Tempering Martensite

Metallogr. Microstruct. Anal.
(2014) 3:70–90
DOI 10.1007/s13632-013-0117-1
CLASSIC PAPER
Tempering of Martensite
Leonard E. Samuels
ASM International 2014
Our main concern in this chapter is the effects that can be • Carbide Precipitation. A discrete carbide phase in the
observed by light microscopy when a steel in the hardened form of extremely small (*2 nm diameter) uniformly
condition, the structure of which then consists of martensite distributed spherical particles is precipitated when steels
and possibly some retained austenite, is heated at subcritical containing more than about 0.2% C are tempered at
temperatures. This is the heat treatment known as tempering. temperatures between 100 and 250 C, leaving a mar-
The term is also sometimes, although inappropriately, tensite with a saturated carbon content of about 0.2%.
applied to a treatment in which other primary structures, such The precipitate was first identified as epsilon carbide
as pearlite and bainite, are heated at subcritical temperatures, (e carbide, or Fe2.4C) with a close-packed hexagonal
but the structural changes in these instances comprise only crystal structure [3]. Later it was established that it had an
redistribution of cementite and other carbides. Chapter 7 orthorhombic structure isomorphous with transition
discusses such changes in pearlitic structures, referred to as metal carbides of the M2C type [4, 5]. This phase was
‘‘spheroidization.’’ The corresponding changes in bainitic designated eta carbide (g carbide). The difference
structures are somewhat similar, but we do not discuss them between the two structures is slight, concerning only a
because they are not of great practical importance. sublattice of the carbon atoms. At temperatures above
It has taken investigation by many techniques, of which 250 C, however, a phase known as the chi carbide
light and electron microscopy are only two, to establish the (v carbide) forms, but subsequently transforms to the
sequence of phenomena that may occur during tempering theta carbide (h carbide) that is known as cementite
[1, 2]. These phenomena are as follows: (Fe3C, orthorhombic). Cementite precipitates first as
lath-like platelets within martensite plates (at transfor-
• Carbon Segregation. Carbon is redistributed to low-
mation twins and at the midrib) and at plate interfaces.
energy sites such as vacancies, dislocations and cell
The platelets later agglomerate to form spheroidal
boundaries. This redistribution may occur extensively
particles that coarsen rapidly at high temperatures.
while the material is cooling below the Ms temperature
• Decomposition of Retained Austenite. Retained austen-
during quenching to form the martensite, this being
ite decomposes to a mixture of ferrite and cementite at
more probable the higher the Ms temperature of the
tempering temperatures between 200 and 300 C. The
steel. About 90% of the carbon in martensites with
transformation occurs only after the transition carbides
carbon contents below 0.2% is likely to be redistributed
are well established.
in this way. Above 0.2% C, the sites become saturated
• Recovery and Recrystallization. Recovery of the defect
and the excess carbon is held in the defect-free regions
structure of martensite becomes significant at temper-
of the martensite crystals.
ing temperatures above 400 C. Recrystallization of the
ferrite matrix can occur at 600–700 C.
All of these phenomena are time dependent and occur
Reprinted from Light Microscopy of Carbon Steels, chapter 10, both isothermally and during continuous heating. They
Leonard E. Samuels, ASM International, Materials Park, OH (1999).
overlap, and some can occur simultaneously.
123
Metallogr. Microstruct. Anal. (2014) 3:70–90 71
Fig. 10.1 Basic effects of

tempering. (a–d) 1%C–1%Cr
(1.10C–0.25Si–0.30Mn–1.0Cr,
wt%). (a) Austenitized at
1,000 C, water quenched. 760
HV. 40 vol% austenite. Etched
with (b) and (c) under strictly
comparable conditions. Picral.
2,0009. (b) Austenitized at
1,000 C, water quenched,
tempered 30 min at 250 C. 670
with (a) and (c) under strictly
2,0009. (c) Austenitized at
1,000 C, water quenched,
with (a) and (b) under strictly
2,0009. (d) Austenitized at
1000 C, water quenched,
HV. 0 vol% austenite. Picral.
2,0009. (e–h) 1.9%C (1.86C–
0.10Si–0.16Mn, wt%).
(e) Austenitized at 1,150 C for
1 h, water quenched, tempered
30 min at 210 C. 2% nital.
2,0009. (f)Austenitized at
1,150 C for 1 h, water
quenched, tempered 30 min at
240 C. 2% nital. 2,0009.
(g) Austenitized at 1,150 C for
1 h, water quenched, tempered
30 min at 250 C. 2% nital.
2,0009. (h) Austenitized at
1,150 C for 1 h, water
quenched, tempered 30 min at
240 C. Electron micrograph of
gold–palladium shadowed
replica. 2% nital. 25,0009
123
72 Metallogr. Microstruct. Anal. (2014) 3:70–90
Fig. 10.2 Basic effects of

tempering. 0.4%C (0.43C–
0.22Si–0.86Mn, wt%). These
micrographs are the same
specimens as those shown in
Fig. 10.3–10.5. (a) Austenitized
at 850 C, water quenched,
tempered at 200 C for 30 min.
630 HV. Arrows indicate typical
carbide particles. Transmission
electron micrograph of a foil.
540 HV. Arrows indicate
typical carbide particles.
Transmission electron
micrograph of a foil. 20,0009.
(c) Austenitized at 850 C,
water quenched, tempered at
400 C for 30 min. 470 HV.
Arrows indicate typical carbide
particles. Electron micrograph
of a platinum-shadowed carbon
replica. 1% nital ? picral.
carbide particles. Electron
micrograph of a platinum-
shadowed carbon replica. 1%
nital ? picral. 10,0009.
(e) Austenitized at 850 C,
600 C for 30 min. 305 HV.
particles. Electron micrograph
of a platinum-shadowed carbon
replica. 1% nital ? picral.
10,0009. (f) Austenitized at
carbide particles. Electron
micrograph of a platinum-
shadowed carbon replica. 1%
nital ? picral. 10,0009.
(g) Austenitized at 850 C,
600 C for 5 min. 1% nital.
tempered at 600 C for 16 h.
particles. 1% nital ? picral.
2,0009
123
Fig. 10.3 Tempering of a

medium-carbon (0.4%)
hypoeutectoid steel. 0.43C–
0.18Si–0.75Mn (wt%). These
micrographs constitute part of
the series shown in Fig. 10.3–
10.6. The specimens were
etched together under identical
conditions. The hardness of the
steel quenched was 750 HV.
(a) Austenitized at 850 C,
water quenched, tempered
30 min at 200 C. 630 HV.
Picral. 2509. (b) Austenitized
HV. Picral. 1,0009.
30 min at 300 C. 540 HV.
Picral. 2509. (d) Austenitized
HV. Picral. 1,0009.
30 min at 400 C. 470 HV.
Picral. 2509. (f) Austenitized at
HV. Picral. 1,0009.
30 min at 500 C. 360 HV.
Picral. 2509. (h) Austenitized
HV. Picral. 1,0009.
(i) Variation with tempering
temperature of the hardness of a
water-quenched 0.4% C steel
tempered for 30 min. The
relevant microstructures are
shown in Fig. 10.3 and 10.4
Precipitation of e carbides is sometimes known as the Carbide Precipitation

first stage, decomposition of retained austenite as the sec-
ond stage, and precipitation of cementite as the third stage Precipitation of e carbides (or g carbides) can be detected
of tempering. only indirectly: martensite in which the precipitation has
occurred etches more darkly than untempered martensite
(compare Fig. 10.1a and b).
Effects Observable by Light Microscopy This is most clearly seen in high-carbon steels, such
as the steel illustrated in Fig. 10.1(a) and (b)–particularly
Carbon Segregation if a large volume fraction of retained austenite is present,
in which case the individual martensite plates can be
No effects due to carbon segregation can be observed by seen in contrast with the background of austenite. It can,
light microscopy. however, also be observed in steels of lower carbon
123
Fig. 10.3 continued
content if the etching conditions are made adequately formed at intermediate tempering temperatures can be
comparative. The presence of the particles of e carbide detected by electron microscopical examination of replicas
responsible for this increased etching reactivity can be of specimens prepared as for examination by light
confirmed by transmission electron microscopy of thin microscopy (Fig. 10.2c and d). The cementite particles
foils; an example is given in Fig. 10.2(a), an illustration formed during tempering at the lower temperatures of the
of this specimen being shown by light microscopy in range are plate shaped (Fig. 10.2c), but those formed at
Fig. 10.3(a and b). higher tempering temperatures become increasingly sphe-
Formation of cementite at low temperatures also can be roidal in shape with increasing temperature (Fig. 10.2d–f).
detected only indirectly, by the progressively increasing The particles can also be seen to decrease in number and
reactivity of the martensite to etchants (Fig. 10.3 and 10.4), increase in size with increasing tempering temperature or
and often by development of a mottled appearance after tempering time, and it is only when this growth process is
tempering at higher temperatures in the range concerned reasonably well advanced that the cementite particles
(Fig. 10.1a and 10.3d). Discrete particles of cementite become large enough to be resolved discretely by light
cannot be detected by light microscopy until tempering microscopy (compare Fig. 10.2d and 10.3h, 10.2e and
temperatures approaching 600 C are reached (Fig. 10.4a 10.4b, and 10.2f and 10.4d).
and b). The cementite particles mostly are located at the
The small particles of cementite precipitated at low boundaries of ferrite grains formed by recovery or recrys-
tempering temperatures can again be detected only when tallization of the martensite (see below) or, less frequently,
thin foils are examined by transmission electron micros- at austenitic grain boundaries. Consequently they are not
copy (Fig. 10.2b; compare with the light micrographs in truly spheroidal, their shape being modified by the inter-
Fig. 10.3c and d), and the cementite can be distinguished facial energies of the various phases involved (Fig. 10.2f
from e carbide in this instance by electron diffraction and h), as was discussed in connection with Fig. 7.1.
experiments. The somewhat larger particles of cementite Moreover, most particles are found to be located at the
123

medium-carbon (0 .4%)
10.6. The specimens in (a–
d) were etched together with
those illustrated in Fig. 10.3
under identical conditions.
30 min at 600 C. 305 HV.
Picral. 2509. (b) Austenitized
HV. Picral. 1,0009.
30 min at 700 C. 235 HV.
Picral. 2509. (d) Austenitized
HV. Picral. 1,0009.
600 C for 5 min. 310 HV.
Picral. 1,0009. (f) Austenitized
tempered at 600 C for 1 h. 270
HV. Picral. 1,0009.
600 C for 4 h. 260 HV. Picral.
210 HV. Picral. 1,0009
123
grain boundaries existing at any stage of tempering (com- clearly because of the confusion caused by the presence of
pare, for example, Fig. 10.2e and f), from which it follows many partly resolved particles of cementite; Fig. 10.2(g) is
that the particles relocate by diffusion as the positions of an indication of the best resolution of the structure that can be
the boundaries change. A consequence that can be achieved by light microscopy at this stage.
observed by light microscopy is that the distribution of the The ferrite grains grow with further tempering. They at
cementite particles changes from somewhat linear arrays to first have a distinctly platelike morphology, but the aspect
random arrays as tempering progresses in its advanced ratio of the plates decreases progressively. This can be seen
stages (Fig. 10.4f–h). most clearly by electron microscopy of replicas
(Fig. 10.2c–f), but can be recognized by light microscopy
Decomposition of Retained Austenite (Fig. 10.6). Discrete grains of ferrite can be recognized
during the later stages of the sequence after etching by
When steels containing a significant volume fraction of appropriate techniques (Fig. 10.6e–h); the sodium bisulfite
retained austenite are tempered at temperatures around etching reagent is most effective in this respect (Fig. 10.6h).
350 C, the volume fraction of austenite can be seen by light The ferrite grains continue to grow and gradually assume
microscopy to decrease; eventually, austenite can no longer an equiaxed shape, but this requires long times (tens of hours)
be detected (Fig. 10.1b–d). The austenite is replaced by at high temperatures (approaching the A1 temperature). By
dark-etching material that cannot be distinguished from the this stage, the grains will have grown to a size that can be
interleaving regions of tempered martensite (Fig. 10.1d). readily resolved by light microscopy, if only at the highest
The extent to which the decomposition of retained effective magnifications (Fig. 10.2h). The structural
austenite can be observed by light microscopy is illustrated arrangement then is not too different from that obtained by
by an experiment, illustrated in Fig. 10.1(e–h), carried out annealing at subcritical temperatures after cold rolling
with a steel of very high carbon content (1.86%) which, in (Fig. 4.7), or by heating a pearlitic parent structure for long
the quench-hardened condition, had a large volume fraction periods at subcritical temperatures (Fig. 7.1). However, the
of retained austenite (Fig. 10.1e). Transformation of the ferritic grain size tends to be smaller for the case under con-
austenite commenced at the surfaces of the plates of mar- sideration than for the other two routes by which the structure
tensite (Fig. 10.1f) and progressed by migration of a rela- may be obtained. The volume fraction of cementite present
tively smooth interface into the austenite (Fig. 10.1f and g). will of course vary with the carbon content of the steel.
The transformed austenite could be distinguished clearly The above recovery and recrystallization stages occur
from untransformed austenite because it etched compara- less rapidly the higher the carbon content of the steel,
tively darkly. It could also be distinguished from the now- because carbide particles inhibit migration of the grain
tempered martensite when the two were in close juxtapo- boundaries and because more carbide particles are present
sition, the transformed austenite etching the more darkly of in steels of higher carbon content.
the two, but it would perhaps be difficult to distinguish
between the two when present independently. The product Changes in Hardness
of transformation of the austenite could not be resolved by
light microscopy, but electron microscopy of replicas Precipitation of carbides may at first result in a slight increase
showed that the structure was duplex and that the mor- in hardness, but always eventually results in a significant
phology of the minor phase was different from that in the decrease. Transformation of austenite itself contributes an
lightly tempered martensite (see Fig. 10.1h; the central increase in hardness, but at best this only delays the decrease
diagonal band in this photograph is the tempered mar- due to carbide precipitation even when large volume frac-
tensite, and the two side bands are the decomposition tions of retained austenite are present. Normally, therefore,
product from the austenite). hardness decreases steadily with increasing tempering tem-
perature in the manner illustrated in Fig. 10.3(i). The effects
Recovery and Recrystallization of the various phenomena discussed consequently cannot be
distinguished by hardness tests.
When tempering is carried out at temperatures of 400 C or The hardness of untempered martensite depends almost
higher, removal of carbon from the martensite, together with solely on carbon content, but the influence of carbon on the
annihilation of cell boundaries and random dislocations, hardness of tempered martensite decreases markedly as the
results in formation of grains of ferrite. These grains at first tempering temperature is increased. The influences of other
have either a lathlike or a platelike morphology, depending alloying elements then become increasingly important, but
on the morphology of the parent martensite. This morphol- the interrelating effect of the various elements, the tem-
ogy can be vaguely distinguished by light microscopy, pering temperature, and the tempering time is a very
although discrete grains of ferrite cannot be recognized complex one [6, 7].
123
Hypoeutectoid Steels cementite particles is noticeably larger in the eutectoid

than in the medium-carbon steel. The volume fraction
A Medium-Carbon Steel (0.4% C) of cementite could be related with carbon content,
however, only by making laborious quantitative
The characteristic structures developed in a medium-car- estimates.
bon hypoeutectoid steel are illustrated in Fig. 10.3–10.5.
The increasing reactivity to the etchant caused by the Effects of Plastic Deformation
earlier stages of carbide precipitation is illustrated in
Fig. 10.3, and the development of resolvable particles of Plastic deformation of martensite after quenching but
cementite in Fig. 10.4 and 10.5. The variation of hardness before tempering suppresses precipitation of e (or g) car-
with tempering temperature at constant tempering time is bides, and consequently the increased reactivity to etching
shown graphically in Fig. 10.3(i). after tempering at low temperatures, illustrated in
The later stages of the growth of cementite particles are Fig. 10.3, is suppressed [8]. The effect is observed after
illustrated in considerable detail in these micrographs tempering at temperatures in the range 150–250 C (it
because this is the phenomenon that can be followed most virtually disappears after tempering at 300 C or higher) in
effectively by light microscopy. The growth of the steels containing more than about 0.4% C, and for either
cementite particles with time at a constant high tempera- tensile or compressive strains exceeding about 0.5%
ture is illustrated in Fig. 10.4(e–h), and an indication of the (engineering strains) [9–12]. Note that the temperature
effects of time at a range of high temperatures is given in attained during mounting of the specimen in a thermoset-
the micrographs in Fig. 10.5. Figure 10.2(g and h) is also ting plastic usually is sufficiently high to induce this effect.
relevant in this context. The strong time dependence of the The suppression of tempering can be detected, espe-
phenomena is apparent, as well as the interchangeability of cially in locally deformed regions, after etching by any
temperature and time in producing a given final micro- standard method (see Fig. 10.10a–c, where the light-etch-
structure. Comparison of Fig. 10.5(d, f, and h) gives an ing areas are the plastically deformed regions), although
indication of the variation with temperature of the time the deformation patterns developed by the various etching
required to produce approximately the same degree of methods differ in detail. Picral is the most satisfactory
cementite particle growth and spheroidization. standard reagent for this purpose (Fig. 10.10b). However,
A selection of these specimens etched with nital and an extensive survey [12] has shown that a picral–ferric
with the bisulfite reagent is illustrated in Fig. 10.6. The chloride reagent is the most satisfactory for revealing the
morphology of the newly formed grains of ferrite is more maximum extent of the plastically deformed zone
apparent after etching with these reagents than after etching (Fig. 10.10d), this etchant being particularly useful as a
with picral, particularly so after etching with the bisulfite macroetch. Note, incidentally, that the general shape of the
reagent. Note, however, that the cementite particles are not locally deformed region surrounding the hardness inden-
as clearly revealed. tations is similar to that revealed in Fig. 6.8(f) for a low-
carbon steel in the annealed condition.
Other Hypoeutectoid Steels The decrease in hardness that normally accompanies an
increase in tempering temperature also is retarded by
A selection of hypoeutectoid steels tempered at an early plastic deformation between quenching and tempering; the
stage of carbide precipitation is shown in Fig. 10.7 and effect is detectable for tempering temperatures between
10.8. These micrographs should be compared with those in 150 and 300 C and being at a maximum at about 250 C
Fig. 10.3(c) and (d) for the 0.4% C steel, and with the [9].
micrographs in Fig. 9.17 for the same steels in the as- The most likely explanation for the suppression of
quenched condition. Darkening during etching under tempering by plastic deformation is that vacancies intro-
comparable conditions is in all cases greater than that for duced into the structure of the martensite by plastic
as-quenched material, but the variation in contrast with deformation attract surrounding carbon atoms, tie them up
variation in carbon content is still small. The basic mor- with high binding energy, and effectively take them out of
phology (i.e., lath or plate) of the parent martensite can still solution [9]. The system is then no longer unstable with
be discerned, particularly after etching in nital. respect to e carbides when a sufficient number of vacancies
The kinetics of cementite growth and spheroidization has been introduced, and consequently precipitation of e
in a eutectoid steel are illustrated in Fig. 10.9, which carbides does not occur during tempering at low tempera-
can be compared with Fig. 10.5 for a 0.4% C steel. The tures [8, 9]. The system is, however, still unstable with
processes are the same, and the kinetics approximately respect to cementite, and at higher temperatures the tem-
the same, in the two cases. The volume fraction of pering processes proceed as usual.
123

10.6. (a) Austenitized at 850 C,
270 HV. Picral. 1,0009.
190 HV. Picral. 1,0009.
HV. Picral. 1,0009.
HV. Picral. 1,0009
123

10.6. (a) Austenitized at 850 C,
200 C for 30 min. 630 HV.
Nital. 1,0009. (b) Austenitized
630 HV. Bisulfite. 1,0009.
400 C for 30 min. 470 HV.
Nital. 1,0009. (d) Austenitized
500 C for 30 min. 360 HV.
Nital. 1,0009. (f) Austenitized
700 C for 30 min. 235 HV.
Nital. 1,0009. (h) Austenitized
235 HV. Bisulfite. 1,0009
123
Fig. 10.7 Tempering of a range

of hypoeutectoid steels. All
picral etched. The specimens
were etched under strictly
comparable conditions. This is,
however, a somewhat lighter
etch than that used for the
specimens shown in Fig. 10.3.
The same specimens etched in
nital are shown in Fig. 10.8.
(a and b) 0.25% C (0.23C–
0.06Si–0.52Mn, wt%).
Austenitized at 925 C, water
quenched, tempered 300 C for
30 min. 430 HV. (a) 2509.
(b) 1,0009. (c and d) 0.4% C
(0.39C–0.22Si–0.75Mn, wt%).
30 min. 515 HV. (c) 2509.
(d) 1,0009. (e and f) 0 .6% C
(0.61C–0.08Si–0.60Mn, wt%).
30 min. 580 HV. (e) 2509.
(f) 1,0009. (g and h) 0.8% C
(0.78C–0.1Si–0.6Mn, wt%).
30 min. 655 HV. (g) 2509.
(h) 1,0009
123
Fig. 10.8 Tempering of a range

of hypoeutectoid steels. All 1%
nital etched. The specimens
were etched under strictly
comparable conditions. The
same specimens etched in picral
are shown in Fig. 10.7. (a and
b) 0 .25% C (0.23C–0.06Si–
0.52Mn, wt%). Austenitized at
tempered 300 C for 30 min.
430 HV. (a) 2509. (b) 1,0009.
(c and d) 0.4% C (0.39C–
0.22Si-0.75Mn, wt%).
30 min. 515 HV. (c) 2509.
(d) 1,0009. (e and f) 0 .6% C
(0.61C–0.08Si–0.60Mn, wt%).
30 min. 580 HV. (e) 2509.
(f) 1,0009. (g and h) 0.8% C
(0.78C–0.1Si–0.6Mn, wt%).
30 min. 655 HV. (g) 2509.
(h) 1,0009
123
Fig. 10.9 Tempering of a 0.8%

C eutectoid steel. 0.81C–
0.22Si–0.32Mn (wt%).
305 HV. Picral. 1,0009.
210 HV. Picral. 1,0009.
HV. Picral. 1,0009.
HV. Picral. 1,0009
123
Fig. 10.10 Effects of plastic

deformation on tempering. 0.4%
C (0.43C–0.18Si–0.75Mn,
wt%). This is the same steel as
that shown in Fig. 10.3–10.5.
(a–d) Austenitized at 850 C,
water quenched, indented,
tempered at 200 C for 30 min,
sectioned. (a) Nital. 2509.
(b) Picral. 2509. (c) Bisulfite.
2509. (d) Picral–ferric chloride.
2509. (e–h) Austenitized at
tempered at 200 C for 30 min,
indented, tempered at 200 C
for 30 min, sectioned. (e) Nital.
2509. (f) Picral. 2509.
(g) Bisulfite. 2509. (h) Picral–
ferric chloride. 2509
123
Fig. 10.11 Revealing the

austenitic grain size in a 0.3% C
quenched-and-tempered steel.
0.25C–0.06Si–0.52Mn (wt%).
water quenched. Picral–HCl.
2509. (b) Austenitized at
860 C, water quenched.
Nital ? aqueous picral. 2509.
30 min at 150 C. Picral–HCI.
2509. (d) Austenitized at
tempered 30 min at 150 C.
2509. (f) Austenitized at
2509. (h) Austenitized at
Nital ? aqueous picral. 2509
123
Fig. 10.12 Revealing the austenitic grain size in a 0.5% C quenched- (d) Austenitized at 960 C, water quenched, tempered at 150 C for
and-tempered steel. 0.50C–0.06Si–0.7Mn (wt%). (a) Austenitized at 30 min. Nital ? aqueous picral. 2509. (e) Austenitized at 960 C,
960 C, water quenched, tempered at 150 C for 30 min. Picral–HCI. water quenched, tempered at 325 C for 30 min. Nital ? aqueous
2509. (b) Austenitized at 960 C, water quenched, tempered at picral. 2509. (f) Austenitized at 960 C, water quenched, tempered at
325 C for 30 min. Picral–HCI. 2509. (c) Austenitized at 960 C, 550 C for 30 min. Nital ? aqueous picral. 2509
water quenched, tempered at 550 C for 30 min. Picral–HCI. 2509.
Differential etching of plastically deformed zones is also energy sites for carbon rather than for e carbides. The e (or
observed when the martensite is first tempered at about g) carbides precipitated during the first temper conse-
200–300 C, deformed, and then retempered at 150–200 C quently tend to redissolve during the second temper, and the
(preferably not above 200 C) [8, 10, 12]; the differential redissolved carbon atoms become bound to the vacancies.
etching contrast obtained in this instance is illustrated in The deformed martensite then reassumes its comparatively
Fig. 10.10(e–h). The differentially etched zones are much light-etching characteristic.
less extensive than those illustrated in Fig. 10.10(a–d),
although the plastically deformed zones surrounding the Revealing Austenitic Grain Size
indentations would have been at least as extensive. This
implies that the threshold strain for differential etching to be There is commonly a need to determine the grain size of
observed is larger when the deformation is carried out the parent austenite of a quenched-and-tempered steel, for
between tempering treatments. The four etchants are which purpose it must be possible to recognize the sites of
approximately equally effective in this instance. The phe- the grain boundaries of the parent austenite. Moreover, it
nomenon probably is again a consequence of the fact that must be possible to do this at low magnifications, because
vacancies introduced by deformation are the preferred low- quantitative estimates of austenitic grain size usually are
123
Fig. 10.13 Revealing the

austenitic grain size in a 0.8% C
quenched-and-tempered steel.
0.79C–0.01Si–0.3Mn (wt%).
water quenched. Picral–HCI.
2509. (b) Austenitized at
Picral–HCI. 2509.
350 C for 30 min. Picral–HCI.
2509. (d) Austenitized at
Picral–HCI. 2509
carried out at a magnification of 1009. (Methods for another. The first possibility for distinguishing the parent
estimating grain size are described in Refs. [13, 14].) austenite grains therefore is to use an etching technique that
This is simple to do if a small volume fraction of a high- clearly delineates the colonies of martensite laths or plates.
temperature transformation product has formed during Standard etching methods are of limited effectiveness in this
quenching, because the transformation product invariably respect (see, for example, Fig. 10.8), but more satisfactory
forms preferentially at the austenitic grain boundaries; results can be obtained with an etchant known as Vilella’s
examples are given in Fig. 4.17(c), 4.17(g), 6.2(a), 6.3(c), reagent (picral with the addition of hydrochloric acid).
9.2(c), 9.3(a), and 9.4(a). Moreover, particles of cementite Examples of the effectiveness of this etchant in devel-
precipitated during tempering may under some circum- oping contrast between the grains of parent austenite are
stances decorate the austenitic grain boundaries, in which given in Fig. 10.11–10.13 for steels consisting of lath
event the grain boundaries can be delineated clearly by many martensite, a mixture of lath and plate martensite, and plate
etching methods (see, for example, Fig. 10.5b and 10.14c); martensite, respectively. The technique is quite effective
this occurs, however, only with higher tempering tempera- for steels consisting largely of lath martensite, irrespective
tures and longer tempering times than those normally used in of tempering treatment (Fig. 10.11). It is reasonably
practical quench-and-tempering heat treatments. effective in steels consisting of a mixture of lath and plate
In the more usual case confronting a microscopist, sig- martensite, more so after tempering at high temperatures
nificant amounts of a second phase are not present at the (Fig. 10.12). However, it is not very effective in steels
austenitic grain boundaries. The use of special etching consisting entirely of plate martensite, irrespective of
techniques then becomes necessary to distinguish the par- tempering treatment (Fig. 10.13).
ent austenitic grains. There are two methods available for There is a group of etchants that can delineate the au-
doing this. stenitic grain boundaries themselves, and an extensive
We saw in the discussion of Fig. 9.7 and 9.17 that each survey has shown that the most effective of this group is an
parent grain of austenite is filled during quenching by colo- aqueous solution of picric acid containing a small amount
nies of martensite laths or plates, and that the groups of of a wetting agent [15, 16] (it is desirable to precede
colonies formed in adjoining grains are different from one treatment in this etchant by a short etch in nital). This
123
Fig. 10.14 Temper brittleness.

(a and b) 0.5% C Ni–Cr alloy
(0.48C–0.19Si–0.84Mn–
1.35Ni–0.82Cr, wt%).
(a) Austenitized at 860 C, oil
quenched, tempered 1 h at
625 C, cooled slowly. Ethereal
picral. 2509. (b) Austenitized at
860 C, oil quenched, tempered
1 h at 625 C, water quenched.
Ethereal picral. 2509. (c) 0.4%
C (0.43C–0.18Si–0.75Mn).
Compare with Fig. 10.5(b).
quenched, tempered 48 h at
500 C. Ethereal picral. 2509
etchant delineates the austenitic grain boundaries well in the sodium alkyl sulfonate or sodium tridecylbenze sulfo-
both 0.2% C steels (lath martensite) and 0.5% C steels (lath nate originally used in this etchant [16], although their
and plate martensite) after tempering at high temperatures effectiveness varies slightly. A variant has also been
such as 550 C (Fig. 10.11h and 10.12f), and moderately developed in which a small amount of hydrochloric acid is
well after tempering at intermediate temperatures, such as added. The pre-etch in nital is eliminated, and the surface is
350 C (Fig. 10.11f and 10.12e). It does not delineate the polished lightly after etching to just remove the general
austenitic grain boundaries at all in either of these steels in etching effects [17]. Repeated etch–polish–etch cycles may
the as-quenched condition or after tempering at low tem- be used if judged desirable; the sought-after result is for the
peratures (Fig. 10.11b, d and 10.12d; the structural features austenitic grain boundaries to be delineated by a continu-
shown in these fields are largely those developed by the ous network of dark lines.
nital pre-etch). In a 0.7% C steel, the reagent is barely
effective after tempering at 550 C and not effective at all Temper Brittleness
after tempering at 350 C. It is not effective in 0.8% C steel
tempered at any temperature; consequently, illustrations of Temper brittleness is a term applied to a decrease in notch
the results obtained by this etchant are not included in toughness (more accurately, an increase in the ductile–
Fig. 10.13. It is not known what structural feature is brittle transition temperature in a notched impact test) that
responsible for the delineation of austenitic grain bound- may occur in steels heated in, or slowly cooled through, the
aries by these etchants. The most likely appears to be temperature range of about 375–575 C. The path of the
segregation of a solute to the grain boundaries, and perhaps fracture in embrittled material is largely along the prior
a phenomenon of this nature precedes the precipitation of austenitic grain boundaries. Although most commonly
cementite particles of the type illustrated in Fig. 10.5(a and encountered in steels in the quenched-and-tempered con-
b). Note also that the etchant is effective in developing dition, it also can occur in steels in the bainitic and ferritic
evidence of general segregation in higher-carbon steels conditions. The complexities of this phenomenon have
(Fig. 10.12e and f). been reviewed in Refs. [18, 19].
A number of proprietary wetting agents (also known as Temper brittleness generally does not occur in plain
surface-active agents or surfactants) can be substituted for carbon steels unless they contain more than about 0.5%
123
Fig. 10.15 Tempering of a 1%

C hypereutectoid steel. 1.05C–
0.24Si–0.25Mn (wt%).
355 HV. Picral. 1,0009.
275 HV. Picral. 1,0009.
HV. Picral. 1,0009.
HV. Picral. 1,0009
123
Mn. It is, in fact, basically a phenomenon of alloy steels— brittleness are, if anything, made more obvious by similar
particularly those containing both nickel and chromium but repolishing; the specimens shown in Fig. 10.14(a, b) were,
no molybdenum, and so warrants only brief mention here. in fact, briefly repolished, which is recommended practice
It is caused by segregation, without precipitation of a dis- for the picric–ether etching technique. Very long polishing
crete second phase, of certain specific elements to the au- treatments are necessary to remove the true temper-brittle
stenitic grain boundaries, the most effective of these etching effects.
elements being antimony, tin, phosphorus, and arsenic (in It is worth repeating that the networks of dark lines in
that order). The presence of only several hundred parts per Fig. 10.14(a) are grooves etched at austenitic grain bound-
million of antimony or tin is needed; larger amounts of aries at which segregation of phosphorus has occurred [20].
phosphorus are required, but the phosphorus contents of They are not indicative of a precipitate of a second phase. It
commercial steels of the highest attainable purity are still should also be recalled that there is one other effect that can
sufficient to induce the effect; amounts of arsenic not likely be responsible for development of intergranular facets on
to be encountered in practice are needed to induce signif- fracture surfaces produced by notched impact—namely,
icant brittleness. Segregation of the sensitizing elements to overheating; this was discussed in connection with
the austenitic grain boundaries was first deduced from Fig. 8.14–8.17. The two types of effects can be readily dis-
indirect evidence, but later was established positively by tinguished, however, by their respective responses to the
Auger electron emission spectroscopy analysis of fracture specific etching techniques that have been described.
surfaces produced in embrittled materials [20–22]. The
concentration of phosphorus at embrittled grain bound-
aries, for example, has been shown to be more than 200 Hypereutectoid Steels
times as high as the bulk; moreover, it has been established
that this high concentration of phosphorus is confined to a Hypereutectoid steels normally are tempered at tempera-
layer only about 10 A (i.e., several atom diameters) thick. tures that result only in precipitation of e carbides and
No embrittling phase has been detected at the grain perhaps also in transformation of retained austenite. The
boundaries, however, by even the most advanced metal- effects then observed by light microscopy are illustrated in
lographic techniques. Fig. 10.1(a–d).
It is somewhat remarkable, therefore, that this phe- Nevertheless, the effects observed due to cementite
nomenon can be detected at all by light microscopy. Cer- precipitation and growth after tempering at high tempera-
tain reagents etch grooves at the austenitic grain boundaries tures have been illustrated in Fig. 10.15 for comparison
of steels in the temper-brittle condition, but not in the same with Fig. 10.5 and 10.9. The phenomena involved are the
steels in the ductile condition, the severity of the etched same as for hypoeutectoid steels except, of course, that the
grooves corresponding in a general way with the severity volume fraction of cementite precipitated is larger.
of embrittlement. These etchants will do this, however,
only when segregation of phosphorus has occurred; they
are not effective for the other sensitizing elements. The
most effective of these reagents is a solution of picric acid References
in ethyl ether saturated with zephiran chloride. An example
of the grooves developed by this etchant in an embrittled 1. G.R. Speich, Tempered ferrous martensitic structures, in Metals
Handbook, vol. 8, 8th edn. (American Society for Metals, Metals
steel is given in Fig. 10.14(a), which can be compared with Park, 1973), p. 202
the result for the same steel in the ductile condition shown 2. G. Krauss, Steels: Heat Treatment and Processing Principles
in Fig. 10.14(b). Note that the severity of etched grooves in (ASM International, Materials Park, 1990)
Fig. 10.14(a) varies across the field shown; this presumably 3. K.H. Jack, Structural transformations in the tempering of high-
carbon martensitic steels. J. Iron Steel Inst. 169, 26 (1951)
corresponds with longitudinal bands of general phosphorus 4. Y. Imai, Phases in quenched and tempered steels. Trans. Jpn. Inst.
segregation. Met. 16, 721 (1975)
Some tempering treatments of a nonembrittling type 5. Y. Ohmori, Chi-carbide formation and its transformation into
may produce nearly continuous precipitates of cementite cementite during the tempering of martensite. Trans. Jpn. Inst.
Met. 13, 119 (1972)
along the austenitic grain boundaries (see, for example, 6. R.A. Grange, C.R. Hribal, L.F. Porter, Hardness of tempered
Fig. 10.5a and b). These precipitates are attacked strongly martensite. Met. Trans. A 8A, 1775 (1977)
by the picric–ether reagent, developing networks at the 7. C.R. Brooks, Principles of the Heat Treatment of Plain Carbon
austenitic grain boundaries (Fig. 10.14c), and are not to be and Low Alloy Steels (ASM International, Materials Park, 1996)
8. D.V. Wilson, Role of grain boundaries in the discontinuous
confused with those due to temper brittleness. The effects yielding of low-carbon steels. Acta Met. 5, 293 (1967)
due to cementite precipitation are removed by compara- 9. M.F. Miller, N.N. Breyer, Influence of deformation on the tem-
tively short repolishing, whereas those due to temper pering of 1045 martensite. Trans. Am. Soc. Met. 62, 891 (1969)
123
10. E.T. Stephenson, M. Cohen, The effect of prestraining and re- 17. A.W. Brewer, K.A. Elven, G. Krauss, Etchings and image ana-
tempering on AISI type 4340. Trans. Am. Soc. Met. 54, 72 (1961) lysis of prior austenite grain boundaries in hardened steels. Mater.
11. N.N. Breyer, N.H. Polakowski, Cold drawing of martensitic Charact. 27, 53 (1991)
steels to 400,000 psi tensile strength. Trans. Am. Soc. Met. 55, 18. C.J. McMahon, Jr., Temper Brittleness in Steels, ASTM Special
667 (1962) Technical Publication No. 407 (ASTM, Philadelphia, 1968),
12. W.R. Warke, J.L. McCall, An etching technique to reveal strain p. 127
in high-strength steel. Trans. Q. Am. Soc. Met. 59, 991 (1966) 19. C.J. McMahon Jr., Grain Boundaries in Engineering Materials
13. ASTM, Standard Methods for Estimating the Average Grain Size (Claiters Publishing, Baton Rouge, 1975), p. 525
of Metals (ASTM, Philadelphia, 1963) 20. I. Olefjord, Temper embrittlement, review 231. Int. Met. Rev. 23,
14. G.F. Vander Voort, Metallography, Principles and Practice 149 (1978)
(McGraw-Hill, New York, 1984) 21. C.J. McMahon, E. Furubayashi, H. Ohtani, H.C. Feng, A study of
15. D.R. Barraclough, Etching of prior austenite grain boundaries in grain boundaries during temper embrittlement of a low carbon
martensite. Metallography 6, 465 (1973) Ni–Cr steel doped with antimony. Acta Met. 24, 695 (1976)
16. G.F. Vander Voort, Wetting agents in metallography. Mater. 22. R. Viswanathan, Temper embrittlement in a Ni–Cr steel con-
Charact. 35, 135 (1995) taining phosphorus as impurity. Met. Trans. 2, 809 (1971)
123

Tempering Martensite

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

Tempering Martensite

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tempering Martensite

Uploaded by

Copyright:

Available Formats

Metallogr. Microstruct. Anal.

ASM International 2014

Fig. 10.1 Basic effects of

Fig. 10.2 Basic effects of

Fig. 10.3 Tempering of a

Precipitation of e carbides is sometimes known as the Carbide Precipitation

Fig. 10.3 continued

Fig. 10.4 Tempering of a

Hypoeutectoid Steels cementite particles is noticeably larger in the eutectoid

Fig. 10.5 Tempering of a

Fig. 10.6 Tempering of a

Fig. 10.7 Tempering of a range

Fig. 10.8 Tempering of a range

Fig. 10.9 Tempering of a 0.8%

Fig. 10.10 Effects of plastic

Fig. 10.11 Revealing the

Fig. 10.13 Revealing the

Fig. 10.14 Temper brittleness.

Fig. 10.15 Tempering of a 1%

You might also like