_’—_— __4__— __'—_. _—__——. 0.4 pg .4 I (1301—3 SOME DILATOMETRIC OBSERVATIONS OF SEVERAL COMMERCIAL STEELS ' THESIS FOR THE DEGREE OF M. S. Cabs: M‘ Heath 11931 . .A l." ..‘(u? my. , . . v ._. . . 0‘. .x x . ..ufi. . . . D. .. . . . ..Q....m..s4....h 5)»..-th 1“ .‘J | \. .\ . v .. up A g \ i .17., .’....r.u..nn. .91 ff. 2.); I. , L... . K .....t .. . .. M‘ P“, o ..h..... «Emma ....T J... ‘5’". .OJNffi-Lw ‘ b . ‘I‘I.'_"'Vt ""1‘ A‘.I'm‘. )? ‘ fjé 3‘5, "“3“" Eta-:3. -:<‘._'.';_f :_. P. D $ I‘M". {A if .v“ 4': 1' 921%" 3'“ Y". 1." 1c 4'5" N V177)". 0‘. ._| z‘ _ ' 1;“; V win 0 | ’A'.‘ A, Zu‘.“,‘i .NrdC'.‘ H 7".“ ‘.A.‘.‘ \E , - . ‘ Jq.’ ' "v ‘ l l i L l a O- .I ’1} :fll ‘ ‘t'n‘éizf, £- ‘3‘? a,'. - n53” _ fl ' -( g . Out? 5 r:' . " w J", {'5 “We? it". ,;< x '., ..r‘q “liw‘ . 0351-“ fiteé“ ba;" I‘Tfi- &I 4 ‘I7 , n‘ (3“ v :n' w.” K. . p .‘Kv: I?“ 1:“ J {’ch T .“ J . 'ré§-"\.1'\‘ ’ ' ‘1' ‘4 . '-. Oi“ .. ' m'fzfif ‘ 4" MO ”SO? Er. E "2 an»; .5»; 95* fI- ’01 X‘ ‘ .O ..g'h f ' I: ‘c OWE? n v ‘6‘! 266$? . I 89}: . J "r - . w,» _ - ' u-U . ._ w 9 3 .‘ .°,‘4‘5,'-"q1“. ' '31 SOME DILATOXETRIC OBSERVATIOXS OF SEVERAL COMMERCIAL STEELS Some Dilatometric Observations of Several Commercial Steels Thesis Submitted to the Faculty of Michigan State College In Partial Fulfillment of the Requirements for a Degree of Master of Science Carlos M. Heath —— June, 1931 T T... F ..' The writer wishes to express his appreciation and to thank Professor h. E. rublow for the valuable direction and assistance that ne nas generously given on this problem. . He is also grateful to Gordon Stunpf, of the R60 motor Car Company, for the machined and carburized samples in this work. 101182 A .3“:l.ll‘.bu.¥l ...? |t\,4.’l'..wfotobuna.hi.:.tifi ABSTRACT In order to understand sone of the reasons why there is shrinkage in case carburized stoca upon heating througn the Acl critical range, a general study has been made here of the dilatation of iron, S.A.E. 1020, 1015, and 4615 steels, both carburized and non— I carburized. A study of sone quenched samples was also made. All these dilatation curves were taken on the Chevenard Industrial Thermal Analyser. It has been found that all steels have a tendency - to shrink upon returning to room temperature after having been heated above the Acl range and slow cooled. On any of the above steels except iron and carburized stock, if all strains are removed by a slow heat, a soak, and a slow cool through the critical ranges, then the piece may be heated again, up and down , slowly through the critical range and return to its original size.. Any oiece that is under a mechanical strain or one due to heat treating strains will invariably decrease in size after heating through the critical range. A piece of steel that has been quenched will reduce in size, if it is only heated to 10000. Any steel, if it is in the pearlitic state seems to return to its original size, if it is not heated over the Acl range. Case carburized stock and low carbon iron will shrink no matter what condition it is in, if it is heated above the Acl range and slow cooled. Quenching of course will expand steel. If these two stocns are oeriectly annealed from above the criticalso as to remove work strains and any other strains present, the niece will not shrink nearly so much uoon reheating through the critical ranges. It is known in oracfice that case carburized stock tends to shrink and warp upon heating_through the critical ranses. The dilatation of two carburized . 1015 and 4615 were investigated to find f11 stocks, S.A. out what causcd the shrinkage and how to orevent it, if possible. A Chevenard Industrial Thermal Analyser was used for the purpose. The preliminary results showed that the carburized stock always shrank upon heating throuqh the critical range and slow cooling, or at least cooled slowly enough to keeo it in the oearlitic state. We not only found out that the carbur- ized stock shrank but also the non-carburized stock had tendency to shrink although not so marked and very erratic. This led to a general study of non-carburized stock together with a very low carbon, in order to explain the more conolicated dilatation of the case carburized stock. Fisure l. Chevenard Lilatoneter APPARATUS The Chevenard Industrial Therual Analyser used is a simple but delicate fort of dilatometer. It consists of a silica tube holding a sanple. Its expansion upon heatinais transmitted by a small silica rod to tne machine proper where it is enlarfed approximately seventy tines by a system of levers. This is recorded on a drum moved by clocK worn around a vertical axis. The tenperature is automatically recorded at the same time and on the same drum by the expansion oi the standard piece of non—corrosive alloy Known as pyros. Pyros has a fairly constant rate of expansion, at least to 100090. the limit of the machine. A photograWh of the instrument is shown in figure 1 with a furnace covering the silica tube. Figure 2 shows a side elevation of the apparatus without the furnace and with a cut in the silica tube to snow the piece inside. The steel sanple is a cylinder, 55 millimeters in length and sixteen millineters in dia- neter. The standard oyros is fifty millimeters in lennth and flour nillineters in diaxeter. This is placed within the cylindrical hole which runs through the steel sample horizontally or rather longitudinally. goth pieces rest firnly against the end of the tube. t. The expansion is transmitted fron the two pieces by Figure 2 twp NSGK V de QQVZM> NIH .. “.V/éV/V/V/fi/K/Z I. w 7 ,_.////..,//////////,n | silica rods to two brass rods running tnrough a bush- ne. These Ha ing in the cast iron framework of the nacn irods transnit the motion to two arms, whicn are sus- pended on points, by means of a moving contact. This motion is in turn transnitted to the pen arms also by sliding contacts. The mechanism is kept taut by two springs, as shown in the drawing. The novenent of these arms is measured by a circular scale which is calibrated in 10000. units on one side and expansion in 2 milli- 7 meters per millimeter tines lO-Jon the other. These units of expansion are so large that }2' of each space may also be measured down to 5 degree units. This apparatus has checked results very well measurenents taken on aninterfcroneter. An auxillary silica tube was used with the ap- paratus in IfllS experinent which was much easier to handle and was just as accurate in most respects. This other tube was made for unmachined specimens. The tube in the far end was cut away a little longer than the length of the specimen and to the axis of the tube. A silica platform supported the specimen and the pyros which are now placed side by side. The pyros used here is 50 millimeters long and has an 8 milli— meter square cross section. The sample is made sowe- where near the sane cross section and is also 55 milli- meters in length. A half circular cut out of the tube \xili lllllll is used to cover the piece. The specimens may be cooled more rapidly and nay be removed without first removing the silica tube. In either case oure powdered charcoal is inserted in the tube to Keep away oxidati>n, which would natur- ally introduce errors into the results, if the silica is eaten away by the iron oxide. The pieces are heated with a horizontal type, tube muffle, electric furnace. The current is controlled with a finely graded rheostat, so that the rate of heat may be controlled very well. A small amount of heat may be used while cooling, so that most any cooling rate may be obtained down to an air quench. A very fast heating rate may be obtained by heating the furnace up before putting the furnace over the silica tube. As stated before the apoaratus is fairly accurate but errors are introduced, if the proaer care is not used in placing the wedges behind the rods to adjust the height of the pen arms. Beside the deterioration of the silica tube there may also be a slight amount of slippage, if the piece is not placed properly. This results in one or both of the hands coming baca below normal on cooling. As stated before there are two curves on the drum graph. The upper is is ruled off in 5000. lines and the lower curve is the dilatation of the piece. The -5- dilatation readinrs are taken for every lOOOC. temoerature and points where there was an change in either curve. These points were plotted, dilatation on the ordinate and temperature on the abscissa. The heating curve is solid and cooling curve is dashed. All 1020 pieces are hot rolled, flat, bar stock. The iron is hot rolled iron thermocouple wire with a @arbon content of about .026 %. The S.A.E. 4615 and 1015 steel are hot and cold rolled rod stock, respect~ ively o S.A.E. 1090 Stock Plain eutectoid CQFbQJ steel is used as a standard for calibrating this machine. This is used as it has only one critical range to go thru or rather the majority of the piece has only one critical to go ,hrouxh, as ma" be seen in figure 3, of the equilibrium diagram of Iron and Iron Carbide. This steel may be heated up to the lower critical and will expand to a limit, as shown by the solid line of curve 3 in figure 4. At the lower critical the steel will change complete— ly over to gauna iron, and the cenentite going into solution without a change in tenoerature. This is not true with other steels as may be seen by the other curves on tie same sheet. upon cooling the eutectoid steel shrinks until it hits the A critical again. Tne only 1 i, chanfie here until the iron is changed to alpha is tn rise of about 5 or 10 degrees that tahes place due to the giving off of heat of transformation. The pyros being inside the steel and very snail, is very quicx to record this evolution of heat. This is one advant- age of the closed tube over the open one. It will be noticed that when the oiece returns to its original size, as this does, that the change in expansion at the criticals is about the sane for the change from alpha to gamma as it is from gahna to alohr. 0n cooling the curve returns practically back *0 the heating curve -3- Figure )0 “e Equililn'ium Diagram ' of Iron anel Iron Cal-Lille De Cbn 15:29 Fahrg‘ef‘zgif 1600 fle/fa [ran /7 A Mm Iron in Homer [moor 0 z 74.. — 2800 [.500 00 \ A/ 4 z —" 2700 We a,- z- 1 0 — 2500 0063/- ”)3 l (j ' j” . ”00 Magma-\- 2:: ' 7””J’j%\ ~2500 an . 1., ’ 2‘0 . 500 Ame/”[9 NUS/7y flage ——l———Jo/"-7’dz>< 0 3400 . AUSfE/Jffe if) Nofber ll‘qz/or \ a“? cows 2300 1200 \ f! _l 03;“ 2300 50/1'0' 5o/Ufl'on °o f Sol/dos, infect/c Freezes 5 Y) ”40 °_C 2/00 [/00 — Carbide in Gamma fron / —— 2000 (Al/5f ll 'z‘e) 6“ ' ’ [00 8 l $6: 60? C . f : I900 fiofif T Amie/life , Zea’ebur/‘fe and Q 9”]sz e _ 1500 G ‘0‘ ®Awlemfe Cemenfiz‘e 3‘ i’ a /700 9m x, b“: ‘ s. gledeborlre: e a and if a —— moo HO?” £4)? ‘76? ‘OCemenh'le If. i: 500 600 ta??? efilfifififtm-Arrin—d ‘ofng—néfée I'm-73‘”: BIZ/1822’??— FLA---.— ‘ £38? fin”; Teqmerafm 41,5 Alp/78 Ira/2 :: Gamma Iron fik : I400 700 -__—‘-—-§:15_74r,, 1300 ~11200 600 —-»//00 - ' 500 Pearl/'1‘ e Pear/ire Cemenh'fe’, Pear/ife ‘1/000 and and 3170’ — 900 err/Te 7° '\ Cane/7W? Transformed Zeo’ebz/rife __ 500 400 E «g a t ‘4 700 “s ‘3 300 -———'~ ‘& a 600 2 —— 500 A. [170’ of Na net/em in Cemenh'z‘e an Heal/'12 _, 200 _" ' —" ' -----._..__l£_-___-__-.---._.____.9 '_ 'L—" 400. 100 33 3% if 200 [V ' 6 I: V. —* IOU 0 u, 9 .q I e to. a. Lo 0. n a. I Q N N . ' (\z to m \r \r ‘n l *— Per Cem‘ Carbon —~ ' ire—m 4 — — ——5fee/3 ——————— T— ———————————— White Casf Irons -- —- -- - — -— — — —- ——’{ Delta Iron Regions according to Ruer, Ferrum, Vol. 11, 1914. Liquidus according to Ellis, Metals & Alloys, April, 1930. Solidus according to Kaya, Science Papers, Vol. 2, No. 4, Japanese Institute of Physical & Chemical Re- search. Top Quenching and Annealing Temperatures from recommended practices, National Metals Handbook, 1930 Edition. Reprinted from M eta! Progress, September, 1930. Burning line according to Jominy, 1929 Transac- tions, American Society for Steel Treating. As. Al‘s, A1, and Ar. according to Hoyt and Dowdell, “Metals and Common Alloys,” 1921. Line GPN according to Yensen, 1929 Transactions, American Institute of Mining & Metallurgical Engineers. Line SE according to Honda and Endo, 147th Re- port, Imperial Japanese Research Institute. Magnetic Lines according to Honda, 60th Report, Imperial Japanese Research Institute. Figure 4. by the time the expansion of gamma to alpha is complete. This 'ust be true, if the coefficient of expansion is the sane on heating and cooling, to have the piece return to its original size. S.A.E. 1015 & 1020 Hone of the annealed steels shrink below their oririnal size when heated anywhere below below the A01 critical. C‘Jnr‘ve l, fi:1_lr'e 5, SLiO"'."S +liL lie,"til ClirHJ-t: (solid line) and the cooling; curve (ohm—shed line) 0.3.x; ‘u . 1‘ I' . ~“ "" f‘ L . r r-v‘ “/‘. g x ‘5'! A15 \t 1’. J': ~ . ' ‘. ’17? '\‘ ‘ :lL all olefintha e-r-i:p oi earn; other. (juive;2?.i1 the sa“e sheet is the dilatation of a iece of S.A.E. 1320 '0‘ brake Crun stock run on the platform silica tube. - . ,1. . memos After cole QTESClflfi a prciialnary anneal of lsjd r- r- 1 a ‘ -1 l— O _ for four HCUPS. This was heateo to 930 C. ano slow coolec. This piece being only one fourth oearlite does not show much reaction at the lower critical rwn e. This nay ..., \ be followed also in the equilibrium diagram, in figure 9. As the heat reaches this ranxe fljlgIAV hanoens for awhile. The tezoerature continues to rise until there is enoush enervy built up to throw the iron in the pearlite into solution with the ceaentite, forming islands teel aac can in; the slight aaoun O) of austenite in the (I) of shrinkage. This accounts for the snail nob on the l O A 1 ‘ +- heoting curve at 740 0. From there the temperature continues to rise, the austenite continuing to absorb more ferrite to form more gamma iron, consequently more Figure 5. shrinks e. This takes place until all tne iron or ferrite is in solution in the austenite. Then the austenite on further heatiwe exuands as shown by the dilatation curve after 37093. is reached. The reverse takes place on cooling. Nothing happens when the temperature lowers to the theoretical, upper critical. The tenoerature lowers 7030. below the ACE, in this particular instance, before enough energy is built up to throw ferrite out of solution. It then throws out quite a quantity of ferrite, accounting for th decided rise in the curve. As the temperatures continue Ito lower the less ferrite can be held in solution accordin: to the equilibriua diagram and collaborated by the dilatation curve, which shows a continual expansion to alohairon as the tenperature lowers. when the lower critical ranne is again reached notming happens. The mass, being all chanred to aloha except the ootential pearlite, begins to shrink, and all at once there is the final change accounting for the austenite changing to pearlite. Then the metal is heating through the criticals all parts not changing from alpha to gamma or that have changed are expanding enen though the total mass is contracting. The curve reoroouced is the result- ant of all the reactions. The reverse is true wnen cooling through the criticals. -19- The third curve shows a similar niece or netal heated up in the sale manner. Curve 1, in figure 4, snows an S.A.L. 1015 steel heated up in the same tanner in a closed silica tube. This piece snows slight shrinkage, h wever. In any plain carbon steel tne ferrite throfin out of solution is not sure iron but at least a solid solution of iron and carbon in sone torn to about .OBfi. Upon lowering to roon tenoerature the carbon content lowers to .Oii. In nost steels tnis effect is not 3, noticeable but it has a :ore apprecieole effect on iron c" ."I containing in the neighborhood of .Oino, as will be shown. Iron Pure iron seems to snrink more than plain carbon steel for some reason or other. It isr't due to oxidation either, as olenty of charcoal was used although the oven silica tube has to be used due to the fact that the low carbon iron was in the form of no. 8, thermo- couple wire. Figure 6, curve 1, shows this iron heated in the reeion of tie lower critical. Lue to the fact that solid solution ronee in ferrite is so snall and the critical is so vertical, there is no sudden change in volume before before the eanna range is reached. Besides thereis no change fronalpna to ganna unless the iron in the cementite is gamma. all—a Figure 6. When iltxi is heatec in) throufin.tmm3icritical ranies as shown by curve 2, on the sane sheet, there is a 1 - . . - ‘—~ - “‘ O ciiferent story. As the-temperature runs over poo Co, the exoansion slacas off sradualiy, as the ferrite rradually becomes leaner in dissolved carbon according to the steel diagram. This heals that there are snail oaroicles oi austeiite formed which are yanua aid are shrinriag as the temperature is increased age the resis— ual ferrite is sudcenly changed to ganua iron forming a weik austenite. This is shown by the steep decrease after‘EBCfaj. is reached. The reverse is true in cooling of iron also. The najority of the weak austenite is changed to ierrite instantly and then there is a slow decrease for a while until all the remaining austenite is changed back to ferrite. Then the decrease is more rapid. This sluggisn period is reoresentefi in temperature on the equilibrium oiagran by the distance between tne intersections of a n .anfi Carbon and the A7 vertical line representing 2 g: and the line PG. Figures 7 and 8 show the iron, that has been on . ‘ . + annealed at 1000 o. for one hour ano slow cooleo, at 100 afld 503 diameters, resaectively. The latter shows the small particles of cementite that are thrown out of solution, as the temperature passes tnrough the range PN, on tue equilibrium diatran. This happened to be one of the richer areas in carbon ciscovered- Microscopically F i gure 7 . Iron Thermocouole Wire 100 K Figure 8. Iron Thermocouple Wire 500 X . .5. . ..r $1” .. LII/M321... ,. .\/ J Nymw Lylhfim- the carbon content varied quite a bit. 8. A. E. 4615 The alloy steels are not Quite true to the iron— carbon equilibrium diagram of course. The heating curVe is very similar to tha of the 1015 steel exceot for the spreading out of the he“ range, shown by the 3 flatness of the curve between 7500 and 82000. on curve 2, figure 9. The Acl is also lowered at least aOOG. On cooling the action is very such the sane as may be seen by the two curves in figure 9. The cooling is even more sluggish, as the Ar3 is at least 4003. lower. The Arl is lower or at least_spread out so that it still has some effect at room temperature. This spreading out of the Arl, probably due to the sluggishness of nickel, changes the coefficient of expansion, as also may be seen in the diagram. V Figure 10, shows the curves of five consecutive rapid heats and coolings. It may be seen how heating below the critical has very little efiect on the shrinkage and as the tenoerature is elevated into the critical ranges, the greater becones the shrinnage. Cther Peculiarities of heat Treating. There are a great many variables to be taken into acCount wnen studying the dilatation of steel. Unless these are taken into accoun, results obtained will seem very unreliable. That is why we studied the plain carbon steels aling with the carburized stocx, thus -13- Figure 9. .1 .o .111. 9‘ 1"... Figire lO. llst‘Q-ll“ | I- - - . - - . , I elininating some of the variables. Two great effects are, rates of heating and cooling, and other internal strains produced by necnanical work. ‘ Everyone knows that quenching in water or oil has a marked effect on the shape and strains of a piece of steel. Figure ll, curve 1, is the dilatation curve of a piece of S.A.E.*4615 steel that has been quenched from 76006., in brine. These are both carburized' pieces shown here. The second curve is a piece that has been furnace cooled on carburizing. Lvidently the quenching eXpands the piece because when the quenching strains are relieved it should return somewhere near the normal size minus one normal annealing shrinkage. The start and finish of curve 1 must represent the quenched and nornal size of the piece plus the shrinkage of an ordinary anneal, resoectively. The shrink in curve 1 minus the shrink in curve 2 ahould give sonewhere near the expansion on quenching. Figure 12 shows the curve of a piece that had been quenched from the box. The first heat was to 19000. The second curve is a subsequent heat to 33000. and slow cool. On curves 1 of figure ll, it will be noticed that there are two changes in the coefficient of exoansion on heating. These same Jogs would be in the curve in sheet eleven except not so marked as the quench was not so drastic. The upper jog is present but the lower -14- Figure ll. Figure 12. _ a . _ _ . W _ _ . . . . a _ h _ . . w n _ . _ 3O, 1.“. I. .- V- o . A 4 N a. .Qufl \E .3 elm a . jog was removed by the 19000. draw. Even less drastic heats and cools have sone effect on the critical temperature and the amount of expansion. This is especially true on larger pieces. If a piece is heated too rapidly the outside will heat up much more rapidly than the inside. This will cause a lowering of the total expansion of the piece at the lower criti— cal. ,As the piece approaches the lower critical the outside will begin to change to gamma iron while the inside is still expanding. This will havea tendency to keep the inside 0 f the piece from expanding. Hence tn total expansion may be lower at the Acl. Also at the Ac5 the dip may nat be so great due to the action of O l case again. On cooling the outside will again reach the Ar3 first and the inside will be held from shrinking its full amount, as the outside will be expanding again to alpha. Tie same damping effect will take place at the Arl. The net effect of heating and cooling is to flatten out the expansion changes in the critical range on fast heating. Of course this effect would probably not take effect on these small samples used but there is a pos— sibility. Probably the greatest factor causing shrinkage in non-carburized pieces besides quenching is strains set up by mechanical work. Figure 13 has two curves of iron thermocouple wire heat treated as received. And figure 6 has three curves of the sane wire that has been -15- Figure 15. heat treated at 10030. for one hour. The grain of the two irons are shown in the next two figures. Figure14 is, as received, and figure 15 is the annealed piece. The first curve in figure 13 shows no shrinxage when heated below the Acl and the second curve shows quite a bit of shrinkage. The fact that the shrinxagge to gamna iron is so much greater taan the return to alpha, made us wonder, if some of the shrinkage was not due to hot rolling or some other mechanicalwork. We annealed some of this iron at 100000., as shown. Tnis showed a shrinkage when hea+ed both below and above tne criticals. We blame this on the fact that the first two curves on sheet 6 were obtained from pieces that were straightened. slightly before using. The third curve was run fron iron that was not strained in any manner after annealing and the shrinkare as it will be noticed, is a lot less. If time oernitted this would have been checked. Figures 16, 5, 17, and 13 show the sane thing for S.A.E. 1020 non—carburized stoca. Figure 17 shows the hot rolled bar stock, as received, and F gure lSshows _ .oCo the same after it has been annealed for 4 hours at 1000 In comparing sheets 16 and 5, it will be noticed that the shrinkage of the unannealed stocx is much greater. Also the maximum expansions at the A01 are inclined to be greater. In looking over all these curves it will be seen ~16- Figure 14. Iron Thernocouole Wire As Received 100 X Figure 15. Iron Thermocouple Wire Annealed 100 K , l . . I, 1“! . . J r V \ a ’ ‘v‘h r o . . \V“ ‘9 - 1‘" ‘1' f 1"" V __v ‘13:! ‘ ' ‘ .6"; a'd“ V3- ‘ I *4- "‘1 M ... I b ' ’ W “‘1' ' ‘A I rl ..t "95' . 3 . . \‘ .h ; ‘a'Qél . I) £ _.. . . ‘ ‘ \ ‘ Figure 16. l . 7?. 9"- '0. Figure 17. S.A.E. 1020 As Received 100 X Figure 13, s.A.t. 1020 Annealed at 100000. 100 X \- 1.4m 1 h cl}, that there is a nreat deal of uncerta nty about these curves, as to the anounts of dilatation in either 1 1 direction. Part of tgis way he one, especially in iron, to the fact that the line GEN is so nearly vertical in figure 5. The composition nay vary from nearly zero carbon content to over .026fi, wnich is the average carbon cont nt. Lifferent pieces nignt act quite differ— ently on annealing, as a consequence. This effect may be noticed in steel, the lower its carbon content. For once in a while S.A.E. 1615 stocx will have a tendency, even though it has been annealed previously, to shrink. D e L It is difficult to say how much effect any of the‘ variables will have on a piece, without a much more detailed study on each different phase, such as, rates of heating and cooling, temperatures gradients and nechanical strains. Carburized Stock All 1015 and 4615 carburized stock has been cooled in the box and in the furnace, except for one piece that was quenched from the box in oil, at 92000. All pieceswere originally cold and hot rolled. The 4615 were rolled hot andthe 1015, cold. Figure 19 snows two curves fron a piece of carburized,1015 stocx, one heated just to the critca and the other well through the criticals. As may be seen t ere is plenty of shrink- age in the annealed piece. The shrinxage of carburired -17- Figure 19. pieces is consistent in all cases, also. In fact several oieces ha.ve been hes ted up and down through the criticals for a number of times and the total shrinkage is usually the exact nultiole of the shrinkage of one single neat. Going back to figure 4, we may see why this shrinkage takes place. On curve 1 you will see that the eutectoid constituent has little effect. As this car- burized piece is heated up toe steel begins to snrink o at about 730 C. The core stops shrinking but the case keeps shr rinkiis as much as the core will letit, as sho"n on curve 2.1“;1e core ii: 1ally gets the upper hand and the piece s ,oos sirinkin; nargedlv. The temperature increases, the case wishes to expand but the core is still chanatng to ganna iron. Evidently the shape of tne curves is due to t.he stre1-on or the phase change balanced against the strength of the dilatation of tne piece. However, when t1e aloha is all changed to gamma the steel expands aiain. Upon cooling, the cwse is still trying to shrink while the core is trying to expand after the Ar} is reached. The curve is flattened out quite a bit. At the lower critical the crzsr wishes no err-round and the 1‘ ~, 4‘ -L ‘ «.1 N ' « “.\.~~ rs. ' o -~. N I '1 ’v . r‘. "1 ‘. r‘ . -1 r .... ,w . Core is 111Ho11in, s .u.‘1x7'talb eKQJ.HolQJLo also " \ "1 A " ’1 j ' ‘ rd 7—. r1 ' \ Z " .‘ 1 "‘ H e1n i re: e- . iuie t.om1eso:fl3 o: i it a: supixrccis;ons ‘ \ ' ‘ ‘-' ‘L ‘. ‘ . ‘ ‘L ‘ ~ ,' ' P. , * ,—, - ‘ , - .1 r... . V - , . 1 ~ :1" ~ \ 'r ‘ . " 1 ‘ - . ~ - .2. beinl that nae piece as a vhoie shrinus. Anotner tnilu '- - '. r" 1 1 . ‘-- '- -‘ - cvv -‘ x -— 1‘ ‘- —- (WV- ‘. ‘ ~‘- ‘7 -~ (fi'fi H 1‘ very notLCeJole is the lonelin; oi the i Aanfl eiolnsion at the As although the teioeiature of the Ac is ~+~ |_J sane in either case. The coefficient of expansion is 1 lower than the non—carburized 1015 stocg from tne lowest tenperatures. It seess to follow ann avera;e between the two c rves until 730 C. is reacgcd and tnei it fails ,, _- 2“,- '1‘_:... n,‘ 1:, 1. . , 11,... _ 1,“. 1-1--.1... off‘zrzyit r . '111s ifilllik- s .1-.c 4t 1L1e c): e bt.mJtr3fitLrC ‘ ' \‘. 1.1“ a': ' " : "V‘ . " 1‘\1‘ ‘ 7 'fl . gzs tne {LAlflcolflolflfit stocs.1,ut 91,!1 lJMEl"golnt cm 1—. ', ., ‘: , 4.1.. -. 1 my. 4 x .‘ ,. . . K sf exizlsioi cue to .ne case. ins case Ltlmg tne saie conoosition as 1390 stock (eutectoid), acts as nearly like curve 3 in fitrre 4 as the core will let it. Consequently the the case will not reacn its saxigiuin heifnt as the core is .xpanding hardly at all for about 20 C. before the Ac is reached. This nignt be also 1 be effected by a ttnoerature gradient but it is doubtful. Possibly tnere is a better exolanation. - L Figure 29 brinjs out tne sane effects on 4615 carburized stock. It also has the sage shrinkage and lowering of tie Acl dilatation. The carburized case brings out very well the subnerging of the Arl. This critical was not shown in tne son—carburized stocx due .to the lack of eutectoid material. Figure 21 shows the carburized 4515 stocx heated to just below the Acl and tnrougn tne ch. Altnougn the Acl exoansion is lowered tne cefficient of expansion is the sane on heating and cooling, wnen heated below tne lower critical, as there is no shrinkage. Figure 2? gives a conoarison in the dilatation curves of carburized 1015 and 4615 steels, again snowing tnat 4615 is a much better steel to use for carburiziag as far as dilatation is concerned. Of course tnis curve -19- Figure 20. _- -.. o _——.—— --- ..— -r—A-u ---. _ Figure 21. Figure 22. we? , m \ ....E J, \ \3 M _ \ M w . _ . f .._T..:.-.TTT_TT --T-..--.....,V-.,. T w h w. .- - f . _ 3.3g» Shgwsgwu _ w R T f ., _ . L. . _ _ x x ~ 1834...: 1 T , Tr Tr T ‘1‘ T . . TL—fl. u: wfiT ~—T L... - - —.——.~. 9T T _ _ m ~ M . . _ . 3T . Ti. T- T. 1. _ _ h . . a . . . fl . W m . m w . fl . n n x - HTTTTJTIN. TL \J/ ... \ ... \ _ .. lté\ o .9 ~‘4AOA:’.AI‘AW;' v... 4 11 . ------- h w . m _ 2‘ A 3 . T. .W .. . W. .v~.. *4 . . . . w. . _ T w t _ - ,-TT T4. _ 3w. Tm ._ J; . _. ._ _ :2: WWW .f.. .rL¢I.. ...LTH . .-...- r._..,TY a g . . .. . T H. . . »~..» » _. Mmbb of carburized 4615 stooK is tne fesultant dilatation curve of the case ano core. We were unable to obtain a any eutectoid molybdenum—nicgel stooa to find out how the case acted alone. C ; l‘ICLU ICT’S So far we have found nothing to prevent a case carbur— ized steel from diminishing in size upon heating through the critical ranges and cooling slowly enougn to keep the steel in the pearlitic state. All Kinds of trest~ ments were tried when heating above the critical to bring it back to its original size. The treatnents tried were; a very slow heatingand cooling, rapid heating and cooling, combinations of both and also changes in the rates of heating and cooling during the run. ~Cf course the piece nay be exoanded by quenching but this sonetiues gives the wrong properties, and anyway on the draw, if J the .raw is high enough, the piece will again shrink. In order to reduce shrinkage as nuchas possible by heat treating, one should elininate all the variables, such as cooling strains, working strains, strains set up due to too rapid heating and cooling during the run, and temperature gradients. This treatment also holds for pure iron and will eliminate shrin‘age in non-carburized stock/ ‘ Alloying combinations may be also developedand studied, that would tend to reduce this shrinfiage or eliminate it entirely. The 4615 steel has done this to a great extent. Of course to Keep away fron eXpense, it might be better to calculate for the shrinka e in the machining and forging operations. Of course this would not eliminate warpage and strains set up in the piece. -21- ROOM USE ONLY . fl 1 37"." ;~ . I “I"; )R I’r 4' ‘ ‘ nfi'.>‘&t t . . "S‘n “ . a .; ,--' i.» - 73’. r.\_‘."*~ .. -. - ’ ' ." .."' I v v.51. .. . .o- ‘. .I: -I ‘ .VJ. .. ‘ 'élrfi‘ 4 PI "...? fiffififir ("3‘4" ”(L S" V‘ I ‘ I' Q I“ ' ‘ . “‘0‘ ”‘31§ .§ . a . ll]... I: I (‘ AVI‘ -’ . - V .I ‘ ,‘ , ~ .‘ ;~'V‘._“_I .- . Q‘ I . U""f1"'_,.,' I) ‘\ o“ I .I .4 . v - . . . , .' ' & II.-.I “I ' I . ’ «WP. .‘ . v 44‘ .. . ‘ II . - . , . . .I . . .. . . N d I . t.vI 'I- T. no. . an”... '. u a I s; 'l ‘ | 1“. t . ‘l" U ' :‘ 233‘}. :/ ~,3\>. L? ' "'.'C"\ 2 .- . . . - . . .I I: 1"" . . I. . II n." zip-‘g'r'dg' ltmr'?’ ' "7 ‘ -' ' i" ‘ - -' . ' -- " . ' ’ ',. " .. ‘ f . ' '- . :9 . ~ " ”$1.31yz-'_-1’I -.-.- - n. . ' , '. . .- - . n. - . - -. . . .. I, ' ‘ H ‘7" ’-" ' ' ‘ . _. ' ' ‘ ' -' H ‘ '. . ' w m- 4?. w{‘ |‘\ ‘\ .: I f t I‘D : L - ..A'U , _-" . _ . . I . l ' . a n ‘ ' :. f . A ‘R ' s p l" I ‘ III ‘. 1:1]; ‘IJ ‘."" ,I. «VLI‘ I I .- II , b, ‘b I I . .... . I I, _ . IIV . , a . , . i‘ - 1. .-3‘2- 1 V'?’.‘ . ‘L '. »I '5'" ”Q '3 ‘ . . , . - , l . . . . , v . 'a . . ‘. CI , .I.I ~, ‘ ‘. .I I‘II.gII..II. , . ... r. IE4 ‘ . ' .. -.' ' I' I .v)"". ' | "4' ,I ' .’ II ‘ - . ~. I . .c p I ' .- a ‘ '.-' ‘ 3 “ c L .i HJC"‘1’I""~'~ ‘ I 'h ’- ' "3:" "if/”J " R"; ,I .' .O' . ' to. .-_ ' a {5‘1”};- :-\'.. ‘- .: v IJIv .' V" ‘J‘ . ' _ ‘. 7 I ' ‘ I. . -.". ‘: i.u$‘éfi;a‘¥l€‘fi?’ L‘Ir'fi- I flifif‘s.‘ ‘ ‘1‘ 'N ~ 'u~'.:r‘-r',~.7;fi',&.’-.‘-g 3y? 'n .‘ .r. v 1..» «I. ' nut: 1. 7-- "n‘z, ‘?'*.-’.-*9’7M 'r-‘- - I 1' n. I", .‘ '. I! t“ 'Ilfiu’ I " AI"“"‘J A 'a".( I no I. ' ' "i” «it ST“? 74.: "#4.- .~ v! [34% ' . .I .I.‘ '3’, -.“:I 4 La) .a-‘iI '- ”:9. RI". “if! 12"»?‘I. . ' I 1' $1 fl'vf‘fx 3‘“. ’ if”. Q“! ’g‘fi}! ~ 1 O 'C. “ ' . 4 ’ -< 0 ‘I- 104 “q K .c a .'_ “5' 1. A‘T' '- n." - . set..." ;. “ .r.'.‘I‘ ' ‘:/‘|. .-~:V,m. ...;ax...3, : - w“- .. 3M. -. ~ .- .‘. _‘ '9 - ."IJLM. ‘ J) I ’2'} I 1 o It. _I IN} J" III. I. . ‘ . 4 .. . 1.. . . a» -. .»¢~ .. ‘ ' ' » ’. 'l? - r u ‘. ’ ‘. ‘ .,§‘{\.I" I‘ ’I‘“( .i I i. 'o i . 1 - - I -,...|_I I- ,II\ “ — fl is -, u - £1” w: {'3 "i ..., ..1 'I6 I, _ -"I .- IoI . ...‘.I 3'“ fillgiitrfl‘” I «1‘- AM (II-II. .I . -. 3"" a w 1'. I If 234,134.-. ... JIM?“ .-II..II . y ‘5 Iv." ...“tn‘rfi‘ '1 ' ‘?~ ‘ " . ' ‘3. will fl . .a‘ I ”v ‘ -‘ :2 ‘0' I, ." _ 0' III. 7‘. .II -_ I I, I "."4. .‘ I . I I I I I II. ... «\- 1’7"; “£51.; , .139.- 1"272: 7, . 4. . .. .II :- ”it . V ‘L.'.' .. ’ ”Ilfix‘r‘ "xvi-1:9 'III I I v I . III .f ’ '. . -* '..' . :' . ' ' ' . 'j a“! 'r.'I.'§g';;I ‘ .- ~ . .4.‘. .‘.. r. velé. I" I I . >§A1£k !.J.I .l 0 \ 6 43*.- ‘x. I. a‘ . v5..."«5*~’~f‘,‘“ii. . aw I ’1‘, E 5.15}:- .‘:[m ¥ n k "L331; 985:1:- v .' W‘ " ’17:" 1‘ .¥f( . "'-9'1"i., u; ,_ r u . I .. .‘~ 9‘. g .‘D r - I‘,,;’tt:'7g -..'~' I“ .l. A . ‘ ~ I, \ ~fl" III‘ . I ' " ~ " 3'" 7«'"*‘-“-".‘~. ‘.--‘t'.&J'C"+i".‘.-¥."“ "3. 9"‘-§...*:.~.t.-:-ti . : . \ $1" ".J_’r KK‘III': I.IIE‘%£..HIF4? I. I.:§/“_‘I, ~(.:.III~.III!II ’; . '- I ' ." . . 'a'jub'A 8:”.‘7'1‘j'fi‘x ',‘. I.2’-£\'.',"'I “1"? .‘vk. {.- Ami/\I-I‘I ._ .; I. Ir Vy)‘ 1?: Awerkfigv‘,s‘1 .:;'I:.“J‘II.I fi!‘.“,§"‘l"i‘q“¢-T I I' G.‘ ‘ {z Jr‘s" .' ._§I.-T£‘gg.-. .r-.-".‘..',’I_.' ‘ I I _ I ’ I‘ .“i"“‘:l .;;.-; ’A ' ‘ P' ‘5 {@11’7. I.“ ‘ ‘. a u ' n1, 0 l ' J 1.3.3? o .I :v I' 44$ p I-. . I. I" .. .‘ _ .II _ III.. ' ‘I IJIIII. II'I; .' . ‘f ‘ It I IIJM .I‘L» _. I'v'.;» .‘ ' - _ H ."'\ A . .Q'T-z .“II' ~ . . . V _I. 1.". . '4“. 1 .’.! I , v II . . » A, .. =~ “mafia .-. «may... . 23,3... ad». 1 “i t: - . 0" ..9If:,t.o:fi"-"i ‘ .‘i’PS Mart"? ’- V' " "«'£='-f'i"v.~-".?.-. v '_ ‘ .1 ~ . . ' ._‘ _ ~ . , rw ’ r v A)! :— ‘ f. ‘ ‘ A I?!‘ .‘: ‘ i 3 i 5? ,v ._ . q. .-‘ ‘ I fiIQE In t I: ' m»? aflé |||||||||l||H Ill lll' l‘| Ill ll| ll l I I'll llll II |