The Iron Age 1923-09-27: Vol 112 Iss 13

ane oe THE IRON AGE New York, September 27, 1923 ESTABLISHED 1855 VOL. 112, No. 13 Cause of Red Hardness of High-Speed Steel New Facts and Theories—Changes Due to Heat Treatment X-Ray Crystal Analysis—Slip Interference BY EDGAR C. BAIN AND DR. ZAY JEFFRIES HE high-speed steels considered in this article are essentially iron, tungsten, chromium and carbon alloys containing approximately: Per Cent NNN sg 6a ak wae bike ee eae Sete ee 18.0 COG 6 cA TN Rok wa eO at eee cae 4.0 CORR: icici joes teadtideeteune es 0.7 They are called high-speed steels because they can be used as cutting tools at a speed which produces a temperature corresponding to a dull red heat. The amount of metal which can be removed from the work by these high-speed steels in a given time is, therefore, much greater than that which could be removed by car- Fg. (1300° Fig. 1—High-Speed Steel, Thoroughly Annealed, x 500 bon steel tools which soften at a temperature far below that admissible for high-speed steel. An average of 50 crucible high-speed steels shows the following composition: Per Cent ere Te 0.67 ee OE OES Re 0.24 BE. Kecincvisds ccavaanvacnachents 0.28 ee, Det CROs. «cane ncmcds amie 0.03 Phosphorus, less than............+.>. 0.03 COPE: 's. «.6 panne s tee eeusneeeneny 3.99 i. cas dt ot hace Wasa 18.36 Be a oe ee EES 0.96 Cobalt, practically nil. Nickel, practically nil. At the outset we may place manganese, silicon, va- 805 2——- ( Below )— High-Speed Steel, Quenched From 2400 Deg. Fahr. nadium, cobalt and nickel in a class of elements of sec- ondary importance. Sulphur and phosphorus seem to be somewhat less harmful than in carbon steels, but nevertheless, in good practice they are kept quite low. Nickel is conceded to be detrimental to high-speed steel, and it is, therefore, somewhat surprising that extraor- dinary results are often cited in the special high-speed steels containing 3 to 4 per cent of cobalt. Vanadium, aside from its cleansing action seems to impart a cer- tain additional hardness to high-speed steel and in recent years has been generally used in amounts of about 1 per cent. Although iron, tungsten and carbon in Oil, Fig. 3—High-Speed Steel, Quenched From 2400 Deg. Fahr. (1300° C) in Oil, x 1200 might seem to be the chief elements of a steel possess- ing high temperature hardness, no high-speed steel is on the market which does not contain a_ substantial amount of chromium. Chromium is, therefore, to be classed as one of the mbre important elements of high- speed steels. Steels exhibit good high-speed properties with the tungsten as low as 14 per cent. In European practice a tungsten content of over 22 per cent has been used. Perhaps over nine-tenths of the total tonnage of high- speed steel has a tungsten content between 16 and 20 per cent. These figures merely tend to emphasize the PSs eo pe cgi $ OO acelin tile a tat nS! a” — __ Masry | ya eee ee eee ape A 806 THE IRON AGE fact that the thermal and mechanical history of a bar of steel is so much more significant than its composi- tion (within certain limits) that the effect of small variations in composition is relatively unimportant. Molybdenum can be substituted for tungsten, but so far it does not seem to have complied with all of the necessary factors for commercial success. The same general procedure in melting, easting, éogging and rolling which produces a fine grained, clean and homogenéous steel of any other class must be ap- plied in the production of high-speed steel. The ingot shows the usual columnar structure with carbide en- velopes about the grains, so there must be sufficient mechanical working to refine the cast structure. After mechanical working and annealing, the microscope re- veals a matrix of a sort of finely spheroidized pearlite made up of ferrite and globular, particles of carbide. The carbide particles seem to vary greatly in size. A small percentage of the particles are large and give evidence of being primary carbide particles, whereas the smaller particles are apparently of secondary origin. That is, the large fragments have never been entirely in solid solution in the matrix at any stage of the previous working and heat treatment, whereas the smaller particles at some previous stage have’ been in solution in the matrix and have been precipitated dur- ing the anneal. This structure is shown in Fig. 1. There is no difficulty in differentiating, at least in ex- treme cases, between the primary and secondary car- bides. This annealed high-speed steel is relatively soft and malleable. Most high-speed steel is delivered from the mills in this condition. Changes Caused by Heat Treatment In order to form a comprehensive idea of the changes. taking place in high-speed steels during the various steps in heat treatment, the authors have con- sidered the published work on microstructure, hardness, expansion and shrinkage, magnetic measurements, chemical behavior with reagents used as electrolytes, etching characteristics, and the spontaneous genera- tion of heat at room temperature after quenching. The authors have supplemented the above information by the study of the crystal structure of the constituents of high-speed steel as revealed by X-ray crystallograms. The results of the X-ray examinations of high-speed steel are published here for the first time, although one of the authors has previously given a resumé of the information in lectures. It is the purpose of the authors to set forth briefly the outstanding facts derived from the various methods of experimentation and to offer an interpretation of the cause of the high temperature hardness of high-speed steel. The authors have approached the problem from the standpoint of the basic and fundamental cause of hardness—slip interference’. In other words, high- speed steel is hard at a red heat because the slip inter- ference is effective at a red heat. To discover possible sources of slip interference which persist at elevated temperature, let us consider the various properties as modified by heat treatment. Microstructure When high-speed steel of the average composition considered above is quenched from a temperature near that of incipient fusion, we obtain the structure as shown in Fig. 2. At low magnifications the polyhedral grains are almost structureless. Some undissolved carbide particles are seen to be present. Most of the carbide shown in Fig. 1 would be dissolved at a tem- perature near incipient fusion. The carbide particles shown in Fig. 2 are, therefore, residual particles. After re-annealing, these particles would be designated as primary in contradistinction to the smaller secondary particles. At higher magnifications there is visible a September 27, 192° discernible structure within the polyhedral grains. W, now know these grains to consist of a mixture of ma: tensite and austenite. As the quenching temperatur. is lowered these grains etch more readily and th structure more nearly resembles that of carbon stee! martensite. Upon quenching below 1800 deg. Fahr (980 deg. C.) there is no microscopic evidence of th. retention of austenite. The matrix etches so rapidly as to suggest the presence of troostite or sorbite. In terms of the metallography of ordinary steel, high-speed steel at a temperature just above the A...-, transformation is markedly hyper-eutectoid, that is, it contains a large excess of carbide. This carbide is more soluble in the matrix, the higher the tempera- ture. All of the carbide does not go into solution in thé matrix at any temperature at which the steel is com- pletely solid. The higher quenching temperatures are, therefore, conducive to the solution of the maximum amount of carbide and hence to the retention of a min- imum amount of residual carbide. The matrix con- tains more carbon and tungsten and chromium, the higher the quenching temperature. The quenched matrix, therefore, has different properties in accord- ance with its composition at the time of quenching. Apparently a high carbon-tungsten-chromium content of the matrix is conducive to the retention of austenite by quenching. Fig. 3, magnified 1200 diameters, shows the microstructure of the high-speed steel quenched from near incipient fusion. This steel is known to con- tain both martensite and austenite. Photomicrographs of high-speed steel are unsatisfactory for even quali- tatively determining austenite. Sauver’ states that high-speed steel becomes much more resistant to etching the higher the quenching temperature. In one case an annealed steel was suffi- ciently etched in two minutes, whereas, after a high temperature heat treatment,.an immersion of 14 min. in the etching solution was required. It is stated that the rate of etching in certain reagents can be used as an indication of the previous heat treatment. The car- bides of either annealed or quenched high-speed steel are darkened by immersion in a solution of hydrogen peroxide and sodium hydrate. Fe,C (ordinary cement- ite) is not darkened by this solution. This offers chem- ical evidence that the visible carbide in high-speed steel] is not cementite. The development of a structure in high-speed steel, corresponding to globular pearlite in ordinary carbon steel, requires a temperature of about 1450 deg. Fahr. (790 deg. C.). In other words, it re- quires a higher annealing temperature to produce car- bide particles of a given size in high-speed steel than in carbon steel. Hardness Measurements of hardness showing the effect of reheating temperature on specimens quenched at va- rious temperatures have been made by a number of in- vestigators. Fig. 4, taken from Scotts® work, gives the general information on these points. The increase in hardness observed after tempering at 1100 deg. Fahr. (595 deg. C.) on specimens quenched from 2100 deg. Fahr. (1150 deg. C.) and 2280 deg. Fahr. (1250 deg C.) is known as secondary hardness. It appears that very little secondary hardness is obtainable with a quench below about 1900 deg. Fahr. (1040 deg. C.). The po- tentiality for the development of secondary hardness increases with the quenching temperature above about 1900 deg. Fahr. (1040 deg. C.). It will be noted that there is a decrease in hardness on tempering at tem- peratures up to 950 deg. Fahr. (540 deg. C.). Recent and careful hardness measurements indicate that the hard- ness of quenched high-speed steel is reduced by temper- ing temperatures up to about 840 deg. Fahr. (450 deg. C.). If the quenching temperature is below about 1900 deg. Fahr. (1040 deg. C.) the hardness may be still lower Catteni ee “= ated Ec a September 27, 1923 on tempering at 1100 deg. Fahr. (595 deg. C.) but it will probably not show a smooth curve through the tempering range 840 deg. Fahr. (450 deg. C.) to above 1100 deg. Fahr. (595 deg. C.). In other words, there will either be a halt in the rate of hardness reduction at a tempering temperature near 1100 deg Fahr. (595 deg. C.) or an increase in hardness depending upon the previous heat treatment. ——a | BRINELL Ba scciendlbcacnlabicncoesiai sol | | | it | 2 02 2 \ —_—— —_—_—s — Oo oe — r 379207 _ wt My ~ 7742? wa. ‘\ NOL ; | _ ss am VV . — T —, | | | | * 90+ - | “E 80 a | S 10] lesienslbrseiiibeiadtonteae | | a “| C2065 W217.5 Cr=3.4, V=Q73 ® Quenced in Oil fromm | | | wo 60 eg 1142 °F — nee eee eee EN *. e 2/02 °F | | | | WY | 4 2282°F | | | | 50}+_+—_|__— babendil Pog Ae; Bol | | | | BO med declan —1___t__| at; —t— 4+ V 100 400 600 800 1000 1200 Tempering Temperature Fig. 4—Relation of Hardness in Quenched High-Speed Steel to Drawing Temperature.—Scott If the quenching temperature is slightly below about 1900 deg. Fahr. (1150 deg. C) there will be a local rise in the hardness-tempering curve correspond- ing to the secondary hardness, but the value obtained by tempering at 1100 deg. Fahr. (595 deg. C.) may not be so high as that obtained before tempering. With quenching temperatures above about 1900 deg. Fahr. (1150 deg. As C.) the hardness after secondary hardening may be greater than the hardness before tempering. The 0.0010 secondary hardening effect reaches a maximum with short time of tempering at about 1100 deg. Fahr. (595 deg. C.). It should be mentioned that the decrease in hardness occasioned by tempering up to 840 deg. Fahr. (450 deg. C.) is not always grad- ual, and often a small hump in the curve toward a maximum has been observed at about 575 deg. Fahr. (300 deg. C.). This hump is strongly marked in the special high-speed steels containing cobalt and is encountered in high chromium steels containing no tungsten, there being, of course, nO maximum at the higher temperature in this latter type of steel. This slight discontinuity in the hardness-tempering relations is somewhat more pro- nounced after quenching from relatively lower tem- peratures, Expansion and Shrinkage Throughout this discussion, the authors have delib- erately avoided the mention of time in any of the heat- treating schedules; this has been done in the interests of simplicity, and because of the variations in indi- f) } 2)Or mA 7A ta j Juenched at300°F 500 00 00 Ann THE IRON AGE 807 vidual practiees, but it is safe to assume that 30 min. at drawing temperature would duplicate the results of the averages recorded by the authors. Interesting results on the linear change of dimen- sions of high-speed steel under various conditions have been reported by Grossmann‘. The results are in quali- tative accord with those published by Edwards’, and by Gill and Bowman’. If the linear dimensions are plotted directly, or density inversely, with drawing temperature the curves very much resemble the hardness tempering curves, as may be seen in Fig. 5. The steel expands on quench- ing, the expansion being greater the 9 « higher the quenching temperature ~ up to 2300 deg. Fahr. (1250 deg. C.). With successively higher tempering \ _| lto9 = temperatures the steels contract and \\ == the density increases up to about a c 800 deg. Fahr. (425 deg. C.). Be- | 400 5 yond this temperature expansion sets in, the volume reaching a sharp local maximum at about 1100 deg. —300 Fahr. (595 deg. C.). Above 1100 deg. Fahr. (595 deg. C.) the density again increases. A final draw at about 1500 deg. Fahr. (815 deg. C.) restores the approximate original density of the annealed bar. Magnetic Measurements Honda’ has made preliminary studies on the magnetism-tempera- ture relations of high-speed steel. 1400 He has determined the presence of cementite in high-speed steel by means of the reversible magnetic transformation which occurs in cementite at about 430 deg. Fahr. (215 deg. C.). He has also noted an additional rever- sible transformation at about 750 deg. Fahr. (400 deg. C.) which he attributes to the double carbide of iron and tungsten. This latter transformation is weaker than the cementite transformation and practically dis- appears with certain heat treatments. Draw? Fig. 5—Change in Dimensions of High-Speed Steel with Drawing Temperatures.—Grossmann With reference to experiments on the magnetization of high-speed steel it has been observed" that in steels quenched from a high temperature the maximum in- duction drops below 4000 gausses. Reheating increases the value slowly until a temperature of 930 deg. Fahr. (500 deg. C.) is reached. Tempering above this point rapidly restores the magnetization to the value found for thoroughly annealed steel. Electrolytic Corrosion When a completely annealed specimen of high-speed steel is made annode in a dilute acid solution nearly all of the ferrite can be removed without dissolving the carbide particles. One of the specimens treated by one of the authors retained the original shape of the 808 THE IRON AGE specimen until nearly three-quarters of its weight had been removed. The residual carbide falls to powder when thoroughly dried. Great care is necessary to wash the carbide completely and then dry it without atmospheric attack. This powdered carbide was first prepared to obtain a sample for the X-ray examination. A sample of this carbide was analyzed and gave results shown in the first column of the following table: Analysis of carbide Analysis of carbide Residue by Smith ® Residue by Arnold ~ 9 929 9 2 Carbon ; ‘ 2.5 2.33 eee _ 28.1 23.96 Chromium ; 6.5 3.13 Tungsten ; 60.1 64.24 Vanadium . ‘ 2.8 4.81 An analysis of a similar residue has been reported by Arnold and is given in the second column. Both of the above samples were obtained from annealed high- speed steels. In the course of the decomposition of the steel in the dilute hydrochloride acid the tungsten, if in a con- dition to be attacked, is converted to insoluble tungstic acid. Sometimes the precipitation does not seem to take place at once, but instead, a colloidal solution of the tungsten results. However, this is not persistent, and the final filtrate is free from tungsten. This yellow precipitate is, of course, mingled with carbides and can F t The Crystallogram Lines of the Three Principal Constituents of High-Speed Steel be separated by solution in ammonia. It is very in- teresting to note that the carbides from thoroughly annealed high-speed steel show only a very slight (about 2 per cent) loss by this ammonia treatment. On the other hand, carbide residue from freshly quenched high-speed steel, shows a different result. In this case as much as 33 per cent of the carbide residue mixture is soluble in ammonia, indicating that about 26 per cent of all the tungsten in the steel had been converted to tungstic acid by the electrolytic decom- positior ; Tests show that the carbides are at most only very slightly attacked by dilute hydrochloric acid, hence we may be reasonably sure that nearly all of the tungsten is located in the carbide phase after annealing and that much of the tungsten is in another form in the freshly quenched or slightly drawn condition. The inference is very strong that the tungsten must be atomically dis- persed, or at least in an extremely fine state of aggre- gation, in order to form the tungstic acid in the hydro- chloric acid solution. Spontaneous Generation of Heat in Quenched Condition C. F. Brush” has water-quenched high-speed steel from a temperature not far below fusion and within an hour after the quenching has observed a tempera- ture change in the steel when placed in a sensitive calorimeter. He performed many similar experiments with high carbon tool steels and found a generation of heat which persists to a measurable extent after sev- eral weeks. There was a contraction accompanying the heat generation in the tool steel. It is assumed that a contraction would take place also in the high- speed steel at room temperature. Dr. Brush remarks: “It is seen that heat generation in the tungsten steel is the same in character as in the carbon steel although much less in amount and somewhat more per- September 27, 1923 sistent.” The authors would infer that the spontaneous shrinkage would be less than in the carbon steel also. X-ray Crystal Analysis of High-Speed Steel By means of crystal analysis” carried out along the lines of the Hull method, it has been shown that iron having the body-centered cubic arrangement of atoms at low temperature undergoes an atomic rearrange- ment at the critical range and changes to the face- centered cubic crystal lattice. The patterns of the dif- fraction lines recorded on a photographic film are very distinctive for these two crystal types. The changes from austenite (face-centered cubic atomic arrange- ment) to ferrite (body-centered cubic atomic arrange- ment) may be very easily followed. In the case of car- bon steel, the cementite present does not yield a pat- tern strong enough to reveal its presence or disappear- ance with either gamma iron or alpha iron as a matrix. However, the carbide of high-speed steel contains a high percentage of the heavy tungsten atoms and is, therefore, capable of producing X-ray patterns at least when present in considerable amounts, and in particles of a size corresponding to about 50 or 75 atomic layers (0.0000015 cm.). Annealed high-speed steel as shown in Fig. 1 shows a strong ferrite pattern and slightly weaker carbide pattern. The carbide pattern is the same, however, as that obtained by the use of isolated carbide par- ticles obtained from electrolytic decomposition. When the steel is quenched from successively higher tempera- tures, beginning at about 1470 deg. Fahr. (800 deg. C.) the intensity of the carbide pattern gradually decreases. At about 2350 deg. Fahr. (1290 deg. C.) the carbide pattern becomes so faint that it is scarcely discernible. When the steel is quenched from 1900 deg. Fahr. (1040 deg. C.) a faint face-centered cubic crystal lattice is re- vealed proving the retention of austenite. The inten- sity of the austenite lines in the crystallogram in- creases and the intensity of the ferrite lines decreases as the quenching temperature is raised above 1900 deg. Fahr. (1040 deg. C.). High-speed steel quenched from about 2350 deg. Fahr. (1290 deg. C.) shows both fer- rite and austenite patterns (as suspected by Scott) well developed with the ferrite pattern being only slightly more intense. The austenite retained on the quench is very resis- tant toward martensitic transformation. A specimen was repeatedly immersed in liquid air without causing the austenite to transform to martensite. Twenty cycles of temperature change from room temperature to —310 deg. Fahr (—190 deg. C.) failed to have any observable effect on the austenite content of the steel. A sample quenched from 2350 deg. Fahr. (1290 deg. C.) shows such a faint carbide pattern that it may be regarded as essentially a ferrite-austenite combina- tion, the ferrite in this case being, of course, of marten- sitic nature. Furthermore, when quenched steel is re- heated about 16 hr. at 950 deg. Fahr. (515 deg. C.) the austenite changes to martensite and the carbide particles have not yet reached a sufficient combination of size and number to produce an effective X-ray pat- tern. A steel so treated shows essentially a ferrite pattern. Heating for a moderate time at 1200 deg. Fahr. (650 deg. C.) increases the intensity of the car- bide pattern. The pattern lines as recorded on the crystallogram for the three principal constituents are shown in Fig. 6. In spite of the fact that the literature is full of re- ports indicating the existence of several carbides, the authors have found only one carbide pattern in an- nealed high-speed steel. This pattern is too compli- cated to decipher at the moment in order to construct the space lattice, but it is distinctive and unlike Fe,C, W:C or FeW. The X-ray results indicate that the stable carbide is the iron-tungsten carbide which no doubt is capable of retaining varying amounts of chro- Clo hs bo eal ee a ela a September 27, 1923 mium and vanadium without changing the type of crys- talline lattice. General Considerations and Conclusions From all of the above facts, the authors have ar- rived at certain conclusions regarding the changes tak- ing place in high-speed steel with various heat treat- ments and offer an interpretation as to the cause of the red hardness of high-speed steel. Many of the statements enumerated below are well known and cor- roborated by many investigators, while others are be- lieved to be new: 1. Ordinary. high-speed steel contains sufficient carbon, tungsten, and chromium to be classed as a markedly hyper- eutectoid steel. There is a considerable amount of excess carbide at a temperature just above the Ag-2-; ' transformation. 5 I500r This fact is demonstrated clearly =: nT Ferrite by microscopic examination and a 1400 - ‘ is typically shown in Fig. 1. so ‘ While contemplating the behavior = 1300 of high-speed steel, it may be well ee 1200 to bear in mind that the metallic c - atoms of the elements are present = 1100 in the ratio of about 140 iron, Co 10 tungsten, 10 chromium and vanadium together and 6 carbon. That is to say, in an entirely representative cube of 10 atoms on the cube edge, we would have about 840 iron atoms, 60 tungsten i 100 atoms, 60 chromium and vana- S dium atoms and 36 carbon atoms. o 2. The excess carbide is more - soluble in the austenitic matrix, — the higher the temperature. The greatly reduced quantity of car- bide shown in Fig. 2 gives suffi- cient evidence of this point. 3. The amount of alloying 3 elements in ordinary high-speed « steel is sufficient to form more = carbide than is soluble even at = incipient fusion of the steel. This = fact is demonstrated by micro- ‘ scopic examination after quench- — ic ing high-speed steel from incipient ee 0 fusion. = 4 4. At temperatures appreci- @ 7 ably above the lower critical point me 12 only one crystal structure in the carbide is present. This carbide is essentially an _ iron-tungsten carbide, but no doubt contains varying amounts of vanadium and chromium with the reten- tion of the same type of crystal structure. This ability to retain crystalline structure intact with change in composition iS an outstanding property of intermetallic compounds. It is, of course, conceivable that with certain chemical composition the carbon and tungsten would be present in such propor- tions that either of them might be substantially used up in the formation of the iron-tungsten carbide and leave an excess of the other element. In such a case an additional compound would be expected to form when conditions per- mitted. It is also submitted that the iron-tungsten carbide is the most stable carbide in high-speed steels above about 1100 deg. Fahr. (595 deg. C.). 5 As the iron-tungsten carbide dissolves in austenite at high temperatures, the constituents of the carbide, no doubt, are dissolved in the austenite as individual atoms. The evidence against the existence of the intermetaliic com- pound as such in solution in a space lattice has been con- sidered at length in the literature cited. Before the quench the high-speed steel consists of some residual, undissolved iron-tungsten carbide embedded in a matrix of austenite which contains a considerable proportion of the tungsten, carbon, chromium and vanadium in solid solution 6. The quenching of the high-speed steel leaves the residual carbide particles unchanged and produces varying amounts of martensite in accordance with the quenching temperature and the remainder persists as austenite. Fresh- ly quenched high-speed steel, therefore, consists of. a small amount of iron-tungsten carbide and the remainder is com- posed of martensite and austenite in both of which the atoms other than iron are atomically dispersed. Grain growth in the austenitic matrix of high-speed steel at the quenching temperature is profoundly discouraged by the presence of the residual carbide particles. 7. Some of the austenite which exists at the high tem- THE IRON AGE 809 perature changes during the quench to martensite. The amount of austenite whieh remains unchanged varies with the temperature of the quench, the higher the quench the greater the proportion of austenite. This is probably due in part to the greater solubility of carbon, tungsten and chromium in gamma iron at higher temperatures—these elements acting to restrain the conversion. 8. In the martensite portion of the freshly quenched material, the only atoms which can migrate easily at room temperature or a little higher are the carbon atoms In all probability the tungsten atoms are held in fixed positions in the iron lattice, and likely this is true also of chromium and vanadium Iron atoms are available on every hand for combination with these carbon atoms; hence, even at room temperature FesC begins to form. The heat of formation of FesC is probably the heat evolution observed by Brush. 9. At progressively higher temperatures the chromium and vanadium will be capable of diffusion in the ferrite space Carbides ¢ ontal roe HO ' Turas te > v Martensite > 4 ‘ LULIQST CH VV Fig. 7—The Relative Proportions of the Principal Constituents (Abcissae) of High- Speed Steel After Various Treatments (Ordinates) lattice and carbides will form. The crystalline structure of the ferrite becomes more perfect and the carbide particles grow by coalescence to a super-critical size so as to soften the martensite at 850 deg. Fahr. (450 deg. C.). The carbide precipitation in the order of availability of atoms accounts in a large measure for the softening, heat evolution, shrinkage and magnetic point. The order of atomic volumes from smallest to largest of the principal elements involved is—C, Fe, Cr, Va, W. 10. At about 850 deg. Fahr. (450 deg. C.) the larger tungsten atoms are capable of slight diffusion and at this temperature the formation of carbides will follow the order of carbide stability in preference to the sequence of atom availability. Iron-tungsten carbide is the most stable one and forms to the elimination of the earlier formed carbides. 11. The iron-tungsten carbide particles reach approxi mately the size for critical dispersion after a short reheat at 1100 deg. Fahr. (595 deg. C.). Within the grain growth temperature range carbide particles are present to obstruct the growth and hence the grains of the matrix are not per- mitted to grow excessively. It is the combination of re- tention of small grains and precipitation of the tungsten iron carbide particles in critical size at 1100 deg. Fahr. (595 deg. C.) which accounts for the red hardness in high-speed steel. 12. The austenite present after the quench may not undergo any change important to the physical properties of the steel up to the temperature of martensite transformation at from 850 deg. Fahr. to 1100 deg. Fahr. (450 deg. C. to 595 deg. C.). Some excess cementite (FesC) may precipitate from austenite below this temperature, but if so, we would regard it as only an incidental feature. 13. Between 850 deg. Fahr. and 1100 deg. Fahr. (450 deg. C. and 595 deg. C.) the austenite transforms into martensite with expansion and increase in hardness. At 810 THE IRON AGE this temperature the iron-tungsten carbide can form, and does form, in critical dispersion. Its presence helps to keep the grain size of the new martensite small and to key the slip-planes of the ferrite grains 14 As the temperature is raised above 1100 deg Fahr (595 deg. C.) grain growth of ferrite and particle growth of the iron-tungsten carbide produce rapid softening similar the corresponding process in carbon steel at lower tem- peratures Fig. 7 is a chart which is intended to show the hange in relative proportion (abcissae) of the constituents rf gh speed steel with heat treatment (ordinates) It should be noted that the cementite, FesC, is included in the martensite (ferrite) area and that nly the change in haracterist stable ron-tungsten carbide S i 1 The double cart iins its a s in a relatively dense state When double carbide dissolves in austenite the density f the steel is lowered and when double carbide pitates, the density increases There is also the usual a eas I l€ Ss whe Ss inges ai SiL€ This is largely due to the wer density of alpha iron as mpared to gamma Cause of Red Hardness Briefly Stated ] The cause of the red hardness of high-speed steel might be briefly stated as follows: The changes which cause martensite of carbon steel to soften are grain growth of the ferrite and growth of the carbide par- ticles above critical size. Similar changes in high- speed steel take place only at a red heat. The out- standing reasons for the retention at red heat of fine grains in the ferrite of high-speed steel are the in- creased resistance to growth due to the elements in atomic dispersion in the ferrite and the copious pres- ence of obstructing carbide particles. The reason for retention at red heat of carbide particles of critical size is the great stability of the iron-tungsten carbide and the large size of the tungsten atom. The great stabil- ity of this double carbide forces its formation to the exclusion or elimination of other carbides when the September 27, 1923 necessary atoms are available. The large size of the tungsten atom prevents its diffusion in the ferrite space lattice until a temperature corresponding to a red heat is reached. The double-carbide is an intermetallic com- pound which owes its existence entirely to crystalliza- tion. The formation of a particle of this carbide, there- fore, requires a number of tungsten atoms which must be supplied by diffusion through the ferrite lattice. The precipitation and growth of the double carbide in quenched high-speed steel at a dull red heat is, there- fore, somewhat comparable to the precipitation and growth of cementite in quenched carbon steel below 300 deg. C. REFERENCES 1 “The Slip Interference Theory of the Hardening of Metals.” Zay Jeffries and R. S. Archer. Chem. € Met. Eng., June 15, 1921. Vol. 24, No. 24, p. 1057. 2 “The Metallography and Heat Treatment of Iron and Steel.” Albert Sauveur, p. 361. 3. “High Temperature Treatment of High-Speed Steel.” Howard Scott. Transactions of the American Society for Steel Treating. Vol. 1, No. 9, p. 520. 4. “The Change in Dimensions of High-Speed Steel in Heat Treatment.” Marcus A. Grossmann. Transactions of the American Society for Steel Treating. Vol. 2, No. 6, May 1922, p. 691. 5. “The Physico-Chemical Properties of Steel.” C. A. Edwards, Chas. Griffin & Co., 1920, p. 216. 6. “The Metallurgy of High-Speed Steel.” J. P. Gill and L. D. Bowman. Transactions of the American Society for Steel Treating. Vol. 2, No. 3, Dec. 1921, p. 192. 7. “On Magnetic Analysis as a Means of Studying the Structure of Iron and Steel.” Honda, Jour. Iron and Steel Institute, Vol. 98, No. 2, 1918, p. 375. 8. Howard Scott. Loc. cit. 9. Analysis furnished by Oscar R. Smith, chief chemist, Atlas Steel Corporation, in connection with unpublished work of Grossmann and Bain. 10. C. F. Brush, “Some Thermal Relations in the Treat- ment of Steel.” Trans. A. I. M, E. Pyrometry Volume, 1920, p. 590. 11 “Studies of Crystal Structure with X-rays.” Edgar C. Bain. Chem. & Met. Eng. Oct. 5, 1921, p. 663. Dr. Arne Westgren independently obtained the same information in Sweden. “X-ray Data on Martensite Formed Spontaneously from Austenite.” Edgar C. Bain, Chem. € Met. Eng. Mar. 22, 1922, p. 543. Heat-Treated Steel Parts in Service Effect of Long-Time Drawing Conditions or Further Heat Treatment Under Use—Stability of a Steel BY GEORGE K. ELLIOTT* S commonly understood, heat treating consists of A\ passing a metal or alloy through one or more changes of temperature for the purpose of impart- ing to it definite physical properties. Applied to steel it generally involves raising the temperature of the steel to some point above its critical temperature and cool- ing in some way so as to reach eventually a room tem- perature. If the cooling be conducted slowly, the process is known as annealing; if rapidly, it is called hardening, the degree of hardening being determined largely by the amount of carbon in the steel and the speed of cooling. All this is a matter of elementary knowledge to heat treaters. Hardening usually is paralleled by an increase in strength, elastic limit and brittleness, which last prop- erty customarily calls for a supplementary treatment in the form of tempering or heating to a relatively low temperature and cooling, for the purpose of de- creasing it to a safe degree. Hardened Steel Is Unstable The writer wishes to dwell upon the fact that steel in the hardened condition is often unstable and readily slips back to some form of greater stability whenever its existing temporary equilibrium is disturbed by an increase in temperature. Especially is this true of carbon steels and of alloy steels of a relatively mild nature which may be classed as “near” carbon steels. *Metallurgical engineer, Lunkenheimer Co., Cincinnati Ohio ; Rudimentary as this fact is to the metallurgist, it has been found to lie beyond the pale of knowledge of not a few engineers who design mechanical devices and specify materials for them. They overlook the fact that under certain working conditions a treated steel may automatically undergo further heat treat- ment in service with possible harmful results. Here reference is made to modern industries which are adopting a number of processes and practices which demand higher temperatures than formerly could sanely be even hinted at; frequently these in addition involve rather severe working stresses. In order to meet these stresses we occasionally find engineers specifying steels and strengths in the steel that are at- tainable only through a hardening heat treatment. Attention should be called to the fact that such treat- ment is liable to leave the steel in a more or less un- stable condition. If a steel thus treated be exposed for long periods of time to temperatures of 600, 300 or 1000 deg. Fahr., there is a dangerous possibility that it may take advantage of the circumstance and recede to a more stable state with a corresponding de- crease in physical properties. Therefore it would seem safer for the engineer to figure on using his steel for high temperature, in its most stable form, the form least susceptible to thermal influences. In order to accomplish this at least two roads lie before him—either he can stick to the carbon steel with which he is best acquainted, have it in a thor- ougly annealed or normalized condition, and design 9 San ttle ee i a September 27, 1923 the steel parts accordingly heavier; or if he desires to reduce to a minimum the weight of his structure, he must look for some alloy steel which, in its most stable state, will have the needed physical properties. A piece of steel that has been heated above its critical point, quenched and drawn at perhaps 800 deg. Fahr. may seem to the casual inquirer to be perma- nent enough for service under stress and at 800 deg. temperature, but if it be considered that the service perhaps will extend through months and even years there arises a grave doubt that the cumulative effect of time and temperature will not sooner or later effect a “drawing” of the steel which will be much more drastic than would be deduced from the temperature alone. Demand for Steel Used at High Temperatures The petroleum industry, ultra-modern high pres- sure and temperature power plants, and several chemi- cal processes are creating a demand, limited as yet but growing, for steel parts which will withstand high pressures at temperatures of 600, 800, 1000 deg. Fahr. and perhaps higher, and the metallurgical engineer must heed very carefully the material that is supplied to these wants. Especially does it seem important that the steel used be of such composition and in such state as to be of maximum stability under the thermal con- ditions of the service. It seems highly desirable that extensive investiga- tions be made concerning the true effect of long-time THE IRON AGE 811 drawing at moderate temperatures of carbon and alloy steels which have first been Su™#mitted to a variety of heat treatments. These tests’ Would involve exposing the test samples for many m s at various indus- trially important temperatures a noting’ the effect upon the physical properties igfexampe, one series of tests might be conducted™at 400 deg. Fahr., and others at 600, 800 and 1000 deg. At higher than 1000 deg. the drawing completes itself much more rapidly so that the results either are already known or are more readily predicted. It is now known that drawing a steel at 350 deg. for a long time will give approximately if not exactly the same results as drawing at 450 deg. for a much shorter time, but metallurgical literature does not yield much information as to whether drawing at 350 deg. can be prolonged sufficiently to attain results equal to drawing at 800 or 1000 deg. for a short time. In other words there is need of knowledge as to how far various commercial steels and others can vroceed toward greater stability when exposed to any of the moderate temperatures just mentioned and how the physical properties are affected by the stabilizing. The question of stability at these temperatures also embraces a study of alloy steels. In fact there seems to lie the greatest promise of a solution because it is well known that many of the alloying elements such as nickel, chromium, manganese, etc., have an effect upon the thermal properties of steel which serves toward greater stability. Some Causes for File Steel Failures Suggested Means for Prevention—Conditions Prevailing in the Rolling, Forging and Heat Treatment of the Steel BY ARTHUR W. F. GREEN* open-hearth furnaces. The relative merits of the products of these three steel-making processes are not to be discussed. The purpose of this paper is to describe briefly and illustrate some of the causes for failure in files made from carbon steel, and to suggest possible means for their prevention. File steels made from carbon steel have a carbon content ranging from 1.10 to 1.40 per cent, the average being about 1.35 per cent. File steel as cast has a struc- ture as shown in photomicrograph Fig. 1, which shows plainly the free cementite both as the network around the grains and as needles within the grains. (All photo- micrographs are X150 and etched in 10 per cent nitric acid.) Steel containing so high a percentage of carbon is not a particularly hard material to make in no mat- ter what kind of furnace employed, but, because of that fact, it sometimes happens that usual care is not exer- cised. A wild heat or one that is not correctly handled may cause material to be made that will not be uni- “wee The John Illingworth Steel Co., Philadel- phia. | JILE steels are made in the crucible, electric and form, in that there will be areas of carbon segregation with the result that files made from the steel will not be uniformly hard. Some idea of the way in which carbon segregates can be seen in photomicrograph, Fig. 2. This shows the carbon segregation around a blow hole in a large ingot of crucible steel. The heavy white network being free cementite. Photomicrographs, Figs. 3 and 4, show two sections taken from the same file, which failed to pass inspection because of lack of uniform hardness. In Fig. 3 a hard spot is shown in the tooth, while in Fig. 4 the tooth was soft as shown by the troostitic structure, and the lack of free cementitic globules. It will also be noted that the top of the tooth shown in Fig. 4 is burred over. This occurred when the file was tested. The failure was due primarily to carbon segre- gation. Another cause for failure of files due to improper melting conditions is slag inclusions. Defects of that kind may cause soft spots, and very often cracking dur- ing quenching for hardening. Photomicrograph, Fig. 5, shows a slag inclusion in a file which broke in harden- 812 THE IRON AGE September 27, 1922 The rolling of file steel into the various commercial shapes does not give rise to any unusual conditions, and with proper care in heating, coupled with correct roll design it is possible to produce rolled flats or tapers that will meet the most exacting needs. However, there is one problem that faces all file steel manufacturers, and that is decarburization of the steel. Because of the high carbon of the material, this is an easy condition to produce, especially where little attention is given to furnace atmospheres, etc. Photomicrographs, Figs. 6, 7 and 8, show some ( phases of decarburization of file steel. Figs. 6 and 7 Oo is were made from the same piece of steel. In Fig. 6 is shown the surface of the piece. Practically no carbon remains on the surface. Going in from that point the carbon content gradually increases until a wide area is reached containing 0.8 per cent to 0.9 per cent carbon. This eutectoid area is several thousandths of an inch wide, and is distinguished by the lack of either free iron or cementite. The true condition of the steel is shown in Fig 7. File steel decarburized as deeply as that just described is useless, since it would be necessary to grind so great an amount from the surface to render the file usable that the cost of grinding alone would be prohibitive, Fig. 6 as well as making the piece too small. Photomicro graph, Fig. 8, shows the decarburized edge of a rolled file steel, which is not as bad as that shown previously. However, because of the width of the eutectoid area it would be difficult to produce good files without doing a great deal more grinding than usual. Photomicro graph, Fig. 9, was made from a bar of file steel show- ing normal decarburization from rolling. Since all file manufacturers count on grinding several thou- sandths of an inch from all file blanks before cutting and hardening, such decarburization is not serious. Mill decarburization can best be regulated by keep- ing as nearly as possible a reducing atmosphere in the furnace used for heating, and also by designing the Fig. 9 rolls so as to give as great a number of scaling passes as is possible. The mill in which the writer is em- ployed has been quite successful in combating decar- burization by installing oil-burning furnaces using low-sulphur oil, and by using specially designed rolls. It is not to be assumed from the foregoing that all the failures of file steel are to be laid at the door of the mill producing steel. Very often the cause or causes for failure of files occur in the file manufac- turers’ shops. Many files are doomed for failure and the manu- facturer’s scrap heap because of lack of control of the temperatures used for forging the tang and shaping a + spn oe ae Ly i September 27, 1923 THE IRON AGE 813 Fig. 13 Fig. 14 Fig. 15 the file. Most of that work is done on a piece work basis. Consequently, the workers very often feel that by using high temperatures they can produce more work, because it has been heated quickly. The results of too quick and too high heating may be seen by refer- ring to photomicrographs, Figs. 10 and 11, Fig. 10 showing the edge of the sample severely decarburized, and Fig. 11 the structure of the steel near the center of the piece. This is in contrast with the structure shown in photomicrograph, Fig. 12, taken from a sec- tion of the unforged end of the same piece of steel. Material so mistreated and then made up into files will result in the files being brittle and very often cracking Fig. 16 Fig during quenching. The structure of the forged end poorly worked and then hardened is shown in photo- micrographs, Figs. 13 and 14. The use of correct forging temperatures usually improves the structure of the part of the file worked. This is shown in photomicrographs, Figs. 15 and 16. It will be noted that the grain is smaller and more re- fined in the forged section shown in Fig 15 than in the unforged section shown in Fig. 16. The annealing of the file blanks after the tang has been forged and the pieces properly shaped by forging or rolling is a process which, in the writer’s experi- ence, is carried on in a different way in each shop. Some manufacturers anneal by packing the blanks in suitable boxes that can be readily sealed to keep out air and furnace gases. Charcoal or other carbonaceous material is sometimes added to the charge in the box to guard against decarburization. Some anneal suc- cessfully by tying a number of blanks into bundles and placing a number of bundles in the furnace at the same time. Others merely pack the furnace with loose blanks. It has been found that wood, coke, oil, coal, gas and electricity are used for fuel for the annealing furnaces. One shop using a method of packing the blanks in the furnace without protection, experienced great dif- ficulty in producing good files, and examination dis- 17 F'g. 18 closed that those files which had formed the upper layer and sides of the bundle had been severely decar- burized. A section from one of the blanks is shown in photomicrograph, Fig. 17. A great deal of grinding was done, but even then the files would come from the hardening with soft spots. This can be readily under- stood by noting the way in which the free cementite had segregated to the grain boundaries while the amount of spheroidized cementite within the grains gradually decreased and only normal pearlite re- mained. Photomicrograph, Fig. 18, shows the decarburized edge on a file blank after having been ground in the usual way. This blank had been annealed in an old Fig. 19 Fig. 20 Fig. 21 a ~_— 814 THE IRON September 27, 1923 AGE Fig. 23—Quenchéd From 1400 Deg Fig. 24—Quenched From 1420 Deg Fahr. Fahr. wood-burning furnace with little care being taken to trude above the bridge-wall in side-fired furnaces, or protect the material from excessive oxidation. It has been found that the most uniform results in annealing have been obtained in those shops using boxes or other suitable containers, which had been carefully sealed, and in which charcoal or other car- bonaceous material had been used to pack around the in the box. Even with these precautions slight decarburization may shown in photomicro- Fig. 19, showing a section taken from a blank which had been carefully ground before having been This small amount is, however, not a serious blanks occur as graph, annealed. matter The fact that file steel blanks have been packed in yoxes and carefully sealed, and a pyrometer has been employed to indicate the temperature of the furnace used, is not a criterion that the steel will be properly annealed as is shown in photomicrographs, Figs. 20 and 21. These were taken from two separate’ pieces of steel, taken from different boxes which had been in the same furnace at the same time. The furnace employed was a side-fired oil-burning type. The sam- ple from which the photomicrograph was made was taken from a box which rested of the floor of the fur- nace. The other sample was in a box that protruded inches above the bridge-wall of the furnace where it had been exposed to a very hot flame. The pyrometer used showed that a temperature of 1440 Fahr. had been used.