Available online at www.sciencedirect.com

Chemosphere 71 (2008) 306–313 www.elsevier.com/locate/chemosphere

Analysis of perfluorooctanoate (PFOA) and other perfluorinated compounds (PFCs) in the watershed in N-

Robert Loos *, Giovanni Locoro, Tania Huber, Jan Wollgast, Eugen H. Christoph, Alfred de Jager, Bernd Manfred Gawlik, Georg Hanke, Gunther Umlauf, Jose´-Manuel Zaldı´var

Institute for Environment and Sustainability (IES), Joint Research Centre (JRC) of the European Commission, Via Enrico Fermi, 21020 Ispra, TP 290, Italy

Received 12 June 2007; received in revised form 28 August 2007; accepted 2 September 2007 Available online 23 October 2007

Abstract

C7–C11 perfluorinated carboxylates (PFACs) and perfluorooctansulfonate (PFOS) were analysed in selected stretches of the River Po and its major tributaries. Analyses were performed by solid-phase extraction (SPE) with Oasis HLB cartridges and methanol elution followed by LC–MS–MS detection using 13C-labelled internal standards. High concentration levels (1.3 lgl1) of perfluorooctanoate (PFOA) were detected in the Ta´naro River close to the city . After this tributary, levels between 60 and 337 ng l1 were measured in the Po River on several occasions. The PFOA concentration close to the river mouth in was between 60 and 174 ng l1. Using the river discharge flow data in m3 s1 at this point (average 920 m3 s1 for the year 2006), a mass load of 0.3 kg PFOA per hour or 2.6 tons per year discharged in the has been calculated. PFOS concentration levels in the Po River at Ferrara were 10 ng l1. 2007 Elsevier Ltd. All rights reserved.

Keywords: River Po; Perfluorooctanoate (PFOA); Perfluorooctansulfonate (PFOS); Solid-phase extraction (SPE); Liquid chromatography tandem mass spectrometry (LC–MS–MS)

1. Introduction in biota from remote areas like the Arctic (Giesy and Kan- nan, 2001; Martin et al., 2004a). Contrary to other classical Concern about perfluorinated organic compounds persistent organic pollutants (POPs), these chemicals are (PFCs) is growing because they are globally distributed, primarily emitted to water, they accumulate in surface environmentally persistent, bioaccumulative, and poten- waters, and water is the major reservoir of PFCs in the tially harmful. Organofluorine molecules have unique environment, as well as the most important medium for physical, chemical, and biological properties; the high- their transport (Prevedouros et al., 2006; McLachlan energy carbon–fluorine bond renders PFCs resistant to et al., 2007). hydrolysis, photolysis, microbial degradation, and metabo- The two PFCs most commonly used and found in the lism by vertebrates (Giesy and Kannan, 2001; Prevedouros environment are perfluorooctansulfonate (PFOS) and per- et al., 2006). They have been detected in human blood fluorooctanoate (PFOA). They are widely employed in dif- worldwide (Kannan et al., 2004; Houde et al., 2006; Mida- ferent industrial processes such as in protective coatings for sch et al., 2006), in human milk (Ka¨rrman et al., 2007), and carpets, textiles (rain gear; e.g. Goretex), leather, paper, paints, adhesives, grease repellents in popcorn bags and * Corresponding author. Tel.: +39 0332 786407; fax: +39 0332 785355. food-wrapper coatings, wiring insulation for telecommuni- E-mail address: [email protected] (R. Loos). cations, in fire-fighting foams, as specialty surfactants in

0045-6535/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2007.09.022 R. Loos et al. / Chemosphere 71 (2008) 306–313 307 cosmetics, aerospace, electronics (semiconductors), and begins in the upper reaches of the Western (Monviso medical use (Martin et al., 2004b; Prevedouros et al., mountain), flowing west to east across northern Italy in a 2006). In particular, PFOA is used as adjuvant in the pro- highly populated and industrialized region, ending in the duction of fluoropolymers such as PTFE, Teflon or simi- Adriatic Sea. It is 652 km in length, and the median annual lar products, and occurs in these applications as aqueous discharge flow is around 1.540 m3 s1 (Wikipedia, URL 2). and gaseous process emission (Fricke and Lahl, 2005; It is, in terms of water flow and length, among the 10 larg- Davis et al., 2007). est European (wordatlas, URL 3). The objective of Point source manufacturing facilities of fluorochemicals this study was to identify the PFOA source tributary in are one of the largest sources of emissions for PFCs (Preve- the River Po watershed. douros et al., 2006; Davis et al., 2007). Municipal and industrial wastewater treatment plant (WWTP) effluents 2. Experimental are minor sources of PFC emissions (Boulanger et al., 2005a,2005b; Prevedouros et al., 2006; Schultz et al., 2.1. Chemicals and reagents 2006; Sinclair and Kannan, 2006). Other research suggests that indirect sources such as volatile atmospheric precur- The analytical standards were delivered by the Sigma– sors (alcohol fluorotelomers) may be important contribu- Aldrich group, and the 13C-labelled internal standards for 13 13 tors to PFC levels in the environment (Ellis et al., 2004). PFOA ( C4) and PFOS ( C4) by Wellington Laboratories PFCs, especially PFOS and PFOA, have been found in (Guelph, Ontario, Canada). Methanol (SupraSolv), aceto- surface waters in Japan (Saito et al., 2004a, 2004b), the nitrile (LiChrosolv for HPLC), and acetone (SupraSolv) USA (Hansen et al., 2002; Boulanger et al., 2004, were obtained from Merck (Darmstadt, ) and 2005a,2005b), and Europe (Skutlarek et al., 2006; Loos Sigma–Aldrich. et al., 2007; McLachlan et al., 2007). PFOA has been found Single standard stock solutions of the analytes in the in relatively high concentration levels near fluorochemical high mg l1 range were prepared by weighing mg amounts manufacturing facilities in the USA, 500 ng l1 in the Ten- of the compounds and dissolving in 10 ml methanol. The nessee River (Hansen et al., 2002), in the Ohio River (Davis working standard solutions were prepared by further dilut- et al., 2007), and in Germany up to 56 lgl1 in the River ing the stock standard solutions with methanol. Alz (near a fluorochemical factory), and 33.9 lgl1 in the Creek Steinbecke. In addition, it has been found in drinking 2.2. Samples and sample pretreatment waters, up to 519 ng l1 in Germany (Schaefer, 2006; Skut- larek et al., 2006; Schulte, 2007), and in ground water near Water samples from the major River Po tributary rivers the Ohio River, USA (Davis et al., 2007). Moreover, PFOS were investigated (Table 1). The samples were taken in and PFOA have also been analysed in ocean waters with February–March 2007. Not sampled were the smaller trib- detected concentrations up to 100 pg l1 and marine utaries Pellice (22.3 m3 s1), Dora Riparia (26 m3 s1), mammals (Kannan et al., 2002; Yamashita et al., 2004, Orco (28 m3 s1), (23 m3 s1), (40 m3 s1), 2005; Schulte, 2007; Hauka˚s et al., 2007). (30 m3 s1), (11.3 m3 s1), Enza (12 m3 s1), An OECD hazard assessment from the year 2002 and (37 m3 s1) (Wikipedia, URL 2). (OECD, 2002) identified PFOS as a PBT-chemical (persis- Water samples were collected in 1 l glass bottles (Schott tent, bioaccumulative, toxic). It was proposed to include Duran) and stored at 5 C in the laboratory until further PFOS as a candidate in the Stockholm convention on persis- processing. The bottles were washed in a laboratory dish- tent organic pollutants (POPs). In December 2006, the washer and then heated in an oven at 450 C. The water European Parliament and the Council decided to restrict samples were not filtered. From the 1 l water samples col- marketing and use of PFOS with a few exceptions by amend- lected 500 ml water was taken (decanted), the internal stan- 1 13 13 ing Council Directive 76/769/EC on dangerous substances dard added (10 ng l for PFOA C4 and PFOS C4), and for PFOS (EC, 2006). It is currently being discussed if PFOA then 400 ml water automatically extracted by SPE. should be included in this Directive. In addition, it is inves- tigated to introduce PFOS and PFOA as so-called priority 2.3. Solid-phase extraction substances regulated by the Water Framework Directive (WFD). PFOA has a lower bioaccumulation potential in The water samples were extracted by solid-phase extrac- biota (e.g. fish) than PFOS (Hauka˚s et al., 2007). There is tion (SPE). The SPE procedure for the clean-up and concen- weak evidence for cancerogenic properties of PFOS and tration of water samples was performed automatically using PFOA (Kennedy et al., 2004; Saito et al., 2004). an AutoTrace SPE workstation (Caliper Life Sciences). In a European study carried out by the Perforce consor- 200 mg (6 ml) Oasis HLB columns (Waters, Milford, tium (PERFORCE project, URL 1) considering the major MA, USA) were used. The cartridges were activated and EU rivers, the Po watershed was identified as the dominant conditioned with 5 ml methanol and 5 ml water at a flow- source of PFOA in Europe. The Po accounted for two rate of 5 ml min1. The water samples (400 ml) were passed thirds of the total PFOA discharge of the rivers studied through the wet cartridges at a flow-rate of 5 ml min1, (McLachlan et al., 2007). The Po River, Italy’s longest river the columns rinsed with 2 ml water (flow 3 ml min1), and 308 R. Loos et al. / Chemosphere 71 (2008) 306–313

Table 1 Sample details and PFC concentrations in the river water samples Point on the Sampling Location River River flow PFOA PFHpA PFNA PFUnA PFOS map date tributaries [m3 s1] [ng l1] [ng l1] [ng l1] [ng l1] [ng l1] 2 20.02.07 Torino di 32 2 n.d. 1 1 4 Lanzo 5 20.02.07 Saluggia Dora Baltea 110 1 n.d. 1 n.d. 1 6 20.02.07 Vercelli Sesia 80 5 1 7 2 2 8 01.03.07 Rivarone Ta´naro 132 1270 18 6 2 2 9 07.02.07 Ponte della Becca Ticino 350 4 1 2 1 7 11 01.03.07 San Zenone al Po Olona 30 11 2 1 1 25 13 07.02.07 Lambrinia Lambro 40 20 2 13 2 21 01.03.07 16 3 9 2 21 14 06.03.07 Pizzighettone Adda 190 10 3 3 1 25 15 06.03.07 Marcaria 137 3 1 2 n.d. 3 16 06.03.07 Governolo Mincio 60 2 n.d. 1 n.d. 1 18 06.03.07 Quistello Se´cchia 42 1 n.d. n.d. n.d. n.d. River Po (in Ferrara) 1 20.02.07 Torino (centre) Po 875 2 n.d. 1 1 4 3 20.02.07 San Mauro Po 2 n.d. 1 n.d. 2 Torinese 4 20.02.07 Po 2 n.d. 1 1 2 7 01.03.07 Casale Po 814 3 1 2 1 2 Monferrato 10 01.03.07 Portalbera Po 337 5 3 1 2 12 07.02.07 Monticelli Pavia Po 811 91 2 2 1 6 06.03.07 Governolo Po 744 70 2 3 1 12 (Mantova) 17 06.03.07 San Benedetto al Po 78 2 2 1 8 Po 19 29.06.06 Pontelagoscuro Po 230 160 1 n.d. 1 12 (Ferrara) 05.09.06 Po 685 174 5 1 1 10 11.10.06 Po 1512 60 1 n.d. n.d. 7 Not sampled were the smaller river tributaries Pellice (22.3 m3 s1), Dora Riparia (26 m3 s1), Orco (28 m3 s1), Scrivia (23 m3 s1), Trebbia (40 m3 s1), Taro (30 m3 s1), Parma (11.3 m3 s1), Enza (12 m3 s1), and Panaro (37 m3 s1); n.d. = not detected (below LOD). PFDA was not analysed because its MRM transition was confused with that of PFUnA. the cartridges dried for 30 min using nitrogen at 0.6 bars. The eluants used for the separations of the target ana- Elution was performed with 6 ml methanol. Evaporation lytes were water and acetonitrile. The water phase used of the extracts with nitrogen to 300 ll was performed at a was acidified with 0.1% acetic acid (pH 3.5). The flow-rate temperature of 40 C in a water bath using a TurboVap was 0.25 ml min1. The gradient started with 90% water II Concentration Workstation (Caliper Life Sciences). and proceeded to 90% acetonitrile over 25 min, conditions hold for 5 min, returned back to the starting conditions over 2.4. Liquid chromatography tandem mass spectrometry 5 min, and followed by 5 min equilibration. The injection (LC–MS-MS) volume was 5 ll; injection was performed by the autosampler. Analyses were performed by reversed-phase liquid chro- Instrument control, data acquisition and evaluation matography (RP-LC) followed by electrospray ionization (integration and quantification) were done with MassLynx (ESI) mass spectrometry (MS) detection using atmo- software. Nitrogen is used as the nebulizer gas and argon spheric-pressure ionization (API) in the negative ionization as the collision gas. Capillary voltage was operated at mode with a triple–quadrupole MS–MS system. LC was 2.8 kV in the negative mode, extractor lens at 1.0 V, performed with an Agilent 1100 Series LC system consist- and RF lens at 0.0 V. The source and desolvation temper- ing of a binary pump, vacuum degasser, autosampler and atures were set to 80 and 150 C when working with syringe a thermostated column compartment. LC separations were injection for MS–MS optimisations and to 120 and 350 C performed using a Hypersil column (Thermo under chromatographic HPLC conditions. Cone and desol- Electron Corp., 100 · 2.1 mm, 3 lm particles). Tandem vation gas flows were 50 and 600 l h1, respectively. The mass spectrometry was performed on a bench-top triple– applied analyser parameters for MRM analysis were: LM quadrupole Quattro micro MS from Waters-Micromass 1 and HM 1 resolution 11.0, ion energy 1 1.0, entrance (Manchester, UK) equipped with an electrospray probe 1 (negative mode), 2 (positive mode), exit 1, LM 2 and and a Z-spray interface. HM 2 resolution 10.0, ion energy 2 2.0, multiplier 600 V. R. Loos et al. / Chemosphere 71 (2008) 306–313 309

The MRM inter-channel delay was 0.05 and the inter-scan 2.6. Quality assurance delay 0.15. Quantitative LC–MS-MS analysis was performed in the Good performance of the method was demonstrated by multiple reaction monitoring (MRM) mode. Collision- successful participation in a dedicated intercomparison induced dissociation (CID) was carried out using argon exercise on PFCs with ITM (Stockholm University, Swe- at approx. 3.5 · 103 mbar as collision gas at collision den, Urs Berger) and NILU Norwegian Institute of Air energies of 7–40 eV. The optimized characteristic MRM Research (Tromsø, Norway; Dorte Herzke). In this inter- precursor ! product ion pairs monitored for the quantifi- comparison, four water samples were analysed by each lab- cation of the compounds together with the cone voltage oratory, two identical from the River Rhine and two from and collision energy are given in Table 2. the Po River. Good agreement was achieved for the three The compound-dependent method detection limits labs in this exercise (McLachlan et al., 2007). (MDLs) referring to the MRM quantitation for the SPE- LC–MS–MS procedure were determined from real water 3. Results and discussion samples, at a signal-to-noise ratio of 3; 400 ml water was extracted and concentrated to 300 ll (enrichment factor 3.1. Method development 1333). The injection volume was 5 ll. Low blank values (<1 ng l1) were identified for the PFCs (Loos et al., The applied LC gradient of the chromatographic sepa- 2007). No special precautions were necessary to avoid sam- ration started with 90% water (containing 0.1% acetic acid) ple contamination during analysis. and was going up to 90% acetonitrile because also other earlier eluting compounds (e.g. diclofenac, ibuprofen, ben- 2.5. Identification and quantification tazone) were analysed together with the PFCs. Under these chromatographic conditions the PFCs are eluting relatively The compounds were identified by retention time match late, after 24 min (see Table 2); the PFCs were analysed in a and their specific MRM transitions. The first mass transi- time-scheduled LC–MS/MS detection window between 22 tion in Table 2 was used for quantification, and the second and 35 min. Good peak shapes were achieved under these only for confirmation purposes. The PFCs were quantified chromatographic conditions and relatively long retention 13 using C-labelled internal standards; The C7–C11 PFACs times. This improves the previously published method 13 were quantified with the C8 carboxylate ( C4 PFOA), (Loos et al., 2007) where the PFCs were analysed sepa- 13 and PFOS with C4 PFOS. PFDA was not analysed in rately, since less time and organic solvents are consumed. the real samples because its MRM transition (513 > 469) When analyzing only PFCs, the gradient should start with was confused with that of PFUnA (563 > 519), which approximately 60% water. was realized only after the analyses. The recoveries were determined with spike experiments in the concentration 3.2. River Po and tributary water samples range of 10–100 ng l1 using Milli-Q water (replication n = 4); they were around 60% (see Table 2 and (Loos In the year 2006 we found during three exploratory cam- et al., 2007)). Quantification of the PFCs was not affected paigns at the Po River on three occasions relatively high by matrix effects, since the comparison of external with PFOA concentration levels in the River Po in Pontelagos- internal quantification showed very good agreement. curo, Ferrara (160, 174 and 60 ng l1; see Table 1), around

Table 2 Analytical details for the PFCs Compound Formula MRM Cone Coll. Ret. time [min] Recovery [%] MDL [pg/l] PFHpA FxC6–COO 363 > 319 14 10 24.2 61 ± 14 50 363 > 169 PFOA FxC7–COO 413 > 369 14 10 27.5 60 ± 11 100 413 > 169 13 PFOA C4 417 > 372 14 10 27.5 61 ± 9 PFNA FxC8–COO 463 > 419 14 10 29.7 56 ± 9 100 463 > 169 PFDA FxC9–COO 513 > 469 14 11 31.5 60 ± 8 50 PFUnA FxC10–COO 563 > 519 14 11 33.0 61 ± 9 50 PFOS FxC8–SO3 499 > 80 60 47 31.1 56 ± 10 100 499 > 99 13 PFOS C4 503 > 80 60 47 31.1 57 ± 10 13 13 PFHpA: Perfluoroheptanoate, PFOA: Perfluorooctanoate, PFOA C4: Perfluoro-n-[1,2,3,4- C4]octanoate, PFNA: Perfluorononanoate, PFDA: Per- 13 13 fluorodecanoate, PFUnA: Perfluoroundecanoate, PFOS: Perfluorooctansulfonate, PFOS C4: Perfluoro-1-[1,2,3,4- C4]octansulfonate. MRM = multiple reaction monitoring, parent > product transitions, Coll. = collision energy, Recovery = SPE recovery rates from 400 ml Milli-Q water at neutral pH with 200 mg Oasis HLB cartridges at 100 ng l1 spike level (n = 4), MDL = method detection limit of the whole SPE-LC–MS–MS procedure. 310 R. Loos et al. / Chemosphere 71 (2008) 306–313

70 km upstream the river mouth delta. These measure- After this tributary PFOA levels between 60 and 337 ng l1 ments have been confirmed by two laboratories of the were measured in the Po River on several occasions (see PERFORCE consortium (McLachlan et al., 2007). Table 1). Fig. 2 gives LC–MS–MS chromatograms for 13 Therefore, in 2007 an initiative to identify the major PFOA and the C4 labelled internal standard for three PFOA source(s) in the Po River basin was started. The water samples, the River Po at Portalbera (down- major tributary rivers and the Po at different locations were of the Ta´naro tributary influent, point 10 on the map), sampled and the water samples analyzed for PFCs and the River Ta´naro at Rivarone (point 8), and the River Po other organic substances (e.g. herbicides, pharmaceuticals, at (upstream the Ta´naro tributary). nonylphenol) (see Fig. 1). Fig. 3 shows LC–MS–MS chromatograms of PFCs close All tributaries except the Ta´naro (Alessandria) showed to the LOD. relatively low PFOA levels. In the Ta´naro River in Riva- In 2006, PFOA was the organic compound with the rone a PFOA concentration of 1270 ng l1 was detected. highest concentration determined by us in the Po River

Fig. 1. Po River basin with sampling points. Map based on CCM River and Catchment Database for Europe (Vogt et al., 2003).

100 27.70 1.24e4 417 > 372 Area 13C -IS

% 4

0

100 27.70 PFO A 337 ng l -1 3.97e5 River Po, Portalbera Area 413 > 369 %

0

27.62 100 1.26e4 13 417 > 372 Area C4-IS %

0

100 -1 27.62 PFO A 1270 ng l 1.48e6 River , RivaroRivarone Area 413 > 369 %

0

28.10 100 1.40e4 13C -IS 417 > 372 Area 4 %

0

100 28.12 -1 4.18e3 413 > 369 River Po, Casale Monferrato PFO A 3 ng l Area %

0 8.00 12.00 16.00 20.00 24.00 28.00 32.00 mins

Fig. 2. LC–MS–MS chromatograms of PFOA. R. Loos et al. / Chemosphere 71 (2008) 306–313 311 water. In 2007 in San Benedetto al Po (PFOA 78 ng l1) A higher average value of 1540 m3 s1 had been reported a higher level was only detected for atrazine (251 ng l1). (Wikipedia, URL 2). Fig. 4 shows the discharge flow Other substances detected were: caffeine (68 ng l1), terbu- data for the year 2006, with our three sampling dates indi- tylazine (53 ng l1), carbamazepine (45 ng l1), bentazone cated. From these data, the mass load of PFOA discharged (45 ng l1), bezafibrate (44 ng l1), and sulfamethoxazole in the Adriatic Sea was calculated. For the first sampling (24 ng l1). on 29.06.2006, it was 0.132 kg PFOA per hour, for the sec- In the year 2006, the River Po had an average annual ond on 05.09.2006 0.429 kg h1, and for the third on discharge flow of 920.5 m3 s1 at Pontelagoscuro (Ferrara) 11.10.2006 0.327 kg h1. From these data an average mass measured by the Provincia di Ferrara (ARPA, URL 4). load of 0.3 kg PFOA per hour discharged in the Adriatic

-1 2.19e3 33.47 100 PFUnA 1 ng l Area 563 > 519 % 0

31.50 100 PFOS 13C 2.25e4 503 > 80 4 Area % 0

-1 2.52e3 31.50 100 PFOS 2 ng l Area 499 > 80 % 0

-1 100 3.42e3 30.25 463 > 419 PFNA 2 ng l Area % 0

1.40e4 28.10 13 417 > 372 100 Area PFOA C4 % 0

-1 100 4.18e3 28.12 413 > 369 Area PFOA 3 ng l % 0

-1 2.21e3 100 25.00 363 > 319 PFHpA1 ng l Area % 0 10.00 14.00 18.00 22.00 26.00 30.00 34.00 38.00 mins

Fig. 3. LC–MS–MS chromatograms of PFCs close to the LOD; River Po at Casale Monferrato (before the Ta´naro River influent).

5000

4500

4000

3500 ] -1 s 3 3000

2500 11.10.2006 1512 m3 s-1 2000 Discharge flow [m 1500 05.09.2006 685 m3 s-1 29.06.2006 1000 230 m3 s-1

500

0

6 6 6 6 6 6 06 /06 06 06 /06 06 06 06 0 /0 06 /0 1/0 4/06 5/06 6/06 7/06 7/ 8/06 9/06 0/06 0 2/06 2/ /0 /01 /02/06 /02/ /03/ /0 /04/0 /0 /06 /0 /0 /07/ /08/ /0 /09/ /0 /1 /1 /11/06 /11/0 /1 /1 /12 0 8 7 01/01/ 15 29 12 26 12/03/0626 09 23 07 21/05/0604 18/0 02 16 3 13 27 10 24 0 22 05 19 03 1 31

Fig. 4. River Po discharge flow in Pontelagoscuro (Ferrara) for the year 2006. Data by courtesy of Provincia di Ferrara (URL 4). 312 R. Loos et al. / Chemosphere 71 (2008) 306–313

Sea can be calculated, which corresponds to 2.6 tons per References year. The concentration levels for the other PFACs and PFOS Boulanger, B., Vargo, J., Schnoor, J.L., Hornbuckle, K.C., 2004. were in comparison to PFOA relatively low (see Table 1). Detection of perfluorooctane surfactants in Great Lakes water. 1 Environ. Sci. Technol. 38, 4064–4070. In the Ta´naro River some PFHpA (18 ng l ) and PFNA Boulanger, B., Vargo, J.D., Schnoor, J.L., Hornbuckle, K.C., 2005a. 1 (6 ng l ) were detected as well. The highest PFNA levels Evaluation of perfluorooctane surfactants in a wastewater treatment were found in the Lambro River (9 and 13 ng l1). PFOS system and in a commercial surface protection product. Environ. Sci. inputs in the Po River are mainly coming from the rivers Technol. 39, 5524–5530. Ticino (7 ng l1), Olona (25 ng l1), Lambro (21 ng l1), Boulanger, B., Peck, A.M., Schnoor, J.L., Hornbuckle, K.C., 2005b. Mass 1 budget of perfluorooctane surfactants in Lake Ontario. Environ. Sci. and Adda (25 ng l ). PFOS concentration levels in the Technol. 39, 74–79. 1 Po River in Ferrara were around 10 ng l , which is around Davis, K.L., Aucoin, M.D., Larsen, B.S., Kaiser, M.A., Hartten, A.S., ten times lower than for PFOA. 2007. Transport of ammonium perfluorooctanoate in environmental The PFOA detected in the Po River is probably coming media near a fluoropolymer manufacturing facility. Chemosphere 67, from industrial sources. However, more research is needed 2011–2019. Ellis, D.A., Martin, J.W., De Silva, A.O., Mabury, S.A., Hurley, M.D., to establish the exact origin. It has to be stressed that Sulbaek Andersen, M.P., Wallington, T.J., 2004. Degradation of industry is working on PFOA emission reductions and fluorotelomer alcohols: A likely atmospheric source of perfluorinated major fluoropolymer producers are investigating alterna- carboxylic acids. Environ. Sci. Technol. 38, 3316–3321. tive manufacturing processes that slash emissions of the European Commission, Directive 2006/122/EC of the European Parlia- chemical by 97% (ES&T policy news, URL 5). Davis ment and of the Council of 12 December 2006 amending for the 30th time Council Directive 76/769/EEC on the approximation of the laws, et al. (Davis et al., 2007) have studied the transport path- regulations and administrative provisions of the Member States ways of PFOA from a fluoropolymer manufacturing facil- relating to restrictions on the marketing and use of certain dangerous ity into environmental media (air, soil, ground and river substances and preparations (perfluorooctane sulfonates), Off. J. water). Europ. Commun., L 372/32, 27.12.2006. Fricke, M., Lahl, U., 2005. Risk evaluation of perfluorinated surfactants as contribution to the current debate on the EU Commission’s 4. Conclusions REACH document. Z. Umweltchem. O¨ kotox. 17 (1), 36–49. Giesy, J.P., Kannan, K., 2001. Global distribution of perfluorooctane sulfonate in wildlife. Environ. Sci. Technol. 35, 1339–1342. Relatively high levels of PFOA have been detected in the Hansen, K.J., Johnson, H.O., Eldridge, J.S., Butenhoff, J.L., Dick, L.A., Ta´naro River. The exact origin cannot be established but it 2002. Quantitative characterization of trace levels of PFOS and PFOA is likely coming from industrial sources in the Ta´naro/Bor- in the Tennessee River. Environ. Sci. Technol. 36, 1681–1685. mida watershed. PFOA could gain relevance as river basin Hauka˚s, M., Berger, U., Hop, H., Gulliksen, B., Gabrielsen, G.W., 2007. specific pollutant of the Po watershed. The source of con- Bioaccumulation of per- and polyfluorinated alkyl substances (PFAS) in selected species from the Barents Sea food web. Environ. Poll. 148, tamination in the Ta´naro/ watershed should be 360–371. investigated in more in detail to trace the origin of the Houde, M., Martin, J.M., Letcher, R.J., Solomon, K.R., Muir, D.C.G., PFOA detected. Furthermore, as a further measure, we 2006. Biological monitoring of polyfluoroalkyl substances: a review. propose the analysis of tap (drinking) water along the Environ. Sci. Technol. 40, 3463–3473. rivers Po and Ta´naro to assess possible infiltration of Kannan, K., Corsolini, S., Falandysz, J., Oehme, G., Focardi, S., Giesy, J.P., 2002. Perfluorooctanesulfonate and related fluorinated hydrocar- PFOA into ground water aquifers, since it is known that bons in marine mammals, fishes, and birds from coasts of the Baltic polar persistent compounds such as PFOA are of impor- and the Mediterranean Seas. Environ. Sci. Technol. 36, 3210–3216. tant drinking water relevance (Loos et al., 2007). Little is Kannan, K., Corsolini, S., Falandysz, J., Fillmann, G., Kumar, K.S., known about possible human health impacts of PFOA Loganathan, B.G., Mohd, M.A., Olivero, J., Wouwe, N.V., Yang, and additional information to establish scientifically sound J.H., Aldous, K.M., 2004. Perfluorooctanesulfonate and related fluorochemicals in human blood from several countries. Environ. environmental quality standards are urgently needed. Sci. Technol. 38, 4489–4495. PFOS concentration levels in the Po River at Ferrara were Ka¨rrman, A., Ericson, I., Van Bavel, B., Danerud, P.O., Aune, M., Glynn, 10 ng l1, which is around ten times lower than for A., Lignell, S., Lindstrvm, G., 2007. Exposure of perfluorinated PFOA. The Rivers Ticino, Olona, Lambro, and Adda chemicals through lactation: levels of matched human milk and serum have been identified as major PFOS sources to the Po and a temporal trend, 1996–2004, in Sweden. Environ. Health Perspect. 115, 226–230. River. 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 of perfluorooctanoate. Crit. Rev. Toxicol. 34, 351–384. Acknowledgement Loos, R., Wollgast, J., Huber, T., Hanke, G., 2007. Polar herbicides, pharmaceuticals, perfluorooctansulfonate (PFOS), perfluorooctanoate Michael McLachlan and Urs Berger (ITM, Stockholm (PFOA), nonylphenol and its carboxylates and ethoxylates in surface University) are acknowledged for their scientific advice. and tap waters around Lake Maggiore in Northern Italy. Anal. Furthermore, we would like to thank the Servizio Risorse Bioanal. Chem. 387, 1469–1478. Martin, J.W., Smithwick, M.M., Braune, B.M., Hoekstra, P.F., Muir, Idriche e Tutela Ambientale della Provincia di Ferrara (Sil- D.C.G., Mabury, S.A., 2004a. Identification of long-chain perfluori- vano Bencivelli, Rita Gregatti) for their support and for nated acids in biota from the Canadian Arctic. Environ. Sci. Technol. the retrieval of the River Po discharge flow data. 38, 373–380. R. Loos et al. / Chemosphere 71 (2008) 306–313 313

Martin, J.W. et al., 2004. Environ. Sci. Technol. 38, 249A–255A. chromatography tandem mass spectrometry-characterization of muni- McLachlan, M., Holmstro¨m, K.E., Reth, M., Berger, U., 2007. Riverine cipal wastewaters. Environ. Sci. Technol. 40, 289–295. discharge of perfluorinated carboxylates from the European Conti- Sinclair, E., Kannan, K., 2006. Mass loading and fate of perfluoroalkyl nent, Environ. Sci. Technol, in press. surfactants in wastewater treatment plants. Environ. Sci. Technol. 40, Midasch, O., Schettgen, T., Angerer, J., 2006. Pilot study on the 1408–1414. perfluorooctansulfonate and perfluorooctanoate exposure of the Ger- Skutlarek, D., Exner, M., Fa¨rber, H., 2006. Perfluorinated surfactants in man general population. Int. J. Hyg. Environ. Health 209, 489–496. surface and drinking waters. Environ. Sci. Poll. Res. 13 (5), 299–307. OECD report ENV/JM/RD(2002)17/FINAL, 21 Nov. 2002. Hazard URL 1: http://www.science.uva.nl/perforce/index.htm?http%3A//www. assessment of perfluorooctanesulfonate (PFOS) and its salts. science.uva.nl/perforce/projecto.htm. Prevedouros, K., Cousins, I.T., Buck, R.C., Korzeniowski, S.H., 2006. URL 2: http://it.wikipedia.org/wiki/Po. Sources, fate and transport of perfluorocarboxylates. Environ. Sci. URL 3: http://worldatlas.com/webimage/countrys/euriv.htm. Technol. 40, 32–44. URL 4: ARPA, Provincia di Ferrara, http://www.arpa.emr.it/sim/ Saito, N., Harada, K., Inoue, K., Sasaki, K., Yoshinaga, T., Koizumi, A., ?idrologia. 2004a. Perfluorooctanoate and perfluorooctane sulfonate concentra- URL 5: ES&T policy news, 7 March 2007, http://pubs.acs.org/subscribe/ tions in surface water in Japan. J. Occup. Health 46, 49–59. journals/esthag-w/2007/mar/business/rc_dupont.html. Saito, N., Sasaki, K., Nakatome, K., Harada, K., Yoshinaga, T., Vogt, J.V., et al., 2003. CCM river and catchment database for Europe, Koizumi, A., 2004b. Perfluorooctane sulfonate concentrations in Version 1.0. EC-JRC, EUR 20756 EN. surface water in Japan. Arch. Environ. Contam. Toxicol. 45, 149–158. Yamashita, N., Kannan, K., Taniyasu, S., Horii, Y., Okazawa, T., Schaefer, A., 2006. Perfluorinated surfactants contaminate German Petrick, G., Gamo, T., 2004. Analysis of perfluorinated acids at parts- waters. Environ. Sci. Technol. 40, 7108–7109. per-quadrillion levels in seawater using liquid chromatography-tandem Schulte, C., 2007. Newsletter 2 of the EU NORMAN project, URL http:// mass spectrometry. Environ. Sci. Technol. 38, 5522–5528. norman.ineris.fr/index_php.php?module=public/others/newsletter. Yamashita, N., Kannan, K., Taniyasu, S., Horii, Y., Petrick, G., Gamo, Schultz, M.M., Barofsky, D.F., Field, J.A., 2006. Quantitative determi- T., 2005. A global survey of perfluorinated acids in oceans. Marine nation of fluorinated alkyl substances by large-volume-injection liquid Poll. Bull. 51, 658–668.