Removal and transformation of pharmaceuticals in wastewater treatment plants and constructed wetlands

Since trace organic compounds such as pharmaceuticals in surface water have been a relevant threat to drinking water supplies, in this study removal of pharmaceuticals and transformation of pharmaceuticals into metabolites were investigated in the main source of micropollutants such as WWTPs and engineered constructed wetlands. Pharmaceuticals were e ff ctively removed by di fferent WWTP processes and wetlands. Pharmaceutical metabolites with relatively low log D value were resulted in the low removal e fficiencies compared to parent compounds with relatively high log D value, indicating the stability of metabolites. And the constructed wetlands fed with wastewater e ffluent were encouraged to prevent direct release of micropollutants into surface waters. Among various pharmaceuticals, di fferent transformation pattern of ibuprofen was observed with significant formation of 1-hydroxy-ibuprofen during biological treatment in WWTP, indicating preferential biotransformation of ibuprofen. Lastly, transformation of pharmaceuticals depending on their structural position was investigated in terms of electron density, and, the electron rich C 1=C2 bond of carbamazepine was revealed as an initial transformation position.


Introduction
For several decades, pharmaceuticals and personal care products (PPCPs) have been noticed as emerging problematic compounds (Ternes et al., 1998;Snyder et al., 2003).In order to reduce residual concentrations of PPCPs, various advance treatments have been studied by many research groups (Lee and Gunten, 2010;Rosal et al., 2010 and many others).However, high levels of PPCPs are still detected in wastewater effluent, surface waters, and drinking waters (Kim et al., 2007;Lee et al., 2012;Yoon et al., 2010;Benotti et al., 2009).Since those trace organic compounds have been detected even in treated drinking waters by Benotti et al. (2009), control of micropollutants has been important especially in the wastew-ater treatment plant, the main source of the micropollutants in the aquatic environments.
Meanwhile, constructed wetlands have been introduced as an alternative to wastewater treatment for micropollutants removal (Matamoros and Bayona, 2006).And few reports have been focused on the relationship between log D and removal of PPCPs in wetland systems.In previous study (Lee et al., 2011), removal efficiency in constructed wetlands has been investigated using corresponding octanol-water partitioning coefficient of pharmaceuticals.
And few studies have been reported with regarding to behaviors of pharmaceutical metabolites in various environments (Stumpf et al., 1998;Quintana et al., 2005).The Published by Copernicus Publications on behalf of the Delft University of Technology.
potential ecotoxicological effect of pharmaceutical metabolites is still not known, thus, there is necessity to study further about the fate of pharmaceutical metabolites in various wastewater treatment processes.
In this study, removal and transformation of pharmaceuticals has been investigated with regarding to physicochemical and structural properties of pharmaceuticals and their metabolites in various environments such as WWTPs and constructed wetlands receiving wastewater effluent.

Sample collection
In June 2010, water samples were collected from each process (influent and effluents of the various unit operations) in three different wastewater treatment plant.One is Gwangju primary municipal wastewater treatment plant, which is operated with the nitrogen and phosphorus removal treatment system (600 000 m 3 day −1 ).Second sampling sites was Gwangju secondary municipal wastewater treatment plant, which had two different treatment trains (Conventional Activated Sludge and Modified Ludzack-Ettinger (MLE) process), having total capacity up to 120 000 m 3 day −1 .In those two different Gwangju municipal WWTPs, sodium hypochlorite was added as a disinfectant and final effluent samples were collected after disinfection process.Additionally, water samples from both Damyang wastewater treatment plant and Damyang constructed wetlands were collected and studied.There was no disinfection system in Damyang WWTP and final effluent was collected after secondary treatment.Damyang constructed wetlands connected to Damyang wastewater treatment plant is free surface flow constructed wetlands, which have two different ponds, containing Aporus ponds followed by Typha ponds.The hydraulic retention time of wetlands is approximately 6 h and flow rate is 1800 m 3 day −1 .Every sample was spiked with a biocide sodium azide and ascorbic acids to quench any residual oxidant in the field.

SPE and LC-MS/MS analysis
After filtration using glass fiber membrane filter, all analytes were extracted by using AutoTrace automated solid phase extraction (SPE) system (Caliper Corporation, Hopkington, MA), as depicted by Vanderford and Snyder (2006).Briefly, the 6 mL, 500 mg hydrophilic-lipophilic balance (HLB) glass cartridges (Waters Corporation, Milford, MA) were preconditioned in the following order: 5 mL of MTBE, 5 mL of methanol, and 5 mL of deionized water.500 mL of samples, spiked with the addition of standards for internal calibration, were loaded onto the cartridges at 15 mL min −1 in duplicate, after which the cartridges were rinsed with 5 mL of deionized pure water, and then dried with a steam of air for 50 min.The cartridges were eluted with 5 mL of methanol, followed by 5 mL of 1/9 (v/v) methanol/MTBE.The eluted solution was concentrated in a water bath at 40 • C with a gentle stream of air to a final volume of 500 µL, which was a concentration factor of 1000.The levels of pharmaceuticals and their metabolites were then measured using a Water 2695 Separations Module (Waters, Milford, MA) coupled with a Micromass Quattro Micro triple quadrupole tandem mass spectrometer (Micromass, Manchester, UK) in electrospray ionization mode (ESI).A 20 µL sample loop and 150 × 2.1 mm SunFire C18 column with a particle size of 3.5 µm (Waters, Milford, MA) was employed for analyte separation.A binary gradient, consisting of 0.1 % formic acid (eluent A) and 100 % acetonitrile (eluent B), was used at a flow rate of 0.2 mL min −1 .Selected PPCPs and their metabolites was analyzed using two different gradients.The gradient used for the PPCPs and most of the metabolites was: gradient with 15 % of B was held for 4 min, increased linearly to 80 % for www.drink-water-eng-sci.net/6/89/2013/ Drink.Water Eng.Sci., 6, 89-98, 2013 OH a log P, log D, and pK a value were calculated from the Software Calculator Plugins.b N.E: nonexistent at pH range 1-14.
6 min, held for 3 min with 80 % of B and then linearly increased to 100 % for 7 min.The gradient used for the carbamazepine metabolites was: gradient with 10 % of B increased linearly to 40 % for 15 min, increased linearly to 90 % for 10 min, and held until 30 min (Kang et al., 2008).A 5 min equilibration step with gradient of 10 % B was used at the beginning of each run.Detail LC-MS/MS analysis condition and analytical parameters of target compounds are shown in Table 2.

Molecular orbital calculations
Molecular orbital were calculated single determinant (Hartree-Fock) for optimization bearing the minimum energy obtained at the AM1 level.All semi-empirical calculations to obtain the point charge and electron density for pharmaceuticals and metabolites were performed in MO-G with a SCI-GRESS package version 7.7 (Fujitsu Co. Ltd.) (Watanabe et al., 2003).3. Unfortunately, the contribution of different transport mechanisms such as biodegradation, adsorption to sludge and sediments, and oxidation during disinfection was not considered here.Gwangju WWTP was found as a major source, releasing tons of micropollutants to the Yeongsan River water, and, the concentrations of PPCPs in Gwangju WWTPs (∼ µg L −1 ) were generally much higher than in Damyang WWTP (∼ ng L −1 ), presumably resulting from the dense populations in Gwangju area.
Particularly, high concentrations of caffeine and acetaminophen in WWTP influents reflect frequent use and ingestion of those compounds in this urban area.Most of PPCPs exhibited high removal efficiency (> 90 %) during the WWTPs processes, except for carbamazepine removal in Gwangju secondary WWTP (removal efficiency at 74 %).After discharge of WWTP effluent into connected constructed wetlands, levels of some PPCPs (carbamazepine, sulfamethoxaolze, diclofenac, and TCEP) slightly decreased.This result indicates that operation of wastewater stabilization ponds or wetlands would be helpful to prevent release of micropollutants into surface water.Our research group has previously been studied the Damyang constructed wetlands with respect to the control of organic micropollutants, and, reported the efficiency of wetland treatments depending on the wetland characteristics and properties of micropollutants (Park et al., 2009).Many other studies have also been supported the necessity of additional wetland treatments for mi-cropollutants control (Matamoros and Bayona, 2006;Conkle et al., 2008).

Transformation of pharmaceuticals and formation of metabolites in WWTPs
Table 4 shows occurrence of parent pharmaceuticals and their metabolites in WWTPs and wetlands system.In the WWTPs influent, most metabolites were detected at high level of concentrations with range of 100 ∼ 10 000 ng L −1 compared to WWTP effluents, indicating dominant transformation pathway of metabolite resulting from human body rather than microbial transformation.However, after passing the WWTPs, concentrations of most metabolites were observed higher than that of parent compounds, indicating the structural stability of the metabolites during the WWTP process relative to their parent compounds.The stability of metabolites may be related to the difference of the log D value between parent pharmaceutical compounds and pharmaceutical metabolites.As pharmaceuticals transformed into their metabolites, their log D value at pH 7 slightly decreased (Table 1).Considering the dominant removal mechanism of micropollutants in WWTP process (i.e., sorption and biodegradation), decrease of log D may result in the reduction of sorption of pharmaceutical metabolites on the sludge surface.Consequently, pharmaceutical metabolites would be detected at higher levels than parent compounds in WWTP effluents.
In case of sulfamethoxazole metabolite, even though the concentrations of N-acetyl-sulfamethoxazole were low enough in wastewater effluents, by considering retransformation of acetylated metabolite back into its parent compound, N-acetyl-sulfamethoxazole should be assumed as a pharmaceutically active parent compound (Göbel et al., 2005).And based on this reason, monitoring of N-acetyl-sulfamethoxazole is important and necessary for the effective control of micropollutants in aquatic environment.
During the engineered constructed wetland treatments, there was no specific change of pharmaceutical metabolites in exception with ibuprofen metabolites.As can be seen in Fig. 1, ibuprofen was differently transformed depending on the treatment process.In WWTPs influent, 2-hydroxyibuprofen, the well-known human metabolite of ibuprofen was dominant, whereas, after the activated sludge treatment, concentrations of 1-hydroxyibuprofen indicate the highest level in all WWTPs and wetlands.Significant formation of 1-hydroxyibuprofen in WWTPs might be possibly explained by preferential microbial metabolism in ac- tivated sludge treatment process.In the additional experimental results conducted in river waters, completely different composition of ibuprofen metabolites was observed.2hydroxy-ibuprofen was found to be dominant in both river waters, Yeongsan River and Seomjin River (data are not shown).This might infer that 2-hydroxyibuprofen is much more persistent and stable than 1-hydroxyibuprofen and even than parent compound, ibuprofen, as previously mentioned by Weigel et al. (2004).
Even though the levels of 10,11-epoxy-carbamazepine was low, the eco-toxicological effect of CBZ-EP on the environment is worthy of further examination due to its pharmaceutically active property like its parent compound, carbamazepine.The most dominant carbamazepine metabolite was 10-hydroxy-carbamazepine.The predominant formation of CBZ-10OH might be due to the different distribution of electron density of atoms in carbamazepine.

Effect of electron density on transformation of carbamazepine into metabolites
In previous studies, electron density distribution of chemicals has been used to find the initial positions of OH radical attack in oxidation (Watanabe et al., 2003;Jung et al., 2010;Heimstad et al., 2009).A higher electron density indicates more electrons in the bonds, resulting in electrophilic reaction (Horikoshi et al., 2004;Kaneco et al., 2006).In this study, electron density of pharmaceutical was examined to find the initial transformation position in various transformation mechanisms including biological process and photochemical oxidations.In previous study, Park et al. (2009) suggested that hydrolysis reaction at the amide and urea functional groups may lead to the biological transformation of carbamazepine.Frontier electron density of carbamazepine calculated by MOPAC, Scigress software is shown in Table 5. Radical frontier density, the averaged value of nucleophilic frontier density and electrophilic frontier density, are used in this study.Based on the electron density of carbamazepine, the carbon bonds between C 1 and C 2 showed the largest frontier electron density (0.208 and 0.214, respectively).This electron rich carbon bond of the olefin structure may provide the initial position of oxidation.Prevalent occurrence of specific carbamazepine metabolites such as CBZ-EP, CBZ-10OH, and CBZ-DiOH in environment are presumably related with this oxidation pathway.However, other metabolites were difficult to find a relationship between transformation and electron density.The reason for this low relationship is thought that there might be some other different preferred transformation pathways depending on the intrinsic property of chemicals rather than electron density derived transformation.However, electron density of chemicals is still believed as a key parameter to elucidate unknown pathway in various processes, especially in oxidation process.

Conclusions
Removal and transformation of pharmaceuticals in WWTPs and constructed wetlands were extensively investigated in this study.Pharmaceuticals were effectively removed by different WWTP processes and wetlands.From this study, the additional operation of wastewater management wetlands was encouraged to prevent direct discharge of micropollutants into surface waters.And additionally, pharmaceutical metabolites were found to be more stable than the parent compounds during WWTP processes due to the lower log D value of metabolites than parent compounds.Different transformation pattern of pharmaceuticals was also observed, especially in transformation of ibuprofen.1-hydroxyibuprofen was dominantly formed during biological treatment in WWTP, indicating preferential biotransformation of ibuprofen.At last, electron density of carbamazepine was examined to elucidate the transformation pathway.The electron rich C 1 = C 2 bond in olefin structure of carbamazepine was revealed as an initial transformation position.

Table 1 .
Tested parent compounds and metabolites.Table 1.Tested parent compounds and metabolites.

Table 3 .
Concentrations of selected PPCPs in WWTPs and wetlands.

Table 4 .
Concentrations of parent pharmaceuticals and their metabolites in WWTPs and wetlands.
Removal of PPCPs in municipal wastewater treatment plants and constructed wetlands were investigated and summarized by comparing the concentrations of PPCPs in the influent and final effluent of each WWTP in Table

Table 5 .
Calculations of frontier electron density for carbamazepine.