Why Might PFAS Last “Forever”? A Quantum Chemistry Investigation

Perfluoroalkyl and polyfluoroalkyl substances (PFAS) are known as “forever chemicals” because of their long biodegradation times attributed to their strong carbon-fluorine bonds; they are used in a wide variety of products for their water, grease, and stain-resistant properties. The goal of the this project, entitled Why Might PFAS Last “Forever”, is to apply machine learning to predict the chemical properties of PFAS. To achieve this, we followed the Open (Quantitative) Structure-activity/property Relationship App (OPERA) model from the Environmental Protection Agency (EPA)—which uses the k-nearest neighbors (KNN) algorithm—using pre-screened testing and training sets and new PFAS data; and added quantum descriptors. Our results showed that the testing set did better without partial charges and that the PFAS set did better with all features.



1. Mansouri, K.; Grulke C. M.; Judson, R. S.; Williams, A. J. OPERA models for predicting physicochemical properties and environmental fate endpoints. Journal of Cheminformatics 2018. https://doi.org/10.1186/s13321-018-0263-1.