Driven by the dual requirements of carbon neutrality and lightweighting in advanced manufacturing, there is a growing demand for high-performance structural materials that offer an excellent synergy of strength and toughness. Austempered Ductile Iron (ADI), a member of the broader family of ductile iron casting, stands out due to its unique multiphase microstructure consisting of bainitic ferrite and carbon-enriched retained austenite. This microstructure imparts a superior combination of mechanical properties, making ADI an ideal candidate for critical components such as high-speed train brake discs, automotive gears, and wind turbine parts.
However, the traditional research and development paradigm for ADI faces significant challenges. The final properties are governed by a complex, non-linear interplay of multiple parameters including carbon and silicon content, alloying element ratios (e.g., Mn, Cu, Ni), and austempering temperature and time. Key mechanisms, such as the synergistic control of austenite stability, bainitic transformation kinetics, and martensite formation, are not yet fully elucidated. Furthermore, the conventional trial-and-error approach, involving the full cycle of melting, casting, heat treatment, and mechanical testing for each parameter set, is prohibitively time-consuming and costly for exploring the vast design space. There is an urgent need to develop a more scientific and efficient design methodology.
In this context, the CALculation of PHAse Diagrams (CALPHAD) method has emerged as a powerful computational tool. It is currently the only method capable of accurately calculating phase equilibria in multicomponent systems while meeting the precision requirements for practical engineering applications. By integrating thermodynamic equilibrium calculations and transformation kinetics simulations, CALPHAD-based high-throughput computation enables the automated screening of billions of potential composition and process parameter combinations. This represents a paradigm shift from “experience-driven” to “computation-driven” material design.
This work leverages the CALPHAD method to systematically investigate the microstructural evolution of ADI under various austempering conditions. We construct a three-stage phase transformation model to simulate the key steps in ADI heat treatment: austenitization, bainitic (ausferritic) transformation, and martensitic transformation. A high-throughput computational framework is developed to perform batch calculations, focusing on the volume fractions of graphite, bainitic ferrite, retained austenite, and martensite. Based on the computational results, we analyze the influence of key alloying elements (C, Si, Mn, Cu, Ni) and the austempering temperature. Furthermore, we propose key microstructural feature parameters related to mechanical performance and establish a comprehensive criterion to intelligently optimize the composition and process for ADI, paving a new, efficient path for the design of high-performance ductile iron casting.
The heat treatment of Austempered Ductile Iron typically involves three critical stages: austenitization, austempering (isothermal quenching), and final cooling. The developed computational model addresses each stage to predict the final microstructure.
Stage 1: Austenitization. The as-cast ductile iron casting is heated to a temperature within the austenite region (typically 840–950 °C) and held for a sufficient time (1–2 hours). This process aims to achieve a homogeneous, carbon-saturated austenitic matrix. For the high-throughput calculations, we assume equilibrium conditions at the austenitization temperature of 900 °C for 2 hours. Using the phase equilibrium module of Thermo-Calc software, we calculate the two-phase equilibrium between austenite and graphite. The outputs—volume fraction of graphite, volume fraction of austenite, and its carbon content—serve as the initial state for the subsequent transformation simulations.
Stage 2: Austempering (Bainitic Transformation). After austenitization, the component is rapidly quenched to an isothermal holding temperature (Austempering Temperature, AT) between 230 °C and 400 °C to prevent pearlite formation. During the isothermal hold (e.g., 2.5 hours), part of the austenite transforms into bainitic ferrite via a displacive mechanism. This transformation rejects carbon into the surrounding austenite, enriching it. The reaction proceeds until the carbon concentration in the untransformed austenite reaches a critical value, often associated with the T0 curve, where the driving force for the displacive transformation diminishes. This stage is simulated using appropriate models within Thermo-Calc to calculate the amounts of bainitic ferrite and untransformed austenite, along with the carbon content of the latter.
Stage 3: Final Cooling (Martensitic Transformation). After the austempering hold, the component is cooled to room temperature. The stability of the carbon-enriched, untransformed austenite determines whether it will remain as retained austenite or partially transform into martensite during this cooling step. This stage is simulated by assessing the martensite-start (Ms) temperature of the austenite with its final composition from Stage 2. The potential martensite volume fraction is then estimated.
The high-throughput calculation workflow was automated using the TC-Python interface. For the initial screening, the primary elements C, Si, and Mn were varied within typical ranges for ADI, while Cu was fixed at 0.5 wt.% and Ni at 0 wt.%. Eight different austempering temperatures (AT) were selected. This resulted in 864 unique composition-process combinations. The calculation conditions are summarized below:
| Parameter | Lower Limit | Upper Limit | Step Size | Number of Values |
|---|---|---|---|---|
| C (wt.%) | 3.0 | 4.0 | 0.2 | 6 |
| Si (wt.%) | 2.0 | 3.0 | 0.2 | 6 |
| Mn (wt.%) | 0.2 | 0.6 | 0.2 | 3 |
| AT (°C) | 230 | 400 | Variable* | 8 |
* Step size: 30°C for AT < 350°C; 20°C for 350°C ≤ AT < 390°C; 10°C for AT ≥ 390°C.
The Pearson correlation coefficients between the calculated microstructural features and the input variables were analyzed. The results, presented as a heatmap, clearly show that the austempering temperature (AT) has the most significant overall correlation with the microstructure of the ductile iron casting, far exceeding the influence of individual alloying elements in this initial screening. Specifically, AT shows a strong positive correlation with the volume fraction of bainitic ferrite, retained austenite, and its carbon content, while exhibiting a strong negative correlation with martensite content. The correlation of C, Si, and Mn with the microstructure is more complex and varies with the AT level.
A detailed analysis reveals that the effect of composition differs markedly between low (e.g., 290 °C) and high (e.g., 370 °C) austempering temperatures. At 290 °C, graphite content is dominantly controlled by carbon, with a correlation coefficient of 0.99. In contrast, bainitic ferrite, martensite, and retained austenite content and stability are primarily influenced by Si and Mn. Silicon shows a stronger effect than Mn, with correlation coefficient magnitudes generally above 0.7. The relationships are summarized by the following trends:
$$ V_f^{Graphite} \propto [C] $$
$$ V_f^{Bainite}, V_f^{RA}, C_{RA} \propto \frac{1}{[Si]}, \frac{1}{[Mn]} $$
$$ V_f^{Martensite} \propto [Si], [Mn] $$
where $V_f$ denotes volume fraction and $C_{RA}$ is the carbon content in retained austenite.
At 370 °C, the influence patterns change. While carbon still solely controls graphite formation, manganese becomes the primary factor affecting bainitic ferrite amount and retained austenite carbon content (correlation ≈ -1). Silicon predominantly governs the amounts of martensite and retained austenite. This stark contrast underscores the decisive role of austempering temperature in mediating the effect of alloying elements in ductile iron casting.
Subsequently, the influence of Cu and Ni was investigated by fixing several promising base compositions from the first screening and varying Cu (0.5-1.5 wt.%) and Ni (0.5-2.0 wt.%) contents. A total of 480 new combinations were calculated. The analysis shows that, under the dominant influence of the base C-Si-Mn matrix, Ni generally has a stronger effect on most microstructural features than Cu. For instance, at 290 °C, Ni significantly influences graphite, bainitic ferrite, retained austenite, and its carbon content, while Cu’s effect is more pronounced on martensite. The correlation coefficients for Cu are typically below 0.3, indicating a relatively modest influence compared to the primary elements and Ni.
| Microstructural Feature | Primary Influencing Element(s) at Low AT (~290°C) | Primary Influencing Element(s) at High AT (~370°C) |
|---|---|---|
| Graphite Content | C (r ≈ 0.99) | C (r ≈ 0.99) |
| Bainitic Ferrite Content | Si, Mn (Negative correlation) | Mn (r ≈ -0.99) |
| Martensite Content | Si, Mn (Positive correlation) | Si (r ≈ -0.95) |
| Retained Austenite Content | Si, Mn (Negative correlation) | Si (r ≈ 0.94), Mn |
| Carbon in R. Austenite ($C_{RA}$) | Si, Mn (Negative correlation) | Mn (r ≈ -1.0) |
To guide the optimization of composition and process, three key microstructural and compositional feature parameters were defined based on physical metallurgy principles for ductile iron casting. These features are closely linked to the final mechanical properties of ADI.
1. Carbon Equivalent (CE): Closely related to the castability and soundness of the iron. A CE near the eutectic point (approximately 4.3-4.5) is desirable to ensure good fluidity while avoiding defects like shrinkage porosity or graphite flotation. The simplified carbon equivalent is calculated as:
$$ CE = C + \frac{1}{3}Si – 0.03Mn $$
A target value of $CE_{target} = 4.5$ was used. The first feature parameter $F_1$ measures the deviation from this ideal:
$$ F_1 = |CE – 4.5| $$
Lower $F_1$ values are preferred.
2. Retained Austenite Stability: The stability of retained austenite is crucial for the TRIP (Transformation Induced Plasticity) effect, which enhances ductility and toughness. Stability can be proxied by the product of the retained austenite volume fraction ($V_f^{RA}$) and its carbon content ($C_{RA}$). A higher product indicates greater stability. Therefore, the second feature parameter is defined as:
$$ F_2 = \frac{1}{V_f^{RA} \times C_{RA}} $$
Lower $F_2$ values indicate higher austenite stability and are preferred.
3. Martensite Content: Martensite, being a hard and brittle phase, can increase strength but severely compromise ductility and impact toughness if present in significant amounts. For a balanced property profile, its content should be minimized. The third feature parameter is simply:
$$ F_3 = V_f^{Martensite} $$
Lower $F_3$ values are preferred.
To enable a unified optimization across these three criteria, the individual parameters were normalized, and a Comprehensive Criterion (CC) was formulated as the Euclidean distance from the ideal point in this three-dimensional feature space:
$$ CC = \sqrt{ (F_1)^2 + (F_2)^2 + (F_3)^2 } $$
A lower CC value signifies a composition and process combination that is closer to the ideal balance of good castability, high austenite stability, and low martensite content—all indicative of potential for superior comprehensive mechanical properties in the ductile iron casting.
A partial correlation analysis was conducted to identify the factors most strongly affecting the Comprehensive Criterion (CC). The results confirm that the austempering temperature is the most influential parameter, followed by the carbon content. This aligns with the earlier observation that AT dramatically alters microstructural evolution paths. By ranking all calculated composition-process combinations based on their CC value, the optimal combination identified was: Fe-3.6C-2.6Si-0.6Mn-1.0Ni-0.5Cu (wt.%) with an austempering temperature of 320 °C.
This optimized composition and process are in good agreement with high-performance ADI grades reported in the literature. For instance, a composition of Fe-3.59C-2.72Si-0.18Mn-0.47Ni-0.68Cu-0.15Mo subjected to austempering at 320 °C has been shown to yield an excellent combination of strength (UTS ~1281 MPa) and elongation (~6.1%). The proximity of our computationally optimized composition to this experimentally validated one underscores the efficacy of the CALPHAD-driven high-throughput design approach for ductile iron casting.
This work demonstrates a successful computation-driven design strategy for Austempered Ductile Iron. By constructing a three-stage CALPHAD-based phase transformation model and implementing a high-throughput computational framework, we systematically mapped the effects of key alloying elements (C, Si, Mn, Cu, Ni) and austempering temperature on the final microstructure of this ductile iron casting.
The results conclusively show that the austempering temperature is the paramount factor controlling microstructural evolution, significantly modulating the influence of alloying elements. Carbon primarily dictates graphite formation, while silicon and manganese are crucial in governing the amounts and stability of bainitic ferrite, retained austenite, and martensite, with their specific roles being temperature-dependent.
By defining and integrating three performance-critical feature parameters—carbon equivalent deviation, retained austenite stability inverse, and martensite content—into a single Comprehensive Criterion, we established a quantitative metric for optimization. The optimal composition and process identified through this criterion, Fe-3.6C-2.6Si-0.6Mn-1Ni-0.5Cu with austempering at 320 °C, is rationally supported by the analysis and aligns well with known high-performance ADI grades.
This study provides a new, efficient pathway for the design of high-performance ductile iron casting, transitioning from traditional empirical methods to a targeted, calculation-guided paradigm. The framework and methodology can be extended to optimize other grades of cast iron or to incorporate additional constraints and objectives, such as hardenability, wear resistance, or cost.