystal boundaries. The particles of pure metal slowly migrate and coalesce together, so as to form little islands surrounded by the impurity. Accordingly, in the solidified mass we find the crystals of pure metal enveloped by a film of the metal associated with the foreign substance. This investing membrane separates the crystals of pure metal one from the other. Ob- viously the mechanical and physical properties of the alloy tenacity, ductility, elasticity, electrical con- ductivity will depend upon the character of the film. The mass of pure metal, for example, may be quite ductile like gold, while the mass of metal with the FIG. 35. Pure Swedish Iron. FIG. 36. Cubic Crystals of Iron. (J. E. Stead.) [To face p. 68. FIG. 37. Pure Copper. (After J. 0. Arnold.) FIG. 38. Copper-bismuth Alloy. (After J. O. Arnold.) [ To face p. 69. 30 CRYSTALLIZATION OF IRON AND STEEL 69 impurity may be quite brittle, as Arnold l found to be the case with an alloy of gold with 0*2 per cent, of bismuth; and copper containing 0*5 per cent, of bismuth. A diagrammatic representation of the latter alloy is shown in Fig. 38. The light bands surround- ing the grey patches of pure copper consist of an alloy of bismuth and copper, and there is a distinct line of nearly pure bismuth between each band. The in- dividual crystals, when separated from the investing membrane of bismuth and gold, were quite as ductile as gold, while the metal as a whole was as brittle as glass. Similarly with hypereutectic steels containing an excess of cementite above that required for the for- mation of pearlite. The pearlite behaves like a pure metal, and rejects the excess of cementite to the boundaries, so as to form a network of cementite in a groundwork of pearlite, as shown in Figs. 39 and 40. The former contains 1*29 per cent, of carbon, the latter 1*8 per cent. In Fig. 40 the enveloping walls of cementite are much thicker than in Fig. 39, as you would expect from the percentage composition of carbon. The reason the network of cementite is light in Fig. 39 and dark in Fig. 40 is due to the fact that the illumination is direct in the former case, oblique in the latter. The whiteness or blackness of the lines depends upon the mode of illumination when the photograph was made. Sometimes the heavier constituents of the cooling alloy settle at the bottom, and the lighter near the top. Ice, for example, rises to the 1 J. O. Arnold and J. Jefferson, Engineering, 61. 177, 1896 ; T. Andrews, ibid., 66. 411, 541, 733, 1898 ; Metattographist, 2. 105, 1899; W. G. McMillan and B. H. Housman, Nature, 54. 171, 1896 ; P. Osmond and W. C. Roberts-Austen, Phil. Tram., 187. 423, 1896. 70 CRYSTALLIZATION OF IRON AND STEEL 30 surface of freezing water, and graphite collects in the form of beautiful plates known as " kish " at the surface of cooling ultra-grey cast iron. 31. Grain Size and Fracture The fracture of a metal, or the broken surface which the metal presents, may be fibrous or crystalline. Experts can deduce a good bit of information from the appearance presented by, say, a fractured pig of cast iron. Each variety of steel has its own peculiar fracture. Wrought iron has a fibrous fracture, while high carbon steel has a characteristic porcelain-like fracture. ^ The same piece of metal, however, may be broken so as to present very different fractures. Wrought iron, for example, if nicked on one side and gradually bent, gives a well-defined fibrous fracture ; while, if nicked on all four sides and suddenly broken, the fracture will be crystalline. But these are well-known "tricks of the trade." The fracture, when performed under definite con- ditions, furnishes a true indication of the coarseness of the crystallization. The degree of coarseness of the fracture, or the average size of the crystalline grains when a suitably prepared specimen is examined under the microscope, is called the grain size of the specimen. As a general rule, the smaller the grain size the better the steel, and we naturally ask : Is there any relation between the grain size and the highest tem- perature to which the steel has been heated in annealing ? For practical purposes the physical properties of steel may be taken to depend upon FIG. 39. Iron with 1-3 per cent. Carbon. (F. Popplewoll.) FIG. 40. Iron with 1-8 per cent. Carbon. (F. Popplewell.) [To face p. 70. FIG. 41. Grain Size acquired at 900. (The Iron Age.) FIG. 42. Grain Size acquired at 1200. (The Iron Age.) [_To face p. 71. 3i CRYSTALLIZA TION OF IRON AND STEEL ^\ (1) Chemical composition, i.e., the relative amounts of other elements present ; (2) Distribution of constituents, i.e., the relative proportions of ferrite, cementite, etc., present ; (3) Size of grains. Here are a few tests, by Sauveur, on the relation between the size of the grain and the physical pro- perties of the same piece of steel : l Size of grain in 0*0001 sq. mm. Tensile strength kilogrm. per sq. mm. Elongation per cent, of length. Redaction per cent. 148 118 62 69-6 70-3 77-7 15-0 19-0 22-5 20 22 35 The relation between the average area and the tenacity is Tenacity = 75'5 - 0-004 A. An experienced man can generally give a sur- prisingly accurate guess of the temperature to which the steel has been heated, from the fracture, or from the microstructure of the metal. Tschernoff and Brinell have observed that the higher the temperature of annealing the larger the size of the grain. This will be evident on examination of Campbell's diagrams, Figs. 41 to 43, which represent the appearance of soft steel when heated to the temperatures indicated : Fig. 41 to about 900; Fig. 42 to 1200; and Fig. 43 to 1 N. Ljamin, Chem. Zeit. y 21. 205, 1899 ; Baumaterialien, 3. 105, 1899, finds the tenacity in different steels varies directly as the size of the pearlite grains at the same finishing temperature ; H. Jiiptner von Jonstorff, Metcdlographist, 2. 222, 1899 ; Stahl und Eisen, 19. 237, 278, 1899 ; F. Osmond, AnnaUs des Mines, [9], 8. 153, 1900. 72 CRYSTALLIZATION OF IRON AND STEEL 31 about 1300. A similar result has been observed during the crystallization of brass. 1 The following are the more important generalizations which have been made : I. The higher the temperature above the Aci point from which steel cools the larger the size of the grains, while if the Ac point is the highest temperature attained, then the steel will have the finest possible structure which it can assume. Howe and Sauveur 2 seem to believe that there is a definite relation between the size of the grain and the highest temperature to which the steel has been heated. For a steel containing 11 per cent, of carbon, this relation can be represented by the formula T = 680 + 281,250 A where A denotes the actual area of the grain in square millimetres ; T is the highest temperature reached in the annealing furnace. TschernoiFs experiments do not quite tally with this formula at the more elevated temperatures, say 1400, although at lower tempera- tures the agreement is satisfactory. The more exact law is given by the curve in Fig. 44. Let us illustrate 1 A. H. Coote, Technics, 2. 290, 1904 ; H. M. Howe, Iron, Steel, and other Alloys, Boston, 250, 1904. 2 H. M. Howe and A. Sauveur, Eng. and Mining Journ., 60. 537, 1895 ; A. Sauveur, Trans. Amer. Inst. Mining Eng., 26. 863, 1896 ; Metallographist, 2. 264, 1899; H. Juptner von Jonstorff, Stahl und Eisen, 19. 237, 278, 1899 ; I. A. Brinell, Journ. Iran and Steel Inst., 29. i. 365, 1886 ; Metallographist, 2. 129, 1899 ; H. Fay and S. Badlam, ibid., 4. 31, 1901 ; J. E. Stead, Journ. Iron and Steel Inst., 53. i. 145, 1898; 54. ii. 147, 1898; Metallographist, 1. 289, 1898; 2. 85, 1899; C. H. Kisdale, ibid., 3. 64, 1900 ; Journ. Iron and Steel Inst., 53. i. 220, 1898; 56. ii. 102, 1899; E. J. Ball, ibid., 37. i. 85, 1890; 39. i. 103, 1891 ; J. O. Arnold, Proc. Inst. Civil Eng.,1 123, 1895 ; D. Tschernoff, Proc. Inst. Mech. Eng., 152, 286, 1880. FIG. 43. Grain Size acquired at 1300. (TJie Iron Aye.} [To face p. 72. 31 CRYSTALLIZATION OF IRON AND STEEL 73 its application to the specimen of steel for which the curve Ac : P was determined ; SP, that is OM, repre- sents the size of the grain at the temperature OS. If the size of the grain is smaller than it should be at any given temperature, then the grain will grow in size until it reaches the normal size indicated by the above law. On the other hand, if the grain is larger than that which is characteristic of any particular temperature, the grain will not shrink to its normal size. For example, if the steel has the grain size SP, Fig. 44, at the temperature OR, the size of the grain will remain OM, and will not shrink to ON. The slope of the curve will be different for different steels ; the steel represented in Fig. 44 contained 1-1 per cent, of carbon, and traces of silicon and manganese. S R N M 6 rain, size Fia. 44. The average maximum size of the grain at any temperature is the average size of the grain when the full sectional area is presented. In some places only part, or one corner of the grain is shown. Sauveur l 1 A. Sauveur, Trans. Amer. Inst. Mining Eng., 22. 546, 1893; Journ. Iron and Steel Inst., 56. ii. 195, 1899 ; MetaHographist, 2. 264, 1899 ; K. G. Morse, ibid., 3. 130, 1900 ; Trans. Amer. Tnst. Mining Eng., 29, 1900. 74 CRYSTALLIZATION OF IRON AND STEEL 31 uses the camera lucida planimeter for measuring the size of the grains. Risdale recommends a comparison of photographs made on the same scale. Morse counts the number of grains in a measured area on the photograph, and takes the average of as many different grains as possible. This is reduced to actual size on dividing by the magnification and the number of grains counted. II. If hardened or unhardened steel be heated to the Aci critical point, all previous crystalline structure, how- ever coarse and distorted, is obliterated and replaced by the finest possible structure which the metal can assume. This rearrangement of the size of the grains is called heat refining. The breaking up of the old structure is due to the change of cement into hardening carbon, or of cementite to martensite, and to the diffusion of the carbon after these changes. The transformation of cement to hardening carbon is sudden ; while the passage from hardening to cement carbon takes place gradually, and is accompanied by the evolution of heat. 1 The law only applies to hard steels. With hypo- eutectic steels, the old coarseness is not quite obliterated when the Ac 3 point is reached ; and as the old struc- ture being destroyed between the Aci and the Ac 3 points, a new growth sets in. Consequently, hypo- eutectic steels cannot be refined so completely as eutectic, or hypereutectic steels, because the old structure cannot be effaced without permitting, at the same time, a considerable growth of the new. 1 J. E. Stead, Journ. Iron and Steel Inst, 53. i. 145, 1898 ; 54. ii. 137, 1898 ; MetaUograplitot, 1. 289, 1898 ; I. A. Brinell, ibid., 2. 129, 1899. 31 CRYSTALLIZATION OF IRON AND STEEL 75 Stead l has shown that with very soft steels, con- taining 0'25 per cent, to 012 per cent, of carbon, the ferrite grains grow larger as the temperature rises above 500, and, instead of being refined, the grains continue growing as the temperature passes the Aci (700) point, and continue growing until the Ac 2 point is reached. No material increase in size then occurs until the Ac 3 (900) point is reached, when the former granular structure is broken up, and the steel is refined. This shows how necessary it is to avoid heating soft steels for any length of time at a dull red (about 500). HI. If steel be heated for some time to a temperature just above 1000, the metal, even mild steel, becomes very brittle, and acquires a coarse crystalline structure. The size of the crystals may be considerably reduced by quickly cooling the steel from the temperature of overheating although the brittleness still persists. Steel so affected is said to be overheated. The disease can be cured by heating; the steel for a few days near the Ac 3 (850-900) critical point. 2 There is an aggravated form of overheating pro- duced when the metal is heated to a temperature so near its melting-point that an evolution of gas occurs. This gas forces the crystalline grains apart, and so destroys the continuity of the mass. The gases are probably carbon monoxide, formed by the union of the occluded oxygen with the carbon of the steel, as well as hydrogen and nitrogen. The disease is called burning, and the steel is said to be burnt. Burnt 1 J. E. Stead, Journ. Iron and Steel Inst., 53. i. 289, 1898. 2 E. Heyn, Journ. Iron and Steel Inst., 62. ii. 73, 1902 ; W. Camp- bell, ibid., 64. ii. 359, 1903. 76 CRYSTALLIZATION OF IRON AND STEEL 31 steel is brittle both hot and cold ; it has a coarse crystalline fracture, and a coating of oxide is frequently found on the faces of the crystals at the fractured surface. The cooling curve of pure iron shows critical points at about 400- The effects are due to the hydrogen occluded in the metal. 1 It is interesting to notice that the presence of hydrogen gas diminishes the ductility of the metal, as you can see by dipping two pieces of steel wire in dilute sulphuric acid (one of acid to ten of water by weight) for half an hour, and keeping a third piece for comparison. The pickled wires are quite brittle. One of the pickled wires may be heated to 100 to show that when the hydrogen gas is driven off by heating, the metal regains its ductility. Burning is an incurable disease at least so far as heat treatment is concerned. Heat treatment can only induce allotropic or chemical changes in the alloy ; it cannot close up actual fissures in the metal. Mechanical kneading and compression by hammering and rolling is much more effective in closing up the cracks, especially if the oxidation is not very great. Why do not ingots and castings burn as they pass through the burning range on cooling down from the casting temperature, whereas if a steel bar be heated to this temperature and then cooled quickly or slowly it is incurably burnt ? We have seen that burning is caused by the oxidation of the crystal faces by the inward diffusion of oxygen. When a casting is cooling, the hydrogen dissolved by the metal is working outwards in such a way as to prevent the entry of oxygen; but steel bars made from these ingots, having lost their hydrogen, have nothing to 1 E. Heyn, Eisen und Stahl, 20. 837, 1900 ; Metallographist, 6. 39, 1903 ; L. Romanoff, ibid., 2. 247, 1899 ; Stahl und Eisen, 19. 265, 1899 : for oxygen in steel. 32 CRYSTALLIZATION OF IRON AND STEEL 77 prevent the infiltration of atmospheric oxygen between the faces of the crystals. 32. Influence of Mechanical Work The normal crystallization of steel generally leaves the metal in a state not well adapted for industrial requirements. But this structure may be broken up by the mechanical work of rolling and hammering the metal. The forging may be performed under two conditions : i. Hot ivork when the metal is forged above the critical temperature Aci. ii. Cold work when the forging is performed below the Ari critical point. If the temperature is rising from below the Ari point, then at temperatures between the Ari and the Aci points, the metal is being cold worked ; and hot worked if the temperature is falling from above the Aci point. The following are the more important structural changes which take place during the mechanical work- ing of the metal : I. If the work is sufficiently vigorous to affect all parts of the mass, no crystallization takes place ivhile the steel is being worked. 1 II. Hot work lias no direct action upon the structure of the steel, but, as it retards crystallization till a lower temperature is reached, it may influence structure in this way. Howe 2 explains the relation between the tempe- rature of the hot work and the size of the grain, 1 B. Job, Metallographitt, 5. 177, 1902; S. S. Martin, ibid., 5. 191, 1902 ; A. Sauveur, ibid., 5. 197, 1902. H. M. Howe, Iron, Steel, and other Alloys, Boston, 263, 1903. 78 CRYSTALLIZATION OF IRON AND STEEL 32 somewhat as follows : If the steel is worked at a temperature OB (Fig. 45), the rollers break down the grain ; after that, the crystalline grains grow along the line BC as the metal cools down. If the cooling mass be again rolled, the grain will be again broken up, and the size of the grain on subsequent cooling will be reduced to RE. The grain goes on growing as the metal cools along the line EF, only to be again broken up when the metal is rolled again. This alternation of changes goes on as the steel passes Ac, J Gmuisva FIG. 45. through and through the rollers. The result is the zig-zag BGEF . . . HIJ. If the rolling ceases when the curve HI crosses the normal line OA, the size of the grain in the finished product will be OJ. Hence, the size of the grain will le smaller the lower the finish- ing temperature. The finishing temperature is the last effective rolling or hammering. The object of rolling is to expel slag ; close up the blowholes and other irregularities. The metal will be softer the hotter it is, and less power will then be required to do the work. Hence it it is desirable to work the metal as hot as possible. But we must set against this the facts just brought under notice FIG. 47. (J. E. Stead.) [To face p. 79. VNIYERS/TY 1 or ./ . / . c O H H ^,J> J '"" 33 CRYSTALLIZATION OF IRON AND STEEL 79 the higher the finishing temperature the coarser the grain, and the more brittle the metal. Hence the higher the finishing temperature above the Aci point the poorer the quality of the steel. The practical application will be obvious. Work the metal at as high a temperature as possible, so as to save power during rolling, and then reduce the temperature by, say, a current of cold water, so that the rolling may take place at a low temperature. This is done, for instance, in tire and rail rolling. In welding, too, after the two pieces are actually joined, the smith continues hammering until the temperature has fallen low enough to insure a relatively fine grain in the finished piece. The parts to be welded are first thickened, so as to allow of this " hammer " refining. III. Cold work distorts the grain, or flattens and elongates the crystals in the direction of rolling. The lower the temperature the more pronounced the effects of cold working. The structural changes of cold working can be obliterated by heat treatment up to or over the Aci critical point. The adjoining diagrams (Figs. 46 and 47) show a specimen of rolled steel before and after annealing. Notice the fibrous struc- ture of the sheet (Fig. 46) after rolling at 600, and the crystalline structure of the same sheet after annealing (Fig. 47). The fractured surface of the rolled sheet had a fibrous appearance, and the sheet was not brittle; the second broke with a crystalline fracture after severe punishment. 33. Influence of Other Elements A large number of empirical observations with the testing machine have been made upon the influence of 8o CRYSTALLIZATION OF IRON AND STEEL $33 elements like manganese, sulphur, silica, chromium, nickel, and phosphorus upon the properties of the iron-carbon alloys. A full description of their effects must be sought in the regular text-books. 1 Some elements increase, others diminish the tenacity. I. The tenacity is lowered by elements like silicon, phosphorus, and sulphur, which promote the formation of graphitic carbon or cementite; by elements like phosphorus and copper, which cause the formation of other separations ; by elements which increase the size of the grain, e.g. phosphorus and manganese ; by non- metallic elements which induce the formation of thick cell-walls, or which separate the crystalline grains one from the other ; for instance, gases, large amounts of silicon, compounds of iron and sulphur, and of manga- nese, and sulphur, slag, and oxides. II. The tenacity is increased by elements like nickel and cobalt, which possess a high tenacity ; by small amounts of elements like manganese, chromium, tung- sten, titanium, which hinder the separation of graphitic carbon ; by small amounts of elements like silicon and aluminium, which hinder the formation of blow-holes ; and by elements like nickel and chromium, which favour the formation of firmly locked crystals. The internal changes by which these effects are produced are little understood. The " impurities " or "foreign substances" may not only react upon the iron, but also upon the carbon, and upon each other. Among the secondary reactions we have the possible 1 A. H. Hiorns, Steel and Iron, London, 1903 ; or, F. W. Harbord, The Metallurgy of Steel, London, 1903; A. Carnot and E. Goutal, Annales des Mines, [9], 18. 263, 1900 ; Metallographist, 4. 286, 1901 ; H. Juptner von Jonstorff, ibid., 2. 222, 1899 ; Stahl und Eisen, 20, 939, 1899 ; T. Turner, The Metallurgy of Iron, London, 1904. 33 CRYSTALLIZATION OF IRON AND STEEL 81 formation of manganese carbides Mn 3 C ; Fe 3 C.4Mn 3 C when manganese is present ; with chromium we may have the double carbide 3Fe 3 C . Cr 3 C2 ; if silicon be present along with manganese we may have manganese silicide, MnSi, produced; and with sulphur, manga- nese sulphide, MnS; while phosphorus may produce many varieties of phosphides of iron. 1 Besides the formation of compounds of this type, there is also the possible existence of allotropic forms of the foreign elements nickel, cobalt, say which modify the properties of the alloy in a specific manner. There is thus a vast field of work before the metal- lurgist. What is the effect of the various secondary components upon the physical properties of the final product? Under what conditions are they formed? How can the formation of deleterious secondary products be hindered ? EXERCISES. (1) Why should the finishing temperature of un- hardened steel be at the Ac! critical point ? Ans. The smaller the grain the more ductile and tough the netal, etc. (2) How can the structural deformation brought about by cold work be obliterated ? (3) Why have the central portions of thick pieces larger crystalline grains than the outer portions ? Ans. In practice, thick pieces cannot be finished at a uniform temperature. The central portions of thick pieces leave the rollers at a higher temperature than the outer layers, etc. (4) If chisel steel which has been tempered at 300 be heated to 230, what changes will take place in the temper of the metal? Ans. None, if the brake of passive resistance had relaxed to its limit at 300. But see 27. 1 J. E. Stead, Journ. Iron and Steel Inst., 58. ii. 60, 1900 ; MetaHo- graphist, 4. 89, 199, 332, 1901. THE INFLUENCE OF STRESS AND STRAIN 34. Intercrystalline or Intergranular Weakness THE popular idea of the qualities of steel has crys- tallized into the phrase "as true as steel," but metallurgists have now to face the uncomfortable fact that steel which has safely passed all the mechanical tests occasionally develops an extraordinary tendency to become brittle, and breaks under stresses far below the elastic limit of the metal. The condition of pur- chasing steel rails is that they stand without fracture the blow of a ton weight falling 30 feet. But the director of one of the largest railways in England, the late Sir Lowthian Bell, has said that "occasionally rails for no apparent reason break with a blow from the height of a half or a quarter of that ; and some- times after rails have been in use for a short time they positively break in two with a fall of less than one foot." I am also informed that the United States Government have gone back to wrought iron anchors for their ironclads, on account of the treacherous and mysterious fractures which occur with approved types 82 34 THE INFLUENCE OF STRESS AND STRAIN 83 of cast steel kedge anchors. 1 It is therefore of interest to examine this question in a little more detail. 2 Just as the shaking of a superfused solution of sodium thiosulphate might induce the crystallization of the substance, so is it conceivable that, if the con- stituents of steel were in an unstable state of equili- brium, a sudden shock, or the continued recurrence of a series of vibrations, might so relax the passive resist- ance that the internal structure would be modified in some way, and brittleness result. The disease may not be due to any alteration in the crystalline form of the metal. Steel is normally in a crystalline state. In many cases the source of weakness is the joints between the crystalline grains. When such a metal is fractured the line of fracture follows the junction of the grains. Stead calls this ail- ment intergranular or intercrystalline weakness (inter = between). We have had examples. Arnold's work on the influence of bismuth on copper and on gold. One per cent, of sulphur arranged as a mesh of iron sulphide will entirely destroy the ductility of the iron, reducing the ultimate stress from 20 to 2 tons per square inch. Sulphide of iron is only found in appreciable quantities in iron low in manganese. "Manganese," says Arnold, "seems almost entirely to prevent the formation of cell walls of iron sul- phide." 3 The network of cementite which envelops the 1 W. M. Carr, Metallographirt, 5. 58, 1902. 2 We shall not consider weakness due to the presence of blowholes or slag. See C. H. Kisdale, Journ. Iron and Steel Inst., 64. ii. 232, 1903. 3 J. O. Arnold and G. B. Waterhouse, Metattographist, 6. 302, 1903 ; Journ. Iron and Steel Inst., 63. i. 136, 1903. 84 CRYSTALLIZATION OF IRON AND STEEL 34 crystal grains of steel containing over 1 per cent, of carbon are the principal lines of weakness. The metal, when fractured, generally breaks through the centre of this brittle envelope. The coefficient of contraction of the cementite cell walls is greater than of the cell contents. Pearlite cells, for example, bound together by thick cementite walls (Fig. 40), are liable to rupture, because the coefficient of contraction of the cementite cell walls is greater than the cell contents. The mass is, in consequence, very feebly held together, and a sudden blow will easily fracture the metal. 1 Intergranular weakness resembles the weakness of a brick building with faulty mortar; the whole might collapse, as jerry-built houses sometimes do when put to a severe test. There is another type of intergranular weakness which is due to imperfect union of the crystal grains. This is particularly marked in phosphorous steels. The crystal grains, on cooling, contract unequally, and tend either to draw the grains away from each other, or to leave the mass in a state of unnatural tension. The fracture then follows the granular junctions. Thick plates and bars are frequently brittle, because comparatively little work has been done on them. The crystals are not interlocked one with another as in steel which has been well worked. Intergranular weakness may, therefore, be of two kinds (1) Brittle envelope surrounding the crystal grains. (2) Imperfect union of the crystal grains. Intergranular weakness may often ..be detected under the microscope. Steels which exhibit " loose " 1 J. O. Arnold, Metalloyraphist, 5. 267, 1902. 35 THE INFLUENCE OF STRESS AND STRAIN 85 intercrystalline ferrite joints, 1 or a meshed or cellular structure of any kind, are to be labelled " dangerous." 35. Intracrystalline or Cleavage Weakness Stead has pointed out another type of weakness in sheet steels which has to do with the crystals them- selves without reference to the union of one crystal with another. It is a kind of intracrystalline weak- ness (intra = within). It is characteristic of some crystals to break more readily in some directions than in others. For example, in a cubical crystal, say of G l FIG. 48. Cleavage Planes. DEF, GHI). Iron crystallizes in the cubic system, rock salt, the crystal splits up more readily in the three directions ABC, DEF, GHI (Fig. 48), parallel to the three crystal faces, and at right angles to one another. This property of crystals is called cleavage. The directions in which the crystal splits are called cleavage planes. A cubical crystal of rock salt has the three cleavage planes shown in Fig. 48 (ABC, 1 That is, a junction which develops rapidly on slight etching with nitric acid, as indicated on p. 100. 86 CRYSTALLIZATION OF IRON AND STEEL 35 and therefore every separate grain in a crystalline mass of iron is more liable to split up in these three directions than in any other. If a bar of iron could be cut from a single crystal, that bar would have three lines of weakness in the direction of the three cleavage planes; while if the bar were built up of a number of crystals whose cleavage planes were all in the same direction, that bar would be more readily broken in the direction of FIG. 49. FIG. 50. its cleavage planes, neglecting for the moment inter- granular weaknesses. Still further, if a number of adjacent crystals at one particular spot in a bar of iron were built up in this manner, then the bar would be more liable to rupture along the cleavage planes at this spot when subjected to the influence of a properly directed force. On the other hand, if the cleavage planes of the adjacent crystals are inclined at con- siderable angles to one another, the bar would be less liable to break than one in which the crystals were arranged symmetrically. Figs. 49 and 50 will make this clear. The dotted lines, ab, Fig. 49, represent the 35 THE INFLUENCE OF STRESS AND STRAIN 87 cleavage planes across a sheet of iron when the crystals are arranged symmetrically; while in Fig. 50 the crystals are arranged in an irregular manner. The cleavage plane of Fig. 49 run along parallel lines, and the sheet would therefore be more liable to rupture than the sheet shown in Fig. 50, where the lines of weakness are not in the same direction, and this in spite of the fact that Fig. 49 has a finer grain. Other things being equal, a fine-grained structure is stronger and tougher than a coarse-grained piece. Figs. 49 and 50 show that this order of things may be reversed. Fortunately the crystals of iron and steel do not in general grow symmetrically. But the development of small into larger grains in the anneal- ing ovens shows that the cleavage planes of a number of adjacent crystals may assume the same direction. Wherever there are large crystals there must be large cleavage planes, or lines of weakness. Intercrystalline or cleavage weakness of this kind is comparatively rare, and its presence may be detected by the usual mechanical tests. A remarkable fact connected with the brittleness of soft steel is, that the fracture is nearly always at an angle of 45 to the direction of rolling, 1 and at right angles to the surface of the steel. But this is also the orientation of the crystals of the metal made brittle by annealing. So long as this material is bent at right angles to that direction, there is no danger of breaking, but if the bending is attempted at an angle of 45 to the direction in which the plates were rolled fracture takes/ place. The rolling seems to impart a tendency to crystallization in certain fixed directions which develop during annealing. 1 J. E. Stead, Journ. Iron and Steel InsL> 54. ii. 137, 1898. 88 CRYSTALLIZA TION OF IRON AND STEEL 36 36. Birth, Growth, and Structure of Crystals It is a fascinating and wonderful power which causes substances in passing from the liquid to the solid state to assume definite and characteristic shapes. I dare not say much upon the birth and growth of crystals. The first visible sign of crystallization in a fluid is the appearance of a globular nucleus, called a globulite, 1 about which the integral parts of the crystal are deposited in a regular manner, so as to produce a typical crystal. The series of drawings shown in Fig. 51 are adapted from a set of "cinemato- graphic " photographs of a growing crystal of potassium iodide under a magnification of 100 diameters, and exposed 0*25th of a second. 2 The little crystal at the bottom right-hand corner of 2, Fig. 51, has a globular form, characteristic of the primitive " globulite." The globular shape, however, is really due to the fact that, while the crystal is small, its growth is so rapid that the borders of the growing mass does not seem to possess a sharply defined outline. Every true crystal is made up of smaller crystals, each of which is a perfect replica of the original. These crystals are made up of smaller crystals, and these in turn, made up of still smaller crystals, and so on. The small crystals called crystallites are arranged in a perfectly definite and regular manner. 1 A. Fock, An Introduction to Chemical Crystallography. Oxford, 1895 ; J. W. Judd, Nature, 44. 83, 1891 ; G. D. Liveing, tWd.,44, 156, 1891. 2 T. H. Richards and E. H. Archibald, Phil Mac)., [6], 2. 488, 1901 ; Am&r. Chem. Journ., 26. 61, 1901. $ 6 ^ 6 ^ a a a d a lS B 8 / 2 3 4- Jg Hartungskohlenstoff ; F. carbon de trempe.) Cellular (cellula = a little cell.) Containing irregular spheroidal or ellipsoidal cavities. Applied to net-like or a meshed structure. (G-. zellig / F. cellulaire.) Cementite, A name given by H. M. Howe to the iron carbide which, according to the works of Abel, Miiller, and Osmond, has the formula Fe 3 C. The term is used in a general way for all the carbides which exist in cast iron and steel, whether these carbides contain manganese chromium, etc. Cementite exists in granules, thin plates, or in comparatively hard masses. Its hardness on Mphs' scale is 6. Is not coloured by polishing or etching with dilute nitric acid, iodine solution, hydrogen chloride, or alcohol, etc. The cementite which is a constituent of pearlite, has been called " segregated " cementite (Sauveur) ; the cementite which occurs independently is then called " free " cementite. It is, perhaps, better to call the former " pearlite-cementite," the latter " excess cementite." (GL Cementit ; F. cementite.) Cleavage (Old Eng.). The property possessed by crystals and crystal grains of splitting more readily in certain directions than in others. These directions are called cleavage planes. The cleavage planes are not necessarily related to the external faces of the crystals. (G-. Spaltbarkeit ; F. clivage.) Cohesion (coharere = to stick). The force of attraction by which particles are held together. (Gr. Kohasion / F. cohesion.) Cold-short. Cold-short steel is steel which is weak and brittle when cold. (G. Jcaltbriichig ; F. cassant a froid.) Conchoidal ("dyx 1 ? = shell, ^s = form). Shaped like APPENDIX 115 a shell. Applied to the concave and convex fractures of some alloys of zinc and copper, tin and copper, glassy slags, etc. (Gr. muschlig ; F. conchoidal.) Congealed solution. A homogeneous solution after solidification, irrespective of whether the constituents of the solution in the latter state form a mixture or a solid solution in the true sense of the word. Do not confuse with solid solution, q.v. (Gr. errsturrte Lb'sung ; F. solution congelee.) Constituent (constituere = to constitute). The structural parts of which alloys and metallic substances are built. Alloys containing only two or more elements or com- ponents may have three or more constituents. (G-. Bestandtheile, Gefugelestandtheile, Gefugebildner ; F. constituent.) Cooling curve. A graphic representation of the thermal changes which occur when liquid or solid substances cool from a higher to a lower temperature, and in which time and temperature are co-ordinates. Cooling curves may be represented in several ways : 1. Take temperature, 0, and time, t, reckoned from the commencement of cooling, as co-ordinates. 2. Take temperature, 0, and the time required for cooling down from that temperature through a definite number of degrees, as co-ordinates. 3. Take temperature, 0, and the difference between the time required for the cooling of the metal under investigation, and the time necessary for cooling some other metal (say, platinum), as co-ordinates. These are sometimes called differential cooling curves. (Gr. AbkvJUungsJnerve, Kilhlungkurve ; F. courle de refroi- dissement.) Critical points. The points at which a physical or chemical change takes place. For instance, in the cooling of pure iron or steel from 900 C. an evolution of heat occurs at the points designated Ar 3 and Ar 2 . This indicates that a physical change occurs at these temperatures. If the change occurs not at one fixed point, but extends over several degrees, the interval is called a zone. (G. Tcritische Punkte, Haltepunkte; F. points critiques.) n6 APPENDIX Cryptocrystalline (K/DVTTTOS = hidden). When the crystalline form of a substance is so fine that only traces of the crystalline structure can be detected, even under the microscope. Crystal (/cpvoraXAos = crystal, ice). This term, as dis- tinguished from mere geometrical solids, is the invariability of the angles between corresponding faces in different individuals of the same substance. The term is sometimes used for particles of crystals, allo- triomorphic crystals, pseudomorphic crystals, and crystallites. (G. Krystall ; F. cristal.) Crystal grain. An allotriomorphic crystal, or a frag- ment of crystal devoid of its crystal faces and angles. (G. KrystallTcorn ; F. grain cristallin.) Crystalline. In crystallography the term refers to the physical properties of crystallized matter, and is applied to any body, or portion of a body, which possesses these properties, without regard to the external form. In petrology the same word refers to mineral aggregates which consist of crystallized sub- stances, whether in the form of perfect crystals, or merely as grains possessing the physical properties of crystals. (G. Tcrystallin ; F. crystalling Crystallite. A word used with several different mean- ings. For example, it is used for all indefinitely crystalline or incipient forms of individualization of minerals ; imperfect crystals in which the plane faces and angles are not developed ; and for the crystal elements which build up the crystal. (G. Krystallit ; F. cristallite.) Crystallized. Refers to any substance, whatever the external form, which has been produced by a process of crystallization, and which possesses the physical properties of crystallized matter. (G. krystallisirt ; F. cristallise.) Cuboidal. Cube-shaped. Applied to crystals which appear to have the form of a cube in certain micro- sections of alloys. (G. iviirfelformig; F. apeupres de la forme ffun cule.) Cuneiform or Cuneate (cuneus ~ a wedge). Wedge- shaped (G. Tceilformig ; F. cuneiforme.) APPENDIX 117 Cuspidate (cuspis = a point). Spear-shaped ; tapering abruptly. (G. speerfwrnig ; F. termine en pointed) Dendritic (8e/8pos = a tree). Tree-like. See *' Arbor- escent." (G. dendritisch, verashlt ; F. dendritique.) Druse (druse = bonny Ger.). A cavity having its interior surface studded with crystals. (G. Druse; F. geode.) Ductility (ducere = to lead). The property of metals to elongate and bend. "Those metals and alloys in which sliding can take place along the cleavage planes without separation occurring." (G. dehnlarkeit ; F. ductilite.) Elastic limit. The maximum stress a substance will bear without suffering permanent deformation, dis- tortion, or fracture. Usually expressed in tons per square inch, or in kilograms per square millimetre. (G. Elasticitatsgrenze ; F. limit cCelasticite.) Elasticity (eXavVeiv = to drive). The property of a substance in virtue of which it tends to recover its size and shape after distortion. (G. Elastiritat ; F. elasticite.) Elasticity, Modulus of. If a rod or wire of length I be stretched until it becomes I + 1, then y is the ex- tension per unit of length ; and if the force causing the extension be / units, and a units be the sectional area of the rod or wire, then the force or stress per unit of cross-section will be Then Young's modulus is defined , , stress a fl Young's modulus = -^ = ^ = - I In iron and steel Young's modulus is about 13,000 tons per square inch. In other words, a stress of one ton per square inch will produce an extension of 13 Q 00 th ii8 APPENDIX of the original length. (G. Elasticitdtsmodul ; F. co- efficient d'elasticite.*) Embrittling. A term used by Howe for those elements which produce brittleness when added to steel. Equilibrium curve. A curve showing the relation be- tween temperature and composition, temperature and pressure, pressure and volume, etc., of any system. For alloys an equilibrium curve shows the relation between the temperature and composition of an heterogeneous system, in which the phases present are in equilibrium with each other. An approximation to the curve can be obtained by plotting a set of cooling curves ; but in order to obtain a true equilibrium curve in this way it would be necessary for the equilibrium curve to be indefinitely slow. (G-. Gleichgewichtskurve ; F. courle d'equilibre.') Etching (cietzen = to corrode Ger.) In order to develop the constitutional and crystalline structure of metals and alloys, the polished surfaces are subjected to the action of suitable reagents nitric acid, iodine, sulphuric acid, hydrochloric acid, etc. This action is termed etching. (G. Aetzen ; F. attaque chimique.) Etching figures. Figures, generally hollowed-out pits, obtained by suitably etching crystalline surfaces, polished surfaces, or the surfaces of cleft crystals. They have a close connection with the orientation of the surfaces of the crystals. If the crystals are equivalent in form and orientation, the figures will be uniform ; if not, the figures will be different. Etching figures are very useful for deciphering the crystalline structure of a " crystalline body." In amorphous bodies there are no etching figures. Very often their size and form depends upon the duration of the etching process, and on the etching medium employed ; hence etching figures are not considered to be identical with the crystal elements which build up the crystal. (G. Aetzfiguren ; F. contours a" 'attaque.) Eutectic alloy (CVT^KTO? = easily melted or liquefied). A term used by Guthrie, in 1875, to indicate that mixture of two substances which has a lower melting- point than any other mixture of the same constituents. APPENDIX 119 All eufcectics have a lower melting-point than that of the mean of their constituents. Eutectic alloys have usually one of three different structures " curved plates," "flat plates," and honeycombed or cellular. As a rule, after polishing and etching, eutectics have a mother-of-pearl or opal appearance when viewed by oblique illumination. Many eutectics take on crys- talline forms during solidification. (GL euteJctisch ; F. eutectique.) Eutectic point. The common point of intersection of two inclined branches, and an approximately horizontal line in the freezing-point curves. The horizontal line is called the eutectic line. The alloy which has the composition corresponding with the eutectic point is called an eutectic alloy ; before solidification, a eutectic solution; after solidification, a eutectic mixture; the temperature at which the eutectic alloy solidifies is called the eutectic temperature. (Gr. euteMscJier Punkt; F. pointe eutectique.") Face. A bounding plane surface of a crystal. Fatigue (fatigare = to fatigue). When a metal is strained just beyond the elastic limit it may recover its original form on standing some time. The metal is then said to be fatigued. (G. elastische Nachwir'kung ; F. fatigue.) Ferrite (ferrum = iron). Term proposed by H. M. Howe for iron. The term is now used for that part of steel or iron containing no carbide, or at least not more than a trace in solid solution. It covers, therefore, iron which may or may not contain silicon, manganese, nickel, etc., which form solid solution or isomorphous crystallized mixtures with iron. Ferrite is the softest structural constituent. It is unmistakably recognized as the production of cuboidal etching figures after treatment with nitric acid or copper-ammonium chloride (1 : 12) solution. (G. Ferrit ; F.ferrite.) Fibrous. Thread-like. Composed of fibres or threads. Fractured surfaces of some metals may indicate a false fibrous structure. The fibrous fractured surface may be the result of tension, which draws out the crystal 120 APPENDIX grains into threads or fibres. (G. fasrig, sehnig ; F. fibreux.) Fissile (fissus, fromfindere = to split). Can be split into laminae. (G. Hating spalibar ; F. clivaUe.) Foliated (folium = a leaf). Composed of thin layers or plates. (G-. Udttrig ; F. lamellaire.) Fracture (frangere == to break). The broken surface of a metal or alloy. (G. Brmh ; F. cassure.) It may be conchoidal (r) = form). A term applied to crystals similar in form. It is also applied to substances which crystallize together to form a homogeneous whole, even though the two constituents do not crystallize in the same form. See " Mixed crystals." (G. isomorph ; F. isomorphe.) Isomorphous mixture. Bodies which crystallize to- gether to form a homogeneous whole. Jog (O.E.). A term used by Howe to express the sudden limited extensions which certain metals undergo at different tenacities. The term was derived from the peculiar shape of the curve obtained when the metal is subjected to tension. Under proper conditions, iron and steel of certain classes, if they have previously APPENDIX 12$ been neither worked cold, nor hardened by qnenching, give string diagrams with a well-marked bend where serious permanent set first occurs. This bend, of which the shape varies much, is called the " jog." Jogless. Refers to iron^r metals whose stress-strain diagram has not this peculiar inflexion. The term " smooth " might be misunderstood. Joint (O.E.). The planes of junction of crystalline grains. (G. Fuge^ Kornyrenze ; F. joint.) Lamella (lamina = a plate). A thin plate ; e.g. plates of graphite in grey pig iron, and of cementite in pearlite. (G. Pldttchm, Bldttchen; F. lamelk.) Lamellar (lamina == d plate). Divisible into thin plates. (G. Udttrig F. lamellaire.) Lamina. A thin leaf -like plate. (G. Bldttchen ; F. lame.) Lanceolate (lanceola = a little lance). Lance-shaped ; tapering gradually. (G. lanzettlich ; F. lanceoU.) Lap (O.E.). A portion of iron or steel folded over on itself ; the walls are oxodized and cannot unite. A lap is caused by careless hammering, or by badly por- portioned grooves in rolls, or by careless rolling, or projections on the ingots. (G. Falz ; F. crique.) Lattice structure. A structure developed on etching Hadfield's manganese steel and certain other metals. (G. gitterformig Structur, Netziverk ; F. treillis.) Lenticular (lens = a lentil). Shaped like a lens. (G. linsenformig ; F. lenticulaire.) Liquation (iiquare = to melt). The flowing out from partially solid metals or alloys of a portion of the still fluid mass. (G. Saigerung ; F. liquation.) Lustre (lustrare = to purify). A term used in describing the character of the reflections obtained from the fractured surfaces of minerals and rocks. (G. Glanz ; F. eclat.) There are several kinds : adamantine (G. Diamant- ; F. adamantin) ; greasy (G. Fett- ; F. gras) \ metallic (G. Metall- ; F. metalliqm) ; pearly (G. Perl- mutter- ; F. nacre) ; resinous (G. Harz- ; F. resineux) ; silky (G. Seiden-; F. soyeux) ; vitreous (G. Glas- / F. vitreuz) ; waxy (Wachs- ; F. cireux). 126 APPENDIX Macles {macula - a spot). Synonymous with twin crys- tals. (G. Zwillinge ; F. macles.) Macroscopic (/xaK/xfe = large). A term used in contra- distinction to microscopic, to imply that the character in question is visible to the naked eye. (G. mafcro- skopisch; F. macroscopique.) Margarite (/xapyaptV^s = a pearl). Vogelsang's name (Die Krystalliten, Bonn, 19, 1875) for the linear arrange- ment, like strings of beads, assumed by crystallites. Martensite (in honour of A. Martens). A constituent of steel produced when small samples are quenched in cold water. The structure characteristic of polished and etched hardened steels. It has the appearance of interlacing rectilinear fibres frequently arranged so as form triangles. The structure is very fine in 1 per cent, carbon steel, and coarse in 0'45 per cent, carbon steel. Some regard martensite as a solid solution of carbon in iron ; Jiiptner considers it to be a solid solution of iron carbide in iron ; Campbell gives it the composition C 2 Fe 6 . (Q . Martensit ; F. martensite.) Massive. A term used in contradistinction to "strati- fied." The term has been used for free cementite as distinct from that in pearlite. The term does not imply that the material is homogeneous. (G. massiv ; F. massif.) Matrix (mater = mother). This term embraces " mother substance " and eutectic. (G. Orundmasse ; F. magma.) Megascopic (/xe'yas = great). See " Macroscopic." Metallography (metallum = a metal). A description of the structure of metals and their alloys. The science which deals with the composition, constitution, struc- ture, and physical properties of : metals and alloys, but does not include the art of metallurgy. That branch dealing with iron and its alloys has been called " siderology." (G. Metallographie ; F. metallographie.) Metallurgy. The art of working metals comprising the whole process of extracting them from their ores, smelting, refining, etc. (G. Hiitenkunde ; F. metal- htrgie.) Melting-point curve. Identical with the freezing-point APPENDIX 127 curve, q.v. (G. SchmehpunktsTcurve ; P. course des points de fusion.) Mixed crystals. A term used for two or more sub- stances which crystallize together as a homogeneous whole. The homogeneous solution of two or more crystallized bodies while in a solid state. From the point of view of the phase rule these form a single phase. (G. Mischkrystalle ; P. cristaux meles.) Mother liquid. During the process of congelation of any system consisting of several substances in a state of solution, a portion of the solution remains in a fluid state after the separating out of solid crystals; this liquor is termed "mother liquor/* or "mother liquid." In systems consisting of two constituents with a eutectic point, the composition of this liquid approximates as congelation proceeds to that of the eutecbic mixture, and quite attains to this composition before the whole mass passes into the solid state. (G. Mutterlange ; P. liqueur 'mere.) Natural hardness. The original degree of hardness possessed by steel before quenching above Ar x . (G. Naturhdrte ; P. durete naturelle.) Needle. See " Acicular." (G. Nadel; P. aiguille.) Non-eutectic cementite. That part of cementite in iron and steel which is external to the pearlite in high carbon steel. Orientation (prire = to rise). The relative direction of the axes of crystals or of the axes of elasticity in two or more crystals, or the relative position of these axes with regard to a certain surface or line ; for instance, the polished surface or a cleavage plane, etc. (G. Orientirung ; F. orientation.) Osmotic pressure ( troostite -> sorbite -> pearlite. It is softer than martensite, and more easily attacked by acids. It gives a brown colour when APPENDIX 133 treated by the polish attack. It is coloured dark by a solution of 1 c.c. of hydrochloric acid in 100 c.c. of alcohol, while martensite is not affected. It is con- sidered, by some, to be a solution of Fe 9 C 3 in iron. (G. Troostit; F. troostite.) Twinned (O.E.). A crystal is twinned when two por- tions of the same individual, or two different individuals are related to one another according to a definite law. (G. verzwillingt, zwillingsUldung ; F. hemitrope, made.) Vesicular (vesicula = a little blister). Containing small holes. (G. mil Blaschen besetzt ; F. vesiculaire.) Widmanstaten figures. Certain patterns which appear on etched surfaces of meteoritic iron. Named after A. B. Widmanstatten. J. 0. Arnold and A. McWilliam, Nature, 71. 32, 1904. INDEX Abkiihlungskurve, 115 Aciculaire, 111 Acicular, 111 Adamantin, 125 Adamantine lustre, 125 Adoueir, 121 Aetzen, 118 Aetzfiguren, 118 Aetzpolieren, 129 Aggregation, States of, 5 Aiguille, 126 Air-hardening steels, 56 Air pits, 111 Alcohol of crystallization, 40 Aliament, M., 14 Alliage metallique, 112 Allotriomorph, 111 Allotriomorphe, 111 Allotriomorphic, 111 Allotropic forms of mercury iodide, 6 sulphur, 5 tin, 6 theory, 62 Allotropie, 111 Allotropy, 5, 111 Alloy, 51 Alpha-eisen, 112 Alpha, fer, 112 iron, 13,14, 15, 54, 112 Alternating stress, Action of, 92 Alumina polishing powder, 97 Aluminium, 15, 19 Ammonium nitarate etching fluid, 102 Amorph, 112 Amorphe, 112 Amorphous, 112 Andrews, T., 18 Anlassen, 123 Anlassfarben, 132 Anlauffarben, 123, 132 Annealing, 58 carbon, 84 fluids, 34 Theories of, 62 Antimony, 19, 20, 22 copper alloys, 44 Fusibility of, 21 Aphanitic, 112 Aphanitique, 112 Aphanitisch, 112 Apparatus for microscopic work, 108 Apparent equilibrium, 4 Arborescent, 112 Archibald, E. H., 88 Arnold, J. O., 37, 39, 40, 58, 64, 69, 72, 83, 84, 101, 133 Arsenic, 15 Atoms, 2 Attaque chimique, 118 Contours d', 118 Ausscheidung, 130 Austen, W. C. Roberts, 20, 23, 41, 52, 69, 112 Austenit, 112 Austenite, 41, 103, 112 Bacillaire, 112 Bacillar, 112 Badlam, S., 72 Bagshaw, W., 107 Ball, E. J., 72 Bands, Slip, 90 Barrett, W. F., 56, 57 The numbers refer to pages. '35 i 3 6 INDEX Beck, J., 108 Behrens, 104, 122 Beilby, G. T., 98 Bell,_Sir L., 82 Bestandtheile, 115 Beta-eisen, 112 Bdta, fer, 112 iron, 13, 14, 15, 54, 112 Bienenwabig, 123 Bismuth, 19, 20, 69, 83 Bivariant system, 46 Blair, A. A., 31 Bliischen, 133 Blasen, 113 hohlraume, 113 Blasig, 123 Blattchen, 125 Blattrig, 120, 125 spaltbar, 120 Blau, 132 Bleu, 132 Blowholes, 112 Boudouard, 0., 13 Boundaries of crystals, 68 Boynton, H. C., 43, 54 Brass, 72 Braun, 132 Breuil, P., 94 Brinell, I. A., 58, 72, 74, 122 Brittleness, 113 of steel plates, 86 Bro'cklig, 120 Brown, W., 57 Bruch, 120 Briichigkeit, 113 Brule, 113 Brun, 132 Buchanan, J. Y., 18 Burning of steel, 75, 112 Calcite, 89 Campbell, E. D., 128 W., 19, 37, 71, 75, 126 Cancellated, 113 Capillary, 113 Carbidekohle, 113 Carbo-allotropic theory, 63 Carbon, 15, 113 Annealing, 34, 113 Carbide, 113 Cement, 30 Cementite, 30 Carbon, Combined, 113 de recuit, 113 Graphitic, 30, 113 Hardening, 30, 114 Influence on tenacity of iron, 49 iron alloys, 22, 27 Solidification of, 33 Solution theory of, 33 Missing, 114 Temper, 34, 113 Carbure, carbon de, 113 Carnot, A., 80 Caron, 30 Carpenter, H. C. H., 36, 108 Carr, W. R., 83 Cartaud, G., 90, 102 Cassant a chaud, 129 froid, 114 Cassure, 120 Cast iron, Grey, 37, 47 Malleable, 38 White, 38, 47 Ultra grey, 38, 70 Cellulaire, 114 Cellular, 114 Cement carbon, 30, 113 Cementit, 114 Cementite, 23, 28, 30, 47, 49, 69, 84, 103, 114 Charpy, G., 13, 22, 37 Chartaud, G., 67 Chatelier, H. le, 13, 19, 20, 23, 24, 35, 55, 56, 63, 105, 107 Chaud, Cassant a, 129 Chauffage, Oxydation par, 123 Chernoff, See Tschernoff. Chrome steels, 31 Chromic acid etching fluid, 102 Chromium, 15, 31, 56, 80 and iron carbide, 81 Cireux e'clat, 125 Cleavage, 85, 90, 114 weakness, 85 Clivable, 120 Clivage. 114 Plan de, 121 Cobalt, 80/81 Cohen, E., 7 Cohesion, 114 Cold-short, 114 Cold work, 77 Colour names for temperatures, 24 Components, 44 Compound, 31 Conchoidal, 114 The numbers refer to pages. INDEX 137 Congealed solution, 115 Congele'e, solution, 115 Constituent, 115, 131 Contours d'attaque, 118 Cooling curve, 9, 115 Differential, 115 of copper, 9 of iron, 12 of sodium thiosulpbate, 10 of steel, 12 of water, 10 Influence of rate of, 51 Rate of, 58 Coote, A. H., 72 Copper, 15, 69, 80, 83 antimony alloys, 26, 27, 44 Fusibility of, 21 Cooling curve of, 79 silver alloys, 18, 47 Corpuscles, 2 Couleurs de recuit, 132 Couture, 130 Crique, 125 Cristal, 116 Reseau du, 130 Cristallin, 116 grain, 116 Cristallise, 116 Cristallite, 116 Cristaux meles, 127 Critical points, 13, 64, 115 Critiques, Points, 115 Cryohydrate, 17 Cryosel, 17 Cryptocrystalline, 116 Crystal, 1, 67, 116 boundaries, 68 grain, 116 Crystalline, 116 Crystallite, 88, 116 Crystallization, Alcohol of, 40 Iron of, 40 Water of, 40 Crystallized, 116 Cuboidal, 116 Cnneate, 116 Cuneiform, 116 Cuneiforme, 116 Curie, S., 13 Cuspidate, 117 Dachziegelartig, 123 Dahms, A., 19 Degradation of energy, 4 Degre de recuit, 132 Degree of freedom, 45 Dehnbarkeit, 117 Dendritic, 117 Dendritique, 117 Dendritisch, 117 Dependent variable, 45 Diamantglanz, 125 Diamantine, 96 Dillner, G., 31 Direction de solution, 131 Druck, Osmotischer, 128 Druse, 117 Ductilite, 117 Dumas, L., 57 Dumont, E., 56 Durchdringnng, 124 Durete-, 122 da verre, 120 naturelle, 127 E Eben, 120 Ebner, 131 Ecailleuse, 120 Echauffement, courbe d', 123 Eclat, 125 Eggertz's color test for carbon, EinschlUsse, 124 Elastic limit, 91, 117 Elasticity, 117 Elastic! til tsmodul, 118 Elasticite, 118 Co-efficient, 118 limit, 117 Elasticity, 117 Modulus of, 117 Young's, 117 Elastische nachwirkung, 119 Electrons, 2 Elements, 2 Embrittling, 118 Energy, 3 Degradation of, 3 Enfermees, Scones, 130 d'Equilibre, Courbe, 118 Equilibrium, 4 Apparent, 4 Curve of, 118 False, 4 Erhitzungskurve, 123 The numbers refer to pages. 138 INDEX Errstarrte Lo'silng, 115, 131 Erstarrungspunkteskurve, 120 Etching, 98, 118 figures, 89, 118 fluid, 98 Etoile, 132 Eutectic alloys, 118, 119 - line, 119 mixtures, 17, 32, 44, 119 point, 119 solution, 119 steels, 28, 29 temperature, 119 Eutectique, 119 pointe, 119 Eutektisch, 119 Eutektischer Punkt, 119 Ewing, J. A., 90, 93, 94 Face, 119 False equilibrium, 4 Falz, 125 Faraday, M., 30 Fasrig, 120 Fatigue, 90, 119 Fay, H., 72 Ferrit, 119 Ferrite, 23, 28, 40, 47, 49, 103, 119 crystal, 67 Feste Losungen, 131 Fettglanz, 125 Fibreux, 120 Fibrous, 119 Filing for microscope, 96 Findlay, A., 48 Finishing temperature, 78 Fissile, 120 Flusseisen, weiches, 132 barter Stahl, 132 Mittelhartes, 132 Fock, A., 88 Foliated, 120 Fondamehtale, masse, 121 Fracture, 70, 120 Freedom, Degree of, 45 Freezing of aqueous sodium chloride, 16 point curve, 120 water, 47 Fremont, C., 90 Friable, 120 Froid, Cassant a, 114 Fuge, 125 Fusibilite, courbe de, 120 Fusibility curve, copper-antimony alloys, 18, 21 gold-silver alloys, 19 Fusion, Courbe des points de, 127 Gamma-eisen, 120 Gamma fer, 120 iron, 13, 14, 15, 54, 120 Gas, 2 in steel, 75, 76 Gautier, A., 20 Gefrierpunktskurve, 120 Gefiigebestandtheile, 115 Gefiigebildner, 115 Geode, 117 Gestrickte Formen, 130 Gibbs, J. W., 44 Gitterformig, 113 Struktur, 125 Glanz, 125 Glasglanz, 125 Glasharte, 120 Glass hardness, 120 Gledhill, J. M., 56 Gleichgewichtskurve, 118 Gleitflache, 121 Gliding plane, 120 Globulaire, Structure, 131 Globulite, 88, 121 Gold, 15, 19, 22, 69, 83 silver alloys, 20 Gore, G., 56 Goutal, E., 80 Grain, 121 cristallin, 116 size, 70 Granite, 48 Granular, 121 Granulitic, 121 Granulitique, 121 Granulitisch, 121 Graphite, 30, 114 temper, 34 Graphitic carbon, 30, 113 Gras eclat, 125 Greasy lustre, 125 Grenet, L., 13, 37 Grenu, 121 Grey cast iron, 37, 47 Ultra, 38, 70 The numbers refer to pages. INDEX 139 Grey tin, 6 Grinding, 121 Groundmass, 121 Grunduiasse, 121 Guillaume, A. E., 56 Guthrie, F., 16 H Hackly, 121 Hadfield, R. A., 56 Haltepunkte, 115 Harbord, F. W., 80 Hardening, 58, 121 carbon, 30, 114 Theories of, 62 Hardenit, 121 Hardenite, 39, 40, 121 Hardness, 122 Scales of, 122 Hart, 132 Harte, 122 Harten, 121 Hartungskohlenstoff, 114 Harzglanz, 125 Heat refining, 74 tinting, 104, 123 tints, 123 Heating curve, 13, 123 Hemitrope, 133 Heusler. F., 57 Heusler's magnetic alloys, 58 Heycock, C. T., 19, 22, 129 Heyn, E., 23, 30, 40, 75, 76 High-speed steels, 56 Hiorns, A. H., 80, 108 Holocrystallin, 123 Holocrystalline, 123 Honeycombed, 123 Hopkinson, 56, 92 Hot-short, 123 Hot-work, 77 Housman, R. H., 69 Howe, H. M., 19, 49, 50, 55, 58, 62, 72, 77, 114, 118, 124, 128 Humfrey, J. C. W., 90, 93 Hutenkunde, 126 Hydrochloric acid etching fluid, 101 Hydrogen in steel, 76 Hypereutectic steels, 28, 29, 49, 123 Hypoeutectic steel, 123 Hysteresis, 123 Ice, 16 Idiomorphe, 123 Idiomorphic, 123 Idomorph, 123 Igewsky, 101 Illumination of micro-objects, 106 Oblique, 31, 106 Vertical, 71, 106 Imbricated, 123 Imbrique', 123 Inclusions, 123 Independent variable, 45 Inegale, 120 Intercellulaire, 124 Intercristallin, 124 Intercrystalline, 124 weakness, 83 Intergranular, 124 weakness, 83 Interpenetration, 124 Intracrystalline, 85, 124 Invar alloys, 56 Invariant system, 46 Iodine etching fluid, 99 Iron carbide, 23 of chromium and, 81 carbon alloys, 22, 27 solidification of, 33 solution theory of, 33 Cast, Grey, 37, 47 Malleable, 38 Ultra grey, 38, 70 mite, 38, 47 Cooling curve of, 12 influence of carbon on, 49 of crystallization, 40 subcarbide, 37, 40 sulphide, 83 Isomeric, 124 Isomerique, 124 Isomerisch, 124 Isomorph, 124 Isomorphe, 124 Isomorphous, 124 mixture, 32, 124 Jaune paille, 132 Jefferson, J., 69 Job, R., 77 Jog, 124 The numbers refer to pages. 140 INDEX Jogless, 125 Joint, 125 Judd, J. W. } 88 Jtiptner von Jonstorff, H., 30, 35, 66, 71, 72, 80, 112, 126, 128, 131 Kaltbruchig, 114 Keeling, B. F. E., 36 Keilformig, 116 Kemp, A. W., 92 Kish, 70 Kohasion, 114 Kohlungsgrad, 132 Korn, 121 Korngrenze, 125 Koraig, 121 Kreuzpointer, P., 94 Krystall, 116 Krystallin, 116 Krystallisirt, 116 Krystallit, 116 Krystallkorn, 116 Krystallskelette, 130 Kritische Punkte, 115 Kiihlungskurve, 115 Lame, 125 Lamella, 125 Lamellar, 125 Lamellaire, 120, 125 Lamelle, 125 Lamina, 125 Lanceolate, 125 Lanceole, 125 Lanzettlich, 125 Lap, 125 Lattice structure, 125 Lead, 19 Ledebur, A., 39, 63, 113 Lenticulaire, 125 Lenticular, 125 Levy, M., 121 Liegierung, 112 Lines, Slip, 90 Linsenformig, 125 Liquation, 125 Liqueur-mere, 127 Liquid, 3 Liquorice root etching fluid, 102 Liveing, G. D., 88 Ljamin, N., 71 Longmuir, P., 59 Losung, Errstarrte, 115, 131 Losungflache, 131 Lowry, T. M., 55 Luftgriibchen, 111 Lugfestigkeit, 132 Lunker, 128 Lustre, 125 M Macles, 126, 133 Macroscopic, 126 Macroscopique, 126 Magma, 126 Magnetic alloys, 58 Magnetism, 57 Makroskopisch, 126 Malleable cast iron, 38 Manganese, 15, 36, 56, 73, 80, 81, 83 carbides, 81 silicide, 81 sulphide, 81 Margarite, 126 Martens, A., 33, 104, 124 Martensite, 33, 39, 40, 42, 47, 103, 126 Mass action, 60 Masse fondamentale, 121 Massif, 126 Massiv, 126 Massive, 126 Matrix, 126 McMillan, W. G., 69 McWilliam, A., 39, 132 Mechanical work, Influence on steel, 77 Megascopic, 126 M616s, cristaux, 127 Mellor, J. W., 62 Mercury iodide, 6, 9 Allotropic forms of, 6 Metallglanz, 125 Metallic lustre, 125 Metallique 6clat, 125 Metallographie, 126 Metallography, 126 Metallurgy, 126 Melting-point curve, 126 Microscope, 105 Microscopic work, 95 The numbers refer to pages. INDEX 141 Mischkrystalle, 127 Mittelhart, 132 flusseisen, 132 Mixed crystals, 32, 127 Modulus of elasticity, 117 Young's, 117 Mobs 1 scale of hardness, 122 Molecular state of dissolved sub- stances, 65 weight, 6G Molecules, 1, 2 size of, 1 Molybdenum, 56 Morse, R. G., 73 Mother liquid, 127 Mounting (microscopic), 104 Multivariant system, 46 Muschlig, 115 Muscovite, 48 Mushet, R., 56 steel, 56 Mutterlange, 127 N Nache, 121 Nachwirkung, Elastische, 119 Nackig, 121 Nacre eclat, 125 Nadel, 127 Xadlig, 111 Naht, 130 Natural hardness, 127 Naturelle durete, 127 Naturharte, 127 Needle, 127 Needle-shaped. See "Acicular, " Prismatic." Netzartig, 129 Netzformig, 129 Netzwerk, 125 Neville, F. H., 19, 22, 129 Nicholson, J. T., 56 Nickel, 15, 56, 63, 80, 81 Nitric acid etching fluid, 100 Non-euteche cementite, 127 Non-expansive alloys, 56 Oberflachenspannung, 132 Oblique illumination, 71, 106 Octahedral sulphur, 5 Orientation, 127 Orientirung, 127 Orthoclase, 48 Oscillatory stress, Action of, 92 Osmond, F., 12, 13, 14, 35, 40, 41, 57, 58, 63, 67, 69, 71, 90, 99, 102, 104, 108, 112 Osmotic pressure, 127 Osmotique pression, 128 Osmotischer Druck, 128 Overheating steel, 75 Oxalic acid etching fluid, 102 Oxydation par chauffage, 123 Passive resistance, 4, 8, 60 Pearlite, 23, 28, 47, 69, 84, 103, 128 Pearly constituent, 128 lustre, 125 Pene'tration, 124 Perlmutterglanz, 125 Phase, 44, 128 Loi de, 128 rule, 43 Phasenregel, 128 Phosphide steels, 104 Phosphorus, 15, 80, 81 Photography, 107 Picric acid etching fluid, 101 Pipe, 128 Plastic sulphur, 5 Plasticity, 90 Plates, Brittleness of steel, 86 Plattchen, 125 Polieren, 129 Polish attack, 102, 128 etching, 128 Polishing for microscope, 96, 129 in relief, 97 powder, 97 Polissage, 129 attaque, 129 en bas relief, 128 Polyhedral, 129 Polymorphism, 129 Polymorphisme, 129 Polymorphismus, 129 Ponsot, A., 17, 18 Popplewell, F., 51, 70 Potassium iodide, 88 Pouillet, C. S. M., 24 Preserving polished specimens, 105 Pression osmotique, 128 The numbers refer to pages. 142 INDEX Prismatic, 129 sulphur, 5 Prismatique, 129 Prismatisch, 129 Pseudomorpb, 129 Pseudomorphe, 129 Pseudomorphous, 129 Quartz, 48 50 Radiating, 129 Rapid-cutting steels, Rate of cooling, 58 Influence of, 51 Recalescence, 12, 129 Recalescenz, 129 Recuit, carbon de, 113 Colorations de, 123 Couleurs de, 132 Degrede, 132 Red-short, 129 Refining, Heat, 74 Refroidissement, Courbe de, 115 Relief polieren, 128 polishing, 97 Reseau du cristal, 130 Resineux eclat, 125 Resinous lustre, 125 Resistance, Passive, 4, 8, 60 Retassure, 128 Reticulated. 129 Reticule^ 129 Richards, T. H., 88, 108 Risdale, C. H., 72, 74, 83, 107 Romanoff, L., 76 Roozeboom, H. W. B., 35, 90 Rotation effect, 129 Rothbriichig, 129 Rothmund, V., 32 Saigerung, 130 Sampling for microscope, 95 Saniter, E. H., 56 Saturated steels, 28, 29 Saulenformig, 129 Sauveur, A., 21, 28, 39,54, 55, 71, 72, 73, 100, 114 Scalibleux, 130 Scaly, 130 Schillerization, 130 Schlackenartig, 130 Schlackenartiger Korper, 130 Schlackeneinschliisse, 130 Schlackig, 130 Schleifen, 121 Schlerometer, 130 Schlerometre, 130 Schmelzpunktskurve, 127 Schuppig, 130 Schwindungshohlraum, 128 Scoriace, 130 Scoriaceous, 130 Scoriae, 130 Scories, 130 enfermees, 130 Seam, 130 Segregation, 130 Sehnig, 120 Seidenglanz, 125 Self-hardening steels, 56 Short, 130 Siderology, 130 Silicon, 15, 36, 73, 80, 81, 89 Silky lustre, 125 Silver, 15, 19 copper alloys, 18, 26, 27, 47 gold alloys, 20 Size of grains, 70 of molecules, 1 Skelerometer, 130 Slag inclusions, 130 Slip bands, 90, 131 lines, 90 Sodium chloride, 44, 47 freezing aqueous solutions, 16 thiosulphate, 4, 11, 47 cooling curve of. 10 Solid, 3 solutions, 31, 32, 131 Solides solutions, 131 Solidification, molten iron-carbon alloys, 33 Solidified solution, 32, 131 Solution, 32 congelee, 131 direction de, 131 plane, 131 solid, 31, 32, 131 supersaturated, 11 theory of iron-carbon alloys, 33 Sorbit, 131 Sorbite, 42, 53, 54, 131 The numbers refer to pages. INDEX 143 Sorby, H. C., 23, 27, 42, 128, 131 Soyeux eclat, 125 Spaltbar, Blattrig, 120 Spaltbarkeit, 114 Speerformig, 117 Spharolithische Struktur, 131 Spheralitic structure, 131 Spiller, F. W., 101, Splittrig, 120 Stabchenformig, 112 Stahl Flusseisen, Harter, 132 Stansfield, A. S., 33 State of aggregation, 5, 131 Stead, J. E., 19, 22, 54, 67, 68, 74, 75, 79, 81, 83, 87, 89, 104, 107, 108 Steady stress, Action of, 90 Steel chrome, 31 Cooling curve of, 11 Stellate, 131 Sternformig, 132 Stress, Action of alternating, 92 oscillatory, 92 steady, 90 vibratory, 92 Strohgeib, 132 Struktur, Spharolithische, 131 Subcarbide of iron, 37, 40 theory, 64 Sulphur, 5, 6, 15, 36, 80 Allotropic, 5 Octahedral, 5 Plastic, 5 Prismatic, 5 Sulphuric acid etching liquid, 102 Superficial tension, 132 Superficielle, tension, 132 Supersaturated solution, 11 Surface tension, 132 Surfusion, 11, 37, 132 Svedelius, G. E., 13 Tannenbaumformig, 112 Taylor, F. W., 24 Tellurium, 15 Temper, 132 carbon, 34 graphite, 34 Temperatures, colour names for, 24 Finishing, 75 Transition, 6 Tempering, 58 colours, 132 'emperkohle, 113 Tenacite, 132 Tenacity, 132 of iron, 70 Tension, Superficial, 132 superficielle, 132 Pennine en pointe, 117 Tiemann, H. P., 41 Tin, 6, 7, 19 Allotropic forms of, 6 - Grey, 6, 7 -pest, 7 White, 6, 7 Tinting, heat, 104 Transition temperatures, 6 Trellis, 125 Forme en, 113 Trempe, 121 Carbon de, 114 Troost, L., 43, 132 Troostite, 43, 103, 132 Tschernoff, D.. 14, 67, 72 Tungsten, 15, 56 Twinned, 133 Uberschmelzung, 132 Ultra grey cast iron, 38, 70 Unie, 120 Univariant system, 46 Unsaturated steels, 28, 29 Variables, Dependent, 45 Independent, 45 Velocity of transformation, 7 Venteux, 123 Verbrannt, 113 Verastelt, 117 Verre, durete du, 120 Vertical illumination, 71, 106 Verzwillingt, 133 Vesiculaire, 133 Vesicular, 133 Vibratory stress, Action of, 92 Violet, 132 Violett, 132 Vitreous lustre, 125 Vitreux eclat, 125 Vogelsang, 126 Vollkrystallin, 123 The numbers refer to pages. 144 INDEX W Wabig, 123 Wachsglanz, 125 Water, Cooling curve of, 10 Freezing of, 47 of crystallization, 40 Waterhouse, P. B., 83 Waxy lustre, 125 Weakness, Cleavage, 85 Intercrystalline, 83 Intergranular, 83 Intracrystalline, 85 Weich, 132 Weiches Flusseisen, 132 Werth, J., 63 White, M., 24 cast iron, 38, 47 tin, 6 Widrnanstaten, A. B., 133 figures, 133 Wohler, A., 92 Work, Cold, 77 Hot, 77 Influence of mechanical, 77 Wiirfelformig, 116 Yield point, 91 Young's modulus of elasticity, 117 Zellig, 114 Zinc, 19 Zone, 115 Zwillinge, 126 Zwillingsbildung, 133 Hie numbers refer to pages. THE END PRINTED BT WILLIAM CLOWES AND SONS, LIMITED, LONDON AND BECCLE3. RETURN TO the circulation desk of any University of California Library or to the nn REGIONAL LIBRARY FACILITY 00, Richmond Field Station University of California Richmond, CA 94804-4698 ALL BOOKS MAY BE RECALLED AFTER 7 DAYS Renewals and recharges may be made 4 days prior to due date BELOW DD20 15M 4-02 YB 156 UNIVERSITY OF CALIFORNIA LIBRARY