Combined Molecular Dynamics and Phase Field Simulation of Crystal Melt Interfacial Properties and Microstructure Evolution during Rapid Solidification of TI-NI Alloys

Phase field method has become a popular tool to investigate the microstructure evolution during the solidification. Quantitative predictions using this method is still limited, and in this dissertation, we try to study this problem from different perspectives. Most of the phase field models consider the solid-liquid interface to be in local equilibrium. Solidification during some manufacturing processes like selective laser melting, and electron beam additive manufacturing is rapid and far from equilibrium which can result in supersaturated solid solutions, segregation-free crystals, or metastable phases. Before obtaining any conclusions from the phase field simulations, we must know the answer for “which phase field model works for rapid solidifications?” Solidification involves multiple special and temporal scales ranging from picoseconds to seconds, and from nanometers to micrometers. For this, phase field method considers interface width and the characteristic dissipation time scale in between those two limits. However, the questions that should be addressed is “how using diffuse interface phase field models, especially for the rapid solidification, effects the quantitative predictions of this method?” The phase field parametrization requires knowing multiple material properties, and experiments are capable of calculating some of these them. The rest of these parameters are either estimated by analytical methods or molecular dynamic (MD), like the solid-liquid interfacial free energies, or considered to zero and non-effective for solidification, like solid-liquid kinetic coefficient. However, it is critical to know “How the phase field simulation results, especially for rapid solidification, are effected by MD-calculated material properties?” To address these questions, in this research, we performed combined MD and phase field simulations to study the solidification of pure Ti and Ti-Ni alloys. MD simulations are used to calculate interfacial properties, namely anisotropic kinetic coefficient and solid-liquid interface free energy. The first step for obtaining reliable MD simulation results is having an accurate interatomic potential. For this, we developed a new MEAM interatomic potential predicting the high-temperature solid liquid coexistence. We also performed MD simulation to yield a detailed understanding of the kinetic processes that occur during rapid solidification and the results are compared with different phase field simulations to test the consistency of these two simulation methods