Strength and ductility of gradient structured copper obtained by

Materials and Design 105 (2016) 89–95

Contents lists available at ScienceDirect

Materials and Design journal homepage: www.elsevier.com/locate/matdes

Strength and ductility of gradient structured copper obtained by surface mechanical attrition treatment Zhe Yin a, Xincheng Yang a, Xiaolong Ma b, Jordan Moering b, Jian Yang a, Yulan Gong a, Yuntian Zhu b,c, Xinkun Zhu a,⁎ a b c

Faculty of Materials Science and Engineering, Kunming University of Science and Technology, Kunming, Yunnan 650093, China Department of Materials Science & Engineering, North Carolina State University, Raleigh, NC 27695, USA School of Materials Science and Engineering, Nanjing University of Science and Technology, Nanjing 210094, China

a r t i c l e

i n f o

Article history: Received 6 January 2016 Received in revised form 8 April 2016 Accepted 6 May 2016 Available online 07 May 2016 Keywords: Gradient structure Mechanical properties Mobile dislocation density Synergetic strengthening Pure copper Surface mechanical attrition treatment

a b s t r a c t By using surface mechanical attrition treatment (SMAT) at room temperature, a gradient structure (GS) is generated in the surface layer of bulk pure copper samples, which exhibits good uniform elongation and high yield strength simultaneously. Changing SMAT processing time leads to different gradient structures with various component fractions and therefore tune their mechanical properties. The yield strength of the SMAT samples is much higher than the sum of standalone GS layer and coarse-grained (CG) matrix, indicating a synergetic strengthening. Repeated stress relaxation tests were performed to characterize the evolution of mobile dislocations. It was found that the relative mobile dislocation density of SMAT processed sample first drops and then increases with increasing tensile strain. The evolution of mobile dislocations correlates well with strain-hardening evolution. These observations provide insight for the superior combination of high strength and good ductility in SMAT samples. © 2016 Elsevier Ltd. All rights reserved.

1. Introduction Homogeneous nanostructured metals usually exhibit disappointingly low ductility under quasi-static tensile tests at room temperature, which becomes a major bottleneck for their applications [1]. Ductility is measured under tensile load and can be enhanced by improving the work hardening ability to prevent early emerging plastic instability, i.e. necking [2]. A variety of strategies have been reported to improve the strain hardening and ductility in the literature [3]. By producing a high density of nanotwins (NT) in pure copper samples, Lu et al. [4,5] synthesized a NT-Cu sample with strength about 10 times higher than that of coarse-grained copper and a high elongation-to-failure value of 13.5%. The interaction between twin boundaries and dislocations is the main reason for the high strain hardening [4]. Wang et al. [6] described a thermomechanical process of Cu, which results in a bimodal grain size distribution, leading to a high tensile ductility-30% uniform elongation. Zhao et al. simultaneously improved the strength and ductility by introducing high density of second-phase particles [7], preexisting deformation twins, deformation twinning, and increasing the fraction of high-angle grain boundaries (HAGBs) [8]. Other methods such as transformation induced plasticity (TRIP) and twining induced plasticity (TWIP) have also been developed for improving the ductility ⁎ Corresponding author. E-mail address: [email protected] (X. Zhu).

http://dx.doi.org/10.1016/j.matdes.2016.05.015 0264-1275/© 2016 Elsevier Ltd. All rights reserved.

of nanocrystalline (NC) metals [9,10]. A strain non-localization mechanism that avoids strain/stress localization and delays the catastrophic failure by introducing multiple strengthening and toughening features was presented by Kou et al. [11], who have successfully produced three-dimensional hierarchical nanotwins (3D HNTs) in bulk nanostructured TWIP steels to achieve extremely high strength and high ductility. Recently, surface nanocrystallization (SNC) technology has been generally considered as a new approach to enhance the global behavior of the materials by optimizing the surface structure [12]. A nanocrystalline surface layer with high thermal stability was generated in pure Al by means of surface mechanical attrition treatment (SMAT), with which the yield strength of samples enhanced to about two times higher than that of the coarse-grained counterpart [13]. Researches have shown that generating a gradient structure (GS) with a continuous increase in grain size as a function of depth produces a superior combination of strength and ductility [14–16]. However, this strategy hasn't been systematically investigated in terms of the effect of processing parameters on tuning microstructures and properties. The deformation mechanism of such gradient microstructure is not fully understood as well, despite limited reports. To explore the deformation mechanism during tension, stress relaxation experiment is widely utilized to reflect the dislocation characteristics of metals. Analysis of stress relaxation will shed light on the evolution of mobile dislocation density and the corresponding

90

Z. Yin et al. / Materials and Design 105 (2016) 89–95

deformation mechanisms of metals. It provides a method to explore the relationship between mechanical properties and dislocation behaviors. Wang et al. [17] found that the mobile dislocation exhaustion observed in NC Ni is associated with the work hardening rate by means of stress relaxation tests. By investigating the work hardening and the apparent activation volume (Va) of equal channel angular extrusion (ECAE) processed copper, Torre et al. [18] concluded a rate-controlling deformation mechanism demonstrated by the interaction between dislocations and cell walls or subgrain/grain boundaries. Lu et al. [19] studied the deformation kinetics of nano-twinned (NT)-Cu and discussed the relationship between exhaustion rate of mobile dislocations and twin thickness, which paved the way to explain the high strength and good ductility of NT-Cu. In this work, two gradient structured surface layers sandwiching a coarse-grained (CG) central layer was processed in the commercial pure copper samples by surface mechanical attrition treatment (SMAT) at room temperature. For simplicity, such samples with gradient structure are called SMAT samples. Microstructure analyses and stress relaxation tests were applied to evaluate the effects of the gradient structure (GS) on its mechanical performance. The results provide some insights on the remarkable combination of high strength and good ductility obtained in the SMAT samples.

2. Experimental details The copper samples were prepared by rolling commercially pure copper (99.995 wt%) to a thickness of 3 mm and were cut into square samples (100 × 100 mm). The sectioned plates were then annealed in vacuum at 873 K for 2 h to obtain homogeneous coarse grains. The samples were polished to a mirror finish before SMAT treatment. SMAT have been described in detail previously [20]. In brief, a large number of balls impact the sample surface repeatedly from various directions, leading to strain induced grain refinement at the surface. 180 stainless steel balls with diameter of 8 mm were placed at the bottom of a cylinder-shaped chamber vibrating with a frequency of 50 Hz. Both sides of the samples were processed under vacuum at room temperature for 5 min, 15 min and 30 min to produce varying GS structures. After the SMAT process, dog-bone-shaped tensile specimens, which have a gauge length of 15 mm, a width of 5 mm and a thickness of 3 ± 0.1 mm, were cut from SMAT processed samples by wire-electrode cutting. Tensile tests were performed on a SHIMADZU Universal Tester at room temperature and at a strain rate of 5.0 × 10−4 s−1. Stress relaxation experiments were performed at six initial applied strains at room temperature, i.e., 1.5%, 3%, 5%, 8%, 12%, and 16%. The specimens were first strained up to the required strain at a strain rate of 5.0 × 10−4 s−1, then the cross-head was stopped to maintain a constant strain while the stress was recorded as a function of time for 50 s. After the first relaxation, the specimens were reloaded to the same onset stress level of the previous cycle for the next relaxation, and five relaxation cycles were conducted at each strain [19]. The stress relaxation cycles were performed with the same testing parameters. Three specimens were tested at each condition to ensure good repeatability of the results. The cross-sectional hardness profiles of the SMAT samples were measured using a Vikers microhardness tester with a load of 10 g and a duration of 15 s and were determined by averaging the value of 8 indentations at each depth of the GS layer. Microstructure observations of cross sectional samples of the GS layer were prepared with conventional polishing techniques. The samples were etched in a solution containing FeCl3 (5 g), HCl (50 mL) and H2O (100 mL) for 30 s. Then the cross-sectional microstructure was analyzed by scanning electron microscopy (SEM, Hitachi S 3500 N). A global view of gradient structure from top fine grains to the coarse grain in the matrix is performed by ion channeling contrast imaging under a FEI Quanta 3D FEG dual-beam instrument. The cross-sectional

specimen for this imaging is prepared by focused ion beam (FIB) with Ga+ ions at 30 kV, followed by lower-current thinning. The TEM foil was prepared by unidirectional polish from the matrix side to ~30 μm so that the SMAT processed surface was retained for observation. Finally, the TEM foil was further thinned and perforated by ion mill. The milling procedure was protected by liquid nitrogen so that the grain growth is largely prohibited. TEM images were acquired in a JEM-2010F microscope operated at 200 kV. 3. Results Fig. 1 shows the hardness of 3 mm thick samples treated by SMAT for different times (5, 15, 30 min) and the annealed CG samples, respectively. Hardness increases from about 0.62 GPa in the CG matrix to about 1.25 GPa (B), 1.34 GPa (C), and 1.35 GPa (D) in the top treated surface layer, respectively. For convenience, we define the top surface layer as the nano-structured (NS) layer. The whole gradient structure (GS) layer includes the NS layer and the deformed CG layer with the dimension of grain sizes changing from sub-micro to micrometers [21]. Hardness values reach the as-received state at about 120 μm (B) and 140 μm (C and D) depth, respectively, which defines the border between GS layer and CG matrix. The micro-hardness of the top treated surface layer is approximately two times of that of the CG matrix. Formation of this gradient structure can be understood in terms of the strain and strain rate distribution during the SMAT process. The topmost surface layer undergoes plastic deformation with the largest strains and strain rates [22]. Fig. 2a shows the engineering stress-strain curves of the 3 mm thick samples treated by SMAT for different time (5, 15, 30 min) and the annealed CG samples, respectively. From the average of three tensile tests, we observe yield strength (0.2% offset) of 165 ± 10 MPa (B), 189 ± 12 MPa (C), and 203 ± 9 MPa (D), respectively, which is nearly four times that of the CG samples (53 ± 5 MPa). The SMAT samples exhibit a uniform elongation of 29 ± 3% (B), 27 ± 3% (C) and 23 ± 3% (D), which are a little lower than that of the CG tensile sample (35 ± 3%). Some studies have shown that standalone GS layer always exhibits low ductility [14,15], but our SMAT samples with a GS layer show very good ductility, suggesting that the early-emerging strain localization and failure of GS layer has been successfully suppressed by the CG matrix during the tension. Fig. 2b shows the work hardening rate (Θ = dσ/dε) as a function of true strain. For each SMAT sample, the work hardening rate steeply decreases in the elastic–plastic transition stage (true strain is less than 2%). However, in the plastic deformation stage (true strain is larger than 2%),

Fig. 1. Variation in microhardness as a function of depth from the surface in the copper samples treated by SMAT, for 5 min (B), 15 min (C), and 30 min (D), respectively. The horizontal dashed line represents the microhardness value of the annealed sample (A).

Z. Yin et al. / Materials and Design 105 (2016) 89–95

91

morphology extended to depths exceeding 50 μm (Fig. 3a). Two important deformation modes of metallic materials are dislocation activity and deformation twinning [22]. Dislocation tangles and specific subgrain structures (dislocation cells, walls, geometrically necessary boundaries and incidental dislocation boundaries) are generated, which lead to grain refinement when a metal is subjected to severe plastic deformation [25].A transition from coarse grains to ultrafine grains was seen in the gradient structured layer at a depth of 10 μm (Fig. 3d). Cross-sectional FIB image of the SMAT-5 sample (Fig. 3e) shows the variation in grain size as a function of depth from the treated surface. A gradient structure with a transition of grain sizes from nano-scale to submicron and micron scales in the treated surface layer was generated. Within the top 5 μm thick layer, grain sizes are less than 100 nm. Comparing the SMAT-5 sample and the SMAT-15 sample, we found that the GS layer in SMAT-15 sample is thicker than that in SMAT-5 sample, suggesting that the SMAT processing time has great influence on the thickness of GS layer. The microstructure of the very top SMAT surface was obtained by TEM. Fig. 4 is a TEM micrograph of SMAT-30 sample surface and the corresponding selected area electron diffraction (SAED). Grain size of nano-grains at SMAT top surface is about 53 nm based on the TEM observation in Fig. 4a. SAED pattern indicates that the nano-grains had random crystallographic orientations. Other studies on SMAT processed low carbon steel showed a strong texture at the surface [26], which is not the case in our work. This is probably related to the dynamic recrystallization during the highly energetic impact procedure. The imparted kinetic energy and resultant temperature increase depends on the frequency of oscillations and the translational amplitude of the specific balls used. As such, an analogous temperature increase in the metal substrate typically occurs for these high-energy processes [23]. Deformation under elevated temperature makes it easy for dynamic recrystallization and therefore randomly reorients the grains. 4. Discussion 4.1. Synergetic strengthening of gradient structure Fig. 2. (a) Tensile engineering stress–strain curves for coarse-grained copper (A), SMAT-5 sample (B), SMAT-15 sample (C), and SMAT-30 sample (D) at a quasi-static strain rate of 6 × 10−4 s−1, respectively. (b) Work hardening rate (Θ) as a function of true strain for coarse-grained copper (A), SMAT-5 sample (B), SMAT-15 sample (C), and SMAT-30 sample (D), respectively. Inset is an image of true stress-strain curves.

the SMAT samples show a slower work hardening rate reduction than that of the CG samples. The work hardening rate of SMAT samples decreases with the increasing SMAT processing time. It is remarkable in the SMAT-30 sample that the work hardening rate has a slight upturn characteristic, which is described in recent report of GS IF steel structures [15]. Note that this feature is not found in the other two SMAT samples with less processing time, which is consistent with the previous observation of minimum SMAT processing time for up-turn hardening phenomenon. Generally, the gradient structures show marked improvements in strain hardening when compared to nanocrystalline counterpart standalone [23]. Although the detailed mechanism of this phenomenon is not well understood, the popular belief is to attribute the superiority to the interaction between nanocrystalline and coarse-grained layers during the incompatible plastic deformation. As the CG matrix begins to deform plastically, the GS layers still deform elastically, leading to complex stress states, and a quick increase of geometrically necessary dislocations to accommodate the large strain gradient near the migrating elastic/plastics interfaces and later the migrating table/unstable interfaces [15,24]. Fig. 3a–d are the SEM cross sectional images of the SMAT-15 sample, showing the variation in grain size as a function of depth from the treated surface. A high density of etch pits were observed at the depths 20 μm from the treated surface (Fig. 3b and c) and deformed grain

Gradient structures in engineering materials can produce an intrinsic synergetic strengthening, which is considered much higher than the sum of individual gradient layers, i.e. beyond prediction of the ROM. We calculated yield strength by using the ROM (σROM): σROM = VGSσGS + VCGσCG, where, VGS is the volume fraction of the GS layers, VCG is the volume fraction of the CG layers, σGS is the yield strength of GS layer, and σCG is the yield strength of CG sample at 0.2% strain. Volume fraction is estimated by the thickness percentage. The thickness of GS layer in SMAT-30 sample is about 140 μm. An empirical relationship Hv = 3σy can be used to calculate the average yield strength of standalone GS layer at such thin thickness [27]. Based on the hardness profile in Fig. 1, the estimated average yield strength of GS layer is about 317 MPa, therefore yield strength of SMAT samples calculated by ROM is 77.6 MPa. The results show that the measured yield strength of SMAT samples is much higher than the sum of standalone GS layer and CG matrix, indicating that the ROM is inadequate in estimating the yield strength of the composite structure when the interaction between layers is strong. And this extra increment of yield strength may be attributed to the synergetic strengthening [16]. 4.2. Grain softening and reinforcing effect of GS layer The relationship between yield strength and grain size follows the Hall–Petch relation: σ y ¼ σ 0 þ kD−1=2

ð1Þ

where σy is the yield stress, σ0 is the lattice frictional stress, k is the Hall– Petch constant, and D is the grain size. For pure Cu, the typical

92

Z. Yin et al. / Materials and Design 105 (2016) 89–95

Fig. 3. (a–d) Cross-sectional SEM images of the SMAT-15 sample. (e) Cross-sectional FIB images of the SMAT-5 sample.

parameters σ0 = 25.5 MPa and k = 0.11 MPa m1/2 [28]. Based on these relations, the grain size of the top surface layer can be calculated. The grain size in the top surface of SMAT-30 sample calculated by formula above is 67 nm, slightly larger than average grain size obtained from the TEM images. Previous studies indicate that grains smaller than 165 nm will grow under tension due to the mechanically driven grain boundary migration for Cu [29]. With strains increasing, the energy

release due to numerous defect annihilations will causes grain coarsening and softening. On the other hand, grains above 165 nm continue to refine, inducing hardening. The tension-induced softening and hardening has made obvious contribution to the ductility. Another major factor that is responsible for the good ductility could be attributed to the suppressed strain localization in GS sample [14]. Due to the varying grain size of GS layer, plastic deformation in GS

Fig. 4. TEM observations of the top surface layer. (a) A typical TEM bright-field image of the SMAT-30 sample. (b) The corresponding selected-area electron diffraction (SAED) pattern.

Z. Yin et al. / Materials and Design 105 (2016) 89–95

sample is more concordant and better accommodated, which could lead to the release of intergranular stress between adjacent grains and avoid strain localization in the sample and formation of voids or cracks at the grain boundaries effectively [30]. Furthermore, the strain energy stored during deformation creates a stress field gradually accumulated along the GS/CG interface, blocking the extension of cracks initiated in the sub-surface layer and preventing plastic instability in a high stress range. In addition, a high compressive residual stress introduced by SMAT processing also can further restrain the crack propagation [11]. The suppressed strain localization and crack propagation, and the delayed plastic instability finally lead to a considerable tensile ductility in GS samples. This crack-stopping reinforce layer (GS layer) has been used to explain the high strength and pronounced work hardening in ultrafine-grained Ti with gradient structure [31]. It is in accordance with the results of our previous work in pure copper treated by SMAT at cryogenic temperature [32]. 4.3. Exhaustion of mobile dislocation density To further investigate the mechanism of good ductility observed in the SMAT samples, stress relaxation experiments were conducted on both SMAT-30 samples and CG samples. Fig. 5a shows an engineering stress-strain curve of SMAT-30 sample from the stress relaxation test at six starting strains. The inset shows the stress-relaxation time curve selected to cover all the six starting strains range. Five relaxation tests were performed at each strain. According to D. Caillard and J.L. Martin [33], the relative mobile dislocation density (Re = ρm/ρm0) is given by: ρm ¼ ρm0



Cr t þ Cr

β=ð1þβÞ

ð2Þ

93

where ρm0 is the initial mobile dislocation density of the transient, Cr is a time constant, and β is a dimensionless immobilization parameter. Cr and the apparent activation volume (Va) can be obtained by fitting the stress relaxation time curve according to: Δτ ðt Þ ¼ −

  kB T t ln 1 þ Va Cr

ð3Þ

where Δτ(t) is the shear stress drop at time t, kB is the Boltzmann constant and T is the temperature. β can be calculated by equations below: β¼

Ω −1 1 þ K=M

ð4Þ

where Ω = Va/V*, K is the work-hardening coefficient calculated from the stress–strain curve, and M is the elastic modulus of the specimenmachine assembly. The physical activation volume V* is determined by:   i f V  ¼ kB T ln γ_ 2 =γ_ 1 Δτ

ð5Þ

where γ_ 2 is the shear strain rates at the onset of the second relaxation i

cycle, and γ_ 1 is the shear strain rates at the end of the first relaxation cycle. The applied shear stress consists of two components: the athermal stress τu and the effective stress τ⁎. Δτ⁎ is the change of thermal component of the shear stress during relaxation. It can be described as: f

Δτ  ¼ ð1 þ K=MÞΔτ

ð6Þ

Fig. 5. Evolution of mobile dislocation density. (a) Engineering stress-strain curve of SMAT-30 sample from the stress relaxation test at six starting strains. Five relaxation tests were performed at each strain. The inset shows the stress-relaxation time curve. (b) Evolution of the mobile dislocation density in CG Cu sample at varying strains. (c) Evolution of the mobile dislocation density in SMAT sample at varying strains. (d) Re = ρm/ρm0 after the first stress relaxation (50 s) as a function of strain.

94

Z. Yin et al. / Materials and Design 105 (2016) 89–95

Based on these formulas, the relative mobile dislocation density after relaxation can be calculated. Fig. 5b and c show the evolution of the mobile dislocation density with relaxation time in CG Cu sample and SMAT-30 sample at varying strains, respectively. Fig. 5d reveals the Re varies with tensile strain as a function. As illustrated in Fig. 5, with the increasing strain, the Re of SMAT sample first drops and then increases, which is very different from that of CG sample. The calculated mobile dislocation density of the SMAT sample is reduced to 92% of its onset value at the initial stage of plastic deformation after 50 s relaxation. It continues to decrease at the strain of 3%, indicating a quick exhaustion of mobile dislocations at the low strains. In this stage, strain gradient induced by gradient grain-size under tension leads to the redistribution of interior stress and generation of more geometrically necessary dislocations (GNDs) [15]. The stress gradually accumulated along the GS/CG interface, causing more dislocations interacted with GNDs and accumulated at the border of GS/CG layer. Dislocation structure annihilates and disentangles at the same time, leading to the reduction of Re. With increasing strain, multiaxial stresses in the samples lead to additional slip systems to be activated and more dislocation propagated, resulting in a lower exhaustion rate of mobile dislocations. This evolution of mobile dislocations probably improves the ductility and work-hardening ability of the material. 4.4. Superior combination of strength and ductility With the growing interests in UFG and NS materials, numerous processes have been proposed to synthesize these structures [34]. Cold rolling (CR) [35–37], accumulative roll bonding (ARB) [38], equal channel angular pressing (ECAP) [39,40] and high-pressure torsion (HPT) [41,42] are the most promising methods for producing bulk UFG metals while avoiding the consolidation step that often brings artifacts into the samples. UFG or NS metals typically exhibit high strength but poor ductility. In our study, the SMAT samples have superior combination of high strength and good ductility compared to other literature values [7,14,36, 37,41,43–45] (Fig. 6), exhibiting extraordinary mechanical properties. The SMAT processed gradient structures shown above have a unique strengthening characteristic, which leads to various combinations of strength and ductility. This also opens up another question about how to control the gradient layer thickness and tune for the superior mechanical properties. Comparing the hardness of three SMAT samples (Fig. 1), we found that the thickness of GS layer increase when the SMAT processing time increasing from 5 min to 15 min, but the thickness of GS layer no longer increases when the SMAT processing time increasing from 15 min to 30 min, indicating that the limit value of GS layer thickness is 140 μm under our facility conditions. In order to

Fig. 6. The superior combination of strength and ductility in the SMAT processed pure copper samples compared with their counterparts.

study the influence of gradient structure, we have compared the samples with different gradient structure volume fractions (VF) in gradient Cu. Based on our research, it was found that samples have the best combination of strength and ductility when VF is in the range of 0.08–0.1. 5. Conclusions Surface mechanical attrition treatment of bulk pure copper samples at room temperature produced gradient structure with gradually refined grain sizes near the surface. The SMAT copper samples showed a fourfold improvement in the yield strength and slight decrease of the uniform elongation compared with CG copper samples. The additional strengthening in the gradient structure is believed caused by the mechanical incompatibility between the CG and GS layers. The hardness of the treated surface is higher than 1.2 GPa and the maximum GS layer thickness is 140 μm, which corresponds to ultrafine and nanocrystalline structures. The SMAT processing time has great influence on the thickness of GS layer and therefore the resulted mechanical properties. The resulted high strength and good ductility in the gradient structured samples is believed related to the mismatch and interaction between the CG and GS layers. An indirect evidence of such interaction induced mobile dislocation activity is provided by quantitative stress relaxation tests. By controlling the component fraction of gradient structures, it's also possible to tune the strength and ductility to achieve the superior properties. Acknowledgements The authors would like to acknowledge financial supports by the National Natural Science Foundation of China (NSFC) (grant, 51561015), and the introduction of talents fund project of Kunming University of Science and Technology (grant KKSY201407100). Y.T. Zhu is supported by the US Army Research Office under the Grant Nos. W911NF-09-10427, W911QX-08-C-0083 and by the Nanjing University of Science and Technology. References [1] R.Z. Valiev, I. Sabirov, A.P. Zhilyaev, T.G. Langdon, Bulk nanostructured metals for innovative applications, JOM 64 (2012) 1134–1142. [2] Y.T. Zhu, X.Z. Liao, X.L. Wu, Deformation twinning in nanocrystalline materials, Prog. Mater. Sci. 57 (2012) 1–62. [3] C.C. Koch, Optimization of strength and ductility in nanocrystalline and ultrafine grained metals, Scr. Mater. 49 (2003) 657–662. [4] K. Lu, L. Lu, S. Suresh, Strengthening materials by engineering coherent internal boundaries at the nanoscale, Science 324 (2009) 349–352. [5] L. Lu, Y.F. Shen, X.H. Chen, L.H. Qian, K. Lu, Ultrahigh strength and high electrical conductivity in copper, Science 304 (2004) 422–426. [6] Y. Wang, M. Chen, F. Zhou, E. Ma, High tensile ductility in a nanostructured metal, Nature 419 (2002) 912–915. [7] Y.H. Zhao, X.Z. Liao, S. Cheng, E. Ma, Y.T. Zhu, Simultaneously increasing the ductility and strength of nanostructured alloys, Adv. Mater. 18 (2006) 2280–2283. [8] Y.H. Zhao, J.F. Bingert, X.Z. Liao, B.Z. Cui, K. Han, A.V. Sergueeva, A.K. Mukherjee, R.Z. Valiev, T.G. Langdon, Y.T. Zhu, Simultaneously increasing the ductility and strength of ultra-fine-grained pure copper, Adv. Mater. 18 (2006) 2949–2953. [9] X. Wu, N. Tao, Y. Hong, J. Lu, K. Lu, γ → ε martensite transformation and twinning deformation in fcc cobalt during surface mechanical attrition treatment, Scr. Mater. 52 (2005) 547–551. [10] Y. Ma, J.-E. Jin, Y.-K. Lee, A repetitive thermomechanical process to produce nanocrystalline in a metastable austenitic steel, Scr. Mater. 52 (2005) 1311–1315. [11] H. Kou, J. Lu, Y. Li, High-strength and high-ductility nanostructured and amorphous metallic materials, Adv. Mater. 26 (2014) 5518–5524. [12] K. Lu, J. Lu, Surface nanocrystallization (SNC) of metallic materials-presentation of the concept behind a new approach, J. Mater. Sci. Technol. 15 (1999) 193–197. [13] Y. Liu, B. Jin, J. Lu, Mechanical properties and thermal stability of nanocrystallized pure aluminum produced by surface mechanical attrition treatment, Mater. Sci. Eng. A 636 (2015) 446–451. [14] T.H. Fang, W.L. Li, N.R. Tao, K. Lu, Revealing extraordinary intrinsic tensile plasticity in gradient nano-grained copper, Science 331 (2011) 1587–1590. [15] X.L. Wu, P. Jiang, L. Chen, F.P. Yuan, Y.T.T. Zhu, Extraordinary strain hardening by gradient structure, Proc. Natl. Acad. Sci. U. S. A. 111 (2014) 7197–7201. [16] X.L. Wu, P. Jiang, L. Chen, J.F. Zhang, F.P. Yuan, Y.T. Zhu, Synergetic strengthening by gradient structure, Mater. Res. Lett. 2 (2014) 185–191.

Z. Yin et al. / Materials and Design 105 (2016) 89–95 [17] Y.M. Wang, A.V. Hamza, E. Ma, Activation volume and density of mobile dislocations in plastically deforming nanocrystalline Ni, Appl. Phys. Lett. 86 (2005) 241917. [18] F.H. Dalla Torre, E.V. Pereloma, C.H.J. Davies, Strain hardening behaviour and deformation kinetics of Cu deformed by equal channel angular extrusion from 1 to 16 passes, Acta Mater. 54 (2006) 1135–1146. [19] L. Lu, T. Zhu, Y.F. Shen, M. Dao, K. Lu, S. Suresh, Stress relaxation and the structure size-dependence of plastic deformation in nanotwinned copper, Acta Mater. 57 (2009) 5165–5173. [20] K. Lu, J. Lu, Nanostructured surface layer on metallic materials induced by surface mechanical attrition treatment, Mater. Sci. Eng. A 375–377 (2004) 38–45. [21] X.X. Huang, N. Hansen, N. Tsuji, Hardening by annealing and softening by deformation in nanostructured metals, Science 312 (2006) 249–251. [22] K. Wang, N.R. Tao, G. Liu, J. Lu, K. Lu, Plastic strain-induced grain refinement at the nanometer scale in copper, Acta Mater. 54 (2006) 5281–5291. [23] C. Suryanarayana, Mechanical alloying and milling, Prog. Mater. Sci. 46 (2001) 1–184. [24] X.L. Ma, C.X. Huang, W.Z. Xu, H. Zhou, X.L. Wu, Y.T. Zhu, Strain hardening and ductility in a coarse-grain/nanostructure laminate material, Scr. Mater. 103 (2015) 57–60. [25] M.A. Meyers, A. Mishra, D.J. Benson, Mechanical properties of nanocrystalline materials, Prog. Mater. Sci. 51 (2006) 427–556. [26] J. Moering, X.L. Ma, G.Z. Chen, P.F. Miao, G.Z. Li, G. Qian, S. Mathaudhu, Y.T. Zhu, The role of shear strain on texture and microstructural gradients in low carbon steel processed by surface mechanical attrition treatment, Scr. Mater. 108 (2015) 100–103. [27] L. Yang, L. Lu, The influence of sample thickness on the tensile properties of pure Cu with different grain sizes, Scr. Mater. 69 (2013) 242–245. [28] M.A. Meyers, K.K. Chawla, Mechanical Behavior of Materials, Prentice-Hall, Upper Saddle River (NJ), 1999. [29] T.H. Fang, N.R. Tao, K. Lu, Tension-induced softening and hardening in gradient nanograined surface layer in copper, Scr. Mater. 77 (2014) 17–20. [30] K. Lu, Making strong nanomaterials ductile with gradients, Science 345 (2014) 1455–1456. [31] D.K. Yang, P. Cizek, D. Fabijanic, J.T. Wang, P.D. Hodgson, Work hardening in ultrafine-grained titanium: multilayering and grading, Acta Mater. 61 (2013) 2840–2852.

95

[32] X. Yang, X. Ma, J. Moering, H. Zhou, W. Wang, Y. Gong, J. Tao, Y. Zhu, X. Zhu, Influence of gradient structure volume fraction on the mechanical properties of pure copper, Mater. Sci. Eng. A 645 (2015) 280–285. [33] D. Caillard, J.L. Martin, Thermally Activated Mechanisms in Crystal Plasticity, Pergamon, Amsterdam, 2003. [34] H. Gleiter, Nanostructured materials: basic concepts and microstructure, Acta Mater. 48 (2000) 1–29. [35] R. Kumar, S.M. Dasharath, P.C. Kang, C.C. Koch, S. Mula, Enhancement of mechanical properties of low stacking fault energy brass processed by cryorolling followed by short-annealing, Mater. Des. 67 (2015) 637–643. [36] X.Y. San, X.G. Liang, L.P. Cheng, C.J. Li, X.K. Zhu, Temperature effect on mechanical properties of Cu and Cu alloys, Mater. Des. 35 (2012) 480–483. [37] H. Bahmanpour, A. Kauffmann, M.S. Khoshkhoo, K.M. Youssef, S. Mula, J. Freudenberger, J. Eckert, R.O. Scattergood, C.C. Koch, Effect of stacking fault energy on deformation behavior of cryo-rolled copper and copper alloys, Mater. Sci. Eng. A 529 (2011) 230–236. [38] Y. Estrin, A. Vinogradov, Extreme grain refinement by severe plastic deformation: a wealth of challenging science, Acta Mater. 61 (2013) 782–817. [39] R.Z. Valiev, T.G. Langdon, Principles of equal-channel angular pressing as a processing tool for grain refinement, Prog. Mater. Sci. 51 (2006) 881–981. [40] F. Salimyanfard, M.R. Toroghinejad, F. Ashrafizadeh, M. Hoseini, J.A. Szpunar, Investigation of texture and mechanical properties of copper processed by new route of equal channel angular pressing, Mater. Des. 44 (2013) 374–381. [41] K. Edalati, T. Fujioka, Z. Horita, Microstructure and mechanical properties of pure Cu processed by high-pressure torsion, Mater. Sci. Eng. A 497 (2008) 168–173. [42] M. Kawasaki, T.G. Langdon, Review: achieving superplasticity in metals processed by high-pressure torsion, J. Mater. Sci. 49 (2014) 6487–6496. [43] S. Ranjbar Bahadori, K. Dehghani, F. Bakhshandeh, Microstructural homogenization of ECAPed copper through post-rolling, Mater. Sci. Eng. A 588 (2013) 260–264. [44] K. Hanazaki, N. Shigeiri, N. Tsuji, Change in microstructures and mechanical properties during deep wire drawing of copper, Mater. Sci. Eng. A 527 (2010) 5699–5707. [45] L. Lu, L.B. Wang, B.Z. Ding, K. Lua, High-tensile ductility in nanocrystalline copper, J. Mater. Res. 15 (1999) 270–273.