- Email: [email protected]
High leaded tin bronze processing during multi-directional forging: Effect on microstructure and mechanical properties
Materials Science & Engineering A 654 (2016) 282–291
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
High leaded tin bronze processing during multi-directional forging: Effect on microstructure and mechanical properties Rahul Gupta a,n, Sanjay Srivastava a, Nand Kishor Kumar b, Sanjay K. Panthi c a
Department of Material Science and Metallurgical Engineering, MANIT, Bhopal 462003, India Department of Metallurgical and Materials Engineering, Indian Institute of Technology Kharagpur, 721302, India c Advanced Materials and Processes Research Institute, Bhopal 462024, India b
art ic l e i nf o
a b s t r a c t
Article history: Received 28 October 2015 Received in revised form 19 December 2015 Accepted 21 December 2015 Available online 22 December 2015
The high leaded tin bronze alloy has been studied during multidirectional forging (MDF) process with differing crystallite size (D) in the scale of 240–90 nm. The effect of MDF strain ( εMDF ∼ 0.25, 0.50, 0.75) after 3, 6 and 9 passes respectively on the development of homogeneity and refinement in terms of dislocation density, twin spacing, twin lamellae thickness and their division have been studied using high-resolution transmission electron microscopy (HRTEM) and micro-hardness measurements using Vickers micro-hardness. X-ray diffraction peak broadening investigation has revealed that the crystallite size decreases down to 90 nm after 9 pass ( εMDF ∼ 0.75). The tensile fracture surfaces have been studied using scanning electron microscope (SEM) of received (AR) and MDF specimens. The results of crystallite size and dislocation density on the strength of refined microstructure and nanostructure have been correlated with mechanical flow stress. & 2015 Elsevier B.V. All rights reserved.
Keywords: Multi-directional forging High leaded tin bronze Microstructure Flow stress X-ray diffraction
1. Introduction In last decade, the production of bulk nanostructured metals and alloys using severe plastic deformation (SPD) processes has been progressed rapidly with the aim of developing novel mechanical and functional properties in the materials [1]. SPD techniques like as multidirectional forging (MDF), equal channel angular pressing (ECAP), high pressure torsion (HPT), accumulative roll bonding (ARB) have been used to produce ultrafine grained material by applying very large plastic strain in various materials [2–4]. During 1990 MDF was introduced for the first time to fabricate ultrafine grained structure in bulk billets [5,6]. MDF is a repetition of free forging operation multiple times including setting and pulling operations with the changes of the axes of applied load [7]. Many researchers have been reported to study microstructure and mechanical behavior of copper and copper alloy after SPD [8– 13]. Parimi et al. [8] reported SPD of copper and Al–Cu alloy by using multiple channel-die compression process. The nano size grain structured copper was achieved along with increased flow stress and hardness during compression. Kundu et al. [9] applied multiple compressions in a channel die at different temperature range on copper and result enhanced flow stress and strain rate n
Corresponding author. E-mail address: [email protected] (R. Gupta).
http://dx.doi.org/10.1016/j.msea.2015.12.068 0921-5093/& 2015 Elsevier B.V. All rights reserved.
sensitivity. Sarkar et al. [14] deformed interstitial free steel through ECAP to assess the microstructure by different XRD peak profile analysis using Williamson–Hall plot technique and the variance method as a function of strain (ε). The dislocation densities estimated from XRD were also correlated with experimental yield strength. Gubicza et al. [15] studied the microstructure of face centered cubic (FCC) metal and alloys processed through SPD methods and also correlate the yield strength with dislocation densities calculated using Taylor's equation. Thereby they reported the main strengthening mechanism for both metals and alloys were the interaction between dislocations. Ungar et al. [16] determined the particle size and dislocation structure in nanostructured copper by high resolution XRD profile analysis method using modified Williamson–Hall plot and modified Warren Aver Bach’s plot and also reported that the twin boundaries found good agreement with TEM micrograph. Makhlouf et al. [17] studied the microstructure of Al alloy after ECAP using XRD and TEM. The high dislocation density along with improved mechanical behavior was obtained in the alloy. The high leaded tin bronze (copper based alloy) was used in the present study finds application in automotive, agricultural, rail road, mining, off highway equipments [18]. The high leaded tin bronze has lower strength which makes this alloy to only use under moderate/light loading conditions and against unhardened shafts. The trouble with the target alloy was that it has not been specified to use under high loading or impacts applications. Since these alloy operated under moderate loads and high speeds only.
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
But the situation for high loading and against hardened shafts was required the strength of the alloy to be improved. Therefore, the introduction of new mechanical properties in the alloy through MDF was aimed, along with its microstructural studies, so that the alloy functional conditions has been varied, such as used under high speeds, under high-load, and impact situations [19,20]. The production of bulk nanostructured metals and alloys (such as high leaded tin bronzes) has been carried out by using severe plastic deformation (SPD) with the aim of developing novel mechanical and functional properties. Generally, Cu-based alloys have been used in ancient times such as Romans and other civilizations to produce artistic and functional artifacts. Now days, we have been obtained more novel mechanical and functional properties, which as reported by many researcher from long time ago. The lowest grain size was detailed to be 400–600 nm for SPD processed pure Al [21], grain size between 100–450 nm for Cu alloy [22–24] also 70–110 nm for Cu–30 wt% Zn alloy [25]. In addition, post SPD recrystallization fabricates superior microstructure with ∼4.5 μm grain size for Cu–5 wt% Al and for Cu–5 wt% Zn ∼7.5 μm respectively [26]. Twinning is one of the key deformation methods for metals and alloys with small stacking fault energy (SFE), for silver (25 mJ/m2), stainless steel (8–45 mJ/m2), and α-brass (14 mJ/ m2) respectivley [27–29]. While, inclusion of Zn in Cu, 30% Zn alloy decreases the stacking fault energy (SFE) of pure Cu from 78 mJ/m2 to 14 mJ/m2 [27,29,30]. The dislocation density of Cu–0.7% Cr was 38 74 1014 and hardness was 804 738 Hv through severely
283
deformed by HPT at room temperature (RT) and a rate of 1 rpm under a pressure of 4 GPa [31]. The hardness was 173.2 74.3 Hv and yield strength was 640 720 MPa at CR095 for Cu–30 wt% Zn [27]. The present investigation is aimed to study the microstructure and mechanical properties of high leaded tin bronze processed by MDF at room temperature. The characterization of MDFed microstructure has been carried out in terms of twin spacing (dtwin), crystallite size, twin lamellae thickness (λ) and their division at different level of MDF, to investigate the fundamental mechanism of microstructural modification. The result of strengthening due to twin spacing, dislocation density, and crystallite size on the mechanical flow stress of differently MDF specimens have been compared in this research work.
2. Experimental procedures The material used in this study was high leaded tin bronze alloy (Cu–17 wt% Pb–6 wt% Sn–4.5 wt% Zn). Initially the alloy was cut into the measurement of 60 mm 55 mm 55 mm from square section and heat treated at 573 K (300° Celsius) for 1 hour using muffle furnace followed by water quenching. The process has been applied at room temperature and inside the MDF die [CAD model is shown in Fig. 1(a)] between top punch and bottom punch and pressed vertically downward number of times by changing the
Fig. 1. Schematic showing (a) Computer aided design model of MDF die (Exploded View), Top punch, Bottom punch, and Left and Right die halves with Allen bolts. (b) Schematic of MDF processing of high leaded tin bronze for different passes.
284
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
direction of loading from pass to pass along three mutually perpendicular axes i.e. in order of z to x, x to y, y to z and so on… as shown in Fig. 1(b) [32]. The process has been operated at constant pressing speed of 10 mm min 1 with MDF capacity press of 200 t. The total equivalent strain of 0.75 was applied during 3 MDF cycles after 9 passes. MoS2 was used as a lubricant between the samples and die contacts during compression. X-ray diffraction (XRD, Philips PANalytical PW 3373, Netherlands) was performed for structural investigation with Cu Kα radiation. Thin slices have been prepared by Ar-ion milling (PIPS-691, Gatan). Microstructural characterization was performed with HRTEM (JEM-2100, JEOL, Japan) operated at 200 kV. Vickers micro-hardness were taken out using a Bühler MicroMet 5103 hardness tester with 15 s dwell time at 50 gf load of AR (As-received) and MDF specimens. Hardness has been evaluated at each location on specimens were estimated by averaging the hardness of four surrounding points positioned at the edge of a 500 μm 500 μm size square. Parallelepiped specimens (2 2 4 mm3) were cut from the AR and MDF bars and were tested using Tinius Olsen H50KS (Germany) universal testing machine. The different surfaces of the specimens were polished carefully to retain uniaxiality throughout the compressive loading at a strain rate of 8 10 4/s under room temperature. The tensile fracture behavior of the MDF alloy has been examined using scanning electron microscope (JEOL JSM 6390 A) operated at an accelerating voltage of 25 kV.
single fcc phase directing that no phase alteration had occurred during MDF. On the other hand, the width of all the peaks became wider in MDF specimens than that of AR. Usually crystallite size, lattice strain and the instrumental broadening are the chief aspects, which add on peak broadening [16]. The instrumental line broadening has been calculated by a standard Si-disc (PW3132/62) of thickness and diameter, 2 mm and 32 mm respectively. The specimens’ peak broadening (β) has been précised by using the following Eq. (1) [33].
β=
2 (βobs − βSi2)
(1)
where, βobs and βSi are the integral breath of the equivalent (h k l ) reflections by the specimen and standard silicon disc respectively. The average nanocrystalline size was determined by using DebyeScherrer's formula:
D=
Kλ βhkl cos θ
(2)
where D¼ crystalline size, K ¼ shape factor (0.9), and λ ¼ wavelength of Cu Kα radiation. The crystal imperfection and distortion have been developed the strain form the process in the material; calculated by using the formula expression:
ε=
βhkl (3)
4 tan θ
3. Results
By rearranging the above equation, we get the following expression:
3.1. Peak broadening analysis of MDF using X-ray diffraction
βhkl cos θ =
XRD patterns of AR and MDF specimens have been shown in Fig. 2. All these patterns analysis has disclosed the existence of
The above equations are defined as Williamson–Hall (W–H) equations. Then after, the lattice strain and crystallite size have
Kλ + 4ε sin θ D
Fig. 2. XRD patterns of AR, 3 pass, 6 pass and 9 pass specimens showing peak broadening upon MDF.
(4)
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
Table 1 1
Results of XRD analysis such as crystallite size (D) and lattice strain ( 〈ε2〉2 ) as estimated through classical and modified Williamson–Hall method.
The dislocation densities (ρd) have been estimated by the lattice strain, according to the following equation [34]:
2 3 ε2 Sample
AR εMDF ∼ 0.25 εMDF ∼ 0.50 εMDF ∼ 0.75
Classical Williamson–Hall
Modified Williamson–Hall
D (nm)
D (nm)
240 175 150 90
〈ε
1 〉2
2
0.00001 0.01200 0.01220 0.01030
〈ε
225 165 110 81
1 〉2
2
0.000015 0.01300 0.01330 0.01050
285
ρd =
1 2
Db
(5)
where, Burger’s vector b is 0.26 nm, for Cu–17 wt% Pb–4.5 wt% Zn– 6 wt% Sn. The mean value of dislocation density (ρd) has been initiate to rise from 9.14 1014/m2 ( εMDF ∼ 0.25) up to 1.52 1015/ m2 ( εMDF ∼ 0.75) using classical Williamson–Hall for D [35–37]. 3.2. Microstructural characterization after MDF under HRTEM
been estimated by using classical and modified Williamson–Hall technique from the XRD peak broadening [33]. The crystallite size of the AR specimen has been estimated to be ∼240 nm, which was reduced down to 90 classical Williamson–Hall and 81 nm for modified Williamson–Hall after 9 passes ( εMDF ∼ 0.75) specimen shown in Table 1. 1
On the other hand, average lattice strain ( 〈ε2〉2 ) increases from 1.0 10 5 (AR) to 10.3 10 3 ( εMDF ∼ 0.75) by using classical Williamson–Hall and 10.5 10 3 for modified Williamson–Hall.
The microstructures of AR and MDF specimens have been shown in Fig. 3, characterized under HRTEM. In the microstructures of AR and εMDF ∼0.25 specimens, the Cu– rich phase has been indicated in Fig. 3(a) and (b). Lead was extremely low solid solubility with copper and tin and dispersed throughout the matrix as discrete particles. The primary α-phase provided support to copper tin intermetallic compound which have more hardness and strength [38]. Moreover, the dislocation has been increased with the deformation, the Cu – rich phase, dissociated with others grain
Fig. 3. TEM images (a) AR–bright field (BF), (b) 3 pass–BF, (c) 6 pass–BF, and (d) 6 pass– dark field (DF).
286
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
Fig. 4. TEM images of 9 pass specimen (a) BF, (b) HR, (c) DF, and (d) SAED dotted ring pattern.
further deformation The existence of curved twin boundaries (TBs) has been often observed in Fig. 3(c) and (d). It has been described that during intense deformation of Cu and its alloys, dislocations either build up at TBs or spread across TBs as well as high densities dislocation flux glide [24,39–41]. The dislocated TBs decreased their excess energy throughout forming steps principal to the creation of curved TBs. These faceted tracks are a low energy situation; also thermodynamically have more steady than that of dislocated TBs [39–41]. The high deformed MDF specimen has been characterized which has been observed twin during process shown in Fig. 4(a) bright field (BF) image as well as in Fig. 4(c) dark field (DF) images. In addition, a nearer examination of the HRTEM images has disclosed that the microstructure contains ∼25–30 nm twin lamellae after 9 pass at ( εMDF ∼0.75) has been shown in Fig. 4 (b). Additionally, high twin densities with decreased twin spacing (dtwin) has been observed in MDF specimens. The dense dislocations have been observed due to the accumulation of dislocations during MDF. The density of dislocations results due to the accumulation of dislocations and dislocation restructuring for minimizing the total energy state. The orientation of dislocations walls with different angles has been seen after 3 pass, 6 pass and 9 pass after εMDF ∼0.25 εMDF ∼0.50 and εMDF ∼0.75 strains, respectively in
Fig. 3(b) and (c) and Fig. 4, which may be due the application of strain by changing the direction of loading during MDF. The density of dislocation was also refined into the finer size and the lamellar structures consist of roughly equiaxed sub-grains were formed between the layers. The lamellar shaped structure has been also separates the different constituent phases present in the alloy. In series to describe the microstructure significantly during MDF at room temperature, twin lamellae thickness (λ) and twin spacing/matrix lamellae thickness (dtwin) were approximated from a large quantity of BF and DF images. The SAED pattern has been observed of MDF specimen is shown in Fig. 4(d), as dotted ring pattern. Generally, the dotted ring pattern confirms about polycrystalline that the grain size has been reduced after MDF at different specimen due to increasing the twins in higher strain specimen. The grain size has been found the nano scale ∼90 nm at 9 pass ( εMDF ∼ 0.75), is nanostructured. It has been also observed at εMDF ∼0.50, the numbers of twin has been increased over that εMDF ∼0.25. It has been observed that the average values of λ and dtwin decrease and their distribution become narrow with the increase of MDF strain. The TEM has also been estimated after 3 pass in Fig. 3(b) reveals dislocation. The dislocation density (ρd) for AR, 3 pass, 6 pass and 9 pass has been enlisted in Table 2. The density
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
287
Table 2 Result of hardness, yield strength, ultimate strength, strain and dislocation density. Sample
Hardness (Hv) Yield strength (MPa)
AR 757 3 εMDF ∼ 0.25 1157 5 εMDF ∼ 0.50 1507 8 εMDF ∼ 0.75 1757 2
80 125 140 185
Ultimate strength (MPa)
Strain (%) Dislocation density (ρd)
199 201 261 346
24.2 4.59 3.80 6.39
5.55 9.14 1.08 1.52
1011 1014 1015 1015
of dislocation has been increased with the applied equivalent strain during MDF at ( εMDF ∼0.50). Finally, the crystallized size has been correlated by XRD as well as twin lamellae thickness (λ) under HRTEM with increasing strain at different MDF. Clearly, the microstructure has been observed that the MDF strain has been increased with decreased the twin lamellae thickness as well as crystallite size also has been decreased using evolution of XRD. Finally, the microstructure has been refined during MDF process and increased the materials strength. 3.3. Microstructural characterization under high resolution optical microscope The optical microstructure of AR and MDF specimen at different passes has been shown in Fig. 5. The optical images have been revealed primary α phase, Cu–Sn intermetallic compound and Pb particles as marked in Fig. 5(a). The lead particles were insoluble in the copper and tin, and solidify as a discrete particle has been shown in Fig. 5. The primary α phase has been soft and ductile was a solid solution of tin in copper and provided support to copper tin intermetallic compound which was hard and carries load [38]. The
Fig. 6. Hardness homogeneity profile of MDF specimens respectively showing distance from center and homogeneity at 9 pass specimens.
microstructure has been gained smaller alloy constituents elements with increasing MDF strain. It has been gained more homogeneous microstructure after MDF as shown in Fig. 5(a)–(d). The optical microstructure has been achieved good crystallite refinement during MDF at different specimens; also it has been confirmed with XRD. 3.4. Micro-hardness measurements of MDF Hardness values has been estimated by averaging the hardness of four surrounding points positioned at the edge of a 150 μm 150 μm size square distance from centre [27]. The initial
Fig. 5. High resolution optical microscopic images of (a) AR, (b) 3 pass, (c) 6 pass and (d) 9 pass specimen.
288
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
Fig. 7. Stress vs strain curve during compression test at room temperature of AR and MDF specimens respectively.
Vickers micro-hardness of the alloy 75 Hv was improved to 175 Hv after 9 passes. The microstructural homogeneity has been evaluated by Vickers micro-hardness measurements results are shown in Fig. 6 and enlisted in Table 2. A few trends have been observed as follows: (a) the average hardness of all specimen increases regularly with the increase of MDF strain, and (b) irregular variations of micro-hardness have been monitored over a span of ∼2.5– 5 mm. In addition, the amount of variations increases with increase of MDF strain and has been observed to achieve to a maximum value of 17572 Hv at εMDF ∼0.75, compare to that of 757 3 Hv of AR. On the other hand, the variations decreases with the increase of MDF strain. The decrease in the error range of the micro-hardness reading indicates the evolution of homogeneity at εMDF ∼ 0.75, as observed in Fig. 6. 3.5. Results of compression test after MDF The stress–strain curves of AR and MDF specimens have been shown in Fig. 7. The yield strength (sy) of AR specimen (80 MPa) is lower than that of MDF specimens. Moreover, sy has been increased considerably with increasing of MDF strain and reaches up to sy ¼ 185 MPa in εMDF ∼ 0.75 after 9 pass, which is higher to that of 125 MPa (εMDF ∼ 0.25) after 3 pass and 140 MPa (εMDF ∼ 0.50) after 6 pass, Besides good elongation (εp∼24.2%) in AR, the compressive plastic strain has been decreased with increasing of MDF strain, as evidenced in Fig. 7 and enlisted in Table 2. As an example, εp increased from 3.8% (εMDF ∼ 0.50) to 6.39% (εMDF ∼ 0.75). Therefore, the flow stress have been increased with increasing of MDF strain accompanied by without decrease of compressive plasticity at 9 pass [24], but other specimen has been followed by a reduction of compressive plasticity [27]. The mechanical results have been observed in Fig. 8. The hardness has been increased with increasing MDF as well as crystallite size decreased shown in Fig. 8 (a). The yield strength and dislocation densities have been increased with increasing the MDF strain shown in Fig. 8(b). The deformation behavior has been observed that ultimate compression strength has been increased and ductility or % of strain has been decreased with increasing the MDF strain shown in Fig. 8(c). The fractography has been observed after tensile test of different MDF specimen as well as AR in Fig. 9. The strengthening of the alloy during MDF has been validated using Hall-Petch relationship as shown in Fig. 10.
Fig. 8. Results of (a) hardness and crystalline size (W–H method), (b) yield strength and dislocation density and (c) ultimate strength and strain vs. number of pass of AR, 3 pass, 6 pass and 9 pass specimens respectively.
4. Discussion 4.1. Microstructural refinement during MDF process The MDF process has been refined as eradication of defects owing to dynamic recovery restricts during MDF at room temperature [29,33]. Although, MDF deformation has been minimized dynamic restoration decreases the opportunity of de-twinning effects in the matrix at room temperature, and thus assists in holding the defect structure in alloy [29,36]. It is understood from XRD estimation that the crystallite size reduces from ∼240 nm (AR) to ∼90 nm ( εMDF ∼ 0.75) after MDF process. The microstructural processing depends on chemical composition homogeneity, dislocation density (ρd) and a gradual recovery method during plastic deformation [34]. The cross-slip or climb of dislocations of low-stacking fault energy FCC metals/alloys, [211]
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
289
Fig. 9. SEM (SE) images of alloy tensile fractured surfaces shown in (a) AR (interfacial region marked by arrow sign), (b) 3 pass, (c) 6 pass, and (d) 9 pass respectively.
Fig. 10. Hall–Petch strengthening mechanism of the alloy during MDF.
dislocations were separate into partials [121] and [110] composing a broad stacking fault ribbons [42]. Consequently, deformation twinning turn into a favorable method of deformation owing to its low stacking fault energy (∼14 mJ/m2) and expected to form in high leaded tin bronze upon MDF. The dissociation of the microstructure has been continues by the creation of matrix regions and twinned during the early stage of MDF at room temperature, has been studied in the present work. The thickness has been reduced of λ and dtwin a thin twin/matrix distribution to advance modification in the microstructure engages creation of newer twins in the matrix during MDF. XRD investigation has been shown that dislocation density (ρd) increases from 9.14 1014/m2 ( εMDF ∼ 0.25) to a maximum value of 1.52 1015/m2 at ( εMDF ∼0.75). It is deduced that a bulk dislocation flux rush throughout the grains during MDF to increase the accumulated stalkless dislocation density as well as
local dislocation density. The MDF specimens were activated through the movement of dislocation and accumulation of dislocations during process. These activities lead to the formation of dislocation walls which was decreased further with the application of strain during MDF. The development of such dislocation configurations results in sectioning of coarse grains into subgrains which was separated by the dislocation. The dislocation wall spacing is directly related to the dislocation cell dimensions (L) formed in the coarse grains by the relation L¼ 10 Gb/τ, where G is shear modulus, b is burger vector and τ is shear stress [43]. The dislocation density has been related to the shear stress directly that leads to the smaller dislocation wall density with the increase in shear strain during MDF at different passes. The optical microstructure has been observed the application of strains during MDF, the coarser lead particles, α-phase and copper tin compound dispenses into finer size shown in Fig. 5. The number of composite alloy constituent elements increases with decreased size within the same area of observation due to the uniformity of strain applied during MDF. 4.2. Evolution of homogeneity upon 9 passes ( εMDF ∼ 0.75) The micro-hardness measurements have been disclosed two major tendencies in differently MDF specimens. The first significant examination is that, the mean Vickers micro-hardness has been increased with the increasing of MDF strain. HRTEM investigations disclosed that twin spacing has been decreased with increasing in MDF strain and also increased the portion of twin boundaries (TB) area, which perform as an obstruction to dislocation movement. Second, the variation of Vickers micro-hardness measurements along breadth path has been increased with increasing of MDF strain up to ( εMDF ∼0.50) as shown in Fig. 5. A uniform hardness allocation along the specimen has been monitored in case of AR (75 73 Hv) and εMDF ∼0.75 (175 72 Hv)
290
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
specimens. The microstructure of the AR composed of equiaxed crystals of ∼240 nm size, and is accountable for the uniform hardness division with little error value. Moreover, the microstructure is divided into matrix and twin upon MDF at room temperature. A good variation in the values of ρd, λ, and dtwin has been found at the initial stage of MDF ( εMDF ∼0.50). As λ odtwin, Thus, the hardness evidence in twin rich sections will be soaring than that of matrix region. Accordingly, the variation in the microhardness contour has been observed for MDF specimens. Conversely, ρd, λ, and dtwin have been decreased and λ values move towards to dtwin resulting a slight allocation of the two, twin and matrix lamellae at larger MDF. Therefore, a uniform nanostructure has been obtained at ( εMDF ∼0.75) resulting consistent hardness allocation. The TEM observation reveals the decrease in twin spacing with the increasing in MDF strain, which proceed as a wall for dislocation movement. Therefore the value of hardness of MDF sample has been followed by hall–Petch Eq. (7) with the decrease in twin spacing (d), where H0 and kH are constants [44]. The hardness has been increased with decreasing the crystallite size of different MDF specimen at room temperature shown in Fig. 8(a). 4.3. Evolution of mechanical behavior correlation with flow stress on post-MDF refinement The increase in strength of the alloy with the decrease in crystallite size (D) could be explained by well known hall–Petch relations,
σy = σ0 + kHD
−1 2
Hv = H0 + kHD
−1 2
(6)
(7)
Which state that the yield strength sy and room temperature hardness Hv both are inversely related to the average crystallite size (D) of the material [32]. The yield strength of MDF nanostructured has been increased after 9 passes (εMDF ∼ 0.75), which has been shown in Fig. 8(a) and (b), with decreasing crystallite size. The crystallite size of the MDF alloy has been estimated by W–H and modified W–H and correlated with twin matrix lamellae thickness as well. The grain refinements have been contributed to the increase in ultimate strength of the alloy. In addition, the reason for the increased strength was due to the decreased twin spacing during MDF. It is known from the literature [24] that decreasing twin spacing during deformation results increased in both strength and ductility after 9 passes. The ultimate strength, yield strength of the alloy was gradually improved during MDF and ductility was almost maintained. Which identify that MDF process might be useful to enhance the mechanical properties of the alloy. In order to more clear the view of increased in strength with the decrease in grain size of the alloy, the fractography of the samples AR and MDF specimens has been shown in Fig. 9. The sample AR reveals deep coarser dimples that become finer in size has been shown in Fig. 9(a) to (d), which results due to refinement. It is well known from the literature [13,45,46] that during SPD due to the grain refinement the tensile fractured sample surfaces results finer sized dimples compared to the dimples reported without SPD and hence the strength of the alloy increases with the crystallite size decrease and followed by hall-petch relationship [32]. The yield strength has been increased with increasing the dislocation density (ρd) of different MDF specimen at room temperature shown in Fig. 8(b). Whereas, the ultimate compression strength has been increased with decreasing the ductility upto ( εMDF ∼ 0.50), but after 9 passes the ductility has been increased with increasing MDF strain shown in Fig. 8(c). The local dislocation density ( ρn ) of bronze has been correlated with estimated by using XRD of AR specimen. The
equation is following:
ρn =
γ2 1 * G2b12 (nα′b − b1)2
(8)
where γ ¼ 35 10 J/m2 for bronze [46], α′ ¼0.5, b = ( 2 /2)a , b1 = ( 6 /6)a, n ¼ 2, G ¼ 44 GPa for bronze, a ¼ 3.6 10-10 m, Local dislocation density ( ρn )∼3.4 1013 /m2 [22]. The yield strength (sy) was 80 MPa of AR, which is lower than that of sy ¼ 185 MPa after 9 pass as evidenced in Fig. 7. The hardness has been increased with increasing MDF as well as grain size decreased shown in Fig. 8(a). The yield strength and dislocation densities have been increased with increasing the MDF strain shown in Fig. 8(b). Firstly, the dislocation density has been increased with increasing the yield strength, because higher strain MDF specimens have been decreased the plastic strain after MDF. Secondly, the dislocation density has been increased with increasing the hardness, because number of dislocation increased after MDF. 4.4. Comparative study of fractography The AR specimen has been shown Fig. 9(a) bigger dimples and brittle failure at some location but, after ( εMDF )∼0.25–0.75 strain the coarse dimples refined into much finer scale with increased in number which corresponds to refined during MDF. The interfacial region between lead particle and matrix (primary α-phase) [marked by arrow sign] has been considered as sites for nucleation of fracture for all the samples. The lead particles have been settled as a discrete particle due to poor solid solubility with copper/tin causes weak bonding between lead and matrix, hence the interface region between lead particle and matrix becomes possible section for the micro-crack to initiate and grow along the way of dislocation accumulated during MDF. The contact of such dislocations with lead particles makes easy initiation of cracks and the cracks when joined caused fracture to occur in the alloy [47]. The sample without MDF also consists of many macro-voids as shown in Fig. 8(a) that becomes finer and eliminated due to the applied strain during MDF has been shown in Fig. 9(b), (c) and (d). 4.5. Correlation between hardness and yield strength The improvement in strength sy and hardness Hv with decrease of crystalline size D has been correlated with hall petch strengthening mechanism. Fig. 10 shows the mechanism of strengthening of the alloy, which contains calculated strength as a function of inverse of square root of the calculated crystalline size of the alloy. The strength, hardness and crystalline sizes values after 3 pass, 6 pass and 9 pass was very closely matched with linearity and hence validates hall petch relationship between mechanical properties and crystalline size of the alloy.
5. Conclusions The microstructure and mechanical properties of high leaded tin bronze processed by MDF were studied. The study reported variation in mechanical properties of the alloy along with its mechanism through microstructural studies via HRTEM, and OM and texture studies via XRD. A number of general examinations have been emerged that established MDF route a better method to achieve improved mechanical strength which was required to vary the functional applications of the alloy. With the current efforts, we have been contributed an approach to the scientific advancement for the particular alloy under study to be a flexible alloy. Since the alloy found applications in automotive, agricultural, rail
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
road, mining, off highway equipments, which are any ways, related to our society various needs; hence it forms the link between science and society. The current effort and outlook have been built up a substantial base of information and lead to a vast selection of functional applications of alloy, particularly not able to use under high loading or impacts applications problems and against hardened shafts have been supposed. The present findings in some order have been to increase the knowledge related to the alloy, methods of changing its functional properties and to simplify socially essential problems to the commercial enterprises. The major findings are as following:
During MDF the crystallite size of high leaded tin bronze alloy
has been decreased down to nano-meter ∼90 nm after 9 pass from ∼240 nm of AR as calculated by XRD technique. The MDF process involves microstructural refinement by the creation of 25–30 nm size sub-grains within the twin lamellae, creation of twins, and separation of the microstructure into matrix and twinned regions by the flux of density dislocation. The average micro-hardness and compressive strength of the alloy have been enhanced to about 2–2.5 times upon MDF, however the limited micro-hardness variations have been occurred over a length of 2–3 mm, which was saturated at the 0.75 strain i.e. 9 pass. The creation of fresher twins in the matrix increased the dislocation. The uniform distributions of twins at 0.75 strains of 9 pass assist in the development of improved hardness homogeneity. During MDF the decrease of crystallite size and the increase of dislocation densities have been elevated the flow stress pursue by a loss of compressive strength. The fractography of the alloy has been shown ductile and brittle mode of tensile failure. The interfacial region between lead and matrix was considered as a site for the nucleation of cracks for the samples AR and MDF. The deep coarser dimples were refined of AR and smaller size dimples structure during MDF, accordingly to crystallite size of alloys which has been refined during the process.
References [1] R.Z. Valiev, Nat. Mater. 3 (2004) 511. [2] T.C. Lowe, R.Z. Valiev., J. Min. Met. Mater. Soc. 56 (2004) 64. [3] A. Azushima, R. Kopp, A. Korhonen, D.Y. Yang, F. Micari, G.D. Lahoti, P. Groche, J. Yanagimoto, N. Tsuji, A. Rosochowski, A. Yanagida., CIRP Ann. Man. Tech. 57 (2008) 716. [4] J. Xing, H. Soda, X. Yang, H. Miura, T. Sakai., J. Jpn. Inst. Light. Met. 54 (2004) 527.
291
[5] O.R. Valiakhmetov, R.M. Galeyev, G.A. Salishchev., Fiz. Met. Metall. 72 (1990) 204. [6] R.M. Galeyev, O.R. Valiakhmetov, G.A. Salishchev., Russell Met. 4 (1990) 97. [7] S.V. Zherebstov, G.A. Salishchev, R.M. Galeyev, O.R. Valiakhmetov, S. Y. Mironov, S.L. Semiatin., Scr. Mater. 51 (2004) 1147. [8] A.K. Parimi, P.S. Robi, S.K. Dwivedy, Mater. Des. 32 (2011) 1948. [9] A. Kundu, R. Kapoor, R. Tewari, J.K. Chakravartty., Scr. Mater. 58 (2008) 235. [10] O.F. Higuera-Cobos, J.M. Cabera., Mater. Sci. Eng. A. 571 (2013) 103. [11] F. Salimyanfard, M.R. Toroghinejad, F. Ashrafizadeh, M. Hoseini, J.A. Szpunar., Mater. Des. 44 (2013) 374. [12] M. Shaarbaf, M.R. Toroghinejad., Mater. Sci. Eng. A. 473 (2008) 28. [13] S. Pasebani, M.R. Toroghinejad., Mater. Sci. Eng. A. 527 (2010) 491. [14] A. Sarkar, A. Bhowmik, S. Suwas., Appl. Phy. A 94 (2009) 943. [15] J. Gubicza, N.Q. Chinh, T. Csanadi, T.G. Langdon, T. Ungar., Mater. Sci. Eng. A. 462 (2007) 86. [16] T. Ungaâr, S. Ott, P.G. Sanders, A. Borbeâly, J.R. Weertman., Acta Mater. 46 (1998) 3693. [17] T. Makhlouf, A. Rebhi, J.P. Couzinié, Y. Champion, N. Njah., Int. J. Miner. Metall. Mater. 19 (2012) 1016. [18] J.R. Davis, Copper and Copper Alloys, ASM International, Ohio, 2001. [19] 〈http://www.copper.org/applications/industrial/bronze_bearing.html〉. [20] W. Bolton, Newnes Engineering Materials Pocket Book, Heinemann Newnes, Oxford, 1989. [21] N. Rangaraju, T. Raghuram, B.V. Krishna, K.P. Rao, P. Venugopal., Mater. Sci. Eng. A. 398 (2005) 246. [22] C.X. Huang, K. Wang, S.D. Wu, Z.F. Zhang, G.Y. Li, S.X. Li., Acta Mater. 54 (2006) 655. [23] L. Lu, R. Schwaiger, Z.W. Shan, M. Dao, K. Lu, S. Suresh., Acta Mater. 53 (2005) 2169. [24] M. Dao, L. Lu, Y.F. Shen, S. Suresh., Acta Mater. 54 (2006) 5421. [25] Y. Li, Y.H. Zhao, W. Liu, C. Xu, Z. Horita, X.Z. Liao, Y.T. Zhu, T.G. Langdon, E. J. Lavernia., Mater. Sci. Eng. A. 527 (2010) 3942. [26] V.S. Sarma, K. Sivaprasad, D. Sturm, M. Heilmaier., Mater. Sci. Eng. A. 489 (2008) 253. [27] N.K. Kumar, B. Roy, J. Das., J. Alloy. Comp. 618 (2015) 139. [28] G.E. Dieter, Mechanical Metallurgy, McGraw-Hill, New York, 1988. [29] Y.S. Li, N.R. Tao, K. Lu., Acta Mater. 56 (2008) 230. [30] A.M. Hodge, Y.M. Wang, T.W. Barbee., Scr. Mater. 59 (2008) 163. [31] S.V. Dobatkin, J. Gubicza, D.V. Shangina, N.R. Bochvar, N.Y. Tabachkova., Mater. Lett. 153 (2015) 5. [32] J. Xing, H. Soda, X. Yang, H. Miura, T. Sakai., J. Jpn. Inst. Light. Met. 46 (2005) 1646. [33] V.D. Mote, Y. Purushotham, B.N. Dole., J. Theo., Appl. Phys. 6 (2012) 1. [34] Y.H. Zhao, Z. Horita, T.G. Langdon, Y.T. Zhu., Mater. Sci. Eng. A. 474 (2008) 342. [35] Y. Zhong, T. Sakaguchi, F. Yin., Mater. Sci. Eng. A. 492 (2008) 419. [36] Y. Zhang, N.R. Tao, K. Lu., Acta Mater. 56 (2008) 2429. [37] L. Balogh, R.B. Figueiredo, T. Ungar, T.G. Langdon., Mater. Sci. Eng. A. 528 (2010) 533. [38] B.K. Prasad., Wear. 257 (2004) 110. [39] J. Wang, Q. Yu, Y. Jiang, I.J. Beyerlein., JOM 66 (2014) 95. [40] J. Tu, X. Zhang, J. Wang, Q. Sun, Q. Liu, C.N. Tome., Appl. Phys. Lett. 103 (2013) 051903. [41] J. Wang, I.J. Beyerlein, C.N. Tome., Int. J. Plast. 56 (2014) 156. [42] J. Das., Mater. Sci. Eng. A. 530 (2011) 675. [43] D.K. Wilsdorf, J.H.V. Merwe., Mater. Sci. Eng. 55 (1982) 79. [44] A.S. Taha, F.H. Hammad., Phys. Status Solidi (a). 119 (1990) 455. [45] M.R. Rezaei, M.R. Toroghinejad, F. Ashrafizadeh., J. Mater. Proc. Technol. 211 (2011) 1184. [46] Y.H. Zhao, Y.T. Zhu, X.Z. Liao, Z. Horita, T.G. Langdon., Appl. Phys. Lett. 89 (2006) 1. [47] B.K. Prasad, A.H. Yegneshwaran, A.K. Patwardhan., J. Mater. Sci. 32 (1997) 1169.
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
High leaded tin bronze processing during multi-directional forging: Effect on microstructure and mechanical properties Rahul Gupta a,n, Sanjay Srivastava a, Nand Kishor Kumar b, Sanjay K. Panthi c a
Department of Material Science and Metallurgical Engineering, MANIT, Bhopal 462003, India Department of Metallurgical and Materials Engineering, Indian Institute of Technology Kharagpur, 721302, India c Advanced Materials and Processes Research Institute, Bhopal 462024, India b
art ic l e i nf o
a b s t r a c t
Article history: Received 28 October 2015 Received in revised form 19 December 2015 Accepted 21 December 2015 Available online 22 December 2015
The high leaded tin bronze alloy has been studied during multidirectional forging (MDF) process with differing crystallite size (D) in the scale of 240–90 nm. The effect of MDF strain ( εMDF ∼ 0.25, 0.50, 0.75) after 3, 6 and 9 passes respectively on the development of homogeneity and refinement in terms of dislocation density, twin spacing, twin lamellae thickness and their division have been studied using high-resolution transmission electron microscopy (HRTEM) and micro-hardness measurements using Vickers micro-hardness. X-ray diffraction peak broadening investigation has revealed that the crystallite size decreases down to 90 nm after 9 pass ( εMDF ∼ 0.75). The tensile fracture surfaces have been studied using scanning electron microscope (SEM) of received (AR) and MDF specimens. The results of crystallite size and dislocation density on the strength of refined microstructure and nanostructure have been correlated with mechanical flow stress. & 2015 Elsevier B.V. All rights reserved.
Keywords: Multi-directional forging High leaded tin bronze Microstructure Flow stress X-ray diffraction
1. Introduction In last decade, the production of bulk nanostructured metals and alloys using severe plastic deformation (SPD) processes has been progressed rapidly with the aim of developing novel mechanical and functional properties in the materials [1]. SPD techniques like as multidirectional forging (MDF), equal channel angular pressing (ECAP), high pressure torsion (HPT), accumulative roll bonding (ARB) have been used to produce ultrafine grained material by applying very large plastic strain in various materials [2–4]. During 1990 MDF was introduced for the first time to fabricate ultrafine grained structure in bulk billets [5,6]. MDF is a repetition of free forging operation multiple times including setting and pulling operations with the changes of the axes of applied load [7]. Many researchers have been reported to study microstructure and mechanical behavior of copper and copper alloy after SPD [8– 13]. Parimi et al. [8] reported SPD of copper and Al–Cu alloy by using multiple channel-die compression process. The nano size grain structured copper was achieved along with increased flow stress and hardness during compression. Kundu et al. [9] applied multiple compressions in a channel die at different temperature range on copper and result enhanced flow stress and strain rate n
Corresponding author. E-mail address: [email protected] (R. Gupta).
http://dx.doi.org/10.1016/j.msea.2015.12.068 0921-5093/& 2015 Elsevier B.V. All rights reserved.
sensitivity. Sarkar et al. [14] deformed interstitial free steel through ECAP to assess the microstructure by different XRD peak profile analysis using Williamson–Hall plot technique and the variance method as a function of strain (ε). The dislocation densities estimated from XRD were also correlated with experimental yield strength. Gubicza et al. [15] studied the microstructure of face centered cubic (FCC) metal and alloys processed through SPD methods and also correlate the yield strength with dislocation densities calculated using Taylor's equation. Thereby they reported the main strengthening mechanism for both metals and alloys were the interaction between dislocations. Ungar et al. [16] determined the particle size and dislocation structure in nanostructured copper by high resolution XRD profile analysis method using modified Williamson–Hall plot and modified Warren Aver Bach’s plot and also reported that the twin boundaries found good agreement with TEM micrograph. Makhlouf et al. [17] studied the microstructure of Al alloy after ECAP using XRD and TEM. The high dislocation density along with improved mechanical behavior was obtained in the alloy. The high leaded tin bronze (copper based alloy) was used in the present study finds application in automotive, agricultural, rail road, mining, off highway equipments [18]. The high leaded tin bronze has lower strength which makes this alloy to only use under moderate/light loading conditions and against unhardened shafts. The trouble with the target alloy was that it has not been specified to use under high loading or impacts applications. Since these alloy operated under moderate loads and high speeds only.
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
But the situation for high loading and against hardened shafts was required the strength of the alloy to be improved. Therefore, the introduction of new mechanical properties in the alloy through MDF was aimed, along with its microstructural studies, so that the alloy functional conditions has been varied, such as used under high speeds, under high-load, and impact situations [19,20]. The production of bulk nanostructured metals and alloys (such as high leaded tin bronzes) has been carried out by using severe plastic deformation (SPD) with the aim of developing novel mechanical and functional properties. Generally, Cu-based alloys have been used in ancient times such as Romans and other civilizations to produce artistic and functional artifacts. Now days, we have been obtained more novel mechanical and functional properties, which as reported by many researcher from long time ago. The lowest grain size was detailed to be 400–600 nm for SPD processed pure Al [21], grain size between 100–450 nm for Cu alloy [22–24] also 70–110 nm for Cu–30 wt% Zn alloy [25]. In addition, post SPD recrystallization fabricates superior microstructure with ∼4.5 μm grain size for Cu–5 wt% Al and for Cu–5 wt% Zn ∼7.5 μm respectively [26]. Twinning is one of the key deformation methods for metals and alloys with small stacking fault energy (SFE), for silver (25 mJ/m2), stainless steel (8–45 mJ/m2), and α-brass (14 mJ/ m2) respectivley [27–29]. While, inclusion of Zn in Cu, 30% Zn alloy decreases the stacking fault energy (SFE) of pure Cu from 78 mJ/m2 to 14 mJ/m2 [27,29,30]. The dislocation density of Cu–0.7% Cr was 38 74 1014 and hardness was 804 738 Hv through severely
283
deformed by HPT at room temperature (RT) and a rate of 1 rpm under a pressure of 4 GPa [31]. The hardness was 173.2 74.3 Hv and yield strength was 640 720 MPa at CR095 for Cu–30 wt% Zn [27]. The present investigation is aimed to study the microstructure and mechanical properties of high leaded tin bronze processed by MDF at room temperature. The characterization of MDFed microstructure has been carried out in terms of twin spacing (dtwin), crystallite size, twin lamellae thickness (λ) and their division at different level of MDF, to investigate the fundamental mechanism of microstructural modification. The result of strengthening due to twin spacing, dislocation density, and crystallite size on the mechanical flow stress of differently MDF specimens have been compared in this research work.
2. Experimental procedures The material used in this study was high leaded tin bronze alloy (Cu–17 wt% Pb–6 wt% Sn–4.5 wt% Zn). Initially the alloy was cut into the measurement of 60 mm 55 mm 55 mm from square section and heat treated at 573 K (300° Celsius) for 1 hour using muffle furnace followed by water quenching. The process has been applied at room temperature and inside the MDF die [CAD model is shown in Fig. 1(a)] between top punch and bottom punch and pressed vertically downward number of times by changing the
Fig. 1. Schematic showing (a) Computer aided design model of MDF die (Exploded View), Top punch, Bottom punch, and Left and Right die halves with Allen bolts. (b) Schematic of MDF processing of high leaded tin bronze for different passes.
284
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
direction of loading from pass to pass along three mutually perpendicular axes i.e. in order of z to x, x to y, y to z and so on… as shown in Fig. 1(b) [32]. The process has been operated at constant pressing speed of 10 mm min 1 with MDF capacity press of 200 t. The total equivalent strain of 0.75 was applied during 3 MDF cycles after 9 passes. MoS2 was used as a lubricant between the samples and die contacts during compression. X-ray diffraction (XRD, Philips PANalytical PW 3373, Netherlands) was performed for structural investigation with Cu Kα radiation. Thin slices have been prepared by Ar-ion milling (PIPS-691, Gatan). Microstructural characterization was performed with HRTEM (JEM-2100, JEOL, Japan) operated at 200 kV. Vickers micro-hardness were taken out using a Bühler MicroMet 5103 hardness tester with 15 s dwell time at 50 gf load of AR (As-received) and MDF specimens. Hardness has been evaluated at each location on specimens were estimated by averaging the hardness of four surrounding points positioned at the edge of a 500 μm 500 μm size square. Parallelepiped specimens (2 2 4 mm3) were cut from the AR and MDF bars and were tested using Tinius Olsen H50KS (Germany) universal testing machine. The different surfaces of the specimens were polished carefully to retain uniaxiality throughout the compressive loading at a strain rate of 8 10 4/s under room temperature. The tensile fracture behavior of the MDF alloy has been examined using scanning electron microscope (JEOL JSM 6390 A) operated at an accelerating voltage of 25 kV.
single fcc phase directing that no phase alteration had occurred during MDF. On the other hand, the width of all the peaks became wider in MDF specimens than that of AR. Usually crystallite size, lattice strain and the instrumental broadening are the chief aspects, which add on peak broadening [16]. The instrumental line broadening has been calculated by a standard Si-disc (PW3132/62) of thickness and diameter, 2 mm and 32 mm respectively. The specimens’ peak broadening (β) has been précised by using the following Eq. (1) [33].
β=
2 (βobs − βSi2)
(1)
where, βobs and βSi are the integral breath of the equivalent (h k l ) reflections by the specimen and standard silicon disc respectively. The average nanocrystalline size was determined by using DebyeScherrer's formula:
D=
Kλ βhkl cos θ
(2)
where D¼ crystalline size, K ¼ shape factor (0.9), and λ ¼ wavelength of Cu Kα radiation. The crystal imperfection and distortion have been developed the strain form the process in the material; calculated by using the formula expression:
ε=
βhkl (3)
4 tan θ
3. Results
By rearranging the above equation, we get the following expression:
3.1. Peak broadening analysis of MDF using X-ray diffraction
βhkl cos θ =
XRD patterns of AR and MDF specimens have been shown in Fig. 2. All these patterns analysis has disclosed the existence of
The above equations are defined as Williamson–Hall (W–H) equations. Then after, the lattice strain and crystallite size have
Kλ + 4ε sin θ D
Fig. 2. XRD patterns of AR, 3 pass, 6 pass and 9 pass specimens showing peak broadening upon MDF.
(4)
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
Table 1 1
Results of XRD analysis such as crystallite size (D) and lattice strain ( 〈ε2〉2 ) as estimated through classical and modified Williamson–Hall method.
The dislocation densities (ρd) have been estimated by the lattice strain, according to the following equation [34]:
2 3 ε2 Sample
AR εMDF ∼ 0.25 εMDF ∼ 0.50 εMDF ∼ 0.75
Classical Williamson–Hall
Modified Williamson–Hall
D (nm)
D (nm)
240 175 150 90
〈ε
1 〉2
2
0.00001 0.01200 0.01220 0.01030
〈ε
225 165 110 81
1 〉2
2
0.000015 0.01300 0.01330 0.01050
285
ρd =
1 2
Db
(5)
where, Burger’s vector b is 0.26 nm, for Cu–17 wt% Pb–4.5 wt% Zn– 6 wt% Sn. The mean value of dislocation density (ρd) has been initiate to rise from 9.14 1014/m2 ( εMDF ∼ 0.25) up to 1.52 1015/ m2 ( εMDF ∼ 0.75) using classical Williamson–Hall for D [35–37]. 3.2. Microstructural characterization after MDF under HRTEM
been estimated by using classical and modified Williamson–Hall technique from the XRD peak broadening [33]. The crystallite size of the AR specimen has been estimated to be ∼240 nm, which was reduced down to 90 classical Williamson–Hall and 81 nm for modified Williamson–Hall after 9 passes ( εMDF ∼ 0.75) specimen shown in Table 1. 1
On the other hand, average lattice strain ( 〈ε2〉2 ) increases from 1.0 10 5 (AR) to 10.3 10 3 ( εMDF ∼ 0.75) by using classical Williamson–Hall and 10.5 10 3 for modified Williamson–Hall.
The microstructures of AR and MDF specimens have been shown in Fig. 3, characterized under HRTEM. In the microstructures of AR and εMDF ∼0.25 specimens, the Cu– rich phase has been indicated in Fig. 3(a) and (b). Lead was extremely low solid solubility with copper and tin and dispersed throughout the matrix as discrete particles. The primary α-phase provided support to copper tin intermetallic compound which have more hardness and strength [38]. Moreover, the dislocation has been increased with the deformation, the Cu – rich phase, dissociated with others grain
Fig. 3. TEM images (a) AR–bright field (BF), (b) 3 pass–BF, (c) 6 pass–BF, and (d) 6 pass– dark field (DF).
286
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
Fig. 4. TEM images of 9 pass specimen (a) BF, (b) HR, (c) DF, and (d) SAED dotted ring pattern.
further deformation The existence of curved twin boundaries (TBs) has been often observed in Fig. 3(c) and (d). It has been described that during intense deformation of Cu and its alloys, dislocations either build up at TBs or spread across TBs as well as high densities dislocation flux glide [24,39–41]. The dislocated TBs decreased their excess energy throughout forming steps principal to the creation of curved TBs. These faceted tracks are a low energy situation; also thermodynamically have more steady than that of dislocated TBs [39–41]. The high deformed MDF specimen has been characterized which has been observed twin during process shown in Fig. 4(a) bright field (BF) image as well as in Fig. 4(c) dark field (DF) images. In addition, a nearer examination of the HRTEM images has disclosed that the microstructure contains ∼25–30 nm twin lamellae after 9 pass at ( εMDF ∼0.75) has been shown in Fig. 4 (b). Additionally, high twin densities with decreased twin spacing (dtwin) has been observed in MDF specimens. The dense dislocations have been observed due to the accumulation of dislocations during MDF. The density of dislocations results due to the accumulation of dislocations and dislocation restructuring for minimizing the total energy state. The orientation of dislocations walls with different angles has been seen after 3 pass, 6 pass and 9 pass after εMDF ∼0.25 εMDF ∼0.50 and εMDF ∼0.75 strains, respectively in
Fig. 3(b) and (c) and Fig. 4, which may be due the application of strain by changing the direction of loading during MDF. The density of dislocation was also refined into the finer size and the lamellar structures consist of roughly equiaxed sub-grains were formed between the layers. The lamellar shaped structure has been also separates the different constituent phases present in the alloy. In series to describe the microstructure significantly during MDF at room temperature, twin lamellae thickness (λ) and twin spacing/matrix lamellae thickness (dtwin) were approximated from a large quantity of BF and DF images. The SAED pattern has been observed of MDF specimen is shown in Fig. 4(d), as dotted ring pattern. Generally, the dotted ring pattern confirms about polycrystalline that the grain size has been reduced after MDF at different specimen due to increasing the twins in higher strain specimen. The grain size has been found the nano scale ∼90 nm at 9 pass ( εMDF ∼ 0.75), is nanostructured. It has been also observed at εMDF ∼0.50, the numbers of twin has been increased over that εMDF ∼0.25. It has been observed that the average values of λ and dtwin decrease and their distribution become narrow with the increase of MDF strain. The TEM has also been estimated after 3 pass in Fig. 3(b) reveals dislocation. The dislocation density (ρd) for AR, 3 pass, 6 pass and 9 pass has been enlisted in Table 2. The density
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
287
Table 2 Result of hardness, yield strength, ultimate strength, strain and dislocation density. Sample
Hardness (Hv) Yield strength (MPa)
AR 757 3 εMDF ∼ 0.25 1157 5 εMDF ∼ 0.50 1507 8 εMDF ∼ 0.75 1757 2
80 125 140 185
Ultimate strength (MPa)
Strain (%) Dislocation density (ρd)
199 201 261 346
24.2 4.59 3.80 6.39
5.55 9.14 1.08 1.52
1011 1014 1015 1015
of dislocation has been increased with the applied equivalent strain during MDF at ( εMDF ∼0.50). Finally, the crystallized size has been correlated by XRD as well as twin lamellae thickness (λ) under HRTEM with increasing strain at different MDF. Clearly, the microstructure has been observed that the MDF strain has been increased with decreased the twin lamellae thickness as well as crystallite size also has been decreased using evolution of XRD. Finally, the microstructure has been refined during MDF process and increased the materials strength. 3.3. Microstructural characterization under high resolution optical microscope The optical microstructure of AR and MDF specimen at different passes has been shown in Fig. 5. The optical images have been revealed primary α phase, Cu–Sn intermetallic compound and Pb particles as marked in Fig. 5(a). The lead particles were insoluble in the copper and tin, and solidify as a discrete particle has been shown in Fig. 5. The primary α phase has been soft and ductile was a solid solution of tin in copper and provided support to copper tin intermetallic compound which was hard and carries load [38]. The
Fig. 6. Hardness homogeneity profile of MDF specimens respectively showing distance from center and homogeneity at 9 pass specimens.
microstructure has been gained smaller alloy constituents elements with increasing MDF strain. It has been gained more homogeneous microstructure after MDF as shown in Fig. 5(a)–(d). The optical microstructure has been achieved good crystallite refinement during MDF at different specimens; also it has been confirmed with XRD. 3.4. Micro-hardness measurements of MDF Hardness values has been estimated by averaging the hardness of four surrounding points positioned at the edge of a 150 μm 150 μm size square distance from centre [27]. The initial
Fig. 5. High resolution optical microscopic images of (a) AR, (b) 3 pass, (c) 6 pass and (d) 9 pass specimen.
288
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
Fig. 7. Stress vs strain curve during compression test at room temperature of AR and MDF specimens respectively.
Vickers micro-hardness of the alloy 75 Hv was improved to 175 Hv after 9 passes. The microstructural homogeneity has been evaluated by Vickers micro-hardness measurements results are shown in Fig. 6 and enlisted in Table 2. A few trends have been observed as follows: (a) the average hardness of all specimen increases regularly with the increase of MDF strain, and (b) irregular variations of micro-hardness have been monitored over a span of ∼2.5– 5 mm. In addition, the amount of variations increases with increase of MDF strain and has been observed to achieve to a maximum value of 17572 Hv at εMDF ∼0.75, compare to that of 757 3 Hv of AR. On the other hand, the variations decreases with the increase of MDF strain. The decrease in the error range of the micro-hardness reading indicates the evolution of homogeneity at εMDF ∼ 0.75, as observed in Fig. 6. 3.5. Results of compression test after MDF The stress–strain curves of AR and MDF specimens have been shown in Fig. 7. The yield strength (sy) of AR specimen (80 MPa) is lower than that of MDF specimens. Moreover, sy has been increased considerably with increasing of MDF strain and reaches up to sy ¼ 185 MPa in εMDF ∼ 0.75 after 9 pass, which is higher to that of 125 MPa (εMDF ∼ 0.25) after 3 pass and 140 MPa (εMDF ∼ 0.50) after 6 pass, Besides good elongation (εp∼24.2%) in AR, the compressive plastic strain has been decreased with increasing of MDF strain, as evidenced in Fig. 7 and enlisted in Table 2. As an example, εp increased from 3.8% (εMDF ∼ 0.50) to 6.39% (εMDF ∼ 0.75). Therefore, the flow stress have been increased with increasing of MDF strain accompanied by without decrease of compressive plasticity at 9 pass [24], but other specimen has been followed by a reduction of compressive plasticity [27]. The mechanical results have been observed in Fig. 8. The hardness has been increased with increasing MDF as well as crystallite size decreased shown in Fig. 8 (a). The yield strength and dislocation densities have been increased with increasing the MDF strain shown in Fig. 8(b). The deformation behavior has been observed that ultimate compression strength has been increased and ductility or % of strain has been decreased with increasing the MDF strain shown in Fig. 8(c). The fractography has been observed after tensile test of different MDF specimen as well as AR in Fig. 9. The strengthening of the alloy during MDF has been validated using Hall-Petch relationship as shown in Fig. 10.
Fig. 8. Results of (a) hardness and crystalline size (W–H method), (b) yield strength and dislocation density and (c) ultimate strength and strain vs. number of pass of AR, 3 pass, 6 pass and 9 pass specimens respectively.
4. Discussion 4.1. Microstructural refinement during MDF process The MDF process has been refined as eradication of defects owing to dynamic recovery restricts during MDF at room temperature [29,33]. Although, MDF deformation has been minimized dynamic restoration decreases the opportunity of de-twinning effects in the matrix at room temperature, and thus assists in holding the defect structure in alloy [29,36]. It is understood from XRD estimation that the crystallite size reduces from ∼240 nm (AR) to ∼90 nm ( εMDF ∼ 0.75) after MDF process. The microstructural processing depends on chemical composition homogeneity, dislocation density (ρd) and a gradual recovery method during plastic deformation [34]. The cross-slip or climb of dislocations of low-stacking fault energy FCC metals/alloys, [211]
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
289
Fig. 9. SEM (SE) images of alloy tensile fractured surfaces shown in (a) AR (interfacial region marked by arrow sign), (b) 3 pass, (c) 6 pass, and (d) 9 pass respectively.
Fig. 10. Hall–Petch strengthening mechanism of the alloy during MDF.
dislocations were separate into partials [121] and [110] composing a broad stacking fault ribbons [42]. Consequently, deformation twinning turn into a favorable method of deformation owing to its low stacking fault energy (∼14 mJ/m2) and expected to form in high leaded tin bronze upon MDF. The dissociation of the microstructure has been continues by the creation of matrix regions and twinned during the early stage of MDF at room temperature, has been studied in the present work. The thickness has been reduced of λ and dtwin a thin twin/matrix distribution to advance modification in the microstructure engages creation of newer twins in the matrix during MDF. XRD investigation has been shown that dislocation density (ρd) increases from 9.14 1014/m2 ( εMDF ∼ 0.25) to a maximum value of 1.52 1015/m2 at ( εMDF ∼0.75). It is deduced that a bulk dislocation flux rush throughout the grains during MDF to increase the accumulated stalkless dislocation density as well as
local dislocation density. The MDF specimens were activated through the movement of dislocation and accumulation of dislocations during process. These activities lead to the formation of dislocation walls which was decreased further with the application of strain during MDF. The development of such dislocation configurations results in sectioning of coarse grains into subgrains which was separated by the dislocation. The dislocation wall spacing is directly related to the dislocation cell dimensions (L) formed in the coarse grains by the relation L¼ 10 Gb/τ, where G is shear modulus, b is burger vector and τ is shear stress [43]. The dislocation density has been related to the shear stress directly that leads to the smaller dislocation wall density with the increase in shear strain during MDF at different passes. The optical microstructure has been observed the application of strains during MDF, the coarser lead particles, α-phase and copper tin compound dispenses into finer size shown in Fig. 5. The number of composite alloy constituent elements increases with decreased size within the same area of observation due to the uniformity of strain applied during MDF. 4.2. Evolution of homogeneity upon 9 passes ( εMDF ∼ 0.75) The micro-hardness measurements have been disclosed two major tendencies in differently MDF specimens. The first significant examination is that, the mean Vickers micro-hardness has been increased with the increasing of MDF strain. HRTEM investigations disclosed that twin spacing has been decreased with increasing in MDF strain and also increased the portion of twin boundaries (TB) area, which perform as an obstruction to dislocation movement. Second, the variation of Vickers micro-hardness measurements along breadth path has been increased with increasing of MDF strain up to ( εMDF ∼0.50) as shown in Fig. 5. A uniform hardness allocation along the specimen has been monitored in case of AR (75 73 Hv) and εMDF ∼0.75 (175 72 Hv)
290
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
specimens. The microstructure of the AR composed of equiaxed crystals of ∼240 nm size, and is accountable for the uniform hardness division with little error value. Moreover, the microstructure is divided into matrix and twin upon MDF at room temperature. A good variation in the values of ρd, λ, and dtwin has been found at the initial stage of MDF ( εMDF ∼0.50). As λ odtwin, Thus, the hardness evidence in twin rich sections will be soaring than that of matrix region. Accordingly, the variation in the microhardness contour has been observed for MDF specimens. Conversely, ρd, λ, and dtwin have been decreased and λ values move towards to dtwin resulting a slight allocation of the two, twin and matrix lamellae at larger MDF. Therefore, a uniform nanostructure has been obtained at ( εMDF ∼0.75) resulting consistent hardness allocation. The TEM observation reveals the decrease in twin spacing with the increasing in MDF strain, which proceed as a wall for dislocation movement. Therefore the value of hardness of MDF sample has been followed by hall–Petch Eq. (7) with the decrease in twin spacing (d), where H0 and kH are constants [44]. The hardness has been increased with decreasing the crystallite size of different MDF specimen at room temperature shown in Fig. 8(a). 4.3. Evolution of mechanical behavior correlation with flow stress on post-MDF refinement The increase in strength of the alloy with the decrease in crystallite size (D) could be explained by well known hall–Petch relations,
σy = σ0 + kHD
−1 2
Hv = H0 + kHD
−1 2
(6)
(7)
Which state that the yield strength sy and room temperature hardness Hv both are inversely related to the average crystallite size (D) of the material [32]. The yield strength of MDF nanostructured has been increased after 9 passes (εMDF ∼ 0.75), which has been shown in Fig. 8(a) and (b), with decreasing crystallite size. The crystallite size of the MDF alloy has been estimated by W–H and modified W–H and correlated with twin matrix lamellae thickness as well. The grain refinements have been contributed to the increase in ultimate strength of the alloy. In addition, the reason for the increased strength was due to the decreased twin spacing during MDF. It is known from the literature [24] that decreasing twin spacing during deformation results increased in both strength and ductility after 9 passes. The ultimate strength, yield strength of the alloy was gradually improved during MDF and ductility was almost maintained. Which identify that MDF process might be useful to enhance the mechanical properties of the alloy. In order to more clear the view of increased in strength with the decrease in grain size of the alloy, the fractography of the samples AR and MDF specimens has been shown in Fig. 9. The sample AR reveals deep coarser dimples that become finer in size has been shown in Fig. 9(a) to (d), which results due to refinement. It is well known from the literature [13,45,46] that during SPD due to the grain refinement the tensile fractured sample surfaces results finer sized dimples compared to the dimples reported without SPD and hence the strength of the alloy increases with the crystallite size decrease and followed by hall-petch relationship [32]. The yield strength has been increased with increasing the dislocation density (ρd) of different MDF specimen at room temperature shown in Fig. 8(b). Whereas, the ultimate compression strength has been increased with decreasing the ductility upto ( εMDF ∼ 0.50), but after 9 passes the ductility has been increased with increasing MDF strain shown in Fig. 8(c). The local dislocation density ( ρn ) of bronze has been correlated with estimated by using XRD of AR specimen. The
equation is following:
ρn =
γ2 1 * G2b12 (nα′b − b1)2
(8)
where γ ¼ 35 10 J/m2 for bronze [46], α′ ¼0.5, b = ( 2 /2)a , b1 = ( 6 /6)a, n ¼ 2, G ¼ 44 GPa for bronze, a ¼ 3.6 10-10 m, Local dislocation density ( ρn )∼3.4 1013 /m2 [22]. The yield strength (sy) was 80 MPa of AR, which is lower than that of sy ¼ 185 MPa after 9 pass as evidenced in Fig. 7. The hardness has been increased with increasing MDF as well as grain size decreased shown in Fig. 8(a). The yield strength and dislocation densities have been increased with increasing the MDF strain shown in Fig. 8(b). Firstly, the dislocation density has been increased with increasing the yield strength, because higher strain MDF specimens have been decreased the plastic strain after MDF. Secondly, the dislocation density has been increased with increasing the hardness, because number of dislocation increased after MDF. 4.4. Comparative study of fractography The AR specimen has been shown Fig. 9(a) bigger dimples and brittle failure at some location but, after ( εMDF )∼0.25–0.75 strain the coarse dimples refined into much finer scale with increased in number which corresponds to refined during MDF. The interfacial region between lead particle and matrix (primary α-phase) [marked by arrow sign] has been considered as sites for nucleation of fracture for all the samples. The lead particles have been settled as a discrete particle due to poor solid solubility with copper/tin causes weak bonding between lead and matrix, hence the interface region between lead particle and matrix becomes possible section for the micro-crack to initiate and grow along the way of dislocation accumulated during MDF. The contact of such dislocations with lead particles makes easy initiation of cracks and the cracks when joined caused fracture to occur in the alloy [47]. The sample without MDF also consists of many macro-voids as shown in Fig. 8(a) that becomes finer and eliminated due to the applied strain during MDF has been shown in Fig. 9(b), (c) and (d). 4.5. Correlation between hardness and yield strength The improvement in strength sy and hardness Hv with decrease of crystalline size D has been correlated with hall petch strengthening mechanism. Fig. 10 shows the mechanism of strengthening of the alloy, which contains calculated strength as a function of inverse of square root of the calculated crystalline size of the alloy. The strength, hardness and crystalline sizes values after 3 pass, 6 pass and 9 pass was very closely matched with linearity and hence validates hall petch relationship between mechanical properties and crystalline size of the alloy.
5. Conclusions The microstructure and mechanical properties of high leaded tin bronze processed by MDF were studied. The study reported variation in mechanical properties of the alloy along with its mechanism through microstructural studies via HRTEM, and OM and texture studies via XRD. A number of general examinations have been emerged that established MDF route a better method to achieve improved mechanical strength which was required to vary the functional applications of the alloy. With the current efforts, we have been contributed an approach to the scientific advancement for the particular alloy under study to be a flexible alloy. Since the alloy found applications in automotive, agricultural, rail
R. Gupta et al. / Materials Science & Engineering A 654 (2016) 282–291
road, mining, off highway equipments, which are any ways, related to our society various needs; hence it forms the link between science and society. The current effort and outlook have been built up a substantial base of information and lead to a vast selection of functional applications of alloy, particularly not able to use under high loading or impacts applications problems and against hardened shafts have been supposed. The present findings in some order have been to increase the knowledge related to the alloy, methods of changing its functional properties and to simplify socially essential problems to the commercial enterprises. The major findings are as following:
During MDF the crystallite size of high leaded tin bronze alloy
has been decreased down to nano-meter ∼90 nm after 9 pass from ∼240 nm of AR as calculated by XRD technique. The MDF process involves microstructural refinement by the creation of 25–30 nm size sub-grains within the twin lamellae, creation of twins, and separation of the microstructure into matrix and twinned regions by the flux of density dislocation. The average micro-hardness and compressive strength of the alloy have been enhanced to about 2–2.5 times upon MDF, however the limited micro-hardness variations have been occurred over a length of 2–3 mm, which was saturated at the 0.75 strain i.e. 9 pass. The creation of fresher twins in the matrix increased the dislocation. The uniform distributions of twins at 0.75 strains of 9 pass assist in the development of improved hardness homogeneity. During MDF the decrease of crystallite size and the increase of dislocation densities have been elevated the flow stress pursue by a loss of compressive strength. The fractography of the alloy has been shown ductile and brittle mode of tensile failure. The interfacial region between lead and matrix was considered as a site for the nucleation of cracks for the samples AR and MDF. The deep coarser dimples were refined of AR and smaller size dimples structure during MDF, accordingly to crystallite size of alloys which has been refined during the process.
References [1] R.Z. Valiev, Nat. Mater. 3 (2004) 511. [2] T.C. Lowe, R.Z. Valiev., J. Min. Met. Mater. Soc. 56 (2004) 64. [3] A. Azushima, R. Kopp, A. Korhonen, D.Y. Yang, F. Micari, G.D. Lahoti, P. Groche, J. Yanagimoto, N. Tsuji, A. Rosochowski, A. Yanagida., CIRP Ann. Man. Tech. 57 (2008) 716. [4] J. Xing, H. Soda, X. Yang, H. Miura, T. Sakai., J. Jpn. Inst. Light. Met. 54 (2004) 527.
291
[5] O.R. Valiakhmetov, R.M. Galeyev, G.A. Salishchev., Fiz. Met. Metall. 72 (1990) 204. [6] R.M. Galeyev, O.R. Valiakhmetov, G.A. Salishchev., Russell Met. 4 (1990) 97. [7] S.V. Zherebstov, G.A. Salishchev, R.M. Galeyev, O.R. Valiakhmetov, S. Y. Mironov, S.L. Semiatin., Scr. Mater. 51 (2004) 1147. [8] A.K. Parimi, P.S. Robi, S.K. Dwivedy, Mater. Des. 32 (2011) 1948. [9] A. Kundu, R. Kapoor, R. Tewari, J.K. Chakravartty., Scr. Mater. 58 (2008) 235. [10] O.F. Higuera-Cobos, J.M. Cabera., Mater. Sci. Eng. A. 571 (2013) 103. [11] F. Salimyanfard, M.R. Toroghinejad, F. Ashrafizadeh, M. Hoseini, J.A. Szpunar., Mater. Des. 44 (2013) 374. [12] M. Shaarbaf, M.R. Toroghinejad., Mater. Sci. Eng. A. 473 (2008) 28. [13] S. Pasebani, M.R. Toroghinejad., Mater. Sci. Eng. A. 527 (2010) 491. [14] A. Sarkar, A. Bhowmik, S. Suwas., Appl. Phy. A 94 (2009) 943. [15] J. Gubicza, N.Q. Chinh, T. Csanadi, T.G. Langdon, T. Ungar., Mater. Sci. Eng. A. 462 (2007) 86. [16] T. Ungaâr, S. Ott, P.G. Sanders, A. Borbeâly, J.R. Weertman., Acta Mater. 46 (1998) 3693. [17] T. Makhlouf, A. Rebhi, J.P. Couzinié, Y. Champion, N. Njah., Int. J. Miner. Metall. Mater. 19 (2012) 1016. [18] J.R. Davis, Copper and Copper Alloys, ASM International, Ohio, 2001. [19] 〈http://www.copper.org/applications/industrial/bronze_bearing.html〉. [20] W. Bolton, Newnes Engineering Materials Pocket Book, Heinemann Newnes, Oxford, 1989. [21] N. Rangaraju, T. Raghuram, B.V. Krishna, K.P. Rao, P. Venugopal., Mater. Sci. Eng. A. 398 (2005) 246. [22] C.X. Huang, K. Wang, S.D. Wu, Z.F. Zhang, G.Y. Li, S.X. Li., Acta Mater. 54 (2006) 655. [23] L. Lu, R. Schwaiger, Z.W. Shan, M. Dao, K. Lu, S. Suresh., Acta Mater. 53 (2005) 2169. [24] M. Dao, L. Lu, Y.F. Shen, S. Suresh., Acta Mater. 54 (2006) 5421. [25] Y. Li, Y.H. Zhao, W. Liu, C. Xu, Z. Horita, X.Z. Liao, Y.T. Zhu, T.G. Langdon, E. J. Lavernia., Mater. Sci. Eng. A. 527 (2010) 3942. [26] V.S. Sarma, K. Sivaprasad, D. Sturm, M. Heilmaier., Mater. Sci. Eng. A. 489 (2008) 253. [27] N.K. Kumar, B. Roy, J. Das., J. Alloy. Comp. 618 (2015) 139. [28] G.E. Dieter, Mechanical Metallurgy, McGraw-Hill, New York, 1988. [29] Y.S. Li, N.R. Tao, K. Lu., Acta Mater. 56 (2008) 230. [30] A.M. Hodge, Y.M. Wang, T.W. Barbee., Scr. Mater. 59 (2008) 163. [31] S.V. Dobatkin, J. Gubicza, D.V. Shangina, N.R. Bochvar, N.Y. Tabachkova., Mater. Lett. 153 (2015) 5. [32] J. Xing, H. Soda, X. Yang, H. Miura, T. Sakai., J. Jpn. Inst. Light. Met. 46 (2005) 1646. [33] V.D. Mote, Y. Purushotham, B.N. Dole., J. Theo., Appl. Phys. 6 (2012) 1. [34] Y.H. Zhao, Z. Horita, T.G. Langdon, Y.T. Zhu., Mater. Sci. Eng. A. 474 (2008) 342. [35] Y. Zhong, T. Sakaguchi, F. Yin., Mater. Sci. Eng. A. 492 (2008) 419. [36] Y. Zhang, N.R. Tao, K. Lu., Acta Mater. 56 (2008) 2429. [37] L. Balogh, R.B. Figueiredo, T. Ungar, T.G. Langdon., Mater. Sci. Eng. A. 528 (2010) 533. [38] B.K. Prasad., Wear. 257 (2004) 110. [39] J. Wang, Q. Yu, Y. Jiang, I.J. Beyerlein., JOM 66 (2014) 95. [40] J. Tu, X. Zhang, J. Wang, Q. Sun, Q. Liu, C.N. Tome., Appl. Phys. Lett. 103 (2013) 051903. [41] J. Wang, I.J. Beyerlein, C.N. Tome., Int. J. Plast. 56 (2014) 156. [42] J. Das., Mater. Sci. Eng. A. 530 (2011) 675. [43] D.K. Wilsdorf, J.H.V. Merwe., Mater. Sci. Eng. 55 (1982) 79. [44] A.S. Taha, F.H. Hammad., Phys. Status Solidi (a). 119 (1990) 455. [45] M.R. Rezaei, M.R. Toroghinejad, F. Ashrafizadeh., J. Mater. Proc. Technol. 211 (2011) 1184. [46] Y.H. Zhao, Y.T. Zhu, X.Z. Liao, Z. Horita, T.G. Langdon., Appl. Phys. Lett. 89 (2006) 1. [47] B.K. Prasad, A.H. Yegneshwaran, A.K. Patwardhan., J. Mater. Sci. 32 (1997) 1169.