Unveiling Coordination Effects for Hydrogen Evolution Reaction Using an Interpretable Machine Learning Model.

Area of Science:

• Materials Science
• Computational Chemistry
• Catalysis

Background:

• The hydrogen evolution reaction (HER) is crucial for clean energy production.
• Discovering efficient and cost-effective HER catalysts is a significant challenge.
• Transition metal-doped graphene offers potential as a catalytic material.

Purpose of the Study:

• To accelerate the discovery of novel hydrogen evolution reaction (HER) catalysts.
• To develop an interpretable machine learning (ML) framework for catalyst design.
• To identify promising transition metal (TM)-doped graphene configurations for HER.

Main Methods:

• Combined density functional theory (DFT) calculations with interpretable machine learning (ML) models.
• Computed hydrogen adsorption energies (${\text{Δ}E}_H^{\ast }$) for 140 single-atom configurations.
• Utilized XGBoost for prediction and SHapley Additive exPlanations (SHAP) for feature importance analysis.

Main Results:

• Identified XGBoost as the optimal ML model for predicting hydrogen adsorption energies.
• Revealed key structure-property relationships governing hydrogen binding using SHAP analysis.
• Highlighted Cu and Ag (C₃N₁) and La (C₀N₄) as promising candidates for HER catalysis.

Conclusions:

• Established a data-driven, interpretable framework for designing graphene-supported single-atom HER catalysts.
• Provided coordination guidelines for prioritizing experimental catalyst development.
• Demonstrated the synergy between DFT and ML in accelerating materials discovery.