Environmental occurrence and risk of organic UV filters and stabilizers in multiple matrices in Norway

Environment International 80 (2015) 1–7

Contents lists available at ScienceDirect

Environment International journal homepage: www.elsevier.com/locate/envint

Full length article

Environmental occurrence and risk of organic UV filters and stabilizers in multiple matrices in Norway Katherine H. Langford ⁎, Malcolm J. Reid, Eirik Fjeld, Sigurd Øxnevad, Kevin V. Thomas Norwegian Institute for Water Research (NIVA), Gaustadaléen 21, NO-0349 Oslo, Norway

a r t i c l e

i n f o

Article history: Received 21 October 2014 Received in revised form 6 March 2015 Accepted 14 March 2015 Available online xxxx Keywords: UV filters UV stabilizers Sunscreen Bioaccumulation Wastewater Landfill leachate

a b s t r a c t Eight organic UV filters and stabilizers were quantitatively determined in wastewater sludge and effluent, landfill leachate, sediments, and marine and freshwater biota. Crab, prawn and cod from Oslofjord, and perch, whitefish and burbot from Lake Mjøsa were selected in order to evaluate the potential for trophic accumulation. All of the cod livers analysed were contaminated with at least 1 UV filter, and a maximum concentration of almost 12 μg/g wet weight for octocrylene (OC) was measured in one individual. 80% of the cod livers contained OC, and approximately 50% of cod liver and prawn samples contained benzophenone (BP3). Lower concentrations and detection frequencies were observed in freshwater species and the data of most interest is the 4 individual whitefish that contained both BP3 and ethylhexylmethoxycinnamate (EHMC) with maximum concentrations of almost 200 ng/g wet weight. The data shows a difference in the loads of UV filters entering receiving water dependent on the extent of wastewater treatment. Primary screening alone is insufficient for the removal of selected UV filters (BP3, Padimate, EHMC, OC, UV-234, UV-327, UV-328, UV-329). Likely due in part to the hydrophobic nature of the majority of the UV filters studied, particulate loading and organic carbon content appear to be related to concentrations of UV filters in landfill leachate and an order of magnitude difference in these parameters correlates with an order of magnitude difference in the effluent concentrations of selected UV filters (Fig. 2). From the data, it is possible that under certain low flow conditions selected organic UV filters may pose a risk to surface waters but under the present conditions the risk is low, but some UV filters will potentially accumulate through the trophic food chain. © 2015 Published by Elsevier Ltd.

1. Introduction There is increasing concern over the release of chemicals originating from the use of personal care products (PCPs) such as pharmaceuticals, fragrances and preservatives, into the environment (Bu et al., 2013; Brausch and Rand, 2011). Due to the continuous release of PCPs from wastewater treatment works, they have been termed pseudopersistent, irrespective of their PBT (persistent, bioaccumulative and toxic) characteristics (Barceló, 2007). The occurrence of synthetic musk compounds (Chase et al., 2012; Lee et al., 2010) and siloxanes (Borgå et al., 2012) for example, are well studied, and there is growing interest in the occurrence of UV (ultraviolet) protective compounds also used in PCPs. The increase in public awareness of the dangers of overexposure to UV radiation has lead in an increase in the availability of UV protection products containing organic UV filters (Lodén et al., 2011). Organic UV protecting chemicals all absorb UV light (inorganic formulations reflect UV light) and in general can be loosely divided into 2 categories; UV filters used in PCPs to protect hair and cutaneous membranes from sun damage, and UV stabilizers used in technical ⁎ Corresponding author. E-mail address: [email protected] (K.H. Langford).

http://dx.doi.org/10.1016/j.envint.2015.03.012 0160-4120/© 2015 Published by Elsevier Ltd.

products such as plastics and paints to protect polymers and pigments against photodegradation and to prevent discoloring. Many of the compounds are used for both purposes and frequently used in combination to extend the UV protection range provided. The composition of cosmetic products in Europe is regulated by Directive 76/768/CEE (EC, 2009), and the maximum content of the compounds in PCPs is between 8 and 10% (e.g. 10% benzophenone (BP3), 10% octocrylene (OC), 8% Padimate and 10% ethylhexylmethoxycinnamate (EHMC)). The UV stabilizer polymer additives, such as the benzotriazoles; UV-234, -327, -328, and -329, are regulated by REACH in Europe. It is widely reported that UV filters used in PCPs enter the aquatic environment indirectly via sewage effluent discharges and directly from water sports activities causing them to wash directly from skin into receiving waters (Amine et al., 2012; Buser et al., 2006; Díaz-Cruz et al., 2008; Langford and Thomas, 2008; Fent et al., 2010a; Santos et al., 2012). The first reported environmental occurrence of an organic UV filter was over 30 years ago when a benzophenone was detected in the Baltic Sea (Ehrhardt et al., 1982), although PCPs were not identified as the source. UV filter occurrence can be season and weather dependent, higher concentrations were detected in wastewater influents in summer than in winter (Tsui et al., 2014) and receiving waters have demonstrated the same patterns of distribution with higher concentrations in

2

K.H. Langford et al. / Environment International 80 (2015) 1–7

summer than in autumn (Langford and Thomas, 2008). Previous studies have mainly focused on the occurrence of UV filters such as BP3, OC and EHMC and there are limited data available on the environmental occurrence of benzotriazole UV stabilizers in Europe. Certain UV filters are known to be toxic to aquatic organisms and exhibit effects consistent with endocrine disruption. Paredes et al. (2014) demonstrated that species from different trophic levels are more sensitive to some UV filters than to others, for example, microalgae was the most sensitive species studied, in particular to BP3 but sea urchin larvae were most sensitive to EHMC. The resulting risk quotients from this study concluded that BP3 poses a potential risk to aquatic organisms, but EHMC poses no environmental risk, although a study involving Daphnia concludes that EHMC does pose a potential environmental risk (Fent et al., 2010b). EHMC has demonstrated multiple hormone activities in fish with gene expression profiling showing antiestrogenic activity compared to estrogenic/antiandrogenic activity using vitellogenin induction (Fent et al., 2008; Christen et al., 2011). UV filter mixtures showed synergistic interaction in vitro and additive antagonistic activity in vivo (Fent et al., 2008; Kunz and Fent, 2009). BP3 showed potential antagonistic response after metabolic activation with a rat liver preparation but benzotriazoles showed no estrogenic activity (Morohoshi et al., 2005). The purpose of this study is to address the paucity of data on the occurrence of organic UV protective compounds, assess the environmental risk of these compounds and establish if they are pseudopersistent and released continuously into the Norwegian aquatic environment. Our study investigated the occurrence of 4 organic UV filters and 4 organic UV stabilizers (Table S1 shows structures and log Kow values): BP3, Padimate, EHMC, OC, UV-234 (2-(2H-benzotriazol2-yl)-4,6-bis(2-phenyl-2-propanyl)phenol), UV-327 (2-(5-chloro-2Hbenzotriazol-2-yl)-4,6-bis(2-methyl-2-propanyl)phenol), UV-328 (2-(2H-benzotriazol-2-yl)-4,6-bis(2-methyl-2-butanyl)phenol) and UV329 (2-(2H-benzotriazol-2-yl)-4-(2,4,4-trimethyl-2-pentanyl)phenol) in potential sources, receiving waters, and through different common species in a marine and a freshwater environment. For the purposes of this manuscript the term UV filters will be used for all organic UV protecting compounds.

ISI was in operation from 1972 until 2002. Leachate mixed with incoming groundwater, flows through a discharge tank downstream of the landfill and leachate is discharged to VEAS WWTW for treatment. Lindum is a fully operational landfill site and the leachate is heavily influenced by incoming groundwater, with high rainfall events significantly increasing leachate discharge rates. The leachate flows through an aerated lagoon with subsequent sedimentation prior to discharge to a WWTW (not included in this study).

2. Materials and methods

2.2. Analysis

2.1. Sample collection

All standards (BP3, Padimate, EHMC, OC, UV-234, UV-327, UV328, UV-329) and internal standards (BP-d10, naphthalene-d8, and chrysene-d12) and HPLC grade solvents were purchased from SigmaAldrich (Germany). Bulk primary secondary amine (PSA) sorbent (Supelco, SuperClean) was supplied by Sigma Aldrich (St. Louis MO, USA) and Hydromatrix was supplied by Varian.

Section 2 of the supplementary information contains detailed sample information. 2.1.1. Wastewater and sludge Automated 24 h composite effluent samples and 5 grab samples of sludge were collected on 5 consecutive days from 3 different wastewater treatment works (WWTWs). VEAS is Norway's largest WWTW receiving municipal wastewater from a population of 580,000. Approximately 100 million m3 of wastewater is treated annually by mechanical, chemical and biological (post-denitrification) treatment before the effluent is discharged into the Oslofjord at a depth of 50 m. Sludge is treated by anaerobic digestion and drying. HIAS receives wastewater from a population of 52,000. The wastewater is treated mechanically, biologically (no nitrogen removal) and chemically before it is discharged into Lake Mjøsa at a depth of 15 m, 250 m from the shore. Sludge is treated by thermal hydrolysis at 160 °C prior to anaerobic digestion at 38 °C. Tomasjord receives wastewater from a population of 38,400. The wastewater undergoes primary treatment before discharge into Tromsøysundet. 2.1.2. Landfill leachate Automated 24 hour composite samples were collected from 2 landfill sites over a period of 3 days.

2.1.3. Sediment Sediment samples were collected from the Oslofjord in a 5 point transect from the VEAS discharge point following the main current flow (Table S2). A stainless steel Van Veen grab was used to collect 3 subsamples of the top 2 cm of sediment at each location. Sediment samples were collected in a southerly transect from HIAS discharge point in Lake Mjøsa (Table S4). A stainless steel gravity corer with a core tube and a retractable sediment stopper was used to collect 3 subsamples of the top 2 cm of sediment at each location. 2.1.4. Biota Atlantic cod (Gadus morhua) and Northern shrimp (Pandalus borealis) were caught in the inner Oslofjord by trawling. Individual livers were removed from the Atlantic cod (n = 15) (Table S3). Shrimps were peeled and split into 15 bulk samples each comprising of 50–60 individuals. Common shore crab (Carcinus meanas) were caught north of VEAS WWTW by snorkeling. Tissue from 180 crabs was split into 15 bulk samples comprised of 10–13 crabs. Burbot (Lota lota), perch (Perca fluviatilis) and whitefish (Coregonus lavaretrus) were caught using gillnets deployed in the vicinity of the HIAS discharge point in Lake Mjøsa (Table S5). Soft tissue (homogenates of muscle, stomach, intestines and liver) samples from 15 individuals of each species were prepared for analysis. All sludge, sediment and biota samples were frozen within 4 h of collection and stored at − 20 °C until analysis. Wastewater samples were stored at 4 °C until analysis which was within 48 h of collection.

2.2.1. Sample preparation and extraction 2.2.1.1. Solid samples (biota, sediment and sludge). All solid samples were extracted by Accelerated Solvent Extraction (Dionex ASE 200 system, Sunnyvale CA, USA). Sediment and sludge samples were freeze dried prior to extraction and biota samples were extracted wet. Approximately 1 g of PSA was added to the ASE cells (22 ml) to aid the clean-up of fatty acids and other matrix interferences. Samples were mixed with hydromatrix sorbent to improve the solvent flow through the ASE cell and the mixture composed the second layer in the ASE cell. The ASE extraction solvent was hexane/dichloromethane (50/50, v/v) at a temperature of 100 °C. The static time was 5 min, and the purge time was 2 min with 3 static cycles. The ASE extracts were reduced to approximately 1 ml under a stream of nitrogen (35 °C) before further clean-up via Gel Permeation Chromatography (GPC). GPC was carried out on an Alliance 2695 system (Waters, Milford MA, USA) with two sequential Envirogel (Waters, Milford MA, USA) GPC clean-up columns (19 × 300 mm and 19 × 150 mm) and dichloromethane (DCM) as a mobile phase. The 12.1–20.0 minute fraction was collected.

K.H. Langford et al. / Environment International 80 (2015) 1–7

The GPC fraction was subsequently reduced to 2 ml under a stream of nitrogen (35 °C). PSA sorbent (approximately 100 mg) was added to each extract to further remove matrix interferences. Extracts were centrifuged (21,000 g, 10 min) and the supernatant was transferred to vials for analysis via LC-HRMS and GC-HRMS. Note that extracts for LC-HRMS were first solvent exchanged to acetonitrile by addition of acetonitrile before evaporation of remaining DCM. 2.2.1.2. Water samples. Wastewater samples (1 L) were extracted via solid phase extraction (SPE) on Oasis HLB (200 mg, 6 ml) cartridges (Waters Corp, Milford MA, USA). SPE cartridges were pre-washed with DCM (6 ml) and methanol (6 ml) before equilibration with water (12 ml) prior to sample-loading. SPE cartridges were eluted with 20 ml ethylacetate/DCM (50/50) and the eluent reduced to 2 ml under a stream of nitrogen (35 °C) and transferred to vials for analysis via LC-HRMS and GC-HRMS. Note that, as with solid phase samples, extracts for LC-HRMS were solvent exchanged to acetonitrile before injection on the LC system. 2.2.2. Sample analysis 2.2.2.1. GC-HRMS. Samples (1 μl) were injected into an Agilent gas chromatograph fitted with a 30 m × 0.25 mm, 0.25 μm film thickness DB5MS column (Agilent Technologies) with helium carrier gas. Splitless injection at 250 °C was used. The initial temperature of 60 °C was held for 2 min, followed by an increase of 15 °C/min to 120 °C, followed by 5 °C/min to 280 °C and held for 5 min. The high-resolution time-offlight mass spectrometer (GCT Premier, Waters Corp, Milford MA, USA) was operated in full scan positive electron impact mode with a scan range of 100–450 m/z. Accurate mass spectra to 3 decimal places were used for peak identification with an error threshold of 5 mDa. See Table S7 in supplementary information for details. 2.2.2.2. LC-HRMS. Analysis was carried out on an Acquity UPLC system with a Xevo G2-S QTOF mass spectrometer as detector (both UPLC and MS from Waters Corp, Milford MA, USA). Chromatography was performed on a Waters Acquity BEH C8 column (2.1 × 50 mm) running a 7 min gradient from 50% methanol in 10 mM ammonium acetate to 100% methanol.

3

Mass spectrometry was performed in positive electrospray mode (0.7 kV capilliary and 20 V cone). Data acquisition was in MSE mode with the low energy (LE) function having a 5 V collision, and the high energy (HE) function having a collision ramp from 15–45 V. The LE function provides accurate mass detection of the parent ions (MH+), while the HE function provides time-aligned accurate mass fragment information. See Table S7 in supplementary information for details. Method validation data, detection limits and extraction recovery details are presented in Table S6 in supplementary information. Limits of quantification were calculated by 9 × s/n, and were in the range 10–50 ng/g for biological samples, 4–10 ng/g for sediment and sludge, and 5–10 ng/L for effluent samples. 3. Results and discussion 3.1. Effluent and sludge Total UV filter concentrations in WWTW effluent were between 300 and 8900 ng/L with OC, BP3 and EHMC dominating the effluent streams (Table 1, Fig. 1). Total UV filter concentrations in sludge were however, 2 orders of magnitude higher, between 5 and 51 μg/g with OC, EHMC, UV-327 and UV-328 dominating. Lui et al. (2012) also noted higher concentrations of BP3 in effluent than in sludge and the reverse for OC with a negligible amount in the effluent relative to sludge concentrations. The same distribution pattern was observed at VEAS which has the most advanced treatment process in the present study. A difference in log Kow may in part explain this with OC (log Kow 7.3) more likely to be removed via sorption to sludge than BP3 (log Kow 3.8). The concentrations of OC in effluent from Tomasjord (median concentration 2 167 ng/L) were over an order of magnitude higher than in samples from VEAS or HIAS (median concentrations of 258 and 158 ng/L respectively) which both have more advanced treatment processes. There was no removal of OC observed in an Australian study (Lui et al., 2012) during primary treatment, but during secondary treatment some removal was observed which, in combination with the high concentrations detected in this study at Tomasjord, indicate that mechanical treatment alone is not sufficient for OC removal. In the present study, the concentrations of UV-327 determined in sludge from VEAS and HIAS (Table 1, Fig. 1) were an order of magnitude

Table 1 Concentrations (range and median) UV filters detected in wastewater effluent (ng/L) and sludge (ng/g dry weight), landfill leachate (ng/L) and sediment (ng/g dry weight). BP3

Padimate

EHMC

OC

UV-324

UV-327

UV-328

UV-329

Total

81–598 293 b10 – 104 10–438 233 824–2116 1218 148 374–1915 721 169

b5 – b4 – nd b5 – b10 – nd b5 – nd

b5 – 551–793 714 nd b5 – 2501–4689 – nd 4.3–37 10 3.1

181–538 258 3448–12661 6257 97 7 +- 227 158 26823–41610 35973 70 1701–6969 2167 561

b5 – b11 – nd b5 – b14 – nd 4.5–5.6 5 1.3

b10 – 30.4–77.1 44 nd b10 – 83.3–159.9 89 nd b10 – nd

b5 – b11 – nd b5 – b25 – nd b5 – nd

b5 – 1172–3075 1789 nd b5 – 1493–3303 2362 nd b5 – nd

330–1136 474 5215–16606 9414 168 89–596 460 32683–51020 42882 218 2741–8921 2838 794

Range Median Range Median

b10–372 18 32–646 114

b5 – b15 –

b5 – 26–85 81

b5–40 b5 2318–21160 10557

b5 – b5–19 b5

b10 – b10 –

b5 – b5 –

b5 – b5 –

Range Median Range Median

b5 – b5 –

b4 – b5 –

8.5–16.4 11.2 9.9–19.8 14.1

b7–82.1 b7 b7 –

b15 – b15 –

b4–8.1 4.8 b65 –

3.2–25.1 12.5 b25 –

b15 – b15 –

Wastewater treatment works VEAS Effluent (ng/L)

Range Median Sludge (ng/g) (dry wt.) Range Median Median daily loading (ng/day/1000 people) HIAS Effluent (ng/L) Range Median Sludge (ng/g) (dry wt.) Range Median Median daily loading (ng/day/1000 people) Tomasjord Effluent (ng/L) Range Median Median daily loading (ng/day/1000 people) Landfill leachate (ng/L) ISI Lindum

Sediment (ng/g) (dry wt.) Oslofjord Mjøsa

16.6–126.5 25.6 9.9–19.8 14.1

4

K.H. Langford et al. / Environment International 80 (2015) 1–7

Effluent

Concentration (ng/L)

BP3 7000

EHMC

6000

OC

5000

UV-234

4000 3000 2000 1000

0 1

2

3

4

5

1

VEAS

2

3

4

5

1

2

HIAS

3

4

5

Tomasjord

Sludge BP3 EHMC OC UV-327 UV-329

45000

Concentration (ng/g)

40000 35000

30000 25000 20000 15000 10000 5000 0 1

2

3

4

5

VEAS

1

2

3

4

5

HIAS

Fig. 1. UV filters detected in wastewater effluent and sludge on 5 consecutive days.

higher than those reported in WWTW sludge from China (Zhang et al., 2011) while BP3 measured in sludge from HIAS was 2 orders of magnitude higher. UV-326, -327 and -328 have been reported to be completely removed during treatment at a WWTW in Spain and over 30% removal was observed at another WWTW in Portugal (Carpinteiro et al., 2012). In the present study, benzotriazoles were not detected in any effluent samples with the exception of 3 samples from Tomasjord (4.6–5.6 ng/L) where only primary screening is implemented, UV-327 and UV-329 were however detected in the sludge from HIAS and VEAS where more advanced treatment is in place (sludge is not produced at Tomasjord). UV-327 was detected in the range 30.4– 159.6 ng/g and UV-329 was more than an order of magnitude higher (1172–3 075 ng/g). The total UV filter mean daily effluent loading (Table 1) at Tomasjord (794 ng/day/1000 people) was approximately 5 × higher than that of

VEAS or HIAS (168 and 218 ng/day/1000 people respectively), both WWTWs having more advanced treatment than Tomasjord, highlighting the potential for negative environmental impact of wastewater effluent discharge after only primary screening. Trace concentrations of EHMC and UV-234 were detected at Tomasjord where none was detected at VEAS or HIAS, again highlighting the potential impact of minimal treatment processes. The largest contribution to the effluent load at Tomasjord is OC. Kupper et al. (2006) observed very limited removal of OC during primary treatment, however secondary treatment resulted in some degradation of OC and partitioning of approximately 50% to the sludge phase. This is in line with results from the current study showing high effluent loading of OC at Tomasjord (561 ng/day/1000 people) where there is no secondary treatment, and lower effluent loadings at VEAS (97 ng/day/1000 people) and HIAS (70 ng/day/1000 people) but high concentrations in sludge (medians of 6 and 36 μg/g respectively).

K.H. Langford et al. / Environment International 80 (2015) 1–7

3.2. Leachate The concentration of UV filters in the leachate collected from the active landfill at Lindum was higher than those from ISI that has been closed for 11 years (Table 1). The total daily load of UV filters discharged from ISI was 71 mg/day compared to 45,080 mg/day from Lindum. The largest contribution to the loading to Lindum leachate was OC with a median concentration of almost 11,000 ng/L compared to b 5 ng/L from ISI. Other UV filters were also present in higher concentrations at Lindum than at ISI. BP3 was in the range b10–372 ng/L and 32– 646 ng/L at ISI and Lindum respectively and the same was observed for EHMC where none was detected at ISI and it was detected in all 3 samples from Lindum (26–85 ng/L). This may be partly explained by the suspended solids loading in the effluent streams. ISI was analysed as a total sample due to very low particulate content. Lindum however had very high particulate content and the aqueous and solid fractions were analysed separately, and data combined. 100% of BP3, EHMC and UV-234 were found in the particulate fraction, and 84–100% of OC was in the particulate fraction. All 4 UV filters have log Kow values greater than 4 so partitioning to solids would be expected. The dissolved organic carbon content (raw data not shown) of ISI effluent (32.5– 37.6 mg/L) was an order of magnitude lower than was measured at Lindum (376–565 mg/L) indicating a greater potential for compounds with high log Kow values to sorb to the solid phase at Lindum compared to ISI. The higher particulate loading and associated UV filters could be expected from Lindum which is still in operation, compared to ISI which has been out of operation for over 10 years. 3.3. Sediment All of the sediments from the receiving waters contained organic UV filters, with EHMC being detected in all samples (Table 1). UV-328 (3.2–25.1 ng/g) was the dominant benzotriazole found in Oslofjord sediment and has been previously reported in sediments in Japan (6.3 ± 4 ng/g) (Nakata et al., 2009a). UV-327 was also detected in the Oslofjord but at lower concentrations (b 4–8.1 ng/g) and with less frequency (4 out of 5). Sediments collected from Lake Mjøsa contained only EHMC (9.9–19.8 ng/g) above the detection limit. 3.4. Biota In the marine biota collected from the Oslofjord, the concentrations for most of the analysed UV filters were typically below LoQ (Table 2), and only BP3 in shrimp and cod liver, and OC in cod liver, had concentrations above LoQ for greater than approximately half of the samples (Fig. 3). Padimate was not detected in any biota sample (n = 45) in the present study with the exception of 1 individual cod liver (37 ng/g). BP3 was detected in 6 out of 15 cod livers in the range b 30–1037 ng/g, and in 8 out of 15 prawns in the range b 30–69 ng/g. The median concentrations of b20 for cod liver does not show a true picture of the concentrations measured and is lower than expected. However the 90th percentile values for cod liver of 700 ng/g may be more representative.

5

Due to the low detection frequency, any meaningful conclusion of their relationship with trophic level is not possible although the concentration ranges and 90th percentile values (700 ng/g for cod liver and 62 ng/g for prawn) do indicate the possible biomagnification of BP3. OC was detected in 80% of the cod liver samples but only 1 of the 15 prawn samples so it is not possible to suggest a similar trophic accumulation for OC as for BP3. The concentration range of OC in cod liver was b20–11875 ng/g. The median value was 115 ng/g, and the 90th percentile was 1500 ng/g which is a factor of 8 lower than the maximum concentration, suggesting that the individual sample with a concentration of almost 12 μg/g should perhaps be considered an outlier. Removal of this value reduces the median concentration of OC in cod liver to 86 ng/g. Biomagnification has been observed in other studies (Fent et al., 2010a) and must be considered as a possibility in the present study despite the limited data set. Uptake of 6 different UV filters (BP3, Padimate, EHMC, OC, UV-327 and UV-328) by 4 different species (cod, prawn, whitefish and perch) was observed. BP3 and OC are discharged from VEAS (Table 1) into the Oslofjord so cod can be exposed via water as well as through food. No cod livers were free from all measured UV filters and the median concentration for total UV filters was 250 ng/g (90th percentile; 1700 ng/g) and even with the removal of the outlying individual liver, the median concentration was 216 ng/g. Selected individuals contained low concentrations (less than 40 ng/g in each case) of EHMC or UV-328 and 1 of the samples containing UV-328, also contained 21 ng/g Padimate. In addition to the BP3 and OC detected in prawns, 52 ng/g of UV-327 was detected in one sample. None of the UV filters were detected in any crab samples. Filter feeding and sediment dwelling organisms contained elevated concentrations in Asian marine waters, indicating sorption to particulates is a likely sink for some benzotraizole UV stabilizers. UV-326 (not included in the present study), UV-327 and UV-328 dominated higher trophic level species, such as shark and porpoise (Nakata et al, 2009a,b) and here we report comparable data in cod from the inner Oslofjord. The occurrence of UV filters in freshwater species in Lake Mjøsa demonstrates a different pattern to the marine environment and detection frequency and concentrations were lower (Table 2). Only, BP3, EHMC and OC were measured above detection limits in any species. The UV filters were not detected in any of the burbot while BP3, EHMC and OC were detected in 3 individual perch samples at concentrations of 6.5, 36 and 2.1 ng/g respectively. Perhaps an interesting observation is the correlation between BP3 and EHMC in whitefish. BP3 and EHMC were both detected in the same 4 individuals, it is not possible to draw any conclusions from such limited occurrence (4 out of 15 samples) but the correlation is nonetheless worthy of note, particularly since the ratio of the 2 compounds is similar in each case, with BP3 being the dominating compound at approximately double the concentration of EHMC. This pattern could be an indication of the same source of contamination. The median concentration was less than detection limits for both compounds but the maximum concentration was found in the same individual, 182 and 117 ng/g (wet wt.) for BP3 and EHMC respectively

Table 2 UV filter concentrations in marine and freshwater biota (ng/g) with detection frequencies (n) in 15 samples of each matrix. Cod (liver)

BP3 Padimate EHMC OC UV-324 UV-327 UV-328 UV-329

Prawn (whole)

Crab (whole)

Burbot (filet)

Whitefish (filet)

Perch (filet)

Range

Median

n

Range

Median

n

Range

Median

n

Range

Median

n

Range

Median

n

Range

Median

n

b20–1037 b20–21.3 b30–36.9 b20–11875 b10 b50 b10–19.5 b25

b20 b20 b30 115 b10 b50 b10 b25

7 1 3 12 0 0 3 0

b30–68.9 b20 b20 b10–23.1 b10 b10–51.8 b10 b25

45.2 – – b10 – b10 – –

8 0 0 1 0 1 0 0

b30 b20 b10 b10 b10 b10 b10 b25

– – – – – – – –

0 0 0 0 0 0 0 0

b5 b20 b5 b2 b10 b50 b10 b25

– – – – – – – –

0 0 0 0 0 0 0 0

b20–182 b20 b5–117 b2 b10 b50 b10 b25

b20 – b5 – – – – –

4 0 4 0 0 0 0 0

b5–6.5 b20 b5–35.7 b2–2.1 b10 b5 b10 b25

b5 – b5 b2 – – – –

1 0 1 1 0 0 0 0

6

K.H. Langford et al. / Environment International 80 (2015) 1–7 90

Concentration (ng/g dry wt)

80 70 60 EHMC 50

OC

40

UV-327

30

UV-328

20 10 0

Fig. 2. Concentrations of UV filters in sediment (ng/g dry weight).

and as with cod liver data, 90th percentiles may be a better representation at 76 and 38 ng/g, respectively. In a survey of the occurrence of UV filters in rivers in Switzerland in 2006–2007, Fent et al. (2010a) report of average concentrations of EHMC 50–170 ng/g lipid in muscle samples of fish (Chub, Leuciscus cephalus; barb, Barbus barbus; brown trout; Salmo trutta) and a total range of b LOD − 337 ng/g lipid. Biomagnification between Gammarus and chub, and chub and barb to cormorant, is suggested. In comparison, using the average lipid content of 1.5% (raw data not shown) of whitefish in the present study, the median concentration of EHMC was 170 ng/g of lipid, and the 90th percentile 2500 ng/g of lipid which is an order of magnitude higher than the concentrations detected in Switzerland.

4. Potential risk Using the measured environmental concentrations (MEC) for effluent and leachate predicted environmental concentrations (PEC) were calculated for receiving surface waters using a dilution factor of 10. Published PNEC (predicted no effect concentration) from ECHA are available for BP3, EHMC, OC and UV-329. The surface water PECs were compared to the published PNECs with the resulting risk quotients b 1 for all of the detected UV filters (Table 3). Based upon the assessment of the individual chemicals, the risk to the recipient from these chemicals is low. It is difficult to generalize with regard to the environmental risks associated with the levels detected in sludge since it is very much dependent on how the sludge will be used or disposed of.

Fig. 3. Concentrations of UV filters in cod liver from the Oslofjord (ng/g wet weight).

K.H. Langford et al. / Environment International 80 (2015) 1–7

7

Table 3 Predicted no effect concentrations (PNEC) for selected UV filters. Compound

Predicted no-effect concentration (PNEC) (ng/L) Effluent (ng/L)

BP3 EHMC OC UV-327 UV-329

Leachate (ng/L)

PEC/PNEC Surface water (ng/L)

Effluent (ng/L)

Median

Max.

Median

Max.

Marine/freshwater

381

1915

195

372

258

6969

40

380

670/6700⁎ 1000/10,000⁎ ⁎2300

1 × 106⁎ 1 × 106⁎ 1 × 106⁎

⁎1 × 104/1 × 105

1 × 106⁎

Surface waterLeachate

Surface waterEffluent

Effluent

Median

Max.

Median

Max.

Median

Max.

0.06

0.29

0.03

0.06

0.00004

0.0019

0.01

0.3

0.002

0.02

0.00003

0.007

⁎ PNEC (ng/L) taken from European Chemicals Agency (ECHA; www.echa.europa.eu). Surface water PEC calculated from MEC using a dilution factor of 10.

A detailed risk assessment based on different scenarios would be required. BP3, EHMC, OC, UV-327 and UV-329 were all detected in sludge, however it is not possible to evaluate whether they pose a risk to soil dwelling organisms when sludge has been applied to land. The occurrence of BP3, EHMC, ODBAPA, OC, UV-328 and UV-237 in selected biota samples suggests the potential to bioaccumulate. The accumulation of EHMC, OC, UV-327 and UV-328 was also observed in sediments. Acknowledgments The authors gratefully acknowledge the financial support from the Norwegian Environment Agency (Report M-176/2014). Christian Vogelsang, Merete Schøyen, Janne Gitmark and Pawel Krzeminski are acknowledged for sample collection, and Kine Bæk for her administrative role. Appendix A. Supplementary data The supporting information contains information on compound structures, sampling methodology and sample locations, and analysis methodology including method performance details. Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.envint.2015.03.012. References Amine, H., Gomez, E., Halwani, J., Casellas, C., Fenet, H., 2012. UV filters, ethylhexyl methoxycinnamate, octocrylene and ethylhexyl dimethyl PABA from untreated wastewater in sediment from eastern Mediterranean river transition and coastal zones. Mar. Pollut. Bull. 64, 2434–2442. Barceló, D., 2007. Pharmaceutical residue analysis. TrAC 26, 454–455. Borgå, K., Fjeld, E., Kierkegaard, A., Mclachlan, M.S., 2012. Food web accumulation of cyclic siloxanes in Lake Mjøsa. Environ. Sci. Technol. 46, 6347–6354. Brausch, J.M., Rand, G.M., 2011. A review of personal care products in the aquatic environment: environmental concentrations and toxicity. Chemosphere 82, 1518–1532. Bu, Q., Wang, B., Huang, J., Deng, S., Yu, G., 2013. Parmaceuticals and personal care products in the aquatic environment in China: a review. J. Hazard. Mater. 262, 189–211. Buser, H.R., Balmer, M.E., Schmid, P., Kohler, M., 2006. Occurrence of UV filters 4methylbenzylidene camphor and octocrylene in fish from various Swiss rivers with inputs from wastewater treatment plants. Environ. Sci. Technol. 40, 1427–1431. Carpinteiro, I., Ramil, M., Rodríguez, L., Nogueira, J.M.F., 2012. Combining stir-bar sorptive extraction and large volume injection-gas chromatography–mass spectrometry for the determination of benzotriazole UV stabilizers in wastewater matrices. J. Sep. Sci. 35, 459–467. Chase, D.A., Karnjanapiboonwong, A., Fang, Y., Cobb, G.P., Morse, A.N., Anderson, T.A., 2012. Occurrence of synthetic musk fragrances in effluent and non-effluent impacted environments. Sci. Total Environ. 416, 253–260. Christen, V., Zucchi, S., Fent, K., 2011. Effects of the UV-filter 2-ethyl-hexyl-4trimethoxycinnamate (EHMC) on expression of genes involved in hormonal path-

ways in fathead minnows (Pimephales promelas) and link to vitellogenin induction and histology. Aquat. Toxicol. 102, 167–176. Díaz-Cruz, S., Llorca, M., Barceló, D., 2008. Organic UV filters and their photodegradates, metabolites and disinfection by-products in the aquatic environment. TrAC 27, 873–887. EC, 2009. Regulation (EC) No 1223/2009 of the European Parliment and of the Council of 30 November 2009 on Cosmetic Products. Ehrhardt, M., Bouchertall, F., Hopf, H.P., 1982. Aromatic ketones concentrated from Baltic Sea water. Mar. Chem. 11, 449–461. Fent, K., Kunz, P.Y., Gomez, E., 2008. UV filters in the aquatic environment induce hormonal effects and affect fertility and reproduction in fish. Chimia 62, 368–375. Fent, K., Zenker, A., Rapp, M., 2010a. Widespread occurrence of estrogenic UV-filters in aquatic ecosystems in Switzerland. Environ. Pollut. 158, 1817–1824. Fent, K., Kunz, P.Y., Zenker, A., Rapp, M., 2010b. A tentative environmental risk assessment of the UV-filters 3-(4-methylbenzylidene-camphor), 2-ethylhexyl-4-trimethoxycinnamate, benzophenone-3, benzophenone-4 and 3-benzylidene camphor. Mar. Environ. Res. 69, S4–S6 (Suppl.). Kunz, P.Y., Fent, K., 2009. Estrogenic activity of ternary UV filter mixtures in fish (Pimephales promelas) — an analysis with nonlinear isobolograms. Toxicol. Appl. Pharmacol. 234, 77–88. Kupper, T., Plagellat, C., Brandli, R.C., de Alencastro, L.F., Grandjean, D., Tarradellas, J., 2006. Fate and removal of polycyclic musks, biocides and UV filters during wastewater treatment. Water Res. 40, 2603–2612. Langford, K.H., Thomas, K.V., 2008. Inputs of chemicals from recreational activities into the Norwegian coastal zone. J. Environ. Monit. 10, 894–898. Lee, I.S., Lee, S.H., Oh, J.E., 2010. Occurrence and fate of synthetic musk compounds in the water environment. Water Res. 44, 214–222. Lodén, M., Beitner, H., Gonzalez, H., Edström, D.W., Åkerström, U., Austrad, J., Buraczewska-Norin, I., Matsson, M., Wulf, H.C., 2011. Sunscreen use: controversies, challenges and regulatory aspects. Brit. J. Dermatol. 165, 255–262. Lui, Y.S., Ying, G.G., Shareef, A., Kookana, R.S., 2012. Occurrence and removal of benzotriazoles and ultraviolet filters in a municipal wastewater treatment plant. Environ. Pollut. 165, 225–232. Morohoshi, K., Yamamoto, H., Kamata, R., Shiraishi, F., Koda, T., Morita, M., 2005. Estrogenic activity of 37 components of commercial sunscreen lotions evaluated by in vitro assays. Toxicol. in Vitro 19, 457–469. Nakata, H., Murata, S., Filatreau, J., 2009a. Occurrence and concentrations of benzotriazole UV stabilizers in marine organisms and sediments from the Ariake Sea, Japan. Environ. Sci. Technol. 43, 6920–6926. Nakata, H., Murata, S., Shinohara, R., Filatreau, J., Isobe, T., Taskahashi, I., Tanabe, S., 2009b. Occurrence and concentrations of persistent personal care products, organic UV filters, in the marine environment. In: Obayashi, Y., Isobe, T., Subramanian, A., Suzuki, S., Tanabe, A. (Eds.), Interdisciplinary Studies on Environmental Chemistry—Environmental Research in Asia. Terrapub, Japan, pp. 239–246. Paredes, E., Perez, S., Rodil, R., Quintana, J.B., Beiras, R., 2014. Ecotoxicological evaluation of four UV filters using marine organisms from different trophic levels Isochrysis galbana, Mytilus galloprovincialis, Paracentrotus lividus, and Siriella armata. Chemosphere 104, 44–50. Santos, A.J.M., Miranda, M.S., Esteves da Silva, J.C.G., 2012. The degradation products of UV filters in aqueous and chlorinated aqueous solutions. Water. Res. 46, 3167–3176. Tsui, M.M.P., Leung, H.W., Lam, P.K.S., Murphy, M.B., 2014. Seasonal occurrence, removal efficiencies and preliminary risk assessment of multiple classes of organic UV filters in wastewater treatment plants. Water. Res. 53, 58–67. Zhang, Z., Ren, N., Li, Y.F., Kunisue, T., Gao, D., Kannan, K., 2011. Determination of benzotriazole and benzophenone UV filters in sediment and sewage sludge. Environ. Sci. Technol. 45, 3909–3916.