Environ. Sci. Technol. 2005, 39, 7469-7475
Carbamazepine and Its Metabolites in Wastewater and in Biosolids in a Municipal Wastewater Treatment Plant XIU-SHENG MIAO, JIAN-JUN YANG, AND CHRIS D. METCALFE* Water Quality Centre, Trent University, Peterborough, Ontario K9J 7B8, Canada
Pharmaceutically active compounds (PhACs) are discharged into the environment from domestic wastewater treatment plants (WWTPs). In this study, we determined the distribution of the anti-epileptic drug, carbamazepine (CBZ), and its major metabolites and caffeine in both aqueous and solid phases through different treatment processes of a WWTP. A method was developed to extract samples of biosolids using pressurized liquid extraction (PLE), coupled with cleanup of extracts using solid-phase extraction. Samples of biosolids and wastewater were analyzed for caffeine and CBZ and five of its metabolites, 10,11-dihydro-10,11-epoxycarbamazepine (CBZ-EP), 11-dihydro10,11-epoxycarbamazepine (CBZ-DiOH), 2-hydroxycarbamazepine (CBZ-2OH), 3-hydroxycarbamazepine (CBZ3OH), and 10,11-dihydro-10-hydroxycarbamazepine (CBZ10OH). The analytes were quantified using liquid chromatography-electrospray ionization tandem mass spectrometry. The recoveries of the analytes were 82.191.3% from raw biosolids and 80.1-92.4% from treated biosolids, and the limits of detection were 0.06-0.50 and 0.060.40 µg/kg on a wet weight basis for raw and treated biosolids, respectively. The behavior of carbamazepine and its metabolites, together with caffeine as a marker of domestic inputs, was investigated in the WWTP for the City of Peterborough, ON, Canada, which utilizes secondary sewage treatment technologies. CBZ, CBZ-2OH, CBZ3OH, and CBZ-DiOH were detected at concentrations of 69.6, 1.9, 1.6, and 7.5 µg/kg (dry weight), respectively, in untreated biosolids and at concentrations of 258.1, 3.4, 4.3, and 15.4 µg/kg (dry weight), respectively, in treated biosolids. However, CBZ-EP and CBZ-10OH were not detected in any of the biosolid samples. CBZ and its five metabolites were detected in all wastewater samples collected from four different stages of treatment. The results showed that 29% of the carbamazepine was removed from the aqueous phase during treatment in the WWTP, while the metabolites were not effectively removed. Concentrations of caffeine were reduced by 99.9% in the aqueous phase, which appeared to be due primarily to degradation. Caffeine was also detected at concentrations of 165.8 and 7.6 µg/kg (dry weight) in raw and treated biosolids, respectively. Because of differences in hydrophobicity, CBZ is the primary analyte in biosolids, while CBZ-DiOH is the primary analyte in the aqueous phase * Corresponding author phone: (705)748-1011, ×7272; fax: (705)748-1569; e-mail:
[email protected]. 10.1021/es050261e CCC: $30.25 Published on Web 09/07/2005
2005 American Chemical Society
of the wastewater. A mass balance calculation showed that the majority of CBZ and its metabolites exist in the aqueous phase (i.e., wastewater), rather than in the biosolids, 78 g of CBZ and its metabolites enters the Peterborough WWTP daily, and 91 g is discharged from the WWTP daily in the combined suspended solids and aqueous phases of the wastewater. The calculated daily inputs into the WWTP are somewhat less than the inputs of 192 g estimated from Canadian annual sales data for CBZ.
Introduction Pharmaceutically active compounds (PhACs) are an emerging environmental issue, due to their presence in the aquatic environment and potential for impacts on wildlife and humans (1, 2). Some metabolites of pharmaceuticals are still bioactive, and they may have high stability and mobility in the environment. Therefore, the fate of metabolites is a very important component of assessing the environmental risks associated with the release of PhACs. Municipal wastewater treatment plants (WWTP) are one of the major sources of PhACs in the environment (3, 4). Many chemical, physical, and biological factors may affect the fate of PhACs in WWTPs, including adsorption/desorption on biosolids (i.e., sludge), pH, the ionic strength of the sewage, microbial decomposition rates, and the physical and chemical properties of the PhACs (e.g., polarity, photostability, volatility, etc.). Hydrophilic compounds that are resistant to degradation may remain dissolved in the aqueous phase of the WWTP effluent, or more hydrophobic substance may bind to the biosolids. Thus, these compounds may enter the environment through the discharge of WWTP effluents into receiving waters or they may enter the environment in association with biosolids that are deposited in landfills or spread onto agricultural land for soil amendment. Pharmaceutical compounds discharged into surface water have the potential to contaminate sources of drinking water (5, 6). In addition, some pharmaceuticals have been detected in groundwater (7-9). Since WWTPs provide the first and perhaps most important opportunity for removing pharmaceuticals that are destined for discharge into the environment, it is important to characterize the fate of pharmaceuticals during the treatment of domestic wastewater. The most important process that governs whether pharmaceuticals will primarily enter the aquatic or terrestrial environments is whether these compounds partition from aqueous sewage into sludge. Some studies have been conducted on the distribution of PhACs within sewage in WWTPs (10-12), but there have been only few studies on PhACs and their metabolites in both sewage and sludge. In this study, we analyzed carbamazepine (CBZ) and its major metabolites in both aqueous and biosolid phases of a municipal WWTP. CBZ is an important drug for the treatment of epilepsy, as well as for various psychotherapy applications. In Canada, approximately 28 tons of carbamazepine was sold as prescriptions in 2001 (13). Studies in Europe and North America have shown that CBZ is one of the most frequently detected pharmaceuticals in WWTP effluents and in river water (3, 8, 14, 15). We previously detected CBZ and five of its metabolites in the effluents of WWTPs in Canada and CBZ and one of these metabolites, 10,11-dihydro-10,11-dihydroxycarbamazepine (CBZ-DiOH), in surface water (16). CBZ undergoes extensive hepatic metabolism by the cytochrome P450 (CYP) system (17, 18). Thirty-three meVOL. 39, NO. 19, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7469
TABLE 1. Molecular Structures, Abbreviations, Formulas, Molecular Weights, and Log Kow of Carbamazepine and Its Metabolites and Caffeine
a Chemical Abstracts Service registry number. b Molecular weight (MW) was calculated for the lowest isotopomer. c Ref 32. of carbamazepine and its metabolites were calculated with ACD/log P. e Ref 34.
tabolites of CBZ have been identified from human and rat urine (19). The main metabolic pathway for CBZ is oxidation to 10,11-dihydro-10,11-epoxycarbamazepine (CBZ-EP), then hydration to CBZ-DiOH and conjugation of CBZ-DiOH with glucuronide. The hydrolysis of CBZ-EP to CBZ-DiOH is catalyzed by microsomal epoxide hydrolase (20). The metabolism of CBZ to CBZ-EP appears to be catalyzed by CYP3A4 and CYP2C8. Lesser pathways include oxidation to 2-hydroxycarbamazepine (CBZ-2OH) and 3-hydroxycarbamazepine (CBZ-3OH), which appear to be catalyzed by CYP1A2, as well as oxidation to 10,11-dihydro-10-hydroxycarbamazepine (CBZ-10OH). The most important metabolites are CBZ-DiOH and, to a lesser extent, CBZ-EP. The latter compound has been shown to possess similar anti-epileptic properties to CBZ, and it may cause neurotoxic effects (21, 22). In some cases, clinical toxicities parallel CBZ-EP concentration (23). Despite being chemically stable under physiological conditions, CBZ-EP is converted to the CBZ-DiOH metabolite by epoxide hydrolase. Bernus et al. (24) investigated the metabolism of CBZ in women during pregnancy and found that CBZ, CBZ-EP, CBZDiOH, CBZ-acridan, CBZ-2OH, and CBZ-3OH accounted for 0.5, 2.1, 34.6, 3.2, 2.3, and 3.7% of total concentrations in urine samples, respectively. The CBZ-DiOH metabolite is not pharmaceutically active. Caffeine has been used as a chemical marker for human excretory products discharged in domestic wastewater (14, 25, 26). Canadian annual consumption of caffeine is close to 2.2 × 103 tons a year, or 240 mg per person per day (27), which is largely excreted in urine and transported into domestic sewage. 7470
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 19, 2005
d
Values of log Kow
We monitored the concentrations of CBZ and its major metabolites and caffeine in the aqueous phase of wastewater and in biosolids at various stages of treatment in the WWTP for the City of Peterborough, ON, Canada. To conduct this study, it was necessary to develop methods for the analysis of the CBZ and metabolites in biosolids. We developed a method for extraction of analytes using pressurized liquid extraction (PLE) and a method for cleanup of the extracts using solid-phase extraction (SPE) cartridges. Extracts were then analyzed using liquid chromatography-electrospray tandem mass spectrometry (LC-ES-MS/MS).
Experimental Procedures Chemicals. The formulas, abbreviations used in figures and tables, molecular weights, and log Kow values for the analytes are summarized in Table 1. CBZ, CBZ-EP, CBZ-DiOH, CBZ2OH, CBZ-3OH, and CBZ-10OH were provided by Novartis Pharma AG (Basel, Switzerland). Caffeine (1,3,7-trimethylxanthine) and ammonium acetate (98%) were purchased from Sigma-Aldrich Canada (Oakville, ON, Canada). A surrogate standard of caffeine (trimethyl-13C3, 99%) was purchased from Cambridge Isotope Laboratories (Andover, MA). Acetonitrile and methanol were purchased from Caledon Laboratories (Georgetown, ON, Canada). Formic acid (90%) and HPLC grade water were purchased from EM Science Industries (Gibbstown, NJ). Hydromatrix was purchased from Varian Canada (Mississauga, ON, Canada). Sampling. The WWTP investigated in the small city of Peterborough, ON, Canada employs secondary (i.e., activated sludge) sewage treatment technologies (Scheme 1). Second-
SCHEME 1. Flow Chart of the Peterborough WWTP, Showing the Sampling Sites (A-F)
ary sewage treatment is the most common technology used in WWTPs in Canada. The Peterborough WWTP serves a population of about 75 000 people. Its design average flow capacity is 60 000 m3/day, and its average handling flow is 46 000 m3/day (77% of its design capacity). The influent flow rate consists of about 25% industrial wastewater and 75% domestic sewage. The wastewater entering the WWTP is first treated mechanically using grit removal and screening and goes to the primary clarifier. The primary effluent is directed to an activated sludge system for biological treatment in the aeration tank and then goes to the second clarifier. Full aeration is provided for complete nitrification in the aeration tank. After settling in the secondary clarifier, the wastewater is disinfected by UV irradiation at 30 mW s-1 cm-2. Following irradiation with UV, the treated effluent is discharged to the Otonabee River. The hydraulic retention time for the wastewater in the WWTP is 12-18 h. As for the biosolids, some activated sludge from the secondary clarifier is returned to the inlet of the primary clarifier, and the remainder goes to the digester (anaerobic) and is combined with primary sludge (Scheme 1). The biosolid treatment process prior to input into the digester is operated at ambient temperatures with a retention time of 4-6 days, which varies with wastewater flow and treatment conditions. In the digester, the biosolids are stabilized by thermal treatment at 37 °C for approximately 30 days. The treated biosolids are then ready to be shipped from the WWTP for application to agricultural fields. Samples of wastewater were collected as flow proportional composites using automated samplers that collected defined volumes every hour over a 24 h period. Samples of biosolids were collected as grab samples. On April 28, 2003, samples were collected of untreated wastewater (site A), treated wastewater after UV (site D), untreated biosolids (site E), and treated biosolids (site F), according to the sites shown in Scheme 1. On July 14, 2003, composite samples were collected of untreated wastewater (site A), primary wastewater (site B), treated wastewater before UV treatment (site C), and treated wastewater after UV treatment (site D). After collection, the wastewater samples were filtered through 0.45 µm glass filters and immediately transported to the laboratory, where they were stored at 4 °C in a cold room until preparation for analysis, which generally occurred within 24 h of collection. The biosolid samples were concentrated by centrifugation at 2500 rpm for 5 min at room temperature and frozen in amber jars until analysis. The moisture content of the biosolids was expressed as the ratio of the mass of water to the dry weight of the biosolids sample, after drying in an oven at a temperature of 105 °C for 24 h. Preparation of Wastewater and Biosolid Samples. Thawed samples of centrifuged biosolids (10 g) were extracted by PLE using an ASE 300 accelerated solvent extractor (Dionex, Sunnyvale, CA) equipped with 34 mL stainless
extraction cells. To dry the sample, reduce particle clumping and solvent channeling in the extraction cell, and reduce the void volume, Hydromatrix was mixed with biosolids. Hydromatrix is a specially cleaned and sieved diatomaceous earth and is capable of adsorbing and retaining up to twice its weight of aqueous media (i.e., aqueous capacity is 2 mL water per g of Hydromatrix). The amount of Hydromatrix added (2-10 g) depended on the water content of the sludge sample. The extraction procedure was optimized for extraction temperature, extraction time, and the extraction solvent using sludge samples spiked with the analytes. For all PLE experiments, a fixed extraction pressure of 1500 psi was used. The other optimized PLE conditions were prefill method: solvents, acetone/water (3:7); equilibration, 5 min; static time, 5 min; flush volume, 60%; purge time, 60 s; static cycles, 3; and temperature, 80 °C. The flush volume amounted to 100% of the extraction cell volume, and the total solvent use was approximately 90 mL. After PLE extraction, the extract was poured into a 250 mL boiling flask, and the acetone was evaporated using a rotary evaporator at 40 °C. After evaporation, the remaining aqueous matrix was diluted with 100 mL of HPLC water. This matrix was cleaned up by SPE using HLB Oasis extraction cartridges (Waters, Milford, MA). SPE cartridges were conditioned, and the sample extraction was conducted as previously described for the extraction of aqueous samples (16). A volume of 0.5 L of wastewater was extracted using the same SPE extraction protocol that was used for cleanup of extracts from biosolids. Analysis. Mass spectrometry was performed using a Quattro LC tandem quadrupole mass spectrometer (Micromass, Manchester, UK) equipped with a Z-Spray electrospray ionization source. Analyte separations were conducted with an Alliance 2695 liquid chromatograph (Waters, Milford, MA) with a Genesis C8 column (150 mm × 2.1 mm i.d., 3 µm) purchased from Jones Chromatography, Hengoed, Mid Glamorgan, UK. The conditions for LC-ES-MS/MS in selected reaction monitoring (SRM) mode for CBZ and its metabolites have been reported previously (16). SRM channels of m/z 195 > 138 and 198 > 140 were used for analysis of caffeine and a 13C3-labeled caffeine surrogate, respectively. Organic coextractives are responsible for matrix effects with electrospray ionization mass spectrometry, especially for the determination of trace contaminants in environmental samples. To compensate for matrix effects, a surrogate of 13C -labeled caffeine was used as an internal standard to 3 quantify caffeine. At the time of this study, there was no stable isotope surrogate for CBZ commercially available. Therefore, for the analysis of CBZ and its metabolites, the analytes were quantified using calibration by standard additions, as reported previously (16). Briefly, this method involves spiking five subsamples with a dilution series of the analytes, then constructing a calibration curve for the analytes VOL. 39, NO. 19, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7471
TABLE 2. Mean Recoveries with Standard Deviations and Limits of Detection (LODs) and Limits of Quantification (LOQs) of CBZ and Its Metabolites and Caffeine Spiked into Samples of Raw and Treated Biosolids (µg/kg, Wet Weight)a raw biosolids
treated biosolids
analyte
recovery (%)
LOD
LOQ
recovery (%)
LOD
LOQ
CBZ CBZ-EP CBZ-DiOH CBZ-2OH CBZ-3OH CBZ-10OH caffeine
91.3 ( 5.8 86.1 ( 8.4 83.6 ( 4.4 84.0 ( 6.9 83.8 ( 7.9 88.4 ( 4.8 82.1 ( 5.6
0.15 0.06 0.11 0.08 0.07 0.10 0.50
0.50 0.22 0.32 0.26 0.22 0.34 1.70
92.4 ( 3.7 87.7 ( 4.5 82.8 ( 8.1 89.1 ( 3.6 80.1 ( 6.5 90.4 ( 5.1 83.2 ( 6.1
0.17 0.07 0.10 0.07 0.06 0.08 0.40
0.58 0.23 0.34 0.22 0.20 0.28 1.35
a Recoveries are the average of three determinations at spiked concentrations of 10 µg/kg (wet weight).
within the sample matrix. Three other unspiked subsamples were analyzed at the same time to yield triplicate analyses of each sample.
Results and Discussion Evaluation of Sample Preparation Methods for Biosolid Samples. Extraction efficiency by PLE is mainly governed by the solubility of the analytes in the extraction solvent. Because organic solvents cannot be passed through the SPE following PLE extraction, acetone was chosen as the extraction solvent since it can be easily removed by evaporation at moderate temperatures (i.e., 40 °C) prior to the SPE cleanup step. Different proportions of acetone and water were investigated for maximum extraction efficiency by PLE. A high percentage of acetone caused more organic coextractives to be extracted and resulted in clogging of the SPE cartridge and complex chromatograms. Extraction efficiencies for the analytes were poor if the solvent mixture contained more than 80% water. A 3:7 mixture of acetone and water optimally extracted CBZ and its metabolites and caffeine. However, temperature also influences PLE extraction efficiency. Higher temperatures decrease the viscosity of water, thus allowing better penetration of the sample matrix, but higher temperatures may increase the amounts of unwanted coextractives and result in degradation of thermally labile analytes. A solvent temperature of 80 °C was found to be optimal for extraction of all analytes. However, failures in PLE were experienced as a result of the high back-pressure when extracting wet biosolids. Therefore, Hydromatrix was mixed with biosolid samples to absorb water and improve permeation of the sample matrix by the extraction solvent. To evaluate the extraction method for samples of biosolids, samples of raw and treated biosolids (wet) were spiked with the analytes. After the biosolid samples were allowed to stabilize for 2 h, they were extracted using the optimized PLE method. Table 2 lists the percent recoveries of the analytes, which ranged from 82.1 to 91.3% in raw biosolids and from 80.1 to 92.4% in treated biosolids, with standard deviations of 3.6-8.4%. The limits of detection (LODs) and limits of quantification (LOQs) of the analytes for combined PLE, followed by SPE cleanup, were determined by spiking sludge
samples with the analytes (Table 2). The LOD and LOQ were defined as concentrations in a sample matrix resulting in peak areas with signal-to-noise ratios (S/N) of 3 and 10, respectively. The mean water contents of raw and treated biosolids samples in this study were 84.4 and 81.7%, respectively. Distribution of CBZ and Its Metabolites. Previous studies have shown that CBZ is relatively persistent within WWTPs (3, 4). Analytical data for wastewater and biosolids collected from the Peterborough WWTP in April 2003 (Table 3) are generally consistent with these previous studies. The mean concentration of CBZ declined from 356 ng/L in the untreated wastewater to 251 ng/L (29% decline) in the treated wastewater. The metabolite, CBZ-DiOH, was the predominant analyte in the aqueous phase (Table 3), which is consistent with our earlier studies (16). No loss of the metabolites of CBZ was observed in the aqueous phase (Table 3). Indeed, the concentrations of CBZ-2OH and CBZ-DiOH increased significantly in the treated wastewater relative to concentrations in the untreated wastewater. The hydroxylated metabolites of CBZ occur primarily in conjugated forms in human biological fluids (28), and these conjugated metabolites could have been transformed to the free form by microbial activity during treatment, as was described previously for estrogens (29). More work is required to determine the mechanism involved in this observed response. Note that the samples of wastewater represent composite samples collected over a 24 h period, so these data should integrate daily variations. Composite sampling over several days (e.g., 4-5 days) would have integrated variations over a longer period of time within the WWTP, but this sampling scheme was rejected because of concerns over the potential for degradation of some of the analytes in the wastewater matrix over a long-term sampling period. In the biosolid samples collected in April 2003, only CBZ and three of the metabolites, CBZ-2OH, CBZ-3OH, and CBZDiOH, were detected, with CBZ as the predominant analyte (Table 3). The metabolite that was dominant in the wastewater, CBZ-DiOH, did not partition into the biosolids to high concentrations due to its hydrophilicity. Two of the metabolites that were detected in the aqueous phase, CBZ-EP and CBZ-10OH, were not detected in the biosolid samples. Many PhACs have relatively low log Kow values, relative to legacy contaminants, such as PCBs and polynuclear aromatic hydrocarbons, which have log Kow values between 4.5 and 8.5 (30). Compounds with log Kow < 2.5 are assumed to have a low potential for adsorption onto particulates (31). As shown in Table 1, the metabolites of CBZ have log Kow values between 0.13 and 2.41, in comparison to the log Kow for CBZ of 2.67 (or 2.25) (32). The concentrations of CBZ and metabolites increased on a dry weight basis between untreated and treated biosolids (Table 3). This may be due to the different composition of untreated and treated biosolids, which show different adsorption behaviors. Another explanation may be the recycling of excess return activated sludge back to the untreated sludge, leaving behind the more highly concentrated biosolid substrate as the treated material. In any event, more studies are required to determine the mechanisms for this observation.
TABLE 3. Mean ((Standard Deviation in Brackets) of the Concentrations of CBZ and Its Metabolites in Wastewater (ng/L) and in Biosolids (µg/kg Dry Weight) at Various Stages of Treatment in the Peterborough WWTPa sample and units
CBZ
CBZ-EP
CBZ-2OH
CBZ-3OH
CBZ-10OH
CBZ-DiOH
untreated wastewater (ng/L) treated wastewater untreated biosolids (µg/kg dry weight) treated biosolids
356.1 ((5.8) 251.0 ((6.3) 69.6 ((2.2) 258.1 ((4.7)
39.2 ((1.2) 19.1 ((1.5) n.d. n.d.
59.0 ((2.1) 70.4 ((2.3) 1.9 ((1.1) 3.4 ((0.9)
55.4 ((6.1) 69.2 ((3.1) 1.6 ((0.8) 4.3 ((0.9)
22.2 ((2.5) 32.5 ((2.4) n.d. n.d.
1001.2 ((12.5) 1081.2 ((13.0) 7.5 ((0.7) 15.4 ((1.3)
a
All samples were collected on April 28, 2003 and were analyzed in triplicate. n.d. ) not detected.
7472
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 19, 2005
TABLE 4. Related Data for Mass Balance Calculations in the Peterborough WWTP during the Sampling Period in April 2003 parameter flow (m3) suspended solids water content
FIGURE 1. Mean concentrations (ng/L) of CBZ and its metabolites in the aqueous phase of wastewater through different stages of treatment in the Peterborough WWTP. Samples were collected on July 14, 2003, and sampling sites A-D are marked in Scheme 1. Samples were analyzed in triplicate. To more fully investigate the fate of CBZ and its metabolites in the aqueous phase of the wastewater in the Peterborough WWTP, samples were collected in July 2003 as untreated wastewater, primary wastewater (after the primary clarifier), and treated wastewater before and after UV disinfection. Figure 1 illustrates the profile of CBZ and its metabolites in the aqueous phase of sewage at different stages of treatment. Once again, CBZ and all metabolites were detected in the wastewater samples, and CBZ-DiOH was the predominant analyte. The concentration of CBZ in the aqueous phase of the wastewater declined from 651 ng/L in the untreated wastewater to 353 ng/L (46% decline) in the primary sewage, indicating effective removal of this compound from the aqueous phase by the primary clarifier. However, the concentration of CBZ in the treated effluent (after UV) subsequently increased to a final concentration of 463 ng/L. Removal over the total treatment process was 29%, which is higher than the removal rates of less than 10% reported for WWTPs in Germany (3). Further investigations should be conducted to determine the specific mechanisms for removal of CBZ within the WWTP. Figure 1 indicates that the concentrations of CBZ, CBZDiOH, and CBZ-10OH increased significantly after UV disinfection, in comparison to the samples collected before UV disinfection. In particular, the concentration of CBZ10OH increased from 843 ng/L in the treated sewage before UV to 1065 ng/L in the sewage after UV (Figure 1). We have previously observed this response to UV treatment, specifically, an increase in the concentrations of synthetic musks after UV treatment in the Peterborough WWTP (33). It is possible that UV irradiation mediates the conversion of the conjugated forms of the hydroxyl metabolites of CBZ to the free form and/or modifies the dissolved organic matrix in the treated wastewater so that analytes are released from the bound to the dissolved phase. Obviously, more work is required to elucidate the mechanism involved in this response, including whether the level of UV irradiation is sufficient to alter the wastewater matrix. Caffeine was used in this study as a marker of human excretory inputs into the domestic sewage. In samples collected in April 2003, the mean concentration of caffeine in the untreated wastewater was 63.2 µg/L, which is far higher than the ng/L concentrations of CBZ and metabolites in the treated wastewater. Approximately 99.9% of the caffeine was removed from the aqueous phase by the treatment
untreated sewage
treated sewage
51196 87 mg/L
59806 8 mg/L
untreated sludge
treated sludge
3.9% 84.4%
2.4% 81.7%
process, as the concentration of caffeine was 0.068 µg/L (i.e., 68 ng/L) in the treated wastewater. This observation is consistent with previous data that show high rates of elimination of caffeine in WWTPs (25, 26). Despite caffeine being hydrophilic (34), this compound was detected at concentrations of 165.8 and 7.6 µg/kg (dry weight) in raw and treated biosolids, respectively, for a 95.4% rate of elimination of caffeine from biosolids. These data for caffeine contrast to the data for CBZ and its metabolites, which appear to be relatively resistant to degradation in WWTPs. Degradation appears to be a major mechanism for the elimination of caffeine from municipal wastewater. Mass Balance of CBZ and Metabolites. To evaluate the fate of CBZ and its metabolites in the sewage in the Peterborough WWTP, a mass balance approach was used to evaluate the distribution of the analytes in the aqueous and suspended solids phases of the wastewater. We assumed that the concentrations of CBZ and its metabolites in the suspended solid phase of the raw and treated wastewater were approximately equal to their corresponding concentrations in raw and treated biosolids, respectively (Table 3). Related data for mass balance calculations such as flow rates and the total suspended solid content of the wastewater were collected at the Peterborough WWTP during the sampling period and are listed in Table 4. The daily inputs and outputs of the analytes in the aqueous and solids phases of the WWTP were calculated using eqs 1 and 2, respectively. For the aqueous phase
Mi ) CiQj/1000000
(1)
where Mi is the daily input or output of analyte i in the aqueous phase, g; Ci is the concentration of analyte i in raw or treated wastewater (ng/L); and Qj is the flow of raw sewage or treated wastewater (m3/day). For the suspended solid phase
Mi ) CiQjSj/1000 000 000 000
(2)
where Mi is the daily input or output of analyte i in the suspended solid phase (g), Ci is the concentration of analyte i in the raw or treated biosolids (ng/kg dry weight), Qj is the flow of raw or treated wastewater (m3/day), and Sj is the suspended solid content in raw or treated wastewater (mg/L). Table 5 presents the results of the mass balance calculations for CBZ and its metabolites in the aqueous and suspended solid phases of the wastewater in the Peterborough WWTP. The results show that CBZ and CBZ-DiOH were the predominant compounds in the wastewater. The majority of CBZ and metabolites exists in the aqueous phase of the wastewater, and the proportion of CBZ in the suspended solid portion of the wastewater only accounted for 0.45 and 0.15% of the total amount (aqueous + suspended solids) in the raw and treated wastewater, respectively. The daily input and output of CBZ and its metabolites in the sewage of the Peterborough WWTP was 79 and 91 g, respectively. It is acknowledged that caution must be used in the interpretation of these data since the mass balances are based on single 24 h composite samples for wastewater and grab samples of biosolids. VOL. 39, NO. 19, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7473
TABLE 5. Mean Daily Inputs and Outputs of CBZ and Its Metabolites in the Aqueous and Suspended Solids Phases of the Wastewater in the Peterborough WWTP during the Sampling Period of April 28, 2003, as Determined by Mass Balance Calculations input (g) aqueous phase CBZ CBZ-EP CBZ-2OH CBZ-3OH CBZ-10OH CBZ-DiOH subtotal total
18.23 2.01 3.02 2.84 1.14 51.26 78.49 78.85
output (g) solid phase 0.31
