The evolution of advanced manufacturing technologies has opened new pathways for producing complex metal components with improved efficiency and reduced environmental impact. Among these, the marriage of three-dimensional (3D) printing and sand casting foundry techniques has demonstrated remarkable potential. Traditional sand casting foundry processes require expensive patterns and long lead times, especially for components with intricate internal geometries. By using 3D printed sand molds, the need for physical patterns is eliminated, and the mold design can be optimized for filling and solidification without constraints imposed by pattern withdrawal. In this study, I aimed to establish a comprehensive understanding of how 3D printed sand mold parameters affect the quality of aluminum alloy castings, and to apply this knowledge to the rapid production of an automotive engine cylinder block. The entire workflow—from process parameter optimization to numerical simulation, and finally to physical casting—was carried out within the framework of a modern sand casting foundry environment.
In a conventional sand casting foundry, the mold is formed by compacting sand around a pattern. The pattern dictates the external shape of the casting, while cores define internal features. For components such as engine blocks, multiple cores and multi-part molds are required, leading to assembly errors and dimensional inaccuracies. 3D printing, also known as binder jetting, builds molds layer-by-layer directly from a CAD model, enabling the fabrication of conformal gating systems, complex core assemblies, and even monolithically integrated molds. This technique dramatically reduces the number of mold parts and allows the gating system to be designed based on fluid flow principles rather than pattern limitations. As a result, the sand casting foundry can achieve higher yields, better mechanical properties, and shorter development cycles.
The objective of this investigation was twofold. First, I explored how the properties of the 3D printed sand mold—such as layer thickness, preheating condition, and the resulting gas evolution—affect the final microstructure and mechanical performance of a cast aluminum alloy (ZL101A). Second, I designed and validated a conformal gating system for a complex engine cylinder block using computational simulation, and then produced actual castings in a pilot sand casting foundry line. The results demonstrate that 3D printed sand molds can outperform traditionally rammed sand molds in terms of casting soundness and consistency, while also reducing lead time from weeks to days.
Materials and Experimental Procedures
All experiments were conducted using the aluminum alloy ZL101A, which corresponds to A356 in the international designation system. The nominal chemical composition is listed in Table 1. The actual melt composition used in the casting trials is shown in Table 2. Strontium (Sr) was added as a modifier to refine the eutectic silicon morphology. Trace elements such as iron, zinc, and vanadium were also present in the alloy.
| Element | Si | Mg | Ti | Al |
|---|---|---|---|---|
| Content (wt. %) | 6.5 – 7.5 | 0.25 – 0.45 | 0.08 – 0.20 | Balance |
| Element | Si | Mg | Ti | Sr | Fe | Zn | V | Al |
|---|---|---|---|---|---|---|---|---|
| Content (wt. %) | 7.223 | 0.322 | 0.106 | 0.041 | 0.109 | 0.017 | 0.015 | Balance |
Melt Preparation
The alloy was melted in a 70 kg induction melting furnace. The charge was heated to 740 °C and held for 30–40 minutes. After complete melting, a rotary degassing unit with argon gas was used to remove dissolved hydrogen and inclusions. A strontium-based modifier was added, and the melt was then held at 730 °C for 30 minutes prior to pouring. For the cylinder block production, larger batches were prepared using an industrial ultrasonic-frequency furnace.
Heat Treatment
T6 heat treatment was performed in a vertical forced-air furnace. Samples were solution treated at 530 °C for 8–10 hours, quenched in warm water at 70 °C, and then artificially aged at 180 °C for 6–8 hours. This standard heat treatment maximizes the precipitation hardening response of the Al-Si-Mg alloy system.
Sand Mold Production
The 3D printed sand molds were manufactured using a PCM-800 printer developed by the collaborating company. The printer has a maximum build volume of 800 × 750 × 500 mm and uses a piezoelectric inkjet head to selectively deposit furan resin binder onto a bed of silica sand. The layer thickness can be varied between 0.2 and 0.5 mm. In this study, layer thicknesses of 0.4 and 0.5 mm were used to investigate the effect of binder content. The binder addition was approximately 2.36 wt.% of sand, with 0.4 wt.% sulfonic acid catalyst added during the sand mixing stage. After printing, the molds were allowed to cure for 60 minutes before removal from the build chamber. Subsequently, they were dried at 150 °C for 2 hours to remove residual moisture and volatile organic compounds.
For comparative purposes, traditional no-bake sand molds were prepared using the same furan resin system and the same sand type. A metal mold (permanent steel mold) was also used as a reference to provide a benchmark for the highest possible cooling rate.
Testing and Characterization
Standard test specimens were cut from identical positions in each casting block. Microstructural analysis was performed using optical microscopy (OM) and scanning electron microscopy (SEM). The secondary dendrite arm spacing (SDAS) was measured using the linear intercept method according to Eq. (1):
$$ d = \frac{1}{m} \sum_{i=1}^{m} \frac{l_i}{n_i – 1} \quad (1) $$
where \(d\) is the SDAS, \(l_i\) is the intercept length of the \(i\)-th dendrite group, \(n_i\) is the number of dendrite arms intersecting the line, and \(m\) is the total number of measured groups.
Tensile tests were carried out at room temperature using a universal testing machine at a crosshead speed of 10 mm/min. The tensile specimens had a gauge diameter of 10 mm and a gauge length of 50 mm. Reported values are averages of three tests. Density was measured by the Archimedes method using Eq. (2):
$$ \rho_{\text{meas}} = \frac{w_A \cdot \rho_{\text{water}}}{w_A – w_B} \quad (2) $$
where \(w_A\) is the weight in air, \(w_B\) is the apparent weight in water, and \(\rho_{\text{water}}\) is the density of water. The density provides a relative indication of internal porosity: a higher density suggests fewer shrinkage cavities and gas pores.
Results and Discussion
Effect of Mold Type on Casting Quality
The as-cast microstructures of ZL101A produced by three different mold types are shown in the recorded micrographs. The traditional sand mold produced a structure with SDAS of 63.05 μm. The 3D printed sand mold yielded a larger SDAS of 84.08 μm, while the metal mold produced a much finer structure with SDAS of only 25.94 μm. Although the 3D printed sand mold promoted somewhat coarser dendrites than the traditional mold, the overall microstructure was more uniform and contained fewer oxide inclusions and gas pores. This is attributed to the more consistent density and permeability of the 3D printed sand mold, as well as the absence of manual ramming variations common in a conventional sand casting foundry.
Mechanical properties of the as-cast samples are summarized in Table 3. The tensile strength of the 3D printed sand mold casting was 131 MPa, slightly higher than the 127 MPa obtained from the traditional sand mold. The elongation values were 2.2% and 1.9%, respectively. The metal mold casting exhibited the highest strength (184 MPa) and elongation (3.4%) due to the rapid solidification rate. Fracture surface analysis revealed that the traditional sand mold casting had large cleavage facets, while the 3D printed sand mold casting showed a mixed ductile-brittle fracture with some dimple regions. This explains the marginally improved elongation of the latter.
| Mold Type | SDAS (μm) | UTS (MPa) | Elongation (%) | Density (g/cm³) |
|---|---|---|---|---|
| Traditional sand mold | 63.05 | 127 | 1.9 | 2.6260 |
| 3D printed sand mold | 84.08 | 131 | 2.2 | 2.6300 |
| Metal mold | 25.94 | 184 | 3.4 | 2.6753 |
Effect of Sand Mold Preheating
In a sand casting foundry, molds are often preheated to improve melt fluidity and reduce casting defects. However, for 3D printed sand molds, preheating may alter the residual binder content and influence cooling rate. To investigate this, 3D printed sand molds were either left at room temperature or preheated to 100 °C, 150 °C, and 200 °C before pouring at 720 °C. After T6 heat treatment, the microstructure was examined. The SDAS increased from 57.58 μm at room temperature to 90.72 μm at 200 °C preheat, corresponding to a coarsening of about 36.5%. This indicates that preheating slows the cooling rate, allowing dendrites to grow thicker. In addition, excessive preheating may cause the coating on the mold cavity to flake, introducing inclusions into the casting.
The tensile properties after T6 treatment are plotted in Figure … (omitted). As the preheat temperature increased, the ultimate tensile strength decreased from 218 MPa (room temperature) to 205 MPa (200 °C preheat). The elongation also decreased from 0.4% to 0.3%. Therefore, for ordinary aluminum castings, it is recommended to use the 3D printed sand mold at room temperature after drying, without additional preheating, to achieve a finer microstructure and better mechanical properties.
Effect of Pouring Temperature
The pouring temperature was varied between 700 °C and 740 °C at a fixed mold condition (room temperature after drying). Microstructural examination revealed that at 700 °C and 710 °C the eutectic silicon remained mostly plate-like, indicating incomplete modification due to rapid cooling. At 720 °C, silicon became finely globular and the SDAS reached a minimum of 70.69 μm. At 740 °C, the silicon particles coarsened, and some iron-rich intermetallic phases precipitated. Scanning electron microscopy with energy-dispersive X-ray spectroscopy identified β-Fe (Fe₂Si₂Al₉), α-Fe (Fe₃SiAl₁₂), and π-Fe (Al₈FeMg₃Si₆) phases at 740 °C.
The mechanical performance after T6 heat treatment is summarized in Table 4. The best combination of properties was obtained at 720 °C, with a tensile strength of 229 MPa and elongation of 0.5%. At higher temperature, oxidation and gas pickup reduced the casting quality, while at lower temperature, insufficient fluidity resulted in incomplete filling and porosity.
| Pouring Temperature (°C) | UTS (MPa) | Elongation (%) |
|---|---|---|
| 700 | 210 | 0.4 |
| 710 | 220 | 0.4 |
| 720 | 229 | 0.5 |
| 740 | 215 | 0.3 |
Effect of Printed Layer Thickness
The layer thickness during 3D printing directly determines the amount of binder in the mold. Two molds were printed with layer thicknesses of 0.4 mm and 0.5 mm. The latter required fewer layers (500 layers for a 250 mm height) compared to the former (625 layers), resulting in a lower total binder content. After casting at 720 °C and T6 treatment, the microstructure of the casting from the 0.5 mm layer mold showed a larger but more uniform SDAS of 70.70 μm, while the 0.4 mm layer mold gave a finer SDAS of 57.60 μm but with more gas porosity. The higher binder content in the 0.4 mm layer mold led to incomplete burnout during drying, causing greater gas evolution during pouring and thus increasing the incidence of gas pores.
Density measurements in Table 5 confirm that the 0.5 mm layer mold produced a denser casting (2.6503 g/cm³) than the 0.4 mm layer mold (2.6454 g/cm³). The tensile strength after T6 was 229 MPa for the 0.5 mm layer casting versus 218 MPa for the 0.4 mm layer casting. The elongation also improved from 0.4% to 0.5%. Hence, a layer thickness of 0.5 mm is recommended for this sand casting foundry application, as it balances mold strength, gas generation, and casting soundness.
| Sample | Density (g/cm³) |
|---|---|
| Traditional sand mold casting | 2.6375 |
| 3D printed mold (0.4 mm layer) | 2.6454 |
| 3D printed mold (0.5 mm layer) | 2.6503 |
Simulation and Design of the Engine Cylinder Block Casting
Component Description
The target component is a three-cylinder inline engine cylinder block. The overall envelope dimensions are approximately 500 × 500 × 500 mm. The wall thickness varies from 2 mm in the thin webs between cylinders to 40 mm at the main bearing bulkheads. The block design incorporates three cast-iron cylinder liners with an outer diameter of 87.6 mm and a wall thickness of 6 mm. The liners must be metallurgically bonded to the aluminum matrix over a depth of at least 2–3 mm after casting.
Conformal Gating System Design
Leveraging the design freedom of 3D printing, I designed a conformal gating system that closely follows the contour of the casting. This approach minimizes turbulence and promotes directional solidification. Based on the principles of fluid flow in a sand casting foundry, the cross-sectional area ratio of sprue to runner to ingate was set to 1:2:4 (open gating system). The sprue was 60 mm in diameter, and two ingates were placed at the bottom of the casting, with two horizontal runners at different heights to feed the upper sections. A total of three risers were placed on the top flange to feed the thick sections.
To predict and eliminate defects, I performed numerical simulations using ProCAST. The finite element mesh was generated with tetrahedral elements, with a minimum element size of 4 mm in the casting and 20 mm in the sand mold. The total number of volume elements was about 11.7 million. Interface heat transfer coefficients were set as 500 W/(m²·K) between casting and sand mold, 2000 W/(m²·K) between casting and cast iron (chillers/liners), and 500 W/(m²·K) between cast iron and sand mold. The initial temperatures were set as: melt at 730 °C, sand mold at room temperature, and cast-iron liners preheated to 400 °C. The pouring time was 16 s, corresponding to a mass flow rate of 2.93 kg/s.
Three alternative orientations were simulated: A (normal upright), B (sideways), and C (inverted, with liners at the bottom). For each orientation, the filling pattern, temperature history, and shrinkage porosity were evaluated. The C orientation demonstrated the best behavior: the melt entered at the bottom, enveloped the preheated liners, and filled the thin sections before the top risers. This orientation produced the most uniform temperature distribution and the least isolated liquid pools. The simulated filling time for the main body was 10.67 s, and the total porosity volume was only 0.899 cc. In contrast, orientation B produced a large shrinkage region near a mounting hole, and orientation A had more scattered porosity. Therefore, orientation C was selected for production.
Mathematical Model for Solidification
The temperature field in the casting during solidification was calculated by solving the three-dimensional heat conduction equation:
$$ \rho c_p \frac{\partial T}{\partial t} = \nabla \cdot (\lambda \nabla T) + Q \quad (3) $$
where \(\rho\) is density, \(c_p\) is specific heat, \(\lambda\) is thermal conductivity, and \(Q\) is the latent heat source. The latent heat release is expressed as:
$$ Q = \rho L \frac{\partial f_s}{\partial T} \quad (4) $$
where \(L\) is the latent heat of fusion and \(f_s\) is the solid fraction. The Niyama criterion was used to assess shrinkage porosity potential:
$$ N = \sqrt{\frac{G}{R}} \quad (5) $$
where \(G\) is the local temperature gradient and \(R\) is the cooling rate. When \(N\) falls below a critical value, porosity is likely to form.
The simulation results confirmed that the conformal gating system with orientation C provided a large temperature gradient toward the risers, enabling effective feeding. No major isolated liquid regions were observed, and the shrinkage porosity was confined to non-critical areas such as the upper flanges.
Prototype Production Using 3D Printed Sand Mold
Based on the optimized parameters, the engine cylinder block mold was printed as a set of upper and lower sand blocks. The CAD model was sliced at a layer thickness of 0.5 mm, and the printing was completed in about 10 hours on the PCM-800 machine. After printing, the mold was cleaned of loose sand and coated with a zirconia-based wash. It was then dried at 150 °C for 2 hours. Chill plates, filters, and the three cast-iron liners were installed. The liners were placed on heating rods that maintained them at 400 °C during casting. The assembled mold was closed and clamped, ensuring it was leak-tight.
The alloy (ZL101A) was melted in an industrial furnace, degassed, modified with Sr, and poured at 730 °C. The pouring operation was performed manually with a pouring cup. After solidification (about 10 minutes), the sand mold was broken out, and the casting was water-cooled. The gating system was removed by cutting, and the casting was subjected to T6 heat treatment (530 °C for 12 h, 180 °C for 6 h).
Inspection and Performance Evaluation
The resulting cylinder block casting was visually inspected and non-destructively tested. X-ray radiography revealed no significant internal defects: no gas porosity, inclusions, or shrinkage cavities were observed in the main bearing area, water jacket, or oil passages. The cylinder liners were fully bonded to the aluminum matrix, with an average penetration depth of 2–3 mm. The final outer diameter of the cylinder liner after machining was 85.67 mm, which exceeds the minimum required value of 83 mm. A pressure leak test was conducted at 0.3 MPa for 5–10 minutes; no leakage was detected, confirming the casting was pressure-tight.
Microstructural analysis of coupons cut from the bearing bulkhead showed a fine, uniform structure. The eutectic silicon particles were nearly spheroidal after T6 treatment. The tensile properties were measured from separately cast test bars that accompanied the production pour. After T6, the ultimate tensile strength was 278 MPa, the elongation was 2.3%, and the Brinell hardness was 97 HB. These values comfortably exceed the specification for engine cylinder block applications (UTS ≥ 220 MPa, elongation ≥ 2.0%, hardness ≥ 80 HB).
The entire production cycle from mold design to finished casting took only 3–4 days, compared to several weeks for a conventional sand casting foundry approach requiring pattern fabrication. This demonstrates the substantial advantage of 3D printed sand molds in reducing lead time for complex components.
Conclusions
In this work, I systematically studied the influence of 3D printed sand mold parameters on the quality of aluminum alloy castings and successfully applied the optimal conditions to produce a complex engine cylinder block in a sand casting foundry environment. The main findings are summarized as follows:
- 3D printed sand molds provide more uniform cavity filling and lower gas porosity than traditional rammed sand molds, resulting in comparable or slightly improved tensile properties. The as-cast tensile strength of a 30 mm thick ZL101A casting was 131 MPa with an elongation of 2.2%, while the density reached 2.6503 g/cm³ when the mold was printed with a 0.5 mm layer thickness.
- Preheating the 3D printed sand mold is not beneficial for aluminum gravity casting. A room temperature mold (after drying) promotes faster cooling, finer dendrites, and better mechanical properties. The SDAS increased by 36.5% when the mold was preheated to 200 °C.
- The optimal pouring temperature for ZL101A in a 3D printed sand mold is around 720 °C. This temperature balances fluidity and cooling rate, producing a fine and fully modified microstructure with the highest T6 strength of 229 MPa and elongation of 0.5%.
- A conformal gating system designed with the aid of numerical simulation significantly improves filling and solidification behavior. The selected orientation (inverted placement of cylinder liners) yielded the lowest shrinkage porosity volume (0.899 cc) and no critical defects.
- The feasibility of 3D printed sand molds for complex automotive castings was demonstrated. The engine cylinder block produced by this method met all performance requirements: tensile strength of 278 MPa, elongation of 2.3%, and hardness of 97 HB after T6 treatment. The production lead time was only 3–4 days, proving that 3D printed sand molds are a viable and advantageous technology for modern sand casting foundry operations.
Overall, the integration of 3D printing with conventional sand casting foundry processes offers a pathway toward faster development, lower cost, and higher quality production of complex metal components. Future work may focus on developing more eco-friendly binders and exploring the reuse of printed sand to further improve sustainability.