PNNL

Abstract

Atomic-scale understanding of important geochemical processes including 5 sorption, dissolution, nucleation, and crystal growth is difficult to obtain from 6 experimental measurements alone and would benefit from strong continuous progress 7 in molecular simulation. To this end we present a reactive neural network potential-based 8 molecular dynamics approach to simulate the interaction of aqueous ions on mineral 9 surfaces in contact with liquid water, taking Fe(II) on hematite(001) as a model system. 10 We show that a single neural network potential predicts rate constants for water exchange 11 for aqueous Fe(II) and for the exergonic chemisorption of aqueous Fe(II) on 12 hematite(001) in good agreement with experimental observations. The neural network 13 potential developed herein allows one to converge free energy profiles and transmission coefficients at ab initio level accuracy 14 outperforming state-of-the-art classical force field potentials. This suggests that machine learning potential molecular dynamics 15 should become the method of choice for atomistic studies of geochemical processes.