Long-Term Comparison of Trace Organics Removal Performances Between Conventional and Membrane Activated Sludge Processes

By Zuehlke, S; Duennbier, U; Lesjean, B; Gnirss, R; Buisson, H

ABSTRACT: The removal of trace organic compounds through membrane bioreactors (MBR) compared with a conventional wastewater treatment plant (WWTP) in a long-term study was investigated. Two MBR pilot plants were operated in parallel to a full-scale WWTP fed with the same municipal raw wastewater. Polar compounds (phenazone-type pharmaceuticals. their metabolites, and carbamazepine) and less polar estrogenic steroids (estradiol, estrone, and ethinylestradiol) were quantified.

The removal rate of phenazone, propyphenazone, and formylaminoantipyrine by the conventional WWTP was less than 15%. Significantly higher removal rates (60 to 70%) started to be clearly monitored in the pilot plants after approximately 5 months. Higher removal rates coincided with higher temperatures in the summer.

The conventional WWTP removed, on average, more than 90% of the natural steroids estrone and estradiol and approximately 80% of the synthetic ethinylestradiol. Approximately 99% of estradiol and estrone and approximately 95% of ethinylestradiol was eliminated by the MBR processes. Water Environ. Res., 78, 2480 (2006).

KEYWORDS: endocrine-disrupting compounds, pharmaceutical residues, wastewater treatment, membrane bioreactor, steroids.

doi:10.2175/106143006X111826

Introduction

In recent years, the occurrence and fate of pharmaceutical residues and endocrine-disrupting compounds (EDCs) in the aquatic environment has been recognized as an important issue in environmental chemistry (Daughton and Jones-Lepp, 2001; Daughton and Temes, 1999; Halling-Srensen et al., 1998; Heberer, 2002; Johnson and Sumpter, 2001; Kmmerer, 2001). Pharmaceuticals applied in human medical care are not completely eliminated in the human body, and they are excreted with only slightly transformation or even unchanged and mostly conjugated to polar molecules. Previous investigations have shown evidence that pharmaceutical residues and EDCs are often not or not completely eliminated during wastewater treatment (Andersen et al., 2003; Daughton and Ternes, 1999; Heberer, 2002; Ternes, 1998). Thus, wastewater treatment plants (WWTPs) act, for some trace organic compounds, as significant point sources of surface water pollution. Though only inadequate data are available about the effects of pharmaceutical residues in the aquatic environment (Jones et al., 2004), studies reported a higher reproductive abnormality in fish if rivers receive WWTP discharges (Harries et al., 1996; Purdom et al., 1994). This was accounted for by the presence of steroid compounds. Thus, the discharge of wastewater effluents must be considered a source of environmental EDC pollution. However, it is questionable whether classical standardized tests give reliable data needed for environmental risk assessment of pharmaceutical residues (Kmmerer et al., 2004). Only a few ecotoxicological studies for the assessment of the environmental risk of these substances are available. Algal tests and acute Daphnia tests were performed, and measured EC^sub 50^ differed with values in the milligram-per-liter range (clofibric acid and carbamazepine) (Cleuvers, 2002). The substances followed the concept of concentration addition, which means that compounds can contribute to mixture toxicity, even in concentrations beneath their individual effect threshold (Cleuvers, 2004).

Because the Berliner Wasserbetriebe (Germany) recharges groundwater by artificial surface water infiltration, wastewater treatment must be of a high standard, and the fate of trace organic compounds (endocrine disrupters and pharmaceuticals) must be carefully studied.

The technology of membrane activated sludge or membrane bioreactor (MBR) has already been implemented in full-scale WWTPs since the early 1990s. It was progressively adapted for nitrogen and phosphorus removal. As a result of their specific design and operation characteristics, resulting from the use of a micro- or ultrafiltration membrane to separate the sludge (physical selection depending on size), instead of a sedimentation tank (gravity separation), MBRs are also considered as promising technologies to achieve further removal of micropollutants. Through its specific features, MBR technology may improve three of the four elimination mechanisms in the activated sludge process-adsorption, biodegradation, and separation. Volatilization, the last removal mechanism, should be identical between the two technologies. On the other hand, a drawback of the MBR technology could be the shorter hydraulic retention time (HRT), characteristic of some MBR plants and resulting from high biomass concentration and small reactor volumes. For these reasons, it was frequently reported that bulk organics (effluent organic matter) or residual soluble chemical oxygen demand (COD) is less after MBR treatment than after conventional activated sludge (CAS) technology. Available information on the performance of MBRs regarding removal of micropollutants, however, is currently limited (Kimura et al., 2005; Lesjean, Gnirss, Dnnbier, and Gibert, 2002). These few results are often based on a very small number of samples and are therefore not statistically representative, and little information is given on the MBR operation conditions. A long-term investigation, with various sludge ages, was envisioned to help understand the elimination mechanisms within MBR technologies, and it enabled determination of the best operating conditions to remove trace organic compounds. The investigation, with two MBR plants, was undertaken in parallel to a conventional WWTP, which was operated under a relatively constant sludge age. This enabled the direct comparison of treatment performances between membrane and conventional activated sludge plants. Carbamazepine, phenazone, propyphenazone, and several metabolites of phenazone-type pharmaceuticals were selected as indicators of high polar pharmaceutically active compounds. The estrogenic steroids estrone (E1), 17β-estradiol (E2), and 17α-ethinylestradiol (EE2) were chosen as less polar compounds of most ecotoxicological concern.

Methodology

Description of the Membrane Bioreactor Systems. Two multistage MBR pilot plants (PPl and PP2) were operated in parallel to a conventional enhanced biological phosphorus removal (EBPR) WWTP (Berlin-Ruhleben, Germany), designed with standard predenitrification. The pilot plants were fed with raw waste water of the WWTP (Ruhleben) and screened through 1-mm punch holes (rotative drum). This plant receives the wastewater from a large combined sewer catchment located in the central western districts of Berlin. It is therefore prone to storm water effects and is combined with industrial effluents. Detailed characterization of the raw wastewater and the conventional WWTP is given in Gnirss, Lesjean, and Buisson (2003).

The pilot facility included a pretreatment step (1-mm screening), two parallel biological reactors, two membrane units, and a control- and-acquisition system. The PP1 and PP2 featured the two process configurations presented in Figure 1 (EBPR with pre- and post- denitrification without carbon addition, respectively). The filtration systems were developed by Memcor (South Windsor, Australia). They included one immersed hollow fiber polyvinylidene fluoride membrane module each, set up in a membrane tank (pore size approximately 0.2 m and membrane area approximately 9 m^sup 2^).

The MBR system has a membrane fouling problem, which makes the system less competitive. In general, membrane fouling occurs more seriously on hydrophobic membranes than hydrophilic ones because of hydrophobic interaction between solutes, microbial cells, and membrane materials. Membrane fouling can also be attributed to the adsorption of organic species, precipitation of less soluble inorganic species, and adhesion of microbial cells at the membrane surface. Lower fouling can be achieved by lowering activated sludge concentration and membrane flux, and the activated sludge should be discharged periodically. The combination backwashing of pressured air and permeated effluent is to remove the cake layer deposited on the membrane surface, decrease membrane fouling, and recover the membrane flux effectively; thus, it is effective for prevention of membrane fouling. In this special MBR, a major influence is the concentration of suspended extracellular polymeric substances (EPS); the higher the suspended EPS concentration, the lower the filtration index (Rosenberger and Kraume, 2003). If necessary, chemical cleaning of blocked membranes was performed with hydrogen peroxide. This procedure was not applied during sampling periods.

Each configuration included successive anaerobic, anoxic, and aerobic zones. Depending on the configuration tested, the anoxic and aerobic zones consisted of three or four reactors. The volume ratio between anoxic and aerobic zone was 50:50 for configuration 1 and 55:45 for configuration 2. The size of the anoxic zone in configuration 1 was guided by experience from full-scale MBR plants. A larger anoxic volume was allowed for in configuration 2, as lower deni\trification velocities were expected. The sludge return and recycle percentages were set up in the two units, at 100% from the anoxic to the anaerobic reactor (R2), 400 to 500% from the membrane reactor to the aerobic zone (R3), and 400% for the sludge recycle from the aerobic to the anoxic reactor (R1, configuration 1 only).

Net flow for the two MBR pilot plants was 120 to 190 L/h, taking into account the excess sludge flow. For comparison, the food-to- microorganism (F/M) ratios of the WWTP (approximately 10 000 m^sup 3^/h) were between 0.13 and 0.18 kgCOD/kgTS . d. The two pilot plants (2.1 and 1.9 m^sup 3^) were always operated in parallel, under equal solid retention times (SRT) and HRTs. After seeding with return sludge coming from the main plant in September 2001, they were first operated, for 6 months, with an SRT of 26 days and HRT of 18 hours. During this period, stable conditions were reached in PP1 and PP2, with average sludge concentrations in the aerated reactors (TS^sub AE^) of 13 and 12.5 g/L and excess sludge of approximately 1060 and 975 g/d, respectively. The sludge age was then lowered stepwise, to 8 days (HRT of 11 hours), over the following year. The flows were accordingly increased stepwise, to ensure that TS^sub AE^ remained in the range 8 to 16 g/L. As a result, the HRT and contact time in all reactors decreased, and the F/M ratio increased, from 0.10 to approximately 0.17 kgCOD/kgTS . d. Given the relatively short trial period of 2 months, stable values of TS^sub AE^ cannot be provided. From January to April 2003, the two MBR plants were operated under a constant sludge age of 8 days, and the following stable conditions were reached in PP1 and PP2, respectively: TS^sub AE^ of 9.1 and 8.4 g/L and excess sludge of approximately 2070 and 1770 g/d. It is important to note that no coagulant was added over the trial period, and the two MBR pilot plants were continuously operated, with only short periods of filtration interruption (<5% of entire time) in the case of membrane chemical cleaning or equipment defects. Water samples were not taken in the 24 hours following an interruption of operation. The conventional WWTP was operated with a sludge age of approximately 15 days and a constant average HRT of approximately 18 hours within the biological stage.

Selection of Pharmaceutical Residues. The antiepileptic drug carbamazepine has frequently been detected in municipal wastewater and surface water samples (Heberer et al., 2001; Ternes, 1998). Investigations of influent and effluent samples from different municipal WWTPs have shown that carbamazepine is not significantly removed during waste water treatment (Ternes, 1998). Phenazone-type pharmaceutical s are widely used as analgesic and antipyretic drugs. Among other things, biotransformation of metamizol in human bodies leads to acetylated and formylated metabolites. These compounds, designated as FAA (fonmylaminoantipyrine) and AAA (acetylaminoantipyrine), can be found in wastewater samples in the low microgram-per-liter level. Phenazone and propyphenazone were detected in their native form in raw and treated wastewater. A metabolite of the analgesic dimethylaminophenazone, 1-acetyl-1- methyl-2-dimethy loxamoyl-2-phenylhydrazide (AMDOPH; Reddersen et al., 2002), was also monitored. All the studied pharmaceutical residues were not sorbed to sediment or wastewater sludge (Zuehlke, 2004).

Selection of Endocrine-Disrupting Compounds. Estrogenic steroids are among the most potent EDCs causing effects in aquatic organisms, even at trace-level concentrations. The synthetic steroid hormone EE2 (prescribed as an oral contraceptive for birth control or estrogen-substitution therapies), the natural hormone E2, and its main metabolite E1 were excreted from the human body into the municipal wastewater and the aquatic environment. Steroid hormones released into the aquatic system were identified as having the highest endocrine-disrupting potential (Johnson and Sumpter, 2001; Snyder et al., 2001; Thorpe et al., 2003). In vivo investigations have also shown that exposure of fish down to 1 ng/L E2 or 0.1 ng/L EE2 provoke feminization in some species (Purdom et al., 1994; Routledge et al., 1998). Because of their ecotoxicological relevance and their described moderate biodegradability, the estrogenic steroids were selected for this investigation as guide compounds with a less polar chemical structure.

Sampling. Twenty-four-hour composite samples were taken from both the pilot plant effluents and the influents and effluents of the municipal WWTP in Ruhleben. The influent and all effluent samples were collected with a time shift, taking into account the HRT of the wastewater in the WWTP or MBRs.

After 3 months of pilot-plant stabilization (December 2001), regular analyses of the pharmaceuticals (carbamazepine, phenazone, propyphenazone, and several metabolites of phenazone-type pharmaceuticals) started with 24-hour composite samples of raw wastewater, conventionally treated wastewater, and the effluents of both MBR pilot plants. Analyses of the estrogenic steroids started in July 2002. In average, pharmaceutical residues and steroids were analyzed in 6 to 16 samples per 2 months and for each type of water.

Analytical Methods. Two analytical methods were applied for the measurement of the estrogens and pharmaceutical residues. Complete validation of both methods has already been described (Zuehlke et al., 2004, 2005). The determination of trace organics was performed by the laboratory of the Berliner Wasserbetriebe. Both methods were based on a solid-phase extraction (SPE) of the analytes on RP-C18 materials. The extraction for the determination of the estrogens also included advanced cleanup steps during the SPE. All samples of treated effluents were extracted unfiltered, whereas raw wastewater was filtered to avoid blocking of the SPE columns. Liquid Chromatographic separation coupled with mass spectrometry was used for the detection of the analytes. To obtain both high sensitivity and selectivity, tandem mass spectrometry (recording multiple reaction monitoring chromatograms) was applied. All methods are suitable for raw and treated wastewater and for surface, ground, and drinking water. The limits of quantification and recoveries of the steroids and the selected pharmaceutical residues are shown in Table 1 (Zuehlke et al., 2004, 2005).

Results

Chemical Oxygen Demand and Nutrient Removal. The average values of the monitored parameters (i.e., nitrate, ammonia, and orthophosphate [o-PO^sub 4^]) in the influent (after prescreening) and effluent of the two pilot plants, over more than 3 months of stable operation with a sludge age of 8 days, are given in Table 2. The influent COD concentration varied between 600 and 1000 mg/L. As expected, effluent COD concentrations of both pilot plants were approximately the same. A comparison with the WWTP Berlin-Ruhleben showed that the effluent concentrations for COD, total nitrogen, and total phosphorus were slightly lower for the MBR process, as a result of the complete retention of paniculate and colloidal fractions by the membranes (Gnirss, Lesjean, Buisson, Adam, and Kraume, 2003). The removal rates are shown for COD, total nitrogen, and total phosphorus (Table 2). The main difference between the pilot plants was the nitrogen removal, which was 84% in PPl (predenitrification) compared with 96% in PP2 (postdenitrification).

Global Organic Parameters. To compare the results on trace organics with the overall removal of the complex organic matrix, it was necessary to monitor a global and standard indicator of organic pollution. Therefore, dissolved organic carbon (DOC) was analyzed in the filtered raw wastewater, and total organic carbon (TOC) was analyzed in the three effluents. Mean DOC in the raw wastewater was 85 mg/L (range 25 to 112 mg/L). The average TOC concentration was 14 mg/L in the WWTP effluent (84% removal) and 12.3 and 12.5 mg/L in the permeates produced by PPl and PP2 (approximately 86% removal), respectively. This slight difference, probably a result of the presence of particles and colloids in the conventional WWTP effluent and rejected by the microfiltration in MBRs, was constant throughout the year and apparently independent of the operation conditions (sludge age, sludge concentration, HRT, temperature, etc.).

Trace Organics. In the following, the average concentrations of trace organics are reported by periods of 2 months, from the start of the analysis in November 2001 (pharmaceutical residues) and July 2002 (steroids). In general, between 6 and 16 series of samples were analyzed during each period of 2 months. This enabled sound statistical evaluation and interpretation of the gathered data and assessment of the differences between the conventional WWTP and the investigated MBRs, with regard to possible seasonal variations or other operation conditions. Besides the mean values, 10 and 90% percentiles are given in Figures 2 and 4.

Pharmaceuticals and Their Metabolites. The behavior of phenazone- type pharmaceutical and their metabolites during wastewater treatment could be observed precisely during this study. The mean concentrations of carbamazepine, AAA, and phenazone are given in Figure 2. These results demonstrate the persistent behavior of carbamazepine through conventional and membrane activated sludge processes. Both treatments were unable to remove the compound. Marginal lower concentrations were even monitored in the raw wastewater, and this could be attributed to a broader deviation resulting from the organic matrix in the raw wastewater. Also, AMDOPH was monitored in the raw wastewater, with values between 0.4 and 1.1 g/L. This compound shows, similar to carbamazepine, persistent behavior during wastewater treatment.

Other phenazone-type pharmaceuticals were partly removed by the conventional WWTP (up to 30% for phenazone, up to 15% for propyphenazo\ne and FAA, and 15 to 40% for AAA, Figure 2). These four compounds were all better removed by the MBR processes. The PPl removed 25 to 70% of phenazone, up to 65% of propyphenazone, 10 to 65% of FAA, and 35 to 65% of AAA, respectively. The PP2 removed 25 to 40% of phenazone, 30% of propyphenazone, 5 to 20% of FAA, and 30 to 60% of AAA, respectively.

Endocrine-Disrupting Compounds. The three analyzed steroids were all detected at the nanogram-per-liter level in the degritted wastewater. The El was always beyond the quantification limit (0.2 ng/L) in effluent samples of the WWTP and the two MBR pilot plants, and EE2 in all samples of the conventional WWTP was higher than 0.4 ng/L. The E2 was often below the quantification limit (0.4 ng/L) in all three effluents.

The relative standard deviation was approximately 50% for El and E2 and approximately 85% for EE2 in both the influent and the effluent samples. This demonstrates the variability of the estrogen loads at the WWTP inlet, especially for EE2, even though 24-hour composite samples were taken. This also justifies the decision to display the data by statistically representative periods of 2 months.

The El and E2 were present in the degritted raw waste water with average concentrations of 108 to 231 ng/L and 11 to 43 ng/L, respectively. The EE2 was found with mean concentrations of 5 to 23 ng/L. These values were comparable with investigations of Andersen et al. (2003) (El: 55 to 77 ng/L; E2: 12 to 20 ng/L; and EE2: 6 to 10 ng/L). The estrogens were present in the effluent of the conventional WWTP at average levels of only 4.4 ng/L for El, 0.8 ng/ L for E2, and 1.3 ng/L for EE2. During nearly the entire period of analyses (except January to June 2003 with PP2), the elimination performances of the MBR plants for El were as good as or better than the conventional plant (Figure 4).

The E2 is partially converted to El in the wastewater network before entering the biological activated plant. Despite the relatively low concentrations of E2 at the plant inlet, this process is further continued in the wastewater biological treatment. The removal rate of E2 was above 90% for all treatment plants (Figure 4), and average effluent concentrations below 2 ng/L were achieved most of the time. The two MBR pilot plants showed slightly better performances than the conventional plant (except March to June 2003 with PP2). However, this was not that significant, as the concentrations after conventional wastewater treatment were already very low.

The removal of the synthetic hormone EE2 ranged between 59 and 82% for the WWTP, 82 and 94% for PPl, and 83 and 92% for PP2 (Figure 4). Therefore, the MBR technology achieved significant improved removal of this steroid. It is important to note that, in most cases, the MBR technology resulted in EE2 concentrations of below or approximately 1 ng/L. The results of March/ April 2003 are also remarkable, as the MBR processes could cope much better with the increasing load than the conventional plant, despite the low sludge age of 8 days applied during this period. The MBR reactors continue to perform high elimination beyond 80%, whereas the conventional plant could not remove more than 60% of the compound.

Discussion

Pharmaceuticals and Their Metabolites. Carbamazepine was present in the range 1.5 to 2 g/L and AMDOPH in the range 0.4 to 1.1 g/L in the influent. No significant removal of the two compounds could be observed through the conventional and membrane activated sludge technologies. This long-term study confirmed that carbamazepine is a very mobile, persistent, and nonbiodegradable compound, as reported elsewhere (Heberer, 2002).

Within the MBRs and the conventional WWTP, AAA exhibited seasonal variations of their removal rates (Figure 3). In the summer months (June to August), removal efficiency increased in all investigated plants. This effect was not directly related to the influent concentration (Figure 2). In the conventional WWTP, operated under relative constant sludge age, and the two MBR plants, operated under varying sludge age, it could be demonstrated that the elimination rate was dependent from the temperature.

Phenazone and propyphenazone are very polar, hardly biodegradable, and known to be mostly persistent through conventional activated sludge WWTPs. Propyphenazone was not significantly removed by the WWTP, and the removal rate of phenazone was 15 to 30%. In the MBRs, the removal rate of the two compounds was monitored up to 70% with the pilot plant in predenitrification mode (PPl) and 40% with the postdenitrification (PP2). The FAA showed comparable behavior as phenazone and propyphenazone in all investigated processes. Removal in the conventional WWTP was approximately 10% and, during membrane activated sludge treatment, up to 65%. Significant effects of temperature were not observed for these three compounds at the MBRs and the conventional WWTP.

It was also observed that the two MBR pilot plants, although operated in parallel under equivalent operation conditions (SRT, HRT, F/M, etc.), exhibited different removal rates of phenazonetype compounds, with generally greater performances observed with the predenitrification configuration (PPl). Until now, these differences could not be seriously explained. Both plants differed only by their configuration design (pre- versus postdenitrification) and the slightly different distribution of anoxic and aerobic volumes (50:50 for PPl and 55:45 for PP2). This observation tends to infer the benefit of improved aeration volume and/or predenitrification configuration to achieve enhanced removal of phenazone-type compounds. The aerobic degradation of propyphenazone could be shown by Mersmann et al. (2003), where only less anaerobic degradation in soil-column experiments could be observed. The better removal of some phenazone-type pharmaceutical residues could also be observed under aerobic conditions during bank filtration (Massmann et al., 2006). Other studies reported that, compared with CAS, MBRs exhibited much better or comparable removal of pharmaceuticals (i.e., naproxen and ketoprofen) (Kimura et al., 2005). This may be a result of activated sludge deposition onto the membrane surface (Urase et al., 2005).

Endocrine-Disrupting Compounds. The concentrations of the estrogenic steroids monitored in the degritted wastewater of the WWTP (Ruhleben) were approximately 100 ng/L for El, 10 to 40 ng/L for E2, and 5 to 20 ng/L for EE2. In the effluents of six investigated Italian WWTPs, the concentrations of El and E2 were measured with 9.3 and 1 ng/L, respectively, and EE2 was found at 0.5 ng/L (Baronti et al., 2000). These results were comparable with those WWTP Ruhleben effluent, although the concentration of EE2 increased to 10 ng/L (March/April 2003). This high concentration seems to result from an increase of the influent concentration (approximately 60%). Our investigation confirmed the results of Baronti et al. (2000) and Andersen et al. (2003), that operational parameters, such as HRT or temperature, do not seem to explain significant differences in elimination rates of the estrogenic steroids.

After the adaptation period, both MBR plants showed very low concentrations of El and E2 (approximately 99% removal) and EE2 (approximately 90% removal). These removal rates were achieved with different sludge concentrations in the aerated tank ( 13 to 9 g/L) and different sludge ages (20 to 12 days). A higher influent concentration of EE2 (Mar/Apr 2003) led to lower removal rates within the conventional WWTP. The MBR processes could cope much better with the increased concentration of EE2 in March/April 2003. The MBR reactors continued to perform high elimination rates beyond 80%, whereas the conventional WWTP could not remove more than 60% of this compound.

On closer examination, the comparison of the elimination of El between July/August 2002 and May/June 2003 is very informative; despite similar influent concentrations, a different absolute elimination is obvious for both MBR plants. It appears that lower performances were achieved with a low sludge age of 8 days. This should not occur if sorption was the only removal process, as daily excess sludge withdrawal increased by approximately 90% between both periods. These results tend to confirm the effect of biological processes in the removal of steroid compounds (Andersen et al., 2003).

Conclusions

This long-term comparative study on the removal of trace organic compounds by conventional activated sludge treatment technology and the MBR process demonstrated that, for most of the studied compounds, the MBRs showed significantly better removal rates than the conventional WWTP. However, the MBR plants exhibited chronological performance variation, a probable consequence of the change of operating conditions (i.e., temperature). Several observations indicate the significance of biodegradation in the removal mechanism of polar (pharmaceuticals) and less polar (steroids) trace organics. These observations were dependent on operational parameters, such as reactor temperature, loading rates, length of the period required for process improvement and stabilization, and at least the demonstration that no pharmaceutical compound was adsorbed on the biomass.

No removal of carbamazepine or AMDOPH was observed during conventional and membrane activated sludge treatment. Only slight removal of phenazone, propyphenazone, and FAA was monitored by the conventional WWTP (below 15%), but removal was significant higher by the MBRs. The removal of the drug metabolite AAA during conventional treatment was below 30%. In comparison, the removal of these compounds in both permeates reached 70%. Higher removal rates coincided with higher temperatures during the summer months. Increased MBR performances resulted mainly from improved biodegradation mechanisms. This is an indirect effect of the use of micro- or Ultrafiltrat\ion membranes, which warrant the complete physical retention of the microorganisms and therefore enable the cultivation and enrichment of slow-growing metabolic specialists. It was also observed that unusually long adaptation phases (approximately S months) were required for this slow biological acclimation to reach full completion.

Additional, the two pilot plants, operated in parallel under equivalent operation conditions, showed different performances. Larger aerobic volumes and/or predenitrification systems could favor the development of specialist microorganisms, performing the aerobic degradation of the pharmaceutical residues.

Wastewater treatment (CAS and MBR) reduces the concentration of most of the investigated pharmaceutical residues and estrogenic steroids. The MBR technology was shown to achieve enhanced elimination of trace organics compared with conventional activated sludge treatment. However, it should be pointed out that the MBR process can achieve only partial (phenazone-type compounds) or no elimination (such as carbamazepine or AMDOPH). Comparison with CAS treatment, removal by the MBR is only slightly better. Thus, the MBR technology does not provide any stand-alone solution when extensive or complete removal is required for all compounds, for example, in the case of direct or indirect potable water reuse.

References

Andersen, H.; Siegrist, H.; Halling-Srensen, B.; Ternes, T. A. (2003), Fate of Estrogens in a Municipal Sewage Treatment Plant. Environ. Sci. Technol., 37 (18), 4021.

Baronti, C.; Curini, R.; D’Ascenzo, G.; Di Corcia, A.; Gentili, A.; Samperi, R. (2000) Monitoring Natural and Synthetic Estrogens at Activated Sludge Sewage Treatment Plants and in a Receiving River Water. Environ. Sci. Technol., 34 (24), 5059-5066.

Cleuvers, M. (2002) Aquatic Ecotoxicology of Selected Pharmaceutical Agents-Algal and Acute Daphnia Tests. Umweltwissenschaften Schadstoff-Forschung, 14 (2), 85-89.

Cleuvers, M. (2004) Mixture Toxicity of the Anti-Inflammatory Drugs Diclofenac, Ibuprofen, Naproxen, and Acetylsalicylic Acid. Ecotoxicol. Environ. Saf, 59 (3), 309-315.

Daughton, C. G.; Jones-Lepp, T. (Eds.) (2001) Pharmaceutical and Personal Care Products in the Environment: Scientific and Regulatory Issues, Symposium Series 791; American Chemical Society: Washington, D. C.

Daughton, C. G.; Ternes, T. A. (1999) Pharmaceuticals and Personal Care Products in the Environment: Agents of Subtle Change? Environ. Health Perspect., 107 (6), 907.

Gnirss, R.; Lesjean, B.; Buisson. H. (2003) IMF-Immersed Membrane Filtration: Cost-Effective Biological Treatment in Membrane Bioreactor for Decentralized Areas, Wasser Berlin 2003, International Trade Fair and Congress Water and Waslewater, Berlin, Germany, April 7-11; Messe Berlin: Germany.

Gnirss, R.; Lesjean, B.; Buisson, H.; Adam, C.; Kraume, M. (2003) Enhanced Biological Phosphorus Removal with Post-Denitrification in Membrane Bioreactor, Atlanta, Georgia, March 2-5; American Water Works Association: Denver, Colorado.

Halling-Srensen, B.; Nielsen, N.; Lansky, P. F.; Ingerslev, F.; Hansen, L.; Liitzhft, H. C.; Jrgensen, S. E. (1998) Occurrence, Fate and Effects of Pharmaceutical Substances in the Environment-A Review. Chemosphere, 36 (2), 357.

Harries, J. E.; Sheahan, D. A.; Jobling, S.; Matthiessen, P.; Neall, P.; Routledge, E. J.; Rycroft, R.; Sumpter, J. P.; Tyler, T. (1996) A Survey of Estrogenic Activity in United Kingdom Inland Waters, Environ. Toxicol. Chem., 15 (11), 1993.

Heberer, T. (2002) Occurrence, Fate and Removal of Pharmaceutical Residues in the Aquatic Environment: A Review of Recent Research Data. Toxicol. Lett., 131 (1-2), 5.

Heberer, T.; Fuhrmann, B.; Schmidt-Bumler, K.; Tsipi, D.; Koutsouba, V.; Hiskia, A. (2001) Occurrence of Pharmaceutical Residues in Sewage, River, Ground and Drinking Water in Greece and Germany. In Pharmaceuticals and Personal Care Products in the Environment: Scientific and Regulatory Issues, Symposium Series 791, Daughton, C. G., Jones-Lepp, T. (Eds.); American Chemical Society: Washington, D.C., 70-83.

Johnson, A. C.; Sumpter, J. P. (2001) Removal of Endocrine- Disrupting Chemicals in Activated Sludge Treatment Works. Environ. Sd. Technol., 35 (24), 4697.

Jones, O. A. H.; Voulvoulis. N.; Lester, J. N. (2004) Potential Ecological and Human Health Risks Associated with the Presence of Pharmaceutically Active Compounds in the Aquatic Environment. Crit. Rev. Toxicol., 34 (4), 335-350.

Kimura, K.; Hara, H.; Watanabe, Y. (2005) Removal of Pharmaceutical Compounds by Submerged Membrane Bioreactors (MBRs). Desalination, 178(1-3), 135-140.

Kmmerer, K. (2001) Drugs in the Environment: Emission of Drugs, Diagnostic Aids, and Disinfectants into Wastewater by Hospitals in Relation to Other Sources-A Review. Chemosphere, 45 (6-7), 957.

Kmmerer, K.; Alexy, R.; Huttig, J.; Scholl, A. (2004) Standardized Tests Fail to Assess the Effects of Antibiotics on Environmental Bacteria. Waler Res., 38 (8), 2111-2116.

Lesjean, B.; Gnirss, R.; Dnnbier, U.; Gibert, M. (2002) Removal of Trace Organics Through Membrane Bioreactor Processes-Presumptive Advantages, Review, and Discussion on Appropriate Indicators. Proceedings of the 2nd World Water Congress of the International Water Association, Melbourne, Australia, April 7-12; International Water Association: London.

Massmann, G.; Greskowiak, J.; Duennbier, U.; Zuehlke, S.; Knappe, A.; Pekdeger, A. (2006) The Impact of Variable Temperatures on the Redox Conditions and the Behavior of Pharmaceutical Residues During Artificial Recharge. J. Hydrol., 328 (1-2), 141-156.

Mersmann, P.; Scheytt, T.; Heberer, T. (2003) Column Experiments on the Transport Behavior of Pharmaceutically Active Compounds in the Saturated Zone. Ada Hydnochimica et Hydmbiologica, 30 (5-6), 275- 284.

Purdom, C. E.; Hardimann, P. A.; Bye, V. J.; Eno, N. C.; Tyler, C. R.; Sumpter, J. P. (1994) Estrogenic Effects of Effluents from Sewage Treatment Works. Chem. Ecol., 8 (4), 275.

Reddersen, K.; Heberer, T.; Duennbier, U. (2002) Identification and Significance of Phenazone Drugs and Their Metabolites in Ground- and Drinking Water. Chemosphere, 49 (6), 539.

Rosenberger, S.; Kraume, M. (2003) Filterability of Activated Sludge in Membrane Bioreactors. Desalination, 151 (2), 195-200.

Routledge, E. J.; Sheahan, D.; Desbrow, C.; Brighty, C. G.; Waldock, M.; Sumpter, J. P. (1998) Identification of Estrogenic Chemicals in STW Effluent. 2. In Vivo Responses in Trout and Roach. Environ. Sci. Technol., 32 (11), 1559.

Snyder, S. A.; Villeneuve, D. L.; Snyder, E. M.; Giesy, J. P. (2001) Identification and Quantification of Estrogen Receptor Agonists in Wastewater Effluents. Environ. Sci. Technol., 35 (18), 3620.

Ternes, T. A. (1998) Occurrence of Drugs in German Sewage Treatment Plants and Rivers. Water Res., 32 (11), 3245.

Thorpe, K. L.; Cummings, R. I.; Hutchinson, T. H.; Scholze, M.; Brighty, G.; Sumpter, J. P.; Tyler, C. R. (2003) Relative Potencies and Combination Effects of Steroidal Estrogens in Fish. Environ. Sci. Technol., 6 (6), 1142.

Urase, T.; Kagawa, C.; Kikuta, T. (2005) Factors Affecting Removal of Pharmaceutical Substances and Estrogens in Membrane Separation Bioreactors. Desalination. 178 (1-3), 107-113.

Zuehlke, S. (2004) Behavior of Phenazone-Type Pharmaceuticals and Their Metabolites as Well as Carbamazepine and Estrogenic Steroids During Drinking, Surface and Wastewater Treatment. PhD-thesis, Technical University of Berlin, Germany, http://edocs.tu-berlin.de/ diss/2004/ zuehlke_sebastian.htm.

Zuehlke, S.; Duennbier, U.; Heberer, T. (2005) Determination of Estrogenic Steroids in Surface and Wastewater Applying Liquid Chromatography-Electrospray Tandem Mass Spectrometry. J. Sep. Sci., 28 (1), 52-58.

Zuehlke, S.; Duennbier, U.; Heberer, T. (2004) Determination of Polar Drug Residues in Sewage and Surface Water Applying Liquid Chromatography-Tandem Mass Spectrometry. Anal. Chem., 76 (22), 6548- 6554.

Acknowledgments

Credits. This research project was supported jointly by the Berliner Wasserbetriebe (Germany) and Veolia Water (Berlin, Germany) in the frame of the KompetenzZentrum Wasser Berlin.

Authors. S. Zuehlke is an academic researcher at the Institute of Environmental Research, Dortmund, Germany. U. Duennbier is an academic researcher, and R. Gnirss is a project coordinator at Berliner Wasserbetriebe, Germany. B. Lesjean is a project coordinator, and H. Buisson is a senior project manager at Veolia Water, Berlin, Germany. Correspondence should be addressed to Sebastian Zuehlke, Institute of Environmental Research, University of Dortmund, 44221 Dortmund, Germany; e-mail: S.Zuehlke@infu. uni- dortmund.de.

Submitted for publication November 22,2004 ; revised manuscript submitted March 9, 2006; accepted for publication March 10, 2006

The deadline to submit Discussions of this paper is March 15, 2007.

Copyright Water Environment Federation Dec 2006

(c) 2006 Water Environment Research. Provided by ProQuest Information and Learning. All rights Reserved.