Machine learning force field development and basic physical property studies for molten salt reactor fuel salt LiF-BeF2-UF4.

Abstract

As one of the most promising technological pathways for Generation IV advanced reactors, molten salt reactors (MSRs) rely on the fuel salt LiF-BeF2-UF4 (FLiBeU), whose microstructural characteristics and fundamental physical properties determine the reactor's thermal-hydraulic behavior and safe operating limits. In response to the experimental challenges posed by the high temperature and high radioactivity of this molten salt system, this study adopts the deep potential molecular dynamics (DPMD) method combined with an active learning strategy to construct a high-precision machine learning force field model, whose accuracy is validated by density functional theory (DFT) and experimental results. Based on this force field model, the microstructure (radial distribution function, coordination number, angular distribution function, and network structure), thermophysical properties (density and heat capacity), and transport properties (self-diffusion coefficient, electrical conductivity, and shear viscosity) of the FLiBeU molten salt system were systematically investigated over a wide temperature range (773-1173 K) and a wide composition range (UF4 3-50 mol%). The study reveals the patterns of how temperature and UF4 concentration influence the structural evolution of the molten salt. In particular, it identifies the key mechanisms by which high UF4 concentration leads to a decline in ionic diffusion capacity, a sharp nonlinear increase in viscosity, and reductions in heat capacity and electrical conductivity. This research provides rich physical property data and microscopic mechanistic insights for both fundamental research and engineering applications of FLiBeU fuel salt, while also establishing a robust framework for studying high-temperature molten salt systems using machine learning methods.