To provide a complete and accurate assessment of the environmental risk of PFAS, it is necessary to have reliable tools to evaluate their fate-determining physico-chemical properties (PChPs). However, the experimental determination of these PChPs for more than a few PFAS is infeasible due to lack of pure standards, analytical complexity, and other operational complications (e.g., formation of micelle-like aggregates during standard experimental protocols for determination of Kow). Furthermore, any efforts to measure PChPs will be outpaced by the continuously growing number and diversity of PFAS classes/families that are known to be of concern. The only realistic approach to overcoming this challenge is by using in silico methods that can predict PChPs from their chemical structures.

The overall objective of this project is to provide estimates of PChPs for most PFAS. To do this, the research team will work to extend and apply the in silico methods that they have been developing for predicting properties of insensitive munitions constituents/compounds (IMCs), and chlorinated solvents, to PFAS. This study will integrate and reconcile mined data with computed estimates, in order to provide greater confidence in the reliability of the recommended property values. The research team plan to deliver model estimates and curated data for whole families of PFAS congeners in the form of a searchable database. Since the models will be based on the molecular structure of the compounds of interest, they will have the flexibility to accommodate a range of diverse structural variations. In this way, the estimation tools developed will have a wide applicability domain, thereby allowing for estimates of newly identified PFAS or PFAS alternatives.

The approach juxtaposes experimental data mining and modeling tasks within each Objective, so that each can benefit from the results of the other. Objective 1 (Structure versus Properties) will determine the relationship between PFAS molecular structure and the values of PChP proxies for which measured data are readily available (mainly chromatographic retention time). Using the proxy data, models that are based on molecular descriptors (computed with quantum chemical methods) will be developed to explain the trends in PChPs with molecular structure (e.g., level of fluorination and functionalities). Objective 2 (Property Estimation) will adapt and use existing quantitative structure property relationship (QSPR) and polyparameter linear free energy relationship (pp-LFER) models to predict PChPs for PFAS, using the validated molecular descriptors. The scope will include most of the commonly studied PFAS (e.g., U.S. EPA 24 PFAS list) as well as PFAS in novel families.

The estimates of PChPs for PFAS using the predictive models will allow much more complete and systematic assessment of PFAS occurrence, fate, effects, and treatment technologies by enabling a wide range of specific tasks such as (i) prioritizing PFAS congener toxicity studies based on bioavailability, (ii) matching specific PFAS structures to chromatographic retention times, and (iii) evaluating the potential for removal of PFAS from solid matrices using chemical methods. In this way, this project is responsive to several research priorities of SERDP, including improving the fundamental understanding of environmental processes controlling the fate and effects of PFAS and the design and assessment of remedies for PFAS contaminated sites. (Anticipated Project Completion - 2023).

Murillo-Gelvez, J., O. Dmitrenko, T.L. Torralba-Sanchez, P.G. Tratnyek, and D.M. Di Toro. 2023. pKa Prediction of Per- and Polyfluoroalkyl Acids in Water Using In Silico Gas Phase Stretching Vibrational Frequencies and Infrared Intensities. Physical Chemistry Chemical Physics, 25:24745-24760. doi.org/10.1039/D3CP01390A.

Torralba-Sanchez, T.L., D.M. Di Toro, O. Dmitrenko, J. Murillo-Gelvez, and P.G. Tratnyek. 2023. Modeling the Partitioning of Anionic Carboxylic and Perfluoroalkyl Carboxylic and Sulfonic Acids to Octanol and Membrane Lipid. Environmental Toxicology and Chemistry, 42(11): 2317-2328. doi.org/10.1002/etc.5716.