(84k) Quantification of Self-Interaction Errors in Selective Catalytic Reduction of NOx in Zeolites

Zeolites are porous materials that are utilized as selective catalysts in various industrial reactions. To understand the adsorption energies and barrier heights of these reactions, researchers commonly use periodic density functional theory (DFT). However, DFT suffers from self-interaction errors (SIE), resulting in errors in calculated energies. For example, over-stabilization of delocalized densities in transition state structures results in underestimation of reaction barriers by about 40 kJ/mol and adsorption energies by MAE of 20 kJ/mol. This raises a need to remove SIE to make accurate property predictions. To our knowledge, the extent of SIE on reaction barriers for zeolite-catalyzed reactions has never been assessed.

The Perdew–Zunger self-interaction correction (PZSIC) and locally scaled self-interaction correction (LSIC) improve the prediction of barrier heights of chemical reactions, with LSIC giving better accuracy than PZSIC on average. These methods employ an orbital-by-orbital correction scheme to remove the one-electron SIE. Fermi–Löwdin Self-Interaction correction (FLOSIC) is a method to implement the Perdew-Zunger self-interaction corrections (PZSIC) energy using a Fermi-orbital approach.

In this study, we assess the extent of SIE in the catalytic reduction of NOx using a Cu-SSZ-13 zeolite cluster model. We compare the accuracy of commonly used DFT theory and our PZSIC and LSIC energies to study the thermodynamics and kinetics of NOx reactions. Our hypothesis is that PZSIC and LSIC can improve the predictions of reaction barriers and adsorption energies for these reactions. This study is particularly interesting because of the transition-metal based zeolites, where the transition metal changes oxidation states during the reaction, which has not been previously studied using the FLOSIC method. We have evaluated the density-driven errors by comparing the LDA and PBE energies at self-consistent PZSIC energy and functional-driven errors for these reactions. Our findings will be shared for barrier heights and reaction energies for the NOx reactions.