In our production of as-cast ductile iron connecting rods for light vehicle engines, we have encountered various sand casting defects that significantly affect the mechanical performance and reliability of the final product. The connecting rod is a thin-walled, small-component casting weighing approximately 590 g, with a large end wall thickness of 3.0–7 mm, a small end wall thickness of 3.0–7 mm, and a minimum web thickness of only 4 mm. The abrupt section changes at the junction between the rod shank and the large end, as well as between the shank and the small end, create hot spots prone to shrinkage cavities and porosity. Additionally, the thin sections are susceptible to gas porosity and inclusions. Through multiple experimental trials and process optimizations, we have systematically analyzed the root causes of these sand casting defects and implemented effective countermeasures. This article presents our findings in a structured manner, focusing on shrinkage-related defects, gas porosity, and inclusions, with detailed tables and mathematical formulations to guide defect prevention.
Shrinkage Cavity and Porosity
Shrinkage defects are among the most critical sand casting defect categories for ductile iron connecting rods. The localized metal accumulation at the shank-large end and shank-small end junctions creates thermal centers that solidify last, leading to macro-shrinkage cavities or micro-shrinkage porosity. The formation mechanism is governed by the carbon equivalent (CE), inoculation effectiveness, and mold rigidity. We controlled the CE within a narrow range of 4.5%–4.8% to balance graphitization expansion and liquid contraction. The CE is computed as:
$$
CE = C + \frac{Si}{3}
$$
If CE is too low (<4.5%), insufficient graphite precipitation increases solidification contraction and reduces iron liquidity, thereby worsening shrinkage tendency. Conversely, if CE exceeds 4.8%, excessive graphitization expansion forces the green sand mold to deform (since the squeeze-line mold has limited rigidity), causing mold wall movement and enlarging the cavity, which then leads to shrinkage porosity despite the expansion. Therefore, precise CE control is essential to minimize this sand casting defect.
Inoculation plays a vital role. Inadequate inoculation results in massive carbide precipitation, which increases contraction because carbides have a higher density than graphite. We adopted a composite inoculation method: primary inoculation in the spheroidizing ladle and secondary inoculation during pouring. This ensured sufficient nucleation sites for graphite, promoting a eutectic solidification mode with self-feeding characteristics.
Two gating system designs were tested to address hot-spot shrinkage. The first was a top-gating side-riser scheme (as described in our earlier trials) where molten iron enters from the top of the large end, allowing early stabilization and earlier graphitization expansion. The small end core hole was changed from through-hole to blind-hole design to shift the hot spot to the center hole, which could later be removed by machining. The second approach was a vertical bottom-gating system with the ingate at the small end weight-reduction boss, ensuring smooth upward filling and utilizing the small end riser pressure plus graphitization expansion for feeding, while a separate riser at the large end provided supplemental hot metal. Both methods effectively eliminated shrinkage cavities in production.
Table 1 summarizes the causes and countermeasures for shrinkage-related sand casting defect.
| Cause | Countermeasure |
|---|---|
| Low carbon equivalent (<4.5%) | Maintain CE = 4.5%–4.8% |
| High carbon equivalent (>4.8%) causing mold expansion | Control CE upper limit and improve mold rigidity |
| Insufficient inoculation | Composite inoculation: ladle + stream inoculation |
| Inadequate feeding at hot spots | Use top-gating side-riser or vertical bottom-gating with dual risers |
| Through-hole at small end intensifies hot spot | Redesign to blind-hole to shift defect to machinable area |
Gas Porosity
Gas porosity is a common sand casting defect in thin-walled ductile iron castings. For our connecting rod, gas pores can appear as surface blowholes, subsurface pinholes, or internal gas cavities. The rod shank, with only 4 mm thickness, is particularly vulnerable because the high cooling rate traps gas bubbles before they can escape. We classified gas porosity into three types and addressed each with specific process controls.
Surface and Subsurface Pinholes
These defects often result from nitrogen, hydrogen, or oxygen entrapped during mold filling. We strictly controlled green sand permeability between 100 and 140 (AFS). Permeability below 100 increases the risk of mold gas entrapment, while above 140 causes metal penetration into sand pores, creating rough surfaces and metallic penetration defects. The mold surface hardness was maintained at 85–95 to balance compaction and venting.
The vertical bottom-gating system was adopted to promote laminar filling from bottom to top, allowing mold gases to escape through vents and the top riser. This design significantly reduced entrainment of mold gases compared to top-gating.
We also minimized the gas content in the molten metal. All charge materials were clean and dry, free from moisture, oil, or rust. Inoculants and nodulizers were preheated before use. The pouring temperature was controlled: the first ladle temperature ≥1400°C, and the last pouring temperature ≥1320°C. Higher temperatures improve bubble floatation and re-melting of any cold metal splashes.
Subsurface Pinholes (Reaction Pinholes)
The most troublesome sand casting defect was subsurface pinholes caused by the reaction between magnesium and water vapor at the mold-metal interface. Magnesium in the nodularizing alloy reduces water vapor:
$$
\mathrm{Mg} + \mathrm{H_2O} \rightarrow \mathrm{MgO} + 2[\mathrm{H}]
$$
The atomic hydrogen dissolves into the iron layer adjacent to the mold, and during solidification, it precipitates to form pinholes just beneath the casting surface. The critical residual aluminum content for wet-sand ductile iron is 0.030%–0.050% — within this range the reaction rate is highest. To avoid this, we reduced the addition of nodulizer and inoculant (both contain Al) and limited their Al content. Residual Al was kept below 0.030%.
Another reaction involves magnesium sulfide (MgS) slag. If not removed completely, MgS floats to the interface and reacts with water:
$$
\mathrm{MgS} + \mathrm{H_2O} \rightarrow \mathrm{MgO} + \mathrm{H_2S}\uparrow
$$
The H₂S gas forms pinholes surrounded by poorly spheroidized graphite (often flake graphite). We adopted double slagging (first after spheroidization, second after transfer to pouring ladle) and used a more effective slag coagulant instead of wood ash to achieve clean metal.
We also optimized molding sand moisture. Statistical data from our production line showed that when moisture exceeded 5%, the subsurface pinhole rejection rate increased dramatically. We set the moisture target at 3.8%–5.0% (higher in summer/autumn, lower in winter/spring). Although foreign literature suggests 3.3%–3.8% for squeeze-line molding, our equipment limitations required the broader range; the results were still acceptable.
Adding coal dust (3%–5% by mass) to the molding sand helped. During pouring, coal dust decomposes to form lustrous carbon, which deposits on the sand–metal interface, blocking the reaction between magnesium and water. However, excessive coal dust (>8%) increases gas evolution and may cause other sand casting defect like blowholes and surface carburization.
Table 2 summarizes the causes and countermeasures for gas porosity.
| Type | Cause | Countermeasure |
|---|---|---|
| Surface/entrapped gas | Low permeability, high moisture, turbulent filling | Permeability 100–140; bottom-gating; pouring T≥1320°C |
| Subsurface pinhole (reaction) | Mg + H₂O → H; residual Al in 0.030%–0.050% | Keep Al <0.030%; double slagging; moisture 3.8%–5.0%; coal dust 3%–5% |
| Subsurface pinhole (MgS slag) | MgS + H₂O → H₂S; incomplete slag removal | Double slagging; use effective slag coagulant; reduce nodulizer addition |
| Low pouring temperature | Bubbles cannot escape | First pouring ≥1400°C; last pouring ≥1320°C |
Inclusions
Inclusion defects in our connecting rod production manifested as metallic shot (iron beads), unmelted alloy particles, and non-metallic slag entrapment. Each type had distinct formation mechanisms and required targeted process modifications to eliminate this sand casting defect.
Metallic Shot (Iron Beads)
Iron beads form when molten iron splashes during mold filling, forming droplets that oxidize and become enveloped by a thin oxide film. The oxide reacts with carbon in the iron:
$$
\mathrm{FeO} + \mathrm{C} \rightarrow \mathrm{Fe} + \mathrm{CO}\uparrow
$$
The CO gas tries to escape but the droplet solidifies, creating a porous bead. To prevent this, we redesigned the gating system by adding a splash cup (well) at the bottom of the sprue to absorb the kinetic energy of the falling metal. We also strictly controlled sand moisture to avoid steam explosions that can cause splashing. Pouring temperature was kept high (≥1400°C) to allow any small droplets that do form to re-melt upon contact with the bulk metal.
Unmelted Alloy Particles (Metallic Inclusions)
Nodulizers and inoculants that are too coarse may not fully dissolve in the molten iron, leaving particles that become entrapped in the casting. Conversely, particles that are too fine burn off excessively, generating oxides that also act as inclusions. We selected an optimal size distribution: nodulizer (FeSiMg8RE3) 5–12 mm, inoculant (75FeSi) 2–6 mm. This ensured complete dissolution while minimizing oxidation losses. The tapping temperature was maintained above 1450°C to promote rapid melting.
Slag/Dross Inclusions
Non-metallic inclusions (dross) originate from melting slag, furnace/bowl lining erosion, and reaction products from nodularization. The most effective countermeasure was the vertical gating system with integrated slag traps. The figure below (inserted earlier) shows the design: molten iron enters from the small end, flows upward through the cavity, and carries slag into the blind riser at the large end. Additionally, we placed ceramic foam filters (pore size 0.8 mm) in the gating system to capture coarse inclusions. Melt cleanliness was improved by: (1) using clean charge materials, (2) allowing a 5-minute settling time after tapping to float out non-metallic particles, (3) raking slag thoroughly before pouring, and (4) covering the melt surface with a slag coagulant. The pouring ladle was a teapot-type design to prevent slag from entering the pouring stream.
Table 3 summarizes the inclusion-related sand casting defect and countermeasures.
| Type | Cause | Countermeasure |
|---|---|---|
| Iron beads (metallic shot) | Metal splashing; low temperature; high moisture | Splash well at sprue base; T≥1400°C; moisture control |
| Unmelted alloy particles | Oversize nodulizer/inoculant; low tapping temperature | Optimal particle size (nodulizer 5–12mm, inoculant 2–6mm); Ttap≥1450°C |
| Slag/dross | Slag from melt; lining erosion; reaction products | Vertical gating with slag trap; ceramic foam filter; slagging; teapot ladle |
Process Integration and Results
By integrating all these countermeasures — precise carbon equivalent control, composite inoculation, vertical bottom-gating with slag traps, moisture and permeability management, and careful handling of nodulizer and inoculant — we were able to drastically reduce the rejection rate due to sand casting defect. Our production data showed that shrinkage-related defects dropped from 12% to <1%, gas porosity from 8% to <0.5%, and inclusion defects from 5% to negligible levels. The final connecting rods passed rigorous mechanical testing (tensile strength ≥700 MPa, elongation ≥3%, hardness 217–269 HB) and microstructural examination (nodularity ≥80%, ferrite content ≥70%, no carbides or flake graphite). The elimination of these sand casting defect also resolved the potential safety hazard of fatigue fracture during engine operation.
We continue to monitor the process through statistical process control (SPC) charts for key parameters such as CE, moisture, permeability, and pouring temperature. Any deviation triggers immediate corrective action. This systematic approach has made the production of ductile iron connecting rods for light vehicle engines both reliable and economical.
Conclusion
The sand casting defect of shrinkage cavities, gas porosity, and inclusions in ductile iron connecting rods can be effectively controlled through a combination of metallurgical and process engineering measures. Key actions include: maintaining carbon equivalent within 4.5%–4.8%; using composite inoculation; adopting a vertical gating system with risers at both ends; controlling green sand moisture (3.8%–5.0%) and permeability (100–140); minimizing residual aluminum (<0.03%); double slagging; using coal dust (3%–5%); and implementing filtration and slag traps. These measures have proven successful in our plant, ensuring the production of defect-free connecting rods suitable for demanding automotive applications. Continuous monitoring and refinement are essential to maintain low defect levels and adapt to changes in raw materials and equipment.