Regular Solution & Scheil-Gulliver Lab · binary phase diagrams and non-equilibrium solidification, from first principles

What this is actually computing. Both phases are modeled as regular solutions, G(x,T) = Ω·x(1−x) + RT[x ln x + (1−x) ln(1−x)], with the liquid anchored to the solid reference state through each pure component's melting point (ΔG_fus(T) = ΔH_fus(1 − T/T_m), i.e. ΔC_p = 0 is assumed). At every temperature the two curves are sampled on a fine composition grid and the equilibrium phases are read off the lower convex hull of the combined points, the same minimize-the-Gibbs-energy principle a full CALPHAD solver uses, just without sublattices, multicomponent systems, or an assessed database behind it. No diagram topology (eutectic, isomorphous, miscibility gap) is hard-coded; whatever shape the hull finds for the chosen Ω/T_m values is what renders. teaching-scale model

What this is actually computing. This mode does not use Mode 1's Gibbs-energy model at all; it directly integrates the Scheil-Gulliver mass-balance assumption (infinitely fast diffusion in the liquid, zero diffusion in the solid, local equilibrium at the interface, and a constant partition coefficient k, i.e. linearized liquidus/solidus). Real solidification almost always has some solid-state back-diffusion, which is exactly what Thermo-Calc's Diffusion Module (DICTRA) and its own back-diffusion Scheil variant add on top of this idealization using an actual mobility database, a materially harder numerical problem than what's running here. teaching-scale model