Provided for non-commercial research and educational use only. Not for reproduction or distribution or commercial use.

Transcription

1 Provided for non-commercial research and educational use only. Not for reproduction or distribution or commercial use. This article was originally published by IWA Publishing. IWA Publishing recognizes the retention of the right by the author(s) to photocopy or make single electronic copies of the paper for their own personal use, including for their own classroom use, or the personal use of colleagues, provided the copies are not offered for sale and are not distributed in a systematic way outside of their employing institution. Please note that you are not permitted to post the IWA Publishing PDF version of your paper on your own website or your institution s website or repository. Please direct any queries regarding use or permissions to wst@iwap.co.uk

2 2221 Government of Canada 2014 Water Science & Technology Removal of pharmaceuticals and personal care products in a membrane bioreactor wastewater treatment plant M. Kim, P. Guerra, A. Shah, M. Parsa, M. Alaee and S. A. Smyth ABSTRACT Ninety-nine pharmaceuticals and personal care products (PPCPs) were analyzed in influent, final effluent, and biosolids samples from a wastewater treatment plant employing a membrane bioreactor (MBR). High concentrations in influent were found for acetaminophen, caffeine, metformin, 2-hydroxy-ibuprofen, paraxanthine, ibuprofen, and naproxen ( ng/l). Final effluents contained clarithromycin, metformin, atenolol, carbamazepine, and trimethoprim (>500 ng/l) at the highest concentrations, while triclosan, ciprofloxacin, norfloxacin, triclocarban, metformin, caffeine, ofloxacin, and paraxanthine were found at high concentrations in biosolids (>10 3 ng/g dry weight). PPCP removals varied from 34% to >99% and 23 PPCPs had 90% removal. Of the studied PPCPs, 26 compounds have been rarely or never studied in previous membrane bioreactor (MBR) investigations. The removal pathway showed that acetaminophen, 2-hydroxyibuprofen, naproxen, ibuprofen, codeine, metformin, enalapril, atorvastatin, caffeine, paraxanthine, and cotinine exhibited high degradation/transformation. PPCPs showing strong sorption to solids included triclocarban, triclosan, miconazole, tetracycline, 4-epitetracycline, norfloxacin, ciprofloxacin, doxycycline, paroxetine, and ofloxacin. Trimethoprim, oxycodone, clarithromycin, thiabendazole, hydrochlorothiazide, erythromycin-h 2 O, carbamazepine, meprobamate, and propranolol were not removed during treatment, and clarithromycin was even formed during treatment. This investigation extended our understanding of the occurrence and fate of PPCPs in an MBR process through the analysis of the largest number of compounds in an MBR study to date. Key words biosolids, membrane bioreactor, pharmaceuticals and personal care products, wastewater treatment M. Kim P. Guerra A. Shah M. Parsa M. Alaee S. A. Smyth (corresponding author) Science and Technology Branch, Environment Canada, 867 Lakeshore Road, P.O. Box 5050, Burlington, Canada, L7R 4A6 shirleyanne.smyth@ec.gc.ca INTRODUCTION Membrane bioreactors (MBRs) integrate aerobic biological processes with membrane separation and have become an alternative to conventional activated sludge (CAS) processes for wastewater treatment (Trinh et al. 2012). The advantages of MBR operations include effective capture of suspended solids and elevated biomass concentrations to achieve efficient chemical oxygen demand, suspended solid, nutrient, and pathogen removals (Li et al. 2009). Pharmaceuticals and personal care products (PPCPs) are present in wastewater; however, studies on the effectiveness of MBR processes in removing PPCPs are limited compared to studies in CAS processes (Trinh et al. 2012; Verlicchi et al. 2012). Previous investigations have demonstrated that PPCP removals from MBR processes ranged from no removal to complete elimination depending on operating conditions and PPCP characteristics (Verlicchi et al. 2012). These studies provided useful information, albeit relatively few PPCPs (<40) were studied under different operating conditions. Thus, the comparison of the removal pattern of different PPCPs in similar conditions was limited. In this study, 99 PPCPs in influent, final effluent, and biosolids samples were collected from an MBR plant employing hollow-fibre membranes. Removal efficiencies of these compounds were determined using influent and final effluent concentrations. The target PPCPs included 11 analgesic/ anti-inflammatories, 47 antibiotics, two antidiabetics, four antifungal/antibacterials, seven antihypertensive drugs, two drugs affecting the respiratory system, three lipid regulator agents, seven psychiatric drug, and three psychomotor stimulants. Thirteen PPCPs used as H1 receptor blockers, doi: /wst

3 2222 M. Kim et al. PPCP removal in an MBR wastewater treatment plant Water Science & Technology diuretics, and H2 histamine receptor blockers were grouped in an other category. The aim of this study was to investigate the fate of 99 PPCPs in liquid and solid stream of a wastewater treatment plant (WWTP), developing the largest data set of PPCPs studied in MBR processes. samples were sub-sampled into 1,000 ml high-density polyethylene bottles and shipped to the laboratory on ice by overnight courier. This sampling protocol provides information about variability in wastewater composition over 3 days, as discussed by Ort et al. (2010). MATERIALS AND METHODS WWTP operation and sample collection Influent, final effluent, and biosolids samples were collected from a Canadian full-scale WWTP equipped with an MBR serving 24,000 inhabitants. The plant consisted of a fine screen (2 mm mesh), primary clarification, a bioreactor, and a membrane tank. The bioreactor had an anoxic zone ahead of the aeration zone, and the dissolved oxygen concentration in the aeration zone was maintained at 2.5 mg/l. Ferric chloride was added to primary effluent for phosphorus removal. The membrane tanks employed submerged membrane modules made of hollow-fibre membranes (GE Zeeweed 500d, 0.04 μm pore size). Four cassettes of membrane modules were placed in each of the six trains. Scour air was supplied to mitigate membrane fouling. For maintenance, membranes were regularly cleaned with sodium hypochlorite and citric acid. MBR permeates were disinfected by ultra-violet treatment. Excess sludge was thickened by membranes and incinerated. During the sampling campaign, the following operational conditions were observed: average flow rate of 8,800 m 3 /d, influent/effluent temperature of 21 W C, average hydraulic retention time (HRT) of 11 h, solids retention time (SRT) of 6 8 days, mixed liquor suspended solids (MLSS) concentrations of 5,700 mg/l, and sludge generation of 90 m 3 /day (4.2% solids). Influents and effluents were collected for three consecutive days using Hach Sigma 900 refrigerated autosamplers (Hach Company, Loveland, CO, USA) to obtain 24-h equal volume composite samples at 400 ml every 30 minutes. Biosolids were obtained by grab samples from the holding tank that blended primary sludge and membrane-thickened waste activated sludge. The use of grab sampling for biosolids is considered to be representative because the biosolids were predominantly composed of thickened waste activated sludge that was stabilized during biological wastewater treatment. The treatment decreased variability in solid characteristics, producing a more homogeneous mixture. This sampling method is comparable to a composite in short-term period sampling (Guerra et al. 2014). Wastewater and biosolids Chemical analysis Ninety-nine PPCPs (Table S1, available online at were analyzed by AXYS Analytical Services (Sidney, BC, Canada) based on EPA method 1694 (US EPA 2007). Twenty-seven compounds were added to the analytical protocol, which are marked in Table S1. This method is based on acidic and basic extractions starting with 500 ml of previously filtered wastewater and 0.5 g of dry biosolids. In short, each sample was adjusted for acidic and basic conditions and was extracted using Waters Oasis HLB 20, 1 g 60 µm particle size. The quantitative determinations were accomplished using a Waters 2795 liquid chromatograph (Milford, MA, USA) coupled to a Micromass Quattro Ultima triple quadruple mass spectrometer (Manchester, UK) operated in multiple reaction monitoring mode. Target compounds were quantified using isotope dilution technique. Details of the analysis are provided in the supporting information and analytical conditions are presented in Tables S2 and S3 (available online at Mass balances Fractions (%) of mass loading of compounds that were degraded/transformed, partitioned to biosolids, and discharged through final effluent were estimated by Equations (1) (5) Influent ðmg=dþ ¼ sorption ðmg=dþþfinal effluent ðmg=dþ þ degradation=transformation ðmg=dþ (1) Degradation=transformation ðmg=dþ ¼ influent ðmg=dþ sorption ðmg=dþ final effluent ðmg=dþ (2) Degradation=transformation ð% Þ ¼ Degradation=transformation ðmg=dþ 100=influent ðmg=dþ (3) Sorption onto sludge ð% Þ ¼ sorption ðmg=dþ 100=influent ðmg=dþ (4)

4 2223 M. Kim et al. PPCP removal in an MBR wastewater treatment plant Water Science & Technology Final effluent ð% Þ ¼ final effluent ðmg=dþ 100=influent ðmg=dþ (5) Equation (1) assumed that compounds in influent degraded/transformed, sorbed to sludge, or remained in final effluent. No accumulation of PPCPs was assumed in order to simplify the mass balances. Degradation/transformation mass loading (Equation (2)) was estimated by subtracting influent mass loading from combined mass loadings of sorption and final effluent. However, some compounds showed negative mass loading of degradation/transformation because mass loading in biosolids was higher than influent mass loading, particularly for compounds that show a strong affinity to solids, marked with asterisk in Table 1. Possible reasons can be: (1) particulate portions in influent and final effluent were excluded in our analysis to simplify estimation of sorption coefficients; (2) some compounds are produced through transformation of parent compounds; and (3) generally sample variations were wider in influent samples than solid samples (Song et al. 2006). In this case, it was assumed that combined mass loading of sorption and effluent was close to actual influent mass loading, replacing influent value in Equations (3) (5). Additionally, degradation/transformation mass loading was considered negligible in Equation (3). RESULTS AND DISCUSSION During the 3 sampling days, this MBR plant showed good performance for conventional parameters. Water qualities of influent wastewater and final effluent (average ± standard deviation) were, respectively, 420 ± 87 and 17 ± 6 mg/l for chemical oxygen demand, 38 ± 0.6 and 0.46 ± 0.2 mg/l for total Kjeldahl nitrogen, 5.3 ± 0.7 and 0.4 ± 0.1 mg/l for total phosphorus, and 190 ± 68 and <2 mg/l for total suspended solids. Combined average concentrations of nitrite and nitrate nitrogen in influent and effluent were 0.14 ± 0.0 and 22 ± 1.6 mg/l, respectively; this considerable increase from influent to effluent indicated substantial nitrification during wastewater treatment. Concentrations of PPCPs in influent, final effluent and biosolids Fifty PPCPs were not detected in influent wastewater, with detection limits ranging from 1.1 to 2,200 ng/l (median 24 ng/l) (Table S4, available online at wst/069/145.pdf). Median concentrations of the 49 detected PPCPs are presented in Figure 1(a). Ranked from highest to lowest, acetaminophen, caffeine, metformin, 2-hydroxy-ibuprofen, paraxanthine, ibuprofen, and naproxen were found at the highest concentrations with median levels (n ¼ 3) ranging from 10 4 to 10 5 ng/l. The combined concentrations of these compounds ( ng/l), which belong to analgesic/anti-inflammatories, psychomotor stimulants, and antidiabetic categories, accounted for 96% of total concentrations ( ng/l) of detected PPCPs in influent, indicating that these seven compounds were the most prevalent PPCPs in this WWTP. The high concentrations of acetaminophen, caffeine, ibuprofen, metformin, and naproxen were also consistent with previous reports (Radjenovic et al. 2007; Ziylan & Ince 2011; Trinh et al. 2012). The second highest group of PPCPs found in influent ranged from 10 3 to 10 4 ng/l. These compounds, ranked from the highest to the lowest, include atenolol, triclosan, diphenhydramine, codeine, diltiazem, ciprofloxacin, cotinine, and sulfamethoxazole. Antibiotics in influent were found at <10 3 ng/l, similar to a recent review ( ng/l) (Michael et al. 2013). This study agreed with previous research in that ciprofloxacin, sulfamethoxazole, and clarithromycin were compounds with high concentrations among antibiotics (Michael et al. 2013). One important factor to PPCP occurrence in influent wastewater could be input from medical facilities. However, this studied WWTP sewer shed did not receive discharges from hospitals, which are known to be a significant point source of PPCPs (Michael et al. 2013). Instead, there were several small medical facilities including a nursing home (<200 residents), medical clinics, and veterinary clinics. It is presumed that these inputs did not exert significant influence on PPCP concentrations in this study. In final effluent, 58 PPCPs were not detected, with detection limits ranging from 0.3 to 310 ng/l (median 7 ng/l) (Table S4). The 41 PPCPs detected in final effluent are presented in Figure 1(b). Median concentrations of these PPCPs ranged from 10 1 to 10 3 ng/l. Among the different therapeutic categories, of the total concentration provided by 41 compounds (6,800 ng/l), 40% was from 11 antibiotics, 22% from five antihypertensive drugs, 13% from one antidiabetic, and 11% from seven psychiatric drugs. Compounds with concentrations greater than 500 ng/l were clarithromycin, metformin, atenolol, and carbamazepine, all of which also showed higher effluent concentrations in a previous study (Trinh et al. 2012). The combined concentration of these four compounds was 3,500 ng/l, accounting for 51% of the overall concentration. Thirteen compounds were present at concentrations between 100 and 500 ng/l: trimethoprim, sulfamethoxazole, diltiazem, metoprolol,

5 2224 M. Kim et al. PPCP removal in an MBR wastewater treatment plant Water Science & Technology Table 1 Fractions of degradation/transformation, sorption, and residuals in final effluent Fractions (%) Therapeutic class Compounds Total removal (%) Degradation/transformation (%) Sorption onto sludge (%) Final effluent (%) Analgesic/anti-inflammatories Acetaminophen Ibuprofen Naproxen Hydroxy-ibuprofen Codeine Methylprednisolone Oxycodone Antibiotics Tetracycline * Epitetracycline * Doxycycline Norfloxacin * Ciprofloxacin * Azithromycin Sulfamethoxazole Ofloxacin * Erythromycin-H 2 O Trimethoprim * Clarithromycin * Antidiabetics Metformin Antifungal/antibacterial Triclosan * Miconazole * Triclocarban * Thiabendazole * Enalapril Furosemide Atenolol Diltiazem Metoprolol Propranolol Drugs affecting respiratory system Albuterol Lipid regulator agents Atorvastatin Gemfibrozil Psychiatric drugs Amitriptyline Paroxetine * Diazepam Fluoxetine * Carbamazepine Alprazolam * Meprobamate * Psychomotor stimulants Caffeine Paraxanthine Cotinine Others Digoxin Diphenhydramine Warfarin Cimetidine Ranitidine Hydrochlorothiazide * Notes: (i) Compounds that were rarely or never previously studied in MBR processes are highlighted. (ii) Methods estimating fractions of compounds with asterisk are described in the Mass balances section.

6 2225 M. Kim et al. PPCP removal in an MBR wastewater treatment plant Water Science & Technology hydrochlorothiazide, oxycodone, erythromycin-h 2 O, ranitidine, ciprofloxacin, azithromycin, diphenhydramine, sulphanilamide, and ofloxacin. The combined concentrations of these 13 compounds represented 41% of overall concentrations. Sixty PPCPs were not detected in biosolids, with detection limits ranging from 2 to 610 ng/g dry weight (dw) (median 30 ng/g dw) (Table S4). Thirty-nine PPCPs were detected at median concentrations between 1 and 10 4 ng/g dw, with ciprofloxacin being the highest (Figure 1(c)). Figure 1 Median concentrations of (a) 49 PPCPs in influent wastewater (n ¼ 3), (b) 41 PPCPs in final effluent (n ¼ 3), (c) 39 PPCPs in biosolids (n ¼ 3) (abbreviations of category names are AG/AI for analgesic/anti-inflammatories, ABO for antibiotics, AD for antidiabetics, AF/AB for antifungal/antibacterials, AH for antihypertensive drugs, DR for drugs affecting respiratory system, LR for lipid regulator agents, PC for psychiatric drugs, PM for psychomotor stimulants, and OT for others). (continued)

7 2226 M. Kim et al. PPCP removal in an MBR wastewater treatment plant Water Science & Technology Figure 1 Continued. Total PPCP concentration in biosolids was 52,000 ng/g dw. Triclosan, ciprofloxacin, norfloxacin, triclocarban, metformin, caffeine, ofloxacin, and paraxanthine were present at greater than 10 3 ng/g dw and the combined concentration of these substances was greater than 87% of the total concentration. In early investigations at 74 US WWTPs (US EPA 2009) and 11 Canadian WWTPs (Hydromantis et al. 2010), high concentrations of ciprofloxacin, triclocarban, triclosan, and ofloxacin (>10 3 ng/g dw) were also found in biosolids, indicating significant accumulation of these compounds in biosolids. Among different therapeutic categories, the 11 antibiotic compounds were most concentrated at 24,000 ng/g, which contributed to 46% of overall PPCP concentration in biosolids. Particularly, among the antibiotics, ciprofloxacin, norfloxacin, and ofloxacin were predominant at 92% of the 24,000 ng/g. Four antifungal/antibacterial chemicals (triclosan, tricloban, miconazole, thiabendazole) constituted the second largest group in biosolids at 33% of the overall concentrations with a predominance of triclosan and triclocarban. Overall, comparing influent, final effluent, and biosolids, the predominant therapeutic categories were analgesic/antiinflammatories and psychomotor stimulants for influent, antibiotics and antihypertensive drugs for final effluent, and antibiotics and antifungal/antibacterial drugs for biosolids. Removal of PPCPs To evaluate MBR performance, removal efficiencies (%) of PPCPs were estimated using influent and final effluent concentrations: (PPCP influent PPCP effluent ) 100/PPCP influent. In this calculation, when final effluent concentration was below detection limit, the non-detects were substituted with half the value of the detection limit (Kim et al. 2013). Median removal efficiencies for 49 PPCPs are shown in Figure S1 (available online at wst/069/145.pdf). Removals ranged from 34% to >99% (median 88%); 23 PPCPs had greater than 90% removal. Moderate removals (50 90%), poor removals (10 50%), and no removal (<0 10%) in this study were seen for 14, 6, and 6 PPCPs, respectively. The 26 PPCPs that were rarely or never studied in past MBR investigations were removed by between 6 and >99%. Among the therapeutic category, analgesic/anti-inflammatory compounds showed 7 to >99% removal (median >99%) and oxycodone was resistant to removal. Removals of acetaminophen, ibuprofen, naproxen, and codeine were similar to or better than those reported from previous CAS and MBR studies (Verlicchi et al. 2012). 2-hydroxy-ibuprofen, betamethasone, codeine, and oxycodone were investigated for the first time in MBR studies, and showed a wide range of removals. Antibiotic compounds were removed at 34 to 97% with negative removals for trimethoprim and clarithromycin. A

8 2227 M. Kim et al. PPCP removal in an MBR wastewater treatment plant Water Science & Technology wide range of antibiotic removals was also previously reported (<0to >99%) (Verlicchi et al. 2012). Tetracycline, 4-epitetracycline, doxycycline, and norfloxacin were studied for the first time in an MBR process and showed better removals (>90%) than most previous CAS studies (Verlicchi et al. 2012). The negative removals of trimethoprim and clarithromycin in this study were also reported in previous studies (Gobel et al. 2005; Michael et al. 2013). Negative removal refers to higher concentrations in effluent than influent. It could occur because of the nature of samples; due to variability of PPCP disposal, influent sample characteristics could be more variable than effluent samples that were obtained after a stabilization step (Jelic et al. 2011). More importantly, increase in compound concentration during treatment could possibly be due to the transformation of metabolites; for instance, unmeasured human metabolites and transformation products can be converted to their parent compounds during treatment (Jelic et al. 2011). Additionally, macrolides such as clarithromycin can be released from feces particles during treatment (Gobel et al. 2005). Metformin, an antidiabetic, showed >99% removal, similar to a previous study (Trinh et al. 2012). Due to its high influent concentrations, metformin is the second highest concentrated compound in effluent despite having the highest removal efficiency. Among antifungals and antibacterials, removals of triclosan, miconazole, and triclocarban exceeded 94% while thiabendazole was low at 8%. Removal of the six antihypertensive drugs ranged from 19 to 99%; particularly, enalapril and furosemide removals outperformed previous CAS studies while propranolol removal was lower than other CAS and MBR studies (Verlicchi et al. 2012). Removal of albuterol, a respiratory drug, was low (55%) while atorvastatin and gemfibrozil, and lipid regulator agents were effectively removed (>98%). Psychiatric drug removals ranged from 6 to 85%. Carbamazepine and fluoxetine showed <35% removals. Carbamazepine s poor removal was possibly due to poor degradability and cleavage of the parent compound (Kasprzyk-Hordern et al. 2009). Psychomotor stimulants (caffeine, paraxanthine, cotinine) were highly removed (>98%). Compounds belonging to the other category were removed at 5 94% with lowest removal for hydrochlorothiazide. Removals of digoxin, diphenhydramine, warfarin, and dimetidine, compounds that were included in an MBR study for the first time, were >80%. Previous studies showed that PPCP removals are influenced by operational conditions such as temperature, degree of nitrification, HRT, and SRT (Verlicchi et al. 2012). In the present study, these conditions were 21 W C, over 99%, 11 h and 6 8 days, respectively. Good removals of many PPCPs in this study indicate that the operational conditions of this MBR plant were effective. Summer temperature of 21 W C likely influenced the removal of some compounds, particularly atenolol, enalapril, furosemide, ibuprofen, and sulfamethoxazole, which showed better removals in the summer in a previous study (Castiglioni et al. 2005). Regarding effects of nitrification, a previous study reported that ibuprofen removal was enhanced under nitrifying conditions (Suarez et al. 2010). Hence, it is presumed that complete removal of ibuprofen (100%) in this study was also related to the complete nitrification. HRTs and SRTs are important factors to enhance PPCP removal by lengthening these parameters to elevate degradation and sorption rates (Verlicchi et al. 2012). Most former MBR studies for PPCPs (Trinh et al. 2012; Verlicchi et al. 2012) were operated at >15 days of SRT, longer than the present study (6 8 days). A recent MBR study (Trinh et al. 2012) was also conducted at longer SRT (10 15 days), longer HRT (1 day), and higher MLSS concentration ( g/l) compared to this study (HRT 11 h and MLSS 5.7 g/l). Ibuprofen, naproxen, metformin, caffeine, triclocarban, and triclosan were highly removed in both studies. Moderate removals of carbamazepine, diazepam, fluoxetine, and sulfamethoxazole were also similar in both investigations. However, atenolol and trimethoprim removal, which were 77% and 2% in this study, respectively, were higher in the previous study (atenolol 90% and trimethoprim 30%) (Trinh et al. 2012). In contrast, amitriptyline and gemfibrozil removals, which were >85% in this study, were 20 40% higher than the previous study (Trinh et al. 2012). Although the reason was unclear, it can be presumed that microbial conditions in the previous study might not be effective to remove these two compounds. The previous investigation operated at lower dissolved oxygen concentrations (<1 mg/l) to achieve simultaneous nitrification and denitrification. This could develop different microbial activity than this study running at 2.5 mg/l dissolved oxygen concentrations (Suarez et al. 2010). The comparison between these two studies suggested that the shorter HRTs and SRTs in this study were still effective to treat PPCPs that are well and partially removed. However, SRTs of 6 8 days in this study may not be sufficient for removal of trimethoprim ( 2%) and clarithromycin ( 34%), which previously showed higher removal (<88%) at longer SRTs (16 70 days) (Verlicchi et al. 2012). Nevertheless, this investigation showed high removal for many compounds, indicating that the studied operational conditions can influence PPCP removal from influent to effluent.

9 2228 M. Kim et al. PPCP removal in an MBR wastewater treatment plant Water Science & Technology Removal mechanisms Removal of PPCPs predominantly occurs through two pathways: degradation/transformation and sorption ( Jelic et al. 2011). In order to investigate the different fate of the studied PPCPs, percentage of degradation/transformation, sorption, and residuals in effluent was estimated. For this calculation, mass loading in influent, final effluent, and biosolids was obtained by multiplying compound concentrations by daily flow rate (m 3 /day) for influent and final effluent and daily biosolids production rate (kg total solids/day) for biosolids. Using the mass loadings, the amounts and fractions of degradation/transformation, sorption, and residuals estimated according to Equations (1) (5) are described in the supporting information. Results are summarized in Table 1. Compounds with the most significant degradation/ transformation (>97%) were acetaminophen, 2-hydroxyibuprofen, naproxen, ibuprofen, codeine, metformin, enalapril, atorvastatin, caffeine, paraxanthine, and cotinine. Furosemide, gemfibrozil, diphenhydramine, and digoxin also showed high degradation/transformation (80 92%). Moderate degradation/transformation (34 70%) was seen for methylprednisolone, sulfamethoxazole, diltiazem, ranitidine, warfarin, albuterol, azithromycin, and metoprolol. The remaining compounds were considered very resistant to degradation/transformation. The biological degradation/ transformation proportions of ibuprofen, naproxen, azithromycin, sulfamethoxazole, trimethoprim, enalapril, atenolol, metoprolol, furosemide, gemfibrozil, cimetidine, and ranitidine in this study were similar to previous studies ( Joss et al. 2005; Gobel et al. 2007; Jelic et al. 2011; Ziylan & Ince 2011). Fluoxetine was the exception, with <1% degradation/transformation in this study, substantially lower than the 80 90% achieved in another study (Suarez et al. 2010); however, the latter study operated at much longer SRTs (50 days) with enhanced nitrifying conditions. PPCPs with strong sorption to sludge (sorption fraction >85%) were triclocarban, triclosan, miconazole, tetracycline, 4-epitetracycline, norfloxacin, ciprofloxacin, doxycycline, paroxetine, and ofloxacin. Fluoxetine, amitriptyline, alprazolam, diazepam, and warfarin showed moderate sorption tendency (30 70%). To further understand their sorption tendency, solids liquid distribution (log K d ) coefficients of the compounds were estimated by dividing their concentrations (ng/kg) in biosolids by their respective concentrations (ng/l) in final effluent. These coefficients ranged from 2.0 to 5.8 (Figure S2, available online at Log K d values of the aforementioned 10 compounds with a high sorption fraction (>85%) were , substantiating their strong sorption tendency. Among previous compounds with high degradation/transformation, gemfibrozil (5.2), acetaminophen (5.0), ibuprofen (4.8), naproxen (4.5), and 2-hydroxy-ibuprofen (4.3) also showed high K d values. It indicated that these compounds are both degradable and partitioning onto solids. Our studied K d values were comparable to those in a recent study by Guerra et al. (2014) showing log K d values from 4.1 to 4.6 for compounds with strong sorption tendency. In contrast, in the present study log K d values of compounds with high degradable/transformation were higher than the recent study that reported for acetaminophen, ibuprofen and naproxen. This indicated that different WWTP operational conditions and degree of degradation may result in different K d values. Lastly, PPCPs that persisted through final effluent (71 93% in final effluent, Table 1) were trimethoprim, oxycodone, clarithromycin, thiabendazole, hydrochlorothiazide, erythromycin H 2 O, carbamazepine, meprobamate, and propranolol, indicating that these compounds are slow to degrade/transform and have less affinity to sorption. This classification of PPCPs with respect to their fate indicated that removals of some compounds, particularly compounds with total removal of >90%, predominantly depended on either degradation/transformation or sorption. The remaining compounds were assumed to have moderate degradation/transformation tendency and/or moderate sorption capacity or strong persistency. For instance, among PPCPs with moderate total removals (50 70%), some showed partial degradation/transformation (20 74%), such as azithromycin, atenolol, diltiazem, ranitidine, sulfamethoxazole, albuterol, and diazepam. Removal of these compounds could be enhanced under different operating conditions such as longer HRT and SRT than this study. CONCLUSIONS This study extended our understanding of the occurrence and fate of 99 PPCPs in an MBR plant. In influent wastewater the predominant PPCPs belonged to the analgesic/antiinflammatory, psychomotor stimulant, and antidiabetic therapeutic categories. In contrast, in final effluents antibiotics, antihypertensive drugs, and antidiabetics were largely present while in biosolids antibiotics and antifungal/antibacterials were most concentrated. Removal efficiencies of PPCPs by MBR were diverse, ranging from 34 to >99%. Among them, the majority of the highly concentrated compounds in influent was significantly removed. PPCPs exhibiting

10 2229 M. Kim et al. PPCP removal in an MBR wastewater treatment plant Water Science & Technology significant removals from wastewater (>90%) did so through either degradation/transformation or sorption. Other PPCPs showed partial degradation/transformation, partial sorption capacity, and strong persistency. Further studies on MBR performance at longer HRTs and SRTs may need to be investigated for those compounds that showed partial degradation/transformation and sorption. ACKNOWLEDGEMENTS Funding for this investigation was provided by the Chemicals Management Plan (CMP, Health Canada). Mingu Kim and Paula Guerra thank the CMP for financial support on postdoctoral research. The authors are also grateful to the WWTP manager and operators for their participation in this study, assistance during sampling, and for providing plant information. REFERENCES Castiglioni, S., Bagnati, R., Fanelli, R., Pomati, F., Calamari, D. & Zuccato, E Removal of pharmaceuticals in sewage treatment plants in Italy. Environmental Science & Technology 40 (1), Gobel, A., Thomsen, A., McArdell, C. S., Joss, A. & Giger, W Occurrence and sorption behavior of sulfonamides, macrolides, and trimethoprim in activated sludge treatment. Environmental Science & Technology 39 (11), Gobel, A., McArdell, C. S., Joss, A., Siegrist, H. & Giger, W Fate of sulfonamides, macrolides, and trimethoprim in different wastewater treatment technologies. Science of The Total Environment 372 (2 3), Guerra, P., Kim, M., Shah, A., Alaee, M. & Smyth, S. A Occurrence and fate of antibiotic, analgesic/antiinflammatory, and antifungal compounds in five wastewater treatment processes. Science of the Total Environment , Hydromantis, Inc., University of Waterloo, and Trent University 2010 Emerging Substances of Concern in Biosolids: Concentrations and Effects of Treatment Processes. Canadian Council of Ministers of the Environment. ca/assets/pdf/pn_1445_biosolids_esoc_final_e.pdf. Jelic, A., Gros, M., Ginebreda, A., Cespedes-Sanchez, R., Ventura, F., Petrovic, M. & Barcelo, D Occurrence, partition and removal of pharmaceuticals in sewage water and sludge during wastewater treatment. Water Research 45 (3), Joss, A., Keller, E., Alder, A. C., Gobel, A., McArdell, C. S., Ternes, T. & Siegrist, H Removal of pharmaceuticals and fragrances in biological wastewater treatment. Water Research 39 (14), Kasprzyk-Hordern, B., Dinsdale, R. M. & Guwy, A. J The removal of pharmaceuticals, personal care products, endocrine disruptors and illicit drugs during wastewater treatment and its impact on the quality of receiving waters. Water Research 43 (2), Kim, M., Guerra, P., Theocharides, M., Barclay, K., Smyth, S. A. & Alaee, M Parameters affecting the occurrence and removal of polybrominated diphenyl ethers in twenty Canadian wastewater treatment plants. Water Research 47 (7), Li, F., Wichmann, K. & Otterpohl, R Review of the technological approaches for grey water treatment and reuses. Science of The Total Environment 407 (11), Michael, I., Rizzo, L., McArdell, C. S., Manaia, C. M., Merlin, C., Schwartz, T., Dagot, C. & Fatta-Kassinos, D Urban wastewater treatment plants as hotspots for the release of antibiotics in the environment: a review. Water Research 47 (3), Ort, C., Lawrence, M. G., Reungoat, J. & Mueller, J. F Sampling for PPCPs in wastewater systems: comparison of different sampling modes and optimization strategies. Environmental Science & Technology 44 (16), Radjenovic, J., Petrovic, M. & Barcelo, D Analysis of pharmaceuticals in wastewater and removal using a membrane bioreactor. Analytical and Bioanalytical Chemistry 387 (4), Song, M., Chu, S., Letcher, R. J. & Seth, R Fate, partitioning, and mass loading of polybrominated diphenyl ethers (PBDEs) during the treatment processing of municipal sewage. Environmental Science & Technology 40 (20), Suarez, S., Lema, J. M. & Omil, F Removal of pharmaceutical and personal care products (PPCPs) under nitrifying and denitrifying conditions. Water Research 44 (10), Trinh, T., van den Akker, B., Stuetz, H., Coleman, H. M., Le-Clech, P. & Khan, S. J Removal of trace organic chemical contaminants by a membrane bioreactor. Water Science & Technology 66 (9), US EPA 2007 Method 1694: Pharmaceuticals and Personal Care Products in Water, Soil, Sediment, and Biosolids by HPLC/ MS/MS. US Environmental Protection Agency, epa.gov/scitech/methods/cwa/ppcp/method1694-qna.cf. US EPA 2009 Targeted National Sewage Sludge Survey Statistical Analysis Report. US Environmental Protection Agency, Washington, DC. Verlicchi, P., Al Aukidy, M. & Zambello, E Occurrence of pharmaceutical compounds in urban wastewater: Removal, mass load and environmental risk after a secondary treatment a review. Science of The Total Environment 429, Ziylan, A. & Ince, N. H The occurrence and fate of antiinflammatory and analgesic pharmaceuticals in sewage and fresh water: treatability by conventional and nonconventional processes. Journal of Hazardous Materials 187 (1 3), First received 9 October 2013; accepted in revised form 5 March Available online 20 March 2014