See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/362029171

Crystal plasticity simulation of in-grain microstructural evolution during

large deformation of IF-steel

Article  in  Acta Materialia · July 2022

DOI: 10.1016/j.actamat.2022.118167

CITATIONS READS

0 479

6 authors, including:

Karo Sedighiani Jilt Sietsma

Tata Steel Netherlands Delft University of Technology
17 PUBLICATIONS   233 CITATIONS    507 PUBLICATIONS   12,235 CITATIONS   

SEE PROFILE SEE PROFILE

Dierk Raabe
Max Planck Society
2,674 PUBLICATIONS   70,956 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Machine Learning and Data-Driven Methods for Material Simulation View project

Intermetallic and precipitates in metals and alloys View project

All content following this page was uploaded by Karo Sedighiani on 17 July 2022.

The user has requested enhancement of the downloaded file.

 Crystal plasticity simulation of in-grain microstructural evolution during large deformation of IF-steel

Journal Pre-proof

Crystal plasticity simulation of in-grain microstructural evolution

during large deformation of IF-steel

Karo Sedighiani, Konstantina Traka, Franz Roters, Jilt Sietsma,

Dierk Raabe, Martin Diehl

PII: S1359-6454(22)00548-1
DOI: https://doi.org/10.1016/j.actamat.2022.118167
Reference: AM 118167

To appear in: Acta Materialia

Received date: 2 March 2022

Revised date: 16 June 2022
Accepted date: 11 July 2022

Please cite this article as: Karo Sedighiani, Konstantina Traka, Franz Roters, Jilt Sietsma,
Dierk Raabe, Martin Diehl, Crystal plasticity simulation of in-grain microstructural evolution during large
deformation of IF-steel, Acta Materialia (2022), doi: https://doi.org/10.1016/j.actamat.2022.118167

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.

© 2022 Published by Elsevier Ltd on behalf of Acta Materialia Inc.

aphical Abstract
1 Crystal plasticity simulation of in-grain microstructural evolution during large
2 deformation of IF-steel

3 Karo Sedighiania,b,∗, Konstantina Trakaa,b , Franz Rotersa , Jilt Sietsmab , Dierk Raabea , Martin Diehlc,d
4 a Max-Planck-Institut für Eisenforschung, Max-Planck-Str. 1, 40237 Düsseldorf, Germany

5 b Department of Materials Science and Engineering, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands
6 c Department of Materials Engineering, KU Leuven, Kasteelpark Arenberg 44, 3001 Leuven, Belgium

7 d Department of Computer Science, KU Leuven, Celestijnenlaan 200 A, 3001 Leuven, Belgium

8 Abstract

High-resolution three-dimensional crystal plasticity simulations are used to investigate deformation heterogeneity and

microstructure evolution during cold rolling of interstitial free (IF-) steel. A Fast Fourier Transform (FFT)-based

spectral solver is used to conduct crystal plasticity simulations using a dislocation-density-based crystal plasticity

model. The in-grain texture evolution and misorientation spread are consistent with experimental results obtained

using electron backscatter diffraction (EBSD) experiments. The crystal plasticity simulations show that two types

of strain localization features develop during the large strain deformation of IF-steel. The first type forms band-like

areas with large strain accumulation that appear as river patterns extending across the specimen. In addition to these

river-like patterns, a second type of strain localization with rather sharp and highly localized in-grain shear bands is

identified. These localized features are dependent on the crystallographic orientation of the grain and extend within

a single grain. In addition to the strain localization, the evolution of in-grain orientation gradients, misorientation

features, dislocation density, kernel average misorientation, and stress in major texture components are discussed.
9 Keywords: Crystal plasticity, Microtexture, Shear bands, Dislocation density, Polycrystalline materials, DAMASK

∗ Corresponding author
Email address: k.sedighiani@tudelft.nl (Karo Sedighiani)

Preprint submitted to Elsevier July 15, 2022

10 Contents

11 1 Introduction 3

12 2 Microstructure evolution in steels 4

13 3 Experimental methods 5

14 4 Crystal plasticity simulations 5

15 4.1 Remeshing procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

16 5 Results and discussion 7

17 5.1 Texture evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

18 5.2 Misorientation spread . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

19 5.3 In-grain orientation gradients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

20 5.4 Shear localization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

21 5.5 Effect of multi-step mesh refinement method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

22 5.6 Dislocation density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

23 6 Conclusions 22

2
24 1. Introduction

25 Plastic deformation of polycrystalline metals is significantly heterogeneous in terms of strain, stress, and crys-

26 tal reorientation. Characterization and understanding of these deformation heterogeneities and the associated mi-

27 crostructural evolution during plastic deformation play a critical role in identifying the underlying mechanisms behind

28 many physical phenomena. For instance, a thorough physical understanding of the mechanisms leading to recrys-

29 tallization is not feasible without properly characterizing microstructural evolution during plastic deformation [1–7].

30 Moreover, damage formation, fracture, and failure in metals are also often related to deformation localization and

31 microstructures formed during deformation [8–11]. Besides, the material properties of crystalline materials depend on

32 the microstructures developing during the deformation including crystallographic texture [12, 13]. As a result, over

33 the past years, a considerable amount of attention has been devoted to understand and characterize the intragranular

34 deformation heterogeneity and microstructural evolution during plastic deformation [14–21].

35 Deformation microstructures are mainly quantified using experimental techniques such as scanning [22–26] and

36 transmission [27, 28] electron microscope based techniques. Choi and Jin [22], Choi and Cho [23] used electron

37 backscatter diffraction (EBSD) analysis to examine the orientation dependency of the stored energy of deformed

38 grains in cold-rolled low carbon steels. Allain-Bonasso et al. [25] studied the effect of orientation distributions and

39 grain size on the development of deformation heterogeneities during tensile deformation of interstitial free (IF-)

40 steel by EBSD. Wang et al. [26] investigated shear band formation in cold-rolled Ta-2.5W alloy at various thickness

41 reduction levels using EBSD analysis. Li et al. [27] investigated the microstructural evolution during cold rolling of

42 IF-steel at different thickness reductions using scanning and transmission electron microscope techniques.

43 Despite the extensive amount of knowledge that can be developed from experimental studies, most experimental

44 techniques suffer from several drawbacks: (i) The microstructure and microtexture can be characterized solely based

45 on the specimen surface, and determining the variation of microstructure through the thickness is only possible using

46 destructive or highly advanced, expensive approaches. As a result, the actual neighboring and boundary conditions of

47 the material points cannot be precisely determined. (ii) It is difficult to track the region of interest, especially during

48 large deformation. (iii) A large number of experiments are required to acquire complete knowledge of the deformation

49 history, which is extremely time-consuming and economically inefficient. (iv) The experimental procedures usually

50 contain uncertainty due to sensitivity to sample contamination and random errors.

51 Crystal plasticity simulations [29] have been established as an alternative to model and quantify deformation

52 processes at the microstructural level and the associated complex mechanical fields [30–37]. These models are

53 developed based on physical mechanisms such as glide of dislocations on preferred slip systems and the interaction

54 of dislocations with various defects. Crystal plasticity simulations have been compared with experimental data in

55 several works. For instance, Raabe et al. [38] compared the spatial distribution of the plastic strain from channel-

56 die experiments with crystal plasticity simulations for a sample of pure aluminum with 18 grains after an 8%

57 sample thickness reduction. However, comparisons between simulations and experiments are mostly limited to

3
58 small/moderate strains (strains less than 0.4) [38, 39] or simple microstructures (e.g. bicrystals) [40–43].

59 Problems involving the deformation of solid materials are usually formulated in a Lagrangian context, in which

60 the mesh is attached to the deformable body and deforms with a change in the material’s shape. As a result, the

61 mesh gets distorted due to the heterogeneity of the deformation [44, 45]. When the mesh distortion becomes too

62 large, the simulation fails to converge. At the same time, modeling of in-grain localized deformation features requires

63 a high-resolution crystal plasticity simulation, even up to hundreds of thousands of elements per crystal [44, 46]. A

64 higher simulation resolution allows capturing more detailed localized deformation features, which results in earlier

65 mesh convergence issues.

66 This paper presents a computational study to investigate the evolution of in-grain deformation heterogeneity

67 during cold rolling of IF-steel over a wide range of strains up to 77% thickness reduction. The large-deformation

68 crystal plasticity simulations are conducted using the remeshing technique proposed by Sedighiani et al. [44]. This

69 approach enables conducting high-resolution large-deformation crystal plasticity simulations and overcoming the

70 associated mesh distortion problem due to the strain localization. The in-grain orientation spreads obtained using

71 simulation are compared with two EBSD measurements performed on a cold-rolled IF steel sample after 77% thickness

72 reduction. In addition, the orientation dependency of deformation heterogeneity, strain localization, and dislocation

73 evolution are numerically investigated over a wide range of strains.

74 2. Microstructure evolution in steels

75 In bcc metals, dislocation slip occurs along the h1 1 1i direction, where the slip plane can be {1 1 0}, {1 1 2},

76 and at higher temperatures also {1 2 3} [47, 48]. The lattice rotates due to the shear strain and reaches a preferred

77 orientation based on the loading conditions and the initial orientation of the lattice. There are numerous works

78 devoted to studying the evolution of texture in bcc metals and especially steel [27, 45, 49–56, 56, 57]. The rolling

79 texture in bcc metals belongs mainly to two families of α (RDkh1 1 0i) and γ (NDkh1 1 1i) fibers. In general, the

80 rolling texture developed in bcc metals is noticeably affected by the degree of deformation. By increasing thickness

81 reductions, the main features of the texture remain similar. However, at large strains, the α-fiber becomes stronger

82 and more prominent at the expense of weakening the γ-fiber[51].

83 The evolution of the in-grain microstructure in bcc metals has been a motivation for many experimental works

84 [25, 27, 50, 51, 54, 57]. It was found that crystals with an initial rotated cube orientation {0 0 1}h1 1 0i show minimal

85 deformation heterogeneity [51, 52]. The deformation heterogeneity increases for orientations further along the α-

86 fiber, such as {1 1 2}h1 1 0i, and the hard orientations of {1 1 1}h1 1 0i and {1 1 0}h1 1 0i contain the largest internal

87 misorientations [58]. However, the soft α-fiber orientations also sometimes show texture dispersion when they are

88 located near a hard grain [51, 59]. Generally, a stronger tendency to form in-grain misorientation spread is observed

89 in the γ-fiber compared to the α-fiber.

90 Under plane-strain deformation, a soft grain cannot store energy alone by concentrating the deformation into

4
91 localized shears [60]. On the contrary, the formation of in-grain shear bands for hard grains is energetically more

92 favorable than deforming the grain homogeneously under large stress boundary conditions. Therefore, in-grain shear

93 bands are observed more frequently in hard γ-fiber orientations, whereas intergranular deformation heterogeneities are

94 more common in softer α-fiber orientations [51]. In-grain shear bands are more prevalent in {1 1 1}h1 1 2i orientations

95 [61, 62]. In steel, they are most frequently formed at an angle of approximately 35◦ with the rolling direction

96 [60, 62, 63]. These bands can be explained by their high degree of texture softening [64]. Barnett [62] observed that

97 for {1 1 1}h1 1 2i oriented grains, the orientation inside the in-grain shear bands rotate toward the texture component

98 of {5 5 4}h2 2 5i.

99 3. Experimental methods

100 The material used in the present study is an IF-steel with chemical composition as given in Table 1. The initial

101 microstructure and crystallographic texture of the material before cold rolling were measured on the RD-ND (rolling

102 direction-normal direction) plane perpendicular to the transverse direction (TD) using electron backscatter diffraction

103 (EBSD), see [65] for more information. The grain structure is almost completely equiaxed, and the material exhibits

104 a mild texture commonly observed for hot-rolled IF steel.

Table 1: The chemical composition of the IF-steel considered in this study.

element C Mn S Ti N Al Cr Fe
wt. (%) 0.002 0.095 0.006 0.045 0.002 0.05 0.02 balance

105 We performed two EBSD measurements on two nearby areas of the industrially cold-rolled IF-steel sample sub-

106 jected to 77% thickness reduction (ε = 1.47). The two scanning areas are located at the mid-thickness of the rolling

107 plane (ND-RD plane). Therefore, these areas experience a near plane-strain compression deformation mode, i.e.

108 shear deformation modes are negligible. The first EBSD map of a scan area of 600 µm × 600 µm was measured with

109 a step size of 0.3 µm. The second EBSD map of a scan area of 594 µm × 438 µm was measured with a step size of

110 0.6 µm. Standard metallographic techniques were used to prepare the specimens for characterization. Analysis of

111 the EBSD data was performed using the TSL OIM software.

112 4. Crystal plasticity simulations

113 In the computational example presented in this study, we use a high-resolution RVE consisting of 36 grains to

114 investigate the deformation patterning and misorientation features evolving during large strain deformation. The

115 crystallographic orientations of the grains are sampled from the EBSD map of the undeformed hot-rolled sample

116 (see [65] for more information) using the approach presented by Eisenlohr and Roters [66]. The RVE is subjected to

117 plane-strain compression at a strain rate of 100 s−1 up to a total thickness reduction of 77% (ε = 1.47). The initial

118 number of elements at the beginning of the deformation is 80 × 48 × 320, i.e. around 34 000 elements per grain on

119 average. The number of elements is gradually increased to 1280 × 48 × 320 during the deformation using a multi-step

5
 Table 2: Model parameters of IF-steel used for crystal plasticity simulations.

variable description units value

3
ρα
0 initial mobile dislocation density m/m 1.0 × 1012
3
ρα
d0 initial dipole dislocation density m/m 1.5 × 1012
v0 dislocation glide velocity pre-factor m/s 1.4 × 103
∆F activation energy for dislocation glide J 1.57 × 10−19
p p-exponent in glide velocity – 0.325
q q-exponent in glide velocity – 1.55
τ0∗ short-range barriers strength at 0 K MPa 454
parameter controlling dislocation
Cλ – 50
mean free path
Canni coefficient for dislocation annihilation – 2

120 mesh refinement method as described in Section 4.1. Therefore, at the final stages of the deformation, each grain

121 is discretized using around 550 000 elements. Considering an average initial grain size of 50 µm, the approximate

122 element sizes at the beginning and at the end of the simulation would be 1.6 µm and 0.4 µm, respectively. Such

123 a large number of elements per grain allows for predicting the development of a strain gradient and deformation

124 heterogeneities within the grain.

125 We use a Fast Fourier Transform (FFT) based spectral method [67, 68] implemented in DAMASK [69] to conduct

126 the crystal plasticity simulations using a dislocation-density-based constitutive law [65, 70]. The material parameters

127 used in this study are based on the parameters identified in [65] for IF-steel using the approach presented in [71].

128 However, the hardening-related parameters (Cλ and Canni ) are revised based on the hardening behavior of the

129 material at large strains, see Table 2.

130 4.1. Remeshing procedure

131 We use a remeshing technique presented by Sedighiani et al. [44] to overcome the mesh distortion problem in the

132 high-resolution crystal plasticity simulations. This employed method is based on replacing the distorted mesh with

133 a new undistorted mesh. The variables from the deformed stage are mapped onto the newly created mesh using a

134 nearest-neighbor mapping algorithm. Finally, the simulation is restarted as a new simulation in which the initial

135 state is set based on the last deformation state that had been reached.

136 During large deformation, the elements aspect ratio, i.e. the ratio of the element size in the stretching direction

137 to the element size in the compression direction, can become very large. Extensively elongated elements introduce

138 errors in the simulation and, more importantly, can prevent strain localization. A multi-step mesh refinement method

139 is used for updating the mesh density during the deformation at each remeshing step [44]. The idea behind this

140 approach is to keep the number of elements in the compression direction constant and adjust the number of elements

141 in the stretching direction accordingly to keep the elements close to a cubic shape. The multi-step mesh refinement

142 approach leads to a gradual rise in the number of elements and the simulation resolution during the deformation.

6
143 Keeping the number of elements in the most compressed direction constant ensures a minimum information loss

144 during mapping with a minimal increase in the number of simulation points [44].

145 5. Results and discussion

146 We use a combined simulation and experimental study to investigate the deformation heterogeneity developed

147 within grains at large strains. For this purpose, the results from the large-deformation crystal plasticity simulation

148 with high resolution (Section 4) are compared with the results from two EBSD measurements (see Section 3).

149 5.1. Texture evolution

150 Figure 1a shows the inverse pole figure (IPF) color maps parallel to the loading (vertical) direction in the mid-

151 surface of the 3D simulation for different thickness reductions. The 3D IPF color map for the same RVE after 77%

152 thickness reduction is shown in Figure 1b. The ϕ2 = 45◦ section of the orientation distribution function for the 3D

153 simulation is shown in Figure 1c. Plastic deformation leads to changes in the grains’ orientation and the development

154 of deformation textures. The simulation results show that a strong α-fiber and a slightly weaker γ-fiber are developed

155 after 77% thickness reduction. Figure 3a and Figure 4a show the IPF color map parallel to the loading (vertical)

156 direction for two EBSD measurements on two nearby areas of the industrially cold-rolled IF-steel sample subjected

157 to 77% thickness reduction (see Section 3). The corresponding ϕ2 = 45◦ sections of the orientation distribution

158 function for these two EBSD maps are shown in Figure 3d and Figure 4d, respectively.

159 Despite some differences in the predicted texture, there is a good agreement between the three sets of results.

160 The first EBSD measurement shows almost the same strength for the two fibers, while the second EBSD data shows

161 a strong α-fiber and a much weaker γ-fiber. The main reason for the different observations between the three sets

162 of the results is that neither of the data sets is statistically fully representative of the material microstructure. It is

163 challenging to conduct simulations and experiments at such resolutions in a way that full statistical representation

164 of the material microstructure is obtained. Therefore, statistical differences are expected between the three sets of

165 results, i.e. the two EBSD measurements and the simulation result.

166 5.2. Misorientation spread

167 This section investigates the deformation heterogeneity in terms of misorientation spread developed within the

168 RVE. For this purpose, the data is divided into low KAM (kernel average misorientation) regions, i.e. points whose

169 KAM value belongs to the lower 20% of the data distribution, and high KAM regions, points whose KAM value

170 belongs to the higher half of the data distribution. For calculating the KAM of a point, the disorientation, i.e. the

171 misorientations considering the cubic symmetry of the material, to all first-order neighboring points is calculated.

172 Then, the average value is calculated. The IPF color maps parallel to the loading (vertical) direction for low KAM

173 and high KAM regions are shown respectively in Figures 2b and 2c, and the corresponding orientation density maps

7
 0%

20%

40%

60%

77%

(a)

(b) (c)

Figure 1: (a) IPF color maps parallel to the loading (vertical) direction in the mid-surface of the 3D simulation at different thickness
reductions. (b) The 3D IPF color maps and (c) Orientation density f(g) maps (ϕ2 = 45◦ ) for the same RVE after 77% thickness
reduction.

174 are shown respectively in Figures 2e and 2f. For the sake of comparison, the IPF color map and the corresponding

175 orientation density map of the full mid-surface are shown respectively in Figures 2a and 2d.

176 The tendency to form in-grain misorientation spread is smaller in crystals belonging to the α-fiber. Specifically,

177 grains close to the rotated cube orientation, {0 0 1}h1 1 0i, show minimal deformation heterogeneity, i.e. a very

178 small misorientation spread. This is in accordance with the experimental observations (e.g. [51, 52]) that, for

179 this texture component, the orientation does not change noticeably during the rolling deformation (the in-grain

180 orientation remains below 15◦ misorientation), and a uniform microstructure without significant orientation gradients

181 is formed during deformation. The inverse Brass orientation, {1 1 2}h1 1 0i, shows similar behavior, and grains close

182 to this crystallographic orientation show generally small misorientation spread. However, the in-grain misorientation

183 spread is slightly larger compared to that of the rotated cube component. In general, the deformation heterogeneity

184 increases for orientations further along the α-fiber, such as {1 1 2}h1 1 0i and {1 1 1}h1 1 0i. The crystal plasticity

185 simulation shows a stronger tendency to form in-grain misorientation spread in the γ-fiber compared to the α-fiber.

186 The results from the two EBSD measurements confirm similar behavior, as shown in Figures 3 and 4. Regions

8
 (a) full section

(b) low KAM region

(c) high KAM region

(d) full section (e) low KAM region (f) high KAM region

Figure 2: Simulation results showing the low and high KAM regions in the mid-surface of the 3D simulation after 77% thickness reduction.
(a–c) IPF color maps parallel to the loading (vertical) direction, (d–f) orientation density f(g) maps obtained from the ODF
section ϕ2 = 45◦ .

187 with small deformation heterogeneity and misorientation spread belong mainly to the α-fiber, while the components

188 belonging to the γ-fiber show the strongest misorientation spread. These observations show significant similarities

189 with experimentally reported results in the literature for bcc metals, e.g. [49, 51, 52, 54, 72].

190 5.3. In-grain orientation gradients

191 In this section, we investigate the deformation heterogeneity in terms of in-grain orientation gradients (orientation

192 variations within a grain). We quantify the in-grain deformation heterogeneity using grain orientation spread (GOS)

193 and grain average misorientation (GAM). The GOS is the average of disorientation angles of all points within a grain

194 to the grain mean orientation:

Ni
1 X
GOSi = ωij , (1)
Ni j=1

195 where Ni stands for the number of points belonging to grain i, and ωij is the disorientation angle between point

196 j and grain i ’s mean orientation considering the cubic symmetry of the material. The GAM is the average of the

9
 (a) full section (b) low KAM region (c) high KAM region

(d) full section (e) low KAM region (f) high KAM region

Figure 3: Results showing low and high KAM regions of EBSD measurement 1 for a 600 µm × 600 µm scan area. (a–c) IPF color maps
parallel to the loading (vertical) direction, (d–f) orientation density f(g) maps obtained from the ODF section ϕ2 = 45◦ .

(a) full section (b) low KAM region (c) high KAM region

(d) full section (e) low KAM region (f) high KAM region

Figure 4: Results showing low and high KAM regions of EBSD measurement 2 for a 594 µm × 438 µm scan area. (a–c) IPF color maps
parallel to the loading (vertical) direction, (d–f) orientation density f(g) maps obtained from the ODF section ϕ2 = 45◦ .

10
197 KAM for all points within a grain:

Ni
1 X
GAMi = KAMj , (2)
Ni j=1

198 The GAM parameter can be seen as a local measure for the orientation variation inside the grains. Larger GOS and

199 GAM values reflect a higher degree of plastic deformation heterogeneity.

200 Figure 5a shows the GOS and GAM values for different grains. The grains are defined and numbered based on

201 the initial microstructure before deformation as shown in Figure 5b. The inverse pole figure distribution maps with

202 respect to the loading direction (z ) and stretching direction (x ) for a few selected grains are shown in Figure 6. The

203 large reorientation spread for the grains reveals that in addition to the initial orientation, other factors like spatial

204 constraints of neighboring points also play a crucial role. This is because the resolved shear stresses on various slip

205 planes are determined by the compression load combined with the spatial constraints applied by the neighboring

206 material. In general, the GOS and the GAM values are noticeably smaller in grains belonging to the α-fiber (e.g.

207 grains 18, 29, 31, 32, and 33) than in grains belonging to the γ-fiber (e.g. grains 3, 4, 13, and 35). The smallest

208 values of GOS and GAM belong to grains with orientations close to the rotated cube orientation.

GOS
40 GAM
Magnitude (degree)

30
20
10
0
0 10 20 30
Grain number
(a) (b)

Figure 5: (a) GOS and GAM for all individual grains. (b) Grains are defined and numbered based on the initial microstructure before
deformation.

209 Figure 7 shows the correlation between the GOS and the GAM with volume fraction of various texture compo-

210 nents. There is a strong negative correlation between the volume fraction of the points belonging to the α-fiber with

11
 (a) Grain 18 (b) Grain 33 (c) Grain 13 (d) Grain 2

Figure 6: Inverse pole figure distribution map with respect to the loading direction (z ) and stretching direction (x ) for four selected
grains. (a) A Grain close to the rotated cube orientation with minimal deformation heterogeneity; (b) A grain belonging to the
α-fiber showing a small tendency to form in-grain misorientation spread; (c) A grain belonging to the γ-fiber showing a strong
tendency to form in-grain misorientation spread compared to the α-fiber; (d) A grain fragmented into regions of distinctly
different orientations. The grains are defined based on the initial microstructure before deformation (see Figure 5b), and the
blue stars show the initial orientations.

211 GOS and GAM, which means that a more uniform microstructure without significant orientation gradients is formed

212 in grains with a larger volume fraction of α-fiber orientations. On the other hand, there is a weak positive correlation

213 between the volume fraction of the γ-fiber with GOS and GAM, which implies noticeable variations in the orientation

214 spread for grains with a similar fraction of orientations belonging to the γ-fiber. For example, the GOS & GAM

215 values for grains 6 and 15 are respectively 14.2 & 9.8 and 34.0 & 15.1. This result indicates that local effects and

216 the local stress field considerably affect the reorientation of crystal points rotating towards the γ-fiber. Therefore,

217 altering the grain shape or orientation of a neighboring grain can suppress or promote grain fragmentation in these

218 crystals more strongly than in the α-fiber grains. Between the γ-fiber components, we observed that the volume

219 fraction of {5 5 4}h2 2 5i component has the strongest correlation with the GOS and GAM. On the other hand, the

220 volume fraction of {1 1 1}h1 1 0i has the weakest correlation with the GOS and GAM. These outputs indicate that

221 the in-grain orientation spread is orientation-dependent. The analysis presented here is based on how the initially

222 uniform grains reorient during plastic deformation and form in-grain orientation gradients. It is one of the main

223 advantages of simulations to allow simple tracking of the reorientation history for all points, while such an analysis

224 for experimental methods is often impossible.

225 There is a significant difference between the GOS and GAM values for some grains, e.g. grain 2. A 2D section of

226 grain 2 at the mid-surface is shown in Figure 8. As can be seen, this grain has been fragmented into two regions of

227 distinctly different orientations. As a result, the misorientation angle to the grain mean orientation is noticeably large

228 for most points within the grain. However, since the GAM measures local variation in the orientation, it captures

229 the bifurcation. For such cases, looking at the GOS alone may be misleading.

230 5.4. Shear localization

231 This section investigates the evolution of intergranular and intragranular shear localization during the deformation

232 of the IF-steel over a wide strain range. Figure 9 shows the development of different variables (i.e. equivalent von

233 Mises strain, equivalent von Mises stress, dislocation density, KAM, and Taylor factor) at the mid-surface of the

234 RVE after a 20% thickness reduction (ε = 0.22). Non-crystallographic band-like deformation regions with large strain

235 accumulation that appear as river-like patterns are formed during the deformation, which pass through several grains

12
 Figure 7: Correlation matrix showing the dependence between different pairs of variables. The subplots show a scatter plot of a pair of
variables with a linear regression fit. The displayed values present the correlation coefficient between the GOS and GAM with
volume fraction of various texture components. Each diagonal subplot contains the distribution of a variable as a histogram.

Figure 8: A 2D section showing the IPF color maps parallel to the loading (ND) direction at the mid-surface of grain 2.

236 and extend across the specimen (see Figure 9a). The local strain in these regions is, on average, around two times

237 higher than the applied strain. These macroscopic bands are initially formed at an approximate angle of ±40◦ − 45◦

238 with the rolling direction. The accumulation of the plastic strain in these macroscopic river-like patterns results in

239 a significantly lower plastic strain in the neighboring regions.

240 One of the factors that promote such macroscopic strain localization is the mechanical contrast in the plastic

241 behavior among neighboring grains (e.g. yield stress). Strain localization typically initiates near grain boundaries

242 of soft grains with low strain hardening rates. The formation of an area plastically deformed decreases the effective

243 load capacity of the sheared section and results in local mechanical instability. Consequently, the strain localization

244 propagates through the specimen and creates the river-like patterns. The sharpness of the river-like patterns is

245 higher when the contrast in the deformation behavior of the grains is larger (see region A1). However, a cluster

13
 (a) εvM (b) σvM (MPa) (c) log(ρ) (m−2 ) (d) KAM (degree) (e) Taylor factor

Figure 9: Simulation results at the mid-surface after a 20% thickness reduction; (a) equivalent strain εvM , (b) equivalent stress σvM (MPa),
(c) log(ρ) (m−2 ), (d) KAM (degree), (e) Taylor factor with respect to the loading (vertical) direction.

246 of soft grains close to the areas of strain localization results in the broadening and homogenizing of the river-like

247 patterns (see region A2). The high deformation concentration at these bands leads to a significantly lower strain in

248 the surrounding regions. For example, for the grain shown in region A1, the strain in the rest of the soft grain is

249 notably small.

250 In addition to the river-like patterns which pass through several grains, a small number of rather sharp and highly

251 localized in-grain shear bands can also be observed. These bands extend inside a single grain, and it seems that they

252 originate at the grain boundaries (see region B1). They exist only in a very limited number of grains at this strain

253 level and have developed mainly in hard grains, interrupting the macroscopic river-like patterns. They are formed

254 at an angle of approximately ±30◦ − 35◦ with the rolling direction.

255 The stress values are generally higher in harder grains, i.e. orientations with higher Taylor factor. However, the

256 macroscopic river-like patterns affect the stress distribution inside soft and hard grains considerably. The regions

257 with high stress are mainly located inside hard grains next to the river-like patterns. The in-grain deformation also

258 influences the stress distribution inside the grain. In general, the average stress inside grains with such features is

259 notably higher than the total average stress.

260 With an increase of thickness reduction to 40% (ε = 0.51), the river-like patterns rotate down to inclination

261 angles of ±30◦ − 35◦ to the rolling direction (Figure 10). In addition, the contrast between the strain in areas with

262 large strain accumulation and the surrounding regions increases, which results in sharper river-like patterns. On the

263 other hand, the in-grain shear bands are more clearly visible, and they develop in a larger number of grains. These

264 bands still have an angle of around ±30◦ − 35◦ with the rolling direction, so unlike the river-like patterns, they do

265 not rotate notably towards the rolling direction. The grains that contain such microstructure features pertain mostly

266 to the γ-fiber.

267 After 60% thickness reduction (ε = 0.92), the river-like patterns rotate significantly towards inclination angles of

14
 (a) εvM (b) σvM (MPa) (c) log(ρ) (m−2 ) (d) KAM (degree)

Figure 10: Simulation results at the mid-surface after a 40% thickness reduction; (a) equivalent strain εvM , (b) equivalent stress
σvM (MPa), (c) log(ρ) (m−2 ), (d) KAM (degree).

(a) εvM (b) σvM (MPa)

(c) log(ρ) (m−2 ) (d) KAM (degree)

Figure 11: Simulation results at the mid-surface after a 60% thickness reduction; (a) equivalent strain εvM , (b) equivalent stress
σvM (MPa), (c) log(ρ) (m−2 ), (d) KAM (degree).

268 ±15◦ − 20◦ to the rolling direction (Figure 11). On the other hand, the angle for the in-grain shear bands is still

269 around ±30◦ − 35◦ , revealing that the inclination of in-grain shear bands is not noticeably dependent on the applied

270 strain. After 60% thickness reduction, the in-grain shear bands are well developed and established in some grains

271 that mainly belong to the γ-fiber.

272 After 77% thickness reduction (ε = 1.47), the whole microstructure is severely deformed, and it is difficult to

273 trace the original grain boundaries (Figure 12). However, it is still possible to identify and trace some of the features

15
274 previously observed at lower strains. The river-like patterns rotate further towards the rolling direction. In some

275 places, they rotate to inclination angles of ±5◦ − 10◦ to the rolling direction, while in some other places, they are

276 almost parallel to the rolling direction. The localized strain in the river-like patterns can reach a value of 4.0, around

277 three times higher than the applied strain.

278 The in-grain shear bands also rotate slightly, and their inclination angle with the rolling direction rotates to

279 around ±25◦ − 30◦ . However, bands with an angle of around ±15◦ are also observed. The highly rotated in-grain

280 shear bands are mainly those developed at the early stages of deformation, i.e. those already visible at 20% thickness

281 reduction. Moreover, the fraction of in-grain shear bands has increased notably with increasing deformation. The

282 in-grain shear bands are orientation-dependent, i.e. they form preferentially in γ-fiber (high Taylor factor) grains.

283 However, not all grains belonging to the γ-fiber contain such features. Similar experimental observations have been

284 reported for steels [50, 55, 73].

285 A 3D view of the river-like patterns after 60% and 77% thickness reductions is shown in Figure 13a and Figure 13c,

286 respectively. These patterns are aligned parallel to the transverse direction and extended through the depth of the

287 RVE. Since these macroscopic band-like areas pass through several grains and extend across the specimen, they are

288 comprised of regions belonging to both, the α-fiber and γ-fiber. The regions belonging to the γ-fiber are areas with

289 relatively high local misorientations and stress. On the contrary, regions belonging to the α-fiber deform relatively

290 homogeneous, and their stress value is relatively low. Therefore, the orientations belonging to the γ-fiber have

291 significantly higher stored energy and misorientation than the α-fiber orientations. This facilitates the growth of

292 recrystallized volumes of γ-fiber orientations, i.e. these areas act as successful nucleation sites for recrystallization.

293 This leads to the well-known γ-fiber recrystallization texture appearing after annealing of the cold-rolled IF-steel

294 sheet [6, 51, 54].

295 Figure 14 shows the in-grain shear bands developed in a selected grain, i.e. grain number 24 (see Figure 5b). It can

296 be seen that the orientation inside the in-grain shear bands is displaced from the matrix orientation of {1 1 1}h1 1 2i

297 toward the orientation of {5 5 4}h2 2 5i. This supports the role of in-grain shear bands in the nucleation of the

298 {5 5 4}h2 2 5i recrystallization texture appearing after annealing of the cold-rolled IF-steel sheet [54, 62].

299 Figure 15a shows the evolution of the average equivalent strain with deformation in major texture components.

300 The mean equivalent strain is larger in grains with smaller Taylor factors, i.e. orientations close to the rotated cube

301 component. The mean equivalent strain decreases in orientations belonging to the γ-fiber, i.e. orientations with

302 larger Taylor factors.

303 Figure 15b compares the cumulative distributions of equivalent strain (the probability of strain being below a

304 certain value) in major texture components. The figure illustrates the data for all elements after 77% thickness

305 reduction. In addition to the notable difference in the median values, there is a significant difference between the

306 distribution range for different texture components. The median value of the distributions decreases with an increase

307 in the orientation’s hardness. The lower bound of the distributions also follows almost the same trend as the median

16
 (a) εvM

(b) σvM (MPa)

(c) log(ρ) (m−2 )

(d) KAM (degree)

Figure 12: Simulation results at the mid-surface after a 77% thickness reduction; (a) equivalent strain εvM , (b) equivalent stress
σvM (MPa), (c) log(ρ) (m−2 ), (d) KAM (degree).

308 values. However, the upper bounds are not orientation-dependent; it is almost the same for all orientations. This

309 leads to a narrower distribution for the soft grains, i.e. orientations close to the rotated cube component. The

310 distribution becomes wider for orientations further along the α-fiber, and it becomes noticeably wide for hard grains

311 belonging to the γ-fiber. The widest distributions are found for the {1 1 1}h1 1 2i and {1 1 1}h1 2 3i components. This

312 wide distribution is due to the formation of in-grain shear bands in γ-fiber grains. The in-grain shear bands are areas

313 with large strain accumulation. The presence of these bands serves to redistribute plastic strain and significantly

314 reduces the extent of the strain in regions next to these bands, resulting in a large contrast in the strain distribution

315 in these grains.

17
 ND
RD

TD
(a) (b)

(c) (d)

Figure 13: IPF color maps parallel to the loading (vertical) direction for the river-like patterns after (a) 60% thickness reduction and (c)
77% thickness reductions. The corresponding orientation density f(g) maps (ϕ2 = 45◦ ) after (b) 60% thickness reduction and
(d) 77% thickness reductions.

316 5.5. Effect of multi-step mesh refinement method

317 This study uses a multi-step mesh refinement method to adjust the number of elements during the simulation.

318 To further investigate the advantages of this meshing strategy, we have conducted an additional simulation using

319 the same RVE investigated in this section (see Figure 1a). However, the simulation is performed without using the

320 multi-step mesh refinement method, i.e. the mesh density is kept constant as 80 × 48 × 320. Figure 16 shows the

321 results after 77% thickness reduction. During the simulation, the element size in the loading direction decreases

322 with increasing the deformation, while the element is elongated in the stretching (rolling) direction. Therefore, the

323 element becomes rectangular, and the element aspect ratio, the ratio of the element size in the stretching direction

324 to the element size in the compression direction, considerably increases. For this example, the elements’ aspect

325 ratio reaches a value around 19 after 77% thickness reductions. An element represents the average response of a

326 section in the discretized space. When the element size is larger than the localized features, it will capture only the

327 homogenized behavior and not the morphology of the localized deformation. Therefore, these elongated elements

18
 (a)

(b)

(c)

(d) (e)

Figure 14: A 2D section of grain 24 showing IPF color maps parallel to the loading (vertical) direction for (a) the full cross-section, (b)
the in-grain shear bands. (c) Strain distribution at the in-grain shear bands. Orientation density f(g) maps obtained from the
ODF section ϕ2 = 45◦ for the (d) grain matrix and (e) in-grain shear bands.

Thickness reduction (%)

50 70 77
2.00 {001} 110 1.0
{114} 110
1.75 {112} 110
Average equivalent strain

Cumulative probability

{111} 110 0.8

1.50 {111} 112
{111} 123 {100} 011
1.25 0.6 {114} 011
{112} 011
1.00 {111} 011
0.4 {111} 112
0.75 {111} 132
0.50 0.2
0.25
0.0
0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Strain Equivalent strain
(a) εvM (b)

Figure 15: (a) Evolution of the average equivalent strain with deformation in major texture components. (b) Cumulative distribution of
equivalent strain in major texture components after 77% thickness reduction.

328 prevent strain localization from occurring. This behavior can be seen clearly in Figure 16b, where these highly

329 elongated elements substantially have reduced the level of strain localization. Although the simulation can still

330 capture the global shape of the river-like patterns, it is no longer possible to recognize the details inside and close to

331 these patterns. In addition, it is no longer possible to identify the in-grain shear bands. One should note that the

332 resolution of this simulation (i.e. approximately 3400 elements per grain) is still relatively high compared to typical

19
 (a) εvM

(b) εvM

Figure 16: Simulation results at the mid-surface after 77% thickness reduction; (a) equivalent strain εvM , (b) equivalent stress σvM (MPa).
The simulation is conducted without using the multi-step mesh refinement method and the mesh density is kept constant as
80 × 48 × 320. Compare with Figure 1a and Figure 12a for results including mesh refinement.

333 crystal plasticity simulations. Hence, the use of the multi-step mesh refinement approach is critical for capturing the

334 in-grain microstructure evolution.

335 5.6. Dislocation density

336 Orientation dependency of dislocation density. Figure 17 shows the IPF color map parallel to the loading (vertical)

337 direction at the mid-surface. The data are divided to low dislocation density regions, i.e. points whose dislocation

338 densities belong to the lower 20% of the data distribution, and high dislocation density regions, points whose dis-

339 location densities belong to the higher half of the data distribution. Grains close to the rotated cube component

340 generally have the lowest dislocation density. The dislocation density increases for orientations further along the

341 α-fiber, similar to what was previously observed for KAM. On the other hand, regions that belong to the γ-fiber have

342 the highest dislocation density.

343 Figure 18 shows the development of the dislocation density with deformation for major texture components. At

344 20% thickness reduction, the difference between the dislocation density of different orientations is relatively small.

345 At larger strains, the difference between the dislocation density of γ-fiber and α-fiber components becomes more

346 pronounced. After 77% thickness reduction, the average dislocation density of γ-fiber and α-fiber are 1.45 × 1015

347 and 1.6 × 1015 m−2 , respectively. It should be noted that {1 1 1}h1 1 0i is a common component between γ-fiber and

348 α-fiber. A similar trend for geometrically necessary dislocations (GND) has been reported experimentally, e.g. [55].

349 Correlation between dislocation density with KAM, εvM , and σvM . It is generally accepted that GNDs are related

350 to the in-grain misorientation. Although the crystal plasticity model used here does not differentiate between the

351 GNDs and SSDs (Statistically Stored Dislocations), it is still expected that the total dislocation density correlates

352 with the local misorientation and KAM. Figure 19a shows the correlation between the dislocation density with other

353 main variables at different strain levels. There is a small positive correlation between the dislocation density and

20
 (a) low dislocation density regions

(b) high dislocation density regions

(c) low dislocation density regions (d) high dislocation density regions

Figure 17: Simulation results showing low and high dislocation density regions. (a,b) IPF color maps parallel to the loading (vertical)
direction, (c,d) orientation density f(g) maps obtained from the ODF section ϕ2 = 45◦ .

Thickness reduction (%)

50 70 77
Average dislocation density (×1014 m 2)

16 {001} 110
{114} 110
{112} 110
14 {111} 110
{111} 112
12 {111} 123
10

2
0.2 0.4 0.6 0.8 1.0 1.2 1.4
Strain
Figure 18: Average dislocation density in major texture components.

354 KAM at small strains (0.21). The correlation coefficient increases rapidly, and it reaches a significant value of 0.60

355 after a 77% thickness reduction. The high correlation value indicates that the dislocation density and the KAM are

356 strongly correlated at large deformation. Figure 19b shows a 2D probability density plot of the dislocation density

357 and the KAM after 77% thickness reduction.

358 The correlation between the dislocation density and the equivalent strain follows an opposite trend. There is a

21
 Thickness reduction (%) 15.8
20 50 70 77 0.25

0.6 15.6 0.20

Correlation coefficient

Log (m 2)
0.5 15.4 0.15
KAM
0.4 vM 15.2
vM 0.10
0.3
15.0
0.05
0.2
14.8
0.1
0.2 0.4 0.6 0.8 1.0 1.2 1.4 0 10 20 30 40 50
Strain KAM (degree)
(a) (b)

Figure 19: (a) Correlation between the dislocation density and main variables. (b) 2D probability density plot of the dislocation density
and KAM. The blue line shows the linear regression line.

359 strong positive correlation between these two variables at small strains (0.66). However, the correlation coefficient

360 substantially decreases with increasing deformation, and it reduces to 0.13 after 77% thickness reduction. There is a

361 moderate correlation between the equivalent stress and the dislocation density at small strains (0.42). The correlation

362 coefficient steadily increases to 0.58 at 77% thickness reduction.

363 6. Conclusions

364 High-resolution 3D crystal plasticity simulations are used to investigate the in-grain microstructural evolution

365 and deformation heterogeneity during large deformation of IF-steel. An RVE consisting of 36 grains is subjected to

366 plane-strain compression up to a thickness reduction of 77%. At this deformation level, around 550 000 elements on

367 average are used for the discretization of each grain. The results reveal that the crystallographic orientation of a

368 grain is a critical factor in determining whether a grain deforms relatively homogeneously or heterogeneously. The

369 orientations close to the rotated cube orientation show the minimal tendency to form orientation gradients. The

370 deformation heterogeneity increases for orientations further along the α-fiber. On the other hand, the orientations

371 belonging to the γ-fiber show the strongest misorientation spread.

372 We observe that band-like areas with severe strain localization extending across the specimen (which appear

373 as river patterns) are formed during deformation. These river-like patterns are initially formed at an approximate

374 angle of ±40◦ − 45◦ with the rolling direction. With increasing the thickness reduction, the river-like patterns

375 rotate significantly to inclination angles almost parallel to the rolling direction. The localized strain in the river-like

376 patterns can reach a value that is three times the overall applied strain value. In addition to the macroscopic river-

377 like patterns, several regions with rather sharp and highly localized in-grain shear bands are identified. These bands

378 extend inside only one grain. The in-grain shear bands are formed at an angle of approximately ±30◦ − 35◦ with

379 the rolling direction, and their inclination angle is not noticeably dependent on the applied strain for a wide range of

22
380 deformation. The in-grain shear bands are orientation-dependent, and they form mainly in grains belonging to the

381 γ-fiber. However, not all grains belonging to the γ-fiber contain such features.

382 The results reveal that the difference between the dislocation density of different orientations is relatively small at

383 small strains. However, with increasing strain, the dislocation density in the γ-fiber becomes notably larger than in

384 the α-fiber. Although dislocation density and KAM are weakly correlated at small strains, the correlation coefficient

385 increases rapidly with increasing strain. The correlation between the dislocation density and the equivalent strain

386 follows an opposite trend.

387 Acknowledgments

388 This research was carried out under project number S41.5.15572a in the framework of the Partnership Program
389 of the Materials innovation institute M2i (www.m2i.nl) and the Technology Foundation TTW (www.stw.nl), which
390 is part of the Netherlands Organization for Scientific Research (www.nwo.nl). The industrially cold-rolled IF-steel
391 sample have been provided by Tata Steel Ijmuiden, The Netherlands. The authors are also grateful to Prof. Leo
392 Kestens for helpful discussions.

393 References

394 References

395 [1] L. Chen, J. Chen, R. A. Lebensohn, Y. Z. Ji, T. W. Heo, S. Bhattacharyya, K. Chang, S. Mathaudhu, Z. K. Liu,
396 and L. Q. Chen. An integrated fast Fourier transform-based phase-field and crystal plasticity approach to model
397 recrystallization of three dimensional polycrystals. Computer Methods in Applied Mechanics and Engineering,
398 285:829–848, mar 2015. ISSN 00457825. doi: 10.1016/j.cma.2014.12.007.

399 [2] P. Zhao, T. Song En Low, Y. Wang, and S.R. Niezgoda. An integrated full-field model of concurrent plastic
400 deformation and microstructure evolution: Application to 3D simulation of dynamic recrystallization in polycrys-
401 talline copper. International Journal of Plasticity, 80, 2016. ISSN 07496419. doi: 10.1016/j.ijplas.2015.12.010.
402 [3] D.K. Kim, W. Woo, W.W. Park, Y.T. Im, and A. Rollett. Mesoscopic coupled modeling of texture formation
403 during recrystallization considering stored energy decomposition. Computational Materials Science, 129, 2017.
404 ISSN 09270256. doi: 10.1016/j.commatsci.2016.11.048.
405 [4] K. K. Alaneme and E. A. Okotete. Recrystallization mechanisms and microstructure development in emerging
406 metallic materials: A review, mar 2019. ISSN 24682179.
407 [5] M. Diehl and M. Kühbach. Coupled experimental-computational analysis of primary static recrystallization in
408 low carbon steel. Modelling and Simulation in Materials Science and Engineering, 28(1), 2020. ISSN 1361651X.
409 doi: 10.1088/1361-651X/ab51bd.
410 [6] K. Traka, K. Sedighiani, C. Bos, J. Galan Lopez, K. Angenendt, D. Raabe, and J. Sietsma. Topological aspects
411 responsible for recrystallization evolution in an IF-steel sheet Investigation with cellular-automaton simulations.
412 Computational Materials Science, 198:110643, oct 2021. ISSN 09270256. doi: 10.1016/j.commatsci.2021.110643.

413 [7] V. Shah, K. Sedighiani, J.S. Van Dokkum, C. Bos, F. Roters, and M. Diehl. Coupling crystal plasticity and
414 cellular automaton models to study meta-dynamic recrystallization during hot rolling at high strain rates.
415 Materials Science and Engineering: A, 849:143471, aug 2022. ISSN 09215093. doi: 10.1016/j.msea.2022.143471.
416 [8] C. C. Tasan, J. P M Hoefnagels, M. Diehl, D. Yan, F. Roters, and D. Raabe. Strain localization and damage
417 in dual phase steels investigated by coupled in-situ deformation experiments and crystal plasticity simulations.
418 International Journal of Plasticity, 63:198–210, 2014. ISSN 07496419. doi: 10.1016/j.ijplas.2014.06.004.
419 [9] M. Diehl, M. Wicke, P. Shanthraj, F. Roters, A. Brueckner-Foit, and D. Raabe. Coupled Crystal PlasticityPhase
420 Field Fracture Simulation Study on Damage Evolution Around a Void: Pore Shape Versus Crystallographic
421 Orientation. Jom, 69(5):872–878, 2017. ISSN 15431851. doi: 10.1007/s11837-017-2308-8.

23
422 [10] H. Hallberg, S. K. Ås, and B. Skallerud. Crystal plasticity modeling of microstructure influence on fatigue crack
423 initiation in extruded Al6082-T6 with surface irregularities. International Journal of Fatigue, 111, 2018. ISSN
424 01421123. doi: 10.1016/j.ijfatigue.2018.01.025.
425 [11] H. S. Oh, K. Biggs, O. Güvenç, H. Ghassemi-Armaki, N. Pottore, and C. C. Tasan. In-situ investigation of
426 strain partitioning and microstructural strain path development up to and beyond necking. Acta Materialia,
427 215, 2021. ISSN 13596454. doi: 10.1016/j.actamat.2021.117023.
428 [12] E. K. Cerreta, I. J. Frank, G. T. Gray, C. P. Trujillo, D. A. Korzekwa, and L. M. Dougherty. The influence of
429 microstructure on the mechanical response of copper in shear. Materials Science and Engineering A, 501(1-2),
430 2009. ISSN 09215093. doi: 10.1016/j.msea.2008.10.029.
431 [13] G. Zhou, M. K. Jain, P. Wu, Y. Shao, D. Li, and Y. Peng. Experiment and crystal plasticity analysis on plastic
432 deformation of AZ31B Mg alloy sheet under intermediate temperatures: How deformation mechanisms evolve.
433 International Journal of Plasticity, 79, 2016. ISSN 07496419. doi: 10.1016/j.ijplas.2015.12.006.
434 [14] A. J. Beaudoin, H. Mecking, and U. F. Kocks. Development of localized orientation gradients in fcc polycrystals.
435 Philosophical Magazine A: Physics of Condensed Matter, Structure, Defects and Mechanical Properties, 73(6),
436 1996. ISSN 01418610. doi: 10.1080/01418619608242998.
437 [15] M. Sachtleber, Z. Zhao, and D. Raabe. Experimental investigation of plastic grain interaction. Materials Science
438 and Engineering A, 336(1-2):81–87, 2002. ISSN 09215093. doi: 10.1016/S0921-5093(01)01974-8.
439 [16] D. Raabe, Z. Zhao, and F. Roters. Study on the orientational stability of cube-oriented FCC crystals under
440 plane strain by use of a texture component crystal plasticity finite element method. Scripta Materialia, 50(7),
441 2004. ISSN 13596462. doi: 10.1016/j.scriptamat.2003.11.061.
442 [17] K. S. Cheong and E. P. Busso. Effects of lattice misorientations on strain heterogeneities in FCC polycrystals.
443 Journal of the Mechanics and Physics of Solids, 54(4), 2006. ISSN 00225096. doi: 10.1016/j.jmps.2005.11.003.
444 [18] C. Zhang, H. Li, P. Eisenlohr, W. Liu, C. J. Boehlert, M. A. Crimp, and T. R. Bieler. Effect of realistic 3D
445 microstructure in crystal plasticity finite element analysis of polycrystalline Ti-5Al-2.5Sn. International Journal
446 of Plasticity, 69, 2015. ISSN 07496419. doi: 10.1016/j.ijplas.2015.01.003.
447 [19] J. Oddershede, J. P. Wright, A. Beaudoin, and G. Winther. Deformation-induced orientation spread in individual
448 bulk grains of an interstitial-free steel. Acta Materialia, 85, 2015. ISSN 13596454. doi: 10.1016/j.actamat.2014.
449 11.038.
450 [20] D. Wang, M. Diehl, F. Roters, and D. Raabe. On the role of the collinear dislocation interaction in deformation
451 patterning and laminate formation in single crystal plasticity. Mechanics of Materials, 125, 2018. ISSN 01676636.
452 doi: 10.1016/j.mechmat.2018.06.007.
453 [21] J. Cappola, J. C. Stinville, M.A. Charpagne, P. G. Callahan, M. P. Echlin, T. M. Pollock, A. Pilchak, and
454 M. Kasemer. On the Localization of Plastic Strain in Microtextured Regions of Ti-6Al-4V. Acta Materialia,
455 204:116492, feb 2021. ISSN 13596454. doi: 10.1016/j.actamat.2020.116492.
456 [22] S. H. Choi and Y. S. Jin. Evaluation of stored energy in cold-rolled steels from EBSD data. Materials Science
457 and Engineering A, 371(1-2), 2004. ISSN 09215093. doi: 10.1016/j.msea.2003.11.034.
458 [23] S. H. Choi and J. H. Cho. Primary recrystallization modelling for interstitial free steels. Materials Science and
459 Engineering A, 405(1-2), 2005. ISSN 09215093. doi: 10.1016/j.msea.2005.05.093.
460 [24] S. Zaefferer, N.-N. Elhami, and P. Konijnenberg. Electron backscatter diffraction (EBSD) techniques for studying
461 phase transformations in steels. Phase Transformations in Steels, pages 557–587, jan 2012. doi: 10.1533/
462 9780857096111.4.557.
463 [25] N. Allain-Bonasso, F. Wagner, S. Berbenni, and D. P. Field. A study of the heterogeneity of plastic deformation
464 in IF steel by EBSD. Materials Science and Engineering A, 548, 2012. ISSN 09215093. doi: 10.1016/j.msea.
465 2012.03.068.
466 [26] S. Wang, Z. H. Wu, C. Chen, S. K. Feng, R. Liu, B. Liao, Z. H. Zhong, P. Lu, M. P. Wang, P. Li, Jan W.
467 Coenen, L. F. Cao, and Y. C. Wu. The evolution of shear bands in Ta-2.5W alloy during cold rolling. Materials
468 Science and Engineering A, 726, 2018. ISSN 09215093. doi: 10.1016/j.msea.2018.04.059.

24
469 [27] B. L. Li, A. Godfrey, Q. C. Meng, Q. Liu, and N. Hansen. Microstructural evolution of IF-steel during cold
470 rolling. Acta Materialia, 52(4):1069–1081, 2004. ISSN 13596454. doi: 10.1016/j.actamat.2003.10.040.
471 [28] D. A. Hughes and N. Hansen. The microstructural origin of work hardening stages. Acta Materialia, 148, 2018.
472 ISSN 13596454. doi: 10.1016/j.actamat.2018.02.002.
473 [29] F. Roters, P. Eisenlohr, L. Hantcherli, D. D. Tjahjanto, T. R. Bieler, and D. Raabe. Overview of constitutive
474 laws, kinematics, homogenization and multiscale methods in crystal plasticity finite-element modeling: Theory,
475 experiments, applications. Acta Materialia, 58(4):1152–1211, 2010. ISSN 13596454. doi: 10.1016/j.actamat.
476 2009.10.058.
477 [30] C. Reuber, P. Eisenlohr, F. Roters, and D. Raabe. Dislocation density distribution around an indent in single-
478 crystalline nickel : Comparing nonlocal crystal plasticity finite-element predictions with experiments. Acta
479 Materialia, 71:333–348, 2014. ISSN 1359-6454. doi: 10.1016/j.actamat.2014.03.012.
480 [31] M. Khadyko, S. Dumoulin, and O. S. Hopperstad. Texture gradients and strain localisation in extruded alu-
481 minium profile. International Journal of Solids and Structures, 97 98:239–255, oct 2016. ISSN 00207683. doi:
482 10.1016/j.ijsolstr.2016.07.024.
483 [32] M. Diehl, D. An, P. Shanthraj, S. Zaefferer, F. Roters, and D. Raabe. Crystal plasticity study on stress and
484 strain partitioning in a measured 3D dual phase steel microstructure. Physical Mesomechanics, 20(3), 2017.
485 ISSN 19905424. doi: 10.1134/S1029959917030079.
486 [33] M. Diehl, M. Groeber, C. Haase, D. A. Molodov, F. Roters, and D. Raabe. Identifying StructureProperty
487 Relationships Through DREAM.3D Representative Volume Elements and DAMASK Crystal Plasticity Simula-
488 tions: An Integrated Computational Materials Engineering Approach. JOM, 69(5), 2017. ISSN 15431851. doi:
489 10.1007/s11837-017-2303-0.
490 [34] H. Zhang, J. Liu, D. Sui, Z. Cui, and M. W. Fu. Study of microstructural grain and geometric size effects
491 on plastic heterogeneities at grain-level by using crystal plasticity modeling with high-fidelity representative
492 microstructures. International Journal of Plasticity, 100, 2018. ISSN 07496419. doi: 10.1016/j.ijplas.2017.09.011.
493 [35] M. Diehl, J. Niehuesbernd, and E. Bruder. Quantifying the contribution of crystallographic texture and grain
494 morphology on the elastic and plastic anisotropy of bcc steel. Metals, 9(12), 2019. ISSN 20754701. doi:
495 10.3390/met9121252.
496 [36] T. F.W. van Nuland, J. A.W. van Dommelen, and M. G.D. Geers. Microstructural modeling of anisotropic
497 plasticity in large scale additively manufactured 316L stainless steel. Mechanics of Materials, 153, 2021. ISSN
498 01676636. doi: 10.1016/j.mechmat.2020.103664.
499 [37] J. L. Dequiedt and C. Denoual. Localization of plastic deformation in stretching sheets with a crystal plasticity
500 approach: Competition between weakest link and instable mode controlled process. International Journal of
501 Solids and Structures, 210-211, 2021. ISSN 00207683. doi: 10.1016/j.ijsolstr.2020.11.021.
502 [38] D. Raabe, M. Sachtleber, Z. Zhao, F. Roters, and S. Zaefferer. Micromechanical and macromechanical effects in
503 grain scale polycrystal plasticity experimentation and simulation. Acta Materialia, 49(17):3433–3441, oct 2001.
504 ISSN 13596454. doi: 10.1016/S1359-6454(01)00242-7.
505 [39] K. Thool, A. Patra, D. Fullwood, K.V.M. Krishna, D. Srivastava, and I. Samajdar. The role of crystallographic
506 orientations on heterogeneous deformation in a zirconium alloy: A combined experimental and modeling study.
507 International Journal of Plasticity, 133, 2020. ISSN 07496419. doi: 10.1016/j.ijplas.2020.102785.
508 [40] S. Zaefferer, J. C. Kuo, Z. Zhao, M. Winning, and D. Raabe. On the influence of the grain boundary misorien-
509 tation on the plastic deformation of aluminum bicrystals. Acta Materialia, 51(16), 2003. ISSN 13596454. doi:
510 10.1016/S1359-6454(03)00259-3.
511 [41] A Ma, F Roters, and D Raabe. A dislocation density based constitutive model for crystal plasticity FEM
512 including geometrically necessary dislocations. Acta Materialia, 54(8):2169–2179, may 2006. ISSN 13596454.
513 doi: 10.1016/j.actamat.2006.01.005.
514 [42] A. Ma, F. Roters, and D. Raabe. On the consideration of interactions between dislocations and grain boundaries
515 in crystal plasticity finite element modeling Theory, experiments, and simulations. Acta Materialia, 54(8):2181–
516 2194, may 2006. ISSN 13596454. doi: 10.1016/j.actamat.2006.01.004.

25
517 [43] C. Rehrl, B. Völker, S. Kleber, T. Antretter, and R. Pippan. Crystal orientation changes: A comparison between
518 a crystal plasticity finite element study and experimental results. Acta Materialia, 60(5), 2012. ISSN 13596454.
519 doi: 10.1016/j.actamat.2011.12.052.
520 [44] K. Sedighiani, V. Shah, K. Traka, M. Diehl, F. Roters, J. Sietsma, and D. Raabe. Large-deformation crystal
521 plasticity simulation of microstructure and microtexture evolution through adaptive remeshing. International
522 Journal of Plasticity, 146(May):103078, nov 2021. ISSN 07496419. doi: 10.1016/j.ijplas.2021.103078.
523 [45] D. K. Kim, J. M. Kim, W. W. Park, H.W. Lee, Y. T. Im, and Y. S. Lee. Three-dimensional crystal plasticity
524 finite element analysis of microstructure and texture evolution during channel die compression of if steel. Com-
525 putational Materials Science, 100(PA):52–60, apr 2015. ISSN 09270256. doi: 10.1016/j.commatsci.2014.09.032.

526 [46] H. Lim, C. C. Battaile, J. E. Bishop, and J. W. Foulk. Investigating mesh sensitivity and polycrystalline RVEs
527 in crystal plasticity finite element simulations. International Journal of Plasticity, 121:101–115, oct 2019. ISSN
528 07496419. doi: 10.1016/j.ijplas.2019.06.001.
529 [47] M. Rhee, H. M. Zbib, J. P. Hirth, H. Huang, and T. De La Rubia. Models for long-/short-range interactions and
530 cross slip in 3D dislocation simulation of BCC single crystals. Modelling and Simulation in Materials Science
531 and Engineering, 6(4), 1998. ISSN 09650393. doi: 10.1088/0965-0393/6/4/012.
532 [48] A. Patra and D.L. McDowell. Crystal plasticity-based constitutive modelling of irradiated bcc structures.
533 Philosophical Magazine, 92(7):861–887, mar 2012. ISSN 1478-6435. doi: 10.1080/14786435.2011.634855.
534 [49] D. Raabe and K. Luecke. Rolling and annealing textures of bcc metals. Materials Science Forum, 157-6(pt 1),
535 1994. ISSN 02555476. doi: 10.4028/www.scientific.net/msf.157-162.597.

536 [50] M. R. Barnett and J. J. Jonas. Influence of ferrite rolling temperature on microstructure and texture in deformed
537 low C and IF steels. ISIJ International, 37(7), 1997. ISSN 09151559. doi: 10.2355/isijinternational.37.697.
538 [51] B. Hutchinson. Deformation microstructures and textures in steels. Philosophical Transactions of the Royal
539 Society A: Mathematical, Physical and Engineering Sciences, 357(1756):1471–1485, 1999. ISSN 1364503X. doi:
540 10.1098/rsta.1999.0385.
541 [52] D. Raabe, Z. Zhao, S. J. Park, and F. Roters. Theory of orientation gradients in plastically strained crystals.
542 Acta Materialia, 50(2), 2002. ISSN 13596454. doi: 10.1016/S1359-6454(01)00323-8.
543 [53] P. Van Houtte, S. Li, M. Seefeldt, and L. Delannay. Deformation texture prediction: From the Taylor model to
544 the advanced Lamel model. International Journal of Plasticity, 21(3):589–624, mar 2005. ISSN 07496419. doi:
545 10.1016/j.ijplas.2004.04.011.
546 [54] L. A. I. Kestens and H. Pirgazi. Texture formation in metal alloys with cubic crystal structures. Materials
547 Science and Technology, 32(13):1303–1315, sep 2016. ISSN 0267-0836. doi: 10.1080/02670836.2016.1231746.
548 [55] N. Deeparekha, Aman Gupta, Murat Demiral, and Rajesh K. Khatirkar. Cold rolling of an interstitial free (IF)
549 steelExperiments and simulations. Mechanics of Materials, 148, 2020. ISSN 01676636. doi: 10.1016/j.mechmat.
550 2020.103420.
551 [56] R. Pokharel, J. Lind, A. K. Kanjarla, R. A. Lebensohn, S. F. Li, P. Kenesei, R. M. Suter, and A. D. Rollett. Poly-
552 crystal plasticity: Comparison between grain - Scale observations of deformation and simulations. Annual Review
553 of Condensed Matter Physics, 5(1), 2014. ISSN 19475462. doi: 10.1146/annurev-conmatphys-031113-133846.

554 [57] Q. Z. Chen and B. J. Duggan. On cells and microbands formed in an interstitial-free steel during cold rolling
555 at low to medium reductions. Metallurgical and Materials Transactions A: Physical Metallurgy and Materials
556 Science, 35 A(11), 2004. ISSN 10735623. doi: 10.1007/s11661-004-0178-5.
557 [58] D. Vanderschueren, N. Yoshinaga, and K. Koyama. Recrystallisation of Ti IF Steel Investigated with Electron
558 Back-scattering Pattern (EBSP). ISIJ International, 36(8):1046–1054, 1996. ISSN 0915-1559. doi: 10.2355/
559 isijinternational.36.1046.
560 [59] P. Van Houtte, L. Delannay, and I. Samajdar. Quantitative Prediction of Cold Rolling Textures in Low-Carbon
561 Steel by Means of the Lamel Model. Textures and Microstructures, 31(3):109–149, jan 1999. ISSN 0730-3300.
562 doi: 10.1155/tsm.31.109.

26
563 [60] J. Gil Sevillano, P. van Houtte, and E. Aernoudt. Large strain work hardening and textures, 1980. ISSN
564 00796425.
565 [61] M. Hatherly and A. S. Malin. Shear bands in deformed metals. Scripta Metallurgica, 18(5), 1984. ISSN 00369748.
566 doi: 10.1016/0036-9748(84)90419-8.

567 [62] M. R. Barnett. Role of in-grain shear bands in the nucleation of 111 //ND recrystallization textures in warm
568 rolled steel. ISIJ International, 38(1), 1998. ISSN 09151559. doi: 10.2355/isijinternational.38.78.
569 [63] K. Ushioda and W. B. Hutchinson. Role of Shear Bands in Annealing Texture Formation in 3%Si-Fe (111)[112]
570 Single Crystals. Isij International, 29(10), 1989. ISSN 09151559. doi: 10.2355/isijinternational.29.862.
571 [64] W. B. Lee and K. C. Chan. Symmetry requirement in shear band formation. Scripta Metallurgica et Materiala,
572 24(6), 1990. ISSN 0956716X. doi: 10.1016/0956-716X(90)90289-S.
573 [65] K. Sedighiani, K. Traka, F. Roters, D. Raabe, J. Sietsma, and M. Diehl. Determination and analysis of the
574 constitutive parameters of temperature-dependent dislocation-density-based crystal plasticity models. submitted
575 to Mechanics of Materials, 2021.

576 [66] P. Eisenlohr and F. Roters. Selecting a set of discrete orientations for accurate texture reconstruction. Compu-
577 tational Materials Science, 42(4), 2008. ISSN 09270256. doi: 10.1016/j.commatsci.2007.09.015.
578 [67] P. Eisenlohr, M. Diehl, R.A. Lebensohn, and F. Roters. A spectral method solution to crystal elasto-
579 viscoplasticity at finite strains. International Journal of Plasticity, 46:37–53, jul 2013. ISSN 07496419. doi:
580 10.1016/j.ijplas.2012.09.012.

581 [68] P. Shanthraj, P. Eisenlohr, M. Diehl, and F. Roters. Numerically robust spectral methods for crystal plasticity
582 simulations of heterogeneous materials. International Journal of Plasticity, 66:31–45, mar 2015. ISSN 07496419.
583 doi: 10.1016/j.ijplas.2014.02.006.
584 [69] F. Roters, M. Diehl, P. Shanthraj, P. Eisenlohr, C. Reuber, S.L. Wong, T. Maiti, A. Ebrahimi, T. Hochrainer,
585 H.O. Fabritius, S. Nikolov, M. Friák, N. Fujita, N. Grilli, K.G.F. Janssens, N. Jia, P.J.J. Kok, D. Ma, F. Meier,
586 E. Werner, M. Stricker, D. Weygand, and D. Raabe. DAMASK The Düsseldorf Advanced Material Simulation
587 Kit for modeling multi-physics crystal plasticity, thermal, and damage phenomena from the single crystal up to
588 the component scale. Computational Materials Science, 158:420–478, feb 2019. ISSN 09270256. doi: 10.1016/j.
589 commatsci.2018.04.030.
590 [70] A. Ma and F. Roters. A constitutive model for fcc single crystals based on dislocation densities and its appli-
591 cation to uniaxial compression of aluminium single crystals. Acta Materialia, 52(12):3603–3612, jul 2004. ISSN
592 13596454. doi: 10.1016/j.actamat.2004.04.012.
593 [71] K. Sedighiani, M. Diehl, K. Traka, F. Roters, J. Sietsma, and D. Raabe. An efficient and robust approach
594 to determine material parameters of crystal plasticity constitutive laws from macro-scale stressstrain curves.
595 International Journal of Plasticity, 134:102779, nov 2020. ISSN 07496419. doi: 10.1016/j.ijplas.2020.102779.

596 [72] D. Raabe. Investigation of the orientation dependence of recovery in low-carbon steel by use of single orientation
597 determination. Steel Research, 66(5), 1995. ISSN 01774832. doi: 10.1002/srin.199501116.
598 [73] M. Calcagnotto, D. Ponge, E. Demir, and D. Raabe. Orientation gradients and geometrically necessary dislo-
599 cations in ultrafine grained dual-phase steels studied by 2D and 3D EBSD. Materials Science and Engineering
600 A, 527(10-11), 2010. ISSN 09215093. doi: 10.1016/j.msea.2010.01.004.

27
eclaration of Interest Statement

Declaration of interests

☒ The authors declare that they have no known competing financial interests or personal relationships
that could have appeared to influence the work reported in this paper.

☐The authors declare the following financial interests/personal relationships which may be considered
as potential competing interests:

View publication stats