STEEL AND ITS HEAT TREATMENT BY DENISON K. BULLENS * Consulting Metallurgist FIRST EDITION FIRST THOUSAND NEW YORK JOHN WILEY & SONS, INC. LONDON: CHAPMAN & HALL, LIMITED 1916 Copyright, 1916 BY DENISON K. BULLENS .1/Z rights r?serocd THE SCIENTIFIC PRESS ROBERT DRUMMOND AND COMPANY IN MEMORY OF MY FATHER 331168 PREFACE MODERN Heat Treatment should be considered as an art or trade, since it certainly requires knowledge, skill and judgment for its proper performance. These, in turn, necessitate at least some knowl- edge of heat, of steel, and of the effect of heat upon steel. And all three factors are linked together by the " human element." The author has therefore endeavored to bring together the theoretical and practical sides of the general subject of steel and its heat treatment in such a manner as will, he hopes, be understandable by that " human element." It has been the author's attempt to make the chapters dealing with the heating problem more of a " heat talk " than of a " furnace talk"; of heat application rather than details of construction; of the importance of the human element and scientific efficiency rather than the elimination of the human element through scientific management; and finally, of viewing the heating problem as an engineering prop- osition, adapting each fuel to proper furnace design and operation to meet the requirements of the problem in hand, and by so doing aim for the adoption of the standard heating unit in terms of finished product " the cost of a unit of quantity of given quality." He has attempted to make as practical as possible those chapters relating to steel and the effect of heat upon steel. Theories have been advanced only so far as has been thought necessary for a clear understanding of principles. Wherever possible, illustrations in the form of photomicrographs and charts have been given. The data given under the various types of heat-treated steels have been checked as far as possible and every effort has been made to be correct. v vi PREFACE To the many friends who have aided him in the preparation of this book the author would express his sincere appreciation. Effort has been made to give due credit for cuts and data at the proper place, and for such as may not have been made, acknowledgment is hereby rendered. DENISON K. BTTLLENS. PHILADELPHIA, October 1, 1915. CONTENTS CHAPTER PAGE I. THE TESTING OF STEEL 'I'- ll. THE STRUCTURE OF STEEL 16 III. ANNEALING 39 IV. HARDENING 65 V. TEMPERING AND TOUGHENING 96 VI. CASE CARBURIZING . 112 VII. CASE HARDENING: THERMAL TREATMENT 154 VIII. HEAT GENERATION 173 IX. HEAT APPLICATION 192 X. CARBON STEELS 229 XI. NICKEL STEELS 257 XII. CHROME STEELS 295 XIII. CHROME NICKEL STEELS 306 XIV. VANADIUM STEELS 335 XV. MANGANESE, SILICON AND OTHER ALLOY STEELS 344 XVI. TOOL STEEL AND TOOLS 357 XVII. MISCELLANEOUS TREATMENTS 386 XVIII. PYROMETERS AND CRITICAL RANGE DETERMINATIONS. . . 408 vii STEEL AND ITS HEAT TREATMENT CHAPTER I THE TESTING OF STEEL Growth of Heat Treatment. Probably no one division in the metallurgy of steel has taken such wonderful strides in recent years as has the art of heat treatment. Twenty years ago the scientific knowledge and technical application of heat treatment were but very limited. Such as it was, it usually consisted in " heating to a red heat " for annealing, or perhaps the instructions called for " harden at a bright red and temper to a straw color." Then it was an art guarded with much secrecy and confined for the most part to makers of tools and a certain few specialties. Practically all alloy steels require treatment of one sort or another. In the " natural " state very few steels present their full value, so that heat treatment is not only advisable but often mandatory. Necessity for Heat Treatment. Take for example the steels used in the automobile industry. The frame requires resistance to vibratory stresses occasioned by rough roads, as well as strength and toughness. Rear axles must have great tortional resistance; front axles must withstand vibrations. The steering parts must be strong, tough and without brittleness; the springs must neither sag nor break. Crank-shafts must be able to resist impact, besides being stiff. Gears are subject to wear and must be capable of withstanding this action if a smoothly running transmission is to be had. And so each separate part might be named, all having a more or less severe duty to perform and requiring steels possessing various degrees of strength, toughness, resilience, endurance, shock-resisting and wearing qualities. Testing. These various combinations of static and dynamic strength} are obtained by adjusting and correlating both the chemical 2 STEEL AND ITS HEAT TREATMENT composition and heat treatment of the steel. Certain chemical components intensify the static properties of the material; others may affect the dynamic qualities. Thus by coupling with a steel of suitable chemical combination the proper heat treatment, there arises a product with physical properties most adapted for the work in hand. Similarly, having once produced a suitable article, it then remains to duplicate it. To this end all rational heat treatment must be aligned and standardization of results be obtained. In order to accomplish these specific requirements, the influence of definite chemical composition and definite treatment must be known, as will be described in later chapters. The guide to this work is frequent and constant testing, and a definite knowledge of the vari- ous components should be possessed by every heat-treatment man. Thus we may say that the purpose of practical testing is (1) to sup- ply information as to suitable material and its qualities for different purposes, both for the manufacturer of the material and for the designer or user, and (2) to test the specified uniform quality of the material. Stresses and Strains. Testing resolves itself into a determination of the strength of the material, which in turn is measured by the application of a force and its resultant effect. The force put upon a body is termed the stress, and the deformation resulting from that force is the strain. Upon the method of applying that force depends the nature of the test. Thus we may conveniently classify such stresses under the following headings: A. Steady or constant loads static stresses; B. Repeated static stresses and accelerated stresses fatigue stresses; C. Suddenly applied loads impact stresses; D. Repeated impact or vibratory stresses dynamic stresses; E. Miscellaneous tests such as resistance to penetration, wear, etc. Tensile Strength. The most common test for static strength, that is, the strength of the steel under constant load and without shock or vibration, is the tensile test. Thus we may call the tensile strength the absolute strength of the metal under tension, i.e., the force actually required to pull the metal asunder. A standard test piece is gripped between the upper and lower jaws of a testing machine and the total resistance to rupture is measured. Knowing the area of the cross-section of the test piece and the load required to break THE TESTING OF STEEL 3 it, the strength per square unit may then be calculated. The tensile strength is usually given in pounds per square inch (American), tons per square inch (British), or kilograms per square millimeter (metric system; 1 kg. per mm. 2 =1422. 32 Ibs. per square inch). The accuracy of the tensile test is dependent not only upon the conditions under which the test is made, such as the rate of pulling, alignment of test piece in the machine, etc., and which are more or less influenced by the human element, but also upon the metal itself. The higher the tensile strength and brittleness of the steel, the greater the possibility of error; differences of several thousand pounds per square inch are often encountered in the same piece of high-tensile, heat-treated steel, even in the absence of brittleness. Test pieces are generally taken half way between the center and the outside of the piece, and longitudinally or " with the grain." Occasionally it is necessary to take tests transversely or " across the grain"; in this case the results will be lower than in the longitudinal test, the exact amount depending upon the composition and treat- ment of the steel. The static strength increases in direct proportion to the carbon content of the steel. For the ordinary basic and acid open-hearth steels, without heat treatment, Campbell gives the following formulae by which the tensile strength of such steels may be roughly deter- mined. These results apply for steel " in the natural." Acid open-hearth steel : Tensile strength = 40,000 +1000C + 1000P +xMn. Basic open hearth steel: Tensile strength = 41,500+7700 + 1000P+?/Mn. In these formulae, C equals each one point (0.01 per cent.) of carbon as determined by combustion, P equals each 0.01 per cent, of phosphorus, Mn equals each 0.01 per cent, of manganese, and x and y are given in the table on page 4. Elastic Limit (Tension). The term " elastic limit " has probably been more ill-used than any other common technical testing name, with the possible exception of " hardness." Among its many definitions the two which stand out pre-eminently are (1) the least stress at which the material retains a permanent deformation or " set " after the removal of the stress; and (2) the least stress under STEEL AND ITS HEAT TREATMENT Percentage of Carbon. On Acid Steel. X Lbs. per Sq. In. On Basic Steel. y Lbs. per Sq. In. 0.05 not 0.10 80* 130 0.15 120 150 0.20 160 170 25 200 190 0.30 240 210 35. 280 230 0.40 320 250 0.45. 360 0.50 400 55 440 0.60 480 * Beginning only with 0.4 per cent, manganese, t Beginning only with 0.3 per cent, manganese. which ductile material exhibits a marked yielding sometimes denoted as the " yield point." The determination of the true elastic limit should always be taken from a curve plotted, using an extensometer, from a series of careful observations, as otherwise sets caused by non-homo- geneity and initial stress might be obtained which do not repre- sent the plasticity of the material. This method of determining the elastic limit is but little used commercially, as the amount of labor involved is too great. The yield point, or commercial elastic limit, is obtained by noting the stress at which the test piece first begins to " give " or elongate. This may be obtained by means of two prick-punch marks and observing the first signs of elongating by means of dividers held on these points; or by noting the drop of the weighing beam or halt in the load indicator (" jockey ") ; or by means of the general appear- ance of the test piece. In its practical application the elastic limit may be called the working strength of the material, for in most cases the steel or machine part becomes useless when strained beyond its elastic limit. This is particularly true of automobile construction, in which the value of a car is dependent upon the correct adjustment and alignment of its several working parts, such as in transmissions and transmission suspensions. All tests given in this book, unless otherwise noted, refer to the commercial elastic limit or yield point. THE TESTING OF STEEL 5 The relation existing between the elastic limit and the tensile strength is too broad a subject for discussion here, as the varying chemical compositions and heat treatments exert such a tremen- dous influence; a study of the results given in following chapters will show a proportionality of 40 per cent, and upward. Elongation. The elongation is measured in per cent, of the original test section and is commonly the amount of stretch which will occur in the material when pulled apart by tension. It is usually measured in relation to an initial distance of 2 or 8 in., or 100 mm. when the metric system is used, but other specifica- tions as used in Europe give a definite relation of original gauge- length to the thickness or diameter of the specimen. Reduction (or Contraction) of Area. The reduction of area refers to the area at the point of rupture, usually reported in per cent, reduction of the original area that is, the original area of the test piece minus the area of the smallest cross-section after frac- ture; this divided by the original area is the percentage reduction of area. Ductility. The percentage elongation and percentage reduction of area are a measure of the " ductility " of the material, usually varying inversely with the tensile strength. The true measure of the ductility of the steel cannot be taken alone from either the elonga- tion or reduction of area, as the results obtained in either case will depend in a large measure upon the size of the test piece, the method of testing, etc. Many engineers regard the reduction of area as the more reliable; this is offset by the fact that many steel specifications make no mention of the reduction of area, but particularly specify the percentage elongation. Ductility may also be defined as the amount of distortion of the material before final rupture. Compressive Strength. The compressive strength of material is its resistance to crushing. The test is generally carried out upon a small cylinder or 1-in. cube of the metal, using the same machine as for the tensile test. Care must be used to see that the line of strain passes exactly through the axis of the specimen, and that the plates above and below the piece have a greater resistance to pene- tration than the metal to be tested. The application of the term elastic limit is similar to that in the tensile test. Torsional Strength. As its name implies, the torsion test is used to determine the resistance to twisting. This test is very largely used to-day for automobile steel and is measured in inch-pounds with the amount of distortion given in degrees. The elastic limit is 6 STEEL AND ITS HEAT TREATMENT obtained as in a tension test, using either a tropometer or an auto- graphic attachment. The usual comparison is by calculating the shearing stress in pounds per square inch. Endurance. The computation and understanding of such static stresses as have been previously outlined are comparatively simple. The requirement to be fulfilled in designing is that the working stress shall not exceed the elastic limit of the material, whether it be in tension, compression or torsion. Numerous every-day failures, however, which cannot be accounted for by the limited information given by such tests, have forced investigators to probe more deeply into the complicated kinematic forces which seem to have such a great influence upon the " life " of the metal. It is now a well- known fact that, if a stress is applied a great number of times, i.e., repeated, each application being made before the material has had time to recover from the preceding stress, the material will event- ually break even though the stress is below the elastic limit of the material. These repeated stresses upon steel cause a gradual dis- turbance of the structure and its component particles, which greatly weakens the material, and is called fatigue. The resistance to fatigue and its numerical test value may be termed the endurance of the steel. The stresses embodied under the heading of fatigue may be broadly classed as repeated static stresses and acceleration stresses ranging from zero to maximum or from a negative maximum to a positive maximum alternating stresses. Fatigue Stresses. These stresses are produced in a machine part by an external force or forces of varying strength and direction acting upon the part. When the force is produced by a continu- ously varying acceleration or retardation of masses taking part in the movement of the machine part, they may be conveniently termed acceleration stresses. 1 Typical stresses of this category are the revolving shaft stress on a loaded wheel or machine axle, the piston pressure and the acceleration pressure of the movable parts in the piston rod and crank-shaft of high-speed steam and oil engines. These perpetual stresses or so-called fatigue stresses are the essential ones' in the movable parts of most high-speed machines, and a knowl- edge of the capacity of the material to resist them should serve as a basis for the selection of the material and design. Rotary Bending. Such static endurance tests may be carried out in a machine of the rotary bending type, such as the Wohler or the White-Souther machines. From a study of a large number of 1 J. O. Roos af Hjelmsaeter, Int. Assoc. Test. Mat., 1912, Vol. II, No. 9. THE TESTING OF STEEL 7 experiments made on a rotary bending machine of the Wohler design, Foos concludes that such endurance tests are not suitable as specification tests, but are of great value in the selection of material and the heat treatment for various purposes. On the other hand, the real value of the rotary bending test as a criterion of the brittleness-fatigue endurance has been of late greatly questioned. That the results usually obtained are largely indicative of the elastic limit alone is probably more in accord with our present- day knowledge. The results from a series of tests conducted by Foos upon a Wohler type rotary bending machine with steels of 0.11, 0.40 and 0.65 per cent, carbon, given in the following table, would tend to support the latter theory, as one would naturally expect from past experience that the 0.40 per cent, carbon steel would have a greater fatigue-resisting strength than the 0.65 carbon steel. ROTARY BENDING TESTS, WOHLER MACHINE Chemical. Static Properties. Endurance Limit. Fiber Stress d o> ta fl o ,0

>*: / / / / Pea rlite / / ~/ 0.2 0.4 0.6 0.8 1.0 1,2 1.' Per Cent Carbon FIG. 1. Ferrite-pearlite-cementite Diagram. (Sauveur.) ratio, we see the first appearance of the free cementite between the pearlite crystals. In Fig. 11 we have this whole range represented by means of case carburizing a " dead soft " steel. Physical Properties Dependent upon Constituents. Upon the relative proportions of these constituents will depend the physical properties of the slowly cooled steel, neglecting for the time being their relative arrangement. Each of these components ferrite, pearlite and cementite has certain physical characteristics with which we must be familiar in order to gain some idea of the proper- ties of such steels. THE STRUCTURE OF STEEL 19 FIG. 2. 0.06 per cent. Carbon. Approximately Pure Ferrite. X75. (Ord- nance Dept.) v > *-'V i - .V FIG. 3. 0.18 per cent, Carbon. Ferrite (White) and Pearlite (Dark). X75. (Ordnance Dept.) 20 STEEL AND ITS HEAT TREATMENT IG. 4. 0.32 per cent. Carbon. Ferrite (White) and Pearlite (Dark). (Ordnance Dept.) X75. JP IG 5, 0.49 per cent. Carbon. Ferrite (White) and Pearlite (Dark). X75. (Ordnance Dept.) THE STRUCTURE OF STEEL 21 0.57 per cent. Carbon. Ferrite and Pearlite. X75. (Ordnance Dept.) ^ M FIG. 7. 0.71 per cent. Carbon. Ferrite and Pearlite. X75. (Ordnance Dept.) 22 STEEL AND ITS HEAT TREATMENT FIG. 8. 0.83 per cent. Carbon. Pearlite. X485. (Ordnance Dept.) FIG. 9. 1.46 per cent. Carbon. Pearlite and Cementite (White), nance Dept.) X75. (Ord- THE STRUCTURE OF STEEL 23 Ferrite. Ferrite is soft, ductile and relatively weak. It has a Tensile strength of approximately 40,000 to 50,000 Ibs. per square inch, with an elongation of about 40 per cent, in 2 ins. Ferrite in itself has no hardening power as applicable to industrial purposes. It is magnetic and has a high electric conductivity. Its appear- ance under the microscope has been shown in the photomicro- graphs previously mentioned that is, as polyhedral crystals in the low carbon steels. Pearlite. As previously mentioned, the common occurrence of pearlite in slowly cooled steels is in the lamellar formation, is being FIG. 10. Laminated Pearlite and First Appearance (as Veins between Grains) of the Excess Cementite. XlOO. (Titanium Alloys Mfg. Co.) composed of alternate plates of ferrite (showing dark under the microscope) and cementite (showing white under the microscope). As will be shown later, under different rates of cooling pearlite may exist in other formations and dependent upon the relative arrange- ment of the ferrite and cementite of which it is composed; some of these various modifications are shown in Figs. 8 and 10. Normal pearlite, that is, interstratified bands of ferrite and cementite such as shown in Fig. 8, has a tensile strength of approximately 125,000 to 130,000 Ibs. per square inch, with an elongation of about 10 per cent, in 2 ins. 24 STEEL AND ITS HEAT TREATMENT Cementite. The properties of cementite are very little known with the exception of its great hardness and brittleness, which are a maximum. Free cementite, that is, unassociated with ferrite to form pearlite, probably does not have a tensile strength much greater than 5,000 Ibs. per square inch. Its ordinary occurrence in slowly cooled steels (carbon greater than 0.9 per cent.) is either as a network, such as we have seen, or as spines and needles. Pearlite and Cementite Pearlite Pearlite and Ferrite 3$. > * FIG. 11. Case-carburized Steel, Showing nearly Carbonless Steel (Bottom) Gradating into High-carbon Steel (Top). (Weyl.) Static Strength. We may now sum up these facts in their relation to the static strength of slowly cooled steel as follows: Free ferrite has a minimum tensile strength with maximum ductility; pearlite has a maximum tensile strength with low ductility; free cementite confers added hardness and brittleness, with a consequent lowering of the tensile strength. In other words, by increasing the amount of pearlite in the steel, we increase the static strength but with a THE STRUCTURE OF STEEL 25 corresponding decrease in the ductility. And as an increase in the amount of pearlite necessarily means an increase in the amount of carbon, the effect of increased carbon will give the same results. This is shown graphically in the diagram of Fig. 12. Heat Treatment. Heat treatment in general consists in chang- ing or regulating the structure of the steel by various methods of Per Cent. Elongation. 8 8 S Tensile Strength, Lbs. per Sq. Inch ]S 8 8 8 i jt> ^^" 4 Y/ '/ / fJNl <:::: ^^ << x *Y\ ' f/ Jl / \ / / / / > / ^ V > / / / / / / / X \ 'Carbon 0.2 0.4 0.6 0.8 1.0 1.2 1. Per Cent. Carbon. FIG. 12. Approximate Influence of Carbon upon the Strength and Ductility of Steel. heating and cooling. By the term " structure " is meant (1) the metallographic constituents, among which are those just described; (2) the size of the grain; (3) the net-work. In order to understand the nature of these changes and their application it will be necessary to have a clear understanding of the mechanism by which these changes are brought about. Critical Points. The nature of steel, as explained before, is complex. The structure of any particular steel may be modified 26 STEEL AND ITS HEAT TREATMENT or entirely changed by various degrees of heating, and all of which take place in the steel while it is in the solid condition. These structural changes take place at temperatures known as the " critical points " or " critical ranges " of the steel. These critical ranges are denoted by the letter "A," followed by the letter "c" (abbreviation for the French word " chauffage," signifying " heating ") or the 1800 / 1700 / 4 / 1 g 1600 \ . / \ J* /* ^3 Vf i ^< 5 s \ o b i r >oo V > a \ / g \ 4 3 A2 \ s 1 2 ^ ^. j/ A 1-2-3 1300 la - v lb 1200 1100 ) 0.20 0.40 0.60 0.80 1.00 1.20 1.40 Per Cent. Carbon. FIG. 13. Critical Range or Carbon-Iron Diagram. letter " r " (" refroidissement " or " cooling ") These signs, Ac or Ar, are further modified by the numerals 1, 2 or 3, indicating the particular point referred to. Thus Acl would mean the first critical range passed upon heating the steel beyond a certain temperature, and so forth. These critical points or ranges are indicated graphic- ally in Fig. 13. THE STRUCTURE OF STEEL 27 In considering this diagram let us devote our attention to a certain specific case, such as a low-carbon steel with about 0.2 per cent, carbon. We will also assume that the steel is in the normal condition resulting from slow cooling, in that it consists of about 25 per cent, pearlite and about 75 per cent, free ferrite. We will also first consider what is the influence which these changes occurring during the critical ranges have upon the constituents of the steel. In the first place, practically no change in the constituents occurs during heating until a temperature corresponding to the lower critical range, Acl, is reached, which is equivalent to about 1330 F. In passing through this critical range there is a complete change in the nature and structure of the pearlite, it being converted into an entirely new constituent with new characteristics. This is tech- nically known under the generic term of a " solid solution," micro- graphically called " Austenite." The excess ferrite remains unchanged. Solid Solutions. To understand better the nature of this new component let us consider the interaction between salt and ice. When these two substances are placed in contact with each other, we know that under suitable conditions of temperature the salt and ice merge into one another and so pass from the state of two separate substances or mechanical mixture into that of one separate substance or brine solution. A similar process takes place in the case of the pearlite, except that the resultant solution is solid instead of being a liquid. The individual plates of ferrite and cementite which have characterized the pearlite grains now merge into one another, form- ing this new substance or constituent, known as a solid solution. This new constituent, save that it is a solid and not a liquid, has all the properties of a liquid solution. Its original components are merged into a single entity, giving a complete indefiniteness of com- position, and with entirely new characteristics. Absorptive Power of Austenite. Just as the brine solution can dissolve more salt or ice with increased temperature, so this solid solution of iron and iron carbide possesses the power of absorbing more free ferrite or free cementite. Therefore, as the temperature is raised above that of the lower critical range (Acl), and there being an excess of ferrite in this particular steel (0.2 per cent, carbon), the solid solution or austenite begins to absorb this ferrite. This continues progressively with increased temperature until the upper critical range, Ac3, is reached, or, for this particular steel, a tempera- ture of about 1525 F. At this temperature the last of the remaining 28 STEEL AND ITS HEAT TREATMENT excess of ferrite is absorbed by the austenite, so that above the upper critical range of the steel the steel is composed entirely of austenite the solid solution. These changes are illustrated graphically in Fig. 14. It will be seen that the initial pearlite, comprising about 25 per cent, of the normal steel, changes into austenite (the solid solution) at a tempera- ture corresponding to that of the lower critical range, Acl, and then progressively absorbs the free ferrite until at a temperature corre- sponding to that of the upper critical range, Ac3, the whole steel con- sist^ of austenite. AcS Y Free Fcrrite FIG. 14. Change of Pearlite and Free Ferrite into Austenite during Heating. Carbon about 0.2 per cent. These same changes are shown microscopically in Figs. 15, 16, 17 and 18. The first photomicrograph shows the normal condition of the steel, being made up of a small proportion of pearlite (dark), and a large amount of free ferrite (light) . The three other structures were obtained by heating this same steel to temperatures above the lower critical range and then " fixing " the structure obtained at those temperatures by " quenching." Fig. 16 shows the structure representative of a temperature between that of the Acl and Ac2, the solid solution l (dark) having increased considerably in amount 1 Strictly speaking, the dark areas thus referred to as the " solid solution " are not austenite, but its transitional stage, martensite. In the ordinary carbon steels austenite as such cannot be retained by the ordinary methods of quenching THE STRUCTURE OF STEEL 29 over that of the original pearlite as in the previous figure. Fig. 17 represents the structure obtained at a temperature somewhat under that of the upper critical range, Ac3; in this case it will be noted that the solid solution covers nearly the whole field, there being but a small amount of the free ferrite (white). The structure representa- tive of heating to slightly above the upper critical range is shown in Fig. 18; it will be seen that the free ferrite has now been entirely absorbed by the solid solution. Also note the extremely refined FIG. 15. Normal Low-carbon Steel as Rolled. X60. (Bullens.) structure, as we shall have occasion to refer to this particular feature a little later. Allotropic Modifications of Ferrite. Associated with these critical ranges there is also a change in the allotropic 1 form of the ferrite (iron). Thus pure ferrite (as distinguished from the ferrite (as will be explained under the chapter on Hardening), but changes into martens- ite. Martensite, however, is also a solid solution, and for the purposes of explana- tion in this chapter in order not to complicate matters we will consider it permissible to use the term as indicated. 1 Sauveur defines allotropy as " suggesting marked and sudden changes in some of the properties of a substance occurring at certain critical temperatures, without any change of state or of chemical composition." 30 STEEL AND ITS HEAT TREATMENT FIG. 16. Low-carbon Steel Quenched between Acl and Ac2. X60. (Bullens.). FIG. 17. Low-carbon Steel Quenched a Little below Ac3. X60, (Bullens.) THE STRUCTURE OF STEEL 31 associated with cementite to form pearlite) in its normal condition is called " alpha "-ferrite or " alpha "-iron, and is characterized by extreme ductility and magnetic properties. Upon heating this alpha-ferrite to a little over 1400 F., corresponding to the critical range Ac2, the iron becomes practically non-magnetic and is then known as " beta "-ferrite or " beta "-iron. Upon further heating to a temperature above the upper critical range, Ac3, there is still another change in the allotropic modification of the iron, it being known as " gamma "-ferrite; this gamma-iron is slightly softer than the beta modification. Gamma-iron has the property of being able FIG. 18. Low-carbon Steel Quenched above A3. X60. (Bullens.) to dissolve carbon or iron carbide, a characteristic which is not held by alpha-iron. Merging of the Critical Points. Now by referring to the carbon- iron diagram in Fig. 13 it will be noted that at the eutectoid ratio of carbon, that is, at about 0.9 per cent, carbon, 1 the three critical ranges Al, A2, and A3, merge into one. That is, steels consisting of pearlite alone, when heated to a temperature beyond this point, will change directly into the solid solution austenite, which will consist of a solution of carbide (or carbon, according to some authorities) in 1 The eutectoid ratio on the chart is given as 0.85 per cent, carbon. Accord- ing to the authority selected this ratio will vary between 0.8 and 0.9 per cent, carbon; but the more recent tendency is to adopt 0.90 per cent. 32 STEEL AND ITS HEAT TREATMENT gamma-iron. Similarly, as normal pearlite always represents this eutectoid ratio, the same change of pearlite into a solid solution of carbide in gamma-iron will always occur at this temperature in ordi- nary carbon steels irrespective of the carbon content of the steel as a whole. Changes in Heating Different Steels. With this explanation clearly in mind, we may now refer back to the example of the 0.2 per cent, carbon steel and more fully explain the changes which take place in the constituents. Under normal conditions, this steel will consist of pearlite plus alpha-ferrite. Upon heating through the Acl range, the pearlite will change into austenite, the iron of which will be in the gamma modification; the free ferrite will still remain in the alpha condition. Upon further heating through the zone marked " 2 " on the diagram Fig. 13, the austenite will begin to absorb the free ferrite. Upon passing through the Ac 2 range the balance of the free ferrite will pass from the alpha modification into that of beta-f errite ; the steel as a whole will be hard and non-mag- netic. Upon further heating (zone 3) the remnant of the beta free ferrite will be gradually absorbed, so that on passing through the critical range, Ac3, the whole steel will be in the condition of austen- ite (zone 5), or a solution of iron carbide (or carbon) in gamma-iron. In a similar manner we might explain the changes in constituents which take place upon heating normal steel with any carbon up to that of the eutectoid ratio. With a carbon content somewhere between 0.3 and 0.4 per cent, (varying according to different author- ities) it will be noted that the A2 and A3 ranges merge into one, known as A2-3. In a manner analogous to the absorption of free ferrite by the solid solution in the hypo-eutectoid steels, the free cementite will be absorbed in the case of the hyper-eutectoid steels, the final solution taking place at a temperature range indicated by the line Acm. The only difference, and that a practical one, is that the solution of the free cementite takes place more sluggishly than the solution of the free ferrite of the lower carbon steels. The Ar Ranges. Corresponding critical changes take place upon cooling slowly from above the upper critical range, except that they occur in the reverse order and with opposite effect. On account of the molecular inertia, however, we find that these critical ranges (of cooling, Ar3, Ar2, Arl, etc.) are a number of degrees below the temperatures at which they appeared on heating. This difference is dependent upon length of exposure and the temperature to which THE STRUCTURE OF STEEL 33 the steel was subjected, the rate of cooling, and, more particularly, upon the influence of the alloying elements which may have been added to the steel. Some of the alloys, if present in sufficient amount, will cause the recalescent points to fall below normal temperatures, and are the basis of air-hardening steels and similar compositions. Changes on Slow Cooling. Upon slow cooling from above the upper critical range, the solid solution will commence to reject the excess ferrite (or, of course, the excess cementite in the case of hyper- eutectoid steels) as the temperature decreases from Ar3 to Arl. The reverse changes in the physical nature and properties of the iron occur at the critical ranges during cooling as those previously noted under heating. When the lower critical range is reached, the excess ferrite or cementite will have been entirely rejected, and as the steel passes downward through this range (or point), the solid solution now containing 0.9 per cent, carbon will change into pearlite. Under similar conditions of cooling, the original steel and the present heated and cooled steel will have the same structure. Refinement. Before leaving the subject of the influence which heating through these various critical ranges has upon the structure of the steel, there are a few points which we wish to mention briefly concerning refinement. Again assuming that the steel is in the normal condition, no change will take place in the structure until the temperature has been raised at least to that of the lower critical range. At this temperature the original pearlite grains are com- pletely changed and will possess that maximum refinement which the formation of the austenite can impart that is, complete refine- ment. If the steel has a carbon content other than that of the eutec- toid ratio (i.e., contains free ferrite or free cementite), the steel as a whole will not be refined; the excess ferrite or cementite will remain unaltered and the steel will retain its original grain-size. This is brought out by a comparison of Figs. 15 and 16. Complete refine- ment of the steel as a whole will not result until the steel has been heated to a temperature slightly over that of the upper critical range, as a comparison of Figs. 17 and 18 will prove, and as is evident from previous discussion. A clear understanding of these principles must be had, as they form the basis of many of the heat treatment pro- cesses which will be later developed. Grain-Size Beyond Ac3. As the temperature is progressively raised above the critical range, a gradual coarsening of the aus- tenite grains occurs. This increase size is not only a function of the temperature, but also of the length of time at which the high tern- 34 STEEL AND ITS HEAT TREATMENT perature is maintained. The practical application of the principles noted in this and the previous sections will be considered in the chap- ter on Annealing. Network. The third factor in the structural changes taking place upon heating is the effect of temperature upon the network. All hypo-eutectoid steels in the normal condition are made up of pearlite with a varying amount of excess ferrite, the latter decreasing with the increase in carbon content. From our study of the inter- FIG. 19. Microstructure of Cast-steel Ingot as Cast. X75. (Ordnance Dept.) Tensile Strength, 77,000. Elastic Limit, 39,000. Elongation, 10.5. Red. of Area, 16.9. nal mechanism by which the constituents of the steel are formed by slow cooling, we know that the pearlite forms the basis of the structure, the ferrite being rejected by the solid solution (pre-pearlite). Being thrown out to the boundaries of these austenitic grains, the excess ferrite forms a network around these grains. Upon reheat- ing, this network is gradually absorbed, its final absorption taking place upon passing the upper critical range. This change is similar to that explained previously under the description of the action of the excess ferrite. THE STRUCTURE OF STEEL 35 FIG. 20. Microstructure of Cast Steel Ingot Forged to 1450 F. X75. (Ord- nance Dept.) Tensile Strength, 83,500. Elastic Limit, 50,500. Elongation, 27.5. Red. of Area, 43.3. FIG. 21. Microstructure of Steel Subjected to Cold Work, and Showing Dis- tortion of Grain. X50. (Ordnance Dept.) 36 STEEL AND ITS HEAT TREATMENT FIG. 22. Hammer-hardened Steel, 0.46 per cent. Carbon. X300. (Savoia.) FIG. 23. Effect of Cold Rolling on 0.20 per cent. Carbon Steel. X60. (Bullens.) THE STRUCTURE OF STEEL 37 FIG. 24. Effect of Punching upon Structure of &-in. Chrome-nickel Steel Plate. Hole Downwards and at Right. X50. (Bullens.) FIG. 25. Machining Strains on Surface of Mild Steel. (Brearley.) 38 STEEL AND ITS HEAT TREATMENT The Effect of Work on Grain-Size. 1 Steel cooled slowly and undisturbed from a high temperature will show a coarsely granular or crystalline structure, and the size of the grain is a function of the temperature and time during which the material is held at the maxi- mum temperature, and the rate at which the material is cooled. In large masses of material the structure will be coarser in the center than at the surface, due to the difference in rate of cooling. In order to overcome this difference and at the same time produce a homogeneous, uniform material, the steel is worked during the period at which grain growth would ordinarily take place. Steel which has been hot- worked down to the Arl point will show a finer grain, and will be stronger than the same steel slowly cooled without work, and will at the same time show high ductility. Examples of steel worked and unworked are shown in the photomicrographs of Figs. 19 and 20. Steel which has been worked below the Arl range that is, cold- worked will show considerable distortion of grain, as is illustrated by Fig. 21, and may even become hardened, Fig. 22. Cold rolling frequently develops a weak, laminated structure, as is shown in Fig. 23. Even punching or machining operations may greatly affect the structure, examples of which are given in Figs. 24 and 25. 1 In part from Bulletin 1961, Ordnance Dept. CHAPTER III ANNEALING Annealing. Annealing, in its commercial application, may have for its purpose any or all of the following aims: (1) to " soften " the steel and thus put it in condition for machining or to meet certain physical specifications; (2) to relieve any internal stresses or strains caused by previous hardening or elaborating operations; (3) to obtain the maximum refinement of the grain in combination with large ductility. Thus, depending upon the results desired, commercial annealing will consist of a heating operation carried to some predetermined temperature although not necessarily over the critical range to produce the results desired in items 1 and 2 previously noted, and followed by a moderately slow cooling of the metal from that tem- perature. True or full annealing requires a heating to above the upper critical range of the steel. Elemental Considerations. In the abstract, annealing would appear to be but a suitable correlation of the following elements: 1. Rate of heating; 2. Temperature of heating; 3. Length of heating; 4. Rate of cooling. But in actual practice the success to be attained in annealing (or in any heat treatment process, for that matter) must depend upon the judgment and skill of the furnace operator in applying the basic principle's which may be derived from a consideration of the above factors. Thus it is the man who determines the manner of placing the charge in the furnace, of regulating the flow and composition of the hot gases, of determining when the steel has been uniformly and thoroughly heated, and similar fundamentals. For while it is advisable and perhaps necessary to understand the theory behind the actual work itself the " wherefore " the " wherewithal " is largely a personal equation and should be borne in mind throughout every theoretic discussion of principles or practice. 39 40 STEEL AND ITS HEAT TREATMENT Heat Application. The manner in which the steel is placed in the furnace is a factor of supreme importance. It may even be said that three different kinds of annealing may be produced in the same furnace operating at the same indicated pyrometer reading, dependent simply upon the method of placing the sto\;k in the furnace. Particular stress should be laid upon the necessity for getting heat to the center and bottom of the charge, not only for the sake of uniform annealing, but also to shorten the time of absorption and lessen the time of exposing the top and outside edges of the charge to the heat and influences of the chamber atmosphere. Thus it has been shown that a uniform chamber temperature does not necessa- rily mean a uniformly annealed product; that a circulation of heat through the mass is more desirable than the mere application from the outside; that, with the same chamber uniformity, it is possible to vary the quality of the anneal by the manner in which the stock is placed in the annealing zone. It is advisable to raise the charge above the furnace floor or hearth upon suitable blocks or supports, to separate each piece from the other, and to avoid localized heating through over-loading. It is only by such means that there will be provided an opportunity for the circulation of the hot gases through the charge. Pre-Heating. Slow, careful and uniform heating is always advisable regardless of the chemical composition or physical condi- tion of the steel. Heating to such temperatures as are common in general annealing practice necessarily results in more or less change of physical condition or molecular readjustment, and the greater the hardness, brittleness and amount of internal strain in the metal, the greater will be the deleterious effect of such heating. Thus objects of intricate design, or with varying cross-sections, or steel in a hard, brittle condition, should be given the greatest care in heat- ing in order that the release of any strains shall not cause warpage or otherwise injure the metal. Such pieces should never be placed directly in a hot furnace, but should be given a careful pre-heating. ANNEALING HYPO-EUTECTOID STEELS Microscopic Changes. In the previous chapter we have explained that, in the ordinary cast, rolled or forged sections (pearlitic in character), there is virtually no change in grain size or in constit- uents during the heating to a temperature below that of the lower critical range Acl. That is, there is no refinement of the steel. ANNEALING 41 As the temperature passes the Acl range there occurs the complete change of the pearlite to the solid solution, giving the maximum refinement to the austenite. Passing through zone 2 (Refer to Fig. 13) the excess ferrite is progressively absorbed by the solid solution, causing an apparent decrease in the grain-size of the steel as a whole. This absorption is the slower the greater the carbon until the carbon nears 0.85 per cent., but is offset by the fact that the amount of free ferrite decreases as the eutectoid ratio is approached. Upon passing through the critical range Ac2 we have the forma- tion of beta iron with no apparent change between the relative grain-size of the alpha ferrite and beta ferrite grains. The same absorption of the excess ferrite continues progressively, but with increased sluggishness (due to the supposed properties of beta ferrite) . This applies to steels with say 0.12 to 0.30 per cent, carbon. In the very low carbon steels Howe 1 sums up the probable changes during this period in a provisional proposition that t>) if initially fine-grained the steel coarsens, though only very slowly; (b) if initially coarse-grained it refines slowly; (c) to coarsen again upon long exposure to these temperatures. The changes taking place through zone 2 continue through zone 3, although more slowly. If the rate of heating through this range of temperatures is comparatively slow, there will be a complete absorption of the remaining ferrite just before Ac3 is reached. Under ordinary circumstances final absorption will occur on passing through the Ac3 range. The Upper Critical Range. As the steel passes the upper critical range there is the complete refining of the grain, it becoming very fine and almost amorphous. As the temperature increases beyond this range the grain-size coarsens, causing a diminution in the strength of the steel. The effect upon the physical properties of the steel is great. The tensile strength is increased somewhat as the temperature advances. The elastic limit rises until a point is reached about 175 to 200 F. over the upper critical range, after which it then decreases. The elongation and reduction of area decrease very rapidly. These changes in the physical properties are shown graphically in Fig. 26 in which the results obtained by heating a 0.40 per cent, carbon, basic open-hearth steel to a definite temperature and then slow-cooling with the furnace are plotted 1 H. M. Howe, "Life History of Network and Ferrite Grains in Carbon Steel," Proc, A, S, T, M,, Vol. XI, 1911. 42 STEEL AND ITS HEAT TREATMENT against the temperatures. It will be noted that the softest and most ductile steel is obtained at approximately 1475 F., which is about 50 over the upper critical range. Heating Over the Upper Critical Range. The effect of heating beyond the critical range is well developed by the series of photo- graphs (by Howe) shown in Figs. 27, 28, 29, 30 and 31. The steel (0.40 per cent, carbon, 0.16 per cent, manganese) was heated to the 1,200 1,300 1,400 1,500 1,600 1,700 1,800 Degrees Fahrenheit FIG. 26. Effect of Annealing Temperature on Physical Properties. temperatures indicated, held at those temperatures for ten minutes, and then cooled in air. There is a difference in grain-size between that cooled from 1472 F. and from 1652 F., showing that anneal- ing should never be carried very far beyond the upper critical range or Ac3 point unless for special reasons. As the high temperatures were successively raised to 1832 and 2012 the grain-size becomes noticeably larger, until at 2192 the steel is " burnt." These photomicrographs also exhibit the effect of air cooling upon the structure, in that it develops a distinct net-work or cellular structure. The effect of heating beyond the upper critical range is also brought ANNEALING 43 EFFECT OF HEATING BEYOND Ac3, 0.40 per cent. Carbon Steel Heated at Temperatures Indicated for Ten Minutes and AIR COOLED. FIG. 27. 1472 F. X40. (Howe.) FIG. 28. 1652 F. X40. (Howe.) FIG, 29, 1832 F. X40. (Howe.) FIG. 30. 2012 F. X40, (Howe.) FIG. 31. 2192 F. X40. (Howe.) 44 STEEL AND ITS HEAT TREATMENT out in an analogous manner by Figs. 32, 33, 34, 35 and 36, except that in this case the steel has been cooled very slowly (furnace cooled) from the specific temperatures. Use of the Microscope for Checking Structural Changes. From previous theoretical discussion, it is evident that in order to fulfill the true or full annealing operation, it is necessary to heat the metal to over the upper critical range of the steel in order to obtain the com- plete change of structure with the smallest grain-size possible. The microscopic changes which take place during such heating of a 0.28 per cent, carbon steel are given as follows: Temperature. Structure. 1325 F. Very coarse ferrite and pearlite similar to the original bar. 1375 F. Laminae of ferrite strong, ground-mass refined. 1425 F. About 25 per cent, ferrite laminae left. 1475 F. Trace of coarse ferrite unabsorbed. 1500 F. Complete refining. No coarse ferrite. 1550 F. Structure similar. 1650 F. Refined but grain-size coarsening. These experiments l were carried out with a view to discover the cause of failure of an eye-bar (carbon 0.28 per cent.) when placed in service. The original steel had been annealed several times at temperatures under the upper critical range, but a microscopic study showed that these heatings had simply refined the pearlitic ground- mass. In other words, it was found that the proper annealing temperature necessary to obtain a completely refined steel was beyond the upper critical range. In this steel it would seem to be about 1500 F. The lower critical range is shown by the refining of the ground-mass which occurred between 1325 and 1375. Diffusion. We have repeatedly stated that complete absorption of the excess ferrite takes place at the upper critical range of the steel. Although this statement is true, there is another phase of this absorption to be considered, and a full understanding of which will probably clear up many of the questions which have perplexed those unfamiliar with the theory of annealing. This phenomenon may be called " diffusion." Let us hark back to our former simile i Wm. Campbell, " Further Notes on the Annealing of Steel," Proc. A. S. T. M., Vol. X, 1910. ANNEALING 45 EFFECT OF HEATING BEYOND Ac3. 0.40per cent.'Carbon Steel Heated at Temperatures Indicated for Ten Minutes and FURNACE COOLED. FIG. 32. 1472 F. X40. (Howe.) FIG. 33. 1652 F. X 40. . (Howe.) FIG. 34. 1832 F. X40. (Howe.) FIG. 35. 2012 F. X40, (Howe.) FIG. 36. 2192 F. X40. (Howe.) ^46 STEEL AND ITS HEAT TREATMENT of the salt and brine solution. When a grain of salt is dissolved by the brine, it is the solution in the immediate neighborhood of the salt crystal which acts as the solvent and not the entire volume of the brine solution. In time, however, the dissolved salt will eventu- ally diffuse through the whole body of brine and the brine will then be of equal composition throughout. Now a similar process is going on in the steel when the solid solution (austenite) is absorbing the excess ferrite, and it will be found that complete absorption may not mean complete diffusion or equalization. The process of equalization goes on with the rise in temperature. If the passage through tem- peratures under that of the upper critical range is only slow enough, a large part of the diffusion will have occurred by the time Ac3 is reached. In order that there may be complete diffusion, and there- fore complete grain-refining, the sojourn at a temperature approxi- mating Ac3 must be long enough for this complete diffusion of the absorbed excess ferrite and therefore of the solid solution. Although exposure to a higher temperature would naturally hasten this diffu- sion, it would be at a cost of coarsening the austenite grains. The effects of non-equalization will be discussed in a later part of the chapter. Rate of Heating. Studying the rate of heating from the practical aspect there is also another factor to be considered that of bringing the whole mass of the steel to the proper temperature evenly. It is self-evident that the center of a large mass of steel, such as loco- motive axles or steel blooms, will lag in temperature behind the exterior. In other words, it is the tendency of the core to be con- siderably lower in temperature than the shell or outside of the steel. It is then a common procedure to raise the temperature of the fur- nace beyond the proper annealing heat in order to drive the heat to the center of the piece to be annealed. This is a great mistake. It is far better to take the extra time required to heat more slowly as the proper temperature is neared, thus bringing the steel to an even temperature throughout. If this were not done, the exterior of the piece might be carried beyond the proper temperature and, in general, a needlessly high temperature is injurious and tends to recoarsen the grain. Expressing this question in a different way, we may say that the furnace in which the metal is being heated for annealing should in no case be run at a higher indicated temperature than the maximum temperature to which the metal itself is to be heated. To illustrate: a piece of steel heats, cools and decarbonizes on the corners first. ANNEALING 47 The life of the entire piece of steel is no greater than the life of the corners. If the steel is placed in a hot furnace, the corners are apt to be heated long before the major part of the mass. If the temper- ature is high, the corners are overheated before the center of the mass is saturated. From this commonplace example there should be indicated the necessity for slow, soaking heats in order to prevent overheating the corners of the metal, and further, the necessity of soft hazy heats to prevent oxidation or decarburization of the exposed edges. Temperature of Heating. Assuming that the proper degree of care has been used in heating the steel, the next question is the degree of heat necessary. Reduced to lowest terms, the true or full anneal- ing operation requires the production of an entirely new crystalline structure, the constituents of which shall be of the smallest grain- size attainable; this operation should also eliminate all internal strains and stresses. As previously described, this new structure is given birth at a temperature known as the " upper critical range " of the steel. The exact temperature 1 will depend upon the chemical composition of the steel, and, more particularly, upon the carbon content. As this transformation does not occur suddenly, but usually covers a range of some 25 to 50 it is customary to adopt a temperature of about 50 over the upper critical range as the proper annealing heat. For straight-carbon steels these may be roughly given as shown in the chart in Fig. 37. The upper critical range is approximately located by the dash line on the chart. The temperatures recommended by the American Society for Testing Materials 2 are as follows : Range of Carbon Content. Range of Annealing Temperature. Less than 0.12 per cent. 0.12 to 0.29 0.30 to 0.49 0.50 to 1.00 1607 to 1697 F. 1544 to 1598 F. 1499 to 1544 F. 1454 to 1499 F. 1 Methods for determining the critical ranges are described in Chapter XVIII. 2 It will be noticed that the temperatures recommended by A. S. T. M. are distinctly higher especially for the tool-steel grades than those advised by the author. In the light of my own experience, and that of others, I believe that the lower the temperature which can be used to give the desired results, the greater will be the maximum efficiency of the annealed steel. 48 STEEL AND ITS HEAT TREATMENT Length of Heating. Ordinarily the underlying practice of this part of the operation is to heat the steel until the whole mass has been heated uniformly throughout at the proper temperature. This will of course depend upon the size of the object. This full heating is generally sufficient to give birth to the new grain structure and relieve all internal stresses. The proper rate of cooling should then maintain the steel in that condition. If the steel should be quenched in some hardening bath such as oil or water, this new grain-size and rear- iroo 1200 o.i 0.3 0.4 0.5 0.6 Carbon Content, Per Cent. 0.9 FIG. 37. Annealing Range for Carbon Steels, rangement of the structure would be kept. The annealing operation should theoretically bring about approximately the same results as to grain-size, neglecting for the moment the effect of slow cooling through the transformation ranges. From the standpoint of prac- tice, however, much difficulty is experienced in this regard, par- ticularly in cases where the mechanical work upon the steel has been severe, and also in alloy steels. It seems that the greater the internal stress upon the steel the greater is the amount of intermolecular lag or final release of this ANNEALING 49 stress behind the actual change of constituents. That is, even though a totally new structure may be set up by the annealing temperature, there remains for a considerable length of time a tendency of the new FIG. 38. Frame Steel as Rolled. X60. (Bullens.) structure to return, upon slow cooling, to the stressed condition of the original, even though the constituents themselves may be those born at the new temperature. FIG. 39. Frame Steel Partly Annealed. X60. (Bullens.) This point is illustrated in Figs. 38, 39 and 40. These are photomicrographs taken from tests made upon chrome nickel steel plates for automobile frames : Fig. 38 shows the structure of the steel 50 STEEL AND ITS HEAT TREATMENT as rolled; Fig. 39 shows the steel after a short annealing at a tem- perature above the upper critical range; and Fig. 40 shows the same steel after a long anneal at the same temperature. It will be noticed that the steel in Fig. 39 has taken on approximately the same struc- tural constituents as in the fully annealed piece as shown in Fig. 40, but that it still remains in the "stressed condition of Fig. 38, even though the annealing temperatures were the same in both cases. It is important, therefore, if a soft steel, free from all internal strains and stresses is desired, that a sufficient length of time be allowed for the permanent elimination of these intermolecular strains, before and after cooling. In the case of the steel plate just referred to it re- FIG. 40. Frame Steel Fully Annealed. X60. (Bullens.) quired some twelve hours for the complete change or equalization to take place! " Milky-Ways." We have previously explained this same phenomenon under the heading of " Diffusion," as this is the scientific principle underlying it. The reoccurrence or reformation of these laminations or other stressed structures is due to the fact that the complete effacement by equalization had not taken place. In other words, it means that where these stressed areas occur the carbon content as a whole is less than in the rest of the mass. Where ferrite predominates, as in the lower carbon steels, there will the mass more easily coalesce into what may be termed " milky-ways " (Howe). In order to equalize the steel as a whole the length of time of the sojourn at or slightly above Ac3 should be inversely proportional to the time occupied in reaching that temperature. ANNEALING 51 Alloy Steel. Alloy steel is particularly an example of the re- tarded transformation as described above, although the author has repeatedly found it in carbon steels cold worked. Most notable of the alloy steels exhibiting this peculiarity are chrome, chrome- nickel and chrome-vanadium steels. Many users and even manu- facturers of these steels contend that annealing will not give entirely satisfactory results. The oft-encountered " hard-spots " would seem to bear out this dispute. From the experience of the author the proposition develops into a simple question of time. The alloy- ing metals add to the density of the grain, so that a longer time is needed to complete the change in entirety. It was found that a certain 3-inch rolled-round approximating 0.50 per cent, carbon, 1.50 per cent, nickel and 0.50 per cent, chrome required sixteen hours for this complete change, together with the elimination of hard- spots, to take place ; high-carbon high-chrome steel often takes days for a complete anneal. Time of Heating. Assuming a proper rate of heating, it therefore remains to determine the required length of heating by means of experimentation, taking into consideration such points as have been mentioned above. Along these lines some interesting experiments have been carried out by Mr. M. E. Leeds l for the determination of the variations in rates of heating of specimens of different sizes to various furnace temperatures and which in some degree answer the oft- repeated question " How long shall we heat this piece of steel? " The experiments were made with round specimens of nor- mal open-hearth carbon steel approximating 0.5 per cent, carbon, and ranging in size from 2 ins. to 12 ins. in diameter, by 24 ins. long. Each specimen was heated to four temperatures, namely, 1000, 1200, 1400, and 1600 F. During the time of heating a continuous record was kept of furnace temperatures, the temperature of the sur- face of the specimen, and of one to three points in its interior. While the results obtained are necessarily of relative value only on account of the varying furnace conditions which might be found elsewhere, there are, nevertheless, several interesting conclusions of value which were drawn from these experiments: 1. "Variation in Time of Heating with Size. As would be expected, the smaller specimens heat more rapidly than the larger. In curves (Fig. 41) the relation between the size of specimen and time of heating to various temperatures are brought out. Except in a very general way, this information could not be used as a guide 1 M. E, Leeds, A. S. T. M., June Meeting, 1915. 52 STEEL AND ITS HEAT TREATMENT to heating practice, as the rates would vary with the size of furnace and probably with other conditions. 2. " Relation between Time of Heating and Furnace Tempera- ture. The time of heating for a specimen of any size is less when it is brought up to 1600 F. than when brought up to 1200 F., and less I 5 o I 300 X / 4" 8" Size of Section. FIG. 41. Curves Derived from Rate of Heating. thrup Co.) (Courtesy of Leeds & Nor- F. for 1200 F. than for 1000 F., although it is greater for 1400 than for any other temperature. " It is more difficult to account for the fact that the higher temperatures are attained more rapidly than the lower ones. This fact, however, appears to be clearly demonstrated. It may be that ANNEALING 53 the specimens received a large amount of their heat by radiation from the furnace walls. The heat transfer by radiation between two bodies at different temperatures is proportional to the difference between the fourth powers of their absolute temperatures, and so for a 100 dif- ference in temperature between furnace wall and test specimens, at 1600 F., the heat transfer would be at a higher rate than for the same temperature difference at lower temperatures. 3. " Relation between Surface and Interior Temperatures. From all of the curves, it is deduced that there is no large difference in temperature of the points inside of the specimen. This was quite surprising, as it was expected that the 12-in. specimen would show considerable differences of temperature between a point 2 ins. from the surface and the center. " All of the runs show that the contact couple is at a higher temperature than any of the interior couples until the specimen has attained the temperature of the furnace. It cannot properly be assumed that the temperature shown by the contact couple is exactly that of the surface of the specimen. " When the contact couple attains the furnace temperature, all parts of the specimen have also attained that temperature. This suggests a practical method of using contact couples in conjunction' with furnace couples, namely, by means of the furnace couple the furnace should be held at the temperature at which it is desired to treat the specimen, and the contact couple should then be used to determine when the specimen has assumed the desired temperature. 4. " Contact Couple Shows Time of Transformation. The curves (not given here) showing the heating of the 12-in. and 8-in. specimens to 1400 and 1600 F. show that the transformation point is clearly shown by the couples inside of the specimen, and that it is also shown by the contact couple. The interior couples show, with approximate correctness, the temperature at which the trans- formation takes place. The contact couple shows a corresponding flexure in its curvature, at the same time as the interior couples, though not at the same temperature." The close correspondence in time between the flexures of the contact couple and the interior couples points to what Mr. Leeds believes is an important new method of determining when a piece of steel has been heated through its transformation point. Rate of Cooling. We know from our study of the previous chapters that in hypo-eutectoid steels the solid solution rejects the excess ferrite upon cooling through the transformation range. This 54 STEEL AND ITS HEAT TREATMENT ferrite will form either a network around the grains of solid solution or pearlite, or will coalesce into irregular masses, the same being dependent upon the rate of cooling. A moderately slow cooling will develop the cellular or network structure without breaking it up. A very slow cooling will break up the network structure, giving ample time for the ferrite to coalesce into large masses. The slower the cooling through the transformation ranges, the greater also will be the size of the grains. Effect of Cooling. The effect of the varied rate of cooling is illustrated in the photomicrographs of a 0.45 per cent, carbon steel FIG. 42. Network Structure, 0.45 per cent. Carbon Steel. X100. (Bullens.) shown in Figs. 42, 43 and 44, all taken at the same magnification. All three pieces were heated to a temperature somewhat in excess of the full annealing temperature. The steel of Fig. 42 was cooled quite rapidly (air-cooled) ; that of Fig. 43 was cooled rapidly through the upper part of the transformation range, but slowly through the lower critical range ; that of Fig. 44 was cooled with the furnace. Thus we have the network structure in the first case, showing a comparatively small grain-size. In the second instance the network is coarse and the pearlite is fairly well developed. A very slow cooling, as in the third case, has resulted in a coalescence of the ferrite into large grains, intermingling with the coarse pearlite. The ferrite in all three photomicrographs is represented by the white constituent. ANNEALING 55 Some very interesting facts might be drawn from a study of these photomicrographs in comparison with those previously mentioned in the series of Figs. 27 to 31, and Figs. 32 to 36. FIG. 43. Coarse Network Structure, 0.45 per cent. Carbon Steel. (Bullens.) X100. FIG. 44. Coalesced Ferrite and Pearlite, 0.45 per cent. Carbon Steel. (Bullens.) xioo. 56 STEEL AND ITS HEAT TREATMENT Effect of the Rate of Cooling upon the Pearlite. Not only does the rate of cooling from the annealing temperature have a very great effect upon the network and grain structure, but also upon the char- acteristics of the pearlitic constituents of the steel. The rate of MlCROSTRUCTURE. SEGREGATION STAGES. I. Sorbite or " sorbitic pearl- ite." Cementite emulsi- fied. MECHANICAL PROPERTIES. Tensile strength about 150,000 Ibs. per sq. in. Elongation about 10% in 2 ins. II. Sorbite passing into nor- Tensile strength about 125,000 malpearlite. Semi-segre- Ibs. per sq. in. gated cementite. Elongation about 15% in 2 ins. III. Finely laminated pearlite. IV. Laminated pearlite. Com- pletely segregated ce- mentite. Tensile strength about 100,000 Ibs. per sq. in. Elongation about 10% in 2 ina. Tensile strength about 85,000 Ibs. per sq. In. Elongation about 8% in 2 ina. Cementite white Ferrite black. V. Laminated pearlite pass- Tensile strength about 75,000 ing into massive pearl- Ibs. per sq. in. i t e . Cementite and Elongation about 5% in 2 ins. ferrite each coagulating FIG. 45. Pearlitic Development. cooling through the lower critical range, at which the transformation of the solid solution into pearlite is effected, will so change the arrangement of the ferrite and cementite constituents of the pearlite that widely varying physical results may be obtained in this manner. ANNEALING 57 As we will explain later, the austenUe does not directly change into pearlite, but passes through a series of transition constituents with varying physical properties. The majority of these, however, are not retained in the steel through methods of cooling other than quenching (which may or not be followed by a reheating), so that we need consider only the very last transition, sorbite. This com- ponent sorbite represents the last stage of the transition austenite to pearlite, and in which the individual particles of ferrite and cementite are just on the verge of coalescing. Sorbite, or sorbitic-pearlite, is noted for its combination of high tensile strength (i.e., in comparison with the later phases of pearlite) and ductility. Sorbite is generally formed by air cooling through the lower critical range, and is shown in Fig. 45. This figure also illustrates the different phases of the pearlite, together with their approximate physical characteristics. From this it will be evident that the rate of cooling must necessarily have a great influence upon the physical properties of the slowly cooled or annealed steel, and that the operation must be adjusted accordingly. Definite Cooling. Thus we see that having obtained the per- manent release of all internal strains and stresses and brought about the formation of an entirely new grain-size and structure by means of proper heating, it now remains to adjust the physical properties by means of regulating the rate of cooling. In general, there are three methods of cooling as used in the annealing process. These are: (1) Cooling in and with the furnace, (2) removing the steel from the furnace and covering with some blanketing substance such as lime, sand, ashes, etc., (3) cooling in air. Cooling by means of quenching is not a true annealing operation, and will therefore be considered under the subject of " Hardening." Furnace Cooling. Cooling in and with the furnace will generally give the slowest cooling of the steel which is possible if the furnace is of heavy construction and can be tightly closed. Furnace cooling will give a maximum " softness " and ductility that is, the tensile strength and elastic limit will be at a minimum, and the elongation and reduction of area will be large. Steel in this condition will be in a suitable condition for ordinary machining, and will also have the quality of resisting a small number of severe distortions. Slow Cooling. In cases where the objects are of large size, an approximation of furnace cooling may be obtained by removing the steel from the furnace and covering with some blanketing substance and slow conductor of heat such as lime, sand or ashes. This will 58 STEEL AND ITS HEAT TREATMENT also permit the recharging of the furnace for another heat without loss of time. Pit Annealing. Where a large tonnage of steel must be annealed, a pit lined with brick or concrete and suitably fitted with cover plates is sometimes made. The hot steel is immediately delivered from the annealing furance to the pit and covered with ashes. Cool- ing by this method of pit-annealing is often slower than cooling in the furnace itself if the latter is not properly constructed so that no cold air can find its way in. Size, of Object. It is readily realized that the size of the object has a great bearing upon the rate of cooling. Under the same con- ditions a smaller object will cool much more rapidly and will there- fore be harder and less ductile than a piece of considerable size. The rate of cooling must therefore be proportioned to the size of the object. Air Cooling. If the steel is removed from the furnace and allowed to cool in air the physical properties will be proportional to the dimensions of the piece and also dependent upon the carbon content. Thin objects and those with high carbon content cannot stand so rapid a cooling as thick and low carbon ones, lest their ductility be too greatly sacrificed. In this regard the American Society for Testing Materials recommends the following: " Thick objects with less than 0.50 per cent, of carbon may be cooled completely in air, of course protected from rain or snow. Objects with 0.50 per cent, of carbon or more, and thin objects with from 0.30 to 0.50 per cent, of carbon may be cooled in air if their cooling is somewhat retarded, as, for instance, by massing them together, as happens in the case of rails." This more rapid cooling will give great strength and high elastic limit, but less ductility. Combination Air and Furnace Cooling. Besides the regular air or furnace cooling there are a number of different combinations of the two which have given great success in innumerable cases. We will give them briefly as follows: 1. Heat to slightly over Ac3, air cool to just over Arl, return to a furnace which is held at that temperature (about 1350 F.), heat until uniform, and then cool slowly. The latter heating should not be any longer than is possible. This method will tend to prevent the formation of large amounts of free ferrite, but will affect the pearlite, as there will be slow cooling through the Arl range. (f* 1 *^. Heat to slightly over the Ac3 range, air cool to just under the i Arl range, return to a furnace and heat at 1350 F. and slow cool. ANNEALING 59 This method will effect a greater " toughening " if the temperature has not been prolonged too greatly at the second heating. & 3. Heat to slightly above Ac3, air cool to below Arl, return to a furnace heated at a temperature slightly below Arl (about 1200 to 1250 F.), hold at this temperature until uniformly heated, and slow cool. In fact, the last cooling may be made in the air if desired, as there will be little or no change in cooling from under the lower Critical range. Fine Grain Annealing. Of the three special methods given, the third is the preferable, as well as the most uniform and certain in results. By permitting the steel to air cool to a temperature below the lowest transformation, advantage is taken of any " hardening effect " or retardation in the transformation of austenite into a conglomerate of pearlite and ferrite. This effect will increase with the percentage of carbon and the smaller the size of the piece. The reheating to a temperature below the lower critical range, if not prolonged, will neither change the grain size nor allow, of the coalescing of the excess ferrite or of the individual constituents of the pearlite, but will form a mass of irresolvable and intermixed pearlite and ferrite known as " sorbite." At the same time, however, it will give the maximum combination of large ductility, good strength and excellent machining properties. This method is of particular value in the annealing of tool steels, in which it has given most excellent results. The main objection to both of the other two methods is that a varying duration of heating above the lower critical range will cause corresponding changes in the results, so that no absolutely definite result capable of commercial duplication can be obtained. The methods, however, find many applications in a general way, particu- larly in steels of medium carbon content which have been severely stressed by previous mechanical elaboration. The second method especially will give the advantage of having had a double heating through the lower critical range (besides the minimum grain-size conferred by fairly rapid cooling from the upper critical range), and thus breaking up the previous structure. Double Annealing. The next variable which may be used is that of heating to a temperature considerably in excess of the upper critical range, air cooling to under the lower critical range, and reheating to slightly above the lower or upper critical range. As; an example of this the author will cite a case which was success- fully solved by this method. Certain medium-carbon steel plates 60 STEEL AND ITS HEAT TREATMENT had been finished at a temperature considerably under the proper temperature for hot-rolling and thus had been considerably stressed in fact, the ordinary annealing method would not relieve this con- dition. The plates were put in a furnace with a car-bottom, heated thoroughly at about 1700 F. (that is, considerably over the Ac3 range), air cooled until black, and then reheated and slow cooled from a temperature slightly over the lower critical range. The laminations occasioned by the rolling were entirely eliminated by the high temperature, and their reformation prevented by the rapid cooling in air. The second anneal then thoroughly softened the steel and put it in good condition for the following forming opera- tions. This steel might, of course, have been reannealed at the Ac3 range (instead of the Acl range) and an even better product obtained. Tool Steel Annealing. The annealing of hypo-eutectoid tool steel may be broadly grouped under two headings, dependent upon the initial condition of the steel and upon the results desired. Tool steel which has been carefully hammered is undoubtedly strength- ened by this mechanical elaboration; a full annealing that is, heat- ing at a temperature over the critical range will entirely destroy the results of the forging operation. If it is therefore desired simply to anneal the steel in order to put it in suitable condition for machine work that is, to soften it and at the same time to retain the bene- ficial effects of the forging the annealing operation should be carried out at a temperature less than that of the critical range, or in the neighborhood of 1200 to 1250 F. On the other hand, if it is desired to obtain the finest grain size possible, the maximum softness, and to entirely eliminate any previous heating or forging work, the annealing should be carried out at a temperature slightly over that of the critical range, or in the neighborhood of 1400 F., dependent upon the composition of the steel in question. k*. Protection of Steel. One of the vital points in obtaining a satis- factory steel after annealing is the protection of its surface. Steel when heated beyond a low-red heat exhibits a great tendency to oxidize or scale, this action increasing, in the presence of oxygen, with the temperature and the length of time involved. This con- dition will exist in furnaces operated so as to produce sharp heats, instead of soft, slightly hazy, reducing atmospheres. Decarburiza- tion to a depth of J to \ inch is not a rare occurrence where improper combustion and heat application is the rule. If, due to poor furnace design and worse operation, such conditions do exist ANNEALING 61 produce a clean surface it will be necessary to protect the exposed surface of the steel in some manner. Tool steel is often annealed by placing in a tube, packing carefully with charcoal, and then clos- ing the ends of the tube with caps or luting with clay. On the other hand, the prevention of oxidation or the scaling of the metal during the heating process is a simple thing with the proper furnace design and operation. Assuming such a design, if the furnace is operated so as to produce soft, hazy heats such as we have previously mentioned, there should be no occasion for packing the steel in charcoal or other such substances. This statement is made not as one of theory, but as one of actual practice. Under-fired fur- naces are being run to-day on brass cartridge case work in which there is less oxidation and decolorization of the metal than in other fur- naces in which the metal is packed in charcoal; not only is a better product being obtained, but at less operating copt. \~ Box-Annealing. For the protection of larger masses or a number of smaller pieces, " box-annealing " is often resorted to. This par- ticularly applies to cases where a finished surface must not be injured. The steel is placed in a rectangular pot or box made of cast iron or of plates riveted together. This box may or not be lined with some refractory substance such as silica brick. The metal is then carefully packed with some material such as ground mica, sand, charcoal, charred bone or leather, lime, etc. If the steel is low in carbon a carbonaceous or carbon monoxide generating sub- stance must not be used, for a slight case-hardening action would take place. In the case of higher carbon steels, and especially of tool steels, reducing agents may be used, although it is better to mix the charcoal with clean ashes. Sand and ground mica are probably the most satisfactory of the simple non-reducing, refractory materials. The cover is then placed on the box and' the box with its contents is charged into the furnace and given the proper degree and duration of heating.. The box should be raised from the floor of the furnace so that the hot gases may have opportunity for circulation around it. When properly heated throughout, the box may be removed from the furnace and allowed to cool to atmospheric temperature. Stead's Brittleness. We have previously stated that practically no change occurs below the Acl range if no previous hardening of the steel has taken place. The one exception is that of very low-carbon steels and is due to the fact that steels very low in carbon behave more like pure or carbonless iron, there being but small percentages of cementite (and therefore pearlite) to influence the grain-size. 62 STEEL AND ITS HEAT TREATMENT Upon heating such steels through the upper part of zone la (refer to diagram in Fig. 13), a distinct coarsening of the ferrite grains occurs, this being a function of time as well as of temperature. Steels of such carbon held at say 1100 F. for a considerable length of time will develop such coarsening of grain-size as to make the steel unfit for commercial use if any degree of strength is required. This phenomenon is known as " Stead's Brittleness." With steels of greater carbon content the increased pearlite so operates upon the molecular structure of the steel that practically no change occurs until the Acl range is reached. ANNEALING HYPER-EUTECTOID STEELS Critical Ranges. Strictly speaking, hyper-eutectoid steels have two critical ranges: the A 1.2.3, at which on heating the pearlite changes into the solid solution; and the A cm range, at which on heating there is the final solution of the excess cementite just as in hypo-eutectoid steels the Ac3 range represents the solution of the last of the excess ferrite. However, on account of the relatively small proportion of free cementite in the ordinary hyper-eutectoid steels, and also because there is a large increase in grain-size upon heating to the Ac. cm range the temperature position of the latter increasing very rapidly with increase in the carbon content the Ac. cm range requires .but little practical consideration and the majority of the annealing operations are more intimately connected with the principal critical range Ac 1.2.3. Commercial Annealing. Similarly to hypo-eutectoid steels, the annealing of high-carbon steels may have for its object any or all of the following factors: (1) the release of internal strains and stresses set up by previous operations, (2) the softening of the steel to place it in a suitable condition for machining, (3) the entire change of structure. The first item may be accomplished by a simple reheating at temperatures below those of the critical range. The second and third items are more complex in their solution, as the form in which the excess cementite may exist is one of the governing factors. If the mass of the steel is in the sorbitic state, as may generally be expected in the usual tool steel, satisfactory results (the softening of the steel for machining, and relieving the internal strains) may be obtained by an annealing at a temperature slightly under that of the principal critical range, or at about 1250 to 1300 F. This heating ANNEALING 63 should not be prolonged for such length of time as may cause the excess cementite to coagulate, but only until the steel has been thor- oughly and uniformly heated throughout. On the other hand, if it is desired to obtain the complete change of structure, and to refine the grain (previously coarse), it will be necessary to heat to a temperature at least in excess of the Ac 1.2.3 range (about 1340 F.). For steels with a carbon content approx- imating 0.9 per cent., such heating will accomplish the complete change of structure and give the finest grain-size obtainable through annealing. For steels with a carbon content considerably in excess of the eutectoid ratio the annealing may be done at similar tempera- tures, provided, however, that the excess cementite is more or less in solution in the sorbite. Incidentally, if the condition stated under (3) is desired, and it will warrant the expense, the best method is first to oil quench from a temperature somewhat over the Ac 1 .2.3 range, and subsequently anneal at a temperature just below that range. Normalizing. If the steel to be annealed has the free cementite existing as network or spines, which would make the steel difficult of machining, annealing at the usual temperatures (Ac 1.2.3) will not affect this cementite: it will simply refine the ground-mass. In order to eliminate this free cementite, it will be necessary first to normalize or quench the steel from a temperature above that of the Ac. cm range. That is, air cooling from a temperature of say 1750 or 1800 F. will not permit of the reformation of coagulated cement- ite. The second annealing may then be carried out at a temperature of 1375 with the refining of the grain size and complete softening of the steel as a whole; this second heating should be just as short as possible in order to prevent the reformation of the free cementite. Spheroidizing the Cementite. The above method may be further modified by reheating to a temperature slightly under the lower critical range instead of over it. The objection to this is that the steel will not be refined, but will possess the large grain size charac- teristic of the high temperature. On the other hand, the lower annealing temperature will entirely prevent the formation of the free cementite as either spines or as a network. Instead, it will be found that the excess cementite will be thrown out, under these conditions, as little nodules or " spheroids " if the reheating temperature is just about at the end of the lower critical range; or, under certain con- ditions, the whole mass of the steel may be called " granular," if such 64 STEEL AND ITS HEAT TREATMENT / a term is permissible. Further reference to this spheroidal forma- tion of cementite, as obtained by a double " quenching," is given under Chapter VII. Spheroidal cementite in annealed steels may also be obtained by very slow cooling through the end of the Arl transformation: cementite in this condition is a great help in the machining of high-carbon steels. CHAPTER IV HARDENING Hardening. Fundamentally, the operation of hardening in- volves two operations of change in temperature : heating and cooling. The function of the heating is (1) to obtain the best refinement, and (2) to obtain the formation of the " hard " constituents of the steel. Having done this, the steel must then be held in this condition by very rapid cooling that is, by quenching in some medium such as water or oil. Associated with both the heating and rapid cooling there must be as great a degree of uniformity as is possible. Changes on Heating. Steel, when properly hardened, should show no trace of the original structure, such as coarse grain size, network, unabsorbed ferrite (in hypo-eutectoid steels), or any other peculiarities of untreated steel. If such are present in the hardened steel it goes to prove that the operation was not properly carried out. Further, if the structure of the steel has not been suitably changed or developed by the heating operation, it most assuredly will not be altered for the better by subsequent quenching. The.most that such quenching can do is to retain the characteristics which the heating has developed. An attempt has been made graphically to illustrate these facts in the chart in Fig. 46. Column 1 (at the left) represents a normal, 0.4 per cent, carbon, pearlitic steel (at the bottom of the column), and the structural changes taking place in that steel as it is pro- gressively heated through and beyond the critical ranges. For the present ,it is assumed that the structure and micrographic constit- uents obtained by heating to various temperatures, such as A to E, may be retained by quenching, as illustrated by columns II to VI. Thus heating to a temperature A, under that of the lower critical range, will produce no change in the original steel, which consists of pearlite (the cross-hatched circles) and ferrite (the black area). The quenching likewise will produce no change, as is illustrated by Column II. Heating to a temperature B, slightly over the lower critical range, will change the pearlite to the solid solution (represented by 65 66 STEEL AND ITS HEAT TREATMENT the dotted area), but without affecting the free ferrite. Quenching, column III, will therefore produce a semi-hardened steel since the solid solution is the " hard " constituent with a refinement of the " ground-mass " (the original pearlite) only. Heating to a temperature C, between the lower and upper critical ranges, will effect a progressive absorption by the solid solution of ii in IV FIG. 46. Changes in a 0.4 per cent. Carbon Steel on Heating and Quenching. the remaining free ferrite. Quenching, column IV, will therefore produce a " harder " steel than in case III, but nevertheless without complete refinement of the steel as a whole. Heating to a temperature D, slightly over the upper critical range, if prolonged for a length of time sufficient to effect complete diffusion and equalization, will entirely refine the steel, giving it the smallest grain size possible. Quenching, column V, will retain this condition and give the maximum hardness possible. HARDENING 67 Heating to a temperature E, considerably over that of the upper critical range, will tend to increase the grain size; and quenching, column VI, will retain this condition, giving a more brittle steel. Relation of Hardening to Annealing. Thus it will be seen that during the heating operation the changes taking place in the micro- scopic constituents and the structure as a whole are similar in both hardening and annealing. The main difference in the final results of the two processes is due to the rate of cooling through the critical ranges, and, therefore, upon the nature of the micro-constituents which are thereby retained in the steel when cold. The effect of slow cooling through the critical ranges, which is characteristic of true annealing, has been discussed; in brief it may be said that the austenite or solid solution shifts its carbon content through generating pro-eutectoid ferrite (or cementite) to the eutec- toid ratio of about 0.85 to 0.9 per cent, carbon, and then transforms with increase of volume at Arl into pearlite, with which the ejected ferrite (or cementite) remains mixed. This change or decomposition of the austenite, however, does not take place suddenly or spas- modically, but develops by stages; and that these intermediary stages between austenite and its final constituents may be recognized and identified under the microscope as martensite, troostite, osmond- ite and sorbite is generally accepted. Hardening is but the result of obstructing this transition, thereby retaining in the steel the " hard " austenite or its early decomposition products martensite or troostite. Austenite. Austenite is only obtained with difficulty in the ordinary carbon steels, and even then is usually decomposed in part into martensite. The two agents 1 rapid cooling and carbon tending to obstruct this transition must be grouped in suitable pro- portions that is, the carbon content must be high, and the cooling take place with extreme rapidity. With about 1.5 per cent, carbon steel, such as is generally used in corrugating and roll-turning tools, when quenched in brine or very cold water from about 1400 F., about one-half of the austenite will remain unaltered. When the carbon is about 1.1 per cent. which may be regarded as about the minimum limit, although the author has succeeded in obtaining some austenite with water-quenched 0.9 per cent, carbon steel the cooling must be done in iced solutions from a temperature of 1800 F. or more. 1 Alloys are also obstructing agents in the sense that, if present in the proper amount, they lower the temperature at which the transition will commence. 68 STEEL AND ITS HEAT TREATMENT The hardness of austenite, as preserved in hardened high-carbon steels, does not fall very far short of that of the accompanying mar- tensite, probably because the austenite is partly transformed into martensite in cooling. On the whole, however, austenite may be regarded as being considerably softer than martensite, and also much tougher; the austenite as obtained in high manganese and high nickel steels is but moderately hard. Martensite. Martensite is the chief characteristic constituent of hardened carbon steels when cooled rapidly in water from a tem- perature above the A3 range. In very high-carbon steels, rapidly FIG. 47. Martensite. X75. (Ordnance Dept.) cooled, the martensite is associated with austenite. In the lower carbon steels hardened in water, in high-carbon steels hardened in oil, or in thick pieces of high-carbon steel hardened in water, mar- tensite is associated with troostite and with some pro-eutectoid ferrite or cementite. Of the transition constituents austenite to pearlite martensite is the hardest and also the most brittle, having extremely high tensile strength with little or no ductility. Microscopically martensite is characterized by a needle-like structure as is shown in Fig. 47. Troostite. Troostite is obtained by cooling through the trans- formation range at an intermediate rate, as in small pieces of steel HARDENING 69 when quenched in oil, or quenched in water from the middle of the transformation range, or in the center of larger pieces quenched in water from above the critical range. The early appearance of troos- tite in tool steel is shown in Fig. 48. The hardness of troostite is intermediate between that of the martensitic and pearlitic state corresponding to the carbon content of the specimen. In general, the hardness increases, the elastic limit rises, and the ductility decreases, as the carbon content increases. H '" j -'> '. v4 v r >< *^ * NW > H v-v N 7'.^ FIG. 48. Troostite (Dark) in Hardened Carbon Tool Steel. X100. (Bullens.) Sorbite. Sorbite, when obtained by hardening, is ill defined and almost amorphous; it is 'softer than troostite for a given carbon content. Dependent upon the carbon content, sorbite may be obtained by quenching small pieces of steel in oil or in molten lead, or even by air-cooling them; or it may be obtained by quench- ing in water from just above the bottom of the Arl range. Sorbite, and to some extent, troostite, are more characteristic of tempered 70 STEEL AND ITS HEAT TREATMENT steels than of hardened steels. The transformation of troostite into troosto-sorbite is shown in Fig. 49. Temperature for Hardening. As a general rule, hardening is carried out from a temperature of about 50 F. above the line A3-A2.3 -Al.2.3 in Fig. 13. This is done to obtain the best refinement of the steel as well as the maximum hardening effect. Rapid cooling of medium and low-carbon steels from a temperature just above the FIG. 49, -Trooste-Sorbite (Dark), X100. (Bullens.) bottom of the critical range Al, will not bring out the maximum hardening effect. The general temperatures most applicable for individual steels are given in subsequent chapters. When the maximum results, both as affecting the structure and also the physical results, are to be obtained, experiments should be made to determine exactly how far over the critical range the steel should be taken. In some steels it will be found that approximately 50 F. will accomplish this purpose; in others it may be necessary HARDENING 71 to raise the temperature to even 150 F. or more. This effect will also be influenced by the size of the piece to be hardened, as will be shown under " Tool Steel." Heating for Hardening. The general rules for heating for hardening may be simply stated, but their fullest comprehension and application may be obtained only in the light of experience. This heating requires much more care than heating for annealing (if such be possible) , on account of the diametrically opposite functions which are indicative of the two operations. Heating for annealing is followed by slow cooling and the gradual release of all stresses and strains: heating for hardening is followed by the most severe test to which steel can be put very rapid cooling, accompanied by the setting up of a condition of stress and strain. In general, we may say that the heating for hardening should be slow, uniform, and thorough, and to the lowest temperature which will give the desired results. Non-uniformity in heating must of necessity result in lack of uniformity in cooling, which in turn is the genesis of most of the troubles in the hardening process. Hardening cracks are more often the result of uneven heating than of a defect in the steel. Heating requires time and care. The peculiarities of each steel and article must be thoughtfully studied; experiments must be made; and the clear judgment of experience applied to each individual case. It has been well said that " Steel is mercurial and delicately responsive to heat; its records appear in its own structure." Lowest Quenching Temperature the Best. The lowest heat which will give the results desired should always be used in hardening. This point can be brought home in no better way than to give the results of two tests made which illustrate exactly this principle. Two automobile gears were made from the same bar, by the same man, and in all other ways as nearly alike as possible. The tests were made by a disinterested third party. Number 1 was quenched in oil from 1450 F., annealed at 1400, hardened in oil from 1450, and tempered at 475 in oil. It gave a sclerescope hardness of 76 to 78. It withstood 48 blows of a 10-lb. hammer dropping 30 ins. before a tooth could be broken out, or 8 blows of a 10-lb. hammer dropping 48 ins. Number 2 was quenched in oil from 1400, which was just over the critical range, and determined by when the magnet " let go." It was annealed at just under that temperature, followed by hardening in oil from 1400 and tempering in oil at 475. The 72 STEEL AND ITS HEAT TREATMENT hardness was the same as with No. 1. In this case, however, it required 200 blows of the 10-lb. hammer falling 30 ins., or 78 blows with a fall of 48 ins. The effect of the increase of only 50 in the hardening temperature is self-evident: it meant a difference in efficiency in the ratio of 48 to 200. Temperature of Quenching. It is not only the uniformity of heating of the steel object which is necessary for uniform and proper hardening, but also and equally important the uniformity of temperature of the piece at the moment of quenching. A piece of steel may be properly heated at the moment previous to its with- drawal from the furnace; but that same piece may have wide dif- ferences of temperatures in different parts of the mass at the moment of quenching. Non-uniformity of heat saturation at the latter instant must inevitably result in non-uniformity of hardening with the attendant possibility of warping, cracking and similar fea- tures. Whether or not the indirect cause of such a condition is due to the shape and size of the object or to the method of handling the stock between the furnace and bath is immaterial as far as the basic principle outlined above is concerned. These are but a mani- festation of the ever-present "personal element." Overheating. Overheating is probably one of the most com- mon sins of the hardening shop. Unfortunately, many "practical " men still believe in the efficacy of high temperatures for greater hardening effect. Although this may to a very limited extent be true, the weakening of the steel by the increase in grain size and greater hardening strains as obtained by high temperatures more than offset the questionable production of greater hardness. Fully 80 per cent, of the complaints of " bad steel " which have been brought to the author's attention have been the direct result of over- heating. Both theory and practice support the old rule that " the lowest heat which will give the desired results is the best heat." The Magnet in Hardening. It will be recalled that steel becomes non-magnetic (for all practical purposes) in passing through and beyond those temperatures represented by the heavy black line in Fig. 50. For steels with about 0.35 per cent, carbon and upwards this temperature line also corresponds to the best refinement of the steel in heating. We therefore have a very simple and practical means of determining the proper temperatures for hardening such steels. All that is necessary is to apply an ordinary horse-shoe magnet, suspended from a suitable rod, against the hot steel. When HARDENING 73 the correct hardening temperature has been reached there will be no attraction between the magnet and the steel. For steels with less than about 0.3 per cent, carbon the rise in temperature between the Ac2 and Ac3 ranges may be estimated and the steel hardened when it has reached the latter temperature. It will be found that the use F. 1800 1700 1600 1500 1400 1300 1209 0.2 O.i 0.6 0.8 1.0 1.2 FIG. 60. Carbon-iron Diagram Showing Temperatures at which Ordinary- Carbon Steel Loses Its Magnetism on Heating. of the magnet will be of great value to those not having the proper pyrometer control over their heating operations and who have to depend entirely upon the eye for gauging such. An instrument more convenient for this purpose than the ordinary magnet is made by magnetizing a small, elongated, diamond-shaped piece of steel, 74 STEEL AND ITS HEAT TREATMENT and supported between two pins in the end of a forked rod, as is shown in Fig. 51. Motion during Hardening. When a piece of steel is quenched, movement should be given to the steel, to the quenching medium, or to both. This is for two reasons: (1) in order to cool the steel as rapidly as possible, and (2) to break up any tendency toward the formation of a distinct line between the hardened and unhardened parts of a differentially hardened forging. The first factor should be self-evident: agitation of the bath or movement of the hot steel will lower the temperature of the oil or water which is cooling the steel, will prevent the formation of vapor or steam around the steel, and in other ways more rapidly cool the metal. In the second place, if a piece of hot steel, such as a chisel or die-block, were to be immersed to a certain point and held there quietly, the rapid cooling would harden the steel up to the point of immersion and no further; in other words, there would be a sharp line of demarkation between the hard- FIG. 51. A Magnet Used in Hardening. ened and unhardened parts, and which in turn would be a source of great weakness and possible fracture. So, in quenching, move the piece up and down in the bath; or if it is to be only partly immersed, agitate the bath so that there will be no distinct line of hardening: avoid straight-line hardening. Furnace Equipment. The general principles of heat application will be discussed elsewhere. Some of the main points to be con- sidered are uniformity of heated product, quality, and economic efficiency, and to conduct all operations with these in view. The material to be heated should be so arranged in the furnace that there is ample room for the circulation of the heat through the mass. The furnace should be so designed as to suit the class of work to be heated, and so operated that the heat shall be evenly distributed and of the same temperature from floor to roof and side to side. Only under special circumstances, such as in connection with automatic furnaces, is it desirable to operate the furnace at a temperature higher than the maximum temperature desired in the steel. The Human Element arid Basic Heat Treatment Conditions. What is it or who is it that determines when the charge in the cham- HARDENING 75 her is saturated with heat to the temperature indicated by the pyrometer? What determines the manner in which the charge is placed in the furnace and the room for circulation throughout the mass?' What determines when the bottom and center of the mass are at the same temperature as the top and outside? What regulates the flow and composition of gases in the chamber around the stock and the discharge of heat from the chamber? What determines if each piece is heated like every other piece and is uniform throughout, and whether each piece goes into the quenching bath at the same temperature as all the others and at the temperature indicated by the pyrometer? What is it that controls the flow of air into the furnace or the flow of gases from the furnace, and by so doing determines whether the atmosphere surrounding the stock is oxidizing or neutral, and whether the fuel is conserved or wasted? These are some of the elements that affect proper heat treat- ment, and they are determined, not by a pryometer nor a furnace nor by similar apparatus nor by any mechanical means, but by the same methods that govern the quality of the products of the kitchen the judgment and skill of the operator. Heating Baths. Despite the efficiency of design of many furnaces, and not mentioning those of poor design their often inefficient operation has a general tendency towards non-uniformity in heating and oxidation. In the effort to solve these two problems at once that is, to surround the object to be heated with a constant and uni- form heat on all sides, and to avoid contact with the air the appli- cation of various molten baths has come about. Chief of these heat- ing mediums are molten lead and certain salts. On account of the operating cost and necessarily small capacity, however, their use is largely confined to the heating of tools and other articles requir- ing particular care, uniformity, and freedom from oxidation. The principaKuse for such baths, in the author's opinion, should be for the retention of a bright surface on the metal after hardening, and not for uniformity in heating; any furnace, properly designed for the work in hand, heated with the right fuel, and correctly oper- ated, should give entire uniformity of heating. Heating in Lead. Before the advent of the modern heat treat- ment furnace, heating in molten lead represented the most practical method of obtaining uniform heating. With a reasonable amount of care and attention the typical lead bath may be maintained at 76 STEEL AND ITS HEAT TREATMENT such temperatures as the ordinary hardening operation requires and with a satisfactory degree of uniformity. Its use, however, presents many difficulties. The bath must be frequently agitated to preserve a uniform temperature. When heated to over 1200 F. lead begins to volatilize, giving off fumes which are both offensive and poisonous ; suitable ventilation, such as may be obtained with a properly de- signed hood, should be provided to remove these fumes. Further, the bath must be covered with powdered charcoal to reduce the oxides or dross which are formed in the molten lead. Many plants will not use lead baths if temperatures greater than 1475 or 1500 F. are necessary. On account of its high specific gravity, heating in lead requires some method of holding the steel beneath the surface, as otherwise the tool would float on the surface of the bath and thus be unevenly heated. One of the most troublesome difficulties with lead baths is the tendency of the lead to stick in the holes, threads, or even to the surface of the tool when it is removed for quenching, so that uniformity of cooling is sometimes materially affected. Al- though this particular difficulty has been largely eliminated by the use of a paste, the trouble may simply be aggravated in case this coating has not been carefully and properly applied. Salt Baths. Many of the difficulties encountered in the use of lead for heating may be overcome by the substitution of different salts. Their lower specific gravity permits of a more uniform circulation and there is no tendency of the tool to float on the sur- face of the bath. At the usual temperatures used for hardening there is little or no vaporization. Although lead may prevent the steel from oxidation while the steel is being heated, as soon as the tool comes in contact with the air on removal from the lead bath, a thin film of oxide is formed; with the salt bath, on the contrary, the steel receives a thin and uniform coating of molten salt, which protects the surface of the metal. The minimum temperature of the salt bath may be very closely estimated without the use of a pyrometer. Common table salt has a freezing-point of 1472 F., and if it should be melted with potassium chloride (freezing-point 1325 F.) or other salts, the freezing-point of the melt may be quite accurately adjusted over a wide range of temperatures. By keeping the bath very near its freezing-point by a suitable regulation of the heat, overheating of the steel may be entirely overcome. Further, if the composition of the salt bath has been so adjusted as to approximate the proper hardening temper- ature, when the steel is removed from the bath it may be quenched HARDENING 77 just at the time when the salt film begins to solidify or at exactly the correct temperature. BATHS FOR QUENCHING General Properties of Quenching Media. The main thought in selecting a proper bath for quenching is the rapidity with which the heat is removed from the hot steel. This property of transference or withdrawal of heat from the solid by and to the liquid, will depend upon the specific heat of the liquid, its conductivity, viscosity and volatility. That is, the specific heat will indicate the heat-absorptive power of the liquid; the conductivity will measure its capacity for transferring the heat thus absorbed to the cooler part of the bath ; the viscosity affects the motion of the liquid and thus influences the uniformity of cooling; and the volatility indicates the temperature at which the liquid will become gaseous, thus forming a vapor around the steel and preventing the quick removal of the heat from the steel. By obtaining a suitable combination of these various prop- erties a bath giving the desired effect may be obtained. Temperature of the Bath. The continuous use of any bath for quenching will gradually and progressively raise the temperature of the liquid used. As a general rule, differences in the temperature of the bath will give rise to varying results in the actual hardening taking place the higher the temperature of the bath, the less its cooling efficiency. This is especially noticeable with water; a change of 50 or 100 F. will often entirely alter the physical properties of the quenched steel. The effect is less marked with oils, and with some oils may be almost negligible for certain classes of work. On the whole it is decidedly better practice to maintain as nearly as possible a standard temperature in the quenching bath. Quenching Speed of Different Media. It is evident that the cooling medium used, its temperature and condition will affect the rate of cooling. Matthews 1 and Stagg have devoted considerable time to investigating numerous commercial media which are in use in typical hardening plants of the country at the present time. Their method was as follows : A suitable test piece was machined from a solid bar, and a hole drilled through one end to within an equal dis- tance from each side and bottom of the test piece. Into this hole a calibrated, platinum-rhodium couple was inserted and the leads connected to a calibrated galvanometer. The test piece was then i Matthews and Stagg, " Factors in Hardening Tool Steel," A. S. M. E., 1915. 78 STEEL AND ITS HEAT TREATMENT immersed in a lead pot, and the lead pot was maintained at a tem- perature of 1200 F. When the couple inside the test piece was at 1200 F., and the couple in the lead pot also read 1200 F., the test piece was removed and quenched in 25 gals, of the quenching medium under consideration. At the start the quenching medium was at room temperature. The time in seconds that it took the test piece to fall from a temperature of 1200 F. to a temperature of 700 F., was noted by the aid of a stop-watch. It is clear that immersing the test piece in the quenching medium raised the temperature of the medium. The test piece was then replaced in the lead, heated to 1200 F., quenched into the medium at this higher temperature and the time again taken with the stop-watch. These operations were continued until the quenching medium, in the case of oils, had attained a temperature of about 250 F. The results obtained, time in seconds, for a fall from 1200 F. to 700 F., were plotted against the temperature of the quenching medium and a series of curves as shown in Fig. 52 l were obtained. The various curves represent the following quenching media : W. Syracuse city water. B. Brine. Sec. 1. New fish oil; average of readings from 80 to 250 F. . 85 2. No. 2 lard oil 87 3. Prime lard oil in use two years 99 4. Boiled linseed oil 101 5. Raw linseed oil 102 6. New extra-bleached fish oil 106 7. New yellow cottonseed oil 107 8. New tempering oil; 60% cottonseed, 40% mineral. . . . 122.6 9. New mineral tempering oil 130 10. No. 1 dark tempering oil 157 . 3 11. Special " C " oil 164.7 A consideration of the results is interesting. Pure water (curve W) has a fairly constant quenching rate up to a temperature of 100 F., where it begins to fall off. At 125 the slope is very marked. Brine solutions (curve B) have both a quicker rate of cooling and are more effective at higher temperatures than water. The curve does not begin to fall off seriously until a temperature in the neighborhood of 150 is reached. Where water at 70 cooled the test piece in 60 sec., the brine solutions cooled it in 55 sec. 1 For the sake of brevity and clearness the numerous curves as given by Mat- thews and Stagg have here been grouped under one plot. HARDENING 79 As is well known, the oils are slower in their quenching powers than water or brine solutions, but the majority of them have a much more constant rate of cooling at higher temperatures than water or brine. The curves shown in 10 and 11 are for thick viscous oils similar to cylinder oils. These curves are particularly interesting in that they have slower quenching abilities at low temperatures than at higher temperatures. A comparison of curves 2 and 3 shows the variation in quenching power of the same oil due to continued ser- 100 150 Time jaSeconds FIG. 52, Quenching Power of Liquids. vice. The differences in quenching rates may well account for different results from the same steel in different shops, or in the same shop due to change of oil used. Water Spray for Hardening. Water sprayed under pressure is the quickest agent for rapid cooling in common use, exceeding in its hardening qualities either brine or water baths. The main point to be noted is that there shall be sufficient volume and pressure to prevent the formation of a blanket of steam between the hot steel and the 80 STEEL AND ITS HEAT TREATMENT^ spray. Its most common use is for such tools as sledges and others requiring a differential hardening, and for armor plate. Brine. Brine is used only in certain particular lines, such as file- hardening, for which an extremely hard surface is required. Unless the steel has been most carefully heated, and is of a proper chemical composition, quenching in brine is almost certain to crack the steel. This is particularly true of large sections, for in these the very sudden cooling of the outer surface, while the center is still hot, will set up stresses and strains which will not be relieved or equalized in the short time allowed, and with the inevitable results. Water Quenching. The author is a firm believer in the use of oil for quenching, rather than water, and would recommend its use whenever conditions permit. Water cools the steel more rapidly, but its more drastic action increases the internal strain and consequent liability to fracture. For the low-carbon steels, and for small and comparatively simple sections of the higher carbons, water quenching may be used without much danger. Of course in cases where it is required that the surface shall be glass-hard, or that the maximum tensile strength be obtained, water quenching is man- datory. On the other hand, if the steel is to be given a full heat treatment (i.e., quenching and toughening), the difference in hard- ness as obtained by the two baths may usually be nearly equalized by using a lower drawing temperature for the oil-quenched piece; that is, if a 0.40 per cent, carbon steel forging is quenched in water and toughened at say 1200 F., approximately the same static prop- erties may be obtained by oil quenching and a subsequent reheating to say 1050 F. The principal objections to the latter method are that the lower drawing temperature is not so easily recognized by its color, nor will the dynamic properties probably be quite as high though this last point is questionable. Generally speaking, however, oil quenching is more desirable than water quenching. Oil Tempering. The term " oil tempering," referring to the quenching in oil, is one which has become current in the trade, so that the term, " hardening " often refers to quenching in water only, or in some medium which will give an equivalent or greater hardness. Strictly speaking, the use of " tempering " in this sense is a mis- nomer, for it should be used as indicative of a slight reheating or " softening " of the quenched steel. Special Quenching Methods. It often happens that especially high tensile results are desired in certain large forgings of such size and chemical composition that direct quenching in water is deemed HARDENING 81 unwise, and yet in which it is desired to obtain as near the maximum effect of water cooling as possible. A method which has proven in a large measure successful is to use a bath of oil resting upon an equiva- lent or greater volume of cold water. The forgings, when heated to the proper temperature, are lowered into the oil for a few seconds and thence into the water. The oil forms a film on the surface of the steel, so that the sudden effect of the water is somewhat diminished or retarded. The rapidity of cooling may be controlled by the dura- tion of the oil quenching. It is obvious that in using this method there must be a sufficient volume of water under the oil to prevent the formation of steam and its consequent danger. For small tools or thin instruments such as saws, the above method may be so modified as simply to have a film of oil upon the surface of the water, the oil in this case consisting of some animal or vegetable oil. The heated tool is plunged directly and evenly through the oil film so that it enters the water with a thin coating of burnt oil which protects it from the direct action of the water and lessens the risk of fracture. The amount of oil may of course be increased as desired. The main objection to these methods is the lack of uniformity in hardening unless the operator has had more or less experience. A method which is extensively used in some tool works is that of using a combination of water and oil quenching, that is, first plunging the tool into water until a certain amount of heat has been removed, and then transfer to the oil, where it remains until cold. Molten lead is sometimes used as a quenching medium for small sections in which great toughness and only a moderate degree of hard- ness is desired. Although dependent upon the carbon content, steel subjected to this process will generally be sorbitic. Such treatment will require no further reheating. Other Aqueous Quenching Media. Hardeners, at one time or another, have tried about everything under the sun in the attempt to discover some new and wonderful quenching medium which would accomplish the phenomenal. The results, for the most part, do not warrant the addition of expensive chemicals; and if the experi- menters do claim the marvelous, the " gold-brick " scheme is gen- erally revealed by thorough investigation. Some substances, such as lime, soap, etc., may be added to form a protective coating around the steel. Calcium chloride will raise the boiling-point of water to a considerable degree, so that the solu- tion may be used at a temperature up to 150 or 175 F. without 82 STEEL AND ITS HEAT TREATMENT danger, and at the same time give many of the advantages which oil hardening possesses. Some salts increase the hardening effect of water; others purify the water or soften it. One of the most inter- esting (and wonderful?) combinations which has come to the author's attention contained by addition ammonia, glycerine, sal-ammo- niac, spirits of nitre, ammonium sulphate, alum and zinc sulphate! Differential Hardening. In certain tools, such as anvil faces, die blocks, edge tools, and the like, it is desired to obtain a very hard outer part, surface or edge, to be " backed " by a less hard and tougher steel. That is, the steel is gradually and progressively to change from extreme hardness to the opposite, or what we may term differential hardening. This phase may be obtained either by heating the whole mass of the steel as in die blocks, or by heating only part of the article as in chisels; in either case that part which is to have the greatest hardness is immersed or quenched. By this method the heat is gradually withdrawn from the part not immersed through that part which is being subjected to the cooling bath, so that the mass of steel as a whole will become progressively softer or tougher from the hardened face or edge to the opposite side. Precautions must be taken to avoid straight-line hardening, Cooling the Water Bath. Where water is used as the quenching medium it is customary to maintain a flow of fresh, cold water into the quenching tank so as to keep a uniform temperature and purity. Water which has been used for any length of time without renewal goes " stale " with a corresponding loss in cooling efficiency. If the cost of water is such that it is inadvisable to dispose of the overflow from the tank, the hot water may be cooled by spraying, cooling towers, etc., aerated, and then returned to the tank. Cooling the Oil Bath. The common methods for cooling the oil- quenching bath may be broadly classified as follows: (1) The cir- culation of cold water around, or through coils in the bath; (2) the circulation of the oil itself; (3) by the use of compressed air. One of the simplest methods for cooling the oil when in small tanks and not too constantly used, is to place the oil tank within a larger tank, with a space of say 2 to 6 ins. between the two tanks. This space is kept filled with cold water. As in all these systems, the intake should be at the bottom of the tank, with the outlet or overflow at the top. The main objection to this method is the fact that the heat in the oil must penetrate through the walls of the tank before it can be conducted away by the water. The next type of cooling makes use of coils or radiators placed HARDENING 83 within the oil tank and the circulation of cold water through these pipes. These water lines are placed close against the side of the tank so that they may not interfere with the work being treated. From his own experience, the author does not feel that the radiator J L a! ' rj- II r a i IV II . 01 I n- J-n ft- FIG. 53. Radiator Type of Cooling System. type as shown in Fig. 53 gives as great efficiency as the simple coil system of Fig. 54. With the difference of temperature of the oil in the bottom of the tank, as contrasted with the hotter oil at the top, it is difficult to obtain a thorough circulation of the cold water VJ 1 N 1 1 1 D ) ^1 L^ ( C II 1 r V 1 r\ 1 1 i p ) /I - . ' - [y 1 S i 1 i r M r\ A u ' ?J (t- n- \ j L \ \j .... _ L s FIG. 54. Coil Type of Cooling System. through all sections of the radiator. Further, this same difference in temperature has the tendency towards unequal expansion of the top and bottom pipes, which may cause a leakage of water into the oil and its attendant dangers. In the coil system there is of necessity a complete circulation, together with the elimination of expansion 84 STEEL AND ITS HEAT TREATMENT dangers. These pipes vary in size from about 1 J ins. to 3 ins. diam- eter; the latter size has given excellent satisfaction in a tank approximately 8 ft. wide by 16 ft. long and with a working capacity of about 8000 gals, of oil. Guide strips should be placed at intervals along the coils from top to bottom to prevent any articles from catching against the pipes while the quenched material is being raised out of the tank. Circulation and Cooling of the Oil Itself. The best results for keeping down the temperature of the oil bath are undoubtedly to be had when the oil itself is circulated. The circulation is continually bringing cold oil into the vicinity of the hot metal, removing the hot oil from the tank, as well as giving a more uniform temperature to the bath as a whole. In the previous systems the heat must be taken away by gradual and progressive transference from the region of the hot steel towards the sides of the tank, and at the best is a slow procedurethis is assuming that the oil is not kept in motion by compressed air. In the present system, the heat is taken away from the quenching bath by the actual removal of the hot oil itself. The usual methods are to pump the hot oil from the tank and then through coils which are cooled by suitable means; or by maintaining large supply tanks in which the oil will have sufficient time to cool before being returned to the quenching tank. In the former pro- cedure the coils containing the hot oil may be cooled by refrigerating such as the ammonia process, etc. or by placing the coils in a water tank, or by cooling the coils with a continual stream or spray of water. Where the size of plant will permit the installation of a refrigerating system, such a method is by far the most satisfactory; the heat may be removed very quickly, and the temperature of the oil controlled at any desired temperature by the regulation of its flow through the cooling coils. As an example of the water-bath method, one steel company pumps the oil from the quenching tank holding some 12,000 gals. through 3-in. pipes and thence through coils placed in a large water tank used for the mill supply. The cold oil then returns to the quenching tank by gravity. For smaller plants the coils may be most conveniently cooled by the use of tiny streams of water trickling over the coils. On the whole, this is probably the most satisfactory system of all for small plants. In one case (in which the question of the cost of water was important) this method was found to be both cheaper and to give a higher cooling efficiency than could be obtained by setting the coils HARDENING 85 in a small water tank. In the latter case the heat is removed by transference from one part of the water to that further removed from the coils, so that unless a very good flow is maintained, the cooling will be comparatively slow. Further, the water removed from the tank is, on the whole, but lukewarm, and therefore but imperfectly accomplishes its mission. On the other hand, in the drip system a small amount of cold water is always in contact with the coil, giving a maximum cooling efficiency with a minimum expense. A recent heat treatment installation l attacks the problem of keeping the quenching medium at a uniform and low temperature by the maintenance of a large and separate supply of oil. The hardening is done in special quenching tank cars, as shown in Fig. 55, and which are wheeled to any furnace desired. Just before quenching commences the valve in pipe K is turned on and a 2-in. stream of cold oil is kept flowing into the tank. The hot oil passes out through the overflow pipe L, through the hole in the floor and into a pipe that conducts it into an underground tank. This underground pipe is made very large, so that there will be no danger of its clogging, which would necessitate tearing up the floor. Each furnace through- out the 400-ft. length of the shop is provided with a similar inlet pipe and floor hole connection to the pipe which carries away the overflow. From the underground tank the oil is pumped to upright tanks close to the outside of the building; from these tanks the oil flows by gravity to the tank cars. Use of Compressed Air. The advisability of using compressed air in the -quenching tank is a much debated point. If applied intelligently, however, it undoubtedly renders great assistance in the hardening and cooling operations. In systems in which the oil is kept in constant and fairly rapid circulation, it is neither required nor advised. 2 But if the oil is cooled by the circulation of water in pipes, the use of compressed air is often mandatory in order to obtain the maximum, as well as uniform, cooling efficiency of both oil and water. In any case, the air must not be allowed to come in contact with the hot steel, as soft spots would result; neither should it be used in too great quantities nor pressure, especially with the heavier and low-grade oils, as it may cause the precipitation of 1 " A Modern Heat-Treatment Plant,' 7 Machinery, Sept., 1914. 2 The cold oil forced into the quenching tank may be distributed under pres- sure to different parts of the tank, thus providing excellent circulation, and accomplishing the same results as compressed air. 86 STEEL AND ITS HEAT TREATMENT certain constituents of the oil, or cause the formation of a scum or foam on the surface of the oil. When the air sets up a fairly efficient circulation of the oil (or water, if water is the quenching bath), it has accomplished its mission. Compressed air should rarely be used with animal or vegetable oils on account of oxidation. HARDENING 87 Size of Quenching Tank. The volume of the quenching medium to be used, and hence the size of the tank, depends principally upon the size and number of the pieces to be hardened, and also upon the method used for cooling the quenching bath. The tank should always be of sufficient size to take with ease the maximum size stock to be treated, besides a generous allowance on all sides for a suffi- cient body of oil or water, for rapidity in handling the material, and for circulation. Further, the size of the tank should be proportioned to the degree to which the solution can be kept cooled when the hard- ening department is operating at maximum capacity; the more efficient the cooling system, the smaller the size of tank necessary. On the whole, it is decidedly preferable to have the tank too large than too small. CRACKING AND WARPING Influence of Non-uniformity of Section on Cracking. One of the main causes of steel breaking in hardening is from the unequal con- traction and expansion in different parts of the steel. If it were pos- sible to get every particle of the steel cold at the same moment there would be an end to danger of this sort. But as this is a physical impossibility, we must approach such a condition as near as we can. This danger of cracking is particularly emphasized in forgings or tools of unequal thickness. If the thinner part should be first immersed in the quenching bath (e.g., water), it would become cool much sooner than the heavier sections; that is, the thin part would be cold or " fixed " while the thicker part of the article was still contracting from loss of heat. Hence the thin part in its then hard and brittle state cannot " give " and will consequently break; or, if it does not break at the time of hardening, the steel is held in such a state of stress that it is ready to break when applied to the work, or even when being tempered. These influences are the more marked with the greater the rapidity of cooling and hardening effect of the bath, as well as with the increase in carbon content and alloys. Influence of Bulk of Section on Cracking. Further, the danger of cracking is dependent upon the bulk of the article, even though it be of uniform section. Its effect is repeatedly illustrated by large forgings such as locomotive axles, crank-pins, etc., of rather high carbons quenched in water. This point is illustrated by the case of a locomotive crank-pin which had been hardened in water and then toughened. A thorough examination of the forging before shipment to the railroad company revealed no external evidences 88 STEEL AND ITS HEAT TREATMENT of any crack; but when it had been in service but a very short time it fractured badly. Examination then showed that it had evidently been in a state of stress within its center, with the development of an embryo crack; the dynamic stresses to which it had been sub- jected in service were sufficient to raise the tension beyond what the steel would stand, with the resultant internal fracture and its pro- gressive development into complete rupture. Expansion and Contraction. In view of recent research work this phenomenon of cracking may be explained in a theoretical manner along the following lines. We know that when a piece of steel is heated through the critical range the formation of austenite takes place with a decrease in volume; and a somewhat corresponding and opposite increase in volume occurs when it is cooled through the same critical range. Now if a large forging of considerable diameter is quenched rapidly, the outer sections will be held in the hardened condition, and therefore rigid and stressed. Meanwhile the interior of the steel, being cooled much less rapidly, will in all probability actually pass from the austenitic-martensitic condition into that of pearlite, accompanied by the increase of volume noted above. If the outer portion or surface of the steel is unable to withstand this expansive force, rupture must necessarity occur. Illustrative of this, the author has seen heavy locomotive axle forgings, after removal from the oil-hardening bath, actually break open with a tremendous report. However, if the forging has not been hardened too drastically, and is removed from the quenching bath before entirely cold, an immediate reheating or toughening process will generally relieve these stresses before any actual damage takes place. Hollow Boring. In order to avoid such dangers, there appears to be a decided tendency toward requiring the drilling of axles, shafts and heavy forgings of large diameters to provide for heat treatment and to remove defective material. It is undoubtedly the fact that heat treatment will not attain its full effects in the core of a large section. With a solid axle, the heat, upon quenching, is removed by a flowing from the center to the outside and thence to the hard- ening bath; the amount of heat is so great, however, that at the best the core will be but semi-hardened, and in most cases will but be grain-refined, or annealed. This point was well illustrated by one company in its experiments: it split open a large, heat-treated driving axle; the fracture showed that the heat treatment had penetrated the ends to a depth of about 6 or 8 ins., and on the HARDENING 89 sides to a depth of about one-half the radius; the fracture of the core was similar to that of annealed steel. Again, the loss of duc- tility and failure of the heat-treatment process thoroughly to pene- trate the core of semi-hardened steel are shown by the following results obtained from a 12-in. axle, heat treated, and taken at regular intervals from the center to the outside: Tensile Strength. Lbs. per Sq. In. Elastic Limit, Lbs. per Sq. In. Elongation, per Cent in 2 Ins. Reduction of Area, per Cent. 1. (Center) 95,000 60,000 7.5 9.6 2. 99,750 60,000 15.0 35.3 3. 104,500 65,000 17.5 35.7 4. 104,500 65,500 19.0 40.3 5. (Outside) 106,500 70,000 21.5 47.7 Treatment: Quenched in water from 1580 F.; toughened at 1100 F. Analysis: Carbon, 0.35; manganese, 0.56; phosphorus, 0.020; sulphur, 0.024; nickel, 1.19; chrome, 0.31. By means of drilling a hole through the axle, the quenching solu- tion is able to remove the heat from both the inner and outer part of the axle at the same time. Hollow-bored axles should be quenched vertically whenever possible, and a constant flow of the oil or water through the bore be supplied. The American Railway Master Mechanics' Association in its proposed specifications for alloy steel locomotive forgings (June, 1914) calls for " drilling forgings over 7 ins. in diameter, unless otherwise specified by the purchaser. The committee has found a great tend- ency among users of quenched and tempered steel to require drilling of parts over 7 ins., and this practice is advocated by steel-makers. In the case of axles and crank-pins particularly, drilling takes away practically nothing from the strength of the part; it removes the material from the center where defective material is most likely to exist and where it is least subject to the beneficial effects of heat treatment, and it allows the forging to adapt itself to expansion and contraction due to heating and cooling." Warping. Warping is but another manifestation of the effect of unequal contraction and expansion, originating mainly in incorrect heating or neglect in the manner of quenching, rather than in the more drastic effect of the bath itself. Non-uniform heating must inevitably result in warping, for if some parts are hotter than others when the steel is quenched, it is evident that the rate of cooling over the entire length of the piece cannot be the same. The general 90 STEEL AND ITS HEAT TREATMENT tendency will be for bars to buckle or twist, due to unequal contrac- tion during hardening. Take, for example, a bar which has been placed upon the relatively cold floor of a heating furnace in which the main heat application comes from above. Under these condi- tions the tendency will be for the bar to become more heated along the upper surface than in that in contact with the cold floor. If the bar should now be quenched, the under part being lower in temperature would contract first (provided it were heated and quenched from a temperature over the critical range) and thus become bowed. But if the temperature in the cooler part of the bar were under the critical range, the tendency would be to bend in the opposite direction. Other variations in heating might give a double bend; certain localized heating might even cause twisting or tor- sional strains. Manner of Quenching. Uniformity of quenching is requisite to good hardening work. As a general rule, objects should be quenched vertically in the direction of their greatest length. Like all rules, there are certain exceptions which must be made to this general statement such as in the case of half-rounds and articles of a corresponding design, as well as in such cases where economic handling requires other methods, as with shafts, small axles, plates, etc. But where no special facilities have been designed for uniform quenching, the above rule will be found worthy of adoption for symmetrical sections, and especially with unskilled workmen. The reasons for this may be best explained by taking small automobile drive-shafts as an example. In pulling the piece out of the furnace with the tongs, the tendency is to grasp it nearer the end than at the middle; consequently, in the general haste to get the steel into the quenching bath as soon as possible, the average work- man is very apt to drop or plunge it into the oil or water at an angle that is, one end of the piece strikes the quenching solution before the remainder of the steel. Hence, initial hardening strains are set up which usually result in a bent shaft when it is removed from the tank. It is very difficult, in the space of a second or two, to get hold of the bar exactly at the middle and also to lower it into the water or oil so that both ends are immersed at the same identical moment which this method of quenching demands. Now if the workman was to aim at immersing the piece end foremost, as in Fig. 56, grasping it near the end (as usual) with his tongs, the weight of the shaft would automatically tend to bring the shaft to the normal, and the quenching would be more nearly uniform, Axles and HARDENING 91 forgings of a similar nature should be quenched vertically whenever possible, as less strains are set up in the axle by this mode of quench- ing. Extensive investigations by one locomotive builder would tend to show that axles quenched horizontally (as is customary) develop a series of stresses which, when plotted, appear as an oval around the axis of the axle instead of as a circle. FIG. 56. Proper Method of Quenching Small Round Bars. Hollow forgings, such as guns, hollow tools, etc., should always be quenched vertically, so that the quenching medium may have a free flow through the bore, and also to prevent the pocketing of any steam or vapor which may be formed by the contact of the hot steel and the solution. Round Sections. The hardening of round sections without cracking or bending, and without undue labor cost, presents a problem 92 STEEL AND ITS HEAT TREATMENT which has attracted much study. The danger of fracture, especially of internal origin whether actual or potential is always greatest in the circular section. This is largely due to the fact that all the stresses and their subsequent strains are grouped symmetrically and converge upon the central axis. Both the square bar with its corners, and the plate or sheet with its larger surface exposure, can more easily yield to the internal stresses and afford relief either in cooling during hardening or in the reheating for tempering or toughening than can the circular section. Further, there is greater danger of bending and twisting due to non-uniform cooling in the long, round bar than in almost any other common section. As has been noted, short lengths of rounds of small diameter should always be quenched vertically. But when it comes to the handling of large numbers of larger bars, either of greater length or diameter, this method is obviously at a disadvantage. Yet if the bars are simply dropped into the bath by hand, even if every effort is made to have the axis of the bar parallel to the surface of the quenching medium, general unsatis- factory results are obtained, due to non-uniform cooling. One satisfactory method for quenching such bars is shown in Fig. 55, in which automobile shafts are handled. The bars, after careful heating, are pulled out with long rods which have a hooked end, across the inclined steel fore-hearth J, whence they drop on to a jointed rack in the oil tank and are quenched. By starting the bars with their axes parallel to the surface of the oil, they must neces- sarily be held in the same relative position as they pass down the rack into the oil. The rolling also effects a more uniform cooling of the shaft in relation to its central axis. Fig. 57 shows how the traveling crane lifts one side of this jointed rack to raise the shafts out of the oil and dumps them on to the truck at the side. i v An improvement on this method to give further uniformity in cooling, and which has been used on finished shafts with almost the entire elimination of bending, is illustrated in principle in Fig. 58. The apparatus consists of a number of inclined planes or racks (similar to that shown in Fig. 57), made from small bars or old rails which are held in position by suitable cross-pieces. The hot shaft is started down the first plane and passes into the oil or water; thence it drops to the next, and so on until it reaches the bottom, and is removed by suitable methods. Notice that the change from one plane to the next causes a reversal in the direction of rolling, so that any stresses set up by one plane are practically counteracted by the next plane, giving a maximum uniformity in cooling. The HARDENING 93 angle of incline has a great deal to do with the practical working out of the procedure, and should be varied according to the diameter of the bar, its chemical composition, and the nature of the quenching medium. The rate of travel down the incline should not be too rapid, but should nevertheless be sufficient to allow the reversing 94 STEEL AND ITS HEAT TREATMENT action of the several planes to take its effect before the steel is to'd cold. The angle of the planes may be increased as greater depth in the solution is reached. The bar should be cold when it reaches the bottom of the tank. The angle of the first incline is the most important, and should be determined by experiment; it will gen- erally be in the vicinity of 10 or 15. In one plant in which this method was used the number of shafts requiring straightening was reduced from a very high percentage to less than 1 per cent, of the total number treated. i FIG. 58. Rough Sketch of Inclined Racks for Quenching Rounds. Double Quenching. The effect of a double quench is, as a general rule, to raise the elastic limit and tensile strength without diminishing the ductility. This is for the most part due to the higher degree of refinement which this double quenching makes possible, thus putting the steel in the best possible condition. If the steel is in good con- dition (i.e., refinement) before the first quenching, the influence of the second quenching will be the less in proportion. It is often cus- tomary first to quench from a temperature 100 or 200 F. over the critical range, and then, for the second quenching, to heat just enough over the critical range to obtain the degree of hardness desired. HARDENING 95 For high-carbon steels the double quenching is not to be recom- mended except under unusual conditions such, for example, when the steel has been greatly overheated in some previous operation. The hardening of high-carbon steels is at best a difficult operation, and the less heating to which such steel is subjected the better. Manganese on Hardening. As we have previously mentioned, the presence of manganese causes a greater hardening effect, due to its obstructing the austenite transition. This increase in hardness in ordinary carbon steels with less than 1.75 per cent, manganese is commonly thought to be associated with an increase in brittle- ness, 1 and with the danger of cracking during or immediately sub- sequent to . quenching. Forethought must therefore be used in obtaining the proper combination of manganese, carbon, and rate of cooling to avoid the latter difficulty. The general limits of safety for practical work may be broadly (but not invariably) set somewhat as follows: water quenching is always dangerous when the mangan- ese content runs up around 1.50 per cent., even in low-carbon steels; with approximately 1.00 per cent, manganese water quenching may be used although not advised with mild forging steels; with the progressive increase in carbon the manganese content should be rapidly lowered, so that in tool steels for water hardening the mangan- ese is under 0.40 per cent., and with very high-carbon tools is not over 0.25 per cent. Dependent upon the size and general shape (design) of the piece, as well as the condition (refinement) of the steel, oil quenching is generally safe up to 1.75 per cent, manganese with 0.60 per cent, carbon in fact, one well-known oil-hardening tool steel analyzes about 0.90 per cent, carbon with 1.60 per cent, man- ganese. The subject of high manganese steels will be considered under a separate chapter. 1 Refer to Chapter XV for a further discussion of this point. CHAPTER V TEMPERING AND TOUGHENING TEMPERING Tempering. When a piece of carbon tool steel is heated to a red heat and quenched in water (i.e., hardened), the steel becomes hard, brittle, and is held in such a state of stress that its use except in a few particular cases would be highly inadvisable. This hardening operation has arrested the austenitic transition at the martensitic stage, and prevented it from advancing further, as into troostite, etc. Under these circumstances, the application of heat will now accomplish two results: (1) it will relieve the hardening strains, and (2) permit the transition to proceed. By properly adjusting the temperature of this reheating process, any desired stage in the martensite-troostite transition may be obtained. And by permitting just the right amount of the hard, brittle martensite to go over into the softer and tougher troostite, any desired combina- tion of physical properties within the capacity of that particular steel may be realized. This process of " letting down " or softening is called tempering. Troostite. If the steel has been fully hardened so that it consists entirely of martensite, troostite will begin to form at somewhere in the vicinity of 400 F., or possibly lower. As the tempering tempera- ture is progressively raised, the troostite increases in amount until at about 750 F. it begins to change into sorbite. Thus steel in the tempered condition is usually characterized by the presence of more or less troostite, dependent upon the degree of hardening and upon the tempering. Just as martensite may be said to represent the condition of hardened steel, or pearlite that of annealed steel, so troostite is indicative of a tempered steel whether it be obtained by water quenching and reheating, or by quenching in some less drastic medium such as oil but with no reheating. The question of whether troostite represents a complete step in the transformation is not definitely known, and as far as practical heat-treatment work 06 TEMPERING AND TOUGHENING 97 is concerned is but a question of scientific value; the value of troostite in its influence upon the hardness and allied properties of tempered steel is, however, definitely recognized. Hardening Strains. It should be always remembered that tempering not only softens the steel through the influence of troostite, but also relieves the strains set up in hardening. This last factor should not be lost sight of, for although the proper degree of hard- ness is requisite for specific work, no tool will eventually prove of much value if it retains the state of strain occasioned by rapid cooling. This statement applies not only to water quenching, but also to oil quenching (or oil tempering) . Even the influence of boil- ing water is often sufficient to relieve more or less of these strains, if it is not desired to further soften the steel by higher reheating. Naturally, however, the higher the softening temperature the better will be the condition of the steel in this regard. Temper Colors. Nature has provided a useful and more or less empirical indication of the degree to which tempering has affected the steel through the formation of a surface film of oxide colors (oxide of iron). If a piece of hardened steel is brightened with emery paper or other suitable means, and is then slowly heated with expo- sure to the air, the brightened surface will take on characteristic " temper colors." These commence with a very faint yellow and progressively change with increase of temperature through varying degrees of yellow, brown, purple and blue. That these colors bear a definite relation to, and are closely indicative of, a known tempera- ture, under certain conditions, is now a generally accepted fact. Although a difference in distinguishing the various shades of color is bound to occur on account of the " personal equation," the follow- ing table is fairly representative : Temper- Temper- ature, De- frees Color. ature, De- frees Color. ahr. * ahr. 420 Very faint yellow 510 Brown 430 Yellowish-white or light straw 520 Brown purple (peacock) 440 Light yellow 530 Light purple [450 Pale yellow straw 540 Purple '460 Straw 550 Dark purple 470 Dark Yellow 560 Light blue 480 Deep straw 570 Blue 490 Yellow brown 600 Dark blue 500 Brown yellow , 625 Blue tinged with green 98 STEEL AND ITS HEAT TREATMENT Limitation of Color Method. The previous statement regarding the relation of tempering colors to temperature is true in its entirety only under certain definite conditions of heating, and which are largely dependent upon the time element. So long as the heat of the steel is being progressively raised that is, so long as the temper- ature of the fire, furnace or tempering plate is greater than the temperature of the steel the temper colors indicate the temperature of that part of the steel most affected the surface. But when the steel is being kept at a definite tempering temperature for any length of time, the colors do not represent the actual temperature. This point is readily illustrated by heating a small piece of hardened steel at a constant temperature for a considerable period of time. Thus, in one instance, a straw color was produced in about a minute, but changed to a brown in about ten minutes, and to a purple in about forty minutes; and yet the temperature of the steel was never higher than 460 F., representative of the straw color. In other words, the time element has developed a new set of conditions w.hich may greatly affect the depth of oxidation or color. On the other hand, it is a debatable point as to whether or not these temper colors represent the actual condition (not the temper- ature) of the steel itself. Some tool makers maintain that the efficiency of the tool both in hardness and in other properties is the same whether the color has been obtained by a short heating at a high temperature, or a longer heating at a lower temperature. That is, the ultimate results are indicated by the temper color, independent of the method of obtaining it. Others aver that such is not the case. Tempering for Depth. It is obvious that the temper color is at the best but a surface indication. For some tools or articles which require a specific superficial hardness only, and in which the condition of the center of the tool is of little consequence, it probably does not matter a great deal in the ultimate results whether the temper color a straw color for example has been obtained by a few min- utes' heating at 460 F., or by heating for a longer period at say 360 F. Contrariwise, if the tool or part is to be subjected to stresses of such nature as demand the best that the steel is capable of, the greatest degree of uniformity and release of hardening strains is requisite. Such may only be obtained by a thorough heating at a specified temperature, and which may be entirely independent of the color indication. In such cases, to use the above temperatures, the thorough heating at 360 it more uniformly affecting the whole mass of the steel might prove immeasurably better than the TEMPERING AND TOUGHENING 99 incidental surface heating to 460. And as will be mentioned later, a continued heating at 460 would again be an improvement over either color method. Quenching after Tempering. The method of tempering by color indication inherently requires immersion when the specified color is reached to prevent any further rise in temperature, or in the blacksmith's phrase, to " set the grain." Although it is possible so carefully to heat the steel that the maximum effect is just to develop the color desired and no further, such methods take so much time and patience that they are rarely carried out in practice. The necessity of such immersion or quenching, even in the hands of an experienced hardener, is the source of many troubles. Not only does the quenching probably induce further strains into the steel, but it is also entirely inconsistent with uniformity of results. If the object is of considerable size, or varies greatly in dimension of adjoining sections to be similarly tempered, or is of intricate design, the difficulty in obtaining the same temper throughout even on the surface (to say nothing of the interior of the steel), will be greatly magnified. If the proper color is reached on one part before another, there will be a corresponding difference in hardness. And thus the difficulties multiply ad infinitum. Use of Liquid Baths. Later methods involving the use of liquid baths for heating overcome the difficulties in color tempering, eliminate as a general rule the necessity for quenching, and further give complete uniformity of heating throughout the whole mass of the steel and the maximum elimination of hardening strains as can be obtained at the temperature used. By maintaining the bath at the proper temperature there can be no overheating, the heat must penetrate all parts of the steel alike, and the " personal equation " is as nearly eliminated as is possible. This method has the further advantage of cutting down labor costs and increasing the output, since a number of pieces may be heated at the same time, and while one lot is being tempered another bath may be charged or discharged. Comparison of Physical Properties Obtained. An excellent example of the efficiency of bath tempering is illustrated in auto- mobile gears. On account of the relatively thin section of the teeth as compared with the mass of the gear, exact tempering by the ordi- nary temper-color practice is rather difficult. The teeth, which should be the hardest, take the temper first, and are therefore the softest part of the gear as a whole. If the gears were to be tern- 100 STEEL AND ITS HEAT TREATMENT pered by revolving on a hot bar much better results would be ob- tained than by ordinary tempering, but the time and cost elements would prove excessive where hundreds of pieces were to be handled. By the use of a suitable liquid tempering bath thorough uniformity could be obtained throughout. Where by the color method, the core of the gear would have the tendency to be too hard, the teeth per- haps too brittle or soft in places, and only the surface of the gear as a whole affected by the temper-color representing say 475 F., by the more modern method the whole mass of the gear would have the physical properties as characterized the drawing temperature of the 475 F. Exact Temperatures. Too much attention cannot be given to the necessity of obtaining exact temperatures in the tempering opera- tion. For the average run of carbon tools the tempering range is very narrow, probably within a hundred degrees for the great majority. The tempering action takes place extremely rapidly and often a difference of 15 or 20 may cause much trouble. Trying to temper tools over an open fire may be all right in isolated cases, but it spells failure if made a general practice. Tempering Methods. The procedure to be employed in temper- ing must necessarily depend upon the nature of the tool or part. Methods must be developed to satisfy the individual requirements and are too numerous to discuss here. Briefly, however, the more common practices may be covered by the tempering plate, the sand bath, and such liquid baths as oil, lead and alloys, and molten salts. Tempering Plate. The tempering plate generally consists of an iron casting planed smooth on top, and heated from beneath by suitable means, such as gas, oil, or even a coal or coke fire. The steel articles are placed on the plate and moved about until they have attained the proper temper color and then quenched. Fig. 59 shows a characteristic equipment for heating, hardening and temper- ing dies; Q represents the discharge end of the heating furnace, R the quenching tank, and T the tempering plate, the latter being- heated by oil burners from beneath. Sand Bath. In order to effect more uniform tempering of small tools, a pan of clean, well-dried sand may be placed on a suitable hot-plate, or in a furnace. The sand is held at the desired tem- perature, which may be determined by the insertion of a ther- mometer or pyrometer couple, and may be protected by covering with a suitable hood. The oxide colors on the steel may also be TEMPERING AND TOUGHENING 101 used as a measure of the tempering, as there is of course free access of air between the particles of the sand. Oil Baths. For much of the ordinary tempering work an oil bath will probably prove as satisfactory as any method for temper- atures up to about 500 F. or even higher. The chief requisites are a tank holding an ample supply of oil, a suitable furnace or method of heating by which accurate and constant temperatures may be obtained, and a mercury thermometer for determining the tempera- ture of the oil. Mineral oil with a flash-point of some 600 F. is FIG, 59, Quenching and Tempering Dies, (" Machinery.") generally used for the bath ; certain of the animal and vegetable oils are also occasionally used. Handling the Material. Oil baths, and similarly the salt baths, are provided with a wire basket in which the pieces to be tempered are placed and which is then lowered into the oil. By this method a number of pieces may be tempered at once, besides preventing the steel from coming in contact with the sides or bottom of the tank, which is apt to be hotter than the oil. It is advisable, whenever possible, to allow the hardened steel to 102 STEEL AND ITS HEAT TREATMENT come up gradually to the desired temperature, and not to immerse in the oil when the latter is already at the highest heat. Rather put the steel in the oil when the latter is about 200 to 300 F. and let the two heat up together. The reason for this is that the pre- heating if it may be thus termed allows the heat to penetrate more gradually, softening the outer portion of the steel in such a way that the inner and stressed part may be more gradually relieved and thus avoiding the danger of fracture. Sudden heating has the tendency to set up new stresses which must in turn be overcome. The length of time allowed for the tempering to take place will depend upon the size and nature of the piece under treatment; fifteen minutes or so after the maximum temperature has been reached will generally be sufficient for the average run of small tools, gears, etc., while larger parts require more time in proportion. If large and small parts are tempered at the same time it will do no harm to the small pieces if they are not removed until the larger pieces are ready, although on general principles long-continued heating is never desirable after the steel has responded to the desired heating. When the full effect of the tempering has been attained, the pieces may then be removed from the oil and allowed to cool off in the air, for if the steel has been thoroughly heated at the maximum temperature of the tempering operation, no further change will take place in the ordinary steels; each phase of the transition is represented by a definite temperature for each steel, so that no further step in the transition will occur unless the temperature is raised with the possible theoretic exception of very long-continued heating. For some large work, such as die blocks, large cutters, etc., the steel is allowed to cool off in the oil in order to procure the greatest elimination of strains. Salt Baths. If higher drawing temperatures than those possible with oil are desired, a bath of salts may be used. A combination of two parts of potassium nitrate and three parts of sodium nitrate melts at about 450 F. and may be used up to about 1000 F. Methods of heating and using are similar to those with oil baths, and described under Hardening Baths. The use of nitrate salts instead of the chloride salts is necessary on account of the lower tempera- ture desired. Lead Baths; Alloys. Lead, having a melting-point of about 610 to 630 F., may also be used for tempering where temperatures higher than its melting-point are required. The disadvantages are similar to those noted under its use for heating for hardening. The TEMPERING AND TOUGHENING 103 melting-point may be lowered by alloying the lead with tin, and temperatures suitable for ordinary tempering may be obtained approximately as follows: l Lead Parts. Tin Parts. Approx. Melt- ing Temp. F. Lead Parts. Tin Parts. Approx. Melt- ing Temp. F. 14 8 420 28 8 490 15 8 430 38 8 510 16 8 440 60 8 530 17 8 450 96 8 550 18 5 8 460 200 8 560 20 8 470 Melted lead 610 to 630 24 8 480 The use of these various alloys of predetermined melting-points for tempering is similar to that previously explained when selecting a combination of salts with certain melting-point in the hardening operation. TOUGHENING Sorbite. As the reheating or drawing temperature is increased still further beyond the tempering range we find that another stage in the austenitic transition commences the change of troostite into sorbite. Like the change from martensite to troostite, the formation of sorbite does not take place spontaneously throughout the whole steel, but increases gradually and progressively. Most writers believe that sorbite is essentially an uncoagulated conglomer- ate of irresoluble pearlite with ferrite in hypo-eutectoid (less than about 0.85 per cent, carbon), and cementite in hyper-eutectoid steels respectively, but that it often contains some incompletely trans- formed matter. Its components at all times tend to coagulate into pearlite. On higher heating, sorbite changes into sorbitic pearlite, then slowly into granular pearlite, and probably indirectly into lameliar pearlite. Sorbite differs from troostite in that it is softer for a given carbon content, and in usually being associated with pearlite instead of martensite, and from pearlite in being irresoluble into separate particles of ferrite and cementite. Importance of Sorbite. The main importance of sorbite is due to its physical properties. Although slightly less ductile than pearl- itic steel for a given carbon content, its tenacity and elastic limit are so high that a higher combination of these three properties can be 1 Table by O. M. Becker, using melting-point of lead as 610 F. 104 STEEL AND ITS HEAT TREATMENT had in sorbitic than in pearlitic steels. Steels which are so treated as to contain sorbite are often called " toughened " steel. Toughening Range. The transition of troostite the chief characteristic of tempered steel, into sorbite characteristic of toughened steel, is gradual, and progresses with the increase and duration of the reheating. At some point, depending upon the composition of the steel and the degree to which the steel has been affected by the hardening process, sorbite is formed. If we accept sorbite as the characteristic constituent of toughened steel (and which it undoubtedly is), we may then consider as the lower limit of the toughening range that temperature which will produce sor- bite. In fully hardened steel of the medium forging and higher carbon analyses, characteristic sorbite begins to form at about 750 F. At about 1250 to 1300 F. the sorbite coagulates into pearlite, which is distinctive of annealed steel. With these facts in view we may then consider, in a general way, that the toughening range lays approximately between 750 and 1250 F. It must be remembered, nevertheless, that these temperatures are in no sense definite, but are arbitrarily taken as representative of a class of heat- treatment work: differences in chemical composition, the degree of hardening, the size of work, etc., all play their part. Influence of Toughening. When a piece of hardened steel is reheated for toughening, each specific temperature has a certain definite influence upon the steel. The results of this toughening process are interpreted by the ability of the steel to do certain work, to withstand the application of stated loads, or as measured by standard methods of testing. On account of the almost universal use of the last named for purposes of comparison, we will deal briefly with (1) the static strength, as measured by the tensile strength and elastic limit, (2) the ductility, as measured by the percentage elongation and reduction of area, and (3) the dynamic strength, as measured by the alternating impact test. Effect of Increased Temperature. Each increase in the toughen- ing temperature lowers the tensile strength and elastic limit, but with a corresponding increase in the ductility and dynamic endurance. With the majority of ordinary carbon, nickel, chrome and vanadium steels the ratio of the elastic limit to the tensile strength remains very nearly constant throughout the sorbitic range (which we assumed to be approximately from 750 to about 1250 F.). Be- yond these temperatures, and coincident with the formation of pearl- ite, the values for the elastic limit and tensile strength of each par- TEMPERING AND TOUGHENING 105 106 STEEL AND ITS HEAT TREATMENT ticular steel begin noticeably to diverge until they reach their smallest ratio in fully annealed steel. Up to near the end of the sorbitic range the graphs obtained by plotting the elastic limit and tensile strength against the drawing temperatures are, for general purposes, straight lines, but beyond this range curve towards the horizontal, as is represented in Fig. 60. Effect on Ductility. These changes are accompanied by reverse changes in the ductility, as measured by the reduction of area and elongation. As interpreted by the research work of others, and from his own experimental work, the author is inclined to the belief that these two factors differ from each other in that the reduction of area generally reaches a maximum at about the end of the sorbitic range and then decreases, while the elongation does not attain its maximum until the steel is fully annealed or in the pearlitic condition. Be this as it may, through the sorbitic stage at least, each increment of decrease in tensile strength and elastic limit is associated with, and counterbalanced by, an increase in the reduction of area and elongation. This combination of static strength and ductility is further almost directly proportional to the toughening temperature. Impact Strength. The effect of toughening upon other properties and especially in relation to the impact strength, is shown in Fig. 60, rearranged from the work of Grard. The steels, approximating 0.15, 0.40 and 0.50 per cent, carbon, were hardened and then re- heated to temperatures varying from no tempering up to 2200 F. The impact strength curves present some extremely interesting facts. We find that the greatest resistance to shock to be obtained from a toughening, after hardening, at a temperature about 100 F. below the upper critical range (Ac3); annealing at a temperature superior to the Ac3 range gives a lower impact strength. Further, as the temperature is raised more and more and overheating results, there is a marked diminution in the impact strength. Increase in the carbon content, assuming the same heat treatment, diminishes the impact strength. Tempering (reheating up to say 600 F.) has little or no effect upon the impact strength. As a general prop- osition we may sum up by stating that it is preferable, in order to obtain the greatest impact strength, to keep the 'carbon content as low as possible and to have a high drawing tempe- rature. Capacity of the Steel. Thus it will be seen that by changing the drawing temperature the grouping of these factors may be varied TEMPERING AND TOUGHENING 107 through a considerable range and limited only by what we may call, for want of a better phrase, the " capacity of the steel." This quantity is defined largely by the chemical composition, the method of manufacture, the size of the piece to be treated, and by other subordinate factors. With these qualifying conditions in mind, we may further define the capacity of the steel as the limiting ratio of strength to ductility. Each steel, as qualified above, has certain definite limits within which the physical properties may be varied. At one end of the see-saw, as in hardened steel, there is a maximum tensile strength with minimum ductility; and at the other extreme, as in fully annealed or sorbitic-pearlitic steel, there will be a mini- mum tensile strength with maximum ductility. Following out the simile of the see-saw, we may place tenacity on one end and ductility on the other; when one is up, the other must be down; both cannot be up nor both down at one and the same time; raise one and the other must fall. The heat-treatment man now stands on the middle of the board and by means of his reheating temperature can adjust the opposing factors to that position which he desires; but he cannot change the maximum and minimum of either, because they are fixed by the limitations previously mentioned at the beginning of the paragraph, and over these he has no control as far as the individual stael is concerned. Duplication of Results. Happily for the heat-treatment man, each grouping is distinctive of a definite toughening temperature, other conditions being the same. When he has once determined the relation existing between static strength, ductility, and temperature, for a given size piece of work made from a steel of specific analysis, he knows that he can approximately duplicate his results under like conditions at any time. Not that he can absolutely and ultra- scientifically obtain results within a few pounds elastic limit or hun- dredths of a per cent, elongation for such are neither necessary nor expected but that he can reasonably expect to get a com- mercially acceptable duplication. It is with this thought in mind that the subsequent chapters have been developed, giving under each steel many of the results and details which have been obtained in practice and experiment, and which should prove advantageous to the average heat-treatment man as a time-saver. Slow Cooling and Stresses and Strains. It is one of the incon- trovertible facts of heat-treatment work that slow cooling predicates the release of internal stresses and strains. Not only is this true of the full-annealing process as indicative of slow cooling from a 108 STEEL AND ITS HEAT TREATMENT temperature above that of the critical range, but also of the toughen- ing operation. In fact, the very nature of the usefulness of tough- ened steel depends upon the absence of a state of strain just as much as upon specific static or dynamic properties. Strange as it may seem, some of the failures in locomotive forgings may be traced back to the lack of slow cooling after toughening; and this trouble is coming to be recognized in many specifications by the require- ment of cooling in the furnace after toughening. Just as the dangers in hardening increase with the rapidity of cooling, carbon content and size of section, so are they likewise magnified in cooling after toughening although on a smaller scale. If these factors become noticeably important, cooling in air from the toughening tempera- ture may set up such a new series of cooling strains that many of the real advantages of toughening may be invalidated. Use of Furnace Cooling. The greater part of hardened and toughened work, such as automobile and other small forgings, may not require furnace cooling, besides being economically impracticable. But even with these it is desirable that the pieces should be piled together after removal from the furnace so that the cooling will be retarded. For forgings of section greater than 3 or 4 ins., such as heavy machine parts, ordnance, etc., cooling in the furnace is always desirable. It may be said that such slow cooling never did any harm, and it may do a world of good in relieving strains. Effect of Furnace Cooling on Physical Properties. Contrary to the opinion held by some, the author does not believe that slow cooling in the furnace has any noticeable tendency to further " soften " the usual straight carbon or alloy steels to which the toughening process is generally applied. That is, for similar pieces of the same steel treated alike, equivalent physical test results would be obtained in the forging which had been furnace cooled as in the one which had been allowed to cool in the air the tests being taken from the same relative position. In making this statement there is, however, one other necessary qualification: it is assumed that the whole mass of the steel has been thoroughly heated at the tough- ening temperature. Otherwise the effect of the toughening would not be so great in the air-cooled piece as in the slowly cooled piece, for the latter would have greater opportunity to be affected by the heat of the furnace during the furnace cooling. During the tough- ening range the effect of the heat upon the transition, except for very large pieces, practically ceases as soon as the source of heat is removed as by air cooling. TEMPERING AND TOUGHENING 109 High vs. Low Toughening Temperatures. On the hypothesis that either of two specified analyses would prove equally satisfactory, under suitable treatment, for the same piece of work, but that on account of the difference in chemical composition one steel would require toughening at say 1200 F. and the other at say 800 or 900 F., the selection of the higher drawing-point steel should be ma/le. Such conditions often arise in heat-treatment plants handling a variety of commercial work and it may be well to sum up briefly the reasons for the above conclusion. The more stable the state of equilibrium which exists between the transition constituents the more lasting and effectual will be the treatment. Further, the smaller the amount of internal strains which may remain in the steel from the previous hardening operation the better. Both of these conditions are more nearly brought about by the higher drawing temperature. As there is also a decided tendency for the dynamic strength to reach a maximum at about 1200 to 1300 F. it is probable that the higher drawing temperature steel will have a greater dynamic strength than the other steel, provided that there is not too much difference between the chemical compositions of the two steels. From the furnace man's point of view the temperatures around 1200, being of characteristic visible reds, are decidedly more easily recognizable than those temperatures around 800 to 1000, since with these lower temperatures there is very little visible heat color. The higher drawing temperatures therefore aid in the efficiency of judging the heating operation and lead to greater uniformity of control and of results. Quenching Medium vs. Toughening Temperature. There is another phase of the high or low toughening temperature proposition which cannot be solved by any general rule, but only after due consideration of all the circumstances involved; this relates to the condition of affairs when there is no opportunity for the choice of steel, but depends more upon the selection of the quenching medium in relation to the toughening temperature. As we have noted, water quenching gives a harder steel than oil quenching. It natu- rally follows that, in order to obtain approximately the same physi- cal results, the oil-quenched piece must be drawn at a lower temper- ature than the water-quenched piece. The arguments regarding water vs. oil quenching, and low vs. high drawing temperatures have been previously discussed. If the solution were to be developed entirely along these lines it is orobable that in the majority of cases; 110 STEEL AND ITS HEAT TREATMENT the oil quenching (giving less hardening strains) and lower drawing temperature would be employed. In other words, the difficulties to be encountered with water quenching the hardening operation being the more drastic of the two would more than outweigh the the disadvantages of the lower toughening temperature. This is a question in which the personal element and experience of the heat- treatment man would be paramount. Influence of the Carbon Content. In respect to the selection of the steel in relation to the treatment there remains the consideration of the influence of the carbon content. Carbon not only intensifies the effect of the rapid cooling (hardening), but it also directly augments the brittleness of the steel. Or, to put it in other words, the greater the carbon content the greater the hardening strains, and the lower the ductility which can be obtained with a stated tensile strength. It is therefore usually desirable to provide a steel with as low a carbon content as will give the desired results. Toughening vs. Annealing. It is only within comparatively recent years that the toughening process with its attendant sorbitic structure has been used and understood. Previously, annealing was generally the cure-all for brittleness and a strained condition of the steel. Pearlite produced by annealing on account of its entangled structure, gives a large measure of ductility; but also gives a minimum tenacity. The appearance of sorbite, however, is even more entangled than pearlite; sorbite is far superior to pearlite in tensile strength and especially in elastic limit. Thus by obtaining a sorbitic steel by suitable treatment, almost as much ductility, greater working strength, greater dynamic strength, and by being able to use a lower carbon steel less brittleness may be obtained than in a pearlitic or annealed steel. Standardization of Results. With the same degree of hardening, and if the reheating has been uniiorm and thorough at a given tem- perature, the physical results will be comparatively the same for material of equivalent section and the same composition. That is, the product will be standard for standardized treatment. Fur- ther, in order to get standard results with steel purchased under the same general specifications (i.e., each chemical constituent within certain limits) , the toughening temperatures may be varied according to the chemical composition. To illustrate: the following heats of steel of varying chemical composition and made by several steel companies were manufactured into a certain product which, when heat treated, required an elastic limit of 85,000 to 95,000 Ibs. per TEMPERING AND TOUGHENING 111 square inch, and an elongation of not less than 16 per cent, in 2 ins. In spite of the varying carbon, manganese, chrome and nickel contents, the toughening temperatures (maintained within 5 F. under or over) were so adjusted as to give the desired results. Thousands of pieces, some weighing as much as 200 Ibs., all ful- filled, by actual test, the standard physical specifications. Carbon. Manga- nese. Phosphorus. Sulphur. Chrome. Nickel. Toughen- ing Temp. Deg. Fahr. 0.16 0.43 0.015 0.017 0.62 1.82 1050 .185 .44 .010 .015 .57 1.56 975 .20 .43 .009 .016 .64 1.74 1025 .20 .46 .011 .017 .40 1.56 950 .21 .48 .015 .015 .60 1.77 1050 .21 .50 .017 .014 .67 1.84 1075 .23 .50 .016 .018 .65 1.73 1075 .245 .53 .015 .020 .62 1.79 1120 .25 .50 .011 .019 .64 1.40 1140 .26 .43 .010 .018 .60 1.65 1100 27 .49 .015 .021 .63 1.79 1150 .28 .51 .008 .011 .41 1.57 1120 Quench-Toughening. A process which has been used consider- ably for the treatment of large forgings of uniform section, such as heavy axles, is that of heating as usual for hardening and then quenching in oil for a specified number of seconds, followed by air cooling. The oil quenching affects the steel to a certain depth, but still leaves a considerable amount of heat in the forging when removed from the bath. As the forging cools in the air this heat from within will toughen or " soften " the steel affected by the quenching. In order to obtain equivalent results under varying conditions the number of seconds required for immersion in the oil of a piece of given size must be determined by experiment and strictly adhered to. Forgings treated by this process are characterized by a soft or annealed core, with a progressively toughened outer part. Physical Results. In subsequent chapters will be given results obtained in actual practice by the use of various toughening tem- peratures for different grades of steel, CHAPTER VI 1 CASE CARBURIZING Object of Case Hardening. The object of case hardening or partial cementation is the production of a hard wearing surface (the " case ") on low carbon steel, and at the same time the retention or increase of the toughness of the " core " of the metal. The process may be roughly divided into two distinct periods. First, the car- burization or impregnation of the surface by which the carbon con- tent is sufficiently raised dependent upon the demands of the work so as to give a steel capable of taking on very great surface hardness. Second, suitable heat treatment which shall develop the properties of both case and core. The complete operation should not only result in the obtaining of a very hard case, but also and simultaneously in the realization of special mechanical properties in the core more especially that of non-brittleness. Briefly, the aim is to have a piece of steel which shall possess a minimum fragility and a maximum surface hardness. Requirements for Case Carburizing. In order to obtain a case rich in carbon, the metal is heated in the presence of a body which is capable of delivering this carbon, by more or less complex reactions, which is then dissolved by the steel. Aside from the use of gases in the newer processes involving such factors as pressure, quantity, etc., there are four main factors which must be considered in the carburizing operation: 1. The solvent: that is, the steel; 2. The product to be dissolved, or more exactly, the compound capable of delivering the carbon, i.e., the cement; 3. The temperature; 4. The time of contact between the steel and the carburizing agent. 1 Cuts by Giolitti from " The Cementation of Iron and Steel," by courtesy of McGraw-Hill Book Co.; references made in this chapter to investigations by Giolitti are also from the above. 112 CASE CARBURIZING 113 THE STEEL The Steel. The character of the initial steel used for case car- burizing depends largely upon the fact that one of the main desires is to eliminate brittleness in the core. We have seen that any increase of carbon, other conditions being equal, will increase the brittleness, particularly when the carbon content is raised to over about 0.25 per cent. Further, as practically all commercial car- burizing processes involving case hardening are followed by one or more hardening operations, it follows that the use of a steel with a higher carbon content will also increase the brittleness through quenching. For these reasons it is therefore necessary to keep the carbon content of the steel to be carburized quite low, prefer- ably under 0.25 per cent, for straight carbon steels. In fact, the best French practice is to demand a carbon content of not over 0.12 per cent., further qualified by the specifications that the core after quenching shall give a tensile strength of about 54,000 Ibs. per square inch and not to exceed 85,000 Ibs., together with an elongation of 30 per cent, in 100 mm. (3.94 ins.). However, one of the important and often unsatisfactory results of using an extra-soft steel is the difficulty encountered in machining (before carburizing) . If the carbon is extremely low the steel is very apt to tear, and thus increasing the amount of grinding after hardening in order to obtain a perfectly smooth surface. For this reason, the general American practice is to adopt a carbon content about midway between the extreme upper and lower limits and specify a steel with about 0.16 to 0.22 per cent, carbon. The higher carbons also give increased stiffness to the core which, in some cases, is necessary. It is generally recognized that the carbon content, at least up to some 0.50 per cent., has no influence upon the velocity of penetration of the carburization, i.e., the depth of carburization which will be obtained for a given length of exposure. On the other hand, the initial carbon content of the steel will have a decided influence upon the maximum carbon content which will be obtained in the case; the higher the initial carbon, the higher the maximum carbon concentration in the case. Manganese. It is considered the best practice, in general, to require a low manganese content with about 0.30 to 0.35 per cent, as the maximum. It should be remembered that the case which will be formed during the carburization will be characteristic of 114 STEEL AND ITS HEAT TREATMENT a high-duty tool steel and will have the properties of such. Thus manganese will increase the hardness of the case (and also of the core) and will make the steel as a whole more sensitive to rapid cool- ing. In spite of this, it is often customary, especially in British practice, to use a manganese content of about 0.70 per cent. and in some cases even up to 0.90 per cent. in order to obtain greater stiffness in the core. Manganese at such percentages also increases the brittleness produced by long heating during carburization, and diminishes the efficacy of the regenerative quenching. These last named points are also true when the silicon is much over 0.30 per cent. Other Impurities. It is self-evident that the content of phos- phorus and sulphur in the initial steel should be just as low as is possible. Slag, blow-holes, segregation, and all other impurities and imperfections should be entirely absent from steels for case hardening. THE CEMENT Direct Action of Carbon. Carburization by its very nature requires the presence of free carbon in some form or other, either as a solid body, or as some gas which will produce free carbon by its decomposition. The mere presence of free carbon in contact with iron, however, will not satisfy the conditions necessary for commercial carburization. Although it has been shown scientifically that car- bon alone, without the intervention of any gas, will carburize iron if it is kept in contact with it for a sufficiently long time and at a sufficiently high temperature, this direct action, as far as industrial results are concerned, is negligible. That is, the ordinary forms of solid carbon, such as wood charcoal, sugar charcoal, etc., exercise directly on iron but a very slight carburizing action in the absence of gases. Action of Gases. It will be noted that emphasis has been laid upon the " direct action " in the " absence of gases." This at once leads to the question as to what is meant by the action of gases, and which, in turn, involves the mechanism of cementation itself. It is a well-known fact that when steel is heated, the " pores of the steel are opened " to use the vernacular expression it becomes easily permeable to gases, and the surrounding gases diffuse into the steel. This is true whether the steel is heated in the ordinary atmosphere, when the gases consist of nitrogen and oxygen, or whether it is heated in some specially prepared atmosphere, such as carbon monoxide, CASE CARBURIZING 115 illuminating gas, etc. The main fact to be realized is that the gases do penetrate into the steel, although the effect of the gases will depend upon the composition of the gas, besides such other factors as pressure, temperature, and so forth. Thus, recognizing that the direct action of carbon that is, the carburizing results obtained by mere contact of carbon with iron is commercially negligible in the absence of gases, it is evident that carburization must be intimately related to the presence of gases. In other words, the gases (or, more exactly, certain gases) must in themselves act as the carrier or vehicle for the carbon. That this carrier action, or transporting of the carbon, has not been definitely recognized or determined until recently has been due to the fact that practically all of the solid cements generate the necessary gases through their own decomposi- tion and interaction with the occluded air. Further, the intense and critical study of this action has been developed only by the research work in connection with the newer processes of case carburizing by means of gases alone. Action of Oxygen. As a typic 1 example of this diffusion and its effect we may consider any ordinary carburization process in which wood charcoal is used as the base cement. When the carburizing material and articles are packed in the carburization boxes there is necessarily a considerable quantity of air also occluded with the particles of the cement. Under the influence of heat the oxygen of the occluded air will react with the carbon or charcoal to form car- bon monoxide gas, which has the symbol CO. Then, as the tem- perature of the box and contents increases to the temperature of the carburization proper, these gases of carbon monoxide permeate or diffuse through the surface and outer section of the steel. At the same time, by catalytic action, the carbon monoxide gas decom- poses when it comes in contact with the steel and sets free a part of the carbon it contains. This decomposition may be represented by the reversible reaction 2CO < C0 2 + C carbon monoxide^carbon dioxide (gas) -+- carbon (solid). Thus, as the gas diffuses into the mass of the steel it continues to decompose, setting free new quantities of carbon within the steel. This carbon, at the proper temperatures of carburization, passes directly into solution in the steel and forms a true steel proper. The reaction above, being reversible as might be shown will continue indefinitely under suitable conditions, the charcoal regenerating the 116 STEEL AND ITS HEAT TREATMENT supply of carbon monoxide. Further, while it is a well-known fact that carbon monoxide, acting alone on iron, will deposit free carbon on the surface of the iron, this action takes place only at temperatures lower than those ordinarily used for commercial cementation. In other words, the carburizing action of charcoal as used in practice is not due to the direct action of the carbon, but is due (under the conditions named, which of course may be modified by the presence of other gases or components of the cement) entirely to the specific action of carbon monoxide as a gas. Nitrogen. The action of the oxygen of the occluded air being accounted for, the accompanying constituent nitrogen must be con- sidered. Although it has been shown that during carburization the nitrogen may and will diffuse in small amounts into the steel, it is now certain that the presence of pure nitrogen does not increase, except to a minimum extent, the carburizing action of free carbon. In fact, instead of nitrogen being requisite as many still believe it may even exert a pernicious effect. LeChatelier has suggested that the increase in brittleness sometimes observed in those parts of the steel subjected to cementation, but which the carburization has not even reached, may be due to this nitrogen. It might be added that this deleterious nitrogenizing theory is further supported by experi- ments along other lines particularly in the apparent cleansing effect for nitrogen of the titanium additions to steel during manufacture. Another general effect of nitrogen gas is to reduce the cementing action of the carbon monoxide mentioned by its diluting the car- burizing gas. For practical purposes of carburization, however, the action of nitrogen in the presence of free carbon is too slight to influence commercially the results obtained with a given cement, unless actually added (in gaseous cementation) as a diluent. Carbonates. The ash of the carbonaceous matter may also contain carbonates of the alkali or alkaline-earth metals. Or these carbonates, such as barium carbonate, may be added directly to the cement. In the light of the most reliable and recent researches it would appear, contrary to previously accepted theories, that the activity of these carbonates is not due to the formation of volatile cyanides by the action of the nitrogen of the occluded air, but exclu- sively to the formation of carbon monoxide produced by the action of the hot carbon on the carbon dioxide produced through the dis- sociation of the carbonates. Thus the effect of such carbonates is similar to that produced by carbon monoxide under similar con- ditions. CASE CARBURIZING 117 Cyanides. The most maligned constituents of cements are the cyanogen group. In the past it has been thought that the deriva- tives of this group played the chief part in carburization processes. This, however, has been strongly disproved by Giolitti, who ad- mirably explains the matter as follows: That cyanogen and the more or less volatile cyanides can cement iron intensely is beyond doubt. Moreover, it is well known that fused potassium and potas- sium ferrocyanide are used in the pure state to obtain thin and strongly carburized zones (as in superficial carburization or cyanide hardening) . In industrial practice the cyanides do not exist already formed, but may be formed in very small quantity by the action of the nitrogen of the air (occluded in the cement) on the carbon used and on the small quantities of alkali constituting a part of the ashes of this carbon. Although the formation of small quantities of alkali cements cannot therefore be wholly avoided in industrial car- burization with carbon as a base, the part which is played by these traces of volatile cyanides is certainly negligible in comparison with that of the carbon monoxide formed by the action of the air on the carbon used as cement. He then submits conclusive proofs to substantiate these statements. Carbon Monoxide Gas. Carburization carried out by the use of carbon monoxide gas alone will give a mild or gradual carburization in which the maximum carbon content is comparatively low not usually reaching the eutectoid ratio even at the periphery and which diminishes progressively and in a uniform and slow manner passing from the surface of the case toward the interior of the car- burized piece. Carburized zones of this type correspond always and only to carburization carried on with pure carbon monoxide, a concentration-depth diagram of which is shown in Fig. 61. On account of its definite chemical composition and simplicity of action, the general behavior of carbon monoxide is known within almost exact limits. The carburizing action is easily regulated, and the case may be obtained with certainty with any kind of steel in com- mercial use. When working under suitable conditions, carbon monoxide either alone or with a mixture in which the carbon monoxide can exer- cise its maximum carburizing action will give the greatest velocity of carburization, i.e., the depth reached in a given time by the car- burized zone. This depth is also a direct function of the time or length of exposure. All other conditions being equal, the higher the temperature of 118 STEEL AND ITS HEAT TREATMENT carburization using carbon monoxide, the smaller will be the maxi- mum carbon content of the case. Similarly, the lower the pressure of the carbon monoxide, the smaller the maximum carbon content; and the greater the quantity of pure carbon monoxide gas coming in contact with a unit of surface, the greater the carbon concentra- tion. Under suitable conditions, carbon monoxide gas will deposit no carbon on the surface of the steel being carburized, so that there is little difficulty in keeping the surface bright. Further, the use of carbon monoxide reduces to a minimum the deformations and varia- tions in volume due to the carburizing processes. Carbon monoxide also lends itself in obtaining a good protection of the parts of the steel which it is not desired to carburize. i.o 0.8 0.6 0.4 0.2 0.5 1.5 2.5 3 MM. FIG. 61. Carburization at 2010 F. for Ten Hours with Carbon Monoxide. (Giolitti.) Hydrocarbons. Most of the forms of solid carbon used in prac- tical carburization are not pure, but may contain organic residues not wholly decomposed, or considerable proportions of ash rich in cer- tain carbonates. Thus charred bone, charred leather and similar organic products often used, will, under the influence of heat, evolve hydrocarbons. These hydrocarbons, by more or less complex reactions, deposit the excess of finely divided carbon which they con- tain on the surface of the metal; and this, in turn, being in perfect contact with the metal, at high temperatures may cause a direct carburization by contact. But further and vastly more important than this direct action of the carbon deposit on the surface of the metal, is the carburization by means of the specific action of the gas itself, although of course depending more specifically upon the exact conditions of carburization. In a manner somewhat analogous to CASE CARBURIZING 119 that of the decomposition of the carbon monoxide within the steel, yielding carbon directly to the steel, the hydrocarbon gases will also diffuse into the steel and there yield carbon. Hydrocarbons there- fore also act as carriers for the carbon and effect a carburization due to the specific action of the gas. Carburization with pure hydrocarbon gases give cases of a type corresponding to Figs. 62 and 63, and to Fig. 64. These are characterized on slow cooling by (1) a layer or zone of hyper-eutectoid steel consisting of free cementite and pear lite ; (2) by a layer of eutec- toid steel, generally quite thin; and (3) by an internal layer of hypo-eutectoid steel. The main points to be noticed are, that 5.' ...dr. .."-*.,'-. ?- * V."^"iV. >ty FIG. 69. Carburization of 0.11 per cent. Carbon Steel at a Temper- ature under the Upper Critical Range with BaCO 3 and Char- coal, for One Hour. (Nolly and Veyret.) >' ** * '!A** V * ' ^ "* V * V*'^ ! K"*\ > * " '^ f 'C^i^ ^ *" * V; "'' ^*SJ y *' 4 ^*^'"''^^f FIG. 70. Carburization of 0.11 per cent. Carbon Steel at a Temper- ature Considerably Over the Upper Critical Range with BaCOa and Charcoal, for One Hour. (Nolly and Veyret.) perature before being quenched. This variation consists of a true liquation of the cement ite (and of the ferrite) during their segrega- tion front the solid solution. Take, for example, the diagrams in Figs. 71 and 72, representing the results obtained by carburizing a 0.26 per cent, carbon steel for four hours at 1830 F. in ethylene, with the difference, however, that the carburized steel represented by Fig. 71 was cooled during 32 minutes to a temperature of 1380 F. and then quenched, while that of Fig. 72 was quenched immediately following Carburization from that temperature (1830). Comparing the two diagrams we see that, while the concentration of the carbon in Fig. 72 decreases continuously and uniformly as we 130 STEEL AND ITS HEAT TREATMENT proceed from the surface towards the core, that in Fig. 71 shows a marked increase, followed by a very rapid decrease, before it exhibits *c. 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 / / \ ^ ^ \ ^ - _,___ """"- \ *N^^ \. ^_ 1.4 1.0 1.8 2.2 MM. FIG. 71. Liquation of Cementite through Slow Cooling. (Giolitti.) that gradual decrease which characterized the first case. As the carburization proper was identical in both cases, it is evident that JfC. JL6 2 1.0 0.8 0.6 *& - - - &/.: t -"^: *^ FIG. 74. Carburization with Char- coal for One Hour at 1560F., 0.11 per cent. Carbon Steel. (Nolly and Veyret.) ?-i/vv: y.vtj": r r ..,, - -v. ^'Kisi '^'tit/rfv:^^^^^ FIG. 75. Carburization with Char- coal for One Hour at 1925 F., 0.11 per cent. Carbon Steel. (Nolly and Veyret.) charcoal with definite proportions of these animal charcoals, the carburizing action may be roughly adjusted between the minimum value to be obtained with wood charcoal and the maximum value of the animal charcoals. Of the more common " mild " cements thus obtained we may mention the following: Parts a. Powdered oak charcoal 5 Powdered leather charcoal 2 Lampblack 3 b. Wood charcoal 7 Animal charcoal .3 CASE CARBURIZING 137 Parte c. Powdered beech charcoal 3 Powdered horn charcoal 2 Powdered animal charcoal 2 Common Salt. Common salt (sodium chloride) is used in many works in addition to charcoal, it seeming to give better results than wood charcoal alone. Exactly what is its specific action is not thoroughly understood. Thus we have the mixture: Parts Wood charcoal 7 to 9 Common salt 3 to 1 Barium Carbonate. One of the best solid cements for general use is that consisting of: Parts Barium carbonate . . 40 Powdered wood charcoal 60 Its action is well known and is as we have previously described. For cases of small depths it gives carburized zones markedly more homogeneous than those furnished by other solid cements. Giolitti sums up its advantages as follows: " In general, the maximum con- centration of the carbon in the cemented zones obtained with carbon and barium carbonate at temperatures between 1650 and 2010 F. varies from a minimum of about 0.7 per cent., for the very thin zones obtained near 1650, to a maximum of about 1.3 per cent, for the zones thicker than 1 mm. (0.04 in.) obtained near 2010 F. " Another advantage of this cement lies in its property of being ' regenerated ' easily and spontaneously when it is left exposed in a thin layer to the air, after having been used in the usual manner. This process of regeneration is due to the fact that the barium oxide formed during cementation by the dissociation of the barium car- bonate, absorbs carbon dioxide from the air, again forming barium carbonate. After a certain number of alternating cementations and regenerations it is necessary to add some wood charcoal to the cement to replace that burned during the cementation and during the dis- charging of the boxes. " The preparation of this cement consists simply in finely grind- ing and intimately mixing the wood charcoal and barium carbonate. If the natural barium carbonate (witherite) is used, it is necessary to powder it carefully before adding it to the carbon; the finely 138 STEEL AND ITS HEAT TREATMENT divided precipitated barium carbonate, on the contrary, can be mixed directly with the granulated carbon and the one operation of grinding the carbon can be used for preparing the mixture." It is, of course, not always necessary to use the above mixture ratio of 40-60, although this combination has been shown to give about as good results as may be obtained. The conditions of heat- ing, temperature, size of the pieces, type of carburization box and method of packing, etc., will alter each individual carburization, and experiments should be made to determine as exactly as possible the proper combination of the different factors of carburization which 14 16 24 6 8 10 Duration of the Heating (Hours) FIG. 76. Carburization with Common Carburizing Compounds. (Scott.) enter into consideration. One of the governing factors which is often overlooked is the action of the charcoal, dependent upon its composi- tion. Thus much of the ordinary commercial charcoal still contains considerable volatile or organic matter (hydrocarbons) which may distinctly alter the effect of the carburizing. In order to reduce the intensifying action of such constituents, and to reduce the forma- tion of the hyper-eutectoid zone, it is always advisable first to calcine the charcoal before using. The general relation existing between the depth of penetration due to charcoal, charred leather, and the usual 40-60 barium car- bonate-charcoal mixture is graphically shown in Fig. 76, obtained by Scott in the carburization of soft steel bars at 1650 F. CASE CARBURIZING 139 Gradual Cements. The cements which we have just enumerated are generally classed as " gradual," for reasons previously given. Yet on the other hand, these same cements, under different con- ditions of carrying out the carburization, act as " sudden " or " quick " cements. Thus the barium carbonate mixture when used at low temperatures, or for the carburization of pieces of large dimension which heat up slowly, may furnish cemented zones in which the maximum carbon concentration may not be over the eutectoid ratio (0.9 per cent.). The same mixture, on the con- trary, may become a sudden cement at the very high temperatures and in carburizing objects of small dimensions. Other Solid Cements. In addition to the use of charcoal plus the animal charcoals, barium carbonate and common salt, the other agents which may be added are innumerable. To give a list of them would occupy several pages, besides leading to the inevitable con- clusion that the efficacy of the majority of them is small, or might be even detrimental. For it should be stated with emphasis that the more simply and more chemically definite a cement can be made, the greater will be the industrial advantages. Sudden Cements. The nature of the most important of these additions is to make the mixture act quickly, giving rise to a thin cemented zone of high carbon in a very short interval. Thus we have the use of coke saturated with mineral oil, of the saturation of the charcoal in solutions of cyanides or ferrocyanides, and of the presence in greater or less quantities of the concentrated salts of cyanogen as specific additions. Of those used in practice, the follow- ing example is extremely interesting; 11 Ibs. prussiate of potash, 30 Ibs. sal soda, 20 Ibs. coarse salt, 6 .bushels powdered hickory charcoal, 30 quarts water. Grenet recommends the following cements which have given good results in practice: Parts. a. Powdered wood charcoal 1 Salt i Sawdust 1 140 STEEL AND ITS HEAT TREATMENT Parts 6. Coal with 30 per cent, volatile matter 5 Charred leather 5 Salt 1 Sawdust 15 c. Charred leather 10 Yellow prussiate 2 Sawdust : 10 The velocity of carburization increases gradually from the first to the third of these cements. The sawdust, by making the mass more porous, increases the activity of the gases. Size of the Carburizing Box. The selection or general design of the box or container for carburizing is worthy of more attention than is frequently given to it. In the attempt to get a uniform case, much thought and research has been given to the selection of the steel, the carburizing mixture and the degree and duration of heating; and yet in many instances it has all proven unavailing. It must be remembered that the inside of a small box takes quite a while to come up to the temperature of the furnace; and that if a large box is used, the material in the center may, and does, lag behind the indicated furnace temperature several hours or its time equivalent several hundred degrees. The greater the size of the box, the larger will be this error, and the greater the actual difference in the thick- ness of case taken on by steel near the sides of the box as compared with that near the center of the box. No manipulation of the furnace can change this effect; it can only be remedied by altering the dimensions of the box itself. Here, then, lies one explanation of many unexplained failures. The box should not be larger than is absolutely necessary, even where large quantities are to be carburized in it. It should be narrow in at least one dimension so that the heat has a chance to penetrate quickly at least from two sides and reach all the contents at about the same time. Further, the boxes should not be made too deep in proportion to their other dimensions, as it makes it more difficult to pack the parts into them if so made. Whenever possible, the design of the box should follow the outline of the piece to be carbur- ized, allowing about 1 to 2 ins. all around for clearance and packing, so that the surfaces may be uniformly heated and carburized alike. Material for Boxes. Malleable iron probably gives as good satisfaction as any of the materials used in making the boxes. Cast- CASE CARBURIZING 141 iron boxes, although of comparatively small initial cost, will not stand reheating very many times, and have the further objectionable feature of being somewhat porous. Soft-steel plates and wrought iron may also be made up into good boxes. The thickness of the wall forms an important feature of the box, for if it is too thin it easily burns through, and if too thick it offers too much resistance to the penetration of the heat to the interior. For the ordinary size boxes, wall thicknesses of J to J in. are common practice. The boxes should be provided with feet so that the heat may circulate all around them. The cover should be as close fitting as is practicable, and should also be provided with ribs along the top to prevent excessive warping. Ribs along the side also add to the service of the box, besides making handling with grappling irons more easy. The sides of the boxes should taper slightly towards the bottom so that the contents can be the more quickly dumped out. Packing. Carefulness in packing is fundamental to good practice and uniformity of results, just as much as carefulness in heating or treatment. The method of packing should be such as will insure as nearly as possible the even heating and uniform carburizing of all pieces in the same box. The method of packing must necessarily vaiy with each type of article to be handled. Heavy pieces, or pieces of regular shape, do not require the care and patience which should be used with pieces of intricate design or with those which on account of their size and shape may be readily influenced by high temperatures. The packing of such pieces must be individualized. For example, long, slender pieces should always be packed vertically, so that the pieces will be held in position by the carburizing material and cannot sag under the influence of the high temperatures. Again, gears and similar pieces may be most suitably packed in tubes, so that the same amount of carburizing material and the same degree and length of heating may influence all parts of the periphery in equal proportion. In carburizing screws and bolts it is well to distribute them in the box in two opposite rows, each row having the head of the screw towards the side of the box and the stem towards the center. New compound may be used at the sides and old compound in the center. Owing to the difference in heat and the difference in the carburizing power of the compounds, this will cause a much deeper carburization of the heads than of the stems which is exactly what is desired. Again, should a narrow or low box not be available in connection with small work, carburizing compound which has been used once 142 STEEL AND ITS HEAT TREATMENT before may be put next to the sides of the box while the new com- pound is placed in the center. In this way the difference in car- burizing which might result from the different temperatures in various parts of the box may be offset. The first step in the general operation of packing is to cover the bottom of the box with the compound to a depth of 1J to 2 ins., tamping it solidly into place. The parts to be carburized are then placed firmly upon this bed so that the compound and work are in close contact with each other. The pieces should in no case touch the sides of the box, but should be placed about 1 to 1J ins. away from it. Further, the articles should be separated from each other by at least \ to 1 in., dependent upon their size and the depth of case desired. If the articles should touch one another, it is evident that the carburizing action will have less influence at that particular point with resulting soft spots. Non -uniformity of case may also result if there is not sufficient carburizing material in the box; it is better to err by using too much than too little. After the first layer of work has been placed in the box, it is entirely covered with the carburizing compound. This should be packed and tamped down around and over the pieces so as to have the particles of cement in close contact with the steel, but yet not so tightly as to prevent the free circulation of the carburizing gases which are generated during the heating process. When the first layer has thus been suitably packed and covered, the same procedure is repeated until the box is nearly filled. The point to be kept in mind is that each and every piece should be surrounded on all sides by a suitable amount of the carburizing compound. At least 2 ins. of the compound should form the top blanket over the last layer of work. Some shops adopt the following, with the aim of further preventing the escape of the gases : about 2 ins. from the top of the box sheet-steel strips about -^ in. thick are laid over the last layer of the carburizing material and these, in turn, are covered with about 1 in. of powdered charcoal. When the box is finally packed, the cover is placed on the box and the edges are care- fully sealed with fire-clay or asbestos cement. The box is now ready for the heating operation. Type of Furnace. It is not our intention to recommend any particular type of furnace for carburizing work, but rather to emphasize the necessity of designing the furnace to suit the work. And, as is evident, the conditions will vary greatly from one plant to another. CASE CARBURIZING 143 There are three main points which should be taken into account for case-hardening furnaces. (1) The furnace shall be capable of attaining easily the maximum temperature which shall be necessary for the carburizing work, and which temperature may be as high as 2000 F. (2) It must be possible to obtain a thoroughly uniform heat application at any of the intervening temperatures, and of maintaining that uniform heating with little or no variation hour after hour. (The effect of oscillating temperatures has been de- scribed.) (3) The atmosphere in the furnace shall be non-oxidizing, in order to protect the carburizing boxes from the intense oxidation which would otherwise occur at the high temperatures necessary. The Heating. The two principal points to be mentioned under this heading are: the heating, at least up to 1300 F., should be gradual; (2) the heating beyond this temperature should be uniform over all parts of the carburizing box. It has been shown by several experimenters that the energetic liberation of gases commences very strongly at temperatures somewhat under 1300 F. for the majority of solid cements, and it is advisable to diminish this factor as much as possible in order to obtain a more gradual cementation. Furthermore, it gives more opportunity for the steel to adjust itself to the effect of heating. The second point made is self-evident: non-uniformity of heating must necessarily result in non-uniformity of product. Sulphur Diffusion. The influence of sulphur contained in the cements is an extremely important factor in carburization carried on with solid cements. Grayson l has produced uncontrovertible evidence that sulphur will diffuse into iron at the temperatures ordinarily used for carburization with such substances as charred leather (which, under the conditions of his case-hardening experi- ments, contained 0.55 per cent, total sulphur), and that this sulphur combines with the manganese and iron to form manganese and iron sulphides. Thus in Fig. 77, which is a photomicrograph of a piece of 0.17 per cent, carbon steel carburized for six hours at 1650 to 1750 F. with charred leather, it will be noticed that on the edge are present, in large quantities, sulphide of manganese, also sulphide of iron with ferrite crystals intermingled. That this is sulphide was later proven by means of silver prints and by analysis which showed 2.10 per cent, of sulphur increase in the first 0.0025 inch. !S. A. Grayson, Inst. Journ., No. 1, 1910. 144 STEEL AND ITS HEAT TREATMENT This sulphur diffusion is a very serious matter, because when the surface is saturated, as in this figure, it tends to produce a soft skin, FIG. 77. Soft Case Due to Sulphur Diffusion. (Grayson.) FIG. 78. Sulphides Diffusing Further into Case with Higher Temperatures of Carburization. (Grayson.) FIG. 79. Sulphide Globules in Carburized Steel after Hardening. (Grayson.) and even if present in smaller proportions it will weaken the structure considerably, thus making it very " chippy," consequently causing CASE CARBURIZING 145 two effects which must essentially be avoided in any case-hardened work. In Fig. 78, being a similar steel carburized at 1750 to 1830 F., the sulphide is again present, but not in such a large proportion; thus the higher temperature has volatilized still more of the sul- FIG. 80. American Gas Furnace Co.'s Carburizing Machine. phur from the carburizing material. Fig. 79 shows the same car- burized piece as in Fig. 77, but afterwards reheated and quenched in water from 1380 F. In this reheating the sulphide tends to " ball " itself up, and, if anything, diffuse further in. Thus it may be seen that, for proper carburizing, the solid cements should be as free from sulphur as is possible. On the other 146 STEEL AND ITS HEAT TREATMENT *b f CASE CARBURIZING 147 hand, the barium carbonate mixtures generally used do not contain sulphur, and this sulphur diffusion cannot take place. American Gas Furnace Process. The apparatus for carburizing with gas, as devised by the American Gas Furnace Company, is shown in Fig. 80. The carburizing machine consists of a carburizing retort enclosed by a cylindrical furnace body in which it rotates, 148 STEEL AND ITS HEAT TREATMENT together with suitable arrangements for charging and discharging the work, burners for securing a proper distribution of the fuel, and supply pipes for gas and air. The machine shown has a space available for work of 30 ins. in length by 7 ins. in diameter. It is suitable for work not over 6 ins. in diameter or 20 ins. in length; for shafts, tubes, mandrels and bars of nearly equal thickness through- out of not over 24 ins. in length or 5 ins. in diameter; or for small pieces such as screws, washers, discs, etc., of a charge of about 100 pounds. The machine uses ordinary illuminating gas for both heat- and carburizing. The vertical section, Fig. 81, through the center lengthwise, shows the heavy wrought-iron retort A, which is slowly rotated on the rollers BB by the gear C, in contact with worm D, propelled by a sprocket and chain belt. The reference letters EE show air spaces in the retort formed by the two pistons /, between which the work is confined, to the properly heated central section of the retort. Letters FF indicate the heating space surrounding the retort, into which the fuel gas and air are injected under pressure, from two rows of burners indicated in the upper half of the casing by the letter G. The cover H, closing the retort, is connected with the piston-like disc marked /, by the pipe J, which is the vent of the retort. The cover H and disc I are withdrawn to charge the retort and replaced after the work is inserted. Carburizing machines connected with an automatic quenching bath are shown in Fig. 82. SUPERFICIAL HARDENING Superficial hardening differs from case carburizing in that in the former method the outer and higher carbon section constitutes a " skin " of only a few thousandths of an inch in thickness, while in the case-carburizing process the carburized zone forms a case of noticeable thickness. Exactly the same principles apply, however, in both instances, and which have been previously explained. Processes. The superficial hardening processes may be grouped under the headings of " cyanide hardening " and " pack hardening. " The cyanide hardening processes are essentially used for the pur- pose of obtaining an extreme degree of surface hardness (wear) on low-carbon or machinery steel, and in which it is not necessary to obtain high resistance to shock, etc. On the other hand, pack hardening is essentially a method of heating used particularly for fine threaded tools and other tool- CASE CARBURIZING 149 steel work. The process, when correctly carried out, permits of uniform heating with the entire elimination of oxidation by surround- ing the steel with a carbonaceous packing. But further, by prolong- ing the duration of heating at the hardening temperature, a very thin skin of higher carbon content may be formed, so that pack hardening may develop, either intentionally or otherwise, into a superficial hardening process. Cyanide Hardening. In cyanide hardening the superficial car- burizing and hardening may be effected by one of two general methods: (1) immersion of the object in a bath of liquid potassium cyanide or other mixture with cyanogen as the base, followed by quenching; (2) coating or sprinkling the surface of the object with an adhesive mixture of finely pulverized carburizing cyanogenous salt or " varnish," heating the steel to the proper hardening tempera- ture and thus melting the cyanide and hardening as usual. The first or " immersion " process is by far the most efficient, both as to unifcrmity of the carburized zone and simplicity and uniformity of operation. Further, this first method has the tendency to reduce deformation and oxidation during heating and quenching, since, as previously explained, heating in any molten bath has this effect. The Immersion Method. The method of cyanide hardening by immersion is quite simple. The salt, usually potassium cyanide (KCN), is melted in a suitable pot-furnace, and is maintained at a temperature a little over the upper critical range of the steel to be carburized and hardened. This temperature, for ordinary machin- ery steel, is about 1550 to 1600 F. The steel is then immersed in the molten cyanide and kept there until it has been uniformly heated; or this heating may be somewhat prolonged in order to obtain a greater depth of skin. In general, however, it is not advis- able to heat for a length of time much greater than ten or fifteen minutes, or at temperatures much over the critical range, since such heating will tend to give non-uniform and high-carbon zones which, after quenching, are intensely brittle and may chip off in service. Quenching is usually done in lime water in order to neutralize the cyanide remaining on the steel. Some concerns adopt the method of immersing the steel in the cyanide as soon as it has become molten, permitting the steel to heat up with the bath, and then quenching as soon as the desired temperature of say 1550 to 1575 F. has been attained. It is absolutely necessary to remember that cyanogen compounds are deadly poisonous, and every precaution should be adopted when 150 STEEL AND ITS HEAT TREATMENT using them. Furnaces should be supplied with hoods which have strong draft. Gloves should be used in handling all work, for if cyanide gets into a fresh cut or scratch it will prove deadly. In some cases, when working at the furnaces, it is even advisable to use face masks and to cover up any exposed parts of the body. Cyanide Hardening Plant. A battery of twenty cyanide fur- naces is shown in Figs. 83 and 84. l In front of the first pair of FIG. 83. Battery of Cyanide Furnaces' Special Quenching Machines for Clutch Rings in Foreground. ("Machinery.") furnaces in Fig. 83 two special machines are shown which suddenly cool or quench the work as fast as it can be heated and removed from the furnace. They are used for hardening the steel ring discs shown at U. These alternate with brass discs in a multiple-disc clutch on the engine of an automobile. Each pair of furnaces shown in these two figures is covered with a hood to convey the poisonous fumes to the outer atmosphere through pipes extending through the roof. In addition to this, sheet-metal shields are located in front of 1 E. F. Lake, in " Machinery," Sept., 1914. CASE CARBURIZING 151 the furnace openings shown at V to carry away from the workmen any fumes that might come through these openings. (These shields were removed for photographing.) At the end of the cyanide fur- o b I I I 1 . . . acteristics 0.18 to 0.25 0.40 to 0.80 1500 to 1550 800 to 900 70,000 to 85,000 45,000 to 60,000 35 to 25 65 to 45 Auto, lever 0.18 0.40 1650 800 70,030 45,400 32 64 Pressed auto, frame. 0.22 0.40 1530 800 71,950 43,400 29 56 Engine forging 26 0.28 1650 1025 77,210 52,200 28 65 Old rolled Y in. plate 0.24 0.60 1525 900 93,300 65,250 20.5 51 The above remarks apply mainly to the smaller sections up to 2 ins. in thickness, but are nevertheless applicable in part to heavy work. With the increase in sectional area, the effect of hard- ening decreases, and for particularly heavy work may result only in a refinement of grain. Thus, for heavy, oil-treated forgings, toughening may not be considered a necessity; such reheating will, however, relieve the strains which are always inherent to quenched steels. Large forgings thus treated will show an elastic limit of 30,000 to 50,000 Ibs. per square inch, with an elongation of 35 to 25 per cent, in 2 ins. Annealed. There is probably more disagreement and argument as to the proper annealing temperatures for this range of carbon steel than for any other. Opinion and practice are divided over the use of a comparatively high temperature 50 to 100 over the upper critical range or a lower temperature laying somewhere between the Acl and Ac3 ranges. In this group the Acl and Ac3 ranges are widely separated and the influence of the carbon-mangan- ese content is rapidly increasing. The high annealing temperature, 1550 to 1600 F. or more, will give ample opportunity for the absorption of the excess ferrite, for diffusion and for equalization. On the other hand, there is according to some authorities a marked increase in grain size from 1350 or 1375 F. and upwards. The whole question really depends upon the condition of the steel before annealing. If the " breaking-down " during elaboration either rolling or forging has been severe, if high temperatures have been used, and if the finishing temperature has not been just right, a high annealing temperature may be necessary to entirely relieve the strains and equalize the steel. On the other hand, if the steel CARBON STEELS 235 has been carefully worked and the micrographic structure is fairly good, the lower temperatures will probably be entirely satisfactory. Much must be left to the operator and his own particular problem. The main point to bear in mind is that the lowest temperature should be used which will produce the desired results. If we assume as average figures for annealed steel of this cer- bon range: Tensile strength, Ibs. per sq. in 58,000 to 65,000 Elastic limit, Ibs. per sq. in 28,000 to 35,000 Elongation in 2 ins., per cent over 30 and compare these with the results of a tensile test taken from the steel to be annealed, a very good idea of the degree and length of heating may be obtained. For example, the following results from If -in. rounds for gun barrels show that a high annealing tem- perature was not necessary in this case, inasmuch as the original steel was in excellent condition. Gun barrel steel, If -in. rounds. Carbon, 0.18 per cent. Manganese, 0.50 per cent. Phosphorus, 0.070 per cent. Sulphur, 0.055 per cent. Silicon, 0.055 per cent. Treatment. Tensile Strength. Lbs. per Sq. In. Elastic Limit. Lbs. per Sq. In. Elongation. Per Cent. In 3 Ins. Reduction of Area. Per Cent. As Rolled 66,750 33,820 33.3 57.6 Annealed at degrees F. for minutes 1360-1400 30 64,960 34,050 38.0 61.0 1500 20 65,180 32,930 38.3 58.3 1500 105 64,060 33,150 39.1 62.3 1830 15 62,940 31,810 35.7 56.3 2120 5 61,150 31,580 33.8 53.1 On the other hand, the following cold-rolled automobile-frame steel was particularly " hard " before annealing and required a tempera- ture of 1550 F. to relieve thoroughly the effect of the cold work: Carbon, 0.24 per cent. Manganese, 0.38 per cent. Phosphorus. 0.028 per cent. Sulphur, 0.038 per cent. 236 STEEL AND ITS HEAT TREATMENT Tensile Strength. Lbs. per Sq. In. Elastic Limit. Lbs. per Sq. In. Elongation. Per Cent, in 2 Ins. Before annealing 100 400 68,500 18.6 After annealing at 1550 F 66,000 38,100 37.0 For the average run of annealing work for this range of carbon, a temperature of about 1500 F. will be found to give satisfactory results; individual cases must be treated as such. 0.25-0.35 CARBON STEEL Steel containing from 0.25 to 0.35 per cent, carbon is known as soft-forging steel and is used principally for structural ptirposes in infinite variety. It responds in a most satisfactory manner to welding, forging and machining, and may be vastly improved by proper heat treatment. Under skillful treatment, the variety of combinations of -strength and ductility are to be had in probably no other range of carbons. Relative to static strength, some really wonderful results for straight carbon steels in the way of high tensile strength with high ductility have been obtained from heat-treated (oil quenched and toughened) forgings of 0.30 to 0.35 per cent, carbon. The follow- ing results, obtained from the center of a 5-in. electric car, heat- treated axle, the #xle being selected at random from a group of about one hundred forgings, give an idea of the extent to which proper heat treatment may develop the physical properties; Electric Car Axle, 0.32 Carbon, Acid Steel Tensile strength, Ibs. per sq. in 91,700 Elastic limit, Ibs. per sq. in 61,620 Elongation, per cent, in 2 ins 33.5 Reduction of area, per cent 48.1 In the hardened condition without subsequent tempering these steels may be used for gears. In the toughened condition these steels present the maximum resistance to fatigue and other dynamic stresses, as represented by alternating impact and other tests, over any of the straight carbon steels; the dynamic strength probably apexes at about 0.30 per cent, carbon, as far as the author can judge from his own researches and from the work and conclusions of others. CARBON STEELS 237 Untreated. In the untreated condition, with standard man- ganese, phosphorus and sulphur, the average tensile strength of these steels will be about as follows: Carbon. Acid Steel. Basic Steel. 0.25 to 0.30 67,000 to 78,000 63,000 to 72,000 0.30 to 0.35 69,000 to 83,000 65,000 to 74,000 Rolled plates, from 2 to 4 ins. thick, made of basic steel with 0.25 to 0.35 per cent, carbon and about 0.40 per cent, manganese, will usually fulfill the following specifications : [Tensile strength, Ibs. per sq. in 65,000 to 75,000 Elastic limit, Ibs. per sq. in 33,000 to 37,000 Elongation, per cent, in 2 ins 30 to 25 Reduction of area, per cent 50 to 36 These results may also be considered as generally applicable to untreated steel of this analysis, but which has had more or less elaboration or working. Heat Treated. The upper critical range decreases from about 1500 F. for 0.25 per cent, carbon, to about 1425 F. for the 0.35 per cent, carbon steel. Practical experience has shown that a quenching temperature of 1500 to 1525 F. for the lower carbons of this range, and 1450 to 1500 F. for the higher carbons will give satisfactory results under ordinary conditions. If the heating has been conducted uniformly and not too rapidly especially when approaching the maximum temperature the original structure of the steel should be entirely eliminated, as the temperatures recom- mended are distinctly above the upper critical range. Never- theless, some metallurgists prefer to quench these steels from a higher temperature, say 1575 to 1600 F., in order to make cer- tain of the complete change in structure and to obtain a maximum hardening effect. In either case, intelligent furnace operation and heat control will probably be the governing factor rather than the indicated furnace temperature or mere theorizing. lor forgings in which especially high qualities are desired, double quenching will produce a refinement of grain and correspondingly higher elastic limit and ductility than are usually obtained by the single treatment. The temperatures recommended for this range of carbons are : 1. Jirst quenching 'from 1600 F., or from 1500 to 1550 F. if the higher quenching should prove too drastic. 238 STEEL AND ITS HEAT TREATMENT 2. Second quenching from 1425 to 1450 F., followed by 3. Suitable toughening according to the size of piece and physical properties desired. The results to be obtained from heat treatment will vary largely for this range of carbon in particular, due to such influence as the increase of a few points in the carbon content (particularly noticeable in these mild steels), the size of the section, the quenching medium, and so forth. The results given under the 0.15 to 0.25 carbon range, and under the 0.35 to 0.45 carbon range to follow, may be used as a general measure of the carbons under discussion. Stated roughly, these carbons will give elastic limits ranging from 35,000 to 80,000 FIG. 146. 0.28 per cent. Carbon Steel. X39. (Campbell.) Ibs. per square inch, with corresponding elongations of 30 to 10 per cent, in 2 ins. Annealed. As has been previously explained, heating for anneal- ing to just above the Acl (lower) critical range will refine the ground- mass only, while complete refinement is shown by the disappearance of the ferrite and network beyond the upper critical range (Ac3). As an example of this, examine the photomicrographs of a basic open-hearth steel containing 0.28 per cent, carbon and 0.52 per cent, manganese, as shown in Figs. 146, 147 and 148. The first photo- graph shows the original steel with its coarse, weak structure. Fig. 147 shows the same steel annealed at 1425 F., or just over the Acl range; the pearlitic ground-mass has been entirely refined, but there still remains the unabsorbed and undiffused excess ferrite. Fig. CARBON STEELS 239 148 shows the same still heated to 1520 F. and slow cooled in the same manner; but in this case the structure has been entirely changed and refined by heating to a temperature over the upper critical range. . . FIG. 147. 0.28 per cent. Carbon Steel Annealed at 1425 F. X39. (Campbell.) FIG. 148. 0.28 per cent. Carbon Steel Annealed at 1520 F. X39. (Campbell.) Practical experience has shown that a temperature of 1500 to 1525 F. will give excellent results for the full annealing of steels within this range of carbons. On account of the hardening effect of air cooling steeis with over 0.20 per cent, carbon when in small 240 STEEL AND ITS HEAT TREATMENT sections, these steels should be slow cooled, either in the furnace, in lime or in ashes. In regard to the physical properties to be obtained from the annealing of these steels, the lower carbons of this range should always meet tthe U. S. Government specification of: Tensile strength, Ibs. per sq. in 60,000 Elastic limit, Ibs. per sq. in 30,000 Elongation, per cent, in 2 ins 30 while the higher carbons will usually give : Elastic limit, Ibs. per sq. in 35,000 to 45,000 Elongation, per cent, in 2 ins 22 to 32 Reduction of area, per cent 30 to 60 0.35-0.45 CARBON STEEL Straight carbon steels with 0.35 to 0.45 per cent, carbon are particularly suited to medium and heavy forgings for which the lower carbons would not give sufficient strength, and for which it is also not desirable to use water quenching on account of the possi- bility of starting incipient cracks or strains. This steel is commonly used for high-duty and moving machine parts; for axles, side bars, crankpins and other locomotive forgings ; for guns and gun forgings ; for crank shafts, driving shafts and similar automobile parts; and for general structural purposes requiring the combination of maxi- mum strength with minimum brittleness. It has excellent dynamic strength, although probably not quite so much as the previous class of 0.25-0.35 carbon. Steel with 0.40 carbon according to Robin x presents the greatest resistance to abrasive action (wear) . These steels are easy to machine when in the annealed or soft- toughened condition, but should not be used for screw machine stock. The upper critical range temperature of this steel is about 1425 F. to 1400 F. Untreated. The average untreated American open-hearth steel with standard manganese, phosphorus and sulphur will average about as follows in tensile strength: Carbon. Acid Steel. Basic Steel. 0.35 to 0.40 78,000 to 92,000 70,000 to 78,000 0.40 to 0.45 87,000 to 100,000 76,000 to 89,000 1 J. Robin, Inst. Journ., II, 1910. CARBON STEELS 241 Annealed. Remarks. C. Mn. Phos. Sul. Tensile Strength. Lbs. per Sq. In. Elastic Lir it. Lbs. per Sq. In. Elong- ation. %in 2 Ins. Red. of Area. Per Cent. General limits. . . . 0.35 to 0.45 not over 0.70 under 0.045 under 0.045 70,000 to 85,000 38,000 to 50,000 28 to 20 55 to 40 Forged gun jacket acid steel 0.35 0.25 0.038 0.019 77,080 39,500 27 Forged gun jacket basic steel 0.43 0.22 tr. 0.023 78,180 43,100 25.5 8-in. axle acid steel annealed at 1400 F 0.42 0.51 78,420 47,460 28 54.5 Heat-treated. Large sections, when quenched in good mineral oil from 1400 to 1500 F., and toughened at 900 to 1200 F. (according to the carbon content and largest section), should always meet the specification of 85,00050,0002245. The following tests taken from large forgings show the variety of combinations of strength and ductility which may be obtained: Forging. Carbon. Treatment. Tensile Strength. Lbs. per Sq. In. Elastic Limit. Lbs. per Sq. In. Elonga- tino. Per Cent, in 2 Ins. Red. of Area. Per Cent. Gun jacket Axle 0.35 0.41 1500-0/1200 1450-w/lOOO 109,560 90,250 65,090 54,575 16.5 25 4 52 4 Gun jacket Shaft 0.43 0.42 1500-0/1200 1525- 0/1300 111,100 82,040 69,700 57,060 17.0 29.0 55.0 o = oil. w = water. It is always advisable to keep the drawing temperature as near 1200 to 1250 F. as possible, not only because it is easier for the furnace operator to obtain more accurate temperature control at these more readily distinguished " reds," but also on account of the greater dynamic strength which is obtained by the use of the higher drawing temperatures. The results obtained from the water quenching from 1450 F. and subsequent toughening of small rounds of 0.40 per cent, carbon steel are given in the chart in Fig. 149. 242 STEEL AND ITS HEAT TREATMENT ""g*"* OOS0 ^1 D 1 ^ (g) aequinx saonp-rcH nauug 1 1 1 1 1 Annealed i 5 1 1 i I i i l / Chemical Analysis: Critical Ranges: Size of Section Treatment: C. 0.40 Acl 1330 1 inch round Quenched in Mn. 0.60 Ac2 water from P. 0.02 Ac3 1410 1450, and S. 0.03 drawn as -. given. FIG. 149. Normal Characteristics of 0.40 Carbon Steel, Heat Treated. / \ 1 / \ ffl 1 \ / \ /! 1 \ / / \ / \ \ / i \ \ / I \ / A \ \ / / / \ / ^ \ / \ / / ^ V / ' / X \ ]!:ml.md 1 i 1 1 1 1 1 3 ' 1 tpni 9Jimbg aod spunoj '(z) iitniq o^stJig PUB '(x) m3naj?s ^Ijanax "= o ^ s ' s s CARBON STEELS 243 LOCOMOTIVE AXLES Locomotive axles and other heavy forgings used in locomotive construction are illustrative of the treatment of pieces of large section of the above carbon range. Heat-treated Axles. For heat-treated axles the carbon content will range between 0.35 and 0.50 per cent., and with such steel the treatment may generally be adjusted to meet the standard specification of: Tensile strength 85,000 Ibs. per sq. in. Elastic limit 50,000 Ibs. per sq. in. Elongation 22 per cent, in 2 ins. Reduction of area. . . . 45 per cent. The quenching of the axles, usually from temperatures of 1400 to 1500 F., is mainly a proposition of correct heat application and efficient handling. Both oil and water are used extensively for hardening axles. Water will bring out the full effect of heat treat- ment by giving the highest tensile test properties of which the steel is capable; with the same ductility, oil quenching will give lower tensile values than water quenching. A steel with lower carbon content may more logically be used with water quenching than with oil. Water requires no expensive cooling nor circulation system, and has practically no cost of upkeep or replenishment, as all these may be regulated by the intake of fresh, cold water. On the other hand, many engineers severely condemn the use of water in that it is too harsh in its action upon large masses of steel as in axles, that cracks are more liable to develop, and that internal strains are set up which often are not always entirely relieved by the reheating or toughening. Oil is much the safer quenching medium to use for axles, will give more uniform results, and should only be replaced by water for economic reasons. The hardened axles are charged while still warm into the re- heating furnace, which is maintained at such temperature as will relieve all strains set up in the hardening and at the same time give the physical properties desired. This temperature will vary between 900 and 1200 F., depending upon the chemical composi- tion of the steel. The higher the drawing temperature, the more ductile the steel and apparent coarseness of the grain, due to the transformation of the transition constituents troostite and sor- bite into pearlite plus free ferrite. Straight carbon steels quenched in water and drawn at 1000 or 1100 F. will be entirely sorbitic, but 244 STEEL AND ITS HEAT TREATMENT at 1200 may show considerable free ferrite. Too much emphasis cannot be given to the necessity of keeping a uniform temperature and allowing sufficient time for the heat thoroughly to penetrate the axle. It is then preferable for the axles to cool with the furnace, rather than to remove them while still at the toughening temperature. The temperatures used by one manufacturer of acid steel axles and other large forgings to meet the standard A. S. T. M. specifica- tions are as follows: after quenching in oil from 1450 F., reheat as below and air-cool or cool in ashes. 0.42 to 0.45 per cent, carbon 1175 F. 0.38 to 0.42 " " " 1125 F. 0.33 to 0.38 " " " 1075 F. 0.28 to 0.33 " " " 1000 F. The following are characteristic tests from Open-hearth steel locomotive axles (Penna. R.R.) : 0.41 Carbon t? nrtrfM * Treated, orged . j 500 o water/ ! 000 F. Tensile strength 73,627 90,250 Elastic limit 31,505 54,575 Elongation 31.6 25.4 Reduction of area. . 43.6 52.4 0.50 Carbon Treated, wa ter/1200 F. Tensile strength .............. 83,430 90,092 Elastic limit ................. 34,370 53,655 Elongation ................... 22.6 27.1 Reduction of area ............. 30.0 52.8 " Tempered " Axles. For want of a better name, under this heading might be included such as are treated by the " Coffin " or similar processes. The main principle consists in heating the axle as usual for hardening, then immersing in the quenching bath for a cal- culated number of seconds, immediately withdrawing, and allowing the heat from the interior of the axle to " temper " the part which had been hardened by the short immersion in the oil or water. This process has been developed to such a nicety that surprisingly uniform results may be obtained if the test is always taken from the same relative place, such as half-way between the center and the outside, The main arguments for the process are that it is simple CARBON STEELS 245 and that the axle will have a tough (annealed) core, and, at the same time, a hard wearing surface. Non-uniformity of structure is the principal argument of those condemning the process, and on account of the inevitable " human equation " which enters into it, would seem to be not without justice in many instances. Annealed Axles. The axles are heated up slowly and uniformly to a temperature slightly in excess of the upper critical range, main- tained at this temperature for sufficient time for the steel to respond to the heat, and then cooled with the furnace. If the working and the finishing temperature during forging have been adjusted so as to give a fine grain to the steel, besides good physical test re- sults, it will be found that heating to a temperature over the critical range may not be necessary in many instances. Where it is neces- sary to anneal large numbers of heavy axles, the hot axles may be removed quickly to a pit and covered with lime or ashes. Annealing alone will generally overcome the strains set up in the previous proc- esses of manufacture, but it does not bring out the higher physical properties of which the steel is capable. Annealed axles will show pearlite and free ferrite, the apparent size of the ferrite grains de- pending upon the rate of cooling and the time thus given for the ferrite to separate out from the matrix. In order to obtain the required tensile strength upon anneal- ing it is necessary to use a steel of higher carbon content than that used for full heat treatment. A tensile strength of 80,000 Ibs. per square inch will require a 0.50 carbon steel or higher. A temperature of 1500 F. is generally recommended (A. S. T. M.) for annealing 0.40 to 0.60 per cent, carbon steel, but since the critical range of this steel is about 1400 F. or a little under, an annealing temperature of 1400 to 1450 F. will give a better fracture, together with a better combination of tensile strength and ductility. The author has had much better results with the lower temper- ature, although the time required for the annealing generally is longer. The following results from acid open-hearth steel will show the effect of the lower temperature anneal : Heat 4261 Annealed at 1400 Carbon, 0.42 per cent Manganese, 0.51 per cent . . . Phosphorus, 0.034 per cent . . Sulphur 0.028 per cent Tensile strength Elastic limit Elongation in 2 ins ... Reduction of area. . . . 78,420 Ibs. per sq. in. 47,460 Ibs. per sq. in. 28 per cent. 54.6 per cent. 246 STEEL AND ITS HEAT TREATMENT Failures of Heat-treated Axles. Aside from piping, segrega- tion, and other impurities in the steel, improper heat treatment is the active cause of failure of both heat-treated carbon and alloy steel axles. Unequal or insufficient heating in either the hardening or toughening processes will produce unequal stresses, which in turn will sooner or later result in failures. These failures are always transverse, and never longitudinal. Water quenching large sections has a strong tendency to produce cracks, often not appear- ing on the surface, and which may open up when subjected to the heavy duty and " pounding " when placed in service. Such defects may be sometimes discovered by the drop test, but its expense prevents many railroads from using it for the test of every axle. Heat-treated axles, when given ample reduction in the forging operation, carefully and uniformly heated to the proper tempera- tures, and held at those temperatures for a time sufficient for the steel to respond throughout, should prove vastly superior to un- treated or annealed axles. On the other hand, the engineering departments of many rail- roads have become considerably alarmed over the frequent failures of so-called " heat-treated " axles, and many have absolutely refused to have anything to do with axles which have been oil or water quenched. This really serious phase of the axle question has led to the investigation of the possibilities of hardening such forgings in air or steam. Surprisingly good results have been obtained by methods based upon this system and from the technical stand- point are indeed remarkable, since a considerable toughening heat is necessary with even 0.40 per cent, carbon steel. 0.45-0.60 CARBON STEEL Treatment of Large Sections. As the carbon content is pro- gressively increased beyond 0.45 per cent., its effect becomes quite noticeable in the added brittleness of the steel. This is strongly illustrated by the fact that a general study of heat-treatment prac- tice will show that there is very little quenching of large sections when the carbon content exceeds the 0.50 per cent. mark. The dangers to be encountered, both in the treatment itself and by possible fracture in service, almost prohibit such treatment of large sections. Any increase in static strength which can be obtained by quenching and toughening is most certainly acquired with the ever-present danger of cracking, or of starting incipient cracks. For these reasons it is, therefore, apparent that the full heat treat- CARBON STEELS 247 ment of large sections, even though it may bring out higher physical characteristics in the steel as is shown by the subsequent figures obtained from the treatment of a 0.50 per cent, carbon axle is becoming less and less of a factor in steels of these carbons. 0.50 Per Cent. Carbon Axle Forged. Quenched in Water from 1400 F. Toughened at 1200 Tensile strength, Ibs. per sq. in Elastic limit so in 83,430 34370 90,090 53655 Elongation per cent, in 2 ins 22 6 27 1 Reduction of area, per cent 30.0 52.8 Tempering, and Small Sections. On the other hand, the harden- ing and tempering (as distinguished from toughening) of the smaller sections, such as gears, dies, etc., begins to take an important place in heat-treatment work with these carbons. In such cases the increased-carbon content brings about an inherently possible wearing hardness which is developed by hardening and tempering. The medium and smaller size sections may be satisfactorily hardened in water with but a small proportion of the danger which would inevitably result from the water quenching (or even oil quenching) of larger sections. And, by varying the reheating temperatures, the following approximate physical results may be obtained: Elastic limit, Ibs. per sq. in.... . . Elongation, per cent, in 2 ins. .. Reduction of area, per cent . 50,000 to 110,000 20 to 5 50 to 15 Annealing. The commrecial annealing of steel of say 0.50 to 0.60 per cent, carbon will give a variety of results which in them- selves have proven a stumbling block for many a heat treater. This is largely due to the prominent part and effect of different rates of cooling in relation to the size or mass of the steel. To illus- trate: 6X6 in. billets of 0.50 to 0.55 carbon which have been heated to 1400 F. and furnace cooled will, in general, meet the specifica- tions of Tensile strength 80,000 Elastic limit 40,000 Elongation 22 Reduction of area . 35 248 STEEL AND ITS HEAT TREATMENT On the other hand, smaller sections, annealed in the same manner and in the same furnace, will according to their size, give physical results varying anywhere between Elastic limit 45,000 to 60,000 Elongation 20 to 15 Reduction of i area 40 to 30 In other words, the extreme variability in the rate of cooling, as dependent upon the size of section and mass of the steel, its rela- tion to the size of the furnace, the degree to which the cooling of the furnace may be controlled, and numerous other related factors make the commercial annealing of these steels an individual problem as far as actual physical results are concerned. It is therefore always advisable, if specific physical results must be obtained by annealing (used in the broad interpretation of the term) , to take first a preliminary test of the steel in the condition as received. From such results it will then be evident how much the steel must be " let down," and the proper reheating temperature may be judged from previous experience or by experiment. Al- though annealing at a temperature under the critical range will not change the general structure of a pearlitic steel, it will relieve the strains and stresses, and thereby improve the steel. But fur- ther, the previous elaboration, such as rolling or forging, which the steel has undergone, will, in a majority of cases in actual practice, leave the steel in more or less of a sorbitic state. Under such conditions, a reheating or commercial annealing will actually change the physical results, even though the annealing temperature is under the critical range. Such commercial annealing or reheating temperatures may vary from 900 F. and upwards through the upper critical range. In the author's experience there is little or no change in the physical test results through the annealing of such steel at temperatures under 900 F. or thereabouts. But from this temperature upwards the sorbitic constituents will gradually coagulate into the pearlite and ferrite, with a corresponding lowering of the static strength and increase in the ductility. Consequently, by regulating the commercial annealing temperature, the physical results may be " let down " to the desired limits. If it is desired to change entirely the structure of the steel and to obtain the finest grain size possible, with maximum ductility, it CARBON STEELS 249 will be necessary to anneal the steel at a temperature slightly in excess of the upper critical range, followed by slow cooling. The influence of the rate of cooling, as exerted by air cooling, is manifested in the peculiar statement that the tensile strength of these hard forging steels may be actually raised by annealing (as distinguished from quenching). It is a well-known fact that steel of such carbon content when cooled in air at a more or less rapid rate through the critical range will take on a noticeable degree of hardness. The author has found that this fundamental principle may be applied to great advantage in the treatment of axles with, of course, certain modifications and that it is even necessary to reheat or toughen in order to lower the tensile strength and obtain the proper ratio of static strength to ductility. Such a process is now being developed by a large manufacturer of axles, and will in all probability have an influence upon the heat treatment of axles and other forgings of large section. SHRAPNEL Shrapnel are illustrative of this range of carbon and of pieces of medium section. The current specifications for foreign shrapnel cover a wide range of physical properties, varying between 80,000 and 140,000 Ibs. per square inch in tensile strength, with 20 to 8 per cent, elongation. The chemical composition of the steel used will be approximately between 0.50 and 0.60 per cent, carbon, although the extreme limits are 0.35 to 0.8 per cent., 0.4 to 1.0 per cent, manganese, phosphorus, sulphur and silicon about normal, and with or without the addition of chrome or nickel. Thus, according to the physical or chemical specifications worked under, some shrapnel manufacturers have been able to meet their particular specifications without any treatment except perhaps cooling the cases in lime after forming; others have had to anneal, or harden and temper; while still others have add to carry out all three heating operations. There is nothing unusual in the heat treatment required. The proposition in short is merely one of proper heat application in fur- naces of correct design and construction; and yet one may see almost any and every kind of a furnace being operated in almost any and every kind of a way except the right one, with the result of large rejections. The latest and a very efficient type of furnace for this work is 250 STEEL AND ITS HEAT TREATMENT designed on underfired principles, with a continuous and auto- matic charging and discharging of the shrapnel. The cycle for the complete hardening and drawing is as follows: one man places the rough-formed shells on the charging platform of the hardening furnace, as shown in Fig. 149a; an automatic device takes the shrapnel into and through the heating chamber at a specified and predetermined rate; the heated shrapnel are then discharged con- tinuously from the furnace into an oil-quenching bath, as shown in FIG. 149a. Automatic Hardening Furnaces, Charging End, Working on 3-in. Shrapnel. Fig. 1496, from which they are removed by a conveyor system and delivered to a table in front of the second or drawing furnace. Here another man takes them from the table and places them on the charging table, as shown in Fig. 149c; the shrapnel then follow the same course as in the first furnace, except that they are received from the furnace into wheelbarrows or similar means of taking them away, as shown in Fig. 149d. The actual temperature of the shrapnel in the case of one plant was 1500 F. for hardening and 850 F. for the tempering or drawing operation. CARBON STEELS 251 FIG. 1496. Hardening Furnaces, Discharge End; Oil Quenching Tanks at the Left. FIG. 149c. Tempering Furnaces Charging End Conveyer from Quenching Tank. 252 STEEL AND ITS HEAT TREATMENT In connection with the above equipment a comparison between the old hand method and the new automatic appliances is extremely interesting. The following are actual tests made in the same plant on 3.3-in. shrapnel: Old Practice Two furnaces, 6 men to each furnace: One furnace, 6 men, 910 shells in 10 hours One furnace, 6 men, 1207 " " " Total, two furnaces, 12 men, 2117 " " " Average of 176 shells per man per 10 hours. FIG. 149d. Tempering Furnaces Discharge End. New Practice Two furnaces, with 3 men total: One furnace, 2508 shells in 10 hours One furnace, 2603 " " " Total, two furnaces, 3 men, 5111 " " " " Average of 1704 shells per man per 10 hours. In addition to obtaining some 1000 per cent, increased output per man with the new underfired, automatic furnace, it has been shown that in the old method the average rejections were running around 15 to 20 per cent., and were only about 3 per cent, for the new practice. CARBON STEELS 253 CARBON STEELS WITH OVER 0.60 CARBON Treatment in General. The treatment of high-carbon steel develops into the two propositions of hardening and annealing. Toughening, as referring to high reheating temperatures subse- quent to hardening, is but very little used, due to the fact that these steels are too brittle for ordinary structural purposes. Similarly, tempering is governed entirely by the degree of hardness required by the tool and is dependent, not only upon the chemical analysis, but in a larger measure upon the result of the hardening operation. Hardening. The precautions to be adopted in hardening may be repeated in the following general summary: (1) Use the lowest temperature which will give the desired results. (2) Heat slowly and uniformly. (3) The higher the carbon content, the greater is the degree of care which must be used, and, in general, the more narrow the hardening temperature limits. The temperatures to be used in hardening are largely governed by the carbon content, and which, in turn, influences the position of the critical range. We may sum up these factors as follows: Carbon Content. Per Cent. Critical Range. F. Hardening Temperature. F. 0.60 1340-1380 1400-1460 0.70 1340-1375 1400-1450 0.80 1340-1365 1390-1450 0.90 1340-1360 1375-1450 1.00 1340-1360 1375-1450 1.10 1340-1360 1375-1430 1.20 1340-1360 1375-1430 1.30 1340-1360 1375-1420 1.40 1340-1360 1375-1420 In giving the above hardening temperatures we have assumed that the previous mechanical and heating operations have left the free cementite (in hyper-eutectoid steels greater than 0.9 per cent, carbon) well distributed, or emulsified, throughout the steel. This will generally be true when the proper finishing temperatures, either in rolling or in forging, have been used. In such cases, therefore, it will not be necessary to heat to above the Ac. cm range in order to emulsify the free cementite, and following with a sub- sequent quenching from slightly above the principal critical range. 254 STEEL AND ITS HEAT TREATMENT If, however, the previous heating operations have left the free cementite in the form of spines or network, it will be mandatory to use the double-quenching method in order to spheroidalize this free cementite and thus obtain the maximum wearing and cutting hardness; for details of such procedure the subject matter in Chap- ter VII should be studied. Annealing. The general subject of annealing hyper-eutectoid steels has been discussed in Chapter III. A series of physical test results of experiments carried out by Fdbry 1 upon the annealing of steels with carbon contents of 0.58 to 1.36 per cent., the size of bar being 1.18 ins. square, and the selected annealing temperatures being maintained for three hours, are given in the following tables: 0.58 PER CENT. CARBON STEEL ANNEALED Treat- ment. Tests. Hard- ness. Microscopic. Annealed Deg. Fahr. Tensile Strength, Lbs. per sq. In. Elastic Limit, Lbs. per sq. In. Elonga- tion, per cent In 3.15 Ins. Red. of Area, per cent Brlnell No. Structure. Notes. 1110 99,540 15.8 43.4 196 Free ferrlte and pearllte. Ferrlte reticulated, meshes filled with grainy pearl- lte. 1200 98,420 45,510 17.7 49.0 183 Ferrlte begins to change Into pearllte. 1290 84,200 39,820 20.7 59.2 174 Smaller ferrlte crys- tals and pearllte. Structure essentially dif- fering from other speci- mens, because the ferrlte Is uniformly distributed. 1380 93,860 36,980 18.6 43.4 176 1470 96,860 39,820 19.1 36.8 . 187 Free ferrlte and pearllte. Ferrlte forms a network; pearlite partly grainy, partly lamellar. 1560 96,710 39,820 17.9 35.6 183 1650 98,130 39,820 18.6 36.8 185 Network of large ferrlte crystals filled with pre- dominantly grainy pearl- lte. 1740 93,860 36,980 16.7 33.6 187 1830 100,700 39.820 13.1 25.2 196 Critical range Ac commences at 1337, maximum at 1355. 1 Zs. Fabry, " The Variation in the Mechanical Properties and Structures of a Few Special Tool Steels Annealed between 600 and 1000 C. " Int. Soc. Tes. Mat., 1912. CARBON STEELS 0.81 PER CENT. CARBON STEEL ANNEALED 255 Treat- ment. Tests. Hard- ness. Microscopic. Annealed Deg. Fahr. Tensile Strength, Lbs. per sq. In. Elastic Limit, Lbs. per sq. In. Elonga- tion, per cent In 3. 15 ins. Red. of Area, per cent Brlnell No. 1110 102,950 13.1 37.6 212 1200 106,400 42,670 14.0 35.6 207 1290 99,540 39,820 17.8 43.4 187 1380 100,700 31,290 14.6 29.4 183 1470 102,840 31,290 12.5 14.8 196 1560 1650 105,250 36,980 13.1 23.0 187 100,400 31,290 13.1 19.4 203 1740 98,980 31,290 10.2 14.0 207 Critical range Ac commences at 1328, maximum at 1337 C 0.92 PER CENT. CARBON STEEL ANNEALED Treat- ment. Tests. Hard- ness. M Icroscopic. Annealed Deg. Fahr. Tensile Strength, Lbs. per sq. In. Elastic Limit, Lbs. per sq. In. Elonga- tion, per cent In 3.15 Ins. Red. of Area, per cent Brlnell No. Structure. Notes. 1110 122,900 12.0 23.0 228 Euctectlc. Grainy pearllte with larger grains. 1200 120,600 42,670 13.0 25.2 217 1290 98,420 36,980 11.5 33.6 163 Structure perfectly homo- geneous and essentially differing from those of other specimens. 1380 91,030 34,130 17.8 43.4 174 1470 113,500 31,290 10.5 14.8 212 Grainy pearllte. 1560 112,100 9.1 14.0 207 Lamellar pearllte. 1650 112,500 31,290 8.7 14.8 216 1740 105,800 31,290 9.0 11.6 214 1830 123,450 36,980 6.8 9.2 228 Indications of overheated structure. Critical range Ac begins at 1346, maximum at 1355. 256 STEEL AND ITS HEAT TREATMENT 1.11 PER CENT. CARBON STEEL ANNEALED Treat- ment. Tes ts. Hard- ness. Microscopic. Annealed Deg. Fahr. Tensile Strength, Lbs. per sq. in. Elastic Limit, Lbs. per SQ. In. Elonga- tion, per cent in 3.15 Ins. Red. of Area, per cent Brinell No. 1110 12, 550 9.7 20.8 248 1200 126,300 51,200 12.6 23.0 235 1290 108,650 54,040 10.3 23.0 185 1380 88,180 39,820 19.6 49.0 170 1470 91,590 36,980 16.7 36.8 178 1560 96,700 25,600 10.3 18.6 196 1650 105,100 29,580 6.1 10.0 207 1740 100,700 25,600 6.6 9.2 202 1830 116,050 36,980 6.0 6.8 228 Critical range Ac begins at 1337, maximum at 1346. 1.36 PER CENT. CARBON STEEL ANNEALED Treat- ment. Tests. Hard- ness. Microscopic. Annealed Deg. Fahr. Tensile Strength, Lbs. per sq. in. Elastic Limit, Lbs. per sq. in. Elonga- tion, per cent in 3.15 Ins. Red. of Area, per cent Brinell No. Structure. Notes. 1110 132,550 6.2 11.6 262 Free cementlte and pearl ite. Cementite reticulated, meshes filled partly with lamellar, partly with grainy pearllte. 1200 129,700 67,980 8.5 14.0 255 Free cementlte begins to change Into pear lite. 1290 123,450 48,355 9.6 16.2 288 Fine-grained cementite. Structure appears uniform. 1380 93,300 45,510 14.6 37.6 192 As before, grains finer. 1470 90,310 47,500 17.3 36.8 187 Free cementlte and grainy pearllte. Cementite concentrated again into small crystals. 1560 93,300 42,670 13.7 27.4 187 1650 102,100 32,710 4.5 5.0 209 Free cementlte, grainy and lamel- lar pearllte. Cementlte crystals larger, pearlite partly in lamellae. 1740 95,580 28,446 4.6 6.8 196 Free cementite and and lamellar pearllte. Structure essentially al- tered. Cementite form;? a network with large meshes. Steel Is over- heated. 1830 101,830 2.6 4.4 223 Critical range Ac begins at 1345, maximum at 1355. CHAPTER XI NICKEL STEELS Nickel Steel. Nickel may well be said to have been the pioneer among the common alloys now used in steel manufacture. Origi- nally added merely to give increased strength and toughness over that obtained in the ordinary rolled structural steel, the development and possibilities of heat treatment have greatly enhanced its value, so that nickel steel holds a premier position in alloy steel metallurgy. The chief difficulties attendant upon its use have been its tend- ency to develop a laminated structure and its liability to, seams. But when care is used in its manufacture and rolling, and it is not made in too large heats or ingots, and when piping and segregation are avoided by confining the finished product to that produced from the bottom two-thirds of the ingot, an admirable product for many purposes is obtained. Nickel steel has remarkably good mechanical qualities when subjected to suitable heat treatment and is an excellent steel for case hardening. In machining qualities it usually takes first place among the alloy steels. Strength and Ductility. Nickel primarily influences the strength and ductility of steel in that the nickel is dissolved directly in the iron or ferrite, in contradistinction to such elements as chrome and manganese which unite with, and emphasize the characteristics of the cement it ic component. Thus, for the forging grades of ordinary nickel steer in the natural condition, the addition of each 1 per cent, of nickel up to about 5 per cent, will cause an approximate increase of 4000 to 6000 Ibs. per square inch in the tensile strength and elastic limit over that of the corresponding straight carbon steel, without any decrease in the ductility. This influence of nickel upon the static strength of steel also increases to some degree with the percentage of carbon. To illustrate the effect of nickel upon steel in the natural condition: a steel with 0.25 per cent, of carbon and 3.5 per cent, nickel will have a tensile strength equivalent to that of a straight carbon steel with 0.45 per cent, carbon, a propor- 257 258 STEEL AND ITS HEAT TREATMENT tionately greater elastic limit, and the advantageous ductility of the lower carbon grade. Necessity for Heat Treatment. On the other hand, and in con- nection with the use of alloy steels in general, it should be borne in mind that such steels should be used in the heat-treated condition only that is, not in either an annealed or natural condition. In the latter conditions there is a benefit, as compared with straight car- bon steels and as illustrated above, but often is not commensurate with the increased cost. In the heat-treated condition, however, there is a very marked improvement in physical characteristics. And closely associated with this is the finely divided state of both ferrite and pear lite which characterizes heat-treated nickel steel. Nickel vs. the Critical Ranges. One of the most interesting phenomena connected with nickel steel is the effect of nickel upon the position of the critical ranges. Nickel, like carbon, has the property of lowering the points of the allot ropic transformations of iron, but in a more marked degree. Just as we have seen, in the chapter on Hardening, how " rapid cooling " and " carbon " are " obstructing agents " in preventing the transformation of the austenite into martensite into pearlite, so likewise does nickel act as an obstructing agent. The effect of nickel is obtained through a lowering of the Ar ranges, so that the temperatures of the critical ranges on cooling may even be brought below atmospheric temper- atures. Thus we may have a steel which, without quenching, may be pearlitic, troostitic, martensitic or austenitic, dependent upon the relative percentages of nickel and carbon. Hence, such steels containing nickel may be classified according to their microscopic constituents which are obtained upon slow cooling from a high temperature. Classification of Nickel Steels. In Fig. 150 there is shown graphically the influence of the nickel-carbon ratio upon the struc- ture of nickel steels as cast, or as moderately cooled from a high temperature. The " Pearlitic " nickel steels are those in which the critical ranges are all above the ordinary temperatures, so that such steels as slow cooled from a high temperature will consist of pearlite plus ferrite (or cementite). These are the ordinary commercial nickel steels, and are represented by the lower triangle of Fig. 150. " Martensitic " nickel steels contain that percentage of nickel and carbon which will so lower the position of the critical ranges on cooling that only the partial transformation may proceed. That NICKEL STEELS 259 is, the austenite is transformed into martensite, but no further the steel being too rigid to allow a more complete transformation at the low temperatures involved. These steels correspond to the mid- dle triangle in Fig. 150. Nickel steels martensitic throughout have no practical value, as it is impossible to work or machine them. On the other hand, great importance is attached to the use of cer- tain pearlitic nickel steels which can become martensitic upon case carburizing due to the increased carbon content. 20 Austenltic Martensitic 10 Pearlitic 0.40 0.80 1.20 1.60 FIG. 150. Influence of the Nickel-carbon Content upon the Structure of Nickel Steels as Cast. A still further increase in the nickel or carbon content will cause the critical range on cooling to fall below atmospheric temper- atures, so that such steels will be characterized by an " austenitic " or " polyhedral " structure, and are known under these names. Micrographic Structure. These changes in structure are illus- trated by the series of photomicrographs (by Savoia) given in Figs. 151 to 158, all the steels being in the natural condition, having approximately the same carbon content (0.25 per cent.), but with increasing percentages of nickel. 260 STEEL AND ITS HEAT TREATMENT FIG. 151. Steel with 0.25 per cent. Carbon, 2 per cent. Nickel. X650. (Savoia.) *^^^w aft* ^5 *6s L*$*1S FK;. 152. Carbon, 0.25 per cent. Nickel, 5 per cent. X650. (Savoia.) NICKEL STEELS 261 FIG. 153. Carbon, 0.25 per cent. Nickel, 7 per cent. X650. (Savoia.) FIG. 154. Carbon, 0.25 per cent. Nickel, 10 per cent. X650. (Savoia.) 262 STEEL AND ITS HEAT TREATMENT FIG. 155. Carbon, 0,25 per cent. Nickel, 12 per cent. X650. (Savoia.) FIG. 156. Carbon, 0.25 per cent. Nickel, 15 per cent. X650, (Savoia.) NICKEL STEELS 263 FIG. 157. Carbon, 0.25 per cent. Nickel, 20 per cent. X650. (Savoia.) FIG. 158. Carbon, 0.25 per cent. Nickel, 25 per cent. X650. (Savoia.) 264 STEEL AND ITS HEAT TREATMENT It will be seen that the structure of the 2 per cent, nickel steel (Fig. 151) is similar to that of a corresponding straight carbon steel, but is finer and more homogeneous. The 5 per cent, nickel steel (Fig. 152) shows a still finer and denser structure, in that the pearlite is more divided and distributed. With the 7 per cent, nickel (Fig. 153) the ferrite and pearlite are still seen, but they are distributed in a special manner as if disturbed by the approach of a transformation. A tendency to orientiate, somewhat like martensite, is also noticeable. Pearlitic Martensitic Austenilic 200,000 150,000 100,000 FIG. 159. Comparative Physical Properties of Nickel Steels with 0.25 per cent. Carbon, The steels with 10 and 12 per cent, nickel (Figs. 154 and 155) are both wholly martensitic. With the 15 per cent, nickel (Fig. 156) intensely white con- stituents appear amidst the martensite and probably represent the first appearance of austenite. The latter increases quite noticeably in the steel with 20 per cent, nickel (Fig. 157), taking on its poly- hedral form. At 25 per cent, nickel (Fig. 158) the whole steel is characterized by large polyhedra of gamma-iron. Physical Properties with Increasing Nickel. The physical properties of these same cast nickel steels are plotted graphically in NICKEL STEELS 265 the chart in Fig. 159. It will be noted that the abrupt changes in the curves correspond very closely with the theoretic structure given by the upper abscissae ; and that these same physical properties are indicative of the essential characteristics of pearlite, martensite and F. 1600 1500 1400 1300 1200 \ C. 0.20 0.40 0.60 0.80 FIG. 160. Critical Changes on Heating 3 per cent. Nickel Steel. 1.0 austenite, as denoted by the tensile strength, ductility (elongation) and resistance to shock. Critical Range of Pearlitic Steels. For nickel steels with less than 5 to 7 per cent, nickel, each 1 per cent, nickel lowers the crit- ical range (Acl) on heating by about 15 to 20 F., and also lowers the Arl range (cooling) by about 30 to 40 F., below those of the 266 STEEL AND ITS HEAT TREATMENT corresponding ranges for straight carbon steels of the same carbon and manganese contents. Similarly, there is a corresponding lower- ing of the other critical ranges. This is graphically shown in Fig. 160, which the author has care- fully plotted from a series of observations obtained with 3 per cent, nickel steels of various carbon contents. It will be seen from this curve that the critical ranges on heating are about 60 F. below the corresponding straight carbon steels. With the very low carbons there appears to be a tendency for the Ac3 curve to flatten out; this is further substantiated by results with steels containing 5 to 7 per cent, nickel. Beyond the eutectoid ratio of carbon it was found that the Ar range would begin to drop quite rapidly (not shown in the diagram) below its normal value, as might be expected from the fact that an increase in carbon in these steels act in an analogous manner to an increase in nickel. The approximate temperatures of the Acl and Ar ranges for nickel steels are as follows; Per Cent. Nickel. Acl, F. Ar, F. 1340 1280 1.0 1320 1240 2.0 1300 1200 2.5 1290 1180 3.0 1280 1160 3.5 1270 1140 4.0 1260 1120 4.5 1250 1100 5.0 1240 1080 6.0 1220 1040 7.0 1200 1000 It must be borne in mind, however, that the Acl temperatures may vary considerably from steel to steel but those given above will probably be about the average of those obtained in practice, and will in any event be within 25 F. On the other hand, the Arl temperatures given are liable to an even greater variation, as the maximum temperature attained in heating, the length of time occupied in both heating and cooling, the effect of the higher car- bon contents, and many other experimental factors, all tend to change the position of the Arl range. From these figures, and from the critical range diagram given for 3 per cent, nickel steels, it will be observed that the hardening of NICKEL STEELS 267 nickel steels may be carried out at temperatures considerably lower than those required by the corresponding straight carbon steels. The Eutectoid for Nickel Steels. The effect of additional nickel, or at least up to 7 per cent., is to reduce the eutectoid carbon ratio below that of the 0.9 value for straight carbon steels. That is, a nickel steel with 3 per cent, nickel will be saturated, having neither excess ferrite nor excess cementite (on slow cooling), at about 0.75 to 0.8 per cent, carbon; while in 7 per cent, nickel steel the eutec- toid ratio appears to be about 0.6 per cent, carbon. This fact is of great importance in case-hardening work, in that it not only permits of a shorter duration of the carburization in order to obtain the maximum carbon concentration necessary in the case, but also reduces the carbon content over which it is likely that enfoliation may occur. Heat Treatment of Pearlitic Nickel Steels. The heat treatment of pearlitic nickel steels presents some very interesting phenomena which are quite distinctive from ordinary straight carbon steels. One would naturally assume that the treatment of pearlitic nickel steels would be carried out in an analogous manner to that of the ordinary carbon steels that is, the quenching should be done at a temperature slightly in excess of the upper critical range, provided that the duration of heating at the maximum temperature has been sufficient to effect the entire solution of the previous components, together with their diffusion and the equalization of the steel as a whole. Similarly, as in carbon steels, we would assume that we might replace the length of heating at the proper quenching temper- ature by a higher temperature in order more quickly to effect the equalization of the steel; provided, however, that this new and more elevated temperature shall not produce too great a deteriora- tion in the metal through increase in grain-size, etc. or that this higher quenching is followed by a quenching at the proper tempera- ture. In straight carbon steels the change of structure by heating slightly above the upper critical range takes place quickly as a general rule; and the coarsening or embrittling of the steel also occurs rapidly when higher temperatures are used. The influence of nickel in the steel, however, often necessitates a modification, or permits a simplification, of these general principles, both in regard to the temperature of quenching and the length of heating. In the first place, the addition of nickel appears to make the solution of the ferrite or cementite and the equalization of the steel as a whole take place more slowly than in the ordinary carbon 268 STEEL AND ITS HEAT TREATMENT steels. Thus, if we take a steel containing some 4 or 5 per cent, nickel, and a mild or medium carbon content, and quench it after a normal heating at a temperature some 50 F. over the critical range, the transformation is often incomplete and the martensite not uniformly distributed nor equalized. In such an event, which is usually characteristic of nickel steel which has either undergone a more or less severe elaboration or work- ing, or has been finished at too low a temperature, or has been sub- jected to a prolonged heating at some high temperature, there are then four methods of procedure available: (1) Prolonged heating at the proper quenching temperature to effect the necessary transformation, followed by quenching; (2) Heating to a higher temperature than in (1), and quench- ing; (3) Heating to the higher temperature, cooling to a temper- perature a little above the Ar temperature, and then quenching; (4) Quenching, or air cooling, from the higher temperature, followed by a normal reheating to a temperature slightly hi excess of the Ac range, and quenching. These propositions at once evoke a discussion of further char- acteristics which the presence of nickel involves. If an ordinary carbon steel be heated for a considerable duration of time at a temperature even slightly over the critical range, the grain-size will begin to increase, with a corresponding decrease in both the static and dynamic strength of the material. On the other hand, if a nickel steel be subjected to a length and temperature of heating equivalent to that of the carbon steel, the pearlite and ferrite grains will remain (after slow cooling) considerably finer, more uniformly distributed, and much more subdivided than the carbon steel. This characteristic permits the greater duration of heating as required under the first proposition, without any per- ceptible deterioration such as would be noticeable in a straight car- bon steel with the same prolonged heating. However, such treat- ment when required is disadvantageous from the commercial standpoint, as it decreases the capacity of the heat treatment plant, with a corresponding increase in the cost of production. Again, the increased brittleness due to a more or less prolonged heating at temperatures in excess of the upper critical range is con- siderably less for nickel steels than for ordinary carbon steels. This NICKEL STEELS 269 fact is well illustrated by the following results upon a straight carbon steel in comparison with a 2 per cent, nickel steel of the same carbon content, taken from a memoire l by Guillet: Length of Heating Resistance to Shock. at 1830 F. Ordinary extra-soft steel. Extra-soft steel with 2 per cent nickel. Normal heating 20 kgms. 60 kgms. (not broken) Four hours 4 . 5 kgms. 60 kgms. (not broken) Six hours 4.0 kgms. 60 kgms. (not broken) If, in order to obtain the full equalization of the steel and also to avoid a prolonged heating at the lower and theoretic temperature, it shall be necessary to heat and quench from a higher temperature, such operation may be undertaken without that fear of greatly increasing the brittleness which would most probably occur in a straight carbon steel. Although it is granted that a heating to this high temperature may be necessary, a quenching from this same high temperature would not be entirely logical if this were to be the only quenching, and also if viewed from the standpoint of the best product. In such high temperature quenchings there is the ever-present danger of cracking and warping. Further, it is a generally admitted fact that no change in the molecular arrangement of the steel occurs in cooling such a steel until the upper critical range on cooling (Ar3) is reached. Assuming this to be true, we may then modify the previ- ous treatment (proposition 2) by first cooling the steel after heat- ing to the high temperature to a temperature slightly above that of the Ar3 range, and then quench, as stated under proposition 3. This treatment will retain all the benefits which may accrue from the original high-temperature heating, but at the same time will diminish to a considerable degree the dangers of cracking and warping. And as the critical ranges on cooling in nickel steels are even further below those of the Ac ranges in comparison with straight carbon steels, this quenching temperature will be reason- ably low. Objections which may be offered to this method are that the quenching from just over the Ar range may not give the maximum *M. L. Guillet, " Traitements thermiques des aciers speciaux," Rev. de Met., July, 1910. 270 STEEL AND ITS HEAT TREATMENT hardening effect unless the quenching temperature has been gauged just rightly, or if the carbon content is low. The first objection may be overcome by suitable temperature control; if the quenching temperature should fall too low, the difference in the hardening effect, for forgings or full-heat treated work, may be later corrected by using a lower toughening or drawing temperature. By the use of exact methods, such as one furnace maintained at the high tem- perature, and then another furnace (into which the steel may be subsequently placed) maintained at the temperature a little over the Ar range, the first objection may be entirely cleared away. The second objection may also be at once overruled by the fact that the treatment of the low-carbon steels is generally limited in com- mercial work to the obtaining of a suitable toughness and absence of brittleness (regeneration), and that it is usually not desired to obtain the maximum hardness. In brief, it does not matter whether the same mechanical prop- erties in a pull test are obtained by a quenching made at a very high temperature, or by a quenching at a lower temperature follow- ing the return. As these results in the mechanical properties are practically the same, the treatment under proposition 3, as compared with Number 2, is always more advantageous from the point of view of non-brittleness and probably also from the point of view of the strength of the piece. The most serious objection to the treatment in either (2) or (3), however, is the increase in brittleness which is liable to occur if the high temperature heating is unduly prolonged. Although the presence of nickel tends to diminish such a condition, the effect of high heating is always towards the increase in grain-size, and coarse martensite generally corresponds to a diminution in the strength of the steel. Assuming that a temperature considerably in excess of the upper critical range is mandatory, any ill effects resulting therefrom may be entirely overcome by a double heating and cooling, and yet also retaining the benefits of such high temper- ature heating. That is, by cooling the steel but not quenching, unless the original structure is very bad indeed; or unless the most perfect structure is desired from the high temperature to a tem- perature under that of the Al range, in order to impress the effect of the high temperature upon the steel, followed by a reheating to a temperature slightly in excess of the upper critical range, and then quenching. Such a hardening treatment, either with air cool- ing and a subsequent single quenching, or with a double quenching, NICKEL STEELS 271 is the best, although the most expensive. As this method has been discussed in its relation to carbon steels, and as its influence is approximately the same with pearlitic nickel steels, it will not be necessary to dwell further upon it. In general, the treatments (for the best quenching effect) given under (1), (2) and (3) will suffice for ordinary commercial practice, but with the preference given to (1) or (3). That under (4) is best if the higher cost is permissible. Moreover, it should be borne in mind that, in perhaps even a majority of cases, the regular and normal quenching from a tem- perature slightly in excess of the upper critical range (Ac3), after a thorough and uniform heating at that temperature, will generally suffice and especially for small work. But in order more fully to explain the difficulties which are sometimes met with in the treat- ment of nickel steels, the author has entered into the foregoing explanations. As a safe and general fundamental principle, re- peatedly urged, it is always advisable to quench from the lowest temperature which will give the desired results. The tempering and toughening of pearlitic nickel steels is carried out exactly as with straight carbon steels, previously explained, and is dealt with in more detail later on in the chapter. CARBURIZATION OF NICKEL STEELS The general principles of the carburization of nickel steels are similar to those which apply to straight carbon steels, and should not require repetition. There are, however, certain peculiarities, presenting both advantages and disadvantages, which should be mentioned. (1) Nickel steels show less susceptibility to brittleness due to prolonged heating at the high temperatures often used in carburiza- tion than do the corresponding carbon steels. This important fact not only gives a steel better able to withstand shock, but also gives a well-defined means of simplifying the subsequent heat treatment if desired. Such advantages may be readily obtained by the addi- tion of even 2 per cent, of nickel, and largely compensate for the slightly higher cost. (2) The variations in the concentration of the carbon in the carburized zones are more gradual and uniform in nickel steels than in carbon steels. This better distribution of the carbon therefore tends towards the prevention of a distinct line of demarkation be- tween the different zones, and thus to eliminate the chipping and 272 STEEL AND ITS HEAT TREATMENT flaking off of the case. Similarly, the phenomenon of " liquation, "- a principal factor in such enfoliation is less marked, under equal conditions, in nickel steels than in carbon steels. (3) Although it is true that carburization proceeds with greater slowness with some solid carburizing compounds, referring to their use with nickel steels with less than say 3.5 per cent, nickel, the use of a mixed cement (carbon monoxide plus carbon) will effect a car- burization with a rapidity equal to that with ordinary carbon steels. (4) Under the same conditions, the depth of penetration of the carburized zone for a given time, using a mixed cement, is even slightly higher for nickel steels than for carbon steels. (5) The lesser hardness which, with the same treatment, is possessed by the carburized zones in nickel steel as compared with the carburized zones in carbon steels under identical conditions, is due not only to the different effects of different quenchings, but also to the smaller concentration (especially for less than 3 per cent, nickel) of carbon in the carburized zones. This disadvantage may be eliminated by raising the carbon in the carburized zone by a suitable change in the conduct of the carburization. (6) When the nickel content exceeds 3 per cent, the maximum concentration of the carbon in the carburized zones decreases with an increase in the percentage of nickel contained in the steel. The following table, from Giolitti, 1 contains data relative to the maximum concentration reached by the carbon in the carburized zones when carburizing, under various conditions, steels with varying nickel contents : Condition of Carburization. Nickel Content. 2% 3% 5% 25% 30% Carbon monoxide: 5 hours at 1740 F 0.38 0.35 1.53 0.23 0.35 0.93 1.28 0.70 0.80 0.73 0.74 0.83 0.15, 0.17 0.39 0.63 0.67 0.40 5 hours at 1920 F Ethylene: 5 hours at 1740 F 1.12 5 hours at 1920 F 0.84 0.64 0.59 0^73 Mixed cement: 2 hours at 1830 F 0.70 1.12 0.83 0.92 1.07 5 hours at 1830 F. 5 hours at 1920 F 2 hours at 2010 F 5 hours at 2010 F 1 F. Giolitti, " The Cementation of Iron and Steel.' NICKEL STEELS 273 (7) By employing a nickel steel of the proper nickel content, and carburizing in such a manner as to attain a definite maximum carbon concentration in the case, a steel characterized by a tough core and a martensitic structure in the case may be obtained with- out subsequent quenching. The approximate maximum carbon con- centration in the case which it is necessary to obtain for steels with definite percentages of nickel in order to produce a martensitic struc- ture without quenching, may be given about as follows: Per cent. Nickel. Per cent. Carbon. Per cent. Nickel. Per cent. Carbon. 2 1.50 5 0.95 3 1.30 6 0.85 4 1.10 7 0.75 Such methods eliminate the necessity for subsequent heat treat- ment, if so desired, and effect corresponding reductions in the cost of the process, besides obviating, in a large measure, such important factors as warping, grinding, etc. Further, by extending the car- burization so as to reach a maximum of 1.5 per cent, carbon at the periphery, for steels containing 5 to 7 per cent, nickel, there can also be obtained a superficial layer, superimposed upon the mar- tensitic zone, containing austenite, which easily admits of polishing without loss. (8) The lower temperatures at which the critical ranges are located, in the pearlitic nickel steels, permit a lower temperature to be used in case carburizing, which is an important factor in intricate or exact work. THERMAL TREATMENT AFTER CARBURIZATION In general, the thermal treatment of nickel steels, subsequent to case carburizing, may be classified according to the structure of the case after slow cooling from the temperature of carburization that is, whether it is pearlitic or martensitic. Although this struc- ture depends primarily upon the conduct of the carburization and the maximum carbon concentration thus obtained in the case, the procedure as practically carried out in commercial work will usually give (upon slow cooling after carburization) (1) a pearlitic structure in the case for steels with nickel contents under 4 per cent, and (2) a more or less martensitic case for steels with 4 to 7 per cent, nickel. 274 STEEL AND ITS HEAT TREATMENT As explained in Chapter VII, the best treatment which can be given any case-carburized pearlitic steel is that involving a double quenching. Each quenching for regeneration and for hardening should be carried out at the most suitable temperature, and which is fixed by the transformation points of the core and case respectively. These treatments for nickel steels with 2 to 2J and 3 to 3J per cent, nickel are approximately as follows: Carburization. Carburize at the desired temperature, usually 1600 to 1750 F. Cool slowly in the carburizing material (assuming solid cements). Thermal Treatment. Nickel Content, Per cent. 2 to '* l|. 3 t( ) 3*. Carbon Content, Per cent. 0.10 to 0.15 0.15 to 0.20 0.10 to 0.15 0.15 to 0.20 Regenerative quenching (a) 1550-1600 1500-1550 1475-1525 1450-1500 Hardening quenching or (6) 1600 1325-1375 1600 1325-1375 1600 1300-1350 1600 1300-1350 The steel may be removed from the quenching bath as soon as it loses its red color during the regenerative quenching, and imme- diately reheated for the second quenching. Practice differs as to the temperature to be used for the regenerative quenching, some preferring to quench from a temperature slightly above the Ac3 range, as under (a), while others prefer to use the higher temper- ature (6). The reasons for these have been discussed in previous sections. On the other hand, for 2 per cent, nickel steels, Guillet recom- mends temperatures distinctly higher than those given above which probably coincide with the best American practice for the regenerative quenching, and which he gives as follows: Regenerative quenching 1760 to 1800 Hardening quenching 1365 to 1420 The hardening quenching should be conducted at the lowest pos- sible temperature at which the metal of the case will become glass- hard. In many instances it will be found that temperatures some- what lower than those given in the above table can be used. For NICKEL STEELS 275 example, the critical curve given in Fig. 161 for a steel with 0.13 per cent, carbon, 0.49 per cent, manganese and 3.35 per cent, nickel, shows the Acl range to be about 1250 F., so that a temperature under 1300 to 1350 might be used effectively for the hardening quenching. The effect of different treatments upon Quillet's 2 per cent, nickel steel in its resistance to shock is shown in the following table: Treatment. Resistance to Shock. Steel with 2 per cent, nickel and 0.1 per cent, carbon: Heated to 1700 F. and air cooled 33.4 kgms. Quenched in water from 1700 F 34 . 5 Same steel cemented at 1830 F. for 1.2 mm.: Slow cooled 31 .0 Quenched in water from 1830 F 33.5 Double-quenched in water, 1830 and 1380 36.0 Quenched in water from 1380 F , , , , , , 32, FIG. 161, Critical Range Diagram. Simplified Thermal Treatments after Carburization. On account of the fact that the brittleness of the core (with nickel steels) is not greatly increased by the heating during carburization if the tempera- ture of that operation is not too high, and as the Ac3 range of the ordinary nickel steels is considerably lower than that of the corre- sponding straight carbon steel, it makes it possible, as we have seen, to effect a regenerative quenching at a temperature in the neighbor- hood of 1500 -1550 F, Further, as the nickel steel case can be 276 STEEL AND ITS HEAT TREATMENT hardened at a temperature considerably above the normal Acl without losing too much of its hardness or increasing too largely in brittleness, it follows that, in many instances, the regenerative quenching may also serve as a hardening quenching. This permits the simplification of the treatment to a single quenching for nickel steels, unless the piece is to be subjected to exceptional stress. The practical usefulness of this method is obvious. It is evident, however, that the double quenching will always give considerably better results for both core and case. This is particularly shown in the depth and degree of hardness obtained by the lower quenching over the higher quenching temperature by the following experiments by Guillet on 2 per cent, nickel steels: Shore Hardness Numbers. Treatment. Maximum. Minimum. Mean. Cemented pieces, not quenched 40 39 39.37 Cemented pieces, quenched from 1830 F. 84 69 78.05 Cemented pieces, quenched from 1380 F. 88 85 86.56 Case Hardening by Air Cooling. Again, the case-hardening process may be even further simplified by the use of nickel steels with 3.5 per cent, of nickel, or more, and conducting the carburization in such a manner as will produce a maximum carbon content in the case, which will give a martensitic structure on air cooling from the tem- perature of carburization. An example of this is shown in Figs. 162 and 163, representing a case-carburized steel with an initial carbon content of 0.176 per cent., with 3.44 per cent, nickel; the steel was then air cooled directly after carburization. The thickness of the martensitic zone is about 0.5 mm. Under the lower magnification (Fig. 163) a solid troostitic band is seen to separate the martensite and the sorbito-pearlite portions. The principal advantage which this method presents consists of its great simplicity, and also in the fact that it permits the avoidance of deformation which so often accompanies any quenching operation. Nickel steels which are martensitic after air cooling may be troostitic, sorbitic, or even pearl- itic after very slow cooling in the furnace, while they may be austen- itic on water quenching. Case Hardening 5 to 7 Per Cent. Nickel Steels. Advancing another step and using nickel steels with 5 to 7 per cent, nickel, we find that the ordinary carburization and subsequent slow cooling NICKEL STEELS 277 FIG. 162. Nickel Steel. Nickel, 3.44 per cent. Carbon, 0.176 per cent. Case Carburized and Air Cooled. X 100. (Sauveur and Reinhardt.) FIG. 163. Same Steel and Treatment as in Fig. 162. (Sauveur and Reinhardt.) X50. 278 STEEL AND ITS HEAT TREATMENT will produce a case with characteristics varying over a wide range, dependent upon the nickel-carbon ratio in the case. In other words, the transformation range of the metal of the case on cooling may be even further reduced below that of the previous example, giving either a martensitic or martenso-austenitic structure upon slow cool- ing. Therefore, when it is not required to produce an extremely tough core, nor to obtain extreme hardness in the case, the carbur- ized pieces with 5 to 7 per cent, nickel may simply be allowed to cool slowly in the carburizing mixture after carburization. The use of the method just indicated, however, has its dis- advantages. The following table shows the results of scleroscope hardness tests made by Guillet on a steel containing 7 per cent, nickel and 0.12 per cent, carbon, carburized to a depth of 0.1 mm., but not quenched: Treatment and Tests. Shore Hardness Numbers. Mean. Maximum. Minimum. Test made on the surface. 18.5 26.5 24.5 20.2 21 27 25 20 17 26 24 21 Test made at a depth of 0.2 mm Test made at a depth of 0.4 mm ....... Test made at a depth of 0.6 mm From this particular instance it will be seen that the surface zone is partly austenitic, so that a very great hardness is not obtained. In the second place, it is evident that the core of the piece thus treated has not been regenerated, although, as we have said before, nickel steel does not have the maximum brittleness which a straight carbon steel would have under the same conditions of cooling. The structure of a steel containing 4.86 per cent, nickel and 0.115 per cent, carbon, intensely carburized, and air cooled, is shown in Fig. 164. This photomicrograph shows that the effect of such treatment is to produce a case which is largely austenitic. The best practice, however, both American and foreign, specifies the use of a double-quench treatment subsequent to a mild carburiza- tion, and using a soft steel with 4.5 to 6 per cent, nickel. Such steels have many peculiar advantages: the carburization may be con- ducted at a moderate temperature; a maximum carbon content in the carburized zone of only about 0.45 to 0.6 per cent, is necessary to produce a glass-hard surface on oil quenching; and the core becomes exceedingly strong, as well as tough and non-brittle. From NICKEL STEELS 279 these facts it is evident that the lowering of the maximum carbon concentration to a percentage not exceeding that of the eutectoid ratio will almost entirely eliminate the danger of chipping and flaking. The use of moderate temperatures for carburization and of oil for quenching decreases the liability to warping and fracture. The physical characteristics of the carburized zone after the second oil quenching, the steel of the case having an approximate chemical FIG. 164. Nickel Steel. Nickel, 4.86 per cent. Carbon, 0.115 per cent. Case Carburized and Air Cooled. XlOO, (Sauveur and Reinhardt.) composition of 0.45 per cent, carbon and 5.0 per cent, nickel, will be approximately: Tensile strength, Ibs. per sq. in 260,000 Elastic limit, Ibs. per sq. in 250,000 Elongation in 2 ins., per cent 2 Reduction of area, per cent 5 Brinell hardness 490 Scleroscope hardness 74 280 STEEL AND ITS HEAT TREATMENT The following physical results taken from the core of a double- quenched steel analyzing: Carbon.. 0.105 Manganese . 43 Phosphorus 0. 014 Sulphur 0.030 Silicon 0.11 Nickel 5.0 show that the core will have great strength, high ductility (from the reduction of area) , and very slight brittleness (as shown by the shock test) : Tensile strength, Ibs. per sq. in 200,000 Elastic limit, Ibs. per sq. in 170,000 Elongation in 2 ins., per cent 12 Reduction of area, per cent 54 Resistance to shock 75 Brinell hardness 295 The same steel, having an upper critical range of 1425 F., and annealed at 1475 F., gave: Tensile strength, pounds per square inch. . . . 90,600 Elastic limit, pounds per square inch 60,160 Elongation in 2 ins., per cent 20 Reduction of area per cut 60. 5 Resistance to shock 116 Brinell hardness . 179 The regenerative quenching for these steels may be carried out either at a temperature slightly in excess of the upper critical range - or at about 1475 F., or, in order to more fully equalize and effect the regeneration of the core, at some higher temperature, such as 1600 F. The hardening quenching temperature should be slightly over the Ac range of the case, or approximately 1275 to 1325 F. Oil may be used for both quenchings. For the characteristic French steel containing about 6 per cent, nickel Guillet recommends the temperatures of 1560 and 1250 F. respectively for the double quenching. 3.5 PER CENT. NICKEL STEEL We have previously discussed some of the factors which enter into the quenching of nickel steels in general. Whether or not it NICKEL STEELS 281 (9) jaqtnn^ asaap-rn jj adoosoaapg a-ioqg g (g) jaqtn Previ to 18 coole Quenc fro draw 33S ipai aaunbg aad spanoj '(z) iniri 8 (f ) wajv J aotaonpaa put! () saqDnj j; a; 282 STEEL AND ITS HEAT TREATMENT (g) Jaqmnjf ssanpanH edoosoaspg oa 8 g g (g) jaqmnfj \ 13335 o o o' o o a a -- D g PP 02 tn ^-, aiwnbg jod epnnoj '(%) li g '(l)qiST: '(5) uaay jo noponpaH pau NICKEL STEELS 283 (9) aaqtunjj ssanpjwjj adoaso-ispg 8 g 7 \/ TV lull. li "S-S I !! J35 \ bg asd spunoj fuaO aaj '(t) weay jo (g) saqoni g u; 284 STEEL AND ITS HEAT TREATMENT (9) jaqranji ssaapJUR adoosoaops 3-iou.s S (g) aaqran^ \ ical Ranges : Size Ac 1240 1 in Ar 1060 ical Analysis C. 0.39 Mn. 0.48 P. 0.009 S. 0.023 Si. 0.10 Ni. 3.36 I 1 ! uoui eavnbg jad spunoj *(z) iniri D !? B 13 P" '(i) qVaea^g 8U 8UO X 2 s NICKEL STEELS 285 (9) jaqtnnjj ssanpavH adoosojspg aaoqg (g) jaq 7 \ / 7 iot$ 1 -2 of a 1400 n as \ Is ^ 1600^ De^r *aJj W,,,.,. 12001 e* Fahr III K 1 l/ 2 2 2'/i 3 Size in Inches FIG. 171. Effect of Mass upon the Hardness of Nickel Steel, Water Quenched. (Matthews & Stagg.) Again, these results represent the treatment of l-in. round sections so that while such results might be duplicated in practice with sections up to 1J ins. in diameter, further increase in the size of section will inevitably lower the physical test results to be obtained under the same treatment. The influence of mass upon the Brinell hardness is shown in Figs. 170 and 171. On the other hand, the normal characteristics for annealed 3J per cent, nickel steel, as given in the diagram in Fig. 172, represent the 288 STEEL AND ITS HEAT TREATMENT average results which are, and should be, obtained in commercial practice in the annealing of the more common and larger sections of nickel steel. They represent, moreover, the minimum requirements which are characteristic of many existing steel specifications for 3J per cent, nickel steel, annealed, for such uses as engine forgings, 100,000 0.35 0.40 Per Cent. Carbon FIG. 172. Normal Characteristics of Annealed 3.5 per cent. Nickel Steel. Large-size Sections of Forgings. Manganese Approx. 0.6 per cent. ordnance forgings, rolled slabs and billets, etc., both for Govern- ment and commercial uses. Similarly, the following physical results for heat-treated work (quenched and toughened) have been taken from various specifica- tions in order to show the minimum results which may be expected in commercial practice. The manganese requirements are approx- imately 0.50 to 0.70 per cent., and the nickel content not less than 3.25 per cent. NICKEL STEELS 289 ! 5 I J o I g f i o O* | g ' Tl ' V g "S o M e3 "H 3 g i 1? 3 . a> "' . 9 O CQ i S | g Sill c 0) H 1 S- BB "3 V a d i H 08 G IO o * 5 13 a> "3 *" "3 (-. ; +j a> a c cs c! 15 o> C ~ cd I a 5 S 3 jt?T w 5 q) ^ o O w W < * ^ H S <5 O < W 1 | 1 g 1 3 CO ^ 4) 0) S &? ? : : S e3 cS . - i < i 09 00 u a a SB OB i OB 0) OB OB d 3 0? ^ d o O 1 i. * ! i 1 ft 8 e H H H H H m ~f. S 9 * a w M w w o K H o> ^ J J j- j O O a Tr g -^ o $ ^ 3 p 00 ^5 ^ ^ H od QQ OQ OQ QQ ^ < S *- fc O O JOQ ad i5 ai of g g' .C 'C c C x x f 1 O d s o 'C o d S d N d I I O o 1 o CO i CO S S o d o ic c o IfiflJ fl!Hi mill ! lul S19 | *S8 oo ^ 5c>o d / / / \ / / \ / 1 / 1 / / i 1 Hardened 1 -1 01 i 1 1 1 1 gill qoaj anmbg i^d spanoj *(j) is^n 3 JW B (3: POB '([) qi3a3ng 8(;saaj, S* S s S ;aao wj (0 Baiy )o aoninpaa pao (?;) s0 (XX) ^210000 : 200 ooo g lyoooo .5 180000 M 170000 SlGOOOO "uo 150 000 ^ 140000 W 130 000 ^ 120000 3110000 100 000 = 80000 5! 70000 02 CO 000 03 o o o o o o o o o o o o 35 o p 1-1 1-1 I-H O T-i T I 1 I 1 I 1 I T i I I I I IO CO C5 C5 er Sq. In. Elastic Limit, Lbs. per Sq. In. Elon- gation, Per Cent in 2 Ins. Reduc- tion of Area, Per Cent. Bend Test, Deg. Alter- nat- ing Str'gth Brinell Hard- ness. Sclero- scope Hard- ness. 13 79,070 41,660 31.1 58.75 180 283 143 16 14 77,060 34,720 36.6 68.23 180 273 143 18 15 83,220 38,640 32.2 57.5 180 215 156 18 16 87,580 45,140 22.2 35.5 171 108 170 20 17 92,290 48,830 16.1 24.0 85 66 187 22 Heat Treated (Hardened in oil, reheated to 1025 F.) 13 82,080 53,760 30.9 65.6 180 239 163 15 14 105,350 75,450 25.3 54.4 180 226 351 30 15 127,400 77,800 21.1 49.2 180 122 444 39 16 247,300 216,830 7.74 23.2 34 33 512 42 17 4 24 512 46 CHAPTER XVI TOOL STEEL AND TOOLS THE problem of selecting a proper grade of steel in relation to the work required is one hitherto met by the steel manufacturer alone. Until recently he has recommended this or that steel for a given requirement, depending more or less upon his general knowl- edge of the purpose for which the tool is to be used, and upon the experience of his customers in the past. But with the entrance of the technical man into manufacturing concerns and the great im- provements resulting therefrom, a fuller knowledge of various steels, their composition, applicability and efficiency has been demanded. This has resulted in a wider dissemination of information regarding the physical, chemical and mechanical properties of steels manu- factured by various steel companies, and a corresponding education of both maker and buyer. Grade. For the aid and information of their customers, the steel maker usually groups his tool-steel products into various " grades " and " tempers." The former term refers to the " quality " of the steel, according to the class of raw material which has been used, together with the skill and care taken in producing the finished material. The highest grades should be used for tools operating under severe working conditions, demanding great endurance and resistance to torsional or other strains, or upon which a large labor cost has been placed. These conditions, such as are found in expen- sive dies, milling cutters, taps, etc., would require a high-grade steel. For such purposes as mill-picks, cheap tools, etc., it would be folly to use any but a lower-quality steel. Wear, the cost of redressing, regrinding and heat treatment are other factors which must be considered in the selection of the proper and most economical steel which will give the greatest efficiency in all senses of the word. With this in mind, the following brief synopsis is given: 1. Finest tools and dies: expense for material the smallest item entering into the cost and upkeep of the finished tool ; 2. Finishing tools for lathe and planer work; special taps, 357 358 STEEL AND ITS HEAT TREATMENT reamers, milling cutters and other similar tools requiring a high- grade steel; wood-working and corrugating tools; 3. General tool purposes; 4. Ordinary purposes, such as chisels, smith and boiler shop work, etc. 5. For rough or heavy work. Expressing this in a different way, we may say that the choice of a grade of tool steel depends upon three factors : 1. The precision of the work required of the tool; 2. The relative cost of the steel in comparison with the labor involved in the manufacture of the tool; 3. The life of the finished tool and its relation to the cost of pro- duction. Temper. Carbon tool steels are further denoted by the " tem- per." In tool-steel sales parlance this refers to the percentage of carbon in the steel and may be denoted by figures or letters. Such classifications generally refer to a 10-point carbon limit thus No. 7 temper may refer to 0.65 to 0.75 per cent, carbon, or it may be represented by whatever the individual company has arbitrarily selected. In this connection it should be noted that this " temper " does not refer to, and should not be confused with the word temper as indicating the operation of " letting down " the steel after hardening. General recommendations for the proper carbon content to use for various tools are given in the following table; these, however, must not be regarded as absolute, for much will depend upon the grade of steel and upon the exact use of the tool, APPROXIMATE CARBON CONTENT FOR ORDINARY TOOLS Carbon, rp i Per Cent. 1.50 Tools requiring extreme hardness. For turning chilled- rolls and tempered gun-forgings. Roll corrugating. 1.40 Hard lathe work generally. Chilled-roll turning. Cor- rugating. Graver tools. Brass-working tools. 1 . 30 General lathe, slotter and planer tools. Razors. Drawing dies, TOOL STEEL AND TOOLS 359 Mandrels, granite points, scale pivots, bush hammers, peen-hammers. Ball-races. Files. Trimming dies. Cutting dies. 1 . 20 Twist drills. Small taps. Screw dies, threading dies. Edge tools generally. Cutlery. Cold stamping dies, leather-cutting dies, cloth dies, glove dies. Nail dies, jewelers' rolls and dies. 1 . 10 Milling cutters and circular cutters of all descriptions. Wood-working tools, forming tools, saws, mill picks, axes. Small punches. Taps. Cup and cone steel. Small springs. Anvils. 1.00 Reamers, drifts, broaches. Large milling cutters, saw swages. Springs. Mining drills, channeling drills. Large cutting and trimming dies. 0.90 Hand chisels, punches. Drop dies for cold work, small shear knives. Chipping chisels. Cutting and blanking punches and dies. 0.80 Large shear knives, chisels, hammers, sledges, track chisels. Cold sets, forging dies, hammer dies, boiler-maker's tools. Vise-jaws. Oil-well bits and jars. Mason's tools. 0. 70 Smith shop tools, track tools, cupping tools, hot sets. Set screws. 0.60 Hot work and battering tools generally. Bolt and rivet headers. Hot drop forging dies. Rivet sets. Flatteners, fullers, wedges. 0.50 Machinery parts. Track bolt dies where water is con- tinually running on dies (hot work). Navy Specifications. The United States Navy specifies the following straight carbon tool steel for its general requirements : 360 STEEL AND ITS HEAT TREATMENT Class. I. II. III. IV. Carbon Manganese 1.25-1.15 0.35-0.15 1.15-1.05 0.35-0.15 0.95-0.85 0.35-0.15 0.85-0.75 35-0 15 Phosphorus Sulphur . . . 0.015-0 0.02-0 0.015-0 0.02-0 0.02-0 0.02-0 0.02-0 025-0 Silicon 0.40-0 10 0.40-0.10 0.40-0 10 40-0 10 Chrome and vanadium optional. Class I. Lathe and planer tools, drills, taps, reamers, screw-cutting dies; taps and tools requiring keen cutting edge combined with great hardness. Class II. Milling cutters, mandrels, trimmer dies, threading dies, and general machine-shop tools requiring keen cutting edge combined with hardness. Class III. Pneumatic chisels, punches, shear-blades, etc., and in general tools requiring hard surface with considerable tenacity. Class IV. Rivet sets, hammers, cupping tools, smith tools, hot-drop forge dies, etc.; tools requiring great toughness combined with necessary hardness. The Navy Department also maintains the requirements as to grade by requiring a steel which will stand rehardening a specified number of times without cracking. General Properties. The following table shows the relative toughness and hardness of tool steel of the different carbon contents : Carbon, Per Cent. . 50 Toughness only. 0.60 Great toughness with properties suitable for hardening and tempering. 0.70 Excellent toughness, but with cutting edge. 0.80 Tough tool steel, withstanding shocks, etc. 0.90 Good cutting edge but with toughness an important factor. 1 . 00 Toughness and cutting edge about equal. 1 . 20 Great hardness combined with some toughness. 1 30 Great hardness in cutting edge. Toughness slight factor. 1.40 Extreme hardness in cutting edge first requirement. Toughness slight factor. Some metallurgists consider that it is safer to select a too hard steel and draw the temper at a higher temperature than to choose a too TOOL STEEL AND TOOLS 361 soft steel with a view to increasing its hardness by a weaker temper- ing. Opposed to this is the fact that the higher the carbon content the more the care which will be required in the harden- ing operation, since the steel becomes more sensitive to overheating. GENERAL TEMPERING COLORS FOR TOOLS Faint yellow: Steel-engraving tools. Light turning tools. Hammer faces. Planing tools for steel. Ivory-cutting tools. Planing tools for iron. Paper-cutting knives. Wood-engraving tools. Light yellow: Milling and other circular cutters for metal. Bone-cutting tools. Scrapers for brass. Shear blades in general. Boring cutters. Leather-cutting dies. Screw dies. Inserted saw teeth. Taps. Rock drills. Chasing tools. Penknives. Straw: Dies and punches in general. Moulding and planing cutters for hardwood. Reamers. Gouges. Brace bits. Plane irons. Stone-cutting tools. Deep straw: Twist drills. Cup tools. Wood borers. Circular saws for cold metal. Cooper's tools. Augers. 362 STEEL AND ITS HEAT TREATMENT Brown. Drifts. Circular cutters for wood. Dental and surgical instruments. Axes and adzes. Saws for bone and ivory. Peacock: Cold sets for steel and cast iron. Hand chisels for steel and iron. Boiler-maker's tools. Firmer chisels. Hack saws. Purple. Moulding and planer cutters for soft wood. Smith tools and battering tools generally. Blue : Screwdrivers. Saws for wood. Springs in general. These colors are for general crucible steel with low manganese. Their applicability to particular work and special steels may be taken 1460 1440 51420 >1400 1380 1360 Diameter in Inches FIG. 199. Temperature-size Curve for Hardening Tools. in a general way, but that temperature must be adopted which will suit the special work or steel in hand. Hardening. We have previously discussed the fact that with any increase in the mass of the steel there is a corresponding decrease in both the maximum surface hardness and the depth of hardness, when quenched from the same temperature. This difference in hardness is due to the difference in the rate of cooling of the small TOOL STEEL AND TOOLS 363 and large sections. In order to produce the same degree of hard- ness in a small and large section, as applied to small tools, it is neces- sary to heat the larger section hotter for hardening than the smaller. To illustrate: Matthews and Stagg have worked out the relation of mass to temperature for one particular grade of the same tool steel in which the sizes varied from Y& m - diameter to f in. diameter, and 1200 1300 1400 1500 1600 FIG. 200. Loss of Hardness Due to High Hardening 1700 1800 Tempera tures . (Shore . ) found that a difference of about 60 F. in heating was necessary to produce the same degree of hardness in the two extreme sizes. Their temperature-size curve is given in Fig. 199. The table on page 365 gives the approximate temperatures for handling general tool steels. Two columns are given under harden- ing temperatures as representing the best practice of two well-known 364 STEEL AND ITS HEAT TREATMENT steel companies. As a general proposition, the lowest temperature should be used for hardening which will give the desired results: the use of abnormally high temperatures will increase the grain size, weaken the steel, and reduce the hardness. These last factors become even the more apparent with increase in the carbon content, as is roughly illustrated by the scleroscope readings as given in Shore's chart in Fig. 200. On the other hand, on account of mass action and other individual and distinctive shop conditions, it is difficult to set the upper limit over which hardening should not be done. Certain classes of work often require temperatures which might prove excessive for other work; thus one instance has come to the author's attention in which the hardening of certain 1 -^-in. rounds of 0.9 per cent, carbon stock are hardened at 1600 to 1620 F. and 80 per cent, more service is being obtained than from the same steel hardened at 1460 F. Again, another well-known company hardens 0.9 per cent, carbon steel of approximately the same size at 1370 F. and obtains better service than when hardened at higher temperatures. Each case, in other words, must be handled separately and those temperatures worked out which will give the best solution of that particular problem. Distortion Factors. Slender pieces of steel, when hot, will bend under the application of a steady, even though slight, load. The weight of the part being heated for hardening is often sufficient to cause noticeable distortion if the tool is placed in the furnace in such a manner that it is not carefully supported. For this reason, such tools are best heated when held in a vertical position, with the point of support at the upper end of the piece, the tool being so held that it automatically comes to the normal position as will a plumb- bob. Distortion may be due to the initial condition of the steel, such as may result from forging, rolling, machining, etc. Any strains which exist in the tool previous to heating for hardening are relieved when the piece is heated, but the readjustment of such strains may cause a bending or twisting of the tool. In making the tool it is advisable to rough down to within about ^ in. of the finished size and then anneal in some non-oxidizing material to relieve the machining strains. If the tools are not straight after annealing, they should be heated, straightened while hot (do not straighten in the cold), and then reannealed. The tools are then finished and are ready for hardening. TOOL STEEL AND TOOLS 365 LO >O O O iO O M a> be ' b b o o o o b o o O _ f*> ^ O O ^ O O (**) ^ O O c3 c3 o3 c3 c3 o3 c3 c3 c3 O Pi CO *O *O iO O CO CO 0^ CN fe M T^ Tfi OOO>OOiOiO ^O >O TtHr^COCOCOCOCO CO CO 00 t^* CO cO CO CO cO CO CO COCOCOCOCOCOCO CO CO CO CO CO CO CO CO CO CO CO COCOCOCOCOCOCO CO CO o 88 1> t^ Jt %m ~~Z ': i g 5 : : : < i 1 ^ ^ o S ^ -, j bC 3 "** tiT S' B be c8 c cj ^ o ' CO CO ^ ^^ C^ C^ *-H -H ^H Q r "O'"O'd'"C ;p d'"O 3 jS'd 3 p 6 'a'a3'a3'a3 t a3 p 73cro H cicrx ^ ^ ^ ^ -^ - q * ^ :::::! o ^ - & -s ; ; i S 1 gS| 3 ^ ; || H - f-, O O : o *-> || 1| I|| | | B- di . -fi 6 S w S K efl-P = o^ O O O O O O O t> 00 C5 O rH CM CO O C > O r-i r-I r-i r4 366 STEEL AND ITS HEAT TREATMENT The influence of annealing or previous hardening operations upon the change in shape, such as increase or decrease in diameter and length, or combinations of these, is very difficult to foretell. As a general rule, however, such changes are the more marked with repetitive hardenings, and with increased percentages of carbon. Changes in Length. Most tool steel has a tendency to contract in length upon hardening, and especially after previous annealing or hardenings. This change, whether it is expansion or shrinkage, is dependent upon the chemical composition and uniformity of the steel, the grain size and influence of the mechanical elaboration and annealing, the uniformity and amount of heat used in hardening, the method and rapidity of cooling, and innumerable other variables. The commercial application of heat-treatment principles when ap- plied to fine tools for their standardization to exact measurements must be along the lines of standardizing each step of the process. The steel must initially be kept uniform in composition and physical test; the limits of treatment must be held to a minimum difference: and having somewhat accomplished these aims, the average change of size for specific dimensions must be studied and the tool made accordingly. Thus in the matter of taps, by obtaining the average contraction lengthwise for a given size tap blank, under standard conditions, the thread may be cut upon lathes having the lead screws so adjusted that the pitch given to the tap before hardening will just come right after hardening. Changes in Diameter. Assuming that the many other variables affecting distortion might be reduced to a constant or standard, it is a generally accepted condition that there is a relationship existing between the amount of distortion by swelling after hardening and the original diameter of the piece. Most slender tools, in ordinary commercial practice, have a tendency to expand in diameter after hardening. In factories where a standard steel and standard methods are in vogue the amount of increase is a very important factor, as is shown by the following data obtained from an exhaustive study 1 of taps: Thus this particular problem was attacked with the view of obtaining, under standard shop conditions, the average increase in diameter due to hardening for each size of tap. With these results it was then possible to determine the necessary angle diameter of the tap before hardening, so that the hardened tap would meet the final requirements. i Woodward. TOOL STEEL AND TOOLS 367 Diameter of Tap. Average Increase in Diameter. Diameter of Tap. Average Increase in Diameter. yjj inch 1^ inch 0025 0.00025 0.0025 i 0.0005 0.001 2 0.003 0.003 1 i 0.0015 0.002 3 0.0035 O.Q035 u 0.002 4 0.004 It was also found, however, that experiments upon the same steel, under apparently the same conditions, showed that there may be very great variations in the effect of hardening upon the diameter; and when the various other factors are taken into account, the difficulty of prognosticating exactly what is going to happen is even more apparent. Heating. Much of the difficulty experienced through distortion or cracking may be largely diminished by the proper application of the principles of heating such as have been discussed elsewhere. The heating should be done slowly, carefully and at a uniform rate. In no case should the temperature of the furnace be greater than the maximum temperature to which the steel is to be heated. After the steel has been thoroughly heated, a further continuance will only tend to weaken the steel by increasing the grain size. The furnace should be of such design, construction and operation that it shall be of uniform temperature over its whole hearth, shall heat at a uniform rate, shall not be greatly affected by the introduction of fresh charges, shall have a neutral or reducing atmosphere, and shall be under exact control. Further, it is not only the temperature to which the steel has been heated in the furnace that counts, but also the temperature (and uniformity of temperature) of the steel when it goes into the quenching bath, HEAT TREATMENT OF TOOLS In the following pages there is given a description of practical methods of treating certain tools. It is not intended that the data shall be comprehensive of all methods or of all tools, but shall give practical hints and methods which may aid the man in the small shop who may have such work to do but occasionally. It should also be remembered that the ideas given are representative of a 368 STEEL AND ITS HEAT TREATMENT type of treatment or tool, and which may find application for many uses not given. The methods given are taken from practical work, and have given satisfaction. Chisels. Chisels belong to that class of tools which can be advantageously hardened without first grinding away the " skin." Chisels are not hardened by heating up the whole tool, but only applying the heat to 2 ins. or so of the cutting end; too short a distance is detrimental, as the long, unhardened shank may bend under heavy blows. In quenching, care should be used to avoid a distinct line of demarkation between the hardened nd unhardened parts, as otherwise the end may break off in service. After harden- ing the cutting end by immersing in water and at the same time mov- ing the chisel up and down, the tool should be removed while suffi- cient heat for tempering still remains in the shank. The cutting edge is brightened with emery cloth. Temper colors will soon appear in the brightened spot, which are caused by the heat from the shank " running down " towards the cutting edge; the temper colors may be " spread out " by holding the end near the fire if it is desired to have a broad band near the edge. When the proper temper color is obtained at and near the edge, the chisel should be immediately plunged into water until cold in order to prevent further softening. The degree of tempering depends upon the steel which has been used, the duties in service and upon the general experience and judg- ment of the hardener. In general it may be said that chisels for metal should be tempered to about a peacock color; for stone to a purple color, although many stone and granite chisels are drawn to only a light straw color; and for soft material to a blue color. If a very tough steel is desired, the chisels may be given a double tempering by heating to the temper color twice, the first color being rubbed off after quenching. The lower the carbon used the lower should be the tempering; in fact, a 0.45 per cent, carbon steel is sometimes used after hardening without subsequent tempering. An excellent plan in treating chisels is to oil treat the blank before forging out, quenching the steel in oil and from a temperature sufficient to harden it well. This will stiffen the shank and keep the head from upsetting during use. The problem of producing a good chipping chisel is simple, and yet has met with more difficulties than the majority of other such tools. The correct adjustment of temperatures plays a most important part in this work. For forging, the steel should be heated TOOL STEEL AND TOOLS 369 up gradually and not much beyond a bright-red heat. 1 The work should be done rapidly, aiming to obtain the greatest reduction to size while the steel is yet near its first heat; as the temperature falls, the blows should be lighter and quicker. It has been the author's experience that added toughness is given to the steel if these light, quick blows are carried on until the steel neare a dull-red heat. If reheating is necessary, adopt the same care in raising the temper- ature as before. The chisel should be immediately hardened and tempered after forging when possible. If the pressure of work will not permit of this, the chisel should be allowed to cool as slowly as possible by sticking the hot end in dry dirt; this latter will tend to eliminate the cooling strains which might otherwise be set up. For hardening heat slowly to the lowest temperature at which the steel will harden (generally about 1350 to 1400 F.), allow the heat to penetrate and harden as usual. If care is used, and sufficient thought has been given to the design of the chisel, a hundred per cent, improvement may be easily obtained over the ordinary " hit or miss " method. An open-hearth steel which has given unusual service for pneu- matic chipping chisels is as follows: Carbon, 0.90 to 1.00 per cent.; manganese .50; phosphorus and sulphur low; chrome, 0.50 per cent. The hardening temperature of this steel is about 1400 F. Die Blocks. The problem of satisfactorily treating die blocks is one which will try the hardener's knowledge and patience, increas- ing in difficulty with the intricacy of design and size. Primarily the requirements of any die are (1) a hard face, (2) sufficient depth of hardening to prevent the impression from sinking, and (3) a tough back or body to take up the shock or blow. Further, the hardening must be conducted in such a manner as will prevent any change of size such as warping besides producing a clean, sharp impression free from scale, pitting or checks. These fundamental requirements, assuming that the right kind of steel is used, involve the factors of proper heating and the most suitable method of cooling. 1 For the convenience of the tool hardener who has to work without a pyrom- eter and must gauge temperatures by the eye, the following table of approxi mate heat colors in moderate diffused daylight is given : White 2200 F. Cherry or full red 1375 F. Light; yellow 1975 Medium cherry 1250 Lemon 1825 Dark cherry 1175 Orange 1725 Blood red 1050 Salmon 1650 Faint red 900 Bright red 1550 370 STEEL AND ITS HEAT TREATMENT More die blocks are warped or cracked through improper heat- ing than through any other cause. It is absolutely essential that the entire mass of the steel of the block shall be heated carefully, uniformly, through and through, to the proper temperature. It is always best to allow the die to heat up with the furnace so that any strains which may exist in the steel may be removed grad- ually and give the mass of steel ample time to adjust itself to the rise in temperature. If the furnace for heating for hardening is already at or near the quenching temperature when the die is ready for heating, it will be advisable carefully to preheat the die in a separate furnace. Equipment is subsequently described for automatically preheating and full-heating in the same furnace. In no case should the temperature of the heating furnace be greater than the maximum temperature from which the steel is to be quenched. The practice unfortunately common of forcing the furnace in order more quickly to heat the steel is to be strongly condemned ; such practice must inevitably result in the overheating of the edges or corners of the die and produce unsatisfactory results. The extra time spent in careful and uniform preheating, and in all subsequent heat- ing operations, will be well worth the expense. The exact temperature best suited for hardening may be men- tioned here only in a general way. The chemical composition of the steel, the size of the die block, the depth of hardness required, the condition of the steel before hardening, and many other factors must be taken into consideration. In general, nickel and chrome nickel steels may be quenched at lower temperatures than those used for the corresponding carbon steels, while vanadium and chrome vanadium steels usually require higher temperatures. Again, the greater the mass of the steel, and the greater the depth of hardness required, the higher is the temperature for quenching. Some die hardeners even find that a temperature which will produce a slightly coarse grain is advisable for certain classes of work, such as dies for cold forming under a heavy drop. In other words, the proper temperature for hardening must be determined by experiment, but the lowest temperature which will produce the desired results is always the best. The next, and a vastly important factor, is the duration of heat- ing at the predetermined temperature for hardening. The point to be emphasized is that the mass of the steel must be uniformly heated throughout and this takes time. Not only must the outer sections attain the necessary degree of heat, but also the very TOOL STEEL AND TOOLS 371 center of the die. Disregard of this fundamental principle is the basis of a large proportion of the failures through cracking or warp- ing, and which are so often attributed to " the steel is no good." If the die block lays directly on the hearth of the furnace (and this, it might be mentioned, is not the best practice, since any work is best heated when the hot gas currents can circulate entirely around it), the penetration of the heat may be roughly determined by moving the block to one side and noting the color of the space va- cated; if this area is not of the same color as the rest of the furnace floor the heat has not thoroughly penetrated to the center of the die. This test, called the " drawrouagh " by old English hardeners, should be followed even though the play of heat colors over the ex- posed portions of the block appear uniform. For protection of the impression from oxidation through contact with the air, the faces of the dies may be packed in carbonaceous material. One of the large manufacturers of silverware packs his dies as follows: A small sheet-iron pan, about 2 ins. high and about an inch or so wider than the die all around, is partly filled with granulated animal charcoal or bone. The die is then pressed firmly upon the charcoal, forming an impression, and is then care- fully removed. This impression is sprinkled with powdered animal charcoal, and with very fine steel filings. The die is carefully replaced in position, surrounded with more granulated animal char- coal to the height of the pan, and the space between the top of the pan and the die carefully luted with fire-clay. Upon heating, the filings and powdered charcoal fuse together upon the surface of the steel, forming a protective coating which eliminates oxidation during heating, but which is washed off during the quenching. A little brushing with oil and emery powder will immediately produce a clean, bright surface. Another method is first to paint the surface of the die with a thick paste made of linseed or cottonseed oil and powdered bone-black; the die is then placed in a shallow pan upon a half-and-half mixture of fresh bone and powdered charcoal, in a suitable pan or box, which is then filled and luted as above. The old-time method of an immediate and total quenching of the block until it is quite cold should be attempted only with the simplest forms and small sizes of dies. Large blocks have a great tendency to warp, bulge, or even crack if a total immersion is adopted, this being caused by the unequal contraction of the metal of the surface and of the core. It may be said, however, that this difficulty may be largely avoided if the block is previously given a special 372 STEEL AND ITS HEAT TREATMENT treatment consisting of oil quenching from just over the critical range, followed by an annealing at a temperature just under the lower critical range. Total quenching also should not be used if an extremely hard face is desired, since the heat cannot usually be removed quickly enough. The best practice for hardening large die blocks consists in first carefully preheating the die, then slowly raising it to the hardening temperature and allowing it to soak at this temperature until it is thoroughly heated. For this work a furnace should be used in which accurate temperatures and uniform heating can be obtained. Large blocks may be most easily handled by the use of a hoist on a swinging run-way, mono-rail or overhead crane, and equipped with tongs or "dogs." The dogs fit into holes which have previously been drilled in opposite sides of the block about half way between the upper and lower faces. When the block is properly heated, it is removed to the front of the furnace, gripped with the dogs, run over to a position above the quenching tank, lowered face-downwards entirely into the water or oil for a few seconds (to prevent warping), and then raised out of the quenching bath until immersed about 1 in. deeper than the depth of the deepest impression in the die. The surface of the bath should be kept in motion, or else the block should be slowly raised and lowered a little so that there will be no one line of hardening. The hardening will be greatly increased if a stream or heavy spray of water (assuming water to be used for hardening) is directed against the face of the block, or into the impression. In the case of blocks which contain a deep impression, such as are used for certain classes of gears, etc., it will be necessary to have a stream of water thus impinge upon the impression in order to harden it; the face of the block will take on great hardness, and the heat from the unsubmerged part will gradually be drawn out. When the face of the block is entirely cold, and the majority of the heat taken out of the other portion of the block (usually at about a very dark red, but dependent upon the size of the block), it is raised out of the water, reversed to face up, and brightened with emery paper. The heat in the hot part of the block will gradually temper the hardened face. When this approaches a good straw color the block is immersed in water or oil until cold; in some instances where a softer block is desired, the block may be allowed to cool in the open without the use of water to stop the temper. In case there is not sufficient heat left in the block after hardening to bring out TOOL STEEL AND TOOLS 373 the desired temper color, the block may be stood in front of the fur- nace, back to the heat, or placed on a hot bar of steel, or laid on top of a low smith fire, until the proper temper is reached. Die blocks will generally give more uniform service when drawn in a tempering bath when such is possible. In case the block is of intricate design and requires very particular tempering in weak spots, this may be done by the local application of heat by means of hot plates, etc. FIG. 201. Intake End of Special Furnace for Hardening Forging Dies. (" Machinery.") Die blocks hardened and tempered as directed above should pro- duce a strong, tough base and core, increasing in hardness as the face is approached. Semi-automatic furnaces for heating die blocks for hardening, in which the preheating as well as the final heating are done in the same furnace, are illustrated in Figs. 201 and 202. Four runways, filled with 3-in. malleable-iron balls, extend throughout the length of the heating floor of the furnace. Castings which fit over the 374 STEEL AND ITS HEAT TREATMENT balls in two of these run-ways and which are of suitable size to carry one of the dies, are placed in position in the end of the furnace shown in Fig. 201. When the cold die is placed on this casting, as shown at 0, one of the pneumatic pushers P is brought into play and the cast- ings act as a cart, carrying the die into the furnace. By following the first casting and its die with others, the furnace is gradually filled, the ram pushing the whole line of dies further into the furnace with each new addition. The furnace is so designed and operated that the temperature at the charging end is low, but gradually FIG. 202. Quenching and Tempering Dies. (" Machinery.") increases up to the maximum near the other end of the furnace; preheating and the final heating are thus obtained in the same furnace. Each furnace is double tracked and heats two rows of dies at once. At the end of the furnace shown in Fig. 202 the hot dies come out on the extension of the run-way marked Q. The faces of the dies are turned downwards so that the dies may be picked up by the traveling crane and lowered into the quenching tank, as shown at R. A stream of water also plays against the impression, as usual. The hot plate shown at T is used for the tempering. TOOL STEEL AND TOOLS 375 Engraved dies for spoons, forks, knives, etc., are treated at one plant by the following method. After packing and heating as described in a previous section, the dies are quenched face up in water at a temperature of about 70 to 80 F., to a depth of within about \ in. of the face. Water at this temperature seems to give the best results in this particular instance colder water is too harsh, while warmer water does not sufficiently distribute the strains nor give sufficient hardness. As soon as the cooling effect just begins to creep towards the face of the die, and which only takes a few FIG. 203. Method of Hardening Engraved Die. seconds; the die is immediately wholly immersed in a vertical posi- tion in the water, with the impression turned toward a heavy stream of water which impinges directly upon it. The arrangement of the quenching bath is shown in Fig. 203: the die (a) rests upon a wire platform (6); the water is supplied under pressure through a IJ-in. pipe (c), flowing out through a J-in. slot (d) which extends from the level of the die support to the top of the pipe. The die remains in the water bath until the " singing " has stopped, about 50 to 90 seconds, and is then cooled in oil until cold. The hardened die is later ^tempered in oil to 435 F. 376 STEEL AND ITS HEAT TREATMENT Many alloy steels have been experimented with in recent years for the purpose of increasing the production of forgings from a given impression, thus avoiding the loss of time and expense incurred in redressing the die-blocks. A chrome nickel steel containing about 0.50 to 0.60 per cent, carbon, 0.50 per cent, chrome and 1.50 per cent, nickel has been found to give most economical results. These die blocks are hardened and tempered in the usual way, using a temperature of 1400 for the hardening heat. If the carbon content runs above 0.60 or 0.65 per cent, it has been the author's experience that cracking during or directly after hardening may result. These blocks are greatly improved, not only in the length of service to be obtained, but also in the elimination of warpage during hardening, and of danger of cracking, by giving the block, before machining, a full heat treatment and toughening or annealing ; blocks which approximate the composition noted should be quenched in oil from a temperature of 1400 to 1450 F., and then full annealed at about 1250 F. Such a treatment gives excellent results, and will also show up any defects such as pipes, seams, etc., before the expen- sive machine work has been done. Dies used in engraving work, and in the jewelry and optical trades, must have a glass finish, both in smoothness and in hardness. If subjected to the usual quenching, followed by sand-blast, acid bath or cyanide, a large amount of stoning and polishing would be required. This may be obviated by the use of borax or boracic acid in the following manner. Fill the matrix with powdered boracic acid and place near a fire until it melts, which temperature is con- siderably below the tempering point or color of the steel. Follow this with a second addition of boracic acid and then harden as usual. Although the salt will generally come off in the quenching, it pro- tects the polished surface of the die and does not interfere with the hardening. In case the salt does not come off in quenching, it may be easily removed by live steam or boiling water. The hardening may be done by complete or partial submersion, depending upon the thickness and general design of the die. Engraving dies are usually tempered to a light straw color. Drills. For occasional work in hardening drills, the following procedure may be used : If an open fire is the only available source of heat for hardening, the points of the drill should be kept out of the hottest part of the fire at first, drawing them in as the upper parts become heated. The heat should extend over a considerable portion of the drill. Quench vertically in water, and keep the drill TOOL STEEL AND TOOLS 377 moving up and down so that there is no abrupt line of demarkation of the hardening. If the drill is held quietly in the water, fracture across the water line is a common occurrence when the drill is placed in service. Allow the drill to remain in the water until the im- mersed part is entirely cold. Remove, brighten, and allow the heat in the shank to run into the hardened part until a dark straw color appears on the cutting edge. The drill should then be immediately and entirely immersed in water. If there is not sufficient heat in the shank to bring out the temper color, use hot ashes, or similar means. The drawing operation upon hardened drills should pref- erably be carried out in an oil or salt bath subsequent to straight- ening; drawing expensive tools to color is poor practice. It is always advisable, however, if an open fire must be used for heating, as noted above, to heat the drill in a pipe or tube to prevent the direct contact of the fire and the steel, or with charcoal to prevent oxidation. The heating should be done slowly, uniformly, and to as low a temperature as is possible and consistent with the desired results. In cases where a large number of drills are to be hardened, it is advisable to use a special hardening tank. The shape of the lands of the drill is such that the steam formed by the contact of the water and the hot metal will in many instances prevent the water from penetrating to the flutes and properly hardening them, besides having a similar influence on the end of the drill, which will become the new cutting edge as the point is ground back. This buffer or blanket of steam may be eliminated by maintaining a constant flow of cold water into the grooves and against the end of the drills. Perforated pipes may be placed up the sides of the quenching tank, and through which the cold water is forced into the grooves; similarly, a jet from the bottom strikes against the end of the drill. For drills for holes under J in. in diameter, the hardening heat should be allowed to penetrate only through the cutting part. The drill should then be quenched entirely and the temper drawn to suit the work. The reason for not allowing the hardening heat thor- oughly to penetrate to the core of the drill is that sudden quenching of a small, slender piece might cause severe strains to be set up in the steel; such drills also require a tough core to be able to withstand the torsional effect in the actual drilling operation. Most of the small drills are quenched in oil. The temper color is usually a dark straw. If the tempering is accomplished by placing the drills upon a heated 378 STEEL AND ITS HEAT TREATMENT bar, the cutting parts must be allowed to project for some distance over the edge of the hot bar, for otherwise the heat will be too sud- denly applied. Milling Cutters. Under this class are included cutters of varying description, such as milling cutters, forming cutters, slotting cut- ters, angle cutters, etc. This consists, in general, of a cylindrical piece of steel with a bore through the center, and teeth on the cir- cumference, sides, or both. The unequal forces of contraction and expansion affect these tools to a large extent. In designing a cutter, as large a mandrel hole as is possible should be used, as larger holes will permit the steel to be hardened more uniformly. If the mandrel holes are standardized, large cutters may have a part of the sides (in the absence of side or angular teeth) dished or paneled out at the place which would tend to garden last, that is, half way between the two circumferences. Great care should be used in heating milling cutters for harden- ing. The heating atmosphere should be neutral or slightly reducing to protect the teeth. If an open fire is used, the fuel should not be allowed to come in contact with the cutter: this may be done by resting the cutter on a fire-brick or plate. If a hearth furnace is used, the cutter should not touch the floor or walls of the furnace, but should be supported by fire-bricks or other suitable methods. If tongs are used in handling, care must be used so that the tongs do not touch the cutting edges; the use of wires is better practice. If the cutter is supported on bricks, or laid on plates, it must be turned repeatedly in heating so as not to leave any unevenly heated spots. The cutter may be conveniently held in the quenching bath by using a small round bar which has three prongs welded to one end, and which extend at right angles to the axis of the bar, by slipping the other end of the bar through the mandrel hole of the cutter ; the latter will rest on the prongs, and then can be conveniently lowered into the quenching bath. Ordinary cutters are best hardened by the use of two small cir- cular plates of a diameter slightly greater than that of the cutter, and with holes bored through the center corresponding to the size of the mandrel hole of the cutter. One plate is placed on each end of the cutter, and the whole placed on the suspension tool as de- scribed above and immersed vertically in the quenching bath. By the use of these plates, the hardening will affect the steel along the entire length of the teeth and at right angles to the center line of the cutter. This will also eliminate the circular fracture or TOOL STEEL AND TOOLS 379 flaking of the teeth which so often characterizes milling cutters subjected to uneven cooling. While in the quenching bath, the cut- ter should be moved up and down and not from side to side ; this will permit the solution to pass through the center hole and give an evenly hardened core. The combined use of water and oil (" broken hard- ening ") in the following manner is good for hardening for large cutters: quench in water until the " singing " caused by the water boiling on the hot steel has stopped, and then immerse in oil until cold; warm the cutter in boiling water to relieve the strains and temper when convenient. Pack-hardening is also used to some extent for milling cutters in order to prevent oxidation; in this case each piece should be quenched separately. Salt baths and lead baths are also used for heating. One of the main points to be observed in quenching milling cutters is that long cutters should be plunged vertically and thin ones edgewise. The tempering of milling cutters is often done by the insertion of a hot rod through the mandrel hole and revolving the cutter on it until the proper temper color is obtained. The most satisfactory results are to be obtained with the use of an oil bath, as an even hard- ness can be best obtained in this manner. Small cutters are tem- pered to a light straw color, or yellowish-white. For medium-sized cutters a good straw color may be used. Very large cutters, on account of the lesser effect of the hardening, may not require temper- ing, but it is always advisable to heat them in boiling water to make them uniform and remove the hardening strains. For hollow mills it is not necessary to heat for hardening very much above the teeth, as it is not required that the back should be hard. Harden with the teeth upwards, working the piece up and down in the quenching bath to get the solution circulating through the hole. T-slot milling cutters should be hardened, not only through the cutting portion, but also through the entire length of the neck, especially if this is of small diameter. In tempering, the cutting portion should be drawn to a straw color and the neck to a blue color. Files. Before the file blanks can be ground and the teeth cut it is necessary to anneal the steel. This is often accomplished by packing the blanks in air-tight oblong boxes and annealing at about 1300 to 1400 F. Lead baths continue to be most used as the heating medium. Salt baths have been tried with varying degrees of success, but in the main have proven unsatisfactory. This is due in a large measure 380 STEEL AND ITS HEAT TREATMENT to the fact that oxide of iron (scale) may settle upon the teeth of the file, causing soft spots when hardened. The method of dipping the file into a solution of ferrocyanide and allowing the coating to dry upon the surface of the steel before heating has been tried. The objections to the use of this method are that a decomposition of the ferrocyanide will yield additional iron oxide and poisonous fumes. Other salts of a harmless character have been tried with little success. The general procedure is to cover the file with a paste which pro- tects the edges of the teeth in the hardening process, heating in lead to the proper temperature (about 1400 F.) and quenching in water in a vertical position. One file-maker uses a paste made of the following base: ground charred leather, 2 parts; table salt, 4 parts; and flour, 3 parts. The file is given a coat of this paste, which is allowed to dry before heating. It is said that the melting-point of this paste will give the proper hardening temperature. After being hardened, and while the file is still warm, it is put through the final straightening process. Half-round files require particular attention on account of their tendency to warp : before hardening, the file is bent back on a fixed template of such form as experience has shown will bring the file to a true line upon hardening; the file is placed again in the template before it is quite hard, strained to the proper degree, and water is thrown on the upper surface of the file to make it quite cold before the strain is relieved; the file is then entirely quenched and will usually return " to the true " after the final hardening. After the final straightening the files are " scrubbed " to remove the paste, and are then washed in lime water and dried by holding them in steam. The tang is then toughened or " blued " by dipping it into a special bath maintained at the proper temperature. File steel will vary in carbon from 0.90 to 1.60 per cent., accord- ing to the size, shape and use of the file; manganese under 0.40 per cent.; low phosphorus and sulphur; and in the case of exceptionally good files, a small percentage of chrome. Nickel is generally con- sidered as detrimental to files. Reichhelm 1 shows the detrimental effect of heat variations in hardening in the microscopic photographs of two fractures of the same file magnified 160 times. This file is one of the highest grade produced in Europe, and Fig. 204 shows the fracture of this file as imported, while Fig. 205 shows a fracture of the same file, a section of which was rehardened, after the exact degree of heat required 1 " Machinery/' Dec., 1914. TOOL STEEL AND TOOLS 381 FIG. 204. Photomicrograph of High-grade Foreign File. (Reichhelm.) X160. FIG. 205. Photomicrograph of Same File Rehardened. (Reichhelm.) X160. 382 STEEL AND ITS HEAT TREATMENT for this particular steel had been experimentally determined. Fig. 204 therefore shows the result of the best hardening practice in Europe, aided by the pyrometer, while Fig. 205 shows the hardening of this identical file by the correct heat automatically maintained. That any number of files, or tools of any kind, can be hardened so as to show uniformly the excellent fracture exhibited in Fig. 204 is due to automatic heat control, as has been demonstrated conclusively in daily practice for over three years. Both of the photographs of fractures have been pronounced excellent by competent judges, but the decidedly finer grain and more even diffusion of the carbon shown in Fig. 205 produced a difference in the durability of the file teeth of nearly 50 per cent., as compared with the section of the file as originally hardened and shown in Fig. 204. Punches and Dies. Similar to all round tools, punches show a great tendency to flake off at the corners, sometimes a whole ring breaking off. Assuming proper heating, this may be overcome to a large extent by means of a water spray. Dies of intricate shape and possessing sharp angles should be most carefully handled. It is often advisable to fill these angles with a little putty or fire-clay to lessen the hardening effect and prevent the formation of quench- ing strains at right angles to the diagonal. A piece of binding wire may also serve for this purpose. Dies should generally be quenched flat, depending upon the shape of the piece. Small punches should not be quenched in real cold water on account of the liability to cracking under sudden cooling an oil bath or lukewarm water is far preferable. Dies or ,any press tools having holes near the edge should always have these holes filled with clay in order to pre- vent cracking or too great hardening; graphite or asbestos may also be used for plugging the holes for stripper or guide screws. Punches and dies are generally tempered to about a straw color, the depth of this varying according to the thickness and hardness of the material to be punched. The tempering may be carried out by setting the hardened pieces in front of a hot furnace, laying on hot plates, in oil baths or in hot sand. Reamers. Reamers may be heated in lead to protect the cutting edge from the direct action of the heat and oxygen. The lead may be prevented from sticking to the tool if the latter is brushed, in the case of small reamers, with a little soft soap. Larger reamers may be protected with a paste made of black lead and water or lampblack and linseed oil, both of which should be allowed to dry TOOL STEEL AND TOOLS 383 on the tool before heating for hardening. If the reamer has been hardened by the use of water alone, and is larger than f in. in diameter, it is advisable to hold it over the fire directly after being removed from the hardening bath, or to set it in hot water for a few moments, in order to remove as far as possible the strains which have been caused by the hardening process. This should always be done in the case of shell reamers and other special reamers of any considerable size, whether the quenching medium has been oil or water. Broken hardening is most excellent for tools of this description. Large fluted reamers require to have only the ribs heated to the proper temperature, and then quenched; temper- ing will not then be required. Ordinary fluted reamers are tempered to a yellowish white or very light straw color. Six-sided or eight- sided reamers may be tempered to a light straw color. Square reamers, triangular reamers and half-round reamers may be tem- pered to a dark straw color, due to the fact that they take hold of the work more deeply and might break if not tempered a trifle softer. Half-round reamers should not be quenched vertically, but with the half-round side at an angle of 20 to 45 degrees to the surface of the bath. If half-round reamers should be quenched vertically, it will be necessary to move them in a horizontal manner in the direction of the half-round side at the same time as immersed ver- tically. The shanks of reamers, taps, drills, broaches and similar tools may be toughened by local lead tempering. Rings. Rings, collars arid hollow tools comprise a class which require great hardness in the inner circumference or bore. Quench- ing is usually done by means of allowing a full stream of water to flow through the bore if it is quite small, or in the case of tools with larger bores the insertion of a small pipe with a series of holes in its circumference and through which a continuous stream of water may be forced, forming a spray. In the first case it is advisable to set the tool upon an asbestos-covered washer in which has been cut a hole slightly larger than the size of the bore of the tool and then apply the flange end of the water-supply line or pipe to the other opening. Rings or collars requiring resistance to frictional wear require no tempering. Eccentric rings cannot be quenched as usual, as the relative thickness and thinness of the opposite sides would tend to give unequal expansion and contraction and cause the hole to become oval-shaped. This may be overcome by binding a small 384 STEEL AND ITS HEAT TREATMENT piece of iron or steel to the thin side, heating, and quenching ver- tically. Rivet Sets. Rivet sets should never be quenched directly by immersion, as this will tend to make the edges of the cup break off, the center to remain soft, and leave a line of great weakness between FIG. 206. Rough Method of Hardening a Rivet Set. the hardened and unhardened parts. A simple and proper method is to hold the cup under or over a stream of water so that the latter will impinge directly upon the bottom of the cup, as shown in Fig. 206. If there are numbers of rivet sets to be hardened, an arrange- ment of clips or holders under each tap or spigot may easily be set up. The tempering may be carried out as in the case of chisels (permitting the heat in the shank to temper the cup) or the shank TOOL STEEL AND TOOLS 385 may be placed in a lead bath and the color allowed to run up into the cup; the rivet set should then be entirely quenched to prevent further softening. Brearley makes the following points, which are of great interest. Rivet sets may have a short life due to the wear on the head, which is as often a failure as that produced by actual fracture. This is pronounced in the case of annealed stock. He advises hardening the head in oil before hardening the cup. Upon reheating for hard- ening the cup, and tempering, a steel of great toughness is obtained, which neither splits nor forms a mushroom head. Saws. Saws may be hardened by either of two methods direct immersion, or press or roll hardening. Circular saws may be heated by enclosing in a sheet-iron case or box between layers of charcoal. Sufficient space for expansion must be allowed to eliminate chance for buckling. Saws may also be heated on the hearth (if level) of any type of hardening furnace; it is advisable, however, to rest the saw on an iron or steel plate so that the heating may be gradual and uniform. The greater part of the secret for the successful hard- ening of saws without buckling is a slow and careful heating. The saws when heated to the proper temperature may be taken out separately with tongs or a J-shaped hook. For direct quenching they should be immersed edgewise and in a perfectly vertical position. It is better to have a thin layer of oil on the surface of the water bath, as the oil will ignite when the hot saw enters it, forming a thin, protective coating on the saw 7 and thus lessening the risk of fracture. Oil alone, or oil with tallow dissolved in it will give suffi- cient hardness for thin saws. The saws may also be placed between lumps of tallow. The latter (tallow) is a better hardener than oil, and therefore gives a greater and deeper hardening. Thin cir- cular saws, and all ordinary saws such as hack saws, hand saws, etc., may be most satisfactorily hardened by means of a press. A com- mon and inexpensive method is to have two cast-iron plates hinged together, with the inner surfaces well oiled with a heavy oil. The hot saw is placed between the plates, which are then clamped to- gether and held until the saw is cold. Thin band saws are often hardened by means of rolls. Circular saws for metal cutting should be tempered to a dark purple color, or to a light blue for wood cutting. Hack saws require tempering to a light purple color, CHAPTER XVII MISCELLANEOUS TREATMENTS THE following examples and discussions of certain heat-treat- ment methods have been selected in an arbitrary manner as repre- sentative of distinct classes of work. Many others might just as well have been taken, but the author feels that those selected will perhaps illustrate in a general way some of the many problems which arise in the course of ordinary heat-treatment work. GEARS Gear-steel Classification. Automobile and similar machine gears may be broadly grouped according to the method of heat treatment, which, of course, is dependent upon the composition of the steel. Thus the three classes are: (1) The case-hardened gear, using a steel of low-carbon content generally less than 0.25 per cent and depending upon the case- carburizing process to give an outer layer of high-carbon steel and upon the subsequent hardening processes to produce the necessary wearing surface of sufficient hardness. (2) The oil-hardened and tempered gear, using a steel of the alloy type of about 0.45 to 0.55 per cent, carbon. (3) The hardened gear (without subsequent tempering), using a steel of an intermediary carbon content about 0.30 per cent. Requirements of Gears. All high-duty gears require that the steel shall be readily forgeable and machineable, and that after treatment it shall have the greatest possible hardness with the least possible brittleness. In this connection it may be said that surface hardness is often more desirable than tensile strength, while the question of brittleness is very important on account of shocks. Case-hardened vs. Oil-tempered Gears. The merits or demerits of each type depend largely upon the point of view and the personal experience of the user. Expert opinion may differ widely, 386 MISCELLANEOUS TREATMENTS 387 as is shown by the following excerpts from addresses by two well- known metallurgists. One says: 1 " Several years of observation and contact with the trade leads me to prefer the case-hardened gear. The result of direct tests upon thousands of gears of both types leads me to the following con- clusions: (1) The static strength of a case-hardened gear is equal to that of an oil-hardened gear, assuming in both cases that steel of the same class and approximate analysis has been used and that the respective heat treatments have been equally well and properly conducted. (2) Direct experiments proved that the case-hard- ened gear resists shock better than the oil tempered. (3) As regards resistance to wear the same type is incomparably better, although perhaps not as silent in action. " One of the leading makers of gears has proved this to his own satisfaction of late by an arrangement of shafts and gears whereby energy is transmitted through two case-hardened gears, in mesh with each other, to two oil-hardened gears. The gears are of the same size. The conditions of the test were severe. Five sets of the oil-hardened gea'rs have already been worn out, while the original case-hardened gears are still in service and show the tools marks. " Upon the part of many there is a strong objection to case hard- ening. In nine cases out of ten this is doubtless due to the fact that the case-hardening operation has not been reduced to a science. The depth of case, the relation of case to core, the time and tem- perature to produce certain results and the exact control of these conditions, together with an accurate knowledge of the material to be treated, are factors that enter into successful case-hardening practice. Further points in favor of this method are easier machin- ing of the blanks, and at least equal static and dynamic properties with kss chance of injury in hardening." Then here is the opposing argument: 2 " For machine tools, hardened high-carbon alloy steel gears appear to be preferable to case-hardened gears for a number of reasons : " 1. Physically they are stronger and tougher and should there- fore be better able to resist sudden impacts and extraordinary loads. They do not show by file and scleroscope test the same 1 J. A. Matthews, " Alloy Steels for Motor Car Construction," Journ. Frank- lin Inst, May, 1909. 2 From a paper by J. H. Parker, before National Machine Tool Builders' Assoc. 388 STEEL AND ITS HEAT TREATMENT degree of hardness as case-hardened gears, but, nevertheless, with proper design, the dense-grained gear-tooth resists wear more satis- factorily, as was demonstrated recently by the examination of a motor-car transmission that had covered over 100,000 miles. The high-carbon steel gears in this car still showed the original tool marks. Not long ago a designer of machine tools commented on the ap- parent softness of some hardened high-carbon gears, but found after several months of hard service that they still showed tool marks, thus proving hardness ample for wear. "2. In service, especially for clash gears, the superiority of these gears is most marked. On the clashing faces, case-hardened gears are likely to have the hard case chipped off, thereby exposing the soft core to the impact of clashing. The hard chips fall into the gearing and may find their way into bearings, thus causing trouble. High-carbon steel gears with a uniform hardness throughout do not chip, nor do they ' dub over.' " 3. The heat treatment of high-carbon steel gears is much simpler than that required for proper case hardening. It is shorter, less costly and produces a more uniform product, and as the gear is heated but once for hardening, as compared with three times for case hardening, the finished gear is certain to be freer from warpage. The cost of proper case hardening is not generally appreciated, but it has been found that a case-hardening steel must cost three to four cents per pound less than a regular high-carbon hardening steel, if finished gears made from both materials are to cost the same. " With all heat-treated gears, little points in design are impor- tant. The gear-teeth should not be undercut, for if the section at the root-line is smaller than at the pitch-line, greater hardness and brittleness is produced where least desired. Great differences in section should be avoided wherever possible, so as to do away with excessive warpage. Sharp edges and angles, even in key-ways, are the cause of internal hardening strains which frequently result in failures; hence, wherever possible, a fillet should be used in place of a sharp angle." Case-hardened Gears Treatment. The steel for a case-hard- ened gear should be low in carbon, preferably under 0.25 per cent.; should be carburized so as to produce a case of a depth of about ^T or ^ inch and contain a maximum carbon concentration of about 0.9 per cent.; and should then be suitably heat treated. Since the principles of case hardening have been described elsewhere, MISCELLANEOUS TREATMENTS 389 it will be necessary here only to outline the process, which is as follows : (Gear blank). 1. Anneal. 2. Rough machine to approximate size (3. Light re-anneal.) 4. Finishing machine. 5. Carburize at about 1600-1650 F. 6. Cool slowly in carburizing box. 7. Reheat and oil quench from 1550-1625 F. 8. Reheat and oil quench from 1350-1425 F. (9. Temper, if desired, to not over 400 F.) The temperatures given are only approximate, depending upon the analysis of the steel, the mass of the steel, the results desired, and various other factors. Nos. 3 and 9 may be omitted if desired. Oil-hardened Gears Treatment. For the higher-carbon steels used for oil-hardened gears it is always advisable to give the gear blanks a preliminary treatment to develop the highest qualities of the alloy steels and the greatest uniformity in their physical properties. This treatment will also give the greatest " softness " of which the steel is capable. This preliminary treatment (before machining) is: 1. Quench in oil from about 150 to 200 F. over the criti- cal range. 2. Quench in oil from about 50 F. over the critical range. 3. Anneal at a temperature about 75 F. under the criti- cal range. If this preliminary treatment is not given, the gears blanks should be given a thorough annealing. The slight reanneal after rough machining and before the final cut is optional; it always helps, however. The final treatment consists in an oil-hardening and tempering process. For the majority of alloy steels this quenching is done from a temperature about 50 F. over the critical range; in the case of chrome vanadium steels, however, the best results are generally obtained by the use of a higher temperature. The temperatures generally used for the standard types of alloy steels for automobile gears, approximating 0.45 to 0.55 per cent, carbon, are about as follows : 390 STEEL AND ITS HEAT TREATMENT Chrome nickel steels: 1.5 per cent, nickel, 0.5 per cent, chrome, 1400 F. 1.75 per cent, nickel, 1.0 per cent, chrome, 1425 3.0 per cent, nickel, 0.75 per cent, chrome, 1375 3 . 5 per cent, nickel, 1 . 5 per cent, chrome, 1400 Nickel steels: 3.5 per cent, nickel 1400 F. 5.0 per cent, nickel 1375 Chrome vanadium steel: Type " D " (1 . per cent, chrome, 0.8 per cent. manganese, 0.16 per cent, vanadium) 1575 F. Silico-manganese steel : 1.5 per cent, silicon, 0.7 per cent, manganese 1550 F. The usual precautions should be observed such as uniform and thorough heating, protection from oxidation, etc. Further, the gear should be quenched in the direction of its axis so that the oil can be made to circulate around the teeth, etc. The notes given under " Milling Cutters " 1 might also be of interest in their bearing upon gear treatment. The tempering is usually done at a temperature of 400 F. or upwards, depending upon the nature of the steel and upon the results desired. It should again be stated that a longer tempering at the lower temperature is preferable to a quicker and shorter tem- pering at a higher temperature. Thus, if a gear were to have the temper drawn quickly, the teeth, which should be the hardest, will be softer than the hub, which will remain brittle; with a longer heating at a lower temperature this will not be the case, since the whole gear will have responded throughout. Similarly, for these reasons, it is inadvisable to temper gears " by color," but to use an oil bath or a mixture of low melting-point salts. For gears made of alloy steel with only about 0.30 per cent, carbon the tempering operation is usually omitted. It is always best, however, to reheat the oil-quenched gears in boiling water for a short time in order to remove the hardening strains; such treat- ment will have little or no influence on the hardness and strength. The quenching temperature for such steels will of course be higher by some 50 or 75 than that given under the 0.45-0.55 per cent, carbon steels. iCf.Ch.XVI. MISCELLANEOUS TREATMENTS 391 SPRINGS The usual analysis for carbon steel springs is approximately: Carbon 0.90 to 1 . 10 per cent. Manganese under 0.40 Phosphorus under 0. 04 Sulphur under 0. 04 Silicon up to 0.25 It is dangerous to allow the percentage of carbon to run up to 1.25 per cent, (as is sometimes done), on account of the possibility of the formation of free cementite, which is an extremely brittle constituent. A crack might easily start in an area of cementite and when once started would follow through the cementite to the outer surface. Lower carbons would preclude the presence of free cement- ite. Finely divided cementite would also be less dangerous, and this could be obtained by hardening at a lower temperature (about 1400 F.), since crystallized and granular cementite can only be obtained by heating for a prolonged time at a high temper- ature. Aside from improper analysis, the majority of spring failures and troubles may be laid to abnormally high temperatures for heating for fitting followed directly by quenching from whatever temper- ature the steel may happen to be at; and then, as if this were not bad enough, to temper by " flashing." From general knowledge it appears that the maker of springs has not kept pace with improve- ments in spring steel and with the increased severity of the duty expected of springs. The old practice of high temperatures and of forming and hardening springs with a single heating cannot be persisted in if maximum quality and service are to be secured. The " practical " spring-fitter generally heats the steel to about as high a temperature as it will take without burning. Its effect upon the structure of of steel has been explained in preceding chapters, and also above in its relation to very high-carbon spring steel. But even assuming that the proper temperatures have been used in fitting, the time taken to go through the forming operation is sufficient to give the steel a chance to cool down to a temperature which will not give the most satisfactory results in hardening. The steel is not of uniform temperature over its length so that, if it be quenched directly after forming, it will probably lock up internal 392 STEEL AND ITS HEAT TREATMENT strains of uncertain magnitude to say nothing of the insufficient hardening if the temperature be under that of the critical range. In other words, the spring should be put back in the furnace again (it being generally preferable that the maximum temperature for forming shall be the same as that required for hardening) and reheated for a few minutes so that it will be heated uniformly throughout at the right temperature. If high temperatures have been used for form- ing it will be advisable to allow the steel to cool to a temperature under that of the Ar range before reheating for hardening; if this is not done the steel will retain the coarse grain-structure character- istic of the high heat for forming. If it is found that the steel departs from its shape at all during this reheating, it may be put through the rolls again previous to quenching, the time occupied being small compared with that for the original bending. The spring should then be quenched in some good, heavy tempering oil. For drawing the temper it is never advisable to use the process known as " flashing." The practice of replacing the steel, after quenching, in a high temperature furnace until the outside of the steel reaches the desired temperature is one which cannot be too strongly denounced, because of the impossibility of uniform treat- ment. No time is allowed for the heat to soak to the center, with the result that the hardness increases from the outside a most undesirable condition. All spring steel should be drawn back in a suitable low-temperature furnace maintained at the proper temper- ature. The steel should be kept in the furnace for a time sufficient to allow of a uniform heating throughout. Lead baths and salt baths are also used considerably for this work. The proper temperatures for treating carbon spring steel have been given considerable attention by the American Society for Testing Materials. Their experiments were made with test speci- mens If by f by 14 ins. long with straight edges, and analyzing about 1.10 per cent, carbon. The results of these tests (1911) are given in the tables on page 394. It is apparent that at a quenching temperature of 1500 F. the maximum results are obtained with a drawing temperature of about 600 F., while with a quenching temperature of 1650 F. the maxi- mum elastic limit was found with a drawing temperature of about 800 F. In the former group, Series A, 1500-600 F., it was found that the angle of bend at rupture showed an average of slightly over 59, there being considerable variation between the specimens; while in the second group, Series B, 1650-800 F., the MISCELLANEOUS TREATMENTS 393 average angle was slightly over 103, without any specimen going below 76. These results are particularly interesting in view of the fact that the critical range of these steels is about 1350 F., and that one would naturally expect that a temperature of about 1400 F., i.e., slightly over the critical range, would give the best results. Whether or not such would show up in vibratory tests is a question which should be given attention. TRANSVERSE, HARDNESS AND BENDING TESTS OF CARBON SPRING STEEL Series A, Quenched in oil from 1500 F. Hardness. Bend Test, Temper Drawn to Elastic Limit, (transverse) Scleroscope. Angle Bent through at Deg. F. Lbs. per Sq. In. T)f*{"t All Rupture, On Flat. On Edge. -tsrineii. Deg. 425 129,137 48.5 47 370 181 600 136,440 46 50.5 388 60 835 131,017 43.5 49.5 351 86 1025 96,852 34.5 39 268 152 1230 105,400 34 37 282 167 Series B, Quenched in oil from 1650 F. 450 130,922 46 52.5 394 90 625 134,232 43 57 371 82 820 141,147 46 56 389 104 1025 126,320 42 50 371 108 . 1210 83,457 31 36.5 260 180 ALLOY STEEL SPRINGS The service conditions to which automobile springs are sub- jected are extremely severe, for they have to sustain the shocks at speed of the irregularities of the ordinary highway, built for slow- moving, horse-drawn vehicles. The necessity for high elastic limit, combined with great toughness and anti-fatigue qualities, make the use of alloy steel almost mandatory. The alloy steels in use are of the same analysis of those previously given under the heading of " Oil-hardened and tempered Gears " (q.v.). The quenching temperatures are likewise the same as there given, but the drawing temperatures are higher generally from 850 to 1025 F. As far as static strength is concerned, the majority of the now common alloy compositions will give about the same test values, approximately : 394 STEEL AND ITS HEAT TREATMENT Tensile strength, Ibs. per sq. in. ... 190,000 to 250,000 Elastic limit, Ibs. per sq. in 170,000 to 225,000 Elongation, per cent, in 2 ins 15 to 6 Reduction of area, per cent . 45 to 20 Some of the alloy steels, and particularly the chrome vanadium type, require annealing before shearing. The chrome vanadium steels used for springs are readily susceptible to " temper," and it is likely that the rapid air cooling of small flats after they leave the rolls will cause them to be brittle, thus giving a great amount of trouble in shearing. The annealing of this chrome vanadium steel is done by bringing the steel up to a full cherry-red heat in the furnace (about 1475 F.) and allowing it to cool slowly after being maintained at this temperature for a sufficient time to allow of uni- form heating. The new steels cannot be handled just like the old carbon steel springs and still obtain from them the maximum development of their powers. However, the new steels, being in general lower in carbon, will stand much abuse in heat treatment and still pro- duce springs of quality undreamed of a decade ago. While as a class spring-makers have been driven to the use of alloy steels, they have not as a class been forced to handle them scientifically. Alloy steels especially should not be heated any higher for form- ing than is absolutely necessary. Then they should always be reheated to the proper temperature for quenching in order to make sure that the entire steel is uniformly heated throughout to that temperature, which must be exact. The same remarks about tem- pering as given under carbon steel springs likewise apply here, and with added emphasis. OIL-WELL BITS Bits used for drilling oil wells, gas wells, etc., represent that class of large implements requiring " end heats." The hardening of these bits is necessarily an operation to be carried out in the field, since the bits require a more or less frequent dressing and must be rehardened after each heating. An extremely hard end and face, together with a strong, tough core and shank are the principal requirements for this work. About 6 or 8 ins. of the bit is carefully heated in the fire (usually a common blacksmith forge), to the proper temperature usually about 1500 F. Higher temperatures should not be used unless absolutely required by the nature of the steel. Any scale should be MICELLANEOUS TREATMENTS 395 carefully and quickly brushed off before quenching. The bit is then removed from the fire and allowed to rest in a bucket of coarse salt for a second or two. This salt treatment may be omitted, but it undoubtedly gives better results; the direct use of brine is gen- erally too severe for most bit steels. A box or trough should previously be fitted with a wooden grating made of slats, the top of which will be about 3 or 4 ins. under the surface of the water in the box. Some drillers add vitriol to the water quenching bath to obtain a greater hardness. The bit should then be quickly lowered vertically into the cold water until it rests upon the wooden grating, and should be allowed to remain there until cold. The precautions to be observed are: (1) Lower vertically, in order to obtain an equal hardness on both faces of the bit; (2) do not quench to a greater depth than 3 or 4 ins. ; (3) do not move the bit nor splash the heated part of the shank with water; (4) allow the steel to remain in the water until cold, generally over night. Although the surface of the water bath may steam, it will generally be found that directly beneath the surface the water is cold, and likewise the end of the bit. Splashing the heated part of the bit with water has a tendency to draw the temper of the faces. Immersion to a greater depth than 3 or 4 ins. is apt to give a soft bit. If these precautions are carefully observed, and the steel is of the right analysis, a bit with a glass-hard surface and a strong, tough core will be obtained. Such bits require no tempering, and should not chip off. Oil-well bit steel will vary between 0.50 and 0.80 per cent, car- bon and manganese, low phosphorus and sulphur, up to 0.25 per cent, silicon, and the addition of about 0.5 per cent, chrome for the lower carbons. The chrome bit steel, if of the proper carbon-manganese- chrome composition, will undoubtedly give the best service. The following analyses are characteristic of American oil-well bits used and giving good service : Carbon. Manganese. Phosphorus. Sulphur. Silicon. Chrome. 0.73 0.61 0.017 0.030 0.14 0.59 0.17 0.010 0.015 0.13 0.83 0.65 0.010 0.021 0.13 0.60 0.51 0.012 0.019 0.01 0.56 0.54 0.53 0.007 0.021 0.006 0.51 0.49 0.56 0.010 0.016 0.008 0.52 396 STEEL AND ITS HEAT TREATMENT SAFE AND VAULT STEEL Safe and vault steel may be taken as representative of that class of material involving different steels welded together, but for which the proper treatment of one analysis will be sufficient for both. Steel for safes and vaults consists of alternate layers of soft and hard steel, and is known to the trade as " three-ply," " five-ply," etc. By having these alternate layers there is obtained, under suitable treatment, a metal which will have sufficient ductility (due to the soft layers) to resist explosive forces, and at the same time be im- penetrable to drilling, sawing or other machine operations (due to the " hard center"). The soft layers are made of ordinary low- carbon or " soft ." steel, while the hard centers will analyze about 0.85 to 1.05 per cent, carbon and manganese, with or without the addition of chrome. The plate is first machined or ground to size and the necessary holes drilled, threaded, and plugged with fire-clay for protection. The plate is then placed in a suitable heat-treatment furnace, and thoroughly heated to 1400 to 1500 F., depending upon the compo- sition of the hard layer. It is extremely important that ample time be allowed for the heat to penetrate and thoroughly heat the high-carbon steel, for it is upon the hardness of these layers that the full value of the finished plate will depend. The major- ity of the cases in which the necessary hardness was not obtained which the author has investigated have been due to an insufficient length of heating rather than to any fault in the analysis of the steel. The plate is then quickly removed from the furnace by a crane or hoist and quenched in cold water. As the hardness is largely dependent upon the rapidity with which the steel is cooled through the critical range, arrangements should be made to obtain a constant supply of cold water in contact with the steel during the quenching operation. If the quenching is done in a tank, the inlet supply should be large enough always to keep the water cold the warm water being taken away from near the top of the tank. In this case the plate is quenched vertically; particular care should be used in getting the whole plate into the water as quickly as possible, and in an ab- solutely vertical position, if warpage is to be avoided. As soon as the initial immersion is accomplished the plate may be swung to and fro in the tank to aid in the heat removal. Other plants quench by means of water sprays, the plate being supported on a horizontal MISCELLANEOUS TREATMENTS 397 rack; with this method of cooling the water supply should be suffi- cient to remove the steam as soon as it is formed. The plates are not tempered or drawn. Specifications require that the best high-speed steel drill shall not penetrate the hard- center layers. STEEL CASTINGS In the mad rush for alloy steels and their heat treatment but little attention has been given to the treatment of steel castings. And yet there is an opportunity for as great, if not greater, improve- ment in these parts as in forged or rolled sections. All steel castings should be annealed or oil treated, not only to remove the casting strains, but also to get the metal into the best possible condition. Due to the method of fabrication, the rapid cooling of thin sections and the slower cooling of adjacent thicker sections must inevitably produce casting strains of a more or less intense nature. Similarly and coincidently , the structure of the metal must inherently be poor : the grain will be coarse instead of fine and " silky," the metal will tend to have low ductility and brittleness, and the physical proper- ties of the steel as a whole will vary considerably. Unlike forgings and rolled sections, castings are not generally subjected to any reheating and elaboration, so that the metal must have those prop- erties characteristic of moderate cooling from high temperatures. Thus the usual specifications for steel castings, in which the low ductility will be apparent, will call for: Tensile strength, Ibs. per sq. in 85,000 Elastic limit, Ibs. per sq. in 45,000 Elongation, per cent, in 2 ins 12 Reduction of area, per cent 18 Even the now common addition of titanium or vanadium will not serve to eliminate entirely the necessity for subsequent treatment. Annealing, or better still, a full heat treatment, is mandatory. Contrary to the ideas held by many " practical " hardeners, the principles of treating steel castings in no wise differ from those of steel forgings of the same section and analysis. The main difficulty encounterad is that caused by the length of time required for the diffusion of the ferrite and the equalization of the metal as a whole. Castings usually require considerable time for this to take place because of the tendency of the metal to return to its original molecular arrangement and structure during slow cooling. Thus 398 STEEL AND ITS HEAT TREATMENT much of the unsatisfactory annealing is, technically speaking, due to the segregation of the ferrite. It is therefore necessary, in annealing steel castings, to (1) heat well over the upper critical range, (2) for a length of time sufficient to obliterate entirely the previous structure and crystallization, and followed by (3) slow cooling. The proper annealing temperature for the ordinary machinery castings will be between 1500 and 1600 F., depending upon the carbon content. If the annealing is preceded by normalizing, i.e., air cooling from a temperature considerably above the upper critical range say 1800 F. the length of time required for the subsequent anneal will be considerably shortened, besides improving the steel. For castings with the carbon on the lower side of 0.25 or 0.30 per cent., or for castings of considerable size, air cooling from about 1600 F. will usually produce good results. The best method, however, is that of oil quenching and annealing or toughening either with or without a previous normalizing. The castings should be heated as directed under annealing, quenched in the proper manner in oil, and then reheated to the temperature which will give the combination of strength and ductility desired. A drawing temperature of 1250 F. will produce the most ductile steel. STEEL WIRE 1 The principal heat treatments used in the manufacture of wire are: 1, annealing; 2, patenting; 3, hardening and tempering. Annealing serves to accomplish three important functions: 1. To remove the effects of hardening due to cold work in wire drawing or cold rolling, thus making the steel ductile and soft. Annealing for this purpose covers principally the low-carbon wires, those with carbon 0.25 per cent, and under. 2. To refine grain- applied principally to the higher-carbon rods and wires, those with carbon 0.30 per cent, and over. 3. To obtain definite structure in the finished material applied principally to the higher-carbon wires, those with carbon 0.30 per cent, and over. When a steel wire rod of the structure shown in Fig. 207 is sub- jected to the wire-drawing process, a marked change in the grain structure takes place. With each successive draft, the grains stretch out in the direction of drafting until a point is reached when the 1 From a paper by J. F. Tinsley, American Iron and Steel Inst., 1914, and The Iron Age, May 28, 1914. MISCELLANEOUS TREATMENTS 399 grains have been elongated to the limit of their ductility. If sub- jected to further strain by further drafting they will part and the wire will break. Before this brittle condition is reached, therefore, it is necessary to heat treat the wire by subjecting it to what is known in the wire business as a " process annealing." The effect of wire drawing in elongating the structural grain of the steel may be seen by comparing Figs. 207, 208 and 209. Fig. 207 shows the structure of the rod before drawing; Fig. 208 shows the structure after a 15 per cent, reduction from the rod; and Fig. 209, the structure after a 60 per cent, reduction from the rod. All of these micrographs represent sections taken from a plane parallel to the axis of tho rod or wire, not cross-sections. The reason for the marked difference in grain shown in Figs. 207 and 209 may be grasped more clearly when it is appreciated that Fig. 209 represents a wire reduced in the wire-drawing process to such a degree that it has become elongated 2J times the original length of the rod. Process or " works " annealing consists in heating the wire to a certain temperature, maintaining that temperature until the entire mass of steel is thoroughly heated through, and finally cooling down. In the most common of all annealing that to remove the effects of cold work such as drawing it is not necessary to reach the critical temperature, which is 1300 F., or higher, depending on the carbon content. A temperature of 1100 F. is entirely sufficient to relieve the strained condition of the grain shown in Fig. 209. Fig. 210 shows the same wire that is depicted in Fig. 209 after annealing at a temperature below the critical range. In the annealing process the strained and elongated grains shown in Fig. 209 break up and rearrange themselves to form a new grain structure as shown in the micrograph. The annealed steel of the structure shown is now in excellent condition to with- stand further cold work in reducing it to finer sizes; or, if already at finished size, is in good condition to meet the demands of annealed wire service. The effect of reduction of section incident to wire drawing on the tensile strength and ductility of steel wire, and the marked change brought about in these characteristics by annealing, as just outlined, is shown in Fig. 216. This table is based on drafting and annealing practice in reducing a low-carbon steel rod in this case 0.10 per cent, carbon to a fine size of wire. It will be noted that between 80 per cent, and 90 per cent, reduction from the rod or annealed wire can be taken before annealing is necessary. 400 STEEL AND ITS HEAT TREATMENT It is found in practice that in cold drawing from a soft rod or annealed wire, the increase in tensile strength is a direct function of the amount of cold work, almost independent of other conditions. FIG. 207. Annealed (0.08 Carbon) Steel. (Tinsley.) Annealing practically brings the rod or wire, regardless of size, back to its original condition with regard to tensile strength and ductility. It will be noted that the final annealing does not bring FIG. 208. Steel Wire (O.OS Carbon) Given One Draft; 15 per cent. Reduction from Rod. (Tinsley.) FIG. 209. Steel Wire (0.08 Carbon) Given Several Drafts; 60 per cent. Reduction from Rod. (Tinsley.) the tensile strength as low as previous annealing. This is due simply to the fact that in annealing the fine sizes it is usual, in order to avoid the mechanical sticking of the wire in coils, to anneal at slightly lower temperatures than in ordinary process annealing. MISCELLANEOUS TREATMENTS 401 The second important function of annealing is that of refining grain, and its practical application in the wire mill covers principally the medium- and higher-carbon steels. The structure of wire rods with regard to size of grain is dependent upon the temperature at which the rods are finished in the hot rolling mill and upon the rate of cooling through the critical temperature of the steel. In steel of low carbon this is not of as much importance as in the higher-carbon steels, for the reason that the ordinary finishing temperature varia- tions of good rolling-mill practice have less effect on grain structure of soft rods, and therefore less effect on their physical properties. In higher-carbon steels a fine grain is important, for it is this struc- ture that makes for such steels their field of usefulness, where high strength, high elastic limit and toughness are required. Theoretically, the ideal structure would be obtained if the entire rod could be finished at about the critical temperature. But this is, of course, impracticable, for the reason that it is impossible to regulate the finishing temperatures so closely, and for the addi- tional reason that there is, necessarily, particularly in rolling very long lengths of very small sections, a marked difference between the finishing temperatures of the first and last end of a rod. The higher the finishing temperatures above the critical range the coarser the grain, and the coarser the grain the more does the steel lack the qualities that give it value. In order to destroy the coarse or uneven structure that may be created as just described, it is necessary to anneal the steel by heating it just above its critical temperature and slowly cooling it down. The effect of overheating in coarsening the grain structure of a 0.45 per cent, carbon steel and the refining influence of this type of annealing is shown in Figs. 211 and 212. The third and last class of annealing to be described that to obtain definite structure is one of comparatively recent develop- ment in the steel-wire industry and one which promises to be of con- siderable value. Annealing of this type is applied principally to the higher carbon wires. Since the structure of such wires can be varied considerably within a small range of annealing temperatures, it covers specific products and not general classes, as would be the case in regard to the two previously described types of annealing, Figs. 213 and 214 illustrate excellently this special type of annealing. These photomicrographs show the structures of two annealed pieces of the same coil of high-carbon wire, in which the annealing temper- ature of the one specimen was 130fr F., and of the other 1250 F. 402 STEEL AND ITS HEAT TREATMENT It is impossible to identify the structure by a simple observation of the fracture, which is the ordinary rough-and-ready method; nor is it possible to regulate annealing temperatures so closely without the use of pyrometers. In passing to the next great class of heat treatment applied to steel wire, patenting, it is interesting to note that we likewise pass to another class of wire as regards grading by carbon content. It naturally covers the medium-carbon steels, being employed chiefly on carbons between 0.35 and 0.85 per cent. In the medium-carbon steel wires strength and toughness are required for both process and finished wire. Patenting makes possible this combination of strength and toughness, and to this process is due in large measure a broad field of application for steel wire. The high strength and toughness of patented wire are due to its carbon condition and to its peculiar structure. The first step in the patenting process is to heat the wire to a temperature above its critical range. The degree of heating is regulated according to the carbon content of the steel, the size of rod or wire, and the time the material is subjected to the heat. After sufficient heating, the next step is to cool the material rapidly below its critical range, the structure obtained depending upon the rate of cooling. In practice, patenting is usually conducted as a continuous operation, the wire being led through the heated tubes of a furnace and cooled by being brought into the air or into a bath of molten lead comparatively cool but seldom under 700 F. A better understanding of the structure of a patented wire may be had by a comparison of the structure obtained by slow and by rapid cooling. If the steel after being heated is allowed to cool slowly through the critical temperature range, the homogeneous pre- existing solid solution of iron and iron carbide separates into a hetero- geneous mixture of two constituents, resulting in the plate-like struc- ture called " pear lite." In a patented wire, part of the carbide of iron is in solid solution and the remainder, while not in solid solution, has not had time to form into plates. The difference in structure between slow and rapid cooling is seen in Figs. 213 and 215. The photomicrograph of the patented wire shows, as a result of the rapid cooling, a structure that might be termed nondescript. Metallo- graphists will recognize the structure as " sorbite," which, in the cooling of the higher-carbon steels from above the critical tempera- ture, is that stage of transition just preceding the pearlitic, the final condition of annealed steel as shown in Fig. 213. The patented MISCELLANEOUS TREATMENTS 403 FIG. 210. Steel Wire (0.08 Carbon) Hard Drawn and then Annealed below the Critical Temperature. (Tinsley.) FIG. 211. Steel (0.45 Carbon) Over heated. (Tinsley.) FIG. 212. Steel (0.45 Carbon) Annealed. (Tinsley.) FIG. 213. Annealed (0.85 Carbon) Steel. (Tinsley.) FIG. 214. Specially Annealed (0.85 Carbon) Steel for Globular Structure. (Tinsley.) FIG. 215. Patented (0.85 Carbon) Steel. (Tinsley.) 404 STEEL AND ITS HEAT TREATMENT fcf | us J 00 CO ^|fi 00 1 s 8 00 2 P if ! * & 2 - < IN CO i (N HO IN BE 5 IN CN CO fe fi CO 1, g R fc < O 00 IN z CO Bfc o |2 1C 00 1C -IN fefl * H 1 (N l IQ CO 1 CO EQ g PROCESS AN- NEALED o o CO O CO (N 4M 00 o IS CO Ho IN Eg ^ w* ?, o IN og ^ CO CONDITION MATERIAL PER CENT. REDUCTION BY DRAWING TENSILE STRENGTH POUNDS PER SQ. IN. PER CENT. ELONGATION IN TEN INCHES ^ ll ~ o S?X3 31 ' o- 5 | w 'S Ll H H CO 218 000 C5 II O CO 184 000 IN Ij 156 000 O ES OQ CO 208 000 00 ii 8" -, CU O-i Ofi H Q CO CO 1 05 . ts o co 156 000 (N Ii - 5jl Q o 128 000 00 t". ^MntHZ^tT 1 ^ f he ^ ffe ^ f cold drawin e and Patenting on the physical properties of highe 5 At the first patenting stage, process annealing gave a tensile strength of 70,000 Ibs. p< is CO 176 000 CO (N Q H H Q I 163 000 IN 1 >-i W CO 00 (N 4 o 115000 (N 00 P s 146 000 00 IN IS s 00 IN 122 000 os O CONDITION OF MATERIAL PER CENT. REDUCTION BY DRAWING TENSILE STRENGTH POUNDS PER SQ. IN. PER CENT. ELONGATION IN TEN INCHES MISCELLANEOUS TREATMENTS 405 wire, therefore, represents an unsegregated condition as against the segregated or coarsely laminated structure of annealed wire. The high tensile strength of patented wire is due to the amount of carbon in solution, and its toughness to the fineness of the grain structure. Patenting serves two important functions in the wire business: 1. In the process of manufacture, the removal of the effects of cold work, such as drawing. 2. In the finished wire to give, in conjunc- tion with cold drawing, the required combination of strength and toughness. Strictly speaking, patenting is not necessary simply to relieve strain, for annealing would serve that purpose, but the structure obtained by patenting permits much further cold drawing than does the structure obtained by annealing. This is due primarily to the increased ductility and toughness of the patented wire. The effect of patenting as just described is shown in Fig. 217. In wire making, hardening and tempering should be conducted usually as a continuous process. In the making of tempered wire the material is first run through the heated tubes of a furnace, then quenched quickly in a bath of oil or water, then run into the tempering bath of, say, molten lead, each wire being in continuous motion from the time it enters the heating furnace until it is wound on a reel. Hardening and tempering apply to the higher carbon steel wires those in which the carbon range is from 0.65 per cent= to 1.00 per cent. With varying tempering temperatures between 500 and 1100 F., the tensile strength runs from about 340,000 Ibs. per square inch to 150,000 Ibs. per square inch. At the lower temper- ature the decrease in tensile strength is, as we should expect, much greater per 100 F. range than at the higher temperatures. From 500 to 600 F. there is a drop of 60,000 Ibs. per square inch, while between 1000 F. and 1100 F. the drop in tensile strength amounts to only about 10,000 Ibs. per square inch. FORGING No small percentage of the difficulty encountered in heat-treat- ment operations is due to improper forging methods, and ofttimes the heat-treatment operation is nothing more than a useless effort or attempt to get something out of a forged piece of steel that is not actually in it. Thus, the steel man is often blamed for the absence of quality in his steel that he actually put in it; and the heat-treat- ment man is blamed for his lack of ability to locate such qualities, which he properly assumes to exist, but which, nevertheless, the forge man took out by poor heating, unknown to himself or the other two. 406 STEEL AND ITS HEAT TREATMENT The strongest language that could be employed in an attempt to describe the general average heat-treatment equipment, the methods of heating, and personnel, as they are actually known to exist, would be altogether too mild and ineffective for a proper description of the heating methods and equipment in the majority of forge shops in the country. As in the case of machine work, the design of the hammers, presses and other machine equipment has made rapid strides forward, but the two most important factors of the operation from the metallurgical end namely, the man and the furnace have either stood still or gone backwards. Many well-informed and experienced men claim that the caliber of forge men to-day is not what it was years ago, and that better quality of work was produced with the old-fashioned coal or coke furnaces, though at a higher cost, than at present with furnaces burning oil or gas. There appears to be something in this statement, particu- larly in view of the high quality work turned out in Europe, where the use of high-speed machines, oil or gas fuel, and efficiency pro- duction methods, are not as prevalent as here. If such a difference actually exists, it can invariably be traced to the personnel of the plant, because, as in most operations involving the skill of the oper- ator as against the fixed movement of a machine, quality reflects the man and his knowledge of the work. But even so, we can and should be able to do better with fuel so closely linked with uni- formity of temperature, steadiness of operation, and ease of control. If we do not, then it is up to the man or the furnace and not to the hammer or the fuel, which is in itself a good argument for improve- ment of the heating and human equations in the operation. Two of the weak links in forging practice, from the metallurg- ical end, are the lack of uniformity and temperature of the heats and the method of handling stock in and out of the furnaces. As a rule, the heats are altogether too high, with the result that, while the surface is apparently hot, there may be actually a " bone " on the inside. It is common practice to see a bar drawn from a furnace that will actually drip, and yet when placed under the ham- mer there will be indications of lack of heating on the inside. It is the inside of the bar that determines the physical properties of the final forging and not the outside; and there is nothing gained in these quick " wash " or surface heats. Slow, soft, soaking heats, affording plenty of time to heat up, are more desirable than the higher quick heats. The idea should be to maintain the temperature of the furnace as near as possible to that actually required to soften MISCELLANEOUS TREATMENTS 407 the steel to the extent necessary for its proper shaping, and to give it plenty of time in the furnace thoroughly to soak at this temper- ature without overheating or oxidizing the outside. The fire should be soft and a little high in carbon, in order to reduce oxidation. The modern alloy steels do not require high, sharp, dripping heats, and the proper handling of them demands the slow, soft, non- oxidizing heats above referred to. The general design of forge furnaces is far below the standard of heat-treating furnaces and is a point usually left to the forge man or to a bricklayer. It is common practice to see furnaces hot on one side and cold on the other. Also, to hear complaints of lack of ability to heat steel properly in a furnace in which the burners blast directly against the stock, which naturally keeps the stock nearest the burner cool and heats the pieces farther away. There are hundreds of such designs in use that have been turned out by furnace builders who ought to know better. CHAPTER XVIII PYROMETERS AND CRITICAL RANGE DETERMINATIONS PYROMETERS 1 Pyrometers in General. The pyrometer has played a basic part in the development of intelligent heat treatment. In hardening rooms where pyrometers are not used, a discussion of any temper- ature treatment and instructions are given as the instructions must have been given in the Tower of Babel. There is no dis- tinction or mutual understanding of terms, and until a pyrometer and an accurate one is in a hardening room, it is not possible for those interested in the heat treatment in that room to talk to each other in a mutually intelligible way. Of course, where one old hardener has been in charge for twenty years and the manage- ment decides to take a chance on his staying with them and living for another twenty years, it may be all right to have everything locked up in his head; but where matters are more extensively and more modernly conducted, it is necessary to have some language in which people can talk; and the pyrometer, by virtue of its tempera- ture scale, which is a conventional scale of denned terms, affords the means of communication in a language that is mutually under- stood. In the same way it permits records to be kept for future reference. Where this is not done, men will be found trying to remember the heats at which they treated this, that or the other lot of steel; they cannot remember, and they are sure to get into trouble if they try to. The pyrometer has changed barbarian methods into civilized methods in a hardening room. There is need for a greatly extended use of pyrometers of the best possible grade, but more especially for an intelligent use of them that will in some measure compensate for the skill in producing them 1 It is the aim of this section to deal more with the rational use of pyrometers rather than with a detailed explanation of the theory and construction of the numerous instruments in commercial use for heat measurement. For a fuller explanation of the latter subject than is subsequently given, the reader is referred to standard reference books on the subject. 408 PYROMETERS AND CRITICAL RANGE DETERMINATIONS 409 and the money involved in their installation. The pyrometer is not all-sufficient, nor it is the cure-all for the troubles of a hardening plant. There should be an education of the man to look upon pyrometers as gauges and indicators of the existence of energy, and as an aid to him in executing his work and not as a means of releasing him from responsibility accompanying the exercise of judgment. The pyrometer has been of inestimable value in affording a means to check temperature, but and aside from the correlation of results its efficiency ends largely with that indication. The uniformity of heated product, however, depends upon the manner of applying the heat, which with the method and cost of operation is primarily a function of furnace design. It is possible to indicate a uniform temperature and yet not produce a uniformly heated prod- uct; and unless the heat is uniformly applied to the stock at the temperature indicated, then a uniform pyrometer reading is mis- leading and inconclusive. Thus an elaborate pyrometer system, with means for signaling variations in temperature and of record- ing these variations, is not conclusive evidence of accurate heating. The development toward better and cheaper results will be brought about by improved heating methods, even though the temperature recorded from any one point in a furnace chamber may be the same as that indicated from a similar point of another furnace less effici- ently designed. The time element is linked inseparably with all heating opera- tions. A piece of steel can absorb heat only so fast and no faster. Only by operating the furnace so that the maximum temperature is maintained for the length of time necessary uniformly to heat the steel throughout to that temperature, is it possible to produce the best results. In other words, the composition and the mass of the steel must be correlated with the time element. First deter- mine the length of time necessary, under standard furnace conditions, to produce the necessary results; then regulate the furnace by the aid of the pyrometer; and finally, place a clock beside the instru- ment and work the two together. The sooner the average heat- treatment man (and his superiors, for that matter) can be brought to realize that a pyrometer is almost valueless without the use of a time clock and common-sense observation of furnace conditions, the better. Thermo-Couples. For the usual operations in heat-treatment work involving temperatures of over 600 or 700 F,, the thermo- 410 STEEL AND ITS HEAT TREATMENT couple system is the most used. The principles upon which its use depends are simple. Expressed briefly, if the ends of two pieces of dissimilar metals (usually as wires) are joined together and one of the junctions (the " hot end ") is heated, the other junction (the " cold end ") being held at a constant temperature, a feeble electric current is generated in the circuit. This electromotive force, aside from being dependent upon the nature of the couple, is, for the thermo-couples in practical use, dependent upon the difference in temperature between the hot and cold ends. In regard to thermo-couples, standard base-metal compositions will generally give satisfaction between 600 or 700 F. and 1800 F.; while above 1800 F. couples of platinum and platinum-rhodium should be used. All base-metal couples should be readily replace- able, and, more emphatically, interchangeable. All couples should be suitably protected with iron pipes from oxidation and rough handling. Position of the Thermo-Couple. The fact that a pyrometer may show that some particular portion of the heating zone is at the proper temperature is no proof that the steel is also at that temper- ature. The hot end of the couple may be so placed that it must inevitably be hotter than the hearth of the furnace, or hotter than any material placed on the hearth. This will be true if the end of the couple is exposed to the direct heat of the flame. It might therefore be concluded that the tip should be as near the work as is possible, so that both may attain the same temperature and which is without doubt advisable in many instances. On the other hand, it has been noted in some cases in which the couples have been placed close to the work that the readings are not in accord with the temperature of the steel because the couples, being of smaller mass, take up readily the high peak of the flame tempera- ture. There are certain instances where it has been found by experience desirable to locate the tip of the couple in a recess in the furnace wall where it was out of the course of the flame and thus dependent for its temperature upon radiation from the main body of the furnace lining and radiation from the work; under some circumstances such a position is preferable. Millivoltmeter vs. Potentiometer. By inserting into the thermo- couple circuit, at the cold junction, a suitable device for measuring the electromotive force, a reading may be obtained in millivolts; or, by suitable calibration, a reading directly in terms of temper- ature. This indicating instrument (the pyrometer) may be of PYROMETERS AND CRITICAL RANGE DETERMINATIONS 411 the galvanometer or millivoltmeter type, or of the potentiometer type. The potentiometer in theory bears much the same relation to the millivoltmeter that the balance-arm scales bears to the spring scales. The constancy of both the spring scales and the millivoltmeter is entirely dependent upon the constancy of springs or of suspensions, and upon the absence of friction. The constancy of the potentiom- eter and of the balance-arm scales is dependent only upon the con- stancy of standard weights in one case and a standard electromotive force in the other. Standard weights are added to or removed from balance scales until a balance between known and unknown is obtained. Similarly in the potentiometer type varying fractions of a known and presumably standard electromotive force are opposed to the electromotive force of the thermo-couple until it is balanced just as a standard weight is moved along a scale arm for balance. This is the potentiometer not as it is, but as we would like to have it. The standard cell will not stay standard if any current is drawn from it and, consequently, the e.m.f. of the standard cell is not opposed to the e.m.f. of the couple in potentiometers as made for any ordinary use. Another cell or battery is brought into use and the e.m.f. of that is opposed to the e.m.f. of the couple. Now this secondary cell varies in e.m.f. from week to week and day to day, and even hour to hour under use, and it is necessary contin- ually to check this service cell against a standard cell and then to adjust for the differences that are creeping in all the time. The balance scales, therefore, instead of being operative with standard weights, have a sort of beaker of boiling water as the weight, which is continually boiling to less mass and which has to be filled up or adjusted every few minutes by comparing it with a standard weight, for the standard weight itself is not trusted on the scales nor is there any other weight, i.e., battery, that can be trusted on the scales that will not vary. Selection of Equipment. The selection of one type or the other is largely a matter for economic and technical consideration. In a word, the purchaser should consider the relation existing between (1) accuracy, sensitiveness and constancy, (2) ease of reading, and (3) the cost both initial and of up-keep. There is also a psycho- logical consideration that goes hand in hand with the above consider- ations and which should not be lost sight of: the millivoltmeter is a direct-reading instrument, which means that it is easy to read; the potentiometer requires a fair amount of manipulation and is 412 STEEL AND ITS HEAT TREATMENT somewhat less easy to read. The question then is: Which instru- ment will the average furnace man read more frequently? No matter how accurate a pyrometer may be, its value is only in the use made of it. Cold-end Temperature. The cold-end temperature is a source of prolific error in some pyrometer installations. It should be remembered that all instruments are calibrated for a definite cold- end temperature, usually 75 F. If the cold end is in a position such that it receives the direct or radiating heat from the furnace, and Copper LeaJs^ groxm A - re n .i pevafiii- e *.V- :; : ; '/ r :'v&2 FIG. 218. Compensating Cold End Temperatures with Auxiliary Couple. (Wilson-Maeulen Co.) therefore varies in temperature, the indicated temperature at the instrument will be incorrect. For this reason the cold end should always be kept cool, and at as near a constant temperature as is possible. This compensation may be accomplished by having the cold end as near the ground as possible; or by letting a small stream of cold water flow over the cold ends; or by connecting an auxiliary couple of the same electromotive force as the furnace couple in oppo- sition to -the couple in the furnace, and running the auxiliary couple to an underground point at the bottom of a pipe driven a few feet into the earth as shown in Fig. 218. The potentiometer type equipment frequently carries the cold end directly to the instrument, PYROMETERS AND CRITICAL RANGE DETERMINATIONS 413 entirely eliminating the effect of fluctuating temperatures near the furnace. Pyrometer Standardization. One of the most important points in connection with pyrometers is the necessity for frequent and regular calibration of the thermo-couples. All base-metal couples should be standardized at least once a week, and oftener if possible. Fur- ther, new couples should always be standardized before use, since errors may frequently be found even in supposedly correct new couples. There are two general methods for standardization or calibration of thermo-couples: (1) Checking against the melting- or freezing- points of known salts or metals, and (2) checking against a standard millivolt meter or pyrometer. Standardization with Common Salt. An easy and convenient method 1 for standardization and not necessitating the use of an expensive laboratory equipment is that based upon determining the melting-point of common table salt (sodium chloride). While theoretically salt that is chemically pure should be used (and indeed this is neither expensive nor difficult to procure), commercial accu- racy may be obtained by using common table-salt such as is sold by every grocer. The salt is melted in a clean crucible of fire-clay, iron or nickel, either in a furnace or over a forge-fire, and then further heated until a temperature of about 1600 to 1650 F. is attained. It is essential that this crucible be clean, because a slight admixture of a foreign substance might noticeably change the melting-point. The thermo-couple to be calibrated is then removed from its protect- ing tube and its hot end is immersed in the salt bath. When this end has reached the temperature of the bath, the crucible is removed from the source of heat and allowed to cool, and cooling readings are then taken every ten seconds on the millivoltmeter or pyrometer. A curve is then plotted by using time and temperature as co-ordinates, and the temperature of the freezing-point of salt, as indicated by this particular thermo-couple, is noted, i.e., at the point where the temperature of the bath remains temporarily constant while the salt is freezing. The length of time during which the temperature is stationary depends on the size of the bath and the rate of cooling, and is hot a factor in the calibration. The melting-point of salt is 1472 F. and the needed correction for the instrument under obser- vation can be readily applied. The curves in Figs. 219 and 220 illus- trate the calibration of a correct and incorrect pyrometer. 1 Carpenter Steel Co. 414 STEEL AND ITS HEAT TREATMENT 180 160 140 120 100 80 60 40 20 / /\ / / / / \ / / 1 I I J / 7 s" r ^ '-""" -' x s* 1050 1600 1550 1500 1450 Degrees Fahrenheit FIG. 219. Diagram Showing the Calibration of a Pyrometer which Reads 45 F. Too High. (Carpenter Steel Co.) 180 100 140 120 100 SO GO 40 20 ] / J\ \ f 1 / / ^ ** ^ << "*' ^ i * 1 ^ x* 1650 1600 1550 1500 1450 .Degrees Fahrenheit FIG. 220. Diagram Showing the Calibration of a Pyrometer which is Correct. (Carpenter Steel Co,) PYROMETERS AND CRITICAL RANGE DETERMINATIONS 415 It should not be understood from the above, however, that the salt-bath calibration cannot be made without platting a curve : in actual practice at least a hundred tests are made without platting any curve to one in which it is done. The observer, if awake, may reasonably be expected to have sufficient appreciation of the lapse of time definitely to observe the temperature at which the falling pointer of the instrument halts. The gradual dropping of the pointer before freezing, unless there is a large mass of salt, takes place rapidly enough for one to be sure that the temperature is constantly falling and the long period of rest during freezing is quite definite. The procedure of detecting the solidification point of the salt by the hesitation of the pointer without platting any curve is suggested because of its simplicity. Complete Calibration of Pyrometers. For the complete calibra- tion of a thermo-couple of unknown electromotive force, the new couple may be checked against a standard instrument, placing the two bare couples side by side in a suitable tube and taking frequent readings over the range of temperatures desired. If only one instrument, such as a millivoltmeter, is available, and there is no standard couple at hand, the new couple may be calibrated over a wide range of temperatures by the use of the follow- ing standards ; Water, Boiling-point 212 F. Tin, under charcoal, Freezing-point 450 Lead, under charcoal, Freezing-point 621 Zinc, under charcoal, Freezing-point 786 Sulphur, Boiling-point 832 Aluminum, under charcoal, Freezing-point 1216 Sodium chloride, Freezing-point 1474 Potassium sulphate, Freezing-point 1958 A good practice is to make one pyrometer a standard; calibrate it frequently by the melting-point-of-salt method, and each morning check up every pyrometer in the works with the standard, making the riecessary corrections to be used for the day's work. By pur- suing this course systematically, the improved quality of the product will much more than compensate for the extra work. Central Switch-boards. For plants in which there are a number of thermo-couples, one indicating instrument with a central switch- board may be used. As many as sixteen couples may be wired to one selective switch, the maximum number simply depending upon the elasticity of the system and the convenience of the operator. 416 STEEL AND ITS HEAT TREATMENT A wiring diagram for such an installation is shown in Fig. 221. By throwing the switch from one contact to another the connection is made with each individual furnace. For large heat-treatment plants the time of one man is generally taken in attending to the system, he signaling the individual operators by means of lights and belts the relative temperatures in the furnace. We have previously commented upon such systems. The Central System. The Chalmers Motor Company operate their system, 1 having two central switch-boards with sixteen furnaces on a switch, as follows; Couple 8 Couple I , FIG. 221. Wiring Diagram Pyrometer and Selective Switch. Showing Four Couples Connected with the Switch, Openings for Four More. (Hoskins Mfg. Co.) " We regulate the heat of the furnaces by a series of lights each furnace having over it a red, blue and green light. These are used as follows: We will say that the temperature of an empty furnace which we are about to use is 1600 F. The loading of the furnace with forgings necessarily reduces the heat by radiation any- where from 100 to 250, depending upon the number of pieces put in the furnace. When we commence to bring the heat up again to the proper place and it gets to about 1575, the man at the switch- board throws on a blue light, which means to the furnace operator that the heat is still considerably too low. When the temperature reaches about 1590 the blue and green light is turned on, which signifies to the operator that the furnace is still not quite hot enough. 1 Personal Correspondence. PYROMETERS AND CRITICAL RANGE DETERMINATIONS 417 When the 1600 point is reached the green light is turned on; this is the O. K. light and means that the temperature is correct. The steel is then allowed to soak at this temperature for the time neces- sary to affect the whole mass. If the heat during the operation gets too high we use signals in an inverse manner, the red and green lights being thrown on. If it shows a dangerous rise in temperature the red light is thrown on. All of these lights are accompanied by the ringing of a loud bell in the heat-treating department, which automatically attracts the attention of the man operating the furnaces, who at once inspect their individual furnace lights to see if their temperature is correct." DETERMINATION OF THE CRITICAL RANGES Critical Ranges. The practical importance of knowing the exact location of the critical ranges of steel to be treated is obvious. Their determination by means of pyrometers is based upon the fact that the changes taking place in the steel at those temperatures involve an absorption of heating during heating (the decalescent points) and a giving out of heat on passing through these ranges on cooling (the recalescent points). Decalescent vs. Recalescent Points. Before discussing methods, it should be stated that, for the majority of heat-treatment work, it is more important to know the location of the decalescent points than that of the recalescent points. This is for several reasons. To effect a complete change of the original structure of the steel, it must be at least heated slightly beyond the Ac3 range, regardless of the position of the Ar3 range. If the steel were to be heated only to the Ar3 range, a complete change in structure cannot take place, because the Ar3 range is always below the temperature of the Ac3 range. Further, the position of the Ar ranges is, experimentally at least, dependent upon the maximum temperature to which the steel is heated, upon the length of heating at that temperature, and upon the rate of cooling from that temperature. It should also be again stated that the determination of the upper critical range is of more importance than that of the lower critical range (Al), since the majority of hardening and annealing work demands a complete change of structure which is obtained only above the upper critical range (Ac3). Temperature Difference Instruments. American-made instru- ments for determining the critical ranges of steel are based either 418 STEEL AND ITS HEAT TREATMENT upon a temperature difference basis, or upon a direct record of a single instrument. The method used by the Leeds & Northrup apparatus, typical of the first class, involves the following points: Two bodies are heated together in the same furnace, the one being the steel under test and the other being a body which will Q&f. .3 Phos,= .033 Mng. .700 Sin. = .253 Sir. = .029 Heating.Curve Cooling Curve Abscissae Temperature Differences between Sample and Non-recalescing Body. FIG. 222. Transformation Curves. (Leeds & Northrup Co.) heat uniformly without undergoing any changes. If the bodies are in sufficiently close contact they will heat at the same rate and, barring changes in one which do not occur in the other, will remain equal in temperature. When, however, the steel undergoes an inter- nal change involving absorption or liberation of heat, its temperature changes relatively to the other body and a temperature difference is set up between the two. Hence the apparatus for the location of PYROMETERS AND CRITICAL RANGE DETERMINATIONS 419 critical points by this method is designed to do two things: first, to measure the temperature of the sample; second, to indicate the temperature relationship between the sample and the unchanging body. A curve using temperatures as ordinates and temperature differences as abscissae is the best way of making use of the results. Temperature Difference Records. Fig. 222 is a reproduction of such a plot. From the start of the test until 1205 the temperature difference is small and constant. When the temperature of 1369 is reached a sudden increase in the temperature difference takes place, the Acl range. As soon as this sudden change ceases (i.e., transformation is completed), the sample and the unknown begin FIG. 223. Critical Range Curve on a Direct-reading Apparatus. Carbon, 0.44 per cent.; Manganese, 0.53 per cent.; Phosphorus, 0.035 per cent.; Sulphur, 0.025 per cent.; Silicon, 0.028 per cent. to equalize in temperature and the record of their decreasing differ- ence follows a typical cooling curve between 1380 and 1455, except at about 1407, where the Ac2 transformation begins to affect the record. At 1407-1410 this Ac2 change is completed. Again at 1450 there is a departure from a smooth curve; this is the beginning of the third transformation, which transformation is not completed until about 1495. This is the Ac3 transformation. On cooling, the reverse takes place, except that the two upper points occur closer together and appear as one. The lowest range is clear cut. Direct-reading Instruments. Fig. 223 shows a record obtained from a Bristol instrument. Leaving aside a discussion of the 420 STEEL AND ITS HEAT TREATMENT scientific pros and cons, the three main objections to this class of instrument are: (1) the small area covered by the record, involving less accuracy; (2) lack of that degree of sensitiveness which is necessary to bring out the upper critical ranges; and (3) a curve showing direct temperatures instead of temperature difference. Practical Method for Determining Critical Ranges. For plants which have to determine the critical ranges but infrequently less costly apparatus may be used. The outfit should consist of a thermo- couple made of small wires so as to respond quickly to any slight variation in temperature; the necessary leads; and a sensitive milli voltmeter or pyrometer with a finely divided scale. This instrument may also be used as a standard, or checking instrument, for calibration work. The specimens to be tested should be small so as to heat uniformly and quickly. These may be either a small cylinder, say f in. diameter by If in. long, or duplicate pieces each 1J in. long by f in. wide by J in. thick. In the former case the end of the couple is inserted in a small hole drilled through the axis of the cylinder to a depth of about \ in.; in the latter case the pieces are clamped together, one on either side of the end of the thermo- couple so as to form a tight contact. The specimen is then heated in any convenient manner, readings being taken every few seconds as the critical ranges are reached. When the indicated temperature is well above the upper critical range, the specimen is removed from the heat, allowed to cool not too rapidly, and readings taken to obtain the Ar ranges. The temperature readings, or difference in readings, should then be plotted against the time to obtain the necessary curves. PYROMETERS AND CRITICAL RANGE DETERMINATIONS 421 TEMPERATURE CONVERSION TABLE BY DR. LEONARD WALDO Reprint from Metallurgical and Chemical Engineering. c. 10 20 30 40 50 60 70 80 90 -200 -100 - F. -328 -148 (- 32 -346 -166 + 14 F. -364 -184 - 4 F. -382 -202 - 22 F. -400 -220 - 40 F. -418 -238 - 58 F. -436 -256 - 76 F. -454 -274 - 94 F. -292 -112 F. -310 -130 32 50 68 86 104 122 140 158 176 194 C. F. 100 200 300 400 500 600 700 800 900 212 392 572 752 932 1112 1292 1472 1652 230 410 590 770 950 1130 1310 1490 1670 248 428 608 788 968 1148 1328 1508 1688 266 446 626 806 986 1166 1346 1526 1706 284 464 644 824 1004 1184 1364 1544 1724 302 482 662 842 1022 1202 1382 1562 1742 320 500 680 860 1040 1220 1400 1580 1760 338 518 698 878 1058 1238 1418 1598 1778 356 536 716 892 1076 1256 1436 1616 1796 374 554 734 914 1094 1274 1454 1634 1814 1 2 3 4 5 6 7 8 9 10 1.8 3.6 5.4 7.2 9.0 10.8 12.6 14.4 16.2 18.0 1000 1832 1850 1868 1886 1904 1922 1940 1958 1976 1994 1100 1200 1300 1400 1500 1600 1700 1800 1900 2012 2192 2372 2552 2732 2912 3092 3272 3452 2030 2210 2390 2570 2750 2930 3110 3290 3470 2048 2228 2408 2588 2768 2948 3128 3308 3488 2066 2246 2426 2606 2786 2966 3146 3326 3506 2084 2264 2444 2624 2804 2984 3164 3344 3524 2102 2282 2462 2642 2822 3002 3182 3362 3542 2120 2300 2480 2660 2840 3020 3200 3380 3560 2138 2318 2498 2678 2858 3038 3218 3398 3578 2156 2336 2516 2696 2876 3056 3236 3416 3596 2174 2354 2534 2714 2894 3074 3254 3234 3614 F. C. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 .56 1.11 1.67 2.22 2.78 3.33 3.89 4.44 5.00 5.56 6.11 6.67 7.22 7.78 8.33 8.89 9.44 10.00 2000 3632 3650 3668 3686 3704 3722 3740 3758 3776 3794 2100 2200 2300 2400 2500 2600 2700 2800 2900 3812 3992 4172 4352 4532 4712 4892 5072 5252 3830 4010 4190 4370 4550 4730 4910 5090 5270 3848 4028 4208 4388 4568 4748 4928 5108 5288 3866 4046 4226 4406 4586 4766 4946 5126 5306 3884 4064 4244 4424 4604 4784 4964 5144 5324 3902 4082 4262 4442 4622 4802 4982 5162 5342 3920 4100 4280 4460 4640 4820 5000 5180 5360 3938 4118 4298 4478 4658 4838 5018 5198 5378 3956 4136 4316 4496 4676 4856 5036 5216 5396 3974 4154 4334 4514 4694 4874 5054 5234 5414 3000 5432 5450 5468 5486 5504 5522 5540 5558 5576 5594 3100 3200 3300 3400 3500 36CO 3700 3800 39CO 5612 5792 5972 6152 6332 6512 6692 6872 7052 5630 5810 5990 6170 6350 6530 6710 6890 7070 5648 5828 6008 6188 6368 6548 6728 6908 7088 5666 5846 6026 6206 6386 6566 6746 6926 7106 5684 5864 6044 6224 6404 6584 6764 6944 7124 5702 5882 6062 6242 6422 6602 6782 6962 7142 5720 5900 6080 6260 6440 6620 6800 6980 7160 5738 5918 6098 6278 6458 6638 6818 6998 7178 5756 5936 6116 6296 6476 6656 6836 7016 7196 5774 5954 6134 6314 6494 6674 6854 7034 7214 C. 10 20 30 40 50 60 70 80 90 EXAMPLES: 1347 C. 2444 F. +12.6 F. = 2456.6 F.: 3367 F. = 1850 C. +2.78 C. : 1852.78 C. INDEX Abrasion, resistance to, 14 Abrasive wear in Mn steels, 346 Air control, 182 Air cooling, 42, 58, 249, 276 Air hardening, 246, 249 Allotropic ferrite, 29 Alloy steel: hard spots in, 51 necessity for heat treatment, 1, 258 Alpha ferrite, 29 Alternating impact tests, 10 American gas furnace process, 147 Animal charcoal. See Charcoal. Annealing. See Ch. III. air cooling after, 58 commercial, 62, 248 definition of, 39 effect of, 44 elemental considerations in, 39 furnace cooling after, 57 hyper-eutectoid steels, 254 length of, 48 pit, 58 rate of cooling after, 53, 57 rate of heating in, 46 size of object, 58 slow cooling after, 58 special methods, 58, 59 temperature for, 47, 231, 234, 238, 247 time experiments, 51 vs. toughening, 110 wire, 398 Armor plate, 319 Arrangement of charge, 214 Ar ranges, 32 Atmosphere in furnace, 189 Austenite, 27, 67, 231, 348 Automatic furnaces: for die blocks, 373 for shrapnel, 249 Automobile steel, 1, 234, 329 Axles, 1, 8, 9, 90, 236, 241, 243, 245, 247, 315, 329, 340 B Ball-bearings, 300 Ballistic tests, 14 Barium carbonate, 127, 133, 137 Baths: heating, 75 . salt, 76 tempering, 99 Best case, 162, 165 Beta ferrite, 31 Bit steel, 394 Bolts, carburizing of, 141 Boring, hollow, 88 Box annealing, 61 Boxes for carburizing, 140 Brains, purchasing of, 184 Brine, 80 Brinell hardness. See Hardness. of carbon steels, 229 of chrome-nickel steels, 318, 320 of chrome-vanadium steels, 337 of nickel steels, 290 Brittleness, 3, 7, 8, 9, 110, 233, 240, 246, 268, 270, 271, 275, 295, 306, 344, 353 Stead's, 61 B.T.U. values, 175 Burners, 186 C Calcium chloride for quenching, 81 Calibration of pyrometers, 415 Capacity of the steel, 106 423 424 INDEX Carbides, 296, 304, 335, 348 Carbon : concentration of, 122, 271 direct action in carburization, 114 maximum in case, 165, 276 plus carbon monoxide, 124 solution of in carburization, 125 Carbon content: for tools, 358 influence of, 3, 110 in manganese steels, 345 in nickel steels, 264 Carbon monoxide, 115, 117, 122 Carbon steel: under 0.15 per cent., 230 0.15-0.25 per cent., 232 0.25-0.35 per cent., 236 0.35-0.40 per cent., 240 0.45-0.60 per cent., 246 over 0.60 per cent., 253 Carbonates, 116 Car bottoms, 219 Carburization : boxes, 140 carbon monoxide plus hydrocar- bons, 122 carbon plus carbon monoxide, 124 depth of penetration, 126 effect of chrome, 297 gas process, 147 heat treatment requirements, 154 object of, 112 of chrome-nickel steels, 308 of nickel steels, 267, 270, 271, 273 requirements of, 112 steel for, 113, 232 temperature of, 125, 132, 133, 134, 155 with carbon monoxide, 115, 117 with simple solid cements, 135 wood charcoal, 115 Case carburizing. See Carburizing. Case hardening: gears, 386-389 maximum efficiency in, 165 treatment of hyper-eutectoid steels, 156 treatment of hypo-eutectoid steels, 155 Case, the best, 162, 165 Castings, 397 Cellular structure, 42, 53 Cementite, 16, 24, 63, 114, 156, 163, 165, 169, 253, 295, 301, 335, 391, 402 Centigrade tables, 421 Central pyrometer systems, 415 Chamber, height of, 202 twin- furnaces, 226 Changes: in diameter, 366 in heating, 32 in length, 366 Charcoal, 133, 135, 136 Charge : height of, 202 influence in heating, 205 influence of arrangement, 205 placement of, 75 Charging, 39 Chemical composition, effect of, 1 Chipping chisels, 368 Chisels, 298, 368 Chrome: influence in carburization, 297, 308 in manganese steels, 347 vs. silico-manganese, 351 Chrome steels: general characteristics, 295 0.5 chrome, low carbon, 296 0.5 chrome, 0.35-0.50 carbon, 297 0.5 chrome, over 0.50 carbon, 297 1.0 chrome, 300 2.0 chrome, 302 high chiome, 302 Chrome-nickel steels: carburization, 309 gears, 390 heat treatment, 309 low Cr, low Ni, 310 0.5 Cr, 2.5 Ni, 322 0.6 Cr, 3.5 Ni, 321 0.75 Cr, 3.0 Ni, 322J 1.0 Cr, 1.75 Ni, 322 l.SCr, 3.5 Ni, 319 Mayari, 329 special analyses, 327 vs. chrome-vanadium, 306, 335 INDEX 425 Chrome-vanadium steels. See Vana- dium. Circulation for cooling oil, 84 Classification of: gear steel, 386 heat treatment after carburization, 155 nickel steel-, 258 Coal furnaces, 218 Coffin process, 244 Cold crystallization, 231 Cold-end temperatures, 410, 412 Cold rolls, 302 Cold rolling, 38 vs. strength, 6 Cold work, effect on structure, 38 Color chart, 369 Colors in tempering, 97 Combustible mixture, 177 Combustion, furnace atmospheres from, 189 Commercial annealing, 39, 62, 248 Commercial data in carburization, 135 Commercial ratio of chrome and nickel , 307 Compensation of pyrometers, 412 Compressed air in quenching, 85 Compressive strength, 5 Conservation of heat, 222 Contact couples in heating, 53 Continuous furnaces, 249, 252, 373 Contraction in hardening, 88 Contraction of area, 5 Conversion, temperature, 421 Cooling, after annealing, 53, 231, 346 Cooling the oil bath, 82 Cooling the water bath, 82 Corrosion, 295 Couples, 409 Cracking, 87, 246. Crank shafts, 1, 329 Critical ranges, 25, 417 changes at, 41 effect of manganese, 345^ effect of nickel, 258, 265~ heating over the, 42 merging of, 31 of chrome steel, 295, 299 of chrome-nickel steel, 309, 327 I Critical ranges of high-carbon steel, 253 of hyper-eutectoid steel, 62 of manganese steel, 350 of tool steel, 363 Cutters, 378 Cyanide hardening, 149 Cyanides in carburization, 117, 139 Dead soft steel, 230 Decalescence, 417 Deck plate, 309, 319 Depth of penetration, 272 Design of furnace, 186 Determination of critical ranges, 417 Diameter, effect on tests, 229 changes in, 366 Die Blocks, 298, 369 Dies, 247, 382. Differential hardening, 82 Diffusion, 44, 50, 232, 267 Distortion, 364 Distribution of carbon, 126 Door heights, 200 Double annealing, 59 Double carbide steel, 304 Double quenching, 94, 163, 237, 254, 276 Double regenerative quenching, 166 Drawing dies, 302 Drawing of wire, 398 Drilling, hollow, 88 Drills, 298, 376 Drop tests, 8 Ductility, 5, 7, 106, 257, 306 Duplex process, 329 Duplication of results, 2, 107 Dynamo sheet iron, 352 Dynamic strength, 2, 110, 236, 306, 307, 319, 337 E Effect of: chrome, 295, 306 manganese, 344 mass, 286, 322, 330, 363 nickel, 257, 306 silicon, 350 vanadium, 335 426 INDEX Elastic limit, 3, 5, 41 Electricity: atmospheres with, 190 for heating, 191 Electromagnets, 352 Elongation, 5, 41 Endurance, 6, 10 Enfoliation, 120, 271, 279 Engine forgings, 232, 234 Engraved, dies, 374 Equalization, 44, 48, 267, 269 Equalizing action of carbon monoxide, 124 Equipment, pyrometer, 411 Eutectoid for nickel steel, 267 steel, 17 Expansion in hardening, 88 Fahrenheit tables, 421 Failures of heat-treated axles, 246 Fatigue, 2, 6, 9, 236, 307 Ferrite, 23, 29, 164, 257, 258, 296 Ferro-cyanides, 127, 139 Files, 298, 302, 379 Fine-grain annealing, 59 Fire-ends, 409 Five-ply steel, 396 Flanges, 232 Flue construction, 222 Force, 2 Forging, 406 Forging temperatures for tool steel, 363 Fragility, 9 Frames, automobile, 1 Fuel: cost of delivering, 180 costs, 175 efficiency, 177 equipment, 181 fluid, the, 178 oil, 182 oil, air control with, 182 selection of, 179 supply, 181 the right, 177 uniformity of, 182 Fuel: vs. furnace design, 209 vs. operations, 178 vs. product, 191 Furnace: atmospheres, 189 batteries, 226 cooling in toughening, 108 design, 186, 209, 215, 407 equipment, 74, 185, 192 guarantees, 196 plans, 223 temperature of in heating, 46 the one, 195 Furnaces : automatic, 250, 373 car-bottom, 219 carburizing, 142 coal, 218 continuous, 250, 373 forge, 407 general considerations, 227 muffle, 207 overfired, 213 perforated arch, 212 practical notes, on, 227 semi-muffle, 209 shrapnel, 250 twin-chamber, 226 underfired, 196, 220, 249 unit system, 225 G Gamma ferrite, 31 Gases, action of, in carburization, 114 regulating the, 39 Gears, 1, 141, 236, 247, 329, 351, 386 Grade, in tool steel, 357 Gradual cements, 133, 139 Grain size, 268 at Ac3, 41 beyond Ac3, 33, 42 by different rates of cooling, 42 effect of work on, 38 in carburization, 134 Gun barrels, 235 forgings, 91, 241,354 INDEX 427 Hardening: cyanide, 149 definition, of, 65 differential, 82 heating for, 71 nickel steels, 267 pack, 148, 152 strains, 97 superficial, 148 temperature for, 70, 232, 237, 241, 253, 274, 286, 336, 370 tool steel, 362 vs. annealing, 67 Hardness, 11, 230, 272, 290, 295, 318, 320, 337, 360 Brinell, 11. Also see Hardness, cutting, 254 due to chrome, 297 scleroscope, 13, 170. Also see Hardness, wearing, 254 Hard spots, 51 Heat, quality of, 188 Heat application, 40, 195, 249 Heat conservation, 222 Heat reservoir, 201 Heat treatment: definition of, 25 growth of, 1 necessity for, 1 Heating: changes on, 32, 40, 65 costs, 173 distinctive conditions in, 173 factors, 174 for carburization, 143 for forging, 406 for hardening, 71 for tools, 367 in lead, 379 in salt, 76, 168 influence of chrome in, 297 large sections, 46 length of, 268, 370 prolonged, 306 rate of, 46, 51, 52 uniform, 196 unit, standard, 174 Heating: with electricity, 191 Height of chamber, 202 Height of charge, 202 High-carbon case, treatment of, 160 High-speed steel, 353 High temperature carburization, 134 High temperatures, effect on springs, 391 Hollow boring, 88 Hollow tools, 383 Hot-ends, 410 Hot work, effect of, 38 Human element, 39, 74, 183, 237, 245, 406, 408 Hydro-carbons, 118, 122 Hyper-eutectoid steel, 17 annealing of, 62 zones in carburization, 120 Hypo-eutectoid steel, 17 annealing of, 40 Impact strength, 2, 8, 9, 106, 134, 170 Impurities in carburization, 114 Influence. See Effect. Intensifies, 337 Intermediary types of carburized zones, 122 Interrupted regenerative quenching, 168 Jar steel, 298 Knives, 298 K Laminations, 257 Lead baths, 75, 102, 379 Ledges, 197 Length, change in, 366 of heating, 39, 370 Levers, 232 Liquation, 129, 272 Locomotive axles, 243 Low-temperature carburization, 133 428 INDEX M Machine parts, 240 Machinery steel, 232 Machining: effect on structure, 38 quality, 257 Magnet steel, 353 Magnet, use in hardening, 72 Manganese: in carburization, 113 on hardening, 95 on machining, 231 steels, 344 M'artensite, 68 Martensitic steels, 258, 302 Mass, influence of, 202, 229, 286, 287, 322, 341, 363 Mayari steel, 329 Mechanical mixture, 16 Mechanical work, effect in annealing, 48, 59 Microscope, use of, 44 Microstructure: of high-carbon steels, 255 of nickel steels, 259 Milky-ways, 50 Milling cutters, 378 Millivoltmeters, 410 Mineral hardness, 306 Mixed cements, 272 Molybdenum steels, 353 Motion in hardening, 74 Muffle furnaces, 207 N Natural alloy, 329 Natural, steel in the, 1, 3 Navy specifications for tool steel, 360 Network, 34, 42, 53, 63 Nickel: effect on physical properties, 264 influence of, 267 influence on critical ranges, 265 Nickel steels: 2 per cent., 264, 274 3.5 per cent., 274, 276, 310, 315 5 per cent., 264, 267, 268, 276, 280, 291 Nickel Steels: 10 per cent., 264 25-35 per cent., 264, 291 carburization of, 270 for gears, 360 Nickel-chrome steel. See Chrome nickel. Nickel-vanadium steel, 342 Nitrogen in carburization, 116 Normalizing, 63 Nuts, 232 Obstructing agents, 57, 258 Oil. Also see Fuel Oil. quenching speed of, 78 vs. water for hardening, 243 Oil baths, 82, 101 Oil burners. See Burners. Oil tempering, 80 Oil-tempered gears, 386, 389 Oil-well bits, 394 Operators, value of, 184 Oscillating temperatures, 132 Osmondite, 67 Overfired furnaces, 231 Overheating, 72, 402 Oxidation, protection from, 371 Oxygen, action in carburization, 115 P Pack hardening, 148, 152 Packing for carburization, 141 Patenting, 402 Pearlite, 16, 23, 27, 56, 110, 258, 265 Penetration, depth of, 126, 272 velocity of with chrome, 297 Perforated arch furnaces, 212 Phosphorus, 3 Physical properties at Ac3, 41 Pins, 232 Pit annealing, 58 Polyhedral steels, 259 Potentiometers, 410 Preheating, 40, 370 Process annealing, 399 Producer gas, 181 INDEX 429 Prolonged heating of nickel steels, 271 Propeller shafts, 329, 354 Protection of steel, 60, 371 Protective deck plate, 309, 319, 324 Punches, 382 Punching, effect on structure, 38 Purchasing brains, 184 Pyrometers, 408 standardization, 413 use of contact couples, 53 Q Quality of heat, 188 Quality of product vs. first cost, 173 Quench-toughening, 111 Quenching : after tempering, 99 baths, 77 best temperatures, 71 double, 94 manner of, 90 media, 77, 109 special methods, 80 speed, 77 tanks, 87 water for, 80 Radiation systems for cooling oil, 84 Eate of cooling, 39, 248 Rate of heating, 39 Razors, 298 Reamers, 382 Recalescence, 417 Reduction of area, 5, 41 Refinement, 33, 39, 41, 65, 160, 231, 232, 237, 238, 248, 401 Regeneration, 159, 160, 180, 274, 280 Relation of austenite to carbide, 347 Relation of physical tests, 10 Requirements of gears, 386 Resilience, 8 Rifle barrels, 354 Rings, 383 Rivet sets, 384 Roller bearings, 300 Rotary bending, 6 Rounds, hardening of, 91, 93 S Safe steel, 396 Salt, use in carburizing, 137 standardization of pyrometers, 413 Salt baths, 76, 102, 168 Sand baths, 100 Saws, 298, 385 Scleroscope, 13, 130, 290, 318, 320, 337 Screw stock, 240 Screws, carburizing of, 141 Seams, 257 Selection of pyrometer equipment, 413 Selection of tool steel, 357 Sensitiveness of manganese steels, 344 Sensitiveness of silico-manganese steels, 351 Shafts, 90, 290 Shock, resistance to, 275 Shore-hardness. See Scleroscope. Shrapnel, 249 Silicon steels, 350, 352 Silico-manganese steels, 351, 390 Size of section, 58, 229, 248 Slow cooling, 33, 42, 57, 67, 107, 157, 167 Soft forging steel, 236 Solid solutions, 27 Solution of carbon in carburization, 125 Sorbite, 56, 59, 69, 103, 110 Specifications for tool steel, 360 Spheroidal cementite, 63, 163, 402 Spheroidal ferrite, 164 Spheroidalizing, 63, 163, 164, 254, 355 Springs, 1, 391 Static strength, 1, 2, 24, 319 Standard heating unit, 174 Standardization of results, 12, 110 Stead's brittleness, 61 Steam hardening, 246 Steel: burnt, 42 castings, 397 for carburization, 113 nature of, 16 Steering parts, 1 Stresses and strains, 2, 6, 39, 97, 107 Structure, definition of, 25 Structure of slowly cooled steel, 17 Sudden cements, 133, 139 430 INDEX Suddenly applied loads, 7 Sulphur diffusion, 143 Summary for case hardening, 169 Superficial hardening, 148 Table for temperature conversion, 421 Tank, size of quenching, 87 Taps, 366 Temper, 358 colors, 97, 93 Temperature : conversion table, 421 effect on grain size, 33 effect on network, 34 relation of surface and interior, 53 Temperature of: annealing, 39, 47 carburization, 128, 132, 155 hardening, 70, 72, 274 pack hardening, 152 quenching bath, 77 toughening, 104 Tempered axles, 244 Tempering: color for tools, 361 definition, of, 96 for depth, 98 gears, 389 handling material in, 101 methods, 100 oil, 80 plate, 100 quenching after, 99 springs, 392 Tensile strength, 2, 41 of cementite, 24 of ferrite, 23 of pearlite, 23 Testing: comparative results, 229 errors in, 307 purpose of, 1 Tests from center, 229 Thermo-couples, 409 Threading, treatment for, 231 Tie rods, 232 Time of heating, 231, 409 Tool steel, annealing of, 59, 60 proper carbon for, 358 selection of, 357 Torsional strength, 1, 5 Tough-hardness, 295 Toughening, 103 high vs. low temperature, 109 range, 1,04 temperature vs. mass, 330 vs. annealing, 110 vs. ductility, 106 vs. impact strength, 106 Toughness, 231, 360 Transference numbers, 12, 230, 290, 318, 320, 337 Transition constituents, 56 Troostite, 68, 96 Tungsten steel, 350, 353 Twin chamber furnaces, 226 U Underfired furnaces, 220 Underfiring, 196 Uniform heating, 196 Unit furnace system, 225 Uses of chrome nickel steel, 329 Value of furnace operator, 184 Valve stems, 232 Vanadium, effect of, 335 Vanadium steels: gears, 390 nickel, 342 Type A, 339 Type D, 340 Type G, 341 Vault steel, 396 Velocity of penetration with chrome, 297 Vents, 198, 199 Vibratory stresses, 1 W Warping, 89, 232 Water bath, cooling the, 82 Water quenching, 80, 246 INDEX 431 Water spray, 79 Water toughening, 349 Water vs. oil for hardening, 243 Wear, 1, 14, 240, 302, 346 Welding of alloy steel, 330 Welding properties of tool steel, 363 Well bits, 298, 394 Wire, 398 Wood charcoal, 115, 127 Work, effect on grain size, 38 Working conditions, influence of, 179 Working strength, 4 Works, annealing, 399 Yield point, 4 FOURTEEN DAY USE RETURN TO DESK FROM WHICH BORROWED This book is due on the last date stamped below, or on the date to which renewed. Renewed books are subject to immediate recall. J9B95J ilVED JUL16'69-1 LOAN DEPT. LD 21-100m-2,'55 (B139s22)476 General Library University of California Berkeley THE UNIVERSITY OF CALIFORNIA LIBRARY