In the production of bearing bushes for railway vehicles, we encountered a persistent quality problem: a high incidence of sand casting defects, specifically scabbing and sand inclusion, which drastically reduced the yield rate. The bearing bush material is ductile iron QT500-7, with a single casting weight of approximately 9 kg. The casting process was carried out on a vertical parting flaskless shoot-squeeze molding line. Initially, the product qualification rate hovered around 60%, with sand casting defects accounting for over 70% of total scrap. These defects appeared on the arc surface and vertical planes of the bearing bush, ranging from 3 mm to 7 mm in depth and varying in area. To address this, we systematically investigated the root causes and implemented both molding system modifications and sand property improvements. This paper presents our findings and the successful elimination of sand casting defects through a combination of gating system redesign and calcium bentonite activation using Na₂CO₃.
Problem Analysis and Initial Investigation
We observed that the original gating system was a closed type, with the sprue directed straight at the arc surface of the sand mold, as illustrated conceptually in Figure 1 (see link below). The concentrated flow of molten iron caused intense local thermal shock on the sand mold surface, leading to rapid heating and moisture evaporation. This high thermal gradient induced thermal expansion stress in the surface layer of the sand mold. Simultaneously, the quality of the molding sand was insufficient to withstand this thermal shock. The combination of localized overheating and inadequate sand strength resulted in surface cracks, bulges, and subsequent sand casting defects on the casting surface.
We identified two primary factors contributing to the sand casting defect: (1) improper gating design causing direct impingement of molten metal on the mold wall, and (2) poor thermal stability and hot-wet tensile strength of the molding sand. The sand used was bonded with Weifang natural calcium bentonite. When exposed to the high temperature of molten iron, the moisture in the sand layer near the surface rapidly migrates inward, forming a high-moisture zone approximately 3–7 mm below the surface. This weakened zone significantly reduces the sand strength. When the internal thermal stress exceeds the strength of the sand at that plane, the surface layer separates and forms a scab or sand inclusion defect. The severity of the defect is directly related to the rate of heat transfer and the temperature gradient.
Improvement of Gating System
To reduce the direct thermal impact of molten iron on the sand mold, we tested three alternative gating arrangements. The first approach was to change the direction of the ingate so that the molten metal enters tangentially along the arc surface. However, due to geometric constraints of the mold cavity, the effect was limited. The second approach was to adopt a bottom-gated top-feeding system, but this caused excessive temperature drop of the molten iron along the long sprue, and the defect rate did not decrease significantly. The third approach, as shown in Figure 2 (not included here), was a top-gated system where the metal enters from the top of the casting. This design balanced filling speed and reduced direct impingement on the mold walls. After implementation, the scrap rate due to sand casting defects decreased from about 40% to approximately 30%. Although this improvement was significant, it was not sufficient to achieve the desired quality. We realized that the fundamental solution lay in improving the sand’s resistance to thermal shock.
Enhancement of Molding Sand Properties through Bentonite Activation
The hot-wet tensile strength of the sand is the key property that resists the formation of sand casting defects. When the mold surface is heated, moisture evaporates and condenses in a subsurface zone. If the sand at that zone has low strength, it will rupture under the thermal expansion stress. The calcium bentonite used originally has a lower swelling capacity and lower hot-wet strength compared to sodium bentonite. We decided to activate the calcium bentonite by treating it with sodium carbonate (Na₂CO₃). The chemical reaction converts calcium montmorillonite into sodium montmorillonite:
This ion exchange increases the interlayer spacing and water absorption capacity, thereby improving the green strength, dry strength, and hot-wet tensile strength of the molding sand. We added 3–5% Na₂CO₃ (by weight of bentonite) to the calcium bentonite during sand preparation. Table 1 compares the key properties of activated and unactivated bentonite, as well as the resulting sand properties.
| Property | Unactivated Bentonite | Activated Bentonite |
|---|---|---|
| Swelling capacity (x times) | 10.5 | 35.7 |
| Methylene blue absorption (g per 100 g bentonite) | 32.80 | 34.20 |
| Green compressive strength of sand (Pa) | (3.4–4.1) × 10⁴ | (5.4–5.9) × 10⁴ |
| Dry compressive strength of sand (Pa) | 4.1 × 10⁵ | 4.4 × 10⁵ |
| Hot-wet tensile strength (relative index) | Low (baseline) | High (70% improvement) |
We also measured the thermal stability of the bentonite by testing the methylene blue absorption after calcination at various temperatures. The results are summarized in Table 2. The activated bentonite retained a significantly higher absorption capacity even after heating to 600°C, indicating better thermal resistance and structural retention.
| Temperature (°C) | Unactivated Bentonite | Activated Bentonite |
|---|---|---|
| 25 | 32.80 | 34.20 |
| 300 | 23.0 | 27.5 |
| 400 | 2.2 | 1.4 |
| 600 | ~0 (almost completely deactivated) | 1.40 |
| 700 | 0 | 0 |
| 800 | 0 | 0 |
The mechanism is that sodium bentonite has a higher bonding strength at elevated temperatures because its layers do not collapse as readily as calcium bentonite. The hot-wet tensile strength, which is the critical parameter for preventing sand casting defects, can be expressed empirically as:
where \(\sigma_{hw}\) is the hot-wet tensile strength, \(\sigma_0\) is a material constant, \(E_a\) is the activation energy for interlayer bond failure, \(R\) is the universal gas constant, \(T\) is absolute temperature, and \(f(M)\) is a function of moisture content. Activation with Na₂CO₃ increases \(\sigma_0\) and possibly reduces \(E_a\) by enhancing the binding forces. In our production tests, the use of activated bentonite allowed us to reduce the total bentonite addition from 10% to 3–5% while still achieving higher green and dry strengths. This also improved the permeability and reduced dust generation.
We performed a series of controlled experiments to quantify the effect of the activation. Table 3 shows the measured hot-wet tensile strength (using a standard sand sample with 3.5% moisture and 4% bentonite) for the two types.
| Bentonite Type | Hot-Wet Tensile Strength (kN/m²) | Standard Deviation |
|---|---|---|
| Unactivated | 12.3 | 1.1 |
| Activated (3% Na₂CO₃) | 21.6 | 0.9 |
| Activated (5% Na₂CO₃) | 23.1 | 0.8 |
The hot-wet strength nearly doubled after activation. In production, this translated directly into a dramatic reduction in sand casting defects. The scrap rate due to scabbing and sand inclusion dropped from over 40% to below 7%, as shown in Table 4.
| Stage | Scrap Rate from Sand Casting Defects (%) | Overall Product Qualification Rate (%) |
|---|---|---|
| Before any improvement | ~40 | 60 |
| After gating change only | ~30 | 70 |
| After sand activation (combined) | ≤7 | 93 |
Discussion of the Mechanism
The formation of a sand casting defect in a vertically parted mold can be described by a simplified thermal-mechanical model. When molten iron at approximately 1400°C contacts the mold surface, the temperature at the surface rises rapidly, creating a steep temperature gradient. The thermal diffusivity of the sand determines the depth of the heated zone. The moisture in the sand evaporates and migrates deeper into the mold, creating a saturated condensation zone. This zone has significantly reduced strength because the water film reduces interparticle friction and bonding. If the thermal expansion stress \(\sigma_{th}\) of the hot surface layer exceeds the tensile strength \(\sigma_{ten}\) of the condensation plane, a delamination occurs, leading to a scab. The critical condition for sand casting defect formation can be expressed as:
where \(\sigma_{th}\) is proportional to the temperature difference \(\Delta T\) and the coefficient of thermal expansion \(\alpha\) of the sand:
Here \(E\) is the elastic modulus of the sand. The tensile strength \(\sigma_{ten}\) is a function of temperature \(T\) and moisture content \(M\). By using activated bentonite, we increased \(\sigma_{ten}\) at the critical plane through stronger interparticle bridges, thereby raising the threshold for defect initiation.
Furthermore, the activation process also improved the thermal stability of the bentonite. As seen in Table 2, at 600°C the unactivated bentonite lost almost its entire binding capacity (methylene blue absorption near zero), whereas the activated bentonite still retained some structure. This means that even under severe thermal exposure, the sand grain-to-grain bonds remained intact longer, preventing crack propagation. The addition of Na₂CO₃ also promotes sintering of the bentonite at a lower temperature, but that is a secondary effect.
Practical Implementation and Quality Control
In the workshop, we implemented the following procedure for sand preparation:
- Add 3–5% Na₂CO₃ (based on bentonite weight) into the calcium bentonite powder.
- Mix the activated bentonite with silica sand in a muller for 3 minutes.
- Add water and continue mulling for another 3 minutes to achieve uniform moisture (target 3.2–3.8%).
- Control the compactability index to 40–45%.
We monitored the hot-wet tensile strength daily using a standard test method. The target value was above 20 kN/m². When the strength dropped, we adjusted the Na₂CO₃ addition or the bentonite content. The gating system was redesigned to a top-gate with a central sprue and multiple ingates to ensure even filling. The combination of these measures virtually eliminated sand casting defects. The economic benefit was substantial: the product qualification rate increased from 60% to over 93%, and the production cost per good casting decreased by 25% despite the additional cost of Na₂CO₃.
We also observed that the activation treatment improved the sand’s resistance to erosion from the flowing metal, further reducing the risk of sand inclusions. The overall cast surface quality improved, reducing the need for grinding and cleaning operations.
Conclusion
Based on our investigation and production trials, we concluded that the elimination of sand casting defects in bearing bush castings requires a dual approach: (1) optimizing the gating system to minimize direct thermal impact on the sand mold, and (2) enhancing the sand’s hot-wet tensile strength through chemical activation of the bentonite binder. The top-gated design reduced the immediate thermal shock, but the key breakthrough came from using Na₂CO₃ to activate calcium bentonite, converting it into a sodium-rich bentonite with superior thermal stability and bond strength. This simple and cost-effective technique increased the hot-wet tensile strength by nearly 100%, dropping the scrap rate from over 40% to below 7%. The process is easy to implement on existing molding lines and does not require expensive equipment. We recommend that foundries dealing with similar sand casting defects in ductile iron castings consider this activation method as a primary remedy.