Insight into the Defluorination Ability of Per- and Polyfluoroalkyl Substances Based on Machine Learning and Quantum Chemical Computations.

Huiming Cao,Jianhua Peng,Zhen Zhou,Yuzhen Sun,Yawei Wang,Yong Liang

DOI: https://doi.org/10.1016/j.scitotenv.2021.151018

IF: 9.8

2021-01-01

The Science of The Total Environment

Abstract:UV-generated hydrated electrons play a critical role in the defluorination reaction of poly- and perfluoroalkyl substances (PFAS). However, limited experimental data hinder insight into the effects of the structural characteristics of emerging PEAS on their defluorination abilities. Therefore, in this study, we adopted quantity structure-activity relationship models based on machine learning algorithms to develop the predictive models of the relative defluorination ability of PEAS. Five-fold cross-validations were used to perform the hyperparameter tuning of the models, which suggested that the gradient boosting algorithms with PaDEL descriptors as the best model possessed superior predictive performance (R-test(2) = 0.944 and RMSEtest = 0.114). The importance of the descriptor indicated that the electrostatic properties and topological structure of the compounds significantly affected the defluorination ability of the PFAS. For the emerging PEAS the best model showed that most compounds, such as potential alternatives of perfluorooctane sulfonic acid, were recalcitrant to reductive defluorination, whereas perfluoroalkyl ether carboxylic acids had relatively stronger defluorination abilities than perfluorooctanoic acid. The theoretical calculations implied that additional electrons on PFAS could cause molecular deconstruction, such as changes in the dihedral angle involved in the carbon chain, as well as C-F bond and ether C-O bond cleavages. In general, the current computational models could be useful for screening emerging PFAS to assess their defluorination ability for the molecular design of fluorochemical structures. (C) 2021 Elsevier B.V. All rights reserved.