Distribution and Behavior of Nonylphenol, Octylphenol, and

Feb 10, 2001 - and bay sediments in the Tokyo metropolitan area were demonstrated. During ... Greater removal for NP (93% on average) than OP (84% on ...
0 downloads 0 Views 142KB Size
Download PDF
Environ. Sci. Technol. 2001, 35, 1041-1049

Distribution and Behavior of Nonylphenol, Octylphenol, and Nonylphenol Monoethoxylate in Tokyo Metropolitan Area: Their Association with Aquatic Particles and Sedimentary Distributions TOMOHIKO ISOBE, HAJIME NISHIYAMA, ARISA NAKASHIMA, AND HIDESHIGE TAKADA* Faculty of Agriculture, Tokyo University of Agriculture & Technology, Fuchu, Tokyo 183-8509, Japan

Distributions of alkylphenols (APs) [i.e., nonylphenol (NP), octylphenol (OP)], and nonylphenol monoethoxylate (NP1EO) in wastewater effluents, river water, and riverine and bay sediments in the Tokyo metropolitan area were demonstrated. During sewage treatments, NP and OP were efficiently removed from the sewage effluents through activated sludge treatments. Greater removal for NP (93% on average) than OP (84% on average) was consistent with their partitioning behavior to particles in primary and secondary effluents. NP concentrations in the river water samples ranged from 0.051 to 1.08 µg/L with higher concentrations in summer and spring than in colder seasons. In the river water samples, ∼20% of NP was found in the particulate phase. Organic carbon-normalized apparent partition coefficients (K′oc) for NP (105.22 ( 0.38) and OP (104.65 ( 0.42) were 1 order of magnitude higher than those expected from their octanol-water partition coefficients (Kow), indicating strong affinity of APs to aquatic particles. Among NP isomers, no significant differences in their K′oc values were suggested. This is consistent with surprisingly uniform isomer peak profiles among the technical standard and all the environmental samples analyzed. NP and OP were widely distributed in the river sediments in Tokyo, and relatively high concentrations (0.5-13.0 µg/g dry) of NP were observed in a long reach (∼10 km) in the Sumidagawa River. In situ production of APs in the river sediment was suggested. Seaward decreasing trend in APs concentration was observed from the estuary to the Tokyo Bay. APs were well preserved in a sediment core collected from the bay. The profile shows subsurface maximum of AP concentrations in the layer deposited around the mid-1970s. The recent decrease in AP concentrations can be attributed to the legal regulation of industrial wastewater in the early 1970s.

Materials and Methods

Introduction Alkylphenol polyethoxylates (APnEO) are commercially important surfactants with industrial, agricultural, and domestic applications. They are comprised of an alkylphenol * Corresponding author e-mail: [email protected]; phone: +8142-367-5825; fax: +81-42-360-8264. 10.1021/es001250i CCC: $20.00 Published on Web 02/10/2001

that is mainly para-substituted with branched alkyl moieties and a hydrophilic ethylene oxide (ethoxylate) chain ether linked at the phenolic oxygen. Nonylphenol polyethoxylates (NPnEO) account for about 80% of APnEO, and octylphenol polyethoxylates (OPnEO) account for most of the remaining 20% (1). Approximately 500 thousand tons of APnEO are produced annually worldwide (1) with 50 thousand tons in Japan (2). These compounds therefore are common constituents of wastewater. Giger et al. (3) found that nonylphenol (NP) is generated through NPnEO degradation during sewage treatments, mainly by anaerobic digestion (3). APnEO are aerobically degraded to lower ethoxymer and subsequently anaerobically degraded to alkylphenols (APs) by microbial activity in the sewage treatment process. Because of their relatively high toxicity on aquatic organisms and persistence in the environment, understanding the environmental distribution and the fate of APs is important. From recent findings, the shorter NPnEO homologues, carboxylated metabolites, NP, and octylphenol (OP) are known to exert estrogenic effects in aquatic organisms and in mammals and birds, while the higher ethoxymers of NPnEO lack estrogenic activity (4). The distributions of APs have been documented through many studies in the United States and Europe. However, in metropolitan Tokyo, which is one of the most industrialized areas in the world involving numerous human activities potentially releasing APs, only little information on APs and related compounds is available (5). The present paper discusses the distribution of NP, OP, and nonylphenolmonoethoxylate (NP1EO) in the Tamagawa and Sumidagawa Rivers in Tokyo and Tokyo Bay. One of the most important factors in determining the distributions of APs in the aquatic environments is the elimination of APnEO and APs during sewage treatment. Thus, in the present study, NP, OP, and NP1EO were measured for the primary and the secondary effluents from five sewage treatment plants (STPs) in Tokyo, and their removal efficiencies were discussed. The association of APs with aquatic particles is one of the important processes controlling the fate of APs in the rivers and coastal environments. Although there are many monitoring studies of APs in the aquatic environment, very few studies focused on partitioning of APs between the dissolved and particulate phases (6). Some researchers (7) reported that NP was not present in particulate phase. However, based on its moderate octanol-water partition coefficient [Kow ) 104.48 (8)] and on observations of NP in the aquatic sediments (ref 9 and references therein), it is reasonable to expect that some portion of NP is present in the suspended solids. In the present paper, distribution between dissolved phase and suspended solids was studied for the river water and wastewater effluent samples, and a significant proportion of APs was found in the suspended solids. APs in the particulate matter are finally deposited to the bottom sediment, which may act as a reservoir and/or ultimate sink of APs. In this paper, the sedimentary distributions of APs in riverine and coastal environment in Tokyo, including their vertical profile in a dated sediment core from Tokyo Bay, were demonstrated.

 2001 American Chemical Society

Site Description/Sampling. The watershed of Tokyo Bay (Tokyo, Saitama, Kanagawa, and Chiba Prefectures) is 7600 km2, and the population is 33 million. Approximately 77% of domestic wastes and some industrial wastewater generated in the area are transported to municipal STPs and treated prior to their discharge to the rivers and Tokyo Bay. The remaining 23% of domestic wastes (gray water) is directly VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1041

FIGURE 1. Sampling locations. discharged into streams and rivers without treatment. Wastewater from large-scale industrial plants is treated in the plants and then discharged to the rivers and the bay. The area of the Tokyo Bay is 980 km2, and the average water depth is 15 m. The Tamagawa and Sumidagawa Rivers comprise ∼30% of the freshwater inflow to the bay. The Tamagawa River is 140 km long with a normal freshwater inflow of ca. 15 m3/s. The population and area of its drainage basin are approximately 3 million and 1240 km2, respectively. The Tamagawa River has eight municipal STPs in the drainage, one of which (STP-1; Figure 1) was studied in the present study. The population served by the STP-1 was 0.6 million. The Sumidagawa River has a drainage basin of 610 km2 and a population of 5.6 million people and is the largest contributor of organic pollution to the bay. Its length is 50 km, and the freshwater inflow is approximately 40 m3/s. There are 10 municipal STPs in the drainage of the Sumidagawa River. The population served by the STPs studied in the present study (STP-3, STP-4, and STP-5; Figure 1) was 1.7, 0.3, and 1 million people, respectively. Activate sludge treatment following mechanical treatment (i.e., settlement) is served in all the plants. The residence time of wastewater in the plants ranges from 7 to 17 h. The generated sewage sludge is incinerated or land-filled after solidification with cement. The river water and the sewage effluent surveys were conducted four times in 1997 (February, May, August, and October). Sampling locations are shown in Figure 1. The river water samples were taken at three locations in the Tamagawa River (TR-9701, TR-9702, TR-9703) and three locations in the Sumidagawa River (SR-9701, SR-9702, SR9704). The effluents from primary settling tanks (i.e., primary effluent) and from the secondary settling pond following aeration tanks for activated sludge (i.e., secondary effluents) 1042

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001

were taken from the five STPs in Tokyo. The effluent samples were grab samples. Additionally, 24-h composite samples (primary and secondary effluents) were collected from one of the five STPs in September 2000. All the water samples were collected using a stainless steel bucket, stored in a 3-L amber glass bottle, and transported to the laboratory in a cooler. The samples were filtered within 8 h after sampling with a prebaked glass fiber filter (GF/F, Whatman). To examine adsorption loss of APs to the glass apparatus during filtration, a filtrate sample was refiltered and analyzed; no decrease in AP concentrations in the double-filtered sample was confirmed. The filtrates were acidified with 4 M HCl to pH < 1 to depress microbial degradation and stored at 4 °C. The sample stability was examined by analyzing a filtrate sample after 23 and 96 days storage and 28% and 29% of NP decrease was observed, respectively. Because filtrates were extracted usually within 1 week, the APs decrease over the storage period was ignored in the present study. The filters were stored at -30 °C until analysis. The surface sediments were collected from the Sumidagawa River (SR-9702-SR9709) in November 1997 and from the estuary of the Tamagawa River (TR-9703-TR-9707) in August 1998 using an Ekman dredge. Additional sediment samples were collected from the Sumidagawa River in December 1999 (SR9909-SR-9911) using the Ekman dredge. The sediment samples were stored in stainless steel containers with a Teflon liner at -30 °C until analysis. One-meter-length sediment cores were collected from four locations in Tokyo Bay using a gravity corer in September 1993 (site F-2), September 1997 (site F-3), April 1997 (site F-5), and December 1998 (station 1). Cores were sliced into 1- or 2.5-cm intervals onboard. The samples were stored in stainless steel or polyethylene containers, which were prerinsed with methanol, and stored at -30 °C until analysis. All the layers of one of the sediment

cores (F-2) were analyzed for the present paper. This core was sliced into1-cm intervals onboard, and composite samples were made in the laboratory by combining every third interval. For sites F-3, F-5, and station 1, only surface layers were analyzed. A mussel sample was collected from a quay in Tokyo Bay (TR-9707) in 1997. The mussels were shucked with stainless steel scalpels, and whole body tissues were stored in Teflon containers at -30 °C until further analysis. Analytical Method. All the solvents were distilled in glass before use. p-Nonylphenol and the NPnEO (n ≈ 2) mixture were purchased from Tokyo Chemical Industry (Tokyo, Japan), and p-tert-octylphenol was from Wako Pure Chemical (Tokyo). The NPnEO (n ≈ 2) mixture consists of nonylphenol ethoxylates having 1-3 ethoxy units, and the average number of ethoxy units is 2. All the glass equipment was baked for 2-3 h at 400 or 500 °C prior to use to remove organic contamination. The aqueous samples (i.e., filtrates) were neutralized to pH 3-5 with 4 M sodium hydroxide just before extraction. NP, OP, and NP1EO were extracted by solid-phase extraction using a disposable cartridge containing 900 mg of ODS resin (SEP-PAK tC18 plus environmental cartridge, Waters) at a flow rate of 20 mL/min under nitrogen pressure. The cartridges had been previously washed with 20 mL each of hexane, dichloromethane, methanol, and distilled water. The volume of filtrate passed through one cartridge was 1 L for river water and 100-500 mL for wastewater. All the analytes were eluted from the cartridge with 20 mL of MeOH. Breakthrough was examined using tandem cartridges for extraction of a river water sample, and no significant amounts of analytes were detected in eluate from the second cartridge. The filter, which trapped particulate matters, were Soxhlet extracted with MeOH or DCM for 12 h (10-15 min/cycle). On the basis of the volume of the samples filtered, the volumebasis concentrations (i.e., µg/L) were calculated for particulate APs. Also, before the extraction, the weights of the particles on the filters were measured; therefore, particulate APs concentrations are reported on a weight basis (i.e., µg/g) as well. Freeze-dried sediment samples were Soxhlet extracted with DCM for 12 h (10-15 min/cycle). Wet mussel tissue (10 g) was macerated/extracted with DCM (100 mL) and prebaked sodium sulfate anhydrous (50 g) in a 200-mL glass centrifuge tube by Polytron (RT2000; Kinematica). All the extracts were purified and fractionated through a 5% H2O deactivated silica gel column (1 cm i.d. × 9 cm). The silica gel (100-200 mesh; FC 923, Davison Chemical) was prepared as follows; baked at 380 °C to remove organic contamination, activated at 200 °C for 5-6 h, and deactivated by adding 5% (w/w) of distilled water. Purification of the sample extracts was accomplished by eluting the column with 20 mL each of DCM/Hex (25:75 v/v), DCM/Hex (40:60 v/v), DCM/Hex (65:35 v/v), DCM/Hex (80:20 v/v), and DCM. The DCM/Hex (65:35 v/v) fraction was collected as APs fraction, which contained NP and OP. The DCM/Hex (80:20 v/v) and DCM fractions were combined and used as the NP1EO fraction. Both fractions were rotary evaporated to reduce the volume. The APs fraction was transferred to a 1-mL glass ampule while the NP1EO fraction was transferred to a 1.5-mL glass screw vial. The solvent in the ampule was evaporated just to dryness under a gentle stream of N2 gas. The APs fraction was directly subjected to GC-MS analysis following the addition of an appropriate volume (50-1000 µL) of injection internal standard solution (5 µg/mL anthracene-d10/isooctane). For the case of the mussel sample, the APs fraction was hydrolyzed with 2 M KOH/MeOH prior to GC-MS analysis to remove polar interference. The NP1EO fraction was trimethylsilylated prior to GC-MS analysis. A total 200 µL of N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA; Wako Pure Chemicals) was added into the vials

FIGURE 2. Gas chromatograms of (a) alkylphenols and (b) NP1EO TMS derivatives in technical standard (upper) and river sediment (bottom). containing the extracts. Following 1-h reaction at room temperature, the solvent in the vial was evaporated just to dryness under a gentle stream of N2 gas, and an appropriate volume (50-1000 µL) of injection internal standard solution (5 µg/mL p-terphenyl-d14/isooctane) was added to the vial. APs and NP1EO were analyzed on Hewlett-Packard 5972A quadrupole mass spectrometer fitted with a HP5890 gas chromatograph. A DB-5 (J&W Scientific) or a HP-5 MS (Hewlett-Packard) fused silica capillary column (30 m, 0.25 mm i.d., and 0.25 µm film thickness) was used with helium as the carrier gas at 100 kPa. GC-MS operating conditions were 70 eV ionization potential with the MS interface at 310 °C and the electron multiplier voltage at ∼2000 eV. The injection port was maintained at 300 °C, and the sample was injected with splitless mode followed by purge 1 min after the injection. The column oven temperature for APs analysis was held at 70 °C for the initial 2 min, then programmed at 30 °C/min to 180 °C, 2 °C/min to 200 °C, 30 °C/min to 310 °C, and held for 10 min. The oven temperature for NP1EO was as follows: the initial temperature was 70 °C for 2 min, then raised to 200 °C at 30 °C/min, to 230 °C at 2 °C/min, to 310 °C at 30 °C/min, and held for 10 min. A selected ion monitoring method was employed after solvent delay of initial 4 min. As shown in Figure 2a, the chromatogram of NP consists of 10 isomer peaks with various branched structures in the nonyl substituent. Furthermore, some of individual NP peaks contain several isomers and do not represent pure isomers. On the other hand, OP consists of a single peak due to one specific structure. They were quantified by comparing the integrated peak area by the summed selected ion monitor (m/z ) 107 + 121 + 135 + 149 + 177 + 220) with the peak area of the injection internal standard (anthracene-d10, m/z ) 188). The peak composition (i.e., concentration of individual peaks) of the NP standard was determined by GC-FID VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1043

TABLE 1. OP, NP, and NP1EO Cocentrations in Aquatic Environments in Tokyo, Japan nonylphenol

n

min-max

octylphenol

n

min-max

Sewage Effluent (µg/L) primary effluent 10 2.04-21.2 10 0.40-1.82 secondary effluent 10 0.08-1.24 10 0.02-0.48

River Water (µg/L) Sumidagawa River 12 0.08-1.08 12 0.01-0.18 Tamagawa River 12 0.05-0.17 12 0.01-0.07

n

Particulate APsa (µg/g)

min-max 0.14a-1.37

4 8 0.21a-2.96 6 0.12a-0.81a 6 0.04a-0.16a

Sumidagawa River 11 0.52-13.0 10 0.05-0.67 6 0.01-3.47 Tamagawa River 5 0.03-1.14 5 0.003-0.05 3 0.06-0.97 Tokyo Bay 4 0.12-0.64 2 0.006-0.01 1 0.03 a Because particulate NP1EO could not be identified due to interfering peaks, the concentrations would be higher than the values indicated.

analysis. The quantification was achieved from calibration curves made for individual peaks using standard solution (1, 2, 3, and 5 µg/mL for NP and 0.1, 0.2, 0.3, and 0.5 µg/mL for OP). All the calibration curves for NP and OP showed high linearity (r 2 g 0.99). The chromatogram of NP1EO-TMS consists of 10 isomer peaks with various branched structures in the nonyl substituent, as shown in Figure 2b. They were quantified by comparing the integrated peak area by the summed selected ion monitor (m/z ) 117 + 135 + 179 + 193 + 237 + 251 + 321 + 336) with the peak area of the injection internal standard (p-terphenyl-d14; m/z ) 244). The peak composition of the NP1EO in the NPnEO (n ≈ 2) standard was determined by GC-FID analysis. All the calibration curves made for individual peaks using standard solution (1, 2, 5, 10, 20, 50, and 100 µg/mL) showed high linearity (r 2 g 0.999). Analytical Precision. Reproducibility of NP, OP, and NP1EO for water samples was examined through four replicate analyses of the river water and sewage effluent filtrates. The relative standard deviation (RSD) was 2%, 14%, and 8%, respectively. The recovery was checked through three or four replicate analyses of the filtrates spiked with NP, OP, and NP1EO standard solution. The recovery was calculated at 122% ( 4% (n ) 3), 81% ( 14% (n ) 3), 99 ( 7% (n ) 4), respectively. The reproducibility of NP, OP, and NP1EO for sedimentary samples was examined through four replicate analysis of 2 g of a sediment sample. RSD was 4.8%, 5.3%, and 3.2%, respectively. The recovery for sediment analysis was checked through four replicate analysis of the sediment samples spiked with NP, OP, and NP1EO standard solution just before Soxhlet extraction. The recovery for NP, OP, and NP1EO was 108% ( 4%, 110% ( 4%, and 91% ( 7%, respectively. Because tests there confirmed satisfactory recoveries, all the data for samples were not corrected by recovery. The procedural blank ran several times with samples. Normally 1.5 ng of NP, 0.3 ng of OP, and 4 ng of NP1EO were found. Quantification limits were defined as 10 times as the procedure blank value. Therefore, the limit of quantification was 15, 3, and 40 ng/L for NP, OP, and NP1EO, respectively, where 1 L of water sample was analyzed. For the 1-g sediment samples, the limit of quantification was 15, 3, and 40 ng/g for NP, OP, and NP1EO, respectively.

Results and Discussions Sewage Treatment Efficiency and Mechanism. The concentration of NP, OP, and NP1EO in the primary and the secondary effluents are listed in Table 1. The concentration of NP in the secondary effluents ranged from 0.08 to 1.24 µg/L. The range is much lower than those reported for Swiss STPs [i.e., 2.2- 44 µg/L (10)] when the usage of NPnEO for 9

primary effluent secondary effluent river water (n ) 8) (av ( SD) (n ) 8) (av ( SD) (n ) 12) (av ( SD)

NP1EO

Sediment (µg/g dry)

1044

TABLE 2. Particulate NP, OP, and K′oc of NP and OP

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001

nonylphenol octylphenol

59.0 ( 62.9 4.36 ( 2.83

9.52 ( 10.26 1.11 ( 1.31

Proportion of Particulate APs (%)b nonylphenol octylphenol

85 ( 7 61 ( 13

nonylphenol octylphenol

4.52c ( 0.25 4.01c ( 0.31

33 ( 21 13 ( 11

Log K′oc

5.02c ( 0.52 4.41c ( 0.61

3.54 ( 1.96 0.33 ( 0.22 23 ( 15 8(8 5.22 ( 0.40 4.65 ( 0.42

a APs in particulate matter trapped on GF/F filter. b Proportion of particulate APs relative to sum of particulate and dissolved APs. c n ) 4.

household detergents was legal. This is consistent with the fact that usage of NPnEO for household detergents in Japan is minimal. The concentration range in Tokyo is also slightly lower than those reported in some other countries (11-13). This is probably due to high removal efficiency of NP in STPs in Tokyo. The elimination efficiency for NP during the secondary treatment was estimated to be an average of 93% with a range of 79-99%. These values are derived by comparing the concentrations in the secondary effluents with those in the primary effluents. All the calculation is based on grab samples, therefore, the calculated efficiencies should carry some error due to diurnal variation in APs concentrations. However, the separate 24-h composite samples gave 88% of the removal efficiency for NP. Conclusively, the removal efficiencies calculation based on the grab samples are still valid. To obtain more reliable information, we are further studying the elimination efficiency using 24-h composite samples. Since NP is also produced by the anaerobic degradation of NPnEO in the plant, actual elimination of this compound should be higher. The removal in STPs in Tokyo seems to be as efficient as that in Italy [85% (11)], the United States [97% (12)], and Canada [88% (14)]. The comparison of AP concentrations between the primary and secondary effluents also indicates that OP removal (84% on average) was less effective than the NP removal (93% on average). A similar trend was reported for a Canadian STP (13), although greater removal for OP was observed in the other Canadian STPs (14). The less effective OP removal was ascribed to smaller amounts of OP adsorbed by sludge due to its less hydrophobic nature (13). The results of the present study support the mechanism. The phase distribution of NP and OP in sewage effluents is summarized in Table 2. In the primary effluents, ∼80% of NP and ∼60% of OP were found in the particulate phase, while in secondary effluents ∼30% of NP and ∼10% of OP were in the particulate phase. The smaller proportion of particulate APs in secondary effluents is explained by the fact that most of the particulate matter is removed in the settling tank after the aeration tank by precipitation and that the effluents from the settling tank (i.e., secondary effluents) contain only small amounts of suspended solids adsorbing APs. This suggests that the removal of APs during sewage treatment process is mainly due to adsorption by sludge particles and their subsequent settling. Both the primary and the secondary effluents less OP were adsorbed on particulate matter than NP. This is consistent with less hydrophobic nature of OP than NP and with less effective removal of OP during treatment than NP. NP1EO concentrations were much lower than that of NP in the primary effluents (0.14-1.37 µg/L), while in the secondary effluents, they were roughly the same level (0.212.96 µg/L). The possible explanation of this observation is the production of NP1EO through aerobic breakdown of NPnEO during secondary treatments.

FIGURE 3. NP concentrations in the river water samples.

FIGURE 4. NP concentrations in the surface sediment samples from rivers in Tokyo and Tokyo Bay. Arrows indicate discharge points of STPs effluents. APs Distribution in Riverine Environment. Total NP concentrations in the river water ranged from 0.051 to 1.08 µg/L (Table 1, Figure 3). These concentrations are less than those observed for some European rivers (15-17) and are similar to the range observed for U.S. rivers (18). Lower concentrations of NP in Japanese rivers than European rivers (e.g., ref 15) may be due to the fact that NPnEO are mainly used as industrial surfactants and rarely used for household applications in Japan. OP concentrations (0.006-0.177 µg/ L) were 1 order of magnitude lower than NP concentrations, similar to observations in other countries (16, 19). The concentration range of APs was higher in the Sumidagawa River than the Tamagawa River (Table 1, Figure 3). A seasonal trend that NP concentrations are higher in warmer season than in colder season was observed (Figure 3). A similar seasonal trend was observed also for OP. Significant concentrations of APs were found in the sediments from the Tokyo metropolitan area. NP concentrations in the surface sediments from the Sumidagawa and Tamagawa Rivers and Tokyo Bay are shown in Figure 4. The concentration range of NP in the river sediments (0.03-13.0 µg/g dry) is similar to those reported for American (12), Canadian (19), U.K. (20), and Korean (21) rivers. In Canadian rivers, high values for sedimentary NP was reported in close proximity to the discharge points of sewage effluents (19) and the distributions of NP were localized to areas close

(e.g., 1 km) to the outfalls (22, 23). In the Sumidagawa River, however, relatively high concentrations (0.52-13.0 µg/g dry) of NP were detected in a long reach (∼10 km). The freshwater flow and tidal current, especially during flood, possibly mix the surface sediments horizontally. A large proportion of the wastewater effluent may also attribute to the ubiquitously high concentrations of APs in the riverine sediments. Effluents from the STPs take 52% of freshwater flow in the Sumidagawa River in normal flow conditions. In addition, industrial effluents from large-scale factories supply APs. On the other hand, once APs are transported to the coastal zone, there appears to be a seaward decrease. NP concentrations in the surface sediments from Tokyo Bay ranged from 0.12 to 0.64 µg/g dry, which was 1 order of magnitude lower than that in the Sumidagawa River sediments. Blackburn et al. also reported the seaward decrease in NP concentrations in the English estuaries (20). Observed APs concentrations in the river waters are 1-3 orders of magnitude lower than reported acute toxicity levels [hundreds of micrograms per liter (24)]. Recently, however, it was reported that much lower concentrations of APs disrupt the endocrine systems of aquatic organisms (25). The lowest concentrations of NP and OP required to induce production of vitellogenin in the plasma of a male rainbow trout, which is a sensitive biomarker of exposure to estrogenic chemicals, are 20.3 and 4.8 µg/L, respectively(25). Naylor et al. (26) reported that no observable effect concentration (NOEC) of NP for length of Mysidopsis bahia at 28 days was 3.9 µg/L. On the basis of these toxicity reports, U.S. and European regulatory standards are 1 µg/L (1). The concentration of NP in the Sumidagawa River might be potentially hazardous to fishes in the river. Further monitoring to cover wide the range of area in Japan and to identify the sources of APs is necessary. Although there are few data on the toxicity of sedimentary alkylphenolic compounds to benthic organisms, the lowest effect concentration reported to date is 26 µg/g for subacute toxicity of NP to shrimp (12). However, the maximum NP concentration observed in the surface sediment from the Sumidagawa River (i.e., 13.0 µg/g dry at SR-9911) is in the same order of magnitude as the concentration that may induce the subacute adverse effects on benthic organism. On the other hand, Hashimoto et al. found reproductive abnormalities on male flounder from Tokyo Bay (27). Further research is needed to investigate whether the observed levels of NP are significantly disrupting the endocrine systems of benthic organisms in Tokyo Bay. Partitioning of APs between Dissolved and Particulate Phases. The present study clearly demonstrated that significant amounts of NP and OP are present in particulate phase in the river water samples, as shown in Table 2. In the river water samples with suspended solid concentrations of 4.2-27.4 mg/L, ∼20% of NP existed in particulate phase, while ∼10% of OP existed in particulate phase. These proportions are similar to the values reported in an American estuary (28) and higher than those in a Canadian river and U.K. rivers (6, 29). As is obvious from Table 2, NP was more adsorbed to particles than OP. This difference can be explained again by difference in their hydrophobicity. It was reported that the log Kow (octanol-water partition coefficient) of NP is 4.48 while that of OP is 4.12 (8). The difference in partitioning between NP and OP suggests that more NP is accumulated in the sediments and/or in the biological tissues than OP, which is consistent with the results of the sediment analysis as described in the following. The in situ organic carbon-normalized particle-water partition coefficient (K′oc) was calculated using the APs concentration in dissolved phase and particulate phase and organic carbon contents in the particles as follows:

K′oc ) Cs/Caq/foc VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

(1) 9

1045

FIGURE 5. Peak compositions of NP in (a) various environmental samples and (b) sediment core samples. Error bar indicates one standard deviation. where Cs is the solid-phase concentration on a unit weight basis (ng/g), Caq is the aqueous phase concentration on a unit volume basis (ng/mL), and foc is the mass fraction of organic carbon on the particles. Log K′oc values for the river water samples are 4.7-6.1 for NP and 4.1-5.6 for OP. The range for NP is comparable to those (4.7-5.6) observed in the Canadian river (6). Johnson et al. (30) measured organic carbon-normalized partition coefficients (Koc) of octylphenol through laboratory batch techniques using suspended and bed sediments collected from English rivers and measured log Koc ranged from 3.52 to 5.59 (30). Our values of log K′oc for OP were within a range of the Koc determined by the laboratory experiment. The partitioning of organic compounds between water and organic matter is generally controlled by hydrophobicity. Schwarzenbach et al. (31) found that there is a linear correlation between Kow and Koc (i.e., organic carbon-normalized particle-water partition coefficient) on a laboratory sorption experiment using natural sediments as follows:

log Koc ) 0.82 × log Kow + 0.14

(2)

If isotherm and equilibrium partitioning are assumed, K′oc of nonpolar compounds can be predicted from Kow, which 1046

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001

is referred to as predicted Koc. The predicted log Koc values calculated from Kow values of NP and OP using eq 2 are 3.81 and 3.52, respectively. All of the K′oc values observed in the river water samples are 1 order of magnitude higher than the predicted Koc. This means that APs partition more to particulate phase than expected from their hydrophobicities. This has also implied that the particulate APs could play a significant role in their transport in the aquatic environments and their incorporation into bottom sediments and could be one of the most important sinks and their fates. In the above discussion, K′oc was calculated in terms of “sum of all the isomers”, although NP consists of many isomers with various branched structures in the nonyl substituent. As shown in Figure 2a, NP consists of 10 peaks on the gas chromatogram. Furthermore, some of the peaks consist of several isomers (32). Many researchers determine sum of all the isomer peaks as “nonylphenol”. Recently, Yamashita et al. (32) indicated that estrogenic activity is different among isomers, which underscores the need to determine the specific NP isomers in environmental samples. However, as shown in Figure 5a, peak composition of NP was fairly constant among the technical standard mixtures and various environmental compartments collected from various locations in Tokyo. This suggests that isomer

composition of source materials is uniform and that isomer selective reaction does not occur during various environmental settings. However, there may appear to be minor differences in the composition between the primary and the secondary effluents, suggesting isomer selective degradation during intensive biological attack (i.e., activated sludge treatment). The peak profile of NP was fairly uniform throughout the sediment core (Figure 5b), suggesting that isomer selectivity in the early diagenesis did not occur and that NP composition in technical mixture is constant during the past 40 years. In the water samples, NP peak profiles are almost identical between dissolved phase and particulate phase (Figure 5a). To discuss quantitatively, K′oc is calculated for each peak. Because some of the peaks contain several isomers, the K′oc values calculated for individual peaks do not represent K′oc for individual isomers but for an individual mixture of isomers. No significant differences in K′oc were observed among isomer peaks, suggesting that partitioning behavior is similar between isomers. This could be one of the major reasons why the isomer peak distribution is constant throughout the various environmental compartments. This may suggest that risk assessment of NP can be accomplished in terms of sum of all the isomers. Possibility of Formation of APs in the River Sediment. It has been proved that AP1EO is transformed to APs during the anaerobic digestion in the sewage treatment (3). So far, however, there has been no reports demonstrating the anaerobic transformation of NPnEO to NP in the river environment. The observations in the present study may be explained as a result of the transformation. APs were accumulated in the river sediments. Especially, NP concentrations in the Sumidagawa River (0.52-13.0 µg/g dry) were higher than those in the Tamagawa River (0.03-1.14 µg/g dry), as shown in Figure 4. Deeper (∼3m) water, slower stream, and invasion of seawater in addition to high inputs of organic matter make the Sumidagawa River anaerobic in its bottom environments, whereas shallow (∼1 m) streams in the Tamagawa River supply oxygen into the river and keep it constantly aerobic. The anaerobic condition in the Sumidagawa River sediments was confirmed by the high contents of organic carbon and the strong smell of hydrogen sulfide during the sampling. The sedimentary NP concentrations in the Sumidagawa River (0.52-13.0 µg/g dry) is in the same range as those in the suspended particles in the river water (2.0-7.0 µg/g dry; Table 2). This is different from the observation in U.K. rivers (29) where NP concentrations in the bottom sediments were 1 order of magnitude lower than those in the suspended particles. Also for another surfactantderived hydrophobic compounds (linear alkylbenzenes; LABs), it was observed that their sedimentary concentrations were 1 order of magnitude lower than those in suspended particles in the Sumidagawa River (33) and was ascribed to the dilution by soils and sand in the river bed (34). In the case of APs, however, their concentrations in the sediment were in the same range as those in suspended solid, despite the dilution by soil and sand. This also can be explained by the in situ transformation of AP1EO to APs in the Sumidagawa River sediment. The decreased ratio of NP1EO to NP in the sediment samples may support the facilitated anaerobic conversion of NP1EO to NP. In the Sumidagawa River, NP1EO/NP ratios in the river waters were 1.2 ( 0.9, whereas the ratios in the river sediments were 0.15 ( 0.24. Furthermore, the lower NP1EO/NP ratio was observed in the Sumidagawa River sediment (0.01-0.09; n ) 6) as compared with those in the Tamagawa River sediment (0.2-0.8; n ) 3). The difference in the sedimentary APs concentrations appears to be related to the water column APs. Higher AP concentrations in water of the Sumidagawa River can be explained by resuspension and/or desorption of APs generated in the bottom sediments, if their in situ formation could

be operative. Also the higher AP concentration is partly due to the fact that the Sumidagawa River receives larger amounts of wastewater from industries such as textile and paper mills that potentially discharge APs (35) than the Tamagawa River whose drainage area is mainly residential. The seasonal trend that APs concentration in the river water are higher in spring and summer than in autumn and winter may be related with the APs formation in the benthic environment. The seasonal trend cannot be attributable to seasonal change in water flow because the water level is normally higher in summer due to higher precipitation in this season than in the other season. The seasonal trend in AP concentrations can be explained by seasonal change in microbial activities that breakdown NPnEO to NP. The water temperatures observed during spring and summer surveys (21.6-30.0 °C) were higher than that of autumn and winter surveys (9.1-18.9 °C). These temperature differences could promote degradation of parent NPnEO to produce lower ethoxymers. The shift of EO distribution to shorter chain length in summer was already reported by Maruyama et al. in the same rivers (5). Higher water temperature could also facilitate the conversion of NP1EO to NP in the benthic environment in the rivers, if the conversion process could be operative. Lower dissolved oxygen concentrations are also observed in spring and summer in the Sumidagawa River (36), which might accelerate the conversion in the warmer season. However, the present study provided only circumstantial evidence for anaerobic conversion of NP1EO to NP in river environments. More direct evidence to support the in situ production of NP in the river sediment is needed. Sediment Core. The dating of the sediment core was achieved by measuring 210Pb, 137Cs, and some molecular markers [i.e., LABs (linear alkylbenzenes) and TABs (tetrapropylene-based alkylbenzenes)]. Vertical distribution of 137Cs is shown in Figure 6b. Activity of 137Cs was first detected at the depth of 40 cm and a reached maximum level at 33 cm depth, corresponding to 1954 and 1963, respectively. The broad peak of the radioactivity indicates vertical mixing and remobilization of the sediment. The dating is consistent with the other geochlonometers, and the details have been described in Sanada et al. (37). The estimated sedimentation date is marked in Figure 6a,b. The historical record of production of APs in Japan (35) is also shown in Figure 6c, although those of APnEO are not available. Because APnEO is a major product from APs, the historical trend of APs production can be regarded as that of APnEO production. As shown in Figure 6a, APs first appeared in the sediment core in the layer (39-42 cm) deposited in late 1950s and increased in their concentration to the 22-24 cm layer corresponding to the middle 1970s. This profile corresponds to the onset and increase in the production of APs and probably APnEO surfactants in Japan (Figure 6c). APs appear to be detected in a deeper layer than the starting year of the APs production in Japan. This may be attributable to penetration of APs in the core due to mixing of sediment. In the upper section of the sediment core, i.e., above 22-24 cm depth corresponding to the mid-1970s or later, their concentrations showed a decreasing trend toward the sedimentwater interface. NP concentration in the most recent sediments (0.07 µg/g dry) was by a factor of 8 less than the maximum concentration (0.53 µg/g dry) in the sediment layer deposited in the mid-1970s. On the other hand, their production has been increasing continuously until today. The gap between the production and the vertical profile in the sediment core could be explained by possible decrease of the input of APs. The reduction is probably related to some legal regulations on industrial wastes implemented around 1970. Water Pollution Control Law was enacted in 1970, where treatment of industrial wastewater is made compulsory. Similar vertical profiles (i.e., maximum around VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1047

FIGURE 6. Vertical profiles of (a) NP and OP and (b) 137Cs in the sediment core from Tokyo Bay (F-2) and (c) alkylphenols production in Japan (data from ref 35). the mid-1970s) were reported for some heavy metals in sediment cores from Tokyo Bay, and it was also explained by the regulations on industrial wastewater in early 1970s (38). Degradation and generation of NP in the sediments should be taken into account. However, neither negligible disappearance nor production of NP were suggested through the lack of change in NPnEO oligomer distribution throughout the sediment cores collected form Canadian coasts (39). Similar subsurface maximum of NP was reported for Swiss lakes (17) and ascribed to the reduced use of NPnEO surfactants due to legal or voluntary restrictions. In Greifensee, NP concentration in the most recent sediments (0.2 µg/g) is by a factor of 5 less than the maximum concentration (1.0 µg/g) in the sediment layer deposited between 1970 and 1985. This decrease is well correlated to the decrease in NP concentrations in rivers flowing into the lake (40), and Schaffner et al. (17) concluded that sedimentary record of NP can be used to evaluate the changes of their input to the aquatic system. More recently, Yamashita et al. (41) reported a different vertical profile for a sediment core collected from Tokyo Bay. In the core, NP concentrations showed steady increase from 30 cm to the surface. However, their reported concentration (i.e., 4 µg/g for the surficial sediment) seems to be extraordinarily highs1 order of magnitude higher than those observed for surface sediments collected from several locations of Tokyo Bay in the present study. This very high concentrations of NP could have been attributed to the interference in the HPLC technique used for the determination of NP, which may cause the bias.

Acknowledgments We are grateful to for Mr. Mohamad Pauzi Zakaria and Dr. Norio Ogura for valuable comments on our manuscripts. We thank bureau of sewage works of Tokyo Metropolitan government for their providing the effluent samples, Dr. Ishimaru and the crew of R/V Seiyo-maru of Tokyo University of Fisheries for their help in collecting the sediment core samples, and Dr. Hashimoto and Mr. Horiuchi for providing the sediment samples from Tokyo Bay. Several graduates and undergraduates in our laboratories provided welcome assistance with the fieldwork. This work was supported by Integrated Research Program for Effects of Endocrine Disrupters on Agriculture, Forestry and Fisheries and Their 1048

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001

Action Mechanisms on Domestic Animals and Fishes (ED00-II-2).

Literature Cited (1) Renner, R. Environ. Sci. Technol. 1997, 31, 316A-320A. (2) Nakamura, Y. J. Jpn. Soc. Water Environ. 1998, 21, 192-196 (in Japanese). (3) Giger, W.; Brunner, P. H.; Schaffner, C. Science 1984, 225, 623625. (4) Routledge, E. J.; Sumpter, J. P. Environ. Toxicol. Chem. 1996, 15, 241-248. (5) Maruyama, K.; Yuan, M.; Otsuki, A. Environ. Sci. Technol. 2000, 34, 343-348. (6) Sekela, M.; Brewer, R.; Moyle, G.; Tuominen, T. Water Sci. Technol. 1999, 39, 217-220. (7) Kannan, N.; Yamashita, N.; Petrick, G.; Duinker, J. C. Environ. Sci. Technol. 1998, 32, 1747-1753. (8) Ahel, M.; Giger, W. Chemosphere 1993, 26, 1471-1478. (9) Takada, H.; Eganhouse, R. In The Encyclopedia of Environmental Analysis and Remediation; Mayers, R., Ed.; Wiley and Sons: New York, 1998; Vol. 5, pp 2883-2940. (10) Ahel, M.; Giger, W.; Koch, M. Water Res. 1994, 28, 1131-1142. (11) Di Corcia, A.; Samperi, R.; Marcomini, A. Environ. Sci. Technol. 1994, 28, 850-858. (12) Naylor, C. G. Text. Chem. Color. 1995, 27, 29-33. (13) Lee, H.-B.; Peart, T. E. Water Qual. Res. J. Can. 1998, 33, 389402. (14) Bennie, D. T.; Sullivan, C. A.; Lee, H.-B.; Maguire, R. J. Water Qual. Res. J. Can. 1998, 33, 231-252. (15) Ahel, M.; Giger, W.; Schaffner, C. Water Res. 1994, 28, 11431152. (16) Blackburn, M. A.; Waldock, M. J. Water Res. 1995, 29, 16231629. (17) Schaffner, C.; Reiser, R.; Albrecht, A.; Sturm, M.; Lotter, A. F.; Giger, W. Presented at the International Conference on Environmental Endorcrine Disrupting Chemicals, Ascona, Switzerland, March 1999; p 52. (18) Naylor, C. G.; Mieure, J. P.; Adams, W. J.; Weeks, J. A.; Castaldi, F. J.; Ogle, L. D.; Romano, R. R. J. Am. Oil Chem. Soc. 1992, 69, 695-703. (19) Bennie, D.; Sullivan, C.; Lee, H.; Peart, T.; Maguire, R. Sci. Total Environ. 1997, 193, 263-275. (20) Blackburn, M. A.; Kirby, S. J.; Waldock, M. J. Mar. Pollut. Bull. 1999, 38, 109-118. (21) Khim, J. S.; Kannan, K.; Villeneuve, D. L.; Koh, C. H.; Giesy, J. P. Environ. Sci. Technol. 1999, 33, 4199-4205. (22) Bennett, E. R.; Metcalfe, C. D. Environ. Toxicol. Chem. 1998, 17, 1230-1235. (23) Bennett, E. R.; Metcalfe, C. D. Environ. Toxicol. Chem. 2000, 19, 784-792. (24) Comber, M. H. I.; Williams, T. D.; Stewart, K. M. Water Res. 1993, 27, 273-276.

(25) Jobling, S.; Sheahan, D.; Osborne, J. A.; Matthiessen, P.; Sumpter, J. P. Environ. Toxicol. Chem. 1996, 15, 194-202. (26) Naylor, C. G. Soap/Cosmetics/Chemical Specialties 1992, 68, 27-31, 72. (27) Hashimoto, S.; Bessho, H.; Hara, A.; Nakamura, M.; Iguchi, T.; Fujita, K. Mar. Environ. Res. 2000, 49, 37-53. (28) Van Ry, D. A.; Dachs, J.; Gigliotti, C. L.; Brunciak, P. A.; Nelson, E. D.; Eisenreich, S. J. Environ. Sci. Technol. 2000, 34, 24102417. (29) Long, J. L. A.; House, W. A.; Parker, A.; Rae, J. E. Sci. Total Environ. 1998, 210/211, 229-253. (30) Johnson, A. C.; Write, C.; Besien, T. J.; Jurgen, M. D. Sci. Total Environ. 1998, 210/211, 271-282. (31) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M. Environmental Organic Chemistry; John Wiley & Sons: New York, 1993; p 275. (32) Yamashita, N.; Kannan, K.; Hashimoto, S.; Miyazaki, A.; Giesy, J. P. Organohalogen Compd. 1999, 42, 121-125. (33) Takada, H.; Ishiwatari, R. Environ. Sci. Technol. 1987, 21, 875883. (34) Takada, H. Ph.D. Dissertation, Tokyo Metropolitan University, 1989. (35) Ministry of International Trade and Industry of Japan. Annual report of statistical data on industries in Japan; Tokyo, 1996 (in Japanese).

(36) Bureau of Environment, Tokyo Metropolitan Goverment. Data of Quality of Riverine and Estuarine Water in Tokyo 1996; Tokyo, 1998 (in Japanese). (37) Sanada, Y.; Sato, F.; Kumata, H.; Takada, H.; Yamamoto, A.; Kato, Y.; Ueno, T. Chikyukagaku (Geochemistry) 1999, 33, 123138 (in Japanese). (38) Matsumoto, E. Chikyukagaku (Geochemistry) 1983, 17, 27-32 (in Japanese). (39) Shang, D.; Macdonald, R.; Ikonomou, M. Environ. Sci. Technol. 1999, 33, 1366-1372. (40) Ahel, M.; Molnar, E.; Ibric, S.; Giger, W. Presented at the 3rd IAWQ Specialized Conference on Hazard Assessment and Control of Environmental Contaminants, Kyoto, Japan, December 1999. (41) Yamashita, N.; Kannan, K.; Imagawa, T.; Villeneuve, D.; Hashimoto, S.; Miyazaki, A.; Giesy, J. Environ. Sci. Technol. 2000, 34, 3560-3567.

Received for review May 10, 2000. Revised manuscript received December 6, 2000. Accepted December 12, 2000. ES001250I

VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1049