Materials Science and Engineering A 438–440 (2006) 300–305

Microstructures and stability of retained austenite in TRIP steels

X.D. Wang

a

,

∗

, B.X. Huang

a

, Y.H. Rong

a

, L. Wang

a

,

b

a

School of Materials Science and Engineering, Shanghai Jiao Tong University, Shanghai 200030, China

b

Baosteel Research and Development Technology Center, Shanghai 201900, China

Received 27 June 2005; received in revised form 13 January 2006; accepted 11 February 2006

Abstract

Transformation induced plasticity (TRIP) steels exhibit a combination of high strength and ductility due to their multi-phase microstructure,

including ferrite, bainite and retained austenite which transforms to martensite under the external stress. The characterization of microstructures
is necessary for understanding the relationship between microstructure and property. In the present work, an effort to determine the volume
fraction of various phases was made for the conventional TRIP steel containing silicon. The microstructures in the TRIP steel were characterized
by optical microscopy, scanning electron microscopy and transmission electron microscopy, especially, an effective method was developed to
identify multi-phase microstructures by atomic force microscopy based on the height difference. Furthermore, the stability of retained austenite
determining TRIP effect was evaluated by electrical resistance tests and tensile tests. The results show that retained austenite does not generate
martensitic transformation at

−80

◦

C and exhibits a good thermodynamic stability, and the transition temperature from stress-induced martensitic

transformation to strain-induced martensitic transformation is determined as about

−5

◦

C, and thus strain-induced martensitic transformation over

−5

◦

C (somewhat lower than room temperature) is favorable for the application of TRIP steels in the automobile industry.

© 2006 Elsevier B.V. All rights reserved.

Keywords: TRIP steel; Microstructure; Retained austenite;

M

σ

s

temperature; Mechanical properties

1. Introduction

Low-carbon multi-phase sheet steels, which were developed

for automotive applications, have attracted a growing interest in
recent years due to their high strength and ductility combination.
These excellent mechanical properties mainly arise from a com-
bination of a ferrite matrix, for ductility, bainite, for strength, and
retained austenite, for uniform elongation produced by marten-
site transformation from austenite under the influence of external
tensile stress

[1–3]

. The multi-phase microstructures are formed

after an intercritical annealing followed by austempering at the
bainite transformation temperature. During the above heat treat-
ment, the austenite phase is stabilized by the enrichment of
carbon due to the presence of alloying elements such as sili-
con or aluminum. From this viewpoint, the volume fraction, the
distribution of various phases, especially the stability of retained
austenite is of great importance to the mechanical properties of
transformation induced plasticity (TRIP) steels. The stability
of retained austenite includes thermodynamic and mechanical

∗

Corresponding author. Tel.: [phone].

E-mail address:

[e-mail]

(X.D. Wang).

stabilities. The M

s

temperature is used to characterize the ther-

modynamic stability of retained austenite against transformation
under cooling, whereas the

M

σ

s

temperature is used to evalu-

ate mechanical stability against stress-assisted transformation.
At the

M

σ

s

temperature, the stress reaches the yield stress for

slip in the austenite phase

[4–6]

. In the present paper, a devel-

oped method, measuring the height of various phases by means
of atomic-force microscopy (AFM) after etching, is built eas-
ily and fast to identify various phases in a TRIP steel, and the
validity of the method is verified by comprehensive analysis by
use of optical microscopy (OM), scanning electron microscopy
(SEM) and transmission electron microscopy (TEM). Moreover,
the volume fraction of various phases is determined by combina-
tion of thermodynamic calculation, metallographic examination
and X-ray diffraction (XRD). The stability of retained austenite
is further investigated by means of the measurement of M

s

and

M

σ

s

temperature, and the different mechanical behaviors in the

TRIP steel are explained.

2. Experimental procedure

The chemical composition of the TRIP steel investigated in

the present work is listed in

Table 1

, together with Ac

1

, Ac

3

0921-5093/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:

10.1016/j.msea.2006.02.149

X.D. Wang et al. / Materials Science and Engineering A 438–440 (2006) 300–305

301

Table 1
Chemical composition of the TRIP steel (mass%)

C

Si

Mn

P

S

Al

N

Ac

1

(

◦

C)

Ac

3

(

◦

C)

T

50

(

◦

C)

0.11

1.19

1.67

0.013

0.006

0.038

0.003

683

861

803

temperatures and T

50

temperature standing for the volume frac-

tions of ferrite and austenite in the intercritical temperature are
50%, respectively, which is calculated by thermo-calc software
system. It was suggested that slow cooling from an intercritical
annealing temperature near Ac

3

to a temperature close to Ac

1

before rapid cooling could lead to a better balance of strength
and ductility

[7]

. Therefore, the intercritical annealing, as shown

a marked “1” line in

Fig. 1

, was carried out at 800

◦

C for 70 s,

then slow cooling to 690

◦

C with a cooling rate of 10 s

−1

, isother-

mally held at 400

◦

C for 4 min followed by water quenching.

In order to check the volume fraction of ferrite phase calcu-

lated by thermo-calc software, a heat treatment was designed
as follows. A sample was treated by intercritical annealing at
800

◦

C for 70 s, then slow cooling to 690

◦

C followed by water

quenching to room temperature, as shown a marked “2” line in

Fig. 1

, and the volume fraction of ferrite in the sample was deter-

mined by quantitative metallographic examination. The volume
fraction of retained austenite in undeformed specimen or ten-
sile specimen was measured with XRD. The specimens for OM
and SEM were prepared according to the traditional procedure,
that is, mechanical grinding and polishing up with 1

␮m dia-

mond paste, followed by etching with 2% nital for 15 s. Three
specimens for AFM were prepared. One of them was prepared
as, same as that of OM, the others were polished up with 1

␮m

diamond paste, then were deformed up to 3% strain, of which
one specimen was further etched with 2% nital for 15 s. Both
deformed and undeformed specimens were observed by TEM.

The thermodynamic stability of retained austenite was deter-

mined by measuring the M

s

temperature using electrical resis-

tance test. In order to characterize the mechanical stability of
retained austenite, the

M

σ

s

temperature of the TRIP steel was

determined by measuring its yield strength. The measurement
method described by Barbe et al.

[6]

, Berrahmoune et al.

[8]

was

conducted, and tensile tests were carried out at a strain rate of
5

× 10

−3

in the rolling direction in temperature range from 20

to

−40

◦

C.

Fig. 1. Heat treatment process.

3. Results and discussion

3.1. Determination of volume fractions of various phases

Microstructure observed with SEM on specimen treated as a

marked “2” line in

Fig. 1

is shown in

Fig. 2

. The volume frac-

tion of ferrite (f

F

) is determined by quantitative metallographic

examination as about 56.1%, which is consistent with thermo-
dynamic calculation. By XRD analysis, the volume fraction of
retained austenite (f

Ar

) in undeformed samples is determined as

about 16%, and about 7% (f

Ar



) in deformed samples. Therefore,

the volume fraction of bainite (f

B

) can be calculated, namely

f

B

= 1

− f

F

− f

Ar

= 27.9%, while the volume fraction of marten-

site (f

M

) in deformed samples due to martensitic transformation

from austenite can be determined as 9% (f

M

= f

Ar

− f

Ar



). It indi-

cates that about 7% of retained austenite does not transform to
martensite during tensile process.

3.2. Phase identification by AFM

OM or SEM observation on specimens etched by nital cannot

be used to distinguish bainite, retained austenite and marten-
site by their contrast, while ferrite matrix is clearly observable,
as black contrast in

Fig. 2

, and white contrast in

Fig. 3

. This

is because the etching solution of nital attacks ferrite grains
and grain boundaries while leaving austenite and martensite

Fig. 2. SEM images treated as line 2 in

Fig. 1

.

302

X.D. Wang et al. / Materials Science and Engineering A 438–440 (2006) 300–305

Fig. 3. Optical micrographs of the TRIP steel.

intact

[9]

, which results in different height of various phases.

The image contrast in OM is different from that in SEM since
their imaging principles are different. In recent years, although
a modified color etching method

[9,10]

was used to distinguish

bainite from ferrite and tempering martensite from austenite in
TRIP steels, this technique is somewhat complicate since the
strain-induced martensite must undergo tempering treatment.

A developed method was presented in this paper, in which

the height of various phases in undeformed samples after etching
was firstly measured by means of AFM with reference to, the
height of retained austenite due to not suffering etching, then the
relative height of martensite with retained austenite in deformed
sample was determined, finally the height of various phases of
the deformed sample was measured. Let the height of retained
austenite be zero, and then the height of martensite is positive
while the height of the ferrite or bainite is negative.

Fig. 4

is an AFM image of various phases in an undeformed

sample. In

Fig. 4

(a), the retained austenite was located in the

cross of three lines, as reference object for measuring the height
of various phases. The relative height of various phases is shown
in AFM image of section analysis (

Fig. 4

(b)), such as the phase

with the relative height

−201.7 nm is ferrite, −97.6 nm is bainite

and +1.5 nm is austenite.

Fig. 5

(a and b) are, respectively, three-

dimensional and two-dimensional images of martensite surface
relief of a deformed sample, and its section analysis is demon-
strated in

Fig. 5

(c). The height of martensite surface relief is

about 50.3 nm.The statistical results of the relative height and
the surface roughness of various phases are respectively listed
in

Tables 2 and 3

for undeformed and deformed samples. The

statistical results indicate that the increasing sequence of abso-
lute height of various phases is ferrite, bainite, austenite and
martensite, and the increasing sequence of surface roughness

Table 2
Statistical results of the surface roughness and height difference in the unde-
formed sample

Phase

Retained austenite

Bainite

Ferrite

Mean height difference (nm)

0

−89.2

−196.4

Surface roughness (nm)

5

9

18

Fig. 4. AFM image of various phases in an undeformed sample: (a) two-
dimensional image and (b) section analysis.

of various phases is austenite, martensite, bainite and ferrite.
Therefore, the multi-phase microstructures in TRIP steel can be
easily identified by the above AFM method.

3.3. Stability of retained austenite

The stability of retained austenite includes thermody-

namic and mechanical stabilities.

Fig. 6

depicts the variation

Table 3
Statistical analysis of the surface roughness and height difference in the deformed
sample

Phase

Martensite

Retained
austenite

Bainite

Ferrite

Mean height difference

(nm)

+58.5

0

−79.7

−189.3

Surface roughness

(nm)

7

5

10

18

X.D. Wang et al. / Materials Science and Engineering A 438–440 (2006) 300–305

303

Fig. 5. AFM image of martensite surface relief on polished surface after deformation: (a) three-dimensional image, (b) two-dimensional image and (c) section
analysis.

of electrical resistance as a function of temperature dur-
ing the thermal cycle from 20 to

−80

◦

C. In the electrical

resistance–temperature curve, there exists no exothermal peak
during cooling, which indicates that martensitic transformation
does not occur. This experiment demonstrates a good thermo-
dynamic stability of the retained austenite.

Fig. 7

shows the

yield and tensile strength–temperature curve of the TRIP steel
in temperatures ranging from 20 to

−40

◦

C. It is worthy to note

that the yield strength increases with decreasing temperature

Fig. 6. Electrical resistance of the sample at different temperatures.

up to

−5

◦

C, then decreasing with decreasing temperature

till

−20

◦

C, further, it increases with decreasing temperature.

This abnormal phenomenon at

−5

◦

C was explained as the

transition point (called

M

σ

s

) from stress-induced martensitic

transformation to strain induced martensitic transformation

[5]

.

The relationship between the yield strength and temperature can
be reasonably explained by on modulus change of TRIP steel.
The modulus of the TRIP steel reflects mean modulus of various
phases, and ferrite and bainite have great weight with mean
modulus. The TRIP steel has the majority of ferrite and bainite
(about 84%) and the minority of retained austenite (about 16%).
With the decrease of temperature the mean modulus of the TRIP

Fig. 7. Yield strength of the TRIP steel at different deformation temperatures.

304

X.D. Wang et al. / Materials Science and Engineering A 438–440 (2006) 300–305

Fig. 8. The stress–strain curves of the TRIP steel.

steel increases, which results in the increase of yield strength
with decreasing temperature. In the range of temperatures from
−5 to −20

◦

C, the onset of the martensitic transformation

due to tensile stress within the elastic range leads to the local
“soft” of modulus in retained austenite, then the drop of yield
strength, such effect continues below

−20

◦

C. However, it is

worthy to point out that below

−5

◦

C (

M

σ

s

), the yield strength

rises with decreasing temperature as the same tendency as in
the range of temperatures from 20 to

−5

◦

C, and this effect

is attributed to the predominant role of ferrite and bainite in
the mean modulus, which is different from single austenite
phase in Fe–20Ni–0.5C steel

[11]

. The TRIP steel studied in

this work exhibits a good thermodynamic stability and about
−5

◦

C of

M

σ

s

(somewhat lower than room temperature), and

thus this is favorable for the application of TRIP steels in
engineering.

The stress–strain relationship at different temperatures

was investigated.

Fig. 8

shows the stress–strain curves at 0

and

−15

◦

C. In the curve for

−15

◦

C, there is a yield-drop,

corresponding to the local “soft” of modulus in retained
austenite. From

Fig. 8

, it can be found that the work hardening

rate at

−15

◦

C is greater than that at 0

◦

C, which results in

the earlier generation of martensitic transformation due to the
decrease of stacking fault energy of austenite with the decrease
of temperature

[12]

. Moreover, the formation of matensite

under stress produces more effective “strain hardening” than
does the ordinary slip mechanism in fcc austenite

[11]

. From the

above analysis, it is suggested that martensitic transformation
in TRIP steel has “strain hardening” effect and “uniform
elongation” effect due to martensitic transformation causing the
redistribution of stress, as a consequence, TRIP steels exhibit
greater tensile strength and ductility than many commercial
steels currently used in the manufacture of automobiles. The
strain-induced martensite in

−15 and 0

◦

C deformation was

verified as twin-type one by TEM observation, as shown in

Fig. 9

. The martensitic transformation from retained austenite

can relax the accumulated stress in ferrite and bainite around the
austenite.

Fig. 9. TEM micrographs of twin-typed martensite of the deformed sample: (a)
bright field and (b) diffraction pattern [1 ¯3 1]

m

//[¯1 3 ¯1]

t

.

4. Conclusions

The microstructure of conventional TRIP steel with Si was

investigated by means of OM, SEM, AFM and TEM. The modi-
fied method presented in this work exhibits an effective approach
to distinguish the multi-phase microstructures. In addition, both
thermodynamic and mechanical stability of retained austenite
were evaluated by electrical resistance measurement and tensile
tests, respectively. Main conclusions can be drawn as follows.

(i) The volume fraction of various phases can be determined

by quantitative metallographic examination and XRD. The
studied TRIP steel contains 56.1% ferrite, 27.9% bainite
and 16% retained austenite while 9% of austenite trans-
forms to martensite during tensile stress.

(ii) AFM topographic analysis (height and roughness) makes

it possible to identify various phases. The statistical results
indicate that the increasing sequence of height of various
phases is ferrite, bainite, austenite and martensite.

(iii) The electrical resistance measurement indicates that when

the retained austenite is cooled to

−80

◦

C, martensitic

transformation does not occur, which shows that the
retained austenite has a good thermodynamic stability,
but it is metastable under stress. The

M

σ

s

temperature is

determined as about

−5

◦

C by yield stress–temperature

X.D. Wang et al. / Materials Science and Engineering A 438–440 (2006) 300–305

305

curve, moreover, the abnormal drop point (

M

σ

s

) in yield

stress–temperature curve and the different change tendency
in stress–strain curve at different temperatures are reason-
ably explained respectively based on the mean modulus of
TRIP steel and martensitic transformation in TRIP steel.

References

[1] V.F. Zackay, E.R. Parker, D. Fahr, R. Bush, Trans. ASM 60 (1967)

253.

[2] K. Sugimoto, M. Kobayashi, S. Hashimoto, Metall. Trans. 23A (1992)

3085.

[3] P. Jacques, E. Girault, P.H. Harlet, et al., ISIJ Int. 41 (2001) 1061.

[4] J.R. Patel, M. Cohen, Acta Metall. 1 (1953) 531.
[5] G.N. Haidemenopoulos, M. Grujicic, G.B. Olson, et al., Acta Metall. 37

(1989) 1677.

[6] L. Barbe, M.D. Meyer, B.C.D. Cooman, in: Proceeding of Interna-

tional Conference on TRIP-Aided High Strength Ferrous Alloys, B.C. De
Cooman (Ed.), Ghent, Belgium, 2001, p. 65.

[7] O. Matsumura, Y. Sakuma, H. Takech, Trans. ISIJ 27 (1987) 570.
[8] M.R. Berrahmoune, S. Berveiller, K. Inal, et al., Mater. Sci. Eng. A378

(2004) 304.

[9] E. Girault, P. Jacques, Ph. Harlet, et al., Mater. Charact. 40 (1998) 111.

[10] K.D. Amar, J.G. Speer, D.K. Matlock, Adv. Mater. Processes 161 (2003)

27.

[11] S.A. Kulin, M. Cohen, B.L. Averbach, Trans. AIME 193 (1952) 661.
[12] L. Remy, A. Pineau, Mater. Sci. Eng. 28 (1977) 99.