Simulating iron in oxygen-containing environments: An improved Fe-O interaction for density-functional tight-binding

The chemistry of iron in oxygen-containing and wet environments plays a central role in corrosion, (electro)catalytic reactions, and several biological processes. These processes hinge on the molecular-level interactions between iron and various oxygen-containing species such as water, molecular oxygen, oxygen radicals, and functional groups such as alcohols or carboxyls. Although the first-principles density-functional theory (DFT) describes these interactions well, DFT is often too slow to simulate the thermodynamics and kinetics of the above-mentioned processes at the necessary time and length scales. Fortunately, second-principles density-functional tight-binding (DFTB) satisfies these traits once properly parameterized for the target systems. Here, we discuss the problems that current DFTB parameterizations have with Fe-O pairwise repulsion, a central contributor to the DFTB performance. We construct an improved Fe-O repulsion by fitting the repulsion to structures relevant for topical research and benchmark it against structures with free and adsorbed Fe interacting with water and other oxygen-containing species. We explore the improved interaction by simulating the dynamics of atomic Fe and FeN4-modified graphene in aqueous environments, demonstrating the applicability of the parameterization to catalytically relevant large-scale simulations.