Per- and polyfluoroalkyl substances (PFAS) are often present as complex mixtures at sites impacted by aqueous film-forming foam (AFFF). These mixtures can arise from products that contain multiple PFAS, the use of AFFF formulations that varied over time, and degradation of primary AFFF components to other PFAS. While most of the mass is typically associated with a handful of specific PFAS, broader and non-targeted analyses show that there can be many other PFAS present in environmental samples in smaller amounts. Rigorous assessment of the ecological risks at AFFF-affected sites must address PFAS mixtures, which, in turn, requires understanding both the potency of individual constituents and the nature of toxic interactions among these constituents.
The hypothesis of this research is that the behavior of PFAS mixtures can be quantitatively characterized through the chronic/sub-lethal testing of representative aquatic species, and that this data can be used to parameterize a robust computational model that can predict the aquatic toxicity of AFFF-relevant perfluorinated compound mixtures, regardless of composition. Such an approach would be suitable for site-specific ecological risk assessment of PFAS mixtures in water, as well as for broader development of environmental guidance, such as water quality criteria. Toward this goal, several research questions will be addressed:
The overall approach for this project is to marry extensive, structured testing of PFAS (individually and in mixtures) with computationally-based modeling to predict the potency and interactive behavior of complex mixtures, even when some components of the mixture may be poorly studied. This project leverages an existing Environmental Protection Agency (EPA) research effort to develop a core set of short-term chronic toxicity data for 11 common PFAS using up to three test organisms. It also relates to an existing SERDP project (ER20-1481) to develop the necessary quantum chemical computational tools.
The initial data set developed pre-project by EPA will be supplemented by additional testing with additional PFAS selected to expand the boundaries of the potency modeling space, thereby ensuring that model parameterization is centered on the key molecular characteristics. In addition, the species examined will be expanded to include short-term chronic toxicity testing with fathead minnow. Endpoints for toxicity testing will emphasize sub-lethal endpoints (growth, development, reproduction) to reflect the demonstrated greater sensitivity of these endpoints compared to short-term lethality.
Toxicity data will be used together with chemical partitioning measurements to validate the quantum chemical methods and will provide a basis for calibrating TSMs for PFAS. The TSM framework has the potential to meet the two basic elements of predicting the toxicity of environmental mixtures of PFAS: (1) A means of quantifying the individual potency of each chemical in the mixture; and 2) a basis for aggregating the individual potencies into a prediction of the overall mixture potency. The TSM is also well suited to address complexities inherent to PFAS, including (1) differential ionization as a function of structure and pH, and (2) differences in uptake and/or potency related to “tail” length and fluorination pattern.
Following initial model development, the project team will conduct additional hypothesis-driven toxicity testing of PFAS mixtures to rigorously document interactive behavior, along with targeted testing of additional PFAS with structures or characteristics that can challenge the underpinnings of the potency model, which may require modification as more is learned.
The expected outcome of this research effort is a comprehensive experimental and modeling approach. It will include the ability to address not only the behavior of AFFF-relevant mixtures with primary components that are well characterized, but also make effective predictions for the potential effects of minor components, or more novel components whose presence may become known but for which there may be limited toxicity data. The approach is also expected to able to predict effects to a range of organisms, thus enabling assessment of broader aquatic ecosystem risk. (Anticipated Project Completion - 2027)