| Abstract Scope |
Conventional phase-field models approximate stoichiometric compounds using parabolic free energy functions, which often lead to artificial non-stoichiometry, thermodynamic inconsistency, and numerical instability. Ji and Chen (2022) addressed this issue by developing a thermodynamically consistent phase-field model that captures the coexistence of stoichiometric and solution phases without introducing artificial compositional dependence. While their approach was demonstrated in binary systems, most engineering alloys involve three or more elements, requiring consideration of multicomponent interactions.
In this study, we extend the phase-field model that describes the growth and dissolution of stoichiometric compounds to ternary systems, enabling the simulation of microstructure evolution involving multiple elements and stoichiometric phase transformations. To improve computational performance, we implement a semi-implicit Fourier spectral method using CUDA-based GPU parallelization, enabling significantly faster simulations and efficient handling of multicomponent systems. |