Tuned layered double hydroxide-based catalysts inducing singlet oxygen evolution: Reactive oxygen species evolution mechanism exploration, norfloxacin degradation and catalysts screen based on machine learning

Abstract

Heterogeneous Electro-Fenton (Hetero-EF) is pursued as a booming technique for real effluent treatments. Here, a novel Simulation-Experiment-Prediction framework has been conceived to reveal atomic-level ROS evolution mechanisms and screen robust catalysts for Hetero-EF reaction. Dynamics-static calculations unveil that the spontaneous conversion pathway of 1O2 generation, which elucidate the transformation from *O2, *HO2, *O2− to *1O2, and confirm optimal Co3Fe2-LDH catalysts with lowest rate-determining step and energy barrier (0.82 eV) of 1O2 generation providing a reasonable and feasible strategy fabricating and tuning Hetero-EF catalysts for realistic contaminants. Subsequently, the Co3Fe2-LDH was verified with best catalytic performance and 1O2 yield, which increased by exceeding 20% on NFXN decay and the treatment time was shortened ∼2 h. Machine learning establishes a Gradient Boosting Regressor model for exploring appreciable catalysts for environmental remediation. Our results inspired the exploration and application of quantum chemistry and artificial intelligence for environmental catalysis, remediation and real wastewater purification.