Composition-transferable machine learning potential for LiCl-KCl molten salts validated by high-energy x-ray diffraction
JC Guo and L Ward and Y Babuji and N Hoyt and M Williamson and I Foster and N Jackson and C Benmore and G Sivaraman, PHYSICAL REVIEW B, 106, 014209 (2022).
Unraveling the liquid structure of multicomponent molten salts is challenging due to the difficulty in conducting and interpreting high- temperature diffraction experiments. Motivated by this challenge, we developed composition-transferable Gaussian approximation potential (GAP) for molten LiCl-KCl. A DFT-SCAN accurate GAP is active-learned from only similar to 1100 training configurations drawn from 10 unique mixture compositions enriched with metadynamics. The GAP-computed structures show strong agreement across high-energy x-ray diffraction experiments, including for a eutectic not explicitly included in model training, thereby opening the possibility of composition discovery.
Return to Publications page