ARTICLE pubs.acs.org/est
Spatially Detailed Survey on Pollution by Multiple Perfluorinated Compounds in the Tokyo Bay Basin of Japan Yasuyuki Zushi,† Feng Ye,† Mamoru Motegi,‡ Kiyoshi Nojiri,‡ Shigeo Hosono,‡ Toshinari Suzuki,§ Yuki Kosugi,§ Kumiko Yaguchi,§ and Shigeki Masunaga*,† †
Graduate School of Environment and Information Sciences, Yokohama National University, 79-7 Tokiwadai, Hodogaya, Yokohama 240-8501, Japan ‡ Center for Environmental Science in Saitama, 914 Kamitanadare Kazo, Saitama Pref. 347-0115, Japan § Tokyo Metropolitan Institute of Public Health, 3-24-1 Hyakuninchou Shinjyuku Tokyo Pref. 169-0073, Japan
bS Supporting Information ABSTRACT: Pollution from 35 perfluorinated compounds (PFCs) in the water of the Tokyo Bay basin was examined. The water in the basin contained relatively high levels of perfluorononanoate (PFNA), perfluorooctanoate (PFOA), and perfluorooctane sulfonate (PFOS) compared to the other PFCs, which were present at concentrations of 20.1 ng/L, 6.7 ng/L, and 5.8 ng/L, respectively. In contrast, the concentrations of their precursors and degradation products were an order of magnitude lower. Sewage treatment plant (STP) effluent in the area also contained high levels of PFNA compared with the river water samples (Mann-Whitney U-test, p < 0.0002). From a spatial aspect, increases in PFC pollution levels correlated with increased urbanization in the study area suggested that there are nonpoint source contributors to the PFC pollution in this area. Branched isomers of the PFCs were also quantified. Samples that contained high concentrations of perfluoroalkyl carboxylates (PFCA) showed lower proportions of its branched isomer. This indicates that the branched isomers are more prominent in the area with lower PFC pollution. This analysis was beneficial for estimating the individual contributions of different PFCA production processes. This survey provided new information on the sources, spatial distribution, and behavioral characteristics of PFC pollutants in this area.
’ INTRODUCTION Perfluorooctane sulfonate (PFOS) and its synthetic starting material, perfluorooctyl sulfonyl fluoride (PFOSF), were designated as persistent organic compounds (POPs) by the Stockholm Convention on POPs in May 2009.1 PFOS, its related compounds such as perfluorooctanoate (PFOA) and perfluorononanoate (PFNA), and their precursors are called perfluorinated compounds (PFCs). These PFCs have attractive properties, such as interfacial activity, resistance to acid and high temperatures, and water and oil repellency, for industrial applications. Consequently, they have been used in industry for over 50 years. This has led to PFC contamination in various environmental matrices. PFOS and PFOA were first recognized as global pollutants in 2001,2 and since then pollution by PFCs has been found in river and ocean water,3,4 sediments,5 wildlife,6 and the human body.7 Restrictions for the production and use of PFCs have been introduced, such as phasing out the production of PFOS, its precursors (i.e., perfluorooctane sulfonamides (FOSAs), perfluorooctane r 2011 American Chemical Society
sulfonamidoethanols (FOSEs)), and PFOSF by 2003.8 PFOA, precursor chemicals that can break down to PFOA, and related higher homologues are regulated by the United States Environmental Protection Agency (US EPA).9 Moreover, several other PFC regulations have been introduced, such as water guidelines by several institutes,10,11 the EU directive from 2008,12 and the Stockholm Convention on POPs from 2009.1 However, PFC pollution has still been reported after the implementation of these regulations.5,13-15 It has been reported that PFOS concentrations are decreasing in human serum.16 However, despite the introduction of strict PFOS regulations, the human serum levels of PFOS did not significantly decrease from 1994-2007/2008 in Busan and Seoul, Korea.13 Moreover, the serum PFOA level significantly Received: July 28, 2010 Accepted: February 14, 2011 Revised: February 2, 2011 Published: March 08, 2011 2887
dx.doi.org/10.1021/es103917r | Environ. Sci. Technol. 2011, 45, 2887–2893
Environmental Science & Technology
ARTICLE
’ MATERIALS AND METHODS
Figure 1. Map of Japan and the study area. Shaded areas represent the Tokyo Bay basin.
increased in Busan from 2000-2008 and in Seoul from 1994-2007. The continuation of PFC pollution after the introduction of PFC regulations may arise from the consumption of stockpiles of PFCs, PFC emissions from commercial distribution of PFC-containing products that were manufactured predating the regulations,4,17 and/ or gradual degradation of acrylate polymers with fluoroalkyl side chains.18 In addition, PFC production, including that of PFOS, in developing countries such as China is still increasing and probably contributes to continued PFC pollution.19 The scientific literature illustrates how difficult it is to reduce PFC pollution, especially in developing countries. In addition, clear decreasing trends have not been observed for perfluoroalkyl carboxylates (PFCAs) in developed countries, and it seems to be difficult to reduce PFCAs even in developed countries. For example, the levels of some PFCAs, which have historically (1974-2007) been measured for PFC evaluations, did not clearly decrease in falcon egg samples.14 Although there are a number of suspected sources of continued PFC pollution, the exact cause has not been clearly explained. The lack of information on PFC sources, their contribution to pollution, and the level/fate of PFCs and their precursors in the environment makes it difficult to evaluate current trends in PFC pollution. Thorough spatially and analytically intensive surveys of PFC pollution are required for full comprehension of PFC pollution and development of appropriate controls and management solutions. PFC pollution analysis is complex because of the many different PFCs. To thoroughly evaluate PFC pollution, PFOS, PFOA, and their precursors need to be included in spatial distribution and behavioral characterization. In this study, we collected river water and sewage treatment plant (STP) effluent samples from strategic locations in the Tokyo Bay basin. These samples were used to evaluate the current status, the spatial distribution, and recent trends in PFC pollution in the Tokyo Bay basin.
Study Area. The Tokyo Bay basin (Figure 1) contains the city of Tokyo, which is one of the most industrialized, urbanized, and populated areas in the world. The total area of the basin is approximately 8000 km2 and contains a population of 29 million. There are six major rivers flowing into the bay. Water from these six rivers makes up most of the total river water discharge into Tokyo bay. The basin mainly consists of the area containing Tokyo, Kanagawa (Southwest in Figure 2), Saitama (Center and North), and Chiba prefectures (East). Detailed information is given in the Supporting Information. Sample Collection. One liter samples of river water and STP effluent were collected in the Tokyo Bay basin between April 3 and May 1, 2009. Samples were collected on days when no rainfall had occurred for two consecutive days. Most of the river water samples were collected from the downstream end of the river in each watershed (n = 50), and these samples should represent the water quality of each watershed. One STP effluent sample was collected at each of 6 STPs in Saitama. The STP samples were mainly composed of domestic sewage and were collected for comparison of their PFC concentrations with those in the river water samples. In addition, six samples were collected from the downstream areas of the six major rivers flowing into Tokyo Bay. These samples were used to estimate the loading of PFCs into the bay. A map of the Tokyo Bay basin and the sampling locations is shown in Figure S-1. Other detailed information is also given in the Supporting Information. Sample Preparation. The water sample extraction method was adapted from a previously reported method.20 The pH of each water sample (500 mL) was adjusted to 4 before extraction by adding 2 mL of 0.5 mol/L tetrabutylammonium hydrogen sulfate (TBA). A small amount of 4 mol/L HCl was added if the pH did not reach 4 after TBA addition. Suspended solids were removed by a glass fiber filter and stored in a freezer at -20 °C until analysis. The surface of any equipment that touched the water sample was washed with 20 mL of methanol to reduce loss of the target compounds by sorption onto the surface. The methanol washings were then combined with the corresponding sample. After that, the samples, which were spiked with 2 ng of labeled internal standards (ISs), were loaded onto a solid phase extraction (SPE) cartridge (150 mg/6 cm3) (WAX, Waters Corp., Milford, MA) by a Sep-Pak concentrator (Waters Corp., Milford, MA) to extract the target PFCs. Three aliquots of 50% aqueous methanol (10 mL) were passed through the cartridge to wash the used materials and the cartridge. The loaded cartridges were then dried and stored at -20 °C until analysis. The target compounds were eluted with 7 mL of 1% ammonia in methanol, and the extract was dried under a gentle stream of nitrogen gas. The residue was dissolved in 200 μL of methanol and passed through a nylon membrane Millex filter unit (pore diameter 0.2 μm, Millipore, Billeirica, MA). The extracted sample was transferred to a polypropylene vial. Instrumental Analysis. The PFCs in the extract were analyzed by liquid chromatography tandem mass spectrometry (LC-MS/MS). Detailed information is given in the Supporting Information. The mobile phase was 10 mmol/L ammonium acetate/methanol at a flow rate of 200 μL/min. The mobile phase gradient began at 15% methanol, was increased to 62% methanol after 5 min, maintained for 10 min, increased to 65% methanol, and then maintained at this level for 10 min. The gradient was then increased to 70% methanol, maintained for 6 2888
dx.doi.org/10.1021/es103917r |Environ. Sci. Technol. 2011, 45, 2887–2893
Environmental Science & Technology
ARTICLE
Figure 2. Spatial distribution of PFCs and the characteristics of PFC composition in the Tokyo Bay basin. The numbers represent the total concentrations of measured PFCs (ng/L). Samples with concentrations below the LOQ were given a value of half of the LOQ, but in this map they are represented as “0” for better visibility. See Table S-1 for abbreviation of each of the PFCs.
min, increased to 100%, and maintained for 5 min. Finally, the gradient was returned to its initial conditions. Forty-five PFCs (Tables S-1 and S-2) were analyzed by LC-MS/MS. The branched isomers of PFOA, PFNA, perfluorodecanoate (PFDA), perfluoroundecanoate (PFUnDA), and PFOS (PFOAisomer, PFNAisomer, PFDAisomer, PFUnDAisomer, and PFOSisomer, respectively) were quantified separately. Standards for some branched PFCA isomers were not available at the time of this study except for isopropyl PFNA. Consequently, the signal response ratios of PFNA to isopropyl PFNA (Wellington Inc., Ontario, Canada) (1:0.45) were used for the determination of the sum of the branched PFCA isomer concentrations. The branched isomers of PFOS were separated into three peaks by HPLC. The first eluted fragment was called PFOSisomer1, the second eluted fragment PFOSisomer2, and the final eluted fragment was identified as linear PFOS. The resolution (Rs), which acts as an index of the extent of the separation of two peaks, between the linear and branched isomers was 0.99, 1.29, 1.46, 1.37, 0.70, and 0.96 for PFOA, PFNA, PFDA, PFUnDA, PFOS (PFOSisomer1-PFOSisomer2), and PFOS (PFOSisomer2PFOS), respectively. Further details concerning the Rs and the quantification of the PFC isomers are described elsewhere.5 QA/QC. Thirty-five of the 45 PFCs of interest (see the Supporting Information for PFC abbreviations) had recoveries
between 50-120% (n = 3, Table S-2). Although the recovery of PFNA was greater than 120% because of high levels of PFNA in the recovery test sample, the results for PFNA were satisfactory based on the level of recovery observed for PFOA and PFDA. The result for the procedural blank was subtracted from the measured concentrations for the water samples. The LOQs for the compounds are presented in Table S-3. Detailed information is given in the Supporting Information.
’ RESULTS AND DISCUSSION Spatial Distributions and Levels of PFCs in the Tokyo Bay Basin. Thirty-five PFCs were quantified in the collected samples.
At each sampling location, the concentration of at least one of these PFCs was above the LOQ in the collected sample (Table S-3), except for one sample that was collected upstream of the Ara river (6S, Figure S-1). The PFC pollution in the rivers of the Tokyo Bay basin is illustrated on the map in Figure 2. High concentrations of PFNA (median 20.1 ng/L), PFOA (6.7 ng/L), and PFOS (5.8 ng/L) (Figure 3) were found in the river water samples collected downstream of the watershed (n = 50). Spatial variation was observed in the patterns of dominant compounds. PFOS and PFNA were dominant in the Tokyo/Kanagawa area, PFOA in the Chiba area, and PFNA in the Saitama area. These 2889
dx.doi.org/10.1021/es103917r |Environ. Sci. Technol. 2011, 45, 2887–2893
Environmental Science & Technology
ARTICLE
Figure 3. Box plots of PFC concentrations in the collected samples in the Tokyo Bay basin. The samples collected from the downstream edge of the watershed (n = 50) are plotted. STP effluents (n = 6) overlap in the box plots. Whiskers show 1.5 IQR (Interquartile range).
results indicate that there are regional differences in the use and emission of PFCs in the Tokyo Bay basin. Overall, PFNA was the most prevalent PFC in the Tokyo Bay basin. This result differs from many earlier reports and is unexpected because PFOS and PFOA are considered the two most industrially utilized PFCs. The prevalence of PFNA might indicate a shift in PFC use because of the introduction of restrictions, as discussed in the Introduction. A recent report indicated that PFNA has been produced selectively in industrial production of PFCA in Japan,21,22 and this might have contributed to the high PFNA levels in the Tokyo Bay basin observed in the present study. The STP effluents also contained high levels of PFNA compared with the river water samples (Mann-Whitney U test; p < 0.0002). Other PFCs, such as PFHpA (p < 0.005), PFOA (p < 0.002), PFDA (p < 0.01), PFOAisomer (p < 0.005), and PFNAisomer (p < 0.002), were also significantly high in STP effluents (Figure 3). In contrast, perfluoroalkyl sulfonates and its precursors showed no significant differences. The relatively high concentration of PFC in STP effluent compared with river water was specific to PFNA. Therefore, PFNA emissions through STPs might have contributed to the river water pollution observed in this study. The similarity in PFC compositions between river water and STP effluent collected from the Saitama area (Figure 2) might also be a reflection of the contribution of STP effluent to river water pollution. The concentrations of PFCAs, such as PFHxA, PFHpA, PFOA, and PFNA, were between one and 2 orders of magnitude higher near the two plants that produce these materials than in other areas. The total PFCA concentrations in coastal water and plant effluent in the Chiba area were 491.1 and 6024.0 ng/L, respectively (Figure 2). These extraordinarily high concentrations of PFCAs might contribute to the PFCA pollution seen in Tokyo Bay. Our observations are consistent with an earlier study, in which the inner section of Tokyo Bay was spatially divided into 20 blocks. This study found that the PFOA concentration was
higher in seawater around the coast of Chiba (25 ng/L in the upper water layer) than in the upper water layer of Tokyo Bay (median 16 ng/L).23 By contrast, the PFOS concentration near the Chiba coast was not high (5.5 ng/L in the upper water layer) within these monitoring blocks compared with the concentration of PFOS in Tokyo Bay (median 5.5 ng/L in the upper water layer).23 The consistency of results from the survey of Tokyo bay and its basin supports the suggestion that PFCA producing plants made large contributions to the total PFCA pollution in the area. The spatial trends in the pollution patterns (Figure 2) suggested that there were nonpoint source contributions to the PFC pollution in the area. PFC pollution was prominent in the areas where buildings were dominant. This is consistent with our previous report, where nonpoint source PFC pollution was identified in areas with dense commercial and transportation activities.4 Statistical analysis will be conducted in the future to identify these sources. PFCAs with perfluoroalkyl chains longer than 10 carbons (PFUnDA) were rarely detected in the river water or STP effluent samples. The median concentrations of PFDoDA, PFTrDA, PFTeDA, PFPeDA, PFHxDA, and PFHpDA were
