Data-Driven Approach to Understanding the In-Operando Performance of Heteroatom-Doped Carbon Electrodes

Doping with heteroatoms such as nitrogen and oxygen has been widely practiced to improve the capacitance of carbon electrodes for supercapacitor. However, the role of different heteroatoms and their local atomic configurations on the supercapacitor performance remains elusive, which hampers the rational design of carbon electrodes to achieve high energy density and power density. In this work, machine-learning models are applied to interpret how the surface chemistry affects the inoperando behavior of various carbon electrodes with different structural features such as the specific surface areas of micro- and mesopores. Quantitative descriptions have been established to predict how the configurations of nitrogen-doping and oxygen-doping influence the capacitance and retention rate. The machine-learning models provide insights into the design and possible routes to the synthesis of nitrogen and oxygen co-doped carbon electrodes that maximize the energy storage capacity.