A linear response DFT+U study of trends in the oxygen evolution activity of transition metal rutile dioxides

Abstract

There are known errors in oxidation energies of transition metal oxides caused by an improper treatment of their d-electrons. The Hubbard U is the computationally cheapest addition one can use to capture correct reaction energies, but the specific Hubbard U oftentimes must be empirically determined only when suitable experimental data exist. We evaluated the effect of adding a calculated, linear response U on the predicted adsorption energies, scaling relationships, and activity trends with respect to the oxygen evolution reaction for a set of transition metal dioxides. We find that applying a U greater than zero always causes adsorption energies to be more endothermic. Furthermore, the addition of the Hubbard U greater than zero does not break scaling relationships established without the Hubbard U. The addition of the calculated linear response U value produces shifts of different systems along the activity volcano that results in improved activity trends when compared with experimental results.