📝 SummarXiv

Abstract

Molten salts are high-temperature ionic liquids whose unique combination of strong Coulombic interactions, large polarizabilities, and high ionic conductivities makes them central to energy storage, metallurgy, and nuclear technology. Understanding their delicate balance of Coulomb forces, short-range repulsion, and electronic polarization, particularly regarding the role that electronic fluctuations play in their structure and dynamics, is critical to predictively designing molten salts for applications of interest. We investigate the importance of electronic fluctuations in molten AgI using density functional theory, a universal machine learning model (Orb), and a classical, empirical pairwise model of interionic interactions. We find that directional polarization fluctuations of iodide ions enhance Ag+ diffusion, manifesting as enhanced force fluctuations and structure in the time-dependent friction experienced by the cations. The coupling between iodide polarization fluctuations and silver diffusion creates a dynamic asymmetry; Ag+ motion is tightly linked to the instantaneous polarization of neighboring I-, whereas I- dynamics are relatively unperturbed by electronic fluctuations. For all structural and dynamic quantities investigated, the Orb model is in excellent agreement with density functional theory-based simulations, highlighting the ability of this universal neural network potential to capture many-body polarization effects. In contrast, the empirical force field fails to reproduce key structural and dynamic quantities involving cations, ultimately because it neglects dynamic electronic fluctuations. Our findings connect liquid=state ionic dynamics with the "electronic paddle-wheel" mechanism of ionic diffusion in superionic solids and motivate further exploration of polarization fluctuation effects in complex electrolytes and ionic liquids.