A Data and DFT-Driven Framework for Predicting the Microstructure of Submonolayer Inverse Metal Oxide on Metal Catalysts.

Metal oxides on metal (inverse) catalysts can selectively drive many important reactions. However, understanding the active site under experimentally relevant conditions is lacking. Herein, we introduce a computational framework for predicting atomic models of stable inverse catalysts and demonstrate it for WO x on Pt(553) and a Pt 79 nanoparticle at variable WO x coverages. An evolutionary algorithm identifies a small (5%) subset of promising atomic configurations on which DFT simulations are performed. We predict a maximum coverage of ∼50% WO x on Pt(553), consisting of small clusters (tetramers and pentamers), which preferentially reside on the terrace, with their oxygen atoms interacting with the Pt step sites. Consistently, WO x does not lie on curved and undercoordinated metal sites of Pt nanoparticles. The oxide clusters prefer a partially reduced oxidation state. Theoretical EXAFS spectra for select configurations provide insights into interpreting experimental spectra of inverse catalysts. The framework applies to other catalysts.