- Email: [email protected]
Effect of dislocation-twin boundary interaction on deformation by twin boundary migration
Materials Science & Engineering A 662 (2016) 95–99
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Short communication
Effect of dislocation-twin boundary interaction on deformation by twin boundary migration Yunchang Xin n, Liangchen Lv, Houwen Chen, Cong He, Huihui Yu, Qing Liu n School of Materials Science and Engineering, Chongqing University, Chongqing 400030, China
art ic l e i nf o
a b s t r a c t
Article history: Received 13 October 2015 Received in revised form 13 March 2016 Accepted 14 March 2016 Available online 15 March 2016
̅ } twin boundaries (TBs) and 〈a〉 dislocations on deformation The effect of interaction between {1012 behavior of TBs migration process was studied. This TB-dislocation interaction greatly enhances activation stress for TB motion and retards TB migration during reloading. Interestingly, a severe TB-dislocation interaction improves strain hardening at the early stage, while reduces the peak hardening rate at the latter stage. High resolution transmission electron microscopy results indicate that severe TB-dislocation interaction greatly damages the coherence of TBs and might induce curving of TBs. Migration of the TBs that intensively interact with dislocations leaves behind low angle boundaries at the initial TBs. & 2016 Elsevier B.V. All rights reserved.
Keywords: Dislocations Twin boundary migration Mechanical properties Strain hardening
1. Introduction Twin boundaries (TBs) play an important role in plastic deformation and ultimately in controlling the forming ability and mechanical properties of many engineering materials [1]. Notable examples are the Mg alloys in which twinning constitutes one of the main deformation mode at room temperature. The formability, yield strength and mechanical anisotropy are closely related to ̅ } TBs have been extensively used to tailor twinning [2–4]. {1012 mechanical properties and working ability in recent studies [5–7]. It was found that yield strength of a hot-rolled AZ31 plate can be significantly enhanced without a drop of ductility by introduction ̅ } TBs [8]. The reorientation by {1012 ̅ } of a large number of {1012 twinning can effectively improve hot rolling ability and reduces tension-compression yield asymmetry [8,9]. Recently, Cui et al. reported that Mg alloys with an enhanced damping capacity can ̅ } TBs [6,8,9]. be prepared by introducing {1012 For nano-structured fcc or hcp metals, dislocations interaction and accumulation at TBs have been reported to significantly affect the strength, ductility and strain hardening. Extensive studies demonstrated that, depending on the characteristics of the dislocations and the driving stress, possible dislocation reactions at TBs included cross-slip into the twinning plane to cause twin growth or detwinning, formation of sessile dislocations at TBs, and transmission across the TBs [10,11]. For metals containing the pre̅ } TBs, TBs migration, i.e. twin growth or detwinning, existing {1012 n
Corresponding authors. E-mail addresses: [email protected] (Y. Xin), [email protected] (Q. Liu).
http://dx.doi.org/10.1016/j.msea.2016.03.061 0921-5093/& 2016 Elsevier B.V. All rights reserved.
is an important (or even the predominant) deformation mode under certain loading conditions. The TBs motion is often accompanied by intensive TB-dislocation interaction during detwinning in nano-structured fcc metals [12] or twinning in hcp metals [5], or cyclic loading of Mg alloys [13]. Generally, a TBs migration process involves twinning dislocations gliding along coherent TBs and subsequently climbing along facets [14]. It is suspected that the TB-dislocation interaction will affect the motion of twinning dislocations, and ultimately vary mechanical properties and strain hardening. Therefore, understanding how the mechanical behavior is influenced by the TB-dislocation interaction during deformation by TBs motion is of great importance. Although the effect of TB-dislocation interaction on mechanical properties and strain hardening during slip predominant deformations has been extensively studied and well understood, that of the TB-dislocation interaction on mechanical behavior of a TB migration predominant deformation is hardly reported and unclear. In the present study, a twinned Mg AZ31 plate was subjected to a slip predominant deformation to allow an intensive interac̅ } TBs. Then, a TBs migration tion of 〈a〉 dislocations with {1012 predominant deformation (detwinning) was initiated by reloading, with the aim to understand the effect of this TB-dislocation interaction on mechanical properties and strain hardening. The corresponding mechanisms were studied and discussed.
2. Experiments and methods A hot-rolled Mg AZ31 thick plate with fully recrystallized grains and a typical basal texture was used. Blocks of 30 mm (ND)
96
Y. Xin et al. / Materials Science & Engineering A 662 (2016) 95–99
(RD) 9 mm (ND) were cut for recompression along the ND to ̅ } twins. A strain rate of 10 3 s 1 was initiate detwinning of {1012 used for all the mechanical tests. Each mechanical test was repeated three times. For comparison, mechanical behavior under compression along the ND of the as-used plate (the designated original plate) and the TD2.0%-RD0% were also measured. To reveal the deformation behavior during reloading, in-situ electron back-scattered diffraction (EBSD) mapping using a step size of 0.5 mm was conducted on a scanning electron microscope (SEM, Zeiss AURIGA) equipped with a HKL-EBSD system. Highangle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and high resolution transmission electron microscopy (HRTEM) conducted on FEI Tecnai G2 F20 TEM were used to further investigate the microstructure of pre-strained samples. For preparation of specimens for TEM examination, the mechanically grinded thin foils were punched into discs of 3 mm in diameter and were subsequently thinned to 50 mm. A twin-jet electro-polishing of the thin foils at 30 °C and 50 V was performed on a Struers Tenupol-5 twin-jet electro-polisher using a solution containing 3 vol% perchloric acid in absolute ethanol. Finally the foils were cleaned for 10 min by a Gatan ion polishing system using an accelerating voltage of 3 kV and at 70 °C.
3. Results and discussion
Fig. 1. (a) A schematic diagram showing the pre-straining to generate intensive ̅ } TB with 〈a〉 dislocation and the preparation of specimen for interaction of {1012 ND reloading and (b) stress-strain curves under compression along the ND.
35 mm(TD) 90 mm (RD) were cut and compressed 2.0% along the TD (the designated TD2.0%-RD0%) to generate a large number ̅ } TBs. Here, RD, TD and ND represent the rolling direction, of {1012 the transverse direction and the normal direction of the as-used plate, respectively. As shown in Fig. 1a, the specimens for tension were cut from the center of the twinned blocks and stretched along the RD to 2.0% (the designated TD2.0%-RD2.0%), 3.5% (the designated TD2.0%-RD3.5%) and 5.5% (the designated TD2.0%RD5.5%), respectively. The RD tension was used to start prismatic ̅ } twins, and, slip of 〈a〉 dislocations in both the matrix and {1012 ̅ } TBs with 〈a〉 hence, generated an intensive interaction of {1012 dislocations. Afterwards, small blocks of 7 mm (TD) 7 mm
Fig. 1b shows stress-strain curves of compression along the ND. The curve of TD2.0%-RD0% has a plateau, a typical manifestation of ̅ } twins. The RD deformation dominated by detwinning of {1012 re-tension significantly increases the yield stress and slightly reduces the elongation, though the peak stress being similar. As seen in Table 1, the yield stress increases from 50 MPa to 137 MPa when the RD strain is up to 5.5%. As seen in Fig. 2a, a continuously decreasing strain hardening rate exists during ND compression of the original plate, which is related to a basal slip predominant deformation [8]. Strain hardening curve of the TD2.0%-RD0% is characterized by an obvious peak. When pre-deformation also includes RD tension, a drop of the peak hardening rate is observed; and when the RD strain is up to 5.5%, this peak in the strain hardening curve is not obvious. In-situ EBSD analyses of the deformation behavior of TD2.0%RD5.5% before and after 1.5% compression along the ND are shown in Fig. 3. As seen in Fig. 3a, many lamellae (the bands in green or ̅ } twins exist in the inverse pole figure blue) identified as {1012 ̅ } twins narrow or map before ND compression. A part of {1012 disappear after ND compression. It is interesting that many low angle boundaries with misorientations of 3–8° (denoted by the ̅ } yellow arrows in Fig. 3b) appear just at the sites where the {1012 TBs disappear. This can be clearly shown in Fig. 3c. The migration ̅ } TBs leaves behind low angle boundaries with a 8° of {1012 misorientation at the initial TB. Generally, for a hot-rolled Mg alloy plate with a strong basal ̅ } twinning pretexture, compression along the TD is a {1012 ̅ } twins are dominant deformation and basal poles of the {1012 close to the compression direction [5]. Therefore, the TDTable 1 Yield stress, peak stress and elongation of samples during compression along the ND.
TD0%-RD0% TD2.0%-RD0% TD2.0%-RD2.0% TD2.0%-RD3.5% TD2.0%-RD5.5%
Yield stress/MPa
Peak stress/MPa
Elongation/%
166 50 89 114 137
283 282 282 285 286
12.3 10.9 10.5 10.3 9.8
Y. Xin et al. / Materials Science & Engineering A 662 (2016) 95–99
Fig. 2. Strain hardening curves of compression along the ND. Here, σ and σ0.2 refer to stress and yield stress, respectively.
compression of AZ31 hot-rolled plate in the present study gen̅ } twins with the (0002) poles close erates a large number of {1012 to the TD. As shown in Fig. 4a, both orientation of the matrix and ̅ } twins after TD-compression are favorable for that of the {1012 the onset of prismatic slip of 〈a〉 dislocations with a Burgers vector 1 ̅ during re-tension along the RD [15]. The RD re-tension of 3 1120
̅ } allows an intensive interaction of 〈a〉 dislocations with the {1012 ̅ } TBs in the pre-strained samples was TBs. Microstructure of {1012 further examined by HAADF-STEM and HRTEM and the results are given in Fig. 4b–d. In a HAADF-STEM image, dislocations generally appear as light lines [16]. As seen in Fig. 4b, many dislocations ̅ } (denoted by the arrows) pile up at the TBs or within the {1012 twins and matrix. This is a clear evidence for the intensive TBdislocation interaction in the pre-twinned and RD tensioned samples. The HRTEM in Fig. 4c shows that the coherent segment of ̅ } TB without intensive interaction with dislocations has a {1012 good coherence with a sharp and straight interface, while that of a
97
̅ } TB experiencing an intensive TB-dislocation interaction {1012 (Fig. 4d) shows evidently rough interface. A curved region even appears as denoted by the arrow in Fig. 4d. This supports the conclusion that the TB-dislocation interaction greatly damages the ̅ } TBs. In addition, the inset figure in Fig. 4d coherence of {1012 clearly shows an obvious deviation of the twin boundary from the twinning plane (about 4.5° deviation). The deviation was mea̅ } twin sured by the Digital Micrograph software. Usually, the {1012 ̅ } and the matrix show a mirror symmetry relationship to the {1012 ̅ } planes both in the plane, so the diffraction patterns of {1012 matrix and the twin should be overlapped. However, it is not this case in this study. This type deviation is often considered to result from TB-dislocation interaction. Interestingly, as seen in Fig. 3, the TBs migration leaves behind many low angle boundaries, indicating that the defects generated by TB-dislocation interaction is sessile during TBs migration. A TB migration process involves twinning dislocations gliding along coherent TBs and subsequently climbing along facets and is suspected to possess a much lower activation stress than that for twin nucleation [17–19]. Therefore, a detwinning predominant deformation in Mg alloys is often reported to have a much lower yield stress [8,13]. This is evident in Fig. 1b, where yield stress of the TD2.0%-RD0% is much lower than that of the original plate during ND compression. As demonstrated in Fig. 4d, the TB-dis̅ } TBs, location interaction greatly damages the coherence of {1012 which is suspected to make it more difficult for twinning dislocations motion. In addition, a considerable plastic deformation along the RD (i.e., RD5.5%) causes not only dislocation pile-up at TBs, but also dislocation accumulation within a-Mg matrix, and thus strain-hardened matrix can also contribute to improve the yield strength of materials. As seen in Fig. 1b, a 5.5% RD tension increases yield stress of the ND compression from 50 MPa to 137 MPa. However, yield stress of the RD2.0%-TD5.5% is still lower than that of the original plate (166 MPa) under ND compression, which is a basal slip predominant deformation [8]. This indicates that yielding during ND compression of TD2.0%-RD5.5% is also induced by detwinning of ̅ } twins. According to the stress-strain curve of TD2.0%-RD0%, {1012 detwinning is exhausted at a plastic strain of about 2.0%. However, the stress of TD2.0%-RD5.5% exceeds 166 MPa at a plastic strain of around 0.75%, implying that slip is already active at low strain. In fact, the stress-strain curve of TD2.0%-RD5.5% looks like a mixture of that of the original plate and the TD2.0%-RD0%. The in-situ EBSD ̅ } examination in Fig. 3 also shows that a large number of {1012
Fig. 3. In-situ EBSD analyses during 1.5% compression along the ND of TD2.0%-RD5.5%: (a) inverse pole figure maps (a) before ND compression and (b) after ND compression; (c) an example showing the presence of low angle boundary after TB migration. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
98
Y. Xin et al. / Materials Science & Engineering A 662 (2016) 95–99
Fig. 4. (a) Schematic diagram showing the orientations of twins and matrix after TD pre-compression; (b) HAADF-STEM image showing the accumulation of dislocations at ̅ } TBs and within twins and matrix of the pre-twinned sample with 5.5% tension along the RD; HRTEM image of (c) a {1012 ̅ } TB without intensive TB-dislocation {1012 ̅ } TB with intensive TB-dislocation interaction. The insert in (b) is the selection area diffraction pattern acquired from position ① in (b). interaction and (d) a {1012
twins still exist after 1.5% ND compression of TD2.0%-RD5.5%. This further confirms the retarding of TBs migration during ND compression. Under ND compression of the pre-twinned samples, orientation of the matrix favors the basal slip and the prismatic 〈a〉 slip within the twins has a high Schmid factor. Therefore, it is suspected that both the two types of slip might start during ND compression of the TD2.0%-RD5.5%. As seen in Fig. 2, an intensive peak appears in the strain hardening curve of TD2.0%-RD0%, a typical feature of deformation ̅ } twins. Wu et al. considered dominated by detwinning of {1012 that a complete detwinning of twins required the activation of slip to accommodate the strain, and this transition from detwinning with a low activation stress to slip that requires a much higher activation stress generated the peak in strain hardening rate [20]. As shown in Fig. 1b, a distinct transition from detwinning to slip exists in TD2.0%-RD0%, while this transition is not obvious in curves of TD2.0%-RD3.5% and TD2.0%-RD5.5%. This is probably the reason for the reduced peak hardening rate in TD2.0%-RD3.5% or the absence of a hardening peak in TD2.0%-RD5.5%. However, in fact, the strain hardening rate during detwinning in TD2.0%-RD0%
(0 o σ−σ0.2 o33 MPa, as denoted by the dash line in Fig. 2) is lower than that of the pre-twinned samples with further RD re-tension. An increased RD strain further enhances the strain hardening rate at this stage. Obviously, the TB-dislocation interaction effectively improves hardening of deformation by TBs migration. This indicate that the TB-dislocation interaction plays an important role in strain hardening of twinning and in cyclic loading of Mg alloys, where an obvious strain hardening with strain exists. Generally, ̅ } TBs are not considered to be strong barriers for dislocation {1012 slips. Thus, TB-dislocation interaction is found to hardly improve the strain hardening rate during a slip predominant deformation in hcp metals [21,22]. In spite of the weak blockage effect of TBs on dislocation slip, the present study shows that dislocations are strong barriers for TBs motion. In a word, the TB-dislocation interaction poses substantially different effects on the strain hardening of deformation by slips and that of deformation by TBs migration. 4. Conclusion
̅ } TBs and 〈a〉 In summary, an intensive interaction of {1012
Y. Xin et al. / Materials Science & Engineering A 662 (2016) 95–99
dislocations was produced by a TD pre-compression and a subsequent RD re-tension. The influence of this TB-dislocation inter̅ } TBs action on mechanical behavior of a deformation by {1012 migration was studied. Several conclusions are reached as follows: (1) TB-dislocation interaction greatly damages coherence of the ̅ } TBs, leading to a much higher activation stress and {1012 retarding of TBs motion. The TB-dislocation interaction improves the strain hardening at the early stage, while reduces the peak hardening rate at the latter stage. (2) The migration of TBs that intensively interacts with dislocations leaves behind many low angle boundaries at the initial TBs.
Prime novelty statement Dislocation-TB interaction has been reported to significantly affect the strength, ductility and strain hardening of metals. Although the influence of TB-dislocation interaction on mechanical behavior of a slip predominant deformations has been well understood, that of the TB-dislocation interaction on mechanical behavior of deformation by TB migration is hardly reported. In the present study, the effect of interaction between TBs and prismatic 〈a〉 slip on detwinning of twins was studied, with the aim to understand the effect of this TB-dislocation interaction on mechanical behavior of TB migration predominant deformation. Our results shows that TB-dislocation interaction damages the coherence of TBs and induces TBs curving, leading to a much higher activation stress and a delay of TBs motion. The TB-dislocation interaction improves strain hardening at the early stage, while reduces the peak hardening rate at the latter stage. Migration of the TBs that intensively interact with dislocations leaves behind low angle boundaries.
99
Acknowledgment This study is co-supported by National Key Basic Research Program of China (2013CB632204) and Natural Science Foundation of China (51371203, 51571041, 51131009 and 51421001), Fundamental Research Funds for the Central Universities (106112013CDJZR13130032 and 106112015CDJXZ138803).
References [1] K. Lu, L. Lu, S. Suresh, Science 324 (2009) 349–352. [2] M.R. Barnett, Mater. Sci. Eng. A 464 (2007) 1–7. [3] A. Levinson, R.K. Mishra, R.D. Doherty, S.R. Kalidindi, Acta Mater. 61 (2013) 5966–5978. [4] N.V. Dudamell, I. Ulacia, F. Gálvez, S. Yi, J. Bohlen, D. Letzig, I. Hurtado, M. T. Pérez-Prado, Acta Mater. 59 (2011) 6949–6962. [5] Y. Xin, M. Wang, Z. Zeng, M. Nie, Q. Liu, Scr. Mater. 66 (2012) 25–28. [6] Y. Cui, Y. Li, S. Sun, H. Bian, H. Huang, Z. Wang, Y. Koizumi, A. Chiba, Scr. Mater. 101 (2015) 8–11. [7] S.H. Park, S.-G. Hong, C.S. Lee, Scr. Mater. 62 (2010) 666–669. [8] Y. Xin, X. Zhou, Q. Liu, Mater. Sci. Eng. A 567 (2013) 9–13. [9] Y. Xin, M. Wang, Z. Zeng, G. Huang, Q. Liu, Scr. Mater. 64 (2011) 986–989. [10] E. Ma, Y.M. Wang, Q.H. Lu, M.L. Sui, L. Lu, K. Lu, Appl. Phys. Lett. 85 (2004) 4932. [11] C.J. Shute, B.D. Myers, S. Xie, T.W. Barbee, A.M. Hodge, J.R. Weertman, Scr. Mater. 60 (2009) 1073–1077. [12] G.J. Tucker, S.M. Foiles, Int. J. Plast. 65 (2015) 191–205. [13] C. Guillemer, M. Clavel, G. Cailletaud, Int. J. Plast. 27 (2011) 2068–2084. [14] J.W. Christian, S. Mahajan, Prog. Mater. Sci. 39 (1995) 1–157. [15] S.R. Agnew, O. Duygulu, Int. J. Plast. 21 (2005) 1161–1193. [16] D.D. Perovic, C.J. Rossouw, A. Howie, Ultramicroscopy 52 (1993) 353–359. [17] S.R.K. Ayman, A. Salem, Roger D. Doherty, Scr. Mater. 46 (2002). [18] Y. Xin, X. Zhou, L. Lv, Q. Liu, Mater. Sci. Eng. A 606 (2014) 81–91. [19] A. Ostapovets, R. Gröger, Modell. Simul. Mater. Sci. Eng. 22 (2014) 025015. [20] L. Wu, S.R. Agnew, D.W. Brown, G.M. Stoica, B. Clausen, A. Jain, D.E. Fielden, P. K. Liaw, Acta Mater. 56 (2008) 3699–3707. [21] B. Song, R. Xin, G. Chen, X. Zhang, Q. Liu, Scr. Mater. 66 (2012) 1061–1064. [22] M. Knezevic, A. Levinson, R. Harris, R.K. Mishra, R.D. Doherty, S.R. Kalidindi, Acta Mater. 58 (2010) 6230–6242.
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Short communication
Effect of dislocation-twin boundary interaction on deformation by twin boundary migration Yunchang Xin n, Liangchen Lv, Houwen Chen, Cong He, Huihui Yu, Qing Liu n School of Materials Science and Engineering, Chongqing University, Chongqing 400030, China
art ic l e i nf o
a b s t r a c t
Article history: Received 13 October 2015 Received in revised form 13 March 2016 Accepted 14 March 2016 Available online 15 March 2016
̅ } twin boundaries (TBs) and 〈a〉 dislocations on deformation The effect of interaction between {1012 behavior of TBs migration process was studied. This TB-dislocation interaction greatly enhances activation stress for TB motion and retards TB migration during reloading. Interestingly, a severe TB-dislocation interaction improves strain hardening at the early stage, while reduces the peak hardening rate at the latter stage. High resolution transmission electron microscopy results indicate that severe TB-dislocation interaction greatly damages the coherence of TBs and might induce curving of TBs. Migration of the TBs that intensively interact with dislocations leaves behind low angle boundaries at the initial TBs. & 2016 Elsevier B.V. All rights reserved.
Keywords: Dislocations Twin boundary migration Mechanical properties Strain hardening
1. Introduction Twin boundaries (TBs) play an important role in plastic deformation and ultimately in controlling the forming ability and mechanical properties of many engineering materials [1]. Notable examples are the Mg alloys in which twinning constitutes one of the main deformation mode at room temperature. The formability, yield strength and mechanical anisotropy are closely related to ̅ } TBs have been extensively used to tailor twinning [2–4]. {1012 mechanical properties and working ability in recent studies [5–7]. It was found that yield strength of a hot-rolled AZ31 plate can be significantly enhanced without a drop of ductility by introduction ̅ } TBs [8]. The reorientation by {1012 ̅ } of a large number of {1012 twinning can effectively improve hot rolling ability and reduces tension-compression yield asymmetry [8,9]. Recently, Cui et al. reported that Mg alloys with an enhanced damping capacity can ̅ } TBs [6,8,9]. be prepared by introducing {1012 For nano-structured fcc or hcp metals, dislocations interaction and accumulation at TBs have been reported to significantly affect the strength, ductility and strain hardening. Extensive studies demonstrated that, depending on the characteristics of the dislocations and the driving stress, possible dislocation reactions at TBs included cross-slip into the twinning plane to cause twin growth or detwinning, formation of sessile dislocations at TBs, and transmission across the TBs [10,11]. For metals containing the pre̅ } TBs, TBs migration, i.e. twin growth or detwinning, existing {1012 n
Corresponding authors. E-mail addresses: [email protected] (Y. Xin), [email protected] (Q. Liu).
http://dx.doi.org/10.1016/j.msea.2016.03.061 0921-5093/& 2016 Elsevier B.V. All rights reserved.
is an important (or even the predominant) deformation mode under certain loading conditions. The TBs motion is often accompanied by intensive TB-dislocation interaction during detwinning in nano-structured fcc metals [12] or twinning in hcp metals [5], or cyclic loading of Mg alloys [13]. Generally, a TBs migration process involves twinning dislocations gliding along coherent TBs and subsequently climbing along facets [14]. It is suspected that the TB-dislocation interaction will affect the motion of twinning dislocations, and ultimately vary mechanical properties and strain hardening. Therefore, understanding how the mechanical behavior is influenced by the TB-dislocation interaction during deformation by TBs motion is of great importance. Although the effect of TB-dislocation interaction on mechanical properties and strain hardening during slip predominant deformations has been extensively studied and well understood, that of the TB-dislocation interaction on mechanical behavior of a TB migration predominant deformation is hardly reported and unclear. In the present study, a twinned Mg AZ31 plate was subjected to a slip predominant deformation to allow an intensive interac̅ } TBs. Then, a TBs migration tion of 〈a〉 dislocations with {1012 predominant deformation (detwinning) was initiated by reloading, with the aim to understand the effect of this TB-dislocation interaction on mechanical properties and strain hardening. The corresponding mechanisms were studied and discussed.
2. Experiments and methods A hot-rolled Mg AZ31 thick plate with fully recrystallized grains and a typical basal texture was used. Blocks of 30 mm (ND)
96
Y. Xin et al. / Materials Science & Engineering A 662 (2016) 95–99
(RD) 9 mm (ND) were cut for recompression along the ND to ̅ } twins. A strain rate of 10 3 s 1 was initiate detwinning of {1012 used for all the mechanical tests. Each mechanical test was repeated three times. For comparison, mechanical behavior under compression along the ND of the as-used plate (the designated original plate) and the TD2.0%-RD0% were also measured. To reveal the deformation behavior during reloading, in-situ electron back-scattered diffraction (EBSD) mapping using a step size of 0.5 mm was conducted on a scanning electron microscope (SEM, Zeiss AURIGA) equipped with a HKL-EBSD system. Highangle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and high resolution transmission electron microscopy (HRTEM) conducted on FEI Tecnai G2 F20 TEM were used to further investigate the microstructure of pre-strained samples. For preparation of specimens for TEM examination, the mechanically grinded thin foils were punched into discs of 3 mm in diameter and were subsequently thinned to 50 mm. A twin-jet electro-polishing of the thin foils at 30 °C and 50 V was performed on a Struers Tenupol-5 twin-jet electro-polisher using a solution containing 3 vol% perchloric acid in absolute ethanol. Finally the foils were cleaned for 10 min by a Gatan ion polishing system using an accelerating voltage of 3 kV and at 70 °C.
3. Results and discussion
Fig. 1. (a) A schematic diagram showing the pre-straining to generate intensive ̅ } TB with 〈a〉 dislocation and the preparation of specimen for interaction of {1012 ND reloading and (b) stress-strain curves under compression along the ND.
35 mm(TD) 90 mm (RD) were cut and compressed 2.0% along the TD (the designated TD2.0%-RD0%) to generate a large number ̅ } TBs. Here, RD, TD and ND represent the rolling direction, of {1012 the transverse direction and the normal direction of the as-used plate, respectively. As shown in Fig. 1a, the specimens for tension were cut from the center of the twinned blocks and stretched along the RD to 2.0% (the designated TD2.0%-RD2.0%), 3.5% (the designated TD2.0%-RD3.5%) and 5.5% (the designated TD2.0%RD5.5%), respectively. The RD tension was used to start prismatic ̅ } twins, and, slip of 〈a〉 dislocations in both the matrix and {1012 ̅ } TBs with 〈a〉 hence, generated an intensive interaction of {1012 dislocations. Afterwards, small blocks of 7 mm (TD) 7 mm
Fig. 1b shows stress-strain curves of compression along the ND. The curve of TD2.0%-RD0% has a plateau, a typical manifestation of ̅ } twins. The RD deformation dominated by detwinning of {1012 re-tension significantly increases the yield stress and slightly reduces the elongation, though the peak stress being similar. As seen in Table 1, the yield stress increases from 50 MPa to 137 MPa when the RD strain is up to 5.5%. As seen in Fig. 2a, a continuously decreasing strain hardening rate exists during ND compression of the original plate, which is related to a basal slip predominant deformation [8]. Strain hardening curve of the TD2.0%-RD0% is characterized by an obvious peak. When pre-deformation also includes RD tension, a drop of the peak hardening rate is observed; and when the RD strain is up to 5.5%, this peak in the strain hardening curve is not obvious. In-situ EBSD analyses of the deformation behavior of TD2.0%RD5.5% before and after 1.5% compression along the ND are shown in Fig. 3. As seen in Fig. 3a, many lamellae (the bands in green or ̅ } twins exist in the inverse pole figure blue) identified as {1012 ̅ } twins narrow or map before ND compression. A part of {1012 disappear after ND compression. It is interesting that many low angle boundaries with misorientations of 3–8° (denoted by the ̅ } yellow arrows in Fig. 3b) appear just at the sites where the {1012 TBs disappear. This can be clearly shown in Fig. 3c. The migration ̅ } TBs leaves behind low angle boundaries with a 8° of {1012 misorientation at the initial TB. Generally, for a hot-rolled Mg alloy plate with a strong basal ̅ } twinning pretexture, compression along the TD is a {1012 ̅ } twins are dominant deformation and basal poles of the {1012 close to the compression direction [5]. Therefore, the TDTable 1 Yield stress, peak stress and elongation of samples during compression along the ND.
TD0%-RD0% TD2.0%-RD0% TD2.0%-RD2.0% TD2.0%-RD3.5% TD2.0%-RD5.5%
Yield stress/MPa
Peak stress/MPa
Elongation/%
166 50 89 114 137
283 282 282 285 286
12.3 10.9 10.5 10.3 9.8
Y. Xin et al. / Materials Science & Engineering A 662 (2016) 95–99
Fig. 2. Strain hardening curves of compression along the ND. Here, σ and σ0.2 refer to stress and yield stress, respectively.
compression of AZ31 hot-rolled plate in the present study gen̅ } twins with the (0002) poles close erates a large number of {1012 to the TD. As shown in Fig. 4a, both orientation of the matrix and ̅ } twins after TD-compression are favorable for that of the {1012 the onset of prismatic slip of 〈a〉 dislocations with a Burgers vector 1 ̅ during re-tension along the RD [15]. The RD re-tension of 3 1120
̅ } allows an intensive interaction of 〈a〉 dislocations with the {1012 ̅ } TBs in the pre-strained samples was TBs. Microstructure of {1012 further examined by HAADF-STEM and HRTEM and the results are given in Fig. 4b–d. In a HAADF-STEM image, dislocations generally appear as light lines [16]. As seen in Fig. 4b, many dislocations ̅ } (denoted by the arrows) pile up at the TBs or within the {1012 twins and matrix. This is a clear evidence for the intensive TBdislocation interaction in the pre-twinned and RD tensioned samples. The HRTEM in Fig. 4c shows that the coherent segment of ̅ } TB without intensive interaction with dislocations has a {1012 good coherence with a sharp and straight interface, while that of a
97
̅ } TB experiencing an intensive TB-dislocation interaction {1012 (Fig. 4d) shows evidently rough interface. A curved region even appears as denoted by the arrow in Fig. 4d. This supports the conclusion that the TB-dislocation interaction greatly damages the ̅ } TBs. In addition, the inset figure in Fig. 4d coherence of {1012 clearly shows an obvious deviation of the twin boundary from the twinning plane (about 4.5° deviation). The deviation was mea̅ } twin sured by the Digital Micrograph software. Usually, the {1012 ̅ } and the matrix show a mirror symmetry relationship to the {1012 ̅ } planes both in the plane, so the diffraction patterns of {1012 matrix and the twin should be overlapped. However, it is not this case in this study. This type deviation is often considered to result from TB-dislocation interaction. Interestingly, as seen in Fig. 3, the TBs migration leaves behind many low angle boundaries, indicating that the defects generated by TB-dislocation interaction is sessile during TBs migration. A TB migration process involves twinning dislocations gliding along coherent TBs and subsequently climbing along facets and is suspected to possess a much lower activation stress than that for twin nucleation [17–19]. Therefore, a detwinning predominant deformation in Mg alloys is often reported to have a much lower yield stress [8,13]. This is evident in Fig. 1b, where yield stress of the TD2.0%-RD0% is much lower than that of the original plate during ND compression. As demonstrated in Fig. 4d, the TB-dis̅ } TBs, location interaction greatly damages the coherence of {1012 which is suspected to make it more difficult for twinning dislocations motion. In addition, a considerable plastic deformation along the RD (i.e., RD5.5%) causes not only dislocation pile-up at TBs, but also dislocation accumulation within a-Mg matrix, and thus strain-hardened matrix can also contribute to improve the yield strength of materials. As seen in Fig. 1b, a 5.5% RD tension increases yield stress of the ND compression from 50 MPa to 137 MPa. However, yield stress of the RD2.0%-TD5.5% is still lower than that of the original plate (166 MPa) under ND compression, which is a basal slip predominant deformation [8]. This indicates that yielding during ND compression of TD2.0%-RD5.5% is also induced by detwinning of ̅ } twins. According to the stress-strain curve of TD2.0%-RD0%, {1012 detwinning is exhausted at a plastic strain of about 2.0%. However, the stress of TD2.0%-RD5.5% exceeds 166 MPa at a plastic strain of around 0.75%, implying that slip is already active at low strain. In fact, the stress-strain curve of TD2.0%-RD5.5% looks like a mixture of that of the original plate and the TD2.0%-RD0%. The in-situ EBSD ̅ } examination in Fig. 3 also shows that a large number of {1012
Fig. 3. In-situ EBSD analyses during 1.5% compression along the ND of TD2.0%-RD5.5%: (a) inverse pole figure maps (a) before ND compression and (b) after ND compression; (c) an example showing the presence of low angle boundary after TB migration. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
98
Y. Xin et al. / Materials Science & Engineering A 662 (2016) 95–99
Fig. 4. (a) Schematic diagram showing the orientations of twins and matrix after TD pre-compression; (b) HAADF-STEM image showing the accumulation of dislocations at ̅ } TBs and within twins and matrix of the pre-twinned sample with 5.5% tension along the RD; HRTEM image of (c) a {1012 ̅ } TB without intensive TB-dislocation {1012 ̅ } TB with intensive TB-dislocation interaction. The insert in (b) is the selection area diffraction pattern acquired from position ① in (b). interaction and (d) a {1012
twins still exist after 1.5% ND compression of TD2.0%-RD5.5%. This further confirms the retarding of TBs migration during ND compression. Under ND compression of the pre-twinned samples, orientation of the matrix favors the basal slip and the prismatic 〈a〉 slip within the twins has a high Schmid factor. Therefore, it is suspected that both the two types of slip might start during ND compression of the TD2.0%-RD5.5%. As seen in Fig. 2, an intensive peak appears in the strain hardening curve of TD2.0%-RD0%, a typical feature of deformation ̅ } twins. Wu et al. considered dominated by detwinning of {1012 that a complete detwinning of twins required the activation of slip to accommodate the strain, and this transition from detwinning with a low activation stress to slip that requires a much higher activation stress generated the peak in strain hardening rate [20]. As shown in Fig. 1b, a distinct transition from detwinning to slip exists in TD2.0%-RD0%, while this transition is not obvious in curves of TD2.0%-RD3.5% and TD2.0%-RD5.5%. This is probably the reason for the reduced peak hardening rate in TD2.0%-RD3.5% or the absence of a hardening peak in TD2.0%-RD5.5%. However, in fact, the strain hardening rate during detwinning in TD2.0%-RD0%
(0 o σ−σ0.2 o33 MPa, as denoted by the dash line in Fig. 2) is lower than that of the pre-twinned samples with further RD re-tension. An increased RD strain further enhances the strain hardening rate at this stage. Obviously, the TB-dislocation interaction effectively improves hardening of deformation by TBs migration. This indicate that the TB-dislocation interaction plays an important role in strain hardening of twinning and in cyclic loading of Mg alloys, where an obvious strain hardening with strain exists. Generally, ̅ } TBs are not considered to be strong barriers for dislocation {1012 slips. Thus, TB-dislocation interaction is found to hardly improve the strain hardening rate during a slip predominant deformation in hcp metals [21,22]. In spite of the weak blockage effect of TBs on dislocation slip, the present study shows that dislocations are strong barriers for TBs motion. In a word, the TB-dislocation interaction poses substantially different effects on the strain hardening of deformation by slips and that of deformation by TBs migration. 4. Conclusion
̅ } TBs and 〈a〉 In summary, an intensive interaction of {1012
Y. Xin et al. / Materials Science & Engineering A 662 (2016) 95–99
dislocations was produced by a TD pre-compression and a subsequent RD re-tension. The influence of this TB-dislocation inter̅ } TBs action on mechanical behavior of a deformation by {1012 migration was studied. Several conclusions are reached as follows: (1) TB-dislocation interaction greatly damages coherence of the ̅ } TBs, leading to a much higher activation stress and {1012 retarding of TBs motion. The TB-dislocation interaction improves the strain hardening at the early stage, while reduces the peak hardening rate at the latter stage. (2) The migration of TBs that intensively interacts with dislocations leaves behind many low angle boundaries at the initial TBs.
Prime novelty statement Dislocation-TB interaction has been reported to significantly affect the strength, ductility and strain hardening of metals. Although the influence of TB-dislocation interaction on mechanical behavior of a slip predominant deformations has been well understood, that of the TB-dislocation interaction on mechanical behavior of deformation by TB migration is hardly reported. In the present study, the effect of interaction between TBs and prismatic 〈a〉 slip on detwinning of twins was studied, with the aim to understand the effect of this TB-dislocation interaction on mechanical behavior of TB migration predominant deformation. Our results shows that TB-dislocation interaction damages the coherence of TBs and induces TBs curving, leading to a much higher activation stress and a delay of TBs motion. The TB-dislocation interaction improves strain hardening at the early stage, while reduces the peak hardening rate at the latter stage. Migration of the TBs that intensively interact with dislocations leaves behind low angle boundaries.
99
Acknowledgment This study is co-supported by National Key Basic Research Program of China (2013CB632204) and Natural Science Foundation of China (51371203, 51571041, 51131009 and 51421001), Fundamental Research Funds for the Central Universities (106112013CDJZR13130032 and 106112015CDJXZ138803).
References [1] K. Lu, L. Lu, S. Suresh, Science 324 (2009) 349–352. [2] M.R. Barnett, Mater. Sci. Eng. A 464 (2007) 1–7. [3] A. Levinson, R.K. Mishra, R.D. Doherty, S.R. Kalidindi, Acta Mater. 61 (2013) 5966–5978. [4] N.V. Dudamell, I. Ulacia, F. Gálvez, S. Yi, J. Bohlen, D. Letzig, I. Hurtado, M. T. Pérez-Prado, Acta Mater. 59 (2011) 6949–6962. [5] Y. Xin, M. Wang, Z. Zeng, M. Nie, Q. Liu, Scr. Mater. 66 (2012) 25–28. [6] Y. Cui, Y. Li, S. Sun, H. Bian, H. Huang, Z. Wang, Y. Koizumi, A. Chiba, Scr. Mater. 101 (2015) 8–11. [7] S.H. Park, S.-G. Hong, C.S. Lee, Scr. Mater. 62 (2010) 666–669. [8] Y. Xin, X. Zhou, Q. Liu, Mater. Sci. Eng. A 567 (2013) 9–13. [9] Y. Xin, M. Wang, Z. Zeng, G. Huang, Q. Liu, Scr. Mater. 64 (2011) 986–989. [10] E. Ma, Y.M. Wang, Q.H. Lu, M.L. Sui, L. Lu, K. Lu, Appl. Phys. Lett. 85 (2004) 4932. [11] C.J. Shute, B.D. Myers, S. Xie, T.W. Barbee, A.M. Hodge, J.R. Weertman, Scr. Mater. 60 (2009) 1073–1077. [12] G.J. Tucker, S.M. Foiles, Int. J. Plast. 65 (2015) 191–205. [13] C. Guillemer, M. Clavel, G. Cailletaud, Int. J. Plast. 27 (2011) 2068–2084. [14] J.W. Christian, S. Mahajan, Prog. Mater. Sci. 39 (1995) 1–157. [15] S.R. Agnew, O. Duygulu, Int. J. Plast. 21 (2005) 1161–1193. [16] D.D. Perovic, C.J. Rossouw, A. Howie, Ultramicroscopy 52 (1993) 353–359. [17] S.R.K. Ayman, A. Salem, Roger D. Doherty, Scr. Mater. 46 (2002). [18] Y. Xin, X. Zhou, L. Lv, Q. Liu, Mater. Sci. Eng. A 606 (2014) 81–91. [19] A. Ostapovets, R. Gröger, Modell. Simul. Mater. Sci. Eng. 22 (2014) 025015. [20] L. Wu, S.R. Agnew, D.W. Brown, G.M. Stoica, B. Clausen, A. Jain, D.E. Fielden, P. K. Liaw, Acta Mater. 56 (2008) 3699–3707. [21] B. Song, R. Xin, G. Chen, X. Zhang, Q. Liu, Scr. Mater. 66 (2012) 1061–1064. [22] M. Knezevic, A. Levinson, R. Harris, R.K. Mishra, R.D. Doherty, S.R. Kalidindi, Acta Mater. 58 (2010) 6230–6242.