Advancements in photocatalysis and machine learning for bisphenol A (BPA) degradation in aquatic systems: A critical review

Water pollution caused by emerging contaminants (ECs) like bisphenol A (BPA) poses significant threats to aquatic ecosystems and human health. This review examines the fate of BPA and its analogues in water bodies, highlighting recent advances in photocatalysis for their degradation and the integration of machine learning techniques to optimize these processes. BPA, which acts as an estrogen mimic, adversely affects cellular structures in aquatic organisms. Reported maximum BPA concentrations include 6469.2 ng/L in bottled drinking water (India), 1950 ng/L in surface water (India), and 37,200 ng/L in raw wastewater (UK). The highest reported BPA concentration in fish was found to be 482 ng/g in tilapias. Traditional water treatment methods are generally ineffective in removing BPA. However, photocatalysis, an advanced oxidation process, shows great promise for BPA remediation by using semiconductor materials and light to generate reactive oxygen species capable of fully degrading BPA. Composite nano-semiconductors and cocatalysts have achieved remarkable BPA removal rates of up to 100 %. Furthermore, the performance of photocatalytic systems is significantly influenced by factors such as pH and light source. Additionally, this review highlights the growing role of machine learning models, particularly artificial neural networks, which have demonstrated high efficacy in predicting and enhancing the efficiency of BPA degradation processes across various methodologies including photocatalysis, enzymatic degradation, and adsorption. Machine learning models have achieved correlation coefficients (R²) as high as 0.99, enabling substantial optimization of process parameters and contributing to sustainable water quality management. This study aligns with the UN's Sustainable Development Goals related to water, health, and biodiversity.