Materials Science & Engineering A 701 (2017) 187–195
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Viscoplastic creep and microstructure evolution of Sn-based lead-free solders at low strain
MARK
⁎
Q.K. Zhanga,b, , F.Q. Hua, Z.L. Songa, Z.F. Zhangb a b
Ningbo Institute of Material Technology and Engineering, Chinese Academy of Sciences, Ningbo 315201, China Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China
A R T I C L E I N F O
A B S T R A C T
Keywords: Viscoplastic creep In-situ EBSD Polygonization Grain boundary sliding Strain concentration
In this study, the viscoplastic creep behaviors of the Sn-4Ag/Cu, Sn-3Cu/Cu and Sn-58Bi/Cu solder joints under shear stress were in-situ observed, and the evolutions in microstructure of the solders were characterized by insitu EBSD. The results reveal that the creep strain of the Sn-Cu and Sn-Ag solders increase linearly with increasing time. Deformation of the Sn-Cu solder concentrates in some parallel shear bands, while deformation of the Sn-Ag solder is relatively uniform, and the deformation concentration along the joint interface is not obvious for the two solders. The strain concentration around the fine Ag3Sn particles in the Sn-Ag solder is not obvious, while the strain concentration around the large Cu6Sn5 grain in the Sn-Cu solder is serious and result in fracture of the Cu6Sn5 grain. Dynamic high-temperature recovery occurs in the Sn-Ag and Sn-Cu solders, result in grain boundary migration, polygonization and formation of a few low-angle grain boundaries. The Sn-Bi solder deforms through grain boundary sliding and phase boundary sliding, the strain increase exponentially with increasing time, and its grain structure is stable during the deformation process.
1. Introduction Soldering in microelectronic package not only provides the electronic connection, but also ensures the mechanical reliability of the joints, which makes the mechanical property of the solder joints one of the major concerns for the integrity of the electronic connection [1]. Therefore, understandings on deformation mechanisms of the solders are of great importance for evaluating the reliability of the solder joints. In the electronic components, the primary strain subjected by the solder joints is resulted from the difference in the coefficients of thermal expansion (CTE). Since the strain is induced by rise and decline of temperature, while the temperature usually changes slowly, the solder joints deform at a low strain rate, and the strain is also very low [2]. The deformation of the solder joints usually concentrates in the solder since the substrate materials and the intermetallic compounds (IMC) have much higher yield strength than the solder. Besides, because the melting temperature and yield strength of the solders are very low, the solders creep through a viscoplastic mechanism when it is deformed at a low strain rate even at room temperature, and dynamic microstructure evolution occur during the deformation process [3,4]. Thus far many studies on creep behaviors of the solders have been reported, the creep stress exponents and activation energies of a series of Sn-based solders were obtained, and creep mechanisms of the solders
⁎
are proposed [5–10]. However, most of these studies focus on the creep curves, parameters and equations, while little attention has been paid to the deformation morphologies and the microstructure evolution during the deformation process. Besides, the strain chosen in most of the studies is quite high, while the strain suffered by the solder joints in the electronic components is usually very low [11]. As both the strain and the strain rate are very low, it is necessary to reveal the deformation mechanisms of the solders at low strain and low strain rate. The discussions above indicate that a comprehensive, intuitionistic understanding on deformation mechanisms of the lead-free solders at low strain and low strain rate is required. Therefore, the creep behaviors of a series of Pb-free solders were investigated in this study. The Scanning Electronic Microscope (SEM) equipped with in-situ tensile stage and Electronic Back-scattered Diffraction (EBSD) system was used to conduct the experiments, which allows visualized observation on morphology of the solder and characterization of the microstructure evolution. The employed specimens are small in scale (1 × 1 mm), in order to make them more like the solder joints in the electronic components. By analyzing the deformation morphologies and the EBSD maps, new understandings on deformation mechanisms of the solders at low strain and low strain rate are provided. Furthermore, it is expected that this research may provide a new approach to reveal the dynamic deformation and microstructure evolution behaviors of the solder.
Corresponding author at: Ningbo Institute of Material Technology and Engineering, Chinese Academy of Sciences, Ningbo 315201, China. E-mail address:
[email protected] (Q.K. Zhang).
http://dx.doi.org/10.1016/j.msea.2017.06.083 Received 26 April 2017; Received in revised form 20 June 2017; Accepted 20 June 2017 Available online 21 June 2017 0921-5093/ © 2017 Elsevier B.V. All rights reserved.
Materials Science & Engineering A 701 (2017) 187–195
Q.K. Zhang et al.
Fig. 1. Back scattered electron images of the (a) Sn-3Cu, (b) Sn-4Ag and (c) Sn-58Bi solders.
2. Experimental procedure The Sn-4 wt%Ag/Cu, Sn-3 wt%Cu/Cu and Sn-58 wt%Bi/Cu solder joints were chosen as examples in this study. The solders were prepared by smelting high-purity Sn, Ag, Cu and Bi metals in vacuum, and the substrate is cold-drawn oxygen-free Cu rods with a yield strength of about 300 MPa. Since the Cu substrate has much higher yield strength than the solders, it only exhibits very slight elastic deformation under the shear loading. The Cu rod was firstly spark cut into small blocks with a step at one end, then the surfaces at the ends were ground and electrolytically polished, a flux was dispersed on the polished area, and the solder sheets were sandwiched between the two Cu blocks. The prepared samples were put in an oven with a certain temperature, kept for 3 min after melting of the solder and then cooled in air. The soldering temperature of the Sn-4Ag, Sn-3Cu and Sn-58Bi solder joints are 260 °C, 260 °C and 200 °C, respectively. After that, the samples were sliced into shear specimens by spark cutting, and their side surfaces were ground and mechanically polished for observation. The back scattered electron images of the Sn-3Cu, Sn-4Ag and Sn-58Bi solders are shown in Fig. 1. It can be found that the Sn-3Cu solder is composed by Sn matrix and some large Cu6Sn5 IMC grains, the Sn-4Ag is composed by Sn matrix and thin Ag3Sn particles, and the eutectic Sn-Bi solder has a fine lamellar structure of Sn-rich phase and Bi-rich phase. The shape, dimension of the test specimens and the loading direction are illustrated in Fig. 2(a). The shear creep tests were carried out by a Gatan MTEST2000ES Tensile Stage equipped on the LEO Super35 SEM, and the creep stresses were chosen according to some early publications [4,12]. The specimens were loaded to and held at the chosen stresses, makes the solder joints deform at a very low strain rate. Each group of the solder joints were deformed at a few stresses. During the deformation process, the dynamic strain and time data were recorded, and the deformation morphologies of the specimens were in-situ observed. The full views of the solder joints were observed firstly to show the macroscopic damage, then deformation morphologies of some local regions were tracked. It should be noticed that the image was not rotated during the
Fig. 2. Illustrations on (a) shape and dimension of testing specimens and (b) position of samples in the stage.
experiment, to make sure that the deformation direction of all images are the same. When the strain increased to a certain value, the stress was decreased to zero and the EBSD data of the solder were collected and analyzed by a HKL Channel software version 5 (HKL Technology, Hobro, DK). All the experiments were conducted at 25 °C in vacuum. Illustrations on the in-situ characterizations are exhibited in Fig. 2(b).
3. Experimental results 3.1. Viscoplastic creep strain-duration curves Fig. 3 presents the creep strain-duration curves of the three solder joints, with the creep stresses labelled on the figure. The strain is calculated by firstly subtracting the elastic displacement of Cu from the 188
Materials Science & Engineering A 701 (2017) 187–195
Q.K. Zhang et al.
3.2. Deformation and microstructure evolution of Sn-Cu/Cu joints The macroscopic deformation morphologies of the Sn-3Cu/Cu solder joint held at a stress of 16 MPa are exhibited in Fig. 4, with the strains tagged in each figure, and the shear direction is indicated by the arrows. It can be found that the deformation is slight, and the strain concentration along the joint interface is not obvious, because the strain and strain rate are very low, the strain can slowly transfer from the interface to the inner of the solder. However, there are some parallel bands appear at the surface of the solder, and become more and more obvious with increasing strain. Similar shear bands were observed in a few groups of alloys processed at high temperature, which were consider to be a high temperature plastic flow [13]. The homologous temperature of the solder is high even at room temperature, so it deforms like other alloys at high temperature and forms similar bands. The magnified morphology of the shear bands in the Sn-3Cu solder are shown in Fig. 5, it is clear that the strain concentration around the bands are serious, and they are parallel to each other. At higher strain, another group of shear bands appear, as in Fig. 5(d). Therefore, it is predicted that formation of the bands is induced by flow instability related to the grain orientation of the solder, because deformation of the solder has a close relationship with the grain orientation when the solder is deformed at low strain rate [14]. At higher magnification, it can be found that there is strain concentration at the Sn/large Cu6Sn5 interface, as in Fig. 6. As the large Cu6Sn5 is hard and brittle, it is easy to fracture and result in unrecoverable cracking initiates at the IMC/Sn phase interface. Fig. 7 presents the SEM image and the corresponding EBSD maps of the Sn-3Cu solder held at 18 MPa, the rectangle in Fig. 7(a) shows the range of the EBSD maps. The different colors in the EBSD maps correspond to solder grains with different orientations, the low-angle grain boundaries (2° < θ < 15°) are indicated with white lines and the highangle boundaries (θ > 15°) are indicated with black lines in the maps. Before the test, the solder consists mainly of 2 coarse grains separated by high-angle grain boundaries, as in Fig. 7(b). With increasing strain, the “mauve” grains become much larger (see Fig. 7(c) and (d)),
Fig. 3. Shear creep strain-duration time curves of the solder joints.
total displacement, and then divided the result with the thickness of the solder, therefore the strain is contributed by deformation of the solder. As in the figure, the strain of the Sn-Ag/Cu and Sn-Cu/Cu solder joints increase linearly with increasing time, while that of the Sn-Bi/Cu solder joints shows an accelerate growth, which should be induced by their different deformation mechanisms. Although the creep stresses are not quite high compared with the shear strength, the strains increase very fast, because the solders exhibit viscoplastic deformation rather than common creep due to their low melting points. Among the three groups of solder joints, the Sn-Cu solder shows the lowest creep resistance at the early stage, i.e. the increase rate of strain is the highest. The creep resistance of the Sn-Bi solder is the highest at first, because the yield strength of the Sn-Bi solder is much higher, but its creep resistance decrease sharply with increasing strain/time. Among the whole process, the Sn-Ag solder joint shows the highest creep resistance.
Fig. 4. Macroscopic deformation morphologies of the Sn-3Cu/Cu solder joints at the strain of (a) 0, (b) 0.0430, (c) 0.0924 and (d) 0.1224.
189
Materials Science & Engineering A 701 (2017) 187–195
Q.K. Zhang et al.
Fig. 5. Microscopic deformation morphology of the Sn-3Cu solder at the strain of (a) 0, (b) 0.0430, (c) 0.0924 and (d) 0.1224.
morphologies at the corner of the Sn-4Ag/Cu solder joint are shown in Fig. 9. Due to the stress concentration at the corner, the deformation at this region forms a few groups of wave-like deformation bands, but not very severe, as in Fig. 9(d). Besides, there is no strain concentration around the Ag3Sn particles. Since the strain rate is low, the creep fracture strain should be very high, so there is no sigh of fracture for the solder joint in this study. Fig. 10 shows the deformation morphology and EBSD maps of the Sn-4Ag solder held at the stress of 22 MPa. As in Fig. 10(a), there are wave-like deformation lines, but there is no blank region in corresponding EBSD map in Fig. 10(d). According to Fig. 7, the regions with serious deformation in the solder are reflect as “blank” regions, because resolution rate of these areas is very low. Therefore, the deformation around the wave-like lines is not very serious, i.e. there is no severe strain concentration around the wave-like deformation bands, so deformation of the Sn-Ag solder is relatively uniform. Comparing Fig. 10(b)-(d), there is only slight change in orientation (i.e. the color) of the solder in general, and a few low-angle grain boundaries appear inside the solder. Therefore, it can be concluded that for the Sn-Cu and Sn-Ag solders, the dislocation movement and dynamic recovery keep occur during the creep process, the strain energy cannot accumulate to a level enough to result in recrystallization or serious grain subdivision.
Fig. 6. Strain concentration and cracking morphology around the Cu6Sn5 IMC in the Sn3Cu solder.
implying that grain boundary migration occurs. However, the grain boundary migration only when the grains have less than six convex boundaries, and the grain structure become relatively stable after the migration completed [4]. The orientation of the solder grain changes little, which fit well with the deformation morphology, i.e. the deformation is slight. There are a little more low-angle grain boundaries in Fig. 7(c) and (d), indicating that there is polygonization in the solder to form a few new low-angle boundaries, but the grain subdivision is far less serious than the solders deformed for a high strain [4,15]. The shear bands in the solder is blank in the EBSD maps, because the local plastic deformation around the bands is too high to be analyzed.
3.4. Deformation and microstructure evolution of Sn-Bi/Cu joints The deformation morphologies of the Sn-Bi solder are presented by Fig. 11, in which the solder joint was deformed at 20 MPa. Before the deformation, the surface of the Sn-Bi solder is flat (see Fig. 1). The strain is 0.0302 in Fig. 11(a), but some streaks have appeared on the surface of the solder. At higher strain, the surface of the Sn-Bi solder become “smooth” (see Fig. 11(b)-(d)). In the microscopic images, it is clearly that the streaks are squiggly bands, and only the deformation localization around the bands is obvious, as in Fig. 12(a) and (b). With increasing strain, the squiggly bands gradually evolve into microcracks. It has been reported that the creep deformation mechanisms of the SnBi solder are related to the stress, the strain rate and the grain size [16,17]. Under nanoindentation loading, the grain/phase boundary related mechanism dominants the deformation behavior Sn-Bi solder
3.3. Deformation and microstructure evolution of Sn-Ag/Cu joints The macroscopic deformation morphologies of the Sn-4Ag/Cu solder joints held at 20 MPa are shown in Fig. 8. As in the figure, the deformation of the solder is not obvious in macro, and changes little with increasing strain. Besides, there is no serous strain concentration along the joint interface, and also no shear bands appear. Only when the strain is relatively high, the deformation at the corner of the solder joint became visible (see Fig. 8(d)). The local deformation 190
Materials Science & Engineering A 701 (2017) 187–195
Q.K. Zhang et al.
Fig. 7. (a) Deformation morphology and (b)-(d) the EBSD maps of the rectangle area of the Sn-3Cu solder at different strains.
Fig. 8. Macroscopic deformation morphologies of the Sn-4Ag/Cu solder joints at the strain of (a) 0, (b) 0.0421, (c) 0.0712 and (d) 0.1877.
191
Materials Science & Engineering A 701 (2017) 187–195
Q.K. Zhang et al.
Fig. 9. Local deformation morphology of the Sn-4Ag solder at the strain of (a) 0, (b) 0.0421, (c) 0.0712 and (d) 0.1877.
Fig. 10. (a) Deformation morphology and (b)-(d) EBSD maps of the rectangle area of the Sn-4Ag solder at different strains.
192
Materials Science & Engineering A 701 (2017) 187–195
Q.K. Zhang et al.
Fig. 11. Macroscopic deformation morphologies of the Sn-58Bi/Cu solder joints at the strain of (a) 0.0302, (b) 0.0906, (c) 0.1533 and (d) 0.3082.
bands in Fig. 11(b) are actually grain boundaries. At higher strain, the relative movement between the Sn-rich phase and Bi-rich phase become obvious (see Fig. 11(d)). The deformation morphology and corresponding EBSD maps of the Sn-Bi solder held at 18 MPa is shown in Fig. 13. In the EBSD maps, the purple colored phase is the Sn-rich phase and the other phase is Bi-rich
creeps when the strain rate and stress are low [16,17], and phase boundary sliding is the major mechanism when the stress is low and the solder has a fine lamellar structure [18]. Since the lamellar structure of the Sn-Bi solder is fine in this study, and the stress and strain rate are very low, it is predicated that grain/phase boundary sliding are also the major creep mechanism under shear loading, and the deformation
Fig. 12. Local deformation morphology of the Sn-58Bi solder at the strain of (a) 0.0302, (b) 0.0906, (c) 0.1533 and (d) 0.3082.
193
Materials Science & Engineering A 701 (2017) 187–195
Q.K. Zhang et al.
Fig. 13. (a) Deformation morphology and (b)-(d) EBSD maps of the rectangle area of the Sn-58Bi solder at different strains.
Because of the high homologous temperature, plastic deformation accompanied dynamic recovery occur during the deformation process. When the deformation and recovery reach a equilibration state, the solders deform at a constant rate, and the strain energy keeps at a certain level. Since the solders are very soft and the strain rate is very low, they exhibit a viscoplastic deformation behavior, i.e. deforms like the fluid in some extent. The strain concentration at the joint interface is not serious at low strain rate, and there is no cracking inside the solder or along the grain boundaries. In contrast, cracking is easier to occur at the grain boundaries or phase boundaries of the solder when it is deformed a higher strain rate, as in Fig. 14, in which the Sn-4Ag solder was deformed at a thermal strain rate of 1 × 10−4 s−1 [19]. At lower strain rate, it should be easier for the solder grains to accommodate the deformation, and cracking is less likely to occur at the grain boundaries. The IMCs in the solder can strengthen the Sn matrix, but the effects
phase. It can be found that the Bi-rich phase is stable, while the Sn-rich phase changes with increasing strain. Since the Bi-rich phase in the eutectic structure is very hard, it can hardly yield during the creep process. However, the soft Sn phase deforms in a dislocation gliding mechanism, so there is microstructure evolution in the Sn phase. Besides, there is also deformation mismatch and relative movement between the Sn and Bi phases, as in Fig. 13(a), but the mismatch will not result in long cracks.
4. Discussions 4.1. Viscoplastic creep mechanisms of solders As exhibited above, when the Sn-Ag and Sn-Cu solders are held at a constant stress, the strain increase linearly with increasing time, indicating that the deformation resistance is stable during that process.
Fig. 14. Cracking at the grain boundaries and phase boundaries of the Sn-3.5Ag-0.7Cu solder after thermal fatigue.
194
Materials Science & Engineering A 701 (2017) 187–195
Q.K. Zhang et al.
avoided.
is affected by the size and distribution of the IMCs [20,21]. Since the Sn-rich phase in the solder has low yield strength, while the Cu3Sn usually exist as large bulks in the Sn-Cu solder (see Figs. 6 and 7), it cannot significantly strengthen the Sn matrix. Therefore, the Sn-Cu solder is easy to deform even at a low stress, and the shear bands can easily occur, result in strain localization. Besides, continuous deformation cannot sustain when large Cu6Sn5 IMC exist, result in serious stress concentration around the IMC/Sn phase interface, so the large Cu6Sn5 is easy to fracture and forms unrecoverable cracking when the strain is relatively high. For the Sn-4Ag solder, since the Ag3Sn IMC in the Sn-4Ag solder exist as uniform distributed thin particles that can strengthen the Sn matrix, its yield strength and creep resistance is relatively high. In another hand, the strengthen effect also restrain the formation of serious shear bands through limiting linear propagation of dislocation movement, therefore only some wave-like deformation lines appear. The strain energy in the Sn-Ag solder can also release through dynamic recovery. Besides, although the deformation mismatch between the Ag3Sn and Sn-rich is non-ignorable, the thin size of the Ag3Sn limits the strain concentration extent, so the Ag3Sn will not fracture. During the viscoplastic creep process, cracking can hardly appear in the Sn-Ag solder. The eutectic Sn-Bi solder has a fine lamellar structure and deforms at low strain rate, and creeps through grain/phase boundary sliding, which are similar to their deformation mechanism under nanoindentation stress at low strain rate [17]. However, the grain boundary sliding and phase boundary sliding do not start at the same time under low strain rate shear loading. As in Fig. 12, the grain boundary sliding occurs firstly when the strain is low, and the phase boundary sliding gradually become obvious with increasing strain. As the boundaries become weaker during the relative sliding process, deformation resistance of the solder decreases, makes the deformation resistance decrease with increasing strain. In addition, grain boundaries and phase boundaries of the Sn-Bi solders do not migrate during the deformation process.
5. Conclusions The viscoplastic creep behaviors and microstructure evolutions of the Sn-4Ag, Sn-3Cu and Sn-58Bi solders were investigated by in-situ observation and in-situ EBSD. Based on the experimental results and discussions, the main conclusions can be summarized as follows: 1. Under a constant shear stress, the strain of the Sn-Cu and Sn-Ag solders increase linearly with increasing time, while that of the Sn-Bi solder increase exponentially. Deformation of the Sn-Cu solder concentrates in some parallel shear bands, while deformation of the Sn-Ag solder is uniform, and the strain concentration along the joint interface is not serious. The Sn-Bi solder deforms through grain boundary sliding and phase boundary sliding. 2. High-temperature recovery occurs in the Sn-Ag and Sn-Cu solder, result in polygonization and formation of a few low-angle grain boundaries, while grain recrystallization or serious grain subdivision do not occur. The grain boundaries and phase boundaries of the Sn-Bi solders do not migrate during the deformation process. 3. The effects of IMCs on deformation behaviors of the Sn-Ag and SnCu solders are different, depend mainly on the size. The strain concentration around the Ag3Sn particles in the Sn-Ag solder is not obvious, while the strain concentration around the large Cu6Sn5 grain in the Sn-Cu solder is serious and result in unrecoverable fracture. Acknowledgements The authors would like to acknowledge W. Gao and L.X. Zhang for sample preparation, in-situ tests and observations. This work was financially supported by the National Basic Research Program of China under grant No. 2010CB631006, the Public Projects of Zhejiang Province under grant No. 2015C31031. References
4.2. Influencing factors on viscoplastic creep deformation Based on the understandings on the viscoplastic creep mechanisms, the major intrinsic factors on the viscoplastic creep should be the grain size of the solder and the IMCs in it. For the Sn-Ag, Sn-Cu and Sn-Ag-Cu solder strengthen by the IMCs, the size, shape and amount of the IMC particles are major factors dominated the deformation resistance of the solder. The thin and uniform distributed IMC particles can significantly improve the viscoplastic deformation resistance of the solder, and also restrain the microstructure evolution induced by dislocation movement. If the IMC coarsen into large bulks, its strengthen effects will decrease sharply, and itself is much easier to fracture. It is widely accepted that the solders composed by fine grains have higher yield strength, and they should have higher creep resistance. For all the Sn-rich solders strengthen by IMC, it is suggested to refine the solder grain by rapid cooling, which can also restrain the coarsen of IMC particles. Besides, addition of Zn, Ag, Ni and In elements have been proved to be able to optimize the microstructure and in turn improve the mechanical properties of the solders [22–26]. The grain-boundary sliding is easier to occur for metallic materials composed of fine grains [27,28]. As the Sn-Bi solder deforms through grain-boundary sliding, the decrease in grain size can weaken its creep resistance, similar phenomenon should also occurs in the other doublephase solders. It is reported that a low cooling rate or a short-term thermal aging can induce an obvious increase in grain size of the Sn-Bi solder [29]. Through heat treatment that increase the grain size, the resistance of the Sn-Bi solder to grain-boundary sliding can be enhanced. Besides, addition of some alloy elements or thin particles have also been used to improve the mechanical properties of the Sn-Bi solder [30–32], but formation of bulk IMCs in the Sn-Bi solder should also
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32]
195
Y. Li, C.P. Wong, Mater. Sci. Eng. R 51 (2006) 1–35. M. Abtewa, G. Selvaduray, Mater. Sci. Eng. R 27 (2000) 95–141. M. Kuna, S. Wippler, Eng. Fract. Mech. 77 (2010) 3635–3647. Q.K. Zhang, Z.F. Zhang, Acta Mater. 59 (2011) 6017–6028. Y. Yao, X. He, L.M. Keer, M.E. Fine, Acta Mater. 83 (2015) 160–168. E.H. Wong, W.D. van Driel, A. Dasgupta, M. Pecht, Microelectron. Reliab. 59 (2016) 1–12. T. Yamamoto, T. Itoh, M. Sakane, Y. Tsukada, Int. J. Fatigue 43 (2005) 235–241. Y.X. Zhu, X.Y. Li, C. Wang, R.T. Gao, Microelectron. Reliab 55 (2015) 1097–1100. P. Rajaguru, H. Lu, C. Bailey, Microelectron. Reliab. 55 (2015) 2371–2381. D. Chicot, K. Tilkin, K. Jankowski, A. Wymysłowski, Microelectron. Reliab. 53 (2013) 761–766. Y.C. Chan, D. Yang, Prog. Mater. Sci. 55 (2010) 428–475. Q.K. Zhang, Z.F. Zhang, Mater. Sci. Eng. A 528 (2011) 2686–2693. D. Sagapuram, K. Viswanathan, A. Mahato, N.K. Sundaram, R. M'Saoubi, K.P. Trumble, S. Chandrasekar, Proc. Math. Phys. Eng. Sci. 472 (2016) 20160167. H.T. Chen, B.B. Yan, M. Yang, X. Ma, M.Y. Li, Mater. Charact. 85 (2013) 64–72. A.U. Telang, T.R. Bieler, M.A. Crimp, Mater. Sci. Eng. A 421 (2006) 22–34. X. Chen, J. Zhou, F. Xue, Y. Yao, Mater. Sci. Eng. A 662 (2016) 251–257. L. Shen, P. Lua, S.J. Wang, Z. Chen, J. Alloy. Compd. 574 (2013) 98–103. L. Shen, P. Septiwerdani, Z. Chen, Mater. Sci. Eng. A 558 (2012) 253–258. Q.K. Zhang, Z.F. Zhang, unpublished work. A.K. Gain, L.C. Zhang, M.Z. Quadir, Mater. Des. 110 (2016) 275–283. A.F. Abd El-Rehim, H.Y. Zahran, J. Alloy. Compd. 695 (2017) 3666–3673. M. Yang, Y.H. Ko, J.W. Bang, T.S. Kim, C.W. Lee, M.Y. Li, Mater. Charact. 124 (2017) 250–259. A.A. El-Daly, A.E. Hammad, G.S. Al-Ganainy, M. Ragab, Mater. Sci. Eng. A 608 (2014) 130–138. A.A. El-Daly, A.Z. Mohamad, A. Fawzy, A.M. El-Taher, Mater. Sci. Eng. A 528 (2011) 1055–1062. A.A. El-Daly, A.E. Hammad, Mater. Des. 40 (2012) 292–298. A.E. Hammad, Mater. Des. 52 (2013) 663–670. F.R.N. Nabarro, Acta Mater. 54 (2006) 263–295. K.A. Padmanabhan, J. Mater. Sci. 44 (2009) 2226–2238. Q.K. Zhang, H.F. Zou, Z.F. Zhang, J. Electron. Mater. 40 (2011) 2320–2328. R.M. Shalaby, Mater. Sci. Eng. A 560 (2013) 86–95. D.L. Ma, P. Wu, J. Alloy. Compd. 671 (2016) 127–136. Y. Ma, X.Z. Li, W. Zhou, L.Z. Yang, P. Wu, Mater. Des. 113 (2017) 264–272.