In my work at a large diesel engine foundry, I have spent years transitioning from traditional clay-bonded dry sand molds to furan resin self-hardening sand (also known as no-bake sand) for producing heavy ductile iron crankshafts. This shift, driven by the need for better surface quality, dimensional accuracy, and energy savings, initially introduced a series of quality problems: shrinkage porosity, gas defects, and slag inclusions. Through systematic experimentation and long-term statistical analysis, I have developed effective countermeasures that allow the sand casting foundry to produce defect-free crankshafts using self-hardening sand. In this article, I will share my findings, supported by extensive tables and mathematical models, to demonstrate that self-hardening sand can fully replace clay-bonded dry sand in sand casting foundry applications for large ductile iron components.
Initial Challenges and Process Adjustments
When I first switched to self-hardening sand in my sand casting foundry, I adopted a horizontal pouring and horizontal cooling (no-riser) process to reduce the risk of mold collapse, because the hot strength of self-hardening sand is much lower than that of dry clay sand. However, ultrasonic inspection soon revealed internal shrinkage porosity at random locations across the crankshaft. The fracture surfaces showed severe shrinkage cavities. In addition, gas defects—often accompanied by “choking” (violent ejection of molten metal from vents)—and slag inclusions became major problems, especially for the 6L/8L series crankshafts with large cross-sections.
To solve shrinkage porosity, I took the following steps:
- Returned to a horizontal pouring and horizontal cooling process, keeping the mold stationary after pouring to avoid mechanical expansion of the casting.
- Combined concentrated and dispersed gating: most liquid iron entered through a blind riser at the small end, while a small portion was introduced via three ingates near the flange and the third and sixth crank throws. This created a favorable temperature gradient: hot at the riser end, cooler in the middle and flange end, so that the riser could feed the hot end while the cooler sections, having less liquid shrinkage, were self-feeding.
- Increased the number and size of chills, especially on the main journal necks adjacent to the riser (first and second crank throws), where overheating was severe. These chills accelerated solidification of the surface layer, forming a rigid shell that could withstand the expansion pressure during graphite precipitation.
- Controlled the carbon equivalent to 4.2–4.3% and optimized inoculation to maximize graphitic carbon, reducing shrinkage tendency.
- Improved mold compaction by hand-ramming with wooden rods while filling the mold, increasing the mold stiffness and preventing mold wall movement.
For gas defects, I increased the number and diameter of vent holes in the cope, inserted half-vents (extending from the mold back to near the cavity but not through) to allow gases to escape, strictly limited resin addition (to keep binder content as low as possible while maintaining strength), controlled the loss-on-ignition and gas evolution of the sand, added iron oxide powder to the coating, and ensured thorough drying of both sand cores and mold coatings.
For slag inclusions, I used a stopper ladle to prevent primary slag from entering, applied cryolite powder on the melt surface to avoid secondary slag, changed the gating system to let most metal enter through the riser (hot end) and only a small amount through the far end to raise the pouring temperature there, increased pouring speed, used foundry coke for melting (lower sulphur), desulphurized the base iron to below 0.02%, and controlled pouring temperature in the range 1360–1400 °C.
These measures, refined over six months of intensive experimentation, eliminated all three defects. In the following two years, no crankshaft was scrapped due to shrinkage porosity.
Comparative Performance of Self-Hardening Sand vs. Clay-Bonded Dry Sand
To fully understand why self-hardening sand can be used successfully in sand casting foundry for large ductile iron parts, I conducted laboratory experiments and analyzed production data over five years. I compared the two mold materials in terms of cooling rate, hot strength, gas evolution, and sulphur contamination.
Cooling Rate
I cast test blocks of four different moduli (0.8 cm, 1.2 cm, 2.0 cm, 2.8 cm) using both self-hardening sand and clay-bonded dry sand. Each test block contained a thermocouple at its center, and I recorded the solidification time from the cooling curves. The results are plotted in Figure 1 (note: I will insert the image later). The data points on a log-log scale fit a straight line remarkably well, allowing me to derive a simple power-law relationship:
$$ t = k \cdot M^n $$
where t is solidification time (min), M is modulus (cm), and k and n are constants. For self-hardening sand, the solidification time was consistently longer than for dry sand. For example, for a crankshaft with a modulus of about 3.5 cm (typical for the 6L series), the estimated solidification time using dry sand was 45 min, whereas with self-hardening sand it was 75 min—a factor of 1.67. For a larger crankshaft (modulus ≈ 5.0 cm), the factor was about 2.0.
The slower cooling rate of self-hardening sand actually helps reduce shrinkage porosity, because it enhances the graphitization expansion stage, which can compensate for solidification shrinkage. However, it also delays the formation of a rigid outer shell. To prevent mold wall expansion, I had to increase the number and size of chills, as described earlier. This balance is critical in sand casting foundry practice.
The following table summarizes the measured solidification times for the test blocks:
| Modulus (cm) | Dry Sand t (min) | Self-Hardening Sand t (min) | Ratio (Self-Hardening/Dry) |
|---|---|---|---|
| 0.8 | 6.2 | 8.9 | 1.44 |
| 1.2 | 12.5 | 19.3 | 1.54 |
| 2.0 | 28.0 | 47.0 | 1.68 |
| 2.8 | 45.0 | 75.0 | 1.67 |
I also measured the density of the castings by the Archimedes method (using an analytical balance) and found no significant difference between dry sand and self-hardening sand castings. The densities were essentially identical, confirming that the increased chill cooling effectively prevented mold expansion.
Hot Strength and Mold Stiffness
Clay-bonded dry sand reaches its maximum hot strength at around 800–1000 °C, while self-hardening sand, due to resin decomposition above 300 °C, loses strength rapidly if it is exposed to air. However, if the sand is in a nearly confined condition (e.g., well-compacted and coated), the interior layers remain relatively cool and can retain sufficient strength. In my sand casting foundry, I achieved this by:
- Increasing ramming during molding to reduce porosity and limit oxygen penetration.
- Using a horizontal pouring/cooling layout that reduced the ferrostatic pressure compared to vertical cooling.
- Applying multiple layers of coating (zircon and graphite) on the core and mold surfaces.
I measured the mechanical properties of specimens cut from the body of crankshafts produced by both methods. The results are shown in Table 2.
| Parameter | Clay-Bonded Dry Sand (Metallurgical Coke, No Desulphurization) | Clay-Bonded Dry Sand (Foundry Coke, Desulphurized) | Self-Hardening Sand (Foundry Coke, Desulphurized) |
|---|---|---|---|
| Number of specimens | 50 | 60 | 80 |
| Tensile Strength (MPa): Range | 680–760 | 750–820 | 740–830 |
| Tensile Strength (MPa): Average | 715 | 785 | 790 |
| Elongation (%): Range | 1.5–2.5 | 2.0–3.0 | 2.0–3.5 |
| Elongation (%): Average | 2.0 | 2.5 | 2.7 |
| Hardness (HB): Range | 240–270 | 240–265 | 235–260 |
| Hardness (HB): Average | 255 | 250 | 248 |
The data indicate that when using high-quality foundry coke and desulphurization, the mechanical properties from self-hardening sand are at least as good as those from dry sand, with slightly higher average tensile strength and elongation.
Gas Evolution
Dry clay sand, after baking at 400 °C for several hours, contains almost no moisture; its gas evolution is negligible. In contrast, self-hardening sand evolves gas from the thermal decomposition of the resin binder. Typical gas evolution values for self-hardening sand are 15–20 mL/g, whereas well-baked dry sand is below 2 mL/g. Moreover, the gas evolution rate is much faster: at 1400 °C, self-hardening sand reaches its maximum gas volume within 5–10 seconds. This can cause “choking” if the venting area is insufficient. I calculated the required vent area using Darcy’s law for gas flow through porous media, but a simpler empirical rule I developed is:
$$ A_{\text{vent}} = 2.5 \times V_{\text{cavity}}^{0.7} $$
where A is total vent cross-section area (cm²) and V is mold cavity volume (dm³). For a typical 6L crankshaft (cavity volume ~200 dm³), this gives about 120 cm² of vent area, compared to only 20 cm² that would have sufficed for dry sand. I also inserted half-vents (blind vents) to increase effective permeability without breaking through the cavity surface.
Sulphur Content and Surface Degradation
Self-hardening sand typically contains 0.3–0.6% sulphur from the acid catalyst (p-toluenesulfonic acid or sulfuric acid ester). Some literature warns that this sulphur can diffuse into the ductile iron surface, causing nodule deterioration. However, in my five years of experience with sand casting foundry, I have never observed such a graphite degeneration layer on any crankshaft, even after deep etching examined under microscope. I attribute this to:
- Heavy coating: zircon wash (first layer) and graphite wash (second layer) applied on both molds and cores, acting as a sulphur barrier.
- Large section size: the total liquid metal volume is large, and the magnesium content in the iron (using a proprietary high-Mg low-RE nodulizer) is sufficient to neutralize any local sulphur pickup through diffusion.
- The limited amount of sulphur available from the sand (only the surface layer, typically <1 mm thick, is affected) and the rapid solidification of the skin due to chills.
I verified this by taking step samples of different thicknesses from castings and polishing the surfaces. No significant degradation was detected even on the thinnest section (15 mm). I also measured the sulphur profile using EDS across the cross-section; the sulphur level at the surface was only 0.02 wt%, not higher than the interior. Therefore, I consider this concern to be negligible in practical sand casting foundry operations with proper coating and nodulizer design.
Summary of Key Process Parameters for Successful Self-Hardening Sand Casting of Large Ductile Iron Crankshafts
Based on my experience, I compiled the following checklist for any sand casting foundry planning to adopt self-hardening sand for heavy ductile iron parts:
| Parameter | Recommended Value / Practice |
|---|---|
| Binder (furan resin) content | 0.8–1.2% based on sand weight; adjust to maintain tensile strength > 0.8 MPa after 24 h |
| Catalyst (p-toluenesulfonic acid) | 40–60% of resin weight, depending on ambient temperature |
| Mold compaction | Hand-ramming or vibration to achieve density > 1.6 g/cm³ |
| Coating | Two coats: first zircon (0.05–0.10 mm dry thickness), second graphite (0.10–0.15 mm dry thickness); dry thoroughly at 200 °C for at least 2 h |
| Chills | Cast iron or steel chills, size 20–40 mm thick, placed on all main journal necks, crank pin journals, and especially near riser contact area; chill coverage ≥ 60% of surface area |
| Pouring temperature | 1360–1400 °C; for far end temperature, ensure > 1300 °C at entry |
| Pouring speed | For a 200 kg crankshaft, fill time 12–15 seconds; for 400 kg, 20–25 seconds |
| Vent area (total) | A_vent = 2.5 × V_cavity^0.7 (cm²); also add half-vents every 100 mm in cope |
| Desulphurization of base iron | S < 0.020% before nodulization |
| Nodulizer | High-Mg low-RE (Mg 5–7%, RE 1–2%, Cu 1%), addition 1.0–1.2% |
| Inoculation | 0.3–0.5% FeSi75 (0.2–0.7 mm), late stream inoculation during pouring |
Conclusion
Through my extensive work in a sand casting foundry, I have demonstrated that furan resin self-hardening sand can successfully replace clay-bonded dry sand for producing large ductile iron crankshafts, provided that adequate measures are taken to compensate for its lower hot strength, slower cooling rate, higher gas evolution, and sulphur content. By implementing a horizontal pour/horizontal cool process with suitable chills, controlled carbon equivalent, extensive venting, and robust coatings, I have achieved internal quality, mechanical properties, and density equivalent to or better than those from dry sand molds. The surface finish and dimensional accuracy are markedly superior. The overall scrap rate decreased by more than 50% after the transition.
I hope that the detailed data, tables, and formulas presented here will assist other sand casting foundry engineers in optimizing their own processes for large ductile iron castings using self-hardening sand. The key is to understand the fundamental differences in solidification behavior and to engineer the mold and process accordingly.