CA1149199A

CA1149199A - Method of heat treating high carbon alloy steel parts to develop surface compressive residual stresses

Info

Publication number: CA1149199A
Application number: CA000331854A
Authority: CA
Inventors: Adam M. Janotik; Charles A. Stickels
Original assignee: Ford Motor Company of Canada Ltd
Current assignee: Ford Motor Company of Canada Ltd
Priority date: 1978-09-13
Filing date: 1979-07-16
Publication date: 1983-07-05
Also published as: US4191599A; DE2967018D1; EP0013063B1; EP0013063A1

Abstract

ABSTRACT OF THE DISCLOSURE
An improved procedure for developing compressive residual stresses at the surface of a hypereutectoid alloy steel is described which is of decreased cost. The procedure does not require special furnace equipment and involves increasing the austenitizing time during heat treated and regulating the austenitizing furnace atmosphere to provide a high carburizing potential.

Description

The present invention relates to the develop-ment of residual stresses in the surface region of ~
articLes, more particularly bearing elements. ~-It is well known that the state or degree of residual stresses present in a machine part subject to bending or contact loading can have a major influence on its service life. Much effort has been devoted toward ;~ developing compressive surface residual stresses by shot peening, surface rolling and by heat treatments such as carburizing, carbonitriding and nitriding. Descriptions of many of these methods are included in the following publications:
(1) J.O. AIman and P.H. Black, "Residual Stress and Fatigue in Metals", McGraw-Hill Book Co., New York, 1963, Chapters 5 and 14.

(2) G.M. Rassweiler and W.L. Grube (Editors), "Internal Stresses and Fatigue in Metals", Elsevier Publishing Co., New York, 1959, pp. 110-119.

3) Metals Handbook Vol. 11, 8th~Edition, Amer ican Society for Metals, Metals Park, Ohio, 1964, "Case Hardening of Steel".

(4) "Carburizing and Carbonitriding", American Society for Metals, Metals Park, Ohio, 1977, pp. 86-92.
Each of the above known methods of developing .
compressive surface residual stresses when applied to high carbon steels have their attendant disadvantages. For `- example, shot peening and surface rolling are disadvanta-geous because of limitations as to (a) material hardness, (b) size and shape of part, and (c) resulting surface finish that cannot meet all requirements. Nitriding at temperatures of 1100F and below is usually economical ~` only as a shallow surface treatment and therefore dis-advantageous. Carbonitriding or nitriding, while steel is in an aust nitic condition, requires simultaneous control of both carbon and nitrogen .

~' 9~9~3 po~entials in the g~s phase; it is difficult to accurately control the potentials and therefore it is fxequently overdone, producing high levels of retained austenite along the part sur~ace which is disad~an~ageous. As ts S carburizing, it i5 generally assumed tha~ it is not possible, by diffusing more carbon into the surface of a high carbon alloyed steel, to produce compressive residual surface strPsses (see 14th Interna~ional Colloqium on Heat Treating, 1972, p. 11).
Carburizing techniques are nearly always applied to low carbon, low alloy steels, such as AISI 8620, 4118 and 4620, which contain 0.1 - 0.3 wt. pct. carbon. For hypereutectoid steels, austenitized at temperatures too low to dissolve all carbides, an effec~i~e equilibrium is established between undissolved carbide and the austenite, which is then saturated in carbon. It is also generally accepted that this saturation prevents such steel from accepting additional dissolved carbon, and thereby prevents an increase in the amount of carbon dissolved in the austenite near the surface. Therefore, an appreciation of carburization wi~h respect to hyper~
eutectoid alloy steels, has remained an unexplored area until this invention.
This is not to say that the prior ar~ has not employed hea~ treatment meth~ds to produce compressive residual surface stresses in hypereutectoid alloy steels, but they have been carried out by methods which have required the addition of ammonia to th~ austenitizing furnace atmosphere, which in turn causes nitrogen to be dissolved in the surface layers o~ the steel. Th goal of usins such atmosphere is to increase the nitrogen conten~ of the austenite surface~ but at the same time avoiding the formations of nitrides of iron or other alloying elements.
Ammonia atmospheres present special equipment requirements which it is desirable to avoid and present control problems as to nitride avoidance.

In accordance with the present invention, there is provided a method of developing compressive residual stresses in the surface region of a hi~h carbon steel alloy article, the method comprising: ta) constituting `;5 the steel alloy article to contain .8-1.6% carbon, .2-5% chromium, and 0-20% alloying ingredients selected .from the group consisting of manganese, vanadium, molybdenum, tungsten, silicon, the remainder being iron;
(b) heating the article to a temperature of 800-950C
. 10 (1472-1742CF) for 1-2.5 hours in a carburizing atmosphere effective to generate a differential in retained austenite, ~ primary carbides and carbon between the surface region and `. core region of the articl~; and ~c) immersing the article in a cooling medium to quench the central core of the ~15 article at a rate sufficiently fast to effectively suppress the formation of non-martensitic austenite decomposition .products, thereby establishing a residual compressive stress gradient proceeding from the surface region of the article to a depth between 0.007 - 0.03 inches, the residual compressive:stress being in the range of 5-40 Ksi.
;This improved me.thod for developing compressive ..residual stresses, when compared with the prior art, is lower in cost and is more convenient to use, in that special furnace equipment is not required. The increased surface hardness of the steel is achieved without a corresponding increase in bulk hardness, so that the toughNess charac-teristic of other lower hardness material is retained.
~.The invention also includes a high carbon steel :alloy article having gradients of compressive residual ~ :
-30 stress, and at least one of a carbon gradient and hard-ness gradient proceeding from the surface region of the article to its core, the article being characterized by a microstructure consisting essentially of tempered martensite, retained austenite and a carbide phase, the :
article having a chemical content consisting essentially of .8 -1.6% carbon, .75 - 25% alloying ingredients including .2 - 5% chromium, the remainder being essen-tially iron, the article having a compressive residual stress level at its surface region of at least 10,000 psi., and tensile stresses at the article core, the article having a hardness differential between its surface and , -",:~..

~f.'~
:

core of at least 2.0 Rc, and a volume fraction of primary carbides at its surface region of at least .1~. -Steels used in ball and roller bearings are of ; the following types: (1) hi~h carbon, low alloy steel, -~ 5 such as AISI 52100 ~1~ C, 1.5~ Cr) through-hardened by heating to typically 825-850C, quenching and tempering, (2) low carbon, low alloy steel such~as AISI 8620, 4118 ' ` and 4620 hardened by carburizing the surface (to maximum surface carbon contents on the order of 1%), quenching and ` 10 tempering, and (3) high carbon, high alloy steel such-as ` M-50, a tool steel, or 440C, a stainless steel, used when elevated surface temperatures or other extreme operating conditions are anticipated.
Type (2) bearing steels, as indicated above, have one distinct advantage in that substantial compressive residual stress can be developed at the part surface as a consequence of carburizing. The favourable residual stress distribution is thought by the prior art to make a significant contribution to the durability of the bear- `-ing. However, compressive residual stresses are not pro-duced when steels of types (1) and (3) are hardened by the indicated conventional through-hardening techniques used by the prior art. Carburizing has not been considered as a means of developing compressive surface residual stresses in types (1) and (3) bearing steels, since it has been generally accepted by the prior art that it is not possible to increase the surface dissolved carbon content by diffusing additional carbon into a hypereutectoid alloy steel from a furnace atmosphere at the usual austen-itizing temperatures. This is evidenced by an articlepresented at the 14th International Colloquim on Heat Treating by Mrs. Stefania Baicu of the Institute of Technological Research for Machine Building in Romania.
In an article entitled "Contributions to the Influence of Compressive Stresses generated by ~eat Treatment on the Fatigue Life of Parts Under Rolling Contact Wear" on page 2 of the conference preprints published in 1972, she , ,~_ states "in a through-hardening steel, in which the martensitic transformation takes place throughout the whole section of the part as a result of its high carbon content, there could not have been question to diffuse S more carbon in the surface layers in order to lower the Ms temperaturen. This means that the possibility of lowering the Ms temperature through the addition o~ carbon to the surface layers (thereby inducing compressive residual stresses) is not possibl~.
Accordingly, the prior art has turned to one other possibility for improving the compressive sur~ace stresses in a high carbon through-hardened steel by heat treatment.
Koistinen (see U.S. Patent 3,117,041, and an article appearing in ASM Transactions,'vol. 57, pp. 581-588, 1964) as well as the aforementioned paper by MrsO Baicu, suggests adding ammonia to the austenitizing furnace atmosphere, causing nitrogen to be dissolvea in the surface layers of a high carbon through-hardened steel, such as 52100 steel, thereby inducing residual surface compressive stresses upon quenching. This is a kind of nitriding process where the goal is to increase the nitrogen content of the austenized surface, but to avoid forming nitrides of iron or other alloying elements.
Although this process has met with some degree of succPss, it carries certain disadvantages such as the cost of adding the ammonia treatment step and the difficulty of controlling the quantity of nitrogen absorbed by the steel to the amount desired. An excess of nitrogen in the surface layers can lead to certain difficulties, e.g., low hardness due to excessive amounts of retained austenite, or, in extreme cases, grain boundary porosity due to internal nitrogen evolution.
The p~ese~t invention pro~ides for an economical 'and controll'able method of increasing the fatigue life of a bearing by (a) providing compressive residual stresses in the surface of the steel specimen by a simple heat treatment in a carburi~ing atmosphexe, (b) providing ~"33 9~
.

:
.. ~
increased retained austenite in said surface zone, (c) ~- providing an increased volume fraction of primary carbides near the surface, and (d) providing higher hardness near the surface, which is i~ part dependent on limiting and controlling the chromium content of the steel. The heat trea~nent can bs carried out at a relatively low tempera-ture in a~carburi2ing atmosphere, and is best conducted :
for critical periods of time between 1 and 2 hours. Theaim of this treatment is to establish a gradient normal to the surface of dissolved carbon in austenite. Since the typical austenitizing temperature is too low to cause all of the carbides initially present in the steel to dissolve, it is not obvious why carburizing shoùld produce residual surface compressi~e stresses. If the austenite is ~aturated in carbon~(because all the carbides cannot dissolve), how is it possi~le to establish a gradient in ~
dissolved;carbon by dissolving more carbon at the surface?
The prior art has been unable to do so or has believed it is not worthwhile trying to do so.
It is theorized in accordance with this i~vention that the austenite in a plain carbon h~pereutectoid steel, a steel with nagligible alloy content, heated to a tempera-ture not high enough to dissolve all of its iron carbides, ~rapidly becom~s saturated in carbon. If carbon is supplied to the steel from the furnace atmosphere, the volume fraction of undissolved carbide (primary carhide) increases, but the amount~of carbon dissolved in the austenite is unchanged. However r a hypereutectoid steel containing an alloying element such as chromium, whose affinity for carbon is greater than the affinity or iron for carbon, ~` held at a temperature high enough to form austenite, but too low t~o dissolve all carbides, slowly redistributes its carbon and chromium between carbide and austenite phases~ After a period of several hours (as opposed to several minutes in a plain carbon steel), effective equilibrium is established and the austenite becomes~
"saturated" in carbon, but saturated only with respect , . ~ , . . .

93L~9 to carbides of the composition with which it coexists.
When carbon is added to the steel from a furnace atmosphere, the voIume fraction of primary carbides increases neàr the sur~ace. As more carbide forms, the remaining austenite becomes~depleted in chromium, because, proportionately,~more chromium than iron goes to form .
the new carbide.: As the chromium content of the austenite ~ is lowered, its solubility for carkon lncreases, thereby :- allowing a surface-to-center gradient in dissolved carbon content to be produced. Within the two phase field ~austenite and cementite) of the C-Cr-~e system, increasiny the carkon content of the system increases the carbon content of the austenite. This effect is especially marked for chromium contents o~ 5% or less.
Another factor contributing to the development of a .
dissolved carbon gradient is the slowness with which the equillbrium~dlstribution of carbon between carbide and austenite is~approached in a steel like AISI 52100. In short ~reabments ~(up to ~ hours at 850C) o well-spheroidized steel, the carbon content of the austenite never attains its quilibrium value. Thus, it is possible according to this invention to establish an even larger surface-to-center difference in dissolved carbon content ~ ~ -than is indicated by the phase diagram. This also has ~5 not been appreciated heretofore.
Carburized high carbon alloy steels containing con-~rolled chromium will contain a larger fraction of primary carbides near the~sur~ace than in the interior. Since the carbide phase exhibits no abrupt volume change on cooling (such as occurs when austenite forms martensite) and since the volume change can be a source of th residual ; stresses which develop, the higher carbide fraction at the surface should moderate any residual stxesses which do develop.
A preferred mode for carrying out the present invention is as follows:

:

: 1. Substrate ~reparation: The steel article to be sub~ected to heat treatment should be selected to have a caxbon content~in the range of ,8-1,6% carbon and should contain chromlum between .8-5~; other alloying ingreaients may be selected~from the group typically consisting of molybdenum,~ vanadium, tungsten, manganese. The total ~lloy content can range'from ~.?5~25%.
2, Aus _ nitizing heat treatment: The substrate or article'is then heated to an'austenit~zing temperature within a carburizing atmosp~ere fox a period of ti~e' between 1 and 2 hours to devel'op the hlghest suxace com~ressive residual-stresses, Longer treatment times ' produce thicker compressively-stressed layers, but stresses o~ less intensity. Th'e'carburizing atmosp~e~e pxeferably should ~ave'a car~on potential sufficiently h~gh to cause carbon saturation in a o . oazs~ thi'ck iron foàl in 30 mi~utes'. The full range~'o~ carbon potential cannot adequately be con~eyed by specifying CO~CO2 ratio because the' equilibrium C02 conten~ ~artes wlth tempera~ure~for~
different carbon potential~ and ~ ary from furnace to '~ ~uxnace,' di~feren't flo~ rates', and the'amount o metal charged. Thus, the shi'm stock empirical tes't method is best, using thin foil tsee'reference 4, pa~e 1 herein~
Such an atmosphere is preferably derived by using an endothermic gas atmosphere, consisting primarily of CO, H2 and N2, generated by ~he partial combustion of a hydrocarbon. The carbon potential can be adjusted by varying the proportions of air and hydrocarbon at the gas genera~or to match~ the carbon content of ~he part. But lt is most impor~ant that such endothermic g~s contain additional hydrocarbon, preferably by the addition of 3-10% methane. The added hydrocarbon in the form of :
methane contributes the necessary carburizing capacity to the furnace atmosphere~ It is not sufficient to 35 merely provide an endothermic gas of a high carbon potential to the austenitizing furnace tcustomaxily re~erred to as an endothermic atmosphere "neutral" to a high carbon steel~, but rather a carburizing gas blend, endothermic gas plus 3-10% methane, for example, mus~ be employed.
When the substrate contains chromium at the hi~h end

5 of the controlled range, it is desirable that the oxygen - content of the gas atmosphere should be reduced so that the formation of chromium oxide on the part surface will - not interrere with carburizing. This may be obtained - by controlling the gas atmosphere to contain nitrogen and methane in the proper proportions for achieving xesul~s equivalent to the results from an endothermic ~as based carburizing atmosphere. ~acuum carburizing ~s another method of carburizing without forming oxides.
3. Quenching: The heated substrate~or artlcle is then subjecte~ to cooling by conventional means, to produce the desired microstructure in the steel, usually martensite, such microstructure depending upon the application for the steel. Since the pres~nt invention is particularly ~ suitable in those applications where rolling contact ; 20 fatigue will be experienced~ the microstructure should be hard and strong. In most instances, quenching in oil maintained at a temperature of about 55C pro~ides a satisfactory cooling rate to achieve such strength and hardness. Slower quenches, e.g., into molten salt, or faster quenches, e.g., into water, may be used in some circumstances. Further cooling of the quenched steel by the use of liquid nitrogen to a temperature of -196~
will reduce the amount of retained austenite, usually producing a further increase in hardness and residual stress.
More essentially, the article is immersed in a cooling medium to quench the central core of said article at a rate sufficiently fast to e~fecti~ely suppress the~
formation o non-martensitic austenite decomposition ; 35 products, thereby establishing a residual compressi~e stress gradient proceeding ~rom the sux~ace region of said article to a depth o~ between 0.007 - .03 inches.

.

4. Tempering: The tempering cycle ca~ be adjusted to suit a wide variety of needs; typically, heating to a temperature of 100-150C and holding for approximately 1-2 hours is satisfactory.
In addition to the other characteristics of the steel alloy article produced in accordance with:this invention, the article may be characterized by resiskance to subsurface crack initiation at hard inclusions and resistance to surface initiated cracking as a result of the high compressive stress distribution in its surface region.
In the following illustrative Examples, reference is made to the accompanying drawings, wherein:
Figure 1 is a graphical illustration of residual stress as a function of normalized distance from the ~ur-face of each of two 0.090" thick specimens, each specimen being heat treated at 850C for 1 hour, one being carburized and the other not;
Figure 2 is an illustration similar to Figure 1 for two other 0.090" thick samples each treated at 875C, one being carburized and the other not;
Figure`3 is a graphical illustration of residual stress as a function of depth below the surface of the : specimen, for three 0.090" thick samples, each being heat treated at 850C for varying periods of time in a carbur-izing atmosphere, quenched and tempered at 150C for 90 minutes;
Figure 4 is a graphical illustration of residual stress as a function of depth below the surface of the samples, the .070" thick samples being heat treated at 800C
for times of l hour and 2 hours, respectively, in a car-burizing atmosphere;
Figure 5 is a graphical illustration of residual stress as a function of depth below the surface of the sample, for three .070" thick samples, the first two of which were heat treated at 980C for 35 minutesJ cooled to develop a pearlitic microstructure, and then heated again :
for a period of 55 minutes at 815C in a carburizing atmosphere, the third sample being heat treated ln a single step at 815C for 55 minutes in the same carburizing atmosphere;

, 11 Figure 6 is a tapered section microphotograph (1000 X) of a specimen which,. in accordance with this inven-~ tion, has been austenitized for 2 hours at 850C and oil .. quenched, the microphotograph being taken of a sur~ace zone;
~: 5 and : ~
~ Figure 7 is also a microphotograph (l000 X) of `~ the interior zone of the specimen in Figure 6. ::
EXAMPLE 1 ~.
; One series of experiments was directed to an analysis of the development of compressive residual stresses in a 52100:steel. The test procedure followed ;
was:
Strips of 52100 steel, which were:spheroidize annealed, having a dimension of 3 inches long~(7.62 cm) .. 15 x 0.5 inches wide (1.25 cm) x:0.090 inches (.23 cm) thick, .`~ were machined from roll-flattened 0.5 inch:diameter wire.
The nominal composition of 52100 steel is 1.0~ carbon, 1.5%
chromium, 0.35~ manganese, 0.25% silicon~and the remainder substantially iron. A total of 12 sample pieces were.
prepared according to the heat~ treat cycles indicated in ~ Table l; those:having an asterisk were copper-plated to ::
: prevent carburlzàtion during heat treatment and thereby :
: equivocate the prior art treatment which would not include carburization, but rather just the heat treatment at the 25 :indicated temperatures in a neutral atmosphere. Those ::
~::
~ samples which are indicated with a double asterisk had the :; copper plate removed after 980C treatment. The heat treat cycles:include a variation of the heating time and : -~
the temperature.~ The samples also were subjected to 30 different quenching treatments,:;.some:being an oil quench ~-~
~: with the oil maintained at 55~C, and~others including an additional~quench with liquid nitrogen. Certain of the ~
specimens were then subjected to a tempering:treatment as ~ ~ :
~ indicated at the temperature and time periods of Table 1. ~ ~.
: 35 . The austenitizing heat treatments were carried . : out in a Lindberg carburizing furnace with an integral :
quench tank.: The gas atmosphere was generated as an endo-thermic gas atmosphere, enriched with methane. A measure . of.the carburizing rate of the furnace atmosphere was obtained ~.-~- :

~ --~j 9~

~12-` by determining the weight gain of a ~oil of 1008 steel, 0.064 mm thick, which was inserted through a sight port into the furnace, held at the temperature fox 30 minutes, then rapidly coolcd. The gas mixture was adjusted prior to each o~ the runs so that the foil carbon content was at least 0.9 wt. pct~ carbon. For most of the runs, foils were also included along with the samples. Because some air was en~rained with the samples when they were charged into the ~urnace, the initial carburizing rate was low.
10 For example, the foil accompanying Sample S, austanitized for only 30 minutes contained only 0.72 wt. pct. carbo~;
in every other case when the austenitizing times were longer, the foils accompanying the samples contained carbon in excess of the amount needed to saturate austenite.
Following each heat treatment cycle, the residual stress distribution in each sample was measured and hardness ~- readings were taken. The residual stress distribution is measured by progressively thinning the strips from one side only by chemical dissolution, measuring the bending of - 20 the strip and analyzing the deflection res~lts using a modification of the method described by R. G. Treuting and W. T. Reed, Jr., Journal of Applied Physics, VolO 22, 1951, pp. 130-134. Average hardness readings were taken for certain samples by a microhardness transverse (Knoop indentor 1 kgm 25 load) throuyh the surface region of the sample.
With respect to the Samples 1 and 2~ Figure 1 shows that the plated specimen, which did not exchange caxbon with the furnace atmosphere, developed a small sur~ace tensile stress, while the unplated piece, which was 30 carburized by the atmosphere, developed a surface compressive stress of about 15,000 psi at the surface, shown as a negative stress in Figure 1. Without carburizing, spe~imens tended to develop tensile surface residual stresses;
therefore, the change (which is a sum of the tensile and 35 compressive values) in residual stress distribution produced by carburizing i~ more substantial than the stress dis-tribution in carburized piecas would sugge~t.

~9~99 ~ -13-:
Samples 3 and 4, which were austenitized at a slightly higher temperature and su~jected to a liquid nitrogen quench following the oil quench, demonstrated a very slight compressive stress for the plated sample at the surface, wh~reas in the unplated sample, the compressive stress was approximately 7,000 psi at the surface. ~he depth o~
compressive residual stress has been increased over that of Sample 2, but the stress intensity is lowered due to the higher austenitizing temperature and the addition of the tempering treatment.
The next three Samples, 5, 6, and 7 were austenitized .
for 30 minutes, l~hour and 2 hours respectively at 850C.
After oil quenching, samples were quenched in liquid nitrogen before tempering. Sample 5 was nearly free of re~idual stress; there was little carbon transfer from the furnace `
atmosphere to the specLmen. The samples held~for longer times (1 hour and 2 hours) developed definite compressive :
residual stresses`at the surface. Distribution o~ the `; stresses is clearly related to the depth of carbon diff~usion.
The ratio of the depth at which the stress changes sign in Sample 7 to the corresponding depth in Sample 6 is 1.48;
this is close to the square root of 2, the value that would be expected if the depth of the compressive stress was related to the depth of carbon penetration from the atmosphere. ~ ~
The distribution of prLmary carbides after 2 hours at 850C is shown in Figures 6 and 7 ~for Sample 7) at 1000 X
(picral etch). These pictures are from a tapered section and polished so that apparent distances noxmal to the surface are magnified by a factor of about 5.5 relative to the distànces tangent to the surface. In Figure 6, grain boundary oxides are found at the specimen surface to a depth of about .004 mm; this is a common occurrence when heat trea~ing chromium-bearing steels in endothermic gas atmospheres. Below the oxide is a carbide-depleted region of about .004 mm, probably the resul~ of migration o~ the chromium to oxides. The carbon content of the austenite, however, must be high in this region. Then a zone appears containing ~18 volume fraction of primary carbides tfrom point counting measurements). This zone extends from .008 mm to .07 or 0.10 mm below the surface.
S The micros~ructural features of the interior, shown in Figure 7~ demonstrate a volume fraction of primary carbides O~ oO8~ about half that near the surface, There seems to be no tendency to form carbide films in austenite grain boundaries in the carburized surface layer; rather the lO existing spheroidal carbides simply grow. The thickness - of the layer under compression increases with increasing austenitizing time, while the magnitude of the surface ~tress decreases somewhat~
The average microhardness of the outer surface region 15 of Sample 7 to a depth of 0.005" was determined to be 947 KHN
tl ~gm load, equivalent to 68-69 Rc). The hardness decreased with increasing distance from the surface until the base hardness of 880 KH~ ~about 66-~7~ RCl was reached at a depth of 0.008 - 0.010". This hardness gradient is an 20 important aspect of the present invention and is attributed to the high carbon content of the martensite in the high carbon surface region, as well as the greater volume ~raction of carbides thereat, more than of~setting the greater volume fraction of retained austenite.
Samples 8 and ~, Figure 4, confirm that a longer austenitizing time produces a deeper case, but a somewhat lower surface stress. This data also shows, by comparison with Sample 2, that the compressive surface stresses are higher with an 800C austeniti~ing temperature than with an 30 850C temperature.
Samples lO-12, Figure 5, show the Pffect of initial carbide size on the intensity of the residual stresses developed. ~educing ~he carbide size increasPs the rate of dissolution at 815C. Sample lO is the baseline for 35 comparison. Pretreating Sample ll at 980C, followed by an air cool, to produce more finely divided carbides, has an adverse efect on the degree to which compressive surface ~15~
stresses can be developed. Thus, Sample 12, with coarse, 510wly dissolving primary carbides, can be treated to produce the highest residual stress.
Average hardness values (KHN) were determined for Samples 11 and 12 at four subsurface regions as follows:
DEPTH ~ELOW SURFACE
Sample 0- 04 mm .04-.08 mm .08-.2 mm Interior 2 928 883 858 82~
~he retained austenite was measured by x-ray method 10 on the carburized surface o Samples 11 and 12 and on their centerlines after they had been thinned to measure residual stress. In both specimens, the average surface retained austenite was 24-26~. On the centerline of Sample 11, the average measurement was 15% retained austenite, and on the 15 centerline of Sample 12, it was 9%. These dif~erences in ~; retained austenite are consistent with the expected differences in dissolved carbon. The differences are also consistent with the observation that quenching carburized spe~imens in liquid nitrogen to lower the 20 retained austenite tends to increase the residual stresses.
EX~MPLE 2 A second series of samples were tested to investigate the ef~ect of differ~nces in chemical composition. ~hree sample materials were obtained with the compositions se~
forth in Table II. Pieces of each material were subjected 25 to a heat treatment cycle which involved heating to 1650F
~9V0C) in a carburizing atmosphere determined as in Example 1, holding at said temperature for about 2 houxs, quenching in oil having a temperature of 55C, tempering at 300F
~149C) for 2 hours, and then air cooling. Sample 13, however, 30 was heated to 1560~ (850C~ with the remainder o~ the pro-cedure the same (this lower temperature is necessitat~d by the lower alloy content). Sample 16B was subjected to a ; different heat treat cycle wherein the material was heated in a carburizing atmosphere, to 17504F (154C) for 2 hours, air 35 quenched, double tempered at 300F ~149C) ~or 2 hoursl and then air cooled.

-16~

Results of the tests (see Table III~ show t~at for Sample 15, no residual compressive stresses were developed at the surface o the article~ It is theoxized that this resulted ~rom the high chromium content of the tool steel 5 ~hich, because of the atmosphere containing CO, caused cxidatiQn of the chromlum which set up a barrier towards carbu~ization of the surface region~ I~ carbon monoxide can be eliminated from the carburizing atmosphere, lt may be possible to eliminate oxidation of such high chromium 1~ content and thus allo~ carburization to proc~ed with the same results g~nerally obtained for Samples 13, 14 and 16.
Samples 14 and 16, li~è l3 C52100 steell each had significant compressive strès`s at the surface consistent with the control of chromium content and carburizing 15 atmosphere.
The surface hardness of Sample 14 was not measurably greater than its interior hardness; the surface layer con-:. tained 14% retained austenite while the interior had 3%
retained austenite. On the other hand, whlle specimen 16A
20 showed a definite increase in surface hardness, there wasno measurable difference between the amount of retained : austenite at the surface and in the interior. All three ~actors - higher surface hardness, higher sur~ace retained austenite and surface compressive residual stress - are 25 împortant characteristics of an optimi~ed carburized layer in these steels; however, either a ~ardness gradient or a gradient in retained austenite content may be absen; ~n a carburized steel that is less than optimized, provided one or more of the other factors are present. A hardness 30 gradient or retained austenite gradient need not always exist, even though carburization ha~ occurred and residual surface compressive stresses de~elop Examples I and II demonstrate that the distribution~of residual stresses ~n ~uenched and tempered steel containing 35 th~ preerred carbon content and alloy range, can be modified by contro~ing the carbon potential o~ the furnace atmosphere during austenitizing. ~xample I shows that by using a 9~9 carburizing atmosphere for austenitizing treatments o~ about 1-2 hours at 815 to 850C, with a 159C temper, will produce compressive residual stresses to a depth of .2-.4 mm below the surface with a maximum surface compressive stress in the ~ 5 order of 70-135 MPa (10-20 SKI). In addition to compressive ; surface residual stresses~ the inventive method increases the amount of retained austenite and the volume fraction of primary carbides at the surface. The increase in surface retained austenite, particularly since the increase is accomplished without coarsening the austenite grains or reducing the hardness, is beneficial to increased contact fatigue life.
~ or short treatments, the depth of carburizing is quite shallow. For example, in the t~ hour treatment of Samples 7 and 9, the depth of the compressive layer is about .016". The amount of metal removed in finishing the bearing components o~ which these substrates may ~e employed, a~ter heat treatment, must be within this thickness, and . ~
~ preferably no more than .002-.004". This is necessary to ; 20 maintain the bene~it of the compressive stresses~

EXAMPLE III
.
The rolling contact ~atigue lives o~ 52100 steel samples were tested. The heat trea~ment was the same as for Sample 12, Table I, except that a tempering temperatuxe of 175C
was used. Group I samples were copper-plated during treatment to prevent carburization: Group II samples were unplated, therefore carburized. The test procedure employed a slmulative test procedure requiring special machines such as that made by Polymet Corp., Model RFC-l in which test bars of steel are tested to fatigue destruckion. The complete test procedure is set forth more clearly in U.S.
Patent 4,02~,988, Column 3, lines 32-68 and Column 4~ lines I-10. For the immediate test, a maxlmum hertzian rolling contact stress of 503 MPa (729,000 psi) was employed. The results axe summarized in Table IV, the st~tis~ical signifi cance of the results was tested by using the nonparametrical ~9~9~

Walsh test described on pp. 83-87 of "Non-Parametric Statistics", S~ Siegel, McGraw Hill, New York, 1956. The Walsh ~est was employed because the alternative, Johnsonls Confidence Method (described in a paper by L. J. Johnson, Industrial Mathematics, vol. 2, 1951, pp. 1-9), is not very accurate or convenient to use at Weibull slopes over 3Ø
The net result of such statistical te~ting was that in 99.5% of the cases, the life of the specimen with compressive surface stresses can be expe~ted to exceed that o~ conven-tional specimens. In ~act, in those cases where enhanced compressive surface stresses are developed, there is a 50% fatigue life increase when compared to a base line group of samples which were not subjected to a carburizing treatment such as in current bearing production. ~he fatigue life improvement extends over the entire range from B-5 to B-50 and beyond.
It is believed that fatigue life is improved by the processing herein because of several factors; (a) residual compressive stresses at the surface, (b) more retained austenite at the surface, (c) a hi~her surface hardness, and (d? a larger volume fraction o~ carbides near the surface. A11 of these factors result from a car~n gradient normal to the surface, and the first two result from a gradient in dissol~ed carbon in austenite. Whethex one of ~
these factors, or all o~ them in combination, are responsible for the contact ~atigue life improvement of this invention, is not known.
The mechanisms by which a dissolved carbon gradient is developed in a hypereutectoid steel ~ere outlined in theory above l~a) increasing carbon content of the alloy system causes an increased carbon solubility in austenite and ~b~ slowness of carbon distribution between carbide and austeni~e]. These mechanisms have been illustrated by the experiments described in the first two examples. In chromium bearing hypereutectoid steels both mechanisms can operate because:
(1) for low alloy compositions, as ~he carbon content of the r-Cr-Fe-C system incxeases, the solubility of carbon in ; austenite~increases, and ~2) relatively large spheroidized carbides, rich in chromium, are slow to dissolve at low austenitizing temperatuxes. In other systems, the Mn-Fe-C
~; 5 system for example, the flrst mechanism w3uld not be - expected to operate, because, according to Ro Benz, J. Fo~
Elliott and J. Chipman, Metallur~ical Transactions, VQl. 4, 1973, pp. 1975-86, increasing the carbon content of the Mn-Fe-C system does not significantly increase the solubility of carbon in austenite for hypereutectoid steels.
~he second mechanism would operate; thus, shallow surface compressive residual stresses of some magnitude could in theory be developed by short time austenitizing treatments in a carburizing atmosphere. In plain carkon hypereutectic steels, carbides dissolve so rapidly that neither mechanism could be expected to produce surface compressiv residu~l stresses.

, ~
.
:

9~
~` :

T BL~_l Sum~ary o`f Experimental Results :
Sa~ile H~at Treatmant ~: 1* : : A--85~C/l:hr, OQ
~- 2 : : Same:
3* ~ 875C/i hr; OQ; LNQ; T-100C/
.~ 1 . 5 hr s : Same A-850C/30 min~ OQ; LNQ; T-150C/
1.5 hrs

6 ;~-850C/1 hr; OQ; LNQ; T-150C/
1.5 hr:s ~ ~

7 . ::~--8S0C/2 hrs; OQ; LNQ; T-150C/
1.5 hrs

8 A-8 00C/l hr; OQ

9 ~ A-800C/2 hrs; OQ ~ :

10* ~ ~-980Cf35 min; AC; A-8ï5C/55 min; oQ; T-150C/l hr

11** Same ::
}2 A-815C/55 min; OQ; T-150C/ 1 hr ,, :
.
. *Copper-plated specimens ~: ** Copper plate removed after 980C treatment ~: A: Austenitize : : 0 : ~ OQ: Quenched~:in ~5C oil : :
~: LNQ: Quenched in liquid nitrogen ~- T: Temper ~-~ : AC: Air cooled .

' ` ':

~149~

,~ :
.

O Q O
~ ~1 OD O
,'~

O
O
~D

Il~ O O
~D O
O O ~
:

~ ~ H O O O O
., ~ ~,1 m ~

'` In In U~ O
rl ~ ~ N ~7 51~ D
~ ~ O O O
~.~
O O O
O O O

' `

O O O
O 1~ O~

~ ~ ~ U~ U~
U~ ~ ~1 ~ ~

9~ .
P~ ~ PC
a~u ~
K ~ n O O U'~ U~
u) a~

C~ O
U ~ ~ ~ o ~ o :1: h o ~o U~

U~
Q) U~

~ o O o I o o O ~
al o o o ~ ~.
~8 a u H
h .~;
m ~ ~ .~ rl .1 U ZO OE~
~ u~ o o o o 4~ a o ~o o o o h L
, ~ ~ O ~1 U~

o _I

a~

t 2 ;2 9~
s~J r~
o U~
~ ~D 1~ ~r , . ~ CO ~ I~
,. ~
, ~n u~ I ~ ~ r-~ m 3 ~
O ~
~1 t~
u~ ~ a~ o 1` ~
~n ~ ~ ~ ~ ~1 ~ . ~
~ ~1 .
, U~
:~ ~ o ~
U~ ' ~
~:

3 ~a ~ 3 ~ -I ~ er O
V
O Id ~J ' ~ O ~ ~ ~ V

. ~

u Q~
~` ~ o ~ O ~ ~ $
~ ~ o nl-~
rl h O
~q H 1-1 g H

rl O

O
O h ~ H s~
L U P~ H 3 R.
cn

Claims

The embodiments of the invention in which an exclusive property or privilege is claimed are defined as follows:

1. A method of developing compressive residual stresses in the surface region of a high carbon steel alloy article, said method comprising:
(a) constituting said steel alloy article to contain .8-1.6% carbon, .2-5% chromium, and 0-20% alloying ingre-dients selected from the group consisting of manganese, vanadium, molybdenum, tungsten, silicon, the remainder being iron;
(b) heating said article to a temperature of 800-950°C
(l472-1742°F) for 1-2.5 hours in a carburizing atmosphere effective to generate a differential in retained austenite, primary carbides and carbon between the surface region and core region of said article; and (c) immersing said article in a cooling medium to quench the central core of said article at a rate sufficiently fast to effectively suppress the formation of non-martensitic austenite decomposition products, thereby establishing a residual compressive stress gradient proceeding from the surface region of said article to a depth between 0.007 -0.03 inches, the residual compressive stress being in the range of 5-40 Ksi.

2. The method as in Claim 1, in which the carburizing atmosphere of step (b) is constituted to be endothermic gas containing 3-10% methane.

3. The method as in Claim 1, in which said article is constituted by having 7% alloying ingredients with about 5%
Cr, so that when said heat treatment is carried out, at about 1650°F, a retained austenite gradient is established.

4. The method as in Claim 1, in which said article is constituted with about 17.5% allowing ingredients, with about 4% Cr, so that when heat treatment at about 1650°F is carried out, a hardness gradient is established.

5. The method as in Claim 1, in which step (c) is comprised of quenching to room temperature, and then additionally quenching in liquid nitrogen.

6. The method as in Claim 4, in which the hardness gradient provides a differential between the hardness of said core and the surface region of at least 2 Rc.

7. The method as in Claim 3, in which the retained austenite gradient provides a differential between the core and surface region austenite of at least 10%.

8. The method as in Claim 1, in which said heating is carried out at a temperature level of 815°C for a period of about 1 hour, and the carbon potential of said carburizing atmosphere is regulated to be sufficiently high to saturate a 0.0025" thick iron foil in 30 minutes.

9. The method as in Claim 1, in which the resulting article possesses a resistance to bending fatigue due to compressive residual surface stresses induced by a gradient in dissolved carbon of at least 10,000 psi in the surface region thereof and tensile stresses in the core thereof.

10. The method as in Claim 1, in which the resulting article is characterized by resistance to rolling contact fatigue having a B10 fatigue life of at least 4.5 million stress cycles at a maximum hertzian contact stress of 729,000 psi.

11. The method as in Claim 1, in which said steel is constituted to contain 1%C, 1.5% Cr, .35% Mn, 0.25 Si and the remainder Fe.

12. The method as in Claim 1, in which said alloying ingredients, other than chromium, are present in an amount of at least 0.50%.

13. The method as in Claim l, in which said heated and quenched article is subjected to a tempering treatment at 100-300°C for 1 - 2 hours.

14. A high carbon steel alloy article having gradients of compressive residual stress, and at least one of a carbon gradient and hardness gradient proceeding from the surface region of said article to its core, said article being characterized by a microstructure consisting essentially of tempered martensite, retained austenite and a carbide phase, said article having a chemical content con-sisting essentially of .8 - 1.6% carbon, .75 - 25% alloying ingredients including .2 - 5% chromium, the remainder being essentially iron, said article having a compressive residual stress level at its surface region of at least 10,000 psi, and tensile stresses at the article core, said article having a hardness differential between its surface and core of at least 2.0 Rc, and a volume fraction of primary carbides at its surface region of at least .18.

15. The steel alloy article as in Claim 14, which is further characterized by a B10 rolling contact fatigue life of at least 4.5 million stress cycles, and a B50 life of at least 8.0 million stress cycles with a hertzian contact stress of 729,000 psi.

16. The steel alloy article as in Claim 14, in which the volume fraction of retained austenite is about 25% at the surface.

17. The steel alloy article as in Claim 14, in which the region of said article extending from the surface to .004 mm contains oxides in the grain boundaries, the region from .004 mm - .008 mm is carbide depleted, and the region from .008 - .1 mm contains .18 volume fraction of carbides.

18. The steel alloy article as in Claim 14, which is further characterized by resistance to subsurface crack initiation at hard inclusions and the resistance to surface initiated cracking as a result of the high compressive stress distribution in its surface region.

19. A method of making bearing elements comprising:
(a) shaping a body of SAE 52100 steel into a desired bearing configuration, (b) heating said body in a carburizing atmosphere to a temperature of about 815°C for about 1 hour, (c) quenching said heated body in oil maintained at about 55°C, (d) tempering said body at 150°C for about 1.5 hours, and (e) finish grinding said heat treated product to a depth of less than .005", said resultant bearing being characterized by a microstructure consisting essentially of tempered martensite, a primary carbide phase which is at least .18 at its surface region and a retained austenite volume fraction of about 25% at the surface.

20. The method as in Claim 19, in which step (c) is further followed by an additional quenching operation employing liquid nitrogen.

21. The method as in Claim 19, in which the resultant bearing body is characterized by an average B10 contact fatigue life of at least 4.5 million stress cycles, and an average B50 life of at least 8.0 million stress cycles at a maximum hertzian contact stress of 729,000 psi.

22. In a method for preparing bearing components wherein a hot-formed shape of low alloy steel containing carbon in the range of .8 - 1.5% and containing alloying ingredients in an amount to achieve a hardening response from heat treatment throughout said shape, said alloying ingredients being selected from a group consisting of Cr, Mn, Ni, Cu and Mo, said shape being subjected sequentially to a spheroidizing-anneal treatment, a rough forming treatment, and a hardening austenitizing treatment, the improvement comprising: (a) carrying out said hardening-austenitizing treatment in a carburizing atmosphere having a carbon potential sufficiently high to saturate a 0.0025"
thick iron foil in 30 minutes and at a temperature level of between 800 - 900°C for a period of time between 1 - 2.5 hours to establish both carbon additions and an increase in the primary carbide and retained austenite phases of said steel in the surface layers of the steel; (b) establishing a compressive residual stress gradient in said steel by quenching said steel at a rate such that the core experiences a cooling rate of at least 300°F per second when it is passing through the temperature zone of 1300 - 700°F, said resultant article having a compressive stresses in the surface region which range between 5 -40 KSI and tensile stresses in its core.