Theoretical tuning of local coordination environment of metal-nitrogen doped carbon catalysts for selective chlorine-evolution reaction

The chlorine-evolution reaction (CER) plays a crucial role in the chlor-alkali process, which is a common industrial method for producing chlorine gas. In this study, we use density functional theory (DFT) calculations and machine learning (ML) approaches to explore how the local coordination of metal in metal-nitrogen doped carbon (MNxCy) catalysts affects CER activity and selectivity. We verify the structural stability of MNxCy catalysts through two stability measures: formation energy and dissolution potential. The CER activity and selectivity of MNxCy catalysts can be effectively tuned by adjusting the metal's local coordination. By replacing N coordination to C coordination in the MNxCy structures, we reduce the overpotential for CER while significantly increasing overpotential for oxygen evolution reaction (OER). The local coordination can modify the electronic states of the metal center, and thus strengthen the Cl binding on metal sites. In addition, we conduct a comprehensive feature importance analysis using ML models, considering electronic parameters of the metal and coordinated atoms (C and N), and the chemical bonds of the metal-coordination as input features. Our results show that the metal atom features, specifically the number of valence electrons and the d-band center, are the most impactful, and coordination-related features also contribute to Cl binding tendencies. By adjusting local coordination, we propose the theoretically optimal MNxCy coordination environment of each metal for maximum CER performance. Our theoretical investigation may facilitate the rational design of efficient CER electrocatalysts.