Zn multilayer processed by accumulative roll bonding (ARB)

Zn multilayer processed by accumulative roll bonding (ARB)

Materials Science & Engineering A 593 (2014) 145–152 Contents lists available at ScienceDirect Materials Science & Engineering A journal homepage: w...

5MB Sizes 0 Downloads 90 Views

Materials Science & Engineering A 593 (2014) 145–152

Contents lists available at ScienceDirect

Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea

Microstructure evolution and mechanical properties of Cu/Zn multilayer processed by accumulative roll bonding (ARB) L. Ghalandari a,n, M.M. Mahdavian b, M. Reihanian c a

Department of Materials Science and Engineering, Shiraz Branch, Islamic Azad University, Shiraz, Iran Department of Materials Science and Engineering, Ahvaz Branch, Islamic Azad University, Ahvaz, Iran c Department of Materials Science and Engineering, Faculty of Engineering, Shahid Chamran University, Ahvaz, Iran b

art ic l e i nf o

a b s t r a c t

Article history: Received 23 July 2013 Received in revised form 24 October 2013 Accepted 11 November 2013 Available online 19 November 2013

Cu/Zn multilayer was processed by accumulative roll bonding (ARB) up to eight cycles. The evolution of microstructure and its correlation with mechanical properties was investigated. Necking and rupture of constituents were observed at intermediate cycles. The intermetallic CuZn5, microcracks and Kirkendall porosities were formed during ARB. After eight ARB cycles, a laminated composite with wavy interfaces and lenticular fragments of Cu was produced. The maximum tensile strength was achieved after four cycles, which was about 1.4 times higher than that of pure Cu. The hardness of individual layers increased continually with increasing the ARB cycle. Tensile fracture surfaces reveal dimples in both Cu and Zn at early ARB cycles. & 2013 Elsevier B.V. All rights reserved.

Keywords: Composites Bulk deformation Mechanical characterization

1. Introduction The manufacture of metal matrix composites (MMCs) can extensively improve many material properties such as specific modulus, tensile strength, fracture toughness, impact behavior and resistance to erosion and corrosion [1,2]. Hence, in the past few years, a significant increase in the application of MMCs has taken place, particularly in the fields of automotive, aerospace and electronics industries. The laminated metal composites with improved properties have also found to be the object of the ancient and modern industries [3]. Modern laminated metal composites can be made by many techniques such as bonding, deposition, and spray forming. Bonding techniques can include other methods such as adhesive bonding, melt bonding, infiltration, diffusion bonding, reaction bonding, and deformation bonding. As a type of solid-state process for joining similar and/or dissimilar metals, cold roll-bonding is considered as a common route for manufacturing laminated metal composites because it allows for the cost-effective continuous production of multilayered clad metals [4]. In recent years, accumulative roll bonding (ARB) has been used as a severe plastic deformation (SPD) technique for producing bulk ultrafine/nano-grained metallic materials. The method was first introduced by Saito et al. [5] as a process to overcome major limitations of other SPD methods, namely, the low productivity and the small work-piece size. The greatest technological advantage

n

Corresponding author. Tel.: þ 98 7116462300; fax: þ 98 7116191626. E-mail address: [email protected] (L. Ghalandari).

0921-5093/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.msea.2013.11.026

of ARB is that it makes use of a conventional rolling facility and has the capability of producing large-scale sheet metals. More recently, ARB has been utilized to fabricate particulate metal–matrix composites such as Al/SiC [6], Cu/Al2O3 [7], Al/Al2O3 [8] and Al/B4C [9]. It has been shown that a uniform distribution of particles can be achieved after imposing a critical reduction [10,11]. ARB has also been employed in order to produce lamellar metal matrix composites such as Al/Mg [12–14], Al/Ni [15–17], Al/Cu [18], Cu/Ag [19], Al/Zn [20], Cu/ Nb [21], Cu/Ni [22] and Ti/Al [23]. It has been reported that a bulk twophase bimetal nanocomposite processed by SPD uniquely possesses simultaneous high strength and high thermal stability [24]. In addition, a few attempts have been made to produce tri-metallic systems such as Al/Ni/Cu [25] and Cu/Zn/Al [26]. Among many bimetal systems, the microstructure development and mechanical properties of Cu/Zn dissimilar alloy combination have not yet been investigated. Accordingly, in the present paper, multilayered Cu/Zn is used as a new bimetal system and is processed by ARB. The evolution of the microstructure and its correlation with mechanical properties is examined and the results are discussed. 2. Material and methods The materials utilized in this study were commercial pure Cu (99.9 wt%) and Zn (98.86 wt%) in the form of strips with a thickness of 1 mm. The whole composition of the materials is presented in Table 1. The strips were cut into 50 mm  150 mm in size and annealed at 500 C for 60 min (for Cu) and at 150 C for 30 min (for Zn). The annealed strips were degreased by acetone

146

L. Ghalandari et al. / Materials Science & Engineering A 593 (2014) 145–152

The process was carried out at room temperature and repeated up to eight cycles without lubrication. ARB experiments were conducted with a rolling machine having the capacity of 20 t and rolling diameter of 170 mm. The rolling speed was set at 12 rpm. Normal direction–rolling direction (RD–ND) sections were prepared for investigation of the microstructure by field emission scanning electron microscope (FESEM) equipped with energy dispersive spectroscopy (EDS). In order to identify the phases, the x-ray diffraction (XRD) pattern of the composite was performed by a diffractometer, operating at 40 kV and 40 mA with Cu Kα radiation and a step time of 58 s.

and scratch brushed with a circular stainless steel brush having a 0.3 mm wire diameter. The sheets were stacked with two Cu layers as the outer surfaces and one Zn layer as the inner surface. To set the prepared surfaces in contact and fixed to each other closely, four holes, which had been drilled near the edges of the strips, were bound tightly by copper wires. The primary sandwich underwent into two processing routes. In the first (zero cycle), it was cold rolled through a 57% reduction in thickness. In the second route (ARB process), sandwiches were cut into two halves, degreased, wire brushed, stacked and fixed by copper wires at the edges. Roll bonding was conducted with a reduction of 50% in thickness.

Table 1 Chemical composition (wt%) of the materials used in this study. Cu (99.9) Zn

Pb

Sn

P

Mn

Fe

Ni

Si

Ag

Co

Mg

Cr

Al

S

As

o0.005

0.029

o0.005

0.003

0.007

o 0.005

o0.005

o 0.005

0.008

0.013

0.002

0.002

0.002

o 0.002

0.004

Pb

Al

Si

Mg

Fe

Ca

Sx

Ta

F

Cd

K

In

Cl

Ge

Mn

Ti

Sc

0.493

0.199

0.131

0.089

0.0728

0.0306

0.0258

0.0239

0.0208

0.0145

0.0128

0.0085

0.0081

0.0067

0.0043

0.0028

0.0017

Zn (98.86)

Zn

Cu

Cu

ND Zn RD

Fig. 1. SEM micrographs of ARB processed Cu/Zn composite after (a) one cycle, (b) three cycles, (c) five cycles and (d) seven cycles.

L. Ghalandari et al. / Materials Science & Engineering A 593 (2014) 145–152

Tensile samples were machined from the ARB processed strips, according to the ASTM E8M standard, with orientation along the rolling direction. The gauge width and length of the samples were 3 and 8 mm, respectively. Tensile test was done at room temperature with a tensile testing machine under a nominal initial strain rate of 10  4 s  1. Vickers microhardness was done using a Koopa apparatus under a load of 10 g and a time of 15 s on the RD–ND sections. Hardness tests were conducted on Cu and Zn layers, separately, in more than six points.

3. Results and discussion 3.1. Microstructure evolution SEM micrographs of Cu/Zn multilayer after various ARB cycles are shown in Fig. 1. It is observed that, as the ARB proceeds, the number of layers increases while their thickness decreases. At early stages of ARB (up to three cycles), the interface between the layers is straight and adherent. At intermediate cycles, some irregularities occur at the interfaces and they become wavy. In addition, necking starts at various locations in the microstructure. Rupturing along the necked regions produces lenticular fragments in the microstructure, which become most prominent at the last stages. The details of microstructure evolution are presented in the following figures and discussions. Fig. 2 shows SEM image of the multilayer at two magnifications after four cycles. The occurrence of necking in Cu layers is obvious after four cycles. In general, due to the difference in flow properties of the constituent phases, the co-deformation of multilayers results in necking and subsequent rupture of the hard phase [27,28]. The plastic instability is controlled by the initial thickness ratio, strength coefficients and strain hardening exponents of the layers [28,29]. In the case of Cu/Zn system, Cu acts as the hard phase [26] and necks at first. The locations of necking in Cu layers are indicated by arrows in Fig. 2. Necking in the hard phase has also been reported in other bimetal systems such as in Al/Ni [16], Al/Cu [18], Al/Zn [20] and Cu/Ni [22]. Inspection of micrographs also indicates some bright phases in Zn layers. Fig. 3 shows the results of EDS analysis used to recognize the type of bright phases in the Zn matrix. Results indicate that the bright phases can be Pb in Zn matrix (since Pb is insoluble in Zn). These results are consistent with the chemical composition of the asreceived Zn (Table 1) which contains the highest value of Pb as an impurity element. Fig. 4 illustrates the SEM image accompanied by the

147

elemental distribution map after five ARB cycles. At this stage, necking in Cu layer is extended until its thickness reduces substantially and two sequential Zn layers get in touch with each other at that location. The Cu layers, in this condition, take apart along the necked regions, producing lenticular fragments and wavy interface in the microstructure. Concurrently, thinning and necking of Zn layers is started at some regions, as indicated by arrows. SEM/elemental maps of the multilayer after six ARB cycles are presented in Fig. 5. It is seen that lenticular fragments of the Cu layers form at more regions due to imposition of large plastic strains. In addition, the extension of necking in Cu layers causes the shearing of a number of Zn layers along about 451 to the rolling direction. These microstructural features are most prominent after eight ARB cycles, as shown in Fig. 6. It is noted that in bimetal systems where the initial thickness of the hard layer is smaller than that of the soft layer, the final microstructure consists of a uniform distribution of fragmented hard layers within the soft matrix [12,18–20,22]. Conversely, in the present investigation, the final structure consists of lamellar layers with wavy interfaces as well as several lenticular fragments of the hard phase. This is because the initial thickness of the hard layer is larger than that of the soft layer (the total initial thickness for Cu is 2 mm and for Zn is 1 mm). This microstructure feature is similar to other bimetal systems where the initial thicknesses of the constituents are comparable [30,31]. Nevertheless, it has been shown

Element Zn Pb

Wt.% 54.89 45.11

Fig. 3. EDS results corresponding to bright phases appearing in Zn layer.

Necking Cu

Necking

Zn

Necking

Necking Cu Necking

Necking

Zn

Cu

ND

At.% 79.40 20.60

RD Fig. 2. SEM micrographs of ARB processed Cu/Zn composite after four cycles taken at (a) low and (b) high magnification.

148

L. Ghalandari et al. / Materials Science & Engineering A 593 (2014) 145–152

Fragmentation

Substantial necking in Cu

Necking in Zn

ND RD Fig. 4. (a) SEM image of Cu/Zn composite after five cycles and the corresponding SEM elemental maps for (b) Cu distribution and (c) Zn distribution.

Shearing

Lenticular fragments

ND

RD Fig. 5. (a) SEM image of Cu/Zn composite after six cycles and the corresponding SEM elemental maps for (b) Cu distribution and (c) Zn distribution.

Fig. 7. The XRD patterns for Cu/Zn composite subjected to (a) five cycles, (b) six cycles, (c) seven cycles and (d) eight cycles.

ND

RD Fig. 6. SEM image of Cu/Zn composite after eight cycles.

that a composite with a uniform distribution of fragments will be achieved, if the multilayer is conducted to relatively high ARB cycles [26].

The XRD patterns for the specimens subjected to five, six, seven, and eight cycles are presented in Fig. 7. All patterns show reflections due to the elements Cu and Zn, as expected. The intensity of Zn peaks after five cycles is comparatively small. At early stages of ARB, the thickness of surface layer is larger than the x-ray penetration depth and hence little Zn can be detected. The increase in intensity of Zn peak particularly at low angles becomes noticeable at higher ARB cycles. The important feature of the patterns is that a small peak of CuZn5 appears after five cycles.

L. Ghalandari et al. / Materials Science & Engineering A 593 (2014) 145–152

149

Zn Cu

Cavity

Cavity

Zn Cu Cavity

Zn ND

Crack

Cu

Crack

RD Fig. 8. SEM micrographs of the Cu/Zn composite after (a) seven cycles and (b) eight cycles.

Engineering stress(MPa)

350 300

Cycle4 Cycle5

250

Cycle2

Cycle6

Cycle3 Cycle1

200 150

Cu-annealed Cycle7 Zn-annealed

100 50 0 0.0

0.1

0.2 0.3 0.4 0.5 Engineering strain(mm/mm)

0.6

300

100

250

80

200 Ultimate tensile stress(MPa) Elongation(%)

150 100

40 20

50 0

60

Elongation(%)

Fig. 9. Engineering stress–strain curves for Cu/Zn composite after different ARB cycles.

Ultimate tensile stress(MPa)

As the ARB proceeds and the layer thickness decreases, the number of interfaces per unit volume is raised. Given a 2 μm penetration depth in typical XRD, there will be a better chance of detecting an interface where intermetallics have formed. As a result, the intensity of intermetallic increases to some extent as the ARB cycles are increased. The formation of intermetallic phases is also reported during ARB of other multilayers [14,17,32–34]. The reason for the formation of intermetallic phases can be attributed to the intense straining during ARB that causes the increase in grain boundaries and dislocation density. This accelerates the elemental diffusion at interfaces and promotes the formation of intermetallic [35]. SEM micrographs of the multilayer after seven and eight ARB cycles are presented in Fig. 8a and b, respectively. The presence of some dark regions (porosities) on the Zn layers is noticeable. The formation of these porosities may be due to Kirkendall effect that occurs in the diffusion couples when the diffusion rates of both species are significantly different [36]. Kirkendall porosities have also been observed in Zn layer during ARB of Cu/Zn/Al multilayer [26]. It is noticed that Kirkendall effect is a diffusional process that normally needs thermal energy to overcome the activation energy for diffusion. Since ARB is carried out at room temperature, the activation energy for diffusion must be provided by another source. During ARB, the lattice strain and dislocation density cause the decrease in activation energy for diffusion and increase pipe diffusion. Further, severe plastic deformation refines the grain sizes and increases the grain boundary area as a high-diffusivity path. Therefore, atoms can diffuse by each other at the interfaces and a critical strain (a number of cycles) is needed to provide the activation energy for the diffusion of the pieces. According to the analysis of stresses during sandwich sheet rolling [28], when the inner layer is softer than the outer layers, the inner layer comes under a state of all compressive stresses. On the other hand, the equilibrium concentration of vacancies decreases under the compression state of stress. Since vacancies have the major role for self-diffusion through vacancy diffusion mechanism, it may be expected that Kirkendall porosities form after exiting the multilayer from the roll gaps. However, the formation of the Kirkendall porosities under the dynamic or static conditions needs further investigating, which is on the margin of the scope of this paper. Inspection of Fig. 8 also reveals several microcracks at the interface of Cu and Zn. The volume changes during intermetallic formation and the expansion mismatch between the intermetallic and other layers can create cracks between the layers [17,37]. It is

1

2

3 4 5 6 Number of ARB cycles

0 7

Fig. 10. Variation in UTS and percent elongation with various ARB cycles.

noticed that the thickness of intermetallic is too low to be visible via SEM. Other techniques such as transmission electron microscopy could help to observe the intermetallic phase. 3.2. Mechanical properties The engineering stress–strain curve of Cu/Zn multilayer for various ARB cycles is shown in Fig. 9. For comparison, the

150

L. Ghalandari et al. / Materials Science & Engineering A 593 (2014) 145–152

engineering stress–strain curve of annealed Cu and Zn is also included. It is noted that the tensile properties are obtained up to seven cycles because during ARB, edge cracks form due to tensile stresses near the strip edges. Therefore, the edges of the rollbonded strips must be trimmed after each cycle and preparation of tensile specimens becomes more difficult for higher ARB cycles.

Vickers microhardness (VHN)

140 120 100 80 60 Cu Zn

40 20 0 0

2

4 6 Number of ARB cycles

8

Fig. 11. Variation of microhardness for individual layers with different ARB cycles.

Cu

Variation in ultimate tensile stress (UTS) and percent elongation with number of ARB cycles are plotted in Fig. 10. It is seen that with increasing the ARB cycle, the UTS of multilayer rises to its maximum value of 280 MPa, which is around 1.4 times higher than that of annealed Cu. UTS drops after five cycles and decreases continually. The strength decrement has also been observed during ARB processing of Mg/Al laminated composite [14]. As ARB proceeds, UTS increases as a result of strain hardening and grain refinement that are typical strengthening mechanisms in SPD methods [38]. The lamellar structure of Cu/Zn multilayer causes additional strengthening effect because the interfaces between constituents can provide an effective barrier for dislocation motions. The drop of strength after five cycles can be attributed to the substantial necking of Cu layers (Fig. 4) that leads to the separation and fragmentation of Cu layers. Generally, the tensile strength of the laminated composites can be dictated by the flow properties of both constituents according to the rule of mixture. Since Cu layers neck substantially part after five ARB cycles, the strength introduced by Cu layers is not adequate to support sufficient load and resist deformation. In addition, the formation of microcracks at the interfaces may destroy the bonding at the interfaces and cause the decrease in strength after five ARB cycles. Formation of Kirkendall porosities in Zn layer can also be another reason for the drop of strength (Fig. 8).

Zn

Fig. 12. Tensile fracture surfaces of Cu/Zn composite after (a) one cycle, (b) three cycles, (c) six cycles and (d) eight cycles.

L. Ghalandari et al. / Materials Science & Engineering A 593 (2014) 145–152

Variation of percent elongation with ARB cycles shows that the percent elongation of the multilayer decreases with increasing the ARB cycle. Briefly, the decrease in elongation is due to strain hardening that causes the increase in dislocation density and accumulation of internal stresses, leading to the nucleation of cracks. Substantial necking of Cu layers that leads to interruption of layers is another reason for decrease in elongation. The drop of elongation is noticeable after five ARB cycles which may be due to the formation of CuZn5 and cracking at the layer interfaces. Variations of hardness for Cu and Zn layers with different ARB cycles are indicated in Fig. 11. Contrary to tensile strength, there is not any drop in hardness variations because hardness measurement is carried out individually for each layer and, thus, the hardness of the layers rises with increasing the ARB cycles. However, at initial stages, the hardness rises with a higher rate. This is a typical trend for hardness during SPD methods in which the dislocation density increases and the average boundary spacing decreases; both contributing to the hardening of material at early cycles. Nevertheless, at final stages, dislocation density and boundary spacing approximately saturates at fixed value, leading to a plateau in hardness variation [39]. Comparison of hardness variation indicates that the hardness of Cu layer is larger than that of Zn layer for all ARB cycles. The difference in hardness can be attributed to the difference in stacking fault energies of the materials. The stacking fault energies of Cu and Zn are 78 and 140 mJ/m, respectively. The hardness of Cu, with lower stacking fault energy, increases at a higher rate. Zn with a hexagonal close-packed structure has limited slip systems and, therefore, stacking fault energy controls the mode of deformation. The difficulty of cross slip reduces the ability of the material to deform plastically by slip, and thus deformation twinning may occur. Thermal expansion coefficient of the constituents can also influence the hardness variations during ARB [25,40]. Plastic deformation and friction between the work piece and die can raise the temperature during ARB. Therefore, thermal stresses will be generated between the layers if the difference in thermal expansion coefficient of the constituents is sufficiently high. The thermal expansion coefficients of Cu and Zn are 17  10  6 and 27.4  10  6 K  1, respectively. As a result, the thermal stresses may be introduced at the interfaces that lead to generation of new dislocations and increasing of the hardness. 3.3. Tensile fracture surface Fig. 12 illustrates the tensile fracture surfaces of the Cu/Zn multilayer subjected to one, three, six and eight cycles. The SEM image and the corresponding elemental maps of fracture surface after one cycle are shown in Fig. 12a. At this stage, the fracture appearance of the multilayer reveals dimples in Cu and Zn layers, indicating that the fracture mode is ductile. Ductile fracture takes place by nucleation and coalescence of microvoids ahead of the main crack. Dimple type fracture has also been observed in Al/Cu [18], Al/Zn [20], Cu/Ni [22] and Al/Ni/Cu [25] systems. After three ARB cycles (Fig. 12b), dimpling is yet evident in both layers. However, the dimples in Zn are shallow and wide that may be attributed to the lower ability of Zn for plastic deformation compared with the Cu layer. Fig. 12c indicates that fracture morphology is changed to some extent after six cycles. Nevertheless, it is not a strong statement that there is a transition from ductile to brittle in Cu or that Zn is undergoing decisively brittle fracture. After eight cycles (Fig. 12d), it is somewhat difficult to identify the individual layers as well as their fracture mode due to increase in layer number and decrease in layer thickness; a higher magnification micrograph would help with this. In addition, delamination or debonding between layers can be observed on

151

the fracture surface of all samples. It has occurred due to generation of tensile stresses during tensile test that causes separation of the layer interfaces. Delamination has also been reported on the fracture surface of 2219/5086 laminates [31] after tensile testing.

4. Conclusions A multilayer of Cu/Zn was processed by accumulative roll bonding (ARB) up to eight cycles and the following results were obtained: 1. The interface between the layers remains straight and adherent up to three cycles. After four cycles, necking starts at various locations in the microstructure and the interfaces become wavy. Substantial necking of Cu layers is observed after five cycles, producing lenticular fragments in the microstructure. 2. During ARB, the intermetallic CuZn5, microcracks (at interfaces) and Kirkendall porosities (in Zn layer) are formed in the microstructure. 3. Tensile strength rises to its maximum value of 280 MPa, and then decreases continually up to seven cycles. The percent elongation decreases continually up to seven cycles. 4. The hardness of the layers increases with increasing the ARB cycles. 5. Dimple type fracture is more evident in both Cu and Zn layers at early stages of ARB.

Acknowledgment The authors would like to thank the financial support of Islamic Azad University (Shiraz Branch). References [1] T.W. Clyne, P.J. Withers, An Introduction to Metal Matrix Composites, first ed., Cambridge University PressUK, 1995. [2] N. Chawla, K.K. Chawla, Metal Matrix Composites, first ed., Springer, USA, 2006. [3] J. Wadsworth, D.R. Lesuer, Mater. Charact. 45 (2000) 289–313. [4] L. Li, K. Nagai, F. Yin, Sci. Technol. Adv. Mater. 9 (2008) 023001. [5] Y. Saito, H. Utsunomiya, N. Tsuji, T. Sakai, Acta Mater. 47 (1999) 579–583. [6] M. Alizadeh, M.H. Paydar, J. Alloys Compd. 477 (2009) 811–816. [7] R. Jamaati, M.R. Toroghinejad, Mater. Sci. Eng. A 527 (2010) 7430–7435. [8] M. Rezayat, A. Akbarzadeh, A. Owhadi, Compos.: Part A 43 (2012) 261–267. [9] A. Yazdani, E. Salahinejad, J. Moradgholi, M. Hosseini, J. Alloys Compd. 509 (2011) 9562–9564. [10] M. Reihanian, E. Bagherpour, M.H. Paydar, Mater. Sci. Technol. 28 (2012) 103–108. [11] M. Reihanian, E. Bagherpour, M.H. Paydar, Mater. Lett. 91 (2013) 59–62. [12] M.C. Chen, H.C. Hsieh, W. Wu, J. Alloys Compd. 416 (2006) 169–172. [13] H. Chang, M.Y. Zheng, W.M. Gan, K. Wu, E. Maawad, H.G. Brokmeier, Scr. Mater. 61 (2009) 717–720. [14] K. Wu, H. Chang, E. Maawad, W.M. Gan, H.G. Brokmeier, M.Y. Zheng, Mater. Sci. Eng. A 527 (2010) 3073–3078. [15] G. Min, J.-M. Lee, S.-B. Kang, H.-W. Kim, Mater. Lett. 60 (2006) 3255–3259. [16] A. Mozaffari, H. Danesh Manesh, K. Janghorban, J. Alloys Compd. 489 (2010) 103–109. [17] A. Mozaffari, M. Hosseini, H.D. Manesh, J. Alloys Compd. 509 (2011) 9938–9945. [18] M. Eizadjou, A. Kazemi Talachi, H. Danesh Manesh, H. Shakur Shahabi, K. Janghorban, Compos. Sci. Technol. 68 (2008) 2003–2009. [19] L. Ghalandari, M.M. Moshksar, J. Alloys Compd. 506 (2010) 172–178. [20] R.N. Dehsorkhi, F. Qods, M. Tajally, Mater. Sci. Eng. A 530 (2011) 63–72. [21] J. Wang, K. Kang, R.F. Zhang, S.J. Zheng, I.J. Beyerlein, N.A. Mara, JOM 64 (2012) 1208–1217. [22] M. Tayyebi, B. Eghbali, Mater. Sci. Eng. A 559 (2013) 759–764. [23] D. Yang, P. Cizek, P. Hodgson, C.E. Wen, Scr. Mater. 62 (2010) 321–324. [24] S. Zheng, I.J. Beyerlein, J.S. Carpenter, K. Kang, J. Wang, W. Han, N.A. Mara, Nat. Commun. 4 (2013) 1696. [25] A. Shabani, M.R. Toroghinejad, A. Shafyei, Mater. Sci. Eng. A 558 (2012) 386–393.

152

L. Ghalandari et al. / Materials Science & Engineering A 593 (2014) 145–152

[26] M.M. Mahdavian, L. Ghalandari, M. Reihanian, Mater. Sci. Eng. A 579 (2013) 99–107. [27] S.L. Semiatin, H.R. Piehler, Metall. Mater. Trans. A 10 (1979) 97–107. [28] Y.-M. Hwang, H.-H. Hsu, H.-J. Lee, Int. J. Mach. Tools Manuf. 36 (1996) 47–62. [29] J.-M. Lee, B.-R. Lee, S.-B. Kang, Mater. Sci. Eng. A 406 (2005) 95–101. [30] S. Roy, B.R. Nataraj, S. Suwas, S. Kumar, K. Chattopadhyay, Mater. Des. 36 (2012) 529–539. [31] S. Roy, B.R. Nataraj, S. Suwas, S. Kumar, K. Chattopadhyay, J. Mater. Sci. 47 (2012) 6402–6419. [32] H. Chang, M.Y. Zheng, C. Xu, G.D. Fan, H.G. Brokmeier, K. Wu, Mater. Sci. Eng. A 543 (2012) 249–256. [33] V.C. Srivastava, T. Singh, S. Ghosh Chowdhury, V. Jindal, J. Mater. Eng. Perform. 21 (2012) 1912–1918.

[34] [35] [36] [37]

S.J. Yoo, S.H. Han, W.J. Kim, Scr. Mater. 67 (2012) 129–132. R. Zhang, V.L. Acoff, Mater. Sci. Eng. A 463 (2007) 67–73. I.D. Choi, D.K. Matlock, D.L. Olson, Mater. Sci. Eng. A 124 (1990) L15–L18. M. Danaie, C. Mauer, D. Mitlin, J. Huot, Int. J. Hydrog. Energy 36 (2011) 3022–3036. [38] R.Z. Valiev, R.K. Islamgaliev, I.V. Alexandrov, Prog. Mater. Sci. 45 (2000) 103–189. [39] M. Reihanian, R. Ebrahimi, N. Tsuji, M.M. Moshksar, Mater. Sci. Eng. A 473 (2008) 189–194. [40] R. Jamaati, S. Amirkhanlou, M.R. Toroghinejad, B. Niroumand, Mater. Sci. Eng. A 528 (2011) 2143–2148.