Comment on “The environmental photolysis of perfluorooctanesulfonate, perfluorooctanoate, and related fluorochemicals”

Abstract

Previous studies have shown that perfluoroalkyl carboxylicacids (PFCAs) and perfluoroalkane sulfonic acids (PFSAs) areextremely persistent in the environment with almost no naturaldegradation pathways (as summarized in risk assessment profilespublished by UNEP (2006) and ECHA (2013)), with the exceptionthat PFCAs may react with OH radicals (Hurley et al., 2004). In con-tradiction to this understanding, Taniyasu et al. (2013) proposedthat long-chain PFSA and PFCA homologues can undergo relativelyrapid photodegradation and form shorter-chain homologues undernatural conditions, based on findings from a combination of fieldand laboratory experiments.We have concerns about the strong conclusions drawn in thepaper by Taniyasu et al. given the absence of essential experimen-tal details (notably the UV light spectrum used in the laboratoryexperiments), internal inconsistencies in the experimental results,and the lack of a mechanism for photolytic degradation. We be-lieve that the observations in the Taniyasu et al. study can beattributed to experimental artifacts. Our three main points of con-cern are outlined below. For a better illustration of our concernsand to help the readers understand them, we compiled thereported percentage reduction of PFSA and PFCA homologues fromTaniyasu et al. after irradiation in the field and laboratoryexperiments in Table 1.[i] A plausible transformation mechanism is essential tounderstand the environmental relevance of their results, butis not provided (the authors stated, ‘‘Our study was not designedto address the mechanism of photolysis’’). When discussing theresults of the field experiments, the authors concluded, ‘‘We be-lieve that both direct and indirect photolysis have played a rolein the photolytic transformation of PFASs in this study’’. However,based on the following reasons, we believe that direct andindirect photolysis under natural light conditions should nothave occurred for (at least) perfluorooctane sulfonic acid(PFOS), perfluorooctanoic acid (PFOA) and perfluorodecanoic acid(PFDA).For the laboratory experiments, the authors did not report thelight conditions used, particularly the UV light spectrum and,therefore, a rational judgment on the environmental relevance ofthese experiments cannot be made. Taniyasu et al. observed signif-icant losses of PFOS, PFOA and PFDA in the experiments usingquartz glass tubes after a short time of irradiation (216 h), whereasno actual losses of these acids were observed in the experimentsusing Pyrex glass tubes after the same irradiation time (Table 1).Considering the lower wavelength cut-off for quartz glass(170 nm) compared to Pyrex glass (275 nm) (http://www2.chemistry.msu.edu/faculty/reusch/virttxtjml/photchem.htm), thephotolysis of PFOS, PFOA and PFDA in the laboratory experimentswith quartz glass was very likely induced by light with awavelength between 170 and 275 nm, which is clearly in the UV-C range. This implication is supported by a further analysis thattakes the settings of the field experiments into account.Regarding direct photolysis: The absorption spectrum of perflu-orooctanoic acid (PFOA) in water has been measured; PFOA showsstrong absorption from the deep UV-region to 220 nm and a weak,broad absorption from 220 to 270 nm (Hori et al., 2004), which isbelow the natural light spectrum at the surface of the Earth(>280 nm). The experimental conditions in the Taniyasu et al.study were different from those in the Hori et al. study, but wedo not believe that the addition of small amounts of methanol tothe water would significantly change the UV absorption spectrumof PFOA. Adding organic solvents such as methanol may change thepolarity of the solution and thus influence the spectra of the targetcompound (Kosower, 1958). However, the very low levels of meth-anol (<2%) in the solutions will have negligible influence on thepolarity of the solutions used in Taniyasu et al. (Kosower, 1958).In addition, it is unlikely that adding less than 2% of methanolhas any effect on the dissociation of PFOA (Sager et al., 1964).Therefore, the low levels of methanol (<2%) in the solutions are un-likely to have changed the absorption spectrum of PFOA comparedto the spectrum that was measured in pure water (Hori et al.,2004).Regarding indirect photolysis: The UV cut-off of methanol is inthe far UV region (<220 nm), i.e. methanol has no absorption oflight with wavelengths above 220 nm (Cheng et al., 2002). There-fore, it is very unlikely that methanol, under the conditions usedby Taniyasu et al. (wavelength above 280 nm, with reduced inten-sity of light from 280 to 315 nm due to Pyrex glass), absorbs andtransfers energy to PFCA and PFSA homologues.[ii] Reported reductions of long-chain PFCAs and PFSAs afterirradiation are not internally consistent between experiments(see Table 1). First, much lower degradation (or no degradationat all) of PFOS, PFOA and PFDA was observed in laboratory exper-iments using the same Pyrex tubes as in the field measurements,