Data-Guided Exploration of Process Control in Carbon-Based Catalyst Design for Two-Electron Oxygen Reduction

Area of Science:

• Electrochemistry
• Materials Science
• Computational Chemistry

Background:

• Electrochemical hydrogen peroxide (H₂O₂) synthesis offers a sustainable alternative to traditional methods.
• Developing efficient carbon-based electrocatalysts for the two-electron oxygen reduction reaction (2e⁻ ORR) is critical for H₂O₂ production.
• Experimental optimization of catalysts and processes is time-consuming and resource-intensive.

Purpose of the Study:

• To develop machine learning models for understanding the impact of process control and carbon-based catalyst design on 2e⁻ ORR performance.
• To identify key parameters influencing H₂O₂ selectivity and current density in the 2e⁻ ORR.
• To provide a data-driven approach for optimizing electrochemical H₂O₂ synthesis.

Main Methods:

• Development of machine learning models to predict H₂O₂ selectivity and current density.
• Analysis of model outputs to determine the influence of catalyst composition (e.g., nitrogen doping, oxygen content, defect density, carbon content) and process parameters.
• Preliminary validation of the developed models using independent experimental data.

Main Results:

• Optimal models achieved high accuracy with R² values of 0.959 for H₂O₂ selectivity and 0.831 for current density.
• Nitrogen doping and oxygen content were identified as significant factors for enhancing H₂O₂ selectivity.
• The ID/IG ratio (defect density) and carbon content were found to be crucial for improving current density.
• Model validation confirmed its accuracy in real-world scenarios.

Conclusions:

• Machine learning provides an efficient approach to optimize electrochemical H₂O₂ synthesis.
• Catalyst design, specifically nitrogen doping, oxygen content, and defect engineering, plays a vital role in enhancing 2e⁻ ORR performance.
• The study offers valuable insights into controlling process parameters and catalyst design for scalable H₂O₂ production.