PFOA - ACS Publications - American Chemical Society

Nov 7, 2008 - Quality Centre, Trent University, 1600 West Bank Drive,. Peterborough, Ontario, K9J 7B8 Canada. Received July 23, 2008. Revised manuscri...
85 downloads 0 Views 171KB Size
Environ. Sci. Technol. 2008, 42, 9283–9288

Experimental pKa Determination for Perfluorooctanoic Acid (PFOA) and the Potential Impact of pKa Concentration Dependence on Laboratory-Measured Partitioning Phenomena and Environmental Modeling D A R C Y C . B U R N S , § D A V I D A . E L L I S , * ,†,‡,§ HONGXIA LI,† COLIN J. MCMURDO,† AND EVA WEBSTER‡ Department of Chemistry, Centre for Environmental Modelling and Chemistry (CEMC), and Worsfold Water Quality Centre, Trent University, 1600 West Bank Drive, Peterborough, Ontario, K9J 7B8 Canada

Received July 23, 2008. Revised manuscript received October 3, 2008. Accepted October 7, 2008.

An accurately measured equilibrium acid dissociation constant (pKa) is essential for understanding and predicting the fate of perfluorocarboxylic acids (PFCAs) in the environment. The aqueous pKa of perfluorooctanoic acid (PFOA) has been determined potentiometrically using a standard water-methanol mixed solvent approach and was found to be 3.8 ( 0.1. The acidity of PFOA is thus considerably weaker than its shorterchain PFCA homologues. This was attributed to differences in molecular and electronic structure, coupled with solvation effects. The pKa of PFOA was suppressed to ∼2.3 at higher concentrations because of the aggregation of perfluorooctanoate (PFO). Often, PFCA partion coefficients are determined at concentrations above those found in the environment. Thus, it was suggested that a pKa correction factor, which accounts for this concentration-dependent shift in acid/base equilibrium, should be applied to PFCA partition efficients before they are implemented in environmental fate models. A pKa of 3.8 ( 0.1 suggests that a considerable concentration of the PFCA exists as the neutral species in the aqueous environment, for example, in typical Ontario rainwater, it is ∼17%. Transport, fate, and partitioning models have often ignored the presence this species completely. The environmental dissemination of PFCAs could, in part, be explained by considering the role of the neutral species.

Introduction Until recently, the environmental fate of perfluorocarboxylic acids (PFCAs) has been almost exclusively associated with the anionic species (i.e., the perfluorocarboxylates, PFCs). For example, ocean water flux, atmospheric decomposition of fluorotelomer alcohol (FTOH) precursors, and atmospheric * Corresponding author e-mail: [email protected]; tel: (705) 748 1011, ext 7898; fax: (705) 748 1625. § Worsfold Water Quality Centre. † Department of Chemistry. ‡ Centre for Environmental Modelling and Chemistry (CEMC). 10.1021/es802047v CCC: $40.75

Published on Web 11/07/2008

 2008 American Chemical Society

marine aerosol transport (1-3) have all been postulated as environmental transport mechanisms for PFCAs; each of these has disregarded transport of the neutral acid species and considered only the anionic carboxylate species. Several modeling studies have attempted to assess the relative importance of these modes of transport (1, 4-6). In each case, it was thought that PFCAs are immediately and almost completely ionized to the corresponding PFC when they partition into water. Thus, transport of the acid was not considered within these models because of its hypothesized low abundance. For example, in modeling the transport of perfluorooctanoic acid (PFOA), which is an environmentally important PFCA, Armitage et al. stated, “...at typical environmental pH of 5-8, >99% of the molecules will be ionized, resulting in environmental fate behaviour that is dominated by the properties of the anion. Therefore, only the environmental fate of PFO is modeled...” (1). In general the acid has been ignored when modeling the fate of PFCAs in the environment, although several researchers have alluded to its potential significance (7-9). It is important to recognize that models for these processes often utilize equilibrium constants such as Kow, Kaw, and Koc to determine partitioning behavior. For PFCAs, which exist in equilibrium between two electronically and structurally unique species (i.e., PFOA and PFO; eq 1), each species may exhibit differing partitioning capacities as dictated by their corresponding unique physicochemical properties. Thus,

movement of these chemicals in the environment will be controlled by the relative proportions of the acid (neutral species) and its conjugate base (anionic species) as well as by the relative magnitudes of the associated partition coefficients. In cases where the difference in relative acid and the base partition coefficient magnitudes are large, partitioning will not be dictated by the relative concentrations of each species. Here, partitioning and the ensuing loss of either species will result in a rapid re-establishment of the acid-base equilibrium. Alternatively, when the relative acid and base partition coefficient magnitudes are similar, partitioning will be determined by the relative concentrations of each species. Thus, the environmental fate of PFCAs, and indeed any chemical that undergoes acid-base chemistry, is governed by the acidity constant (pKa) and also by Le Chatelier’s principle, so that an accurate value of pKa coupled with a knowledge of the physical properties (e.g., partition coefficients) of both species are required for the related environmental fate models. Herein, when discussing the conjugate PFCA and PFC pair together, the acronym PFC(A) is used (e.g., PFO(A)). A recently published study by McMurdo et al. has introduced another potential mechanism for the global dissemination of the conjugate pair, PFO(A), that highlights the relationship between pKa and partitioning phenomena (10). The authors demonstrated, in a laboratory setting, that aqueous aerosols will be enriched in the anion (PFO) at the air-water interface, relative to their source water, because of its surfactant nature. When PFO is protonated at the surface of the water droplet, the neutral acid (PFOA) will readily VOL. 42, NO. 24, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

9283

partition to the gas phase. In short, PFO in bulk water is transported to the atmosphere via aerosol generation, where it is protonated and then rapidly released to the gas phase as PFOA. The authors noted that this will diminish the predicted rainout potential (11), and increase the long-range transport (LRT) potential, of PFOA in the atmosphere via the gas phase. These processes are mainly driven by the remarkable surface activity of PFO and by the magnitudes of pKa and Kaw for PFOA. Despite the fundamental importance of pKa to understanding the environmental fate of PFCAs, there are few reports of experimentally determined values. For instance, only five pKa measurements have been reported for PFOA; each employing different experimental techniques (12-17). These reported values range from 1.01 to 2.8, with the value of 1.01 having been measured for the completely aggregated species and 2.8 for the partially aggregated species. A predicted pKa of zero has also been reported (18). Lo´pezFonta´n et al. reported observing the initial formation of PFO aggregates, which may be dimers or trimers, at concentrations of (4.6 ( 1.4 × 10-12 mol/dm3), that is, 4.6 ( 1.4 pM and premicellar aggregation as commencing at 0.01 mol/dm3, that is, 0.01 M (14). The formation of any aggregated PFO species, regardless of its nature, will reduce the pKa of PFOA, making 4.6 pM the concentration below which the measured pKa is that of the monomeric species. Since there will be little or no aggregation at environmental concentrations (likely less than 4.6 pM, i.e., approximately 2 ng/L), environmental models that use partition coefficients to estimate the capacity for LRT and interpretations of field and laboratory experiments should necessarily incorporate the pKa of the monomeric species rather than those pKa values that have thus far been reported. The objective of this study was to use the well established alkalimetric mixed-solvent titration method (19-25) to accurately determine the pKa of monomeric PFOA and to probe the influence of aggregation on this value. 19F NMR spectroscopy was used as a secondary method to confirm the results. The findings are discussed in the context of the molecular structure and electronics of PFOA. The implications for PFCA transport and fate modeling are presented.

Materials and Methods Hydrochloric acid (HCl), potassium chloride (KCl), potassium hydrogen phthalate (KHP), deuterium oxide (D2O), trifluoroacetic acid (TFA), formic acid, acetic acid, and PFOA (each of purities >99%) were purchased from Sigma-Aldrich and were used without further purification. Potassium nitrate (KNO3) and potassium hydroxide (KOH) (Sigma-Aldrich) were analytical grade and were used without further purification. Methanol (Sigma-Aldrich) was HPLC grade. Each solution was prepared using water that had been filtered through a Milli-Q purification system (18 mΩ, Millipore Corp., Bedford, MA). The pKa of a monoprotic acid can be found potentiometrically using the Henderson-Hasselbalch equations outlined below (eq 2 and 3) Ka )

[H3O+][A-] γH3O+γPFO[HA] γHA

pH ) pKa + log

γA[A-] + log [HA] γHA

(2) (3)

where the bracketed terms correspond to the concentrations of each species and γH3O+, γHA, and γA- are the activity coefficients of the hydronium ion, acid, and conjugate base, respectively (26). This technique, when applied to PFOA, is complicated by two factors. The first is the poor aqueous solubility of the acid (estimated to be 0.026 mg/L, i.e., 6.3 nM 9284

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 24, 2008

(27)), and the second is a tendency for the conjugate base to form aggregates at concentrations exceeding 4.6 pM (14). These two factors preclude the direct determination of monomeric PFOA pKa in a purely aqueous environment. To address these concerns, potentiometric titrations were carried out as indicated below. The use of mixed-solvent pKa measurements, first published in 1925 (19), is a routine method for assessing the pKa of acids with limited aqueous solubility (20-23, 25). Titrations are typically carried out over a range of aqueous cosolvent compositions to yield a series of solvent-dependent apparent pKa values (psKa). The aqueous pKa is extrapolated from a straight-line fit of psKa versus weight-percent organic cosolvent content. Methanol was chosen as the organic cosolvent in this study because its use in mixed-solvent pKa determinations has been examined extensively, and its solvation effects are similar to those of water (22). A detailed methodology outlining the mixed-solvent approach to measuring the pKa of aqueous monomeric PFOA is described in the Supporting Information. It has been suggested (9) that interfacial sorption (e.g., adsorption to glassware) will artificially increase measured pKa values. Consequently, a test for adsorption of PFOA to glassware was carried out whereby total aqueous PFO(A) concentrations were measured as a function of solution volume to glass surface area and as a function of time. The effect of aggregation on pKa was confirmed experimentally by 19F NMR spectroscopy. NMR spectroscopy may be applied to the direct measurement of pKa (28), providing that the acid and base are both soluble at concentrations exceeding the spectrometer detection limit. Details of the method, sample preparation, and NMR spectrometer parameters are given in the Supporting Information.

Results and Discussion pKa Determination for Monomeric PFOA. The methodology and experimental setup for potentiometric titrations of an acid in a mixed-solvent was initially validated using 0.4% acetic acid and 0.2% formic acid solutions (both in 10% v/v methanol/water). These yielded psKa values of 3.78 ( 0.1 for the former and 4.76 ( 0.1 for the latter, which were consistent with the previously published values of 3.745 and 4.758 (29, 30). The same technique was then applied to PFOA (for a representative titration and Gran plot, see Figures S1 and S2 in the Supporting Information). The psKa values of PFOA, which corresponded to the pH at one-half equivalence point (1/2Ve) are summarized as a function of methanol content at different PFOA concentrations in Table S1 of the Supporting Information. As expected, the concentrations of 1-5 mM PFO(A) yielded psKa values that varied nonlinearly with methanol concentration, showing slightly bow-shaped curves. Curves of this shape arise from differences in long-range ion-ion interactions and preferential solvation by either water or methanol at different solvent compositions, and generally lead to skewed estimates of aqueous pKa (22, 31). It is noteworthy that the average psKa was 2.75 ( 0.02 for 5 mM PFOA, that is, very similar to the value reported by Brace (psKa ) 2.80), who performed the titration under nearly identical conditions (15). Yasuda-Shedlovsky treatment of the titration data was applied to all five concentrations to correct for these effects and gain a more accurate pKa measurement for aqueous PFOA. Yasuda-Shedlovsky treatment of mixed-solvent titration data accounts for differences between the solvation effects of water and an organic cosolvent. The technique, when used in combination with proper electrode calibration, yields straight-line relationships between organic cosolvent content and psKa where the y-intercept represents the aqueous pKas

FIGURE 1. Yasuda-Shedlovsky plots showing the variation of psKa with methanol content (expressed as dielectric-1) for PFOA at 5.0 (×), 4.0 (9), 3.0 (O), 2.0 (2), and 1.0 mM (b).

FIGURE 2. Variation of aqueous PFOA pKa with concentration. Values for the concentration-dependent aqueous pKa of PFOA were derived from the y-intercepts of the corresponding Yasuda-Shedlovsky plots. The equation of the line is pKa ) -0.1278 [PFOA] + 3.8437 (R2 ) 0.9059). The pKa of PFOA at infinite dilution (i.e., monomeric PFOA) is 3.8 ( 0.1. (32-34) and are typically in excellent agreement with those would render the solution concentration and changes thereof measured directly in an aqueous environment (22). Here, physically undetectable. In these experiments, the total the resultant Yasuda-Shedlovsky plots for PFOA were linear concentration of analyte (PFO(A)) in solution was monitored and yielded aqueous pKas that varied inversely with acid by LC/MS/MS, both as a function of the solution volume to concentration as shown in Figure 1 and Table S1, Supporting glass surface area and as a function of time. Glass sorption Information. This inverse relationship has also been observed was not detected within the time frame of the experiments for aqueous solutions of perfluoroheptanoic acid (35) (Table conducted (see Supporting Information for experimental S2, Supporting Information) and may be understood in terms details). of aggregation phenomena. At 5 mM, PFO is expected to Relationship between pKa and Aggregated PFO(A). 19F NMR spectroscopy was carried out to further assess the exist in equilibrium between monomeric and aggregated influence of aggregation on the pKa of PFOA. A titration curve states, even in an organic cosolvent such as methanol (36). was generated from twenty-three NMR samples prepared These anionic aggregates bind, sequester, and stabilize the from a series of increasingly acidified buffer solutions. In acid (PFOA) toward hydrolysis (14, 37). The degree of PFO comparing the 19F NMR spectra of the samples it was aggregation is also concentration dependent, such that pKa suppression may be directly related to the concentration immediately obvious that both the chemical shifts and peak profile of PFO aggregation. The aqueous pKa of monomeric intensities varied with solution pH. PFOA was obtained by extrapolating a plot of aqueous pKa The change in chemical shift was consistent with a versus [PFOA] (Figure 2) to infinite dilution and was found reduction in the concentration of the monomeric species to be 3.8 ( 0.1 (R2 ) 0.9). relative to that of the premicellar aggregate (38, 39), indicating As discussed by Goss, the measured pKa of an acid can a correlation between the solution pH and the aggregation be impacted by the neutral acid partitioning to the surface state PFO(A). of the container (9). Goss showed that, in the extreme case, The relative 19F resonance intensities of PFO(A) also diminished as a function of solution pH. The pKa observed where 99.9% of the acid was bound to the surface of the from the inflection point of the curve shown in Figure S5, container, the pKa would be shifted from its true value by up to ∼3 pKa units. The possibility of PFOA adsorption to glass Supporting Information, was 2.3, which was consistent with surfaces was investigated, despite the fact that in the that of an equilibrium-weighted pKa of the monomer and different premicellar aggregates. This was also consistent experiments provided here, a sorption of this magnitude VOL. 42, NO. 24, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

9285

with the observations of Moroi et al., who interpolated a pKa of ∼2.5 for colloidal PFOA from measurements conducted on long- and short-chain PFCA homologues (12). Interestingly, 19F signals corresponding to completely aggregated PFO were still observed at pH 1, which suggested that some fraction of the aggregated species remained in solution even under highly acidic conditions. 19F diffusion-ordered NMR spectroscopy was used to monitor the self-diffusion of PFO and confirmed that the degree of aggregation varied as a function of solution pH (see Supporting Information). Structural and Electronic Basis for the pKa of Monomeric PFOA. The acidity of PFOA, as measured in this study and as reported in the literature, was weaker than that predicted using both a simple QSAR and a comparative analysis approach (9). This can be understood in terms of the structure and electronics of PFO(A) in solution. Fluorination, that is, replacing alkyl hydrogens with fluorine, has been shown to suppress the pKa of linear-chain alkanoic carboxylic acids (40) (Table S2, Supporting Information). Fluorine stabilizes the conjugate base (i.e., the alkanoate) via inductive withdrawal of electron density on the carboxylate anion, so that the acid-base equilibrium is shifted toward the hydronium and alkanoate species (Equation 1). Inductive stabilization of the alkanoate is strongest when multiple fluorine atoms are present and when fluorine is separated from the carboxylate anion by one or two bonds. Fluorine has significantly less effect on carboxylate stability when the number of bonds separating the two moieties increases and the effect is negligible when the separation exceeds four alkyl units (41). These trends are exemplified in the short-chain monomeric PFCAs (C3-C5), whose pKa values are all suppressed relative to their alkanoic acid analogues. The C2-C4 perfluoromethylenyl groups of these acids stabilize the carboxylate anion in a cumulative manner but do not affect the pKa by more than 0.12 units per additional CF2 group (Table S2, Supporting Information). The pKas of the longer-chain (>C5) PFCAs should be influenced by fluorines in the C2-C5 positions of the perfluorinated chain but not by those past the C5 perfluoromethylene moiety (vide supra). By this logic, monomeric PFOA, perfluoroheptanoic acid, and perfluorohexanoic acid are expected to have pKa values that are similar to perfluoropentanoic acid (pKa ) 0.43). However, this is not what has been observed. The pKa of these acids must therefore be mediated by their structures in a manner that necessitates the loss of inductive withdrawal of electron density at the carboxylate head by fluorine. It was first recognized by Bunn, in 1954 (42), that perfluorinated alkyl chains will adopt a helical conformation. This finding has been confirmed on numerous occasions, most recently by vibrational circular dichroism within solution (43). It is generally thought that the unique helicoid structure of these perfluorinated moieties imparts equally unique physicochemical traits, for example, low viscosity, low surface tension, chemical and biochemical inertness, and high gas solubility (36). The chain-length dependent conformation is a direct result of fluorine-fluorine electrostatic interactions within the PFCA tail (43) and is markedly different to the staggered conformation that is observed for nonfluorinated analogues and the small chain length perfluorinated groups. The steeper backbone torsional energy barriers exhibited by long-chain PFCAs (51) and the addition of perfluoroalkyl units in the tail leads to additional backbone rigidification, as confirmed by variable-temperature 19F NMR spectroscopy (44). It has been suggested that conformations of this nature result in the lone-pair fluorine electrons being locked in a delocalized configuration around the alkyl tail, thus altering the molecular orbital energies associated with the perfluorinated backbone. This would be expected to alter 9286

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 24, 2008

the molecular orbital energies, and hence the electron densities, at the site of proton-transfer on these PFCAs. 1H NMR spectroscopy was performed to test for chainlength dependent changes in electron density that are associated with molecular orbitals at the carboxylic acid moiety. In this case, R-COOH chemical shifts (δ) were measured in anhydrous acetone-d6 for PFCAs ranging in length from four to eleven perfluoroalkyl units (Table S3, Supporting Information). An upfield shift of ∼2 ppm was observed for δ as the chain length extended from seven to eight carbons, which was consistent with a significant electronic change in the environment of the carboxylic acid hydrogen. This phenomenon was confirmed using HartreeFock (HF) and density functional theory calculations within our laboratories and remains the subject of continued theoretical investigations. While acidity constants may be predicted using a rigorous quantum mechanical-based approach, the calculation of acid-base parameters remains a challenge even to quantum chemists (45). It is known that pKa often correlates well with quantum mechanical descriptors of local ionization potential, atomic and interatomic charges and highest occupied molecular orbital energies (HOMOs) and the electrons associated with these on the molecule. However, it was noted that these descriptors are not as effective for a number of acid classes. Thus, generalized predictive techniques, such as those used by Goss (9), are often very inaccurate. da Silva et al. (45) have indicated that solvent effects also play a central role in acid-base chemistry because they affect the energy levels of the molecular orbitals associated with the oxygen atoms that are involved in proton transfer. These orbitals have been termed the frontier effective-for-reaction molecular orbitals (FERMOs). FERMO energy levels were calculated and were linearly correlated to pKa, but only when solvent was accounted for in the calculation. Both solvent effects and pKa are directly related to the Gibbs free energy (∆Gaq°) for proton-transfer in aqueous solution (eq 4 and 5). ° ° ° ) Haq - T∆Saq Gaq

(4)

° ) 2.3026RTpKa ∆Gaq

(5)

The solvent effects are manifested in the entropy term (∆S°), which is dominant when acid-base chemistry occurs in solution (46). Consequently, the pKa that has been measured here for monomeric PFOA may also be related to differences in solvation upon rigidification of the long-chain PFCA such that the entropy (∆S°) for proton-transfer is decreased relative to that of the short-chain analogue. The effect of varying backbone rigidity was also observed to impact fluorotelomer alcohol melt entropies (47). The extent to which rigidification affects solvation (∆Saq°) or loss of inductive withdrawal by fluorine (∆Haq°) for PFCAs is currently under investigation. It is clear, however, that the rigid perfluoroalkyl helical twist conformation exhibited by long-chain PFCAs mitigates their acidity, causing an increase in pKa for the monomeric species, and that a fundamental conformational change, brought about by electronic factors, requires the pKa to be different for the short chain PFCAs and their longer chain homologues. The pKa values of monomeric perfluorononanoic acid (C9), perfluorodecanoic acid (C10), and the longer-chain PFCAs (>C10) are expected to be similar or slightly lower than that of PFOA as they are structurally similar. Potentiometric titrations of these acids, the short-chain acids, and the branched-chain acids (which are also not expected to exhibit helicity), along with the appropriate molecular modeling, are currently being conducted. pKa-Based Implications toward the Environmental Fate and Modeling of PFO(A). The pKa of PFOA is an aggregationdependent value. Therefore, a concentration-appropriate pKa

must be employed when interpreting the results of experiments that involve the use of PFOA at concentrations that promote aggregation. Sediment-water distribution experiments have been conducted by Higgins and Luthy (48) at aqueous PFO concentrations of 1-200 nM, that is, 0.4-82.6 µg/L, which are environmentally relevant concentrations and well below concentrations where aggregation is expected to have an effect on the pKa. The organic carbon-water distributions derived from these experiments are expected to be directly applicable to environmental scenarios. Similarly, recent bioconcentration measurements carried out by Martin et al. (49) were conducted at aqueous PFC concentrations of 0.014-1.7 µg/L, that is, about 4 nM. The measured bioconcentration factor (BCF) would reasonably simulate that which occurs at environmental PFC(A) concentrations. However, laboratory experiments may use higher aqueous analyte concentrations to facilitate detection. At these higher concentrations, PFO is partially to fully aggregated, making the measured physical property less environmentally applicable. When laboratory-derived partition coefficients are employed in fate and transport models, a correction factor based on the concentration-dependent pKa of PFOA, must be made to their absolute values. These correction factors should account for differences in aggregation so that the related partition coefficients may be applied to environmental conditions. Models that predict the environmental fate of PFO(A) have, out of necessity, used pKa values (e.g., 1.0-2.8) that have previously been reported in the literature. For example, in 2003, Franklin performed a screening assessment of the potential for PFOA long-range atmospheric transport (LRT) (11). The pKa of the acid was one of the key physical properties that governed its deposition potential via rain out. Franklin assumed a ∼100% homogeneous dissolution of PFO in rainwater based upon a pKa value of 2 for PFOA and then concluded that wet and dry deposition would result in an atmospheric lifetime of a few days at most for PFO(A) emitted from the Earth’s surface. These predictions of low LRT potential, coupled with the observation of PFO contamination in remote regions, further led to the belief that the atmospheric degradation of volatile precursors to PFOA were responsible for the concentrations observed in the arctic (3), so that the deposition of PFO(A) arising exclusively from this source was modeled (6). The atmospheric lifetime of PFO(A) resulting from wet deposition was again considered to be in the order of days. If this study is re-evaluated using a pKa of 3.8, rather than 2, the fraction of PFOA in rainwater is increased by 2 orders of magnitude; clearly, this will result in a significant increase in the expected atmospheric lifetime of gaseous PFOA. For instance, in 2002 the average pH of Ontario rainwater was ∼4.5 and at this pH nearly 17% of PFO(A) would exist as the protonated neutral acid species, which is free to partition to the atmosphere. To compound this, it has been shown that the acid is not taken up homogeneously within the water droplet when it adsorbs to the air-water interface and ionizes, but instead, it remains at the air-water interface because of its surfactant nature, rendering it immediately available for protonation and repartitioning back to the air. As a result, the experimental half-life of PFO(A) adsorbed to an aerosol water droplet was determined to be of the order of seconds (10) with a net flux of PFOA from the water droplet to the air. These factors are all expected to contribute to a reduction in the expected rain-out rate of PFOA. Overall, it can be concluded that PFO(A) rainout rates are likely to be significantly longer than were initially hypothesized, in part, because of the higher than anticipated pKa. Consequently, long-range atmospheric transport mechanisms may be more relevant to the global dissemination of PFOA then previously estimated.

Other global fate models involving PFO(A) have also relied upon previously reported values of pKa to assess atmospheric and oceanic transport potential and their relative contributions to arctic contamination compared to the local degradation of volatile precursors (1, 2, 4, 5). Each of these studies presented the basic assumptions that PFO(A) will be mainly present in its dissociated form because of its low pKa (e.g., ∼0.003% PFOA at a pH of 7 (2)) and that partitioning processes pertaining to the neutral acid can be completely ignored. However, the measured pKa of 3.8 reported here suggests that actually ∼0.1% will be present as PFOA at pH 7 so that even if the physical and partitioning properties that govern the environmental fate and transport of PFOA have less effect than those of PFO, the aqueous concentration of PFOA will be 2 orders of magnitude greater than previously thought. Thus, even at neutral pH, this species should not be ignored in any environmental fate model. Similarly, bioaccumulation of PFO(A) within fish and association with proteins has been assumed to be solely dependent upon the partitioning properties of PFO (49). Again, since the fraction of PFOA present in the aqueous phase is 2 orders of magnitude higher than previously thought, if the partitioning-related properties such as the hydrophobicity of PFOA are of significant magnitude, then the driving force for protein binding may also be dependent upon the neutral species. Furthermore, there are clearly biological situations where PFOA is the predominant form, for example, within the gut, where the pH is ∼2, almost 99% of the total PFO(A) will be present as the neutral acid species. PFO(A) also partitions to sediments with increased sorption occurring as a function of pH, increased organic content, and increased aqueous salinity (48). These observations were explained based upon models that again consider only the conjugate base PFO (50). However, given the increased concentration of PFOA at a pKa of 3.8, the neutral acid species should also be considered in aqueous sediment partitioning phenomena. In conclusion, it is suggested that a model of any aqueousbased system must include the partitioning of both PFO and PFOA coupled with the relative proportions of each given by the pKa of PFOA and pH of the system. A re-evaluation of the general assumptions previously reached, that the presence and influence of PFOA can be excluded from fate models, is warranted.

Acknowledgments We thank the Natural Sciences and Engineering Research Council of Canada (NSERC), DuPont, and the consortium of chemical companies that support research at the Canadian Centre for Environmental Modelling and Chemistry (CEMC).

Note Added after ASAP Publication This paper published ASAP on November 7, 2008 with minor text errors in the Introduction section; the corrected version published ASAP November 11, 2008.

Supporting Information Available Additional materials and methodology describing mixed solvent titrations, the investigation of sorption, and 1H and 19F NMR spectroscopy. This material is available free of charge via the Internet at. http://pubs.acs.org.

Literature Cited (1) Armitage, J.; Cousins, I. T.; Buck, R. C.; Prevedouros, K.; Russell, M. H.; MacLeod, M.; Korzeniowski, S. H. Modeling global-scale fate and transport of perfluorooctanoate emitted from direct sources. Environ. Sci. Technol. 2006, 40, 6969–6975. (2) Prevedouros, K.; Cousins, I. T.; Buck, R. C.; Korzeniowski, S. H. Sources, fate, and transport of perfluorocarboxylates. Environ. Sci. Technol. 2006, 40, 32–44. VOL. 42, NO. 24, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

9287

(3) Ellis, D. A.; Martin, J. W.; De Silva, A. O.; Mabury, S. A.; Hurley, M. D.; ersen, M. P. S.; Wallington, T. J. Degradation of fluorotelomer alcohols: A likely atmospheric source of perfluorinated carboxylic acids. Environ. Sci. Technol. 2004, 38, 3316– 3321. (4) Wania, F. A global mass balance analysis of the source of perfluorocarboxylic acids in the Arctic ocean. Environ. Sci. Technol. 2007, 41, 4529–4535. (5) Schenker, U.; Scheringer, M.; Macleod, M.; Martin, J. W.; Cousins, I. T.; Hungerbuhlert, K. Contribution of volatile precursor substances to the flux of perfluorooctanoate to the arctic. Environ. Sci. Technol. 2008, 42, 3710–3716. (6) Wallington, T. J.; Hurley, M. D.; Xia, J. D. J. W.; Sillman, S.; Ito, A.; Penner, J. E.; Ellis, D. A.; Martin, J. W.; Mabury, S. A.; Neilson, O. J.; Sulbaek Andersen, M. P. Formation of C8F17COOH (PFNA), C7F15COOH (PFOA), and other perfluorocarboxylic acids (PFCAs) during the atmospheric oxidation of 8:2 fluorotelomer alcohol (n-C8F17CH2CH2OH). Environ. Sci. Technol. 2006, 40, 924–930. (7) Ellis, D. A.; Mackay, D.; Butler, J.; McMurdo, C. J. Water Bodies are a Source of PFCAs not just a Sink. Presented at the Society of Environmental Toxicology and Chemistry North America Meeting, Montre´al, Canada, 2006. (8) Yarwood, G.; Kemball-Cook, S.; Keinath, M.; Waterland, R. L.; Korzeniowski, S. H.; Buck, R. C.; Russell, M. H.; Washburn, S. T. High-resolution atmospheric modeling of fluorotelomer alcohols and perfluorocarboxylic acids in the north American troposphere. Environ. Sci. Technol. 2007, 41, 5756–5762. (9) Goss, K. U. The pKa values of PFOA and other highly fluorinated carboxylic acids. Environ. Sci. Technol. 2008, 42, 456–458. (10) McMurdo, C. J.; Ellis, D. A.; Webster, E.; Butler, J.; Christensen, R. D.; Reid, L. K. Aerosol enrichment of the surfactant PFO and mediation of the watersAir transport of gaseous PFOA. Environ. Sci. Technol. 2008, 42, 3969–3974. (11) Franklin, J. Screening Assessment of the Potential for Long-Range Atmospheric Transport of Perfluorooctanoic Acid; U.S. Environmental Protection Agency Docket OPPT-2003-0012-0183; U.S. Environmental Protection Agency: Washington, DC, 2002. (12) Moroi, Y.; Yano, H.; Shibata, O.; Yonemitsu, T. Determination of acidity constants of perfluoroalkanoic acids. Bull. Chem. Soc. Jpn. 2001, 74, 667–672. (13) Ylinen, M.; Kojo, A.; Hanhijarvi, H.; Peura, P. Disposition of perfluorooctanoic acid in the rat after single and subchronic administration. Bull. Environ. Contam. Toxicol. 1990, 44, 46– 53. (14) Lopez-Fontan, J. L.; Sarmiento, F.; Schulz, P. C. The aggregation of sodium perfluorooctanoate in water. Colloid Polym. Sci. 2005, 283, 862–871. (15) Brace, N. O. Long chain alkanoic and alkenoic acids with perfluoroalkyl terminal segments. J. Org. Chem. 1962, 27, 4491– 4498. (16) Igarashi, S.; Yotsuyanagi, T. Homogeneous liquid-liquidextraction by pH dependent phase-separation with a fluorocarbon ionic surfactant and its application to the preconcentration of porphyrin compounds. Mikrochim. Acta 1992, 106, 37–44. (17) Henne, A. L.; Fox, C. J. Ionization constants of fluorinated acids. J. Am. Chem. Soc. 1951, 73, 2323–2325. (18) Goss, K. U. The pKa values of PFOA and other highly fluorinated carboxylic acids. Environ. Sci. Technol. 2008, 42, 5032–5032. (19) Mizutani, M. Die dissoziation der schwachen electrolyse in wasser-alkoholischen losungen. Physik. Chem. 1925, 116, 318– 326. (20) Benet, L. Z.; Goyan, J. E. Potentiometric determination of dissociation constants. J. Pharm. Sci. 1967, 56, 665–680. (21) Albert, A.; Serjeant, E. P. The Determination of Ionization Constants; Chapman and Hall: London, 1984. (22) Avdeef, A.; Comer, J. E. A.; Thomson, S. J. pH-metric Log 3. Glass-electrode calibration in methanol water, applied to pKa determination of water-insoluble substances. Anal. Chem. 1993, 65, 42–49. (23) Rived, F.; Canals, I.; Bosch, E.; Roses, M. Acidity in methanolwater. Anal. Chim. Acta 2001, 439, 315–333. (24) Hemmateenejad, B. Characterization and prediction of solute properties in methanol-water association by chemometrics analysis of solvation data. J. Chemom. 2005, 19, 657–667. (25) Herrero-Martinez, J. M.; Repolles, C.; Bosch, E.; Roses, M.; Rafols, C. Potentiometric determination of aqueous dissociation constants of flavonols sparingly soluble in water. Talanta. 2008, 74, 1008–1013.

9288

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 24, 2008

(26) Harris, D. C. Quantitative Chemical Analysis; 6 ed.; W.H. Freeman and Company: New York, 2003. (27) U.S. Environmental Protection Agency. Estimation Program Interface (EPI) Suite. Available at http://www.epa.gov/opptintr/ exposure/pubs/episuite.htm (accessed June 15, 2008). (28) Berger, S.; Braun, S. 200 and More NMR Experiments: A Practical Course; Wiley-VCH; Weinheim, 2004. (29) Bacarella, A. L.; Grunwald, E.; Marshall, H. P.; Purlee, E. L. The potentiometric measurement of acid dissociation constants and pH in the system methanol-water. pKa values for carboxylic acids and anilinium ions. J. Org. Chem. 1955, 20, 747–762. (30) Johnson, B. J.; Betterton, E. A.; Craig, D. Henry’s law coefficients of formic and acetic acids. J. Atmos. Chem. 1996, 24, 113–119. (31) Bosch, E.; Rafols, C.; Roses, M. Variation of acidity constants and pH values of some organic-acids in water-2-propanol mixtures with solvent compositionsEffect of preferential solvation. Anal. Chim. Acta 1995, 302, 109–119. (32) Shedlovsky, T. The behaviour of carboxylic acids in mixed solvents. In Electrolytes; Pergamon Press: New York, 1962. (33) Shedlovsky, T.; Kay, R. L. Ionization constant of acetic acid in water-methanol mixtures at 25 °C from conductance measurements. J. Am. Chem. Soc. 1956, 60, 151–155. (34) Yasuda, M. Dissociation constants of some carboxylic acids in mixed aqueous solvents. Bull. Chem. Soc. Jpn. 1959, 32, 429– 432. (35) Guo, W.; Brown, T. A.; Fung, B. M. Micelles and aggregates of fluorinated surfactants. J. Phys. Chem. 1991, 95, 1829–1836. (36) Kissa, E. Fluorinated Surfactants and Repellents; 2 ed.; Dekker: New York, 2001. (37) Stainsby, G.; Alexander, A. E. Studies of soap solutions. I. Fatty acid soaps and their hydrolysis in aqueous solutions. J. Chem. Soc., Faraday Trans. 1949, 45, 585–597. (38) Amato, M. E.; Caponetti, E.; Martino, D. C.; Pedone, L. 1H and 19F NMR investigation on mixed hydrocarbon-fluorocarbon micelles. J. Phys. Chem. B. 2003, 107, 10048–10056. (39) Du, D.-Z. 19F NMR study on perfluoroheptanoic acid and perfluorooctanoic acid. Bopuxue Zazhi. 2003, 20, 29–36. (40) Boiadjiev, S. E.; Watters, K.; Wolf, S.; Lai, B. N.; Welch, W. H.; McDonagh, A. F.; Lightner, D. A. pKa and aggregation of bilirubin: Titrimetric and ultracentrifugation studies on water-soluble pegylated conjugates of bilirubin and fatty acids. Biochemistry 2004, 43, 15617–15632. (41) Perrin, D. D.; Dempsey, B.; Serjeant, E. P. pKa Prediction for Organic Acid and Bases; Chapman and Hall: London, 1981. (42) Bunn, C. W.; Howells, E. R. Structures of molecules and crystals of fluorocarbons. Nature 1954, 549–551. (43) Monde, K.; Miura, N.; Hashimoto, M.; Taniguchi, T.; Inabe, T. Conformational analysis of chiral helical perfluoroalkyl chains by VCD. J. Am. Chem. Soc. 2006, 128, 6000–6001. (44) Ellis, D. A.; Denkenberger, K. A.; Burrow, T. E.; Mabury, S. A. The use of 19F NMR to interpret the structural properties of perfluorocarboxylate acids: A possible correlation with their environmental disposition. J. Phys. Chem. A. 2004, 108, 10099– 10106. (45) da Silva, R. R.; Ramalho, T. C.; Santos, J. M.; Figueroa-Villar, J. D. On the limits of highest-occupied molecular orbital driven reactions: The frontier effective-for-reaction molecular orbital concept. J. Phys. Chem. A. 2006, 110, 1031–1040. (46) Edward, J. T. Entropy and equilibriasA reassessment of ionization data for substituted acetic acids. J. Chem. Educ. 1982, 59, 354–356. (47) Liu, J. X.; Lee, L. S. Effect of fluorotelomer alcohol chain length on aqueous solubility and sorption by soils. Environ. Sci. Technol. 2007, 41, 5357–5362. (48) Higgins, C. P.; Luthy, R. G. Sorption of perfluorinated surfactants on sediments. Environ. Sci. Technol. 2006, 40, 7251–7256. (49) Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G. Bioconcentration and tissue distribution of perfluorinated acids in rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 2003, 22, 196–204. (50) Higgins, C. P.; Luthy, R. G. Modeling sorption of anionic surfactants onto sediment materials: An a priori approach for perfluoroalkyl surfactants and linear alkylbenzene sulfonates. Environ. Sci. Technol. 2007, 41, 3254–3261. (51) Borodin, O.; Smith, G. D.; Bedrov, D. A quantum chemistry based force field for perfluoroalkanes and poly(tetrafluoroethylene). J. Phys. Chem. B. 2002, 106, 9912–9922.

ES802047V