The Enhanced Oxygen Evolution Reaction Electrocatalytic Performance of Zr-Substituted nano-CeO2 with Algorithm-Assisted Optimization

Abstract

Immediate research is needed on the possibility of developing noble-metal-free oxygen evolution reaction (OER) based electrocatalysts to develop sustainable, economical water-splitting technology. Improvement of rare-earth oxide and transition metal catalysts by compositional and process modifications is considered to be quite challenging due to the challenge of a perfect combination of properties with dopant concentration and amount of precursors deposited under certain thermal treatment conditions. This paper reports on the electrocatalytic characteristics of the thin-film electrodes that are fabricated using Zr-doped cerium oxide at varying concentrations of Zr, 1-5 per cent by mass. Electrochemical characterisation was done using linear sweep voltammetry (LSV), Tafel slope analysis, electrochemical impedance spectroscopy (EIS), and the stability test of all compositions. The controlled machine learning (ML) model using gradient boosting regression (GBR) was provided with experimentally determined overpotential values as training data in order to obtain a reasonable setting in a minimum amount of time. The ML model displayed a fast prediction speed and gave insight into how 3 mass%Zr doping leads to the best OER activity. The electrode optimised by doping 3mass%Zr in the CeO₂ revealed a low onset potential of 1.39 V against RHE, an overpotential of 319 mV at 10 mA/cm2, and a low Tafel slope of 85.7 mV dec 1, which exceeds that of the other compositions. The composition exhibited long-term stability in operation, having retained 84 per cent of its initial current density after 48 hours of continuous testing, thus presenting enhanced mass activity. The findings illustrate the synthetic influence of Zr on the electronic divide and the surface picture of the CeO₂, showing the eminence of the procedure of info-driven design as a strategy that can be used to discover catalysts. With the guidance of ML, the framework has the potential to assist the researchers in devising a universal algorithm that can be used to develop high-performance OER electrocatalysts based on non-precious metals in the future.