Data-driven design of metal oxide electrocatalysts with low tafel slope in oxygen evolution reaction

In evaluating the oxygen evolution reaction (OER) electrocatalysts, a lower Tafel slope signifies a reduced overpotential requirement to achieve a given current density. Metal oxides have attracted considerable attention as promising OER catalysts due to their natural abundance and high catalytic activity. However, the factors influencing the Tafel slope of metal oxides are inherently complex, encompassing both intrinsic material properties and extrinsic experimental conditions. In this study, we pioneer a data-driven approach integrated with experimental synthesis to investigate the key parameters affecting the Tafel slope of metal oxide OER catalysts. Leveraging a comprehensive database sourced from extensive literature and multiple machine learning (ML) algorithms, we have identified the specific surface area (SSA) as a crucial parameter for optimizing the Tafel slope. This represents a data-driven discovery that has remained unexplored in studies on the OER. Guided by these insights, we developed a sacrificial carbon template method to synthesize 3D mesoporous network NiCo₂O₄ (3D-MN NiCo₂O₄) with structural advantages. This unique architecture achieves high SSA (132.3 m² g⁻¹), uniform nanoscale particles, and enhanced conductivity through its 3D interconnected network, enabling superior mass/charge transport efficiency. The catalyst demonstrates low Tafel slope (38 mV dec⁻¹) across wide current densities, minimal overpotential (0.304 V@10 mA cm⁻²), and good stability (>100 h@10 mA cm⁻²), outperforming RuO₂ and conventional nanoparticles. Remarkably, the synthesis strategy shows generalizability, successfully extending to MnCo₂O₄, FeCo₂O₄, FeNi₂O₄, and high entropy oxides. This work bridges theoretical data analysis with practical catalyst design, establishing a versatile strategy for developing advanced electrocatalysts.