In recent years, the foundry industry has been driven toward “precision”, “lightweight”, and “composite” development, with an ever-increasing demand for near-net-shape and defect-free castings. Complex, thin-walled, and large-scale alloy components are becoming the norm. However, traditional sand casting, die casting, investment casting, and lost foam casting each encounter significant technical bottlenecks when producing large, complex, high-alloy precision castings. The carbon pick-up defect inherent in conventional lost foam castings prevents their use for high-value products such as low-carbon steel and stainless steel. Investment casting, on the other hand, is limited to small parts. The lost foam shell casting (or vacuum shell casting) process, originally proposed by the Steel Castings Research and Trade Association of the UK, combines the advantages of lost foam casting, investment casting shell technology, and vacuum casting. This technique effectively eliminates carbon contamination and enables the production of high-quality, large, thin-wall, complex alloy castings. The coating used in lost foam shell casting plays a critical role in determining the final casting quality. Low coating strength leads to shell collapse or run-out defects, while poor refractoriness causes sand inclusion and slag defects. Therefore, understanding the mechanical behavior of the ceramic shell, especially its bending strength at room and intermediate temperatures, is essential for process optimization. In this study, we systematically investigated the influence of coating layer thickness on the bending performance of lost foam shell casting molds at 25 °C and 300 °C, providing a data-driven reference for coating design and application.
The experimental procedure involved preparing shell specimens with a fixed face coat and varying numbers of back-up coats. The face coat used zircon flour as the refractory filler (particle size 150–800 μm) with silica sol binder, maintaining a powder-to-liquid ratio of 2.5–3.5. The back-up coats used quartz flour (particle size 300–800 μm) with small additions of iron oxide, sodium bentonite, and carboxymethyl cellulose, along with silica sol and a small amount of α-starch, at a powder-to-liquid ratio of 1.5–2.5. The shell configurations are summarized in Table 1. The shell thicknesses were measured after drying, and bending specimens (40 mm wide, specimen thickness as original, length ≥80 mm) were cut from the dried shells. Three-point bending tests were performed on a GNT100 universal testing machine at a crosshead speed of 1 mm/min and a span of 50 mm, at both 25 °C and 300 °C (after holding at 300 °C for sufficient time to remove moisture and volatiles).
| Specimen ID | Number of face coats | Number of back-up coats | Total dried shell thickness (mm) |
|---|---|---|---|
| 1+4 | 1 | 4 | 2.94 |
| 1+5 | 1 | 5 | 3.96 |
| 1+6 | 1 | 6 | 4.42 |
| 1+7 | 1 | 7 | 5.58 |
The force-displacement curves obtained from the bending tests are shown conceptually in the analysis. At 25 °C, the curves exhibited noticeable fluctuations, which we attribute to the bridging effect of the organic binder and short fiber strands within the coating. The binder not only bonds the refractory particles but also improves the room-temperature toughness. When fibers are present, the crack propagation is hindered, resulting in a serrated load-displacement response. In contrast, at 300 °C, the force-displacement curves became smooth and relatively free of jumps. This is because the binder dehydrates and the fibers volatilize during the high-temperature hold, eliminating the fiber-matrix interaction. The macroscopic fracture surfaces also differed: at 25 °C, the fracture surface was silver-grey and uneven, indicating extensive crack path tortuosity and higher energy absorption. At 300 °C, the fracture surface appeared blackened due to the carbonization of organic components, and small pores were observed – a beneficial feature that enhances the permeability of the shell and reduces casting defects.
To quantify the effect of coating thickness on load-bearing capacity, we recorded the maximum bending force (Fmax) for each shell configuration at both temperatures. The results are presented in Table 2 and further illustrated in the following analysis.
| Specimen ID | Fmax at 25 °C (N) | Fmax at 300 °C (N) |
|---|---|---|
| 1+4 | 14.35 | 10.2 |
| 1+5 | 17.0 | 16.9 |
| 1+6 | 22.2 | 28.4 |
| 1+7 | 39.5 | 46.3 |
As the number of back-up coats increased, the maximum bending force exhibited a clear upward trend at both test temperatures. For example, when the shell thickness increased from 2.94 mm (1+4) to 5.58 mm (1+7), the maximum force at 25 °C rose from 14.35 N to 39.5 N, an increase of approximately 175%. At 300 °C, the corresponding increase was from 10.2 N to 46.3 N, i.e., about 354%. Interestingly, the force at 300 °C was slightly higher than that at 25 °C for the thicker shells (1+6 and 1+7), whereas for the thinnest shell (1+4), the opposite trend was observed. This phenomenon can be explained by the competing effects of binder dehydration and thermal sintering. At 300 °C, the moisture and volatile matter are driven off, allowing the refractory particles to pack more densely, and the remaining binder undergoes a degree of sintering that strengthens the particle-particle contacts. However, for very thin shells, the loss of organic binding components may outweigh the sintering benefit, leading to a net reduction in strength. As the shell becomes thicker, the sintering effect dominates, resulting in higher load capacity at elevated temperature.
Beyond the maximum force, we also calculated the bending strength (modulus of rupture, MOR) using the standard three-point bending formula:
$$
\sigma = \frac{3 F L}{2 b t^2}
$$
where $F$ is the applied load, $L$ is the span length (50 mm), $b$ is the specimen width (40 mm), and $t$ is the specimen thickness (the measured shell thickness). The computed bending strengths are summarized in Table 3 and Figure 6 (conceptual). It is evident that for a given shell configuration, the bending strength at 300 °C was consistently higher than that at 25 °C. This aligns with the sintering argument: the particles are more tightly bonded after dehydration and partial sintering, leading to improved cohesion. Notably, the strength did not increase linearly with the number of coats. The most significant jump occurred when moving from 1+4 to 1+5, after which the strength values plateaued. For example, the bending strength of the 1+5 specimen at 25 °C was approximately 3.2 MPa, while that of the 1+6 specimen was only marginally higher at about 3.5 MPa. This suggests that once the shell reaches a critical thickness (~4 mm), the influence of internal defects (microvoids, surface irregularities, uneven coating) is minimized, and the intrinsic mechanical properties of the coating material are approached. Therefore, from a practical standpoint, the optimum coating configuration is 1+5 or 1+6, corresponding to a total shell thickness of 4–5 mm, which balances mechanical performance, production cycle time, and cost.
| Specimen ID | MOR at 25 °C (MPa) | MOR at 300 °C (MPa) |
|---|---|---|
| 1+4 | 2.1 | 2.8 |
| 1+5 | 3.2 | 3.9 |
| 1+6 | 3.5 | 4.5 |
| 1+7 | 3.7 | 4.8 |
To further validate the applicability of our findings, we manufactured ultra-low-carbon stainless steel castings using the optimized shell (1 face + 6 back-up coats, thickness 4.62 mm). The process flow included: lost foam pattern fabrication, coating application (dipping and drying), shell burnout, shell placement in a flask, and vacuum-assisted pouring. The castings included complex valve bodies and curved blades. Visual inspection revealed excellent surface finish with clear part numbers, and chemical analysis confirmed that no carbon pick-up occurred – the carbon content remained within the specified range. This demonstrates that the coating system developed in this study is robust for large, intricate stainless steel castings produced by lost foam shell casting.
In conclusion, our systematic investigation of the mechanical properties of ceramic shells for lost foam shell casting has yielded several key findings. First, increasing the number of back-up coats enhances the maximum bending force that the shell can withstand, but the bending strength tends toward a plateau once the shell thickness exceeds about 4 mm. The optimal configuration is either 1+5 or 1+6 coats, corresponding to a shell thickness of 4–5 mm. Second, the shell strength at 300 °C (typical burnout temperature) is higher than at room temperature due to binder dehydration and partial sintering, which is beneficial for handling and casting operations. Third, the observed force-displacement behavior and fracture morphology provide insights into the role of organic binders and fibers: they increase room-temperature toughness but volatilize at high temperature, leading to smoother load response and improved permeability. These results offer practical guidelines for designing lost foam shell casting coatings and contribute to the broader application of this process for high-quality, carbon-free castings.
Future work should extend the temperature range to simulate actual pouring conditions (e.g., 1000–1500 °C) and investigate the effects of different refractory fillers and binder systems. Additionally, the relationship between shell permeability, mechanical strength, and casting quality warrants further study. The lost foam shell casting process holds great promise for producing near-net-shape components in high-alloy steels, nickel-based superalloys, and other challenging materials, and a deeper understanding of coating mechanics will accelerate its industrial adoption.