(179w) Embedded Correlated Wavefunction Theory for Heterogeneous Catalysis

Density Functional Theory (DFT) has been the most widely used electronic structure theory in simulating heterogeneous catalysis. While DFT balances well between computational efficiency and accuracy, it suffers from a single-reference description of electronic structure, electron self-interaction errors, and energetic delocalization errors associated with approximations of the exchange correlation functional. These pitfalls lead to quantitative and even qualitative errors in describing transition metal chemistry and charge transfer processes, which are problematic for simulating heterogeneous catalysis. Instead, embedded correlated wavefunction theory (ECW) represents a more suitable approach to describe electronic structure and to predict reaction barriers in complex catalytic systems.

Within ECW, the periodic system is partitioned into two subsystems - a cluster containing the important catalytic center and its surrounding environment. The interaction between the cluster and environment can be recovered using an embedding potential obtained through density functional embedding theory (DFET). We can then perform the more computational demanding correlated wavefunction (CW) theory computations on the cluster in the presence of the embedding potential. Here, we couple DFET with several CW methods, including complete active space second-order perturbation theory using reference wavefunctions generated from complete active space self-consistent field theory and the random phase approximation. Here, we test the accuracy of these CW methods by predicting experimentally measurable adsorption energies on metallic surfaces. Our work explores new computational methodologies that goes beyond standard DFT calculations to enable qualitatively and quantitatively reliable modelling of heterogeneous catalysis.