Modeling Liquidus Hypersurfaces through Simplicial Complexes

Abstract

This paper describes an operational pipeline that exploits computational geometry to derive useful knowledge about the crystallization behaviour of materials composed of varying amounts of pure components. Starting from existing knowledge related to the pure components, we compute the Gibbs free energy of all their possible compositions in a given range of temperatures, both in liquid and solid phases. Then, we exploit the convex hull method to derive the coexistence of solid and liquid phases, and model the resulting liquidus hypersurface as a simplicial complex. On such a complex, we propose novel tools to robustly compute descent lines describing the crystallization path induced by heat loss for any initial composition in the system.