In our foundry operations, we have long grappled with casting defects in ductile iron, particularly slag inclusion, subcutaneous porosity, shrinkage cavities, and micro-shrinkage. These issues are intimately linked to the residual magnesium content in the molten iron. The adoption of pressure magnesium addition for nodularization significantly improved magnesium utilization efficiency. However, this advancement inadvertently led to excessively high residual magnesium levels, which in turn exacerbated the prevalence of defects like slag inclusion. This prompted a comprehensive investigation aimed at reducing the magnesium addition, precisely controlling the residual magnesium content, and ultimately minimizing slag inclusion defects. Our systematic trials, conducted under standard production conditions, have successfully developed methods to lower magnesium addition to below 0.15% and effectively control slag inclusion.
The fundamental principle for determining magnesium addition hinges on several factors: the original sulfur content of the base iron, the minimum residual magnesium required for complete graphite spheroidization, and the magnesium loss during treatment and pouring. In practice, due to the multitude of variables affecting magnesium oxidation loss, precise control is challenging. To avoid the costly scrap losses associated with incomplete nodularization, a common practice has been to increase the magnesium addition, often overlooking the consequential rise in casting defects—slag inclusion, subcutaneous blowholes, shrinkage, and porosity. Therefore, we assert that the magnesium addition should be minimized strictly to the level that guarantees complete graphite spheroidization, with the residual magnesium in the iron preferably not falling below 0.03%. Given that different types of nodularizing agents and treatment methods yield vastly different magnesium utilization rates even at the same addition level, the optimal magnesium addition must be determined based on specific production conditions.
Our production environment utilizes acid cupola furnaces for melting. The charge materials include pig iron, returns of ductile iron castings, ferromanganese (75% Mn), ferrosilicon, magnesium ingots (99.5% Mg), and silicon-calcium alloy (60% Si, 30% Ca). The nodularization treatment is performed using a bell-type pressure addition ladle, available in capacities of 500 kg and 1500 kg, as illustrated conceptually below.
The design and operation of this pressure ladle are critical for achieving high magnesium efficiency. We found that a ladle design with a high H/D ratio (height-to-diameter ratio) is advantageous. The “bottle-neck” shape near the mouth reduces the total pressure on the sealing lid, facilitating clamping and sealing, while also promoting fuller interaction between magnesium and the iron melt. Furthermore, using a cast copper bell instead of cast iron prevents premature melting, extends the reaction time, and minimizes magnesium floatation, thereby enhancing utilization.
The charge composition is tailored to promote high carbon and low silicon content, approaching a hypereutectic composition, which improves the iron’s capacity for magnesium absorption and aids graphite spheroidization. A representative charge makeup is presented in Table 1.
| Material | Pig Iron | Ductile Iron Returns | Ferromanganese | Ferrosilicon | Other Alloys |
|---|---|---|---|---|---|
| Percentage (%) | 60-70 | 30-40 | 0.5-1.0 | 1.0-1.5 | As required |
The process flow involves careful control of melting, magnesium treatment, and post-inoculation. The key to controlling and reducing residual magnesium lies in the nodularization treatment step. To minimize magnesium oxidation loss and ensure thorough contact, we strictly control the freeboard volume (the space above the molten iron surface) in the pressure ladle and ensure sufficient immersion depth of the magnesium bell. Operating under these principles, we systematically varied the magnesium addition between 0.10% and 0.20% for treatment. The relationship between added magnesium and resulting residual magnesium is crucial and can be expressed conceptually. The residual magnesium ($Mg_{res}$) is a function of added magnesium ($Mg_{add}$), original sulfur ($S_o$), final sulfur ($S_f$), and process losses ($L$).
$$ Mg_{res} = f(Mg_{add}, S_o, S_f, L) $$
A significant portion of added magnesium is consumed for desulfurization. The stoichiometric relationship for sulfur removal can be approximated. For every 0.01% of sulfur removed, approximately 0.0075% of magnesium is consumed, though this varies with conditions.
$$ Mg_{consumed\ for\ S} \approx k \cdot (S_o – S_f) $$
where $k$ is an empirical coefficient, typically around 0.75 for our conditions (i.e., 0.0075% Mg per 0.01% S).
Therefore, the effective magnesium available for nodularization is the added amount minus the sum of losses and desulfurization consumption. The magnesium utilization efficiency ($\eta$) is a critical performance metric.
$$ \eta = \frac{Mg_{res}}{Mg_{add}} \times 100\% $$
However, a more detailed model accounting for desulfurization gives:
$$ \eta’ = \frac{Mg_{res} + Mg_{consumed\ for\ S}}{Mg_{add}} \times 100\% $$
This represents the total magnesium absorbed/used by the melt relative to addition. In our trials, we focused on the simple residual-based efficiency $\eta$. By fixing most melting and treatment parameters within narrow ranges and varying only the magnesium addition, we established a predictable pattern. Table 2 summarizes production data from controlled trials, showing how residual magnesium tracks with addition.
| Heat No. | Iron Weight (kg) | Mg Addition ($Mg_{add}$, %) | Original S ($S_o$, %) | Final S ($S_f$, %) | Residual Mg ($Mg_{res}$, %) | Mg Utilization ($\eta$, %) | Desulfurization Mg Consumed (%, approx.) |
|---|---|---|---|---|---|---|---|
| 1 | 500 | 0.20 | 0.030 | 0.010 | 0.055 | 27.5 | 0.015 |
| 2 | 500 | 0.18 | 0.028 | 0.009 | 0.050 | 27.8 | 0.014 |
| 3 | 500 | 0.16 | 0.032 | 0.011 | 0.045 | 28.1 | 0.016 |
| 4 | 1500 | 0.14 | 0.029 | 0.010 | 0.040 | 28.6 | 0.014 |
| 5 | 1500 | 0.12 | 0.027 | 0.008 | 0.035 | 29.2 | 0.014 |
| 6 | 1500 | 0.10 | 0.025 | 0.007 | 0.030 | 30.0 | 0.014 |
The data illustrates that as magnesium addition decreases, residual magnesium decreases in a roughly linear manner under constant sulfur levels and treatment efficiency. The utilization efficiency $\eta$ shows a slight increasing trend with lower additions, possibly due to reduced oxidation losses in smaller absolute quantities. Controlling the original sulfur content is foundational for reducing residual magnesium requirements. Using low-sulfur charge materials and optimizing cupola operation (e.g., installing a hot blast system) can lower base sulfur. Table 3 contrasts sulfur levels before and after such an upgrade.
| Condition | Cupola Furnace Number | Blast Temperature (°C) | Average Original Sulfur Content ($S_o$, %) | Remarks |
|---|---|---|---|---|
| Before Retrofit | Multiple heats | Ambient (~20) | 0.030 – 0.035 | High coke consumption |
| After Hot Blast Retrofit | Same furnaces | 250 – 300 | 0.020 – 0.025 | Reduced coke use, higher tap temperature |
Post-inoculation, typically with ferrosilicon, must ensure stable and high silicon recovery with uniform distribution in the iron. We found that adding inoculant to the ladle after slag-off, followed by thorough stirring, yields a recovery rate of around 80-85%, compared to lower and erratic recovery when added in the runner during tapping.
A critical aspect of our work involved dedicated trials to quantify the reduction of slag inclusion defects. The tendency for slag inclusion formation is strongly influenced by residual magnesium. Higher levels promote more vigorous oxidation and slag formation. We conducted production-scale tests where magnesium addition was reduced to 0.10-0.12%, and in some cases, small amounts of silicon-calcium alloy were added post-treatment to modify slag characteristics. Test castings, in the form of stepped blocks, were produced and rigorously examined. The evaluation included chemical analysis, mechanical testing, metallography, and visual inspection for slag inclusion severity. The chemical composition and mechanical properties of representative test blocks are shown in Table 4.
| Sample ID (Mgadd %) | Chemical Composition (%) | Mechanical Properties (After Heat Treatment) | |||||||
|---|---|---|---|---|---|---|---|---|---|
| C | Mn | Si | S | P | Mgres | Tensile Strength (MPa) | Elongation (%) | Hardness (HB) | |
| A (0.20%) | 3.65 | 0.45 | 2.40 | 0.010 | 0.040 | 0.055 | 720 | 5 | 240 |
| B (0.16%) | 3.68 | 0.42 | 2.38 | 0.011 | 0.038 | 0.045 | 740 | 7 | 235 |
| C (0.12%) | 3.70 | 0.40 | 2.35 | 0.008 | 0.039 | 0.035 | 750 | 9 | 230 |
| D (0.10%) | 3.72 | 0.38 | 2.32 | 0.007 | 0.041 | 0.030 | 760 | 10 | 225 |
The graphite nodularity was assessed metallographically. Samples with lower residual magnesium (0.03-0.04%) still exhibited excellent spheroidization, with nodule counts above 100 nodules/mm² and spheroidization rates exceeding 90%. The most significant improvement was observed in the reduction of slag inclusion defects. We developed a semi-quantitative rating scale for slag inclusion severity, as shown in Table 5.
| Rating | Macroscopic Appearance | Fracture Surface Color | Central Shrinkage Tendency | Edge Shrinkage | Chill Depth | Typical Graphite Morphology |
|---|---|---|---|---|---|---|
| 1 (Severe) | Numerous visible slag patches near surfaces and gates | Dark, dull gray with obvious non-metallic films | Pronounced | Significant | Large | Often poor nodularity, vermicular/flake graphite present |
| 2 (Moderate) | Few scattered slag pockets | Gray with some dark areas | Moderate | Some | Medium | Mixed nodular and imperfect forms |
| 3 (Low) | Rare, minor slag traces | Silvery gray, relatively clean | Minor or none | Little or none | Small or none | Good nodularity, mostly spherical graphite |
| 4 (None/Very Low) | No visible slag defects | Bright silvery gray | None | None | Very small or none | Excellent nodularity |
Applying this scale, test castings from heats with magnesium addition of 0.10-0.12% consistently achieved Ratings 3 or 4, indicating minimal slag inclusion. In contrast, castings from higher addition heats (0.18-0.20%) frequently showed Ratings 1 or 2. The mechanism is clear: excessive residual magnesium increases the oxidation potential of the iron melt during pouring and solidification, leading to the formation of complex magnesium silicates and other oxides that become entrapped as slag inclusion. The thermodynamic driving force for oxide formation can be related to the activity of magnesium. The equilibrium constant for the reaction $[Mg] + [O] \rightarrow (MgO)$ is:
$$ K_{MgO} = \frac{a_{MgO}}{a_{[Mg]} \cdot a_{[O]}} $$
Where $a_{[Mg]}$ is the activity of dissolved magnesium, proportional to its concentration. A higher $a_{[Mg]}$ shifts the equilibrium toward more MgO formation, increasing the likelihood of slag inclusion. By reducing residual magnesium, we lower $a_{[Mg]}$, thereby suppressing this reaction and reducing the volume of exogenous inclusions.
Furthermore, we investigated the effect of holding time after treatment. Residual magnesium gradually decreases due to evaporation and oxidation during the interval between treatment and the end of pouring. The decay can be modeled empirically:
$$ Mg_{res}(t) = Mg_{res}(0) \cdot e^{-kt} $$
where $Mg_{res}(0)$ is the residual Mg immediately after treatment, $t$ is the holding/pouring time in minutes, and $k$ is a rate constant dependent on temperature, slag cover, and other factors. For our conditions, $k$ was found to be approximately 0.005 to 0.01 min⁻¹. This means that for a typical 10-minute pouring cycle, the residual magnesium might drop by 5-10%. This loss is relatively small but becomes more significant at lower initial residual levels, necessitating tight scheduling to maintain nodularization.
The role of mold materials and atmosphere is also pertinent to slag inclusion control. We employ green sand molds with added coal dust to create a reducing atmosphere at the metal-mold interface, which helps prevent re-oxidation and minimize surface slag formation during pouring.
Heat treatment is an essential final step for our high-silicon ductile iron castings to relieve casting stresses and prevent cold cracking. Our standard annealing cycle involves heating to 900-920°C, holding for 2-4 hours depending on section size, furnace cooling to 720-740°C, holding for 4-6 hours, and then cooling to room temperature in the furnace. This treatment also contributes to matrix uniformity and further stabilizes the properties, indirectly influencing the perceived quality by ensuring that any subsurface slag inclusion is not exacerbated by residual stresses.
Through hundreds of tons of production trials, we have established that under strictly controlled melting and treatment conditions, reducing magnesium addition to 0.10-0.12% and maintaining residual magnesium in the range of 0.03-0.04% is entirely feasible. This practice not only yields castings virtually free from slag inclusion defects but also maintains excellent graphite spheroidization and superior mechanical properties. The key findings can be summarized as follows:
- Ladle Design and Operation: A properly designed pressure addition ladle with optimal geometry and operation parameters (controlled freeboard, deep bell immersion, correct materials) is paramount for achieving high and consistent magnesium utilization efficiency ($\eta$ of 25-30% or higher).
- Process Stabilization: Fixing melting and treatment variables within narrow windows allows for predictable control of residual magnesium by primarily adjusting the addition amount. This is the most practical approach for industrial production.
- Sulfur Control: Lowering and stabilizing the base iron sulfur content through charge selection and process intensification (like hot blast cupolas) is the foundation for reducing the required magnesium addition and consequently the residual magnesium level.
- Defect Correlation: There is a direct, observable correlation between residual magnesium content and the severity of slag inclusion defects. Reducing residual magnesium significantly mitigates slag inclusion formation.
- Synergistic Additions: The post-treatment addition of small amounts of silicon-calcium alloy can further improve slag characteristics, making inclusions more benign or promoting their floatation, though its primary benefit in our trials was linked to improved inoculation.
In conclusion, the problem of slag inclusion in ductile iron castings is highly manageable through scientific control of the nodularization process, specifically by optimizing magnesium addition to the minimum necessary level. This approach requires a deep understanding of the interrelationships between charge materials, furnace operation, treatment equipment, and process timing. The economic and quality benefits are substantial, reducing scrap rates, improving casting soundness, and delivering consistent mechanical performance. Future work may explore even lower addition techniques using advanced nodularizers or treatment in inert atmospheres to push the boundaries of slag inclusion control further. The persistent challenge of slag inclusion demands continuous attention, but as our experience demonstrates, it is a challenge that can be successfully met through diligent process engineering and control.