Experimental work was undertaken to test whether gaseous perfluorooctanoic acid (PFOA) sorbs to glassfibre filters (GFFs) during air sampling, causing an incorrect measure of the gas-particle equilibriumdistribution. Furthermore, tests were performed to investigate whether deactivation by siliconisationprevents sorption of gaseous PFOA to filter materials. An apparatus was constructed to closely simulate ahigh-volume air sampler, although with additional features allowing introduction of gaseous test compoundsinto an air stream stripped from particles. The set-up enabled investigation of the sorption ofgaseous test compounds to filter media, eliminating any contribution from particles. Experiments wereperformed under ambient outdoor air conditions at environmentally relevant analyte concentrations.The results demonstrate that gaseous PFOA sorbs to GFFs, but that breakthrough of gaseous PFOA on theGFFs occurs at trace-level loadings. This indicates that during high volume air sampling, filters do notquantitatively capture all the PFOA in the sampled air. Experiments with siliconised GFFs showed thatthis filter pre-treatment reduced the sorption of gaseous PFOA, but that sorption still occurred atenvironmentally relevant air concentrations. We conclude that deactivation of GFFs does not allow forthe separation of gaseous and particle bound perfluorinated carboxylic acids (PFCAs) during active airsampling. Consequently, the well-recognised theory that PFCAs do not prevail as gaseous species in theatmosphere may be based on biased measurements. Caution should be taken to ensure that this artefactwill not bias the conclusions of future field studies.

The volatilisation of perfluorooctanoic acid (PFOA) was measured experimentally at a range of pH valuesusing a previously published laboratory method. Water-to-air transfer was studied for five structuralisomers, namely: the linear isomer (n-PFOA) and the four most commonly occurring branched isomers(3-, 4-, 5- and 6-PFOA). The influence of water concentration and water type on the pH-dependent waterto-air transfer was also investigated for n-PFOA. The water-to-air transfer was studied over the course of 48 h at pH values ranging from 0.2 to 5.5. Under all experimental conditions tested, the volatilisation ofPFOA was negligible at pH > 2.5. In experiments performed with MilliQ water, volatilisation increasedwith decreasing water pH. In experiments performed with tap water and lake water, maximum volatilisationwas observed at pH 1. The concentration of PFOA in water had no influence on the pH value atwhich water-to-air transfer was observed (i.e. at pH < 2.5) for the concentration range tested (0.1e50 mg/L PFOA in deionised water). Although the percentage of PFOA volatilised was significantly different forthe four branched isomers at low pH, volatilisation was not observed above pH 2.5 for any branchedisomer suggesting that all PFOA isomers have a low pKa. Overall, these laboratory results demonstratethat volatilisation of any structural isomer of PFOA from water is negligible at environmentally-relevantconditions. It is unlikely that PFOA isomers will be fractionated in the environment as a result of volatilisationbecause it is a process of negligible environmental relevance.

To evaluate the relevance of different proposed sources of perfluoroalkyl acids (PFAAs) to air, PFAA homologue and isomer patterns were analysed in precipitation collected from five geographical locations suspected to be influenced by these different types of sources. In Chinese samples, perflourooctanoic acid (PFOA) dominated the homologue patterns with concentrations exceeding those observed in European samples by a factor of 2-59. Furthermore, the isomer pattern of PFOA in Chinese samples was close to that of technical standards for PFOA manufactured using electrochemical fluorination (ECF). This finding suggests that direct manufacturing emissions of PFOA to air dominate the loading of PFOA in air in these regions of China. The PFAA homologue pattern observed in precipitation sampled in Stockholm was dominated by perfluorohexanoic acid, perfluorononanoic acid (PFNA) and PFOA. The isomer pattern of PFOA was to a large degree linear indicating that telomer-based sources, i.e. most likely atmospheric degradation of fluorotelomer alcohols, dominate in this urban area. Samples from the Azores, a group of islands located in the open Atlantic, were dominated by perfluorooctane sulfonic acid (PFOS), PFNA and PFOA, which displayed an isomer patttern suggestive of input from both ECF and telomer sources. A similar pattern was observed in precipitation collected during winter months at a rural site on the Swedish west coast, while samples taken at this site during summer displayed homologue and isomer patterns more like those observed in Stockholm. We hypothesize that the contribution from ECF PFOA observed in European samples was due to influence from sea spray aerosols and this requires further detailed investigation.