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

Подождите немного. Документ загружается.

Steel Failures due to Tempering and

Isothermal Heat Treatment

Jan Vatavuk, Universidade Presbiteriana Mackenzie

L.C.F. Canale, Universidade de Sa˜o Paulo USP

FAILURE of swords made by early metal-

smiths was a complex phenomenon for blade-

smiths. The repeated working, heating, and

cooling could cause embrittlement, with sword

failure occurring in the most critical moments of

a battle. Likewise, some of the earliest cannons

would break apart after the ﬁrst shots following

manufacture. These problems occurred in the

ferrous alloy application until the beneﬁt of

tempering became recognized. In the middle of

the 18th century, the tempering process (and/or

stress relief) received attention as a fundamen-

tally important process in the heat treatment of

the ferrous components of tools. Some ironsmith

tools were treated by the so-called process of

water annealing, whereby steel was tempered in

the range of 300 to 600



C. The slow cooling

was substituted by water cooling.

At the beginning of the 20th century, Krupp

developed a great number of patents based on

water cooling after tempering of chromium-

nickel steel. This phenomenon received atten-

tion after the start of WWI, when large amounts

of steel were used by the armament industry. In

1917, the term tempering embrittlement was

introduced by Dickenson, having been pub-

lished in papers by Brarley, Hatﬁeld, Philpot,

and Grenet. Some investigators, such as Greves

and his collaborators, began a set of experi-

mental methods using notched bars to determine

the susceptibility of tempering embrittlement.

A relationship between the energy absorbed

after water cooling and annealing was termed

the steel susceptibility ratio. At that time, all

the experiments were performed at room tem-

perature, because no one anticipated that tem-

perature may also have an effect on the results.

The effect of test temperature received atten-

tion in the beginning of 1944, when Jolivet

and Vidal introduced experiments at different

temperatures, resulting in a revision of all for-

mer data.

A very important technological mark was the

development of the beneﬁcial molybdenum

effect on the embrittlement reduction phenom-

enon, through work by Greaves and Jones

(Ref 1). For some time, embrittlement due to the

tempering process has been shown to be an

important failure related to heat treatment. In

this chapter, the causes and cases associated with

problems originated by tempering are reviewed.

However, to provide background on this phe-

nomenon, a brief description of the martensite

reactions and the steel heat treatment of tem-

pering is given to review the different stages of

microstructural transformation.

Martensite

Before describing the solid-state reactions

resulting from the tempering process in the fer-

rous matrix, it is important to deﬁne the mar-

tensitic structure as a function of the alloying

elements, especially for the carbon effect. Fer-

rous martensite is composed of a body-centered

tetragonal crystallographic structure, with lat-

tice parameters (c and a) related to the carbon

contents of its chemical composition, as shown

in the expression (Ref 2):

c=a=1+0:0467 · (wt% C)

The lattice ratio for the tetragonal structure is

approximately 1.0467, with 1 wt% C in solid

solution. As shown in Fig. 1, hardness varies

with carbon content, and that effect is strongly

related to the distortions caused by the carbon

atom in the body-centered tetragonal structure.

Martensite is extremely hard (maximum of 800

to 900 HV) and brittle.

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_285-309.pdf/Chap_09/ 18/8/2008 3:32PM Plate # 0 pg 285

Failure Analysis of Heat Treated Steel Components

L.C.F. Canale, R.A. Mesquita, and G.E. Totten, editors, p 285-309

DOI: 10.1361/faht2008p285

www.asminternational.org

The transformation of martensite from aus-

tenite is a nonequilibrium (athermal) process

that occurs during rapid cooling from the aus-

tenite phase. Unlike the phase transformation

from atomic diffusion at equilibrium tempera-

tures, the martensitic transformation occurs

when many atoms together undergo a shear

displacement. This rapid shear displacement of

atoms results in a rapid change in crystal struc-

ture during cooling. Thus, the martensitic

transformation is referred to as a diffusionless

process, because the transformation involves a

nonequilibrium (athermal) crystal change from a

shear transformation during rapid cooling from

austenite.

Martensite is a nonequilibrium structure and

thus does not appear on the iron-carbon equili-

brium phase diagram. The face-centered cubic

lattice shearing of austenite (Bain’s deforma-

tion) in a martensitic transformation is illu-

strated in Fig. 2. The deformation is large and

rapid over many atoms in the lattice, and the

change in the polycrystalline system is accom-

modated by lattice deformation. This can occur

by slide, mechanical twinning, or even a mixture

of both mechanisms according to the steel che-

mical composition. The crystal change results in

an expansion of the polycrystalline system.

The effect of temperature on martensite for-

mation is directly related to the transformation

temperatures of martensite start (M

) and mar-

tensite ﬁnish (M

). Carbon is the alloy element

that has a higher inﬂuence on M

temperature,

which is mainly responsible for the martensite

morphology of steels. There are several

empirical formulas to calculate M

temperature.

Some examples are reported as follows (Ref

2–4):

=539 432 (%C)  30:4 (%Mn)  17:7 (%Ni)

 12:1 (%Cr) 7:5 (%Mo)

For medium-carbon alloy steels (Ref 4):

=520 320 (%C)  50 (%Mn)  30 (%Cr)

 20 ½%(Ni+Mo)5 ½%(Cu+Si)

In Ref 3:

=561 474 (%C)  33 (%Mn)  17 (%Ni)

 17 (%Cr)  21 (%Mo) (Ref 3)

The higher the transformation temperature,

the higher the probability of the plastic defor-

mation mechanism occuring by dislocation

slide, although the low temperatures provoke a

Fig. 1

Hardness of martensite as a function of carbon content.

Source: Ref 3

Fig. 2

(a) Body-centered tetragonal cell in austenite. (b) Body-centered tetragonal cell before (left) and after (right) the lattice

deformation from austenite to martensite. Source: Ref 2, 3

286 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_285-309.pdf/Chap_09/ 18/8/2008 3:32PM Plate # 0 pg 286

plastic deformation by mechanical twinning.

This way, it is possible to establish a martensite

morphology with respect to the alloy content,

giving special attention to carbon.

Figures 3 to 5 show the morphological aspects

of martensite as a function of carbon content for

steels through optic microscopy. In the case of

lath martensite, the deformation mechanism of

the lattice is dislocation slip. This kind of mar-

tensite is also known as slipped martensite.

Morphologically, this martensite presents lath

packages, which are separated by low-angle

boundaries (Ref 5). In the past, it was thought

that the transformation units happened as lath

packages, although recently it became clear that

each lath is independently formed, and the evi-

dence shows that an austenite ﬁlm exists, which

can be seen in Fig. 6.

Measurements of dislocation density found

in martensite are on the order of 0.3 to

0.9 · 10

cm/cm

of the crystal. This disloca-

tion density is higher than the maximum that can

be obtained by elevating the percentage of cold

plastic deformation. Some observations, made

by transmission electron microscope, show a

very small cellular structure (approximately 0.2

to 0.3 mm) inside the lath (Ref 5).

Figure 4 shows the martensite with a high

carbon content, observed with an optical

microscope. With a high carbon content, the

microstructure has twinned martensite or plate

martensite. With higher carbon contents, higher

volume of retained austenite (Fig. 5) occurs,

because a higher carbon content lowers the M

Fig. 3

Lath (low-carbon) martensite in SAE 8620 alloy steel

(Fe, 0.2% C, 0.8% Mn, 0.55% Ni, 0.5% Cr, 0.2% Mo)

after heat treatment (954



C, or 1750



F, for 1 h, water quench)

Fig. 4 High-carbon martensite (0.8% C). Etched with nital

Fig. 5

Microstructure of quenched 1.3% C steel. Dark nee-

dles of plate martensite and white areas of retained

austenite (white arrow)

Fig. 6

Illustration of the austenite ﬁlm surrounding martensite

laths in a Fe-10Cr-0.2C steel. Source: Adapted from

Ref 2

Steel Failures due to Tempering and Isothermal Heat Treatment / 287

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_285-309.pdf/Chap_09/ 18/8/2008 3:33PM Plate # 0 pg 287

temperature. Twinning density can be seen by

transmission electron microscopy, because

density is high even though twins are very nar-

row (on the order of 10 A

The percentage of slip martensite and twinned

martensite in carbon steel and tool steel can be

experimentally determined, as shown in Fig. 7.

As the carbon content increases, the amount of

lath martensite decreases. The untransformed

austenite increases by the M

temperature

(martensite start temperature) reduction. The

increase in the retained austenite volume frac-

tion can reduce the as-quenched hardness

mainly in the higher content range.

Martensitic transformation causes an increase

in volume and size variations, which contributes

to the residual tension stresses that develop in

the surface after the heat treatment of quenching,

when transformation takes place in all of the

sample cross sections, and transformation be-

tween surfaces and nucleus occurs anachroni-

cally. The volume variation measured during the

transformation from austenite to martensite in a

1% C steel is approximately 4% (the transfor-

mation to pearlite results in a 2.4% expansion)

(Ref 2), decreasing as far as the carbon is added

in the matrix. This occurs because of the dif-

ferent carbon effect in the austenite related to the

martensite. In the ﬁrst, the deformation has a

volumetric character, while in the second, it is

more directional (Ref 7). This behavior can be

seen in Fig. 8.

It can be seen in Fig. 8 that the difference

in the speciﬁc volume between austenite and

martensite is approximately 15% from very low

carbon content to very high carbon content

(2% C). It is also interesting to observe that for

low carbon, the change of volume from an as-

annealed condition to an as-hardened condition

is practically nil. On the other hand, increasing

carbon content raised that difference.

These observations are important during the

component process design. As mentioned ear-

lier, the greater the carbon content, the greater

the embrittlement of the martensite plates

(Ref 8). However, since retained austenite also

increases with carbon content, this fraction of

retained austenite will bring some toughness to

the matrix as well as reduce the volume varia-

tion as shown in the curve “A-FC” in Fig. 8.

This fact results in a lower load to the cold sur-

face because of the incomplete transformation

of the nucleus. It is important to remember

that the nucleus presents a low yield strength

when retained austenite has a low mechani-

cal resistance, decreasing the possibility of

Fig. 7

Effect of carbon content on the lath martensite volume, retained austenite volume fraction, and Ms temperature. Source:

Adapted from Ref 6

288 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_285-309.pdf/Chap_09/ 18/8/2008 3:33PM Plate # 0 pg 288

developing tension stress in the surface of the

component.

It can be seen in Fig. 8 that for 2% C, theo-

retically it could be possible to obtain a complete

transformation to martensite, but, this is not

the case. However, it is interesting to observe

that the retained austenite line matches the aus-

tenite line for very high carbon content. This is

the case for Hadﬁeld steels (Ref 9), which have a

high manganese content (approximately 12%)

that guarantees an austenitic microstructure,

even though the carbon amount is approximately

1.2%. In this situation, it is possible to quench

large components with complex geometry

without the risk of developing cracks, even

while increasing quenchant severity.

The ability to form martensite is described

in terms of hardenability, which is related to

the presence of other alloy elements besides

carbon. For example, molybdenum and manga-

nese increase hardenability, while cobalt lowers

the hardenability of steel. A higher hardenability

allows martensite formation with a slower

cooling rate. This is beneﬁcial for reducing

the tensile residual stresses in the component

surface.

Tempering

Tempering is historically associated with the

heat treatment of martensite in steels. The

resultant microstructure is called tempered

martensite. The main purpose for tempering is to

develop a usable combination of hardness and

toughness. The microstructure and mechanical

properties can be changed when the component

is held isothermally at a temperature where

austenite cannot form.

It is important to emphasize that tempered

martensite usually does not contain martensite.

Instead, it is a structure of ﬁne carbide particles

in ferrite, which has formed from martensite

during the tempering. This structure has a lower

hardness than the martensite, but by proper

choice of temperature and time used, the struc-

ture developed will be ﬁne to give the desired

hardness. Table 1 lists the colors associated with

the tempering heats, and Table 2 illustrates the

times required to reach furnace temperature

during tempering (Ref 10).

Effect on Mechanical Properties

As noted, martensitic structures are too brittle

for most practical applications. However, it

is possible to enhance the structure tough-

ness through tempering. The toughness usually

comes at the expense of a decrease in yield

Fig. 8 Speciﬁc volume (DV/V) of carbon steels relative to room temperature. Source: Adapted from Ref 7

Table 1 Colors of tempering heats

Temperature(a) Temperature(b)

°C ° F Color of oxides °C °F

188 370 Faint yellow 238 460

199 390 Light straw 265 510

210 410 Dark straw 293 560

221 430 Brown 321 610

232 450 Purple 337 640

254 490 Dark blue 349 660

265 510 Light blue 376 710

(a) Temperature held for 1 h. (b) Temperature held for 8 min

Steel Failures due to Tempering and Isothermal Heat Treatment / 289

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_285-309.pdf/Chap_09/ 18/8/2008 3:33PM Plate # 0 pg 289

strength and hardness, as illustrated in Fig. 9

and 10. Figure 11 shows other modiﬁcations

of mechanical properties that occur when an oil-

quenched AISI 4340 steel is tempered at tem-

peratures above 200



The interrelationship between tempering

temperature, steel chemistry, and hardness can

be estimated by the equation (Ref 13):

HB=2:84H

+75(%C)  0:78(%Si)

+14:24(%Mn)+14:77(%Cr)+128:22(%Mo)

 54:0(%V) 0:55T

+435:66

where HB is the Brinell hardness after hardening

and tempering, H

is the Rockwel (HRC) hard-

ness after hardening, and T

is the tempering

temperature in



C. This equation was developed

for the following conditions:

 H

= 20 to 65 HRC and T

= 500 to 600



 C = 0.20 to 0.54%, Si = 0.17 to 1.40%,

Mn = 0.50 to 1.90%, and Cr = 0.03 to

1.20%

An average relation between the hardness

after hardening (H

) and the hardness after

hardening and tempering (H

) can be found

through:

=(T

=167 1:2)H

 17 ½HRC

where H

is the hardness after hardening and

tempering (HRC), and T

is the tempering tem-

perature (



C). This equation is valid for

490



C5T

5610



C and for a tempering time

of 1 h.

The tempering temperature for a speciﬁed

hardness after hardening and tempering is also

possible to calculate when chemical composi-

tion and the degree of hardening are known

(Ref 13):

=647 [S(60

ﬃﬃﬃﬃ

+20)=H

70:9]

1=4

73:45 SH

+(5377561S)(%C)+505S(%V)+219S(%Mo)

+75S(%Cr)+66S(%Si)751 [C



]

where H

is the hardness after hardening and

tempering (HRC), S is the degree of hardening,

Sj1.0, and the alloying elements are given in

weight percent. This formula is valid for a

tempering time of 2 h.

Tempering Reactions

Tempering is a process in which the micro-

structure approaches equilibrium under the

inﬂuence of thermal activation. It follows that

the tendency to temper depends on how far

the starting microstructure deviates from

Table 2 Approximate heating times for tempering

Per inch of diameter or thickness, with furnace maintained steadily at T

max

, and steel having dark or scaled surface

Temperature Heating time, min

°C °F

Cubes or

spheres(a)

Squares or

cylinders(a)

Average

ﬂats(a)

Cubes or

spheres(b)

Squares or

cylinders(b)

Average

ﬂats(b)

121 250 30 55 80 15 20 30

149 300 30 50 75 15 20 30

177 350 30 50 70 15 20 30

204 400 25 45 65 15 20 30

260 500 25 40 60 15 20 30

316 600 25 40 55 15 20 30

371 700 20 35 50 15 20 30

427 800 20 30 45 15 20 30

482 900 20 30 40 15 20 30

(a) In hot air oven, without circulation. (b) In circulation air furnace or oil bath (can be used only in lower temperatures)

Fig. 9

Effect of tempering on the true stress in a carbon steel.

Source: Adapted from Ref 11

290 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_285-309.pdf/Chap_09/ 18/8/2008 3:33PM Plate # 0 pg 290

equilibrium. Martensite microstructure is the

farthest, followed by bainite, ferrite, and

cementite (Ref 14).

When the martensitic structure is metastable,

there is a natural tendency to transform it to a

structure with more stability, and those mod-

iﬁcations are accelerated by increasing the

temperature during the tempering.

The modiﬁcation that occurs during temper-

ing is complex, and the transformations that take

place during the treatment conditions necessary

to produce the best mechanical properties com-

bination are a result of accumulated knowledge,

not just from the academic point of view but also

the practical aspect of observation. Most of the

time, the structures developed during isothermal

heat treatments are inﬂuenced by the low content

of other elements besides iron and carbon.

Tempering stages

Solid-state reactions follow a sequence of

precipitation that is related to variables such as:

 Diffusivity of the involved element

 Surface energy of interfaces produced by the

reactions

 Crystallographic adjustment (coherence

stresses) between the precipitated phases

and the ferrous matrix

 Thermodynamic stability of reactions

During tempering, the martensitic structure is

submitted to a sequence of reactions, often

superimposed and deﬁned as temper stages (Ref

2, 3, 4, 5, 12).

Stage 1. In high-carbon steels, the pre-

cipitation of excess carbon begins with the for-

mation of a transition carbide, such as e (Fe

2.4

C).

The e-carbide can grow at temperatures as low

as 50



C. Martensite is said to be supersaturated

with carbon when the concentration exceeds its

equilibrium solubility with respect to another

phase. However, the equilibrium solubility

depends on the phase. The solubility will be

larger when the martensite is in equilibrium with

a metastable phase such as e-carbide. Approxi-

mately 0.25 wt% C is said to remain in solution

after the precipitation of e-carbide is completed.

Although most textbooks will begin a discussion

of tempering with this ﬁrst stage of tempering,

involving the redistribution of carbon and pre-

cipitation of transition carbides, cementite can

Fig. 10 Effect of carbon content on the hardness of tempered plain steels. Source: Ref 10

Steel Failures due to Tempering and Isothermal Heat Treatment / 291

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_285-309.pdf/Chap_09/ 18/8/2008 3:33PM Plate # 0 pg 291

precipitate directly (Ref 14). This is the case for

the lath martensite structure, where the disloca-

tion density can be as high as 10

to 10

/cm

(Ref 15). Trapped carbon atoms will not pre-

cipitate as transition carbides, but cementite is

more stable than trapped carbon. This stage

begins at room temperature and extends to

250



C. A ﬁne adjustment between the e-

carbides and the ferrous structure is attributed to

the precipitation-hardening effect of martensite

in high-carbon steels tempered between 50 and

100



Stage 2. Tempering at higher temperatures,

in the range of 200 to 300



C, for 1 h induces the

retained austenite to decompose into a mixture

of cementite and ferrite. When the austenite is

present as a ﬁlm, the cementite also precipitates

as a continuous array of particles that have the

appearance of a ﬁlm (Ref 3, 5, 12, 14). The

martensite of the steels with less than 0.5% C

content has a retained austenite amount lower

than 2%, reaching 6% for 0.8% C. There are

some indications that austenite decomposes,

turning into ferrite and cementite, but presently a

consensus does not exist about whether this

structure can be correlated to lower bainite,

typically from the isothermal decomposition of

austenite, in the temperature range of 230 to

300



Stage 3. Tempering at even higher tem-

peratures leads to a coarsening of the cementite

particles, with those located at the plate bound-

aries growing at the expense of the intraplate

particles. This precipitation is responsible for

the embrittlement phenomenon observed at the

temperature of 250 to 400



C. It can be avoided

by adding silicon, which is an insoluble element

in cementite. This allows cementite formation at

Fig. 11 Changes in the mechanical properties of AISI 4340 steel with tempering temperature. Source: Ref 12

292 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_285-309.pdf/Chap_09/ 18/8/2008 3:33PM Plate # 0 pg 292

temperatures where silicon diffusion occurs,

thus slowing the process.

Silicon is a substitutional alloy element that

accumulates in the ferrous matrix adjacent to the

C carbide in the growth process. It increases

the carbon activity in this region, decreasing the

carbon ﬂow to the growing particle and conse-

quently reducing its growth velocity (Ref 5, 16,

17). This silicon effect resulted in the develop-

ment of alloy 300M, which substitutes for 4340

steel in those applications where it is necessary

to use tempering temperatures that cause

embrittlement of tempered martensite, which is

soon deﬁned. This reaction begins to occur at

temperatures on the order of 100



C. Cementite

can also be observed during quenching when the

temperature is elevated, as is the case of steel

with a low carbon content, mainly in the mar-

tensite formed just below the M

temperature.

This phenomenon is known as self-tempering

(Ref 5).

Stage 4. In carbon steels, stage 3 marks the

end of the tempering process. Spheroidization of

C occurs as cementite coalesces. This phe-

nomenon is sometimes called stage 4 of tem-

pering (Ref 2). The lath boundary maintains

stability up to approximately 600



C. Intense

rearrangement occurs between the lath and its

low-angle boundaries above 600



C. This

recovery process is replaced by recrystallization

and coarsening (Fig. 12) at temperatures be-

tween 600 and 700



C (Fig. 13).

Effect of Te mperature and Alloying. The

effect of the tempering temperature on steels

with increasing carbon contents can be inferred

from Fig. 13. During tempering, the continuous

decomposition of martensite to ferrite and car-

bides changes the state of stress because of

continuous dimensional changes. At low tem-

peratures (ﬁrst stage), a volume contraction

takes place as a consequence of e-carbide pre-

cipitation. In the second stage, with the trans-

formation of retained austenite (approximately

300



C), the volume is increased. In stage 3, the

progressive decomposition of martensite leads

to a volume decrease.

It is important to observe that the austeniti-

zation temperature, which determines the

amount of carbon dissolved and the amount of

retained austenite, has a strong inﬂuence on the

expected volume changes (Ref 2, 4, 12, 13).

Table 3 shows the changes in length for various

steels as a function of tempering temperature.

Alloyed steels can also have another stage

with the precipitation of alloy carbides, includ-

ing M

C (molybdenum), M

C, M

(chromium rich), V

, TiC, and so on, where

the “M” refers to a combination of metal atoms.

However, all of these carbides require long-

range diffusion of substitutional atoms. They

can only precipitate when the combination of

time and temperature is sufﬁcient to allow this

diffusion. The alloy carbides grow at the

expense of the less stable cementite. If the con-

centration of strong carbide-forming elements,

such as molybdenum, chromium, titanium,

vanadium, and niobium, is large, then all of the

carbon can be accommodated in the alloy car-

bide, thereby completely eliminating the

cementite.

Figure 14 illustrates the effect of alloying

elements on hardness as a function of tempering

temperature in carbon steels (Ref 5). Increases in

hardness with additions of titanium, vanadium,

molybdenum, and chromium are related to the

alloy carbide precipitation. This phenomenon is

common for tool steels and can affect their

toughness, as illustrated in Fig. 15.

Embrittlement

Hardness decreases with increasing temper-

ing temperature (Fig. 10, 11). Consequently,

yield strength and tensile strength decrease as

well. On the other hand, elongation and ductility

increase. In this general context, a failure related

Fig. 12

Fe-0.17C alloy quenched in water from 900



C and

tempered at 650



C for 5 h. Microstructure shows

ferrite grains and spheroidized Fe

Steel Failures due to Tempering and Isothermal Heat Treatment / 293

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_285-309.pdf/Chap_09/ 18/8/2008 3:34PM Plate # 0 pg 293

to tempering may be attributed to an incorrect

choice of temperature (and/or time of temper-

ing), resulting in an incorrect hardness or low

toughness.

However, most failures are related to

embrittlement phenomena. Quenched and tem-

pered steels are susceptible to a number of dif-

ferent types of embrittlement. Some of them

are due to structural modiﬁcations during tem-

pering, as previously described. However,

there are some due to the interaction of the

environment with the quenched and tempered

microstructures, such as hydrogen embrittle-

ment and liquid metal embrittlement. Examples

of the ﬁrst type of embrittlement are tempered

martensite embrittlement and temper embrittle-

ment, which are described as follows.

Tempered Martensite Embrittlement. It is

well known that tempered martensite embrit-

tlement (TME) is related to tempered martensite

of specimens tempered between 250 and

370



C, as shown in Fig. 16. The impact

toughness after tempering at this temperature

range is lower than that obtained on tempering

at temperatures below 250



C. This type of

brittleness is inherent to some extent in all steels,

including carbon grades. For that reason, med-

ium-temperature tempering is, as a rule, not

employed in practice, although it can ensure a

high yield limit. According to Krauss (Ref 12),

TME may or may not be associated with

impurity atom segregation to prior-austenitic

grain boundaries, but the most common factor,

at least for medium-carbon steels, is the phe-

nomenon that takes place due to decomposition

of retained austenite to cementite in the interlath

Fig. 13 Hardness as a function of tempering temperature for plain carbon steels. Source: Ref 6

Table 3 Length variations related to

metallurgical reactions as a function of

tempering temperature ranges

Stage

Temperature

range, °C Metallurgical reactions

Expansion (E) or

contraction (C)

1 0–200 Precipitation of e-carbide;

loss of tetragonality

2 200–300 Decomposition of retained

austenite

3 230–350 e-carbides decompose to

cementite

4 350–700 Precipitation of alloy

carbides; grain

coarsening

Source: Ref 18

294 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_285-309.pdf/Chap_09/ 18/8/2008 3:34PM Plate # 0 pg 294