Environmental forensics can play a key role in the identification and management of per- and polyfluoroalkyl substances (PFAS) sources to the environment, and in allocating liabilities arising from addressing PFAS in the environment.
Growing disputes and litigation related to PFAS, increasing US state and federal regulations on PFAS such as the possible designation of certain PFAS as hazardous substances under Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA), commonly known as Superfund, the potential widespread occurrence of PFAS in the environment, and the discovery of PFAS in consumer products, can often necessitate the defensible identification of sources.
However, unlike legacy contaminants, PFAS present unique challenges for forensics investigations: evolving analytical methods, PFAS analyte lists, understanding of PFAS behavior in the environment, and the lack of "unique and universal" source fingerprints due to the widespread use of PFAS over time.
To overcome those challenges, PFAS forensics can be successfully and appropriately applied in a site-specific manner using a multiple-lines-of-evidence approach that incorporates chemical fingerprinting, fate and transport assessment, an understanding of PFAS used and different formulations, and potential release mechanisms associated with the alleged sources.
High quality data
Forensic investigations begin with high quality data, and PFAS are no exception. Laboratory methods to analyze PFAS have evolved to include lower detection limits and longer target PFAS analyte lists, which may provide more power for forensic analyses. Inclusion of target PFAS in laboratory analyses – such as replacement compounds, branched and linear PFAS compounds or transformation products – may assist in source identification, evaluating the timing of the releases or manufacturing sources of PFAS.
In the absence of existing high-quality data, forensic studies may involve the collection of spatially and temporally relevant samples, including characterization of affected environmental media, potential sources and background conditions. As PFAS were widely used and given their tendency for mobility and persistence, diffuse non-point sources have been found to contribute to PFAS in the environment, and therefore can be important in distinguishing the source signal from the noise.
Data analysis
The tried-and-true data analysis techniques used in forensics for legacy contaminants can be applied to PFAS as well. As a first step, evaluation of the concentration or mass distributions can reveal whether one or more potential sources affect the environment and can be complemented by evaluation of compositional profiles, or ‘fingerprints
Figure 1. PFAS fingerprints for surface water sample collected upstream (A) and downstream (B) of a wastewater treatment plant (Savoie and Argue 2022). The change in the PFAS fingerprints and an increase in PFAS concentrations suggest that the WWTP may serve as a source of PFAS.
Differences and similarities in PFAS fingerprints among source(s), samples from the affected area and background can be visualized and assessed through qualitative evaluation of chemical fingerprints (Figure 1), or quantitatively by identifying and calculating compound ratios and using statistical multivariate methods, such as hierarchical cluster analysis (HCA) and principal components analysis (PCA), to visualize large datasets (Figure 2). ‘Receptor models’ such as positive matrix fractionation (PMF) can also be used to identify source patterns and contributions to environmental samples.
The outputs of these data analysis techniques should be interpreted in the context of how PFAS move in the environment.
Figure 2. An example PCA plot where soil samples collected from background locations in Vermont (Zhu et al. 2019) are grouped in clusters based on similarities in their PFAS fingerprints using HCA. Cluster 1 is dominated by PFOS, while Cluster 2 is dominantly a mixture of PFOS and PFOA, and Cluster 3 contains mostly PFOS, PFBS and PFHxA.
PFAS fate and transport in the environment is influenced by multiple factors, including the length of the fluorinated carbon chain, the functional ‘head’ group, whether ‘precursor’ PFAS are present or not, and environmental characteristics, among others. These variable characteristics can result in different PFAS moving in the environment at different speeds and in alterations of PFAS signatures in the environment.
Pulling it all together
Putting the data and findings of the data analysis into the context of the historical operations can provide additional strength to the forensic analysis. Forensics investigations often include a reconstruction of historical operations at the affected site and vicinity, including whether and when PFAS-containing materials were used; the nature of PFAS in the materials over time; how and where wastes were disposed over time; and whether releases occurred that could have introduced PFAS into the environment via an analysis of potential migration pathways.
Emerging tools
Several emerging forensic tools are under development, including advanced analytical methods and predictive data analysis tools. Advanced analytical techniques include suspect screening and non-targeted PFAS analysis using high resolution mass spectrometry (HRMS) that can complement data collected with standard analytical methods. While not quantitative now, suspect screening and non-targeted methods may identify PFAS in samples that may be linked with products or are indicative of transformations of precursor PFAS under certain conditions.
Predictive data analysis tools include machine learning that takes advantage of libraries or large datasets of samples of known quality and of known sources. These tools can predict the potential source of PFAS based on the library used to train the tool. However, understanding the origin and accurate association of the training dataset with respect to the source may be important.
Conclusion
Environmental forensics for PFAS is evolving, however, forensic tools developed for legacy contaminants can be applied using a multiple-lines-of-evidence approach in a site-specific manner. Relying on information on historical operations and PFAS use, distribution of PFAS concentrations in the environment and PFAS fingerprints, it can be possible to identify PFAS sources in a defensible manner.
Ramboll is working to evolve analytical and data analysis methods for PFAS, and enhance understanding of how PFAS move in the environment to help our clients address and resolve questions related to PFAS liabilities.
References
Zhu W, Roakes H, Zemba SG and Badireddy AR. 2019. PFAS background in Vermont Shallow Soils. University of Vermont and Sanborn, Head & Associates. Available at https://anrweb.vt.gov/PubDocs/DEC/PFOA/Soil-Background/PFAS-Background-Vermont-Shallow-Soils-03-24-19.pdf
Savoie JG and Argue DM. 2022. Environmental and quality-control data for per- and polyfluoroalkyl substances (PFAS) measured in selected rivers and streams in Massachusetts, 2020: U.S. Geological Survey data release, https://doi.org/10.5066/P967NOOZ
Case studies from Ramboll’s work assisting clients with PFAS forensics
Delineation of PFAS footprint using statistical tools: A client who used polytetrafluoroethylene coatings on devices they manufacture needed assistance after nearby public and private drinking water wells were found to be impacted with PFAS. Ramboll undertook comprehensive site investigation activities including extensive sampling of potential source materials to define the source PFAS fingerprint in addition to sampling of potentially affected environmental media. We then utilized multivariate statistical tools, including HCA and PCA, to identify the likely extent of the client’s impacts to the public and private wells. The analysis indicated that multiple sources were present that were impacting the private and public water supply wells, including those related to nearby fire training activities.
Identifying and attributing the sources of PFAS: The prior owner and operator of a property now owned by Ramboll’s client was deemed responsible for the PFAS contamination at the property by a regulatory agency. However, the prior owner sought to attribute some of the responsibility to the new owner.
Ramboll’s evaluation of existing data, specifically data on branched and linear PFAS compounds, indicated that the data were consistent with the types of PFAS-containing products used by the former property owner and consistent with the ratios present at the known source areas only used by the prior owner.
Using suspect screening to identify PFAS in groundwater: Groundwater samples collected from a former pharmaceutical manufacturing site had a unique PFAS fingerprint, inconsistent with known potential PFAS sources. Ramboll analyzed the groundwater using HRMS and utilized suspect screening to identify PFAS compounds not analyzed using standard targeted methods. The results agreed with those from the targeted method and also showed a presence of PFAS not included in the targeted method that indicated a potential contribution of PFAS from multiple sources, including a potential firefighting foam contribution.
About the author
Dr. Jaana Pietari has more than 20 years of experience in fate and transport of contaminants and leads Ramboll’s New England emerging contaminants growth team. She specializes in environmental forensics to identify sources of contaminants in the environment. Dr. Pietari is an expert witness, providing litigation support in cases involving PFAS, PAHs, PCBs and other contaminants. She assists clients in identifying potential sources of PFAS in consumer products and works closely with industrial clients to develop strategies for addressing PFAS in their operations, including replacement of PFAS-containing firefighting foams.