Computational framework for multi-objective optimization of activated biochar properties using machine learning and...

Area of Science:

• Materials Science
• Chemical Engineering
• Environmental Science

Background:

• Climate neutrality and renewable energy expansion necessitate advanced materials for carbon capture and energy storage.
• Agricultural biochar presents a dual-function opportunity, but traditional pyrolysis methods lack systematic optimization for competing objectives.
• Developing multifunctional materials is crucial for achieving net-zero emissions and sustainable energy solutions.

Purpose of the Study:

• To computationally design multifunctional biochar materials for simultaneous carbon sequestration and electrochemical energy storage.
• To identify optimal pyrolysis process configurations (temperature, time, heating rate, activation chemistry) for maximizing specific surface area, CO2 adsorption, and electrochemical capacitance while ensuring carbon stability.
• To evaluate the potential for large-scale deployment of optimized biochar production in the EU.

Main Methods:

• A simulation-based computational framework was developed, integrating multi-output random forest surrogate modeling with differential evolution algorithms.
• 800 parameter combinations were evaluated across pyrolysis temperature, residence time, heating rate, activation chemistry, and five feedstock types.
• The model's accuracy was validated using test-set R-squared and Root Mean Squared Error (RMSE) for key performance metrics.

Main Results:

• The surrogate model demonstrated high accuracy for specific surface area (R²=0.971) and capacitance (R²=0.942), with moderate fidelity for CO2 adsorption (R²=0.788) and carbon stability (R²=0.497).
• Optimized configurations yielded a specific surface area of 1094 m² g⁻¹, CO2 adsorption of 5.01 mmol g⁻¹, and specific capacitance of 114 F g⁻¹.
• Pyrolysis temperature was the dominant predictor (48% feature importance), and hydrogen-to-carbon ratios below 0.4 indicated suitable materials for long-term sequestration. Optimized processing ensured feedstock-independent carbon stability.
• EU-scale deployment could sequester 53.9 Mt CO2eq year⁻¹, representing 1.2% of total EU greenhouse gas emissions.

Conclusions:

• The study successfully developed a computational framework to optimize multifunctional biochar production for climate and energy applications.
• Identified optimal process parameters and material characteristics pave the way for advanced biochar applications in carbon capture and energy storage.
• The findings suggest significant potential for biochar technology to contribute to EU climate neutrality goals, warranting further experimental validation.