2018 Volume 59 Issue 3 Pages 373-379
Microstructural evolution and changes in the mechanical properties of 6000 series aluminum alloys during multi-directional forging (MDFing) and artificial aging were systematically investigated. The strength gradually increased with increasing cumulative strain. The MDFed sample up to a cumulative strain of ∑Δε = 6 showed a yield strength of 252 MPa and an ultimate tensile strength (UTS) of 282 MPa. MDFing evolved a unique deformation texture at a higher cumulative strain region. Artificial aging at 373 K and 393 K after MDFing caused moderate hardening, although softening took place over 423 K without any obvious hardening. The artificially aged sample at 393 K for 100 ks after MDFing exhibited well-balanced mechanical properties of a 288 MPa yield strength and a 313 MPa UTS with an 18.9% plastic strain to failure. The MDFing and subsequent aging succesfully produced a homogeneous ultrafine-grained microstructure with an average (sub)grain size of 220 nm.
This Paper was Originally Published in Japanese in J. JILM 67 (2017) 277–283.
Ultrafine-grained (UFGed) structures produced by severe plastic deformations (SPDs) have attracted significant research interest because grain refinement effectively improves the mechanical properties in metals and alloys1). In aluminum alloys, various SPD processes such as accumulative roll bonding (ARB)2), equal-channel angular pressing (ECAP)3), high-pressure torsion (HPT)4) and high-pressure sliding (HPS)5) have been applied and UFGed structures have actually demonstrated excellent mechanical properties. The author's research group has investigated the mechanisms of grain refinement and strengthening in several kinds of pure metals and alloys by means of multi-directional forging (MDFing)6,7) which is also one of the major SPD methods. This process can be easily applied to bulky products1). As far as the authors know, however, researches on the MDFing of aluminum alloys are quite limited. For instance, Sitdikov et al. have reported that hot MDFing of A7475 alloy at 763 K produced a fine-grained structure with an average grain size of 7 μm8). However, they did not mention the mechanical properties in detail. Rao et al. have carried out MDFing of A6061 alloy at liquid nitrogen temperature, which was to suppress softening by dynamic recovery. They successfully developed a UFGed structure with an average (sub) grain size of 250 nm and achieved an ultimate tensile strength (UTS) of 472 MPa after artificial aging9). However, loss of age-hardenability in the SPDed aluminum alloys is frequently reported10,11). This is because thermal stability was spoiled in the strain-induced UFGed structures. It is pointed out, therefore, that lower aging temperatures compared with conventional ones would be effective to attain age-hardening in the UFGed aluminum alloys12,13).
In the present study, microstructural evolution and changes in the mechanical properties of a 6000 series aluminum alloy during MDFing at room temperature (RT) are systematically investigated. Even while room-temperature MDFing is important for industrial application, no detailed research on microstructural evolution during MDFing of aluminum alloy at RT has yet been carried out. Furthermore, there were quite few studies on the aging behavior after the SPDs of 6000 series aluminum alloys14). Hence, the influence of aging temperature on the mechanical properties after MDFing was also examined.
The chemical composition of the 6000 series aluminum alloy used in the present study is shown in Table 1. A hot-rolled plate of the alloy was cut into rectangular shaped specimens with dimensions of 15.0 × 18.3 × 22.4 mm3 (aspect ratio of 1:1.22:1.49). The specimens were solution heat treated (STed) at 793 K for 3.6 ks in a salt bath and followed by water quenching. The average grain size of the STed sample was 230 µm. Immediately after ST, the samples were MDFed to a cumulative strain of ∑Δε = 16 at maximum, i.e., 40 passes of MDFing, by means of an Amslar-type mechanical testing machine under the following conditions; RT, pass strains of Δε = 0.4 and an initial strain rate of 3.0 × 10−3 s−1. Lubricant (mineral oil based grease) was used on the forging die surface in order to prevent bulging of the specimens. Both the as-STed and as-MDFed (∑Δε = 6.0) samples were isothermally aged in an oil bath at various temperatures of 373 K, 383 K, 423 K and 448 K. To compensate the effect of natural aging possibly occurring during MDFing (approximately 1.8 ks at maximum), the artificial aging treatment of the as-STed samples was started after keeping at RT for 1.8 ks. Hardness test and microstructural observation were carried out on the planes parallel to final forging axis (F.A.). Hardness was measured using a micro-Vickers hardness tester with a load of 2.942 N for 15 s. The arithmetic mean of 8 measured hardness, where the maximum and minimum values were excluded from 10 experimental data, was recorded as the hardness value. Transmission electron microscopy (TEM) and an electron-backscatter-diffraction-pattern technique on a scanning electron microscope (SEM-EBSD) were used to observe microstructure. TEM foils cut from the center of the rectangular-shaped specimens were prepared by mechanical polishing followed by twin-jet electropolishing with a solution of 25% nitric acid and 75% methanol at 7 V and 243 K. Average grain size was measured by the linear line-intercept method from SEM-EBSD maps and TEM images. Tensile test was conducted on an Instron-type mechanical testing machine at RT with an initial strain rate of 3 × 10−3 s−1. The gauge dimensions of the samples for tensile test were of 2 × 4 × 0.5 mm3. At least three samples in each of the conditions were tensile tested along the perpendicular direction to the final F. A. Therefore, arithmetic mean values calculated from more than three experimental results were used for the evaluation of mechanical properties.
| Si | Fe | Cu | Mn | Mg | Cr | Zn | Ti |
|---|---|---|---|---|---|---|---|
| 0.41 | 0.17 | 0.025 | 0.033 | 0.40 | 0.0040 | 0.0044 | 0.015 |
The STed 6000 series aluminum alloy was successfully MDFed to cumulative strains of ∑Δε = 16 at maximum without any cracking. Figure 1(a) shows a series of nominal stress - true cumulative stain curves obtained during MDFing. Rapid work hardening was observed at low cumulative strain region up to ∑Δε = 1.2, and then, the flow stress gradually increased with increasing cumulative strain at medium to high cumulative strain regions. Periodic fluctuation of the flow stress corresponding to the change in the forging axis, i.e., x → y → z, appeared between ∑Δε = 2 and 10. This is probably due to the mechanical anisotropy induced by initial sharp texture which was formed during hot-rolling and still remained even after ST. The details will be described in section 3.3. In Fig. 1(a), clear yielding appeared in each flow curve. The change in the yield stress during MDFing is plotted in Fig. 1(b). The yield stress increased sharply at low cumulative strain region up to ∑Δε = 1.2, and then gradually from medium to high cumulative strain regions. The change in hardness during MDFing displayed in Fig. 1(c) shows similar tendency with those of the flow stress (Fig. 1(a)) and the yield stress (Fig. 1(b)). The hardness at ∑Δε = 16 was 107 HV, which is approximately twice as high as the 47.7 HV of the as-STed sample. Nevertheless, the amount of hardness change between ∑Δε = 6 and 16 was only 7 HV. Therefore, strain hardening during MDFing was almost saturated after ∑Δε = 6.
Changes in mechanical properties during MDFing of 6000 series aluminum alloy; (a) nominal stress vs. cumulative strain, (b) yield stress vs. cumulative strain and (c) Vickers hardness vs. cumulative strain curves, respectively.
Figure 2 shows the microstructures in the samples STed and MDFed to various cumulative strains. Orientation-image mapping by SEM-EBSD was carried out on the cross-sections parallel to the final F. A. of the MDFed samples. A color code parallel to the final F. A. was employed for easy identification of texture evolution, while observation was carried out normal to the F.A. High-angle grain boundaries, where misorientation angle $ \theta \ge 15^{\circ}$, are indicated by black lines in Fig. 2. In the SEM-EBSD maps at ∑Δε = 6 and 16, some areas were unsuccessfully analyzed due to high dislocation density, high strain energy, as well as UFG size finer than the SEM-EBSD resolution. However, the result that most of the areas in Figs. 2(f) and (g) could be successfully analyzed even after MDFing to such ultra-high cumulative strain regions should imply possible occurrence of restoration and recovery.
SEM-EBSD maps of the multi-directionally forged 6000 series aluminum alloys to various cumulative strains. (e), (f), (g) are the imposed images of (b), (c), (d), respectively. Characteristic microstructures (regions 1, 2, 3) are indicated by arrows in (e) only. R.D. and F.A. indicate the rolling direction and forging axis, respectively.
Shear bands that inclined roughly at around 45 degrees to the final F.A. were observed in the sample MDFed to ∑Δε = 2.4 (Fig. 2(e)). Gradual increase in misorientation angles from sub-boundaries to low/high-angle boundaries with increasing cumulative strain was indicated. From the precise analyses of the microstructures, the microstructures evolved during MDFing can be classified into three specific types. Figure 3 illustrates a typical microstructure developed at medium and high cumulative strain regions. One is an area (Region 1) where sub-boundaries and/or low-angle boundaries with misorientation angles less than 15° were developed in grain interiors. Second is the area (Region 2) where initial grains were fragmented by shear bands with a width of approximately 10 μm. The last one is the area (Region 3) where new grains surrounded by high-angle boundaries with a diameter of several micrometers were evolved. At ∑Δε = 6 and 16, the area fraction of Regions 2 and 3 increased. In particular, at ∑Δε = 16, the initial grains completely disappeared and microstructure was entirely composed of newly formed fine grains. TEM observation of the sample MDFed to ∑Δε = 16 revealed homogeneous and equi-axed (sub) grain evolution with an average size of 220 nm (Fig. 4). Low dislocation density within the grain interior of the evolved UFGs suggests occurrence of dynamic recovery during room-temperature MDFing. Actually, in aluminum alloys, occurrence of continuous dynamic recrystallization due to restoration and recovery irrespective of temperature of SPD has been reported15). Since the area of Region 3 evidently increased after the formation of Region 2, it is clear that shear banding as well as continuous dynamic recrystallization promoted the evolution of UFGs. In fact, stimulated UFG evolution at intersections of shear bands during MDFing of Fe-Cr alloy is reported16).
Schematic illustration of the typical microstructures developed in the multi-directionally forged 6000 series aluminum alloy to medium and high cumulative strain regions.
TEM bright field image of the multi-directionally forged 6000 series aluminum alloy to ∑Δε = 16.
Change in the texture during MDFing is shown in Fig. 5. The texture in the as-STed sample was a typical cube one with {001}<100>. The initial texture, however, was destroyed gradually during MDFing. A unique texture, in which {110} planes are oriented normal to the three forging axes, evolved at high cumulative strain regions (Fig. 5(d)). It is known that uniaxial compression texture evolves in FCC metals where a slip direction of <110> is aligned to the compression axis17). Each pass of MDFing can be considered as uniaxial compression. Furthermore, forging direction was changed for 90 degrees pass by pass during MDFing. In this way, sharp {110} texture was increasingly developed at all the forging planes of the rectangular shaped samples (Fig. 6). The highest peak in Fig. 5(d) is recognized as formed by the final pass of forging. Therefore, it is suggested that orientation changes can occur in UFGed structure at every pass even in the high cumulative strain regions. Figure 7 summarizes the change in the area fraction of the three {110} texture components in Fig. 6. Here, the area fraction was calculated using maximum tolerance misorientation of 15 degrees from the ideal orientations. The area fraction of each orientation, i.e., texture component described in Fig. 6 increases monotonically with increasing cumulative strain and in the order of C, B and A. This is because orientation C is difficult to develop by the final forging, in which direction is not normal to {110} planes as seen in Fig. 6(c). The total area fraction of the three texture components reached 77% at ∑Δε = 16.
Pole figures of the multi-directionally forged 6000 series aluminum alloys to various cumulative strains measured by SEM-EBSD. R.D. and F.A. indicate the rolling direction and forging axis, respectively. N.D. is the normal direction of the rolled plane. The axes of the center and the top of the pole figure are parallel to the F.A. and to the previous forging axis, respectively.
(a), (b), (c) Schematic illustration of three ideal texture components which evolved in the multi-directionally forged 6000 series aluminum alloys and (d), (e), (f) the corresponding pole figures. F.A. is the forging axis.
Plots of area fraction of three texture components which corresponds to Fig. 6 as a function of cumulative strain of the multi-directionally forged 6000 series aluminum alloys.
Figure 8 (a) displays typical nominal stress-nominal strain curves obtained by tensile test and Fig. 8(b) summarized the changes in the tensile properties as a function of cumulative strain. The achieved values of the mechanical properties are also listed in Table 2. The yield stress and the UTS obtained in the tensile test increases monotonically with increasing cumulative strain, corresponding with that of forging flow stress (see Fig. 1). UTS of the as-STed sample (150 MPa) increased up to 300 MPa by MDFing to ∑Δε = 16. While the plastic elongation to failure of the as-STed sample was 31%, it decreased with increasing cumulative strain, i.e., 15% at ∑Δε = 2.4 and 11% at ∑Δε = 16. Uniform elongation was less than 1% in all the MDFed samples and the formation of large necking was observed. For example, the area reduction after fracture was 47% in the MDFed sample (∑Δε = 6). This result suggests that most of the plastic elongation was induced by necking. This necking caused the disappearance of the work-hardening region. Necking is known to appear at the transition region from stable to unstable plastic flow. The plastic elongations attained in the present study are, however, larger than those of the other SPDed materials2–4). It could be because of restoration and recovery during MDFing of 6000 series aluminum alloy, which was confirmed by observations using SEM-EBSD (Fig. 2) and TEM (Fig. 10).
Results of tensile tests of the multi-directionally forged 6000 series aluminum alloys; (a) nominal stress - cumulative strain curves and (b) summarized results of the mechanical properties obtained by tensile tests. UTS, YS and EL indicate ultimate tensile strength, yield strength and elongation, respectively.
| Ultimate tensile strength (MPa) | 0.2% yield strength (MPa) | Elongation to failure (%) | Vickers hardness (HV) | |
|---|---|---|---|---|
| As-solid-solution-treated | 145 | 56 | 31.4 | 47.4 |
| MDFed to ∑Δε = 6 | 282 | 252 | 12.0 | 100 |
| MDFed to ∑Δε = 16 | 300 | 280 | 10.5 | 107 |
| Peak aging at 448 K | 236 | 213 | 17.4 | 98.7 |
| MDFing to ∑Δε = 6 + peak aging at 393 K | 313 | 288 | 18.6 | 109 |
TEM bright field images of artificially aged 6000 series aluminum alloys (a) without and (b), (c) with multi-directional forging in advance. (d), (e), (f) are the imposed images of (a), (b), (c), respectively. Some of the fine precipitates are indicated by black arrows.
The change in the hardness of the samples MDFed to ∑Δε = 6.0 during aging at various temperatures from 373 to 448 K is shown in Fig. 9(a). The aging hardening behavior of the STed sample at 448 K is also presented for comparison. In the cases of the MDFed samples aged at 423 K and 448 K, the hardness appeared almost constant until 3.6 ks, and then, rapidly dropped after 3.6 ks. In contrast, at 393 K, a monotonic increase in hardness was observed within the present testing condition. Hardness value of 108 HV was attained at best after aging, which is approximately 1.1 times higher than that of the as-MDFed sample followed by peak aging. At 373 K, no significant difference was observed compared with the result at 393 K. In summary, the age-hardenability of the MDFed samples was quite low in comparison with that of the STed one. Nevertheless, decrease in aging temperature was rather effective to raise in the hardness.
Changes in (a) the Vickers hardness and (b) the tensile behavior of the multi-directionally forged 6000 series aluminum alloys during aging at various temperatures between 373 K and 448 K. ST indicates solid solution treatment.
This tendency corresponds with reports by Hirosawa et al.13), Deschamps et al.18) and Horita et al.10) Hirosawa et al. demonstrated that the aging of HPTed A6022 alloy at 343 K increased the hardness by 10 HV13). Although the suitable condition to attain sufficient age-hardening of UFGed aluminum alloys is still unclear, it is found that aging at relatively lower temperatures can cause moderate hardening also to MDFed aluminum alloys.
Figure 9 shows the typical stress-strain curves obtained by tensile tests of the samples STed and MDFed to ∑Δε = 6.0, and those followed by peak-aging. Aging of the MDFed sample increased yield strength and UTS by 36 MPa and 31 MPa, respectively. Plastic elongation also increased by aging. For instance, that of the sample MDFed to ∑Δε = 6.0 increased from 12% to 18.6% after peak aging. These mechanical properties surpass those of T832 heat-treated A6063 alloy (yield stress 270 MPa, UTS 290 MPa and elongation 12%)19). This likely due to the combined effects of work hardening and UFG refinement by MDFing, recovery and precipitation hardening during artificial aging at relatively low temperature, which will be mentioned in the next section.
The 313 MPa UTS attained in the present study is, however, much lower than the 472 MPa reported by Rao et al.9) They carried out MDFing of A6061 alloy at liquid nitrogen temperature and, then aged. It would be reasonable, however, because of i) the decrease in forging temperature to suppress dynamic recovery and ii) the different amount of elements added to cause different hardening behavior.
3.6 Microstructural change by artificial aging after MDFingFigure 10 shows typical TEM images of the samples aged after ST and MDFing to ∑Δε = 6.0. The aging time was fixed to be 100 ks. In the sample peak-aged at 448 K after ST, common needle-shaped Mg2Si precipitates elongated in the <100> direction of the matrix were observed. In addition, spherical precipitates with a diameter of about 100 nm were seen in the vicinity of grain boundaries20). On the other hand, in the MDFed samples followed by aging at 448 K and 393 K, spherical precipitates with a diameter of 50–100 nm appeared along the dislocations and (sub) grain boundaries. The spherical precipitates are analogous to those observed at the vicinity of grain boundaries in the sample peak-aged at 448 K after ST. The size and number density of the needle-shaped precipitates in the aged samples after MDFing were smaller and lower than those of the sample peak-aged at 448 K after ST. Instead of them, fine spherical precipitates with a diameter of several nanometers appeared along dislocations and (sub)boundaries as indicated by black arrows in Figs. 10(e) and (f). Watanabe et al. reported that a decrease in the number of precipitates in grain interior and precipitation along grain boundaries occurred in the Mg-rich type Al-Mg-Si alloy processed by HPT and subsequent artificial aging21). They concluded that age-hardening in the UFGed alloy was induced mainly due to the precipitates along grain boundaries while the amount of age-hardening was quite low. Okonogi et al. examined the effect of spherical pearlite distribution on the mechanical properties in medium carbon steel. They concluded that grain-boundary precipitation of spherical particles relatively decreased yield stress22). The precipitates on (sub) grain boundaries and dislocations should essentially contribute to age-hardenability. However, the amount of age-hardening by grain-boundary precipitation appears lower than that by precipitates in grain interior. The age-hardenability in the present MDFed samples should be, therefore, also lowed due to the decrease in the needle-shaped precipitates in grain interiors.
In the sample aged at 448 K after MDFing, equi-axed (sub)grains with a diameter of 200–500 nm were observed to evolve (Fig. 10(e)), which suggested an extensive occurrence of recovery. In contrast, in the sample aged at 393 K after MDFing, a rather elongated dislocation substructure indicated insufficient recovery (Fig. 10(f)). Therefore, it is concluded that the degradation of age-hardenability in the UFGed 6000 series aluminum alloy is attributed to i) the restoration and recovery of the dislocation substructure to derive softening and ii) the decrease in the fine needle-shaped precipitates in grain interiors due to inhomogeneous precipitation of fine spherical particles on dislocations and (sub)grain boundaries.
The microstructure and mechanical properties of multi-directional forged (MDFed) 6000 series aluminum alloys were systematically examined. Furthermore, age hardening behavior of the MDFed alloys was also precisely investigated. Coarse initial grains were fragmented by shear banding in addition to a gradual increase in misorientation angles from sub-boundaries to low/high-angle boundaries with increasing cumulative strain. Consequently, an ultrafine-grained (UFGed) structure with an average (sub)grain size of 220 nm was developed at a cumulative strain of ∑Δε = 16, which possessed a yield strength of 280 MPa and an ultimate tensile strength of 300 MPa. A unique texture, in which the three forging planes were parallel to {110}, developed at high cumulative strain regions. By artificial aging of the sample MDFed to ∑Δε = 6.0, the hardness slightly increased at 393 K, while no hardening appeared at 423 K and 448 K. Therefore, age-hardenability of the UFGed aluminum alloys appears when aged at relatively lower temperatures.
The mechanical properties of the MDFed samples were further improved after aging at 393 K. The tensile strength and the plastic elongation increased by 31 MPa and 8.1%, respectively. The precipitation behavior of the MDFed sample evidently differed from those in the coarse grained samples. That is, inhomogeneous precipitation of spherical particles on dislocations and (sub) grain boundaries instead of fine needle-shaped precipitates in grain interior occurred in the MDFed samples. The decrease in age-hardenability in the MDFed sample could be reasonably understood by the following two factors: i) a softening due to restoration and recovery of dislocation substructures and ii) a decrease in the number of fine needle-shaped precipitates in grain interiors due to inhomogeneous precipitation of fine spherical particles on dislocations and (sub)grain boundaries.
The authors acknowledge the financial supports given by Japan Science and Technology Agency (JST) under Industry-Academia Collaborative R&D Program “Heterogeneous Structure Control: Towards Innovative Development of Metallic Structural Materials” and the Light Metals Educational Foundation, Japan.
