Data-Driven Optimization of Carbon Electrodes for Aqueous Supercapacitors

Doping carbon electrodes with heteroatoms such as nitrogen and oxygen proves effective in improving the performance of aqueous supercapacitors. However, the optimal conditions of N/O doping remain elusive due to the complexity of the porous structure and electrochemical behavior. While physics-based models face challenges in capturing the pseudocapacitance effects, direct empirical correlation of the capacitance with machine-learning (ML) methods may lead to erroneous predictions. In this work, we introduce a Gaussian process regression (GPR) method using a physical model as prior knowledge to limit the coupling effects of different input parameters. The physics-informed GPR proves effective in characterizing the capacitive behavior of N/O-codoped carbon electrodes in both 6 M KOH and 1 M H2SO4 aqueous solutions. Our machine-learning model suggests that the performance of aqueous supercapacitors can be maximized under acidic conditions by enhancing both the mesopore surface area and the O/N doping ratio of carbon electrodes.