- Email: [email protected]
AA6061 composite produced by accumulative roll bonding
Materials Science & Engineering A 559 (2013) 345–351
Contents lists available at SciVerse ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Ultrafine grained AA1050/AA6061 composite produced by accumulative roll bonding Lihong Su a,b, Cheng Lu a,n, Anh Kiet Tieu a, Guanyu Deng a,c, Xudong Sun b a
School of Mechanical, Materials and Methatronics Engineering, University of Wollongong, Australia School of Materials and Metallurgy, Northeastern University, Shengyang, China c State Key Lab of Rolling and Automation, Northeastern University, Shengyang, China b
a r t i c l e i n f o
abstract
Article history: Received 19 May 2012 Received in revised form 21 August 2012 Accepted 22 August 2012 Available online 30 August 2012
In this work accumulative roll bonding (ARB) was used to combine AA1050 and AA6061 sheets to produce an AA1050/AA6061 composite sheet with ultrafine grained (UFG) structure. Two different starting materials were roll bonded as alternate layers up to 5 rolling cycles with 200 1C pre-heating for 3 min before each cycle. Other two types of the UFG sheets with monotonic starting materials, AA1050/ AA1050 and AA6061/AA6061, were fabricated by the same ARB process for comparison. The AA1050/ AA6061 composite sheets with strong bonding between different materials were obtained and substantial grain refinement was achieved after ARB processing. It was found that two different materials in the AA1050/AA6061 composite deformed in a nearly identical way during the first 3 ARB cycles. Afterwards the AA6061 started to neck and eventually fractured. The areas around the interface of two different materials were observed by transmission electron microscopy (TEM) and it was found that the microstructure of the bonded interface was quite complex. Two types of interfacial morphologies were observed along the AA1050/AA6061 interfaces: Type I is the area with direct contact of fresh metals and type II is the area with original metals with surface brittle layers in between. The rule of mixture has been applied to predict the strength of the AA1050/AA6061 composite sheet using the mechanical properties of the AA1050/AA1050 and AA6061/AA6061 sheets. The predicted and the measured tensile strength values are comparable to each other. The hardness values of the AA1050 and AA6061 layers in the composite are close to those measured in the monotonic material laminates, which indicated that both materials accumulated similar strain no matter if they are deformed in the composite sheets or in the single material sheets. & 2012 Elsevier B.V. All rights reserved.
Keywords: Accumulative roll bonding Ultrafine grained materials Laminated composite Bonded interface Rule of mixture
1. Introduction Several severe plastic deformation (SPD) techniques, such as equal channel angular pressing (ECAP), high pressure torsion (HPT) and accumulative roll bonding (ARB), have been widely studied over the last two decades, due to their capability of manufacturing a wide range of bulk metals and alloys with ultrafine grained (UFG) structure [1]. ARB was first proposed by Saito et al. to produce UFG sheet materials [2]. As using the same equipment as conventional rolling, ARB is considered to be one of the most promising methods for manufacturing UFG sheet materials. During ARB, rolling is conducted on two layered sheets which have the exact same dimensions and have been stacked together beforehand. The reduction per rolling pass is 50% and the two sheets are bonded together during the rolling process.
n
Corresponding author. Fax: þ61 2 42213101. E-mail addresses: [email protected], [email protected] (C. Lu).
0921-5093/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.msea.2012.08.109
The bonded specimen of each cycle is conducted to cutting, surface degreasing, brushing, and stacking together for the next cycle [2]. The process can repeat for many times as the geometry of the sheet remains unchanged after rolling. The materials obtained from ARB are quite different from those manufactured by other SPD methods such as ECAP or HPT as the ARB processed materials are more like layered composites [3,4]. The ARB process allows bonding of two different kinds of materials. Al/Mg [5], Al/Cu [6], Al/Ni [7], Al/Zn [8], Al/Ti [9], Cu/Ag [10], Cu/Zr [10], Fe/Cu [11] and other laminated composites have been produced by ARB. It has been found that during the ARB process of two different materials, both materials deform in a similar manner at the first few ARB cycles. Afterwards plastic instabilities in the harder material layers occur earlier than the other and the harder material experiences necking and fracture as the number of ARB cycles increases. This deformation behavior results in homogeneously distributed fragmentation of the hard material in the soft material matrix [5,7–9,12–14]. Mechanical properties of the composite were influenced by the necking and
346
L. Su et al. / Materials Science & Engineering A 559 (2013) 345–351
rupture of the hard layers. Some work reported that the strength of the composite still increased despite the fracture of the hard layers [8,9,13,14], while others reported that the strength of the composite dropped because of the rupture [5]. It is generally believed that the strength of the composite is determined by the mechanical properties of both materials by the rule of mixture. For the composite produced by ARB of dissimilar starting materials, the two materials certainly obtain different properties after ARB deformation. It is reported that the two layers in the composite have different microstructures and different mechanical properties measured mostly by hardness tests in separate layers [5,8,14–16]. Although few studies have been done, the strength of the composite can be predicted by the rule of mixture [9]. It is reported by Yang et al. that the yield strength of the Ti/Al composite fitted well with the value obtained from the rule of mixture [9]. Two grades of the same base metals were also used as starting materials, such as Al/Al(Sc) [17], AA1050/AA5754 [15], AA6014/ AA5754 [15] and AA2219/AA5086 [12]. Combination of different grades of aluminum alloys by accumulative roll bonding could result in materials with combination of the preferential properties of the base materials. AA1050/AA5754 and AA6014/AA5754 laminates were produced by using AA5754 as the core material and AA1050 and AA6014 as lateral materials, respectively. It showed that the composites exhibited combinative positive properties of the single component materials [15]. Quadir et al. processed Al/Al(Sc) layered laminates by using commercial purity aluminum and aluminum with 0.3 wt% Sc as starting materials. It showed that after 350 1C annealing, Al and Al(Sc) layers showed coarse and fine microstructures, respectively [17]. It has been concluded that layered composites can exhibit positive properties of both the starting materials and can produce composites with layered microstructure and mechanical properties distribution. In the present study, commercial aluminum alloys AA1050 and AA6061 were used as starting materials in ARB process to produce laminated composites with AA1050 and AA6061 in alternate layers. Microstructure and mechanical properties of the composites and of the AA1050 and AA6061 layers in the composites were investigated and also were compared with those of AA1050/ AA1050 and AA6061/AA6061 laminates processed by the same ARB process.
2. Experimental Commercial aluminum alloy AA1050 and AA6061 sheets with initial thicknesses of 1.5 mm were used in this work. AA1050 was annealed at 450 1C for 1 h and AA6061 was annealed at 500 1C for 2 h to achieve a fully annealed homogeneous microstructure, resulting in an average grain size of 96 mm for AA1050 and 36 mm for AA6061. The microstructures of the annealed AA1050 and AA6061 are shown in Fig. 1(a) and (b), respectively. Sheet materials with a dimension of 1.5 50 400 mm3 (thickness width length) were cut from the original sheets, with the longitudinal direction parallel to the original rolling direction. The experiments were grouped into three categories according to the starting materials: AA1050/AA1050, AA6061/AA6061 and AA1050/AA6061. Prior to each rolling cycle, the rolls were cleaned by acetone and the roll gap and speed were set to the required setting. One side of the sample was cleaned with acetone and wire-brushed. Two pieces of starting materials were stacked together and weld at one end. The materials were pre-heated in a furnace at 200 1C for 3 min and then rolled with a nominal reduction of 50% under dry condition. This temperature was chosen to achieve good hardening and bonding simultaneously [18]. The rolled samples were cut into two halves and stacked
together by the same method. The above procedure proceeded for 5 cycles to obtain an equivalent strain of 4. Mechanical properties were tested by tensile and hardness tests. Tensile tests were conducted using an Instron 5566 testing machine with an initial strain rate of 10 3/s at room temperature. Tensile specimens with 25 mm gauge length and 6 mm gauge width were processed along the rolling direction of the ARB processed samples. Vickers micro-hardness was measured by a Leco hardness testing machine to show the mechanical behavior before and after each ARB deformation. The hardness values were taken on the RD-ND plane and a load of 25 g for 12 s was used. The initial hardness of the AA1050 and AA6061 sheets were Hv 28.9 and Hv 39.0, respectively. The measurements were taken along the thickness direction of the ARB processed samples with 50 mm distance of two adjacent indents. Average hardness values of the AA1050 and AA6061 layers were obtained from averaging the indentations within the corresponding alloy layer, regardless of the position of the layer. The microstructure was observed by optical microscopy (OM) and transmission electron microscopy (TEM). The optical microstructures were observed with a Leica DMRM microscope, in the longitudinal cross-section after grinding and polishing by a Struers TegraPol-21 polishing machine with an OPS finish and then etching with Barker’s reagent. TEM micrographs were obtained with a JEOL 2011F microscope operating at 200 kV. Thin foils for TEM were prepared by twin-jet electron polishing with an electrolyte of 25% nitric acid in methanol at –20 1C. The graphs were taken on the RD-ND plane.
3. Results and discussion Figs. 1(c)–(e) show the optical micrographs of 1, 3 and 5-cycle ARB processed AA1050/AA6061 composites under polarized light. It can be seen that the grains after ARB have a lamellar structure which are elongated along the rolling direction and the grain size has been decreased substantially as compared to those of the annealed materials before ARB, as shown in Fig. 1(a) and (b). The grain size becomes finer with increasing ARB cycles. It can be seen that the interfaces of AA1050 and AA6061 after 1 and 3-cycle ARB are straight and the thicknesses of both the AA1050 layers and the AA6061 layers are uniform through the whole cross-section. The interfaces become wavy after the fourth cycle (not shown here) and the thicknesses of the AA6061 layers become non-uniform along the rolling direction. After 5-cycle ARB, necking of AA6061 layers continues and fracture happens at various places. Fig. 1(f) shows the micrograph of the 5-cycle ARB processed composite at higher magnification under normal light. The AA1050 and AA6061 layers can be identified by the dense micro-sized precipitates in the AA6061 layers. It is obvious that the AA6061 layers are irregular and wavy in shape, and fractured at various locations. The thicknesses of both the layers vary from point to point. As the hardening rate of AA6061 is significantly higher than that of AA1050, the strength of AA6061 layers in the composite is higher than that of the AA1050 layers, the difference of which is reflected on the relative thinning of the two layers. The AA1050 layers thin down easier than AA6061 layers. At the first three cycles, the AA6061 layers are able to deform in the same way as the AA1050 layers. However, the deformation becomes inhomogeneous and the thickness of AA6061 layers starts to vary along the rolling direction at further ARB cycles. The AA6061 layers start to neck with further deformation and the adjacent AA1050 materials then protrude and fill in the necked and fractured regions. Finally the AA6061 layers fracture at various locations and form the structure shown in Fig. 1(e) and (f).
L. Su et al. / Materials Science & Engineering A 559 (2013) 345–351
ND
347
ND 50 µm
RD
AA1050 AA6061
50 µm
RD
AA6061 AA1050
Interface
ND 200 µm
200 µm
RD
AA6061 AA1050
ND
ND RD
200 µm
20 µm
RD
Fig. 1. Optical micrographs of (a) annealed AA1050, (b) annealed AA6061 before ARB processing and AA1050/AA6061 laminates after (c) 1, (d) 3, (e) and (f) 5-cycle ARB process. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)
Shear zones are observed in the composite processed after 5-cycle ARB, as indicated by different colors in Fig. 1(e). The angles of the shear zones with respect to the rolling direction have the highest value close to the thickness center and decrease when moving towards the surfaces. The angle is 301 at the thickness center and reduces to 171 at the quarter thickness and then further reduces as moving outside. The AA6061 layers close to the surfaces of the composite are not fractured, thus shear bands are hardly seen near the two surfaces. The similar microstructure was also observed by Roy et al. [12]. During the rolling deformation, the material near the sample surface is subject to large shear deformation while the material near the thickness center is subject to large elongation deformation. The shear deformation does not introduce essential tensile stress in the material. However, in the elongation deformation, the soft materials elongate more than the hard materials, which induces tensile stress along the rolling direction in the latter and then breaks the hard materials when the tensile stress exceeds a certain level. This is the reason why the fracture of the hard materials was observed near the thickness center but not close to the surfaces. The average thicknesses of the AA1050 layers and AA6061 layers are obtained by the linear intercept method. The values are listed in Table 1. It is clear that from 1 to 4-cycle, the thickness reduction for both the AA1050 layer and the AA6061 layer is
Table 1 Average thicknesses, reductions and volume fractions of AA1050 layers and AA6061 layers in the ARBed composites. Number of ARB cycles
AA1050 layers
AA6061 layers
Thickness (mm)
Reduction (%)
Thickness (mm)
Reduction (%)
1 2 3 4 5
802.07 3.7 402.37 10.2 193.6 7 12.3 92.17 9.1 32.97 17.8
46.5 49.8 51.9 52.4 64.3
803.5 7 7.8 399.4 7 25.0 194.0 7 8.8 93.9 7 11.1 58.3 7 19.4
46.4 50.3 51.4 51.6 37.9
Volume fraction of AA6061 (%)
50.05 49.82 48.72 50.51 50 (estimated)
around 50%. However, the thickness reduction for the AA6061 layer after 5-cycle ARB is less than 50%. This implies that the accumulated strain in the AA6061 layers is slightly lower than that in the AA1050 layers. It has been reported that once the fracture of the hard layers occurs, the thickness of the hard layers will level off when subjected to further ARB deformation [7,8,13,14], which suggests that in the present study, the thickness of AA6061 starts to level off at the fifth ARB cycle. The thicknesses of the AA6061 layers after 6-cycle ARB were also measured. The average value is 45.8 mm and the reduction is 21.4%. It can be seen that the average
348
L. Su et al. / Materials Science & Engineering A 559 (2013) 345–351
AA1050 AA1050 AA6061
AA6061
1 µm
2 µm
AA6061
AA6061
Interface
AA1050
2 µm
AA1050 2 µm
Fig. 2. (a)–(d) TEM micrographs the interfaces of 5-cycle ARBed AA1050/AA6061 composite and the SAD pattern of (e) AA1050 layer, (f) interfacial area and (g) AA6061 layer in (d).
thickness of AA6061 layers is slightly smaller than that after 5-cycle ARB, which confirms that the thickness of AA6061 started to level off from the fifth cycle on. The volume fraction of the AA6061 layers was calculated by dividing the thickness of the AA6061 layers by the total thickness of the composite. It can be seen that the volume fraction of the AA6061 layers stay around 50% for all the cycles. Note that the volume fraction of the 5-cycle ARB processed composite is an estimated value as the thickness distribution of the AA6061 layers was so nonuniform that it cannot be used to represent volume fraction. TEM observations were conducted close to the bonding interfaces of the 5-cycle ARB processed composite. The micrographs are shown in Fig. 2. It can be seen that there are several morphologies of the interfaces. The AA1050 layer and the AA6061 layer closely contact with each other in Fig. 2(a)–(c), whereas the two layers are separated by oxides and contamination induced small grains in Fig. 2(d). Bonded interfaces consisted of finer grains and oxides and precipitations were also reported in other studies [4,19]. In addition, there are dense micro-sized precipitates in the AA6061 layer (indicated by arrows in Fig. 2), which influence the metal flow during the rolling deformation and form a curved interface (Fig. 2(b)) or induce the rotation of the elongated grains to a certain angle with respect of the interface (Fig. 2(c)). It is believed that the morphologies of the interfaces would influence the bond quality. The finer grains induced by oxides and contamination cause high hardness values close to the interfaces [15,19], which may improve bond toughness. Nano-sized secondary particles were also intentionally added between the interfaces
to enhance bond strength [20,21]. However, large precipitates and particles close to the interfaces may act as origins of little cracks which later induce de-lamination and reduce bond strength. In the present work, the surfaces of the sheets were well brushed and cleaned to ensure that thick oxidation layers and large contaminations were removed. The more refined grains observed close to the interfaces due to thin oxides and contaminations lead to enhanced hardness and hence improve bond quality. During ARB, bonding occurs by fracture of oxide layers to allow the fresh metals contact with each other, which leads to two kinds of areas along the interface. Type I is direct contact of fresh metals (AA1050/AA6061 contact in this case), as shown in Fig. 2(a) and (c); type II is original metal surfaces with thin oxide film layers in between. The oxide films, which are brittle, are easy to break and allow further refinement of the materials around the interface. Therefore, the grain size in this kind of interface is finer than that of the other areas. As can be seen in Fig. 2(d), the AA1050/AA6061 bonding interface is characterized by fine grains which can be distinguished from AA1050 easily but not AA6061 as the grain size is closer to that of the AA6061 layer. It is believed that the grain refinement at the interface is caused by oxide film and other contamination. The fraction of high angle grain boundaries around the interfacial area is also higher than that of the AA1050 layer and AA6061 layer, as indicated by the SAD patterns in Fig. 2(e)–(g). Development of the two kinds of interfacial areas is schematically shown in Fig. 3. Assume that there are thin layers of oxide film at the surface of the starting materials and it distributes evenly along the surface. Some contamination would also exist at
L. Su et al. / Materials Science & Engineering A 559 (2013) 345–351
349
Fig. 3. Schematic diagram showing the development of two types of interface morphologies (a) Initial state: surface oxide and contamination influenced layers are shown as straight, (b) Deformation at low cycles: surface brittle layers break up and induce areas with direct contact of fresh metals (type I) and areas of original metals with surface brittle layers in between (type II), (c) Deformation at high cycles: surface brittle layers allow further refinement of the materials around the interface, (d) Type I and II interfacial areas can be seen in optical micrograph.
the surface even though it has been cleaned and wire brushed before ARB deformation. Both the oxide film and contamination would influence the deformation and the morphologies of the interfacial areas. As the oxide films have much lower ductility than aluminum, it is easy to fracture even at low cycles of ARB deformation, shown in Fig. 3(b). The fresh metals where the oxide films fractured appear and closely bond with each other due to the pressure of the rolling deformation and further refine with the increasing ARB cycles. The areas where the oxide films locate between the metal layers will allow further refinement because the oxide films are harder than the metals. Therefore, the grain sizes of the materials around the type II interfacial areas are smaller than that around the type I interface areas and the metal matrix, shown in Fig. 3(c). Besides the microstructures of the interfacial areas shown in Fig. 2 (type I in Fig. 2(a)–(c) and type II in Fig. 2(d)), the two types of interfacial areas can be observed with optical microscope, shown in Fig. 3(d). Note that the thick interfacial areas are deliberately chosen to show the effect but the thicknesses of the type II interfacial areas are generally much smaller than that shown in Fig. 3(d). Tensile properties of the three kinds of ARB processed sheets are shown in Figs. 4 and 5. The error bars in Fig. 4 are smaller than the symbols. The tensile strength increases with increasing number of ARB cycles for all the sheets and the values after 5-cycle ARB are more than two times greater than that of the initial materials. Uniform elongation, on the other hand, decreases substantially after 1-cycle ARB and stayed almost unchanged with further deformation. The uniform elongations of all the three kinds of materials are less than 3%. It is common to be observed for ARB processed materials that necking starts at an early stage of deformation and results in early fracture and low ductility [2,4,8,9]. It can be observed in Fig. 4 that the tensile strength of the AA6061 laminates are always the highest and that of the AA1050 laminates are the lowest, while the tensile strength of the AA1050/AA6061 composites locates in between two monotonic material laminates. Curve (d) of Fig. 4 is calculated from the values of curves (a) and (b) by the rule of mixture s ¼V1050s1050 þV6061s6061, where V1050 and V6061 are the volume fractions of AA1050 and AA6061 as listed in Table 1, s1050 and s6061 are the tensile strength of the AA1050/AA1050 and AA6061/ AA6061 laminates, respectively. It is obvious that the directly measured tensile strength of the composite fits reasonably well with the value obtained by the rule of mixture. Note that in the
Fig. 4. Tensile strength values of ARBed (a) AA6061/AA6061 laminates, (b) AA1050/AA1050 laminates, (c) AA1050/AA6061 laminates and (d) AA1050/ AA6061 laminates calculated by rule of mixture. (0 ARB cycle corresponding to hardness of annealed materials).
present work the initial values for the rule of mixture are the tensile strength values of the laminates processed by ARB from monotonic starting materials. However in Ref. [9], the values used for the rule of mixture were calculated from the average hardness values measured on the two different layers of the composite. The fact that the measured tensile strength values fit with the calculated ones indicates that both the materials AA1050 and AA6061 undergo similar strain path no matter if they are deformed in the monotonic material laminates or in the composites. Fig. 4 shows that the difference between the measured tensile strength value of the AA1050/AA6061 composite and the value from the rule of mixture for the 5-cycle ARB processed material is greater than that of other cycles, which can be attributed to the fact that the hard layers in the composite have less deformation after 5-cycle ARB, as indicated in Table 1. It is expected that the tensile strength of the composite subjected to more than 5-cycle ARB would move away from the value of the mixture rule. Fig. 6 shows the effect of ARB deformation on strain hardening rate. The strain hardening rate is plotted against true strain. The
350
L. Su et al. / Materials Science & Engineering A 559 (2013) 345–351
effect of ARB deformation can be seen by comparing the strain hardening rate of ARB processed samples for different cycles. As can be seen in Fig. 6(a), the strain hardening rate is increasing with the number of ARB cycles. The similar phenomenon has been observed in Ref. [22]. It can be explained by decreasing grain sizes with the increasing number of ARB cycles due to the increasing accumulative strain. The high strain hardening rate in the samples with high accumulated strain is due to the smaller grain sizes in these samples, which makes it easier for dislocations to encounter grain boundaries, and then resist further deformation. Fig. 6(b) shows the strain hardening rate of 5-cycle ARB processed samples with different starting materials along with the initial annealed materials. It can be seen that the initial hardening rate of the ARB processed samples shows a larger value than the corresponding annealed conventional sized samples, which can be explained by the grain size effect above. However, the annealed samples display positive strain hardening values to significant high strains ( 34% for annealed AA1050 and 24% for annealed AA6061), whereas the strain hardening rate of ARB processed samples drops rapidly below zero after only a small plastic strain subjected to the tensile specimens (less than 3% for all the cases). This explains the limited uniform elongation of the ARB processed samples. It also can be seen in Fig. 6(b) that 5-cycle ARB processed AA6061/AA6061 has high strain hardening
Fig. 5. Uniform elongation of ARBed (a) AA6061/AA6061 laminates, (b) AA1050/ AA1050 laminates and (c) AA1050/AA6061 laminates. (0 ARB cycle corresponding to hardness of annealed materials).
rate than that of the 5-cycle ARB processed AA1050/AA1050 and the initial strain hardening rate of the 5-cycle ARB processed AA1050/AA6061 locates between the values of the two monotonic laminates, which is in good correspondence with the tensile strength values. The average microhardness values of the ARB processed samples are shown in Fig. 7. The evolution of hardness values with the number of ARB cycles is similar to that of the tensile strength. The hardness increases rapidly after the first cycle ARB and then the increment between each cycle becomes less significant. After 5-cycle ARB, the hardness values become around two times greater than that of the initial values for both AA6061 and AA1050. The hardness values of AA6061 are always higher than that of AA1050. The tensile test takes the AA1050/AA6061 composite as a complete material, which makes it hard to evaluate the behavior of the two different layers. However, in hardness test, the values of the AA1050 and AA6061 layers of the composite can be measured separately. Curves (b) and (c) in Fig. 7 denote the average hardness values of the AA6061 and AA1050 layers in the composites, respectively. As can be seen in the figure that the average hardness values of AA6061 single material laminates are very close to that of the AA6061 layers in the AA1050/AA6061
Fig. 7. Hardness values of ARBed (a) AA6061/AA6061 laminates, (b) AA6061 layer in AA1050/AA6061 laminates, (c) AA1050 layer in AA1050/AA6061 laminates and (d) AA1050/AA1050 laminates. (0 ARB cycle corresponding to hardness of annealed materials).
Fig. 6. Strain hardening rate of (a) 1, 3 and 5-cycle ARBed AA1050/AA6061 laminates and (b) annealed starting materials and 5-cycle ARBed laminates with different starting materials.
L. Su et al. / Materials Science & Engineering A 559 (2013) 345–351
laminates, so is that of the AA1050 materials. This confirms the conclusion obtained from the tensile strength values that the materials AA1050 and AA6061 undergo similar deformation no matter if they are deformed in the single material laminates or in the composites for up to 5 ARB cycles.
4. Conclusion
351
strain no matter if they are deformed in the composites or in the monotonic material laminates.
Acknowledgment The author would like to thank Mr. Joe Abbott for the help of ARB experiments. Lihong Su and Guanyu Deng would like to thank the Chinese Scholarship Council for financial support.
AA1050/AA6061 composites were successfully processed by the ARB process. The followings can be concluded. References 1. Laminated composite materials with strong bonding between the AA1050 layers and the AA6061 layers were obtained and substantial grain refinement was achieved after ARB processing. The areas around the interface of two adjacent layers were observed and it was found that the microstructure of the bonded interface was quite complex. 2. During the first 3 ARB cycles, the AA1050 layers and the AA6061 layers in the composite deformed in a nearly identical way. Afterwards the AA6061 started to neck and eventually fractured at various locations. Severe shear bands were observed throughout the cross-section of the 5-cycle ARB processed composite. 3. Two types of interfacial morphologies were observed along the AA1050/AA6061 interfaces: Type I is the area with direct contact of fresh metals and type II is the area with original metals with surface brittle layers in between. Type II interfacial morphologies allow further refinement of the microstructure close to the interface. 4. The hardness and tensile strength of the 5-cycle ARB processed composite became more than two times greater than the initial values whereas the ductility dropped significantly. The tensile strengths of the composite predicted by the rule of mixture were comparable to the measured values. The hardness values of AA1050 and AA6061 in the composite were nearly identical to those in the monotonic material laminates, which indicated that both materials accumulated similar
[1] R.Z. Valiev, T.G. Langdon, Prog. Mater. Sci. 51 (2006) 881–981. [2] Y. Saito, N. Tsuji, H. Utsunomiya, T. Sakai, R.G. Hong, Scripta Mater. 39 (1998) 1221–1227. [3] H. Chang, M.Y. Zheng, K. Wu, W.M. Gan, L.B. Tong, H.G. Brokmeier, Mater. Sci. Eng. A 527 (2010) 7176–7183. [4] S.H. Lee, Y. Saito, T. Sakai, H. Utsunomiya, Mater. Sci. Eng. A 325 (2002) 228–235. [5] K. Wu, H. Chang, E. Maawad, W.M. Gan, H.G. Brokmeier, M.Y. Zheng, Mater. Sci. Eng. A 527 (2010) 3073–3078. [6] H.S. Liu, B. Zhang, G.P. Zhang, Scripta Mater. 64 (2011) 13–16. [7] G.H. Min, J.M. Lee, S.B. Kang, H.W. Kim, Mater. Lett. 60 (2006) 3255–3259. [8] R.N. Dehsorkhi, F. Qods, M. Tajally, Mater. Sci. Eng. A 530 (2011) 63–72. [9] D.K. Yang, P. Cizek, P. Hodgson, C. Wen, Scripta Mater. 62 (2010) 321–324. [10] S. Ohsaki, S. Kato, N. Tsuji, T. Ohkubo, K. Hono, Acta Mater. 55 (2007) 2885–2895. [11] F.H. Kavarana, K.S. Ravichandran, S.S. Sahay, Scripta Mater. 42 (2000) 947–954. [12] S. Roy, B.R. Nataraj, S. Suwas, S. Kumar, K. Chattopadhyay, Mater. Des. 36 (2012) 529–539. [13] L. Ghalandari, M.M. Moshksar, J. Alloys Compd. 506 (2010) 172–178. [14] A. Mozaffari, H.D. Manesh, K. Janghorban, J. Alloys Compd. 489 (2010) 103–109. [15] T. Hausol, H.W. Hoppel, M. Goken, J. Mater. Sci. 45 (2010) 4733–4738. [16] L.H. Su, C. Lu, G.Y. Deng, A.K. Tieu, X.D. Sun, Rev. Adv. Mater. Sci., in press. [17] M.Z. Quadir, O. Al-Buhamad, L. Bassman, M. Ferry, Acta Mater. 55 (2007) 5438–5448. [18] I. Topic, H.W. Heinz, M. Goken, Int. J. Mater. Res. 98 (2007) 320–324. [19] M.Z. Quadir, A. Wolz, M. Hoffman, M. Ferry, Scripta Mater. 58 (2008) 959–962. [20] C. Lu, K. Tieu, D. Wexler, J. Mater. Process. Technol. 209 (2009) 4830–4834. [21] R. Jamaati, M.R. Toroghinejad, Mater. Sci. Eng. A 527 (2010) 4858–4863. [22] C. Kwan, Z.R. Wang, S.B. Kang, Mater. Sci. Eng. A 480 (2008) 148–159.
Contents lists available at SciVerse ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Ultrafine grained AA1050/AA6061 composite produced by accumulative roll bonding Lihong Su a,b, Cheng Lu a,n, Anh Kiet Tieu a, Guanyu Deng a,c, Xudong Sun b a
School of Mechanical, Materials and Methatronics Engineering, University of Wollongong, Australia School of Materials and Metallurgy, Northeastern University, Shengyang, China c State Key Lab of Rolling and Automation, Northeastern University, Shengyang, China b
a r t i c l e i n f o
abstract
Article history: Received 19 May 2012 Received in revised form 21 August 2012 Accepted 22 August 2012 Available online 30 August 2012
In this work accumulative roll bonding (ARB) was used to combine AA1050 and AA6061 sheets to produce an AA1050/AA6061 composite sheet with ultrafine grained (UFG) structure. Two different starting materials were roll bonded as alternate layers up to 5 rolling cycles with 200 1C pre-heating for 3 min before each cycle. Other two types of the UFG sheets with monotonic starting materials, AA1050/ AA1050 and AA6061/AA6061, were fabricated by the same ARB process for comparison. The AA1050/ AA6061 composite sheets with strong bonding between different materials were obtained and substantial grain refinement was achieved after ARB processing. It was found that two different materials in the AA1050/AA6061 composite deformed in a nearly identical way during the first 3 ARB cycles. Afterwards the AA6061 started to neck and eventually fractured. The areas around the interface of two different materials were observed by transmission electron microscopy (TEM) and it was found that the microstructure of the bonded interface was quite complex. Two types of interfacial morphologies were observed along the AA1050/AA6061 interfaces: Type I is the area with direct contact of fresh metals and type II is the area with original metals with surface brittle layers in between. The rule of mixture has been applied to predict the strength of the AA1050/AA6061 composite sheet using the mechanical properties of the AA1050/AA1050 and AA6061/AA6061 sheets. The predicted and the measured tensile strength values are comparable to each other. The hardness values of the AA1050 and AA6061 layers in the composite are close to those measured in the monotonic material laminates, which indicated that both materials accumulated similar strain no matter if they are deformed in the composite sheets or in the single material sheets. & 2012 Elsevier B.V. All rights reserved.
Keywords: Accumulative roll bonding Ultrafine grained materials Laminated composite Bonded interface Rule of mixture
1. Introduction Several severe plastic deformation (SPD) techniques, such as equal channel angular pressing (ECAP), high pressure torsion (HPT) and accumulative roll bonding (ARB), have been widely studied over the last two decades, due to their capability of manufacturing a wide range of bulk metals and alloys with ultrafine grained (UFG) structure [1]. ARB was first proposed by Saito et al. to produce UFG sheet materials [2]. As using the same equipment as conventional rolling, ARB is considered to be one of the most promising methods for manufacturing UFG sheet materials. During ARB, rolling is conducted on two layered sheets which have the exact same dimensions and have been stacked together beforehand. The reduction per rolling pass is 50% and the two sheets are bonded together during the rolling process.
n
Corresponding author. Fax: þ61 2 42213101. E-mail addresses: [email protected], [email protected] (C. Lu).
0921-5093/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.msea.2012.08.109
The bonded specimen of each cycle is conducted to cutting, surface degreasing, brushing, and stacking together for the next cycle [2]. The process can repeat for many times as the geometry of the sheet remains unchanged after rolling. The materials obtained from ARB are quite different from those manufactured by other SPD methods such as ECAP or HPT as the ARB processed materials are more like layered composites [3,4]. The ARB process allows bonding of two different kinds of materials. Al/Mg [5], Al/Cu [6], Al/Ni [7], Al/Zn [8], Al/Ti [9], Cu/Ag [10], Cu/Zr [10], Fe/Cu [11] and other laminated composites have been produced by ARB. It has been found that during the ARB process of two different materials, both materials deform in a similar manner at the first few ARB cycles. Afterwards plastic instabilities in the harder material layers occur earlier than the other and the harder material experiences necking and fracture as the number of ARB cycles increases. This deformation behavior results in homogeneously distributed fragmentation of the hard material in the soft material matrix [5,7–9,12–14]. Mechanical properties of the composite were influenced by the necking and
346
L. Su et al. / Materials Science & Engineering A 559 (2013) 345–351
rupture of the hard layers. Some work reported that the strength of the composite still increased despite the fracture of the hard layers [8,9,13,14], while others reported that the strength of the composite dropped because of the rupture [5]. It is generally believed that the strength of the composite is determined by the mechanical properties of both materials by the rule of mixture. For the composite produced by ARB of dissimilar starting materials, the two materials certainly obtain different properties after ARB deformation. It is reported that the two layers in the composite have different microstructures and different mechanical properties measured mostly by hardness tests in separate layers [5,8,14–16]. Although few studies have been done, the strength of the composite can be predicted by the rule of mixture [9]. It is reported by Yang et al. that the yield strength of the Ti/Al composite fitted well with the value obtained from the rule of mixture [9]. Two grades of the same base metals were also used as starting materials, such as Al/Al(Sc) [17], AA1050/AA5754 [15], AA6014/ AA5754 [15] and AA2219/AA5086 [12]. Combination of different grades of aluminum alloys by accumulative roll bonding could result in materials with combination of the preferential properties of the base materials. AA1050/AA5754 and AA6014/AA5754 laminates were produced by using AA5754 as the core material and AA1050 and AA6014 as lateral materials, respectively. It showed that the composites exhibited combinative positive properties of the single component materials [15]. Quadir et al. processed Al/Al(Sc) layered laminates by using commercial purity aluminum and aluminum with 0.3 wt% Sc as starting materials. It showed that after 350 1C annealing, Al and Al(Sc) layers showed coarse and fine microstructures, respectively [17]. It has been concluded that layered composites can exhibit positive properties of both the starting materials and can produce composites with layered microstructure and mechanical properties distribution. In the present study, commercial aluminum alloys AA1050 and AA6061 were used as starting materials in ARB process to produce laminated composites with AA1050 and AA6061 in alternate layers. Microstructure and mechanical properties of the composites and of the AA1050 and AA6061 layers in the composites were investigated and also were compared with those of AA1050/ AA1050 and AA6061/AA6061 laminates processed by the same ARB process.
2. Experimental Commercial aluminum alloy AA1050 and AA6061 sheets with initial thicknesses of 1.5 mm were used in this work. AA1050 was annealed at 450 1C for 1 h and AA6061 was annealed at 500 1C for 2 h to achieve a fully annealed homogeneous microstructure, resulting in an average grain size of 96 mm for AA1050 and 36 mm for AA6061. The microstructures of the annealed AA1050 and AA6061 are shown in Fig. 1(a) and (b), respectively. Sheet materials with a dimension of 1.5 50 400 mm3 (thickness width length) were cut from the original sheets, with the longitudinal direction parallel to the original rolling direction. The experiments were grouped into three categories according to the starting materials: AA1050/AA1050, AA6061/AA6061 and AA1050/AA6061. Prior to each rolling cycle, the rolls were cleaned by acetone and the roll gap and speed were set to the required setting. One side of the sample was cleaned with acetone and wire-brushed. Two pieces of starting materials were stacked together and weld at one end. The materials were pre-heated in a furnace at 200 1C for 3 min and then rolled with a nominal reduction of 50% under dry condition. This temperature was chosen to achieve good hardening and bonding simultaneously [18]. The rolled samples were cut into two halves and stacked
together by the same method. The above procedure proceeded for 5 cycles to obtain an equivalent strain of 4. Mechanical properties were tested by tensile and hardness tests. Tensile tests were conducted using an Instron 5566 testing machine with an initial strain rate of 10 3/s at room temperature. Tensile specimens with 25 mm gauge length and 6 mm gauge width were processed along the rolling direction of the ARB processed samples. Vickers micro-hardness was measured by a Leco hardness testing machine to show the mechanical behavior before and after each ARB deformation. The hardness values were taken on the RD-ND plane and a load of 25 g for 12 s was used. The initial hardness of the AA1050 and AA6061 sheets were Hv 28.9 and Hv 39.0, respectively. The measurements were taken along the thickness direction of the ARB processed samples with 50 mm distance of two adjacent indents. Average hardness values of the AA1050 and AA6061 layers were obtained from averaging the indentations within the corresponding alloy layer, regardless of the position of the layer. The microstructure was observed by optical microscopy (OM) and transmission electron microscopy (TEM). The optical microstructures were observed with a Leica DMRM microscope, in the longitudinal cross-section after grinding and polishing by a Struers TegraPol-21 polishing machine with an OPS finish and then etching with Barker’s reagent. TEM micrographs were obtained with a JEOL 2011F microscope operating at 200 kV. Thin foils for TEM were prepared by twin-jet electron polishing with an electrolyte of 25% nitric acid in methanol at –20 1C. The graphs were taken on the RD-ND plane.
3. Results and discussion Figs. 1(c)–(e) show the optical micrographs of 1, 3 and 5-cycle ARB processed AA1050/AA6061 composites under polarized light. It can be seen that the grains after ARB have a lamellar structure which are elongated along the rolling direction and the grain size has been decreased substantially as compared to those of the annealed materials before ARB, as shown in Fig. 1(a) and (b). The grain size becomes finer with increasing ARB cycles. It can be seen that the interfaces of AA1050 and AA6061 after 1 and 3-cycle ARB are straight and the thicknesses of both the AA1050 layers and the AA6061 layers are uniform through the whole cross-section. The interfaces become wavy after the fourth cycle (not shown here) and the thicknesses of the AA6061 layers become non-uniform along the rolling direction. After 5-cycle ARB, necking of AA6061 layers continues and fracture happens at various places. Fig. 1(f) shows the micrograph of the 5-cycle ARB processed composite at higher magnification under normal light. The AA1050 and AA6061 layers can be identified by the dense micro-sized precipitates in the AA6061 layers. It is obvious that the AA6061 layers are irregular and wavy in shape, and fractured at various locations. The thicknesses of both the layers vary from point to point. As the hardening rate of AA6061 is significantly higher than that of AA1050, the strength of AA6061 layers in the composite is higher than that of the AA1050 layers, the difference of which is reflected on the relative thinning of the two layers. The AA1050 layers thin down easier than AA6061 layers. At the first three cycles, the AA6061 layers are able to deform in the same way as the AA1050 layers. However, the deformation becomes inhomogeneous and the thickness of AA6061 layers starts to vary along the rolling direction at further ARB cycles. The AA6061 layers start to neck with further deformation and the adjacent AA1050 materials then protrude and fill in the necked and fractured regions. Finally the AA6061 layers fracture at various locations and form the structure shown in Fig. 1(e) and (f).
L. Su et al. / Materials Science & Engineering A 559 (2013) 345–351
ND
347
ND 50 µm
RD
AA1050 AA6061
50 µm
RD
AA6061 AA1050
Interface
ND 200 µm
200 µm
RD
AA6061 AA1050
ND
ND RD
200 µm
20 µm
RD
Fig. 1. Optical micrographs of (a) annealed AA1050, (b) annealed AA6061 before ARB processing and AA1050/AA6061 laminates after (c) 1, (d) 3, (e) and (f) 5-cycle ARB process. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)
Shear zones are observed in the composite processed after 5-cycle ARB, as indicated by different colors in Fig. 1(e). The angles of the shear zones with respect to the rolling direction have the highest value close to the thickness center and decrease when moving towards the surfaces. The angle is 301 at the thickness center and reduces to 171 at the quarter thickness and then further reduces as moving outside. The AA6061 layers close to the surfaces of the composite are not fractured, thus shear bands are hardly seen near the two surfaces. The similar microstructure was also observed by Roy et al. [12]. During the rolling deformation, the material near the sample surface is subject to large shear deformation while the material near the thickness center is subject to large elongation deformation. The shear deformation does not introduce essential tensile stress in the material. However, in the elongation deformation, the soft materials elongate more than the hard materials, which induces tensile stress along the rolling direction in the latter and then breaks the hard materials when the tensile stress exceeds a certain level. This is the reason why the fracture of the hard materials was observed near the thickness center but not close to the surfaces. The average thicknesses of the AA1050 layers and AA6061 layers are obtained by the linear intercept method. The values are listed in Table 1. It is clear that from 1 to 4-cycle, the thickness reduction for both the AA1050 layer and the AA6061 layer is
Table 1 Average thicknesses, reductions and volume fractions of AA1050 layers and AA6061 layers in the ARBed composites. Number of ARB cycles
AA1050 layers
AA6061 layers
Thickness (mm)
Reduction (%)
Thickness (mm)
Reduction (%)
1 2 3 4 5
802.07 3.7 402.37 10.2 193.6 7 12.3 92.17 9.1 32.97 17.8
46.5 49.8 51.9 52.4 64.3
803.5 7 7.8 399.4 7 25.0 194.0 7 8.8 93.9 7 11.1 58.3 7 19.4
46.4 50.3 51.4 51.6 37.9
Volume fraction of AA6061 (%)
50.05 49.82 48.72 50.51 50 (estimated)
around 50%. However, the thickness reduction for the AA6061 layer after 5-cycle ARB is less than 50%. This implies that the accumulated strain in the AA6061 layers is slightly lower than that in the AA1050 layers. It has been reported that once the fracture of the hard layers occurs, the thickness of the hard layers will level off when subjected to further ARB deformation [7,8,13,14], which suggests that in the present study, the thickness of AA6061 starts to level off at the fifth ARB cycle. The thicknesses of the AA6061 layers after 6-cycle ARB were also measured. The average value is 45.8 mm and the reduction is 21.4%. It can be seen that the average
348
L. Su et al. / Materials Science & Engineering A 559 (2013) 345–351
AA1050 AA1050 AA6061
AA6061
1 µm
2 µm
AA6061
AA6061
Interface
AA1050
2 µm
AA1050 2 µm
Fig. 2. (a)–(d) TEM micrographs the interfaces of 5-cycle ARBed AA1050/AA6061 composite and the SAD pattern of (e) AA1050 layer, (f) interfacial area and (g) AA6061 layer in (d).
thickness of AA6061 layers is slightly smaller than that after 5-cycle ARB, which confirms that the thickness of AA6061 started to level off from the fifth cycle on. The volume fraction of the AA6061 layers was calculated by dividing the thickness of the AA6061 layers by the total thickness of the composite. It can be seen that the volume fraction of the AA6061 layers stay around 50% for all the cycles. Note that the volume fraction of the 5-cycle ARB processed composite is an estimated value as the thickness distribution of the AA6061 layers was so nonuniform that it cannot be used to represent volume fraction. TEM observations were conducted close to the bonding interfaces of the 5-cycle ARB processed composite. The micrographs are shown in Fig. 2. It can be seen that there are several morphologies of the interfaces. The AA1050 layer and the AA6061 layer closely contact with each other in Fig. 2(a)–(c), whereas the two layers are separated by oxides and contamination induced small grains in Fig. 2(d). Bonded interfaces consisted of finer grains and oxides and precipitations were also reported in other studies [4,19]. In addition, there are dense micro-sized precipitates in the AA6061 layer (indicated by arrows in Fig. 2), which influence the metal flow during the rolling deformation and form a curved interface (Fig. 2(b)) or induce the rotation of the elongated grains to a certain angle with respect of the interface (Fig. 2(c)). It is believed that the morphologies of the interfaces would influence the bond quality. The finer grains induced by oxides and contamination cause high hardness values close to the interfaces [15,19], which may improve bond toughness. Nano-sized secondary particles were also intentionally added between the interfaces
to enhance bond strength [20,21]. However, large precipitates and particles close to the interfaces may act as origins of little cracks which later induce de-lamination and reduce bond strength. In the present work, the surfaces of the sheets were well brushed and cleaned to ensure that thick oxidation layers and large contaminations were removed. The more refined grains observed close to the interfaces due to thin oxides and contaminations lead to enhanced hardness and hence improve bond quality. During ARB, bonding occurs by fracture of oxide layers to allow the fresh metals contact with each other, which leads to two kinds of areas along the interface. Type I is direct contact of fresh metals (AA1050/AA6061 contact in this case), as shown in Fig. 2(a) and (c); type II is original metal surfaces with thin oxide film layers in between. The oxide films, which are brittle, are easy to break and allow further refinement of the materials around the interface. Therefore, the grain size in this kind of interface is finer than that of the other areas. As can be seen in Fig. 2(d), the AA1050/AA6061 bonding interface is characterized by fine grains which can be distinguished from AA1050 easily but not AA6061 as the grain size is closer to that of the AA6061 layer. It is believed that the grain refinement at the interface is caused by oxide film and other contamination. The fraction of high angle grain boundaries around the interfacial area is also higher than that of the AA1050 layer and AA6061 layer, as indicated by the SAD patterns in Fig. 2(e)–(g). Development of the two kinds of interfacial areas is schematically shown in Fig. 3. Assume that there are thin layers of oxide film at the surface of the starting materials and it distributes evenly along the surface. Some contamination would also exist at
L. Su et al. / Materials Science & Engineering A 559 (2013) 345–351
349
Fig. 3. Schematic diagram showing the development of two types of interface morphologies (a) Initial state: surface oxide and contamination influenced layers are shown as straight, (b) Deformation at low cycles: surface brittle layers break up and induce areas with direct contact of fresh metals (type I) and areas of original metals with surface brittle layers in between (type II), (c) Deformation at high cycles: surface brittle layers allow further refinement of the materials around the interface, (d) Type I and II interfacial areas can be seen in optical micrograph.
the surface even though it has been cleaned and wire brushed before ARB deformation. Both the oxide film and contamination would influence the deformation and the morphologies of the interfacial areas. As the oxide films have much lower ductility than aluminum, it is easy to fracture even at low cycles of ARB deformation, shown in Fig. 3(b). The fresh metals where the oxide films fractured appear and closely bond with each other due to the pressure of the rolling deformation and further refine with the increasing ARB cycles. The areas where the oxide films locate between the metal layers will allow further refinement because the oxide films are harder than the metals. Therefore, the grain sizes of the materials around the type II interfacial areas are smaller than that around the type I interface areas and the metal matrix, shown in Fig. 3(c). Besides the microstructures of the interfacial areas shown in Fig. 2 (type I in Fig. 2(a)–(c) and type II in Fig. 2(d)), the two types of interfacial areas can be observed with optical microscope, shown in Fig. 3(d). Note that the thick interfacial areas are deliberately chosen to show the effect but the thicknesses of the type II interfacial areas are generally much smaller than that shown in Fig. 3(d). Tensile properties of the three kinds of ARB processed sheets are shown in Figs. 4 and 5. The error bars in Fig. 4 are smaller than the symbols. The tensile strength increases with increasing number of ARB cycles for all the sheets and the values after 5-cycle ARB are more than two times greater than that of the initial materials. Uniform elongation, on the other hand, decreases substantially after 1-cycle ARB and stayed almost unchanged with further deformation. The uniform elongations of all the three kinds of materials are less than 3%. It is common to be observed for ARB processed materials that necking starts at an early stage of deformation and results in early fracture and low ductility [2,4,8,9]. It can be observed in Fig. 4 that the tensile strength of the AA6061 laminates are always the highest and that of the AA1050 laminates are the lowest, while the tensile strength of the AA1050/AA6061 composites locates in between two monotonic material laminates. Curve (d) of Fig. 4 is calculated from the values of curves (a) and (b) by the rule of mixture s ¼V1050s1050 þV6061s6061, where V1050 and V6061 are the volume fractions of AA1050 and AA6061 as listed in Table 1, s1050 and s6061 are the tensile strength of the AA1050/AA1050 and AA6061/ AA6061 laminates, respectively. It is obvious that the directly measured tensile strength of the composite fits reasonably well with the value obtained by the rule of mixture. Note that in the
Fig. 4. Tensile strength values of ARBed (a) AA6061/AA6061 laminates, (b) AA1050/AA1050 laminates, (c) AA1050/AA6061 laminates and (d) AA1050/ AA6061 laminates calculated by rule of mixture. (0 ARB cycle corresponding to hardness of annealed materials).
present work the initial values for the rule of mixture are the tensile strength values of the laminates processed by ARB from monotonic starting materials. However in Ref. [9], the values used for the rule of mixture were calculated from the average hardness values measured on the two different layers of the composite. The fact that the measured tensile strength values fit with the calculated ones indicates that both the materials AA1050 and AA6061 undergo similar strain path no matter if they are deformed in the monotonic material laminates or in the composites. Fig. 4 shows that the difference between the measured tensile strength value of the AA1050/AA6061 composite and the value from the rule of mixture for the 5-cycle ARB processed material is greater than that of other cycles, which can be attributed to the fact that the hard layers in the composite have less deformation after 5-cycle ARB, as indicated in Table 1. It is expected that the tensile strength of the composite subjected to more than 5-cycle ARB would move away from the value of the mixture rule. Fig. 6 shows the effect of ARB deformation on strain hardening rate. The strain hardening rate is plotted against true strain. The
350
L. Su et al. / Materials Science & Engineering A 559 (2013) 345–351
effect of ARB deformation can be seen by comparing the strain hardening rate of ARB processed samples for different cycles. As can be seen in Fig. 6(a), the strain hardening rate is increasing with the number of ARB cycles. The similar phenomenon has been observed in Ref. [22]. It can be explained by decreasing grain sizes with the increasing number of ARB cycles due to the increasing accumulative strain. The high strain hardening rate in the samples with high accumulated strain is due to the smaller grain sizes in these samples, which makes it easier for dislocations to encounter grain boundaries, and then resist further deformation. Fig. 6(b) shows the strain hardening rate of 5-cycle ARB processed samples with different starting materials along with the initial annealed materials. It can be seen that the initial hardening rate of the ARB processed samples shows a larger value than the corresponding annealed conventional sized samples, which can be explained by the grain size effect above. However, the annealed samples display positive strain hardening values to significant high strains ( 34% for annealed AA1050 and 24% for annealed AA6061), whereas the strain hardening rate of ARB processed samples drops rapidly below zero after only a small plastic strain subjected to the tensile specimens (less than 3% for all the cases). This explains the limited uniform elongation of the ARB processed samples. It also can be seen in Fig. 6(b) that 5-cycle ARB processed AA6061/AA6061 has high strain hardening
Fig. 5. Uniform elongation of ARBed (a) AA6061/AA6061 laminates, (b) AA1050/ AA1050 laminates and (c) AA1050/AA6061 laminates. (0 ARB cycle corresponding to hardness of annealed materials).
rate than that of the 5-cycle ARB processed AA1050/AA1050 and the initial strain hardening rate of the 5-cycle ARB processed AA1050/AA6061 locates between the values of the two monotonic laminates, which is in good correspondence with the tensile strength values. The average microhardness values of the ARB processed samples are shown in Fig. 7. The evolution of hardness values with the number of ARB cycles is similar to that of the tensile strength. The hardness increases rapidly after the first cycle ARB and then the increment between each cycle becomes less significant. After 5-cycle ARB, the hardness values become around two times greater than that of the initial values for both AA6061 and AA1050. The hardness values of AA6061 are always higher than that of AA1050. The tensile test takes the AA1050/AA6061 composite as a complete material, which makes it hard to evaluate the behavior of the two different layers. However, in hardness test, the values of the AA1050 and AA6061 layers of the composite can be measured separately. Curves (b) and (c) in Fig. 7 denote the average hardness values of the AA6061 and AA1050 layers in the composites, respectively. As can be seen in the figure that the average hardness values of AA6061 single material laminates are very close to that of the AA6061 layers in the AA1050/AA6061
Fig. 7. Hardness values of ARBed (a) AA6061/AA6061 laminates, (b) AA6061 layer in AA1050/AA6061 laminates, (c) AA1050 layer in AA1050/AA6061 laminates and (d) AA1050/AA1050 laminates. (0 ARB cycle corresponding to hardness of annealed materials).
Fig. 6. Strain hardening rate of (a) 1, 3 and 5-cycle ARBed AA1050/AA6061 laminates and (b) annealed starting materials and 5-cycle ARBed laminates with different starting materials.
L. Su et al. / Materials Science & Engineering A 559 (2013) 345–351
laminates, so is that of the AA1050 materials. This confirms the conclusion obtained from the tensile strength values that the materials AA1050 and AA6061 undergo similar deformation no matter if they are deformed in the single material laminates or in the composites for up to 5 ARB cycles.
4. Conclusion
351
strain no matter if they are deformed in the composites or in the monotonic material laminates.
Acknowledgment The author would like to thank Mr. Joe Abbott for the help of ARB experiments. Lihong Su and Guanyu Deng would like to thank the Chinese Scholarship Council for financial support.
AA1050/AA6061 composites were successfully processed by the ARB process. The followings can be concluded. References 1. Laminated composite materials with strong bonding between the AA1050 layers and the AA6061 layers were obtained and substantial grain refinement was achieved after ARB processing. The areas around the interface of two adjacent layers were observed and it was found that the microstructure of the bonded interface was quite complex. 2. During the first 3 ARB cycles, the AA1050 layers and the AA6061 layers in the composite deformed in a nearly identical way. Afterwards the AA6061 started to neck and eventually fractured at various locations. Severe shear bands were observed throughout the cross-section of the 5-cycle ARB processed composite. 3. Two types of interfacial morphologies were observed along the AA1050/AA6061 interfaces: Type I is the area with direct contact of fresh metals and type II is the area with original metals with surface brittle layers in between. Type II interfacial morphologies allow further refinement of the microstructure close to the interface. 4. The hardness and tensile strength of the 5-cycle ARB processed composite became more than two times greater than the initial values whereas the ductility dropped significantly. The tensile strengths of the composite predicted by the rule of mixture were comparable to the measured values. The hardness values of AA1050 and AA6061 in the composite were nearly identical to those in the monotonic material laminates, which indicated that both materials accumulated similar
[1] R.Z. Valiev, T.G. Langdon, Prog. Mater. Sci. 51 (2006) 881–981. [2] Y. Saito, N. Tsuji, H. Utsunomiya, T. Sakai, R.G. Hong, Scripta Mater. 39 (1998) 1221–1227. [3] H. Chang, M.Y. Zheng, K. Wu, W.M. Gan, L.B. Tong, H.G. Brokmeier, Mater. Sci. Eng. A 527 (2010) 7176–7183. [4] S.H. Lee, Y. Saito, T. Sakai, H. Utsunomiya, Mater. Sci. Eng. A 325 (2002) 228–235. [5] K. Wu, H. Chang, E. Maawad, W.M. Gan, H.G. Brokmeier, M.Y. Zheng, Mater. Sci. Eng. A 527 (2010) 3073–3078. [6] H.S. Liu, B. Zhang, G.P. Zhang, Scripta Mater. 64 (2011) 13–16. [7] G.H. Min, J.M. Lee, S.B. Kang, H.W. Kim, Mater. Lett. 60 (2006) 3255–3259. [8] R.N. Dehsorkhi, F. Qods, M. Tajally, Mater. Sci. Eng. A 530 (2011) 63–72. [9] D.K. Yang, P. Cizek, P. Hodgson, C. Wen, Scripta Mater. 62 (2010) 321–324. [10] S. Ohsaki, S. Kato, N. Tsuji, T. Ohkubo, K. Hono, Acta Mater. 55 (2007) 2885–2895. [11] F.H. Kavarana, K.S. Ravichandran, S.S. Sahay, Scripta Mater. 42 (2000) 947–954. [12] S. Roy, B.R. Nataraj, S. Suwas, S. Kumar, K. Chattopadhyay, Mater. Des. 36 (2012) 529–539. [13] L. Ghalandari, M.M. Moshksar, J. Alloys Compd. 506 (2010) 172–178. [14] A. Mozaffari, H.D. Manesh, K. Janghorban, J. Alloys Compd. 489 (2010) 103–109. [15] T. Hausol, H.W. Hoppel, M. Goken, J. Mater. Sci. 45 (2010) 4733–4738. [16] L.H. Su, C. Lu, G.Y. Deng, A.K. Tieu, X.D. Sun, Rev. Adv. Mater. Sci., in press. [17] M.Z. Quadir, O. Al-Buhamad, L. Bassman, M. Ferry, Acta Mater. 55 (2007) 5438–5448. [18] I. Topic, H.W. Heinz, M. Goken, Int. J. Mater. Res. 98 (2007) 320–324. [19] M.Z. Quadir, A. Wolz, M. Hoffman, M. Ferry, Scripta Mater. 58 (2008) 959–962. [20] C. Lu, K. Tieu, D. Wexler, J. Mater. Process. Technol. 209 (2009) 4830–4834. [21] R. Jamaati, M.R. Toroghinejad, Mater. Sci. Eng. A 527 (2010) 4858–4863. [22] C. Kwan, Z.R. Wang, S.B. Kang, Mater. Sci. Eng. A 480 (2008) 148–159.