Biosynthesis and characterization of a novel supported nanocatalyst for the methylene blue dye photodegradation: Machine learning modeling and photocatalytic activity

Abstract

The present work aims to evaluate the photocatalytic activity of an alternative supported nanocatalyst with copper oxide nanoparticles were the photoactive phase and the analcime nanozeolite as the catalytic support (CuO-NPs@nANA), and to carry out a Machine Learning (ML) study to propose a reaction pathway for the MB dye degradation under visible light. Two machine learning algorithms (RF and XGB) were used in the regression model development from scientific papers concerning MB degradation (by GC-MS) to identify the degradation products of the reaction. XGB algorithm resulted in the best predictive model (R² equals 0.88 and 0.89 for training and testing, RMSE = 33–37), confirming the obtention of carbon dioxide (m/z = 44), water (m/z = 18) and low-molar mass compounds at the final of MB degradation reaction. Moreover, 76.93 % MB removal was reported at pH 10, [MB] = 200 mg L−1 and [CuO-NPs@nANA] = 0.5 g L−1 after 180 min under visible light, with k = 0.0088 min−1. Feature importance study revealed that the response m/w was strongly dependent on pH and reaction time. Therefore, this work confirms the potentiality of machine learning algorithms to develop predictive models for the elucidation of the degradation reaction pathway of organic dyes through heterogeneous photocatalysis.