Microstructural and mechanical properties of AA1100 aluminum processed by multi-axial incremental forging and shearing

Multi-axial incremental forging and shearing (MAIFS), as a new severe plastic deformation technique, was successfully applied up to eight passes on the workpieces of commercially pure Al (AA1100). The microstructure evolutions and mechanisms of the grain refinement in the billets deformed through various passes of process were studied using the electron backscatter diffraction (EBSD) analysis. Microhardness measurements and tensile tests were carried out to evaluate the mechanical properties and deformation behavior of the material after successive passes of the MAIFS process. Measured microhardness evolution indicated that while the distribution of hardness was non-uniform after odd-numbered passes up to four passes, but thereafter outstanding deformation homogeneity was achieved when the consecutive MAIFS passes were applied. Tensile tests indicated that yield stress and ultimate tensile strength increased rapidly during the primary pass of process but thereafter there was only a minor increase up to four passes. After that, a little drop could be observed in strength and then it reached to a saturated magnitude. Measured microhardness distribution values exhibited the same trend, viz. it increased through successive passes to a limiting value beyond which it showed a minor decline by disappearance of points having maximum hardness. Some coarsening was taken place and the dislocation walls between the boundaries were reduced significantly in going from four to six passes. It was suggested that the absorption of the dislocations into grain boundaries as an effective recovery process under large deformations and short-range migration of grain boundaries might be significant mechanisms responsible for the softening observed after four passes of process.

Materials Science & Engineering A 639 (2015) 705–716 Contents lists available at ScienceDirect Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea Microstructural and mechanical properties of AA1100 aluminum processed by multi-axial incremental forging and shearing M. Montazeri-Pour a,c, M.H. Parsa a,b,c,n, H.R. Jafarian d, S. Taieban a a School of Metallurgy and Materials Engineering, College of Engineering, University of Tehran, P.O. Box 11155-4563, Tehran, Iran Center of Excellence for High Performance Materials, School of Metallurgy and Materials Engineering, University of Tehran, Tehran, Iran c Advanced Metalforming and Thermomechanical Processing Laboratory, School of Metallurgy and Materials Engineering, University of Tehran, Tehran, Iran d School of Metallurgy and Materials Engineering, Iran University of Science and Technology (IUST), Narmak, Tehran, Iran b art ic l e i nf o a b s t r a c t Article history: Received 14 April 2015 Received in revised form 18 May 2015 Accepted 21 May 2015 Available online 23 May 2015 Multi-axial incremental forging and shearing (MAIFS), as a new severe plastic deformation technique, was successfully applied up to eight passes on the workpieces of commercially pure Al (AA1100). The microstructure evolutions and mechanisms of the grain reﬁnement in the billets deformed through various passes of process were studied using the electron backscatter diffraction (EBSD) analysis. Microhardness measurements and tensile tests were carried out to evaluate the mechanical properties and deformation behavior of the material after successive passes of the MAIFS process. Measured microhardness evolution indicated that while the distribution of hardness was non-uniform after odd-numbered passes up to four passes, but thereafter outstanding deformation homogeneity was achieved when the consecutive MAIFS passes were applied. Tensile tests indicated that yield stress and ultimate tensile strength increased rapidly during the primary pass of process but thereafter there was only a minor increase up to four passes. After that, a little drop could be observed in strength and then it reached to a saturated magnitude. Measured microhardness distribution values exhibited the same trend, viz. it increased through successive passes to a limiting value beyond which it showed a minor decline by disappearance of points having maximum hardness. Some coarsening was taken place and the dislocation walls between the boundaries were reduced signiﬁcantly in going from four to six passes. It was suggested that the absorption of the dislocations into grain boundaries as an effective recovery process under large deformations and short-range migration of grain boundaries might be signiﬁcant mechanisms responsible for the softening observed after four passes of process. & 2015 Elsevier B.V. All rights reserved. Keywords: Severe plastic deformation Multi-axial incremental forging and shearing AA1100 aluminum Mechanical properties Grain reﬁnement 1. Introduction Severe plastic deformation (SPD) of different metallic parts has emerged as an effective and promising method for the production of bulk ultraﬁne grained (UFG) and nanostructured materials over the last two decades [1–3]. Compared to UFG or nanostructured materials processed by alternative methods of nano-powder compaction so-called “bottom-up” approach, bulk samples grainreﬁned through “top-down” techniques such as SPD are without unwanted disadvantages such as residual porosity or impurity, and thus have potential to be utilized in real industrial applications [4]. The general feature of SPD techniques is the simultaneous imposing a large level of shear deformation on the material in the presence of high hydrostatic pressure without any cross-sectional n Corresponding author at: School of Metallurgy and Materials Engineering, College of Engineering, University of Tehran, P.O. Box 11155-4563, Tehran, Iran. Fax: þ98 2188006076. E-mail address: mhparsa@ut.ac.ir (M.H. Parsa). http://dx.doi.org/10.1016/j.msea.2015.05.066 0921-5093/& 2015 Elsevier B.V. All rights reserved. change [5,6]. Various kinds of SPD processes have been developed to date based on these characteristics. The most well-established existing methods are equal channel angular pressing (ECAP) [6], high pressure torsion (HPT) [7] and multi-directional forging (MDF) [8]. Other common process is accumulative roll-bonding (ARB) which can only be applied to sheet materials [9]. It is now generally recognized that the microstructure and mechanical properties of the SPD-processed materials not only depend on the accumulated strain and the strain homogeneity achieved after SPD, but also signiﬁcantly depend on the homologous temperature (and the strain rate) [10,11]. However, most of the widely applied SPD procedures seem to suffer from a nonuniform distribution of imposed strain within the processed bulk metals, which in turn may lead to an unfavorable microstructural non-uniformity and weak mechanical performance of the products [3]. A practical solution would be the optimization of a SPD method itself, to enable producing a more uniform reﬁned microstructure through less possible passes of the process. In this way, the strategies of changing strain path during subsequent 706 M. Montazeri-Pour et al. / Materials Science & Engineering A 639 (2015) 705–716 passes by deﬁnition of different processing routes [6,12], increasing numbers of passes in materials having either high or low, rather than intermediate, stacking fault energies or through changing deformation temperature relative to the homologous temperature of material [10,13–15], and modiﬁcations by manipulating the design of SPD facilities [16–18] were served as the applicable approaches for improving homogeneity in strain distribution and microstructure throughout the SPD-processed workpieces. Furthermore, to improve the operations of SPD processes and to make them more attractive for structural applications by achieving unique properties in the products obtained by SPD, some additional efforts have been made in recent years to introduce new SPD processing methods [16,19]. Multi-axial incremental forging and shearing (MAIFS) as a new deformation technique of SPD has been recently introduced and the applicability of it for deforming materials to high imposed plastic strains was examined [20]. This process consists of incremental forging (or incremental back extrusion) followed by angular pressing (or lateral extrusion). Therefore, it combines some of the main advantages of equal channel angular pressing (ECAP) and multi-directional forging (MDF) techniques to enhance the degree of deformation and to increase the deformation homogeneity of the workpiece after a few passes. In the incremental forging deformation mode, penetration of hard punches into the workpiece is accompanied by forcing the material to shear at localized regions connected to the penetration channels. It has been shown that this deformation mode can break out the material surface whenever the deformation effectively accumulates ahead of a rod penetrating to metal targets of copper or an aluminum alloy [21], thereby leading to the enhanced propensity for shear banding and subsequent grain reﬁnement especially at the back surfaces [22–24]. In the MAIFS technique, the combination of outstanding capabilities of incremental forging and shearing characteristics associated with the angular pressing [25] may provide an opportunity for enhancing the severity of bulk plastic deformation. So far, numerous experimental works have been documented on the severely deformed metals with FCC, BCC or HCP crystalline structures [3,26]. However, aluminum has been recognized to be an ideal model material for investigating the processes of deformation and grain reﬁnement because of its low-cost, simple FCC structure, high stacking fault energy (about 200 mJ/m2), its Table 1 Chemical composition of the commercial purity aluminum (AA1100) studied (mass%). Fe Si Cu Mg Ti Mn V Al 0.52 0.16 0.11 0.02 0.02 0.01 0.011 Bal. potential for producing high strength lightweight products and the long history of research of this material processed by various SPD techniques [1]. The main aim of the current work is thus to examine the effect of processing by the MAIFS method on the properties variations of 1100 aluminum alloy, where deformation homogeneity resulted from microhardness distribution, the nature of microstructure evolution and mechanical performance under tensile test were taken into consideration. 2. Experimental material and procedures Billets of commercial purity aluminum alloy (AA1100) were used as the experimental material. The chemical composition of this alloy is given in Table 1. As-received material was initially machined into the rectangular-shaped specimen with dimensions of 40 20 20 mm3 and then annealed at 400 °C for 2 h to get a fully recrystallized microstructure. The principle of MAIFS processing procedure is represented in Fig. 1, where the facility contains an innovative setup including three movable punches with similar cross-section area and die. The sample is placed into the cavity of die (Fig. 1a) and then pressed by the ﬁrst longitudinal punch so that the material is backextruded into the gap between the second longitudinal punch and die wall (Fig. 1b). In the second step, the workpiece is forged by penetration of second longitudinal punch into the sample similar to the ﬁrst punch (Fig. 1c). In the last step, a third punch is pushed in transverse direction to achieve the initial dimensions of the workpiece via ﬂowing material through the gap between the die and the longitudinal punches (Fig. 1d). A Molybdenum disulﬁde (MoS2) lubricant was applied to the surfaces of workpiece and tools to minimize friction during the process. Fig. 2 displays the deformed shapes of the specimens after each step. It can be seen that compared to some SPD techniques which require complicated design or expensive facilities, MAIFS method needs no special equipment and it offers a great potential for cost-effective processing of various bulk metallic materials. The directions of action of punches during consecutive passes of process are illustrated in Fig. 3a. This route allows repeating the process on all orthogonal faces in such a manner that initial rectilinear shape and direction of metal block is restored after every three passes. The specimens for the microstructural and mechanical investigations were prepared from the central volume part of the deformed samples. The planes considered in this study are schematically deﬁned in the Fig. 3b for a rectangular sample with height of 40 mm that is removed from the die after each pass. The sample is thus rotated through 90° after each pass and reinserted into the die such that the largest dimension of 40 mm ﬁts in the width of die channel and the smallest dimension of 20 mm Fig. 1. Schematic representation for the principle of multi-axial incremental forging and shearing technique: (a) initial state, (b) after step one, (c) after step two, and (d) after step three. M. Montazeri-Pour et al. / Materials Science & Engineering A 639 (2015) 705–716 becomes the height for the next pressing. MAIFS processing was carried out up to eight passes at room temperature using a hydraulic press of 63 t capacity operating at a ram speed of 2.5 mm/s. Microstructure analysis of the as-annealed condition was performed by optical microscopy using polarized light. The microstructure of the material was revealed by grinding, electropolishing and subsequent chemical etching using a solution of 10% HF and 90% H2O. The EBSD examination was carried out on a Zeiss ΣIGMA|VP FEG-SEM equipped with an Oxford Instruments EBSD and X-Max X-ray detectors, and performed at acceleration voltage of 15 kV, 70° tilt and 0.15–0.2 μm scan steps. In order to accomplish this analysis, good surface preparation is crucial and hence the samples surfaces were mechanically grinded and polished by using abrasive papers with different grit sizes progressively. They were then electropolished to remove surface damages. The solution for the electropolishing process was a mixture of 400 ml ethanol, 70 mL distilled water and 30 ml perchloric acid 60%. The electropolishing was performed at room temperature with a stainless steel sheet as the cathode and an applied voltage of 60 V [27]. The Vickers microhardness (kg/mm2) measurements were conducted using a Buehler Micromet hardness tester under a load of 200 g for dwell time of 10 s. The local results were recorded Fig. 2. Photograph of the shape of AA1100 samples subjected to different steps of MAIFS process. 707 following the rectilinear points with regular distances of 4 mm depicting microhardness values on the cross-sectional middle planes and individual reported values were the average of 3 readings with adequate distances near each location. Thus, a pictorial display for evaluation of the microhardness distribution on the cross sections of sample during consecutive MAIFS passes was exhibited. Tensile properties of the as-annealed and MAIFS-processed materials were evaluated using a SANTAM universal testing machine at room temperature and at an initial strain rate of 0.001 s 1. Dog-bone shaped tensile specimens with the gauge section oriented along the longitudinal direction of each rectangular sample (designated as Z plane in Fig. 3b) were extracted from the center portion of the MAIFS-processed materials by using a wire-cutting electric discharge machine. The gauge length and gauge width of the tensile specimens were 4 mm and 3 mm, respectively. The thickness of the tensile specimens was around 2 mm. Specimens were polished to a mirror ﬁnish before tensile testing. 3. Results and discussions Several valuable investigations have been done to date indicating clearly a correlation between the microhardness and the resulting microstructure in materials processed through numerous SPD processes [26,28,29]. Therefore, to better understanding the deformation uniformity, all local microhardness values measured on the cross sections of samples were plotted in the form of colorcoded contour maps as shown in Figs. 4 and 5. These hardness distributions enable estimating of the microstructural homogeneity in the samples processed by different passes. The as-received billet after annealing and prior to MAIFS process has an average Vickers hardness value (Hv) of about 21 which was independent of the location on the cross-section, thereby conﬁrming the existence of a uniform microstructure throughout the sample. Fig. 4 displays the contour maps on the Z plane for samples processed by various passes of MAIFS. As it can be seen, Fig. 3. (a) Scheme of the route and direction of introducing punches during three consecutive passes of the MAIFS process and (b) schematic illustrations of the considered planes for a rectangular sample with the height of 40 mm which is removed from the die after each pass. 708 M. Montazeri-Pour et al. / Materials Science & Engineering A 639 (2015) 705–716 Fig. 4. The coded maps indicating the distributions of the local Vickers microhardness values on the Z plane after processing by consecutive passes of MAIFS: (a) 1 pass, (b) 2, (c) 3, (d) 4, (e) 5 and (f) 6 passes. microhardness for a single pass of MAIFS (Fig. 4a) is signiﬁcantly higher than that of as-received billet and this increase has occurred over the whole area of the section of the billet. However, the distribution of the microhardness values is non-uniform and there is region in the vicinity of the upper surface where the microhardness values are markedly lower than average. The microhardness data shown in Fig. 4a suggest possibility of dividing a specimen processed by single pass MAIFS into two signiﬁcant regions. First region (indicated as A) denotes the area experiencing a high degree of deformation, thereby showing higher microhardness and it is the wide zones around the bottom part of the sample. On the other hand, second region (shown as B) is the area experiencing a relatively lower degree of plastic deformation during MAIFS, thus it contains the zones near top part of the sample. Thus, it can be inferred that initiation of straining from top part of sample, as implied in Fig. 3a, may provide an opportunity to modify the shear planes and shear directions and thereby help to enhance the deformation homogeneity of specimen which is deformed by MAIFS process. Therefore, the improvement inhomogeneity after second pass is achieved by a signiﬁcant reduction in the extent of the area at the upper surface of specimen processed by ﬁrst pass where the hardness values were lower than average. This is consistent with previous model experiments by utilizing a tin-based alloy where it was demonstrated that evolution of deformation homogeneity can be obtained after two passes of the MAIFS process [20]. The distribution after third pass indicated that the extent of hardness declines with increasing distance from the bottom part of sample similar to trend achieved after pass one. The mean magnitude of hardness achieved during the deformation by this pass (Fig. 4c) is about 62 along the bottom regions as compared to the top regions of sample with mean Vickers microhardness value of about 53. This region of low hardness is completely disappeared after four passes (Fig. 4d) to give a reasonably homogeneous microstructure. The spots with Vickers hardness values of about 70 are visible after four passes but they have been absent during processing sample through ﬁve and six passes (Fig. 4e and f). Thus, the evolution of hardness uniformity after pass 5 and 6 is resulted from removing of spots of inhomogeneity where the measured hardness is primarily higher than average. The ECAP-processed pure aluminum [30] and 6061 Al [31] also showed a similar trend of decreasing hardness after a maximum point. Fig. 5 shows the inﬂuence of the number of passes on hardness evolution on the Y plane (across the width of the samples) in experimentally processed alloy. The map shown in Fig. 5a demonstrates that there are a few differences in the Vickers microhardness values from left to right part of sample. These observations indicate that the homogeneity of material deformation patterns in M. Montazeri-Pour et al. / Materials Science & Engineering A 639 (2015) 705–716 709 Fig. 5. The coded maps showing the evolution of the Vickers microhardness values on the Y plane during four consecutive passes of MAIFS process: (a) 1 pass, (b) 2, (c) 3, (d) 4, (e) 5 and (f) 6 passes. transversal center section of samples processed by different passes is better than those seen along longitudinal section (Z plane). Considering hardness data on the Y plane presented here, it can be said that difference in the values of hardness tends to become relatively small after larger numbers of passes because of the overall increase in the microhardness level, similar to the observations made on the Z plane. The overall observed hardness trends point out that the more deformation homogeneity is reasonably achievable during two passes of MAIFS process up to pass four. Hence, it provides evidence that the deformation homogeneity of process may decrease with the deformation in odd-numbered passes but it can increase in even-numbered ones during these early passes. Fig. 6 illustrates an optical micrograph showing the 710 M. Montazeri-Pour et al. / Materials Science & Engineering A 639 (2015) 705–716 Fig. 6. Microstructure of the AA1100 alloy specimen in the as-annealed state. microstructure of the annealed AA1100 alloy before MAIFS process. As is seen, it consists of grains with an average size of about 46 μm. The microstructural characterization by the orientation imaging microscopy (OIM) maps and the corresponding color code key obtained from the EBSD analysis for the samples processed by different passes are shown in Fig. 7. These crystallographic orientation maps were taken from the center portion on the Z plane (Fig. 3b) for the samples subjected to 1, 2, 3, 4, 6 and 8 passes. Complementary statistical information on the grain size distributions taken from the EBSD analysis are given in Fig. 8 as a function of the applied pass. In general, a substructure that consists of cells and subgrains was seen to evolve after pass one (Fig. 7a), as is expected for the case of FCC metals with high SFE. The grain structure is characterized by a substantial inhomogeneity together with regions including some grains which coincide in size with subgrains. However, it is seen from the corresponding OIM map of pass one that most grains form bands which are elongated along the direction of material ﬂow. It was shown that these parallel bands are usually slab-like and also usually extend in the direction of shear on the shear plane [32,33]. The shearing characteristics related to the MAIFS process are somewhat discussed elsewhere [20]. Thus, grains with sizes that are several times greater than the sizes of subgrains are observed, which indicates that the size distribution is substantially wide. This distribution shows that a single pass is not enough to produce a uniform strain in which all of the grains in the material can be reﬁned uniformly. In materials with high SFE such as Al, the rearrangement and annihilation of dislocations take place easily which generally leads to the formation of subgrains [34]. Most of the substructures are separated by low angle boundaries. Therefore, the substantial increasing of hardness values at the bottom center of the sample after one pass can be primarily associated with the development of substructure containing subgrains. However, the microstructure reveals some partially developed and relatively more equiaxed grains lying within or in the vicinity of the elongated grains. The corresponding curve of grain size distribution (Fig. 8a) exhibits a “tail end” at the smaller grains, but the fraction of the area occupied by different grains is maximized for coarser grains with size of about 9.7 μm. It is worth mentioning that the deformation shear component plays the main role in grain reﬁnement during SPD. In fact, the MAIFS process produces the localized shearing at the vicinity of grain boundaries along with different shear bands made by velocity discontinuity occurred during movement of punches [20,35]. These stimulate the preferred dynamic recovery at these strain accumulation inhomogeneties, which in turn may lead to nucleation of grains via bulging mechanism as a result of dynamic recrystallization [36]. The material appears to have begun to elongate uniformly after second pass and hence the corresponding microstructure (Fig. 7b) comprises lamellar high-angle boundaries tend to become somewhat more parallel to each other. These are probably aligned parallel to the shearing plane, together with intersecting boundaries which are mainly of low angle. At the same time, some equiaxed grains are developed from the elongated structure with an average size similar to the width of the elongated grains. The most subgrains formed within the grains are also elongated in the shear direction and tend to be aligned with the grains. Fig. 8b demonstrates that the grain size distribution after pass two of process is shifted toward ﬁner grains. Fig. 7c reveals that part of microstructure after third pass consists of new grains separated by high angle boundaries which are probably developed via a gradual increase in misorientation between preceding subgrains. Another part includes bands of preceding elongated (sub)grains that are approximately aligned in the shearing direction. It is apparent that performing of third pass can result in a bimodal grain size distribution, as seen in Fig. 8c. Distribution curve is characterized by narrow maximums in two regions of grain size and thus the microstructure obtained has a mixture of ﬁne grains with size of about 1.8 μm and coarser grains with size of about 6.6 μm. The corresponding OIM map of pass four (Fig. 7d) indicated that a more homogeneous and effective grain fragmentation has been occurred during this pass of process. The multiple shearing along various directions introduced by MAIFS accelerates the formation of new grains in the interiors of the preceding elongated grains, leading to the propagation of a more equiaxed microstructure through the entire volume after fourth pass. As shown in the graph of Fig. 8d, an increase in the pass numbers of MAIFS up to four led to a signiﬁcant increase in the area occupied by ultraﬁne grains with size of 0.76 μm and to an overall decrease in the size of all grains. Hence, a more profound effect of the grain reﬁnement mechanisms on the larger elongated grains is obtained during this pass of process. It was found that the UFG microstructure slightly coarsened and changed into more equiaxed microstructure after pass six, as seen in Fig. 7e. A recovery simultaneously seems to be occurred at grain interior, so that the dislocation substructures are slightly recognized in the microstructure of sixth pass. The corresponding plot of grain size in Fig. 8e conﬁrms that some coarsening has taken place in all grains such that the fraction of the area occupied by grains with size of 1.6 μm is maximized. Fig. 7f shows such a ﬁne-grained equiaxed microstructure will be unstable with respect to the higher imposed strains up to pass eight. Because the strain paths will vary from one pass to another depending on sample relative orientation with respect to punches, shear takes place on intersecting macroscopic shear planes in different directions. Therefore, elongated grains are again observed after this pass similar to the second pass. Further studies on the detailed shearing characteristics of MAIFS process are needed to understand this behavior fully, which are currently underway. According to Fig. 8f, there are a few reductions in grain sizes after eighth pass; however, the maximum point of grain size distribution curve has not been considerably changed. Fig. 9 illustrates the misorientation distributions of boundaries obtained from the EBSD data for the specimens subjected to various passes. The angles represent the misorientations measured across the various boundaries. As shown in the histogram of Fig. 9a, in the case of the sample subjected to a single pass of M. Montazeri-Pour et al. / Materials Science & Engineering A 639 (2015) 705–716 711 Fig. 7. EBSD maps from the central region of samples on the Z plane after processing by various passes of MAIFS: (a) 1 pass, (b) 2, (c) 3, (d) 4, (e) 6 and (f) 8 passes. The triangle denotes the crystallographic orientations. MAIFS, most angles of misorientation across the boundaries are smaller than 15°, as already observed in the feature of EBSD image of Fig. 7a. The distribution data exhibited a peak at 2° to 5°, in agreement with a predominance of subgrains in the microstructure during ﬁrst passes of different SPD processes [7,11,33]. With increasing strain, however, the distribution of misorientation is shifted toward higher angles, such that the distribution pattern of misorientation becomes bimodal after last pass, as seen in Fig. 9b–f. In the case of the specimen subjected to six passages of MAIFS die, boundaries with high angles of misorientation (above 15°) were observed to be maximized, as in Fig. 9e. However, in the case of the specimen processed by eight passes, an increase in the 712 M. Montazeri-Pour et al. / Materials Science & Engineering A 639 (2015) 705–716 Fig. 8. Statistical variations of the grain size distributions for different passes of MAIFS process: (a) 1 pass, (b) 2, (c) 3, (d) 4, (e) 6 and (f) 8 passes. The average grain size for each pass is mentioned in the upper left side of plot. portion of low angles of misorientation was observed, as in Fig. 9f. Thus, this sample has many high angle boundaries with misorientation angles above 40°, but a large fraction of low angle boundaries, especially boundaries with misorientation angles below 5°, is also present. Such a bimodal misorientation distribution is typical for heavily deformed samples [37–39]. From the results presented above, the generation of submicrongrained structure via formation of grain boundaries in experimental aluminum alloy by performing MAIFS process has been veriﬁed. Grain reﬁnement during SPD processes is usually explained either by continuous dynamic recrystallization or by geometric dynamic recrystallization [36,40–45]. The ﬁne grains that appeared after pass six were more equiaxed than the preceding grains and the microstructure consists mainly of grains surrounded by high-angle boundaries. This indicates that a possible mechanism responsible for the formation of this more equiaxed grain morphology can be the collapse of the initial highly deformed lamellar microstructure and subsequent spheroidization and growth by short-range movement of the grain boundaries which is a continuous dynamic recrystallization [36,44,45]. However, the localized migration of grain boundaries is a diffusioncontrolled process that would not be possible for conventional grain sizes at ambient temperature [46]. But the diffusion kinetics can be radically enhanced by reducing the grain size and/or by introducing lattice defects into the material [47], and one of the most striking features of SPD-processed materials is the creation of a hierarchy of fast diffusion pathways via the non-equilibrium grain boundaries [48]. Such boundaries are characterized by excess grain boundary energy, presence of long range elastic stresses, enhanced free volumes and thus unusual properties [3]. This implies that distorted grains achieved after pass four may be unstable, and that equiaxed morphology and subsequent grain growth allowed a return to a more stable condition during straining [36]. It is assumed that these effects can be caused by the higher rate of absorption of lattice dislocations by the non-equilibrium high-angle grain boundaries at the UFG microstructure [49], which leads to the retardation of the processes of formation of new low-angle boundaries and increase in the angle of misorientation at the existing boundaries, as indicated in the misorientation relationship between grains for sample processed by six passes of process. On the other hand, since the ﬂattening of grains seems inevitable after some passes of MAIFS process such as second and eighth pass, a kind of geometric dynamic recrystallization can be regarded as another possible mechanism for the generation of a ﬁne-grained structure during process, as implied in the literature [42,43,50,51]. From the other point of view, as a result of the high strain induced by MAIFS, the dislocation boundaries would evolve into speciﬁc patterns which can be classiﬁed as geometrically necessary boundaries (GNBs) in a long and continuous scale and incidental dislocation boundaries (IDBs) in a small-scale [38,39]. Dynamic recovery (DRV) as a thermal process then will turn GNBs and IDBs to more equilibrium grain boundaries through increasing their misorientation. In fact, both continuous dynamic recrystallization and probable geometric dynamic recrystallization mechanisms proposed for grain reﬁnement occurred during MAIFS process are forms of a recovery process where the continuous evolution of dislocations is coupled with their interaction with boundaries [43,50]. Thus, it seems that the continuation of the dynamic recovery process at large strains can cause the more absorption of dislocations into grain boundaries and subsequent coarsening of microstructure. Chang et al. [42] reported that M. Montazeri-Pour et al. / Materials Science & Engineering A 639 (2015) 705–716 713 Fig. 9. Histograms of the variations of misorientation angle distributions during different passes of MAIFS process: (a) 1 pass, (b) 2, (c) 3, (d) 4, (e) 6 and (f) 8 passes. dislocation density in the grain interiors of a 1050 aluminum alloy decreases with the increase in strain and most of the grain interiors eventually become free of dislocations after a large deformation. In the present study, the MAIFS processing was conducted at room temperature where the activation of DRV is rationalized considering the relatively low homologous temperature equivalent to 0.32Tm (Tm being the melting temperature) in commercially pure Al [10,52]. In addition, as mentioned, materials processed by SPD processes are thermally unstable due to the huge stored energy in the form of crystalline defects like dislocations and nonequilibrium grain boundaries [53]. Thus, such an increase of lattice defects would also help to enhance the rate of DRV at room temperature by lowering self-diffusion activation energy [54]. The engineering stress–strain curves at room temperature for the as-annealed material and specimens subjected to various passages of MAIFS process are plotted in Fig. 10. A signiﬁcant improvement in the strength of the specimens is observed through the related curves, while the ductility of the specimens processed by MAIFS has been declined. However, the severely deformed specimens keep fairly total tensile elongation of more than 40% up to six passes, but thereafter there is a drop in tensile ductility. On the other hand, after an initial hardening part in curves related to various passes of MAIFS, an abrupt decline in stress is occurred. However, the fall in stress is steep for the curves of ﬁrst and third passes. This steep drop of curves may be attributed to the relatively inhomogeneous distribution of deformation across the related specimens, which in turn leads to development of ﬂow localization easily, as demonstrated by microhardness distribution and the microstructure maps. Fig. 11 summarizes the changes in mechanical characteristics including 0.2% offset yield stress (sy), ultimate tensile strength Fig. 10. The stress–strain curves of the as-annealed and MAIFS-processed samples for various passes. (UTS) and also uniform plastic elongation of different specimens with various grain sizes as a function of applied pass. The high elastic strain related to each sample was accounted and subtracted to obtain the uniform plastic elongation [55,56]. According to Fig. 11a, the yield strength of the as-annealed material with value of 35 MPa is increased by imposing ﬁrst pass of MAIFS up to 108 MPa (an increasing factor of 3.09). One pass MAIFS has also changed the UTS from 56 MPa related to the as-annealed material 714 M. Montazeri-Pour et al. / Materials Science & Engineering A 639 (2015) 705–716 Fig. 11. Variations of mechanical characteristics of the specimens as a function of the applied pass number: (a) yield stress (sY), ultimate tensile strength (UTS), and (b) uniform plastic elongation. to 151 MPa (an increasing factor of 2.7). The yield strength increased to nearly 160 MPa after two passes MAIFS and then improved to about 179 MPa and 189 MPa after three and four passes, respectively. As described, the enhancement in strength is much more pronounced after the ﬁrst pass. The yield stress and UTS reach a maximum after four passes while the uniform ductility decreases from pass two to this pass. From pass four to six, the strength characteristics of the material gradually decrease, whilst the uniform ductility increases. Then after sixth pass, strength exhibits a saturation plateau with number of passes and the uniform ductility of material decreases. The total area under tensile curves as the representative of static toughness [57], was roughly calculated by summing areas of the trapezoids which can cover the whole area of this space via the following formula: Static Toughness = ∫0 ef i=n Sde ≅ ∑ i=1 Si + Si − 1 (ei − ei − 1 ) 2 (1) where n, ei and Si are the number of trapezoids, engineering strain and stress assigned to data points, respectively. This parameter provides guidelines for characterizing the energy absorption during tensile test or maximum strength–ductility product in the specimens which give a view of the overall mechanical performance. The obtained results for the annealed and severely deformed samples were depicted as a function of MAIFS pass number in Fig. 12. As is seen, toughness is increased abruptly following the initial passages through the die such that its highest value has been achieved for the specimen subjected to four passes MAIFS. However, the value related to sample processed by three passes is out of gradual increasing trend during initial passes which is likely to originate from more deformation inhomogeneities developed across the sample after this pass. The calculated values of toughness reached to a relatively saturation state during passes four to six of process, but its value suddenly dropped during the two last passes. In order to evaluate the correlation between the experimental results of tensile tests and the results obtained by microstructural characterizations, the mechanical characteristics of products and the inﬂuences of possible mechanisms on the resultant properties were compared and interpreted in the following. It could be seen that the microstructure of material in severely deformed state resulted from MAIFS would consist of a mixture of high and low angle grain boundaries. Different boundaries existed in the microstructure strengthen the materials either through conventional grain boundary hardening or through dislocation Fig. 12. Approximate static toughness values calculated for the as-annealed state and the AA1100 alloy samples subjected to different passes of MAIFS process. hardening. Low angle boundaries are essentially arrays of dislocations and provide dislocation hardening by well-known Taylor relationship [58,59], whereas higher angle boundaries provide the conventional Hall–Petch strengthening [37]. In aluminum as a cellforming crystalline material which is processed by SPD methods, the major hardening has been estimated to be due to dislocations (and low angle boundaries) rather than grain boundaries [19,58,60]. As was stated, a signiﬁcant reduction in the grain size and a substantial increase in the low angle boundaries were the main microstructural evolutions induced by MAIFS process, which support the improvement of the strength from pass one up to fourth pass. The decreasing rate of increscent in strength during these passes is arisen from enhancing the rate of dynamic recovery and dislocation annihilation which leads to a dynamic equilibrium between the formation and annihilation of dislocations and reaching to a saturation level in the dislocation storage and strength. The decline of uniform ductility from pass two to pass four can be ascribed to the decreasing of hardening capacity of material for homogeneous deformation before the onset of plastic instability during tensile test which is explained in terms of decreasing ability of material to accumulate more dislocations. Nevertheless, as seen in Fig. 11b, it seems that as a consequence of inhomogeneity of deformation and microstructure across the specimens processed by passes of one and three, the occurrence of plastic ﬂow localization can substantially limit the uniform ductility in the deformed zone of material such that measured M. Montazeri-Pour et al. / Materials Science & Engineering A 639 (2015) 705–716 uniform ductility after these passes does not follow the total trend [4,55,61]. From the microstructural point of view, it was demonstrated that the deformed grains became more equiaxed and slightly larger with lower dislocation walls existed between the high angle boundaries due to the operation of recovery mechanisms after sixth pass of MAIFS process. Thus, the increase in the uniform ductility (and decrease in strength after the peak) from pass four to six can be associated with decreasing the boundary volume and the total dislocation walls which cause an increase in the mean free path for movement of dislocations [62,63]. Decreasing uniform and total tensile elongation and subsequently toughness after imposing pass seven may be attributed to the nucleation and coalescence of voids or development of probable cracks under tensile stresses. These defects can be formed due to the lack of ability of material to accommodate the high strains induced by MAIFS process [61,64]. The difference between yield stress and UTS at an identical grain size corresponds with the amount of material hardening during tension [65]. As is seen in Fig. 11a, the difference between these two stresses is minimized in sample processed by four pass. This is due to the fact that at the large imposed strains of fourth pass, the applied aluminum alloy has reached its maximum hardening potential. But thereafter, there is a gradual increase in difference of these two stresses. This suggests that hardening potential of material under uniaxial tension may be enhanced from ﬁfth pass up to later passes. This matter can be conﬁrmed by the uniform ductility variations (Fig. 11b), which show that hardening capacity of material for homogeneous plastic ﬂow from pass four to six is increased. Besides the effect of usual dynamic recovery, by the way, some investigations made suggestions about the effect of heat generation resulted from a high strain rate on the coarsening of the ultraﬁne grains during and after deformation which induces softening [11,41]. Lee et al. claimed that once the generation of the ultraﬁne grain is completed, a gradual extension of the grain boundaries in 1050 Al alloy may occur with the help of both the applied strain energy and the thermal energy induced by a highspeed deformation, resulting in the softening behavior at high strain levels (ε 43) [11]. However, herein, since changes in the microstructures have occurred under the normal deforming speeds of 2.5 mm/s, the temperature rise during MAIFS cannot act as an inﬂuencing important factor for the observed phenomena. Kapoor et al. assigned the softening observed during multiaxial forging of Al using a conﬁned channel die at ambient temperature to the different strengthening contributions of dislocation walls and boundaries. The contribution of dislocation density to strengthening decreases under very large strains as a result of the transformation of the boundary misorientations to high enough values, thereby switching from dislocation strengthening to Hall–Petch strengthening [50]. The present results are also consistent with these theoretical predictions and thus, the softening observed in going from 4 to 6 passes may be related to this effect. However, in this study, it should be pointed out that the fraction of low angle boundaries is again increased in going from 6 to 8 passes while strength is conserved and tensile ductility is decreased, which in turn give an overall reduction in toughness. 4. Conclusions The effect of severe plastic deformation imposed by multi-axial incremental forging and shearing (MAIFS) on an AA1100 Aluminum alloy at room temperature has been investigated. It was conﬁrmed that MAIFS is an effective procedure to provide signiﬁcant grain reﬁnement and a large fraction of high angle grain 715 boundaries. Four passes of MAIFS was sufﬁcient for reﬁning the microstructure of applied alloy into the UFG regime and it reﬁnes the grain size from 46 μm of the original annealed material to a microstructure consisting of dominant grains of around 0.8 μm in size. Based on results achieved from EBSD maps, the microstructure evolved during the early stages of deformation by MAIFS process was more typical of a higher stacking fault energy material, having old grains containing substructural elements. However, some grains which are probably developed due to strain accumulation inhomogenies resulted from velocity discontinuities were also observed following an initial pass. There are many grains elongated along the direction of shear with sizes that several fold exceed the size of subgrains. Very ﬁne roughly equiaxed grains were formed inside the initial elongated grains up to pass four. The grains aspect ratio is progressively reduced up to pass six, and at large strains of pass six, a microstructure with unique features of almost equiaxed grains with a large number of high-angle boundaries therefore evolve relatively homogeneously throughout the microstructure with the operation of dynamic recovery and subsequent dynamic recrystallization mechanisms. To estimate the deformation homogeneity, the microhardness was measured on the cross sections of samples. There is a marked inhomogeneity after a single pass. The improvement in deformation homogeneity with increasing numbers of passes was because of a reduction in the extent of the area at the upper part of sample where the hardness values were lower than average and also to the disappearance of spots of inhomogeneity after four passes, where the measured microhardness was primarily higher than average. The applied route has allowed billet to experience similar deformation modes in different directions with consequent changing of the loading direction through 90° along the three perpendicular axes during three successive passes. Hence, a high imposed strain per pass by more shear planes with different orientations (both in incremental forging and in angular pressing) could progressively result in a more homogeneous grain fragmentation and microhardness evolution during MAIFS process. The trend of yield and ultimate tensile strength obtained from tensile test for specimens processed by different passes showed a transient region of large hardening rate followed by a little strain softening until the stress approaches a saturation value and thus a perfectly plastic behavior with number of passes is achieved through last passes. Acknowledgment The authors thank Dr. Hamed Mirzadeh for useful discussions on the grain reﬁnement mechanisms. References [1] I. Sabirov, M.Yu. Murashkin, R.Z. Valiev, Mater. Sci. Eng. A 560 (2013) 1–24. [2] R.Z. Valiev, I. Sabirov, A.P. Zhilyaev, T.G. Langdon, J. Miner. Met. Mater. Soc. 64 (2012) 1134–1142. [3] R.Z. Valiev, R.K. Islamgaliev, I.V. Alexandrov, Prog. Mater. Sci. 45 (2000) 103–189. [4] M. Dao, L. Lu, R.J. Asaro, J.T.M. De Hosson, E. Ma, Acta Mater. 55 (2007) 4041–4065. [5] M.J. Zehetbauer, H.P. Stüwe, A. Vorhauer, E. Schaﬂer, J. Kohout, Adv. Eng. Mater. 5 (2003) 330–337. [6] R.Z. Valiev, T.G. Langdon, Prog. Mater. Sci. 51 (2006) 881–981. [7] A.P. Zhilyaev, T.G. Langdon, Prog. Mater. Sci. 53 (2008) 893–979. [8] O.R. Valiahmetov, R.M. Galeyev, G.A. Salishchev, Fiz Metall. Metalloved 70 (1990) 204–206. [9] Y. Saito, N. Tsuji, H. Utsunomiya, T. Sakai, R.G. Hong, Scr. Mater. 39 (1998) 1221–1227. [10] K. Edalati, Z. Horita, Mater. Sci. Eng. A 528 (2011) 7514–7523. 716 M. Montazeri-Pour et al. / Materials Science & Engineering A 639 (2015) 705–716 [11] J.-C. Lee, J.-Y. Suh, J.-P. Ahn, Metall. Mater. Trans. A 34A (2003) 625–632. [12] A. Gholinia, P.B. Prangnell, M.V. Markushev, Acta Mater. 48 (2000) 1115–1130. [13] X.H. An, S.D. Wu, Z.F. Zhang, R.B. Figueiredo, N. Gaob, T.G. Langdon, Scr. Mater. 63 (2010) 560–563. [14] C. Xu, M. Furukawa, Z. Horita, T.G. Langdon, Mater. Sci. Eng. A 398 (2005) 66–76. [15] C. Xu, Z. Horita, T.G. Langdon, Acta Mater. 55 (2007) 203–212. [16] 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.-Manuf. Technol. 57 (2008) 716–735. [17] G.J. Raab, R.Z. Valiev, T.C. Lowe, Y.T. Zhu, Mater. Sci. Eng. A 382 (2004) 30–34. [18] R.Z. Valiev, T.G. Langdon, Metall. Mater. Trans. A 42A (2011) 2942–2951. [19] Y. Estrin, A. Vinogradov, Acta Mater. 61 (2013) 782–817. [20] M. Montazeri-Pour, M.H. Parsa, A. Khajezade, H. Mirzadeh, Adv. Eng. Mater. (2015), http://dx.doi.org/10.1002/adem.201400467 (in press). [21] C. Kennedy, L.E. Murr, Mater. Sci. Eng. A 325 (2002) 131–143. [22] T.W. Wright, The Physics and Mathematics of Adiabatic Shear Bands, Cambridge Press, Cambridge, 2002. [23] T. Inoue, S. Torizuka, K. Nagai, Mater. Sci. Technol. 18 (2002) 1007–1015. [24] S.M. Fatemi-Varzaneh, A. Zarei-Hanzaki, S. Izadi, J. Mater. Sci. 46 (2011) 1937–1944. [25] M. Furukawa, Y. Iwahashi, Z. Horita, M. Nemoto, T.G. Langdon, Mater. Sci. Eng. A 257 (1998) 328–332. [26] Y. Huang, T.G. Langdon, Mater. Today 16 (2013) 85–93. [27] G. Petzow, Metallographic Etching, second ed., ASM International, Ohio, 1999. [28] J. Wongsa-Ngam, H. Wen, T.G. Langdon, Mater. Sci. Eng. A 579 (2013) 126–135. [29] S.N. Alhajeri, N. Gao, T.G. Langdon, Mater. Sci. Eng. A 528 (2011) 3833–3840. [30] Y. Iwahashi, Z. Horita, M. Nemoto, T.G. Langdon, Acta Mater. 45 (1997) 4733–4741. [31] W.J. Kim, J.K. Kim, T.Y. Park, S.I. Hong, D.I. Kim, Y.S. Kim, J.D. Lee, Metall. Mater. Trans. A 33A (2002) 3155–3164. [32] Y. Iwahashi, Z. Horita, M. Nemoto, T.G. Langdon, Acta Mater. 46 (1998) 3317–3331. [33] S.D. Terhune, D.L. Swisher, K. Oh-Ishi, Z. Horita, T.G. Langdon, T.R. McNelley, Metall. Mater. Trans. A 33A (2002) 2173–2184. [34] T.G. Langdon, Mater. Sci. Eng. A 462 (2007) 3–11. [35] L. Zaharia, R. Comaneci, R. Chelariu, D. Luca, Mater. Sci. Eng. A 595 (2014) 135–142. [36] F.J. Humphreys, M. Hatherly, Recrystallization and Related Annealing Phenomena, second ed., Pergamon, Oxford, 2004. [37] N. Kamikawa, X. Huang, N. Tsuji, N. Hansen, Acta Mater. 57 (2009) 4198–4208. [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] Q. Liu, X. Huang, D.J. Lloyd, N. Hansen, Acta Mater. 50 (2002) 3789–3802. D.A. Hughes, N. Hansen, Acta Mater. 48 (2000) 2985–3004. A. Belyakov, T. Sakai, H. Miura, K. Tsuzaki, Philos. Mag. A 81 (2001) 2629–2643. N. Tsuji, T. Toyoda, Y. Minamino, Y. Koizumi, T. Yamane, M. Komatsu, M. Kiritani, Mater. Sci. Eng. A 350 (2003) 108–116. C.P. Chang, P.L. Sun, P.W. Kao, Acta Mater. 48 (2000) 3377–3385. M.E. Kassner, S.R. Barrabes, Mater. Sci. Eng. A 410–411 (2005) 152–155. N. Tsuji, Y. Saito, S.-H. Lee, Y. Minamino, Adv. Eng. Mater. 5 (2003) 338–344. F.J. Humphreys, P.B. Prangnell, R. Priestner, Curr. Opin. Solid State Mater. Sci. 5 (2001) 15–21. A. Mishra, B.K. Kad, F. Gregori, M.A. Meyers, Acta Mater. 55 (2007) 13–28. M. Zehetbauer, R. Grössinger, H. Krenn, M. Krystian, R. Pippan, P. Rogl, T. Waitz, R. Würschum, Adv. Eng. Mater. 12 (2010) 692–700. X. Sauvage, G. Wilde, S.V. Divinski, Z. Horita, R.Z. Valiev, Mater. Sci. Eng. A 540 (2012) 1–12. G. Horii, D. Terada, N. Tsuji, J. Phys.: Conf. Ser. 240 (2010) 012114. R. Kapoor, A. Sarkar, R. Yogi, S.K. Shekhawat, I. Samajdar, J.K. Chakravartty, Mater. Sci. Eng. A 560 (2013) 404–412. F.J. Humphreys, Acta Mater. 45 (1997) 4231–4240. C.Y. Yu, P.L. Sun, P.W. Kao, C.P. Chang, Scr. Mater. 52 (2005) 359–363. A. Bachmaier, A. Hohenwartera, R. Pippan, Scr. Mater. 61 (2009) 1016–1019. N.Q. Chinh, T. Csanádi, J. Gubicza, T.G. Langdon, Acta Mater. 58 (2010) 5015–5021. Y.H. Zhao, Y.Z. Guo, Q. Wei, T.D. Topping, A.M. Dangelewicz, Y.T. Zhu, T. G. Langdon, E.J. Lavernia, Mater. Sci. Eng. A 525 (2009) 68–77. Y.H. Zhao, Y.Z. Guo, Q. Wei, A.M. Dangelewicz, C. Xu, Y.T. Zhu, T.G. Langdon, Y. Z. Zhou, E.J. Lavernia, Scr. Mater. 59 (2008) 627–630. D.R. Fang, Q.Q. Duan, N.Q. Zhao, J.J. Li, S.D. Wu, Z.F. Zhang, Mater. Sci. Eng. A 459 (2007) 137–144. J. Gubicza, N.Q. Chinh, T. Csanadi, T.G. Langdon, T. Ungar, Mater. Sci. Eng. A 462 (2007) 86–90. N. Hansen, X. Huang, D.A. Hughes, Mater. Sci. Eng. A 317 (2001) 3–11. Y. Estrin, L.S. Tόth, A. Molinari, Y. Bréchet, Acta Mater. 46 (1998) 5509–5522. M.A. Meyers, A. Mishra, D.J. Benson, Prog. Mater. Sci. 51 (2006) 427–556. F. Dalla Torre, R. Lapovok, J. Sandlin, P.F. Thomson, C.H.J. Davies, E.V. Pereloma, Acta Mater. 52 (2004) 4819–4832. N. Kamikawa, T. Furuhara, J. Mater. Process. Technol. 213 (2013) 1412–1418. V.M. Segal, S. Ferrasse, F. Alford, Mater. Sci. Eng. A 422 (2006) 321–326. N. Tsuji, Y. Ito, Y. Saito, Y. Minamino, Scr. Mater. 47 (2002) 893–899.