- Email: [email protected]
Occurrence and treatment trials of endocrine disrupting chemicals (EDCs) in wastewaters
Chemosphere 61 (2005) 544–550 www.elsevier.com/locate/chemosphere
Occurrence and treatment trials of endocrine disrupting chemicals (EDCs) in wastewaters J.Q. Jiang
a,*
, Q. Yin a, J.L. Zhou b, P. Pearce
c
a
c
School of Engineering (C5), University of Surrey, Guildford, Surrey GU2 7XH, UK b Department of Biology and Environmental Science, School of Life Sciences, University of Sussex, Falmer, Brighton BN1 9QJ, UK Research and Development, Thames Water, Spencer House, Manor Farm Road, Reading RG2 0JN, UK Received 8 October 2004; received in revised form 28 January 2005; accepted 11 February 2005 Available online 12 April 2005
Abstract This study demonstrates that both synthetic and natural endocrine disrupting chemicals (EDCs) (e.g., bisphenol A, estrone and 17b-estradiol) were found in the crude wastewaters from two wastewater treatment works (WwTWs). Conventional biological processes can lower EDCs concentrations to several tens to hundreds ng l1. Since natural EDCs (e.g., estrone and 17b-estradiol) have biological activity and adverse impact on the environment at extremely low concentrations (several tens of ng l1), further treatment after conventional biological processes is required. Preliminary trials with ferrate(VI) and electrochemical oxidation process demonstrated that both processes can effectively reduce EDCs to very low levels, ranging between 10 and 100 ng l1, but the former is more effective than the latter to reduce COD concentration in wastewater for given studying conditions. 2005 Elsevier Ltd. All rights reserved. Keywords: EDCs; Wastewater treatment; Ferrate(VI); Electrochemical oxidation
1. Introduction Since the middle of last decade, a variety of adverse effects of endocrine disrupting chemicals (EDCs) on the endocrine systems of man and animals have been observed, which are of particular environmental concern (e.g., Piva and Martini, 1998; Thorpe et al., 2001). These effects may be cumulative, possibly will only appear in subsequent generations, then the resulting effects may be irreversible, threatening the humanÕs sustainable development. Most EDCs are synthetic organic chemi* Corresponding author. Tel.: +44 (0) 1483 686609; fax: +44 (0) 1483 450984. E-mail address: [email protected] (J.Q. Jiang).
cals being introduced to the environment by anthropogenic inputs, (e.g., bisphenol A) but they can also be naturally generated estrogenic hormones, e.g., estrone and 17b-estradiol, and therefore are ubiquitous in aquatic environments receiving wastewater effluents. The structures and physicochemical properties of EDCs influence their fate and behaviour. As shown in Fig. 1, estrone (E1) and 17b-estradiol (E2) consist three hexagonal rings (A, B, C) and one pentagonal ring (D); both E1 and E2 are characterised by their phenolic ring (A). And bisphenol A consists of two phenolic rings jointed together by a bridging carbon atom, and with OH groups in the para position. Both phenolic ring in E1 or E2 and OH groups in bisphenol A are crucial for high affinity binding to the acceptor sites of the
0045-6535/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2005.02.029
J.Q. Jiang et al. / Chemosphere 61 (2005) 544–550
O
OH C A
545
D
B (a)
HO
(b)
HO
CH3 HO
C
OH
(c)
CH3 Fig. 1. Structures of three EDCs. (a) 17b-estradiol (E2), (b) estrone (E1), (c) bisphenol A.
estrogen receptor, resulting in disrupted endocrine systems. Considering the serious adverse impacts of EDCs on the human health and environment, it is urgent to study an efficient approach for removing such chemicals so as to minimise their entry into natural waters. Due to the adverse effects of EDCs even at low concentrations, preliminary researches have been carried out to remove EDCs in wastewaters. The initial investigation suggests that reverse osmosis may be an effective means among ozonation, microfiltration and nanofiltration of removing a wider range of hormones and pharmaceutically active residuals from treated wastewater (Khan et al., 2004). Obviously, more studies are required for a full evaluation and exploration of potential techniques to remove EDCs. This paper explores the findings from a research, where a preliminary survey was carried out to investigate the concentration levels of EDCs presenting in the influent and effluent of two selected wastewater treatment works (WwTW) and to assess if the current treatment processes (e.g., trickling filter and activated sludge process) could effectively remove EDCs. In addition to this, this paper explores the potential of using alternative technologies, ferrate(VI) oxidation and an electrochemical oxidation, to degrade/remove EDCs. The study results could provide useful information about the potential alternative approaches for removing EDCs and thus to minimise their entry into natural waters.
2. Experimental methods and procedures 2.1. EDCs survey and analysis The selected wastewater samples were collected from different sampling points on various days at two WwTWs, and were delivered to the laboratory for analysing EDCs. WwTW1 has primary sedimentation, two
stage trickling filters, and two stages of post-sedimentation, whilst WwTW2 has primary sedimentation, an activated sludge process with the post-sedimentation, and sand filtration. The samples taken in WwTW1 were raw wastewater, and the effluents of primary sedimentation and of the first post-sedimentation (it is after first trickling filter and is called humas tank). Samples taken from WwTW2 were raw wastewater and effluents from the primary sedimentation, the activated sludge process and the sand filter. All samples were collected three times over a four-month duration. The analysis of EDCs follows an established protocol using solid phase extraction (SPE) together with GC– MS analysis (Liu et al., 2004). Wastewater samples were first filtered through a GF/F filter (0.7 lm) to remove particulate matter. The EDCs were extracted from the filtrates by Waters SPE cartridges (Oasis). All the cartridges were first conditioned with 5 ml of ethyl acetate to remove residual bonding agents, followed by methanol (5 ml) and washed by 5 ml ultrapure water for three times. Then, the samples were extracted through SPE cartridges at a flow rate <5 ml min1, which were subsequently eluted with 10 ml of ethyl acetate. The extracts were blown down to 1 ml under a gentle N2 flow, to which 100 ng each of bisphenol A-d16 and 17b-estradiol-d2 was added as internal standards. The mixtures were then derivatised by the addition of 50 ll each of pyridine (dried with KOH solid) and bis(trimethylsilyl)trifluoroacetamide (BSTFA), followed by heating at 60–70 C for 30 min, before GC–MS analysis. GC–MS analysis was performed using a gas chromatograph (Trace GC 2000, Themoquest CE Instruments, TX, USA) coupled with an ion trap mass spectrometer (Polaris Q, Themoquest CE Instruments, Texas, USA) and an autosampler (AS 2000). A ZB5 (5% diphenyl–95% dimethylpolysiloxane) capillary column of 30 m · 0.25 mm i.d. (0.25 lm film thickness) was used. Helium carrier gas was maintained at a
546
J.Q. Jiang et al. / Chemosphere 61 (2005) 544–550
constant flow rate of 1.5 ml min1. The GC column temperature was programmed from 100 (initial equilibrium time 1 min) to 200 C via a ramp of 10 C min1, 200–260 C at 15 C min1, 260–300 C at 3 C min1 and maintained at 300 C for 2 min. The MS was by electron impact ionisation and operated in selected ion monitoring mode for quantitative analysis. The inlet and MS transfer line temperatures were maintained at 280 C, and the ion source temperature was 250 C. Sample injection (1 ll) was in splitless mode. All data are subject to strict quality control procedures, including the use of extensive recovery tests under varying conditions, with recoveries between 70% and 110% and RSD generally <10%. Procedural blanks were routinely used with each batch of samples and underwent the same process as samples. The method detection limits derived from replicate procedural blanks were 8.5, 2.6, 17.4, 5.6, 11.2, 2.6 and 1.0 ng l1 for 4-tertoctylphenol, 4-nonylphenol, bisphenol A, estrone, 17bestradiol, 17a-ethynylestradiol and 16a-hydroxyestrone, respectively. 2.2. Treatment trials Bisphenol A, 17b-estradiol and estrone were used as targeted EDCs for the treatment feasibility study. For electrochemical oxidation, the working electrodes were made from DSA (dimension stable anode), graphite felt and titanium, whereas the corresponding counter electrodes were titanium, graphite felt and stainless steel respectively. The geometric area of all electrodes was 34 cm2. A double junction Ag/AgCl electrode was used as a reference electrode. The electrode potential or current density was controlled by an Autolab PGSTAT30 with the general purpose electrochemical system (GPES) software. Test solutions containing one of three EDCs at concentrations of either 0.1 or 1 mg l1 were prepared by mixing given amount of EDC with one of electrolytes, either 3 g l1 Na2SO4 solution (which made final solution pH at 7), or 3 g l1 Na2SO4 plus 4 g l1 NaOH solution (which made final solution pH at 13). The wastewater samples from WwTW1 were taken for the treatment trials, which were either primary sedimentation effluent or first humus tank effluent. Before and
after electrochemical treatment, the concentrations of EDCs in test solutions and real samples were analysed. For the treatment of wastewaters, total and soluble COD were also measured. The same test solutions and wastewaters were also used to examine the treatment efficiency of ferrate(VI) oxidation, where a jar test procedure was used to perform the experiments, involving a fast mixing at 400 rpm for 1 min after dosing the ferrate(VI), a slow mixing at 35 rpm for 20 min, and a 60 min period of sedimentation. The pH of test solutions/wastewaters was kept between 7 and 8, and ferrate(VI) doses applied were between 10 and 20 mg l1. After sedimentation, the supernatant was withdrawn for the analysis of EDCs concentrations.
3. Results and discussion 3.1. EDCs survey Both synthetic EDCs (4-tert-octylphenol, 4-nonylphenol, bisphenol A, 16a hydroxyestrone) and natural EDCs (estrone and 17b-estradiol) were found in the raw wastewaters from two WwTWs, and examples of survey results can be seen in Tables 2 and 3. It can be seen from Table 1 that for crude wastewater samples from WwTW1, the concentrations of various synthetic EDCs were between 32 and 451 ng l1, the dominant EDC was bisphenol A (378–451 ng l1), and the concentration of 17b-estradiol was higher than that of estrone. Most EDCs were not significantly removed by primary sedimentation. However, all EDCs in the effluents after trickling filters and post-sedimentation (effluent of humus tank) were significantly reduced. Concentrations of various EDCs in raw wastewaters at WwTW2 were higher than that at WwTW1 (Table 2). The dominant EDCs were bisphenol A (682.8– 890 ng l1) and 4-tert-octylphenol (545.7–611 ng l1). All EDCs in the effluents of activated sludge and sedimentation were greatly reduced; from several hundreds to tens of ng l1. However, sand filter does not help to remove residual EDCs significantly. The natural EDCs (e.g., estrone and 17b-estradiol) were present in the treated effluents at two WwTWs,
Table 1 EDC survey result from WwTW1 (ng l1) Bisphenol A
Estrone
17b-Estradiol
16a-Hydroxyestrone
Crude wastewater
25 06 03 02 07 03
4-tert-Octylphenol 85 112
4-Nonylphenol 32 76
451 378
59 57
224 132
36 44
Settled wastewater
25 06 03 02 07 03
52 710
26 106
359 392
60 45
118 96
23 41
1st humus
25 06 03 02 07 03
59 133
10 37
141 145
21 39
31 32
15 13
J.Q. Jiang et al. / Chemosphere 61 (2005) 544–550
547
Table 2 EDC survey result from WwTW2 (ng l1) 4-tertOctylphenol
4-Nonylphenol
Bisphenol A
Estrone
17-Estradiol
17a-Ethynylestradiol
16a-Hydroxyestrone
Crude wastewater
16 07 03 23 07 03
545.7 611.1
122.3 100.6
890 682.8
81.0 77.8
188.7 182.6
72.4 123.5
71.7 63.7
Settled wastewater
16 07 03 23 07 03
169.7 147.2
98.5 74.1
669 604.6
128.5 77.8
98.3 79.5
26.6 35.6
11.6 35.6
Post-sedimentation effluent
16 07 03 23 07 03
35 41.1
58.9 30.8
15 34.5
87 21.1
48.0 19.5
12.8 14.6
15 13
Sand filter effluent
16 07 03 23 07 03
29.0 48.8
33.2 19.7
12.0 27.6
40.0 22.2
44 10.9
a
a
a
a
a
Data unavailable.
Table 3 Oxidation charge and reduction charge of DSA electrode in 1 mg l1 17b-estradiol solution (pH 7) at sweep at different sweep rates 200 mV s1
100 mV s1
50 mV s1
2 mV s1
0.479 0.4566
0.6516 0.6050
0.8665 0.7151
2.491 0.7106
despite that EDCs concentrations were very low (several tens of ng l1) in the effluents of either the humus tank at WwTW1 or post-sedimentation tank at WwTW2. This is consistent with previous studies (Desbrow et al., 1998; Ternes et al., 1999), where natural estrogenic hormones were demonstrated to be among the most potent of all EDCs and could be stable to survive in conventional wastewater and sewage treatment processes. In addition to this, comparing with synthetic EDCs, which have environmental impact at concentrations of tens and hundreds of lg/l, natural estrogenic hormones are to be biological active at extremely low concentrations (e.g., less than 10 ng l1); this makes them even more threatening and difficult to deal with. A further treatment (after conventional biological treatment processes) has to be applied in order to completely remove natural EDCs. A preliminary trial was thus conducted. 3.2. EDCs degradation with electrochemical oxidation Cyclic voltammetry tests of 17b-estradiol, bisphenol A and estrone at different solution conditions (pH 7 and 13) showed that there were no oxidation current peaks found, i.e., the oxidation charge during sweep in positive direction was approximately equal to the reduction charge during reverse sweep in negative direction. The example of the results can be seen in Fig. 2(a), which shows the voltammograms of DSA electrode in 1 mg l1 of 17b-estradiol solution (pH 7) and at different sweep rates ranging from 2 to 200 mV s1. No current peak for 17b-estradiol oxidation was found in the test. The
30 Current density / A m-2
Oxidation/C Reduction/C
Sweep rate
20
50 mV s-1
100 mV s-1
10 0 -10 2 mV s-1
-20 200 mV s
-1
(a)
-30 30 Current density / A m-2
Charges
20
With magnetic stirring
10 Without magnetic stirring
0 -10 -20
(b) -30 0.2
0.4
0.6 0.8 Electrode potential / V
1.0
1.2
Fig. 2. (a) Cyclic voltammograms of DSA electrode in 1 mg l1 17b-estradiol and sodium sulphate electrolyte of pH 7 at various sweep rates; (b) cyclic voltammograms of DSA electrode in 1 mg l1 17b-estradiol and sodium sulphate electrolyte of pH 7 at sweep rate 100 mV s1.
current varied with time was integrated to obtain oxidation charge and reduction charge in one cycle (listed in Table 3), and the data suggested that when the sweep rate was greater than 50 mV s1, the oxidation charge
548
J.Q. Jiang et al. / Chemosphere 61 (2005) 544–550
in positive direction was approximately equal to the reduction charge in negative direction during a reverse sweep. This implies that the process was essential a reversible electrochemical oxidation and the reduction happened at the electrode surface. However, when the sweep rate was decreased to 2 mV s1, the oxidation charge was increased to 2.491 C, while the reduction charge was almost the same as that at the sweep rate of 50 mV s1. This suggests that the electrode surface was saturated by the oxidation product as did at the sweep rate of 50 mV s1. The further decrease of the sweep rate made the most oxidation product leave the electrode surface and would not be reduced during reverse sweep. Such explanation was supported by the voltammograms obtained with and without stirring in the same electrolyte (Fig. 2(b)). Fig. 2(b) shows that the stirring has very little influence on the oxidation and reduction process, and the electrochemical process was essentially controlled by charge transfer. However, the oxidation of 17b-estradiol might be controlled by mass transport because its concentration in the solution was very low. When the positive potential limit was increased from 1.12 V to 2 V, the oxidation charge in a positive going sweep was four times more than the reduction charge in a reversible negative going sweep, even at a high sweep rate of 200 mV s1, indicating that as the electrode potential became higher, the oxidation product left the electrode surface quicker and the process was less reversible. There was still no current peak for oxidation of 17b-estradiol occurred. However, it did not mean that 17b-estradiol could not be oxidised. When the electrode potential was high enough (e.g., 2.00 V vs. Ag/AgCl, or 2.20 vs. SHE), the rate of oxidation is to be determined by mass transport and the relationship between the mass transport and limiting current density can be expressed by Eq. (1): i ¼ n F k c;
ð1Þ
where i is the limiting current density, n is the number of electrons involved in the reaction, F is Faraday constant, k stands for mass transport coefficient and c for concentration of electro-reactive species. The concentration of 17b-estradiol in the experiment was only 1 mg l1, i.e., 3.67 · 106 M. Assuming that two electrons are involved in the total charge transfer and the mass transport coefficient for 17b-estradiol is 105 m1, the mass transport limiting current density would be only 0.0071 A m2. Therefore, such a small current density occurred after large oxygen evolution would be buried in the water oxidation process. It would be impossible to form a separate visible current peak. Nevertheless, under various studying conditions, the removing efficiency of three targeted EDCs approached to 99.9%. The effluent from primary settling tank was further treated with electrochemical oxidation. For the use of
DSA as anode at current density of 58.8 A m2, and with 3 h electrolysis, majority of EDCs were completely removed, except for 4-nonylphenol and 4-tert-octylphenol. And the solution COD concentrations did not change. However, the UV-254 absorbance was decreased by more than 50%, indicating that the aromatic structure of EDCs have been destroyed but perhaps transformed to form some byproducts. Using graphite felt as anode at current density of 29.4 A m2 could reduce all residual EDCs in the 1st humus tank effluent to very low levels, however, the absorption capacity of graphite felt electrode should be taken into account when considering the general results. Results of using DSA as anode and at current densities of 58.8 and 117.6 A m2 for treating the effluent from first humus tank demonstrated that estrone and 17b-estradiol can be reduced to a very low concentrations (20–100 ng l1), but total and soluble COD concentrations did not change greatly before and after treatment. 3.3. EDCs degradation with ferrate(VI) oxidation Ferrate(VI) ion has molecular formula, FeO2 4 , and is a very strong oxidant. At the acidic conditions, the redox potential of ferrate(VI) ions is greater than ozone (Jiang and Lloyd, 2002). Moreover, during oxidation process, ferrate(VI) ions are reduced to Fe(III) ions or ferric hydroxide, which simultaneously generates a coagulant in a single dosing and mixing unit process. For treating test solutions, ferrate(VI) can effectively reduce three EDCs concentrations to low level (tens of ng l1) at doses ranging 13–17 mg l1 as Fe, and the removal percentage was 99.99%. For treating wastewater effluent from the 1st humus, ferrate(VI) also performed very well at relative low doses (1–5 mg l1 as Fe). As shown in Table 4, all EDCs present in the 1st humus effluent were reduced to a low level, especially for bisphenol A, which has been reduced substantially from 1.2 lg l1 to 46 ng l1. In outline, the optimum ferrate(VI) dose required to achieve the maximum EDCs removal (>99.99%) was 0.10 mg EDC:1 mg ferrate (as Fe) for treating test solutions, and 0.06 mg EDC:1 mg ferrate (as Fe) for treating primary sedimentation effluents. 3.4. Comparative performance in the degradation of EDCs Fig. 3 shows the comparative performance of EDCs reduction with ferrate (sample 2) and electrochemical oxidation (sample 3). It can be seen in Fig. 3(a), that ferrate(VI) can reduce much more bisphenol A, 17bestradiol and 4-tert-octylphenol concentrations than electrochemical oxidation, and should possess greater oxidation potential to degrade EDCs. Fig. 3(b) demonstrates that ferrate(VI) can remove 20% more total COD
J.Q. Jiang et al. / Chemosphere 61 (2005) 544–550
549
Table 4 Analysis of treated sewage samples with ferrate(VI)a (ng l1) Samples a
WwTW1 Supernatant After filtration a b
4-tert-Octylphenol
4-Nonylphenol
Bisphenol A
Estrone
17b-Estradiol
16a-Hydroxyestrone
188.5 51.60
39.90 47.01 15.86
1209 199.67 46.06
30.10 26.25 20.87
52.54 104.2 27.90
63.7
b
b b
The original sewage taken from the effluent of humus tank in WwTW1. Lower than the detect limits.
Concentrations (ng/l)
600
(a)
nanyphenol bisphenol-A
500 400
estrone 17 beta-estradiol
300
4-tert-octylphenol
200 100
COD concentrations (mg/l)
0 200
(b)
Total COD
150
Soluble COD
100
Percentage removal of COD 24% 14% 44% 46%
50
0 1
2
3
Fig. 3. (a) Comparative EDCs concentrations; (b) comparative COD concentrations, 1. wastewater sample taken from the post-sedimentation in WwTW1; 2. treated sample with ferrate oxidation; 3. treated sample with electrochemical oxidation.
and 32% more soluble COD comparing with electrochemical oxidation for giving study conditions. COD represents all organic pollutants existing in wastewaters, a high COD removal efficiency with ferrate(VI) demonstrated that the ferrate(VI) oxidation possesses high capacity of not only degradation of EDCs but also of removing other organic contaminants in wastewaters comparing with electrochemical oxidation.
post-sedimentation can lower EDCs concentrations to a scale of from several hundreds to hundred of ng l1. Since natural EDCs, e.g., estrone and 17b-estradiol, are most potent and have biological activity and adverse impact on the environment at extremely low concentrations (several tens of ng l1), further treatment after conventional biological processes is required. Preliminary trials in treating either EDCs test solutions or wastewaters show that both ferrate(VI) and electrochemical oxidation can effectively reduce EDCs to very low levels, ranging between 10 and 100 ng l1, and the former is more effective than the latter to remove organic contaminants in terms of lowering the effluent COD concentrations. Whilst the results from this study suggest that the ferrate(VI) or electrochemical oxidation could be two potential techniques used for the treatment of EDCs, further investigations should be carried out to examine the treatment efficiency in the use of test solutions with low natural EDCs starting concentrations (e.g., <10 ng l1). In addition to this, the fundamental study on EDCs degradation with other oxidation technologies should be conducted before a full evaluation and assessment are to be conducted in order to select the potential effective technologies for the degradation of EDCs.
Acknowledgements The authors thank the Thames Water, UK, for a grant that provided a research fellowship for Dr. Q. Yin and the financial support for this study. The views expressed in this paper are not necessarily representing that of Thames Water.
References 4. Conclusions This work demonstrates that a range of EDCs were found in the crude wastewaters from two WwTWs in the concentration levels up to 890 ng l1. Primary sedimentation in the processes does not remove EDCs singnificantly, but trickling filter and activated sludge with
Desbrow, C., Routledge, E.J., Brighty, G.C., Sumpter, J.P., Waldock, M., 1998. Identification of estrogenic chemicals in STW effluent. 1. Chemical fractionation and in vitro biological screening. Environ. Sci. Technol. 32, 1549–1558. Jiang, J.Q., Lloyd, B., 2002. Progress in the development and use of ferrate (vi) salt as an oxidant and coagulant for water and wastewater treatment. Water Res. 36, 1397–1408.
550
J.Q. Jiang et al. / Chemosphere 61 (2005) 544–550
Khan, S.J., Wintgens, T., Sherman, P., Zaricky, J., Scha¨fer, A.I., 2004. Removal of hormones and pharmaceuticals in the advanced water recycling demonstration plant in Queensland, Australia. Water Sci. Technol. 50, 15–22. Liu, R., Zhou, J.L., Wilding, A., 2004. Simultaneous determination of endocrine disrupting phenolic compounds and steroids in water by solid-phase extraction–gas chromatography–mass spectrometry. J. Chromatogr. A 1022, 179–189. Piva, F., Martini, L., 1998. Neurotransmitters and the control of hypophyseal gonadal functions: possible implications of endocrine disruptors. Pure Appl. Chem. 70, 1647–1656.
Ternes, T.A., Stumpf, M., Mueller, J., Haberer, K., Wilken, R.-D., Servos, M., 1999. Behaviour and occurrence of estrogens in municipal sewage treatment plants—I. Investigations in Germany, Canada and Brazil. Sci. Total Environ. 225, 81–90. Thorpe, K.L., Hutchinson, T.H., Hetheridge, M.J., Scholze, M., Sumpter, J.P., Tyler, C.R., 2001. Assessing the biological potency of binary mixtures of environmental estrogens using vitellogenin induction in juvenile rainbow trout (Oncorhynchus mykiss). Environ. Sci. Technol. 35, 2476– 2481.
Occurrence and treatment trials of endocrine disrupting chemicals (EDCs) in wastewaters J.Q. Jiang
a,*
, Q. Yin a, J.L. Zhou b, P. Pearce
c
a
c
School of Engineering (C5), University of Surrey, Guildford, Surrey GU2 7XH, UK b Department of Biology and Environmental Science, School of Life Sciences, University of Sussex, Falmer, Brighton BN1 9QJ, UK Research and Development, Thames Water, Spencer House, Manor Farm Road, Reading RG2 0JN, UK Received 8 October 2004; received in revised form 28 January 2005; accepted 11 February 2005 Available online 12 April 2005
Abstract This study demonstrates that both synthetic and natural endocrine disrupting chemicals (EDCs) (e.g., bisphenol A, estrone and 17b-estradiol) were found in the crude wastewaters from two wastewater treatment works (WwTWs). Conventional biological processes can lower EDCs concentrations to several tens to hundreds ng l1. Since natural EDCs (e.g., estrone and 17b-estradiol) have biological activity and adverse impact on the environment at extremely low concentrations (several tens of ng l1), further treatment after conventional biological processes is required. Preliminary trials with ferrate(VI) and electrochemical oxidation process demonstrated that both processes can effectively reduce EDCs to very low levels, ranging between 10 and 100 ng l1, but the former is more effective than the latter to reduce COD concentration in wastewater for given studying conditions. 2005 Elsevier Ltd. All rights reserved. Keywords: EDCs; Wastewater treatment; Ferrate(VI); Electrochemical oxidation
1. Introduction Since the middle of last decade, a variety of adverse effects of endocrine disrupting chemicals (EDCs) on the endocrine systems of man and animals have been observed, which are of particular environmental concern (e.g., Piva and Martini, 1998; Thorpe et al., 2001). These effects may be cumulative, possibly will only appear in subsequent generations, then the resulting effects may be irreversible, threatening the humanÕs sustainable development. Most EDCs are synthetic organic chemi* Corresponding author. Tel.: +44 (0) 1483 686609; fax: +44 (0) 1483 450984. E-mail address: [email protected] (J.Q. Jiang).
cals being introduced to the environment by anthropogenic inputs, (e.g., bisphenol A) but they can also be naturally generated estrogenic hormones, e.g., estrone and 17b-estradiol, and therefore are ubiquitous in aquatic environments receiving wastewater effluents. The structures and physicochemical properties of EDCs influence their fate and behaviour. As shown in Fig. 1, estrone (E1) and 17b-estradiol (E2) consist three hexagonal rings (A, B, C) and one pentagonal ring (D); both E1 and E2 are characterised by their phenolic ring (A). And bisphenol A consists of two phenolic rings jointed together by a bridging carbon atom, and with OH groups in the para position. Both phenolic ring in E1 or E2 and OH groups in bisphenol A are crucial for high affinity binding to the acceptor sites of the
0045-6535/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2005.02.029
J.Q. Jiang et al. / Chemosphere 61 (2005) 544–550
O
OH C A
545
D
B (a)
HO
(b)
HO
CH3 HO
C
OH
(c)
CH3 Fig. 1. Structures of three EDCs. (a) 17b-estradiol (E2), (b) estrone (E1), (c) bisphenol A.
estrogen receptor, resulting in disrupted endocrine systems. Considering the serious adverse impacts of EDCs on the human health and environment, it is urgent to study an efficient approach for removing such chemicals so as to minimise their entry into natural waters. Due to the adverse effects of EDCs even at low concentrations, preliminary researches have been carried out to remove EDCs in wastewaters. The initial investigation suggests that reverse osmosis may be an effective means among ozonation, microfiltration and nanofiltration of removing a wider range of hormones and pharmaceutically active residuals from treated wastewater (Khan et al., 2004). Obviously, more studies are required for a full evaluation and exploration of potential techniques to remove EDCs. This paper explores the findings from a research, where a preliminary survey was carried out to investigate the concentration levels of EDCs presenting in the influent and effluent of two selected wastewater treatment works (WwTW) and to assess if the current treatment processes (e.g., trickling filter and activated sludge process) could effectively remove EDCs. In addition to this, this paper explores the potential of using alternative technologies, ferrate(VI) oxidation and an electrochemical oxidation, to degrade/remove EDCs. The study results could provide useful information about the potential alternative approaches for removing EDCs and thus to minimise their entry into natural waters.
2. Experimental methods and procedures 2.1. EDCs survey and analysis The selected wastewater samples were collected from different sampling points on various days at two WwTWs, and were delivered to the laboratory for analysing EDCs. WwTW1 has primary sedimentation, two
stage trickling filters, and two stages of post-sedimentation, whilst WwTW2 has primary sedimentation, an activated sludge process with the post-sedimentation, and sand filtration. The samples taken in WwTW1 were raw wastewater, and the effluents of primary sedimentation and of the first post-sedimentation (it is after first trickling filter and is called humas tank). Samples taken from WwTW2 were raw wastewater and effluents from the primary sedimentation, the activated sludge process and the sand filter. All samples were collected three times over a four-month duration. The analysis of EDCs follows an established protocol using solid phase extraction (SPE) together with GC– MS analysis (Liu et al., 2004). Wastewater samples were first filtered through a GF/F filter (0.7 lm) to remove particulate matter. The EDCs were extracted from the filtrates by Waters SPE cartridges (Oasis). All the cartridges were first conditioned with 5 ml of ethyl acetate to remove residual bonding agents, followed by methanol (5 ml) and washed by 5 ml ultrapure water for three times. Then, the samples were extracted through SPE cartridges at a flow rate <5 ml min1, which were subsequently eluted with 10 ml of ethyl acetate. The extracts were blown down to 1 ml under a gentle N2 flow, to which 100 ng each of bisphenol A-d16 and 17b-estradiol-d2 was added as internal standards. The mixtures were then derivatised by the addition of 50 ll each of pyridine (dried with KOH solid) and bis(trimethylsilyl)trifluoroacetamide (BSTFA), followed by heating at 60–70 C for 30 min, before GC–MS analysis. GC–MS analysis was performed using a gas chromatograph (Trace GC 2000, Themoquest CE Instruments, TX, USA) coupled with an ion trap mass spectrometer (Polaris Q, Themoquest CE Instruments, Texas, USA) and an autosampler (AS 2000). A ZB5 (5% diphenyl–95% dimethylpolysiloxane) capillary column of 30 m · 0.25 mm i.d. (0.25 lm film thickness) was used. Helium carrier gas was maintained at a
546
J.Q. Jiang et al. / Chemosphere 61 (2005) 544–550
constant flow rate of 1.5 ml min1. The GC column temperature was programmed from 100 (initial equilibrium time 1 min) to 200 C via a ramp of 10 C min1, 200–260 C at 15 C min1, 260–300 C at 3 C min1 and maintained at 300 C for 2 min. The MS was by electron impact ionisation and operated in selected ion monitoring mode for quantitative analysis. The inlet and MS transfer line temperatures were maintained at 280 C, and the ion source temperature was 250 C. Sample injection (1 ll) was in splitless mode. All data are subject to strict quality control procedures, including the use of extensive recovery tests under varying conditions, with recoveries between 70% and 110% and RSD generally <10%. Procedural blanks were routinely used with each batch of samples and underwent the same process as samples. The method detection limits derived from replicate procedural blanks were 8.5, 2.6, 17.4, 5.6, 11.2, 2.6 and 1.0 ng l1 for 4-tertoctylphenol, 4-nonylphenol, bisphenol A, estrone, 17bestradiol, 17a-ethynylestradiol and 16a-hydroxyestrone, respectively. 2.2. Treatment trials Bisphenol A, 17b-estradiol and estrone were used as targeted EDCs for the treatment feasibility study. For electrochemical oxidation, the working electrodes were made from DSA (dimension stable anode), graphite felt and titanium, whereas the corresponding counter electrodes were titanium, graphite felt and stainless steel respectively. The geometric area of all electrodes was 34 cm2. A double junction Ag/AgCl electrode was used as a reference electrode. The electrode potential or current density was controlled by an Autolab PGSTAT30 with the general purpose electrochemical system (GPES) software. Test solutions containing one of three EDCs at concentrations of either 0.1 or 1 mg l1 were prepared by mixing given amount of EDC with one of electrolytes, either 3 g l1 Na2SO4 solution (which made final solution pH at 7), or 3 g l1 Na2SO4 plus 4 g l1 NaOH solution (which made final solution pH at 13). The wastewater samples from WwTW1 were taken for the treatment trials, which were either primary sedimentation effluent or first humus tank effluent. Before and
after electrochemical treatment, the concentrations of EDCs in test solutions and real samples were analysed. For the treatment of wastewaters, total and soluble COD were also measured. The same test solutions and wastewaters were also used to examine the treatment efficiency of ferrate(VI) oxidation, where a jar test procedure was used to perform the experiments, involving a fast mixing at 400 rpm for 1 min after dosing the ferrate(VI), a slow mixing at 35 rpm for 20 min, and a 60 min period of sedimentation. The pH of test solutions/wastewaters was kept between 7 and 8, and ferrate(VI) doses applied were between 10 and 20 mg l1. After sedimentation, the supernatant was withdrawn for the analysis of EDCs concentrations.
3. Results and discussion 3.1. EDCs survey Both synthetic EDCs (4-tert-octylphenol, 4-nonylphenol, bisphenol A, 16a hydroxyestrone) and natural EDCs (estrone and 17b-estradiol) were found in the raw wastewaters from two WwTWs, and examples of survey results can be seen in Tables 2 and 3. It can be seen from Table 1 that for crude wastewater samples from WwTW1, the concentrations of various synthetic EDCs were between 32 and 451 ng l1, the dominant EDC was bisphenol A (378–451 ng l1), and the concentration of 17b-estradiol was higher than that of estrone. Most EDCs were not significantly removed by primary sedimentation. However, all EDCs in the effluents after trickling filters and post-sedimentation (effluent of humus tank) were significantly reduced. Concentrations of various EDCs in raw wastewaters at WwTW2 were higher than that at WwTW1 (Table 2). The dominant EDCs were bisphenol A (682.8– 890 ng l1) and 4-tert-octylphenol (545.7–611 ng l1). All EDCs in the effluents of activated sludge and sedimentation were greatly reduced; from several hundreds to tens of ng l1. However, sand filter does not help to remove residual EDCs significantly. The natural EDCs (e.g., estrone and 17b-estradiol) were present in the treated effluents at two WwTWs,
Table 1 EDC survey result from WwTW1 (ng l1) Bisphenol A
Estrone
17b-Estradiol
16a-Hydroxyestrone
Crude wastewater
25 06 03 02 07 03
4-tert-Octylphenol 85 112
4-Nonylphenol 32 76
451 378
59 57
224 132
36 44
Settled wastewater
25 06 03 02 07 03
52 710
26 106
359 392
60 45
118 96
23 41
1st humus
25 06 03 02 07 03
59 133
10 37
141 145
21 39
31 32
15 13
J.Q. Jiang et al. / Chemosphere 61 (2005) 544–550
547
Table 2 EDC survey result from WwTW2 (ng l1) 4-tertOctylphenol
4-Nonylphenol
Bisphenol A
Estrone
17-Estradiol
17a-Ethynylestradiol
16a-Hydroxyestrone
Crude wastewater
16 07 03 23 07 03
545.7 611.1
122.3 100.6
890 682.8
81.0 77.8
188.7 182.6
72.4 123.5
71.7 63.7
Settled wastewater
16 07 03 23 07 03
169.7 147.2
98.5 74.1
669 604.6
128.5 77.8
98.3 79.5
26.6 35.6
11.6 35.6
Post-sedimentation effluent
16 07 03 23 07 03
35 41.1
58.9 30.8
15 34.5
87 21.1
48.0 19.5
12.8 14.6
15 13
Sand filter effluent
16 07 03 23 07 03
29.0 48.8
33.2 19.7
12.0 27.6
40.0 22.2
44 10.9
a
a
a
a
a
Data unavailable.
Table 3 Oxidation charge and reduction charge of DSA electrode in 1 mg l1 17b-estradiol solution (pH 7) at sweep at different sweep rates 200 mV s1
100 mV s1
50 mV s1
2 mV s1
0.479 0.4566
0.6516 0.6050
0.8665 0.7151
2.491 0.7106
despite that EDCs concentrations were very low (several tens of ng l1) in the effluents of either the humus tank at WwTW1 or post-sedimentation tank at WwTW2. This is consistent with previous studies (Desbrow et al., 1998; Ternes et al., 1999), where natural estrogenic hormones were demonstrated to be among the most potent of all EDCs and could be stable to survive in conventional wastewater and sewage treatment processes. In addition to this, comparing with synthetic EDCs, which have environmental impact at concentrations of tens and hundreds of lg/l, natural estrogenic hormones are to be biological active at extremely low concentrations (e.g., less than 10 ng l1); this makes them even more threatening and difficult to deal with. A further treatment (after conventional biological treatment processes) has to be applied in order to completely remove natural EDCs. A preliminary trial was thus conducted. 3.2. EDCs degradation with electrochemical oxidation Cyclic voltammetry tests of 17b-estradiol, bisphenol A and estrone at different solution conditions (pH 7 and 13) showed that there were no oxidation current peaks found, i.e., the oxidation charge during sweep in positive direction was approximately equal to the reduction charge during reverse sweep in negative direction. The example of the results can be seen in Fig. 2(a), which shows the voltammograms of DSA electrode in 1 mg l1 of 17b-estradiol solution (pH 7) and at different sweep rates ranging from 2 to 200 mV s1. No current peak for 17b-estradiol oxidation was found in the test. The
30 Current density / A m-2
Oxidation/C Reduction/C
Sweep rate
20
50 mV s-1
100 mV s-1
10 0 -10 2 mV s-1
-20 200 mV s
-1
(a)
-30 30 Current density / A m-2
Charges
20
With magnetic stirring
10 Without magnetic stirring
0 -10 -20
(b) -30 0.2
0.4
0.6 0.8 Electrode potential / V
1.0
1.2
Fig. 2. (a) Cyclic voltammograms of DSA electrode in 1 mg l1 17b-estradiol and sodium sulphate electrolyte of pH 7 at various sweep rates; (b) cyclic voltammograms of DSA electrode in 1 mg l1 17b-estradiol and sodium sulphate electrolyte of pH 7 at sweep rate 100 mV s1.
current varied with time was integrated to obtain oxidation charge and reduction charge in one cycle (listed in Table 3), and the data suggested that when the sweep rate was greater than 50 mV s1, the oxidation charge
548
J.Q. Jiang et al. / Chemosphere 61 (2005) 544–550
in positive direction was approximately equal to the reduction charge in negative direction during a reverse sweep. This implies that the process was essential a reversible electrochemical oxidation and the reduction happened at the electrode surface. However, when the sweep rate was decreased to 2 mV s1, the oxidation charge was increased to 2.491 C, while the reduction charge was almost the same as that at the sweep rate of 50 mV s1. This suggests that the electrode surface was saturated by the oxidation product as did at the sweep rate of 50 mV s1. The further decrease of the sweep rate made the most oxidation product leave the electrode surface and would not be reduced during reverse sweep. Such explanation was supported by the voltammograms obtained with and without stirring in the same electrolyte (Fig. 2(b)). Fig. 2(b) shows that the stirring has very little influence on the oxidation and reduction process, and the electrochemical process was essentially controlled by charge transfer. However, the oxidation of 17b-estradiol might be controlled by mass transport because its concentration in the solution was very low. When the positive potential limit was increased from 1.12 V to 2 V, the oxidation charge in a positive going sweep was four times more than the reduction charge in a reversible negative going sweep, even at a high sweep rate of 200 mV s1, indicating that as the electrode potential became higher, the oxidation product left the electrode surface quicker and the process was less reversible. There was still no current peak for oxidation of 17b-estradiol occurred. However, it did not mean that 17b-estradiol could not be oxidised. When the electrode potential was high enough (e.g., 2.00 V vs. Ag/AgCl, or 2.20 vs. SHE), the rate of oxidation is to be determined by mass transport and the relationship between the mass transport and limiting current density can be expressed by Eq. (1): i ¼ n F k c;
ð1Þ
where i is the limiting current density, n is the number of electrons involved in the reaction, F is Faraday constant, k stands for mass transport coefficient and c for concentration of electro-reactive species. The concentration of 17b-estradiol in the experiment was only 1 mg l1, i.e., 3.67 · 106 M. Assuming that two electrons are involved in the total charge transfer and the mass transport coefficient for 17b-estradiol is 105 m1, the mass transport limiting current density would be only 0.0071 A m2. Therefore, such a small current density occurred after large oxygen evolution would be buried in the water oxidation process. It would be impossible to form a separate visible current peak. Nevertheless, under various studying conditions, the removing efficiency of three targeted EDCs approached to 99.9%. The effluent from primary settling tank was further treated with electrochemical oxidation. For the use of
DSA as anode at current density of 58.8 A m2, and with 3 h electrolysis, majority of EDCs were completely removed, except for 4-nonylphenol and 4-tert-octylphenol. And the solution COD concentrations did not change. However, the UV-254 absorbance was decreased by more than 50%, indicating that the aromatic structure of EDCs have been destroyed but perhaps transformed to form some byproducts. Using graphite felt as anode at current density of 29.4 A m2 could reduce all residual EDCs in the 1st humus tank effluent to very low levels, however, the absorption capacity of graphite felt electrode should be taken into account when considering the general results. Results of using DSA as anode and at current densities of 58.8 and 117.6 A m2 for treating the effluent from first humus tank demonstrated that estrone and 17b-estradiol can be reduced to a very low concentrations (20–100 ng l1), but total and soluble COD concentrations did not change greatly before and after treatment. 3.3. EDCs degradation with ferrate(VI) oxidation Ferrate(VI) ion has molecular formula, FeO2 4 , and is a very strong oxidant. At the acidic conditions, the redox potential of ferrate(VI) ions is greater than ozone (Jiang and Lloyd, 2002). Moreover, during oxidation process, ferrate(VI) ions are reduced to Fe(III) ions or ferric hydroxide, which simultaneously generates a coagulant in a single dosing and mixing unit process. For treating test solutions, ferrate(VI) can effectively reduce three EDCs concentrations to low level (tens of ng l1) at doses ranging 13–17 mg l1 as Fe, and the removal percentage was 99.99%. For treating wastewater effluent from the 1st humus, ferrate(VI) also performed very well at relative low doses (1–5 mg l1 as Fe). As shown in Table 4, all EDCs present in the 1st humus effluent were reduced to a low level, especially for bisphenol A, which has been reduced substantially from 1.2 lg l1 to 46 ng l1. In outline, the optimum ferrate(VI) dose required to achieve the maximum EDCs removal (>99.99%) was 0.10 mg EDC:1 mg ferrate (as Fe) for treating test solutions, and 0.06 mg EDC:1 mg ferrate (as Fe) for treating primary sedimentation effluents. 3.4. Comparative performance in the degradation of EDCs Fig. 3 shows the comparative performance of EDCs reduction with ferrate (sample 2) and electrochemical oxidation (sample 3). It can be seen in Fig. 3(a), that ferrate(VI) can reduce much more bisphenol A, 17bestradiol and 4-tert-octylphenol concentrations than electrochemical oxidation, and should possess greater oxidation potential to degrade EDCs. Fig. 3(b) demonstrates that ferrate(VI) can remove 20% more total COD
J.Q. Jiang et al. / Chemosphere 61 (2005) 544–550
549
Table 4 Analysis of treated sewage samples with ferrate(VI)a (ng l1) Samples a
WwTW1 Supernatant After filtration a b
4-tert-Octylphenol
4-Nonylphenol
Bisphenol A
Estrone
17b-Estradiol
16a-Hydroxyestrone
188.5 51.60
39.90 47.01 15.86
1209 199.67 46.06
30.10 26.25 20.87
52.54 104.2 27.90
63.7
b
b b
The original sewage taken from the effluent of humus tank in WwTW1. Lower than the detect limits.
Concentrations (ng/l)
600
(a)
nanyphenol bisphenol-A
500 400
estrone 17 beta-estradiol
300
4-tert-octylphenol
200 100
COD concentrations (mg/l)
0 200
(b)
Total COD
150
Soluble COD
100
Percentage removal of COD 24% 14% 44% 46%
50
0 1
2
3
Fig. 3. (a) Comparative EDCs concentrations; (b) comparative COD concentrations, 1. wastewater sample taken from the post-sedimentation in WwTW1; 2. treated sample with ferrate oxidation; 3. treated sample with electrochemical oxidation.
and 32% more soluble COD comparing with electrochemical oxidation for giving study conditions. COD represents all organic pollutants existing in wastewaters, a high COD removal efficiency with ferrate(VI) demonstrated that the ferrate(VI) oxidation possesses high capacity of not only degradation of EDCs but also of removing other organic contaminants in wastewaters comparing with electrochemical oxidation.
post-sedimentation can lower EDCs concentrations to a scale of from several hundreds to hundred of ng l1. Since natural EDCs, e.g., estrone and 17b-estradiol, are most potent and have biological activity and adverse impact on the environment at extremely low concentrations (several tens of ng l1), further treatment after conventional biological processes is required. Preliminary trials in treating either EDCs test solutions or wastewaters show that both ferrate(VI) and electrochemical oxidation can effectively reduce EDCs to very low levels, ranging between 10 and 100 ng l1, and the former is more effective than the latter to remove organic contaminants in terms of lowering the effluent COD concentrations. Whilst the results from this study suggest that the ferrate(VI) or electrochemical oxidation could be two potential techniques used for the treatment of EDCs, further investigations should be carried out to examine the treatment efficiency in the use of test solutions with low natural EDCs starting concentrations (e.g., <10 ng l1). In addition to this, the fundamental study on EDCs degradation with other oxidation technologies should be conducted before a full evaluation and assessment are to be conducted in order to select the potential effective technologies for the degradation of EDCs.
Acknowledgements The authors thank the Thames Water, UK, for a grant that provided a research fellowship for Dr. Q. Yin and the financial support for this study. The views expressed in this paper are not necessarily representing that of Thames Water.
References 4. Conclusions This work demonstrates that a range of EDCs were found in the crude wastewaters from two WwTWs in the concentration levels up to 890 ng l1. Primary sedimentation in the processes does not remove EDCs singnificantly, but trickling filter and activated sludge with
Desbrow, C., Routledge, E.J., Brighty, G.C., Sumpter, J.P., Waldock, M., 1998. Identification of estrogenic chemicals in STW effluent. 1. Chemical fractionation and in vitro biological screening. Environ. Sci. Technol. 32, 1549–1558. Jiang, J.Q., Lloyd, B., 2002. Progress in the development and use of ferrate (vi) salt as an oxidant and coagulant for water and wastewater treatment. Water Res. 36, 1397–1408.
550
J.Q. Jiang et al. / Chemosphere 61 (2005) 544–550
Khan, S.J., Wintgens, T., Sherman, P., Zaricky, J., Scha¨fer, A.I., 2004. Removal of hormones and pharmaceuticals in the advanced water recycling demonstration plant in Queensland, Australia. Water Sci. Technol. 50, 15–22. Liu, R., Zhou, J.L., Wilding, A., 2004. Simultaneous determination of endocrine disrupting phenolic compounds and steroids in water by solid-phase extraction–gas chromatography–mass spectrometry. J. Chromatogr. A 1022, 179–189. Piva, F., Martini, L., 1998. Neurotransmitters and the control of hypophyseal gonadal functions: possible implications of endocrine disruptors. Pure Appl. Chem. 70, 1647–1656.
Ternes, T.A., Stumpf, M., Mueller, J., Haberer, K., Wilken, R.-D., Servos, M., 1999. Behaviour and occurrence of estrogens in municipal sewage treatment plants—I. Investigations in Germany, Canada and Brazil. Sci. Total Environ. 225, 81–90. Thorpe, K.L., Hutchinson, T.H., Hetheridge, M.J., Scholze, M., Sumpter, J.P., Tyler, C.R., 2001. Assessing the biological potency of binary mixtures of environmental estrogens using vitellogenin induction in juvenile rainbow trout (Oncorhynchus mykiss). Environ. Sci. Technol. 35, 2476– 2481.