Removal of PFOS, PFOA and other perfluoroalkyl acids at water reclamation plants in South East Queensland Australia

Chemosphere 82 (2011) 9–17

Contents lists available at ScienceDirect

Chemosphere

journal homepage: www.elsevier .com/locate /chemosphere

Removal of PFOS, PFOA and other perfluoroalkyl acids at water reclamation plantsin South East Queensland Australia

Jack Thompson a,⇑, Geoff Eaglesham b, Julien Reungoat c, Yvan Poussade d,e, Michael Bartkow f,Michael Lawrence c, Jochen F. Mueller a

a The University of Queensland, National Research Center for Environmental Toxicology (Entox): 39 Kessels Rd., Coopers Plains, QLD. 4108, Australiab Queensland Health Forensic and Scientific Services (QHFSS), Special Services: 39 Kessels Rd., Coopers Plains, QLD. 4108, Australiac The University of Queensland, Advanced Water Management Center (AWMC), QLD. 4072, Australiad Veolia Water Australia, Level 1, 20 Wharf Street, Brisbane, QLD. 4000, Australiae WaterSecure, Level 2, 95 North Quay, Brisbane, QLD. 4000, Australiaf SEQwater, 240 Margaret Street, Brisbane, QLD. 4000, Australia

a r t i c l e i n f o

Article history:Received 18 June 2010Received in revised form 22 September2010Accepted 9 October 2010Available online 3 November 2010

Keywords:Perfluorinated compoundsTertiary treatmentPFOSPFOAReverse osmosisOzonation

0045-6535/$ - see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.chemosphere.2010.10.040

⇑ Corresponding author. Tel.: +61 7 3274 9060.E-mail address: [email protected] (J. Th

a b s t r a c t

This paper examines the fate of perfluorinated sulfonates (PFSAs) and carboxylic acids (PFCAs) in twowater reclamation plants in Australia. Both facilities take treated water directly from WWTPs and treatit further to produce high quality recycled water. The first plant utilizes adsorption and filtration methodsalongside ozonation, whilst the second uses membrane processes and advanced oxidation to producepurified recycled water. At both facilities perfluorooctane sulfonate (PFOS), perfluorohexane sulfonate(PFHxS), perfluorohexanoic acid (PFHxA) and perfluorooctanoic acid (PFOA) were the most frequentlydetected PFCs. Concentrations of PFOS and PFOA in influent (WWTP effluent) ranged up to 3.7 and16 ng L�1 respectively, and were reduced to 0.7 and 12 ng L�1 in the finished water of the ozonation plant.Throughout this facility, concentrations of most of the detected perfluoroalkyl compounds (PFCs)remained relatively unchanged with each successive treatment step. PFOS was an exception to this, withsome removal following coagulation and dissolved air flotation/sand filtration (DAFF). At the secondplant, influent concentrations of PFOS and PFOA ranged up to 39 and 29 ng L�1. All PFCs present wereremoved from the finished water by reverse osmosis (RO) to concentrations below detection and report-ing limits (0.4–1.5 ng L�1). At both plants the observed concentrations were in the low parts per trillionrange, well below provisional health based drinking water guidelines suggested for PFOS and PFOA.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

1.1. Perfluorinated alkyl compounds

Perfluorinated alkyl compounds (PFCs) have received increasingattention in recent years as environmental contaminants due totheir consistent detection in various environmental matrices(Giesy and Kannan, 2002), and their adverse effects in animal tox-icity studies (Kennedy et al., 2004; Lau et al., 2007). The two groupsgiven the most scrutiny to date have been the perfluoroalkyl sulfo-nates (PFSAs) (CnF2n+1SO�3 ) and the perfluorocarboxylic acids(PFCAs) (CnF2n+1COOH), and in particular the eight carbon mem-bers of these groups; perfluorooctane sulfonate (PFOS) and perflu-orooctanoic acid (PFOA). These compounds have been producedcommercially since the 1950s and used in a variety of consumerand industrial applications, including oil and water repellent sur-

ll rights reserved.

ompson).

face coatings for packaging and textiles, surfactants, and aqueousfire-fighting foams (Prevedouros et al., 2006; Paul et al., 2009).

In 2009 PFOS was added to the Stockholm Convention forPersistent Organic Pollutants (Stockholm Convention, 2010).Production of PFOS and similar perfluorooctyl products was phasedout in the USA and Europe 2000–2002 (OECD, 2002), howeverongoing production continues elsewhere (Wang et al., 2009).Perfluorooctanoic acid and its salts continue to be used as process-ing agents in the manufacture of fluoropolymers but efforts havebeen made in conjunction with eight major PFOA manufacturersto reduce emissions from fluoropolymer manufacturing facilitiesby 95% of 2000 levels by 2010. A complete phase-out is soughtby 2015 (USEPA, 2009).

In Australia there is no record of PFCA or PFSA manufacture andimportation and use has been discouraged by the National Indus-trial Chemical Notification and Assessment Scheme (NICNAS,2007, 2008). Despite this, there is potentially still a large stockpileof PFC containing products in inventories and personal ownershipin Australia (e.g. PFC treated carpets). Additionally a large range of

10 J. Thompson et al. / Chemosphere 82 (2011) 9–17

products may contain PFCAs and PFSAs as residual impurities fromtheir manufacture (Washington et al., 2009), others still may con-tain compounds which may degrade to form these in the environ-ment or metabolically (Vestergren et al., 2008). Pooled serum datafrom the general Australian population showed mean concentra-tions of PFOS and PFOA at 15 and 6.4 ng mL�1 respectively (Tomset al., 2009), comparable to those observed internationally (e.g.Calafat et al., 2007). Pharmacokinetic modeling of these concentra-tions has suggested that Australians are exposed to approximately100 ± 37 and 54 ± 15 ng d�1 PFOS and PFOA respectively (Thomp-son et al., 2010), with currently no data describing the specificpathways responsible for these estimated intakes.

Studies of PFCs in primary and secondary wastewater treatmenthave shown them to be inefficiently removed (Sinclair and Kannan,2006; Loganathan et al., 2007), and in some cases increasing ineffluent relative to influent (Schultz et al., 2006; Sinclair and Kan-nan, 2006). Consequently, WWTP discharges have been shown tobe point sources of these compounds to the aquatic environment.The limited efficiency of secondary water treatment technologieshas meant that PFCs have been detected in finished tap water fromvarious countries at parts per trillion ranges, often at concentra-tions similar to those in the source waters (Mak et al., 2009; Qui-nones and Snyder, 2009). Some success has been reportedutilizing activated carbon for removal (e.g. Ochoa-Herrera and Sier-ra-Alvarez, 2008), but there are factors such as contact time andadsorption capacity which need to be considered in adopting thismethod. In studies on more advanced water treatment processes,a number of novel destruction techniques have been shown to besuccessful, such as sonochemical degradation (Cheng et al., 2008)and oxidation with persulfate radicals (Hori et al., 2005). Labora-tory studies of rejection across nanofiltration and reverse osmosismembranes have demonstrated up to 99.9% removal of most PFCsstudied (Tang et al., 2006; Steinle-Darling and Reinhard, 2008).Although there is information available on the response of PFCsto specific treatment processes, relatively little has been publishedon the fate of PFCs in actual full scale reclamation plants underoperational conditions.

In South East Queensland, a generally dry climate and increas-ing demands on water resources has prompted the constructionof several advanced water treatment facilities. The aim of thesefacilities is to reclaim treated wastewater by further treatment,

Fig. 1. Schematic of Plant A, with sampling points marked S1–S7. Note: SRT = slu

producing purified water of a high quality standard for use inindustrial and commercial processes. This helps mitigate the envi-ronmental impacts associated with discharge of WWTP effluent,and also provides potential resources to supplement potable sup-plies if necessary in the future. This study aims to assess the effec-tiveness of two water reclamation plants using different types ofadvanced treatment approaches with regards to PFC removal un-der normal operating conditions. The PFCs studied were the C4–C18 PFCAs, and the C4, C6, C8 and C10 PFSAs.

1.2. Water reclamation plants

The two plants studied use contrasting methods, but share theaim of producing high quality purified recycled water. Schematicsof both facilities, detailing the sequence of treatment steps are pro-vided as Fig. 1 (Plant A) and Fig. 2 (Plant B). Plant A takes secondaryeffluent from a single WWTP, designed to service up to 40 000 EP(equivalent persons), as its influent. It produces up to 8 ML d�1 ofrecycled water for industry consumers and for public irrigationand non-potable household usage (toilets, irrigation) in a dualreticulation system. It is not used to supplement potable suppliesat present, but was designed to meet potable standards. Treatmentconsists of de-nitrification, several stages of ozonation, coagula-tion/flocculation, dissolved air flotation and sand filtration (DAFF)and biologically activated carbon filtration (see Fig. 1).

At the time of sampling Plant B received effluent from 4WWTPs, each taking predominantly residential sewage and de-signed to serve 30 000–185 000 EP. The plant has a 66 ML d�1 pro-duction capacity, and at sampling was producing approximately40 ML d�1 purified water. Treatment at Plant B consists of coagula-tion/flocculation and sedimentation, then ultra-filtration (UF), re-verse osmosis (RO), advanced oxidation (H2O2 + UV) and finalstabilization and disinfection (see Fig. 2). The finished reclaimedwater is piped to industry users including two power stations,which use it as process water. Plant B is also designed to providewater to a nearby dam for indirect potable reuse, if dam levels dropbelow 40%. The reverse osmosis concentrate (ROC) or brine is fur-ther treated to remove nutrients before being discharged into anearby river. Generally at Plant B contaminants removed by ROare concentrated almost 7-fold, with typically 85% of water perme-ating the membrane and 15% going to ROC.

dge residence time, HRT = hydraulic residence time, EP = equivalent persons.

Fig. 2. Schematic of Plant B, with sampling points marked SI–SV.

J. Thompson et al. / Chemosphere 82 (2011) 9–17 11

2. Materials and methods

2.1. Sampling

2.1.1. Plant ATen liters, 24 h composite samples were taken from seven points

in the treatment train (Fig. 1) on two separate sampling occasions.On each occasion and at each of the sampling points, 1 L sub-sam-ples were taken from the total 10 L. The 10 L composite sampleswere collected in a large glass vessel pre-rinsed with methanol, over24 h using a 10 channel continuous flow peristaltic pump (flow7 mL min�1). The sampling vessel was protected from light and keptrefrigerated during the 24 h sampling. One liter sub-samples weretaken from each 10 L sample and stored in pre-rinsed (methanol,distilled water) high density polyethylene (HDPE) bottles. Thesewere further split into two 250 mL replicates in the laboratory toprovide a measure of the methodological variability. Samples werecollected on two occasions, in October and November 2009.

2.1.2. Plant BPre-rinsed (methanol) HDPE bottles were used to collect 1 L

grab samples on three days within a single week of November2008 (Monday 24th and Wednesday 26th in the morning�10:00 am, Friday 28th in the afternoon �2:00 pm), at samplingpoints marked in Fig. 2 (SI–SVI). The 1 L samples were again splitinto 250 mL replicates in the laboratory; with three replicates ana-lysed for these samples. Due to logistical constraints collection ofcomposite samples was not feasible at this facility. The grab sam-pling approach on multiple days was used to estimate the extent oftemporal variation and therefore the usefulness of future grabsampling. The differing times of collection between the three daysmeans samples were taken at times of both low (day 3 afternooncollection) and high flow (days 1 and 2 morning collection).

2.2. Sample extraction and analysis

The 250 mL samples were extracted following the methods ofTaniyasu et al. (2005). In brief, weak anion exchange solid phase

extraction cartridges (OASIS WAX-SPE 5 mg, 6 cc) (Waters, MilfordMA) were preconditioned with 4 mL of 0.1% NH3 in methanol, 4 mLof methanol, and 4 mL of milli-Q water. Samples were loaded ontothe cartridges under vacuum and 4 mL of a 25 mM ammoniumacetate buffer (pH 4) was added to clean cartridges. Cartridgeswere dried under light vacuum and eluted with 4 mL of 0.1% NH3

in methanol. Extracts were concentrated under high purity N2 toa volume of 0.5 mL, and another 0.5 mL of milli-Q water was addedto give a final volume of 1 mL. Some influent samples containedvisible amounts of suspended particulates and were filtered withglass fiber filters. A comparison of results from filtered and unfil-tered secondary effluent samples from a separate WWTP are pro-vided in the supporting information Table S5. While there weresome significant differences in the filtered and unfiltered samples,particularly for PFOS, PFDS and longer chain carboxylic acids, itwas deemed appropriate not to filter all samples as this wouldbe unnecessary given the quality of the water within the reclama-tion plants.

2.3. Analysis and quantification

Analysis was performed using a QTRAP 4000 MS/MS (AB/Sciex,Concord, Ontario, Canada) coupled with a Shimadzu prominenceHPLC system (Shimadzu, Kyoto Japan) using a gradient flow of mo-bile phase of methanol/water with 5 mM ammonium acetate. AGemini C18 column (50 mm � 2 mm i.d. 3 lm 110 Å) (Phenome-nex, Torrance, CA) was used for separation, and an additional col-umn (Altima, C18, 150 mm � 2 mm i.d. 5 lm, 100 Å)(GraceDavison, Deerfield, IL) was installed between the solvent reservoirsand sample injector to separate peaks consistently present in thesystem from those in the samples (e.g. small peaks for PFDoDA(C12 PFCA), and for PFOA present in the mobile phase, and/or fromfluoropolymer components in the LC system). Data was acquired inthe scheduled multiple reaction monitoring mode (SMRM), moni-toring two transitions for most analytes. More details of the instru-mental analysis can be found in the supporting information.

Quantification was done using relative response factors of masslabeled internal standards, via internal standard quantification

12 J. Thompson et al. / Chemosphere 82 (2011) 9–17

methods created in the MS software Analyst (v1.5 AB Sciex). Isoto-pic mass labeled standards were available for a majority of analytes(Wellington Laboratories, Ontario Canada), and the losses of theseduring extraction should reflect those of the native compounds thusaccounting for recoveries in the calculated concentrations. Detailsof the mass labeled standards used, the fortification and quantifica-tion procedures are provided in the supporting information. Detec-tion limits were set at the average blank concentration plus threetimes the standard deviation, or by the average concentration fromrepeated injections (seven injections) of the lowest quantifiablestandard (0.1–0.5 ng mL�1) plus three times the standard deviation.Taking into account sample concentration factors instrumentaldetection limits ranged from 0.2–0.7 ng L�1 and reporting limitswere set at double this, ranging from 0.4–1.5 ng L�1.

2.4. QA/QC

Four blank samples of milli-Q water (250 mL each) were ex-tracted with each batch of samples from Plant A (eight in total),and 11 blanks in total with the Plant B samples. Two SPE cartridgeblanks were obtained by conditioning and eluting cartridges with-out loading a sample. With each instrumental run, the calibrationseries were injected at the start and end of the run to check forinstrumental drift and changes to sensitivity. To provide an appro-priate field blank for Plant A, milli-Q water was pumped throughthe same pump at the same rate and for the same duration as usedin sample collection. NaCl was added to the milli-Q water prior topumping to give a similar conductivity as the sampled water(�780 lS). The same water was analysed both after pumping andwithout pumping to isolate the source of any contamination.

No PFCs were detected in the SPE blanks. During analysis ofPlant A samples, PFOA was detected in one of the eight laboratoryblanks at 0.2 ng L�1. No PFCs were detected in the field blank,either with or without pumping. In the 11 blanks analysed withPlant B samples, very low concentrations of several PFCs were de-tected on at least one occasion, the most frequently detected beingPFOA at concentrations up to 0.5 ng L�1.

Analytical reproducibility was demonstrated by the agreementseen between sample replicates, with% RSD typically below 20%for analytes above reporting limits. Recoveries of internal standardswere generally good, being between 50% and 100% for the mostpart, and recoveries of two native PFC spikes were similar. These re-sults are given in the supporting information (Tables S2–S4).

3. Results and discussion

3.1. Influent concentrations

Influents to both plants come from the treated effluents of pri-mary and secondary WWTPs. As WWTP effluent has been shown

Table 1Concentrations of PFCs (ng L�1) in influent/secondary treated water.

Plant A Plant B

Analyte October 2009 November 2009 Day 1, 24th November 2008 (am

PFBS nq nq 6.4 ± 0.6PFHxS 2.3, 2.1 1.5 ± 0.1 36 ± 2.7PFOS 5.0, 4.5 2.2 ± 0.02 38 ± 1.3PFPeA 5.7, 5.5 4 ± 0.8 8.3 ± 1.3PFHxA 4.4, 4.4 6.3 ± 0.3 13 ± 1.4PFHpA 1.5, 1.4 1.2 ± 0.1 6.2 ± 0.3PFOA 16, 13.6 6.7 ± 0.9 22 ± 3.2PFNA 1.1, 1.1 1.2 ± 0.2 3.3 ± 0.7PFDA 1.2, 1.0 1.1 ± 0.3 1.4 ± 0.4

Note: Plant B average of 250 mL subsample replicates (n = 3) ± standard deviation ofinterference present see text for details.

internationally to be a source of PFCs to the aquatic environment(e.g. Sinclair et al., 2006; Huset et al., 2008), this data provides aninsight into the concentrations of PFCs emitted to Australian aqua-tic environments via this route. Concentrations varied between thetwo plants with 3–10 times higher concentrations of PFCs in theWWTP effluent samples entering Plant B compared to samples ofeffluent water that enters Plant A. Table 1 displays influent (second-ary treated WWTP effluent) concentrations for both plants. At PlantA PFOA was consistently detected at the highest concentrations of6.7 ± 0.9 and 15 ± 1.7 ng L�1 depending on sampling period. Asidefrom PFOA, the majority of PFCs were detected at low concentra-tions around 1 ng L�1, with other exceptions being PFOS (up to3.7 ng L�1) and PFHxA and PFPeA, with maximum concentrationsof 6.7 and 9.6 ng L�1 respectively. At Plant B PFOA and PFOS wereagain the most consistently detected analytes, but in contrast toPlant A PFOS was found in the highest concentrations (20–40ng L�1). Second to PFOS in concentration, PFOA was found at PlantB at 14–29 ng L�1. Other PFCs with relatively high concentrationsin Plant B influent were PFHxA and PFHxS, at concentrations upto 14 and 38 ng L�1 respectively. The influent concentrations aremost likely underestimated at Plant B due to the filtration of somesamples without the corresponding extraction of the filtrate.

At Plant A, PFOA concentrations were approximately threetimes higher than PFOS. ln contrast Plant B had higher PFOS con-centrations than PFOA, and PFHxS concentrations similar to PFOA.In Plant A concentrations of PFOA were approximately 5–7 timeshigher than PFHxS. The variation in PFC profile between the twoplants suggests the possibility of a point source for PFOS and otherPFSAs in the catchments serviced by the WWTPs feeding Plant B.The observed profiles were consistent between sample collectionperiods, despite overall concentrations varying at both sites withsampling time. At Plant B, on day 3, sampling in the afternoonwould have provided a picture of concentrations under low flowconditions, as opposed to the morning sampling on days 1 and 2.Concentrations of all the detected PFCs were lower on this day thanthe previous two, with the exception of PFDA which was slightlyhigher relative to the day 1 sample. For Plant A the variations be-tween sampling occasion were lower than those in Plant B, perhapsdue to the more time integrative nature of the sample collection.The highest variations were seen for PFOA and PFOS, which dif-fered by almost a factor 2 between collection periods. For samplesfrom both plants the consistency between laboratory replicateswas good and less than the inter-day variations observed.

Longer chain PFCs, greater than 10 carbons were not detected inthe influent of either plant. The absence of these compounds is ex-pected given the typically lower concentrations of these encoun-tered in aqueous samples, and the increased likelihood of thelonger chain PFCs partitioning to sludge in the feeder WWTPs (Hig-gins et al., 2005). Additionally, it has been shown that these com-pounds may adsorb to labware during extraction (Taniyasu et al.,

) Day 2, 25th November 2008 (am) Day 3, 28th November 2008 (pm)

4.8 ± 0.4 2.4 ± 0.428 ± 2.1 12 ± 1.738.6 ± 1 23 ± 2.411 ± 2.4 5.8 ± 0.514 ± 0.2 11 ± 0.47.6 ± 0.3 4.6 ± 0.227 ± 1.4 15 ± 1.63.6 ± 0.3 1.8 ± 0.33.1 ± 0.4 2.6 ± 0.3

replicates. Plant A 250 mL duplicates separated by comma. nq = not quantified,

J. Thompson et al. / Chemosphere 82 (2011) 9–17 13

2005). This is evidenced in this study by the poorer recoveries ofthe labeled PFDoDA (C12) standard (average recovery around50%). Despite appearing in samples from later in the treatmenttrain of Plant A, PFBS could not be quantified in influent samples,nor in any samples pre-ozonation due to a broad interfering peakin the quantifying MRM transition (299 > 80). However peaks cor-responding in retention time to PFBS could be observed in its qual-ifying MRM transition (299 > 99) suggesting the presence of thiscompound throughout the treatment process rather than appear-ing post ozonation as Table 3 suggests.

The reclamation plant influent data presented here can be com-pared with multitude of data published on WWTP effluents fromaround the world. Secondary effluents from an urban area in Geor-gia (USA) had reported concentrations of up to 227 ng L�1 PFOA,and 22 ng L�1 PFOS (Loganathan et al., 2007). Also from USA, efflu-ents of 6 WWTPs in New York state had PFOS and PFOA concentra-

Table 3Concentrations of detected PFCs (ng L�1) throughout Plant B.

PFHxA PFHpA PFOA PFNA

Day 1 (morning sampling)SI 13 ± 1.4 6.2 ± 0.3 22 ± 3.2 3.3 ± 0.7SII 23 ± 0.4 12 ± 0.8 43 ± 6.8 6.1 ± 1SIII 14 ± 0.5 7.7 ± 1.1 21 ± 0.8 3 ± 0.6SIV Nd <LOR (0.3) <LOR (0.6) NdSIV a,b 100 ± 13.8 56 ± 8.6 162 ± 24.7 18 ± 1.2SV Nd <LOR (0.3) <LOR (0.5) NdSVI Nd <LOR (0.4) <LOR (0.7) <LOR (0.4

Day 2 (morning sampling)SI 14 ± 0.2 7.6 ± 0.3 27 ± 1.4 3.6 ± 0.3SII 15 ± 2.1 7.8 ± 1.2 32 ± 1.8 3.5 ± 0.2SIII 15 ± 1.4 8 ± 0.3 29 ± 1.5 3.4 ± 0.3SIV Nd Nd <LOR (0.6) NdSIV a,b 84 ± 1.6 46 ± 1.1 116 ± 3.9 12 ± 0.8SV Nd <LOR (0.4) <LOR (1) <LOR (0.4SVI Nd Nd <LOR (0.7) Nd

Day 3 (afternoon sampling)SI 11 ± 0.4 4.6 ± 0.2 15 ± 1.6 1.8 ± 0.3SII 32 ± 0.4 15 ± 1.4 56 ± 9 6.1 ± 0.3SIII 33 ± 3.4 14 ± 0.2 52 ± 2.7 5.3 ± 0SIV <LOR (0.2) <LOR (0.3) 1.4 ± 0.3 <LOR (0.4SIV a,b 103 ± 9.2 61 ± 6.7 165 ± 36.9 17 ± 2.8SV Nd <LOR (0.3) <LOR (1.1) <LOR (0.4SVI Nd <LOR (0.2) <LOR (0.9) <LOR (0.3

Note: Values average of three subsample replicates ± standard deviation of replicates. Nconcentrations < LOR as used in C/Co figures.

Table 2Concentrations of detected PFCs (ng L�1) throughout Plant A.

PFHxA PFHpA PFOA PFNA

October 2009 samplingS1 6.5, 6.0 1.2, 1.1 6.1, 7.4 1.1, 1.4S2 4.3, 7.0 1.3, 1.2 7.5, 7.7 1.2, 1.3S3 5.9, 6.2 1.2, 1.7 7.1, 7.4 1.2, 1.1S4 6.2, 5.1 1.5, 1.1 8.3, 5.9 <LOR (0.8)S5 4.8, 4.8 1.3, 1.7 7.7, 9.8 <LOR (0.7)S6 5.4, 5.9 2.3, 2.1 11.8, 9.3 <LOR (0.4)S7 5.2� 1.6� 7.6� <LOR (0.5)

November 2009 samplingS1 4.4, 4.4 1.5, 1.2 16, 13.6 1.1, 1.1S2 4.2, 3.8 1.3, 1.1 14.4, 13.9 1.0, 0.9S3 4.8, 5.1 1.7, 1.7 16.7, 17.6 1.1, 1.1S4 4.6, 4.8 1.2, 1.3 12.5, 12.4 0.5, 0.7S5 5.1, 4.9 1.4, 1.5 11.6, 12.1 <LOR (0.5)S6 5.9, 6.4 1.7, 1.9 10.4, 10.3 <LOR (0.4)S7 6.0, 6.5 1.8, 2.0 10.9, 12.1 <LOR (0.4)

Note: Values of two subsample replicates (250 mL) separated by comma. Nd = non detetions < LOR as used in C/Co figures. � Only one replicate analysed, second lost during pr

tions up to 68 ng L�1 and 1050 ng L�1 respectively (Sinclair andKannan, 2006). Lower concentrations have been reported in WWTPeffluent from Beijing with PFOA and PFOS both around 5 ng L�1

(Zhao et al., 2007). In Singapore, WWTP secondary effluent concen-trations of 461.7 ng L�1 PFOS and 1057.1 ng L�1 PFOA have been re-ported (Yu et al., 2009), and in Switzerland PFOS and PFOA havebeen detected in effluents at concentrations up to 303 ng L�1 and35 ng L�1 respectively (Huset et al., 2008). Comparison with thisdata shows the current Australian data is relatively low, especiallyat Plant A, but still comparable with that reported internationally.

3.2. Fate of PFCs during treatment

In Plant A PFCs were detected at all sampling points across thetreatment train. In Plant B PFCs were below detection and report-ing limits in samples taken from points after RO treatment. To as-

PFDA PFBS PFHxS PFOS

1.4 ± 0.4 6.4 ± 0.6 36 ± 2.7 38 ± 1.33.4 ± 0.4 9.5 ± 0.7 70 ± 6.9 79 ± 6.21.9 ± 0.1 6.6 ± 1.4 35 ± 2.1 29 ± 4Nd Nd <LOR (0.2) <LOR (0.1)18 ± 10.9 64 ± 6 217 ± 36.4 157 ± 34.1Nd Nd <LOR (0.3) Nd

) <LOR (0.4) <LOR (0.1) <LOR (0.4) <LOR (0.5)

3.1 ± 0.4 4.8 ± 0.4 28 ± 2.1 39 ± 12.8 ± 0.3 6 ± 1.1 31 ± 4.3 41 ± 1.52.5 ± 0.2 5.5 ± 0.9 30 ± 1.5 29 ± 2.4<LOR (0.1) Nd <LOR (0.1) Nd12 ± 1 48 ± 0.8 176 ± 1.1 97 ± 3.9

) <LOR (0.3) <LOR (0.1) <LOR (0.4) <LOR (0.5)Nd Nd <LOR (0.1) Nd

2.6 ± 0.3 2.4 ± 0.4 12 ± 1.7 23 ± 2.45.9 ± 1.1 9 ± 2.9 44 ± 2.2 76 ± 56.1 ± 1.3 9.9 ± 1.5 48 ± 5.8 54 ± 1.1

) <LOR (0.2) <LOR (0.2) <LOR (0.4) <LOR (0.5)8.6 ± 0.3 60 ± 7.2 208 ± 20.9 133 ± 10.7

) <LOR (0.4) <LOR (0.1) <LOR (0.3) <LOR (0.4)) <LOR (0.1) Nd <LOR (0.3) <LOR (0.2)

d = non detect. LOR = limit of reporting, values in brackets are average of estimated

PFDA PFBS PFHxS PFOS

1.2, 1.1 Nd^ 1.5, 1.6 2.2, 2.21.1, 1.1 Nd^ 1.9, 1.4 2.2, 2.41.0, 1.1 Nd^ 1.5, 1.6 2.0, 2.2<LOR (0.2) Nd^ 1.6, 1.1 0.8, 1.0<LOR (0.4) 1.7, 1.6 1.4, 1.4 0.8, 1.0<LOR (0.1) 1.7, 1.8 1.4, 1.4 <LOR (0.3)Nd 1.7� 1.1� <LOR (0.3)

1.2, 1.0 Nd 2.3, 2.1 3.7, 3.61.0, 1.0 Nd 2.2, 1.9 3.2, 3.41.1, 1.0 Nd 2.4, 2.9 3.5, 3.6<LOR (0.4) Nd 1.9, 2.0 1.5, 1.3<LOR (0.4) 0.9, 0.7 1.9, 1.7 1.3, 1.4<LOR (0.1) 0.9, 1.2 1.6, 2.2 0.7, 0.7<LOR (0.2) 0.8, 1.3 1.5, 2.0 0.6, 0.7

ct. LOR = limit of reporting, values in brackets are average of estimated concentra-ocessing. ^ Interference in quantifying MRM transition.

14 J. Thompson et al. / Chemosphere 82 (2011) 9–17

sess removal ratios at both plants, concentrations from a givenpoint were divided by the concentrations in the influent (i.e. C/Co). Fig. 3 shows two C/Co values for Plant A calculated using theaverages of the laboratory replicates on each sampling occasion,and the plotting of the average of these two values (error barsare standard deviation of the same). The same approach was usedto produce Fig. 4 where the average C/Co values were combined forsamples taken from Plant B on days 1 and 2. The samples from day3 were not included as in this figure as the flow regimes were ex-

influe

nt (S

1)

post

de-ni

trifica

tion (

S2)

post

pre-O

3 (S3)

post

flocc

. / DAFF

0

50

100

150

200

PFHxAPFHpAPFOAPFNAPFDA

PFHxSPFOS

C/C

o (%

)

Note: C/Co = concentration at sampling point / concentration in insampling periods, error bars = standard deviation

Fig. 3. Removal of selected P

Influ

ent (S

I)

UF feed

(SII)

RO feed

(SIII)

ROC (*SIVa &

b)

RO p

0

500

1000

1500

C/C

o (%

) *

**

I

*

**

Note:- Boxes used to group sampling points bracketing fast procsamples between each one. Processes; * Ultrafiltration (UF), **

Fig. 4. Removal of PFCAs in Plant B (C/

pected to be different and the concentrations were lower than theprevious days, which had been reasonably consistent. For brevityC/Co plots of the PFSAs are not shown for Plant B as these followedessentially the same pattern as for the PFCAs. For Plant B one mightexpect the probable underestimation of influent concentrationsdue to the filtration of some samples to diminish the observed re-moval efficiencies. However due to the grab sampling approach atthis plant we cannot reasonably assume to be able to measure re-moval efficiencies with any accuracy, except in the case of the pro-

(S4)

post

main-O

3 (S5)

post

BAC (S6)

post

final

O3 (S7)

fluent, using mean of average values from both

FCs in Plant A (C/Co%).

ermea

te (S

IV)

AO (SV)

PFHxA

PFHpA

PFOA

PFNA

PFDA

***

Finish

ed w

ater (S

VI)

***

ess elements, allowing for reasonable comparison of grab Reverse Osmosis (RO), *** Advanced oxidation (AO)

Co%) including ROC waste stream.

J. Thompson et al. / Chemosphere 82 (2011) 9–17 15

cesses occurring relatively quickly (i.e. flow across UF and ROmembranes, and through the advanced oxidation cell). These fasterprocesses are noted with asterisks in Fig. 4.

In Plant A the changes in concentrations between influent andfinished water were compound specific. Concentrations of PFOAand shorter PFCAs remained relatively consistent throughout eachtreatment step (i.e. at each sampling point), whereas removal ofPFOS and the longer PFCAs, PFNA and PFDA is observable duringthe treatment process. The removal of these PFCs was associatedprimarily with the DAFF and BAC filtration steps. These resultsare in keeping with other work suggesting an increase in hydro-phobicity and partitioning to solids with increasing fluorocarbonchain length, and higher log Koc values for PFSAs over PFCAs ofthe same carbon chain length (Higgins and Luthy, 2006). The ozon-ation treatments appeared to have no effect on PFC concentrations;consistent with their high chemical stability.

Adsorption via activated carbon has been reported previously asan effective way of removing PFCs, particularly PFOS, from drinkingand wastewater (Pabon and Corpart, 2002; Skutlarek et al., 2006).However in this instance the biologically activated carbon failedto completely remove any of the PFCs. This may be due to the ageof the carbon, which was 1.5 years old at sampling, and the adsorp-tion capacity potentially exhausted. Additionally it may be relatedto the contact time, with a residence time of only 18 min at thistreatment step, this may not be sufficient to allow complete equilib-rium adsorption. Ochoa-Herrera and Sierra-Alvarez (2008) tested anumber of bio-solids obtained from WWTPs and found there weresignificant differences depending on their characteristics. They alsofound a flattening of the adsorption profile with increasing aqueousconcentrations, suggesting saturation of adsorptive sites is possible.Fig. 3 also suggests a slight increase for some compounds followingthe BAC, possibly related to desorption from the solid phase follow-ing changes in aqueous concentrations.

Table 2 gives the actual concentrations at each sampling point ofPlant A. Using the average of measured concentrations on both occa-sions, and assuming a constant daily flow of 8 ML d�1 through theplant, we can estimate a daily flux of PFCs in and out of Plant A. Tak-ing this approach we obtain input amounts of PFOS and PFOA of 24and 86 mg d�1 (8 ML d�1 � 2.9 ng PFOS L�1 and 10.7 ng PFOA L�1).Calculating estimated outputs using the same method provides uswith values of 4 mg PFOS and 76 mg PFOA d�1.

As stated previously, using the available data to assess removalof PFCs in Plant B is more difficult as the grab samples were takenwithout accounting for hydraulic residence times. On each daythere was an apparent increase between the influent, and the UFfeed water, having undergone flocculation/coagulation. This wasobserved for all analytes on all occasions, with the exception ofPFBA, PFPeA and PFDA which decreased by 5–25% on day 2. The ob-served increases ranged from 2% (PFHpA in day 2 samples) to al-most 300% (PFHxS in day 3 samples) of initial concentrations, andfor all analytes were greatest on day 3 (afternoon sampling). At thistime we have no explanation for these apparent increases, and it islikely at least partially due to the sampling of different parcels ofwater at each point. If pre-cursor degradation, or contaminationintroduced via water additives or leaching from pipes was respon-sible, it would seem unlikely that all PFCs would increase in concen-tration to such an extent. Another plausible explanation for theapparent increase could be the effects of filtration on the influentsamples. Table S5 in the supporting information shows a consider-able portion of many of the compounds detected are potentially lostduring filtration. As the samples taken subsequently in the treat-ment process did not require filtration, this may explain why theUF feed water has apparently higher concentrations.

A decrease in concentrations was observed post UF whichagain could be explained by a sampling strategy not accountingfor HRT. However as movement across the UF membranes is a

relatively fast process this decrease is more likely reflective of ac-tual removal. As with the comparison between influent and UFfeed water (sample points SI and SII) the removal of suspendedparticulates and or colloidal material and PFCs associated withthese may explain the differences. In this instance however it isfiltration in the actual plant responsible, rather than a methodo-logical artifact. Actual concentrations of PFCs measured through-out Plant B are provided in Table 3.

Following UF, the rejection via RO is consistent with what wasanticipated, and also occurs on a fast enough timeframe to allowinterpretation of the grab samples collected. The RO membranesused at the plant are expected to reject compounds > 100–300 Da, and PFOS and PFOA are �499 Da and�413 Da respectively.Although small peaks were still observable in the chromatogramsfollowing reverse osmosis, no PFCs were detected above reportinglimits in the RO permeate. Due to this, the efficacy of the advancedoxidation (AO) step to treat PFC contamination cannot be accu-rately determined here. It is noteworthy that Schröder and Meest-ers (2005) subjected PFOS and other fluorinated surfactants tovarious forms of advanced oxidation and found no significant effecton PFOS concentrations. Concentrations in the ROC were the high-est of anywhere in the plant, for example, PFOS concentrationswere up to 207 ng L�1 and PFHxS concentrations up to 253 ng L�1.Samples taken pre and post nutrient removal treatment of the ROCwater showed no substantial differences in PFC concentrations.

In the final water all PFCs were below reporting and or detec-tion limits, however small peaks for some PFCs were observed inthe chromatograms similar to those observed in the RO permeateand AO samples. These peaks equated to estimated concentrationsof 0.1–1 ng L�1 for PFHpA, PFOA, PFNA, PFHxS and PFOS, but thesemust be considered in light of the uncertainty surrounding peaksclose to baseline noise and blank levels. A rough mass balance ofPFCs in and out of Plant B can be estimated for the period of sam-pling, by multiplying the average concentrations of all three sam-pling days with the approximate amount of water being treateddaily at the time. Using PFOS and PFOA as examples we can esti-mate daily inputs of 1.4 g d�1 and 0.9 g d�1 respectively(44 ML d�1 � 33 ng PFOS L�1 and 22 ng PFOA L�1). Assuming theoutput of water is split 85% and 15% between purified recycledwater and the ROC, the outflows of PFOS and PFOA are estimatedat 0.9 g d�1 and 1 g d�1 in the ROC (7 ML d�1 � 136 ng PFOS L�1

and 152 ng PFOA d�1). Although using this method there is a sur-plus in the outflows of PFOA relative to inflows and a deficiencyin PFOS outflows, we must remember this is a simplistic approachbased on assumed rather than measured water flows, and usingconcentrations derived from grab samples. As a ‘worst case sce-nario’ if we set the concentrations in the finished water to equalhalf their reporting limits, then PFOS and PFOA would be emittedvia the reclaimed water at approximately 0.02 g d�1 each. Despitethe many shortcomings of this approach, it suggests that bothcompounds are almost entirely conserved within the system, andremoved from the finished water following rejection by RO.

As both plants are designed to produce water meeting variousstringent water quality standards, and ideally potable standards,we can compare the finished water from both with concentrationsreported in drinking water internationally. A number of studieshave looked at PFCs in tap water, and aside from those drawn fromcontaminated source waters (e.g. Emmett et al., 2006; Skutlarek etal., 2006), most have found concentrations either below detection,or in the low ng L�1 range (typically < 10 ng L�1) (Mak et al., 2009;Quinones and Snyder 2009). Of the PFCs present in tap water, PFOAappears to be the most prevalent. The data presented here showswater produced in Plant A is at the higher end of the spectrum withregards to previously reported concentrations of PFCs in tap water.It should be stressed that there are no plans to use reclaimed waterfrom Plant A to supplement potable supplies.

16 J. Thompson et al. / Chemosphere 82 (2011) 9–17

3.3. Comparison of results with toxicological and guideline values

Concentrations in both influent and finished water of bothplants were well below the USEPA’s provisional drinking waterguideline value of 200 ng L�1 for PFOS and 400 ng L�1 for PFOA.They were also far lower than the ‘immediate action’ values of500 ng L�1 (exposure to infants and pregnant women) or5000 ng L�1 (exposure to adults) for combined PFOS and PFOA con-centrations suggested by the German Drinking Water Commission(Trinkwasserkommission, 2006).

In terms of environmental risks to the aquatic environmentboth plants are producing finished water that is well below anyavailable guidelines and hence according to the present knowledgethe risks that the PFCs in the finished water would affect aquaticenvironments are considered very low. A recent review by Giesyet al. (2010) also suggested that the NOEC values are typically inthe mg L�1 range which supports the assertion that the currentrisks to the aquatic organisms are extremely low. On the otherhand the bioaccumulation potential of certain PFCs is widely rec-ognized and Giesy et al. recognized that the protection of avianpredators consuming fish in equilibrium with their surroundingsmay require much lower guidelines with a proposed value of47 ng L�1 for PFOS. This guideline would be close to the measuredconcentration in the influent of the water that reaches Plant B andis exceeded in the ROC but the dilution at the outfall would rapidlyresult in much lower concentrations in the receiving water, hencethe risks associated with the release of the water are consideredminimal.

4. Conclusion

The results presented here support the already reported appar-ent ubiquity of PFCs in wastewater, and demonstrate their pres-ence for the first time in Australian wastewater. The influentconcentrations at each reclamation plant suggest several PFCs,including PFOS and PFOA, are not removed in the standard WWTPsproviding influent for each plant. Comparing the two reclamationfacilities, Plant A showed some removal during the adsorption/fil-tration stages. Overall however Plant A failed to completely re-move PFOS and the PFCAs shorter than PFNA in chain length. InPlant B the grab sampling approach means we cannot accuratelydescribe the fate of PFCs in and out of the entire plant, howeverrestricting the discussion to process steps with short residencetimes, all PFCs were reduced following micro-filtration and furtherreduced to below quantification post reverse osmosis. This is inkeeping with previously published results. With regards to humanand environmental impacts, concentrations of PFCs going into andout of the two reclamation facilities were below those whichwould be of immediate concern. The highest concentrations werefound in the waste stream from the RO at Plant B, and at presentthis is discharged into a river. Further work would be beneficialto remove these compounds to prevent further cycling throughthe environment and avoid any potential environmental impacts.Likewise PFCs are still present after treatment at Plant A, whichwould benefit from additional steps to trap and remove these com-pounds. At present the outputs of both facilities represent a routeof transport of PFCs through the environment.

Acknowledgements

We acknowledge financial support from the Urban Water Secu-rity Research Alliance, ARC Linkage (LP 0774925) and industrypartners. The authors also wish to acknowledge the regional coun-cils and treatment plant operators for their support and Steve Car-ter and Shalona Anuj of Queensland Health for access to

instrumentation and assistance with analysis. Entox is jointlyfunded by The University of Queensland and Queensland HealthForensic and Scientific Services.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.chemosphere.2010.10.040.

References

Calafat, A.M., Kuklenyik, Z., Reidy, J.A., Caudill, S.P., Tully, J.S., Needham, L.L., 2007.Serum concentrations of 11 perfluoroalkyl compounds in the US population:data from the national health and nutrition examination survey (NHANES)1999–2000. Environ. Sci. Technol. 41 (7), 2237–2242.

Cheng, J., Vecitis, C.D., Park, H., Mader, B.T., Hoffmann, M.R., 2008. Sonochemicaldegradation of perfluorooctane sulfonate (PFOS) and perfluoorocatnoate (PFOA)in landfil groundwater: Environmental matrix effects. Environ. Sci. Technol. 42(21), 8057–8063.

Emmett, E.A., Shofer, F.S., Zhang, H., Freeman, D., Desai, C., Shaw, L.M., 2006.Community exposure to perfluorooctanoate: Relationships between serumconcentrations and exposure sources. J. Occup. Environ. Med. 48 (8), 759–770.

Giesy, J.P., Kannan, K., 2002. Perfluorochemical surfactants in the environment.Environ. Sci. Technol. 36 (7), 146A–152A.

Giesy, J.P., Naile, J.E., Khim, J.S., Jones, P.D., Newsted, J.L., 2010. Aquatic toxicology ofperfluorinated chemicals. Rev. Environ. Contam. Toxicol. 202, 1–52.

Higgins, C.P., Field, J.A., Criddle, C.S., Luthy, R.G., 2005. Quantitative determination ofperfluorochemicals in sediments and domestic sludge. Environ. Sci. Technol. 39(11), 3946–3956.

Higgins, C.P., Luthy, R.G., 2006. Sorption of perfluorinated surfactants on sediments.Environ. Sci. Technol. 40 (23), 7251–7256.

Hori, H., Yamamamoto, A., Hayakawa, E., Taniyasu, S., Yamashita, N., Kutsuna, S.,Kiatagawa, H., Arakawa, R., 2005. Efficient decomposition of environmentallypersistent perfluorocarboxylic acids by use of persulfate as a photochemicaloxidant. Environ. Sci. Technol. 39 (7), 2383–2388.

Huset, C.A., Chiaia, A.C., Barofsky, D.F., Jonkers, N., Kohler, H.-P.E., Ort, C., Giger, W.,Field, J.A., 2008. Occurrence and mass flows of fluorochemicals in the GlattValley watershed, Switzerland. Environ. Sci. Technol. 42 (17), 6369–6377.

Kennedy, G.L., Butenhoff, J.L., Olsen, G.W., O’Connor, J.C., Seacat, A.M., Perkins, R.G.,Biegel, L.B., Murphy, S.R., Farrar, D.G., 2004. The toxicology ofperfluorooctanoate. Crit. Rev. Toxicol. 34 (4), 351–384.

Lau, C., Anitole, K., Hodes, C., Lai, D., Pfahles-Hutchens, A., Seed, J., 2007.Perfluoroalkyl acids: a review of monitoring and toxicological findings.Toxicol. Sci. 99, 366–394.

Loganathan, B.G., Sajwan, K.S., Sinclair, E., Kumar, K.S., Kannan, K., 2007.Perfluoroalkyl sulfonates and perfluorocarboxylates in two wastewatertreatment facilities in Kentucky and Georgia. Water Research 41, 4611–4620.

Mak, Y.L., Taniyasu, S., Yeung, L.W.Y., Lu, G., Jin, L., Yang, Y., Lam, P.K.S., Kannan, K.,Yamashita, N., 2009. Perfluorinated Compounds in Tap Water from China andSeveral Other Countries. Environ. Sci. Technol. 43 (13), 4824–4829.

NICNAS. NICNAS Alert No. 6 Perfluorooctanoic Acid (PFOA) & PerfluorocarboxylicAcid (PFCA). In: DoHA, Ed. Sydney: Australian Government, 2007.

NICNAS. NICNAS Alert No. 8 Perfluorooctane Sulfonate (PFOS) & PerfluoroalkylSulfonate (PFAS). In: DoHA, Ed. Sydney: NICNAS, 2008.

Ochoa-Herrera, V., Sierra-Alvarez, R., 2008. Removal of perfluorinated surfactantsby sorption onto granular activated carbon, zeolite and sludge. Chemosphere 72(10), 1588–1593.

OECD. Co-operation on existing chemicals: Hazard assessment of perfluorooctanesulfonate (PFOS) and its salts: OECD, 2002.

Pabon, M., Corpart, J.M., 2002. Fluorinated surfactants: synthesis, properties,effluent treatment. Journal of Fluorine Chemistry 114, 149–156.

Paul, A.G., Jones, K.C., Sweetman, A.J., 2009. A first global production, emission, andenvironmental inventory for perfluorooctane sulfonate. Environ. Sci. Technol.43 (2), 386–392.

Prevedouros, K., Cousins, I.T., Buck, R.C., Korzieniowski, S.H., 2006. Sources, fate andtransport of perfluorocarboxylates. Environ. Sci. Technol. 40 (1), 32–44.

Quinones, O., Snyder, S.A., 2009. Occurrence of Perfluoroalkyl Carboxylates andSulfonates in Drinking Water Utilities and Related Waters from the UnitedStates. Environ. Sci. Technol. 43 (24), 9089–9095.

Schröder, H.F., Meesters, R.J.W., 2005. Stability of fluorinated surfactants inadvanced oxidation processes – a follow up of degradation products usingflow injection-mass spectrometry, liquid chromatography-mass spectrometryand liquid chromatography-multiple stage mass spectrometry. J. Chromatogr. A1082 (1), 110–119.

Schultz, M.M., Barofsky, D.F., Field, J.A., 2006. Quantitative determination offluorinated alkyl substances by large-volume-injection liquid chromatographytandem mass spectrometry - Characterization of municipal wastewaters.Environ. Sci. Technol. 40 (1), 289–295.

Sinclair, E., Kannan, K., 2006. Mass loading and fate of perfluoroalkyl surfactants inwastewater treatment plants. Environ. Sci. Technol. 40 (5), 1408–1414.

Sinclair, E., Mayack, D.T., Roblee, K., Yamashita, N., Kannan, K., 2006. Occurrence ofperfluoroalkyl surfactants in water, fish, and birds from New York State.Archives of Environmental Contamination and Toxicology 50, 398–410.

J. Thompson et al. / Chemosphere 82 (2011) 9–17 17

Skutlarek, D., Exner, M., Farber, H., 2006. Perfluorinated surfactants in surface anddrinking waters. Environmental Science and Pollution Research 13 (5), 299–307.

Steinle-Darling, E., Reinhard, M., 2008. Nanofiltration for trace organic contaminantremoval: structure, solution, and membrane fouling effects on the rejection ofperfluorochemicals. Environ. Sci. Technol. 42 (14), 5292–5297.

Stockholm Convention, 2010. <http://chm.pops.int/Convention/ThePOPs/tabid/673/language/en-US/Default.aspx>. (access date 22.08.10).

Tang, C.Y., Fu, Q.S., Robertson, A.P., Criddle, C.S., Leckie, J.O., 2006. Use of reverseosmosis membranes to remove perfluorooctane sulfonate (PFOS) fromsemiconductor wastewaterâ€. Environ. Sci. Technol. 40 (23), 7343–7349.

Taniyasu, S., Kannan, K., So, M.K., Gulkowska, A., Sinclair, E., Okazawa, T., Yamashita,N., 2005. Analysis of fluorotelomer alcohols, fluorotelomer acids, and short- andlong-chain perfluorinated acids in water and biota. J. Chromatogr. A 1093, 89–97.

Thompson, J., Lorber, M., Toms, L.M., Kato, K., Calafat, A.M., Mueller, J.F., 2010. Use ofsimple pharmacokinetic modeling to characterize exposure of Australians toperfluorooctanoic acid and perfluorooctane sulfonic acid. Environ. Int. 36 (4),390–397.

Toms, L.-M.L., Calafat, A., Kato, K., Thompson, J., Harden, F., Hobson, P., Sjodin, A.,Mueller, J.F., 2009. Polyfluoroalkyl chemicals (PFCs) in human blood serum fromchildren and adults in Australia. Environ. Sci. Technol. 43 (11), 4194–4199.

Trinkwasserkommission. Provisional evaluation of PFT in drinking water with theguide substances perfluorooctanoic acid (PFOA) and perfluorooctane sulfonate(PFOS) as examples. In: Health GMo, Ed. Bonn: Federal Environment Agency,2006.

USEPA, 2009. <http://www.epa.gov/oppt/pfoa/pubs/stewardship/index.html>(access date).

Vestergren, R., Cousins, I.T., Trudel, D., Wormuth, M., Scheringer, M., 2008.Estimating the contribution of precursor compounds in consumer exposure toPFOS and PFOA. Chemosphere 73 (10), 1617–1624.

Wang, T., Wang, Y., Liao, C., Cai, Y., Jiang, G., 2009. Perspectives on the Inclusion ofPerfluorooctane Sulfonate into the Stockholm Convention on Persistent OrganicPollutants. Environ. Sci. Technol. 43 (14), 5171–5175.

Washington, J.W., Ellington, J.J., Jenkins, T.M., Evans, J.J., Yoo, H., Hafner, S.C., 2009.Degradability of acrylate-linked, fluorotelomer polymer in soil. Environ. Sci.Technol. 43 (17), 6617–6623.

Yu, J., Hu, J., Tanaka, S., Fujii, S., 2009. Perfluorooctane sulfonate (PFOs) andperfluorooctanoic acid (PFOA) in sewage treatment plants. Water Res. 43, 2399–2408.

Zhao, X., Li, J., Shi, Y., Cai, Y., Mou, S., Jiang, G., 2007. Determination of perfluorinatedcompounds in wastewater and river water samples by mixed hemicelle-basedsolid-phase extraction before liquid chromatography–electrospray tandemmass spectrometry detection. J. Chromatogr. A 1154, 52–59.