Document aBbnrdR8dgJY853j8DJ5Q3w2B

Determination of Perfluorocarboxyl...Impacted by Fire-Fighting Activity http://pubs.acs.org/cgi-bin/jtextd7esthag/asap/abs/es981355+.htmI i. Technol, ASAP/ ;e D ate: July 3, 1999 lx 2531 JE F F E R SO N DAVIS HW Y ARLINGTONVA 2 2 2 4 M 1 M 1999 A m erican C h em ical Society , , 8)01355-8 IN REPLY REFER TO Determination of Pertluorocarboxylates in Groundwater Impacted by Fire-Fighting Activity Cheryl A. Moody and Jennifer A. Field" Department o f Environmental ft Molecular Toxicology, Oregon Slate University, Corvallis Oregon 97331 Received for review December 29, 1998 Revised manuscript received May 12, 1999 Accepted May 21. 1999 Abstract: Perfiuorinated surfactants are used in aqueous film forming foam (AFFF) formulations, which arc used to extinguish hydrocarbon-fud fires. Virtually nothing is known about the occurrence o f perfiuorinated surfactants in the environment, in particular, at fire-training areas and emergency response sites where AFFF entered groundwater without prior treatment. Strong anion exchange Fmpore disks were used to extract periiuorocarboxylat.es from groundwater collected from fire-training facilities located on Naval Air Station Fallon, NV, and Tyndall Air Force Base, Fi,, The earboxylaics were simultaneously eluted from the disks and dcrivaii/ed to their methyl esters for direct analysis by gas chromatography/mass spectrometry. Periluorocarboxylates containing six to eight carbons were detected in groundwater collected from the two field sites with total concentrations ranging from 123 to 7090 /*g/L. The detection of peril uorocarboxylates at field sites after 7-10 years of inactivity indicates their potential utility as markers for delineating groundwater impacted by fire-fighting activity. ' [Tull text in htmll [Full text in pdfI *5 c. Atir4L>. IS '"V u# \ Vi^ d A . X J c A < A V e i *- of o ^ z "tY ^ 2 l v n c?..-/" f i t e 8/5/99 9:41 AM US00002577 Determination of Perfluorocartoxylates in Groundwater Impacted by Fire-Fighting Activity CHE RYL A. MO O D Y AND J E N N I F E R A. F I E L D * Department of Environmental & Molecular Toxicology, Oregon State University, Corvallis, Oregon 97331 Perfluorinated surfactants are used in aqueous film forming foam (AFFF) formulations, which are used to extinguish hydrocarbon-fuel fires. Virtually nothing is known about the occurrence of perfluorinated surfactants in the environment, in particular, at fire-training areas and emergency response sites where AFFF entered groundwater without priortreatment. Strong anion exchange Empore disks were used to extract perfluorocarboxylates from groundwater collected from fire-training facilities located on Naval Air Station Fallon, NV, and Tyndall Air Force Base, FL The carboxylates were simultaneously eluted from the disks and derivatized to their methyl esters for direct analysis by ga s chromatography/mass spectrometry. Perfluorocarboxylates containing six to eight carbons were detected in groundwater collected from the two field sites with total concentrations ranging from 125 to 7090 pql L The detection of perfluorocarboxylates at field sites after 7 -10 years of inactivity indicates their potential utility as markers for delineating groundwater impacted by fire fighting activity. in AFFF formulations (6) is problematic because the sur factants are nonvolatile and may not contain chromophores. As a result, analytical methods for AFFF formulation com ponents are lacking, and therefore it is difficult to assess their occurrence, fate, and transport in AFFF-contaminated groundwater. Because perfluorinated surfactants co-occur with other pollutants (e.g. fuel components, solvents, etc.) in groundwater, it is important to determine ifperfluorinated surfactants affect the transport and biodegradation of other contaminants. Free and emulsified oil, fuel, and AFFF components were shown to adversely affectactivated sludge processes (6, 7) and the performance of anaerobic sludge digesters (8) in wastewater treatment facilities. For this reason, perfluorinated surfactants may have an adverse affect on groundwater microbial populations and their ability to degrade co-contaminants present in AFFF-contaminated groundwater. In addition to fluorinated surfactants use in fire-fighting foams, they are also utilized in herbicides and insecticides, cosmetics, greases and lubricants, and adhesives (3). Flu orinated carboxylic ad d s of industrial significance include perfluorooctanoic acid (PFC8) and perfluorodecanoic acid (PFC10) (31.There is concern regarding the potential toxidty of perfluorinated carboxylic adds. An in vivo study of rat liver response to PFC10 indicated the rapid onset of a lowlevel heptatotoxicity but no detectable damage to the DNA (10). Perfluorodecanoic ad d and PFC8 have been found to inhibit gapjunction intercellular communication in rat liver epithelialcells (11) and may be involved in tumor promotion <S>- . . In this paper, we describe the isolation, identification and quantification ofperfluorinated carboxylates in groundwater impactedby fire-training activitiesa t NavalAirStation (NAS) Fallon, NV, and TyndallAir Force Base, FL. The development of analytical methods is necessary before investigating the occurrence and distribution of perfluorinated surfactants in AFFF-contaminated groundwater and their effect on co contaminant transport and biodegradation. Introduction Aqueous film forming foams (AFFF) are complex mixtures of surfactants and other components used to extinguish hydrocarbon--fuel fires that occur at fire-training sites as well as in emergency situations. Aqueous film forming foams have been commercially available for fire-fighting applica tions since their development by the United States Navy and 3M Co. in the mid-1960s (J). At fire-training areas that routineiy used AFFF mixtures and military emergency response sites, AFFF-laden wastewater that entered surface water and groundwater without treatment has led to groundwater and soil contamination. For example, perflu orinated compounds were tentatively identified in ground water impacted by fire-training activities at Tyndall Air Force Base (2). Unfortunately, definitive identifications of the perfluorinated compounds were not reported. ' Commercial AFFF mixtures are propreitaiy in nature and typically contain fluorinated and nonfluorinated surfactants (1,3--5). Due to the proprietary nature o fAFFF formulations, the chemical structures of the actual perfluorinated surfac tants used in commercial AFFFs are not known outside the companies thatmanufacture them (5). Moreover, the analysis ofanionic perfluorinated surfactants that are known to occur * Corresponding author phone: (541) 737--2265: fax: (541) 737 0497; e-mail: Jennifer.Field@orst.edu. Experimental Section Standards and Reagents. Standards of PFC8 (98%), perfluorododecanoic acid (PFC12) (95%), and the internal standard, 2-chlorolepidine (99%), were purchased from Aldrich Chemical (Milwaukee, WI). Methyl iodide (neat) was used as purchased from Aldrich ChemicaL Field Sites and Sample Collection. From the mid-1950s to 1988, the crash crew training area at NASFallon, NV (Figure la), was used to conduct fire-training activities, which consisted offlooding a fire pitwith flammable liquids, igniting the fluids, and subsequently extinguishing the fire with fire fighting agents including AFFF {121. For a typical trainingexercise, approximately 75--100 L ofAFFF concentrate were diluted with 1200--3200L ofwater according to specifications (3% or 6% solution) and subsequently employed. During the years o f activity at the NAS Fallon site, training exercises occurred on a weekly to monthly basis. At the NAS Fallon site, groundwater samples were collected from four moni toring wells located within a 120 m radius of the fire pit where the water table is located between 2 and 3 m below the land surface. The Tyndall Air Force Base Fire-Training Area FT-23 was used from 1980 to 1992 for similar activities (Figure lb) (13). Four groundwater samples were obtained from wells sur rounding the fire-training area; the water table is located between I and 2 m below the land surface. All samples were collected in high-density polyethylene brown bottles because l0.10T1ies9813S5* CCC: $18,00 PuBlished ori W eb OOOTOOOO X X X * American Chemical Society ' PA G E EST: S.6 V O L xx. NO, xx, xxxx I ENV1RON. SCI. & TECHNOL. A ' r - y^-Tr fT ...x-TOV ; US00002578 FIGURE 1. M ap of (a) Naval Air Station Fallon and (b) Tyndall Air Force Base field sites indicating location of groundwater w ells and direction of regional groundwater flow. perfluorinated carboxylates adsorb to glass (/4). Samples were shipped on ice without preservation and stored at 4 Cprior to analysis. Solid-PhaseExtraction and Derivalization. Samples (55-- 200 mL) were extracted through 25 mm strong anion exchange (SAX) disks in a manner similar to that described by Field and Reed (IS) with the exception that the SAXdisks were pretreated prior to use to remove interfering disk impurities. Fretreatment consisted ofsoaking ihe disks in 12 mM HCl/acetonitrile for 2 days after which the disks were soaked in pure acetonitrile for several hours. Just prior to use, the disks were rinsed with a minimum of 350 mL of deionized: water in order to sufficiently rinse the HC1 from the disks and wet them prior to passing groundwater samples through (hem. Samples (55--200 mL) were passed through the disks under full vacuum, arid the disks were then allowed to dry. The disks containing the exchanged analytes were placed in a 2 mL autosampler vial together with T mL of acetonitrile, 51.2 /gofinternalstandard, and 100.I.ofmethyl iodide. When heated at 80 C for 1 h, the acids were simultaneously eluted from the disk and derivatized to their methyl esters. Spike and Recovery. Spike and recovery experimentswere performed to determine the precision and accuracy of the SAX disk extraction and in-vial elution method. A set of experiments was performed on groundwater samples from NAS Fallon MW 50U and MW 17 that had been previously determined to contain neither PFC8 nr PFC12 above detection. Duplicate groundwater samples from wells MW SOUand MW 17were spiked to contain a final concentration of 1240 fig/L of PFC8 and 560 figIL of PEC12. Standard addition analyses were performed with NAS Fallon groundwater samples that contained measurable quantities ofPFC8; the samples did not contain PFC12 above detection. Known amounts of PFC8 were added to samples to give a final concentration twice that of the background concentration. For example, groundwater from MW 51U and MW 16, which contained background concentrations of6570 and 460 figIL, respectively, were spiked to give final con centrations of 12900 and 1000 fig/L of PFC8, respectively. Each sample also was spiked with 56.4 fig of PFC12. To determine the detection limit of the method, single samples ofgroundwater that contained no perfluorinated carboxylates above detection were spiked to give a range of final PFC8 concentrations from 18 to 54 uglL. Gas Chromatography/Mass Spectrometry. Extracts were analyzed using a Hewlett-Packard Model 5890 Series II Plus gas chromatograph (GC) equipped with a 30 m x 0.32 mm x 4.00 fim SPB-1 SULFUR column (Supelco Inc., Bellefonte, PA). An injection volume of 1 iiL was used under splitless conditions with an injector temperature of 200 C. The GC oven temperature was initially held for 6 min at 60 C, increased by 6 cC/min to 190 C, increased further by 30 "C/min to 270 C, and then held for 5 min. Quantification of perfluorocarboxylate methyl esters was performed using a Hewlett-Packard Model 5972 mass selec tive detector operated in electron impact (El) mode (70 eV). The mass selective detectpr was operated in full scan (50 450 amu) mode and in selected ion monitoring (SIM) mode using a dwell time of 100 ms for each ion. The scanning mode was used for qualitative identification while SIM mode was used for quantification. The ions ofm lz 131 [G.3F5]- ,m !z 169 [C3F7]"**, and m lz 219 [C4F9]+, which are characteristic fragments ofperfluorocarbons (16-18), were used to identify and quantify the methyl esters of perfluorohexanoic acid (PFC6), perfluoroheptanoic acid (PFC7), PFC8, andPFC12, The internal standard,: 2-chlorolepidine, was quantified with the ions at m lz 177 arid m lz 115. The identification of perfluorocarboxylate methyl esters was confirmed by electron capture negative ionization (ECN1) GC/MS, which gave unique molecular ions for each of the perfluorinated carboxylte methyl esters (e.g, m lz 328 for PFC6, m /z 378 for PFC7, m lz 428 for PFC8, and m lz 628 for PFC12). These measurements were performed with a Varan 3400 gas chromatograph interfaced with a Finnigan Model 4023 mass spectrometer. Methane was used as the reagent gas, and the mass spectrometer was operated in full scan mode (100-650 amu).The gas chromatograph was operated with a column and temperature program identical to that used for the El GC/MS. Initially, samples prepared in deionized water were used as the matrixfor constructing calibration curves, and standard recoveries were low. However, when samples prepared in tap water, which contains inorganic cations and anions, were used as the matrix for constructing calibration curves, quantitative recovery of standards was obtained. It is proposed that the 350 mL of deionzed water does not sufficiently rinse the disks of residual HC1 and tap water is required to completely rinse the disks andobtain quantitative recovery of standards. Therefore, calibration curves for quantification of PFC8 were constructed by passing 100 mL of tap water samples that had been spiked with 3.6-1080 fig of PFC8 through 25 mm SAXdisks and derivatizing the acids to their methyl esters using the in-vial elution and derivatlzation technique. The calibration curve for PFC12 was constructed in a similar manner by adding 7.5--113 fig of PFC12 standard to lOOmL of tap water. For all quantitation B ENVIRON. SCI. & TECHNOL. / VOL. xx, NO. xx, xxxx US00002579 1e+6 8e+5 - 0ocrc3oa 6e+5 - XI < 0>rxa: 4e+5 0 tr. 2e+5 PFC8 PFC12 (a) 2-chlorolepktine --r-- 10 -- i-- 15 20 Time (min) 25 ll , I* F 30 35 FIGURE 2. (a) Typical El GC/M S chromatogram of PFCS and PFC12 standards and (b) perfluorinated carboxylates, including PFCB, PFC7, PFC8, and PFC12 (spiked) in Naval Air Station Fallon groundwater. . standards, a total of 51.2 /g of the 2-chloroIepidine internal standard was added to the autosampiervlaljust prior to the ' addition ofmethyl iodide. Both calibration curveswere linear with r2 typically greater than 0.99. Quantification of PFC6 and PFC7 was performed assuming a response factor equal to an equimolar am ount of PFC8. Results and Discussion Gas Chromatography/M ass Spectrometry. A him thickness of4.m (30m x 0.32mmSPB-1 SULFUR; Supelco,Beilefonte, PA) was necessary to obtain sufficient retention times for the methyl esters ofPFC8 and PFC12 to allow for the separation and quantification (Figure 2a). Initial attem pts to separate and quantify the perfluorinated carboxylate methyl esters onathinfilm (0.25/un),30m x 0.25 mmDB-l (J&WScientific, Folsom, CA) column were unsuccessful regardless of the initial column temperature. Note that the stationary phases in the SPB-1SULFUR and DB-1 columns are comparable. A standard of perfluorobutyric acid was not observed under any of the described GC conditions; it is m ost likely that an initial oven temperature less th an 40 C would be required. The El massspectra of methyl PFC8 (Figure 3a) andPFCI2 indicate characteristic perfluoroearbon fragmentation (16, 171 in which the major ions (e.g., 69,119,169,219. etc.) differ by 50 amu, which corresponds to the m assof CFj. Molecular ions were not observed for any of the perfluorinated carboxylate methyl esters under El conditions; however, molecular ions (M]~ were observed under ECNI conditions. For example m iz 428 (in Figure 3b) corresponds to the molecular ion of methyl PFC8. VOL. xx. NO. xx. XXIX I ENVIRON. SCI. 3, TECHNOL mC US00002580 FIGURE 3. (a) El m ass spectrum of methyl PFC8. (b) ECNI m ass spectrum of methyl PFC8. Solid-Phase Extraction.Prior to developing asolid-phase extraction method, initial experiments were conducted using diazomethane as the derivatization reagent. When perflu orinated carboxylates were derivatized using ethanol-based dlazometharie, multiple peaks corresponding to methyl and ethyl esters were detected (unpublished data). Because El GC/MS did not produce molecular ions, ECNI GC/MS was used to verify the formation ofboth methyl and ethyl esters. Consequently, if ethanol-based diazom ethane was used for derivatization in conjunction with El GC/MS. multiple peaks in a chromatogram could be erroneously interpreted as a greater number of perfluorinated compounds than are actually present. In contrast, only the methyl ester was obtained when butyl carbitol (2-(2-butoxyethoxy)ethanol) was used to prepare the diazomethane reagent However, because of the hazards associated with the use of diazo D ENVIRON. SCI. & TECHNOL. I VOL. XX. NO. xx. xxxx methane and the time-consuming nature of diazomethane derivatization, an alternative method was desired. Derivatization ofthe perfluoroearboxylates by solid-phase extraction and th e in-vial elution and derivatization technique gave only a single peak that corresponded to the methyl ester ofeach perfluorinated carboxylate standard; the identification of each methyl ester was confirmed by ECNI GC/MS. In addition, the solid-phase extraction approach combined the steps of isolation an d derivatization, which greatly simplified the procedure and eiiminated the use of diazomethane. Six replicate analyses of blank 25 m m SAX disks that had n o t been prerinsed with 12 mM HCl/acetanitriie prior to tise^ yielded a n average o f 21 1fig (4.8% relative standard deviation (RSD)) of PFC8 per disk. No other perfluorinated carboxylates were present in the disks above the detection limit. The PFC8 is associated w ith the Teflon matrix and not X US00002581 TABLE 1.Recover; oi PFCS andPFC12Spiked into Groundwater Samples fromNaval Air Station Fallon' sample . % recovery PFC8 PFC12 N A S Fallon M W 51U N A S Fallon M W 16 N A S Fallon M W 50U N A S Fallon M W 17 83e 90d 73 74 35 85 77 88 * Duplicate samples w ere analyzed. Sample volume was 100 mL unless otherwise noted. # Sample volume was 55 mL ` Calculated as the final measured concentration divided by backgroundconcentration plus spike concentration and multiplied by 100. The background concentration was 6,570 ug/L. "Calculated as the final measured concentration divided by background concentration plus spike con centration and multiplied by 100. The background concentration was 460,glL. the embedded anion exchange particles (unpublished data). The background PFC8 was successfully removed by rinsing the disks prior to use with 12 mM'HCl/acetonitrile followed by350 mL ofdeionized water. It should be noted that benzoic acid and ethylhexylphthalic arid are also present in the disks as artifacts and are removed by the HCL/acetonitrile prerinse step. Accuracy, Precision, and Detection Limits.Therecoveries ofPFC8 from blank groundwatersamples obtained fromNAS Fallon wells MW SOU and MW 17 were 73 and 74%, respectively, while the recoveries ofPFC12 were 77 and 88%, respectively (Table 1). Because detectable levels of PFC8 occurred in groundwater from MW51U and MW 16,standard addition experiments were performed to determine the recoveries of PFC8. The recoveries of the PFC8 spiked into MW 51U and MW 16 groundwater to give a final concentra tion double that of the background concentration were 83 and 90%, respectively (Table 1). The recoveries of PFC12 from MW 5IU and MW 16 groundwater, which did not contain background concentrations of PFC12, were 35 and 85%, respectively (Table 1). Although the recovery o f PFCS (83%) diffets significantly from that of PFC12 (35%) in groundwaterfrom MW 51U, the recoveriesofPFC8 and PFC 12 were nearly equivalent for the other groundwater samples. Monitoring well 51U is located closest to the fire pit where AFFF agents where applied to burning mixtures of fuels and solvents. Due to its proximity to the fire pit, the groundwater from MW 51U most likely contains the greatest diversity of inorganic and organic constituents, which may adversely affect PFC12 recoveries relative to that of PFC8, Therefore, although the original intent was to use the PFC12 as a surrogate standard because it did not occur in the ground water samples, PFC12 appears more sensitive to matrix interferences compared to PFC8 so that it is an inappropriate choice for a surrogate standard. For this reason, all subse quent quantification was based on the 2-chlorolepidine internal standard. The precision, indicated by the RSD, calculated from five replicate analyses each ofgroundwaterfrom NASFallon MW 16 and Tyndall AFBT11-2 ranged from 3.7 to 14% (Table 2). The detection and quantitation limit of the method was defined as those concentrations of PFCS needed to produce a signal-to-noise (S/N) of 3:1 and 10:1, respectively. The detection and quantitation limits for PFCS were 18 and 36 /zg/L, respectively. Application to GroundwaterSamples. Four groundwater samples from both NASFallonand TyndallAFBwere analyzed for perfluorinated carboxylates. Chromatograms obtained by 0 GC/MS indicated the presence of multiple perfhiorinated compounds ail hairing characteristic periluorocarbon fragmentation (Figure 2b). Analysis by ECNI GC/MS estab lished the identification ofPFC6, PFC7 and PFC8 in ground water obtained from wells MW 51U and MW 16 from NAS Fallon. The molecular ions [Ml- for methyl PFC6 (m lz 328) and methyl PFC7 (m /z378) were observed for peaks eluting 4.7 and 2,3 m in before that ofPFC8 (Figure 4a,b). The ECNI mass spectrum for methyl PFCS in MW 51U was similar to that of the PFC8 standard (Figure 2b). The groundwater samples from NAS Fallon MW 51Uand MW 16 had total perfluorinated carboxylate concentrations of 7090 and 540 /rg/L, respectively (Table 2). The PFC6 detected in NAS Fallon groundwater samples from MW 51U and MW 16 comprised 5.2% and 11%, respectively, of the total perfluorocarboxylates detected.The PFC7 was 2.1%and 3.3% respectively, of die total perfluorinated carboxylates detected in these wells. The dominant perfluorinated car boxylate, PFCS, accounted for 93% and 85%, respectively, of the total perfhiorocarbdxylate concentration. The highest concentrations ofperfluorocarboxylates were observed in groundwater collected from NAS Fallon MW51U, which is the well located closestto the fire-training pit (Figure la). Monitoring well 16, which is located downgradient of MW 51U and the fire-training p it had lower but detectable concentrations o f perfluorocarboxylates. Groundwaterfrom MW SOUand MW 17, which are located off gradient from the fire-training pit, contained no detectable perfluorinated carboxylates. Over the approximate 100 m distance between MW51UandMW l6, the concentrations ofdie perfluorinated carboxylates decreased with increasing num ber of carbons. For example, the concentration of PFC6 decreased 85%over the 100 m compared to decreases of 88% and 93% for PFC7 and PFCS, respectively. The groundwater samples from Tyndall AFBPW-10, PW07, and T ll-2 contained total perfluorinated carboxylate concentrations o f298,159, and 124tig/L, respectively (Table 2). The compositions of Tyndall AFB groundwater collected from the three wells ranged from 46 to 52% for PFC6, from 13 to 15% for PFC7 an d from 34 to 40% for PFC8. In contrast to the groundwater samples from NAS Fallon, the dominant perfluorinated carboxylate in Tyndall AFB groundwater was PFC6. TABLE2.Concentrations of Perfluorinated Carboxylates in Groundwater Samples fromNaval Air Station FallonandTyndall Air Force Base'* sample a PFC6 (ug/L) PFC7 (ugA) PiC&IjiglU total (pg/l) N A S Fallon M W 51U 3 372 4 (1.1%) N A S Fallon M W 16 5 57 8 (14%) N A S Fallon M W 50U 3 nd N A S Fallon M W 17 3 nd T yndall A FB PW -10 2 144 Tyndall A FB PW -07 2 73 Tyndall A FB TT1-2 5 64 4 (6.3%) Tyndall A FB TY22FTA 2 nd 149 5 (3.4%) 18 2 (m o )1 nd nd 38 22? 19 1 (5.3%)e nd 6570 -150 (2.3%) 460 2 0 (4.3%) nd nd 116 64 42 2 (4.8%) nd 7090 160 (2.3%) 540 20 (3,7%) nd nd 298 159 124 8 (6.5%) nd *The relative standard deviation is given in parentheses: "nd: not detected above the detection Iimit. cThe reported value is near the detection limit (SIN 3) and less than the quantitation limit (S'A/ < 10). .The value has been included in the reported total concentration. VOL. XX, NO. XX, xxxx I ENVIRON. SCI. a TECHNOL. E US00002582 25000 20000 - 250 (a) > 15000 - '5 O c 10000 - 278 5000 - 308 [M I' 328 0 150 200 250 300 350 400 450 m/z FIGURE 4. ECN1 m ass spectra of (a) methyl PFCE and (b) methyl PFC7. The highest concentrations of perfluorocarboxylates among the groundwater samples from Tyndall AFB were observed in PW-10 and PW-07.which are the two wells located closest to th e fire-training pit (Figure lb). Monitoring well T11-2, which is located downgradient ofthe fire-training pit, had lower but detectable groundwater concentrations of perfluorocarboxylates- The groundwater collected from a well located north of the fire-training pit, TY22FTA, contained no perfiuorinated carboxylates above the detection limit (18/ig/L). It is not surprising to observe a suite of perfiuorinated carboxylates since the raw materials used in the synthesis of perfiuorinated organic compounds are mixtures (3, 13}. Different ratios of PFC6, PFC7, and PFC8 may result from the use of different AFFF formulations at the two fire-training areas. The observed homologous series consisting of even and odd num ber perfiuorinated carboxylates is indicative of the electrochemical fluorination process used by 3M Co. (3). Otherfiuorination processes, suchastelomerization, produce only even num ber homologues (3). Because o fthe proprietary nature ofAFFFs, it is not known ifperfiuorinated carboxylates are present as one ofthe major surface active agents in AFFF formulations or as unreacted starting materials used in the synthesis of the principal perfiuorinated surfactants used in AFFF formulations. In addition, the carboxylates may be combustion, biological, or nonbiologicai degradation prod ucts of the principal perfiuorinated components in AFFF mixtures. Unfortunately, the exact source and history ofAFFF applications at the two fieldsites are unknown, andtherefore, the relationship between the observed perfiuorocarboxyiate ratios and that of the original AFFF mixtures is unknown. To the best ofour knowledge, very little is known regarding the transport and fate of perfluorocarboxylates in ground water. Adsorption to sludge at wastewatertreatment facilities F ENVIRON. SCI. & TECHNOL. 1 VOL. * * . NO. xx. xxxx X US00002583 is considered a significant process for the removal of perfluorinated surfactants during treatment (3). However, detection of perfluorinated carboxylates at the NAS Fallon and Tyndall AFB sites, which have not been used since 1988 and 1992, respectively, is consistent with the view that biodegradation of the long chain perfiuorocarbon hydro phobe is unlikely (6, 9, 19}. The recalcitrant nature of perfluorinated compounds is attributed in p an to the nu d ity of the perfiuorocarbon chain (9, 20} as well as the strength of the carbon-fluorine bond (3, 9, 21). To the best of our knowledge this is the first definitive identification of perfluorinated carboxylates in groundwater impacted by fire-fighting activity. Further work is needed to determine if additional perfluorinated components are present, such as perfluorooctane sulfonic acid, which is thought to be one of the principle components in some commercial AFFF formulations. In addition, it is of interest to relate the occurrence and distribution of perfluorinated compounds to other site characterization parameters such as dissolved organic carbon, inorganic constituents, and the distribution of co-contaminants and to understand the potential influence of perfluorinated compounds on the biotransformation and transport of other co-contaminants. Acknowledgments The authors thank Donald Hagen and Eric Reiner o f 3M Co. and Mitch Hubert of nsul Inc. for valuable discussions and technical assistance. We would also like to thank ChangJho and Erik Kissaforreviewing draft manuscripts. Ron Hoeppel and Art Fisher from NAS Fallon, Bill Johnson from the University of Utah, and Erica Becvar from Tyndall AFB are gratefully acknowledged for facilitating sample collection. Supelco. Inc. is gratefully acknowledged for the donation of a GC column and vacuum manifolds. This work was financially supported by the Oregon State Department of Chemistry (N. L. Tartar Research Fellowship) and by a grant from the Environmental Protection Agency (OER R821195). The authors acknowledge the Oregon State University Environmental Health Sdenca Mass Spectrometry Core Facility for its support through NIEHS Grant ES00210. Literature Cited (1) O'Brien, A. F. M.S. Thesis, University o f Mayland College Park, Maryland, 1994. (2) Henley, M.: Mayfield. R : Shelley. T. Abstracts of Papers, Pittsburgh Conference, Atlanta, GA, American Chemical Soci ety: Washington, DC, 1997: Abstract 519. (3) Kissa, E. Fluorinated Surfactants: Synthesis, Properties, and Applications-, Marcel Dekken Hew York, 1994. (4) Bass.C, M. Ph.D. Dissertation, Universityo fOklahoma, Norman, OK. 1982. ' (5) Howell, R. D.; Tucket. E. . Am. Environ. Lab. 1996, 12,10. (6) Darwin. R.L;Ottman,R.E.;Norman,E.C.;Gott,J.E.;Hanauska, C. P. Nad. Fire Protect Assoc. 1995,67. (7) Chan. D. B.; Chian, E. S. K. Environ. Prog. 1986, 5, 104. (8) Remde, A4 Debus, K. Chemospbere 1 9 96,52,1563. (9) Key, Bi Howell, R. D.; Criddie, C. S. Environ. Sci. Technol. 1997, 31,2445. (10) Godin. C. S.; Myhr, B. C.; Lawksr, T. E.; Young, 8. R-: Murli, R ; Clfone, M. A. Genotoricily Assessment o f Perfluorodecanoic Acid usingaBattery o f in vitroand in vivo/ln vitro Assays; Harry G. Armstrong Aerospace Medical Research Laboratory: Kens ington, MD, December 1990. (11) Deocampo. N. D.; Upturn, B. L; Trosfco, J, E. Ftindam. Appl. Toxicol. (Suppl) 1996,30, 208. (12) WorkPlan forAssessingthe Feasibility o f IntrinsicRemediation at installation Restoration Program Sites; Oak Ridge National Laboratory: Oak Ridge, TN, October 1996, (13) Contamination Assessment Repent Active Fire-Training Area FT-23 Tyndall Air Force Base Panama City, FL, FDER Facility No. United States Army Corps o f Engineers: Washington, DC, February 1994. (14) Belisle, J.; Hagen, D. F. Anal Biochem. 1 9 8 9 ,101,369. (15) Held, J. A.; Seed, R. L. Environ. ScL Technol 1996, 30, 3544. (16) McLafferty, F. W.: Turecek, F. Interpretation ofMass Spectra, 4th ed.: University Science Books: Sausaiito, CA, 1993. (17) Lyon, P. A.; Tomer, K. B.; Gross, M. L. Anal Chem. 1985, 57, 2984. (18) Hudiicky, M.; Pavlath, A. E. Chemistry of Organic Fluorine CompoundsH: A CriticalReview. American Chemical Society: Washington, DC, 1995. - {19} Key, B. D. Ph.D. Dissertation, Michigan State University. East Lansing. MI, 1996. (20) Asakawa, T,: Mouri, M.; Miyagishi, S.; Nishida, M. Langmuir 1989, 5, 343. (21) Smart, B. E. In MolecularStructure and Energetics; Uebman, J. F., Greenberg, A., Eds.: VCH Publishers: Deerfield Beach, FL, 1986; VoL 3. Received for review December 29, 1998. Revised manuscript received M ay 12,1999. Accepted M ay 21, 1999, ES981355+- PAGE EST: 6.6 YOU xx. NO; xx. xxxx i ENVIRON. SCI. & TECHNOL. O