1 Article

2 Misorientation-dependent twinning induced

3 hardening and texture evolution of TWIP steel sheet
4 in plastic deformation process
5 Ning Guo 1, Chaoyang Sun 1,*, Mingwang Fu 2 and Mingchuan Han 1
6 1 School of Mechanical Engineering, University of Science and Technology Beijing, Beijing 100083, China;
7 guoning19891227@163.com; suncy@ustb.edu.cn; 380376983@qq.com
8 2 Department of Mechanical Engineering, The Hong Kong Polytechnic University, Hung Hom, Kowloon,

9 Hong Kong; mmmwfu@polyu.edu.hk

10 * Correspondence: suncy@ustb.edu.cn; Tel.: +86-010-62333695

11 Academic Editor: name

12 Received: date; Accepted: date; Published: date

13 Abstract: The quantitative contribution of twinning to hardening behavior and its effect on crystal
14 orientation need to be explored in greater depth for design and forming of
15 twinning-induced-plasticity (TWIP) steel products. To address this issue, the characteristics of
16 twinning formation in the plastic deformation of Fe-30Mn-3Si-2Al TWIP steel are investigated in
17 terms of intergranular misorientation distribution using electron back-scattering diffraction (EBSD),
18 which reveals that most deformation twins are adhered to the high-angle grain boundaries
19 (HAGBs) of the face-center-cube (FCC) type TWIP steel. Texture measurements are conducted to
20 show a stable volume fraction of major components including Goss, S and A orientations, while
21 Copper shifts towards Brass orientation. A crystal plasticity finite element (CPFE) model based on
22 virtual polycrystalline microstructure adopting representative volume element (RVE) is employed
23 to simulate the deformation to reveal the correlation between misorientation-dependent twinning
24 and hardening behavior of TWIP steel. The results demonstrate that the proportion of twinning
25 hardening to overall hardening is larger than slip hardening. The stability of texture evolution is
26 simulated to predict the anisotropy of TWIP steel. This research substantiates the twinning induced
27 hardening and texture evolution play in deformation of TWIP steel and thus is essential for
28 accurate prediction of the mechanical behaviors.

29 Keywords: TWIP steel; Crystal plasticity; Twinning; Misorientation; Texture

30

31 1. Introduction
32 In automotive industries, development of new steels with desirable strength, ductility and
33 toughness has been a hot research topic. Recently, low stacking fault energy (SFE) austenitic high
34 Mn steels, which is regarded as a very attractive alloying element, has attracted a great deal of
35 attention driven by the weight reduction and energy and materials saving in this industrial cluster
36 [1-5]. Particularly, the application of a new generation of high Mn steels with Al and Si additions,
37 called twinning induced plasticity (TWIP) steel, has been confirmed to efficiently realize weight
38 reduction and energy saving due to their good mechanical properties to facilitate the down size of
39 part dimension and size [3, 6, 7]. Greatly influenced by the content of chemical composition
40 including Mn, Al and Si, various deformation mechanisms called TWIP and phase transformation
41 induced plasticity (TRIP) effect related to the evolution of microstructure and crystallographic
42 texture components should be considered as an important issue to be explored and addressed in the
43 deformation process [3, 8].

Metals 2017, 7, x; doi: FOR PEER REVIEW www.mdpi.com/journal/metals

 Metals 2017, 7, x FOR PEER REVIEW 2 of 19

44 In terms of low-to-moderate SFE, Grässel et al. [9] observed that the extensive deformation
45 twins play a dominant role for the austenitic high Mn steels alloyed with Si and Al. The TWIP effect
46 significantly influences the hardening behavior at macroscopic level through the interaction of
47 dislocations gliding hindered by twinning boundaries when the Mn content is more than 25 w.t.%;
48 the Al content is about 3 w.t.%; the Si content is between 2 and 3 w.t.%; and the C content is low. On
49 the other hand, the deformation twins, which are regarded as an important lattice defect of TWIP
50 steels, strongly affects the microscopic evolution especially for the crystallographic texture
51 development in plastic deformation [10]. From the perspective of deformation mechanism, a
52 remarkable characteristic associated with the deformation twins is the sudden reorientation of
53 crystallites, leading to the preferred crystallographic orientation or texture component.
54 Subsequently, the evolution of crystallographic texture contributes to strengthening generally via
55 slip activities, known as texture hardening, to enhance the mechanical properties such as high
56 strength and good ductility [11]. Consequently, it is necessary to investigate the texture components
57 and deformation twins’ nucleation position first in the large plastic deformation process to establish
58 a close connection among deformation twins, crystallographic texture and mechanical behavior.
59 However, this relationship has not been explored in greater depth and needs an in-depth
60 investigation to correlate the microstructural and textural evolution with mechanical properties,
61 particularly for the Fe-Mn-Si-Al TWIP steels.
62 Since the anisotropy associated crystallographic texture is important in plastic deformation
63 processing [5, 12], it is indispensable to know more about crystallographic texture evolution and its
64 influence upon microstructure formation and evolution. Currently, most researches on TWIP steels
65 are focused on the relationship between microstructural evolution such as deformation twins’
66 nucleation rate and strain hardening of Fe-Mn-C and Fe-Mn-Si-Al steels [13, 14]. It is well
67 established that the strain hardening contributed by deformation twinning has generally been
68 considered as microstructural refinement called the dynamic Hall-Petch effect. Similar to grain
69 boundaries, twinning boundaries subdivide microstructure and act as obstacles to dislocation
70 motion. However, less attention has been paid to the formation of crystallographic texture and
71 quantitative analysis of twinning induced hardening in the course of deformation. The investigation
72 on the influence of crystallographic texture on deformation twins and hardening behavior in
73 deformation process is critical to be explored. Earlier experimental investigations detailing the
74 uniaxial tensile loading of fine-grained Fe-Mn-C TWIP steel have revealed that the development of
75 the pronounced <111> fiber in the tensile direction facilitates the formation of deformation twins and
76 maintains the strain hardening rate at a high level [15]. Although the misorientation profiles of
77 twinning boundaries have been described, the intragranular misorientation associated with
78 deformation uniformity and the deformation twins’ nucleation position, however, have not yet been
79 given in detail. For the Fe-Mn-Al-C TWIP steels, Souza et al. suggested that the formation of
80 copper-type texture during cold rolling may be attributed to Al addition, which contributes to its
81 low twinning activity compared with low Al alloy or Fe-Mn-C TWIP steels [16]. In addition, the
82 texture transition from copper to brass texture was observed at higher reduction with strong
83 similarity to that found in Fe-Mn-C TWIP steels. Compared with Fe-Mn-C TWIP steels, the texture of
84 Fe-Mn-Si-Al TWIP steels shows a great difference in tensile deformation process due to the lower
85 magnitude of the activated twin systems despite of the fact that the strain-hardening rate curves of
86 the two steels are quite similar [17]. Furthermore, Saleh et al. proposed that the texture measurement
87 of Fe-Mn-Si-Al TWIP steels subjected to the uniaxial tension demonstrated the characteristic double
88 fibre texture for FCC materials, with a relatively stronger <111> and a weaker <100> partial fibre
89 parallel to the tensile axis [18]. However, a detailed explanation of texture transition has not been
90 explored in these researches.
91 To model the relationship between crystallographic texture and deformation twins’ evolution,
92 Dancette et al. *19+ employed a “full-field” and experimental dataset based CPFE analysis to provide
93 an improved prediction of the texture development in uniaxial tension at macroscopic scale, and the
94 relation between the volume fraction of twins inside the individual grains and crystallographic
95 texture at the grain level were confirmed via EBSD measurement and CPFE simulation. The
 Metals 2017, 7, x FOR PEER REVIEW 3 of 19

96 introduced crystal plasticity (CP) model proved that it gives an improved texture prediction
97 compared to the Taylor model for the deformation at higher strain. Based on a classic single crystal
98 plasticity model, Li et al. [20] proposed a mechanistic model for strain hardening by twin thickening
99 and classified the different effective stages of deformation twins and dislocations slip on the texture
100 evolution. Based on the long-range interactions of individual grains with the polycrystalline
101 aggregate [21], another widely used approach for predicting crystallographic texture is the
102 viscoplastic self-consistent (VPSC) plasticity model. Through applying predominant twin
103 reorientation (PTR) schemes [22] and CP model proposed by Kalidindi [23] to deal with the twinning
104 evolution, a VPSC model was presented by Prakash et al. [24] to investigate the influence of
105 deformation twins on the crystallographic texture development after tensile loading. However,
106 compared with Kalidindi model, a limited correspondence between the experimental texture
107 intensity and the simulated one obtained by the simulation using VPSC model and considering PTR
108 scheme was concluded. Consequently, Saleh et al. [18] applied a modified VPSC model to assess the
109 contributions of perfect and/or partial slip, twinning and latent hardening to evolution of the
110 crystallographic texture in TWIP steels and overcame the limited correspondence between
111 experiment and simulation via proper description of hardening parameters of various deformation
112 systems. However, this “mean-field” approach for VPSC model has limitation as there is no
113 sufficient information about the specific interaction between the individual grain and its
114 neighboring grains [25, 26]. Considering this limitation, a non-homogenization scheme based CPFE
115 method was proposed to represent the intragranular misorientation and its distributions for
116 determining the heterogeneous deformation within grains. Furthermore, this approach facilitates
117 taking into account the grain morphology of metallic materials in microstructural simulation [27,
118 28].
119 In the present study, the microstructural and crystallographic texture evolution of
120 Fe-30Mn-3Si-2Al TWIP steel under the uniaxial tensile deformation was studied via EBSD
121 measurement, especially for the deformation twins’ nucleation position and intra- and intergranular
122 deformation. The corresponding CPFE model considering slip and twinning interactions was then
123 proposed. Subsequently, the simulations using Voronoi based polycrystalline RVE microstructure
124 with crystal orientations were conducted to investigate the evolution of hardening behavior
125 associated with slip and twinning activity and further to discuss the effect of twinning on hardening
126 and crystal orientation, as well as the stability of the crystallographic texture development. This
127 study links the mechanical behaviors with microstructural change and texture evolution during
128 deformation thereby provides an accurate prediction of forming shape in deformation processing of
129 TWIP steels.

130 2. Experiments and characterization

131 2.1 Material and experimental procedure

132 The kind of Fe-Mn-Si-Al TWIP steel was selected as the case material. A vacuum induction
133 furnace was used to manufacture the investigated steel with the chemical composition as listed in
134 Table 1. The tensile specimens with a gauge length of 20 mm, gauge width of 10 mm, and the
135 thickness of 0.5 mm were prepared by electron discharge machining of the fabricated steel sheets.
136 The samples were then suitably annealed at the temperature of 800°C for a fixed duration of 1 h and
137 then followed by air cooling.
138 Table 1. Chemical composition (in weight percent) of the Fe-30Mn-3Si-2Al TWIP steel sample.

C Mn Si Al S P Ti Fe
0.11 30.5 2.88 2.34 0.013 0.007 ≤0.01 Bal.
139
140 The tensile tests were conducted on a mechanical testing system with the pre-set constant
141 crosshead speed of 6.9×10-3 mm/s in the tensile direction parallel to the rolling direction (RD). The
142 tests were done with the engineering strain of 0.05, 0.1, 0.2 and 0.4. During the deformation process,
 Metals 2017, 7, x FOR PEER REVIEW 4 of 19

143 a contactless laser extensometer was used to calibrate and measure the deformation strain of the
144 testing samples.
145 Grain morphology and crystal orientation were characterized before deformation by using
146 optical microscopy (OM) and EBSD techniques, respectively. The specimens were mechanically
147 polished using the standard method along the longitudinal section (the plane normal to transverse
148 direction) and etched in an alcoholic solution consisting of 5% nitric acid. OM morphologies
149 demonstrated the equiaxed austenite grains with the average grain size of 30 m (excluding twins)
150 for the homogenized sample shown in Figure 1. The samples were electrochemically polished in 5%
151 perchloric acid alcoholic solution for EBSD preparation. The EBSD scanning was conducted with
152 the step size of 1 m using a field emission Zeiss Auriga scanning electron microscope (SEM) with a
153 HKL camera. Grain boundaries from both the un-deformed and deformed specimens were
154 identified with misorientation greater than 5°, whereas the standard Brandon’s criterion *29, 30+
155 was used to identify the coincident site lattice boundaries. Furthermore, the high-angle grain
156 boundaries (HAGBs) in EBSD orientation maps were defined to have the misorientation larger than
157 15° and the low-angle grain boundaries (LAGBs) have the misorientation between 2-15°. The EBSD
158 data was used for calculation of the pole figures (PFs), inverse pole figures (IPFs), orientation
159 distribution functions (ODFs), and texture components. These results were applied to understand
160 the crystal orientations evolution and the crystallographic texture development in deformation
161 process.

162 2.2 Misorientation-dependent twinning characteristics

163 Figure 1 shows the OM of the as-annealed TWIP steel and the ample annealing twin lamellae is
164 clearly distinguished in the microstructure. Statistical analyses show Fe-30Mn-3Si-2Al-0.11C steel
165 has an average grain size of ~30 m. The subsequently EBSD observations were conducted on
166 RDTD plane of the initial-state sample, providing crystal orientations, grain size distribution, and
167 confirmed that the sample has fine-grained structure as shown in Figure 2 (a)-(b). The morphologies
168 of the annealing twins are lath-shaped characteristically (Figure 2a), while the deformation twins
169 often appear in lenticular and its boundaries in clusters. The experimental (111), (110) and (100) PF
170 plots in Figure 2 (c) indicated that the initial crystallographic orientations of annealed specimen are
171 approximately stochastic distributed scatters, which represents that weak textures exist in the
172 annealed samples.

173
174 Figure 1. Morphology of the Fe-30Mn-3Si-2Al-0.11C steel annealed at the temperature 800 oC for 1h.

(b) 0.05
(a)
0.04

0.03
Frequency

0.02

TD 0.01

0.00
RD 0 10 20 30 40 50 60 70 80 90 100
Grain Size, m
175
 Metals 2017, 7, x FOR PEER REVIEW 5 of 19

176
177 Figure 2. (a) EBSD imaging map cropping from RDTD plane; (b) Statistics of grain size
178 distribution; (c) Experimental (111), (110) and (100) PFs of the annealed TWIP steel samples.

179
180 Figure 3. (a) Distribution of twinning lamella (denoted by arrows) of the specimen with the tension
181 strain of 0.1; (b) Occurrence of a large amount of deformation twinning (partially labeled by black
182 circles) when deformed to 0.4.

183 The IPF maps, which represent the microstructural evolution including twinning distribution
184 in the TWIP steel samples after the uniaxially deformation with the strain of 0.1 and 0.4, are shown
185 in Figure 3. From Figure 3 (a), it is found there is no discernible difference in the grain size and
186 shape except the morphology of annealing twins when compared with Figure 3 (b). It is noted that
187 the deformation twin lamellae was originated mostly from the boundaries of the individual grains
188 with the large grain size such as grains A and B in Figure 3 (a). Within the mostly twinned grains,
189 just one kind of twin lamellae was detected while there are two types of twin lamellae with
190 different crystal orientations measured in grain C. In addition, it is noted that the twin shear
191 associated with twinning in FCC materials is 0.707, which is very high and hence the twin cluster
192 width is very small with the order of few micrometer, as marked by the black ovals in Figure 3 (b).
193 To efficiently determine the position of twin lamellae, it is noted that twin lamellae is often
194 originated from a statistical distribution of defects in the grain boundaries and are activated by the
195 local stress at the grain boundaries [31, 32], which indicates the probability of twin nucleation has a
196 close relation with the parameters of grain boundary, particularly the grain boundary
197 misorientation angle, exhibiting a significant influence on the mechanism for twin nucleation. For
198 the EBSD analysis, as illustrated in Figure 4 (c), the misorientation angles between the adjacent
199 grains of specimen with the tension strain of 0.1 are above 15 o, and mostly above 60o, which
200 illustrates the majority of grain boundaries belonging to HAGBs, including the twinning grain
201 boundaries (TGBs). A visualization of grain boundary misorientation of the sample with the strain
202 of 0.1 is shown in Figure 4 (d). Indeed, HAGBs represented by red color code have a greater
203 magnitude than that of LAGBs highlighted by blue color code. In addition, it is observed that for
204 the uniaxial tensile loading condition, mostly twin lamellae in the FCC-type phase was originated
205 in the HAGBs above 25o.
 Metals 2017, 7, x FOR PEER REVIEW 6 of 19

206
207 Figure 4. (a) Misorientation angles distribution within grains of the specimen with the tension strain
208 of 0.1; (b) the corresponding visualization of the intragranular misorientations; (c) misorientation
209 angles distribution between adjacent grains; and (d) the grain boundary misorientations.

210 With the increase of strain to 0.4, as shown in Figure 5 (c), the misorientation angles above 60o
211 exhibit a decreasing tendency from 55 to 40%, indicating a transition period in which the annealing
212 twins was reduced and the deformation twins emerged. In addition, these deformation twins are
213 attached to the HAGBs of the distorted grains, as shown in Figure 5 (d). It could qualitatively
214 conclude that twin lamellae mostly adhere to HAGBs for a FCC- type TWIP steel undergoing
215 uniaxial tensile deformation. During the plastic deformation process of polycrystalline metallic
216 materials, the substructures within the individual grains are heterogeneous caused by the
217 interaction with neighboring grains, etc. The intragranular misorientations thus evolved with the
218 strain from 0.1 to 0.4, as shown in Figure 4 (a) and 5 (a), respectively. Comparing Figure 4 (a) with
219 Figure 5 (a), it is noted that the major distributions of intragranular misorientation at the strain of
220 0.1 are located in low angles (below 15 o). With increasing tensile strain to 0.4, a similar distribution
221 of intragranular misorientation occurs statistically, which indicates that the lattice rotation is small
222 and the intragranular deformation is approximately homogeneous.

223
 Metals 2017, 7, x FOR PEER REVIEW 7 of 19

224 Figure 5. (a) Misorientation angles distribution within grains of the testing sampling with the strain
225 of 0.4; (b) the corresponding visualization of intragranular misorientations; (c) intergranular
226 misorientation angles distribution; and (d) the grain boundary misorientations.

227 2.3 Crystallographic texture evolution

228 In order to get a complete description of the crystallographic texture evolution, ODFs were
229 calculated based on the experimental PFs. To present the gradual transition of texture of the
230 polycrystalline FCC materials, the partially ideal and important texture components are labeled
231 schematically for the sections of ODFs with2 = 0o, 45o and 60o in Figure 6, as well as the
232 corresponding experimental ODFs with the tensile strain from 0.05 to 0.4.

233

234 Figure 6. Texture development at the strain of 0.05, 0.1, 0.2, and 0.4, the shown sections of ODFs for
235 2 = 0o, 45o, 60o determined by EBSD, and the important texture components in FCC materials.

236 The as-annealed sample has a rather weak grain orientation, as shown in Figure 2 (c). However,
237 as seen in Figure 6 with the strain of 0.05, the starting deformation texture is comprised mainly of A
238 ({110}<556>) orientation, Brass (B) orientation along -fibre, as well as Copper (Cu) and S ({123}<634>)
239 orientation. With the greater tensile strain of 0.1, a pronounced increase in the intensity of the A
240 orientation was found in the -fibre of 2=45o ODF figures. Meanwhile, the B orientation also shows
241 an apparent increase in the intensity with the increasing tensile strain. With the strain larger than
242 0.2, the main textures consist of Goss ({110}<001>) and B orientations. This observation is consistent
243 with the texture evolutions in the conventional FCC materials after uniaxial tension [18]. In addition,
244 a weaken {112}<111> component is also identified. With the increase of tensile strain to 0.4, the main
245 texture components are similar to those observed in the specimen deformed to the strain of 0.2.
246 However, the intensity of several orientations becomes stronger. At this level, an important feature
247 is the remarkable enhancement of the intensity for {112}<111> copper orientation. In addition,
248 another feature is the further decreasing {110}<556> A orientation.
249 To quantify the stability of the main textures in tensile deformation, the volume fraction of
250 each texture component is plotted as a function of tensile strain, as shown in Figure 7. Indeed, the
251 Cu and A components show a large majority comparing with G and B textures. It is noted that the
252 volume fraction of A and G orientations were found to remain nearly stable during the uniaxial
253 tension although there is a slight decrease with the strain of 0.1. However, the Cu orientation
254 exhibits a large decrease in volume fraction while a smaller increase in volume fraction is illustrated
255 for B orientation. This indicates that more grains rotated into B orientation although its volume
256 fraction remains relatively less. This is attributed to the effect of dislocation slip and deformation
 Metals 2017, 7, x FOR PEER REVIEW 8 of 19

257 twinning. In addition, Cu orientations shifted towards brass orientations as a result of dislocation
258 slip. Meanwhile, since dislocations slip was hindered by the twin boundaries and the lattice rotated
259 and favored the formation of B texture component.

0.5
A
Cu
0.4 G
B

Volume fraction 0.3

0.2

0.1

0.0
0.0 0.1 0.2 0.3 0.4 0.5
260 Strain

261 Figure 7. The calculated volume fractions of the important texture components in the tension
262 deformation of TWIP steels.

263 3. Microstructure-based CPFE modeling

264 3.1. Crystal plasticity model including slip and twinning

265 It is well established that dislocations slip is considered as the main physical mechanism of
266 crystal plastic deformation at ambient temperature, as well as deformation twinning. Developed
267 based on dislocations and twinning evolution in crystals, the crystal plasticity (CP) constitutive
268 model is widely adopted to represent the macro- and microscopic evolution of polycrystalline
269 materials, particularly for representing the crystallographic texture development. In this paper, a
270 brief description of this CP model is given.
271 In the proposed CP model, following the work of Kalidindi et al. [33], the decomposition of the
272 deformation gradient tensor F , is formulated as

273 F  Fe  F p , (1)

274 where F e depicts the elastic deformation, while F p represents an irreversible permanent plastic
275 deformation.
276 To explain the evolution of finite deformation kinematics, the plastic deformation gradient F p ,
277 represented by the plastic velocity gradient L p which considers dislocation slip and deformation
278 twinning, is described by the following:
Nslip Ntwin
Lp  F p  F p     SSlip 
1
279 s
  S Twin
t , (2)
 1  1

280 where   (  = 1 to N slip ) denotes the rate of slip shear from each possible slip system,   (  = 1
281 to Ntwin ) represents the rate of twin shear. N slip is the number of slip systems and Ntwin is the
282 number of twin systems. In addition, SSlip
s
and STwin
t
describe the Schmid tensor produced by a unit
283 slip on system s and a unit twin on system t, respectively.
284 The stress based constitutive equation to represent the elastic stretch is given by:
285 T e   : Ee , (3)
 Metals 2017, 7, x FOR PEER REVIEW 9 of 19

286 where  is a fourth-order anisotropic elasticity tensor. E e and T e represent Green elastic strain
287 and the symmetric second Piola-Kirchoff (PK) stress, respectively. These variables are defined in the
288 following equations:
1 eT e
289 Ee  (F  F  I) . (4)
2

290 On the other hand, the stress measure T e is designated as:

 
1 T
291 T e  Fe  det(Fe )  σ  Fe , (5)

292 where σ indicates the Cauchy stress. The resolved shear stress (RSS)   associated with the 
293 slip system is defined as follows:

294   =σ : S . (6)
295 In this paper, the constitutive model is established using rate-dependent approach. In the
296 rate-dependent CP model, the shear rate of slip systems can be obtained directly by the
297 decomposition of shear stress [34]:
1m

298   = 0
s   
sign   , (7)

299 where   denotes the RSS on the slip system  . s is the slip resistance for the slip system  .
300 m is the strain rate sensitivity factor.  0 is a reference slip shear rate, which is considered to be
301 the same for all the slip systems [35].
302 Twin volume fraction (TVF) is treated as a variable to represent the evolution of deformation
303 twinning. The power law is adopted to describe the evolution of TVF as:
1n
 
304  = 0  
 
 , if    0;    0, if    0 , (8)
s 
 tw 

305 where s tw and   represent the twin resistance in a twin system and the RSS of that twin system,
306 respectively. n is the strain rate sensitivity factor.  0 is a reference shear rate of all the twinning
307 systems.
308 The formulation of the hardening relation modeling of the evolution of FCC-type TWIP steel is
309 presented. It is assumed the ratio between twin and slip resistance can be a constant which depends
310 on the twin morphology based on the Hall-Petch explanation. Consequently, in this study, the
311 twinning resistance is considered proportionally to slip resistance. The evolution of slip resistance of
312 the th slip system is formulated as follows:

 s  Nslip 
313 s  hs  1    , (9)
 s   1
 s 

314 where hs and ss represent the hardening rate and the saturated value associated with the slip
315 system  , respectively. The extended hardening equations to capture the complex interactions of
316 slip and twinning are given by [36]:

317 
hs  hs 1  C  f   ,

b
(10)

 f 
0.5
318 ss  ss0  spr 
. (11)
 Metals 2017, 7, x FOR PEER REVIEW 10 of 19

319 The introduced CP constitutive model and integration algorithms are implemented
320 numerically into the commercial finite element code ABAQUS standard via a user-defined material
321 (UMAT) subroutine. A detailed Newton-Raphson iteration scheme could be seen in [7].

322 3.2 Establishment of virtual polycrystalline microstructure

323 To represent the macro-scale response of specimen, a RVE model containing 100 grains, which
324 are approximately consistent with the grain number in EBSD experiments, was employed to
325 reconstruct the polycrystalline microstructure which considers the crystallographic orientation and
326 morphological feature, as shown in Figure 8 (a). This virtual polycrystalline microstructure is
327 generated using an open-source software package: Neper construct as the voronoi tessellation
328 model [37]. In Figure 8 (b), the FE model of this polycrystalline microstructure contains
329 approximately 14200 linear tetrahedral elements (C3D4 in ABAQUS) since the use of higher-order
330 elements does not have a significant influence on the stress-strain field and the prediction of
331 crystallographic texture [38, 39, 40]. The polycrystalline model is then subjected to uniaxial tension
332 up to the strain of 0.4. The initial un-deformed microstructure is illustrated in Figure 8 (b) and the
333 dimensions of the simulation model with 10 mm×10 mm×0.5 mm are given. As shown in Figure 8
334 (b), the simple tensile test was performed by imposing displacement in the X direction (the rolling
335 direction, RD) with the lateral YZ plane fixed. Using trial and error tests, a set of the model material
336 parameters are illustrated in Table 2. With the calibrated material parameters obtained by fitting the
337 stress-strain curves performed with the strain of 0.4, the crystal orientations can be predicted to
338 conduct the simulation.

339
340 Figure 8. Voronoi tessellation subjected to uniaxial tension. Colors represent different grains with
341 different crystallographic orientations. (a) Polycrystalline morphology; (b) C3D4 mesh refinement
342 with 100-grain polycrystals.

343 Table 2. Material parameters in the constitutive relations calibrated for TWIP steel.

Initial hardening rate of slip system hs (MPa) 180

Saturated value of slip resistance without twinning Ss 0 (MPa) 300
Hardening index of twinning b 2
Hardening coefficient of twinning C 10
Effect of Hall-Petch mechanism Spr (MPa) 300
Initial slip resistance S0 (MPa) 120

Initial twinning resistance S (MPa)
0
139.2
Elastic constant C11(GPa) 198
Elastic constant C12(GPa) 125
Elastic constant C44(GPa) 122
Rate sensitivity coefficient m 0.02
Reference shear rate  0 (s-1) 0.001
 Metals 2017, 7, x FOR PEER REVIEW 11 of 19

344 3.3 Evaluation of the CPFE model

345 It is noted that the conventional way to calibrate the CPFE constitutive parameters in
346 simulation is to establish a FE model which assumes that all elements consist of the same number of
347 representative crystallographic orientations. However, the grain morphology such as intra- and
348 intergranular misorientations could not well be reflected. Furthermore, the local deformation
349 induced by individual grain properties also could not be reproduced. In this research, individual
350 grains with orientations is generated and meshed in the FE model. Due to the limit computing
351 power especially for 3D simulations, modeling of a macroscopic specimen containing millions of
352 grains and is then implemented within a FE model is difficult. Consequently, a quasi-3D RVE
353 model containing grain morphology is proposed to represent the mechanical behavior of the
354 macroscopic testing samples.
355 In this RVE model, each grain represents a specific crystallographic orientation and is
356 composed of many elements. Instead of the conventional approach which directly assigning all the
357 EBSD crystal orientations into RVE grain set, the crystal orientations used in this research were
358 generated from discretizing ODFs using MTEX toolbox [41] in order to represent initial texture and
359 to avoid the difficulty of selecting appropriate experimental orientations [42]. Subsequently, these
360 orientations representing the texture were assigned to the RVE grains to describe an
361 inhomogeneous material. Based on the microstructure and texture, the RVE model could capture
362 the anisotropy characteristics of TWIP steel in order to bridge the gap between experiments and
363 simulations.
364 To consolidate the simulated reliability, the comparison between the simulated and
365 experimental intergranular misorientation and intragranular misorientation at the strain of 0.4 are
366 shown in Figure 9. It is noted that a similar distribution of intergranular misorientation obtained by
367 CPFE and EBSD could be observed. Meanwhile, the major boundary misorientation angles are
368 above 15o, indicating there are a large number of twin boundaries, as shown in Figure 9 (a).
369 However, in Figure 9 (b), both the simulated and experimental intragranular misorientation shows
370 a high frequency of the low misorientation angle below 15o, which represents the approximate
371 homogeneous orientation within the individual grain.
(a) 60 (b) 80
Sim. results by CPFE Sim. results by CPFE
70
50 Exp. results from EBSD Exp. results from EBSD
60
40 50
Frequency

Frequency

30 40
30
20
20
10
10
0 0
0 10 20 30 40 50 60 70 0 5 10 15 20 25 30
Intergranular misorientation (degree) Intragranular misorientation (degree)
372
373 Figure 9. Comparison between simulated and experimental (a) intergranular misorientation, and (b)
374 intragranular misorientation at a strain of 0.4.

375 Based on the reliability of the established RVE model, as shown in Figure 10, the comparison of
376 the simulated stress-strain curves with the experimental data provides a strong evidence to validate
377 the proposed CP model in the macroscopic response of TWIP steel. Although there is a slight
378 deviation with the true strain above 0.25, the hardening trend of the numerical simulation captures
379 the main features of the experimental measurement. One factor leading to the deviation may be
380 attributed to the homogeneous misorientation in the individual grains in simulation, where the
381 actual crystal orientation is heterogeneous even in one grain. In addition, since deformation twins
382 increase the work-hardening rate by acting as obstacles for gliding dislocations, TWIP effect plays
383 another significant role leading to additional hardening.
 Metals 2017, 7, x FOR PEER REVIEW 12 of 19

800
Sim. results by CPFE
700 Exp. results from tension test
600
500

Mises, MPa
400
300
200
100
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30
384 Strain

385 Figure 10. Comparison between simulation results by CPFE and experimental results from EBSD.

386 It is well established that twinning is considered as a significant role in the forming process of
387 TWIP steels and its formation was dramatically influenced by the crystal orientations [15, 17]. In
388 addition, the evolution of microstructure (e.g. twinning) and texture components strongly relies on
389 the crystals rotation. On the other hand, crystal orientations could be regarded as a strong evidence
390 to further verify the developed microstructure-dependent CPFE model. Then a comparison of the
391 simulated and measured PFs at different strains has been conducted in Figure 11. In Figure 11 (a), it
392 could be observed that strong scattered distributions of grain orientations are shown in the EBSD
393 measurement at the strain of 0.05, meanwhile, a similar position and intensity could be found in the
394 simulated PFs predicted by the CPFE model. In Figure 11 (b), it illustrates that the proposed CPFE
395 model well captures the main features of grain orientations of TWIP steels after the uniaxial tensile
396 deformation to the strain of 0.4. It confirms that the macroscopic response and microscopic grain
397 orientation distributions can be predicted and simulated by the microstructure-dependent CPFE
398 model via introducing the initial experiment-based crystal orientations.

399
 Metals 2017, 7, x FOR PEER REVIEW 13 of 19

400
401 Figure 11. PFs of TWIP steels obtained by the EBSD measurement and predicted by the
402 microstructure-dependent CPFE model with the strain of (a) 0.05 and (b) 0.4.

403 4. Simulation and discussion

404 4.1 Effect of twinning on hardening evolution

405 The FE simulations of uniaxial tension were employed to validate the CP model for describing
406 the twinning effect on hardening behavior and the evolution of slip and twinning in micro level,
407 respectively. The equivalent von-Mises stress distribution was calculated by the proposed CPFE
408 model. It could be observed that there are non-uniform stress fields caused by various crystal
409 orientations. In order to articulate the contributions of slip and twinning to the hardening evolution
410 in each grain with experiment-based crystallographic orientation, the plot of the evolution of slip
411 resistance is compared with the evolution of twin volume fraction to represent the individual
412 hardening effect of slip and twinning to overall von-Mises stress distribution in the simulated
413 microstructure for the specimen deformed to the strain of 0.4, as shown in Figure 12.

414
415 Figure 12. (a) von-Mises stress distribution (stress concentration within the white dashed line
416 region), (b) slip resistance, (c) twin volume fraction, and (d) twin resistance in the simulated
417 microstructure of the sample stretched to the strain of 0.4.
 Metals 2017, 7, x FOR PEER REVIEW 14 of 19

418 The heterogeneous distribution of slip resistance exhibits a similar appearance compared with
419 that of twin resistance, while the twin resistance shows a higher average intensity than slip
420 resistance, shown in Figure 12 (b) and Figure 12 (d). It is noted that the similar distribution of
421 von-Mises stress and twin volume fraction was observed in simulation, as seen within the white
422 dashed line region in Figure 12 (a) and (c). The total twin volume fraction is statistically about 40%
423 according to the ratio of twinned region to whole grains. Particularly, in Figure 12 (c), the grain
424 with red color represents twin volume fraction shows a relatively higher value than other twinned
425 regions, which shows the stress concentration. It indicated that the deformation twinning plays a
426 significant role in the hardening effect of TWIP steel during uniaxial tensile process.
427 It is noted that in Figure 13, the evolution of slip resistance and twin resistance represent an
428 approximately parallel increasing trend, but the magnitude of twin resistance with initial value is
429 larger than that of slip resistance. In Figure 13, the proportion of hardening due to twinning to
430 overall hardening shows a decreasing trend with the increasing strain, revealing that the twin
431 volume fraction gradually reaches the saturated value. Twins are considered as undeformable hard
432 particles in austenitic matrix, acting as new obstacles for dislocation motions, leading to the
433 so-called dynamic Hall-Petch effect. The obstacles including twinning boundaries impeded the
434 dislocation slip so that the slip increment decreased and caused a reduced proportion of slip
435 hardening to overall hardening.
0.45 200
Twin hardening
Proportion to overall hardening

Slip hardening
0.40 Twin resistance 180
Slip resistance
0.35

Resistance, MPa
160
0.30

140
0.25

0.20 120
0.00 0.05 0.10 0.15 0.20 0.25 0.30
436 Strain

437 Figure 13. Evolution of the simulated slip and twin resistance at microscopic level and the
438 proportion of slip hardening and twin hardening to overall hardening in the simulated
439 microstructure of the sample stretched to the strain of 0.4.

440 Table 3. 12 slip systems and 12 twinning systems of TWIP steel.

Plane Direction Plane Direction

a1 b1 ( ) [ ̅] (̅ ̅ ) [ ̅ ̅]
a2 b2 ( ) [̅ ] (̅ ̅ ) [ ]
Slip a3 b3 ( ) [ ̅ ] (̅ ̅ ) [̅ ]
systems c1 d1 (̅ ) [ ̅] ( ̅ ) [ ̅ ̅]
c2 d2 (̅ ) [ ] ( ̅ ) [̅ ]
c3 d3 (̅ ) [̅ ̅ ] ( ̅ ) [ ]
t1 u1 ( ) [ ̅] (̅ ̅ ) [ ]
t2 u2 ( ) [̅ ] (̅ ̅ ) [ ̅ ]
Twinning t3 u3 ( ) [ ̅ ] (̅ ̅ ) [̅ ]
systems v1 w1 ( ̅ ) [ ] ( ̅ ) [ ]
v2 w2 (̅ ) [ ̅] ( ̅ ) [ ̅]
v3 w3 (̅ ) [ ̅ ] ( ̅ ) [̅ ]
441 Based on the qualitative and quantitative analyses of the micro-deformation mechanisms and
442 hardening behavior, polycrystalline materials such as TWIP steels exhibit deformation twins within
443 the grains and the intragranular microstructure evolution including twin formation and
444 propagation changes greatly in the course of plastic flow, which is mainly responsible for strain
 Metals 2017, 7, x FOR PEER REVIEW 15 of 19

445 hardening. The activity of the slip and twinning systems was performed to investigate the
446 individual contribution of slip and twinning to hardening, respectively. All the slip and twinning
447 systems in the FCC-type TWIP steels are shown in Table 3.
(a) 0.08 a3 (b) 0.015 t1
a2 t3
0.06
a1 t2
Dominant hardening
0.04 c2 v2
c3 0.010 v3
Slip activity

0.02

Twin activity
c1 v1
0.00 d2 w1
d1 w2
-0.02 d3 0.005 w3
-0.04 b3 u1
b2 u3
-0.06 b1 u2
0.000
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
448 Strain Strain

449 Figure 14. (a) Active slip systems and (b) twinning systems of the testing sample predicted by
450 simulation with the strain of 0.4.

451 As shown in Figure 14 (a), it could be observed that there is symmetric slip systems activation
452 since the resolved shear stress of each slip systems meets the critical value during the tensile
453 deformation. However, with the increase of strain up to about 0.3, the magnitude of slip activity of
454 the slip systems designated as d1, a1 and b1 remains at a low level. In Figure 14 (b), it is noted that
455 the twinning systems denoted as v2, v3, u1, u3, t1 and w3 play a dominative role in hardening
456 evolution, and the hardening is attributed to the fact that it is the twinning boundaries of system
457 that impedes the movement of the dislocations slip and results in the dislocations pile-up, leading
458 to an increasing hardening effect.

459 4.2 Effect of twinning on crystal orientation

460 The influence of slip and twinning on hardening was investigated and demonstrated that
461 twinning has a larger effect on overall hardening. To determine whether microstructural evolution
462 such as the change of crystallographic orientation is influenced by twinning, a comparison of PF
463 plots predicted by CPFE model without twinning and considering the effect of twinning is given in
464 Figure 15. It is noted that with twinning effect, the scattered orientations in the (111), (110) and (100)
465 PFs obtained from CPFE model considering only slip rotated 90° clockwisely. The occurrence of
466 twinning induced crystals reorientations is shown in Figure 15 (b). A distinct texture occurred is
467 also shown in the figure. To make a precise prediction of texture evolution, it is essential to
468 introduce twinning into CPFE modeling since twinning has a significant influence on texture
469 evolution.

470
 Metals 2017, 7, x FOR PEER REVIEW 16 of 19

471 Figure 15. Comparison of PF plots obtained from CPFE model (a) without twinning and (b)
472 considering the effect of twinning at the strain of 0.4.

473 4.3 Evolution of texture components

474 As is well known that the anisotropy associated with texture components have a significant
475 influence on the plastic processing [43]. With the increasing tensile deformation at room
476 temperature, texture components would change significantly due to the evolution of twinning and
477 crystal orientations. To investigate the stability of texture components evolution could provide a
478 basis for predicting the anisotropy of Fe-Mn-Si-Al TWIP steel. The simulated textures are presented
479 in terms of 2=45o sections of the ODFs in the space of Euler angles (1, )≤90o via the proposed
480 CPFE model. And the final texture components of the samples after uniaxial tension test are shown
481 in the Figure 16. In the ODF plots, it is noted that the strong A and Brass components along the
482 -fibre could be noticed at the strain of 0.05 and 0.1 in the EBSD observation of TWIP steels,
483 meanwhile, a weak Goss component along the -fibre could be seen in the simulation results at the
484 strain of 0.05, as shown in Figure 6 and Figure 16. However, only the Brass texture component
485 along the -fibre is found at the strain of about 0.4. The Goss, Brass and A orientations are
486 dramatically influenced by twinning formation and its evolution. In the ODF plots of each strain,
487 the 2 section at 45o exposes a strong Copper component along the -fibre, which indicates that the
488 Copper texture is the major part among the deformation textures. Consequently, this CPFE model
489 prediction captures the changing characteristics of the textures named A, Brass and Copper
490 components in measurement and this prediction is similar to those reported previously for the
491 Fe-24Mn-3Al-2Si-1Ni-0.06C TWIP steels [18]. Furthermore, the -fibre <111>//ND, with relatively
492 low intensity about 3.0, could be observed in simulation at the strain of 0.1 but disappear with the
493 tensile strain up to 0.4. So this weak -fibre could be ignored in the analysis of texture evolution and
494 further its influence seems little during the forming process [44]. With the given crystal orientations,
495 the CPFE model describes the main texture components observed in the experimental results
496 although the grain numbers are limited to represent the complete textures. Furthermore, to
497 simulate and predict the anisotropy under uniaxial tension via the texture evolution more precisely,
498 a complex polycrystalline structure with realistic grain properties including grain boundaries must
499 be incorporated in the CPFE simulations.

500
501 Figure 16. Constant ODFs (2=45o) sections predicted by the CPFE model.

502 5. Conclusions
503 By using SEM/EBSD experiments and CPFE model, the characteristics of twinning nucleation
504 and its influence on hardening behavior and crystallographic texture evolution of the Fe-Mn-Si-Al
505 TWIP steels in uniaxial tensile deformation were investigated. The following concluding remarks
506 can be drawn:
 Metals 2017, 7, x FOR PEER REVIEW 17 of 19

507 1. The formation of HAGBs detected by intergranular misorientation distribution could promote
508 the deformation twin lamellae nucleation for a FCC-type TWIP steel in uniaxial tensile
509 deformation. In addition, the low frequency of intragranular local misorientation indicates
510 homogeneity within individual grains.
511 2. Twinning promotes the crystal orientations rotation but has less impact on the intensity of
512 crystal distribution comparing with no twinning effect. Furthermore, the texture components
513 including the Goss, S and A orientations show a stable volume fraction, while the Cu
514 orientation shifted towards Brass orientation.
515 3. The proportion of hardening due to twinning to overall hardening is larger than that due to
516 slip, which could be demonstrated by the evolution of twin volume fraction and twin
517 resistance.
518 Acknowledgments: The authors would like to thank the supports to this research from the National Natural
519 Science Foundation of China (No. 51575039), NSAF (No. U1330121).

520 Author Contributions: Ning Guo and Chaoyang Sun conceived, designed and performed the experiments;
521 Ning Guo, Chaoyang Sun and Mingchuan Han analyzed the data; Ning Guo and Chaoyang Sun contributed to
522 writing and editing of the manuscript; Mingwang Fu polished the English.

523 Conflicts of Interest: The authors declare no conflict of interest.

524 References
525 1. Han, H.N.; Oh, C.S.; Kim, G.; Kwon, O. Design method for TRIP-aided multiphase steel based on a
526 microstructure-based modelling for transformation-induced plasticity and mechanically induced
527 martensitic transformation, Mater. Sci. Eng. A 2009, 499, 462-468.
528 2. Patterson, E.E.; Field, D.P.; Zhang, Y. Characterization of twin boundaries in an Fe-17.5Mn-0.56C twinning
529 induced plasticity steel, Mater. Charact. 2013, 85, 100-110.
530 3. Huang, B.X.; Wang, X.D.; Rong, Y.H.; Wang, L.; Jin, L. Mechanical behavior and martensitic
531 transformation of an Fe-Mn-Si-Al-Nb alloy, Mater. Sci. Eng. A 2006, 438-440, 306-311.
532 4. Jiménez, J.A.; Frommeyer, G. Analysis of the microstructure evolution during tensile testing at room
533 temperature of high-manganese austenitic steel, Mater. Charact. 2010, 61, 221-226.
534 5. Kowalska, J.; Ratuszek, W.; Witkowska, M.; Zielińska-Lipiec, A.; Tokarski, T. Microstructure and texture
535 characteristics of the metastable Fe-21Mn-3Si-3Al alloy after cold deformation, J. Alloys Compd. 2015, 643,
536 S39-S45.
537 6. Kim, H.; Suh, D.W.; Kim, N.J. Fe-Al-Mn-C lightweight structural alloys: a review on the microstructures
538 and mechanical propert ies, Sci. Technol. Adv. Mater. 2013, 14, 014205.
539 7. Sun, C.Y.; Guo, N.; Fu, M.W.; Wang, S.W. Modeling of slip, twinning and transformation induced plastic
540 deformation for TWIP steel based on crystal plasticity, Int. J. Plast. 2016, 76, 186-212.
541 8. Allain, S.; Chateau, J.P.; Bouaziz, O.; Migot, S.; Guelton, N. Correlations between the calculated stacking
542 fault energy and the plasticity mechanisms in Fe-Mn-C alloys, Mater. Sci. Eng. A 2004, 387-389, 158-162.
543 9. Grässel, O.; Krüger, L.; Frommeyer, G.; Meyer, L.W. High strength Fe-Mn-(Al, Si) TRIP/TWIP steels
544 development-properties-application, Int. J. Plast. 2000, 16, 1391-1409.
545 10. Bracke, L.; Verbeken, K.; Kestens, L.; Penning, J. Microstructure and texture evolution during cold rolling
546 and annealing of a high Mn TWIP steel, Acta Mater. 2009, 57, 1512-1524.
547 11. Friedman, P.A.; Liao, K.-C.; Pan, J.; Barlat, F. Texture development and hardening characteristics of steel
548 sheets under plane-strain compression, J. Mater. Eng. Perform. 1999, 8, 225-235.
549 12. Sun, C.Y.; Wang, B.; Politis, D.J.; Wang, L.L.; Cai, Y.; Guo, X.R.; Guo, N. Prediction of earing in TWIP steel
550 sheets based on coupled twinning crystal plasticity model, Int. J. Adv. Manuf. Tech. 2017, 89, 3037-3047.
551 13. Renard, K.; Jacques, P.J. On the relationship between work hardening and twinning rate in TWIP steels,
552 Mater. Sci. Eng. A 2012, 542, 8-14.
553 14. Steinmetz, D.R.; Jäpel, T.; Wietbrock, B.; Eisenlohr, P.; Gutierrez-Urrutia, I.; Saeed-Akbari, A.; Hickel, T.;
554 Roters, F.; Raabe, D. Revealing the strain-hardening behavior of twinning-induced plasticity steels: Theory,
555 simulations, experiments, Acta Mater. 2013, 61, 494-510.
556 15. Barbier, D.; Gey, N.; Allain, S.; Bozzolo, N.; Humbert, M. Analysis of the tensile behavior of a TWIP steel
557 based on the texture and microstructure evolutions, Mater. Sci. Eng. A 2009, 500, 196-206.
 Metals 2017, 7, x FOR PEER REVIEW 18 of 19

558 16. Souza, F.M.; Padilha, A.F.; Gutierrez-Urruti, I.; Raabe, D. Texture evolution in the Fe-30.5Mn-8.0Al-1.2C
559 and Fe-30.5Mn-2.1Al-1.2C steels upon cold rolling, Rem: Rev. Esc. Minas 2016, 69, 59-65.
560 17. Wu, Y.; Tang, D.; Jiang, H.; Mi, Z.; Jing, H. Texture and microstructure evolution during tensile testing of
561 TWIP steels with diverse stacking fault energy, Acta Metall. Sin. 2013, 26, 713-720.
562 18. Saleh, A.A.; Pereloma, E.V.; Gazder, A.A. Microstructure and texture evolution in a
563 twinning-induced-plasticity steel during uniaxial tension, Acta Mater. 2013, 61, 2671-2691.
564 19. Dancette, S.; Delannay, L.; Renard, K.; Melchior, M.A.; Jacques, P.J. Crystal plasticity modeling of texture
565 development and hardening in TWIP steels, Acta Mater. 2012, 60, 2135-2145.
566 20. Li, Y.; Zhu, L.; Liu, Y.; Wei, Y.; Wu, Y.; Tang, D.; Mi, Z. On the strain hardening and texture evolution in
567 high manganese steels: Experiments and numerical investigation, J. Mech. Phys. Solids. 2013, 61, 2588-2604.
568 21. Lebensohn, R.A.; Tomé, C.N. A self-consistent viscoplastic model: prediction of rolling textures of
569 anisotropic polycrystals, Mater. Sci. Eng. A 1994, 175, 71-82.
570 22. Tomé, C.; Lebensohn, R.; Kocks, U. A model for texture development dominated by deformation twinning:
571 application to zirconium alloys, Acta Metall. Mater. 1991, 39, 2667-2680.
572 23. Kalidindi, S.R. Incorporation of deformation twinning in crystal plasticity models, J. Mech. Phys. Solids.
573 1998, 46, 267-290.
574 24. Prakash, A.; Hochrainer, T.; Reisacher, E.; Riedel, H. Twinning Models in Self-Consistent Texture
575 Simulations of TWIP Steels, Steel Res. Int. 2008, 79, 645-652.
576 25. Tari, V.; Rollett, A.D.; Kadiri, H.E.; Beladi, H.; Oppedal, A.L.; King, R.L. The effect of deformation
577 twinning on stress localization in a three dimensional TWIP steel microstructure, Modell. Simul. Mater. Sci.
578 Eng. 2015, 23, 045010.
579 26. Gu, C.F.; Hoffman, M.; Toth, L.S.; Zhang, Y.D. Grain size dependent texture evolution in severely rolled
580 pure copper, Mater. Charact. 2015,101, 180-188.
581 27. Lee, M.G.; Kim, S.J.; Han, H.N. Crystal plasticity finite element modeling of mechanically induced
582 martensitic transformation (MIMT) in metastable austenite, Int. J. Plast. 2010, 26, 688-710.
583 28. Sheikh, H.; Ebrahimi, R.; Bagherpour, E. Crystal plasticity finite element modeling of crystallographic
584 textures in simple shear extrusion (SSE) process, Mater. Des. 2016, 109, 289-299.
585 29. Roach, M.D.; Wright, S.I. Investigations of twin boundary fatigue cracking in nickel and
586 nitrogen-stabilized cold-worked austenitic stainless steels, Mater. Sci. Eng. A 2014, 607, 611-620.
587 30. Cao, S.Q.; Zhang, J.X.; Wu, J.S.; Chen, J.G. Effects of GBCD on cold work embrittlement of high strength
588 interstitial free steels, Mater. Des. 2006, 27, 53-57.
589 31. Galindo-Nava, E.I. Modelling twinning evolution during plastic deformation in hexagonal close-packed
590 metals, Mater. Des. 2015, 83, 327-343.
591 32. Beyerlein, I.J.; McCabe, R.J.; Tomé, C.N. Effect of microstructure on the nucleation of deformation twins in
592 polycrystalline high-purity magnesium: A multi-scale modeling study, J. Mech. Phys. Solids. 2011, 59,
593 988-1003.
594 33. Kalidindi, S.R.; Bronkhorst, C.A.; Anand, L. Crystallographic texture evolution in bulk deformation
595 processing of FCC metals, J. Mech. Phys. Solids. 1992, 40, 537-569.
596 34. Peirce, D.; Asaro, R.; Needleman, A. An analysis of nonuniform and localized deformation in ductile
597 single crystals, Acta Metal.1982, 30, 1087-1119.
598 35. Kalidindi, S.R. Modeling anisotropic strain hardening and deformation textures in low stacking fault
599 energy fcc metals, Int. J. Plast. 2001, 17, 837-860.
600 36. Salem, A.; Kalidindi, S.; Semiatin, S. Strain hardening due to deformation twinning in α-titanium:
601 Constitutive relations and crystal-plasticity modeling, Acta Mater. 2005, 53, 3495-3502.
602 37. Quey, R.; Dawson, P.R.; Barbe, F. Large-scale 3D random polycrystals for the finite element method:
603 Generation, meshing and remeshing, Comput. Method Appl. M. 2011, 200, 1729-1745.
604 38. Ardeljan, M.; McCabe, R.J.; Beyerlein, I.J.; Knezevic, M. Explicit incorporation of deformation twins into
605 crystal plasticity finite element models, Comput. Method Appl. M. 2015, 295, 396-413.
606 39. Knezevic, M.; Drach, B.; Ardeljan, M.; Beyerlein, I.J. Three dimensional predictions of grain scale plasticity
607 and grain boundaries using crystal plasticity finite element models, Comput. Method Appl. M. 2014, 277,
608 239-259.
609 40. Sun, C.Y.; Guo, X.R.; Guo, N.; Yang, J.; Huang, J. Investigation of plastic deformation behavior on coupling
610 twinning of polycrystal twip steel, Acta Metall.Sin. 2015, 51, 1507-1515.
 Metals 2017, 7, x FOR PEER REVIEW 19 of 19

611 41. Bachmann, F.; Hielscher, R.; Schaeben, H. in: H. Klein, R.A. Schwarzer (Eds.), Texture analysis with
612 MTEX - free and open source software toolbox, Solid State Phenom., 2010,160, 63-68.
613 42. Li, L.; Shen, L.M.; Proust. G. A texture-based representative volume element crystal plasticity model for
614 predicting Bauschinger effect during cyclic loading, Mater. Sci. Eng. A 2014, 608, 174-183.
615 43. Li, H.; Yang, H.; Lu, R.D.; Fu, M.W. Coupled modeling of anisotropy variation and damage evolution for
616 high strength steel tubular materials, Int. J. Mech. Sci. 2016, 105, 41-57.
617 44. Haase, C.; Kühbach, M.; Barrales-Mora, L.A.; Wong, S.L.; Roters, F.; Molodov, D.A.; Gottstein, G.
618 Recrystallization behavior of a high-manganese steel: Experiments and simulations, Acta Mater. 2015, 100,
619 155-168.

620 © 2017 by the authors. Submitted for possible open access publication under the
621 terms and conditions of the Creative Commons Attribution (CC BY) license
622 (http://creativecommons.org/licenses/by/4.0/).