Article pubs.acs.org/est
Estimation of the Acid Dissociation Constant of Perfluoroalkyl Carboxylic Acids through an Experimental Investigation of their Water-to-Air Transport Lena Vierke,*,†,‡ Urs Berger,§ and Ian T. Cousins§ †
Federal Environment Agency, Section for Chemicals, Wörlitzer Platz 1, Dessau-Roßlau, Germany Institute of Sustainable and Environmental Chemistry, Leuphana University Lüneburg, Scharnhorst Str. 1, Lüneburg, Germany § Department of Applied Environmental Science (ITM), Stockholm University, Svante Arrhenius väg 8, Stockholm, Sweden ‡
S Supporting Information *
ABSTRACT: The acid dissociation constants (pKas) of perfluoroalkyl carboxylic acids (PFCAs) have been the subject of discussion in the literature; for example, values from −0.2 to 3.8 have been suggested for perfluorooctanoic acid (PFOA). The dissociated anionic conjugate bases of PFCAs have negligible air−water partition coefficients (KAWs) and do not volatilize from water. The neutral acids, however, have relatively high KAWs and volatilization from water has been demonstrated. The extent of volatilization of PFCAs in the environment will depend on the water pH and their pKa. Knowledge of the pKas of PFCAs is therefore vital for understanding their environmental transport and fate. We investigated the water-to-air transfer of PFCAs in a novel experimental setup. We used ∼1 μg L−1 of PFCAs in water (above environmental background concentrations but below the concentration at which self-association occurs) at different water pH (pH 0.3 to pH 6.9) and sampled the PFCAs volatilized from water during a 2-day experiment. Our results suggest that the pKas of C4−11 PFCAs are 10 (controlled with pH-indicator paper). This was necessary in order to prevent further sorption or volatilization of the analytes. Subsequently 20 mL of methanol were added, the tubes were capped, ultrasonicated for 30 min at room temperature and a 200 μL aliquot was transferred into a PP-vial for instrumental analysis. For samples from experiments at pH 0.0 and pH 0.5, it was necessary to centrifuge the tubes before an aliquot was withdrawn due to the formation of a white precipitate. To analyze the PFC(A)s and PFS(A)s sorbed to the vessel walls, the PP-vessels were cut with a knife (Figure 2). The walls of the top and bottom parts of the vessel were extracted separately by rinsing with a 1 mL aliquot of NaOH in methanol (2.5 mol L−1) and with 50 μL IS (2 μg L−1 IS for each analyte except for 6 μg L−1 for 8:2 FTUCA). For instrumental analysis, 100 μL of the extract was transferred into a PP-vial containing 100 μL water. As can be seen from Figure 2, the cutting line was a little higher than the water surface to ensure that no PFC(A)s from the aqueous phase are sorbed to the top part of the vessel, representing the gaseous phase. The amount sorbed from the gaseous phase to the small surface area above the water surface and below the cutting line was expected to be relatively low and introduce only a small error to the mass balance calculations. The cutting and rinsing of each vessel was done right after transferring the water into the tubes and adding the IS. Therefore, exposure of the vessel to air was only approximately 2 min, minimizing contamination and loss from the vessels.
analytes including internal standards (IS, used for quantification in chemical analysis) can be found in the Supporting Information (Table S1 and S2). Except for 8:2 FTUCA, which was dissolved in methanol, all target compounds were available as crystalline standards and were dissolved in methanol (LiChrosolv Merck) and diluted with water; 560 μL of that dilution contained approximately 60 μL of methanol. The dilution was blown down with a stream of nitrogen at 37 °C until 120 μL of solvent had evaporated (60 μL methanol and 60 μL water, determined by weighting). As methanol has a higher vapor pressure than water and water and methanol do not form an azeotrope, it is expected that all methanol was removed from the final standard. Bottled water was used for the experiments (HiPerSolv Chromanorm VWR) and the pH was adjusted with sulfuric acid (95−97% Sigma Aldrich). Sodium hydroxide (98.6% J. T. Baker) solutions were prepared in water (17 mol L−1) and in methanol (2.5 mol L−1). Experimental Setup. In a first pilot experiment we attempted to investigate the transfer of PFCAs from a spiked water reservoir (donor solution) at a range of different pH via the gas phase to a second unspiked water reservoir at neutral pH (acceptor) within two connected polypropylene (PP) vessels. At low pH the PFC(A)s were readily lost from the donor solution but they were not recovered in the acceptor water reservoir. However, the mass balances of the PFC(A)s could be closed within the analytical uncertainties by accounting for compounds sorbed to the vessel walls (both the walls in contact with the air space above the water surfaces and the walls in contact with the donor solution below the surface) and compounds remaining in the donor solution. A simpler setup with only a spiked donor solution in a capped PPvessel was therefore used for subsequent experiments (Figure 2). The vessel walls above the water surface in the top part of
Figure 2. Diagram of the vessels used for the volatilization experiment.
the vessel served as a passive air sampler to monitor the transfer of analytes at different pH from the water phase to the gas phase. A single experimental setup consisted of a rectangular PPvessel (100 mL, bulk dimensions 6.5 × 4.0 × 3.6 cm3) with a screw cap filled with 20 mL water (Figure 2). The water was adjusted to a certain pH with sulfuric acid and spiked with the analytes (nominally 20 ng for each PFC(A) and PFS(A) and 60 ng for 8:2 FTUC(A) in 20 mL). Formation of aggregates, which could influence the pKa of PFCAs, has been reported to occur at concentrations >2 g L−120,25,37 and PFOA dimers can form at ∼1 mg L−1.26 As the nominal concentrations in our experiments were approximately 1 μg L−1, formation of aggregates could be excluded. Targeted pH values were 7.0, 4.5, 3.5, 3.0, 2.5, 1.5, 0.5 and 0.0, and the actual pH was determined with a pH meter (PHM210 Radiometer analytical, pH −9 to 23, ±0.2 pH units). It was not practical to investigate the water-to-air transfer with the current setup at pH values 0.99 for all curves). For 8:2 FTUCA, the calibration standards had concentrations of 0.39 to 63 μg L−1 (in 1:1 methanol/water). The IS concentrations in the calibration solutions were constant and corresponded to the concentrations in the sample extracts. Quantification was undertaken using the internal standard method. Relative response factors (relative to the respective IS) were derived from the calibration curve. Quality Assurance. The lowest quantifiable concentrations in the calibration standards were defined as method quantification limits (MQLs). The MQLs were 0.09 μg L−1 for all analytes except for PFB(A) (0.4 μg L−1), PFHx(A) (0.1 μg L−1) and 8:2 FTUCA (0.3 μg L−1). A MQL of 0.09 μg L−1 corresponded to
