Deformation inhomogeneity in large-grained AA5754 sheets
Guozhen Zhu,Xiaohua Hu,Jidong Kang,Raja K. Mishra,David S. Wilkinson
DOI: https://doi.org/10.1016/j.msea.2011.02.017
2011-01-01
Abstract:Research highlights ► Microstructure and strain relationship at individual grain level was studied. ► ‘Hot spots’ nucleate early and most keep growing throughout deformation stages. ► ‘Hot spots’ are correlated with ‘soft’ grains and soft-evolution grains. ► Grains with high Schmid factors tend to be ‘soft’ grains. ► Grains with the 〈1 0 1〉 direction close to tensile axis tend to become softer. Abstract Models for deformation and strain localization in polycrystals that incorporate microstructural features including particles are computationally intensive due to the large variation in scale in going from particles to grains to a specimen. As a result such models are generally 2-D in nature. This is an issue for experimental validation. We have therefore studied deformation heterogeneities and strain localization behavior of coarse-grained alloys with only two grains across the sample thickness, therefore mimicking 2-D behavior. Aluminum alloy sheets (AA5754) have been investigated by a number of surface techniques, including digital image correlation, slip trace analysis and electron backscattered diffraction, at the individual grain level. Local strain concentration zones appear from the very beginning of deformation, which then maintain sustained growth and lead, in one of these regions, to localization and final fracture. These ‘hot spots’ occur in areas with locally soft grains (i.e. grains with 〈0 0 1〉 or 〈1 0 1〉 close to the tensile direction) and soft-evolution orientations (i.e. grains with 〈1 0 1〉 close to the tensile direction). These grains can be correlated with Taylor and/or Schmid factors. Keywords Strain concentration AA5754 Coarse-grained Digital image correlation 1 Introduction There is growing interest in aluminum alloys within the automotive industry due to their high strength and stiffness to weight ratio, good formability, corrosion resistance, recycling potential and so on. Non-heat treatable Al–Mg sheets (5xxx series) find applications for body-in-white (BIW) internal structural parts and occasionally for outer panels. [1,2] However, raw material and manufacturing costs impede greater utilization of this material. Continuous strip casting (CC) provides a method for fabricating aluminum sheets at lower cost than conventional methods involving direct-chill casting (DC). However, the limited thermo-mechanical processing of CC material, a direct consequence of the low as-cast thickness limits the development of a homogeneous microstructure and further affects the formability of these alloys. [3,4] The formability of Al-Mg sheets is limited by plastic strain localization. The material constitutive behavior is one of the major factors determining the onset of strain localization, in terms of both diffuse and localized necking. Of equal importance, the various microstructural constituents can provide sources of deformation inhomogeneity which can in turn cause premature localized deformation and failure. The understanding of the role of various microstructure factors, i.e. grain structure, on strain localization is still incomplete. The crystallographic texture (e.g. the distribution of crystal orientations (ODF)) is a major cause of deformation anisotropy which can lead to problems such as earring and surface roughening. Earlier experiments on polycrystalline materials [5–7] have demonstrated the non-crystallographic nature of shear bands. However, these results, based on the incomplete information of the overall microstructure, cannot fully describe the effect of grain orientation on the path of propagation of the localization necking bands. To date, both characterization and simulation of the deformation process in a full three-dimensional microstructure with features at different scales ranging from small second phase hard particles to grains up to the much larger specimen scale are still not practical. Simulations of deformation using model microstructures [8,9] have shown that shear band formation is not only sensitive to the variation of statistical texture but also the spatial grain orientation distributions, e.g. the alignment of texture colonies, where a texture colony consists of a cluster of grains with similar grain orientations of a specific texture component. Similarly, with the assumption that there is no substantial change of grain orientations prior to macroscopic strain localization (i.e. localized necking), Hu et al. [10] have represented grain-level inhomogeneities using a constant relative Taylor factors of grains during uniaxial tensile deformation. Strain localization is found to be initiated in soft grains (with small Taylor factors) and propagate through a ‘soft path’ (the arrangement of grains of low Taylor factors) to form a macroscopic localized necking band. Experiments on oligocrystals (i.e. samples with a few grains of millimeter size) have magnified grain-level inhomogeneities to a level that enables more complete characterization and provides a bridge linking to microstructure-based simulation, which for computational reasons are generally two dimensional in nature. In contrast with other strain measurement techniques e.g. synchrotron X-ray diffraction [11] ), digital image correlation (DIC) provides a practical method that can be widely applied to detect local displacement out to large strain with high accuracy [12] . Combined with the surface microstructure characterization techniques (i.e. EBSD), deformation inhomogeneity in large grains can be directly correlated to the microstructure and its evolution. Raabe et al. [13] undertook detailed strain mapping under compression on oligocrystals and proved that macroscopic boundary conditions and grain interactions work together to develop deformation heterogeneities. They also introduced the kinematical Taylor factors to describe the relative hardness of grains and relate them to variation of local deformations between grains. Zhang and Tong [14] suggested that aluminum favors easy slip transmission and accommodation among grains. Zhao et al. [15] recently suggested that grain topology (i.e. the geometry of grain boundaries) and micro-texture play an important role in strain heterogeneities. Strain localization behavior in previous experiments [14–16] on specimens with a single layer of grains can be related to the planar variation of grain orientation, since there is no constraint of deformation by other layers of grains. The objective of the present work is to develop a good comprehension of the grain structure and mechanical properties relationship emphasizing the strain localization process in AA5754 alloys. This is done by the use of samples with approximately two layers of grains with sizes in the millimeter range, in order to understand the influence of though-thickness grain interaction on deformation heterogeneity and development of strain localization and to facilitate the comparison between experiments and modeling. 2 Sample preparation and experimental procedure The material under study is a 1 mm thick AA5754 DC sheet with 0.21 wt.% Fe content. The initial grain structure is nearly equiaxed with an average grain size of around 25 μm. Grain structure was revealed under polarizing light after electrical etching with Barker's reagent (48–50% HBF 4 (conc.) 15 ml + H 2 O (distilled) 90 ml). Optical observation was carried out using a Zeiss Optical Microscope. The coarse-grained structure was obtained through a thermomechanical treatment starting with a small rolling reduction pre-deformation (4–10% thickness reduction) followed by annealing at 575 °C for 3 h. The resulting microstructure consists of thin fine-grained layers on the surface and two or more layers of large pancake-shaped grains in the bulk of sheets ( Fig. 1 ). The fine surface grains are removed by grinding. A quasi-2D microstructure was therefore obtained with two layers through the thickness combined with a center-plane separating the two layers when the rolling reduction ranged from 4% to 7%. Samples reduced by 4% had an average post-annealing grain size of about 1.2 mm while samples reduced 7% led to grain sizes of about 0.6 mm. Tensile bars were machined using electron discharge machining (EDM) to ensure high dimensional precision and minimum surface damage. The dimensions of the tensile bars for materials with 1.2 mm and 0.6 mm grain sizes are shown in Fig. 2 (a) and (b) , respectively. A large transition radius is needed to ensure necking within the gauge length. The gauge area reduction was modified compared with the ASTM standard in order to facilitate EBSD and DIC measurements. The initial crystallographic orientation maps on two surfaces of polished tensile bars were obtained using the electron backscattered diffraction (EBSD) technique. EBSD scanning was done at 35× magnification with 15 μm step size over an area slightly larger than the reduced gauge region on both surfaces. EBSD analysis utilizes Kikuchi patterns, which are collected and indexed using an EDAX-TSL OIM system attached to a LEO Scanning Electron Microscope. The display algorithm removes EBSD data which are indexed with a confidence level less than 95% and shows these as black grains. In order to reduce the number of artificial misorientations at the boundaries during large-area scanning, four or eight scans are performed for each surface and a MatLab program was written to merge these scans. The program can detect non-edge and edge grains and determine a mean orientation of non-edge grains and of each pair of corresponding edge grains which cross the boundaries between scans. The experimental error for crystal orientations is within 2° after such treatment under the assumption that the orientation in a grain is homogeneous. This assumption is statistically reasonable since very few little sub-grain structure was detected in hundreds of millimeter sized grains in these samples, after annealing at elevated temperature. Tensile tests were performed at 1 mm/min cross-head speed on a screw-driven tensile machine. Displacement-controlled mode was selected to avoid abrupt changes in deformation rate during necking. Black and white spots were painted on the specimen surface prior to testing. These are used to form a random speckle pattern on the surface. Two high-speed cameras recorded the speckle patterns on both sides of the sample with a time interval of 2 s. Load–displacement curves were recorded and a two-dimensional technique was used to calculate the strain maps on both sides of the tensile bars. A commercial digital image correlation (DIC) system (Aramis) was used to follow the surface deformation pattern during tensile tests. From this we calculate the micro true strain, i.e. normal strain ɛ xx , ɛ yy and shear strain ɛ xy at the facet level. The deformation strain along the thickness direction is estimated based on the assumption of consistent volume during plastic deformation. (This assumption is valid for strains that exceed the elastic strain by a significant margin as is the case here). From these measurements the von Mises equivalent true strain was obtained and used as an index of micro deformation for each facet. A 7 × 5 pixel facet/step size was used for the strain calculation in Aramis, which is equivalent to a real dimension of 125–300/90–200 μm. The accuracy of calculated strain is limited by the stability of facet centers from stage to stage, which is set as 0.04 pixels, leading to a maximum deviation of strain within 0.005. The evolution of grain orientations was measured on one coarse-grained sample during in situ tensile testing at 0.5 mm/min strain rate run in the SEM. In-situ tests were load-controlled. EBSD maps were collected at approximate strains of around 0.06, 0.13 and 0.19. 3 Experimental results 3.1 Samples with an average grain size of 1.2 mm 3.1.1 Grain-level microstructure Grain structures on both sides of the gauge area for Sample A with 1.25 mm grain size are presented in Fig. 3 with inverse pole figure (IPF) color schemes showing the position of tensile axis in the crystal reference frame of grains. In addition, the orientations of individual grains (numbered 1–16) on both sides are represented in the IPF triangle with respect to the tensile axis. 3.1.2 DIC strain mapping The evolution of micro von-Mises equivalent true strain ( ɛ vm ) on both sides at the gauge areas of the sample during deformation is displayed in Fig. 4 (a)–(g) using a color contours (ranging from blue to red), where, the red part of the colorbar defines the range when strains are larger than half peak value ε vm h = ( ε vm p − ε vm m ) / 2 . ε vm p is peak true strain while ε vm m is the medium true strain value. Regions of local strain concentration are colored in red in these maps, and referred to as ‘hot spots’. These nucleate at very early stages of the deformation, and show sustained growth during almost all of the deformation stages. In contrast, the regions with local strain relaxation are colored blue and referred to as ‘cold spots’. These also keep the same positions during almost all deformation stages and deformation stops in these regions after a certain stage. Strain distribution evolutions of a line scan along the tensile direction describes the nucleation and development of ‘hot spots’ from another view. Fig. 4 (h) shows such evolutions along line l in Fig. 3 on both sides parallel to the tensile axis. Strain distribution along this line keeps similar shapes with deformation until the deformation concentrates on one of the initial sharp peaks leading to necking. The evolutions of von Mises true strain distribution on both sides are quite similar, which implies that the grains in Side A have highly coordinated deformation with corresponding grains in Side B, which may also suggest that surface strains calculated from DIC can give a good evaluation of the bulk strains. 3.1.3 Tensile responses of individual grains In order to gain a better understanding of the deformation processes, detailed analyses for individual grains, especially those in the zones of strain concentration and final failure were carried out on Sample A. The initial grain orientations are presented in the form of overlapping Schmid factors on both sides, and a grey-level legend representing the Schmid factors. Strain partitioning in each individual grain was determined from a network of DIC facet points ( Fig. 5 ). Those points of a grain in Side A were selected for analyses so that their corresponding (overlapping in ( x , y ) coordinates) points on Side B are within another grain(s). Therefore, these points are the overlapping points of a pair of grains with each lying on different sides of the sample in order to show if there is coordinated deformation between the corresponding grains. Micro ( ɛ vm )–macro ( ε ¯ vm ) von Mises true strain curves for individual grains are plotted in Fig. 6 . The micro von Mises true strain ɛ vm indicates the strain for each facet point, while macro von Mises true strain ε ¯ vm is the average true strain of all the DIC facet points on one side in the gauge area. The results can be summarized as follows: (a) The deformation inside each grain is quite uniform ( Fig. 6 (a) and (b)) except for the grains which lie at the sample edges, (see Fig. 6 (c), (d) and (f)). The curves for facet points within one edge grain are quite dispersed, indicating that the deformation is quite inhomogeneous. This is due to the different boundary conditions of these grains compared those interior grains which are surrounded and constrained by other grains in the x – y plane. In addition, the ε vm − ε ¯ vm curves of the edge grain are found to have a curvature at different macro strain levels, indicating large variations of the evolution of grain orientations. However, within most of the interior grains, the micro strains maintain a relatively linear relationship with the macro true strains. It seems that grains try to maintain their ‘hardness’ compared to their initial states, indicating that the relatively hardness of (surface) grains is strongly associated with their initial states. (b) As mentioned in Section 3.1.2 , strain maps are quite similar for both sides. Accordingly, the behaviors of the overlap regions corresponding to grains in Side A and B are almost the same ( Fig. 6 (a) and (b)), indicating that the interaction between the pair of grains is necessary to achieve the strain compatibility. (c) The micro von Mises true strains of grains inside the strain localization region ( Fig. 6 (a)–(d)) increases dramatically after necking. Some points of region A8B14 are inside the localization zone which is the starting position for fracture. The ε vm − ε ¯ vm curves show that the local strains of these points increase dramatically after necking, while those of other points in the region stop changing ( Fig. 6 (d)) at some point after necking, implying that deformation inside the same grain may not always be homogenous after necking. (d) Micro von Mises strains of the regions outside the necking zones ( Fig. 6 (e) and (f)) stop increasing after necking. This confirms the achievement of shear localization in such a coarse-grained sample: strain is only concentrated inside the localization region after localized necking. 3.1.4 Surface slip traces Plastic deformation in high stacking fault energy Al alloys is the result of dislocation slip which leaves slip traces observable on polished sample surfaces. Slip trace observations have been made by observing the side surface of the fractured sample as shown in Fig. 7 . In general, one or two slip traces were observed inside each grain, which implies that most of the plastic deformation of grains is contributed by one or two slip systems. Slip traces with relatively homogeneous interspacing were observed within the body of the grains ( Fig. 7 (c)). Transition areas with mixed slip traces were recorded near grain boundaries. These observations suggest that while plastic deformation is relatively homogeneous in the grain interior, the constraints imposed by grain to grain compatibility demands the involvement of additional slip systems near the grain boundaries. The slip traces are the lines of intersection between the slip planes of activated slip systems and the sample surface. According to the Schmid law, deformation occurs on the slip system with the largest Schmid factor. This is especially true in the case where there is only one set of slip traces parallel to each other on the surface of a grain. The observed slip traces majorly agree with the predicted slip systems. 3.1.5 The evolution of grain orientations The initial surface grain structure and the evolution of grain orientations during an in situ tensile test on Sample B are represented using IPF color maps ( Fig. 8 (a) and (b) ) and IPFs with respect to the tensile axis ( Fig. 8 (c)–(f)). The distribution of grain orientations becomes more random with increasing strain. The division of grain orientation inside one grain is observed in parts with different neighboring oriented grains. In general, the lattice rotates to achieve deformation axis closer to 〈0 0 1〉, 〈1 1 2〉 or 〈1 1 1〉. 3.2 Sample with an average grain size of 0.6 mm 3.2.1 Grain-level microstructure To further understand grain-grain interaction during deformation, a tensile test was performed on Sample C with shape II, which provides more grains in the gauge area compared to Sample A(B) with larger grain sizes. Grain structure for both sides of the gauge region for Sample C with average grain size of 0.6 mm is presented in IPF color maps with respect to tensile axes ( Fig. 9 ). 3.2.2 DIC strain mapping The observations in Sample C were similar to those in Sample A ( Fig. 10 ): a few ‘hot spots’ form at the beginning, and grow in a sustained manner until facture occurs at one of them. In addition, ‘cold spots’ were also observed from the beginning and these eventually see a stop in increasing strain after a certain stage. The half peak value of the micro von Mises true strain is almost the same as that in Sample A for a specific macroscopic strain level, which may due to using the same facet size compared to grain size to calculate strain. 3.2.3 Tensile responses of oriented grain clusters in Sample C The local tensile responses of grain clusters (including a few surface grains) were investigated on Sample C. As shown in Fig. 11 , grain orientations were indexed by the overlapping IPF coloring maps of both sides. According to the spot size of ‘hot and cold spots’, the cluster of grains has a diameter around 700 μm, including 2–4 grains in each side. Strain partitioning of each oriented grain cluster was determined using a network of DIC facet points. Fig. 12 (a) and (b) show the ‘hot spot’ regions, (c) and (f) are within the ‘cold spots’ and (d) and (e) have comparable high strain with their surroundings as marked in Fig. 11 . The initial grain-level microstructure is represented by the overlapping IPF color maps on both sides. The color of grains refers to the position of the tensile axis in the crystal reference frame. It is interesting that strain concentration regions are often associated with grain clusters including green grains (i.e. grains of ∼〈1 0 1〉 direction close to tensile axis) in both sides while the purple grains (i.e. grains of ∼〈1 1 2〉 direction close to tensile axis) always leads to the strain relaxation regions. 3.2.4 Grain orientations of fractured Sample C EBSD maps were obtained on both sides of the fractured sample for Sample C. ( Fig. 13 ) In general terms, grains tend to be in purple and blue within the IPF maps, representing orientations that concentrate near the 〈0 0 1〉–〈1 1 1〉 line after tensile deformation, which could be predicted from the evolution rule observed in Sample B. 4 Discussion Although grain behavior is not solely controlled by crystal orientation, it seems that grains of ∼〈1 0 1〉 direction, e.g. the green grains such as grain No. 11, 14 and 15 on Side A as shown in Fig. 3 , have more possibility to become the region of ‘hot spots’ with large local strains, while the grains of ∼〈1 1 2〉 directions, i.e. the purple grains in Fig. 3 (No. 5 in Side A and No. 4, 6, and 9 in Side B), usually see small strains and become the regions of ‘cold spots’. Furthermore, oriented grain clusters in Sample C that include green grains are the preferred site for a local strain concentration region, while grain clusters including purple grains always show-up as strain relaxation regions. As shown in Fig. 14 , the initial local deformation tends to be larger with decreasing misorientation between tensile axis and the 〈1 0 1〉 axis of a grain. From the above observations, it appears that the grains of ∼〈1 0 1〉 direction are softer compared to all other orientations, while the grains of ∼〈1 1 2〉 direction are harder to deform. The relative strengths of grains during deformation are often correlated with their Taylor factors [16] or Schmid factors [14,17] . A grain is relatively soft with a smaller Taylor factor or larger Schmid factor. Actually the Schmid factor is almost the inversion of Taylor factors for the grains with orientations of high symmetry including 〈0 0 1〉 and 〈1 1 1〉, while there is big variation between the Taylor factor and the inverse of the Schmid factor for grains with ∼〈1 0 1〉 direction (see Fig. 15 ). These two high symmetry orientations are quite stable during the deformation of high stacking fault energy fcc materials [18] . The construction of the yield surface of a textured polycrystalline material is based on the average Taylor factors of a representative cluster of grains which are deformed in different deformation modes [19] . The deformation heterogeneity in a plane strain compression test of a large grained sample has been assessed based on Taylor factors by Raabe et al. [16] , while it has been presented by Zhang and Tong [14] and Tatschl and Kolednik [17] based on Schmid factor in the uniaxial tension test of large grained samples. From the Schmid and Taylor factor contours ( Fig. 15 ) that are plotted on the tensile direction in the IPF triangle, it shows that the grains with both 〈1 0 1〉 and 〈1 1 2〉 directions have relatively high Schmid factors of around 0.42. The Taylor factors of grains with 〈1 0 1〉 orientation (∼3.7), however, is higher than that with 〈1 1 2〉 orientations (∼3), indicating the grains with 〈1 0 1〉 should be harder. The deformation of a grain is also constrained by the surrounding grains and the corresponding grain on the other side. For example, there is quite a large variation of deformations for the grains of similar orientations, e.g. ∼〈1 0 1〉 (see Fig. 14 ), due to this effect. We have shown that there is coordinated deformation between the grain of one layer and the corresponding grain on another layer. It could be that how much the two corresponding grains can deform depends on the overall orientation relationship of these two grains. To illustrate this point, average cross-sectional (TD × ND) Schmid and Taylor factors are plotted ( Fig. 16 ) in reference to the x 1 coordinates along the loading direction (RD), where the Schmid and Taylor factors are averaged over all the points of each TD × ND section along the loading direction. The approximate position of localized necking and final failure is also noted in the curves. It can be seen that the position of localized necking and final failure correlates well with the position corresponding to maximum Schmid factor. In addition, for Sample C, the regions with maximum Schmid factors are located in the same regions with minimum Taylor factors due to statistical effect of more grains. From the above discussion, it seems that the relative ‘hardness’ of grains in uni-axial tension is less correlated with the initial Taylor factors than with Schmid factor as shown in earlier works [14,17] . This may be due to the additional freedom for surface deformation provided by these extra-large grains, leading to only the one or two slip systems that have the largest Schmid factors being activated at the surface. The strain compatibility between the surface grains can be accomplished by the activation of more slip systems at the regions near the grain boundaries. In a typical material that has many grains through the thickness, the internal grains in the bulk are fully constrained by neighbors. More slip systems need to be activated to provide strain compatibility between neighboring grains, in this case the relative ‘hardness’ of the grains may be more strongly coupled with their Taylor factors. It should also be noted that this discussion is based on the initial orientation of grains. During plastic deformation, the grains can re-orient which may change the relative strength of grains leading to geometrical softening and strain localization [20] . Early research [21] on the statistical distribution of grain orientations suggested that texture components see no significant changes before strain localization. However, our analysis based on individual grains demonstrated the preferred evolution trend: grains with 〈1 0 1〉 parallel to the tensile axis will generally rotate toward the 〈0 0 1〉–〈1 1 1〉 line in the IPFs with respect to the tensile axis. Similar observations are obtained from X-ray diffraction results of tensile tested sample [22,23] and SEM in situ tensile testing [17] of polished samples with a single layer of large grains. Furthermore, due to the effect of deformation heterogeneity and anisotropy, the local deformation mode can deviate from the ideal applied uniaxial tension deformation model, e.g. Δ ε = 1 0 0 0 − 0.5 0 0 0 − 0.5 Δ ε 11 which is usually used to calculate the initial Taylor factor for uniaxial tension. Thus, the relative ‘hardness’ of grains can change during deformation. For example, the Taylor factors in the grains with 〈1 0 1〉 close to tensile deformation can become smaller during straining. Although the crystal re-orientation can be measured by in situ EBSD measurement during tension and the deformation mode on the surface can be tracked by DIC, applying both technologies at the same time to a reasonable resolution and precision is still too challenging. These effects, however, can be considered by finite element modeling, which will be discussed in a separate paper where either a Taylor factor-based formulation, i.e. the FE-TBH model [24] , or the conventional viscous crystal plasticity formations (CPFEM) are used to represent the behaviors of crystalline grains. The interaction between grains is quite hard to be analyzed experimentally in a quantitative manner. However, microstructure-based models could trace both orientation and strain evolution of grains and pick up the grain interaction effect. The simulation results of the nucleation and growth of ‘hot spots’ by both the FE-TBH and CPFEM models show reasonable agreement with experimental results and support some new micro-features of strain distribution on the sub-grain level. Fig. 17 shows examples of deformation of individual grains predicted by the CPFEM calculation [25] . The strain distribution inside one grain is relatively even with fluctuations at grain boundaries due to the interaction with neighbors for both layers. 5 Conclusions The micro tensile responses of coarse-grained AA5754 tensile bars have been reported in this paper. Strain distribution evolution maps present the special deformation processes of coarse-grained samples: ‘hot bands’ nucleate very early during deformation and most of them keep growing throughout deformation stages. Two-surface von Mises true strain maps reveal that grains belonging to different layers deform in tandem through the sample thickness. Micro–macro von Mises true strain curves provide evidence for strain localization: strain localization regions can be attributed to dramatically increasing strain after a certain stage while deformation stops in all other regions of the sample. Linking the observed deformation behavior with microstructure, the following conclusions can be drawn: (a) the average initial cross-sectional microstructure (Taylor or Schmid factor) factor can be used to predict several possible sites for strain concentration; (b) ‘hot spots’ occur in local regions with soft grains; (c), those grains with 〈1 0 1〉 close to the tensile direction are favored sites for strain concentration; (d), grains rotate toward the 〈0 0 1〉–〈1 1 1〉 line in IPF triangle of the tensile axis and change the relative ‘hardness’ of grains, which is probably the determining factor for the sustained growth of ‘hot spots’. Acknowledgements We are very grateful to Mr. Robert Kubic, General Motors R&D Center, for helping with the EBSD work. The financial support from General Motors of Canada and the Natural Sciences and Engineering Research Council (NSERC) of Canada is acknowledged. References [1] J. Hirsh Materials Forum 28 2004 15 23 [2] The Canadian Aluminium Industry Technology Roadmap, Industry Canada, report, 2000, edition. [3] J Sarkar T.R.G. Kutty D.S. Wilkinson J.D. Embury D.J. Lloyd Aluminium Alloys: Their Physical and Mechanical Properties 331–333 Pts 1–3 2000 583 588 [4] J. Kang M. Jain D.S. Wilkinson D. Embury S. Kim A.K. Sachdev Tenth International Symposium on Plasticity and Its Current Application 2003 NEAT Press pp. 181–183 [5] L. Anand W.A. Spitzig Journal of the Mechanics and Physics of Solids 28 1980 113 128 [6] A. Korbel P. Martin Acta Metallurgica 36 9 1988 2575 2586 [7] J.C. Huang G.T. Gray III Acta Metallurgica 37 12 1989 3335 3347 [8] W.B. Lee K.C. Chan Textures and Microstructure 14–18 1991 1221 1226 [9] P.D. Wu A. Graf S.R. MacEwen D.J. Lloyd M. Jain K.W. Neale Internaional Journal of Solids and Structures 42 2005 2225 2241 [10] X.H. Hu M. Jain D.S. Wilkinson R.K. Mishra Acta Materialia 56 13 2008 3187 3201 [11] H.-S. Zhang K. Komvopoulos Philosophical Magazine 90 16 2010 2235 2248 [12] Braunschweig, Aramis V4.7 Manual. 2001: GOM mbH. [13] D. Raabe M. Sachtleber Z. Zhao F. Roters S. Zaefferer Acta Materialia 49 17 2001 3433 3441 [14] N. Zhang W. Tong International Journal of Plasticity 20 2004 523 542 [15] Z. Zhao M. Ramesh D. Raabe A. Cuitino R. Radovitzky International Journal of Plasticity 2008 [16] D. Raabe M. Sachtleber H. Weiland G. Scheele Z. Zhao Acta Materialia 51 6 2003 1539 1560 [17] A. Tatschl O. Kolednik Materials Science and Engineering A 342 2003 152 168 [18] W. Mao Y. Yu Materials Science and Engineering A 367 2004 277 281 [19] P.V. Houtte S. Li M. Seefeldt L. Delannay International Journal of Plasticity 21 2005 589 624 [20] I.L. Dillamore J.G. Roberts A.C. Bush Metal Science 13 2 1979 73 77 [21] D. Peirce R.J. Asaro A. Needleman Acta Metallurgica 30 1982 1087 1119 [22] L. Margulies G. Winther Science 291 2001 2392 2394 [23] G. Winther L. Margulies H.F. Poulsen S. Schmidt A.W. Larsen E.M. Lauridsen S.F. Nielsen A. Terry Materials Science Forum 408–412 2002 287 292 [24] X.H. Hu G.A. Cingara D.S. Wilkinson M. Jain P.D. Wu R.K. Mishra Materials and Continua 14 2009 97 122 [25] X.H. Hu, G.-Z. Zhu, D.S. Wilkinson, M. Jain, P.D. Wu and R.K. Mishra, Finite element and experimental studies of tensile deformations of AA5754 sheet alloys with extra-large grains, unpublished.