Materials Science & Engineering A 657 (2016) 115–122 Contents lists available at ScienceDirect Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea Role of Mg in simultaneously improving the strength and ductility of Al–Mg alloys Byeong-Hyeon Lee a, Sung-Hoon Kim a, Jun-Hyoung Park a, Hyung-Wook Kim b,n, Jae-Chul Lee a,n a b Department of Material Science and Engineering, Korea University, Seoul 136-713, Republic of Korea Structural Materials Division, Korea Institute of Materials Science, Changwon 641-010, Republic of Korea art ic l e i nf o a b s t r a c t Article history: Received 18 November 2015 Received in revised form 19 January 2016 Accepted 26 January 2016 Available online 27 January 2016 We found that when Mg is added to Al in small amounts, it results in alloys that exhibit both significantly higher strength and ductility than those of the pure Al counterpart Strength and plasticity are contrasting properties for most crystalline alloys; as such, the simultaneous improvement of these properties seems contradictory from the viewpoint of dislocation motion Comparative studies of Al–xMg alloys containing different amounts of Mg were performed to elucidate how these apparently mutually exclusive properties were improved simultaneously; microstructural observations were made using X-ray diffractometry and transmission electron microscopy to assess the effect of Mg on the strength and ductility on the alloys, while molecular dynamics simulations were performed to comprehensively understand how the resulting high-strength alloys could also exhibit high ductility & 2016 Published by Elsevier B.V Keywords: Al–Mg alloys Mechanical characterization X-ray diffraction Dislocation structures Stacking fault energy Introduction In contrast to their steel counterparts, Al alloys, despite having high specific strength, exhibit low tensile ductility, which limits their widespread application This low ductility stems from the absence of a structure that can mitigate strain localization (or necking) under tension Therefore, methods for synthesizing Al alloys that show both high strength and tensile ductility are centered on designing a structure that is capable of absorbing the mechanical strain while supporting an external load during deformation The stress–strain relationship of an alloy that undergoes uniform deformation can be described using the Hollomon equation, which is as follows: σ = K εn (1) The strain-hardening-related exponent, n, in Eq (1) represents the strain-hardening capacity of the alloy and determines the amount of plastic deformation that can be accommodated prior to the onset of necking; the magnitude of n varies with the value of the stacking fault energy (γSF) [1] Stacking faults are planar defects that are typically formed in crystals with a close-packed structure; these faults form because n Corresponding authors E-mail addresses: hwkim@kims.re.kr (H.-W Kim), jclee001@korea.ac.kr (J.-C Lee) http://dx.doi.org/10.1016/j.msea.2016.01.089 0921-5093/& 2016 Published by Elsevier B.V of the disruptions in the stacking sequence of the crystals resulting from plastic deformation Their formation involves the splitting of perfect dislocations into Shockley partials and is governed by the absolute value of γSF, which is inversely proportional to the distance between the partials Therefore, the γSF value of an alloy is considered an intrinsic property that determines its mode of plastic deformation For example, the activation of slip systems occurs via dislocation pile-up in alloys with low γSF values, whereas cross-slips occur preferentially in alloys with high γSF values [2,3] Therefore, a low γSF leads to an increase in the strainhardening rate, thus mitigating strain localization and promoting the uniform elongation of the material [2] In this context, the γSF value is an important structural parameter that influences the mechanical properties of the alloy, and a decrease in γSF is usually associated with enhanced strength and ductility in face-centered cubic (fcc) metals [4] The γSF value of an alloy is composition sensitive and, therefore, can be adjusted by adding small amounts of various alloying elements [5] For example, elements such as C, Mn, and Al affect the γSF value of Fe-based alloys (or steels) [6–8] The effects of such alloying elements on the deformation mechanisms and the resultant properties of steels have been studied extensively In fact, twin-induced plasticity and transformation-induced plasticity steels are representative examples in which the composition dependency of γSF has been exploited for simultaneously improving both strength and ductility In contrast, the composition dependency of the γSF value of Al alloys has not been studied widely 116 B.-H Lee et al / Materials Science & Engineering A 657 (2016) 115–122 Fig (a) Representative computational cell corresponding to the alloy Al–3Mg used to determine the GSFE curve and γSF (b) Three (111) planes with the stacking sequence ̅ ] and partial [112] dislocations are denoted of an fcc metal, in which the directions corresponding to the complete [ 110 Therefore, the effects of alloying elements on the deformation mechanisms and mechanical properties of Al alloys are less understood than those of their steel counterparts Al is strengthened by adding alloying elements such as Cu, Mg, and Zn to it Of these elements, Cu is known to increase the γSF value of Al, while Mg tends to lower it [9] Therefore, Al–xMg alloys, in addition to being promising materials for high-strength applications, constitute excellent systems for improving our understanding of dislocation motion and its effects on strength and ductility Therefore, in this study, we added various amounts of Mg to Al and found that the resulting Al–xMg (wt%) alloys had both higher strength and ductility than those of their pure Al counterpart Through experiments and molecular dynamics (MD) simulations, we performed comparative studies of these alloys to elucidate why these apparently mutually exclusive properties improved simultaneously Experimental procedure 2.1 Sample preparation As mentioned above, Al–xMg alloys (x ¼0, 3, 5, 7, and 10 wt%) were used as the model alloys in this study; their nominal compositions are listed in Table Appropriate amounts of Al and Mg corresponding to the listed Al–xMg compositions were melted at 680 °C and cast into 150 mm (width) mm (thickness) strips using a laboratory-scale twin-roll strip caster The as-cast strips were subsequently homogenized at 500 °C for h and then rolled at ambient temperature under lubrication at a 30% reduction per pass to a final thickness of mm Before the tensile tests, the coldrolled strips were annealed at 450 °C for h, in order to induce recrystallization; this was followed by cooling in air 2.2 Tensile tests and microstructural observations The samples for the tensile tests were machined from the rolled and annealed sheets, with their long axis being parallel to the rolling direction These samples were tested under tension, in accordance with the ASTM E-8M standard, at an initial strain rate of 1.5 10 s (crosshead speed ¼ mm/min) using a universal testing machine (Model 3367, Instron) The strains were measured using an extensometer attached directly to the gauge part of the test specimen In addition, the contribution of the solution effect to the strength of the alloys was estimated by quantifying the amounts of Mg dissolved in the α-Al phase of the annealed samples; this was achieved by determining the lattice parameters of the alloys from X-ray diffraction (XRD, LYNXEYE, Bruker) measurements These measurements were performed at a scanning rate of 2°/min using filtered Cu–Kα radiation (λ ¼ 1.542 Å) from a source operating at 40 mA and 40 kV Prior to the scanning process, the diffraction peaks were calibrated using standard Si powders The samples were scanned over 2θ values of 30–100°, while being rotated at 15 rpm The lattice parameters were determined from six different peaks using the full-width-at-half-maximum method under linear boundary conditions The measured values were subsequently compared with those obtained from thermodynamic calculations of the Al–xMg alloys; this comparison yielded estimates of the Mg contents of the alloys The microstructural features of the alloys, such as their grain sizes, precipitates, and dislocation structures, were examined using a scanning electron microscopy system (Quanta 3D FEG SEM, FEI Company Inc.) equipped with an energy-dispersive X-ray spectroscopy attachment (Octane Silicon Drift Detector Series, EDAX) These features were also examined via electron backscatter diffraction analyses (S-4300, Hitachi) and transmission electron microscopy (TEM, Tecnai-F20 G2, FEI) 2.3 Evaluation of stacking fault energy (γSF) Table Chemical compositions of the Al–xMg alloy samples used in this study (in wt%) Composition Mg Si Fe Ti Al Al–3Mg Al–5Mg Al–7Mg Al–10Mg 3.1 5.1 7.3 10.2 0.03 0.03 0.04 0.03 0.13 0.13 0.12 0.12 0.02 0.02 0.02 0.02 Bal Bal Bal Bal The effect of Mg on the γSF value of the alloys was evaluated by constructing three-dimensional (3D) binary alloys corresponding to the Al–xMg compositions; these constructions were performed through MD simulations that employed the embedded-atommethod potential developed for the Al–Mg atomic pair [10] Approximately 4000 atoms were packed into the simulation cell to produce the 3D crystals of the Al–xMg alloys, with the x-, y-, and z- B.-H Lee et al / Materials Science & Engineering A 657 (2016) 115–122 ̅ ], and [ 1̅ 11 ̅ ] directions, reaxes being parallel to the [112], [ 110 spectively (Fig 1(a)) This configuration corresponded to a stacking arrangement in which 10 atoms were aligned along the x- and yaxes to form a close-packed (111) plane; 40 layers of the (111) plane were stacked in the sequence ABCABC along the z-axis, thereby forming the fcc crystal The model alloys were then relaxed using an energy minimization method [11] Prior to mechanical testing, a periodic boundary condition was applied along the x- and y-axes, in order to eliminate the surface effects, while the z-axis was designated as the free boundary The ̅ ) as the model crystals were then divided into two parts, with ( 1̅ 11 common plane at K, by shearing the upper 20 blocks along the xdirection, that is, along the [112] direction, which is parallel to the Burgers vector of the partial dislocations (bp) in fcc metals (Fig (b)) The corresponding changes in the energies of the cells were calculated for each Al–xMg alloy and transformed into the generalized stacking fault energy (GSFE) curve and the γSF value [12] Results and discussion 3.1 Tensile properties of Al–xMg alloys Fig 2(a) shows the stress–strain curves of the rolled and annealed Al–xMg sheets containing different amounts of Mg 117 Compared to pure Al, the Al–xMg alloys exhibited two peculiar behaviors The first was the appearance of a stress plateau after yielding, the point after which the stress remains constant even with an increase in the strain This behavior is typically referred to as the yield point phenomenon (or the Piobert–Luders effect) [13] and arises from interactions between the gliding dislocations and the solute atoms; in the initial stage of plastic deformation, gliding dislocations become stuck to Mg atoms Therefore, to sustain the plastic flow at the imposed strain rate, either new dislocations must be generated or the preexisting ones stuck to the solute atoms must be reactivated, thereby leading to the Piobert–Luders effect The second peculiar behavior was manifested in form of a serrated flow curve and is typically referred to as the Portevin–Le Chatelier effect [13–16] This effect stems from a competition between the diffusion rate of the solute atoms (Mg) and the speed of the propagating dislocations That is to say, when Mg atoms diffuse faster than the speed at which the dislocations move, the latter are arrested by the former These arrested dislocations are freed with further plastic deformation, resulting in a decrease in the stress The repeated arrest and release of dislocations is manifested as serrations in the flow curve, which, in turn, can lead to surface roughening of the deformed sheet (i.e., the formation of an orange-peel-like surface) Determining the physical mechanisms governing these peculiar behaviors is essential but was beyond the scope of this study As the stress–strain curves (Fig 2) show, the yield and tensile strengths of the alloys increased linearly from 23 MPa to 170 MPa and 57 MPa to 385 MPa, respectively, with an increase in the Mg content Furthermore, Fig 2(b) shows that the ductility decreased slightly with the increase in the Mg content till 3% and then increased gradually and linearly thereafter The strength and the ductility are, in general, inversely related; hence, these experimental observations were especially interesting This simultaneous improvement in the strength and ductility stemmed from the structural changes associated with the addition of Mg 3.2 Effect of Mg addition on strengthening As shown in the previous section, the addition of Mg to Al resulted in alloys with ductilities and strengths higher than those of pure Al The strength of an alloy is determined by the stress required to overcome the obstacles that hinder the motion of gliding dislocations These obstacles include defects such as solute atoms, precipitates, other dislocations, and grain boundaries Therefore, the yield strength of an alloy ( σy ) can be determined by adding the individual contributions of the defects ( ∆σi ) to the strength of the pure element ( σy ), as follows: Fig (a) Engineering stress–strain curves, as recorded during the tensile testing of the Al–xMg alloy sheets prepared by rolling and subsequent annealing (b) Dependence of the yield strength and uniform deformation of the sheets on the Mg content σy=σy+∆σ (2.1) ∆σ =∆σss +∆σgb +∆σppt +∙∙ (2.2) In Eq (2.2), ∆σss , ∆σgb , and ∆σppt are the increases in the strength attributable to the effects of the solid solution, grain boundaries, and precipitation, respectively The inverse pole figure (IPF) maps in Fig show a comparison of the sizes and distributions of the recrystallized grains in the Al– xMg alloys Pure Al (Fig 3(a)) has a coarse-grained structure with an average grain size of 215 μm However, the grain size was significantly refined, to 30 μm, with the initial addition of Mg and reduced further from 30 μm to 21 μm with the increase in the Mg content to 3–10% This reduction in the grain size resulted in an increase in the yield strength, in accordance with the Hall– Petch relation, as follows: ∆σgb=Ky d−0.5 (3) 118 B.-H Lee et al / Materials Science & Engineering A 657 (2016) 115–122 Fig IPF maps showing the grain structures of the Al–xMg alloys with Mg contents of (a) 0%, (b) 3%, (c) 5%, (d) 7%, and (e) 10% where Ky is a constant ranging from 0.06 MPa m1/2 to 0.28 MPa m1/2, in the case of the 5xxx series Al alloys [17] Based on Eq (3), a strength increment of 30–35 MPa was predicted owing to the effects of grain refinement Therefore, grain-boundary strengthening alone could not account for the strength exhibited by the alloys with the high Mg contents, indicating that other factors had to be evaluated Owing to the high cooling rate, strip-cast Al alloys typically not contain coarse intermetallic particles However, small Mg-rich precipitates formed, especially in the grain-boundary regions of the samples, albeit in low volume fractions and with a large interspacing (Fig 4) This indicated that the Mg existed primarily as a supersaturated solution and that strengthening via the Orowan mechanism did not occur The Mg atoms dissolved in the Al matrix constitute point defects that occupy the substitutional sites and hence act as obstacles to the motion of dislocations Furthermore, the Mg atoms diffuse close to the mobile dislocations during loading and inhibit their motion during loading [13,18] This inhibition, which is manifested as the serration in the flow curves (Fig 2), resulted in a strength increment (∆σss ), which is defined as follows: ∆σss=HCα (4) where C is the solute concentration (in wt%), and α and H are the experimental constants; Theα and H values measured at a strain rate of 10 s for a 5xxx series Al alloy were 1.14 and 13.8 MPa/ wt%, respectively, [19] Because the yield strength of Al–Mg alloys is insensitive to strain rates of 10 4–10 s [20], the aforementioned H and α values were used to estimate the ∆σss value in this study Therefore, the ∆σss value associated with the solution effect could be evaluated by determining the amount (C) of Mg dissolved in the α-Al phase (this was done using the method Fig Optical image showing the spatial distribution and size of the Mg-rich precipitates in the alloy Al–5Mg explained in Section 2.2) When an alloying element is added to a solvent, the lattice parameter of the resulting alloy differs from that of the solvent, as evidenced by the peak shifts in the XRD patterns Fig 5(a) shows the XRD spectra obtained from the 2θ scanning of the Al–xMg alloys As can be seen from the figure, the diffraction peaks were shifted to the left (see the inset of (a)), indicating that the lattice parameter of the α-Al phase increased with the increase in the Mg content The change in the lattice parameter was linearly proportional to the amount of Mg dissolved in the α-Al phase [25]; this change could, therefore, be used for determining the amount of B.-H Lee et al / Materials Science & Engineering A 657 (2016) 115–122 119 Fig Increase in the strength of the Al–xMg alloys resulting because of grain refinement (in yellow) and solution effects (in blue) For comparison, the experimentally determined increase in strength (in brown) is also shown (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) as determined by substituting the experimentally measured lattice parameters for a in Eq (5) A comparison of Tables and reveals that the Mg contents determined from the chemical and XRD analyses differed only slightly These slight differences are probably attributable to the Mg consumed in forming the Mg-rich precipitates (see Fig 4) The ∆σss values obtained by substituting the measured Mg contents (Table 2) for x in Eq (5) were 48–170 MPa Therefore, the net increment in the strength could be determined by adding the increases attributable to the grain boundaries (Δsgb) and the solid solution ( ∆σss ), as shown graphically in Fig Although slightly overestimated, the predicted increase in the yield strength of the alloys was in keeping with the measured values Therefore, it could be concluded that the strengthening of the Al–xMg alloys, while partly attributable to grain refinement, was predominantly because of the solution effects associated with the addition of Mg Fig (a) Analysis of XRD peaks of the Al–xMg alloys The inset shows the superposition of the (220) peaks of the various alloys In order to measure the lattice parameter, the diffraction peaks corresponding to Kα2 radiation were eliminated from the graph (b) Dependence of the lattice parameter of the alloys on the Mg content of the α-Al phase [21–24] The broken line and the equation shown in the figure correspond to the line fitted to the data reported by Ellwood et al [21] and used here to estimate the amount of Mg dissolved in the alloys Mg dissolved in the α-Al phase of the alloys Fig 5(b) shows the dependence of the lattice parameter of the alloys on the Mg content of the α-Al phase Despite minor scattering in the data, the lattice parameter (a) increased linearly with the amount of Mg (x) added; this linear dependence can be described by the following equation: a (Å) = 0.00452x + 4.048 (5) Table shows the amount of Mg in the α-Al phase of the alloys, Table Calculated Mg concentrations in the Al–xMg alloys a (Å) Mg (at%) Mg (wt%) Pure Al Al-3Mg Al-5Mg Al-7Mg Al-10Mg 4.0495 0 4.0644 3.30 2.98 4.0716 4.88 4.42 4.0819 7.16 6.50 4.0945 9.96 9.06 3.3 Effect of Mg addition on plastic deformation From the viewpoint of dislocation motion, the strength of an alloy depends on the ability of the structure to prevent such motion The ductility depends, in contrast, on the ability of the structure to generate and accumulate dislocations in a significant density Strength and ductility are, therefore, contrasting, mutually exclusive properties However, the high-strength Al–xMg alloys also exhibited high ductilities (Fig 2) This may be attributed to the work-hardening ability, which affected the ductility of the alloys; the work-hardening ability is characterized by the n value of the Hollomon relation The magnitude of n was determined from the slope of the logarithmic plot of the true stress ( σt ) versus the true strain ( εt ) obtained from the instantaneous work-hardening curves constructed for the various samples (see Fig 7(a)) However, the log st –log εt plots resulting from the Hollomon analysis of the samples (Fig 7(a)) were not linear but indicated multiple-stage workhardening behavior prior to necking; in the case of pure Al, the slope of the log st–logεt curve increased in the initial stage of plastic flow (region I in the inset of Fig 7(a)) and decreased with further increases in the strain (region II in the inset of Fig 7(a)) The slope of this second work-hardening stage (region II) increased gradually with the increase in the amount of Mg added In fact, for Mg contents greater than 3%, the log st–log εt curves 120 B.-H Lee et al / Materials Science & Engineering A 657 (2016) 115–122 Fig (a) Logarithmic plot of true stress versus true strain for pure Al and the Al– xMg alloy samples, showing the multiple-stage work-hardening behavior associated with plastic flow (b) Dependence of the n value on the Mg content For comparison, the n values reported in [28] are also plotted exhibited three typical stages of work hardening; the work hardening rate increased rapidly in the first stage, gradually in the second stage, and decreased slightly in the final stage The n value determined from the third-stage work-hardening curve (region III in the inset of Fig 7(a)) is particularly important, since a reduction in its magnitude is associated with the annihilation of dislocations; this annihilation results from the dynamic recovery that occurs prior to necking [26,27] As such, the magnitude of the n value corresponding to the third stage has traditionally been used as a measure of tensile ductility; in this study, the n value was determined as a function of the Mg content (Fig 7(b)) As the figure shows, this value increased significantly from 0.21 to 0.32, with the addition of 3% Mg The rate of increase slowed, however, with the addition of more Mg and increased by only 0.05 for a Mg content of 10% This tendency was very similar to those observed by Kang et al [28] The difference in the work-hardening behaviors of pure Al and the Al–xMg alloys (Fig 7) suggests that the dislocation motion associated with deformation in these materials was different The dislocation microstructures of the samples subjected to a deformation, ε, of 10% were examined using TEM Fig shows the different dislocation microstructures that developed in these samples: (1) in the case of pure Al (Fig 8(a)), the deformation promoted the formation of sharply defined cells with interiors that were nearly dislocation free and walls that had a high dislocation density For materials with such a cell structure, the dislocations generated during further deformation not accumulate within the cells but sink through the cell boundaries via cross-slip A cellstructured metal has, therefore, a low dislocation storage capacity, and the dislocation density remains fairly constant even with an increase in the strain This was manifested by the low degree of work hardening (Fig 2(a)), as evidenced by a low n value (Fig (b)), resulting in strain localization or necking (2) The dislocation structure changed gradually with the addition of Mg, such that the dislocation cell formation was suppressed, and dislocations began to tangle in the Al–3Mg alloy sample (Fig 8(b)) This change was also reflected as an increase in the slope of the second-stage workhardening curve in Fig 7(a) and the flow curves in Fig 2(a) (3) For Mg contents greater than 3%, the dislocation structure changed significantly, with the dislocations becoming tangled and forming high-density (dislocation) band structures (see Fig 8(c) and (d)) These observations of the dislocation structure were consistent with the results of previous studies; the dislocation substructure in the deformed Al–Mg alloy samples (with more than 3% Mg) consisted of a fairly uniform distribution of dislocation tangles [29] The change in the dislocation structure with the increase in the Mg content suggests that (1) the Mg dissolved in the Al lattice inhibits cell formation by making cross-slip more difficult and (2) the dislocation tangles or band structures in the Al–Mg alloy are unable to leave their primary slip plane and eventually pile up in this plane Therefore, the tangled dislocation structure led to the work hardening of the deformed region and the dispersion of the deformation to the neighboring regions, thereby reducing the degree of strain localization and promoting spread-out deformation Therefore, Al alloys with high Mg contents, in spite of having high strengths, can accommodate plastic strains through the dispersion of the deformation of the Al grains The tensile test results and TEM observations indicated that the ductility of a material is strongly dependent on the magnitude of the n value and the mode of dislocation motion Furthermore, the mode of dislocation motion varied with the value of γSF [13,36], which is composition dependent This dependence was confirmed by calculating the γSF values of the Al–xMg alloys As can be seen in Fig 9, the magnitude of γSF decreased with the increase in the Mg content, and, in spite of minor deviations, exhibited the same overall trend as that reported in previous studies [9,30–35] The magnitude of γSF determines the ease with the cross-slip of screw dislocations occurs in Al alloys; owing to the high γSF value ( 150 mJ m 2) of pure Al, the partial dislocations that move under a shear stress can easily recombine into perfect dislocations, facilitating the cross-slip of dislocations and hence the formation of dislocation cell structures [13,28] Meanwhile, Mg, when added to Al in small amounts, decreased the magnitude of γSF Owing to the relatively low γSF value, partial dislocations were widely spaced in the Al–xMg alloys with high Mg contents; this wide spacing resulted in the recombination of these partials becoming difficult, thus limiting cross-slip Therefore, the dislocations tended to move along their slip planes and formed both tangled dislocation and dislocation band structures [13] These structures increased the dislocation-storage capacity, leading to the strain hardening of the Al grains, which was manifested as a high degree of work hardening during plastic flow and reduced the tendency for strain localization Conclusions Adding small amounts of Mg to Al resulted in Al–xMg alloys that had significantly higher strength and plasticity than those of pure Al The results of the mechanical tests, microstructural B.-H Lee et al / Materials Science & Engineering A 657 (2016) 115–122 121 Fig TEM images showing the dislocation structure of (a) pure Al and the Al–xMg alloy samples with Mg contents of (b) 3%, (c) 7%, and (d) 10% (e) Magnified view of (d) The micrograph in (a) was recorded along the exact o1104 zone axis, while all the others were recorded from the zone tilted slightly from o 1104 (two-beam condition) accommodates large plastic strains through the spreading of the shear deformation of the Al grains This is what results in a simultaneous improvement in both the strength and the ductility in the case of Al–xMg alloys Acknowledgment This work was supported by the Fundamental R&D Program for Core Technology of Material (Grant no 10047981), which is funded by Korea Evaluation Institute of Industrial Technology, Republic of Korea References Fig Dependence of the γSF value of the Al–xMg alloys on the Mg content For comparison, the experimentally measured γSF values of various Al–xMg alloys are also shown [9,30–35] observations, and atomic simulations performed in this study may be summarized as follows: Mg, when added to Al, occupies substitutional sites such as point defects and acts as an obstacle to gliding dislocations Therefore, Al–xMg alloys exhibit higher strengths than that of pure Al The degree of strengthening, owing to this substitution, varies with the amount of Mg dissolved in the matrix Meanwhile, alloying Al with Mg leads to a decrease in the γSF value and makes cross-slip difficult, as manifested by an increase in the n value during plastic flow This increased workhardening ability, in turn, alleviates the strain localization, which [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] S.R Kalidindi, Int J Plast 14 (1998) 1265–1277 A Rohatgi, K.S Vecchio, G.T Gray, Metall Mater Trans A 32 (2001) 135–145 F Barlat, M.V Glazov, J.C Brem, D.J Lege, Int J Plast 18 (2002) 919–939 Y.H Zhao, Y.T Zhu, X.Z Liao, Z Horita, T.G Langdon, Appl Phys Lett 89 (2006) 121906 R.A Antunes, M.C.L de Oliveira, Mater Des 63 (2014) 247–256 N Medvedeva, M Park, D.C Van Aken, J.E Medvedeva, J Alloy Compd 582 (2014) 475–482 N Cabanas, J Penning, N Akdut, B De Cooman, Metall Mater Trans A 37 (2006) 3305–3315 K.-T Park, K.G Jin, S.H Han, S.W Hwang, K Choi, C.S Lee, Mater Sci Eng A 527 (2010) 3651–3661 T Schulthess, P Turchi, A Gonis, T.-G Nieh, Acta Mater 46 (1998) 2215–2221 M.I Mendelev, M Asta, M.J Rahman, J.J Hoyt, Philos Mag 89 (2009) 3269–3285 E Polak, G Ribiere, Rev Fr Inform Rech Oper Ser Rouge (1969) 35–43 J.-W Jiang, A.M Leach, K Gall, H.S Park, T Rabczuk, J Mech Phys Solids 61 (2013) 1915–1934 G.E Dieter, D Bacon, Mechanical Metallurgy, Metric, McGraw-Hill, New York, 1986 122 B.-H Lee et al / Materials Science & Engineering A 657 (2016) 115–122 [14] Z Kovács, J Lendvai, G Vörös, Mater Sci Eng A 279 (2000) 179–184 [15] H Dierke, F Krawehl, S Graff, S Forest, J Šachl, H Neuhäuser, Comput Mater Sci 39 (2007) 106–112 [16] Y Brechet, Y Estrin, Acta Metall Mater 43 (1995) 955–963 [17] S Thangaraju, M Heilmaier, B.S Murty, S.S Vadlamani, Adv Eng Mater 14 (2012) 892–897 [18] D Hull, D.J Bacon, Introduction to Dislocations, fifth ed, Pergamon Press, Oxford, 1984 [19] Ø Ryen, B Holmedal, O Nijs, E Nes, E Sjölander, H.-E Ekström, Metall Mater Trans A 37 (2006) 1999–2006 [20] T Mukai, K Higashi, S Tanimura, Mater Sci Eng A 176 (1994) 181–189 [21] E Ellwood, J Inst Met 80 (1952) [22] P.P.J.E Dorn, T.E Tietz, J Met (1950) [23] D Poole, H Axon, J Inst Met 80 (1952) [24] W Helfrich, R Dodd, Trans Metall Soc AIME 224 (1962) 757-& [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] V Lubarda, Mech Mater 35 (2003) 53–68 D Kuhlmann-Wilsdorf, Mater Sci Eng A 113 (1989) 1–41 R Madec, B Devincre, L.P Kubin, Scr Mater 47 (2002) 689–695 S.B Kang, C.Y Lim, H.W Kim, Trans Korean Soc Automot Eng (1994) 95–102 G.W.J Waldron, Acta Metall 13 (1965) 897–906 P Dobson, P Goodhew, R Smallman, Philos Mag 16 (1967) 9–22 M Soliman, J Mater Sci 28 (1993) 4483–4488 S Kritzinger, P Dobson, R Smallman, Philos Mag 16 (1967) 217–229 R Kapoor, J Singh, J Chakravartty, Mater Sci Eng A 496 (2008) 308–315 T Morishige, T Hirata, T Uesugi, Y Takigawa, M Tsujikawa, K Higashi, Scr Mater (2011) 355–358 Y Iwahashi, Z Horita, M Nemoto, T.G Langdon, Metall Mater Trans A 29 (1998) 2503–2510 S.S Firouzabadi, M Kazeminezhad, Mater Des 36 (2012) 804–808 ... Table Chemical compositions of the Al? ??xMg alloy samples used in this study (in wt%) Composition Mg Si Fe Ti Al Al– 3Mg Al? ?? 5Mg Al? ?? 7Mg Al? ??1 0Mg 3. 1 5. 1 7 .3 10. 2 0. 03 0. 03 0.04 0. 03 0. 13 0. 13 0.12 0.12... concentrations in the Al? ??xMg alloys a (Å) Mg (at%) Mg (wt%) Pure Al Al- 3Mg Al- 5Mg Al- 7Mg Al- 1 0Mg 4.04 95 0 4.0644 3. 30 2.98 4. 071 6 4.88 4.42 4.0819 7. 16 6 .50 4.09 45 9.96 9.06 3. 3 Effect of Mg addition on... Des 63 (2014) 2 47? ?? 256 N Medvedeva, M Park, D.C Van Aken, J.E Medvedeva, J Alloy Compd 58 2 (2014) 4 75 ? ??482 N Cabanas, J Penning, N Akdut, B De Cooman, Metall Mater Trans A 37 (2006) 33 05? ? ?33 15 K.-T