CAREER: Quantum Embedding of Wave-Function Methods as Path to High-Accuracy Thermochemistry in Heterogeneous Catalysis

Gerald Knizia of Pennsylvania State University is supported by an award from the Chemical Theory, Models and Computational Methods program to develop an accurate and efficient theoretical approach to heterogenous catalysis. The focus is on reactions that take place on a hard surface, such as a metal, which acts as a catalyst. Catalysts are used in the production of household products and almost all industrial chemicals as they help speed up the chemical synthesis. Almost 25% of the global industrial-sector energy consumption is due to the catalytic production of basic chemicals and fuels alone! A sound microscopic understanding of the processes of chemical reactions at the smallest scale, obtained by computational analysis, could help substantially in developing new catalysts or improving existing ones. However, a key challenge is that the present computational methods applicable to surface reactions are not accurate enough to reliably identify which of the many competing reaction pathways are actually taking place in reality. This research addresses this challenge by enabling the use of high-accuracy computational methods from small-molecule theoretical chemistry, for use in the complex environments of realistic catalysis at surfaces. The research activities are integrated with an educational approach aimed at senior undergraduate and incoming graduate students. The goal of the educational activities is to help these students become proficient users of computational techniques. The broader technical impacts of the research may result in improvement in industrial processes (e.g., reducing waste, increasing energy efficiency, reducing dependence on imported precious metals, etc.), and thereby contribute to the economy of independence of the US. The educational materials will be made available to everyone, and may provide disadvantaged but bright students with poor access to educational resources with a starting point for learning powerful computing techniques.

Concretely, the research targets the development of wave-function based electronic structure methods which have sufficient accuracy (~1 kcal/mol in relative energies) to allow for definitive thermochemical calculations on the surfaces of hard materials, such as current industrial heterogeneous catalysts. First, a suitable quantum embedding framework, based on the Density Matrix Embedding Theory (DMET) is developed. The method is specialized for embedding a single target fragment. This allows the polarization functions needed by thermochemical wave function methods and facilitates to be introduced using fast Kohn-Sham DFT for the mean-field description of the environment. Second, high-accuracy local coupled cluster methods will be developed which are capable of being used in the presence of the embedding. Third, the techniques from the first two steps will be adjusted to incorporate a coupling to an actual periodic surface system via imposing fixed boundary condition from ideal bulk/surface systems (as opposed to imposing periodic boundary conditions on the target system itself).

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.