Machine Learning-assisted Optimization Design for Enhanced Oxygen Evolution Reaction Based on Vanadium-doped Nickel-Cobalt Layered Double Hydroxides

Abstract

The increasing demand for sustainable energy has driven significant research into efficient water splitting, particularly the development of electrocatalysts for the oxygen evolution reaction (OER) which is limited by sluggish kinetics. Optimization of the OER process remains, however, a big challenge due to the compositional complexity of multicomponent catalysts and the influences of electrolyte and temperature. In this study, machine learning (ML)-assisted optimization design is performed to enhance the OER performance using vanadium-doped nickel–cobalt layered double hydroxides (NiCoV LDHs) as the catalyst. In the ML framework, a polynomial regression model is systematically trained by experimental datasets to successfully elucidate the correlation between the target feature (overpotential) and the input features (catalyst composition, electrolyte concentration, reaction temperature) with a high coefficient of determination (R2) of 0.842. Based on the optimized input features predicted by the ML algorithm, a superior overpotential of 196 mV is experimentally obtained which is reduced by 21% compared to the best catalytic performance (238 mV) in the original training datasets. Structural and electrochemical characterizations confirm a well-defined layered morphology and efficient charge transfer dynamics for the optimized electrocatalyst. Our results stand as a significant milestone of integrating ML algorithm with experimental synthesis for the rational design and optimization of high-performance, cost-effective OER electrocatalysts.