Ramasubramanian, B., Dutta, R., Koperski, M., Chellappan, V., & Senthilnath, J. (2026). Multi-property optimization for designing carbon electrodes. Journal of Power Sources, 676, 239872. https://doi.org/10.1016/j.jpowsour.2026.239872
Abstract:
In this paper, we propose a novel approach termed as Multi-Property Optimization of Estimated Symbolic Expressions (MPOESE) to systematically optimize carbon electrode properties using a dataset comprising 300 data points compiled from the literature spanning from 2005 to 2025. The dataset captures key input parameters, including precursor type, synthesis method, activation temperature, process time, activation agents, and atmospheric conditions alongside output parameters such as specific surface area, pore volume, and particle size. The proposed approach, MPOESE, leverages the strengths of symbolic regression (SR) to estimate mathematical expressions of multiple outputs and Bayesian Optimization (BO) to determine their optimal values. Using limited data available in the literature, MPOESE estimates nonlinear relations and multivariate dependencies between the variable synthesis conditions that are not readily discernible through traditional trial-and-error experimentation or statistical regression. These estimated functional mappings allow the adopted multi-objective BO technique to explore the material search space and find the optimal set of solutions. Among them, one specific solution using plastic as the precursor material was selected for its ease of availability and was experimentally validated. The results imply that under controlled pyrolysis with KOH activation at 700 °C and moderate dwell time (∼1 h), these feed stocks can yield carbons with exceptionally high microporosity (>1.2 cm3 g−1) and surface areas above 1150 m2 g−1. This study underscores the role of symbolic approximation and multi-objective optimization in navigating complex material design space, enabling rapid evidence-based recommendations for sustainable precursors, and accelerating the rational design of high-performance carbon architectures for electrochemical energy storage systems.
License type:
Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0)
Funding Info:
This research / project is supported by the Agency for Science, Technology and Research - Advanced Manufacturing and Engineering (AME) Programmatic Fund
Grant Reference no. : A1898b0043
This research / project is supported by the Agency for Science, Technology and Research - A-STAR-SINGA
Grant Reference no. : SING-2021-02-0819