Environ. Sci. Technol. 2006, 40, 6131-6136
Fate of a 14C-Labeled Nonylphenol Isomer in a Laboratory-Scale Membrane Bioreactor
Due to its adverse effects on the endocrine system, nonylphenol (NP) is one of the most intensively studied trace pollutants over the past decade. NP is a representative hydrophobic xenobiotic in wastewater, and it is found as the
recalcitrant degradation product of the nonionic surfactant nonylphenol polyethoxylate (NPnEO) (1-3). Since January 2005 the use and production of NP in Europe is restricted to products containing less than 0.1% NP or NPnEO due to their high toxicity and persistence in aquatic ecosystems (4). In contrast, the United States’ demand for this compound is still increasing with a rate of 2% per year (1). NP consists of a mixture containing 211 possible constitutional isomers (5) with a highly branched alkyl chain which is considerably resistant to biodegradation. In addition, NP sorbs to solid particles due to its hydrophobicity (Log Kow ) 4.5) drastically reducing its bioavailability (1, 6). Although many studies performed in laboratory have demonstrated that NP can be biodegraded in activated sludge (7-9), its detection in sewage treatment plant (STP) effluents (up to 343 µg/L) and sewage sludge (50 -200 µg/g) indicates the inefficiency of the STP for its removal (10-12). Membrane bioreactors (MBRs) represent one of the most promising innovations in the field of wastewater treatment due to their high efficiency in the removal of organic compounds and nutrients, in comparison to classical STP (13). The MBR process combines a biological treatment with a filtration step through special membrane modules, usually with microfiltration pore size. The high quality of the effluent is obtained through the complete retention of suspended solids, the almost complete removal of pathogens, and the possibility to increase biodegradation of micropollutants due to higher sludge retention time (SRT) in the MBR (14, 15). Research on removal of NP from wastewater during MBR treatment has been sparsely documented. A more efficient elimination of NPnEO was observed in MBR (removal rate of 87%) than in a conventional STP (74%) (16). A comparative study carried out in MBR and conventional STP resulted in the removal of a nonylphenolic mixture at rates up to 90% for both processes (17). Comparable results have been reported in MBR and conventional STP operated at the same SRT (12 and 25 d), where no higher degradation of NP was obtained with the MBR (18). In a pilot-scale MBR study operated for treatment of dumpsite leachate, approximately 80% of the NP was eliminated (19). The authors assumed that the adsorption of the compound on suspended matter in the bioreactor was the main removal pathway. From all these pilot-scale studies it remains unclear whether NP is really biodegraded or simply removed by withdrawing excess sludge. A reliable determination of NP adsorbed and entrapped into sludge matrix is often difficult to achieve by means of classic analytical methods. These methods are based on the use of internal standards which are used to spike the samples shortly before the extraction step. They neglect the fact that the recovery of the extraction of the freshly added surrogate standard is much better than that of the compound to be analyzed (20). The radioisotope tracing method is a suitable technique to investigate the removal of trace pollutants. It provides the advantageous possibility of calculating mass balance of target compounds. This method has already been applied in many studies concerning the fate of micropollutants in various systems (21, 22).
* Corresponding author phone: +49-(0)241/80-25871 or 80-27260; fax: +49-(0)241/80-22182; e-mail: magdalena.cirja@ bio5.rwth-aachen.de. † Institute of Environmental Research - Environmental Biology and Chemodynamics (BioV), RWTH Aachen University. ‡ Institute of Environmental Research (INFU), University of Dortmund. § Institute of Environmental Engineering (ISA), RWTH Aachen University.
The strategy implemented in the present study is based on the use of a simple NP isomer present in the technical mixture, i.e., 14C- 4-[1-ethyl-1,3-dimethylpentyl]phenol, in order to investigate the fate of NP during MBR wastewater treatment in a laboratory-scale membrane bioreactor developed for experiments under radioactive conditions. 4-[1ethyl-1,3-dimethylpentyl]phenol has one of the highest estrogenic potencies in tests carried out with MVLN cell lines and is one of the most abundant isomers, accounting for
M A G D A L E N A C I R J A , * ,† SEBASTIAN ZU ¨ HLKE,‡ PAVEL IVASHECHKIN,§ A N D R E A S S C H A¨ F F E R , † A N D PHILIPPE F.X. CORVINI† Institute of Environmental Research - Environmental Biology and Chemodynamics (BioV), RWTH Aachen University, Worringerweg 1, 52074 Aachen, Germany, Institute of Environmental Research (INFU), University of Dortmund, Otto-Hahn-Strasse 6, 44221 Dortmund, Germany, and Institute of Environmental Engineering (ISA), RWTH Aachen University, Mies-van-der-Rohe-Strasse 1, 52074 Aachen, Germany
This study aimed at giving a better insight into the possible fate of nonylphenol (NP) during wastewater treatment by using a lab-scale membrane bioreactor (MBR) designed and optimized for fate studies carried out with radiolabeled compounds. After a single pulse of 14C-labeledNP isomer (4-[1-ethyl-1,3-dimethylpentyl]phenol) as radiotracer, the applied radioactivity was monitored in the MBR system over 34 days. The mass balance of NP residues at the end of the study showed that 42% of the applied radioactivity was recovered in the effluent as degradation products of NP, 21% was removed with the daily excess sludge from the MBR, and 34% was recovered as adsorbed in the component parts of the MBR. A high amount of NP was associated to the sludge during the test period, while degradation compounds were mainly found in the effluent. Partial identification of these metabolites by means of HPLC-tandem mass spectrometry coupled to radio-detection showed they are alkyl-chain oxidation products of NP. The use of this MBR and a radiolabeled test compound was found suitable for demonstrating that under the applied conditions, the elimination of NP through mineralization and volatilization processes (both less than 1%) was negligible. However, the removal of NP via sorption and the continuous release of oxidation products of NP in permeate were of relevance.
Introduction
10.1021/es060668z CCC: $33.50 Published on Web 08/30/2006
2006 American Chemical Society
VOL. 40, NO. 19, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6131
FIGURE 1. Lab-scale membrane bioreactor (MBR) system. Legend: 1, bioreactor vessel; 2, membrane plate modules; 3, influent tank; 4, effluent tank; 5, influent peristaltic pump; 6, air pump; 7, level controller; 8, manometer; 9, controlling valve; 10, effluent peristaltic pump; 11, backwashing peristaltic pump; 12, monoethylene glycol flask for the trapping of volatile organic compounds; 13, NaOH flasks for the trapping of CO2; 14, vacuum pump. approximately 20% from the total technical mixture (23, 24). The use of a radiolabeled isomer should allow for an accurate differentiation between the real biodegradation and abiotic physicochemical phenomena such as adsorption and volatilization, which were recently reported to occur during incubation in soil/sludge systems (25). The specific aim is to gain knowledge on NP elimination in MBR and to detect possible degradation products of NP in the effluent.
Experimental Section Lab-Scale MBR Design. The submerged lab-scale MBR is schematically presented in Figure 1 and consisted of a 1.5 L cylindrical aeration tank made of glass, hermetically closed with steel lids as preventive protection against adsorption and losses of the applied radioactivity to the atmosphere. All the required utilities of the system were connected to the lid of MBR: aeration, filtration, feeding, backwashing, and automatic controller. The submerged membrane module consisted of four plates (A3-Abfall-Abwasser-Anlagentechnik GmbH, Germany) of microfiltration cellulose acetate membrane with a total surface area of 0.024 m2. The MBR operation sequence consisted of 10 min filtration with a flow rate of around 4.0 mL/min and 1 min backwashing at 30 mL/min. The feeding rate was controlled by the level sensor to maintain the volume at 1 L. The membranes were regenerated and fouling was avoided by using only backwashing flow provided from the permeate collector. Mixing and aeration of the mixed liquor solid suspension (MLSS) was achieved by using a porous bubble air diffuser supplying a gas flow of 0.09 m3/h and a magnetic lab stirrer. The transmembrane pressure was measured by means of a manometer inserted on-line. The MBR was inoculated with 1 L of sludge from the Soers pilotscale MBR (Aachen Wastewater Treatment Plant, Germany) and the stabilization phase of the process was initiated. As part of the radioactive test compound from MBR can be stripped out, an additional trapping step for the possible volatilized radioactive compounds was connected to the system for safety reasons and in order to perform an exhaustive balancing of the total radioactivity. The equipment consisted of a 1 L flask filled with 500 mL of monoethylene glycol and connected to the gas outlet of the bioreactor (air supply and biomass respiration) via a pipe. By trapping the gas flow in this solution, 14C volatile organic compounds (VOC) are absorbed into the monoethylene glycol phase. After this step the gas flow is sparged further in three similar flasks, each filled with 0.5 L of NaOH 1 M in order to trap the 14CO released from the mineralization of NP. The content 2 of the flasks was replaced before the NaOH solution was saturated with CO2. For detection of the saturation limit, 6132
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 19, 2006
phenolphthalein was added to the NaOH solution. The outlet of these four recipients was connected to a vacuum pump working under a slightly reduced pressure (-0.1 to -0.2 bar). This pump allowed counterbalancing of the positive pressure created inside the MBR by the air supply. The MBR was operated at a solid retention time (SRT) of 25 d, a hydraulic retention time (HRT) of 12 h, and total suspended solids (TSS) were controlled at 10 g/L. During the test period the temperature in the system was within 22 to 27 °C and the concentration of dissolved oxygen was maintained at approximately 6 mg/L. MLSS present in the MBR had a clear, dark brown color due to the composition of the synthetic feed and revealed a homogenized biomass as well as dispersed bacteria, characteristic of MBR sludge (14). The efficiency of the MBR in treating the synthetic wastewater was monitored over a period of 120 d. The synthetic composition was adapted from that of DIN ISO 11733 (DEVL41) (26), autoclaved at 120 °C during 20 min, and applied into the MBR by means of a peristaltic pump (MULTIFIX-Constant MC 1000 PEC, Alfred Schwinherr KG, Germany). The composition was 700 ( 20 mg/L chemical oxygen demand (COD), 40 ( 5 mg/L total nitrogen (TN), and 5 ( 0.5 mg/L total phosphorus (TP). TSS in the activated sludge, flow rates, pH value and temperature were monitored daily in the MBR and excess sludge was removed according to sludge retention time. COD, TP, and TN were measured for each influent charge, replaced every 5 days (see Table SI 1 and “Sampling and Preparation of the Samples” in the Supporting Information). During the radioactive test period all quality parameters of effluent and the dry matter inside of MBR were measured regularly as in the adaptation period. Radiotracer Preparation. A radiolabeled single isomer of NP, i.e., 4-[1-ethyl-1,3 dimethylpentyl]phenol (isomer number 111) (5, 27) was used as radioactive model compound. 4-[1-ethyl-1,3-dimethylpentyl]phenol isomer was synthesized by means of a Friedel-Crafts alkylation (28). The synthesized product was characterized by means of highperformance liquid chromatography coupled to a diode array detector and a liquid scintillation counting detector (HPLCDAD-LSC) and gas chromatography-mass spectrometry (GC-MS). Compound characteristics are given in Table 1. Analyses. Liquid Scintillation Counting (LSC). After sampling, 5 mL of scintillation cocktail was added to each sample, and vials were analyzed by means of a LS 5000 TD liquid scintillation counter (LSC) (Beckman, Germany). The detailed description of the sampling and preparation of the samples is supplied in the Supporting Information.
TABLE 1. Test Compound Characteristics
name: 4-[1-ethyl-1,3-dimethylpentyl]phenol formula: C15H24O molecular weight: 220.36 radioactivity stock solution: 14 MBq/mL specific radioactivity: 1492 MBq/mg concentration: 9.45 mg NP /mL CH2Cl2 radioactive purity: 94.4% chemical purity: 94.7% applied radioactivity: 4 MBq/L
Determination of Non-Extractible Radioactivity. In order to analyze the non-extractible fraction, a 0.1 g sample of predried pellets was packed in cellulose and combusted in a biological oxidizer OX500 (RJ Harvey Instrument Corporation). The resulting 14CO2 was trapped in a scintillation vial containing Carbomax Plus LSC cocktail (Canberra-Packard), and was subjected to LSC. HPLC Analysis. The HPLC analyses of radioactive samples were performed with an Agilent HPLC-System, equipped with a radioisotope detector (Raytest; Germany) with a 3139 quartz cell. A 250/4 Nucleosil 100-5 C18HD column was used (Macherey Nagel) at 35 °C with a flow rate of 1 mL min-1 and the mobile phase consisted of water (A) and acetonitrile (B). Samples were separated using a gradient program as follows: 25% B in A with linear gradient to 90% B during 22 min, after 90% B isocratic for 5 min. Finally, the system returned to its initial conditions (25% B in A) within 10 min, and was kept in this composition for 3 min before the next run started. GS-MS Analysis. The GC-MS studies were carried out as described previously (28). HPLC-MS/MS Analysis. The HPLC analyses of radioactive samples were performed with a Thermo Finnigan surveyor LC-Pump. Compounds were separated on a Synergi 4µ Fusion RP 80A column (150 × 4.6 mm, 3 µm particle size) from Phenomenex (Torrance, CA). The mobile phase consisted of water (A) and acetonitrile (B). Samples were separated using a gradient program as follows: 20% B in A with linear gradient to 30% B during 3 min and to 100% B during another 18 min. After 100% B isocratic for 6 min, the system returned to its initial conditions (20% B in A) within 1 min, and was kept in this composition for 8 min before the next run was started. After splitting of the solvent flow (flow rate of 800 µL min-1) compounds were detected simultaneously via a radioisotope detector (Raytest; Germany) with a 3139 quartz cell and a TSQ quantum ultra AM tandem mass spectrometer (Thermo Finnigan, USA). Full scan mass spectra were obtained using the TSQ quantum equipped with an APCI ion source (Ion Max) operating in negative mode in the m/z range of 92 to 500. Nitrogen was employed as both the drying and nebulizer gas. Product ion spectra were obtained after fragmentation at collision energies of 10, 20, 30, 44, and 55 V.
Results and Discussion Performances in MBR. The MBR system was designed and operated as close as possible to the parameters and performances of a full-scale MBR. The efficiency of the MBR during the whole operation is given by the values in Table SI 1 and Figure SI 1 presented in the Supporting Information. Although parameters remained almost constant after one week, the MBR was operated longer to check the robustness of the membranes and stability of the process. The daily monitoring also allowed registration of typical variations such
as higher nitrification, decrease of pH, or total elimination of COD during some particularly hot days of the summer period. At the end of this equilibration period, the MBR was spiked once with 14C-NP and these parameters were monitored for 34 d in parallel to the radioactivity measurements. The treated effluent achieved the required standard quality for the released effluent of small-scale STP and the removal rates were comparable to those of a pilot MBR (13). Radioactivity Monitoring and Balancing. Radiolabeled NP was applied as single pulse for the MBR at a concentration of 4 MBq/L corresponding to 2.3 mg/L 14C-4-[1-ethyl-1,3dimethylpentyl]phenol. The continuous application with NP dissolved in solvent was not feasible under the applied experimental conditions and design without applying too large an amount of solvent and disturbing the process. During the first day, the radioactivity inside the MBR dropped sharply to around 60% of the initial applied amount and then a slower decrease occurred until 1.8% at day 34 (Figure 2, left). The drastic decrease indicated a strong sorption of 14C-4-[1-ethyl-1,3-dimethylpentyl]phenol inside the MBR. This was corroborated with the low amount of radioactivity found in the mineralization fraction (< 1%), which increased slowly during the first half of the experiment and remained almost constant in the last days, thus precluding the rapid mineralization of NP. As only one pulse of NP was applied to the reactor and because the biomass was grown on a synthetic influent decreasing the chance of microorganisms to adapt to NP, the biodegradation potential might have been underestimated. In the monoethylenglycol fraction containing VOC, a very low level of radioactivity (
