On the Selection of the Optimal Intercritical Annealing

Temperature for Medium Mn TRIP Steel

SEAWOONG LEE and BRUNO C. DE COOMAN

A model is proposed to predict the room temperature austenite volume fraction as a function of
the intercritical annealing temperature for medium Mn transformation-induced plasticity steel.
The model takes into account the influence of the austenite composition on the martensite
transformation kinetics and the influence of the intercritical annealing temperature dependence
of the austenite grain size on the martensite start temperature. A maximum room temperature
austenite volume fraction was obtained at a specific intercritical annealing temperature TM.
Ultrafine-grained ferrite and austenite were observed in samples intercritically annealed below
the TM temperature. The microstructure contained a large volume fraction of athermal mar-
tensite in samples annealed at an intercritical temperature h
igher than the TM temperature.

DOI: 10.1007/s11661-013-1860-2
 The Minerals, Metals & Materials Society and ASM International 2013

I. INTRODUCTION II. EXPERIMENTAL

THERE is increasing interest in the medium Mn
transformation-induced plasticity (TRIP) steel as it is
widely thought that these types of ferrous alloys have a
good chance of being applied in the near future for
automotive car body applications due to their superior
combination of ultimate tensile strength (UTS) and total
elongation (TE).[1–5] It has, however, been known for
some time that their mechanical properties are extremely
sensitive to the annealing temperature. Jun et al.[4]
showed that the mechanical properties are drastically
influenced by the intercritical annealing temperature as
it controls both the volume fraction and the thermody-
namic stability of the austenite. De Moor et al.[6] have
proposed a method to predict the austenite fraction as a
function of the annealing temperature using thermody-
namic data. However, they reported a rather large
discrepancy between the calculated volume fractions
and their experimental data. In the present study, a
modified calculation method is therefore suggested and
it is shown that the method can be used to predict
correctly the room temperature austenite fraction as a
function of the intercritical annealing temperature. In
addition, the proper annealing temperature range for
optimal mechanical properties of medium Mn TRIP
steel is also suggested based on the proposed approach.
SEAWOONG LEE, Graduate Student, is with the Graduate
Institute of Ferrous Metallurgy, Pohang University of Science and
Technology, Pohang 790-784, Republic of Korea. BRUNO C. DE
COOMAN, Professor, Director, is with the Materials Design
Laboratory, Graduate Institute of Ferrous Metallurgy, Pohang
University of Science and Technology. Contact e-mail: decooman@
postech.ac.kr
Manuscript submitted March 27, 2013.
The chemical composition of the medium Mn TRIP
steel studied in the present work was Fe-0.3 pct C-6.0 pct
Mn (in mass pct). An ingot was prepared by vacuum
induction melting, reheated at 1250 C, hot rolled, and
cooled by water spray quenching to avoid formation of
ferrite during cooling. The final microstructure was fully
martensite. The hot-rolled sheet was then cold rolled to a
final thickness of 2 mm. A fully deformed martensitic
microstructure was obtained after cold rolling. The
intercritical annealing temperatures were selected by
means of dilatometry. The Ac1 and Ac3 temperatures
were determined to be 853 K and 993 K (580 C and
720 C), respectively, and the temperature range of 873 K
to 953 K (600 C to 680 C) was selected for the inter-
critical annealing, using a temperature interval of 20 K
(20 C). The annealing times varied from 30 minutes to
2 hours. The annealing treatments were carried out in a
dilatometer and in a box furnace. The dilatometric
samples (5 9 1 9 10 mm3) were used to measure the
amount of retained austenite by means of the magnetic
saturation method after intercritical annealing. ASTM-
E8 standard samples with a 50-mm gage length were used
for tensile testing. The tensile axis of the samples was
aligned parallel to the rolling direction. The samples were
intercritically annealed in a box furnace. The tensile tests
were carried out at a strain rate of 103 s1 in a ZWICK
Z100 universal tensile testing machine. The uniformly
deformed parts of the gage section of the fractured
samples were cut out after tensile testing and used to
measure the volume fraction of retained austenite by
means of the magnetic saturation method.
III. MODEL DESCRIPTION
Figure 1 illustrates the two-stage thermal treatment
used to produced medium Mn TRIP steel. In the first
step, the material is fully austenitized at a temperature

METALLURGICAL AND MATERIALS TRANSACTIONS A

 γ γ
973 (700)
γ+α
873 (600)
773 (500)
Temperature, K (°C)
673 (400)
573 (300)
Ms
473 (200)
373 (100)
α’
273 (0) Mf
Time
(a)
Fig. 1—Principle of the processing of medium Mn TRIP steel: (a) thermal cycle, (b) phase diagram, and intercritical annealing temperature
dependence of the (c) C and (d) Mn content in the austenite phase.
above Ac3 and quenched to room temperature. The
microstructure consists of a large amount of martensite
and a small fraction of retained austenite, as the Ms
temperature is above room temperature and the Mf
temperature is slightly below room temperature as
shown in Figure 1(a). In the second stage, the martensite
is annealed in the intercritical temperature range. The
martensite reverses partially to austenite by nucleation
and diffusion-controlled growth, and the remaining
martensite forms recrystallized ferrite. During the inter-
critical hold, the C and Mn partition to the austenite.
Depending on the position of the Ms–Mf temperature
interval of the intercritical austenite, the retained aus-
tenite will be fully stabilized (Ms < RT), partially
stabilized (Ms > RT > Mf), or fully transformed to
martensite (Mf > RT) at room temperature. The pre-
diction of the room temperature austenite volume
fraction is entirely dependent on the correct determina-
tion of the composition and the grain size of the
intercritical austenite.
The empirical equation proposed by Koistinen and
Marburger[7] was used to determine the volume fraction
of a thermally transformed martensite:
h i
fa0 ¼ 1  exp aðMs  TÞb ½1
Here, fa0 is the volume fraction of martensite. a and b
are coefficients related to the kinetics of the transfor-
mation. Ms is the martensite start temperature in K.
According to Koistinen and Marburger, the constants
a and b are 0.011 and 1, respectively, in the Fe-C
binary alloy system. For steels with a more complex
γ+α
α+ γ + θ
α+ θ
RT
Ms’
0 0.2 0.4 0.6 0 0.2 0.4 0.6 0 10 20
C, mass-% Cγ , mass-% Mnγ , mass-%
(b) (c) (d)
composition, the coefficients of a and b are composi-
tion dependent. The composition dependence of a and
b is given by[8]
a ¼0:0076  0:0182 pct C þ 0:00014 pct Mn
½2
b ¼1:4609 þ 0:4483 pct C  0:0545 pct Mn
Here, pct C and pct Mn are the C and Mn content of the
austenite in mass pct.
It is well known that the Ms temperature of austenite
is affected by its chemical composition and its grain
size.[9] Jimenez-Melero et al.[10,11] have derived an
empirical equation which takes into account the effect
of the grain size on the Ms temperature. This equation
was further modified by Lee et al.[8]:
Ms ð CÞ ¼ 545  423 pct C  30:4 pct Mn  60:5 V1=3
c ;
½3
where pct C and pct Mn are the C and Mn content of
the intercritical austenite, in mass pct, and Vc is the
average austenite grain volume in lm3.
The equilibrium computational thermodynamic soft-
ware package ThermoCalc was used to calculate the
equilibrium composition of the intercritical austenite,
needed to compute the value of a, b, and the Ms
temperature. The phase fraction of the austenite was
also determined as a function of the intercritical
annealing temperature. The room temperature austenite
fraction after annealing was expressed as follows:
fret;c ¼ fTC;c  fa0 ½4
METALLURGICAL AND MATERIALS TRANSACTIONS A
fret,c is room temperature retained austenite fraction,
fTC,c is austenite fraction at the intercritical annealing
temperature calculated by means of ThermoCalc, and
fa0 is the volume fraction of transformed martensite after
annealing, as determined from Eq. [1].
The average austenite grain volume and grain size
were determined as follows. As the reversed martensite
transformation is a nucleation and diffusion-controlled
growth process, the DICTRA software package was
used in conjunction with the MOBFE mobility database
to estimate the half-thickness of the reversely trans-
formed austenite. Figure 2(a) shows the principle of the
calculations. For the DICTRA simulation, a steady-
state solute C and Mn concentration profile was
assumed at the moving interphase boundary. The
detailed solution for the diffusion equation is given in
elsewhere.[10,11] According to Nakaba et al.,[12] austenite
which reversely transforms from martensite has an
equiaxed microstructure when the austenite nucleates
at the prior austenite grain boundaries (PAGBs). It has
an acicular structure when the austenite nucleates at a
lath boundary. The retained austenite grains observed
by transmission electron microscopy in the present work
had an equiaxed microstructure. It is therefore reason-
able to assume that the austenite nucleated mainly at the
PAGBs. Figure 2(b) shows a schematic for the reverse
austenite nucleation mechanism, during which the
austenite is nucleated at the PAGB and grows into
the austenite grain. In this case, the system size for the
DICTRA simulation should be approximately half the
length of a typical martensite lath. Figure 3 shows a
TEM micrograph of a martensite lath in cold-rolled
medium Mn TRIP steel. The lath size is ~5 lm.
Therefore, a system size of 2.5 lm was selected for the
simulation to determine the austenite grain size after
annealing.
IV. RESULTS AND DISCUSSION
Figure 4(a) shows the C and Mn content in the
austenite obtained by Thermocalc as a function of the
annealing temperature. The highest C content is
obtained at about 923 K (650 C). The Mn content
Fig. 3—TEM micrograph of a martensite lath in cold-rolled medium Mn steel.
METALLURGICAL AND MATERIALS TRANSACTIONS A
shows a gradual decrease with increasing annealing
temperatures. Lee et al.[13] showed that the Mn content
of the austenite reached the equilibrium concentration
within an annealing time of 180 seconds, even at low
annealing temperatures. It is therefore reasonable to
assume that the C and Mn contents in the intercritical
austenite are close to their equilibrium values as
calculated by Thermocalc (Figure 4(a)). Figure 4(b)
shows the calculated Ms temperature as a function
of the annealing temperature. The calculated Ms
d/2 at Δ t
Cγ/α
γ α
Cο
Cα/γ
Δt
(a)
C, Mn
γ α
L
γ γ
γ
Stage 1 Stage 2
γ
γ
γ → α’ α’ → γ + α
(b)
Fig. 2—(a) Schematic diagram showing boundary condition for the
DICTRA simulation. Initial solute profile (full line), solute profile at
time Dt (dotted line), Ca/c and Cc/a are interphase solute concentra-
tion, Co is nominal solute concentration. (b) Schematic of the two
stages required to produce the ultrafine-grained microstructure in
medium Mn TRIP steel showing the calculation domain used to
determine the size of the c cell used to compute the diameter of the
intercritical austenite grains. L is the average half-length of the lath.
 0.60 24 defined, as is the intercritical temperature for which
0.55 22 the maximum volume fraction of austenite can be
retained upon cooling to a room temperature of

Mn in austenite, wt%
0.50 20
C in austenite, wt%

0.45 18
298 K (25 C). A sharp drop in the austenite fraction
is observed for intercritical annealing temperatures
0.40 16
higher than TM.
0.35 14 As can be seen from the figure, the austenite volume
0.30 12 fraction after intercritical annealing at 873 K (600 C)
0.25 10 for 1 hour is different from the volume fraction for an
0.20 8 annealing time of 24 hours at the same temperature.
0.15 6 This is due to the fact that the experimentally measured
0.10 4
austenite volume fraction after a long annealing time
773 823 873 923 973 1023 1073 corresponds to the equilibrium austenite volume frac-
(500) (550) (600) (650) (700) (750) (800) tion. Since 1 hour is insufficient time to reach equilib-
Annealing temperature, °C rium, the experimentally measured room temperature
(a) austenite fraction is always slightly lower than the
predicted volume fraction for intercritical annealing
300 temperatures below TM.
Figure 6(a) includes the calculated room temperature
200 austenite fraction taking into account the austenite grain
size effect, assuming a fixed 1-lm grain diameter (dotted
Ms temperature, °C

line), and shows that the predicted austenite is closer to

100
Room temperature
0 grain size effect (full line). The TM temperature of 923 K
-100
-200
773 823 873 923 973
(500) (550) (600) (650) (700)
Annealing Temperature, °C
(b)
Fig. 4—(a) Calculated intercritical annealing temperature depen-
dence of the C and Mn content in austenite from ThermoCalc. (b)
Intercritical annealing temperature dependence of the Ms tempera-
ture for different values of the austenite grain size.
temperature of the austenite is higher than room
temperature after intercritical annealing in the temper-
ature range from 893 K to 1078 K (620 C to 800 C)
when the grain size effect is not considered in the
calculation. This results in athermal martensitic trans-
formation upon cooling. When the influence of the
austenite grain size is considered, there is a decrease of
the Ms temperature for all annealing temperatures. As
the austenite grain size is reduced, the Ms temperature
decreases to a temperature below 298 K (25 C). This
results in the presence of a considerable volume
fraction of retained austenite at room temperature. It
is therefore of prime importance to consider the
austenite grain size to predict the room temperature
austenite fraction.
Figure 6 shows the measured and calculated room
temperature austenite volume fraction after annealing
at different intercritical temperatures. The room tem-
perature austenite volume fraction reaches a maximum
value for an intercritical annealing temperature TM of
~923 K (650 C). The TM temperature is henceforth
the experimental data than the calculated room temper-
ature austenite fraction which does not consider the
No grain size (650 C) is also predicted correctly when the austenite
grain size is considered. There is, however, still a small
d = 1μm
discrepancy between the experimental and the predicted
d = 0.2μm austenite volume fraction for annealing temperatures
higher than 953 K (680 C). This error is due to the fact
1023 1073 that the austenite grain size is assumed to be constant at
(750) (800) all intercritical annealing temperatures. This is, however,
not the case. Due to grain growth the rate of which is
dependent on annealing temperatures, the austenite
grain size is different at each annealing tempera-
ture.[8,12,14,15] Therefore, DICTRA software package
with MOBFE mobility database was used to estimate
the half-thickness of reversely transformed austenite for
an annealing time of 1 hour. Figure 5(a) shows the half-
length of reversely transformed austenite as a function
of the temperature. The diameter of the austenite
increases as the intercritical annealing temperature
increases. The half-length value was used as the radius
of a spherical austenite grain to determine the Ms
temperature taking into account the austenite grain size
at the experimentally tested intercritical annealing tem-
peratures. The Ms temperature is again calculated using
varying grain sizes with temperature. As can be seen
form the Figure 5(b), the Ms temperature drastically
decreases as annealing temperature is decreased when
temperature-dependent austenite grain size is consid-
ered. In addition to this, at high annealing temperature,
the Ms temperature is closed to the temperature that
does not consider the austenite grain size for Ms
temperature calculation. It means that austenite stability
that comes from austenite grain size is no longer acting
as a stabilization mechanism at high annealing temper-
atures.
Figure 6(b) shows the calculated results for a varying
intercritical austenite grain size (dotted line) and for a
grain diameter fixed at a value of 1 lm (full line).
It is clear that the model taking into account the

METALLURGICAL AND MATERIALS TRANSACTIONS A

 1.0
No grain size effect
Grain size effect (d=1um)
0.8

Austenite fraction
Exp. for 24h TM
Exp. for 2h Athermal
0.6 martensitic
Exp. for 1h
Exp. for 30min
0.4

0.2

0.0
773 823 873 923 973 1023 1073
(500) (550) (600) (650) (700) (750) (800)

(a) Temperature, K (°C)

(°C)
(a)
300
1.0
Fixed grain size (d=1um)
200 Varying grain size on annealing temperatures
Ms temperature, °C

0.8 Exp. for 1h

TM

Austenite fraction
100
0.6 γ
Room temperature
γ α
0
α α
0.4
-100
No grain size
d = 1μm
-200 0.2
Temperature dependent d

773 823 873 923 973 1023 1073

(500) (550) (600) (650) (700)
Annealing temperature, °C
(b)
Fig. 5—(a) Simulated radius of austenite as a function of the inter-
critical temperature for an annealing time of 1 h as obtained with
the DICTRA software package. (b) Calculated Ms temperature with
varying grain sizes at different annealing temperatures.
temperature-dependent austenite grain size gives an
excellent prediction of the austenite phase fraction even
at high temperatures.
The schematic diagram in Figure 6(b) shows the
phase content as a function of the annealing tempera-
ture. A ferrite-austenite microstructure is obtained after
annealing below TM. On the other hand, a ferrite-
austenite-martensite phase mixture is obtained after
annealing above TM.
Figure 7 shows transmission electron microscopy
(TEM) micrographs for samples annealed at 893 K
(620 C), i.e., a temperature lower than TM, and at
953 K (680 C), i.e., a temperature higher than TM.
Equiaxed ultrafine-grained (UFG) austenite and fer-
rite are observed in the sample annealed at 893 K
(620 C). Large amounts of athermal martensite are
observed in the sample annealed at 953 K (680C). In
addition, ferrite with a high dislocation density is
also present in the sample annealed at 953 K
(680 C).
0.0
(750) (800)
773 823 873 923 973 1023 1073
(500) (550) (600) (650) (700) (750) (800)
Temperature, K (°C)
(°C)
(b)
Fig. 6—(a) Austenite volume fraction as a function of the annealing
temperature, without taking the austenite grain size into account
(full line), and assuming a single austenite grain diameter of 1 lm
(dotted line). The symbols represent experimental data for an
annealing time of 30 min (e), 1 h (h), 2 h (s), and 24 h (D). (b)
Austenite volume fraction as a function of the annealing temperature
assuming a single austenite grain diameter of 1 lm (full line), and
assuming an annealing temperature-dependent austenite grain size
(dotted line). The pie charts indicate the equilibrium phases fractions
at the annealing temperature.
The stress–strain curves for samples annealed at four
characteristic temperatures are shown in Figure 8. The
sample annealed at 893 K (620 C) (Figure 8(a)) shows
a very limited work-hardening rate due to the absence
of the plasticity-enhancing strain-induced martensite
transformation (TRIP effect). As shown in the Table I,
there is no martensitic transformation in the sample
annealed 893 K (620 C). This implies that the work
hardening is limited to dislocations’ storage in UFG
grains. When the grain size is ultrafine, i.e., smaller than
a critical value of ~1 lm, the classical dislocation
generation and annihilation mechanisms are altered in
such a manner that the dislocation storage is strongly
reduced and, as a consequence, the material has a low

METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 7—TEM micrographs of samples intercritically annealed at 893 K (620 C) (a, b) and 953 K (680 C) (c, d). Micrographs of (a) retained
austenite and (b) recrystallized ferrite in the sample annealed at 893 K (620 C). (c) Lath martensite and (d) ferrite in the sample annealed at
953 K (680 C).
strain-hardening capacity.[16,17] The mechanical behav-
ior of the samples annealed at 933 K and 953 K (660 C
and 680 C) is shown in the Figure 8(c). The sample
annealed at 953 K (680 C) has a high tensile strength,
but a limited ductility. This is mainly due to the
presence of a large volume fraction of athermal mar-
tensite. On the other hand, the sample annealed at
933 K (660 C) has a small volume fraction of athermal
martensite, but a high amount of retained austenite. It
has strength comparable to the strength of the sample
annealed at 953 K (680 C) and a higher elongation.
This is mainly due to the TRIP effect which is activated
during tensile deformation. The sample annealed at
913 K (640 C) (Figure 8(b)) has a remarkably high
work-hardening rate due to the TRIP effect. As indi-
cated in Table I, a large fraction of the retained
austenite is transformed to martensite during tensile
testing in this case. The sample annealed at 933 K
(660 C) shows a much higher work-hardening rate
compared to the sample annealed at 913 K (640 C) at
the start of the deformation. However, since most of the
martensite is formed in the early stage of deformation, it
shows a lower ductility than the sample annealed at
933 K (660 C). Pie charts illustrating the phase content
are also shown in Figure 8. At temperatures below the
TM temperature, the microstructure consists of a
mixture of UFG ferrite and austenite. At higher
temperatures, i.e., 933 K and 953 K (660 C and
680 C), the microstructure contains a large volume
fraction of athermal martensite and a small fraction of
UFG ferrite and austenite.
V. CONCLUSIONS
In conclusion, the austenite stability plays a critical
role in the mechanical properties of the medium Mn
TRIP steels. By using the proposed method to
determine the room temperature volume fraction of
austenite in medium Mn TRIP steel, it is possible to
select an annealing temperature that takes full advan-
tage of strain-induced martensitic transformation
METALLURGICAL AND MATERIALS TRANSACTIONS A
 1800 during tensile testing. The optimal intercritical anneal-
1500 620_1h ing temperature for medium Mn TRIP steel should be
1200 chosen slightly lower than the TM temperature to
900
γ α avoid the presence of athermal martensite in the
600 microstructure.
300
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30
(a) ACKNOWLEDGMENTS
True stress, MPa

1800 This research was supported by the WCU (World

1500 Class University) program through the National
640_1h
1200 Research Foundation of Korea funded by the Ministry
900 of Education, Science and Technology (R32-10147 and
600 γ α
R31-2008-000-10075-0).
300
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30
(b) REFERENCES
1800 1. R.L. Miller: Metall. Trans., 1972, vol. 3, p. 905.
1500
660_1h 2. H. Dong, W. Cao, J. Shi, C. Wang, and M. Wang: HMnS 2011
680_1h Conf. Proc., Seoul, 2011.
1200
900 γ 3. S. Lee, S.-J. Lee, S. Kumar, K. Lee, and B.C. De Cooman: Metall.
Mater. Trans. A, 2011, vol. 42A, p. 3638.
600 α α' 4. H.J. Jun, O. Yakubovsky, and N. Fonstein: HMnS2011 Conf.
Proc., Seoul, 2011.
300
0 5. Y.-U. Heo, Y.-Y. Song, S.-J. Park, H.K.D.H. Bhadeshia,
0.00 0.05 0.10 0.15 0.20 0.25 0.30 and D.-W. Suh: Metall. Mater. Trans. A, 2012, vol. 43A, p. 1731.
6. E. De Moor, D.K. Matlock, J.G. Speer, and M.J. Merwin: Scripta
True strain Mater., 2011, vol. 64, pp. 185–88.
(c) 7. D.P. Koistinen and R.E. Marburger: Acta Metall., 1959, vol. 7,
pp. 59–60.
8. S.-J. Lee, S. Lee, and B.C. De Cooman: Int. J. Mater. Res. (in
Fig. 8—True stress–strain curve of samples intercritically annealed at
press).
(a) 893 K (620 C), (b) 913 K (640 C), and (c) 933 K (660 C) (gray
9. S. Lee, S.-J. Lee, and B.C. De Cooman: Scripta Mater., 2011,
line), 953 K (680 C) (black line). The pie charts indicate the phase
vol. 65, pp. 225–28.
content of the samples after intercritical annealing.
10. E. Jimenez-Melero, N.H. van Dijk, L. Zhao, J. Sietsma, S.E.
Offerman, J.P. Wright, and S. Van der Zwaag: Scripta Mater.,
Table I. Intercritical Annealing Temperature Dependence 2007, vol. 56, pp. 421–24.
11. H. Luo, J. Sh, C. Wang, W. Cao, X. Sun, and H. Dong: Acta
of the Austenite Volume Fraction After Annealing and After
Mater., 2011, vol. 59, pp. 4002–14.
Tensile Testing 12. N. Nakada, T. Tsuchyama, S. Takaki, and N. Miyano: ISIJ Int.,
2011, vol. 51, pp. 299–304.
Retained Retained Percentage 13. S.-J. Lee, S. Lee, and B.C. De Cooman: Scripta Mater., 2011,
Annealing Austenite Austenite Transformed vol. 64, pp. 649–52.
Temperature Fraction Before Fraction After Austenite 14. M. Tokizane, N. Matsumura, K. Tsuzaki, T. Maki, and I.
[K (C)] Deformation Deformation (Percent) Tamura: Metall. Trans. A, 1982, vol. 13A, pp. 1379–88.
15. T. Ogawa, N. Maruyama, N. Sugiura, and N. Yoshinaga: ISIJ
893 (620) 0.28 0.28 0 Int., 2010, vol. 50, pp. 469–75.
913 (640) 0.49 0.10 80 16. S. Lee, Y. Estrin, and B.C. De Cooman: Metall. Mater. Trans. A,
933 (660) 0.69 0.17 75 2013, vol. 44A, pp. 3136–46.
953 (680) 0.21 0.08 62 17. O. Bouaziz, Y. Estrin, Y. Brechet, and J.D. Embury: Scripta
Mater., 2010, vol. 63, pp. 477–79.

METALLURGICAL AND MATERIALS TRANSACTIONS A