In my years of working with ductile cast iron rollers for rolling mill applications, I have encountered numerous challenges related to achieving consistent high quality. The shift towards medium-frequency induction furnaces for melting has brought about both advantages and complexities. While these furnaces offer rapid heating and precise temperature control, they also introduce issues such as difficulty in deoxidation, dephosphorization, and degassing, leading to defects like surface pinholes, slag inclusions, poor nodularization, shrinkage porosity, and insufficient tensile strength in roller necks. This article details my firsthand experiences and the systematic approaches we developed to optimize the production process for large ductile cast iron-cored rollers, specifically those with high-chromium iron outer layers. The core focus is on enhancing the inherent properties of ductile cast iron through meticulous control of every stage, from charge makeup to final solidification.
The primary roller in question had a significant diameter exceeding 600 mm, with a core made of ductile cast iron. Initial production runs revealed several critical failures: internal shrinkage porosity in the upper sections leading to excessive ultrasonic wave attenuation, roller neck tensile strength below 400 MPa, and suboptimal nodular graphite structure in critical cross-sections. My analysis pinpointed several root causes. First, trace elements like Pb, Sb, As, and Bi from impure charge materials can severely interfere with graphite nodulization. Second, improper temperature regimes during melting could lead to excessive carbon loss and silicon pickup via the reaction: $$ \text{SiO}_2 + 2\text{C} \rightleftharpoons \text{Si} + 2\text{CO} $$ This not only alters chemistry but increases slag content. Third, the gating and risering system failed to promote directional solidification, exacerbating shrinkage defects. Addressing these required a holistic quality control strategy.
The foundation of quality lies in correct chemical composition. For thick-section ductile cast iron rollers, carbon equivalent (CE) is paramount. We aimed for a CE between 3.8% and 3.9% to ensure good castability while minimizing the risk of graphite flotation. Within this constraint, we deliberately kept silicon content on the lower side. A balanced composition promotes the formation of the desired bull’s-eye ferrite structure around graphite nodules in the core. The specific compositional ranges we targeted for the roller core are summarized in the table below.
| Element | Target Range (wt.%) | Key Function & Rationale |
|---|---|---|
| C | 3.0 – 3.3 | Primary graphite former; high carbon improves graphitization potential and reduces chilling tendency. |
| Si | 2.0 – 2.2 | Ferritizer and graphitizer; controlled level prevents excessive ferrite and lowers CE to avoid flotation. |
| Mn | 0.4 – 0.6 | Strengthener and pearlite promoter; kept low to minimize segregation and carbide formation in slow-cooling sections. |
| P | ≤ 0.08 | Harmful element; kept minimal to prevent the formation of brittle phosphide eutectics. |
| S | ≤ 0.03 | Detrimental to nodulization; very low levels are essential for efficient magnesium treatment. |
| Mgres | ≥ 0.04 | Residual magnesium from treatment; ensures graphite spheroidization. |
| Ni | 0.3 – 0.5 | Alloying element; enhances strength and hardenability uniformly across thick sections. |
Achieving this chemistry consistently starts with the selection of pure charge materials. We implemented a “high-purity charge” policy. The charge mix was optimized to 25-35% high-quality pig iron (low in Si, Mn, S, P), 55-65% clean internal returns (like risers and feeder heads from previous ductile iron casts), and 5-15% selected steel scrap. This blend provides a stable base with minimal interference from tramp elements. During charging, we layered the furnace: steel scrap and punchings at the bottom to aid early slag formation and protect the lining, followed by pig iron and returns. The melting strategy was “fast melt, fast tap” to minimize the time the iron is exposed to high temperatures in the furnace, reducing carbon loss and lining erosion.
Temperature control is arguably the most critical operational parameter. The medium-frequency furnace’s ability to superheat quickly is a double-edged sword. Prolonged holding above 1510°C accelerates the silica reduction reaction: $$ \text{SiO}_2(\text{s}) + 2\text{C}(\text{dissolved}) \rightarrow \text{Si}(\text{dissolved}) + 2\text{CO}(\text{g}) $$ The generated CO gas can cause pinholes if not allowed to escape, and the reaction depletes carbon while increasing silicon. We found that a short holding period (around 5 minutes) at 1510-1530°C actually improved metal cleanliness by allowing inclusions to float out. However, the tapping temperature was strictly controlled at 1480 ± 10°C. Tapping at this temperature minimizes energy waste, reduces lining attack, and provides an optimal temperature for subsequent treatment and pouring. The relationship between holding time, temperature, and metal quality can be conceptualized by an empirical degradation function: $$ Q_t = Q_0 \cdot e^{-k(T – T_{eq})t} $$ where \(Q_t\) is metal quality (inverse of impurity level), \(Q_0\) is initial quality, \(k\) is a rate constant, \(T\) is holding temperature, \(T_{eq}\) is an equilibrium temperature (around 1470°C), and \(t\) is time. This underscores the need for precise thermal management.
Slag formation and removal are vital. A basic slag is maintained during melting to absorb deoxidation products and eroded lining material. We use sodium carbonate (soda ash) for desulfurization when the base sulfur exceeds 0.02%. The reaction is: $$ \text{Na}_2\text{CO}_3 + \text{FeS} \rightarrow \text{Na}_2\text{S} + \text{FeO} + \text{CO}_2 $$ This is endothermic, so it is performed around 1500°C with 1.5-2.5% addition based on the sulfur level. Frequent and thorough slag raking before tapping is mandatory to achieve a clean, bright metal surface.
The heart of ductile cast iron production is the nodularizing and inoculating treatment. For these large rollers, we adopted a robust combination treatment in the pouring ladle. The treatment alloys are placed in the bottom of a preheated ladle in a specific order:
- Nickel-Magnesium Alloy (80% Ni, 14-18% Mg): Added at 5 kg/ton. Nickel lowers the melting point of the alloy, ensuring a smooth, calm reaction and aiding in magnesium recovery.
- Rare Earth Silicide (RESiFe): Added at 10 kg/ton. Rare earths (Ce, La) counteract the harmful effects of trace elements like Pb and Sb, improve nodule count, and reduce the formation temperature of the oxide/sulfide films that cause dross.
- Zirconium-bearing Inoculant (SiZr): Added at 3 kg/ton (composition: 60-65% Si, 5-7% Zr, 1-2% Ca, 0.75-1.5% Al). Zirconium is a powerful inoculant for thick sections, promoting a uniform, fine graphite structure and delaying nodule growth to optimize the use of expansion during solidification.
The treated metal is then subjected to a post-inoculation as it is poured into the mold, using 1.5 kg/ton of fine (1-3 mm) SiZr inoculant. This dual-stage inoculation is crucial for countering fade and ensuring a high nodule count throughout the slow-solidifying mass. The efficiency of magnesium treatment can be modeled by: $$ \text{Mg}_{\text{recovery}} (\%) = A – B \cdot [\text{S}]_{\text{initial}} – C \cdot T_{\text{tap}} $$ where \(A\), \(B\), and \(C\) are constants derived from practice, \([\text{S}]_{\text{initial}}\) is the pre-treatment sulfur, and \(T_{\text{tap}}\) is the tapping temperature.
Process control extends beyond the furnace. To prevent shrinkage, the timing and magnitude of graphite expansion must be managed. The concept of “feeding demand” versus “expansion supply” is key. The total volume change during solidification (\( \Delta V \)) can be expressed as: $$ \Delta V = V_{L} + V_{G} $$ where \(V_{L}\) is the liquid contraction (negative) and \(V_{G}\) is the graphite expansion (positive). For soundness, we need \( |V_{G}| \ge |V_{L}| \) after the metal has lost fluidity. This requires that graphite precipitation initiates not too early but continues steadily through the eutectic freeze. Our inoculation strategy, combined with a controlled low residual magnesium (aiming for 0.04-0.06%), ensures this. The pouring temperature is maintained at 1360-1380°C. A higher temperature increases liquid shrinkage, demanding more expansion for compensation. After pouring, insulating covers are placed over the mold top within 30 minutes to slow cooling, and the mold is kept intact for over 96 hours to allow full utilization of graphitization expansion for self-feeding.
The microstructural evolution is the ultimate proof of process efficacy. The optimized process transformed the core microstructure. Previously, graphite was unevenly distributed with a high proportion of irregular, vermicular, or compacted forms, resulting in a nodularity of only about 70%. Ferrite content was low and uneven. After implementing the comprehensive controls, the graphite morphology became predominantly spherical (nodular), with nodularity consistently reaching 95% or higher. The ferrite formed perfect “bull’s-eye” rings around the nodules and was uniformly distributed throughout the matrix. This homogeneous structure of spherical graphite in a ferritic matrix is the hallmark of high-quality ductile cast iron, providing an excellent combination of strength and ductility. The improvement directly correlates with the key process parameters, as summarized in the following performance comparison table.
| Performance Metric | Before Optimization | After Optimization | Improvement / Target |
|---|---|---|---|
| Core Nodularity | ~70% | ~95% | Significant increase in graphite spheroidization. |
| Ferrite Volume Fraction | Low, uneven | High, uniform “bull’s-eye” | Enhanced ductility and impact resistance. |
| Roller Neck Tensile Strength | ≤ 400 MPa | ≥ 520 MPa | Increase of over 30%. |
| Surface Hardness (HSD) | Within 72-78 range | Consistently at upper limit (~78 HSD) | Improved wear resistance. |
| Internal Soundness (UT) | Excessive backwave attenuation (>10%) | No significant attenuation | Elimination of major shrinkage porosity. |
| Rolling Tonnage (High-Ni Chill Roll) | ~3,000 tons | ~3,200 tons | ~7% increase in service life. |
| Rolling Tonnage (High-Cr Iron Roll) | ~4,000 tons | ~4,800 tons | ~20% increase in service life. |
The mechanical property enhancement, particularly the tensile strength exceeding 520 MPa in the roller neck, is a direct consequence of the refined microstructure and reduced defect density. The hardness consistency at the upper specification limit indicates not only good wear resistance but also uniform solidification and cooling. The substantial increase in rolling tonnage before re-dressing is the most telling economic benefit, validating the entire quality control regime.
In conclusion, producing high-integrity ductile cast iron for demanding applications like mill rollers requires a system-wide perspective. It is not merely about hitting a chemical specification but about orchestrating raw material purity, thermal history, treatment kinetics, and solidification control. The medium-frequency furnace, while efficient, demands respect for its particular metallurgical dynamics. By rigorously controlling charge makeup, implementing a fast-melt-fast-tap protocol, employing a combined Ni-Mg-RE-Zr treatment system, and meticulously managing pouring and cooling conditions, we can consistently produce ductile cast iron with superior nodularity, mechanical properties, and internal soundness. The principles outlined here—focusing on cleanliness, precise thermal management, and powerful inoculation—are broadly applicable to the production of heavy-section ductile cast iron castings beyond just rollers. The journey from problematic castings to reliable performance underscores the profound impact that integrated process science can have on the quality of ductile cast iron components.