3.2. Effect of DRECE on Dislocation Density
Figure 3 shows the X-ray diffraction profiles of 5754 alloy samples in a different condition. As shown, the main reflections for the fcc Al (111), (200), (220), (311), and (400) were detected. In addition, two supplementary peaks at 2θ = 40.21° and 58.11 were identified to be characteristic of the Mg
2Si phase. The diffraction peaks of the as-deformed samples are broadened, and their intensities change throughout the DRECE process. According to Ungar [
17], this can be due to an increased amount of dislocations accumulated during severe plastic deformation and the small size of the diffracting grains Additionally, the peak shift expresses that the DRECE process causes a substantial amount of distortion in the lattice structure of Al5754 alloy. This can be related to long range stresses generated by severe plastic deformation.
The summarized results of the structural parameters obtained from XRD analysis are presented in
Table 2. Throughout DRECE processing, the average domain size parameter decreased from ~35 nm in an initial state to ~29 nm after six DRECE passes. In addition, after DRECE processing, the dislocation density increased from 3.46 × 10
14 m
−2 in the unprocessed state to 6.47 × 10
14 m
−2 after the first pass. Then, the dislocation density showed a steady decrease with an increasing number of passes, indicating the annihilation of dislocations despite increased strain accumulation [
18], reaching a saturation level of 7.55 × 10
14 m
−2 after six DRECE passes. This result is in accordance with the results of other works for similar Al–Mg alloys. Dinkerl [
19] has documented that the dislocation density of AlMg2 alloy was calculated by XRD to be ~5 × 10
14 m
−2~3 × 10
14 m
−2 and 12 ECAP passes, respectively. According to Liu [
20], the dislocation density of an AA5182 alloy after five turns of the HPT process increased to 12.8 × 10
14 m
−2.
3.3. Evolution of Microstructure
Figure 4a–f shows an optical micrograph (longitudinal plane) of 5754 aluminum alloy subjected up to six DRECE passes. As compared with the initial state of the microstructure (
Figure 2), the grain size is not reduced significantly. Areas of different crystallographic orientations inside individual grains are visible. For the micrographs shown, it appears that the increase in the number of DRECE passes is not accompanied by noticeable changes in the microstructure. Throughout subsequent passes, the grains remain equiaxed and almost constant in size.
Figure 5 shows the EBSD colored inverse pole figure (IPF) maps of the 5754 aluminum alloy samples in an initial state and subjected to one, four, or six DRECE passes, respectively. In these microstructures, the grain color/shade corresponds to the individual grain orientation denoted by the unit triangle legend in
Figure 5b. The red lines indicate the locations of high-angle grain boundaries (θ >15°), while green lines indicate the locations of low-angle grain boundaries (3°> θ >15°). The IPF image of the initial state sample (
Figure 5a) demonstrates that the microstructure is composed of equiaxed grains. The measured average intercept length prior deformation is ~7.916 μm.
Figure 5b shows the microstructure after the first DRECE pass. According to the unit triangle legend used, several orientations can be observed even in the grain interiors. This indicates the variation in crystal orientation in the grain interiors at the early stages of deformation. The average intercept length remains almost unchanged: ~7.862 μm.
Figure 5c shows the effect of four DRECE passes. In this condition, the grain interiors are covered with many elongated bands of cells, having a low-angle grain boundary misorientation. These deformation bands are almost parallel to the TD direction. The measured average intercept length decreased to ~6.872 μm.
Figure 5d shows the microstructure of the sample subjected to six DRECE passes. In this state, the microstructure consists of grains covered with deformation bands having a low angle misorientation. This type of microstructure is typical for aluminum alloys subjected to low strains [
21,
22]. The measured average intercept length after six DRECE passes slightly increases to ~6.958 μm.
Detailed extracted values of the fraction of the grain boundaries are presented in
Figure 6 and
Table 3. In the initial state, the fraction of HAGBs is very high ~88.4% and slightly decreases after the first DRECE pass to ~86.6%. Along with the increase in the number of deformation passes, the fraction of HAGBs decreases. After four DRECE passes, approximately 58.6% of the grains have a grain boundary misorientation of more than 15°, indicating HAG (high-angled grains). Finally, the opposite trend is visible. After six DRECE passes, the fraction of HAGBs (high-angle grain boundaries) slightly increases to 59.6%.
3.4. Mechanical Properties after DRECE Process
Figure 7 presents the indentation Vickers microhardness for an initial state and DRECE processed 5754 alloy samples taken across the thickness of the samples. Additionally, a summary of mechanical properties as a function of DRECE passes is given in
Table 4. The observed high hardness gradient between the surface and center of the initial state sample is due to the inhomogeneity of deformation during rolling. The measured mean microhardness value is ~79 Hv. The hardness of the sheet increases throughout DRECE deformation. After the first DRECE pass, this value raises to about ~92 Hv. In addition, in this state, the hardness gradient between the surface and center increases—the hardness values lie in the range of 82–111 Hv. The observed hardness increase is accompanied by a dislocation density growth, as presented in
Table 2. Along with increasing strain accumulation, the hardness values slightly increase, reaching a maximal value of ~99 Hv after six passes. It is worth mentioning that the hardness gradient between the surface and center of the samples decreases throughout DRECE deformation—the hardness values lie in the range of 92–108 Hv, indicating that the non-homogeneous hardness distribution (non-uniform deformation) in the as-rolled initial sample remains in the DRECE processed samples. This may be attributed to the nature of the rolling process, in which the imposed strain is higher near the surface, while a lower amount of deformation is applied at the center.
The mechanical properties of 5754 alloy after the DRECE process were also tested in a tensile test at room temperature, as shown in
Table 4. It is clear that both yield and tensile strength increase throughout DRECE deformation. For the initial state, the mean yield strength value is 112.3, and this value increases rapidly to 184.7 MPa after the first pass. Then, it rises slightly to 198.2 MPa after six passes. This means that the work-hardening rate reduces with the strain increase. However, the observed strength improvement, which was lower than that with the ECAP method [
23,
24], was achieved at the expense of ductility, which decreased from 30.1% to about 13.2% after the first DRECE pass; then with an increase in strain accumulation, it decreased to 8.4% after the six DRECE passes. Such changes in strength and ductility are typical for severely deformed metals, as proved in numerous research papers [
4,
16,
25,
26].
3.5. Annealing Response of the As-Deformed 5745 Alloy
After the DRECE deformation, a thermodynamically unstable microstructure containing lattice defects and subgrain boundaries was produced. This distorted microstructure increased the mechanical properties. To allow further work operation such as shaping, stamping, or forming, the as-deformed alloy must be softened by annealing. The microstructural changes that occur upon annealing are commonly described with reference to recovery processes, the nucleation of new grains, and their growth. The recovery process reduces the stored energy, but it causes only partial mechanical strength reduction, since the dislocation structure is not rearranged into a stable state.
A further microstructural restoration process—recrystallization—may occur in which a new set of defect-free grains nucleate within the as-deformed microstructure. Although the recrystallization removes some amount of stored dislocations, the material still contains a large fraction of grain/subgrain boundaries, which are thermodynamically unstable. Further annealing causes grain growth, which goes along with the formation of new (defect-free) grains in a lower energy configuration.
Light microscopy characterization of the grain microstructure of the 5754 alloy annealed at 150 °C, 180 °C, 200 °C, 250 °C, 300 °C, and 350 °C is shown in
Figure 8. It is apparent that during annealing at 150 °C, the process of recovery dominates. This is accompanied by a substantial dislocation density decrease from 7.55 × 10
14 m
−2 in the as-deformed condition to 4.68 × 10
14 m
−2, as shown in
Table 5. During annealing at 180–300 °C, the entire microstructure undergoes continuous recrystallization followed by grain growth (
Figure 8b–e). The first effect of annealing is visible in the sample annealed at 180 °C (
Figure 8b). In this condition, small recrystallized grains can be observed in the microstructure. The formation of the new grains is accompanied by a continuous dislocation density decrease to 4.44 × 10
14 m
−2. The DRECE processed structure changes completely into coarse-grained and is approximately equiaxed with an increase in annealing temperature, obviously, as illustrated in
Figure 8e. This is followed by a gradual dislocation density decrease to 2.05 × 10
14 m
−2. It is worth mentioning that for the sample annealed at 350 °C, the microstructure exhibit characteristics of both uniform coarsening and, in several places, of discontinuous recrystallization (
Figure 8f). Thus, the obtained microstructure is bimodal, and the formation of the bimodal microstructure is accompanied by a slight dislocation density growth to 3.18 × 10
14 m
−2. This is due to the presence of the new annealed microstructure that gives rise to compressive stress fields to the surrounding small-sized grains, resulting in a relatively lattice distortion and enhancing the dislocation density.
Figure 9 shows the X-ray diffraction profiles of the DRECE processed and annealed 5754 alloy samples. As shown, only the main reflections for the fcc Al (111), (200), (220), (311), and (400) and Mg
2Si (220) and (400) phases were detected. The diffraction peaks are slightly broadened, and their intensities change depending on the annealing temperature. The intensity and width of the peaks in an XRD pattern can be correlated to the level of residual stresses and dislocation density, which changes after annealing.
The evolution of Vicker’s microhardness of the isothermally annealed 5754 alloy samples is listed in
Table 6. The obtained hardness measurements results are in good agreement with the observed microstructure evolution as well as with structural parameters calculated by XRD.
It is evident, as shown in
Table 6, that the hardness decreases rapidly from an initial value of ~99.5 HV after six DRECE pass samples to 81.5 HV after 30 min of isothermal annealing at 150 °C. It is due to the ongoing recovery processes and is accompanied by a rapid dislocation density decrease from 7.55 × 10
14 m
−1 to 4.68 × 10
14 m
−1. In this condition, the changes in the grain size are almost indistinguishable (
Figure 8), and there is no evidence of the recrystallization processes. With an increase in an isothermal annealing temperature up to 180–200 °C, hardness decreases slightly to ~77 HV. This is accompanied by only a slight dislocation density decrease to 4.16 × 10
14. In addition, the microstructure looks quite similar—there is still no evidence of the recrystallization process, see
Figure 8a–c. The first evidence of the recrystallization and grain growth can be observed in the sample isothermally annealed at 250 °C; see
Figure 8d. This phenomenon is accompanied by a dislocation density reduction and hardness decrease to ~61 HV. Annealing at 300 °C causes a more significant grain growth, as shown in
Figure 8e, as well as the greatest dislocation density drop to 2.05 × 10
14. However, despite the observed microstructural changes, the hardness remains almost unchanged. It should be pointed out that when the material is subjected to isothermal annealing at 350 °C, a slower softening rate is observed. In addition, the obtained microstructure is different: bimodal. This occurs because some grains have a much faster growth rate than others. Such a bimodal microstructure is inhomogeneous and is built of recrystallized and recovered grains and volumes with an unchanged distorted thermodynamically metastable structure; therefore, it combines the effects of the larger grains, which cause softening and the ultrafine grains that increase the mechanical properties.
In general, the difference in softening behavior reported in this study as a function of increasing annealing temperature is in line with competitive recovery–recrystallization kinetics, where a lower isothermal annealing temperature results in a gradual hardness decrease and resembled softening due to recovery, while at higher temperatures where recrystallization processes arise quickly, softening is faster.