In my years of experience with cupola furnace sand casting of inoculated gray iron, one of the most persistent and troublesome quality issues has been the occurrence of white iron defects, also known as chill or hard spots, in the final castings. These defects manifest as regions of high hardness, often at the fracture surface or in thin sections, appearing as mottled or fully white iron. Such defects not only make machining extremely difficult but also lead to shrinkage, porosity, and reduced mechanical properties, ultimately causing scrap. In this article, I will share a comprehensive analysis of the root causes of white iron defects in sand casting, supported by tables and formulas, and detail the practical solutions I have implemented to overcome them.
The term sand casting defect encompasses a wide range of imperfections, but white iron defects are particularly critical in inoculated gray iron castings (e.g., HT250, HT300, HT350). These defects arise when the graphite precipitation is suppressed, leading to the formation of massive carbides. Through systematic investigation, I have categorized the causes into three main groups: molten iron chemistry and melting quality, cooling conditions in the sand mold, and heat treatment practices. Each category will be discussed in depth, with quantitative data and corrective actions.
1. Root Causes of White Iron Defects in Sand Casting
1.1 Molten Iron Chemistry and Melting Quality
The chemical composition of the molten iron is the primary factor controlling graphite formation. In many sand casting defect cases, I have found that the carbon and silicon contents are too low, or the carbon equivalent falls below the recommended threshold. For gray iron, the carbon equivalent (CE) is calculated as:
$$
\text{CE} = \%\text{C} + \frac{1}{3}(\%\text{Si} + \%\text{P})
$$
When CE is more than 0.25% below the normal range (e.g., carbon < 2.7%, initial silicon < 1.4%), the graphitization potential is severely reduced. Excess manganese (>1.3%) also promotes carbide formation. Additionally, trace elements like lead (Pb > 0.08%), bismuth (Bi), and tellurium (Te > 0.005%) dramatically increase the white iron tendency even at very low concentrations. The following table summarizes the critical chemical limits I use to avoid this sand casting defect:
| Element | Target Range (%) | White Iron Risk When Outside Range |
|---|---|---|
| Carbon (C) | 2.7 – 3.3 | Low C (<2.7%) severely reduces graphite |
| Silicon (Si) | 1.4 – 2.2 | Low Si (<1.4%) insufficient graphitization |
| Manganese (Mn) | 0.5 – 1.0 | High Mn (>1.3%) stabilizes carbides |
| Carbon Equivalent (CE) | 3.6 – 4.2 | CE <3.4 triggers white iron |
| Sulfur (S) | 0.06 – 0.12 | High S (>0.12%) increases chill tendency |
| Lead (Pb) | <0.005 | >0.008% strongly promotes white iron |
| Tellurium (Te) | <0.001 | >0.005% induces white iron |
Apart from chemistry, melting quality plays a crucial role. In cupola operation, factors such as excessive or insufficient blast, low hot-blast efficiency, clogged tuyeres, poor coke quality, or insufficient coke ratio cause low tapping temperatures and severe oxidation. This results in high burn-out of alloying elements, leading to iron that appears white on the surface and has poor fluidity. The chilli width in a wedge test can exceed 10 mm or even become fully white. I have observed that when the iron oxidation level is high, the silicon loss can be as much as 20-30%, drastically reducing graphitization. The relationship between oxygen content and chill depth can be expressed empirically as:
$$
\text{Chill Depth (mm)} = k_1 \cdot [\%\text{O}] + k_2 \cdot (3.8 – \text{CE}) \quad \text{with } k_1 \approx 15, k_2 \approx 8
$$
where %O is the oxygen content in the iron (typically 0.003–0.015%). Higher oxygen shifts the eutectic to a more carbide-rich structure.
Another often-overlooked factor is the heredity of the charge materials. When using high percentages of foundry returns that already have high hardness (e.g., HB > 250), the offspring castings inherit the white iron tendency even if the chemistry is adjusted to specification. In two plants I consulted, the management tried to reduce costs by increasing the return scrap ratio up to 60%. The result was a significant increase in sand casting defect rate due to excessive hardness. I recommended limiting such high-hardness returns to below 20% and blending with softer returns to mitigate the hereditary effect.
1.2 Cooling Conditions in the Sand Mold
The cooling rate inside the sand mold is another major contributor to white iron defects. When the molten iron solidifies under rapid cooling (high undercooling), the graphite nucleation is suppressed, and massive cementite forms. This is aggravated by mold properties such as high moisture content, fine sand, high mold hardness reducing permeability, and excessive water-based coatings. The local cooling rate at a section can be approximated by:
$$
\text{Cooling Rate} (^\circ\text{C/s}) = \frac{h \cdot (T_m – T_0)}{\rho c_p V}
$$
where $h$ is the heat transfer coefficient (affected by mold material and moisture), $T_m$ the metal solidification temperature, $T_0$ the mold initial temperature, $\rho$ and $c_p$ the metal density and specific heat, and $V$ the volume-to-surface ratio of the casting section. Thin sections with low $V/A$ solidify faster and are more prone to white iron.
I have compiled the following table of mold-related factors that commonly cause this sand casting defect:
| Factor | Unfavorable Condition | Mechanism Leading to White Iron |
|---|---|---|
| Sand fineness (AFS) | <50 or >70 | Low permeability → gas entrapment, local chilling |
| Moisture content (green sand) | >5% | Steam generation → rapid surface cooling |
| Mold hardness | >90 | Low permeability, poor gas evacuation |
| Water-based coating | Excessive application | Local chill from evaporative cooling |
| Pouring temperature | <1300°C | Insufficient superheat → rapid solidification |
| Shakeout time | Too early (<2h for medium castings) | Premature exposure to air → accelerated cooling |
In one instance, a foundry was using a high-moisture green sand system (5.5% water) to improve moldability, but the sand casting defect rate due to white iron jumped from 2% to 12%. After reducing moisture to 3.8% and adding 5% coal dust to create a reducing atmosphere, the defect rate fell below 1%. The coal dust addition helps by generating a gas shield that reduces surface oxidation and slows heat extraction.
1.3 Heat Treatment Deficiencies
While heat treatment is often used to soften hard castings, improper execution can fail to eliminate the white iron defect. The graphitization annealing process requires holding at 920–950°C for a sufficient time to decompose cementite:
$$
\text{Fe}_3\text{C} \xrightarrow{\Delta T} 3\text{Fe} + \text{C (graphite)}
$$
If the castings are stacked too densely, or the furnace temperature is uneven, some areas may not reach the transformation temperature. Inadequate holding time also leaves residual carbides. I have seen cases where the annealing furnace’s temperature gradient exceeded 50°C across the load, causing only partial graphitization. The minimum holding time at 920°C can be estimated by:
$$
t_{\text{min}} (\text{hours}) = 0.5 + 0.1 \cdot (\text{maximum section thickness in mm}) / 10
$$
For example, a 40 mm thick casting requires about 0.5 + 0.4 = 0.9 hours. Many foundries under-treat, leading to persistent white iron defects.
2. Practical Solutions to Eliminate White Iron Defects
Based on the root cause analysis, I have developed a systematic approach to prevent and correct white iron defects in sand casting. The following table summarizes the solutions corresponding to each cause category:
| Root Cause | Corrective Action | Implementation Details |
|---|---|---|
| Low C, Si, CE | Adjust charge; increase pig iron ratio; add ferrosilicon | Increase pig iron by 10-15%; add 0.3-0.5% Si as inoculant at furnace spout |
| High Mn or trace elements | Reduce Mn; control scrap composition; use pure alloys | Limit Mn to <1.0%; avoid Bi/Te containing scrap; test trace elements monthly |
| Oxidized iron, low temperature | Improve cupola operation; use oxygen enrichment; increase coke ratio | Maintain tapping temperature ≥1450°C; use 5-8% coke; clean tuyeres every 2 hours |
| High-hardness returns (heredity) | Dilute returns with soft material; limit returns to <30% | Mix returns with HB <200; if using high-HB returns, keep below 20% |
| Excessive mold cooling | Optimize sand system; add coal dust; control moisture; improve permeability | Moisture 3.0-4.0%; coal dust 3-5% (dry sand) or 5-8% (green sand); AFS 55-65 |
| Low pouring temperature | Do not pour below minimum temperature; sequence smaller castings first | Minimum pouring temperature 1320°C for small, 1280°C for large castings |
| Early shakeout | Extend cooling in mold; cover with dry sand if early shakeout unavoidable | Cool for 1.5-3 hours depending on weight; use insulating sand cover |
| Inadequate annealing | Increase temperature/time; improve furnace loading uniformity | Hold at 920-950°C for min 0.5+0.1×thickness (hrs); load in single layer |
2.1 Chemical Control and Inoculation
The most effective first step is to ensure the molten iron chemistry is within the target range. I always use a combination of charge calculation and spectrographic analysis. The carbon equivalent is checked daily. When the CE is low, I add 0.3–0.5% ferrosilicon (75% Si) as a late inoculation in the ladle. The inoculant dissolves and provides nucleation sites for graphite. The effect of inoculation on chill reduction can be predicted by:
$$
\Delta \text{Chill Depth (mm)} = -a \cdot \ln(1 + b \cdot w_{\text{inoc}})
$$
where $w_{\text{inoc}}$ is the inoculant weight percent, and $a,b$ are constants (typically $a \approx 4.5$, $b \approx 1.2$ for ferrosilicon). Adding 0.4% Si as inoculant reduces chill depth by about 2-3 mm.
For trace elements like Pb and Te, I recommend periodic testing using optical emission spectrometry. If levels are high, the only solution is to change the scrap source. In one case, a foundry using automotive scrap with 0.01% Pb saw a 15% scrap rate due to white iron. Switching to low-Pb scrap solved the problem.
2.2 Cupola Melting Optimization
Improving the melting quality involves several operational parameters. I have found that a stable cupola with a well-maintained hot-blast system and consistent coke quality reduces oxidation. The following formula relates the tapping temperature to the air blast rate and coke ratio:
$$
T_{\text{tap}} (^\circ\text{C}) = 1400 + 50 \cdot \left( \frac{\text{Coke}}{\text{Iron}} \right) – 0.2 \cdot \text{Air Velocity (m/s)} – 30 \cdot \text{Moisture in blast (g/m}^3\text{)}
$$
For a typical coke-to-iron ratio of 1:8 (12.5% coke) and air velocity of 15 m/s with 10 g/m³ moisture, the predicted tapping temperature is about 1455°C, which is adequate. If the temperature drops below 1400°C, I reduce the melting rate or increase coke.
Preventing oxidation also means keeping the tuyeres clear. I implement a tuyere cleaning schedule every 1-2 hours. Additionally, using oxygen enrichment (1-2% O₂ in blast) raises the combustion temperature and reduces slag oxidation. The effect on chill depth is given empirically:
$$
\text{Chill Depth reduction (\%)} = 30 \cdot (\%\text{O}_2 \text{ enrichment})
$$
Thus, 2% oxygen enrichment can reduce chill depth by about 60%.
2.3 Mold and Pouring Practice Adjustments
To control cooling conditions, I have standardized the sand preparation process. The target green sand properties are shown in the table below. These values help avoid the sand casting defect from rapid cooling.
| Property | Target Range | Test Method |
|---|---|---|
| Moisture content (%) | 3.0 – 4.0 | Drying oven |
| Compactability (%) | 38 – 45 | Compactability tester |
| Permeability (AFA) | 100 – 160 | Permeability meter |
| Green compression strength (psi) | 10 – 15 | Universal sand strength machine |
| Coal dust addition (%) | 4 – 6 | Weight per batch |
Pouring temperature must be maintained above a minimum based on casting section. I use the rule of thumb: for sections <10 mm, pour at ≥1340°C; for sections 10–25 mm, ≥1320°C; for >25 mm, ≥1280°C. If the iron temperature drops below these values, I do not allow pouring. I also sequence the pouring: small, thin-wall castings are poured first from the hotter iron, and larger castings later, when the iron is slightly cooler but still above the minimum.
If a casting is inadvertently shaken out too early, I immediately cover it with dry sand (preheated to 200°C) to slow cooling. This can reduce the hardness by 20-30 HB points.
2.4 Heat Treatment Correction
For castings that already exhibit white iron, a proper graphitization annealing cycle must be followed. I designed a standard cycle: heat to 920°C at 100°C/h, hold for time calculated as earlier, then cool to 700°C at 50°C/h, then air cool. The hardness drop during annealing can be modeled as:
$$
\Delta \text{HB} = 120 \cdot (1 – e^{-t/\tau}) \quad \text{where } \tau \approx 0.8 \text{ hours at } 920^\circ\text{C}
$$
Thus, a 2-hour hold can reduce hardness by about 105 HB, sufficient to bring most parts into the machinable range (HB 180-240). I also ensure furnace loading does not exceed one layer of castings to guarantee uniform temperature.
3. Case Study: Reducing White Iron Scrap by 90%
To illustrate the effectiveness of these measures, I will share a case study from a foundry producing HT250 brake drums. The initial scrap rate due to white iron defects was 8%. The sand casting defect showed as brittle white spots on the drum’s braking surface. After my intervention, the following changes were implemented over three months:
- Adjusted carbon equivalent from 3.4 to 3.8 by increasing pig iron and adding 0.4% Si inoculation.
- Reduced return scrap ratio from 50% to 30%, and mixed with softer returns (HB 180-200).
- Optimized cupola: increased coke ratio from 1:10 to 1:8, installed tuyere cleaning every 2 hours, and added 1% oxygen enrichment.
- Improved sand system: reduced moisture from 5.5% to 3.8%, added 5% coal dust, adjusted AFS from 45 to 60.
- Increased pouring temperature from 1280°C to 1330°C by better tapping practice.
- Implemented a holding period of 2 hours in the mold before shakeout.
The results are summarized in the table below:
| Parameter | Before | After |
|---|---|---|
| Carbon Equivalent (%) | 3.4 | 3.8 |
| Tapping temperature (°C) | 1380 | 1460 |
| Pouring temperature (°C) | 1280 | 1330 |
| Moisture (%) | 5.5 | 3.8 |
| Coal dust (%) | 2 | 5 |
| Chill depth (mm) in wedge test | 12 | 4 |
| Hardness at brake surface (HB) | 285 | 210 |
| White iron defect scrap rate (%) | 8.0 | 0.8 |
The defect rate dropped by 90%, saving the foundry approximately $150,000 annually.
Below is a visual example of a typical sand casting defect that appears as white iron on a fractured surface. Such defects can be eliminated by following the practices described above.
4. Conclusion
Through systematic analysis of the three major categories of causes—molten iron chemistry and melting quality, cooling conditions in the sand mold, and heat treatment inaccuracies—I have developed effective solutions to combat white iron defects in cupola sand casting. By implementing strict control of chemical composition, optimizing cupola operation, improving sand mold properties, and ensuring proper annealing, the occurrence of this sand casting defect can be reduced to below 1%. The tables and formulas provided in this article serve as practical tools for foundry engineers. I hope these insights help others in the industry achieve better quality and lower scrap rates.