Single-atom catalysts based on two-dimensional metalloporphyrin monolayers for electrochemical nitrate reduction to ammonia by first-principles calculations and interpretable machine learning.

Electrocatalytic reduction of nitrate to ammonia (NO 3 RR) is one of the effective methods to treat environmental pollution caused by nitrate and to produce valuable product of ammonia at the same time. Single-atom catalysts based two-dimensional metalloporphyrin monolayers as effective catalysts for NO 3 RR is investigated in this study by first-principles calculations and interpretable machine learning. Trough high-throughput four step screening procedures, three catalysts (NbPP, TaPP and WPP) finally stood out with high stability, catalytic selectivity and activity for NH 3 production (U L = −0.24, −0.28 and −0.33 V respectively). With Δ G *NO3 and d band center as descriptors, volcano curves were construct to explain the activity trend, in which the three catalysts are just located near the top of the volcano. An effective model for the catalyst performance was built by machine learning to illustrate the possible relationship between adsorption free energy of nitrate and the intrinsic properties of the catalyst, which shows Q TM , N, χ and d TM-N have significant effect on catalyst performance. In all, this study screened effective catalyst and provides theoretical guidance for the rational design of SACs for NO 3 RR. • SACs based on 2D metalloporphyrin monolayers for NO 3 RR investigated. • Four step procedures screened NbPP, TaPP and WPP with high-performance. • Effective model for the catalyst performance built with machine learning. • Q TM , N, χ and d TM-N have significant joint effect on catalyst performance. [ABSTRACT FROM AUTHOR]