Prediction of Structural Defects During Solidification of Al-Si-Mg Alloy System by Scheil-Gulliver Method

Abstract

This study applies Scheil-Gulliver non-equilibrium solidification modeling, implemented in ProCAST 13.5.2 with the AlDATA 4.2 thermodynamic database, to predict defect-promoting intermetallic phase formation in an AlSi9Mg alloy (Si 9.2, Mg 0.45, Ni 0.19, Fe 0.2, Mn 0.3 wt.%) derived from automotive scrap remelting in Cuba. The model predicts solidification between 597 and 533 °C a 65 °C non-equilibrium window with a phase formation sequence that places brittle Al3Ni needles first, followed by Mg2Si and the π-phase (Al8FeMg3Si6) during eutectic arrest at ~558 °C. Predicted phase fractions reach 0.022, 0.081, and 0.048 wt.% respectively. All representing upper-bound estimates under the no-back-diffusion assumption. Computed thermophysical properties showed that thermal conductivity dropped 67 W/m·K across the solidification interval, while density decreased by 200 kg/m³, both critical inputs for shrinkage-porosity prediction. The Mn/Fe ratio of 1.5 in this alloy suppresses the formation of β-Al5FeSi in favor of less detrimental Al15(Fe, Mn)3Si2, a result the model captures without additional fitting parameters. Practical thresholds for Ni content, Mg/Si ratio, and cooling rate control are derived from the simulation and validated against published experimental data for comparable alloy systems. The results demonstrate that Scheil-Gulliver modeling with industrial-grade thermodynamic databases provides actionable guidance for secondary aluminum foundry operations where compositional variability is inherent.