378995 Screening for Active Sites on Supported Metal-Oxide Catalysts Via Hybrid- Monte Carlo/Molecular Dynamics Simulations with the Reaxff Potential

Thursday, November 20, 2014: 1:30 PM
308 (Hilton Atlanta)
Thomas Patrick Senftle, Department of Chemical Engineering, Penn State University, State College, PA, Adri C.T. van Duin, Department of Mechanical and Nuclear Engineering, Pennsylvania State University, University Park, PA and Michael J. Janik, Department of Chemical Engineering, Pennsylvania State University, University Park, PA

The rational design of supported metal-oxide catalysts requires a detailed understanding of the catalyst surface structure as a function of temperature and pressure in the reactant gas phase. Generally, the gas phase can rearrange the metal/oxide lattice structure, forming unique surface, subsurface, or metal/oxide interfacial phases - each with catalytic behaviors that greatly vary from the parent system. Additionally, numerous phases or structures can co-exist on the catalyst surface, making it difficult to determine which phases are catalytically active, especially if the active site is attributed to a minority of the possible phases. Here, we demonstrate how ReaxFF-based Monte Carlo and molecular dynamics simulation techniques can be used in tandem to predict phase structure and stability as a function of temperature and pressure in the reactant phase above the catalyst. Using this method, we can generate an extensive set of catalyst structures that are representative of various stable geometries under the specified reaction conditions. These geometries then serve as appropriate models for subsequent kinetic studies assessing catalytic activity (either via quantum reaction barrier calculations or reactive molecular dynamics simulations), thus allowing geometries containing catalytically active sites to be identified. This is initially demonstrated by identifying under-coordinated sites on free Pd nanoclusters that are susceptible to Ar-induced surface reconstructions, in agreement with experimental findings. It is then extended to identifying unique sites on supported (TiO2 and CeO2) Pd clusters, which readily form oxide, hydride, and carbide phases impacting catalytic activity. Although demonstrated for Pd-based catalysts, this approach is readily extendable to other systems, providing a new screening tool for identifying active and selective sites under varying reaction conditions.

Extended Abstract: File Not Uploaded
See more of this Session: Rational Catalyst Design II
See more of this Group/Topical: Catalysis and Reaction Engineering Division