Key Role of Density\nFunctional Approximation in Predicting\nM–N–C Catalyst Activities for Oxygen Reduction

Metal–nitrogen–carbon (M–N–C)\nmotifs\npresent intriguing structural and electronic properties for a number\nof applications, including as oxygen reduction catalysts. However,\ncomputational investigations of M–N–C-catalyzed reactions\nmust grapple with their complex electronic structures. In the present\nstudy, we evaluate the impact of the density functional approximation\non calculated M–N–C catalyst activities for oxygen reduction.\nUsing metalloporphyrins as model catalysts, we find a significant\nsplit between pure (GGA) and hybrid functionals, with hybrid functionals,\nin particular B3LYP, showing greater agreement with DLPNO-CCSD(T)\nreaction energies. Notably, double-hybrids offered no noticeable improvement\nover the much more computationally efficient B3LYP and PBE0. Other\ndiscrepancies between functionals, as well as an in-depth analysis\nof ground state spin and geometry, are also considered in this work.\nFinally, both hybrid and double-hybrid functionals greatly reduced\nthe gas phase errors associated with the main group molecules in the\noxygen reduction reaction relative to GGA calculations, leading us\nto question the application of widely used empirical corrections to\nO<sub>2</sub>.