Canale L.C.F., Mesquita R.A., Totten G.E. Failure Analysis of Heat Treated Steel Components

Подождите немного. Документ загружается.

affected (Ref 43). In addition, subzero treatment

increases the stability and induces microresidual

stresses in any remaining retained austenite (Ref

43). Although subzero treatment does decrease

retained austenite, in view of these complica-

tions, Parrish has suggested that optimization of

the following process variables be considered

before employing this process: quenchant tem-

perature, surface carbon content, steel compo-

sition, and the use of reheat quenching (Ref 43).

Increasing amounts of retained austenite will

produce corresponding decreases in tensile

strength, as shown by Fig. 17, although in-

creasing strains can lead to the transformation of

retained austenite to martensite (Ref 37). How-

ever, conﬂicting test results make it difﬁcult to

predict if the martensite formed by such strain-

induced transformations is beneﬁcial or not,

since one study reported by Parrish stated that

the strain-induced martensite was more ductile,

and another stated that the untempered marten-

site was more brittle.

Koistinen showed that the distribution and

magnitude of residual stresses in carburized

steels was governed by the amount of retained

austenite (Ref 44). Figure 18 illustrates the

magnitude of residual stress as a function of

the amount of retained austenite and position

in the case for carburized SAE 8620 and 5140

and carbonitrided SAE 1118. The maximum

compressive residual stress occurs at the posi-

tion where the ratio of martensite/retained aus-

tenite is maximum (Ref 37).

It is assumed, as Fig. 19 shows, that increasing

the retained austenite content will reduce the

low-stress, high-cycle fatigue limit of carburized

steel (Ref 37, 45, 46).

In a study using a ﬂexural four-point bending

fatigue test and carburized SAE 8620 steel test

specimens, it was shown that increasing the

Fig. 17

Dependence of stress for ﬁrst detectable plastic strain

(approx. 0.0001) on retained austenite content.

AQ, air quenched; OQ, oil quenched; T, tempered

Fig. 18

Residual-stress distribution and retained austenite

content in case-hardened steels

194 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:18PM Plate # 0 pg 194

retained austenite in the case resulted in longer

fatigue life and that fatigue life is directly pro-

portional to the square root of the grain size

0.5

) (Ref 47). Jeddi et al. also showed that

fatigue strength of carburized 14NiCr11 steel

improved with increasing retained austenite con-

tent in the hardened case (Ref 48). In this work, it

was shown that the level of improvement was

related to the microstructure and residual-stress

distribution within j200 mm of the surface. In

addition, although the improvements in fatigue

are attributable to uniform retained austenite

content throughout the microstructure, only a

maximum of approximately 40% of the retained

austenite transformed to martensite at any depth

due to the cyclic loading.

The impact fatigue resistance is also depen-

dent on the amount of retained austenite and the

level of applied stress (Ref 37). Figure 20 shows

that impact fatigue resistance actually increases

with increasing retained austenite at the highest

stress loading, while lowest stress loading

exhibited the opposite effect.

Increasing retained austenite content resulted

in corresponding improvements of carburized

SAE 8620 steel using an abrasive wear test

utilizing a pin-on-disk tribometer, as shown in

Fig. 21 (Ref 47).

Sliding wear tests were conducted on car-

burized SNCM21, which corresponds to SAE

8620 steel, and carburized SCM4, which corre-

sponds to AISI 4140. Although SCM4 is not

typically carburized for the study reported, it

was used to represent an example of high core

strength and case depth. The results are shown in

Fig. 22 (Ref 49). According to these results,

sliding wear resistance increases with increasing

retained austenite at a 40 kg applied load up to a

critical retained austenite level, which, for this

work, was approximately 30%, at which point

the wear resistance decreased.

The presence of retained austenite can also

exhibit a dramatic effect on the scoring resis-

tance of carburized steels. Kozlovskii studied

the scoring resistance of carburized 20Kh2N4A

(0.21% C, 0.62% Mn, 0.20% Si, 3.50% Ni,

Fig. 19

Fatigue limits of plasma- and gas-carburized test

specimens as a function of retained austenite content

Fig. 20

Effect of retained austenite on impact fatigue resis-

tance of a carburized 1.45C-11.5Cr steel

Fig. 21

Effect of retained austenite (RA) on abrasive wear.

Sample A, HRC = 59.7+1.8, RA = 37; sample B,

HRC = 62.7+1.2, RA = 6%; and sample C, HRC = 61.4+1.5,

RA = 23%

Sources of Failures in Carburized and Carbonitrided Components / 195

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:18PM Plate # 0 pg 195

1.42% Cr) using a roller machine test and found

that even a relatively small amount of retained

austenite could exhibit a large decrease in scor-

ing resistance, even if the hardness was affected

only minimally, as shown in Table 8 (Ref 50).

Residual Stress

Residual stress is deﬁned as a tensile or

compressive force within a material such as steel

without application of a thermal gradient or an

external force. Residual stresses are produced by

phase transformation, plastic deformation, or

thermal effects such as contraction upon cool-

ing. Newton’s laws require that compressive

residual stresses at the surface of a material are

balanced by tensile stresses within the material.

Typically, compressive residual stress exhibits

favorable effects such as improved fatigue life

and stress corrosion by inhibiting crack initiation

and propagation. Conversely, tensile stresses

reduce desirable mechanical properties such as

fatigue, fracture, and wear.

Residual stresses are classiﬁed as the mac-

rostress or residual stress of the ﬁrst kind, which

acts over a few grains, residual stresses of the

second kind, and residual stresses of the third

kind. Residual stress of the second kind is the

difference between the average residual stress

with a grain and the residual stress of the ﬁrst

kind. Residual stress of the third kind refers to

stress variations within a grain. Residual stresses

of the second and third kind are microstresses

(Ref 51). In most engineering materials, such as

steel, residual-stress variation between micro-

structural phases is typically more important

than microstresses. The primary focus of this

chapter is on macrostresses, residual stresses of

the ﬁrst kind.

Carburizing and carbonitriding introduce

surface and subsurface compressive stresses as a

result of the formation of a carbon-enriched

case. The increased carbon content in the case

relative to the core signiﬁcantly reduces the M

temperature in the case, as illustrated in Fig. 23

(Ref 25). Because of the depressed M

tem-

perature, austenite-to-martensite transformation

begins in the core before the surface, even

though the surface temperature is lower. The

volumetric expansion of the martensite in the

core can be accommodated by the relatively hot

untransformed austenite nearer the surface.

Upon further cooling, the temperature at the

surface is less than the M

temperature of the

carbon-enriched case, and it begins to transform

to martensite. The martensite that formed ﬁrst in

the core is cooler and stronger than the austenite

that is now transforming, and it resists the

expansion of the higher-carbon-containing sur-

face martensite now forming at the surface,

which puts the surface in compression relative

to the core. To maintain balance, the core is

now in tension. This is illustrated by Fig. 23

(Ref 25).

Fig. 22

Sliding wear rate (at 200 rpm) as a function of

retained austenite content. A, carburized SNCM21,

40 kg load, sliding distance of 864 m; B, carburized SCM4, 40 kg

load, sliding distance of 864 m; C, carburized SNCM21, 20 kg

load, sliding distance of 1728 m; D, carburized SCM4,

20 kg load, sliding distance of 1728 m

Table 8 Effect of retained austenite on scoring resistance of carburized and carbonitrided case

Treatment

Case carbon

content, %

Case depth,

Surface

hardness, HRC

Retained sustenite

Scoring

load, kg/cm

Compressive

stress at

scoring

load, kg/cm

surface

At depth

of 1.5 mm

Carburize at 930



C for 8 h, slow cool to 750



quench in oil, temper at 180



C for 1.5 h

0.81 1.1 57–59 48 60 930 12,020

Carburize at 930



C for 8 h, slow cool to 20



temper at 600



C for 2 h, quench from 830



in oil, temper at 180



C for 1.5 h

0.84 1.1 60–62 8 16 355 7424

196 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:18PM Plate # 0 pg 196

An evaluation was made of 70 steels that were

gas carburized, oil quenched, and tempered

between 150 and 180



C. Case depths were

1 mm or less, and the core carbon contents

varied between 0.15 and 0.20%. The range of

residual-stress proﬁles obtained is shown in

Fig. 24, where it can be observed that the com-

pressive residual stresses acted throughout most

of the case (Ref 25).

It is important to consider the history of resi-

dual stresses due to dimensional changes that

may occur during any stage of the manufactur-

ing process and therefore may contribute to the

ﬁnal residual-stress condition of the heat treated

component. These contributions to the stress

history are illustrated in Fig. 25, which is a

simulation of the production of a carburized

8620 steel cylinder that was subsequently

quenched in unagitated water (Ref 52).

The development of residual stress was

achieved in the case of a 9.5 mm diameter

chromium-molybdenum SCM420 (0.23% C,

0.72% Mn, 1.12% Cr, 0.21% Mo) steel test

specimen that was gas carburized at 930



C until

the case depth shown in Fig. 26 was achieved, oil

quenched, and then tempered for 1 h at 200



(Ref 53). These data show that the compressive

residual stress increases with increasing case

depth. Generally, the magnitude of the surface

compressive stresses will be dependent on

the ratio of the case and core thickness. When

the core is thicker than the case, the surface

Fig. 23

Schematic illustration of carbon content, retained

austenite, and residual stresses in the case of car-

burized steels

Fig. 24 Range of residual stresses obtained for 70 carburized steels

Sources of Failures in Carburized and Carbonitrided Components / 197

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:18PM Plate # 0 pg 197

compressive stresses will be high. When the case

is thicker than the core, the surface compressive

stresses in the case will be lower, and the tensile

stresses in the core will be higher (Ref 36). For

automotive gears, the inversion point occurs

where the surface compressive stresses become

tensile, which was shown to occur at the point

where the hardness becomes equivalent to the

core value (Ref 54). Furthermore, the formation

of surface compressive stresses was shown to

be fundamental to the prevention of fatigue

cracking.

One of the primary reasons for conducting the

carburizing process is to improve the fatigue

performance of a heat treated component.

A surface compressive stress will inhibit the

formation and growth of surface cracks, which

is important since fatigue failures are typically

initiated at the surface (Ref 55). Kanetake, for

example, showed an approximately 40% im-

provement in fatigue strength of the carburized

SCM420 steel (Ref 53). Shot peening is also

used to create residual surface compressive

stresses that will increase fatigue properties of

steel components (Ref 56). Therefore, it has

been of continuing interest to examine the

potential of fatigue property improvements not

achievable by either treatment alone (Ref 55,

57–59). For example, Shaw et al. have reported

up to a 75% increase in fatigue strength by

combining carburizing and shot peening of gears

prepared from gas-carburized 20MnCr5 steel

(Ref 57). Before continuing this discussion, a

brief overview of the shot peening process is in

order.

Shot peening (and shot blasting) involves im-

pinging the surface to be treated with spherical

media called shot. (Shot peening should be dif-

ferentiated from shot blasting. Shot blasting a

process in which an abrasive material is accel-

erated through a pressurized nozzle or cen-

trifugal wheel and directed at the surface of

a part to clean or otherwise prepare the part

surface for further treatment, (Ref 60)). Shot

peening is a cold working process where each

individual spherical ball impinging the surface

acts as a miniature hammer that plastically

deforms and work hardens the surface by

creating a small indentation upon impingement,

as illustrated in Fig. 27 (Ref 61). The indentation

process causes the surface to yield in tension. To

balance the tensile forces involved in indenta-

tion, the subsurface is in a highly stressed

compression state. As the process continues, the

indentations overlap, and a uniform layer of

metal is in residual compressive stress. The

compressive stress is the result of superposition

of residual stress formed by surrounding shots

(Ref 62). The magnitude of the compressive

stresses that are formed is material dependent

Fig. 25

Illustration of the tangential stress history over the ﬁrst 20, of a water-quenched 1.27 cm diameter 8620 carburized steel

cylinder. The carbon gradient and retained austenite content are shown in Table 9.

198 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:18PM Plate # 0 pg 198

and is generally at least

the yield strength of

the material (Ref 56). Some of the variables of

the shot peening process include the shape and

type of shot, size of shot, and impingement

velocity.

The effect of shot peening on contact fatigue

under rolling-sliding conditions was studied

using a carburized and hardened 20CrMnTi steel

(0.21% C, 1.09% Cr, 1.0% Mn, 0.025% Ti,

0.02% P, 0.008% S), and, as a result of that work,

it was shown that failure may produce three

kinds of cracks, which were classiﬁed as: sur-

face, which initiate from 0.015 to 0.05 mm;

shallow sub surface; and deep subsurface, which

initiate from 0.3 to 0.5 mm (Ref 63). Depend-

ing on the speciﬁc contact stresses on the

shot-peened surfaces, cracking, pitting,

shallow spalling, and deep spalling may occur.

In addition to providing substantial improve-

ments in fatigue strength, carburizing and shot

peening offer other beneﬁts, such as reducing the

deleterious effects of internal oxidation. How-

ever, in the absence of surface oxidation and

oxide inclusions, MnS inclusions will then act as

fatigue initiation sites (Ref 57).

If a case-hardened surface is shot peened with

sufﬁcient intensity, a stress-induced transfor-

mation of retained austenite to martensite may

be observed, and the surrounding volume con-

straint may result in a deepening of the surface

compressive residual stress (Ref 55, 58). Nako-

nieczny et al. (Ref 58) have shown that plastic

strain induced by shot peening reduces retained

austenite in tempered martensite and produces

a new e-phase (Fe

C and Fe

C). Peyrac, in

another study, also showed signiﬁcant retained

austenite transformation and increases in

residual stress as a result of shot peening of gas-

carburized 18NCD6 steel (Ref 64). Selected

results from this study are summarized in

Table 10.

It is known that surface structure anomalies,

also known as soft skin layers, which include

internal oxides and nonmartensitic structures,

including retained austenite (Ref 65) near the

surface, will decrease fatigue properties of

Fig. 26

Residual-stress proﬁles of SCM420 steel that was gas

carburized at 930



C, oil quenched, and tempered at

200



Fig. 27

Illustration of the plastic deformation of the surface

and resultant stress distribution after shot peening.

Source: Ref 61

Table 9 Carbon gradient and retained austenite

in the 8620 carburized steel related to Fig. 25

Depth below the surface

Carbon, %

Retained

austenite, %mm in.

0.0 0 1.20 32.0

0.5 0.020 1.10 28.0

1.0 0.040 0.80 15.0

1.5 0.060 0.40 4.0

2.0 0.080 0.25 0

Sources of Failures in Carburized and Carbonitrided Components / 199

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:19PM Plate # 0 pg 199

gas-carburized steel (Ref 10). Structural surface

anomalies act as preferential zones for fatigue

crack initiation under low stress amplitudes.

Shear-type crack growth can occur in ductile

retained austenite in the near-surface region of

the case for gas-carburized steels. Kikuchi et al.

have shown that shot peening is very effective in

improving the fatigue properties of carburized

steels with these surface structure anomalies

(Ref 59, 65). In fact, in their study, the fatigue

properties were essentially comparable for a

20 mm diameter chromium-molybdenum steel

(0.16% C, 0.26% Si, 0.74% Mn, 0.012% P,

0.013% S, 1.01% Cr, 0.18% Mo) bar that was

carburized at 950



C for 1 h with and without

surface structure anomalies (to a depth of ap-

proximately 30 mm) after shot peening. As a

result of this work, it was concluded that since

internal oxides near the surface can also act as

preferential crack initiator sites, it is desirable to

avoid the presence of such defects during gas

carburizing (Ref 65).

Although components such as gears are car-

burized and shot peened to introduce the desired

level of surface compressive stresses, there are

other processes that can be used to introduce

compressive stresses. One such process is pre-

setting. Presetting involves the introduction of

an overload that causes yielding in the area of

maximum stress concentration, such as in the

root area of a gear tooth. When the load is

released, a residual stress is introduced in that

area. The induced stress is compressive on the

side being loaded and tensile on the other side

(Ref 66). Woods et al. evaluated presetting to

improve the bending fatigue of carburized AISI

4120 steel spur gear teeth and found that pre-

setting introduces compressive stresses in

the area of a gear tooth where fatigue cracks

originate. The results of this work showed

that presetting provided substantially longer

fatigue life.

Dimensional Stability

Dimensional stability has two components:

size (distortion) and shape (warpage). Distortion

is deﬁned as “an irreversible change in the

component during heat treatment” (Ref 67).

While changes in shape such as straightness

(warpage) can be corrected by application of

stress or by tempering in the elastic range

(reversible), size changes are irreversible and

cannot be changed in this way. Metallurgically,

distortion may be thermally or transformation-

ally derived. Size distortion typically refers to

dimensional variation due to growth or shrink-

age that is due to volumetric changes attributable

to microstructural phase transformations (Ref

19, 68). Figure 28 shows the effect of the tem-

perature dependence of the speciﬁc phase

volume of different steel transformation phases

(Ref 69).

Variables that affect distortion include (Ref

19, 68, 70):

 Chemical composition and hardenability—

chemical and phase composition as well as

hardenability (distortion increases as hard-

enability increases), as shown in Fig. 29

(Ref 67, 71, 72)

 Steelmaking—grain size and hardenability

 Hot working—hot reduction, length and

direction of ﬁber

 Prior heat treatment—residual stresses,

grain size and uniformity of microstructure

 Geometry—cheese blanks, shafts, rims

 Heat treatment aspects—heat rate, cooling

rate (quench severity), quenching tempera-

ture, jigs and ﬁxtures, plug quenching. Plug

quenching is used to minimize dimensional

change of inside diameters such as round-

ness and taper distortion of ring-shaped

components (Ref 67, 68) (generally, salt

bath quenching yields minimum distortion

Table 10 Effect of shot peening on retained austenite transformation and residual stress of

gas-carburized 18NCD6 steel

Shot peening(a)

Heat

treatment(b)

Depth modiﬁed by

shot peening treatment, mm

Retained austenite

on surface, %

Retained austenite

converted to martensite, % s

max

, MPa

None T1 ... 24.7 ... 300

T2 ... 36.8 ... 300

G1 T1 100 5.7 19.0 1450

T2 100 14.9 21.9 1350

G2 T1 200 12.1 12.6 980

T2 200 22.1 14.6 930

(a) G1, steel shot, BA 300, F 25–30A, overlap rate 150%; G2, steel shot, BA 800, F 55–80A, overlap rate 150%. (b) T1, carburize for 3 h. at 920



C, plateau at 850



oil quenching at 60



C, tempering for 2 h. at 150



C; T2, carburize for 3 h at 960



C, oil quenching at 60



C, tempering for 2 h at 150



200 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:19PM Plate # 0 pg 200

relative to hot oil) (Ref 67). Furnace tem-

perature uniformity, case depth uniformity,

prequench temperature, ﬂuid ﬂow during

the quench (Ref 73), number of times a

part is quenched, carburizing temperature

(Ref 70)

 Racking—Vashchuk et al. reported that

high-temperature deformation of large gears

could occur due to their own weight (Ref

73). They recommended the larger diameter

of a gear wheel should be on the ﬂoor of the

furnace (bottom) to support the end face and

that the smaller diameter should be the free

surface

 Machining—Parts should be machined as

near ﬁnal dimensions so that outer case

will not require grinding after carburizing

(Ref 24, 68). Residual stresses due to prior

machining exhibit a large effect on distor-

tion, and as shaved thickness increases,

potential distortion increases (Ref 67)

 Method of green part manufacture, for

example, parts machined from bar or tubing

or forged from bar or tubing and then

machined (Ref 70)

 Growth of surfaces during carburizing

 Murzin et al. reported that the cooling rate

from the carburizing temperature to the

prequench temperature exhibited one of the

Fig. 28 Variation of the speciﬁc phase volume of different steel transformation phases as a function of temperature. Source: Ref 69

Fig. 29 Effect of steel hardenability on shape distortion

Sources of Failures in Carburized and Carbonitrided Components / 201

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:19PM Plate # 0 pg 201

strongest effects on gear wheel distortion

(Ref 74).

Variables that affect warpage include (Ref 19,

68, 75):

 Stresses resulting from phase transforma-

tions

 Nonuniform residual stresses in the original

blank, such as those due to prior heating,

including stress relief

 Nonuniform heating or cooling—furnace

shape, part shape, heat control

 Insufﬁcient furnace time—undersoaking

 Creep—method of stacking and ﬁxturing of

parts during heating and quenching (hanging

versus standing)

 Internal stresses due to machining

Bulgakov reported that the most effective

method of controlling stresses to reduce or

stabilize warping was by controlling the hard-

enability of the steel (Ref 75). Hardenability

control will permit control of the phase trans-

formation and reduction of volumetric changes

leading to structural stresses and warping. In par-

ticular, it was shown that slow cooling of car-

burized steels in the range of 1100 to 900



after forging was extremely undesirable since

this facilitates the formation of hard-to-dissolve

particles and austenite grain growth during heat

treatment. Optimal warpage control was ac-

hieved with grain sizes not greater than grade 7.

A comparison was made of the effect of

transformation behavior on dimensional stabi-

lity of machine parts, such as gears, constructed

from two different carburized steels: 20KhGR

(0.18% C, 0.82% Mn, 0.24% Si, 1.07% Cr,

0.20% Ni, 0.0031% B) and 12KhN3A (0.14% C,

0.45% Mn, 0.21% Si, 0.78% Cr, 0.85% Ni).

Since only very low cooling rates occur during

carburizing, thermal stresses would be expected

to be minimal, and any stresses that result would

be due to austenitic transformation, which would

be dependent on the steel chemical composition

(Ref 76).

The isothermal transformation diagrams for

both steels before and after gas carburizing are

shown in Fig. 30 (Ref 76). These diagrams

show that austenite is less stable in the 20KhGR

alloy and that relative stability of austenite

compared to the 12KhN3A alloy remains after

carburizing.

Analysis of the kinetics of austenite transfor-

mation of alloys, before and after carburization

Fig. 30

Isothermal transformation diagrams. (a) 20KhGR and (b) 12KhN3A alloys before carburizing (c) 20KhGR and (d) 12KhN3A

alloys after carburizing. Source: Ref 76

202 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:19PM Plate # 0 pg 202

(Fig. 31) over the range of cooling rates used

(0.79 to 175



C/min), showed that the non-

carburized 20KhGR yielded mostly pearlite.

The 12KhN3A alloy yielded predominantly

bainite at the higher cooling rates, and at lower

cooling rates, the pearlitic transformation

predominated for both the carburized and

uncarburized alloy.

When air cooled, the case of the carburized

20KhGR consisted of pearlite and carbides, and

Fig. 31

Dilatometric curves for the transformation of austenite in 20KhGR and 12KhN3A steels. Curves (a) and (b) were after

carburizing, and curves (c) and (d) were before carburizing. (a,c) 20KhGR. (b,d) 12KhN3A. The cooling rates are: 1, 0.79; 2,

1.46; 3, 4.6; 4, 5.0; 5, 70; 6, 175



C/min. Source: Ref 76

Sources of Failures in Carburized and Carbonitrided Components / 203

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:19PM Plate # 0 pg 203