Z. Y.Tang - Deformation twinning and martensitic trans formation and dynamic mechanical properties in Fe–0.07C–23Mn–3.1Si–2.8Al TRIP TWIP steel.pdf

(1880 KB) Pobierz
Materials Science & Engineering A 624 (2015) 186–192
Contents lists available at
ScienceDirect
Materials Science & Engineering A
journal homepage:
www.elsevier.com/locate/msea
Deformation twinning and martensitic transformation and dynamic
mechanical properties in Fe–0.07C–23Mn–3.1Si–2.8Al TRIP/TWIP steel
Z.Y. Tang
a,
n
, R.D.K. Misra
b
, M. Ma
a
, N. Zan
a
, Z.Q. Wu
a
, H. Ding
a
a
b
School of Materials and Metallurgy, Northeastern University, Shenyang 110819, China
Department of Metallurgical and Materials Engineering and Center for Structural and Functional Materials Research and Innovation,
University of Texas at El Paso, 500W, University Avenue, El Paso, TX 79968, USA
art ic l e i nf o
Article history:
Received 20 September 2014
Received in revised form
19 November 2014
Accepted 24 November 2014
Available online 2 December 2014
Keywords:
Microstructure
Dynamic mechanical properties
Martensite transformation
Deformation twin
High strain rate
a b s t r a c t
In the study described here, we have explored the evolution of microstructure and mechanical
properties in Fe–0.07C–23Mn–3.1Si–2.8Al steel with a stacking fault energy (SFE) in the intermediate
range of 15–20 mJ m
À
2
during dynamic deformation and in the strain rate range of 10
1
–10
3
s
À
1
. The
results showed that the transformation induced plasticity (TRIP) effect and twinning induced plasticity
(TWIP) effect coexist during dynamic deformation. The mode of austenite-to-martensite transformation
is
γ
-
ε
,
ε
-
α
0
. With increase in the strain rate, the volume fraction intermediate
ε
-martensite was
increased and
α
0
-martensite remained nearly constant, and the frequency of the intersecting deforma-
tion twins was also increased. This behavior of steel was responsible for a good combination of ultimate
tensile strength of 913 MPa and total elongation of 75.4% at a strain rate of 10
3
s
À
1
. The strength and
elongation increased significantly with an increase in strain rate in the range of 10
1
–10
3
s
À
1
.
The dominant plasticity enhancing mechanisms with increase in strain rate were strain-induced
intermediate
ε
-martensite and intersecting deformation twins.
&
2014 Elsevier B.V. All rights reserved.
1. Introduction
High-manganese transformation/twinning-induced plasticity
steels are potential candidates for automotive steels because they
can reduce the weight of automotive components, improve fuel
efficiency, reduce emission, and enhance the degree of safety
[1–5].
In these steels, the enhancement of strength–high ductility
combination, excellent formability, and fracture toughness is
attributed to the activation of additional deformation mechanisms,
which besides dislocation slip are deformation-induced martensi-
tic transformation and deformation twinning because of low
stacking fault energy of the material
[6–11].
Automotive applica-
tions require steels that exhibit reliable properties even at high
strain rates
[12–14].
In dynamic conditions, the strain rate and
quasi-adiabatic effects play an important role during straining
[15–18].
Research on dynamic deformation behavior is important
for promoting stamping technology and for improving the impact
resistance of materials. Additionally, in order to appropriately
select low carbon and high manganese TRIP/TWIP symbiotic effect
steels and to design an appropriate processing schedule, it is
necessary to understand the high-speed deformation behavior of
these steels. In the study described here, we elucidate the relation-
ship between microstructure and mechanical properties of
Fe–0.07C–23Mn–3.1Si–2.8Al steel in the strain rate range of 10
1
–
10
3
s
À
1
. Moreover, we define the sequence and nature of the
dynamic micro-mechanism which occur at high strain rates. The
study is expected to provide a theoretical basis for the application
of a new generation of light weight and high impact resistance
advanced high strength automotive steels.
2. Experimental
The Fe–0.07C–23Mn–3.1Si–2.8Al alloy (wt%) was prepared by
melting in an induction furnace. It was subsequently hot forged
and hot rolled in the temperature range of 1150–900
1C,
to plates
of
$
3.0 mm thickness, followed by annealing at 1000
1C
for
30 min and cold-rolling to sheets of
$
1.0 mm thickness. Next,
the cold rolled sheets were solution treated at 1000
1C
for 10 min
and water-quenched. Tensile specimens (longitudinal direction)
with a gage length of 20 mm and width of 10 mm were cut by a
wire-cutting machine, and high strain rate tensile tests were
conducted using Zwick HTM 5020 high-speed tensile testing
machine.
Optical microscopy (OM), scanning electron microscopy (SEM),
X-ray diffraction (XRD), and transmission electron microscopy
n
Corresponding author. Fax:
þ86
2423906316.
E-mail address:
tangzy@smm.neu.edu.cn
(Z.Y. Tang).
http://dx.doi.org/10.1016/j.msea.2014.11.078
0921-5093/& 2014 Elsevier B.V. All rights reserved.
Z.Y. Tang et al. / Materials Science
&
Engineering A 624 (2015) 186–192
187
(TEM) specimens were cut from the original specimens and also
from the area close to the fracture surface after the tensile tests.
Specimens were mechanically polished and subsequently electro-
polished at room temperature in an electrolyte containing 95%
CH
3
COOHþ 5% HClO
4
solution to remove any strain-induced
martensite that may have formed during the mechanical polishing.
Phase analysis was carried out using D/Max–Ra XRD. The anneal-
ing and deformation twins and martensite were identified using a
TECNAI G2-20 TEM.
Table 1
Volume fraction of constituent phases estimated by X-ray diffraction (vol%).
Strain rate/s
À
1
0
10
1
10
2
10
3
γ
83.3
67.4
63.7
60.9
α
F
12.6
12.6
12.6
12.6
ε
4.1
6.7
9.9
12.4
α
0
0
13.3
13.8
14.1
3. Results and discussion
3.1. Strain-induced martensite transformation and deformation
twinning
Fig. 1
shows the XRD spectrum of samples prior to and after
deformation at different strain rates. After deformation, the peaks
of 111
γ
, 200
γ
, 220
γ
and 311
γ
for the
γ
-fcc phase were weakened,
while the diffraction peaks of 110
α
, 200
α
and 211
α
for the
α
-bcc
phase and 002
ε
for the
ε
-hcp phase were enhanced. The change in
the phase fraction with strain rate was calculated by XRD and is
presented in
Table 1.
The data imply that the amounts of
ε
-martensite and
α
0
-martensite are increased after deformation,
accompanying the decrease in volume fraction of austenite, which
suggests that the transformation of
γ
-
α
0
or
γ
-
ε
,
ε
-
α
0
occured
during tensile straining. It may be noted that while the trans-
formed volume fraction of austenite is increased with the increase
of strain rate in the range of 10
1
–10
3
s
À
1
, the volume fraction of
ε
-martensite is increased and
α
0
-martensite remains nearly constant.
Fig. 2
shows SEM micrographs of samples prior to and after
high strain rate tensile deformation. As shown in
Fig. 2a,
the
solution treated sample at 1000
1C
for 10 min followed by water
quenching (before deformation) indicated a few annealing twins
in austenite, the annealing twins ran through the entire austenite
grains and the twinning plane of the face-centered cubic crystal
was {111}. The twins within the austenite phase had a certain
orientation, the orientation angle between them was 601 or 1201,
and there was only one set of annealing twins within each
austenitic grain. Deformation twins usually form as thin plates
embedded in the matrix. Some of the thin plate microstructural
Fig. 1.
The X ray diffraction pattern of samples before and after tensile deformation
at different strain rates.
features are shown with the help of a black circle in
Fig. 2b,
which
can be either martensite or deformation twins, and their precise
identification requires the use of diffraction techniques. Based on
TEM studies, the thin plate features observed were confirmed to
be of deformation twins. As shown in the SEM micrographs
(Fig.
2b–d),
there was an abundance of deformation twins
after deformation at high strain rates. The results show that
deformation-induced twins with identical orientation started from
the grain boundary of the austenitic matrix, but did not reach the
other side of the grain. The microstructure of the deformed
specimen mainly exhibited individually oriented deformation
twins at a strain rate of 10
1
s
À
1
(Fig.
2b),
and intersection of
several oriented deformation twins at strain rates of 10
2
and
10
3
s
À
1
(Fig.
2c
and d). These observations suggest that with an
increase in strain rate, multiple twin systems were nucleated with
different orientations, and the grains were divided and refined.
Generally, deformation twins aligned in one direction sub-divide
the original coarse grains into submicron twin-matrix lamellae,
while deformation twin intersections divide two sets of twin-
matrix lamellae into rhombic blocks with a high number of
misorientations, and the sub-divided blocks evolve into randomly
refined grains. In the present work and as shown in
Fig. 2c–f,
with
an increase in strain rate, the multiplication of deformation twins
introduced more intersections, and the grains were refined.
Fig. 3a–c
shows bright-field, dark-field TEM images of
deformation-induced formation of
α
0
-martensite and correspond-
ing to selected area diffraction pattern. The presence of
α
0
-martensite at the intersection of two
ε
-martensite and within
a single
ε
-martensite can be seen. In addition, the formation of
α
0
-martensite directly from austenite without the association of
ε
-martensite was not observed by OM, SEM and TEM. According to
the lattice match relationship of the plane and crystal orientation
between
γ
-fcc,
ε
-hcp and
α
-bcc phases, the lattice match relation-
ship between
γ
-fcc and
ε
-hcp phases is the best, between
ε
-hcp
and
α
-bcc phases is relatively lower, and between
γ
-fcc and
α
-bcc
phases is the worst
[19,20].
Thus, the transformation of
γ
-
ε
,
ε
-
α
0
is more likely to occur relative to the transformation of
γ
-
α
0
.
Fig. 4a
and b shows the TEM micrograph and the corresponding
selected area diffraction pattern (SADP). It can be seen that there
are abundant stacking faults and annealing twins with a straight
boundary. The presence of abundant stacking faults can provide
favorable conditions for the subsequent strain-induced nucleation of
martensite and deformation twins.
Fig. 4c
and d shows the TEM
micrographs of samples after high-strain rate tensile deformation. It
can be seen from
Fig. 4c
that the deformation twins primarily
exhibited single orientation at a strain rate of 10
1
s
À
1
. But when the
strain rate was 10
3
s
À
1
, the intersecting twins appeared within the
austenitic grains (Fig.
4d).
Thus, from these results it can be concluded
that there is a significant effect of strain rate on the morphology of
deformation twins such that with an increase in strain rate, the
twinning was more complex, and the twinning system changed from
single orientation to multiple orientations, consistent with the SEM
results. Generally, twins are nucleated at levels of high stress when
reaching a critical shear stress, followed by the formation of deforma-
tion twins. Under conditions of high strain rate, deformation twins
were nucleated in the interior of grains with a favorable orientation.
188
Z.Y. Tang et al. / Materials Science
&
Engineering A 624 (2015) 186–192
_
_
_
Fig. 2.
SEM micrographs of samples prior to and after deformation with different strain rates. (a) Prior to deformation; (b)
ε
¼
10
1
s
À
1
; (c)
ε
¼10
2
s
À
1
; (d)
ε
¼
10
3
s
À
1
;
(e) partial magnification of (c); and (f) partial magnification of (d).
But, because of the short duration of tensile deformation, the twins
were not fully developed and, thus, the twins did not run through the
entire grain. But with an increase in strain rate, the mutual intersection
of the deformation twins occurred because of the interaction of the
subsequent and previously nucleated deformation twins
[21].
It is well known that the strain-induced martensitic transfor-
mation and twinning in high-Mn steel depends on the SFE. In the
case of dynamic loading, a temperature rise due to adiabatic
heating is enough to affect the SFE. The adiabatic temperature
rise is generally calculated using the following equation
[17]
Z
ε
2
Δ
Q
β
¼
Δ
T
¼
σ
d
ε
ð1Þ
ρ
Cp
ρ
Cp
ε
1
where
Δ
Q
is the fraction of mechanical energy that is converted to
heat energy,
ρ
is the density of steel,
Cp
is the typical heat capacity
of steel,
β
is the coefficient of mechanical energy converted to
thermal energy,
σ
is the true stress, and
ε
is the true strain. For
the experimental steel,
β
is assumed to be 0.9,
ρ
¼7.8
g/cm
3
,
Cp¼0.46
kJ(kg K)
À
1
, and the total mechanical energy generated
under the actual test conditions by integrating the area under
σ
–
ε
curve from
Fig. 5.
The room temperature SFE value was
14.5 mJ m
À
2
for the experimental steel without deformation. The
rise in adiabatic temperature (
Δ
T)
as a function of true strain was
calculated using Eq.
(1),
and in a similar manner the increase in
the SFE caused by
Δ
T
was also calculated according to the relation
of SFE vs. temperature
[17].
The change in SFE as a function of true
strain caused by an increase in temperature is presented in
Fig. 6.
It can be seen that the SFE
r15
mJ m
À
2
when the strain is less
than 0.1, and the SFE increases to 20 mJ m
À
2
when the deforma-
tion level comes to 0.4. The transformation of austenite into
martensite is favored when the SFE is
r15
mJ m
À
2
, deformation
twinning occurs when SFE is
4
20 mJ m
À
2
, and strain-induced
Z.Y. Tang et al. / Materials Science
&
Engineering A 624 (2015) 186–192
189
Fig. 3.
Bright-field and dark-field TEM images of deformed samples at strain rate of 10
3
s
À
1
showing the formation of
α
0
-martensite at the intersection of two
ε-martensite
and within a single
ε-martensite.
(a) Bright-field, (b) dark-field, and (c) corresponding to SADP.
Fig. 4.
Bright-field TEM images of samples prior to and after deformation with different strain rates. (a) and (b) Stacking faults, annealing twins and corresponding to
selected area diffraction pattern (SADP) prior to deformation, (c) one set of deformation twins at 10
1
s
À
1
, and (d) intersection of several sets of deformation twins and the
formation of
α
0
-martensite at 10
3
s
À
1
.
190
Z.Y. Tang et al. / Materials Science
&
Engineering A 624 (2015) 186–192
Fig. 5.
The true strain–stress curves of samples at different strain rates.
increase in strain rate, there is increase in dislocation density,
because of which the interaction between these defects is
increased, the resolved shear stress is higher, stress concentration
at these defects is easier, and lower stress can promote the
nucleation of deformation twins with a favorable orientation.
Simultaneously, the frequency of mutual intersection of deforma-
tion twins increases as shown in
Figs. 2
and
4,
and nuecleation
sites for
ε
-martenstie increase, so the volume fraction of
ε
-martensite increases. The
α
0
-martensite is formed at the inter-
section of two
ε
-martensite and within a single
ε
-martensite as
shown in
Fig. 3,
there is obvious sequence of the occurrence
between the strain-induced
ε
-martensite and
α
0
-martensite, the
transformed temperature of
ε
-martensite is higher than that of
α
0
-martensite, and the required SFE for the presence of
ε
-
martensite is also higher
[6,24,25];
the intersection of
ε
-
martensite is recognized to be a favorable site for
α
0
-martensite
nucleation. However,
α
0
-martensite does not nucleate throughout
the entire intersection but only in specific regions, the
ε
-
martensite in (110)-oriented
γ
can transform into
α
0
-martensite,
and thicker
ε
-martensite can induce
α
0
-martensite due to a
smaller strain accumulation caused by the harder
α
0
-martensite
during transformation
[13,24,25].
Thus, with an increase in strain
rate, more deformation twin boundaries can enhance the forma-
tion of intermediate
ε
-martensite, but only the
ε
-martensite in
specific regions can induce
α
0
-martensite, so
α
0
-martensite
remains nearly constant.
3.2. Mechanical properties
Fig. 8
shows the engineering strain–stress plots of samples at
strain rates of 10
1
, 10
2
and 10
3
s
À
1
. It can be seen that the sample
yields at a very small strain, and stress increases with an increase
in strain during plastic deformation.
Table 2
shows the mechanical
properties of experimental steel at different strain rates. It can be
seen that the experimental steel exhibits excellent mechanical
properties in the strain rate range of 10
1
–10
3
s
À
1
. When the strain
rate was 10
3
s
À
1
, the experimental steel exhibited ultimate tensile
strength (UTS) of 913 MPa, the total elongation (TEL) of 75.4% and
PSE of 68.8 GPa%. For the experimental steel, the strain-induced
martensitic transformations and deformation twinning have both
taken place, strain-induced martenstitc transformation will pre-
ferentially take place in locally formed necking, due to intensive
strain hardening occurring in these areas. The deformation pro-
ceeds alternatively in neighboring areas of the tensile sample.
Twinning is highly effective in terms of enhancing the strain
hardening, stemming from twin boundaries acting as strong
barriers to dislocation motion. The interactions between TRIP
and TWIP effect may contribute greatly to ductility
[26].
It may be noted that the strength of experimental steel
increases with an increase in strain rate and the total elongation
also increases significantly with an increase in the strain rate. It
has been suggested
[27–31]
that the
ε
-
α
0
transformation can
reduce internal stresses generated locally by the blockage of
plastic
flow
in austenite by pre-existing
ε
-martensite. Though
the volume fraction of
α
0
-martensite remains nearly constant,
ε
-martensite fraction is increased with the increase of strain rate
in the range of 10
1
–10
3
s
À
1
. There are different results for the
effect of
ε
-martensite on the properties. Koyama et al.
[32]
evaluated that the amount of contribution to the work hardening
rate for the Fe–17Mn–0.6/0.8C steel can be arranged in the
following order:
ε
-martensite transformation
4deformation
twin-
ning4dynamic strain aging. Yang et al.
[33]
also found that the
Fe–22Mn–0.2C steel with formations of pre-existing
ε
-martensite
had a slightly lower ductility compared with Fe–22Mn–0.4C and
Fe–22Mn–0.6C steel without pre-existing
ε
-martensite. However,
Datta et al.
[34]
proposed that strain-induced
ε
-martensite is a
Fig. 6.
Calculated evolution of the SFE with true strain for Fe–0.07C–23Mn–3.1Si–
2.8Al steel.
martersitic transformation and deformation twinning coexist
when SFE is in the intermediate range from 15 to 20 mJ m
À
2
[6,17,22].
Thus, during the early stages of high strain rate defor-
mation, the metastable austenite transformed to
ε
-martensite, and
α
0
-martensite was formed within the
ε
-martensite plate. With an
increase in strain, the increase in SFE caused by adiabatic tem-
perature rise suppresses the transformation of austenite into
martensite at high levels of stress; however, the direct transforma-
tion of
γ
-
α
0
was observed at high stress levels
[8].
Thus,
combining the lattice match relationship of plane and crystal
orientations between
γ
-fcc,
ε
-hcp and
α
-bcc phases, Fe–0.07C–
23Mn–3.1Si–2.8Al steel shows the deformation induced transfor-
mation of
γ
-
ε
, and
ε
-
α
0
.
The effect of high strain rate on the SFE can also be observed in
Fig. 6.
The change of SFE with an increase in strain rate is slight as
the tensile sample is strained to 0.4. During the stage of strain
ranging from 0.1 to 0.4, the SFE is in the range of 15 to 20 mJ m
À
2
,
and strain-induced martersitic transformation and deformation
twinning coexist.
Fig. 7
shows TEM images and corresponding
SADP of deformed samples at a strain rate of 10
3
s
À
1
; it can be
seen that
ε
0
-martensite is formed at the intersection and boundary
of the deformation twins. Stacking faults and annealing twin and
deformation twin boundaries are important for the formation of
ε
-martensite in Fe–Mn–Si–Al steels because they can act as
nucleation sites for
ε
-martenstie. With an increase in strain rate,
the critical stress for deformation twinning decreases
[23].
Defor-
mation twin nuclei must be linked to dislocation density, stacking
faults, and interaction among annealing twin-dislocation, stacking
fault-dislocation and annealing twin-stacking fault. With an

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika:

Zgłoś jeśli naruszono regulamin