CHALLENGES WITH THE PROVISION OF REFERENCE VALUES FOR PFAS ANALYTES

Abstract

Poly- and perfluoro alkyl substances (PFAS) are of increasing concern due to their wide spread occurrence in the environment, their bio-accumulative properties, slow biodegradation, and public health concerns. Making effective decisions regarding exposure limits and acceptable environmental levels demands high quality and reliable results. Accuracy based proficiency testing (PT) programs using reference values with metrological traceability are a valuable tool to demonstrate that analytical laboratories can deliver results to the required standard. National Measurement Institute Australia (NMIA) conducted a PT program for PFAS (AQA 19-11) using a reference value for perfluorooctanoic acid (PFOA) in drinking water to assess the performance of participants. The reference value was determined by a primary method (Isotope Dilution Liquid Chromatography Tandem Mass Spectrometry) underpinned by a Calibration and Measurement Capability (CMC) in the Key Comparison Database (KCDB) obtained after successful participation in a Consultive Committee for Amount of Substance (CCQM) Key Comparison. Reference values for other PFAS analytes are challenged by the lack of availability of calibrators listed in the KCDB and accepted under the International Committee for Weights and Measures Mutual Recognition Agreement (CIPM MRA). To ensure accuracy of reference value it was important to address the challenges of internal standard equilibration, analyte behaviour during extraction, chromatographic interferences, and matrix effects. The method employed required 48 hours equilibration time for the 13C labelled internal standard, solid phase extraction, and two complimentary chromatographic conditions to demonstrate the absence of matrix interferences and effects. The SI traceable calibrator used was sourced from the National Metrology Institute of Japan (CRM 4056-a). The measurement uncertainty was estimated in accordance with JCGM 100:2008, with component contributions from method precision, gravimetric bias, calibration solution uncertainty, sample density, and potential matrix interferences. The largest contributor to uncertainty was the component associated with the calibration solution, contributing 57% of the total variance. In contrast, analytical precision of the sample measurements contributed to only 14% of the total variance. The expanded uncertainty associated with the reference value was 3.4 % relative. When evaluated against the reference value, 29 of the 30 participants obtained satisfactory z-scores, and 23 participants obtained satisfactory En-scores. While the participant results were generally in agreement with the reference value, those participants with unsatisfactory En-scores may have underestimated their measurement uncertainty. The robust average of participant results is biased low from the spiked value. The reference value confirms that this is only an apparent bias and that the participant methods are generally accurate for PFOA in this sample. The analytical demand for PFAS analysis continues to grow with an ever increasing list of analytes and matrices. The lack of high order SI traceable calibrators constrains provision of accuracy based PT to very few analytes. It will be beneficial for National Metrology Institutes and the International Bureau of Weights and Measures (BIPM) to consider the importance of developing capabilities which would support laboratories world wide and ensure quality of PFAS testing in food and water.