Solidification phase and morphology prediction for Al–Fe castings using a cellular automata nucleation and growth model

Abstract

Non-equilibrium phase formation and the interaction between the nucleation and solidification of competing phases are significant challenges in aluminum alloy design, as key properties such as ductility heavily depend on the grain morphologies and phases present in solidified castings. This manuscript expands upon existing multiphase cellular automata (CA) modeling work to predict hypereutectic Al–Fe alloy microstructures, considering primary and eutectic phase nucleation in both the bulk liquid and at existing solid–liquid interfaces, along with solute diffusion and phase-dependent growth morphologies. The 2D growth approximation for the primary intermetallic θ-Al13Fe4 phase is validated against in-situ radiography data from the literature, and sensitivity of the predicted multiphase grain structure of Al-2.5 wt% Fe to multiple model inputs governing nucleation density, undercooling, and eutectic growth is explored. The predicted primary FCC and eutectic area fractions are found to be particularly sensitive to two highly uncertain inputs — the bulk nucleation undercooling and the eutectic growth coefficient. For cooling rates representative of furnace cooling and rates representative of die casting conditions, the simulations using calibrated input parameters accurately predicted the spatial distribution of primary θ-Al13Fe4, primary FCC-Al, and eutectic phases along with the primary phase grain morphologies. The demonstrated ability to reproduce accurate phase selection and phase fractions for a range of solidification conditions will enable model application for aluminum alloy and casting design with extension to more complex alloys in the future. The simulation results also highlight areas of high experimental uncertainty where collecting additional data could reduce model uncertainty in future work.