Environ. Sci. Technol. 2007, 41, 206-213
Sorption of Selected Endocrine Disrupting Chemicals to Different Aquatic Colloids J. L. ZHOU,* R. LIU, A. WILDING, AND A. HIBBERD Department of Biology and Environmental Science, School of Life Sciences, University of Sussex, Falmer, Brighton BN1 9QJ, United Kingdom
The sorption of seven endocrine disrupting chemicals (EDCs) to aquatic colloids was determined by cross-flow ultrafiltration (CFUF) followed by gas chromatography-mass spectrometry (GC-MS). Results show that the colloidal organic carbon normalized sorption coefficient (Kcoc) of EDCs to different aquatic colloids varies by a factor of 6-12 because such colloids are of different origin. Through characterization of colloidal samples, a significant relationship was established between Kcoc values and the molar extinction coefficient of colloids at 280 nm, whereas no other colloidal properties such as elemental ratios were correlated with Kcoc values. The results are consistent with other reports of the importance of the quality of sorbents such as their aromatic carbon content in sorbing various organic pollutants. The presence of a surfactant was found to increase Kcoc values for estrone (E1) and 17Rethynylestradiol (EE2). The method was subsequently applied for determining EDC concentrations in field samples, where both conventional and truly dissolved EDCs showed higher concentrations close to sewage outfalls than either upstream or downstream, confirming the sourceconcentration relationship. In addition, the truly dissolved EDC concentrations were lower than the conventional dissolved concentrations, confirming that there were interactions between aquatic colloids and EDCs. It is estimated that between 10 and 29% of EDCs are associated with aquatic colloids. As colloids are highly abundant in rivers and ocean, they will therefore play a significant role in the environmental behavior and fate of EDCs.
Introduction Colloids, generally defined as nanoparticles or macromolecules with a size between 1 nm and 1 µm, are ubiquitous in natural environmental systems. Because of weathering reactions and microbiological processes, colloids are produced and transported into natural waters such as lakes, rivers, and sea. It has been widely recognized that colloid concentration is approximately 106-108 particles per liter (1-3), and 10-40% of marine total organic carbon is colloidal in nature (4, 5). Although colloids constitute a relatively small fraction (10%) of total waterborne particle mass (6), the speciation, bioavailability, transport, and ultimate fate of trace pollutants are controlled by their sorption onto colloids (7, 8) because of their large surface area and a great number * Corresponding author phone: 44 1273 877318; fax: 44 1273 677196; e-mail:
[email protected]. 206
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 1, 2007
of sorption sites. There is potentially, therefore, a variety of mechanisms, for example, covalent, electrostatic, and specific sorption through which colloidal particles can interact with solutes in the aqueous phase. Colloidal material has further been implicated to play a crucial role as an intermediary in environmental and biogeochemical processes, for example, N and P cycling, aggregation, sedimentation, as well as trace metal behavior (9, 10). Thus, to understand the role of colloids in the fate and behavior of trace pollutants in the aquatic environment, it is essential to obtain and characterize colloids of different origins for determining their key physicochemical properties. Endocrine disrupting chemicals (EDCs) are pollutants that alter functions of the endocrine system and consequently cause adverse health effects in an intact organism or its progeny. The main evidence leading to disruption of endocrine function comes from changes observed in several wildlife species as well as humans, for example, feminization and hermaphroditism in wildlife and the development of testicular and prostate cancer and decreased sperm reproduction in humans (11-13). These chemicals (e.g., hormones) may originate from natural processes in animals and plants which exert hormonal actions on other animals or plants. In addition, many industrial activities (e.g., agrochemicals, medicines) have produced chemicals such as dichlorodiphenyl trichloroethane, polychlorinated biphenyls, and certain pharmaceuticals being identified as potential EDCs. Such EDCs may be released directly or indirectly to the aquatic environment where interactions with colloidal materials are likely. As a result, the concentration, speciation, bioaccumulation, and toxicity of EDCs are all co-regulated by the presence of colloids (14-16). Direct measurement of the distribution of trace organic pollutants such as EDCs between colloids and dissolved phase has always been a challenge. Currently, fluorescence quenching, solubility enhancement, and radioisotopes have been used in such experimentation (17-21). In our previous study, crossflow ultrafiltration (CFUF) combined with gas chromatography-mass spectrometry (GC-MS) has been successfully developed for determining the partition coefficients of selected EDCs between natural colloids and dissolved phase (22). It has been shown that the partition of selected EDCs onto colloids is relatively independent of the physicochemical properties of EDCs, in particular, their octanolwater partition coefficient values (Kow), and the sorption of relatively polar EDCs by colloids may be attributed to the physical-chemical properties of colloidal materials, for example, a great number of sorption sites with a variety of interaction mechanisms, for example, covalent, electrostatic, and specific binding. The purpose of this study was to extend the application of CFUF-GC-MS technique for determining sorption coefficients of EDCs to various types of natural colloids, to evaluate the key chemical characteristics of those different colloids, and to differentiate between truly dissolved and colloidbound EDCs in field samples close to sewage outfalls.
Experimental Section Materials. The solvents used including methanol and ethyl acetate were of distilled-in-glass grade (Rathburns). EDCs including 17β-estradiol (E2), estrone (E1), 17R-ethynylestradiol (EE2), 16R-hydroxyestrone, E2-d2, and 4-nonylphenol were purchased from Sigma United Kingdom, and bisphenol A, bisphenol A-d16, 4-tert-octylphenol, and bis(trimethylsilyl)trifluoroacetamide (BSTFA) containing 1% of trimethylchlorosilane were supplied by Aldrich (Dorset, United Kingdom). 10.1021/es0619298 CCC: $37.00
2007 American Chemical Society Published on Web 11/29/2006
TABLE 1. Details of the Sampling Sites Used in This Study sampling site
a
purpose
1
spiking EDCs
2 3 4 5 6
spiking EDCs spiking EDCs spiking EDCs spiking EDCs naturally occurring EDCs
site details Longford Stream, a tributary of River Ouse, East Sussex, U.K. River Ouse at East Mascalls, West Sussex, U.K. Horsham STW effluent samples, West Sussex, U.K. River L’Aa at Gravelines, France Seawater in Lancing Beach, West Sussex, U.K. Sheffield Park STW, River Ouse, West Sussex, U.K.
sampling date
pH
COC (mg/L)
salinity
02/10/2004
6.4
2.3
0
01/02/2005 10/04/2005 04/04/2005 23/06/2005 08/04/2004 08/07/2004
7.3 7.1 8.5 8.0 6.8 7.2
3.0 9.2 2.9 0.4 n.d.a 7.5
0 0 0.1 35 0 0
n.d.: not determined.
Separate stock solutions of individual compounds were made up at 1000 mg/L by dissolving an appropriate amount of each substance in methanol. From these standards, a mixture of working standards containing each compound at 100 and 10 mg/L were prepared by diluting the stock solution in methanol and were used to spike the solutions. All standard solutions were stored at -18 °C prior to use. A pure surfactant (dodecylbenzesulfonic acid) was obtained from Aldrich. Highpurity deionized water was supplied by a Maxima Unit from USF Elga, United Kingdom. Sampling Sites. Six sampling sites were selected to provide a range of colloid properties (i.e., river water, treated effluent, and seawater), of which five were used for spiking experiments and one for in-situ analysis (Table 1). Water sample was collected in 50-L stainless steel barrels. Prior to sampling, containers were soaked with 5% Decon-90 overnight and were thoroughly rinsed first with deionized and then highpurity water. Once transported to the laboratory, water samples were filtered within 3 days through precombusted 0.7-µm glass fiber filters (GF/F, Whatman). The filtered waters were stored at 4 °C and were processed within one week. Colloid Isolation and Characterization. Colloid suspensions with a size from 1 kDa to 0.7 µm were isolated from the filtered waters by CFUF in the sampling mode, using a Millipore 1-kDa regenerated cellulose Pellicon 2 module at a concentration factor (cf) of 10-15 (23). The retention and mass balance of colloids by CFUF have been extensively examined (22, 23). Part of colloid suspension was used in the partition experiments following a previously established protocol (22); the remaining part was frozen and freezedried at -50 °C and reduced pressure (Heto PowerDry PL3000) before further characterization. The freeze-dried colloids were subsequently characterized for elemental composition (C, H, N, O) by flash combustion/ chromatographic separation and multidetector techniques (ECS4010, Costech, United States). UV absorbance at 280 nm was measured by dissolving an appropriate amount of each colloid sample in water, followed by analysis using a UV-vis spectrophotometer (UV300, Thermospectronic). Molar extinction coefficients at 280 nm (e280) of each colloid sample were derived by dividing their absorbance values by their respective colloidal organic carbon (COC) concentrations. Simulated Colloid/Dissolved Partition of EDCs. The CFUF method was applied to derive the partition coefficient of EDCs between colloids and dissolved phase of samples from sites 1-5 (Table 1). Different amounts of concentrated colloid solutions were diluted with the
