3.1.1 Initial microstructural analysis

The initial analyses of the microstructure of the AlSi10Mg specimens were performed on the plane perpendicular to the build direction for as-built, T6, and DA specimens, as presented in Fig. 2 (a)(d), (b)(e), and (c)(f), respectively. Fig. 2(a) shows a semi-elliptical melt pool shape (fish scale pattern) along the build direction, attributed to the Gaussian effect of the incident laser beams. In the FE-SEM image, Fig. 2(d) reveals a notably inhomogeneous microstructure, with coarser microstructure at the molten pool boundaries compared to the interior regions. Also, the Si precipitates are located at the cellular structure boundary formed inside the molten pool. After T6 heat treatment, the molten pools and laser traces become almost invisible in the transverse direction, as illustrated in Fig. 2(b) and (e).

During solution heat treatment, significant homogeneity is achieved as the segregated Si around α-Al grains diffuses to form particles that are homogeneously dispersed in the α-Al matrix, as shown in Fig. 2(e). Microstructural homogenization, elimination of molten pools, and the scattered distribution of silicon particles with the formation of an Mg-containing intermetallic phase (Mg₂Si) with a rod/needle-shaped morphology were observed after T6 heat treatment. The DA process indicates that the microstructure of the molten pool with the cellular structure was retained even under the heat treatment conditions applied to the DA alloy, as shown in Fig. 2(c) and (f). Notably, a large amount of fine Si precipitate was generated inside the DA alloy's cellular structure, which is absent in the as-built alloy.

Fig. 2. Optical and FE-SEM initial microstructures of undeformed as-built (a) and (d), T6 (b) and (e), DA (c) and (f) specimens, respectively.

3.1.2 Compression behavior at room temperature

Characteristic compressive stress-strain plots for as-built, T6, and DA samples are depicted in Fig. 3. The curves show no sharp yield point; therefore, the yield strength values were determined using a 0.2% strain offset procedure, as recommended in the ASTM standard E8M. As shown in Fig. 3, the T6 specimen exhibits lower yield strength (YS) and peak strength, as expected. The yield strength values for the DA specimen are 342 MPa, while those for the as-built and T6 specimens are 320 MPa and 183 MPa, respectively. Therefore, the direct-aged specimen has higher compressive strength than the other two specimen conditions.

Fig. 3. Engineering stress-strain compression curves of the as-built, T6 and DA specimens.

3.1.3 Compression test at different elevated temperatures

Fig. 4 (a), (b), and (c) show the true stress-strain curves obtained from the compression tests of as-built, T6, and DA samples conducted at elevated temperatures of 150 ^oC (423K), 200 ^oC (473K), and 250 ^oC (523K), respectively. All three conditions at room temperature indicated that strain hardening predominates; the flow stress increased continuously with significant work hardening. For all compression test conditions, the flow stress increased sharply at the beginning of the plastic deformation stage, which can be attributed to work hardening caused by the increase in dislocation density^[51].

High-temperature compressive flow stress behavior at different temperatures for the selected heat-treated conditions is illustrated in Fig. 5. During high-temperature compression deformation, the flow stress maintains a higher level without significant softening as the strain increases after reaching the peak stress (σ_p). Generally, this softening is caused by either DRV (dynamic recovery) or DRX (dynamic recrystallization). In this investigation, the flow curve shape was restricted, and a dynamic recovery mechanism balanced the strain hardening.

As expected, the high-temperature compressive deformation behavior was sensitive to deformation temperature, i.e., the flow stress decreased with increasing temperature. With direct aging, the compressive strain accumulation was insignificant, and the attainment of dynamic softening was very rapid compared to the other two test conditions. In direct-aged specimens, the existing dislocation density is high, and during high-temperature deformation, the strain causes re-arrangement of the dislocations into a new formation that offers less resistance to deformation. Therefore, dynamic recovery softening occurs very quickly. It may be inferred that the presence of Si particles in direct-aged specimens prevents further deformation or generation of dislocations^[50].

In Fig. 6, the yield strength of compressively deformed specimens at various temperatures is represented as a bar chart. Detailed values of yield strength, strain hardening exponent, and strength coefficient for high-temperature compressively deformed as-built, T6, and DA specimens are displayed in Table 1. Zhang et al^[52]. also reported that the strain hardening coefficient and strength were fundamental mechanical parameters for metallic material performance. The relationship between the strength coefficient (K) and temperature (^oC), as well as strain rate, is similar to that of the strain hardening exponent (n). The strength coefficient decreases with temperature and increases with strain rate. This investigation indicates that the strain hardening exponent and strength coefficient decrease with increasing temperature due to dislocation annihilation and the growth of dynamic recovery.

Fig. 7 illustrates the microhardness values for undeformed and compressively deformed as-built, T6, and DA specimens at room and elevated temperatures. The compressively deformed specimens at room temperature show higher hardness values than the undeformed specimens under all three conditions. During the compression test at room temperature, the flow stress-strain curves display significant strain hardening behavior, which correlates with high hardness values. The hardness values increased sequentially up to 150 ^oC due to sustained flow stress and appropriate work hardening behavior, after which they gradually decreased for high-temperature compressively deformed specimens at 200 ^oC and 250 ^oC. Therefore, under high-temperature compression conditions, strain hardening is balanced by dynamic softening behavior (DRV and/or DRX). Overall, the DA specimens were observed to possess higher hardness than both the as-built and T6 specimens.

Fig. 4. Compression stress-strain curves of the as-built, T6 and DA specimens at different elevated temperatures.

Fig. 5. Flow stress-strain curves of different temperatures (a) room temperature, (b) 150 ^oC (423K), (c) 200 ^oC (473K), and (d) 250 ^oC (523K).

Fig. 6. Compressive properties for as-built, T6 and DA specimens at different elevated temperatures.

Fig. 7. Vickers hardness values of the initial and deformed as-built, T6 and DA specimens at different temperatures.

3.1.4 Microstructural evolution in the deformed specimens

3.1.4.1 Deformation behavior of compression tests at room temperature

To understand the influence of compressive deformation on the microstructural variation of the AlSi10Mg alloy, a series of compression test samples were characterized using optical microscopy. Typical optical microstructures of room-temperature compressively deformed specimens for the as-built, T6, and DA conditions are shown in Fig. 8(a), (b), and (c), respectively. In Fig. 8(a), the as-built deformed sample shows a flattened semi-elliptical melt pool shape (fish scale pattern) due to compression loading. After compression, the Si particles, which were crowded in the undeformed T6 specimens, become segregated in the deformed condition due to compressive deformation, as shown in Fig. 8(b). Additionally, the cracks in the deformed samples clearly propagate through the grain boundaries from one Si particle to another. Notably, the optical microstructures of the deformed specimens in transverse view for the as-built and DA samples show a clearly compressed form of semi-elliptical melt pools, as depicted in Fig. 8(a) and (c), respectively. In both conditions, the original grains/boundaries (melt pools) become elongated due to strain, but no visible cracks are observed.

To better understand the microstructural variations during compressive deformation, the as-built, T6, and DA specimens were further compared with undeformed specimens using FE-SEM images, as shown in Fig. 9. Fig. 9 (a)(b)(c) and (d)(e)(f) show the microstructural images of the as-built (a)(d), T6 (b)(e), DA (c)(f) specimens for undeformed and compressive deformed conditions, respectively. As mentioned earlier, in the as-built and DA specimens, the original grains/boundaries (melt pools) become elongated due to strain, but the sub-grains/boundaries remain more or less equiaxed, similar to the undeformed conditions. Notably, the extent of fine Si particles has relatively increased inside the cellular region of the compressively deformed as-built and DA specimens. Due to compression loading, an insignificant fine crack (circled in yellow) was observed on the cellular boundaries, as shown in Fig. 9(d) and (f). At higher magnification, the coarser Si particles in the T6 specimens appear significantly deformed, as indicated in Fig. 9(e). This also suggests the brittle nature of the coarser Si particles. Primarily, the impact of homogenization on changes in morphology (such as coarsening) is known to adversely affect the strength of the material. The higher compressive strength observed in direct aged conditions is attributed to the formation of fine-sized Si precipitates in the cellular structure.

Fig. 8. Typical optical microstructures of compressive deformed specimens at room temperature for as-built (a), T6 (b), and DA (c) conditions.

Fig. 9. FE-SEM images of (a)(d) as-built, (b)(e) T6, and (c)(f) DA specimens for: (a)(b)(c) undeformed and (d)(e)(f) compressive deformed conditions.

3.1.4.2 Deformation behavior of compression tests at different elevated temperatures

To understand the influence of temperature on compressive deformation and microstructural variation in the investigated specimens, a series of compression tests were performed at various temperatures using optical and FE-SEM analysis. Typical optical and FESEM microstructures of the deformed as-built specimens at different temperatures, 150 ^oC (423 K), 200 ^oC (473 K), and 250 ^oC (523 K), are shown in Fig. 10(a)(d), (b)(e), and (c)(f), respectively. Fig. 8(a) shows the micrograph of the as-built sample tested at room temperature, where deformed fine melt pool boundaries are observed. Fig. 10(a-c) shows the microstructures of the samples tested at elevated temperatures of 150 ^oC, 200 ^oC, and 250 ^oC, respectively. At these elevated temperatures, the microstructure did not show any significant changes compared to the room temperature sample. However, notable microstructural variations were observed at higher magnification. Fig. 10(d-f) shows the FE-SEM microstructures of the deformed as-built samples at 150 ^oC, 200 ^oC, and 250 ^oC, respectively. The Si particles became coarser as the deformation temperature increased from 150 ^oC to 250 ^oC, and this agglomeration is believed to be responsible for the reduced strength of the heat-treated samples.

Optical and FE-SEM microstructures of the deformed T6 specimens at different temperatures, 150 ^oC (423 K), 200 ^oC (473 K), and 250 ^oC (523 K), are shown in Fig. 11(a)(d), (b)(e), and (c)(f), respectively. High-temperature compressive deformation of T6 samples, near the recrystallization temperature, caused significant microstructural changes as shown in Fig. 11(f). Here, the precipitation of Si particles occurred and coalesced into large Si-rich agglomerates (circled in red). At elevated temperatures, the T6 specimens showed a significant reduction in flow stress and enhanced elongation. It has been reported that reduced solid-solution hardening, due to the precipitation and coarsening of Si particles with increasing temperature, leads to reduced strength.

Optical and FE-SEM microstructures of the deformed direct aged specimens at different temperatures, 150 ^oC (423 K), 200 ^oC (473 K), and 250 ^oC (523 K), are shown in Fig. 12(a)(d), (b)(e), and (c)(f), respectively. Ma et al.^[53] suggested that a significant amount of Si exists in the Al phase, and provides a driving force for the growth of more Si particles. Fig. 12(d) shows the fine cellular structure of the DA samples at low temperature, consisting of a fine dispersion of Si particles along the cell boundaries and the Al-Si matrix. In Fig. 12(e) and (f), as the temperature increases, the fine Si particles become coarser and the cellular boundaries expand. This expansion of grain boundaries is more clearly visible, and the equiaxed cellular structure remains stable, unlike in the as-built conditions. The direct-aged specimens exhibit higher compressive strength than the as-built specimens because of the fine Si particles within the fine equiaxed cellular structure.

Illustrative ECCI images were taken to understand the substructural variation during compressive deformation at high temperatures. Fig. 13 shows the observed dislocation variations in the as-built and heat-treated specimens (T6 and DA) subjected to compression tests at room temperature and 250 ^oC. Fig. 13(a) shows the distribution of fine Si particles within the thin cellular region subjected to compressive deformation of the as-built specimen at room temperature. Moreover, Fig. 13(d) clearly shows that the denser dislocation pattern (marked in yellow) and the wider cellular boundaries (marked in red) in the high-temperature deformed specimens are more pronounced compared to those in the room temperature specimens. As mentioned in Section 3.3, in the high-temperature T6 specimen, the precipitated nano Si particles have coalesced into large Si-rich agglomerates, as shown in Fig. 13(e), which is not observed in Fig. 13(b).

Fig. 13(c) and (d) show the ECCI results for the high temperature deformed DA specimen, that confirm the presence of coarser Si particles and expanded cellular boundaries. In the high-temperature deformed DA specimen, a greater number of dislocations are distributed within the fine grain boundary matrix, and the equiaxed cellular structure remains stable, unlike the as-built conditions. This stability is attributed to the DA specimen having higher compressive strength compared to the other two specimens.

Fig. 10. Optical (a, b c) and FE-SEM (d, e, f) microstructures of compressive deformed as-built specimens at (a) (d) 150 ^oC, (b) (e) 200 ^oC, (c) (f) 250 ^oC different temperatures.

Fig. 11. Optical (a, b c) and FE-SEM (d, e, f) microstructures of compressive deformed T6 specimens at (a) (d) 150 ^oC, (b) (e) 200 ^oC, (c) (f) 250 ^oC different temperatures.

Fig. 12. Optical (a, b c) and FE-SEM (d, e, f) microstructures of compressive deformed direct aged specimens at (a) (d) 150 ^oC, (b) (e) 200 ^oC, (c) (f) 250 ^oC different temperatures.

Fig. 13. ECCI images of deformed as-built (a)(d), T6 (b)(e), and DA (c)(f) specimens at room temperature (a) (b) (c) and 250 ^oC (d) (e) (f).

4.1.1 Effect of strain

Fig. 16 illustrates the restoration mechanisms in the as-built, T6, and DA specimens during high-temperature compression. In all three conditions, distinct initial microstructures were observed: as-built—fine equiaxed cellular structure (primary Al-Si eutectic phase), T6—coarser Si particles (primary Si phase), and DA—fine equiaxed cellular structure with sub-grain Si particles (primary Al-Si eutectic phase and secondary Si particles), as shown in Fig. 9. It is well-established that incompatibility arises in the Al-Si matrix as strain increases. Si particles typically lead to a high density of dislocations accumulating around these sites; dislocation motion via climbing and sliding ultimately concentrates stress around these particles^[58].

Three distinct interactions (I, II, and III) between particles and dislocations during the restoration mechanism are indicated in Fig. 16. For the as-built, T6, and DA conditions, dislocations accumulate rapidly during the initial stages of deformation (I). Subsequently, the dislocation density decreases significantly, and stress concentration is generally alleviated by dynamic recovery (DRV), leading to the formation of sub-grains within the deformed grains. Notably, submicron Si particles dispersed in the matrix exert a pinning effect on dislocation motion and grain boundary sliding (II). This pinning effect is crucial to impede both dislocation climb and sliding, and to facilitate the transformation from LAGBs (low angle grain boundaries) to HAGBs (high angle grain boundaries).

Additionally, the DRV/DRX process is effectively suppressed when the pinning force exceeds the driving force for grain boundary growth. At high strain, dislocations rapidly aggregate around irregular primary eutectic phases or secondary submicron Si particles, leading to highly concentrated stress at these sites (III). This concentrated stress makes dynamic recovery more challenging. Consequently, the irregular eutectic phases (cellular boundaries) or submicron particles are readily fragmented along the flow direction. This fragmentation of stress concentration promotes DRX. Given the relatively high stacking fault energy (SFE) of the material, the sub-grain boundaries in deformed Al alloys continuously absorb active dislocations. Ultimately, LAGBs transform into HAGBs, and sub-grains evolve into new DRX grains^[57]. The dynamic restoration mechanism during deformation is significantly influenced by elevated temperatures. The final microstructures at different temperatures (150 ^oC, 200 ^oC, and 250 ^oC) can be schematically illustrated after the restoration process.

Fig. 16. Schematic illustration of DRV mechanisms during the high temperature compressive deformation process: (a) initial microstructure, (b) dislocation aggregated around particles, (c) restoration mechanisms of DRV and (d) microstructure after the deformation process..

4.1.2 Effect of temperature

It has been reported that at lower strain rates the size of sub-grains is strongly influenced by temperature^[58]. As previously mentioned, dislocation density increases significantly, and interactions between dislocations reduce their mobility at room temperature. However, as the temperature rises, the annihilation and nucleation of dislocations become more facile. The sub-grain size for materials deformed at 250 ^oC is larger than those deformed at 150 ^oC and 200 ^oC.

During high-temperature deformation, the material undergoes deformation over an extended period, allowing sufficient time for the arrangement and annihilation of dislocations. This process leads to the formation of well-defined sub-grains, and the increased dislocation mobility at elevated temperatures results in larger sub-grains. Figure 16 provides a schematic representation of the substructural variations with varying temperatures for the as-built, T6, and DA conditions. Following the restoration mechanism at lower temperatures, the general final microstructure for the T6 specimen consists of sub-micron Si particles embedded within partially broken sub-grain boundaries (cellular structure), alongside coarser Si particles. The presence of fine Si particles in the homogeneous Al-Si equiaxed matrix contributes to the material's strength. At 200 ^oC, new sub-grain Si particles are generated in addition to the existing ones, resulting in coarser sub-grain Si particles and Al-Si equiaxed grain boundaries compared to lower temperatures. Notably, the grain boundaries become more fragmented with increasing temperature.

In contrast, in the indirect aged conditions, the sub-grains, and grain boundaries grew coarser without a fragmented equiaxed morphology. This may suggest that the existing fine Si particles in the direct aged specimens offer less resistance to the deformation (restricting further deformation/generation of dislocations). At the high temperature of 250 ^oC, the fine Si particles in the cellular region are agglomerated towards the grain boundaries and transformed into spherical shapes.

In the T6 specimen, the sub-grain nanoscale Si particles are agglomerated on the micro-scale Si particles. Thus, direct aging heat treatment, nucleation, and growth lead to a dense distribution of fine Si particles. However, pre-existing precipitates and finely spaced eutectic Al-Si phase matrix during the high-temperature compression after the deformation supports the growth of existing particles over the nucleation of new ones. Thus, we obtain a coarser distribution of nanoparticles.

However. at high temperature the DA specimen eutectic phase boundaries get coarser, maintaining the proper equiaxed morphology. This is also the reason high strength is retained, compared to the as-built specimen. It may mean that the presence of Si particles does not allow sufficient time for the arrangement and annihilation of dislocations and the fragmentation of cellular boundaries, even at high concentrated stress. The DA samples showed coarsening of existing precipitates and the eutectic cellular boundaries. The cell regions near the eutectic phase boundaries were depleted of particles, and while coarsened, most of the particles dominate the cell centers. These observations imply that large particles (in the cells and the eutectic phase boundary) coarsen at the expense of nano precipitates (Ostwald ripening) within the cell interiors during high-temperature deformation.