Unlocking the Full Potential of Heteroatom-Doped Graphene-Based Supercapacitors through Stacking Models and...

Area of Science:

• Materials Science
• Electrochemistry
• Computational Science

Background:

• Graphene supercapacitors offer superior energy storage.
• Heteroatom-doping enhances graphene electrode properties.
• Optimal doping conditions and synergistic effects remain unclear.

Purpose of the Study:

• To utilize machine learning for predicting graphene supercapacitor capacitance.
• To identify optimal conditions for high-capacitance graphene electrodes.
• To provide a computational tool for electrochemical research.

Main Methods:

• Applied various machine learning models (Light Gradient Boost Machine, Extreme Gradient Boost, Random Forest, etc.) for capacitance prediction.
• Developed a stacking ensemble model to improve prediction accuracy.
• Utilized SHAP values to determine key property influences on capacitance.

Main Results:

• Identified optimal parameters for high-capacitance: large specific surface area (SA), 4-5% nitrogen, 10-15% oxygen, high sulfur content, defect ratio near 1, acid electrolyte, and low current density.
• Developed and validated a predictive machine learning model.
• Quantified the impact of various parameters on capacitance.

Conclusions:

• Machine learning effectively predicts and optimizes graphene supercapacitor performance.
• Established guidelines for designing high-capacitance graphene electrodes.
• The developed model and code offer a valuable resource for future research and applications.