Machine Learning Accelerated Screening Advanced Single‐Atom Anchored MXenes Electrocatalyst for Hydrogen Evolution Reaction

MXene-supported single-atom catalyst (MXene-SACs) systems are considered promising potential electrocatalysts for the hydrogen evolution reaction (HER) due to their excellent conductivity, stability, and hydrophilicity. However, the complex composition of MXene-SACs and the unclear structure-activity relationship limit the rational design of efficient HER catalysts. To address this challenge, a high-throughput screening strategy that integrates theoretical calculations, machine learning (ML), and experimental validation to efficiently identify MXene-SACs with outstanding HER performance is developed. Using a database constructed from theoretical calculations as input, a ML model to screen a batch of potential high-efficiency HER catalysts is built. Based on the distribution pattern of hydrogen evolution barriers predicted by ML and electronic structure analysis, a novel structural descriptor Φ, which can be easily calculated using corresponding properties from the periodic table is derived. The descriptor provides insights into the underlying HER mechanism of MXene-SACs, where electron transfer from surrounding coordinating atoms to the single atom effectively shifts the d-band center to an optimal level (≈ -2.7 eV), minimizing the hydrogen evolution barrier. Guided by this descriptor, Cr₂CO₂-Pt is synthesized, which exhibits outstanding HER performance, achieving a current density of 1 A cm^-2 at an overpotential of 150.7 mV and maintaining long-term stability over 130 h.