Diatomic catalysts (DACs) hold inherent superiority in atomic economy and synergistic catalysis for complex multi-intermediate reactions. However, several fundamental challenges persist, including typically low diatomic loading (21 sites g-1. Through an advanced deep learning-powered diatomic recognition method, we unambiguously identify the heterodiatomic configurations with consistently high pairing ratios (60.5%-70.3%) across the DACs. Mechanistic studies combining experimental and theoretical analyses disclose that the strategic incorporation of soft-base phosphorus in the synthesis effectively diminishes coordination dynamics between hard-acid metals and hard-base nitrogen ligands via the HSAB principle-driven antibonding interactions, thus achieving the superdense diatomic sites. Significantly, the high-loading DACs demonstrate superior electrocatalytic nitrate reduction performance, exhibiting up to a 2.7-fold enhancement in ammonia yield rates over their low-loading counterparts. This work establishes a general coordination chemistry-based design principle for rational construction of advanced diatomic catalysts.
Cai et al. (2026) studied this question.