Untangling product selectivity on clean low index rutile TiO₂ surfaces using first-principles calculations

Computational modeling of metal oxide surfaces provides an important tool to help untangle complex spectroscopy and measured catalytic reactivity. There are many material properties that make rational catalytic design challenging, and computational methods provide a way to evaluate possible structural factors, like surface structure, individually. The mechanism of water oxidation or oxygen evolution is well studied on some anatase surfaces and the rutile TiO₂ (110) surface but has not yet been mapped on other low-index Miller rutile surfaces that are present in most experimental nano-titania catalysts. Here first principles calculations provide new insights into water oxidation mechanisms and reactivity of the most common low-index Miller facets of rutile TiO₂. The reactivity of three surfaces, (101), (010), and (001), are explored for the first time and the product selectivity of multistep electron transfer on each surface is compared to the well-studied (110) surface. Density functional theory shows that a peroxo, O^(p), intermediate is more favorable for water oxidation on all facets. The ˙OH radical formation is favored on the (001) facet resulting in a high overpotential for oxygen evolution reaction (OER). The (101) and (110) facets have low overpotentials, ∼0.3 V, and favor two-electron proton-coupled electron transfer to produce H₂O₂. The only facet that prefers direct OER is (001), leading to O₂ evolution in a four-electron process with an overpotential of 0.53 V. A volcano plot predicts the selectivity and activity of low-index Miller facets of rutile TiO₂, revealing the high activity of the peroxo OER mechanism on the (010) facet.