Machine Learning-Guided Design of Biomass-Based Porous Carbon for Aqueous Symmetric Supercapacitors

Area of Science:

• Materials Science
• Electrochemistry
• Data Science

Background:

• Biomass-derived porous carbons are sustainable, cost-effective materials for supercapacitors.
• Optimizing their physicochemical and electrochemical properties is crucial for high performance.
• Limited studies correlate synthesis parameters with supercapacitor metrics.

Purpose of the Study:

• To develop a machine learning (ML) model for predicting specific capacitance of biomass-derived carbons.
• To identify key synthesis parameters influencing supercapacitor performance.
• To uncover interrelationships between material properties, synthesis conditions, and performance.

Main Methods:

• Utilized ML algorithms to analyze experimental data from prior research.
• Trained a model to predict specific capacitance (F/g) based on material characteristics and processing.
• Investigated factors including biomass type, electrolyte, activating agent, and synthesis temperatures/durations.

Main Results:

• Identified an optimal honeydew peel to H3PO4 ratio (1:4) and activation temperature (500 °C) for porous carbon synthesis.
• Achieved a specific capacitance of 611 F/g at 1.3 A/g using a symmetric device with 1 M H2SO4.
• Demonstrated strong correlation (0.8473) between surface area and pore volume; ML predictions matched experimental results.

Conclusions:

• The ML-assisted framework provides insights into critical parameters for supercapacitor performance.
• This approach facilitates the rational design of advanced energy storage materials.
• Biomass-derived carbons with optimized properties show significant potential for supercapacitor applications.