2017 Volume 58 Issue 12 Pages 1647-1655
In this study, the method of high-pressure sliding (HPS) was applied for grain refinement of an A2024 alloy. Sheet samples with 10–15 mm width and 100 mm length with 1 mm thickness were processed by HPS under pressures of 2–3 GPa. Microstructural observations revealed that the grain size was refined to ～200 nm. Tensile tests showed that the ultimate tensile strength reached 886 MPa with a total elongation of 7.1% and the anisotropy of samples was less developed in the HPS-processed samples. Extra strengthening was attained by aging at 423 K after HPS processing, leading to an ultimate tensile strength of 967 MPa at the peak aged condition. Superplastic elongation of more than 400% appeared when the A2024 alloy processed by HPS for 20 mm was deformed in tension at a testing temperature of 623 K with an initial strain rate of 1.0 × 10−3 s−1.
This Paper was Originally Published in Japanese in J. Japan Inst. Met. Mater. 80 (2016) 593–601. In order to convey the results more precisely, Fig. 3 included the average hardness and the standard deviation for each sample. Miller indices were also added in the selected area electron diffraction patterns in Figs. 4, 10 and 13.
An age-hardenable aluminum alloy, A2024 (called as Super Duralumin), is a T3 heat-treated material widely used as automobile and aircraft components owing to its high tensile strength (480 MPa). Typically, the A2024 alloy is strengthened by a fine dispersion of precipitate particles and solid solution of Cu and Mg. However, there have been reports that the alloy also exhibits great improvements in the strength via grain refinement1–6). Thus, reports are available so that further strengthening is achieved by simultaneous strengthening due to grain refinement and fine precipitation by subsequent aging6). In addition, there have been reports that superplasticity appears in the A2024 alloy by grain refinement7–11).
Severe plastic deformation (SPD) is a well-known process for grain refinement to the submicrometer level12). ECAP (equal-channel angular pressing)13), HPT (high-pressure torsion)14), ARB (accumulative roll bonding)15), and MDF (multichannel directional forging)16) are typical SPD processes; this study is carried out with an SPD process called HPS (high-pressure sliding)17). As shown in Fig. 1, in the HPS process, two samples are placed between upper and lower anvils and a plunger is pushed while high pressure is applied between the two anvils. The intense shear strain introduced in the samples results in ultrafine grains. One advantage of the HPS process is that it can be applied to samples with practical shapes such as sheets and rods; so far, it has been applied to sheets of pure Al (99.99%), Al-Mg alloy, AZ61 Mg alloy, and 7000 series Al alloy18–20), as well as rods of pure Al (99.99%), Al-Mg alloy, AZ61 Mg alloy, 2000 series Al alloy, and 7000 series Al alloy21–23). The capacity of the HPS machine was recently increased to 500 tons24,25), which allows the samples to be scaled up24,25).
Schematic illustration of High-Pressure Sliding (HPS).
Generally, when a rolling process is applied to an aluminum alloy, anisotropy develops such that difference arises in the tensile properties and formability between the direction parallel and perpendicular to the rolling direction. It is anticipated that anisotropy may develop in the A2024 alloy which contains Cu and Mg as major alloying elements, because anisotropy was reported in both Al-Cu26) and Al-Mg systems27–29). In particular, it is more prominent as the Mg inclusion is larger so that the strength is higher in the direction perpendicular to the process27). An earlier prototype HPS machine with a capacity of 50 tons could process samples up to 3–5 mm in width, and hence tensile specimens could be made only in the tensile direction parallel to the longitudinal direction17–20). However, this may be overcome if the HPS machine developed recently is used since the capacity is 10 times higher than the prototype.
In this study, thus, we aim to scale up the sample size by applying the 500 ton capacity HPS machine to an A2024 sheet material and examine the ultrafine structure of the sample in terms of the uniformity and anisotropy of grain refinement as well as the change in mechanical properties according to the strain imposed. We also examine the simultaneous strengthening due to grain refinement and precipitation by aging. In addition, we verify if ultrafine-grained samples produced by the HPS technique show superplasticity when deformed at high temperatures.
HPS samples which were 10 mm and 15 mm in width and 100 mm in length were extracted from an A2024 aluminum alloy sheet (1 mm in thickness) with the composition shown in Table 1. Then, the samples were subjected to solution treatment at 793 K for 1 h followed by water quenching. The average grain size of the solution-treated sample was 62 μm.
| Cu | Mg | Mn | Fe | Si | Ti | Cr | Zn | Al | |
|---|---|---|---|---|---|---|---|---|---|
| A2024 | 4.5 | 1.4 | 0.56 | 0.07 | 0.02 | 0.01 | 0.00 | 0.00 | Balance |
HPS processing was carried out using a 500 ton capacity machine with the anvils, plunger and sample as shown schematically in Fig. 1. The machine was operated at 300 K (room temperature) under the following conditions: applied pressure of P = 2 GPa (for width of 15 mm) and 3 GPa (for width of 10 mm), processing speed of v = 1.0 mm/s, and sliding distance of x = 10 mm or 20 mm. No lubrication was used between the anvils, plunger and samples. The directions of the HPS-processed samples were defined as SD (sliding direction), TD (transverse direction), and ND (normal direction), as indicated in Fig. 2(a). Some of the samples were subjected to aging in an oil bath at 423 K after the HPS processing.
Schematic illustration of (a) sampling location for tensile testing, Vickers microhardness measurement and transmission electron microscopy (TEM), (b) dimensions of tensile specimen, (c) positions for tensile testing.
The HPS-processed samples were subjected to Vickers microhardness measurement, transmission electron microscopy (TEM), and tensile testing. As shown in Fig. 2(a), the hardness measurements were made at the central part of the sample in the longitudinal direction, and the tensile test was carried out in the directions parallel and perpendicular to the longitudinal direction to examine the uniformity of the HPS-processed sample.
The hardness measurement was performed with a Mitutoyo Vickers microhardness testing machine (HM-102). As shown in Fig. 2(a), the measurements were made at intervals of 1 mm on the sample surface, 0.2 mm in the thickness direction of the cross-section, and 0.5 mm in the transverse direction. The load during the hardness measurement was 0.3 kgf for 15 s. The results of the measurements were smoothed and color-mapped using a commercially available software. For aged samples, the hardness was measured at eight different positions and calculated their average and error range.
For TEM observation, the samples were ground with emery papers to ～0.1 mm so that the mid part of the sample in the thickness direction was observed. Thereafter, a thin area for TEM was prepared by a twin-jet electropolishing technique using 20% HNO3-80% CH3OH solution at −20℃, in such a way that the central part of the sample cross-section became thin. TEM observation was performed using a Hitachi-8100 transmission electron microscope at an accelerating voltage of 200 kV. Selected area electron diffraction (SAED) patterns were recorded from an area with 1.3 μm in diameter. Microstructures after aging were examined in the SD-TD, SD-ND, and TD-ND sections.
Tensile specimens were extracted using an electric discharge machine to the dimensions shown in Fig. 2(b) (gauge part: 1.5 mm long, 0.6 mm wide, 0.6 mm thick) in the directions parallel and perpendicular to the longitudinal direction of the sample at the positions given in Fig. 2 (a) and (c) Tensile tests were conducted at 300 K with an initial strain rate of 3.0 × 10−3 s−1 using a 5000 N horizontal tensile testing machine at a constant rate of displacement. Tensile tests were also conducted at 623 K with an initial strain rate in the range of 1.0 × 10−3 s−1–1.0 × 10−2 s−1 in an air atmosphere using the same horizontal type machine but equipped with a heating furnace.
The gauge part after superplastic deformation was observed using a scanning electron microscope (SEM) at an accelerating voltage of 20 kV. A TEM sample for this observation was prepared with a focused ion beam (FIB) system (FEI Quanta 3D 200i) using Ga ion. With this FIB system, it is possible to produce thin areas from specified position so that the TEM observation was feasible at a position 500 μm away from the fractured tip after the tensile deformation.
X-ray diffraction (XRD) analysis was performed on the specimens after superplastic deformation to examine any presence of secondary phase particles. For this analysis, a Rigaku SmartLab was used with the Cu-Kα radiation at a voltage of 40 kV, a current of 30 mA, a scanning speed of 0.2°/min, and a scanning step of 0.01°.
Figure 3 shows the hardness variation on the surfaces and cross-sections of the samples after HPS processing through 10 mm and 20 mm slidings under 2 GPa and through 20 mm sliding under 3 GPa. The color code is normalized by the saturated hardness (equal to 260 Hv)9) achieved by HPT processing on the same alloy under a pressure of 6 GPa and rotation speed of 1 rpm at room temperature. Here, the saturated hardness represents the hardness value at which the hardness no longer changes with further straining. The equivalent strain (εeq) introduced in the sheet sample by the HPS processing was calculated through the following equation17).
| \[\varepsilon _{eq} = \frac{x}{\sqrt{3} t}\] | (1) |
Hardness variations at surface (upper) and cross-section(lower) after HPS processing through (a) 10 mm and (b) 20 mm under 2 GPa, and (c) 20 mm under 3 GPa.
Thus, for Fig. 3(a), (b) and (c), the values of εeq introduced by HPS processing were 5.8, 12.8 and 13.4, respectively. By processing through x = 10 mm, the overall hardness reaches 80% of the saturated hardness and it is uniform throughout the width and cross-section of the sample. The hardness increases as the imposed strain increases (Fig. 3(b)), and reaches a hardness level of 255 Hv through x = 20 mm under 3 GPa, which is almost close to the saturated hardness (Fig. 3(c)). For either of the cases, there appears to be no difference in the average hardness between the surface and the cross-section. This indicates that the HPS processing can introduce intense strain uniformly across its front, center, and rear parts, excluding both ends of the sample. However, a close comparison between Fig. 3(b) and Fig. 3(c) indicates that the sample under the pressure of 3 GPa is higher in hardness than that under 2 GPa despite both having the same sliding distance of x = 20 mm. This is probably because the increased pressure led to more compressive deformation to make the sample thinner and thus, the imposed strain was increased by the difference of the thickness.
Considering the heat generation during the HPS processing, the influence is negligibly small, because the HPT processing operated under the similar constrained condition led to temperature rises of only 6℃, 13℃, and 25℃ for pure aluminum (99.99%), pure copper (99.99%), and pure iron (99.99%), respectively30).
Figure 4 shows TEM microstructures observed on the SD-TD section after HPS processing under 2 GPa. The upper image in each figure is a bright-field image, the lower image is a dark-field image, and the center shows an SAED pattern. The arrows in the SAED patterns indicate diffracted beams used to take the corresponding dark-field images. Since the SAED pattern of the sample processed by 10 mm sliding has a net-like shape as shown in Fig. 4(a), the microstructure consists of subgrains with an average grain size of ～1 μm. When the sliding distance is increased to 20 mm, the average grain size was reduced to ～200 nm as shown in Fig. 4 (b). The SAED pattern exhibits a ring shape, and thus indicates that the microstructure consists of ultrafine grains with high-angle grain boundaries.
TEM bright-field images (upper), dark-field images (lower) and selected area electron diffraction (SAED) patterns (inset) after HPS processing through (a) 10 mm and (b) 20 mm under 2 GPa. Arrows in SAED patterns indicates beams for dark-field imaging.
Figure 5 shows stress-strain curves after tensile testing with an initial strain rate of 3.0 × 10−3 s−1 at room temperature. For comparison, the results after cold-rolling with reduction ratios of 50% (εeq = 0.8) and 70% (εeq = 1.4) are also included along with the solution-treated sample. The ductility was significantly reduced by the cold rolling so that the total elongation decreased to 7.0% for the sample with the 50% reduction ratio and to 6.2% with the 70% reduction ratio. However, when the grains are refined by HPS processing, it is possible to strengthen the alloy while maintaining the total elongation. After HPS processing for 10 mm under 2 GPa, the tensile strength was 693 MPa with the total elongation of 10.6%. The tensile strength further increased as the sliding distance increased to 20 mm. With the HPS processing for 20 mm under 3 GPa, the tensile strength reached 886 MPa while the total elongation of 7.1% was maintained. This signifies that the HPS processing produces a strength at least 1.8 times higher than that after a regular T3 treatment. Thus, the high-strength A2024 aluminum alloy can further increase in the strength by application of HPS processing under higher pressures.
Stress-strain curves obtained at room temperature (R.T.) with initial strain rate of 3.0 × 10−3 s−1 for HPS processed samples including samples after solid-solution treatment and cold-rolling with 50% and 70% thickness reduction.
Figure 6 plots the tensile strength and total elongation after tensile testing of the specimens extracted from three different positions in the transverse direction at the central part along the longitudinal direction. It should be noted that the stress-strain curves delineated in Fig. 5 are those obtained from the center position in the transverse direction. The tensile properties in the transverse direction are almost the same and it turns out to be homogeneous along the transverse direction. The samples processed by HPS for the sliding distance of 20 mm maintains the total elongation of ～10% at all positions in the transverse direction and the tensile strength well more than 800 MPa.
Variations for tensile strength and elongation to failure after tensile testing at three positions along transverse direction.
Figure 7(a) shows stress-strain curves after tensile testing of the sample processed for 20 mm under 3 GPa, where the tensile specimens were extracted as shown in Fig. 2(c) in the directions parallel and perpendicular to the sliding (longitudinal) direction. Figure 7(a) also includes the result in the solution-treated state. The tensile strengths for the parallel and perpendicular directions were 886 MPa and 865 MPa and the total elongations were 7.1% and 5.4%, respectively. Although anisotropy has been reported in a cold-rolled Al-4.6% Cu alloy26), it is little in the HPS-processed sample such that the difference is within 3% even at the high strength level exceeding 860 MPa.
(a) Stress-strain curves obtained at room temperature (R.T.) with initial strain rate of 3.0 × 10−3 s−1 for tensile specimens extracted in longitudinal and transverse directions including samples after solid-solution treatment. (b) Variations for tensile strength and elongation to failure after tensile testing at three positions along longitudinal direction.
Little development of anisotropy after HPS processing is more clearly demonstrated in Fig. 7(b) when the tensile strength and the total elongation are plotted at the front, center, and rear parts along the longitudinal direction of the HPS-processed sample as illustrated in Fig. 2(c). The difference in the tensile strength is as small as ～2% and this is also the case for the total elongation. It is considered that the little anisotropy is attributed to the equiaxed ultrafine-grained structure formed by the HPS processing as confirmed in Fig. 4(b).
3.3 Effect of aging on mechanical propertiesFigure 8 shows a hardness variation with aging time for a sample processed through 20 mm under 3 GPa. The aging was carried out at 423 K for up to 3 h. Hardening occurs with aging and reaches a peak harness of 274 Hv after 35 min. Figure 9 shows results of tensile testing for samples after HPS processing (as-HPS), and subsequently aging for 5 min (under age) and for 35 min (peak age). The values of the tensile strength and total elongation are summarized in Table 2. The tensile strength increases in all specimens by aging treatment irrespective of tensile directions. In the peak-aged condition, the tensile strength reaches 967 MPa with the total elongation of 6.0% in the direction parallel to the HPS processing, and 891 MPa with 10.1% in the perpendicular direction. The anisotropy remained small in the as-HPS condition, but it increased in the peak-aged condition; the tensile strength in the parallel direction was higher by 8.5% than that in the perpendicular direction.
Aging behavior of A2024 after HPS processing under 3 GPa for 20 mm.
Nominal stress vs. elongation for samples deformed in longitudinal and transverse directions after HPS processing and HPS processing followed by aging for 5 min and 35 min.
| Tensile direction | Aging time, τ/min |
Tensile strength, σ/MPa |
Elongation (%) |
|---|---|---|---|
| Longitudinal | As-HPS | 886 | 7.1 |
| 5 | 905 | 8.2 | |
| 35 | 967 | 6.0 | |
| Transverse | As-HPS | 865 | 5.4 |
| 5 | 886 | 7.6 | |
| 35 | 891 | 10.1 |
Close observation shows that yield phenomenon appears in the samples after aging. The yield phenomenon may appear when the grain size becomes very small even for the specimen which does not normally exhibit the yield phenomenon. In fact, this has been reported in pure aluminum when ultrafine-grained structures were produced by ARB31) or ECAP32) processing and subjected to annealing. The possible cause for the advent of the yield phenomenon is because the dislocations generated by both ARB and ECAP processes are eliminated in the grain boundaries during annealing and thus reduces mobile dislocations below the levels required for the initiation of plastic deformation32,33). In this study, the mobile dislocations decreased during the aging at 423 K after the HPS processing and yielding appeared when the dislocations suddenly started to move. This is also explained by the fact that the yield phenomenon appeared more prominently as the aging time is longer.
In order to examine the reason for the anisotropy, TEM was carried out on (a) the SD-TD section, (b) the SD-ND section and (c) the TD-ND section (shown in Fig. 10). It turns out that the grains are equiaxed in the SD-TD section (Fig. 10(a)) with an average grain size of ～130 nm. However, the grains appear to be elongated in the SD-TD section (Fig. 10(b)) with a microstructure as frequently observed in the ARB process31). The grain size was 60 nm wide (in the ND direction) and 280 nm long (in the SD direction). For the ND-TD section (Fig. 10(c)), elongated grains in the TD direction as marked A and equiaxed grains marked B were observed. Hence, rod-shaped grains and pancake-shaped grains are developed in the HPS direction. The anisotropy after aging is therefore attributed to the grains with elongated shapes, and below may be two possible reasons for this.
TEM bright-field images (upper), dark-field images (lower) and selected-area electron diffraction patterns (inset) for (a) SD-TD section, (b) SD-ND section and (c) TD-ND section after aging at 423 K for 35 min.
The first reason is that because the grain shape is elongated, precipitates after aging are also formed with this anisotropy, which then creates a difference in the dislocation movement. Huang et al.34) conducted aging of an ECAP-processed Al-4% Cu alloy having elongated ultrafine-grained structures. Thus, they reported that precipitates nucleated along the grain boundaries had anisotropy. In this study, it was difficult to measure the shape and distribution of the precipitates, but it is probable that the anisotropy in the tensile properties appeared due to the differences in the shape and distribution of the precipitates formed by aging.
The other possible reason is due to an increase in the yield stress caused by a lack of mobile dislocations accompanying heat treatment, which is related to the occurrence of the yield phenomenon as mentioned above. This increase in the yield stress is known as “hardening by annealing” where the yield strength increases with heat treatment even in pure aluminum35). In this study, the grains are elongated in the processing direction, so that the rate of the dislocation generation may be different depending on the direction of deformation, leading to the anisotropy.
It is known that the A2024 alloy exhibits the following precipitation sequence with aging after solution treatment36):
Supersaturated solid solution ⇒ G.P.B zone ⇒ S' ⇒ S (Al2MgCu)
Mohamed et al.6) reported that when aging was attempted at 423 K after HPT processing, the hardness increases while the ultrafine-grained structure remains essentially unchanged because the S phase (Al2MgCu) eventually exists in an over-aged condition. It is reasonable to consider that this increase in the strength should be caused by precipitation of fine particles. The aging temperature used in the present study is the same as that by Mohamed et al.6), and the time to reach the peak age is almost the same as the 30 min as well. Therefore, the increase in the strength due to aging after HPS processing is attributed to fine precipitation of the G.P.B zone or S' phase while maintaining the ultrafine-grained structure as shown in Fig. 10(a). It is then concluded that simultaneous strengthening by grain refinement and fine precipitation was achieved through aging after HPS processing.
There have been two attempts for extra strengthening of the A2024 alloy by aging after SPD processing. Cheng et al.5) reported that, when aging was performed at 373 K with a subgrain structure having grain sizes of 800 nm to 1.5 μm formed after rolling at an extremely low temperature as 123 K, the yield strength increased to 580 MPa. Kim et al.1) reported that aging at 373 K with a subgrain structure having grain sizes of 0.5 to 1.0 μm by ECAP processing gave rise to the tensile strength of 715 MPa. In this study, by contrast, finer grains were produced and the tensile properties are far better than those reported in such earlier ones even in the state immediately after HPS processing, the contribution of grain refinement to the strengthening should be higher than that by other processing methods. The increase in the tensile strength by aging after HPS process were 81 MPa in the sliding direction and 26 MPa in the perpendicular direction. The average value is fairly close to the increase (55 MPa) observed by Kim et al.1) using ECAP processing. The strengthening by aging after SPD processing is thus considered to be equivalent. In the present study, because of the difficulty of measuring the size and distribution of precipitates, it is not possible to evaluate the contribution of the precipitation strengthening, and therefore, it is difficult to evaluate the strengthening by grain refinement using the Hall-Petch relationship.
3.4 High-temperature tensile testsFigure 11 shows stress-strain curves (top) and the appearance of the tensile specimens after deformation to failure (bottom). The tensile tests were conducted at 623 K with an initial strain rate of 1.0 × 10−3 s−1 for the samples after processing under 2 GPa. For high temperature deformation, it is well known that the contribution of grain boundary sliding to the deformation increases when the grains are refined to the submicron level, and this contribution then significantly reduces the flow stress and improves the total elongation to failure. The total elongations after 10 mm sliding were of 270% (W = 2.5 mm), 330% (W = 7.5 mm), and 380% (W = 12.5 mm). Thus, superplasticity was scarcely attained in any specimen extracted along the transverse direction. This suggests the grain refinement was not sufficien for the samples after 10 mm sliding. However, for the samples subjected to 20 mm sliding, superplasticity of up to 430% was attained with almost the same level of total elongations across the transverse direction. The uniform deformation is visible over the gauge length to show that a typical superplastic smooth flow occurred.
Nominal stress vs. elongation (upper) and appearance of tensile specimens (lower) after deformation at 673 K with initial strain rate of 1.0 × 10−3 s−1 for HPS-processed samples through (a) 10 mm and (b) 20 mm including solution-treated samples.
Figure 12 plots the total elongation to failure after high-temperature tensile testing for the specimens extracted from the front, center, and rear along the longitudinal direction of a sample processed by HPS under 3 GPa for the sliding distance of 20 mm. The plots include the results deformed in the directions parallel and perpendicular to the sliding direction. In high temperature deformation, no anisotropy was developed so that no appreciable difference in the total elongation to failure was detected. This is probably because the grain boundary can easily move during the high temperature deformation and, as will be described next, the grains became equiaxed and the formation and growth of stable precipitates (S phase) occurred uniformly.
Variation of elongation to failure after tensile testing at 623 K with initial strain rate of 1.0 × 10−3 s−1 for samples at three positions along longitudinal direction.
Figure 13(a) shows a tensile specimen (W = 7.5 mm) after elongation up to 430% for the HPS-processed sample with x = 20 mm under 2 GPa. Figure 13(b) shows the result of SEM observation near the fracture tip, indicated by ⓑ in (a). Relatively equiaxed grains can be seen on the surface, indicating that the superplastic deformation occurred due to grain boundary sliding37–40).
(a) Appearance of tensile specimen with positions for microstructural observations, (b) SEM micrograph near fracture tip, and TEM bright-field images (upper), dark-field images (lower) and selected-area electron diffraction (SAED) patterns (inset) for (c) gripping part and (d) gauge part of tensile specimen after deformation at 623 K with initial strain rate of 1.0 × 10−3 s−1.
Figure 13(c) and Fig. 13.(d) show TEM micrographs of the gripping part (marked ⓒ) on the SD-TD section and near the fracture tip (marked ⓓ) on the SD-ND section, respectively. Fine secondary phase particles are visible inside the grains and on grain boundaries, as indicated by arrows in the figure. The average grain sizes in the gripping part and deformed part were of 0.8 μm and 1～2 μm, respectively. Dynamic grain growth occurred because of the stress acting on the deformed part during the testing. However, the grain size remains small, which is probably because the fine secondary phase particles inside the grain and on the grain boundaries impeded the grain growth. In order to identify the secondary phase particles, XRD analysis was performed on this sample and the result is shown in Fig. 14. It turns out that the secondary phase particles were identified as Al2CuMg, Al3Fe, and Al20Cu2Mn3. Alhamidi and Horita9) reported that the processing of the same alloy by HPT reduces the grain size to 240 nm, and then annealed at 623 K to detect Al3Fe, Al6Mn, Al2Cu, and Al2CuMg as secondary phase particles. In the present study, however, Al6Mn was not detected; the only identified compound containing Mn was the T phase (Al20Cu2Mn3), which was formed as a dispersoid. As mentioned earlier, the grain size obtained in this study, even after aging treatment, has ultrafine grains with the size of ～130 nm, which is finer than ～240 nm obtained by Alhamidi and Horita9). Therefore, in this study, the T phase had an effective impact on grain refinement by HPS processing. The presence of such secondary-phase particles inhibits significant grain growth during the high-temperature deformation. Thus, the occurrence of superplasticity through grain boundary sliding is substantiated by the observation of equiaxed grains in the deformed part as shown in Fig. 13(d).
XRD profile after tensile deformation at 623 K and 1.0 × 10−3 s−1.
One of the authors (TM) would like to thank Grant-in-Aid for JSPS Research Fellow (No. JP16J07050) and Aluminium research grant program (of Japan Aluminium Association). This work was supported in part by Japan Science and Technology Agency (JST) under Collaborative Research Based on Industrial Demand “Heterogeneous Structure Control: Towards Innovative Development of Metallic Structural Materials”, in part by the Light Metals Educational Foundation of Japan, and in part by Grant-in-Aid for Scientific Research (S) from the MEXT, Japan (No. JP26220909). HPS was carried out in the International Research Center on Giant Straining for Advanced Materials (IRC-GSAM) at Kyushu University.
