top of page

May 24, 2022

TIG Environmental at Battelle Chlorinated and Recalcitrant Compounds Conference

Dr. Bock spoke on the application of statistical fingerprinting to PFAS to identify and assess multi-source contributions.

2022 Battelle Chlorinated Conference

Date:  May 24, 2022

Title: The Unique Challenges Associated with Applying Statistical Fingerprinting to PFAS

Session: F4: PFAS Source and Forensic Considerations (Dr. Bock – Session Chair)

Statistical fingerprinting is routinely applied to chemical mixtures such as PCCD/Fs, PCBs, and PAHs, using methods that are widely published and tested in the courtroom. Per- and polyfluoroalkyl substances (PFAS) are another diverse class of chemicals that have been used in numerous industrial and commercial applications and products (e.g., coatings, AFFF, manufacturing). Because there is increasing evidence for adverse human health and ecological effects from PFAS exposure, there is a critical need to understand the contributions from different sources. Statistical fingerprinting provides a means to explore sources and define the footprints associated with different sources. However, there are serious challenges associated with the application of traditional fingerprinting methods to PFAS. For example, additional compounds are frequently added to the analytical methods. Most multivariate forensic methods require a consistent compound list, forcing a choice between maximizing the number of compounds under evaluation and the number of samples that can be included. Furthermore, even though detection limits are rapidly falling with advances in analytical methods, detected concentrations of many PFAS substances can fall below the detection limit at modest levels of dilution/mixing. Finally, different PFAS substances have different physico-chemical characteristics such as Kow meaning that the relative concentrations of different PFAS compounds can change as a groundwater plume migrates. Dr. Bock and colleagues Nick Rose and Tim Negley sought to develop strategies that can be used to address these various challenges.

Simple simulations were used to model the effects of differing compound lists, dilution/mixing of target PFAS substances to concentrations below the detection limit, and the changes in the relative concentrations of different PFAS substances as a plume travels through the subsurface. Such models demonstrate how these processes manifest themselves during data analysis and provide strategies by which the forensic scientist can identify and account for these processes during interpretation.

Through this work, it was found that simple strategies can be used to compensate for variable compound lists. By conducting analyses that maximize the number of compounds and conducting parallel analyses that maximize the number of samples we can maximize the amount of data and statistical power. By modeling the effects of dilution on sources and considering the absolute concentrations in the receptor samples, we can identify samples in which the original PFAS profile is masked by non-detects and consider this as a confounding factor in the analysis and interpretation. The lack of relevant partition coefficients and retardation factors can result in considerable uncertainty in the modeling of changes in profiles with subsurface transport with groundwater, we can predict the form of these changes for different PFAS classes and molecular weights. We can use these predictions to carefully review changes in PFAS profiles and identify changes in the profiles consistent with this process. Thus, by carefully considering these factors and by making sure we include other lines of evidence such as discharge pathways, concentration gradients, and transport pathways, we can develop a robust interpretation of the data and develop accurate conceptual models of PFAS sources and fate and transport.

TIG Environmental Staff

View bio

Michael Bock, PhD

Managing Director

Environmental Forensics Services Lead

View bio

Timothy Negley

Managing Director

Data Analytics & Visualization Services Lead

View bio

Nicholas Rose

Principal Scientist

bottom of page