CALPHAD, Thermodynamics, Molecular Modeling, and First-Principles Methods for Metallurgical Slags: From Atomic-Scale Structure to Phase Equilibria and Process Prediction

Metallurgical slags simultaneously govern chemical partitioning, heat and mass transfer, refractory interactions, and the separation of valuable metals from oxide phases. Although the CALPHAD (CALculation of PHAse Diagrams) method predicts phase stability and chemical driving forces, its conventional application does not resolve short-range liquid structure or transport kinetics. Conversely, density functional theory (DFT), ab initio molecular dynamics (AIMD), and classical molecular dynamics (MD) can interrogate bonding and liquid networks but cannot independently establish industrial multicomponent equilibria. This critical narrative review examines how these approaches connect across length and time scales, emphasizing thermodynamic descriptions of complex oxide melts, slag structural descriptors, structure–property relationships, experimental validation, and uncertainty-aware process prediction. A section-specific bibliographic framework spans predominantly 2020–2026 literature, supplemented by methodological foundations. The synthesis distinguishes experimentally constrained prediction from plausible but unvalidated scale coupling. It identifies database coverage, redox-dependent speciation, force-field transferability, kinetic departures from equilibrium, and uncertain plant boundary conditions as principal barriers. A staged workflow is proposed in which experimental observations constrain atomistic and thermodynamic models, prediction intervals are tested against unseen conditions, and calibrated outputs inform reactor-scale decisions. Multiscale slag engineering is therefore better viewed as a traceable sequence of validated information transfers than as the simple juxtaposition of modeling techniques.