A. A. Saleh - Texture evolution of cold rolled and annealed Fe–24Mn–3Al–2Si–1Ni–0.06C TWIP steel.pdf

(3396 KB) Pobierz
Materials Science and Engineering A
528 (2011) 4537–4549
Contents lists available at
ScienceDirect
Materials Science and Engineering A
journal homepage:
www.elsevier.com/locate/msea
Texture evolution of cold rolled and annealed Fe–24Mn–3Al–2Si–1Ni–0.06C
TWIP steel
Ahmed A. Saleh
∗
, Elena V. Pereloma, Azdiar A. Gazder
School of Mechanical, Materials and Mechatronic Engineering, University of Wollongong, Wollongong, New South Wales 2522, Australia
a r t i c l e
i n f o
a b s t r a c t
The microstructure and texture evolution of 42% cold-rolled Fe–24Mn–3Al–2Si–1Ni–0.06C TWinning
Induced Plasticity (TWIP) steel is investigated during isochronal annealing at temperatures between
600 and 850
◦
C. In the cold rolled condition, bulk texture returned the distinctive -fibre for low stacking
fault energy materials, with higher intensities for Goss ({1 1 2} 0 1 1 ) compared to Brass ({1 1 1} 1 1 2 ).
A comparison between bulk and micro-textures, showed a significant slip contribution to the devel-
opment of the Brass orientation, along with a possible role for micro-shear banding. Annealing twins
contribute to recrystallisation from the early stages of nucleation and participate in generating new ori-
entations thereafter. Unlike texture studies on other austenitic steels, the F ({1 1 1} 0 1 1 ) and Rotated
Copper ({1 1 2} 0 1 1 ) orientations were detected in this work. The former is due to a more homoge-
neous distribution of nucleation sites, while the latter can be ascribed to second order twinning and
the preferred-growth 30
◦
1 1 1 relation with the Brass rolling component. Based on the microstructural
parameters from Electron Back-Scattering Diffraction (EBSD), the modified Hall–Petch (H–P) relation was
successfully applied to the 0.2% proof stress.
© 2011 Elsevier B.V. All rights reserved.
Article history:
Received 30 November 2010
Received in revised form 9 February 2011
Accepted 18 February 2011
Available online 25 February 2011
Keywords:
TWIP
Recrystallization
Electron Back-Scattering Diffraction (EBSD)
Texture
Hall–Petch
1. Introduction
In order to meet the environmental, economic and safety
demands of the automotive industry, a variety of Advanced High
Strength Steels (AHSS) has been developed such as Dual Phase
(DP), Complex Phase (CP), TRansformation Induced Plasticity (TRIP)
and most recently, TWinning Induced Plasticity (TWIP) steels
[1,2].
TWIP steels are face centred cubic (fcc) austenitic steels with low
stacking fault energy (SFE, 15–40 mJ/m
2
) which instigates twinning
along with dislocation glide during room temperature deformation.
It is well established that materials with high SFE develop
Copper (Cu) type textures, while low SFE leads to Brass (B) type
textures
[3].
The Cu texture is characterised by a strong -fibre
that extends from Cu ({1 1 2} 1 1 1 ) to B ({1 1 0} 1 1 2 ) through S
({1 2 3} 6 3 4 ). With reducing SFE the intensity of Cu decreases
and the B orientation intensifies with a spread towards Goss
(G) ({1 1 0} 0 0 1 ); thus forming the characteristic -fibre
[4].
An intermediate texture between these two extreme cases may
develop depending on the exact SFE value as well as the degree of
∗
Corresponding author at: University of Wollongong, School of Mechanical,
Materials and Mechatronic Engineering, Room No. 105, Bldg. 8, Northfields Av, Wol-
longong, New South Wales 2522, Australia. Tel.: +61 4221 5493;
fax: +61 4221 3662.
E-mail address:
as740@uowmail.edu.au
(A.A. Saleh).
0921-5093/$ – see front matter
© 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.msea.2011.02.055
rolling reduction. The transition from Cu to B-type textures is asso-
ciated with microstructure inhomogeneities in low SFE materials,
namely deformation twinning and shear banding
[5–7].
The latter
two phenomena are briefly presented here due to their relevance
in the development of the cold rolling and subsequent recrystalli-
sation textures.
The influence of twinning on texture evolution in low SFE mate-
rials has been contested ever since Wasserman’s twinning theory
ascribed the texture transition to a volume effect of deformation
twinning
[8].
Accordingly Hirsch et al.
[7]
suggested a microscopic
route for texture transition via the following steps: (i) the Cu
component (which is highly susceptible to twinning) twins to the
Copper Twin (CuT) ({5 5 2} 1 1 5 ), (ii) the CuT rotates to the inter-
mediate F ({1 1 1} 1 1 2 ) orientation, (iii) F subsequently rotates
to the close meta-stable G orientation via shear banding forming
a very fine subgrain structure and finally, (iv) the G orientation
rotates to the B position due to the reoccurrence of homogeneous
slip within subgrains.
On the other hand, Leffers and Ray
[9]
concluded that since the
volume fraction of twins in the bulk microstructure is low
[5,10],
twinning is most likely to influence bulk texture evolution via a
latent hardening effect. Consequently, a volume effect is only con-
sidered in very highly twinned materials. Latent hardening confines
matrix slip to planes parallel to the
{1
1 1} twinning plane in grains
with dense twin bundles. Following this, a composite deforma-
tion pattern at moderate (60–70%) rolling reductions was proposed
4538
A.A. Saleh et al. / Materials Science and Engineering A
528 (2011) 4537–4549
[11,12]
such that: (i) heavily twinned grains deform mainly by sin-
gle slip on the
{1
1 1} matrix planes, (ii) the sparsely twinned grains
may or may not deform by multi-slip, (iii) while the twin-free grains
must deform by multi-slip in order to maintain strain continu-
ity. With increasing rolling reduction, Duggan et al.
[5]
suggested
that latent hardening ultimately leads to the rotation of twin clus-
ters to align with the rolling plane (thus forming the -fibre with
{1
1 1}
u
v
w
orientations), and subsequently, to the formation of
shear bands which gradually consume the twins
[5,6].
Accordingly,
a microscopic route for the B texture evolution with a major role for
shear banding was proposed
[5,6].
However since the development
of shear bands is not always succeeded by the prior formation of
the -fibre
[9],
Leffers and Bilde-Sørensen
[11]
pointed out that
shear banding can be attributed to the difficulty in maintaining
homogeneous, composite deformation.
Most of the aforementioned texture investigations were con-
ducted on low SFE copper alloys, while fewer studies were
performed on austenitic steels. For 316L austenitic stainless steel
(SFE
∼64
mJ/m
2
), Donadille et al.
[13]
emphasised the role of
both, deformation twinning and shear banding in the develop-
ment of B-type texture whereas Chowdhury et al.
[14]
interpreted
such textures following Hirsch et al.
[7]
for
medium
SFE materi-
als. Vercammen et al.
[15]
viewed the formation of B texture in
Fe–30Mn–3Al–3Si TWIP steel (SFE
∼40
mJ/m
2
) as a consequence
of deformation twinning during the early stages of cold rolling
inhibiting the formation of Cu texture and consequently, generat-
ing the B component in agreement with Hirsch et al.
[7]
for
low
SFE materials. Bracke et al.
[16]
could not detect any contribu-
tion from shear banding and attributed texture development in
Fe–22Mn–C–N TWIP steel (SFE
∼15
mJ/m
2
) to both, deformation
twinning (via the volume effect
[7])
and slip.
To this end, recrystallisation textures in low SFE materials are
even more controversial than rolling textures where there is an
agreement on the
end-texture
but a debate as to its root causes.
The origin of recrystallisation textures is usually explained through
oriented nucleation or growth. In the oriented nucleation hypoth-
esis, grains with a particular orientation proliferate via nucleation
more frequently than grains belonging to other orientations. On the
other hand, oriented growth proposes faster growth of the domi-
nant texture than other orientations
[17].
Additionally, for low SFE
materials, recrystallisation twinning may: (i) lead to new orienta-
tions that did not exist in the original deformation texture and/or,
(ii) contribute to nucleation during the early stages of recrystalli-
sation
[18,19].
The SFE is closely related to annealing textures via its effect
on the microstructure and texture of the deformed state. Broadly
speaking, Cu-type rolling textures yield a Cube (C) recrystallisation
texture while B-type rolling textures result in the Recrystallised
Brass (BR) orientation
[20].
However, for the same material the
resulting texture may vary with cold-working strain level, anneal-
ing temperature and the influence of grain growth
[17].
Very few studies are available on the recrystallisation texture of
TWIP steel. For a TWIP chemistry of Fe–30Mn–3Al–3Si, a retained
-fibre was observed
[21].
This retention was ascribed to nuclei
growing inside deformed grains with orientations close to that of
the deformed matrix followed by annealing twinning that leads to
equivalent texture variants. In a 50% cold rolled Fe–Mn–C–N TWIP
alloy, the B rolling texture was retained via random nucleation and
was attributed to the homogeneity of the deformed microstructure
[16].
For Fe–22Mn–0.6C TWIP steel cold rolled to 40–70% reduction,
de las Cuevas et al.
[22]
verified similar B texture retention with
attenuated intensities as grain growth occurred.
The microstructure and mechanical properties of a cold-rolled
and isochronally annealed TWIP steel was characterised previ-
ously via transmission and scanning electron microscopy (TEM
and SEM) and uniaxial tensile testing
[23].
In the current work,
microstructure and texture evolution is investigated via X-ray
Diffraction (XRD) and Electron Back-Scattering Diffraction (EBSD).
The observed cold rolling (CR) bulk texture is correlated to the
microstructure and micro-texture. Finally, in order to evaluate the
strengthening contribution of the different grain boundaries to
the 0.2% proof stress obtained in
[23],
the modified Hall–Petch
(H–P) relation
[24,25]
has been implemented using EBSD
data.
2. Experimental procedure
The nominal composition of the present TWIP steel is
24Mn–3Al–2Si–1Ni–0.06C wt.%. The cast slab was homogenised
at 1100
◦
C for 2 h then 52% hot rolled (at 1100
◦
C) in four passes
with
∼17%
reduction per pass. Thereafter 42% cold rolling reduction
(equivalent to true strain
ε
= 0.55) was undertaken in 11 passes with
∼4.8%
reduction per pass producing a final thickness of 7.3 mm.
The employed roll diameter was 350 mm. This corresponds to a
roll gap geometry (l/h) ranging from 0.85 for the first pass to 1.1 for
the final pass. Here (l) refers to the projected length of the arc of
contact between the roll and strip
[26],
while (h) is the mean strip
thickness. The aforementioned roll gap geometry signifies
homoge-
neous
rolling as both values lie in the intermediate draught range
of 0.5
≤
l/h
≤
5
[27].
Apart from the roll gap geometry needed to
maintain
homogeneous
rolling conditions, the through thickness
texture of the rolled strip is also affected by friction, strain rate and
alloy content
[27].
In order to account for possible heterogeneity
effects, all subsequent investigations were undertaken on normal
direction (ND)–rolling direction (RD) section samples cut from the
middle of the strip cross-section. While micro-texture was esti-
mated from EBSD maps from the centre of the sample thickness,
bulk texture measurements were performed on the same sample
using a 5
×
5 mm
2
spot size to provide average crystallographic
information.
Isochronal annealing between 600 and 850
◦
C included 240 s of
heating to stable temperature followed by 300 s of soaking time and
immediate water quenching. Further processing details are given in
[23].
All microstructure and bulk texture analyses were conducted
on ND–RD sections after polishing up to colloidal silica. Vickers
microhardness with a 500 g load was used to estimate the fraction
softened with annealing temperature.
Three incomplete (0–85
◦
) pole figures
{1
1 1},
{2
0 0} and
{2
2 0}
were collected using a PANalytical Xpert–PRO MRD goniometer
equipped with a Cu tube at 45 kV and
∼40
mA. The orientation
distribution functions (ODFs) were calculated via series expansion
using ResMat. The -fibre intensities were determined in Matlab
with a 10
◦
deviation from the ideal skeleton line.
EBSD was undertaken on a JEOL–JSM7001F field emission gun
(FEG) – scanning electron microscope (SEM), fitted with a Nordlys-
II(S) camera and the Corona Fast Acquisition (FA) software, at 15 kV,
∼5
nA and 15 mm working distance. Step sizes of 0.05 m and
0.4 m were used for the CR sample in order to acquire selected and
large area statistics, respectively. A step size of 0.125 m was used
for the 600
◦
C condition, while a step size of 0.2 m was maintained
constant for all other samples.
In all maps, a minimum of 3 pixels was used to identify sub-
grain/grain structures and misorientations ( ) less than 2
◦
were
disregarded. 2
◦
≤
< 15
◦
are defined as low angle grain boundaries
(LAGBs). The total high angle grain boundaries (THAGBs) comprise
HAGBs (15
◦
≤ ≤
57.5
◦
) and twin boundaries (TBs). First order TBs
are defined as 3 = 60
◦
1 1 1 while second order TBs are 9 = 38.9
◦
1 0 1 . The maximum tolerance of the misorientation angle (
)
from the exact axis-angle relationship was identified following the
Palumbo–Aust criterion (i.e.
−
≤
15
◦ −5/6
)
[28],
yielding toler-
◦
for 3 and 2.4
◦
for 9, respectively. Micro-textures
ance limits of 6
A.A. Saleh et al. / Materials Science and Engineering A
528 (2011) 4537–4549
4539
are depicted after exporting from HKL Channel-5 to MTex
[29]
and
ResMat.
3. Results
3.1. Microstructure characterisation
A representative grain boundary map after CR is shown in
Fig. 1(a).
The cold rolled microstructure is characterised by elon-
gated grains where plastic deformation is accommodated via
various deformation modes. Similar to austenitic stainless steel
(∼20Cr–26Ni)
[30],
untwined areas were detected. Deformation
twins were observed as either: (i) one set of parallel twins (with
some of them rotated towards the RD) (Fig.
1(b,
c)) or as, (ii) inter-
secting primary and secondary twins
[23].
Moreover, microscopic
shear bands (MSBs) were seen in twinned grains indicating that the
imposed strain was not accommodated solely and homogeneously
by slip and twinning. Some of the observed MSBs are inclined at
∼±30
◦
while others are nearly parallel to the RD in accordance with
the sequence of their formation and rotation towards the rolling
plane during CR. Intersecting primary and secondary MSBs are illus-
trated in
Fig. 1(b)
by black and white arrows, respectively. Some of
the detected MSBs were also found to contain deformation twins
as illustrated by the dashed black arrow in
Fig. 1(b).
Following microhardness measurements on the partially recrys-
tallised samples, the softened fractions were estimated as
∼16,
68, 83, 93 and 96% after annealing at 600, 700, 750, 775
and 850
◦
C, respectively. Thereafter, microstructure evolution is
charted as a function of the annealing temperature in
Fig. 2.
In
[23],
recrystallisation was not detected at 600
◦
C using sec-
ondary electron micrographs or in the local areas observed by
TEM. The
∼16%
softening from microhardness measurements was
therefore only associated with recovery. Nonetheless nucleation
events, such as
Fig. 2(a),
bottom inset or in the form of twins
bulging at the boundaries between two grains (Fig.
2(a),
top
inset), were detected via EBSD and represent
∼2%
of the map
area.
Despite the fast progress of recrystallisation at 700
◦
C (∼68%
softened), a greater proliferation of similar new nucleation sites
characterised by twin bulges at grain boundaries were found. In
addition, nucleation events where grain 1 has an orientation close
to the deformed matrix and a 3 relation with grain 2 which in
turn, possesses a
∼
9 boundary with the matrix, were observed in
Fig. 2(b),
top inset.
Recrystallisation proceeds even further with annealing up to
850
◦
C, such that the evolution of a high density of annealing
twins becomes very apparent in the growing, equiaxed grains
(Fig.
2(c–e)).
The change in the misorientation distribution with the anneal-
ing temperature is shown in
Fig. 3(a)
with representative
misorientation axis distributions in the crystal coordinate system
given for the two detected peaks. The intense peak at 60
◦
is asso-
ciated with 1 1 1 and represents 3 or first order TBs, while the
smaller peak at
∼39
◦
is clustered around 1 0 1 due to 9 or second
order TBs.
The aforementioned twin bulges can explain the slight increase
in 3 area fraction from CR to 600
◦
C annealing (Fig.
3(b)).
With fur-
ther annealing and in conjunction with the formation of annealing
twins (Fig.
2(b–e)),
the area fractions of HAGBs and 3 continue to
increase and is associated with the evolution of 9 boundaries and
a decrease in LAGBs. The relatively lower fraction of 9 compared
to 3 and its decreasing area and length fractions upon the com-
pletion of recrystallisation at 850
◦
C should also be noted (Fig.
3(b,
c)).
The overall refining effect of annealing twins on the microstruc-
ture can be seen in
Fig. 3(d),
as they obstruct dislocation glide
during subsequent deformation similar to random grain bound-
aries
[31,32].
A slight increase in recrystallised grain size from
∼3.6
m at 700
◦
C to
∼4.9
m was detected after annealing at
750
◦
C. Thereafter, the average grain size increases significantly to
∼8
m at 850
◦
C due to grain coarsening.
3.2. Texture evolution
Fig. 1.
(a) EBSD grain boundaries map for the CR sample, (b and c) band contrast
maps for two selected twinned areas. LAGBs = grey, HAGBs = black, 60
◦
1 1 1 , 3
TBs = red and rolling direction (RD) = horizontal. In (b) the black and white arrows
indicate intersecting primary and secondary micro-shear bands (MSBs), respec-
tively, while the dashed black arrow denotes a twin within MSB. (For interpretation
of the references to colour in this figure legend, the reader is referred to the web
version of the article.)
3.2.1. After cold-rolling
The ideal orientations of texture components in fcc materials are
shown schematically for the
2
= 0
◦
, 45
◦
and 65
◦
ODF sections in
Fig. 4
and
Table 1.
Although the characteristic -fibre of the B-type
texture has developed after CR (Fig.
5),
the most notable feature is
4540
A.A. Saleh et al. / Materials Science and Engineering A
528 (2011) 4537–4549
Fig. 2.
Representative EBSD grain boundaries maps for the (a) 600
◦
C recovered, (b) 700
◦
C, (c) 750
◦
C, (d) 775
◦
C and, (e) 850
◦
C partially recrystallised samples. In (a), the top
and bottom insets are zoomed-in views of nucleation sites. The top inset in (b) is a zoomed in-view for the recrystallised grain indicated by the dashed arrow. 38.9
◦
1 0 1 ,
9 TBs = blue lines and RD = horizontal. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of the article.)
the almost double intensity of G (f(g) = 5.3) compared to B. The rela-
tively weak texture recorded here is in agreement with Bracke et al.
[16]
where a maximum
f(g)
∼
5 was reported at 40% CR reduction.
The individual contributions of slip and twinning (Fig.
6(a–c))
to
overall texture development are delineated by the micro-texture
of the CR map (Fig.
1(a)).
Twinned grains possess lower internal
misorientations than other grains where slip is more dominant. By
applying an internal misorientation threshold of 1
◦
, the CR map
was deconstructed into: (i) a
slip
fraction where deformation was
accommodated mostly via slip (Fig.
6(b))
and, (ii) a
twin
&
slip
frac-
tion where twinning occurs along with slip (Fig.
6(c)).
The adequacy
of deconstruction is verified by the near absence of twin bound-
aries from the misorientation distribution of the
slip
fraction given
in
Fig. 6(d),
while the intensity variation along the -fibre is shown
in
Fig. 6(e).
Here two points require clarification. Firstly, the division into
slip
and
twin
&
slip
fractions does not mean that dislocation glide
is absent in the twinned areas. Rather, it implies the absence
or scarcity of twinning from the
slip
fraction. Secondly, due to
EBSD resolution limitations, it is possible that fine micro-twins
were missed in the
slip
fraction. In such grains the absence of
detectable twin clusters indicates that twinning was not prominent
and consequently, allows for multi-slip to proceed
[11,33].
Hence
the following discussion focuses on the role of slip as it is the only
clearly separable mechanism. Distinct conclusions regarding twin-
ning are not possible as twined areas are also associated with slip
contributions; as evidenced by the high LAGBs fraction in the
twin
&
slip
subset (Fig.
6(d)).
Lastly, the magnitude of
f(g)
in
Fig. 6(a–c)
should not be compared with
Fig. 5
as the former consists of single
orientations.
A.A. Saleh et al. / Materials Science and Engineering A
528 (2011) 4537–4549
4541
Fig. 3.
(a) Change in misorientation distribution as a function of annealing temperature with representative misorientation axis distributions in the crystal coordinate system
for 3 and 9 angular ranges, (b) grain boundary area fraction as a function of annealing temperature, and (c) evolution of length fraction of 3 and 9 twin boundaries
with recrystallisation. 3 = 60
◦
1 1 1 with 6
◦
deviation, 9 = 38.9
◦
1 0 1 with 2.4
◦
deviation
[28]
and, (d) change in recrystallised grain size with and without considering
TBs with annealing temperature.
Fig. 4.
Schematic representation of the important texture components in fcc materials. The dashed lines depict twin relationships, while the dotted line represents 30
◦
1 1 1
misorientation.

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika:

Zgłoś jeśli naruszono regulamin