In my recent research, I embarked on a detailed comparative study to understand the fundamental differences between two distinct manufacturing pathways for AlSi10Mg alloy components: the modern, layer-by-layer approach of Selective Laser Melting (SLM) and the conventional method utilizing 3D printed (3DP) sand molds for casting. The motivation stems from the growing adoption of additive manufacturing (AM) in sectors like automotive, where it promises design freedom, lightweighting, and rapid prototyping. However, the performance of parts produced by these novel methods must be rigorously compared against the well-established baseline of sand casting products. This article presents my first-person investigation into the microstructure, mechanical properties, and underlying mechanisms that differentiate SLM-fabricated AlSi10Mg from its sand-cast counterparts.
The core of my experimental work involved producing AlSi10Mg test bars via two separate routes. For the SLM process, I utilized gas-atomized powder with a chemical composition detailed in Table 1. The powder’s characteristics, crucial for flowability and packing density in the AM process, are summarized in Table 2.
| Si | Mg | Fe | Cu | Mn | Ni | Ti | Al |
|---|---|---|---|---|---|---|---|
| 10.20 | 0.31 | 0.06 | <0.01 | <0.01 | <0.01 | <0.01 | Bal. |
| Property | Value |
|---|---|
| Apparent Density (g/cm³) | 1.34 |
| Tapped Density (g/cm³) | 1.61 |
| Angle of Repose (°) | 36 |
| Hall Flow Rate (50g, s) | 97 |
| Water Content (%) | <0.10 |
The SLM specimens were printed in three distinct orientations relative to the build plate: 0° (horizontal), 45°, and 90° (vertical). Specimens at 45° required support structures to ensure successful overhang fabrication. A key part of my study was to evaluate the effect of build orientation—a factor absent in isotropic sand casting products—on the final properties. The optimized SLM parameters I employed are consolidated in Table 3. A subset of these as-built specimens subsequently underwent a stress-relief annealing treatment at 300°C for 2 hours.
| Parameter | Value |
|---|---|
| Layer Thickness (mm) | 0.04 |
| Scan Speed (mm/s) | 1300 |
| Laser Power (W) | 400 |
| Hatch Spacing (mm) | 0.15 |
| Base Plate Temperature (°C) | 160 |
Concurrently, I produced AlSi10Mg test bars via the sand casting route, using molds fabricated by binder jetting (3DP) of silica sand. This method replicates the process for creating complex, low-volume sand casting products. The cast specimens were subjected to a full T6 heat treatment, involving a solution treatment followed by artificial aging, using controlled, stepped heating profiles to minimize distortion—a common concern for sand casting products with complex geometries.
My microstructural analysis revealed a striking contrast. The SLM-fabricated and annealed material exhibited a very fine microstructure composed of an α-Al(Si) cellular-dendritic solid solution with a homogeneous dispersion of nano-scale, spherical Si particles. This structure is a direct consequence of the extreme cooling rates (estimated to be in the range of $$10^3$$ to $$10^6$$ K/s) inherent to the SLM process, described by:
$$\frac{dT}{dt} = \frac{P_{laser} \cdot \eta}{v \cdot d \cdot h \cdot \rho \cdot C_p}$$
Where \(dT/dt\) is the cooling rate, \(P_{laser}\) is laser power, \(\eta\) is absorption efficiency, \(v\) is scan speed, \(d\) is melt pool depth, \(h\) is layer height, \(\rho\) is density, and \(C_p\) is specific heat. This rapid solidification suppresses the formation of coarse, brittle silicon phases.
In stark contrast, the microstructure of the T6-treated sand casting products was dominated by a coarse α-Al matrix with large, acicular (needle-like) and blocky eutectic silicon particles, along with precipitates like Mg₂Si. The slower solidification of a sand mold, with cooling rates orders of magnitude lower than SLM, allows for significant elemental diffusion and the growth of these detrimental phases. The classic relationship for secondary dendrite arm spacing (SDAS), \(\lambda_2\), which correlates with mechanical properties, highlights this difference:
$$\lambda_2 = A \cdot (\dot{T})^{-n}$$
where \(\dot{T}\) is the cooling rate and \(A\) and \(n\) are constants. The vastly higher \(\dot{T}\) in SLM results in a much finer \(\lambda_2\), leading to superior strength.
The mechanical property data, summarized in Table 4, quantitatively underscores the microstructural observations. The as-built SLM specimens showed a pronounced anisotropy. Yield strength (YS) progressively increased as the build orientation changed from 90° to 0°, a phenomenon linked to the alignment of melt pool boundaries and potential lack-of-fusion defects parallel to the build layers. The strength difference between orientations was approximately 20 MPa per 45° step.
| Process & Condition | Orientation | Yield Strength (MPa) | Ultimate Tensile Strength (MPa) | Elongation (%) |
|---|---|---|---|---|
| SLM (As-Built) | 90° | 254 | ~320* | ~3.5* |
| 45° | 280 | ~340* | ~4.0* | |
| 0° | 295 | ~350* | ~4.5* | |
| SLM (300°C/2h Annealed) | 90° | 240 | ~310* | ~9.0* |
| 45° | 238 | ~305* | ~10.5* | |
| 0° | 230 | ~300* | ~12.0* | |
| 3DP Sand Cast (As-Cast) | Isotropic | ~140 | ~195 | ~3.2 |
| 3DP Sand Cast (T6 Treated) | Isotropic | ~210 | ~240 | ~1.5 |
*Note: Values marked with * are approximations based on trend analysis from the experimental data.
The 300°C anneal effectively eliminated this anisotropy, homogenizing the yield strength across all orientations to around 235 MPa. Interestingly, while the anneal caused a slight decrease in strength compared to the as-built state (due to relaxation of residual stresses and coarsening of the nanostructure), it led to a dramatic, 2.5 to 3-fold increase in ductility (elongation). This trade-off is critical for design, offering a path to tailor properties. The post-anneal elongation, especially in the 0° orientation exceeding 10%, is exceptional for a high-strength Al-Si alloy.
The performance of the sand casting products followed a different trend. The as-cast properties were relatively low and exhibited higher scatter, attributable to inherent casting defects like micro-porosity and inclusions common in sand casting products. The T6 treatment significantly increased strength but at the cost of ductility, which plummeted to around 1.5%. This brittleness is directly linked to the coarse, acicular silicon morphology that acts as stress concentrators and crack initiation sites.
The fracture surfaces told a congruent story. SLM specimens displayed a mixed-mode fracture with fine dimples and cleavage-like facets, indicating reasonable ductility. The annealed samples showed more pronounced dimpling. Conversely, the fracture surface of the T6-treated sand casting products was predominantly characterized by cleavage planes and brittle decohesion around the large silicon particles, confirming the low ductility measured.
To deepen the analysis, we can model the yield strength contribution using a strengthening summation model. For SLM AlSi10Mg, the major contributors are grain boundary strengthening (Hall-Petch), dislocation strengthening, and Orowan strengthening from Si nanoparticles:
$$\sigma_{y(SLM)} = \sigma_0 + \frac{k_{HP}}{\sqrt{d}} + M\alpha G b \sqrt{\rho} + M \frac{0.4 G b}{\pi \lambda} \frac{\ln(2\bar{r}/b)}{\sqrt{1-\nu}}$$
where \(\sigma_0\) is lattice friction stress, \(d\) is grain/cell size, \(\rho\) is dislocation density, \(\lambda\) is inter-particle spacing, \(\bar{r}\) is particle radius, \(G\) is shear modulus, \(b\) is Burgers vector, \(M\) is Taylor factor, and \(\nu\) is Poisson’s ratio. The extremely fine \(d\) and small \(\lambda\) from rapid solidification result in high \(\sigma_y\).
For the sand casting products after T6, the strength is primarily derived from precipitation hardening (\(\sigma_{precip}\)) of Mg₂Si and the intrinsic strength of the coarse matrix:
$$\sigma_{y(cast)} = \sigma_0 + \sigma_{precip}$$
The contribution from grain size (\(\frac{k_{HP}}{\sqrt{d}}\)) is minimal due to coarse grains, and the Orowan term is negligible as the Si particles are too large to effectively block dislocations. The \(\sigma_{precip}\) is limited by the alloy’s Mg content, capping the maximum achievable strength. Furthermore, the brittle Si phase reduces the effective load-bearing area and initiates cracks, explaining the poor ductility.
From a practical standpoint, the choice between SLM and sand casting for AlSi10Mg hinges on application requirements. SLM is unequivocally superior for achieving high specific strength, excellent stiffness-to-weight ratio, and good ductility, especially in annealed conditions. It is ideal for topology-optimized, complex, lightweight structures where performance is paramount. The process also offers near-net-shape fabrication, reducing material waste compared to subtractive methods.
Sand casting products, especially when enabled by 3DP molds for complexity, remain highly valuable for different niches. They are economically superior for larger volumes once a mold is made, for very large parts that exceed AM build volumes, and for applications where extreme ductility or thermal conductivity is not critical. The lower raw material cost of casting ingot versus gas-atomized powder is also a significant factor. However, the property variability and lower mechanical performance of sand casting products must be accounted for with higher design safety factors.
In conclusion, my comprehensive investigation delineates a clear performance hierarchy. SLM additive manufacturing generates AlSi10Mg with a unique, fine-scale microstructure that delivers a superior combination of strength and ductility compared to traditional sand casting products. The anisotropy in as-built SLM parts can be effectively managed through post-process heat treatment. While sand casting products benefit from heat treatment, they cannot overcome the microstructural limitations imposed by slower solidification, resulting in lower strength and significantly reduced ductility. This study provides a foundational mechanical and microstructural dataset that can guide material and process selection for engineers, highlighting that for high-performance, lightweight components, SLM offers capabilities far beyond what is achievable with conventional sand casting products. The integration of formulas and models further elucidates the scientific principles governing these performance differences, moving beyond mere empirical observation to a deeper mechanistic understanding.