Zinc oxide photocatalysis for pollutant degradation: a review of elemental doping, morphology, and microstructure, integrated with machine learning-based performance modelling

ZnO nanoparticles are effective photocatalysts for the degradation of organic pollutants; however, there are gaps of knowledge about the effects of their morphology, microstructure, doping elements, and estimation of photocatalytic performance. Accordingly, the effects of morphology, microstructure, and elemental doping are reviewed first. Notably, ZnO microstructure is an important parameter influencing its photocatalytic performance. Specifically, the wurtzite phase, which is the most stable phase of ZnO, has been widely utilized in photocatalytic applications. However, the rocksalt phase of ZnO, a metastable phase, could exhibit higher photocatalytic activity than both wurtzite ZnO and anatase TiO2. In addition to the review study, robust machine learning algorithms, including coupled simulated annealing-least square support vector machine, decision tree-particle swarm optimization, random forest- particle swarm optimization, and extreme gradient boosting-particle swarm optimization, were used to predict the photodegradation yield of various pollutants. The prediction is based on various parameters, including the molecular weight of the pollutants and dopants, topological polar surface area, hydrogen bond donor count, hydrogen bond acceptor count of the pollutants, initial concentration of pollutant, solution pH, light source, weight ratio of doping element to Zn, dosage of catalyst, and reaction time, using ZnO-based photocatalysts. A comprehensive dataset of 1176 entries was gathered from 22 different sources. To the best of our knowledge, for the first time in the field of photocatalysis, a web-based model using extreme gradient boosting-particle swarm optimization method has been developed in Python. This model can be easily accessed online to predict the photodegradation efficiency of various pollutants using ZnO-based photocatalysts.