(323c) Probing Shape-Selective Catalysis in Heterogeneous Nanoporous Catalysts

Heterogeneous nanoporous catalysts such as zeolites and metal-organic frameworks comprise a network of channels of different shapes and dimensions, providing distinct local chemistry and unique active site environments. Reactions catalyzed in these confined pore environments may therefore highly depend on the distribution of active sites and their relative reactivities, and are especially challenging to model for large, flexible transition-state complexes. To capture this complexity, we present a method that uses a combination of force field-based sampling and density-functional theory (DFT) based transition state refinement. In this method, molecular mechanics force fields are first used to generate tens of thousands of low energy configurations for articulated transition-state complexes across the entire pore space of a nanoporous catalyst, and depending on available computational resources, condensed to a minimal set of distinct configurations (e.g., based on pore regions or active site identity). An initial reaction pathway is then constructed automatically for each of these configurations and optimized using DFT calculations with the nudged elastic band method. We apply the method to study acid-catalyzed aldol condensation and fission reactions, and compare them to mono- and bi-molecular hydrocarbon transformations. We found that while intrinsic activation energies for bulkier transition states depend sensitively on the location of active sites, more compact transition states such as those for protolytic cracking display much more similar values. The new method allows us to examine site-specific and ensemble-averaged reactivity systematically and provides a way to assess shape selective catalysis for complex reactants in a quantitative manner.