| Abstract Scope |
Exploring the structure and thermodynamics of molten salts (MS) is vital for unlocking their unique properties in emerging energy applications. Spectroscopic techniques play a crucial role in elucidating the structure of MS, with the simulation of Raman spectra offering an accurate interpretation of experiment, enhancing our understanding of structure and speciation. Despite these advancements, significant challenges persist in accurately predicting the densities, mixing enthalpies, and free energies of MS containing multivalent ions, e.g., La(III) and U(III). In this study, we present several cases focusing on computing densities and enthalpies of mixing of MS based on ab initio molecular dynamics simulations using density functional theory and the machine learning interatomic potentials. Since free energies cannot be directly derived from a single molecular dynamics simulation, we describe how we employ thermodynamic integration to compute these quantitates. Finally, we demonstrate the prediction of vapor-liquid phase diagrams of ML using a direct coexistence method. |