Answers are grounded in this paper and cite their sources.
The electrochemical transformation of nitric oxide into ammonia (NH3) presents a promising strategy for sustainable NH3 production, offering an environmentally friendly alternative to address atmospheric NO pollution. This study establishes a database of metalloporphyrins having two-dimensional porphyrinic covalent organic frameworks with isolated transition-metal-N4 sites. The transition metal porphyrin units (TMPh) are interconnected through conjugated linker environments (DX) containing C, N, O, or F atoms, denoted as TMPh-DX (X = C, N, O, and F). We systematically investigate the intrinsic relationships between NORR catalytic performance and key catalyst characteristics using interpretable machine-learning (ML) on the basis of the “four-plus-one” step high-throughput screening results from density functional theory (DFT). Using this framework, we initially screened 15 promising catalysts from a pool of 112 candidates. A comprehensive analysis of these 15 candidates identified six computationally promising NORR candidates with limiting potentials below −0.1 V, suggesting a joint influence of transition metal centers and electron withdrawing groups in ligands. To overcome the limitations of the “black box” in machine learning, we employed SHAP analysis coupled with SISSO symbolic regression, to identify key factors influencing the Gibbs energy changes in the three fundamental steps. Symbolic expressions were derived to predict Gibbs energy variations in the three reaction steps. This workflow can be extended to screen catalysts for other electrochemical reactions efficiently.
No takes yet. Share an insight, caveat, or question.
Li et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: