Designing highly efficient C-coordinated single-atom catalysts for CO₂ reduction through modulating metal center and...

Area of Science:

• Materials Science
• Electrochemistry
• Computational Chemistry

Background:

• Electrochemical CO2 reduction (CO2RR) is vital for carbon cycling and mitigating CO2 emissions.
• Single-atom catalysts (SACs) offer maximal atom utilization and high performance for CO2RR.
• Designing efficient SACs requires understanding structure-activity relationships.

Purpose of the Study:

• To design and screen a large number of single-atom catalysts (SACs) for CO2RR to CH4.
• To identify key descriptors governing catalyst stability and catalytic activity using machine learning.
• To develop a universal descriptor for predicting SAC performance in CO2RR.

Main Methods:

• Computational screening of 520 TM-C3 surface models on N-doped graphene with heteroatom variations.
• Density Functional Theory (DFT) calculations for stability, CO2 adsorption, reaction mechanisms, and catalytic performance.
• Machine learning regression and classification models trained on DFT data for feature importance and descriptor development.

Main Results:

• 460 stable catalysts were identified from the initial 520 models.
• Electronic/structural properties like Ne(TM) and χ(TM) govern stability, while εd(TM), εp(C3), and εp(HAs) regulate catalytic activity.
• A universal descriptor was developed using SHAP analysis to predict SAC catalytic activity for CO2RR to CH4.

Conclusions:

• Key descriptors for stability and activity in SACs for CO2RR were identified.
• A universal descriptor enables efficient prediction of catalytic performance.
• Highly active and selective SACs for CO2RR to CH4 were determined, offering insights for future catalyst design.