Microstructural refinement
The AM-deposited Al–9%Si–3%Cu alloy was effectively deformed using room temperature-HPT in this investigation. The achieved microstructural refinement as obtained by grain size and crystallite size measurements were found to be about 90 and 30 nm, respectively, after 10 turns in HPT, which is ascribed to the influence of the processing temperature on the level of achievable microstructural refinement during severe plastic deformation processes. The current processing was done at room temperature so there was no dynamic recovery, recrystallization or even grain growth that would affect the extent of the grain refinement with the high levels of imposed deformation during HPT.
The severe plastic deformation imposed was higher after a high number of HPT turns, where the grain size and crystallite size were found to be inversely proportional with the preceding deformation. The HPT imposed strain was translated into significant grain refinement of the AM processed alloy from a size of 1 µm to 90 and 30 nm for grain size and crystallite size, respectively. A considerable dislocation density was stored in the alloy, up to 6.2 × 10
14 m
−2, at such levels of imposed strain, while the initial dislocation density of the as-deposited unprocessed alloy was about 1 × 10
14 m
−2 as shown in Fig.
3c.
These findings indicate the importance of room temperature-HPT processing of the investigated alloy, where the hot deformation behaviour has improved significantly as observed from the elongation measurements and strain rate sensitivity values.
The morphology of the AM-deposited alloy has been significantly altered after processing in HPT, where the melt pools of the as-deposited alloy fragmented into a finer nm-sized grain structure. The eutectic silicon continuous network has also fragmented into nanosized particles and their distribution increased as the deformation increased. These nanosized particles were aligned with the direction of torsional deformation and finally showed a fairly uniform distribution within the
α-Al grains. The refinement in the alloy matrix and eutectic phase is attributed directly to the torsional and compressive strains that are imposed during HPT processing, which led to the subdivision of the melt pool into finer structures of nano sizes [
7‐
9,
29‐
31].
Mechanical behaviour at room and elevated temperatures
The stress-elongation curves for the as-deposited and HPT-deformed samples that were tested at ambient and elevated testing temperatures revealed that the ultrafine microstructure achieved after HPT has resulted in large elongations at elevated testing temperature compared to results of testing at ambient temperature, as shown in Figs.
5 and
6. The maximum elongation for the AM processed alloy was 220% at a strain rate of 10
–4 s
−1; which (to the authors’ knowledge) is the highest elongation achieved for such AM build Al–9%Si–3%Cu alloy.
The current results of elongations to failure were remarkably higher than counterparts reported previously for Al–Si–Cu–Mg alloy [
32], where the maximum elongation obtained for the aforementioned alloy was only 12% at a strain rate of 10
–3 s
−1 and a temperature of 573 K. The present results were also higher than the data reported for Al–Si–Cu–Mg alloy with a maximum elongation of 36% [
33] that was obtained using a strain rate of 10
–4 s
−1 and a temperature of 773 K. The current elongation data were also higher than that reported for Al–11%Si alloy [
11] which exhibited an elongation of 34% using a strain rate of 2.3 × 10
–3 s
−1 and a temperature of 573 K.
The present elongations were also higher than reported for Al–11%Si alloy with the previously highest reported elongation of 150% at strain rate of 5 × 10
–4 s
−1 and at a temperature of 788 K [
34]. It is widely reported that Al–Si alloys with ultrafine microstructures can be fabricated at room temperature by severe plastic deformation processes rather than at elevated temperature [
9,
35,
36]. Therefore, these fine microstructures would give higher elongations during forming at high temperatures, where the existence of the fine grains is required for achieving superplastic flow in the polycrystalline materials [
37‐
39].
The higher number of HPT turns has led to finer microstructure of AM-deposited Al–9%Si–3%Cu alloy down to 90 nm for the sample deformed using 10 turns. Thus, the elongations at the elevated temperature of testing (573 K) were increased with increasing number of HPT turns at which the samples were processed in HPT compared to the as-deposited samples as shown in Figs.
5 and
6. The elongations of the deformed samples were lower than in the as-deposited samples when tensile testing was conducted at ambient temperature (298 K) due to the increment in the work hardening that is imposed within the processed samples as the deformation increased with a higher number of HPT turns [
8‐
10].
The ambient temperature-hardenability of the HPT-processed samples was assessed using Vickers microhardness measurements of these samples. It was found that the strength (in terms of hardness measurement) of additively manufactured Al–9%Si–3%Cu alloy has increased from 120 to 240HV after 10 HPT turns. This level of strengthening was associated with a substantial density of dislocations up to 6.2 × 10
14 m
−2 that contributed mainly to a higher level of strain hardening and lower ductility at ambient temperature-tensile testing for all processed samples in comparison with the as-deposited samples as exhibited in Figs.
5 and
6.
The elongation increased with a higher number of HPT turns where finer microstructures have been obtained during the HPT as illustrated in Figs.
6,
7 and
8, where the increase in value of the strain rate sensitivity indicates significant resistance to necking failure and allows remarkable elongations to occur [
40,
41]. The microstructures after the tensile test remained relatively contiguous, especially at slower strain rates (10
–3 and 10
–4 s
−1) as observed along the gauge length regions of the HPT-processed samples after tension as exhibited in Fig.
10, compared to the microstructures of the as-deposited samples at the same strain rate and temperature as seen in Fig.
9.
The migration of grain boundaries during the hot deformation at slower strain rates (10
–3 and 10
–4 s
−1) at temperature of 573 K are seen in Figs.
10 and
11, where the discontinuities at grain boundaries would be associated with lower concentration of stresses. Therefore, the deformation mechanisms that govern the deformation at this stage are creep via glide-dislocation associated with the sliding of grain boundaries. The sensitivity values which were close to 0.3 as shown in Fig.
9 confirms this assumption, where the grains remained relatively equiaxed [
34,
42].
Thermal stability of the microstructure
In the current investigation, the AM-deposited Al–9%Si–3%Cu alloy deformed by HPT at room temperature showed an ultrafine grain structure that exhibited significant thermal stability and plasticity at elevated temperature compared to the behaviour of counterpart alloys such as in Al–Si–Cu–Mg [
32] Al–Si–Cu-Mg cast alloy [
33] Al–11%Si alloy deformed by rotary-die equal channel angular pressing [
11,
34]. This may be ascribed to the effects of grain size and particle size and distribution of the eutectic silicon phase. The current as-deposited unprocessed Al–Si–Cu had grainy melt pool structures of average width of 150 µm, whereas the silicon eutectic phase appeared with a continuous network appearance that agglomerated along the melt pool boundaries as seen in Fig.
1a, c.
Following the HPT processing at room temperature, this morphology has changed significantly where an extensive microstructural refinement has been achieved down to average grain size and crystallite size of 90 and 30 nm, respectively. The eutectic phase has also undergone severe fragmentation into fine particles with an average size of 175 nm with relatively uniform distribution within the alloy matrix with increasing turns of HPT by the virtue of heavily torsional straining during HPT. The fine particles of silicon eutectic phase were distributed homogeneously and appeared as white fine particles as shown in Fig.
2d, for the sample processed in HPT for 10 turns compared to the network morphology of this phase in the as-deposited sample as shown in Fig.
2c.
The existence of eutectic particles with fine sizes and relatively uniform dispersion within the alloy matrix will suppress any rapid grain growth during hot deformation [
38,
43]. Therefore, the ultrafine microstructure has better thermal stability compared to their counterparts with larger grain size during the hot deformation due to activation of different mechanisms of superplasticity that precede the grain growth, resulting in remarkable flow and elongations rather than cavitation failure under the hot deformation conditions [
37,
38].
The melting point of the silicon eutectic phase in the current alloy is 833 K (560 °C) that is somewhat lower than for the alloy itself (873 K, 600 °C) [
1]. Therefore, it is expected that this phase will glide along the grain boundaries at a rate relatively more easily than the grains. The eutectic particles with the lower fraction volume as in the as-deposited alloy were distributed at the grain boundaries and pool boundaries as seen in Fig.
1a, c.
The distribution and volume fraction of eutectic
fine particles has increased significantly as seen in Fig.
1c, d with additional HPT turns in comparison with their counterparts in the as-deposited unprocessed alloy. The localization of these particles on the aforementioned locations added to a pinning effect, where the dislocations accumulated around these nanoparticles and then strengthening of the samples under tension at a temperature of 298 K [
44,
45].
Consequently, the tensile strengths of the additively manufactured Al–9%Si–3%Cu as-deposited and HPT-processed samples at room temperature were significantly better than earlier reports [
11,
14,
15], where the maximum tensile strengths have reached 400 and 700 MPa for the as-deposited and processed samples, respectively. These are considerably higher than counterpart reported values after rotary-die ECAP, e.g. Al–11%Si alloy with 250 MPa [
11], ECAP-Al–10%Si alloy with 234 MPa [
14], and ECAP-Al–7%Si alloy with 250 MPa [
15].
This confirms the uniformity of tensile deformation at room temperature was maintained and assisted by the relatively homogeneous distribution of fine particles of silicon eutectic phase after HPT processing compared to ECAP processing [
44]. The considerably higher values of the tensile strengths of as-deposited samples and processed samples when tested at 298 K compared to reports in [
11,
14,
15], make the combination of HPT processing with additively manufactured Al–Si alloy potentially highly desirable in designing novel alloy processing routes for appropriate industrial applications.
Another factor that plays a significant part in the strengthening of as-deposited samples is the effect of microstructure morphology with regard to the tensile loading direction. The current alloy was built vertically, i.e. along the
z-axis that represents the alloy rod length as schematically shown in [
46], then the HPT disc was cut parallel to the
x–
y plane which is the same orientation as the tensile samples. The elongated grain morphology in terms of melt pool shapes was parallel to the tensile loading, which results in considerably enhanced tensile strength.
The deformation continuity was maintained via melt pools that lie parallel to the loading direction until it reached a point at which the hardening capability was increased by the cross-linking of these pools of different directions. Eventually, cavitation appeared in the as-deposited samples due to the coalescence of lack-of-fusion and gas micropores leading to the failure at relatively lower elongations during tensile testing at room temperature compared to conventional Al–Si alloys [
16].
The nanosized particles enhanced the sliding of grains significantly at a temperature of 573 K for the HPT-deformed samples compared to the as-deposited samples. It has been suggested that the second phase particles act as a lubrication of the grains sliding under hot deformation conditions [
47,
48], where the testing temperature of 573 K corresponds 0.68
\({T}_{m}\) of the silicon eutectic phase that is relatively higher than for the matrix alloy of 0.65
\({T}_{\mathrm{m}}\) [
1]. Hence, the higher elongations achieved for deformed samples by 10 turns in HPT compared to the elongations in the as-deposited samples, were assisted by the high-volume fraction of the fine particles of eutectic silicon phase as seen in Figs.
3,
4 and
5 [
11,
49].
Filaments or fibrous structures appeared at lower strain rates and elevated temperature of 573 K for the processed samples rather than the as-deposited samples as observed in Figs.
11 and
12. These structures were aligned parallel to the tension direction and their role is reconnecting the disconnected grains and grain boundaries, as well as relinking the surface cavities that appear at the final stage of hot deformation. Therefore it seems that the higher values of elongations and alloy flow under conditions of elevated testing temperature and slower strain rate were maintained by the formation of the fibrous structures as observed in Fig.
10 [
47,
48,
50].
The chemical composition of the fibrous structures was analysed using EDS as shown in Fig.
12, which confirmed that these structures are mainly composed of
α-Al matrix grains, as indicated by the weight fractions of elements in the sample processed for 10 turns in HPT and then tested in tension using a strain rate of 10
–4 s
−1 at testing temperature of 573 K with an achieved elongation of 220%. It is worth noting that all samples have been exposed to oxidation as revealed by the oxygen weight ratio in the EDS data that presented in Fig.
12, as all tensile tests were carried out in air.
Fibrous structures were not observed in the as-deposited tensile samples as shown in Fig.
10, that were tested at a temperature of 573 K at all rates of strain. Instead, a cavitation failure was observed at slower strain rates of 10
–3 and 10
–4 s
−1 as seen in Fig.
10. This can be attributed to effects of the larger grain size and particle size and distribution of the eutectic phase in these as-deposited samples.
In the aforementioned conditions of tensile testing, grain growth is expected, where the measured grain size after high temperature testing was about 10 μm in the as-deposited tensile samples compared to the grain size of 5 μm in the tensile samples that were processed earlier for 10 HPT turns. The grain growth in the processed tensile samples was relatively inhibited by the existence of ultrafine grains and nanosized well-distributed eutectic particles [
38,
43].
The ultrafine grains are believed to preferably undergo glide-dislocation creep and slide over each other as indicated by the measurement of strain rate sensitivity. The distribution of nanosized eutectic particles within the
α-Al grains and along the grain boundaries resulted in relative retardation of cavitation and assisted grain sliding due to the softening of this phase at elevated testing temperature [
11,
43].
The non-spherical morphology of the eutectic phase for as-deposited samples, compared to the relatively spherical morphology of the nanosized eutectic particles in the processed samples, operate as sites for crack initiation leading to a reduction in the elongation to failure of as-deposited samples compared to that found in the processed samples at the same conditions of strain rates and testing temperature [
4,
16].