- Email: [email protected]
Environmental fate and partition co-efficient of oestrogenic compounds in sewage treatment process
ARTICLE IN PRESS
Environmental Research 106 (2008) 313–318 www.elsevier.com/locate/envres
Environmental fate and partition co-efficient of oestrogenic compounds in sewage treatment process H.E. Keenan, A. Sakultantimetha, S. Bangkedphol Department of Civil Engineering, University of Strathclyde, John Anderson Building, 107 Rottenrow, Glasgow, Scotland G4 ONG, UK Received 15 November 2006; received in revised form 9 May 2007; accepted 22 July 2007 Available online 29 October 2007
Abstract The presence of residual pharmaceuticals and environmental endocrine disrupters (EEDs) is increasingly significant due to their impact on human health and wildlife. Of the compounds implicated as EEDs, the most potent in their oestrogenic effect are the natural and synthetic oestrogens. As these compounds will be present in the sewage matrix, it is necessary to establish their fate during sewage treatment with a view of removal and safe disposal to avoid unnecessary exposure. Using methodology developed by the author this paper describes the results of a study undertaken to determine both the Kow and the adsorption characteristics of these oestrogens. The experimental values obtained were compared to a computational default model. However, there was disparity between the default model and the values determined experimentally. This was especially the case in the determination of the Koc which impacts directly on the sludge adsorbance potential. The calculated results ranged from log 4.21 for b-oestradiol to log 4.68 for 17a-EE-3-ME, the experimental results were higher log(5.04–log 5.83), respectively. The implications of the findings in terms of water recycling and sewage sludge disposal are also discussed. r 2007 Elsevier Inc. All rights reserved. Keywords: Oestrogens; Environmental fate; Adsorption isotherm; Soil–water partition coefficient (Kd); Organic-carbon partition coefficient (Koc); Octanol–water partition coefficient (Kow); BioConcentration Factor (BCF); Sewage treatment
1. Introduction Many of the widely used chemicals in the environment may be inflicting adverse hormonal effects in human and other animals. Of particular concern are synthetic oestrogens, pharmaceutical preparations deliberately intended to elicit oestrogenic responses. The compounds 17b-oestradiol and oestrone are natural hormones, 17a-ethinyl oestradiol (17a-EE), and 17a-ethinyl-oestradiol-3-methyl ether (17aEE-3-ME) are synthetic hormones widely used in contraceptive and hormone replacement therapy formulations ultimately excreted and are therefore environmental contaminants. This study is particularly concerned with this group of compounds. Exposure to chemicals can occur through digestion, inhalation or skin contact. Many chemicals will ultimately pass through sewage treatment Corresponding author. Fax: +44 141 553 2066.
E-mail address: [email protected] (H.E. Keenan). 0013-9351/$ - see front matter r 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.envres.2007.07.007
works (STWs), either by excretion or directly from industrial effluents. Further exposure can occur due to recycling of effluent for potable water supplies and from disposal of sludge solids to land. Fig. 1 shows how oestrogens may be transported through the environment leading to further exposure. The eventual environmental fate is determined by chemodynamics. Although these compounds may be present in very small quantities, additive, and synergistic effects have to be considered. To date, many studies have concentrated on the adverse physiological effects exhibited by fish exposed to sewage final effluent (IEH, 1999). Little work has been done on the environmental hazards through sewage solids, although the oestrogens would be expected to be sorbed strongly to the solid matrix of the sewage rather than remain in the liquid fraction based on their chemical and physical properties. Therefore, this paper is concerned with routes A, E, and F as shown in Fig. 1.
ARTICLE IN PRESS 314
H.E. Keenan et al. / Environmental Research 106 (2008) 313–318
Fig. 1. Transportation of oestrogens.
2. Theory Many different models have been used in an attempt to predict the environmental fate of chemicals (Samiullah, 1990). A simple approach to environmental partitioning can be explained by the concept of fugacity. This approach although based on empirical values determines partitioning of a compound by ratio, it may be particularly relevant to natural and synthetic oestrogens as the concentration of these compounds at any particular STWs will be variant. The relationship between fugacity and partition coefficients is shown in Fig. 2. This approach (level 3) is used by the US EPA in their EPISUITE environmental fate modeler (http://www.epa.gov/opptintr/exposure/docs/EPISuitedl.htm). Without experimental data, the model produced is useful for prediction but is essentially a default model. A more accurate assessment is possible by measuring the partition co-efficients and entering the experimental values obtained into the model as oppose to the default values. This research determined both the solid/water partition co-efficient Kd (also known as Kp, which is represented as CS in Fig. 2. The Koc was calculated from this) and also, the Kow, the octanol/water partition co-efficient (represented as CO in Fig. 2) experimentally for natural and synthetic oestrogens in a sewage sludge. The actual models obtained from the input of the experimental data were then compared to the default models. The Kow value is defined as the ratio of a chemicals concentration in the octanol phase compared to that of its concentration in the aqueous phase of a two-phase octanol–water system. There is a correlation between Kow and the retention time from a reverse phase high performance liquid chromatographic (HPLC) system. A linear regression between the elution times of each compound and the Kow value can be established. This is a useful parameter in environmental prediction as it is analogous with the lipid/water membrane in humans. It shows strong correlation with the BioConcentration Factor (BCF). One proposed correlation illustrated a linear
Fig. 2. Relationship between fugacity constants and partition coefficients (Samiullah, 1990).
relationship between compounds with a log Kow from 102 to 106 of: log(BCF) ¼ 0.756 log Kow1.415 for human adipose tissue (Geyer et al., 1987). During sewage treatment, a chemical will partition between the solid: liquid fractions (Kd). The chemical would normally have time to equilibrate within the operating time of the STW. Although it is common to measure the Kd, for strongly hydrophobic compounds the partitioning is strongly influenced by the organic content of the solid; therefore, Koc is a better indicator of environmental fate. The Koc was measured using a standard method (ASTM International, 2001).
ARTICLE IN PRESS H.E. Keenan et al. / Environmental Research 106 (2008) 313–318
315
3. Experimental
3.1. Determination of Kow
The oestrogens chosen for the investigation were 17boestradiol, 17a-oestradiol, oestrone, 17a-ethinyl oestradiol, and 17a-ethinyl-oestradiol-3-methyl ether (Mestranol), and the internal standard Diethyl Stibestrol (DES) (Sigma Aldrich). The stock solutions of oestrogens were prepared in methanol, working solution was made by serial dilution in water. Initially, the oestrogens were scanned (Phillips PU8740 UV/Vis spectrophotometer) to determine the optimum wavelength for detection i.e. lmax. All oestrogens showed similar UV characteristics with several absorbance peaks in the spectra. Due to interference from the sewage matrix, the lmax 196 nm could not be used, but l 280 nm proved suitable. However, the loss of sensitivity resulted in a concentration step, this was achieved using solid phase extraction (SPE). Therefore, an SPE–RPHPLC (UV detection) with gradient elution method of qualitative and quantitative analysis was developed (MallinckrodtBaker, Bakerfacts, 11.1, 1996) and subsequently used for the determination of Kd. The Kow analysis required no concentration and was performed at l 254 nm.
The Kow was determined from the retention time of the HPLC analysis. Five polyaromatic hydrocarbons (PAHs) of known Kow (values taken from EPISUITE) were purchased from Sigma Aldrich (Acenapthylene, Fluorene, Phenanthrene, Chrysene, and Benzo[a]pyrene). A mixed 1 mg l1 standard containing the PAHs and oestrogens was prepared in HPLC grade methanol. This was analysed by RPHPLC using the analytical method summarised in Table 1. The log of retention time of the five PAHs were plotted against their Kow values and from the linear equation the Kow of the oestrogens was calculated. 3.2. Determination of Kd and Koc Thermised sewage sludge pellets were obtained from a Regional Council Water Services department. The pellets were produced as a garden fertiliser for use by the general public. Analysis of the pellets showed a total organic carbon (TOC) value of 30.82% this analysis was performed externally. The Kd was determined in accordance with ASTM standard testing method (ASTM International, 2001). The method is in two parts and involves equilibrating the chemical between known amounts of water and solid. Initially, an equilibrium time was established for the oestrogens in the solid:water, then sorption coefficients are established (Fig. 3). Initially, the granules were oven dried at 105 1C to constant weight, then sieved to o2 mm and stored in a desiccator at room temperature until required. Plate counts were carried out on nutrient agar to ensure that the granules were sterile (APHA, AWWA and WEF, 1995). The analytical conditions for the experiments are given in Table 2.
Table 1 Analytical conditions for the determination of Kow Instrument: Perkin-Elmer Series 410 HPLC with LC90 UV spectrophotometric detector (l 254 nm) Column: 125 3.2 mm Envirosep-pp ODS (purchased from Phenomonex) Elution program Time (min)
% Acetonitrile
Water
Gradient
0 1 2 3
– 8.0 8.0 6.5
40 60 100 100
60 40 0 0
– 3.0 3.0 –
3.2.1. Determination of equilibrium time A 1000 mg l1 stock solution of the five oestrogens was prepared in HPLC grade methanol. From this, a 5 mg l1 solution in Nanopure water was made for the experimental 2500
1.4 b-oestradid
1.2 1
y = 39.761x
2000
a-oestradid
2 R = 0.9895
µ x/m µg/g
17aEE Oestrone
0.8
17aEE3me
0.6
y = 33.92x
1500
2 R = 0.985
y = 61.585x
b-oestradiol a-oestradiol 17a-EE Estrone 17a-EEME
1000
0.4 500
60
54
48
39
28
24
20
22
8
6
4
3
2
1.5
1
0
0
0.2 0.5
Ratio of oestrogens sample/DES
Step no.
0
0
Time (hours)
20
40 Ce µg/l µ
Fig. 3. Equilibrium time and sorption isotherms.
60
R2 = 0.9316 y = 48.387x 2 R = 0.9957
y = 218.02x R2 = 0.9268
80
ARTICLE IN PRESS H.E. Keenan et al. / Environmental Research 106 (2008) 313–318
316
reactors. Bottle-point experiments were used to determine each point on the adsorption isotherm in an individual completely mixed batch reactor (CMBR) by equilibrating a given solution of solute with a given quantity of sorbent. Individual samples were prepared in 50 ml polycarbonate screw top centrifuge tubes (Corning). Forty samples were prepared each containing 5.0 g of sewage granules and 40 ml of 5 mg l1 oestrogen solution. Fourteen blank samples were also prepared containing 5.0 g of granules and 40 ml of Nanopure water in order to determine oestrogens already present in the sludge granules that could possibly leach into the aqueous phase. Samples placed on an orbital shaker (EdmundBuhler 7400 tubingen KL2) at 350 rev min1. Samples were removed from the orbital shaker at set time intervals and immediately filtered using a 20 ml polypropylene syringe fitted with a removable filtration unit containing a 25 mm GF/C filter (Whatman). The samples were prepared for analysis using the SPE method, the internal standard was added (Diethyl Stibestrol 5 mg l1) and the samples were analysed. At low concentrations, samples were spiked 1:1 with a 2 mg l1
Table 2 Analytical conditions for the Kd experiments Instrument: Perkin-Elmer Series 410 HPLC with LC90 UV spectrophotometric detector (l 280 nm) Column: 125 3.2 mm Envirosep-pp ODS (purchased from Phenomonex) PAHAqua SPE tubes Elution program Step no.
Time (min)
% Acetonitrile
Water
Gradient
0 1 2
– 3.0 4.5
40 60 100
60 40 0
– 3.0 3.0
standard to verify peak identity and area. Supernatant not analysed immediately was stored in amber bottles at 4 1C for no more than 72 h. An apparent equilibrium was reached at 8 h and maintained to the 60 h limit of the experiment. 3.2.2. Determination of the sorption coefficient The bottle point experiments were repeated with test mixtures of 0.5, 1.0, 2.0, 5.0, and 10 mg l1 oestrogens, each CMBR was removed and analysed at the apparent equilibrium time established in the previous CMBR experiment. From the results, the partition coefficients for each oestrogen were calculated. 4. Results and discussion 4.1. Results for Kow Table 3 shows the results of the experimental values of Kow for the four oestrogens. a-Oestradiol has been omitted from the table as it coeluted with b-oestradiol and has the same values in the EPI suite model. From the remaining experimental values, the Kow and BCFs were calculated. The results for the Kow were compared to those of the default (EPI) and due to the agreement the HPLC was considered a suitable method for determining the partition co-efficient. Toxicokinetic models have been developed that describe and predict the degree of bioconcentration or bioaccumulation of organic pollutants by animals. There appears to be a correlation between the Kow and the BCF, whereby the greater the Kow the greater the BCF, this is described in more detail by Walker et al. (2006). The results show that these compounds have the ability to bioaccumulate in fatty tissue, for example, the 17a-EEME would accumulate to 4100 times that of the ambient
Table 3 Kow values from experimental retention times Compound
Retention time (min)
Log RT
Log Kow (EPISUITE)
Relationship
Acenapthylene Fluorene Phenanthrene Chrysene Benzo[a]pyrene
7.76 12.56 14.73 23.16 27.70
0.90 1.10 1.17 1.37 1.44
3.94 4.02 4.35 5.52 6.11
y ¼ 4.0125 R2 ¼ 0.979
Log Kow calculated from retention time (using linear relationship)
Log Kow
Log BCF
b-Oestradiol
8.71
0.94
Oestrone
6.70
0.83
17a-EE
10.09
1.00
17a-EE-ME
14.09
1.15
3.77 3.94/4.01a 3.13 3.43/3.13a 4.03 4.12/3.67a 4.61 4.68/–a
1.44 2.39a 0.95 1.71a 1.63 2.13a 2.07 2.90a
a
Calculated from EPISUITE.
Log BCF ¼ 0.756 log Kow1.145
ARTICLE IN PRESS H.E. Keenan et al. / Environmental Research 106 (2008) 313–318
317
Table 4 Experimental data for sorption coefficients and calculation of Koc Concentration of test mix (mg l1)
Concentration in aqueous phase at t ¼ 8 h, Ce (mg l1)
% Loss of original concentration
Kd (mg kg1) (gradient)
Koc ¼ Kd 100/TOC% (%TOC ¼ 30.82%)
b-Oestradiol 0.5, 1.0, 2.0, 5.0, 10.0
3.33, 8.43,11.62, 26.15, 48.40
99.3, 99.2,99.4, 99.5, 99.5 (99.4)
39,761
129,010 (log Koc ¼ 5.11)
a-Oestradiol 0.5, 1.0, 2.0, 5.0, 10.0
4.20, 7.69, 9.92, 24.3, 60.6
99.2, 99.2, 99.5, 99.5, 99.5 (99.4)
33,920
110,058 (log Koc ¼ 5.04)
17a-EE 0.5, 1.0, 2.0, 5.0, 10.0
5.84, 6.90, 8.29, 18.32, 28.75
98.8, 99.3, 99.6, 99.6, 99.7 (99.4)
61,585
199,821 (log Koc ¼ 5.30)
Oestrone 0.5, 1.0, 2.0, 5.0, 10.0
2.60, 3.80, 7.13, 18.9, 42.1
99.5, 99.6, 99.6, 99.6, 99.6 (99.6)
48,387
156,998 (log Koc ¼ 5.20)
17a-EE-3-ME 0.5, 1.0, 2.0, 5.0, 10.0
1.9, 2.15, 2.36, 4.87, 8.38
99.6, 99.8, 99.9, 99.9, 99.9 (99.8)
218,020
707,397 (log Koc ¼ 5.85)
Table 5 Comparison of experimental (EXP) values and default calculated (C) values from EPISUITE Name
b-Oestradiol Oestrone 17a-EE 17a-EE-3-ME
Log Koc
Total removal (%)
Sludge adsorbance (%)
C
EXP C
EXP
C
EXP
4.21 4.48 4.67 4.68
5.04 5.20 5.30 5.83
490 490 490 490
23.23 5.82 15.99 53.03
99.40 99.60 99.40 99.80
66.95 32.92 31.23 80.44
The Koc values show the transportation of the oestrogenic compounds are preferentially bound to the organic content in the solid particles rather than being dissolved in the aqueous phase. This suggests that in the treatment of contaminated sewage it would be more effective to treat solid particles rather than the aqueous phase. 5. Conclusions A number of conclusions arise from this work:
concentration. The natural oestrogenic compounds are not as bioaccumulative as the synthetic compounds, due to their greater water solubility (excreted via the kidneys). Thus, the Kow values are suggesting the prediction of the order of accumulation ability of five oestrogens. However, the level of oestrogens accumulate in the organism does not relate to the toxicity due to the different susceptibility of specific organism.
4.2. Results of the determination of Kd and Koc An apparent equilibrium time is reached after 8 h and maintained for the 60 h duration of the experiment. As can be seen the gradient is the [x]solids/[x] liquid i.e. Kd (mg kg1). Values of this experiment along with the calculation of Koc are shown in Table 4. All oestrogens showed 490% removal from the aqueous phase at the equilibrium time. The log Koc values are high indicating that the compounds are strongly sorbed. The experimental values were compared to the computational default values (EPA Draft method/BIOWIN). From Table 5, it can be seen that there is a disparity in the experimental values and the values given from the computational model. This is due to the underestimation of the Koc value in the EPISUITE default compared to the experimental value, this impacts greatly on the model prediction as is seen in Table 5.
Log Kow was successfully measured by RPHPLC, the values obtained were close to those calculated by the EPISUITE programme, this is very useful in assessing environmental fate as RPHPLC is a common analytical tool and also due to the correlation with BCF. The experiments showed that equilibrium was achieved in 8 h and this is sufficient for the normal operational time of a sewage treatment plant. The sorption was independent of initial concentration with 490% of each oestrogen removed by this mechanism. The high Koc values obtained would substantiate the strong sorption potential of the sorbent. Large Koc values are indicative of strong sorption, this is concomitant with recalcitrance and therefore greater environmental persistence. Although this renders most compounds non-bioavailable, the organic carbon may break down through time. Oestrogenic compounds are bioaccumulative with log BCF values generally 41 and thus have the potential to biomagnify through the food chain. Synthetic oestrogens are more problematic than the naturally occuring compounds in terms of bioaccumulation, biomagnification and persistence. The experimental values obtained for sludge adsorbance were much higher than predicted by the BIOWIN model.
ARTICLE IN PRESS 318
H.E. Keenan et al. / Environmental Research 106 (2008) 313–318
Modelling based on chemical and physical properties although a useful tool should never replace realistic experimentation. One of the most significant findings of this work is the underestimate of Koc in the computational model (Table 5). This then suggests that a higher concentration of these compounds will exist in the aqueous phase. This is not the case and more effort is needed in the safe disposal of the solids. There is a danger that scientists may move more towards computational models. Other work by the authors has also shown that the default model was inaccurate and cannot replace experimentation for a more exact prediction of environmental fate (Keenan et al., 2006).
This study shows that sewage sludge disposal must be given careful consideration to prevent further exposure to potentially damaging chemicals. In particular the practice of spreading sewage waste on agricultural land although regulated must cease (Department of Environment, 1989; SEPA, 1998). This practice has the potential to contaminate water supplies from runoff and food supplies from vegetation and grazing animals: both milk and meat (Harms, 1996; Wild and Jones, 1992). Oestrogenic compounds are bioaccumulative and may biomagnify through the food chain resulting in adverse physiological affects. Accumulation into milk may be particularly worrying as it is fed to infants and children and their immune systems are not fully developed, therefore the physiological effects may be more serious. Similarly sewage waste buried in landfill may contaminate the environment through poor management of leachate.
References APHA (American Public Health Association), AWWA (American Water Works Association) and WEF (Water Environment Federation), 1995. In: Eaton, A.D., Clesceri. L.S., Greenberg, A.E. (Eds.), Standard Methods for the Examination of Water and Wastewater. Washington, ISBN 0875532233. ASTM International, 2001. E 1195-01, Standard Test Method for Determining a Sorption constant (Koc) for an Organic Chemical in Soil and Sediment. West Conshohocken, PA, US. Department of Environment, 1989. Code of Practice for Agricultural Use of Sewage Sludge. HMSO, ISBN 185112005X. Geyer, H.J., Scheunert, I., Korte, F., 1987. Correlation between the bioconcentration potential of organic environmental chemicals in humans and their n-octanol/water coefficients. Chemosphere 16, 239–252. Harms, H.H., 1996. Bioaccumulation and metabolic fate of sewage sludge derived organicxenobiotics in plants. Sci. Total Environ. 185, 83–92. IEH (Institute for Environment and Health), 1999. The Ecological Significance of Endocrine Disruption: Effects on Reproductive Function and Consequences for Natural Populations. Institute for Environment and Health, University of Leicester, Leicester, UK ISBN 1899110151. Keenan, H.E., Sentenc, P., Songsasen, A., Sakultantimetha, A., Bangkedphol, S., 2006. Monitoring and modeling of metals and PAH contaminants in Thai: Laos Mekong River. In: Simos, T., Maroulis, G. (Eds.), Recent Progress in Computational Sciences and Engineering, ICCMSE2006, pp. 682–689. Samiullah, Y., 1990. Prediction of the Environmental Fate of Chemicals. Elsevier Applied Science, London/New York ISBN 1851664505. SEPA (Scottish Environmental Protection Agency), 1998. Strategic Review of Organic Waste Spread on Land. Walker, C.H., Hopkin, S.P., Sibly, R.M., Peakall, D.B., 2006. Principles of Ecotoxicology, third ed. CRC, Taylor & Francis, Boca Raton, FL ISBN 084933635X. Wild, S.R., Jones, K.C., 1992. Organic chemicals entering agricultural soils in sewage sludges:screening for their potential to transfer to crop plants and livestock. Sci. Total Environ. 119, 85–119.
Environmental Research 106 (2008) 313–318 www.elsevier.com/locate/envres
Environmental fate and partition co-efficient of oestrogenic compounds in sewage treatment process H.E. Keenan, A. Sakultantimetha, S. Bangkedphol Department of Civil Engineering, University of Strathclyde, John Anderson Building, 107 Rottenrow, Glasgow, Scotland G4 ONG, UK Received 15 November 2006; received in revised form 9 May 2007; accepted 22 July 2007 Available online 29 October 2007
Abstract The presence of residual pharmaceuticals and environmental endocrine disrupters (EEDs) is increasingly significant due to their impact on human health and wildlife. Of the compounds implicated as EEDs, the most potent in their oestrogenic effect are the natural and synthetic oestrogens. As these compounds will be present in the sewage matrix, it is necessary to establish their fate during sewage treatment with a view of removal and safe disposal to avoid unnecessary exposure. Using methodology developed by the author this paper describes the results of a study undertaken to determine both the Kow and the adsorption characteristics of these oestrogens. The experimental values obtained were compared to a computational default model. However, there was disparity between the default model and the values determined experimentally. This was especially the case in the determination of the Koc which impacts directly on the sludge adsorbance potential. The calculated results ranged from log 4.21 for b-oestradiol to log 4.68 for 17a-EE-3-ME, the experimental results were higher log(5.04–log 5.83), respectively. The implications of the findings in terms of water recycling and sewage sludge disposal are also discussed. r 2007 Elsevier Inc. All rights reserved. Keywords: Oestrogens; Environmental fate; Adsorption isotherm; Soil–water partition coefficient (Kd); Organic-carbon partition coefficient (Koc); Octanol–water partition coefficient (Kow); BioConcentration Factor (BCF); Sewage treatment
1. Introduction Many of the widely used chemicals in the environment may be inflicting adverse hormonal effects in human and other animals. Of particular concern are synthetic oestrogens, pharmaceutical preparations deliberately intended to elicit oestrogenic responses. The compounds 17b-oestradiol and oestrone are natural hormones, 17a-ethinyl oestradiol (17a-EE), and 17a-ethinyl-oestradiol-3-methyl ether (17aEE-3-ME) are synthetic hormones widely used in contraceptive and hormone replacement therapy formulations ultimately excreted and are therefore environmental contaminants. This study is particularly concerned with this group of compounds. Exposure to chemicals can occur through digestion, inhalation or skin contact. Many chemicals will ultimately pass through sewage treatment Corresponding author. Fax: +44 141 553 2066.
E-mail address: [email protected] (H.E. Keenan). 0013-9351/$ - see front matter r 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.envres.2007.07.007
works (STWs), either by excretion or directly from industrial effluents. Further exposure can occur due to recycling of effluent for potable water supplies and from disposal of sludge solids to land. Fig. 1 shows how oestrogens may be transported through the environment leading to further exposure. The eventual environmental fate is determined by chemodynamics. Although these compounds may be present in very small quantities, additive, and synergistic effects have to be considered. To date, many studies have concentrated on the adverse physiological effects exhibited by fish exposed to sewage final effluent (IEH, 1999). Little work has been done on the environmental hazards through sewage solids, although the oestrogens would be expected to be sorbed strongly to the solid matrix of the sewage rather than remain in the liquid fraction based on their chemical and physical properties. Therefore, this paper is concerned with routes A, E, and F as shown in Fig. 1.
ARTICLE IN PRESS 314
H.E. Keenan et al. / Environmental Research 106 (2008) 313–318
Fig. 1. Transportation of oestrogens.
2. Theory Many different models have been used in an attempt to predict the environmental fate of chemicals (Samiullah, 1990). A simple approach to environmental partitioning can be explained by the concept of fugacity. This approach although based on empirical values determines partitioning of a compound by ratio, it may be particularly relevant to natural and synthetic oestrogens as the concentration of these compounds at any particular STWs will be variant. The relationship between fugacity and partition coefficients is shown in Fig. 2. This approach (level 3) is used by the US EPA in their EPISUITE environmental fate modeler (http://www.epa.gov/opptintr/exposure/docs/EPISuitedl.htm). Without experimental data, the model produced is useful for prediction but is essentially a default model. A more accurate assessment is possible by measuring the partition co-efficients and entering the experimental values obtained into the model as oppose to the default values. This research determined both the solid/water partition co-efficient Kd (also known as Kp, which is represented as CS in Fig. 2. The Koc was calculated from this) and also, the Kow, the octanol/water partition co-efficient (represented as CO in Fig. 2) experimentally for natural and synthetic oestrogens in a sewage sludge. The actual models obtained from the input of the experimental data were then compared to the default models. The Kow value is defined as the ratio of a chemicals concentration in the octanol phase compared to that of its concentration in the aqueous phase of a two-phase octanol–water system. There is a correlation between Kow and the retention time from a reverse phase high performance liquid chromatographic (HPLC) system. A linear regression between the elution times of each compound and the Kow value can be established. This is a useful parameter in environmental prediction as it is analogous with the lipid/water membrane in humans. It shows strong correlation with the BioConcentration Factor (BCF). One proposed correlation illustrated a linear
Fig. 2. Relationship between fugacity constants and partition coefficients (Samiullah, 1990).
relationship between compounds with a log Kow from 102 to 106 of: log(BCF) ¼ 0.756 log Kow1.415 for human adipose tissue (Geyer et al., 1987). During sewage treatment, a chemical will partition between the solid: liquid fractions (Kd). The chemical would normally have time to equilibrate within the operating time of the STW. Although it is common to measure the Kd, for strongly hydrophobic compounds the partitioning is strongly influenced by the organic content of the solid; therefore, Koc is a better indicator of environmental fate. The Koc was measured using a standard method (ASTM International, 2001).
ARTICLE IN PRESS H.E. Keenan et al. / Environmental Research 106 (2008) 313–318
315
3. Experimental
3.1. Determination of Kow
The oestrogens chosen for the investigation were 17boestradiol, 17a-oestradiol, oestrone, 17a-ethinyl oestradiol, and 17a-ethinyl-oestradiol-3-methyl ether (Mestranol), and the internal standard Diethyl Stibestrol (DES) (Sigma Aldrich). The stock solutions of oestrogens were prepared in methanol, working solution was made by serial dilution in water. Initially, the oestrogens were scanned (Phillips PU8740 UV/Vis spectrophotometer) to determine the optimum wavelength for detection i.e. lmax. All oestrogens showed similar UV characteristics with several absorbance peaks in the spectra. Due to interference from the sewage matrix, the lmax 196 nm could not be used, but l 280 nm proved suitable. However, the loss of sensitivity resulted in a concentration step, this was achieved using solid phase extraction (SPE). Therefore, an SPE–RPHPLC (UV detection) with gradient elution method of qualitative and quantitative analysis was developed (MallinckrodtBaker, Bakerfacts, 11.1, 1996) and subsequently used for the determination of Kd. The Kow analysis required no concentration and was performed at l 254 nm.
The Kow was determined from the retention time of the HPLC analysis. Five polyaromatic hydrocarbons (PAHs) of known Kow (values taken from EPISUITE) were purchased from Sigma Aldrich (Acenapthylene, Fluorene, Phenanthrene, Chrysene, and Benzo[a]pyrene). A mixed 1 mg l1 standard containing the PAHs and oestrogens was prepared in HPLC grade methanol. This was analysed by RPHPLC using the analytical method summarised in Table 1. The log of retention time of the five PAHs were plotted against their Kow values and from the linear equation the Kow of the oestrogens was calculated. 3.2. Determination of Kd and Koc Thermised sewage sludge pellets were obtained from a Regional Council Water Services department. The pellets were produced as a garden fertiliser for use by the general public. Analysis of the pellets showed a total organic carbon (TOC) value of 30.82% this analysis was performed externally. The Kd was determined in accordance with ASTM standard testing method (ASTM International, 2001). The method is in two parts and involves equilibrating the chemical between known amounts of water and solid. Initially, an equilibrium time was established for the oestrogens in the solid:water, then sorption coefficients are established (Fig. 3). Initially, the granules were oven dried at 105 1C to constant weight, then sieved to o2 mm and stored in a desiccator at room temperature until required. Plate counts were carried out on nutrient agar to ensure that the granules were sterile (APHA, AWWA and WEF, 1995). The analytical conditions for the experiments are given in Table 2.
Table 1 Analytical conditions for the determination of Kow Instrument: Perkin-Elmer Series 410 HPLC with LC90 UV spectrophotometric detector (l 254 nm) Column: 125 3.2 mm Envirosep-pp ODS (purchased from Phenomonex) Elution program Time (min)
% Acetonitrile
Water
Gradient
0 1 2 3
– 8.0 8.0 6.5
40 60 100 100
60 40 0 0
– 3.0 3.0 –
3.2.1. Determination of equilibrium time A 1000 mg l1 stock solution of the five oestrogens was prepared in HPLC grade methanol. From this, a 5 mg l1 solution in Nanopure water was made for the experimental 2500
1.4 b-oestradid
1.2 1
y = 39.761x
2000
a-oestradid
2 R = 0.9895
µ x/m µg/g
17aEE Oestrone
0.8
17aEE3me
0.6
y = 33.92x
1500
2 R = 0.985
y = 61.585x
b-oestradiol a-oestradiol 17a-EE Estrone 17a-EEME
1000
0.4 500
60
54
48
39
28
24
20
22
8
6
4
3
2
1.5
1
0
0
0.2 0.5
Ratio of oestrogens sample/DES
Step no.
0
0
Time (hours)
20
40 Ce µg/l µ
Fig. 3. Equilibrium time and sorption isotherms.
60
R2 = 0.9316 y = 48.387x 2 R = 0.9957
y = 218.02x R2 = 0.9268
80
ARTICLE IN PRESS H.E. Keenan et al. / Environmental Research 106 (2008) 313–318
316
reactors. Bottle-point experiments were used to determine each point on the adsorption isotherm in an individual completely mixed batch reactor (CMBR) by equilibrating a given solution of solute with a given quantity of sorbent. Individual samples were prepared in 50 ml polycarbonate screw top centrifuge tubes (Corning). Forty samples were prepared each containing 5.0 g of sewage granules and 40 ml of 5 mg l1 oestrogen solution. Fourteen blank samples were also prepared containing 5.0 g of granules and 40 ml of Nanopure water in order to determine oestrogens already present in the sludge granules that could possibly leach into the aqueous phase. Samples placed on an orbital shaker (EdmundBuhler 7400 tubingen KL2) at 350 rev min1. Samples were removed from the orbital shaker at set time intervals and immediately filtered using a 20 ml polypropylene syringe fitted with a removable filtration unit containing a 25 mm GF/C filter (Whatman). The samples were prepared for analysis using the SPE method, the internal standard was added (Diethyl Stibestrol 5 mg l1) and the samples were analysed. At low concentrations, samples were spiked 1:1 with a 2 mg l1
Table 2 Analytical conditions for the Kd experiments Instrument: Perkin-Elmer Series 410 HPLC with LC90 UV spectrophotometric detector (l 280 nm) Column: 125 3.2 mm Envirosep-pp ODS (purchased from Phenomonex) PAHAqua SPE tubes Elution program Step no.
Time (min)
% Acetonitrile
Water
Gradient
0 1 2
– 3.0 4.5
40 60 100
60 40 0
– 3.0 3.0
standard to verify peak identity and area. Supernatant not analysed immediately was stored in amber bottles at 4 1C for no more than 72 h. An apparent equilibrium was reached at 8 h and maintained to the 60 h limit of the experiment. 3.2.2. Determination of the sorption coefficient The bottle point experiments were repeated with test mixtures of 0.5, 1.0, 2.0, 5.0, and 10 mg l1 oestrogens, each CMBR was removed and analysed at the apparent equilibrium time established in the previous CMBR experiment. From the results, the partition coefficients for each oestrogen were calculated. 4. Results and discussion 4.1. Results for Kow Table 3 shows the results of the experimental values of Kow for the four oestrogens. a-Oestradiol has been omitted from the table as it coeluted with b-oestradiol and has the same values in the EPI suite model. From the remaining experimental values, the Kow and BCFs were calculated. The results for the Kow were compared to those of the default (EPI) and due to the agreement the HPLC was considered a suitable method for determining the partition co-efficient. Toxicokinetic models have been developed that describe and predict the degree of bioconcentration or bioaccumulation of organic pollutants by animals. There appears to be a correlation between the Kow and the BCF, whereby the greater the Kow the greater the BCF, this is described in more detail by Walker et al. (2006). The results show that these compounds have the ability to bioaccumulate in fatty tissue, for example, the 17a-EEME would accumulate to 4100 times that of the ambient
Table 3 Kow values from experimental retention times Compound
Retention time (min)
Log RT
Log Kow (EPISUITE)
Relationship
Acenapthylene Fluorene Phenanthrene Chrysene Benzo[a]pyrene
7.76 12.56 14.73 23.16 27.70
0.90 1.10 1.17 1.37 1.44
3.94 4.02 4.35 5.52 6.11
y ¼ 4.0125 R2 ¼ 0.979
Log Kow calculated from retention time (using linear relationship)
Log Kow
Log BCF
b-Oestradiol
8.71
0.94
Oestrone
6.70
0.83
17a-EE
10.09
1.00
17a-EE-ME
14.09
1.15
3.77 3.94/4.01a 3.13 3.43/3.13a 4.03 4.12/3.67a 4.61 4.68/–a
1.44 2.39a 0.95 1.71a 1.63 2.13a 2.07 2.90a
a
Calculated from EPISUITE.
Log BCF ¼ 0.756 log Kow1.145
ARTICLE IN PRESS H.E. Keenan et al. / Environmental Research 106 (2008) 313–318
317
Table 4 Experimental data for sorption coefficients and calculation of Koc Concentration of test mix (mg l1)
Concentration in aqueous phase at t ¼ 8 h, Ce (mg l1)
% Loss of original concentration
Kd (mg kg1) (gradient)
Koc ¼ Kd 100/TOC% (%TOC ¼ 30.82%)
b-Oestradiol 0.5, 1.0, 2.0, 5.0, 10.0
3.33, 8.43,11.62, 26.15, 48.40
99.3, 99.2,99.4, 99.5, 99.5 (99.4)
39,761
129,010 (log Koc ¼ 5.11)
a-Oestradiol 0.5, 1.0, 2.0, 5.0, 10.0
4.20, 7.69, 9.92, 24.3, 60.6
99.2, 99.2, 99.5, 99.5, 99.5 (99.4)
33,920
110,058 (log Koc ¼ 5.04)
17a-EE 0.5, 1.0, 2.0, 5.0, 10.0
5.84, 6.90, 8.29, 18.32, 28.75
98.8, 99.3, 99.6, 99.6, 99.7 (99.4)
61,585
199,821 (log Koc ¼ 5.30)
Oestrone 0.5, 1.0, 2.0, 5.0, 10.0
2.60, 3.80, 7.13, 18.9, 42.1
99.5, 99.6, 99.6, 99.6, 99.6 (99.6)
48,387
156,998 (log Koc ¼ 5.20)
17a-EE-3-ME 0.5, 1.0, 2.0, 5.0, 10.0
1.9, 2.15, 2.36, 4.87, 8.38
99.6, 99.8, 99.9, 99.9, 99.9 (99.8)
218,020
707,397 (log Koc ¼ 5.85)
Table 5 Comparison of experimental (EXP) values and default calculated (C) values from EPISUITE Name
b-Oestradiol Oestrone 17a-EE 17a-EE-3-ME
Log Koc
Total removal (%)
Sludge adsorbance (%)
C
EXP C
EXP
C
EXP
4.21 4.48 4.67 4.68
5.04 5.20 5.30 5.83
490 490 490 490
23.23 5.82 15.99 53.03
99.40 99.60 99.40 99.80
66.95 32.92 31.23 80.44
The Koc values show the transportation of the oestrogenic compounds are preferentially bound to the organic content in the solid particles rather than being dissolved in the aqueous phase. This suggests that in the treatment of contaminated sewage it would be more effective to treat solid particles rather than the aqueous phase. 5. Conclusions A number of conclusions arise from this work:
concentration. The natural oestrogenic compounds are not as bioaccumulative as the synthetic compounds, due to their greater water solubility (excreted via the kidneys). Thus, the Kow values are suggesting the prediction of the order of accumulation ability of five oestrogens. However, the level of oestrogens accumulate in the organism does not relate to the toxicity due to the different susceptibility of specific organism.
4.2. Results of the determination of Kd and Koc An apparent equilibrium time is reached after 8 h and maintained for the 60 h duration of the experiment. As can be seen the gradient is the [x]solids/[x] liquid i.e. Kd (mg kg1). Values of this experiment along with the calculation of Koc are shown in Table 4. All oestrogens showed 490% removal from the aqueous phase at the equilibrium time. The log Koc values are high indicating that the compounds are strongly sorbed. The experimental values were compared to the computational default values (EPA Draft method/BIOWIN). From Table 5, it can be seen that there is a disparity in the experimental values and the values given from the computational model. This is due to the underestimation of the Koc value in the EPISUITE default compared to the experimental value, this impacts greatly on the model prediction as is seen in Table 5.
Log Kow was successfully measured by RPHPLC, the values obtained were close to those calculated by the EPISUITE programme, this is very useful in assessing environmental fate as RPHPLC is a common analytical tool and also due to the correlation with BCF. The experiments showed that equilibrium was achieved in 8 h and this is sufficient for the normal operational time of a sewage treatment plant. The sorption was independent of initial concentration with 490% of each oestrogen removed by this mechanism. The high Koc values obtained would substantiate the strong sorption potential of the sorbent. Large Koc values are indicative of strong sorption, this is concomitant with recalcitrance and therefore greater environmental persistence. Although this renders most compounds non-bioavailable, the organic carbon may break down through time. Oestrogenic compounds are bioaccumulative with log BCF values generally 41 and thus have the potential to biomagnify through the food chain. Synthetic oestrogens are more problematic than the naturally occuring compounds in terms of bioaccumulation, biomagnification and persistence. The experimental values obtained for sludge adsorbance were much higher than predicted by the BIOWIN model.
ARTICLE IN PRESS 318
H.E. Keenan et al. / Environmental Research 106 (2008) 313–318
Modelling based on chemical and physical properties although a useful tool should never replace realistic experimentation. One of the most significant findings of this work is the underestimate of Koc in the computational model (Table 5). This then suggests that a higher concentration of these compounds will exist in the aqueous phase. This is not the case and more effort is needed in the safe disposal of the solids. There is a danger that scientists may move more towards computational models. Other work by the authors has also shown that the default model was inaccurate and cannot replace experimentation for a more exact prediction of environmental fate (Keenan et al., 2006).
This study shows that sewage sludge disposal must be given careful consideration to prevent further exposure to potentially damaging chemicals. In particular the practice of spreading sewage waste on agricultural land although regulated must cease (Department of Environment, 1989; SEPA, 1998). This practice has the potential to contaminate water supplies from runoff and food supplies from vegetation and grazing animals: both milk and meat (Harms, 1996; Wild and Jones, 1992). Oestrogenic compounds are bioaccumulative and may biomagnify through the food chain resulting in adverse physiological affects. Accumulation into milk may be particularly worrying as it is fed to infants and children and their immune systems are not fully developed, therefore the physiological effects may be more serious. Similarly sewage waste buried in landfill may contaminate the environment through poor management of leachate.
References APHA (American Public Health Association), AWWA (American Water Works Association) and WEF (Water Environment Federation), 1995. In: Eaton, A.D., Clesceri. L.S., Greenberg, A.E. (Eds.), Standard Methods for the Examination of Water and Wastewater. Washington, ISBN 0875532233. ASTM International, 2001. E 1195-01, Standard Test Method for Determining a Sorption constant (Koc) for an Organic Chemical in Soil and Sediment. West Conshohocken, PA, US. Department of Environment, 1989. Code of Practice for Agricultural Use of Sewage Sludge. HMSO, ISBN 185112005X. Geyer, H.J., Scheunert, I., Korte, F., 1987. Correlation between the bioconcentration potential of organic environmental chemicals in humans and their n-octanol/water coefficients. Chemosphere 16, 239–252. Harms, H.H., 1996. Bioaccumulation and metabolic fate of sewage sludge derived organicxenobiotics in plants. Sci. Total Environ. 185, 83–92. IEH (Institute for Environment and Health), 1999. The Ecological Significance of Endocrine Disruption: Effects on Reproductive Function and Consequences for Natural Populations. Institute for Environment and Health, University of Leicester, Leicester, UK ISBN 1899110151. Keenan, H.E., Sentenc, P., Songsasen, A., Sakultantimetha, A., Bangkedphol, S., 2006. Monitoring and modeling of metals and PAH contaminants in Thai: Laos Mekong River. In: Simos, T., Maroulis, G. (Eds.), Recent Progress in Computational Sciences and Engineering, ICCMSE2006, pp. 682–689. Samiullah, Y., 1990. Prediction of the Environmental Fate of Chemicals. Elsevier Applied Science, London/New York ISBN 1851664505. SEPA (Scottish Environmental Protection Agency), 1998. Strategic Review of Organic Waste Spread on Land. Walker, C.H., Hopkin, S.P., Sibly, R.M., Peakall, D.B., 2006. Principles of Ecotoxicology, third ed. CRC, Taylor & Francis, Boca Raton, FL ISBN 084933635X. Wild, S.R., Jones, K.C., 1992. Organic chemicals entering agricultural soils in sewage sludges:screening for their potential to transfer to crop plants and livestock. Sci. Total Environ. 119, 85–119.