Synergistic Effects of Transition Metals and Coordination Environments on Nitrate Reduction in Tetraethynylporphyrin Analyzed by Machine Learning and Verified Using First-Principles Calculations

Electrochemical transformation of nitrate to ammonia (NO 3 RR) offers an up-and-coming way for green ammonia synthesis, simultaneously providing an ecofriendly alternative for addressing nitrate pollution. Analyzed by machine learning (ML) and verified in the light of density functional theory (DFT), four suitable catalysts were selected from 56 metal–organic frameworks (MOFs) based on a single transition metal atom (TM) anchored on the tetraethynylporphyrin (TEP) unit (TM-N 4 /N 3 C 1 -TEP). The relationships between reaction intermediate free energies and catalytic performance are analyzed by machine learning and SHapley Additive ExPlanations (SHAP) global characterization, which show that the cross-descriptor (φ 1 ), the adsorbate electronegativity (φ 2 ), and the TM charge ( Q TM ) are the main contributors to the stable adsorption of nitrate. The synergistic interaction between the TM center and its surrounding coordination environment is identified as a key mechanism for enhancing the NO 3 RR performance. Additionally, the methodology offers a fresh perspective for identifying critical activity descriptors directly from multistep catalyst screening results, offering a direction for breaking the black box of machine learning.