Environ. Sci. Technol. 2006, 40, 7193-7199
Discharge of Three Benzotriazole Corrosion Inhibitors with Municipal Wastewater and Improvements by Membrane Bioreactor Treatment and Ozonation† STEFAN WEISS, JUTTA JAKOBS, AND THORSTEN REEMTSMA* Department of Water Quality Control, Technical University of Berlin, Sekr KF 4, Strasse des 17 Juni 135, 10623 Berlin, Germany
A set of three benzotriazole corrosion inhibitors was analyzed by liquid chromatography-mass spectrometry in wastewaters and in a partially closed water cycle in the Berlin region. Benzotriazole (BTri) and two isomers of tolyltriazole (TTri) were determined in untreated municipal wastewater with mean dissolved concentrations of 12 µg/L (BTri), 2.1 µg/L (4-TTri), and 1.3 µg/L (5-TTri). Removal in conventional activated sludge (CAS) municipal wastewater treatment ranged from 37% for BTri to insignificant removal for 4-TTri. In laboratory batch tests 5-TTri was mineralized completely and 4-TTri was mineralized to only 25%. This different behavior of the three benzotriazoles was confirmed by following the triazoles through a partially closed water cycle, into bank filtrate used for drinking water production, where BTri (0.1 µg/L) and 4-TTri (0.03 µg/ L) but no 5-TTri were detected after a travel time of several months. The environmental half-life appears to increase from 5-TTri over BTri to 4-TTri. Treatment of municipal wastewater by a lab-scale membrane bioreactor (MBR) instead of CAS improved the removal of BTri and 5-TTri but could not avoid their discharge. Almost complete removal was achieved by ozonation of the treatment plant effluent with 1 mg O3/mg DOC.
Introduction Benzo-1,2,3-triazoles are a class of high production volume chemicals (HPVC), which are used as corrosion inhibitors in various industrial processes and in households. 1H-benzotriazole (BTri) and tolyltriazole (TTri, available as technical mixture of the 4- and 5-isomer) form a thin complexing film on metallic surfaces that protects the underlying metal from corrosion. Benzotriazoles are added to many formulations that come in contact with metals, such as aircraft de-icing fluids (ADF), automotive antifreeze formulations, brake fluids, metal-cutting fluids, and dishwashing agents. The annual production of benzotriazoles was reported to be in the range of 9000 tons/year worldwide (confidential source). With household wastewater, via indirect discharge from industry, or by surface runoff collected in combined sewer systems benzotriazoles reach municipal wastewater treatment plants. †
This article is part of the Emerging Contaminants Special Issue. * Corresponding author phone: +49-30-31426429; fax: +49-3031423850; e-mail:
[email protected]. 10.1021/es061434i CCC: $33.50 Published on Web 11/03/2006
2006 American Chemical Society
Benzotriazoles are weakly basic (pKa 8.2 - 8.8) compounds (1) of high polarity and with an only moderate tendency to partition into an organic phase (log Kow 1.23 for BTri and 1.89 for TTri (2)) and may, thus, be very mobile in the aquatic environment. Contrary to N-substituted 1,2,4-triazoles the benzo-1,2,3-triazoles have only limited biological activity: for example acute toxicity of 5-TTri to aquatic organisms is in the low to moderate mg/L range (3, 4). Present knowledge on the occurrence of benzotriazoles in municipal wastewater and their behavior in wastewater treatment is fragmentary. One of the tolyltriazole isomers, 5-TTri, was included in monitoring studies on wastewater treatment plant (WWTP) effluents and surface waters in the United States and positive findings occurred in 45% of the effluents (up to 1.7 µg/L) (5) and 31% of the surface waters (median concentration 0.4 µg/L) (6). Both TTri-isomers have previously been detected in a creek that received surface runoff from an airport, where these compounds were employed as additives in de-icing fluids (3), and in groundwaters underneath an airport (7). In a monitoring study performed in several European countries BTri and TTri were determined in WWTP effluents (4.1 and 1.2 µg/L on average) and also in receiving waters (0.6 and 0.2 µg/L on average) (8). In that study a Water Cycle Spreading Index (WCSI) was proposed that uses the influent and effluent concentration and the extent of removal of a compound in WWTPs to assess its potential to spread in partially closed water cycles. WCSIs calculated for BTri and the sum of both TTri isomers were 21 and 24 µg/L, respectively, and were comparable to that of carbamazepine (23 µg/L) (8). It was recently shown that BTri and TTri are regularly found also in influents and effluents of Swiss WWTP (10 and 1.6 µg/L on average in effluents) (9). All these findings indicate that these benzotriazoles are incompletely removed in municipal wastewater treatment and may be distributed in surface waters. No long term study is available on the occurrence and extent of removal of benzotriazoles in WWTP or on the stability of benzotriazoles after discharge into receiving waters and the potential spreading of these compounds in partially closed water cycles. Moreover, previous studies based on the use of liquid chromatography-mass spectrometry (LC-MS) (8, 9) did not differentiate between the two isomers of TTri. We have recently developed an analytical method for the determination of these three benzotriazoles (BTri, 5-TTri, 4-TTri) using LC-MS, either by direct injection of aqueous samples, or after enrichment by solid-phase extraction, with LOQs in the low ng/L range (10). During method development first indications were received that the two TTri isomers may exhibit different biodegradability (10). Based on these previous studies (8, 10) a long term study was performed on the occurrence and removal of benzotriazoles in a large WWTP. Two approaches to improve municipal wastewater treatment, the use of membrane bioreactors and ozonation, are investigated for their potential to avoid discharges of benzotriazole corrosion inhibitors. Considering the WCSI values of the triazoles and the partially closed water cycle of Berlin, with a high portion of indirect potable reuse of treated municipal wastewater (11), first indications for the fate of benzotriazoles after discharge into receiving waters and through bank filtration are obtained.
Materials and Methods Sampling. Conventional Activated Sludge Treatment (CAS). Samples (24 h composite samples, sampled hourly) were collected weekly from the effluent of the primary and the VOL. 40, NO. 23, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7193
tertiary treatment of a municipal WWTP in Berlin (on Tuesdays, Wednesdays, or Thursdays) between June 2004 and September 2005. The WWTP receives mainly domestic wastewater with an industrial input of about 30%. The plant has a capacity of about 240 000 m3/day at dry weather flow. The activated sludge treatment includes nitrification, denitrification, and enhanced biological phosphorus removal, with an average sludge retention time (SRT) of 15 d and an average sludge concentration (MLSS) of 5 g/L. Sampling of the influent to the biological stage (primary treated wastewater) and its effluent was corrected for the hydraulic retention time (18 h). Samples were filtered over 0.45 µm membrane filters (Sartorius, Goettingen, Germany) and stored frozen at -18 °C until analysis. Membrane Bioreactor. A lab-scale membrane bioreactor (MBR) of 21 L active volume and equipped with a submerged Kubota plate membrane system (nominal pore size 0.4 µm; total surface 0.3 m2) was operated at the same WWTP as mentioned above since October 2003. It was continuously fed with the effluent of the primary treatment. Particles >1 mm were removed by a screen before the effluent entered the MBR. The hydraulic retention time (HRT) ranged from 7 to 14 h, the SRT was from 26 to 102 d, the MLSS was from 15 to 35 g/L, and the membrane flux was between 9.2 and 13.1 L/m2‚h. A total of 56 samplings was performed between June 2004 and September 2005 in parallel to the CAS treatment. Sampling (24 h composite samples) was delayed according to the HRT of the MBR. Samples were stored frozen at -18 °C until analysis. Commercial Products. A technical mixture of tolyltriazoles and several formulations containing benzotriazoles such as an aircraft deicing fluid (ADF Type IV), antifreeze for automotive cooling systems and several dishwashing agents were also analyzed. Biodegradation Study. The biodegradation tests were carried out based on ISO 7827 (12) in 2 L glass bottles containing 1 L of deionized water with a phosphate buffer. The individual test substances (5-TTri and technical mixture of 4- and 5-TTri) were used as sole substrates (47.8 mg) at a dissolved organic carbon (DOC) concentration of 30 mg/L. Approximately 200 mg/L of fresh sludge from the MBR (see above) was added as inoculum. One control batch with aniline as carbon source (DOC 30 mg/L) and an abiotic control poisoned with mercury chloride were operated in parallel. Surface Water. Seven surface water grab samples were collected with a bucket on September 22, 2005. An overview of sample locations is given in Figure S1 in the Supporting Information. Samples were filtered over 0.45 µm membrane filters and analyzed within a few hours by direct injection into the LC-MS system. Bank Filtration. Samples were taken at one day in July 2004 along a transect at a bank filtration site at the south shores of Lake Tegeler See (Figure S1). The site has been described in detail elsewhere (11). In this case 400 mL of each sample was extracted using LiChrolut EN (Merck, Darmstadt, Germany) cartridges at pH 2.7, followed by elution with methanol. Recovery rates of benzotriazoles with this extraction method were independently determined (BTri 73%, 5-TTri 75%, 4-TTri 95%) and the detected concentrations were corrected for the recovery rates. Ozonation. The effluent of the same WWTP (average pH 7.4) was treated with a pilot-scale ozonation plant as described elsewhere (13), using the same installation as in Huber et al. (14). Two experiments with increasing ozone dose were performed on March 1, 2005 (E1) and July 6, 2005 (E2). Samples were filtered over 0.45 µm membrane filters and stored frozen until analysis. Analysis. Benzotriazoles. Only dissolved concentrations (0.45 µm filtred samples) were considered in this study. From these samples the three benzotriazoles were analyzed by 7194
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 23, 2006
liquid chromatography-electrospray ionization-tandem mass spectrometry (LC-ESI-MS/MS) using multiple reaction monitoring (MRM) in the positive ion mode as described in detail elsewhere (10). A few ozonated WWTP effluents were analyzed after enrichment by solid-phase extraction (SPE), while all other samples were analyzed by direct injection (10). Samples from bank filtration were extracted as described above. Analytes were routinely separated in an isocratic run to enable separation of the two TTri isomers. Until December 2004, CAS and MBR samples and ozonation samples were analyzed without separation of the isomers. Quantitation was performed by external sample calibration. Limit of quantification (LOQ) was 0.1 µg/L for direct injection and 10 ng/L after SPE extraction. For SPE extraction of groundwater samples the LOQ was approximately 5 ng/L. Dissolved organic carbon (DOC) was determined by a high TOC analyzer (Elementar, Hanau, Germany). Data evaluation. Quantitative data of CAS and MBR treatment were statistically evaluated using SPSS 10 software (SPSS, Chicago, Il). The significance of differences in concentration between influents, CAS effluent, and MBR effluent was checked by the students T-test (pairwise).
Results and Discussion Occurrence in Municipal Wastewater. The occurrence of dissolved BTri, 5-TTri, and 4-TTri in a municipal WWTP was studied over a period of 66 weeks with almost weekly sampling. All three benzotriazoles were permanently detected (Figure 1) with BTri showing highest concentrations of 4-22 µg/L and a mean concentration of 12 µg/L (Figure 1 and Table 1) in the influent to the biological stage of the WWTP (called “influent” throughout this work). 4-TTri is the second compound in the influents with a mean concentration of 2.1 µg/L, followed by 5-TTri with 1.3 µg/L on average. The isomeric ratio between 5-TTri/4-TTri is 0.62 on average. This is significantly lower than in a technical TTri mixture and formulations containing TTri, where a 5-TTri/4-TTri ratio of 1.2-1.3 is found. Preferential removal of the 5-isomer by biodegradation already in the sewer system or during primary treatment may be the reason for this shift in the isomeric ratio. In a parallel study from Switzerland with an overview on 10 WWTPs 13-39 µg/L of BTri and 0.2-5.6 µg/L of total TTri were found (9). The relative temporal variability of the influent concentrations is comparable for the three benzotriazoles for most of the time (Figure 1). Seasonal influence is visible in concentrations as well as on a mass flux basis: the daily flux of BTri and TTri into the WWTP was significantly higher during the period when frost occurred (3000 and 800 g/d of BTri and TTri) as compared to the summer season (1900 and 450 g/d) (see Table S1). Moreover, in February 2005, a few samples showed a 5- to 10-fold increase of both TTri isomers up to 7 µg/L (Figure 1) (summing up to a total TTri flux of 3600 g/d). Another peak flux of TTri of 19 µg/L (3700 g/d) was identified in early March in the WWTP effluent during the ozonation experiments (see below). Elevated immissions of benzotriazoles into the sewer system during the winter period point at their use as corrosion inhibitors in deicing fluids. An aircraft deicing fluid (ADF) analyzed in this study contained 0.14 g/L 5-TTri, 0.11 g/L 4-TTri, and 0.11 g/L of BTri. This concentration is 1 order of magnitude lower than those previously reported for other ADF products (5-10 g/L of TTri) (1, 2, 15). Similar concentrations of TTri (0.1-0.2 g/L for each isomer) were determined in this study in antifreeze for automotive cooling systems. A part of the increase in benzotriazole concentrations during winter season may be attributable to the use of this ADF at an airport that discharged its wastewater into this sewer system. Even the sporadic peaks in TTri concentration (Figure 1) and flux could orginate from the use of ADF: the
FIGURE 1. Concentrations of benzotriazoles in a primary treated municipal wastewater (“influent”) and minimum daily air temperature in Berlin during that period. Until December 2004 the tolyltriazole isomers were not separately detected.
TABLE 1. Mean Concentrations (and Relative Standard Deviations) of Benzotriazoles in Primary Treated Municipal Wastewater (influent), the WWTP Effluent (CAS), and the Effluent of an MBR (Sampling Period BTri, May 2004-Sep 2005; TTri Isomers, Dec 2004-Sep 2005) CAS effluent
compound
influent µg/L (SD)c
BTri (n ) 39) 5-TTri (n ) 25) 4-TTri (n ) 25)
12.0 (( 3.7) 1.3 (( 1.7) 2.1 (( 1.4)
MBR effluent
µg/L (SD)c
% removal (SD)c
µg/L (SD)c
% removal (SD)c
∆CAS/MBRa µg/L
significanceb
7.7 (( 2.7) 1.2 (( 1.7) 2.2 (( 1.7)
37 (( 17) 11 (( 25) -6 (( 25)
4.6 (( 1.8) 0.5 (( 0.9) 1.7 (( 1.3)
61 (( 12) 61 (( 26) 14 (( 20)
3.1 0.7 0.5
0.000 0.003 0.015
a Difference in mean effluent concentration of CAS and MBR. at the 95% level (T-test, pairwise). c SD: standard deviation.
b
Statistical significance of the difference of CAS and MBR effluent concentrations
increase of about 3000 g/d in total TTri flux compares well with the average use of 3.3 kg/d of TTri during these days, as estimate based on monthly ADF application data from the airport and the analysis of its TTri content. Cleaning agents for dishwashing machines are a source of continuous benzotriazole immission. In two products with declared silver protection properties either BTri (0.1 g/L) or TTri (0.5 and 0.4 g/L of 5- and 4-isomer) was found. Indeed, BTri concentrations of 5-25 µg/L in the wastewater of a residential area were recently reported and it was shown that modeled and measured BTri concentrations agreed well assuming that all BTri originated from dishwashing agents (16). In larger sewer systems such as the one in this study, other sources such as industrial discharge (including airports) and surface runoff (combined sewer system) are also relevant. Removal in Biological Wastewater Treatment. In conventional activated sludge (CAS) wastewater treatment a significant removal of 37% on average is observed over 66 weeks for BTri (Figure 2b; Table 1), resulting in an average effluent concentration of 7.7 µg/L. The extent of BTri removal varies strongly between 5% and 60% and is never stable. This is reflected in the highly variable effluent concentrations (Figure 2a). Of the two tolyltriazoles 4-TTri shows no removal and is the dominant isomer in the effluent with a mean concentration of 2.2 µg/L and the 5-TTri decrease (11% to 1.2 µg/L on average) is statistically insignificant (Table 1). The 5-TTri/4-TTri ratio of 0.54 is only slightly lower than that in the influent (0.62). The lower removal of the slightly
more hydrophobic tolyltriazoles as compared to BTri suggests that the removal of BTri is due to biodegradation rather than to sorption onto particulate matter in the CAS treatment. In a recent Swiss study 13-60% removal for BTri and 23-74% removal for total TTri in isolated samples of 10 WWTP were found (9). Namely the significant removal of TTri is remarkable. Since no differentiation between the isomers was made in that study and the differences in wastewater quality and treatment as compared to this study are not known, these different findings cannot be explained, yet. The long term average concentrations of the benzotriazoles in the effluent of this WWTP are comparable to the average of isolated samples from European (4.1 µg/L of BTri, 1.1 µg/L of total TTri) (8) and Swiss WWTP effluents (10 and 1.2 µg/L) (9). For BTri the WCSI calculated from the mean concentration data (Table 1) confirms the previous calculation (22 µg/L) (8). Owing to the separate determination of 5-TTri and 4-TTri in this study the WCSI previously calculated for the sum of both isomers (24 µg/L (8)) can now be split into the WCSI for 5-TTri (16 µg/L) and for 4-TTri (46 µg/L). These new data suggest that 4-TTri is the most relevant benzotriazole in terms of effluent concentration and stability, followed by BTri and 5-TTri. Membrane Bioreactor. A lab-scale membrane bioreactor (MBR), operated parallel to the full-scale CAS treatment using the same influent, was investigated for its potential to improve the removal of benzotriazoles. As compared to the CAS treatment, MBR can be operated at a higher SRT, which could VOL. 40, NO. 23, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7195
FIGURE 2. (a) Concentrations of BTri in the effluent of the WWTP (CAS) and of a lab-scale membrane bioreactor (MBR) operated parallel to CAS and temperatures of operation for CAS and MBR. No temperature data were available for CAS beyond day 600. (b) Statistical evaluation of concentrations of BTri, 5-TTri, and 4-TTri in WWTP influent (in), WWTP effluent (CAS), and MBR effluent (MBR). Data for BTri cover the whole period shown in Figure 1, data for TTri isomers cover the period since December 2004. Boxes, middle 50% of the data; whiskers, extreme values. favor the development of microorganisms adapted to certain poorly degradable compounds, and at lower sludge loads (F/M ratio) which could support the mineralization of poorly degradable compounds. The potential of MBR to remove poorly degradable trace pollutants more effectively than CAS is still under debate and not clearly proven. In MBR treatment the BTri concentration is reduced by 61% on average to a mean effluent concentration of 4.6 µg/L (Figure 2b and Table 1). This removal is significantly better than that in CAS, where only 37% was removed. Removal in MBR is also more stable (Table 1), resulting in a less variable effluent concentration as compared to CAS (Figure 2a and b). Also 5-TTri is significantly better removed in MBR (61%) as compared to CAS treatment to a mean effluent concentration of 0.5 µg/L. For 4-TTri a slight removal appears in the MBR (Figure 2b) and also the mean concentration was diminished by 14% to 1.7 µg/L (Table 1), but this decrease is not statistically significant. Thus, the 5-TTri/4-TTri ratio in the MBR effluent is 0.29 on average, compared to 0.63 in the influent and 0.54 in the CAS effluent. This difference in the removal of 5-TTri and 4-TTri is, likely, due to a strong influence of the position of the methyl-group (Figure 1) on the oxidative biodegradation by dioxygenases. Although this influence cannot yet be explained mechanistically, it was recently outlined that similar inhibitory effects by substitution in R-position to the angular carbon of two ring aromatic systems have been found for other polar aromatic compounds like naphthalene sulfonates and sulfophthalimides (10). Though visible throughout the whole investigation period, the advantage of MBR in BTri removal is particularly obvious during the winter season (Figure 2a, middle), while it is not statistically significant during the summer periods (left and 7196
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 23, 2006
FIGURE 3. Concentrations of tolyltriazoles and dissolved organic carbon (DOC) in batch degradation studies: (a) pure 5-TTri, (b) technical mix of 5-TTri and 4-TTri. right sections of Figure 2a, Table S2). There are two reasons for these seasonal differences. First, the MBR is able to cope with elevated influent concentrations during winter time (Figure 1) without responding with elevated effluent concentrations. Second, CAS treatment shows a trend toward reduced BTri elimination during the winter period (35% average removal) as compared to the summer period (42%). This may be due to the temperature of the biological treatment, which differs between 22 °C in summer and 1316 °C in winter (Figure 2a). It is reasonable to assume that microbial degradation of BTri in CAS responds more sensitively to this temperature decrease than the more stable MBR process. For the TTri-isomers such a temperature effect is not visible, because there is no removal in CAS at any time. An improved removal of two compounds of intermediate biodegradability (5-TTri, BTri) by MBR as compared to CAS agrees with previous reports concerning bisphenol A (17) or benzothiazoles, another class of industrial chemicals (18), whereas other studies did not observe such an improvement (19, 20). It must be noted that a lab-scale MBR was used in this study. Removal of benzotriazoles by MBR may differ in the case of a full-scale facility. Despite this improvement a total mean concentration of 7 µg/L of these benzotriazoles would be discharged into receiving waters. With this concentration level the benzotriazoles are comparatively prominent polar trace pollutants in WWTP effluents (8). Biodegradation of Tolyltriazoles. The striking difference in the removal of 5-TTri and 4-TTri in MBR treatment could be confirmed by lab-scale biodegradation tests using activated sludge as inoculum (Figure 3). Here, biodegradation of pure 5-TTri starts before day 10 and 5-TTri is completely removed within 17 days (Figure 3a). The DOC data confirm a complete mineralization of 5-TTri, with a final difference between the calculated and the measured DOC of 2-3 mg/ L. Such a final DOC was also found in the aniline batch and indicates dissolved organic matter of bacterial origin. From
day 10 to 17 the difference between measured and calculated DOC is larger (Figure 3a), suggesting intermediate formation of an unknown metabolite that is finally mineralized. In the technical TTri-mixture containing 5-TTri and 4-TTri similar results are obtained for 5-TTri (Figure 3b). For 4-TTri a slight decrease starting during the phase of rapid 5-TTri degradation (day 10-17) is observed. After 28 days about 25% of the 4-TTri is removed but degradation did not continue, as visible from a sample taken after 42 days (not shown in Figure 3b). The DOC balance is in full agreement with these LC-MS results (Figure 3b). This suggests that 4-TTri is only slowly biodegradable under suitable conditions. In this batch test such conditions are found during the rapid mineralization of 5-TTri while in the environment a very active biomass or certain cosubstrates may be required. No transformation of any of the substrates and no DOC removal was observed in the poisoned control experiments confirming that all observed effects were due to microbial activity. In previous reports on groundwater contamination beneath two airports (7, 21) 4-TTri was detected in all contaminated wells, while 5-TTri was not always present. Considering that all technical formulations yet investigated contain both TTri-isomers this was a strong indication that 5-TTri has been preferentially removed by biological means (7). Surface Waters and Bank Filtration. Due to the limited (BTri) or insignificant (5-TTri, 4-TTri) removal of benzotriazoles in WWTP with CAS it was obvious that receiving waters would be polluted by these corrosion inhibitors. This has only recently been shown for different European rivers that are influenced by WWTP discharges, where mean concentrations of 0.6 µg/L of BTri and 0.2 µg/L of TTri were detected (8). Moreover, the WCSI calculation (4-TTri > BTri > 5-TTri) suggested that some of the benzotriazoles may pass other barriers in a partially closed water cycle and eventually reach raw waters used for drinking water production. To verify these previous findings grab samples from different points of the rivers Spree and Havel, upstream and downstream of the city of Berlin and its WWTP discharges and from a bank filtration site (Figure S1), were investigated for the three benzotriazoles (Table S3). In surface waters upstream of the city no BTri or TTri above the LOQ of 0.1 µg/L was detected, but during passage of the city and downstream of a first WWTP concentrations of benzotriazoles in the river Spree increased (Figure S1, site 3). In a distance of 12 km downstream of another WWTP 2.7 µg/L of BTri, 0.2 µg/L of 5-TTri, and 1.2 µg/L of 4-TTri were found (Figure 4a, site 6), showing clearly that WWTPs are significant sources of immission of benzotriazoles in surface water. However, the three benzotriazoles behave differently in this system. Compared to the WWTP effluent BTri and 5-TTri concentrations decrease more strongly after discharge as compared to 4-TTri, as indicated by decreasing 5-TTri/4-TTri and BTri/ 4-TTri ratios (Figure 4a). This behavior corresponds to the observations made in wastewater treatment and in the laboratory degradation tests (see above). Different points of a partially closed water cycle in the North of Berlin were also investigated (Figure S1), including a trench (site 4) that directs a WWTP effluent to Lake Tegeler See, the lake water itself (site 5), four monitoring wells along a transect used for bank filtration (11), and the production well where bank filtrate arrives after a travel distance of 90 m and a residence time of approximately 4.5 months in the subsurface (full data in Table S3). The mean (n ) 3) concentrations found in the lake (3 µg/L Tri, 0.1 µg/L 5-TTri, and 0.3 µg/L 4-TTri gradually decrease during subsurface travel, partly by dilution (approximately 30%) and presumably also by biodegradation (Table S3, Figure 4b). BTri (0.11 µg/ L) and 4-TTri (0.03 µg/L) were still measurable in the
FIGURE 4. Changes in the benzotriazole composition (BTri/4-TTri ratio and 5-TTri/4-TTri ratio) versus the total benzotriazole concentration: (a) from WWTP effluent (data of Table 1) to River Havel 12 km (6) and 18 km (7) downstream; (b) in a partially closed water cycle (trench 4 ()WWTP effluent), Lake Tegeler See (5) and through a bank filtration transect to a production well (4.5 months underground residence time). Site map is provided in Figure S1; complete data are provided in Table S3. production well. Again, concentrations of BTri and 5-TTri are more effectively diminished in this partially closed cycle than 4-TTri. The BTri/4-TTri and the 5-TTri/4-TTri ratios strongly decrease with decreasing total benzotriazole concentration from the trench toward the production well (Figure 4b), where no 5-TTri could be determined at all. It is noteworthy that the concentration profile found in this study for BTri and 4-TTri agrees with that detected previously for carbamazepine at the same bank filtration site (11). This is in line with the prediction made by calculating the WCSI (21 µg/L for BTri, 23 µg/L for carbamazepine) (8). Though these are only the first data on the occurrence of the three benzotriazoles in environmental samples they already show that 4-TTri is the most stable of the three compounds, whereas 5-TTri is the least stable. A proper investigation of the environmental behavior of tolyltriazoles, thus, requires the separation of the two isomers (10). Ozonation. Ozonation is increasingly discussed as an option for improved disinfection and removal of polar trace pollutants from secondary or tertiary treated municipal wastewater before discharge into surface waters as it can remove a wide variety of trace organic compounds (13, 14, 22, 23). For benzotriazoles with their electron rich aromatic system ozonation could also be a viable treatment option. This was investigated in two experiments using a pilot-scale ozonation reactor and real WWTP effluent. Ozonation proved to be very effective in removing the benzotriazoles (Figure 5). With a specific ozone consumption of 0.8 mg O3/mg DOC0 VOL. 40, NO. 23, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7197
Supporting Information Available Additional information as noted in the text: map of sampling locations in the Berlin region (Figure S1), fluxes of benzotriazoles into the WWTP during summer and winter periods (Table S1), BTri effluent concentrations of CAS and MBR (Table S2), and concentrations of benzotriazoles in water samples of the Berlin region (Table S3). This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited
FIGURE 5. Relative concentrations (c/c0) of BTri (black columns and total TTri (white columns) versus the ozone consumption in two ozonation experiments of WWTP effluent. c0 of experiment 1: BTri 16.4 µg/L, TTri (sum of both isomers) 18.6 µg/L, DOC ) 10.0 mg/L. c0 of experiment 2: BTri 6.0 µg/L, TTri 1.8 µg/L, DOC 11.5 mg/L.
a removal of more than 90% is achieved. At 1 mg O3/mg DOC0 all three benzotriazoles are removed to >99% and only BTri can be detected in the range of its LOQ (10 ng/L). No difference in the reactivity of the three benzotriazoles toward ozone was visible in these experiments. The ozone dosage required to remove the benzotriazoles was also adequate to remove acidic and neutral pharmaceuticals from the same wastewater (13). A corresponding dosage of 10 mg O3/L was previously recommended for the removal of pharmaceuticals and personal care products (22), whereas only 2 mg/L was required for a different WWTP effluent (14). These results further imply that ozonation may also be useful to remove benzotriazoles in case they occur in raw waters used for drinking water production. Possibilities to Reduce Pollution by Benzotriazoles. The recently detected benzotriazoles are polar high-production volume chemicals that are incompletely removed in WWTPs and released with their effluents into receiving waters in µg/L concentrations. From there BTri and, namely, 4-TTri may travel along partially closed water cycles to raw waters used for drinking water production. An intensified biological wastewater treatment as in an MBR improves the removal of benzotriazoles only gradually. On the basis of these investigations there are two options to avoid the spread of benzotriazoles in the aquatic environment: (a) WWTP effluents can be ozonated to ensure a complete removal of benzotriazoles, along with other poorly degradable polar pollutants. (b) Alternatively, solely the isomerically pure 5-TTri could be used as corrosion inhibitor, as this compound turned out to be least stable in aquatic environment. Beyond these measures, benzotriazoles may be replaced by biodegradable corrosion inhibitors. At least for the use in aircraft deicing fluids formulations are commercially available that contain corrosion inhibitors that have been claimed to be biodegradable.
Acknowledgments We gratefully acknowledge financial support by the European Union through the project “Removal of Persistent Polar Pollutants Through Improved Treatment of Wastewater Effluents” (P-THREE, EVK1-CT-2002-00116). We thank Peter Bauer and Renata Mehrez for skillful MBR operation, and Steffen Gru ¨ nheid, Carsten Bahr, and Alexandra Hu ¨ tteroth for providing samples. The Berliner Wasser-Betriebe provided continuous and valuable support throughout this project. 7198
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 23, 2006
(1) Gruden, C. L.; Dow, S. M.; Hernandez, M. T. Fate and toxicity of aircraft deicing fluid additives through anaerobic digestion. Water Environ. Res. 2001, 73, 72-79. (2) Hart, D. S.; Davis, L. C., Erickson, L. E.; Callender, T. M. Sorption and partitioning parameters of benzotriazole compounds. Microchem. J. 2004, 77, 9-17. (3) Cancilla, D. A.; Baird, J. C.; Geis, S. W.; Corsi, S. R. Studies of the environmental fate and effect of aircraft deicing fluids: Detection of 5-methyl-1H-benzotriazole in the fathead minnow (Pimephales promelas). Environ. Toxicol. Chem. 2003, 22, 134140. (4) Cornell, J. S.; Pillard, D. A.; Hernandez, M. T. Comparative measures of the toxicity of component chemicals in aircraft deicing fluid. Environ. Toxicol. Chem. 2000, 19, 1465-1472. (5) Glassmeyer, S. T.; Furlong, E. T.; Kolpin, D. W.; Cahill, J. D.; Zaugg, S. D.; Werner, S. L.; Meyer, M. T.; Kryak, D. D. Transport of chemical and microbial compounds from known wastewater discharges: Potential for use as indicators of human fecal contamination. Environ. Sci. Technol. 2005, 39, 5157-5169. (6) Kolpin, D. W.; Furlong, E. T.; Meyer, M. T.; Thurman, E. M.; Zaugg, S. D.; Barber, L. B.; Buxton, H. T. Pharmaceuticals, hormones, and other organic wastewater contaminants in US streams, 1999-2000: A national reconnaissance. Environ. Sci. Technol. 2002, 36, 1202-1211. (7) Cancilla, D. A., Baird, J. C.; Rosa, R. Detection of aircraft deicing additives in groundwater and soil samples from Fairchild Air Force Base, a small to moderate user of deicing fluids. Bull. Environ. Contam. Toxicol. 2003, 868-875. (8) Reemtsma, T.; Weiss, S.; Mueller, J.; Petrovic, M.; Gonzalez, S.; Ventura, F.; Knepper, T. Polar pollutants entry into the water cycle by municipal wastewater: a European perspective. Environ. Sci. Technol. 2006, 40 (17), 5451-5458. (9) Voutsa, D.; Hartmann, P.; Schaffner, C.; Giger, W. Benzotriazoles, Alkylphenols and Bisphenol A in Municipal Wastewaters and in the River Glatt, Switzerland. Environ. Sci. Pollut. Res. 2006, 13, 333-341. (10) Weiss, S; Reemtsma, T. Determination of Benzotriazole Corrosion Inhibitors from Aqueous Environmental Samples by Liquid Chromatography-Electrospray Ionization-Tandem Mass Spectrometry. Anal. Chem. 2005, 77, 7415-7420. (11) Gruenheid, S.; Amy, G.; Jekel, M. Removal of bulk dissolved organic carbon (DOC) and trace organic compounds by bank filtration and artificial recharge. Water Res. 2005, 39, 32193228. (12) ISO 7827. Evaluation in an Aqueous Medium of the Ultimate Aerobic Biodegradability of Organic CompoundssMethod by Analysis of Dissolved Organic Carbon; In German Standard Methods for the Determination of Water, Wastewater and Sludge (L 29); German Water Chemical Society, Ed.; Beuth Verlag: Berlin, 1995. (13) Bahr, C.; Ernst, M.; Reemtsma, T.; Heinzmann, B.; Luck, F.; Jekel, M. Pilot scale ozonation of treated municipal effluent for removal of pharmaceutical compounds and pathogens - The Berlin study. In IOA 17th World Ozone Congress, Strasbourg, 2005. (14) Huber, M. M.; Goebel, A.; Joss, A.; Hermann, N.; Loffler, D.; McArdell, C. S.; Ried, A.; Siegrist, H.; Ternes, T. A.; von Gunten, U. Oxidation of pharmaceuticals during ozonation of municipal wastewater effluents: A pilot study. Environ. Sci. Technol. 2005, 39, 4290-4299. (15) Pillard, D. A.; Cornell, J. S.; Dufresne, D. L.; Hernandez, M. T. Toxicity of benzotriazole and benzotriazole derivatives to three aquatic species. Water Res. 2001, 35, 557-560. (16) Ort, C.; Schaffner, C.; Giger, W.; Gujer, W. Modeling stochastic load variations in sewer systems. Water Sci. Technol. 2005, 52 (5), 113-122. (17) Wintgens, T.; Gallenkernper, M.; Melin, T. Removal of endocrine disrupting compounds with membrane processes in wastewater treatment and reuse. Water Sci. Technol. 2004, 50 (5), 1-8.
(18) Kloepfer, A.; Gnirss, R.; Jekel, M.; Reemtsma, T. Occurence of benzothiazoles in municipal wastewater and their fate in biological treatment. Water Sci. Technol. 2004, 50 (5), 203-208. (19) Joss, A.; Keller, E.; Alder, A. C.; Goebel, A.; McArdell, C. S.; Ternes, T.; Siegrist, H. Removal of pharmaceuticals and fragrances in biological wastewater treatment. Water Res. 2005, 39, 31393152. (20) Clara, M.; Strenn, B.; Ausserleitner, M.; Kreuzinger, N. Comparison of the behaviour of selected micropollutants in a membrane bioreactor and a conventional wastewater treatment plant. Water Sci. Technol. 2004, 50 (5), 29-36. (21) Cancilla, D. A.; Martinez, J.; Van Aggelen, G. C. Detection of aircraft deicing/antiicing fluid additives in a perched water monitoring well at an international airport. Environ. Sci. Technol. 1998, 32, 3834-3835.
(22) Ternes, T. A.; Stuber, J.; Herrmann, N.; McDowell, D.; Ried, A.; Kampmann, M.; Teiser, B. Ozonation: a tool for removal of pharmaceuticals, contrast media and musk fragrances from wastewater? Water Res. 2003, 37, 1976-1982. (23) Westerhoff, P.; Yoon, Y.; Snyder, S.; Wert, E. Fate of endocrinedisruptor, pharmaceutical, and personal care product chemicals during simulated drinking water treatment processes. Environ. Sci. Technol. 2005, 39, 6649-6663.
Received for review June 15, 2006. Revised manuscript received August 31, 2006. Accepted September 25, 2006. ES061434I
VOL. 40, NO. 23, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7199
