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

Подождите немного. Документ загружается.

being carburized does not reach equilibrium with

the carburizing atmosphere. This is illustrated in

Fig. 6. A different, but similar, correlation exists

for various steel alloys and carburizing con-

ditions. From a correlation such as Fig. 6 and

given case depth requirement and carburizing

time, it is possible to correct for the carbon

potential under nonequilibrium conditions to

achieve a speciﬁc surface carbon content, as in-

dicated by the dashed lines in the ﬁgure (Ref 22).

For a speciﬁc temperature, the case depth (d)

will vary with the square root of the carburizing

time (t):

d=Qt

1=2

Values of the depth factor (Q) as a function of

temperature are shown in Fig. 7. This equation is

reported to be valid for low-carbon steel and

some alloy steels (Ref 22). For applications such

as automotive gears, typical case depths are 0.8

to 1.4 mm (Ref 23).

An equation that is often used to calculate the

case depth (in.) as a function of both time and

temperature is the Harris equation (Ref 24):

Case depth (in:)=½31:6 · t

0:5

=10

(6700=T)

where t is the time at the carburizing tem-

perature in hours (carburizing time), and T is the

absolute temperature [Kelvin (K) =



C+273,

Rankine (R) =



F+460]. In metric units, the

case depth (mm) is (Ref 24):

Case depth (mm) =660 · e

8287=T

· t

0:5

At the operating temperature, the carburizing

process may be conducted in two parts. Car-

burizing occurs during the ﬁrst part of the pro-

cess in a high-carbon-potential period when the

enriching gas is added to the furnace atmosphere

to increased the carbon content of austenite (the

carburize-boost period) and the carbon potential

is greater than the desired carbon potential. This

part of the process is typically conducted at a

carbon potential close to the solubility limit of

carbon in austenite, typically between 1.0 and

1.2% C, which is dependent on the temperature

and alloy content of the steel. The time for this

part of the process to occur is called the car-

burizing time. This part of the process is fol-

lowed by a boost-diffuse period, where the

process is operated at the equilibrium carbon

potential, which is reduced to a level that will

maintain surface content, typically 0.8 to 0.9%

C, during which time the carbon will diffuse

deeper into the case and provide a gradual case/

core transition. Together, this is called the boost-

diffuse cycle. The time for this part of the pro-

cess is called the diffusion time. When the

required case depth is achieved, if the compo-

nent is direct quenched, the temperature is

Fig. 6

Carbon dioxide content of the atmosphere required to

produce certain surface carbon levels at different car-

burizing times under a given set of carburizing conditions. The

dashed lines illustrate alternative times and carbon dioxide

contents to produce a single surface carbon content.

Fig. 7

Variation of the depth factor, Q, with carburizing

temperature for low-carbon and certain alloy steels

184 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:17PM Plate # 0 pg 184

lowered to 850



C to reduce the distortion and

then quenched (Ref 25).

Carburizing boost-diffuse cycles conducted

over 2 h are advantageous for case depths

40.50 mm (0.020 in). They are also useful

when relatively deep cases free of carbides or

retained austenite are required. This is important

when the carbon content is greater than the

eutectoid composition, where there is an in-

creased tendency to form carbides and retained

austenite upon quenching. These effects in-

crease with alloy content (Ref 24).

Harris also developed equations to compute

the carburizing and diffusion times to achieve a

speciﬁc case depth and surface carbon content

(Ref 24):

Carburizing time (h)=(C C

)

=(C

 C

)

Diffusion time (h)=Total time Carburizing time

where C is the ﬁnal desired surface carbon

content, C

is the surface carbon content at the

end of the carburizing cycle, and C

is the carbon

content at the core.

The effect of the steel alloy composition on

the carbon gradient is illustrated for AISI 1020

plain carbon steel and AISI 8620 after carbur-

izing at three temperatures in Fig. 8 (Ref 24).

The alloy content will inﬂuence the diffusion

rate, but its greatest effect is on the case carbon

content. Normal carbon gradients, such as those

shown in Fig. 8, can be achieved by maintaining

a saturated austenite condition at the surface

during the entire boost-diffuse carburizing cycle

(Ref 24).

It is important to control the ratio between the

boost and diffuse times and to carefully control

the carbon potential to avoid obtaining a carbon

proﬁle such as that shown in Fig. 9 (Ref 24).

Although the desired surface hardness was

obtained, the lower carbon content at the surface

can lead to a transformation that proceeds simul-

taneously outward from the case-core interface

and at the surface and proceeds inward such that

the last portion of the case to transform is just

below the surface. This will result in an unde-

sirable condition where the surface is in tension

relative to the core as well as a corresponding

decrease in fatigue strength in addition to an

increased potential for cracking (Ref 24).

Boyer reports that a maximum tolerable car-

bon potential for carburizing cycles of up to 10 h

at 925



C is 1.3% to avoid excessive soot for-

mation.

If excessively high carburizing temperatures

are used, the following situations may occur:

 Rapid increase in grain growth and loss of

properties

 Increased energy consumption

 Increased deterioration of the furnace ﬁx-

tures and baskets

When high carbon potentials and long car-

burizing times are used to produce high surface-

carbon content and deep case depths, excessive

retained austenite and/or free carbides may be

obtained as a result. These microstructural pro-

ducts exhibit adverse effects on residual-stress

distribution (which is discussed subsequently).

Therefore, although high carbon potentials may

be used for short carburizing times, substantial

deleterious effects may result if used over pro-

longed carburizing times.

Excessive carbon potentials, gaseous atmo-

sphere composition control problems because of

carbon probe malfunctions or air ingression, and

inadequate furnace purging can lead to excess

free carbon and sooting, which may be sufﬁ-

ciently severe as to leave carbon deposits on the

parts (Ref 24, 26). This can lead to correspond-

ing problems in controlling the carbon potential,

resulting in nonuniform carburizing and dimen-

sional control problems.

Quenching. During carburizing, the steel

microstructure consists of polycrystalline aus-

tenite. Grain coarsening may occur if the car-

burizing times are relatively long. The austenitic

grain size will determine the size and distribu-

tion of martensite that will form as a result of

quenching. In addition, phosphorus segregation

into the grain boundaries may occur during the

carburizing cycle, which has been found to be

directly dependent on phosphorus and carbon

content. Hyde et al. found that fatigue and

fracture toughness decreased slightly when the

phosphorus content increased from 0.005 to

0.017%, and when 0.017 to 0.031% P was pre-

sent, the endurance limit and fracture toughness

decreased substantially (Ref 27).

Phosphorus also affects the degree of carbon

segregation in the form of cementite at the aus-

tenite grain boundaries. During quenching,

small amounts of cementite form at the auste-

nite grain boundaries in the high-carbon case

(Ref 25, 27). This leads to increased sensitivity

to intergranular fracture, which is a major cause

of fatigue crack initiation in carburized steels

(Ref 25).

Sources of Failures in Carburized and Carbonitrided Components / 185

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:17PM Plate # 0 pg 185

Fig. 8

Carbon gradients for gas-carburized 1020 and 8620 steels. The 1020 steel was carburized in a batch furnace, and the 8620 was

carburized in a pit furnace.

186 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:17PM Plate # 0 pg 186

After carburizing, the parts are then either

quenched directly or air cooled and then rehe-

ated and quenched. Quenching is performed to

harden the components. It is most desirable to

develop a martensitic or bainitic case while

controlling the formation of retained austenite to

an acceptable level and simultaneously mini-

mizing proeutectoid and pearlitic structures. The

challenge is to quench sufﬁciently fast to pro-

duce the desired core structure but not so fast

that the higher-carbon and more brittle case

cracks. In addition, the desired hardness gradient

between the surface and the core is critically

important to achieve the desired wear and fati-

gue properties. If the carburized gear, such as a

spiral bevel gear, is not quenched to achieve the

necessary surface hardness and hardness gra-

dient, failures accompanied by micropitting and,

ultimately, fracture may occur (Ref 28).

The morphology of martensite is carbon

dependent, as shown in Fig. 10 (Ref 29). At

lower carbon content, a lath martensitic struc-

ture forms, while plate martensite forms at

higher levels of carbon. The two different mor-

phologies are illustrated in Fig. 11 (Ref 30). Lath

martensite exhibits better toughness than the

higher-carbon plate martensite. Plate martensite,

as the name indicates, forms as lenticular (lens-

shaped) crystals and is sometimes referred to as

acicular (meaning needlelike) martensite or

high-alloy martensite. A characteristic of plate

martensite is the zigzag pattern of smaller plates,

which formed later in the transformation,

bounded by adjacent larger plates that formed in

the beginning of the transformation (Ref 30).

Typically, quenching is performed either di-

rectly from the carburizing process after furnace

cooling to approximately the Ac

temperature

and then quenched, or the parts are air cooled

and then reheated and quenched. Less com-

monly, double reheat quenching may be per-

formed to provide high-durability components

(Ref 19). Some quenching cycles recommended

Fig. 10 Dependence of the martensitic structure on carbon content

Fig. 9 Carbon proﬁle of an incorrectly carburized steel

Sources of Failures in Carburized and Carbonitrided Components / 187

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:17PM Plate # 0 pg 187

by Crucible Steel for carburized 8620 steel

include (Ref 31):

 Direct quench from carburizing: Quench in

oil directly from the carburizing temperature

of 925



C (1700



F). The core is hardened

but unreﬁned. The case is hardened to the

extent that it will be ﬁleproof if the carbon

content is sufﬁciently high.

 Cool, reheat and quench (1): After cooling

from the carburizing temperature (925



or 1700



F), the carburized (but not yet

hardened) steel is reheated to a temperature

above the upper critical temperature, Ac

of the core (835



C, or 1535



F) and then

quenched in oil. The core will be reﬁned and

exhibit maximum strength and hardness.

The case will be hardened and somewhat

coarsened.

 Cool, reheat, and quench (2): After cooling

from the carburizing temperature (925



or 1700



F), the carburized (but not yet

hardened) steel is reheated to a temperature

above the lower critical temperature, Ac

,of

the case (730



C, or 1350



F) and then

quenched in oil to harden and reﬁne the case.

The core will be unreﬁned, soft, and machine-

able, and the case will be hardened.

 Double reheat and quench: The steel is

cooled in the furnace from the carburizing

temperature of 925



C (1700



F). The steel

is then reheated to above 730



C (1350



and oil quenched to reﬁne the core. The steel

is again reheated to 730



C (1350



F) and oil

quenched to reﬁne the case. This double

heating and quenching procedure reﬁnes

both the case and core grain size. The reﬁned

core will be soft and machineable with

maximum toughness and resistance to

impact. The reﬁned case will be hardened for

wear and resistance. In addition to grain

reﬁnement, double reheating and quenching

is reported to improve fatigue properties by

reducing the size and density of microcracks

in the structure (Ref 21).

Of these methods, the most common is direct

quenching. However, there are a number of rea-

sons why reheat quenching is favored for higher-

alloy, case-hardening steels, including (Ref 19):

 To assure grain size and retained austenite

control

 When intermediate subcritical heat treat-

ment is required to condition the carbide

structure within the case or to facilitate

additional machining

 When the parts are to be plug or die quen-

ched for distortion control

 When it is not possible to direct quench, such

as in pit carburizing

Ingham and Clarke compared the results

obtained for carburized 8620 steel with the

carbon gradient shown in Fig. 12, which was oil

quenched from a direct quench following car-

burizing and by a reheat and oil quench cycle

(Ref 32). The results obtained showed that the

direct quench process yielded a higher hardness

than the reheat and quench process, which ex-

hibited a relatively lower as-quenched hardness

due to the presence of bainite in the hardened

case structure.

Fig. 11 Martensite morphology. (a) Lath martensite. (b) Plate martensite. Source: Ref 30

188 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:17PM Plate # 0 pg 188

Quenchants must be selected to provide

cooling rates capable of producing an acceptable

microstructure and hardness gradient through

the case and the core. However, it is not de-

sirable to use quenchants with excessively high-

heat-removal rates, since the propensity to cause

increased distortion or cracking increases with

quench severity. Although a reduction of quench

severity leads to reduced distortion, it may also

be accompanied by undesirable microstructures.

Therefore, it is essential to select optimal quen-

chant and agitation conditions for the required

microstructure, hardness, and strength in critical

sections of the parts for each steel alloy, section

size, and required microstructural and mech-

anical properties. Actual cooling rates or heat

ﬂuxes provided by a speciﬁc quenching medium

are typically unavailable. However, some illus-

trative comparative data are provided in Table 4

(Ref 33). Figure 13 illustrates the comparative

cooling properties of various oil-quenched steel

bars assuming a surface heat-transfer coefﬁcient

of 0.019 cal s

1



1

(Ref 32).

Quench nonuniformity is a signiﬁcant con-

tributor to quench cracking. Quench non-

uniformity can arise from nonuniform ﬂow

ﬁelds around the part surface during the quench

or nonuniform wetting of the surface. Both lead

to nonuniform heat transfer during quenching,

creating large thermal gradients between the

core and the surface of the part. Poor agitation

design is a major source of quench nonuni-

formity, since the purpose of the agitation sys-

tem is not only to take hot ﬂuid away from the

surface and to the heat exchanger but also to

provide uniform heat removal over the entire

cooling surface of all of the parts throughout the

load being quenched.

A wide range of quench media can potentially

be used when quenching carburized parts. Some

comments on quench media selection, provided

by Boyer, include (Ref 34):

 For carbon steels, the most common

quenchants are water and brine. When water

is used as the quenchant, bath temperatures

of 20 to 30



C with agitation are the most

common.

 In the industry, oil quenchants are the most

common, particularly when integral-quench

(sealed-quench) furnaces are used at tem-

peratures of 25 to 70



C. The quench oils

may be classiﬁed as fast, intermediate, or

slow depending on the cooling rate, enhan-

cing additive, and quench oil base stock

being used. When distortion control is cri-

tical, a hot oil that can be used at tempera-

tures as high as 175



C may be used.

Fig. 12 Comparison of direct quenching and reheat and quenching of 10 cm (4 in.) diameter AISI 8620 steel after oil quenching

Sources of Failures in Carburized and Carbonitrided Components / 189

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:17PM Plate # 0 pg 189

 It is possible to use aqueous polymer

quenchants, even in integral-quench fur-

naces, if appropriate structural conditions

are met. The user is advised to consult his

furnace manufacturer prior to use. A wide

range of quench severities is possible by

varying the polymer concentration, bath

temperature, and agitation. In one study, it

was shown that an aqueous polymer quen-

chant produced substantial improvements

in fatigue properties relative to a conven-

tional quench oil, which was attributed to

an improvement in quench uniformity

(Ref 35).

 When distortion control is especially cri-

tical, salt bath quenching may be required.

However, parts should never be transferred

directly from a carburizing bath containing

45% cyanide to a nitrate-nitrite quench

bath, because this will result in a violent

reaction and possibly an explosion (Ref 34).

One often-encountered quenching problem

that may lead to increased dimensional control

problems is contamination. For example, het-

erogeneous quench media caused by water

contamination of oil or oil contamination of

water or aqueous polymer solutions can poten-

tially cause cracking problems. Similarly, salt

contamination, either from salt baths or hard-

metal ion contamination, can lead to problems of

cooling rate control. Solid contamination, such

as sludge or soot contamination in oil or aqueous

media, also may lead to distortion and cracking.

Finally, excessive foaming and air entrainment

of the quench media will lead to nonuniform

cooling, soft spots, increased residual stresses,

and cracking. Therefore, it is essential that the

quench bath be well maintained to assure opti-

mal distortion control and minimize the poten-

tial for cracking.

To develop the optimal residual-stress gra-

dient, it is important to use the proper quenching

Soaking temperature

Bar size, in.

Fig. 13

Centerline cooling curves for oil-quenched steel bars of varying section sizes, assuming a surface heat-transfer coefﬁcient of

0.019 cal s

1



1

Table 4 Comparison of typical heat-transfer

rates

Quench medium

Maximum surface

heat-transfer rate,

W m

2

1

Still air 50–80

Nitrogen (1 bar) 100–150

Salt bath or ﬂuidized bed 350–500

Nitrogen (10 bar) 400–500

Helium (10 bar) 550–600

Helium (20 bar) 900–1000

Still oil 1000–1500

Hydrogen (20 bar) 1250–1350

Circulated oil 1800–2200

Hydrogen (40 bar) 2100–2300

Circulated water 3000–3500

190 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:17PM Plate # 0 pg 190

conditions for the steel grade of interest.

Figure 14 shows the development of the tem-

perature distribution through 12.5 mm diameter

bars carburized to a depth of 0.9 mm for oil-

quenched alloy steel and a water-quenched mild

steel (Ref 36). This ﬁgure shows that the M

and

temperatures are reduced with increasing

carbon content in the case. The isochronal lines

in the ﬁgure illustrate the cooling proﬁle through

the case to the core at speciﬁed time intervals.

The martensitic transformation occurs at the

core/case interface ﬁrst. The case transforms last

along with the corresponding expansion of

martensite in the case. Since the core is already

transformed, this process restrains further ex-

pansion during continued cooling, leading to the

development of surface compressive stresses

and placing the core in relative tension (Ref 36).

As this ﬁgure shows, since the M

is less than

ambient temperature, retained austenite will

accompany this process.

Because carbonitriding is similar to carbur-

izing, except that nitrogen and carbon diffusion

into the case is involved, both processes exhibit

similar microstructural transformation and

evolution of residual stresses upon quenching.

Process temperatures during carbonitriding are

typically lower (800 to 850



C), as are process

times (30 to 60 min), which provides a relatively

shallow case, usually 50.5 mm (Ref 36).

These results show any factor that affects the

cooling proﬁle, such as bath temperature, agi-

tation, or the quenchant selection, will exhibit a

corresponding effect on the thermal distribution

through the carburized case upon hardening

and therefore on the development, type, and

magnitude of residual stresses. To assure opti-

mal distortion control, the following variables

should be carefully monitored and controlled:

 Adequate quality-control procedures of the

quench media should be in place. For

examples, follow ASTM D6710, “Standard

Guide for Evaluation of Hydrocarbon-Based

Quench Oils”, for oil quenchants and ASTM

D6666, “Standard Guide for Evaluation of

Aqueous Polymer Quenchants”, for aqueous

polymer quenchants.

 Carefully control the water content if poly-

mer quenchants are being used.

 Replace the quench mediate at regular use-

level intervals.

 Carefully control the quench bath tempera-

ture.

 Monitor ﬂuid ﬂow variation at critical loca-

tions in the quench tank. Nonuniform quen-

ching has been reported to lead to quench

cracking of a carburized 17CrNiMo6 axle in

a reduction gearbox.

 Monitor hardness and dimensional changes

of the parts being processed to look for

unexpected variance.

Retained Austenite (Ref 1, 7). The quen-

chant temperature is a critically important vari-

able in controlling the amount of retained

austenite in the carburized steel. This is impor-

tant because incomplete quenching and the pre-

sence of retained austenite will often seriously

affect wear resistance and pitting fatigue strength

(Ref 9). With carburized steels, the martensite

start (M

) temperature will decrease with

increasing carbon content. To determine the

impact of the carbon content on the M

tem-

perature from the steel composition, the Steven

and Haynes equation may be used (Ref 37, 38):

(



C)= 561 474C  33Mn  17Ni 17Cr

 21Mo

where C, Mn, Ni, Cr, and Mo are the percent of

the element contained in the steel. This equation

is only accurate for steels containing up to 0.5%

Fig. 14

Temperature distribution and martensitic transfor-

mation during quenching of carburized 12.5 mm

diameter steel bar. The curves (isochronal lines) in the ﬁgure

indicate time in seconds after immersion of the carburized

(0.9 mm case) bar into the quenchant indicated.

Sources of Failures in Carburized and Carbonitrided Components / 191

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:18PM Plate # 0 pg 191

C. For steels with higher carbon content, Fig. 15

should be used to determine a more accurate

value for M

(Ref 37).

Although the degree of transformation

between the M

and martensite ﬁnish (M

)

temperatures is not linear, the difference is es-

sentially constant (M

is approximately 215



lower than the M

temperature) (Ref 37).

Nevertheless, if M

M

is approximately con-

stant, incomplete transformation can be expec-

ted if some part of the transformation occurs at

a temperature lower than the quench bath tem-

perature. Therefore, the volume of untrans-

formed austenite (V

) is related to both the M

and the quenchant temperature (T

). This rela-

tionship is quantitatively deﬁned by the well-

known Koistenen and Marburger equation:

71:10 · 10

)

Using these equations, Parrish demonstrated

the effect of quenchant temperature on retained

austenite on a hypothetical steel. The results of

these calculations are summarized in Table 5,

which show that the amount of retained austenite

is expected to decrease with decreasing bath

temperature.

Trusova studied the formation of retained

austenite after quenching and showed that the

retained austenite content was dependent on the

carbon content and alloying elements and

the content of the carburized case (Ref 39). The

higher the steel temperature prior to quenching,

the greater the decomposition of austenite but

the greater the potential for cracking. To prevent

cracking, steel may be quenched in hot oil, but

the amount of retained austenite increases.

Additional dilatometer examination of an iso-

thermal high-temperature tempering process

with subsequent cooling showed there was

either a volume shrinkage due to decomposition

of tetragonal martensite or an expansion caused

by the decomposition of retained austenite

(Fig. 16) (Ref 39). These transformations oc-

curred during heating, isothermal holding, and

subsequent cooling. As a result of this work,

Trusova showed that reducing the carburized

steel temperature to 800



C prior to quenching

would reduce the potential for cracking and a

double tempering at 580 to 600



C for carbur-

ized case structures containing i1.2% C.

Because of the alloy and high carbon content

in many case-hardened steels, the M

tempera-

tures in the carburized case are typically

between 100 and 200



C or lower. These values

will vary with the carbon content in the case.

Therefore, the M

temperature, which is approxi-

mately 215



C below the M

, is also below the

ambient temperature. Under these conditions,

to reach the M

temperature and therefore

minimize retained austenite content, a subzero

treatment is required. (This treatment is also

known as refrigeration or deep cooling.) To

minimize the possibility for the formation of

subsurface microcracking, tempering at 150 to

175



C prior to cold treatment is commonly

performed to stabilize the retained austenite.

Gulyaev reported that the use of cold treatment

to reduce retained austenite was most effective if

conducted immediately after quenching, as

shown in Table 6 (Ref 40).

Equipment for achieving temperatures as

low as 75



C may be relatively simple, such as

dry ice mixed with kerosene or alcohol in a

bucket. Temperatures down to 100



C can be

Fig. 15

Correlation curves for correcting the Steven and

Haynes M

temperature equation. When the carbon

content is less than 0.9%, an 830



C soak of over 2 h is required

to produce a fully austenitic structure.

Table 5 Effect of quenchant temperature on

retained austenite

Martensite

start (M

)

temperature, °C

Quenchant

temperature

), °C

T

°C

Approximate

retained

austenite, %

Estimated

hardness,

HRC

150 80 70 45 52

150 60 90 35 56

150 40 110 29 57

150 20 130 25 58

192 / 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 192

achieved by mechanical refrigeration. For lower

temperatures, down to 195



C, liquid nitrogen

can be used (Ref 41).

After cold treatment, the presence of retained

austenite can be assessed by comparing the

hardness before and after refrigeration. An

increase in hardness is expected if retained

austenite was present and was transformed to

martensite as a result of the cold treatment. Since

cold treatment increases martensite as a result of

the loss of retained austenite, the carburized

steel must be tempered a second time (150 to

200



C for 1 to 2 h) to reduce the potential for

cracking. Table 7 shows the effect of subzero

treatment after quenching to reduce the presence

of retained austenite (Ref 42).

Although subzero treatment of carburized

parts does provide a reduction of retained aus-

tenite, the degree of transformation at a given

temperature is variable, depending on the

amount of retained austenite at the beginning of

the subzero treatment, the elapsed time between

quenching and subzero treatment, intermediate

thermal treatments such as tempering, the level

of compressive stress, any cold working of the

material, and part design (Ref 41). However, it

has been reported that fatigue resistance is

decreased due to localized residual stresses

imparted by the subzero treatment (Ref 36). The

case ductility also seems to be negatively

Fig. 16

Trusova dilatomer curves for tempering of carbur-

ized steels quenched from 950



Table 6 Effect of time delay between

quenching and cold treatment on retained

austenite reduction

Holding time at room

temperature before subzero

treatment at 183 °C

Retained austenite, %

Steel Kh-12 Steel SKh-12

2–3 min. 24 36.5

24 h 46 54.5

45 days ... 55

60 days 48 ...

Table 7 Effect of subzero cooling after quenching

Heat treatment

after carburizing Condition

Retained

austenite, % Hardness, HRC

Bending

strength

Impact

strength

MPa kg/mm

Oil quenched

from 800



low-temperature

tempered

As-quenched 62 54 1530 156 25.5 2.6

Subzero treated 20 62 1442 147 19.5 2.0

Tempered at 650



oil quenched

from 800



low-temperature

tempered

As-quenched 34 60 1697 173 40 4.1

Subzero treated 10 62 1608 164 ... ...

Air-cooled from

900–750



C, oil

quenched

low-temperature

tempered

As-quenched 90 47 1618 165 59 6.0

Subzero treated 20 60 1353 138 19.5 2.0

Note: Steel is 18Kh2N4VA. Subzero treatment conducted at 120



Sources of Failures in Carburized and Carbonitrided Components / 193

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_177-240.pdf/Chap_06/ 18/8/2008 3:18PM Plate # 0 pg 193