Environ. Sci. Technol. 2008, 42, 7702–7708
Removal of Bromate from Drinking Water Using the Ion Exchange Membrane Bioreactor Concept C R I S T I N A T . M A T O S , †,‡ SVETLOZAR VELIZAROV,† MARIA A. M. REIS,† AND JOA ˜ O G. CRESPO* CQFB/REQUIMTE, Department of Chemistry, FCT, Universidade Nova de Lisboa, P-2829-516 Caparica, Portugal, ´ and IBET, Instituto de Biologia Experimental e Tecnologica, P-2781-901 Oeiras, Portugal
Received April 29, 2008. Revised manuscript received August 2, 2008. Accepted August 8, 2008.
Bromate is a disinfection byproduct with carcinogenic properties that has to be removed from drinking water to concentrations below 10 or 25 µg/L. This work evaluates the applicability of the ion exchange membrane bioreactor (IEMB) concept for the removal of bromate from drinking water, in situations where nitrate is also present in concentrations up to 3 orders of magnitude higher than bromate. The batch results obtained show that the biological reduction of bromate was slow and only occurring after the complete reduction of nitrate. The specific bromate reduction rates varied from 0.027 ( 0.01 mg BrO3 /gcell dry weight · h to 0.090 mg BrO3 / gcell dry weight · h for the studied concentrations. On the other hand, transport studies, using anion exchange membranes showed that Donnan dialysis could efficiently remove bromate from polluted waters. Therefore, the use of a dense, nonporous membrane in the IEMB system, isolates the water stream from the biological compartment, allowing for the uncoupling of the water production rate from the biological reduction rate. The IEMB system was used for the treatment of a polluted water stream containing 200 µg/L of BrO3 and 60 mg/L of NO3 . The concentrations of both ions in the treated water were reduced below the recommended levels. No bromate accumulation was observed in the biocompartment of the IEMB, suggesting its complete reduction in the biofilm formed on the membrane surface contacting the biocompartment. Therefore, the IEMB has proven to be a technology able to solve specific problems associated with the removal of bromate from water streams, since it efficiently removes bromate from drinking water even in the presence of nitrate, a known competitor of bromate biological reduction, without secondary contamination of the treated water by cells or excess of carbon source.
1. Introduction The presence of bromate in drinking water is a concern for public health since it is considered a genotoxic carcinogen (1). A maximum allowed contaminant level of 10 µg/L was * Corresponding author phone: +351 212 948 385; fax: +351 212 948 385; e-mail:
[email protected]. † Universidade Nova de Lisboa. ‡ Instituto de Biologia Experimental e Tecnolo´gica. 7702
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 20, 2008
imposed for bromate by the European Union (2) and the U.S. Environmental Protection Agency (3), while the World Health Organization set a provisional guideline value of 25 µg/L (4). The occurrence of bromate in water results mainly from ozonation of waters that contain high levels of bromide. Ozonation is a disinfection method that destroys microorganisms, reduces color, odor, total organic carbon, and formation of trihalomethanes if compared with chlorination of water. During ozonation, bromide is oxidized to hypobromous acid and hypobromite, which is further oxidized to bromate by ozone and/or free hydroxyl radicals. The formation of bromate depends from the ozonation conditions (e.g., pH, ozone contact time, natural organic matter (NOM) concentration) (5-8). Concentrations in drinking water after ozonation generally range from 0.4 to 100 µg/L (7, 9). However, bromate levels exceeding 2 mg/L were detected in a United Kingdom aquifer, resulting from a chemical production plant spillage (7). Three different approaches can be used to reduce bromate concentration in water: removal of the bromate precursors, such as bromide and natural organic matter before ozonation (1, 10-12); control of bromate formation during ozonation by pH control in a low pH range, by addition of ammonia or hydrogen peroxide and modifications in the ozonation operation (1, 5, 13, 14); and removal of bromate after ozonation. Most of the research on bromate removal has been focused on the use of activated carbon (15-19). Membrane retention by nanofiltration (20), UV degradation (21), and biodegradation (22-26) have been also reported as possible ways for the removal of bromate. Studies on bioreduction of bromate showed its complete reduction to bromide, by nitrate and chlorate reducing microbial cultures, possibly via cometabolism of nitrate reductase and chlorate reductase enzymes (22-25). The reduction reactions of bromate, nitrate, and oxygen in the presence of ethanol as the electron donor can be presented as follows: 3C2H5OH + BrO3 a 3C2H4O + Br-+3H2O∆G° ) - 172.14 kJ ⁄ mol
(1)
+ 5C2H5OH + 2NO3 +2H a
5C2H4O + N2+6H2O∆G° ) -183.93 kJ ⁄ mol
(2)
2C2H5OH + O2 a 2C2H4O + 2H2O∆G° ) -196.48 kJ ⁄ mol (3) The Gibbs free energies associated with each reaction were calculated considering the standard redox potential of each half-reduction reaction (BrO3-/Br- E° ) 0.694V, NO3-/N2 E° ) 0.755V, 1/2O2/H2O E° ) 0.82V, C2H4O/C2H5OH E° ) -0.197V 7, 27). Bromate reduction is thermodynamically less favorable (it releases less energy) compared to nitrate or oxygen reduction. Preferential reduction of nitrate over bromate has been reported by several authors (9, 24-26). This apparent inhibition may be explained by thermodynamic or kinetically related mechanisms (9). Biological reduction of these oxyanions requires the addition of an electron donor and a carbon based donor is typically used in water treatment applications. Therefore, the elimination of bromate by technologies that rely only on biological reduction may induce secondary contamination of the treated water by cells, residual carbon source and/or metabolic byproduct. Since ozonation is usually one of the final drinking water treatment steps, it is important to develop 10.1021/es801176f CCC: $40.75
2008 American Chemical Society
Published on Web 09/10/2008
FIGURE 1. Schematic representation of the ion transport and bioreduction in the IEMB system. a bromate removal technology that can be applied without introducing further contamination to the treated water. The ion exchange membrane bioreactor (IEMB) is a suitable technology for such removal, because it prevents secondary contamination of the treated water by microbial cells, excess of carbon source, and metabolic byproduct (28). The IEMB concept (29), combines the transport of bromate through a dense anion exchange membrane from a polluted water stream to a biocompartment, where an anoxic microbial culture reduces bromate to bromide as illustrated in Figure 1, the transport of the pollutant counter-ions (anions) is governed by the Donnan equilibrium principle and, therefore, it is possible to enhance it by using a more concentrated driving counterion (chloride) added to the biocompartment. Using this approach, the water compartment is physically isolated from the biocompartment in which bioreduction takes place. The IEMB was extensively studied for denitrification of drinking water streams (29-33). These studies demonstrated that polluted waters containing up to 350 mg/L of nitrate were successfully treated at a maximum treated water production rate of 30 L/m2 · h (31, 32). More recent studies demonstrated that the IEMB is equally suitable for the simultaneous removal of nitrate and perchlorate (28, 34). The objective of this study is to evaluate the applicability of the IEMB system for the removal of bromate. Batch tests were performed aiming at studying the reduction of bromate and the effect of nitrate on its reduction. These tests were performed for a bromate/nitrate equimolar solution and under the concentrations usually found in drinking water polluted with the two ions (concentration of nitrate 3 orders of magnitude higher). The removal of bromate was also studied under Donnan dialysis conditions, in order to test the feasibility of transporting bromate through the selected anion exchange membrane. Finally, the capacity of the IEMB to treat water contaminated with bromate in the presence of higher concentration of nitrate was evaluated.
2. Materials and Methods 2.1. Culture Medium. The culture medium, adapted from ref 23, had the following composition: 0.5 g/L (NH4)2HPO4, 1.55 g/L K2HPO4, 0.98 g/L NaH2PO4 · H2O, 0.01 g/L MgSO4 · 7H2O, 0.0025 g/L Na2SeO3 · 5H2O, and 0.1 mL of a micronutrient solution with the following composition: 22 g/L ZnSO4 · 7H2O, 5.54 g/L CaCl2, 5.06 g/L MnCl2 · 4H2O, 4.99 g/L FeSO4 · 7H2O, 1.10 g/L (NH4)2MO7O24 · 4H2O, 1.57 g/L CuSO4 · 5H2O, and 1.61 g/L CoCl2 · 6H2O. Ethanol was added
as the sole carbon source at a concentration of 560 mg/L in the IEMB studies and of 1 g/L in the batch stirred bioreactor studies. 2.2. Contaminated Water. The feedwater (pH 7.4 and conductivity ) 0.21 mS/cm), used in the Donnan dialysis studies, was water from the Lisbon distribution network supplemented with 500 µg/L of bromate. For the experiments performed with the IEMB system, the same water was supplemented with 60 mg/L of nitrate and 200 µg/L of bromate in the form of their sodium salts. 2.3. Microbial Culture. The reactors were inoculated with an enriched culture obtained in a bromate and nitrate medium solution, originally taken from a mixed culture of an anaerobic acidogenic reactor operated for molasses fermentation. The enrichment procedure was performed in sealed, oxygen-free 100 mL flasks with 50 mL of sterile biomedium containing 0.16 mmol/L of nitrate and bromate and 0.022 mol/L of ethanol at 25 ( 1 °C. When the cultures consumed the bromate and nitrate added, they were transferred to new flasks with fresh biomedium. This procedure was repeated four times. The enriched culture thus obtained was then used directly as inoculum for the batch tests and IEMB studies or stored at 4 °C. 2.4. Bioreduction of Nitrate and Bromate in Batch Reactors. Four experiments were conducted with culture medium supplemented with equal molar concentration (0.16 mmol/L) of nitrate (9.92 mg/L) and bromate (20.48 mg/L), and one experiment was carried out with 60 mg/L of nitrate and 200 µg/L of bromate, in the form of their sodium salts. In both cases, a 700 mL reactor was filled with 500 mL of fresh biomedium, with an ethanol concentration of 1 g/L. The reactor was inoculated with 150 mL of the enriched microbial culture (Section 2.3), the initial biomass concentration was 0.09 ( 0.02 gCell dry weight/L. The initial medium pH was adjusted to 7 (with NaOH) and anoxic conditions were maintained by a continuous flux of argon in the headspace of the reactor. The temperature was controlled at 23 ( 1 °C. Samples were taken periodically for nitrate, nitrite, bromate, and bromide analyses and cell dry weight measurements. The maximum specific reduction rates, presented in the results Section 3.1, were determined calculating the maximum slopes achieved when plotting the concentrations of bromate and nitrate against time. 2.5. IEMB Studies. The IEMB was constituted by a flat plate module with an anion exchange membrane Neosepta ACS, manufactured by Tokuyama Soda (Japan), separating two equal rectangular channels, of 260 mm in length, 15 mm in width, and 3 mm in thickness (39 cm2 of membrane area). One of the channels of this module was connected through a recirculation loop to a stirred reactor (biocompartment), filled with 400 mL of fresh biomedium containing 200 µg/L of bromate and 60 mg/L of nitrate, and inoculated with 100 mL of the enriched culture. This reactor was first operated under batch conditions during 3 days for culture growth, after which it was continuously fed with biomedium and operated at an hydraulic retention time (HRT) of 5 days. The other module channel (water compartment) was connected to a recirculation loop, allowing for adjustment of the desired hydrodynamic conditions. Recirculation was maintained in both biomedium and water circuits by two gear pumps ZP 140 (Ismatec) operating at a volumetric flow rate of 97.2 L/h, thus providing Reynolds numbers of 3000 in both channels. After the referred 3 days of culture growth, the water compartment was continuously fed with synthetically concocted polluted water, containing 60 mg/L of nitrate and 200 µg/L of bromate at 0.012 L/h, corresponding to an HRT in the water compartment of 8.3 h. Each study was conducted for 16 days and samples were taken periodically from the polluted water feed, treated water outlet, biofeed and biocompartment for conductivity meaVOL. 42, NO. 20, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7703
FIGURE 2. Evolution of nitrate, nitrite, bromate, and biomass concentrations under batch stirred bioreactor operation conditions, at 23 °C and pH 7, initial concentrations: 200 µg/L of bromate and 60 mg/L of nitrate. surements as well as nitrate, nitrite, bromate, bromide, ethanol, and cell concentration analyses. These experiments were performed at a controlled temperature of 23 ( 1 °C. 2.6. Donnan Dialysis Studies. Two experiments were performed using the same rig described in Section 2.5, but without adding ethanol and microbial culture to the receiving compartment (biocompartment). This compartment was fed continuously with a solution of 100 mmol/L NaCl. The water compartment was fed with water from the Lisbon distribution network supplemented with 500 µg/L of bromate, at two different ratios of water flow rate per membrane area (F/A ) 3.1 L/m2 · h and F/A ) 1.5 L/m2 · h). 2.7. Analytical Methods. Concentrations of BrO3-, Br-, NO3-, NO2-, and Cl-, were determined as previously described (28) using a Dionex ion exchange chromatography system constituted by an ED 50 electrochemical detector, the Ionpac AG9 Guard, and AS9 Analytical (4 mm) columns and an Anion Suppressor-ULTRA (4 mm). The mobile phase used was a 9 mmol/L Na2CO3 aqueous solution at a flow rate of 1 mL/min. At these conditions, the detection limits were 10 µg/L for BrO3- and 0.5 mg/L for Br- (the concentration of bromide in the batch experiment with bromate in the µg/L range and, in the IEMB experiments, was below the quantification limit of the ion exchange chromatography system; therefore data on bromide results are not presented in Figures 2 and 4). For the samples taken from the IEMB biocompartment and from the batch experiments with BrO3in the µg/L range, a mobile aqueous phase of 2.25 mmol/L Na2CO3 was used to guarantee a complete separation of the bromate peak from the other biomedium salt peaks. The ethanol concentration was measured by HPLC using a differential refractometer detector RI-71 and an Aminex HPX-87H column from Biorad. The mobile phase was a 0.01 N of H2SO4 aqueous solution with a flow rate of 0.5 mL/min. Under these conditions the ethanol detection limit was 1 mg/L.
3. Results and Discussion 3.1. Bioreduction of Nitrate and Bromate in a Batch Reactor. In order to investigate the ability of the enriched culture (Section 2.3) to simultaneously reduce nitrate and bromate, experiments were performed in a stirred batch bioreactor at different concentrations of the polluting anions. Four experiments were performed using an equimolar solution (0.16 mmol/L) of nitrate (9.92 mg/L) and bromate (20.48 mg/L), with the objective of comparing their biological reduction rates. An additional experiment was performed 7704
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 20, 2008
FIGURE 3. Evolution of bromate concentration in the receiving compartment (a) and in the water compartment (b), for the Donnan dialysis experiments performed at two different F/A ratios. with 60 mg/L of nitrate (0.968 mmol/L) and 200 µg/L of bromate (1.563 µmol/L), aiming at investigating the ability of the culture to reduce these two oxyanions in the range of concentrations as they typically occur in polluted waters.
TABLE 1. Maximum Specific Reduction Rate (r) and Biomass Growth Rate (µmax) for the Biological Reduction of Bromate and Nitrate Obtained with Different Electron Donors CiBrO3 (mg/L)
CiNO3 (mg/L)
electron donor
rBrO3 (mg/gcell dry weight · h)
rNO3 (mg/gcell dry weight · h)
µmax (h-1)
0.1 0.2 0.2 1.4 20.5
60 30-40 30-40 9.92
hydrogen ethanol glucose glucose ethanol acetate
0.005 0.027 0.099 0.362 0.090 0.333
36.68 14.20 18.91 1.86
0.067 0.020
ref 9 this study 22 22 this study 23
Ci, initial concentration.
Figure 2 shows the results obtained in the experiment performed with 60 mg/L (0.968 mmol/L) of nitrate and 200 µg/L (1.563 µmol/L) of bromate. The data clearly shows that bromate reduction was initiated only after the complete removal of nitrate. Nitrite was temporarily accumulated while nitrate was being reduced. These results are in agreement with findings reported in the literature (25). The maximum specific reduction rates were: 36.68 mg NO3-/gcell dry weight · h (0.64 mmol NO3-/gcell dry weight · h) and 0.027 mg BrO3-/ gcell dry weight · h (0.0002 mmol BrO3-/gcell dry weight · h), respectively. This difference in the specific reduction rates is consistent with the lower concentration of bromate compared to that of nitrate (3 orders of magnitude lower). The specific bromate reduction rate 0.09 ( 0.01 mg BrO3-/gcell dry weight · h (0.0007 mmol BrO3-/gdry weight · h), for the case where the molar concentrations of bromate and nitrate were equal, was two times lower than that of nitrate reduction 1.86 ( 0.4 mg NO3- /gcell dry weight · h (0.03 mmol NO3-/gcell dry weight · h) (Table 1). The presented reduction Gibbs free energies in the presence of ethanol as the electron donor (Section 1) demonstrate that nitrate reduction is thermodynamically more favorable when compared with bromate reduction. As mentioned in the introduction, inhibition of bromate reduction in the presence of nitrate may be explained by thermodynamic or kinetically related mechanisms. The results obtained show a significant increase of the nitrate reduction rate in the experiments performed with 60 mg/L of nitrate and 200 µg/L of bromate, compared with the experiments performed with 9.92 mg/L of nitrate and 20.48 mg/L of bromate. This may be explained by the fact that high bromate concentrations inhibit nitrate reduction (9, 24), and low nitrate concentrations decrease inhibition on bromate reduction (9, 25, 26). The stoichiometric conversion of bromate to bromide was almost one (0.143 ( 0.01 mol/L of BrO3- were reduced and 0.149 ( 0.007 mmol/L of Br- were formed, for the experiment performed with an equimolar solution of nitrate and bromate). Ginkel et al. (23) suggested the formation of bromite as a toxic intermediate species during bromate reduction, thus leading to low bromate reduction rates. However, there has been no experimental evidence supporting the formation of this intermediate, probably due to the fact that at pH lower than 8 it may rapidly decompose to bromide. Table 1 presents the values of maximum specific reduction rates obtained together with values reported in the literature for different bromate and nitrate concentrations and different electron donors, using mixed microbial cultures. Despite the broad range of reported values for specific bromate reduction rates (0.005 -0.362 mg BrO3-/ gcell dry weight · h), which can be attributed to dissimilar microbial communities and different electron donors used, they are all considerably lower compared with those for nitrate. This is in accordance with the results obtained in the present study. The lowest reduction rates were observed when using hydrogen as the electron donor. Hydrogen was used to avoid the secondary contamination of treated water with excess of electron donor (9).
The results obtained suggest that in an IEMB operation, where both pollutants are transported simultaneously through the membrane, bromate could possibly accumulate in the biocompartment since its reduction is slower and only possible after complete reduction of nitrate. This behavior is more likely to be observed in the early stages of the IEMB process, when there is still no sufficiently active biofilm developed on the membrane surface facing the biocompartment and the bioreduction of the two oxyanions may occur preferentially in the bulk solution. 3.2. Donnan Dialysis Studies. The objective of these studies was to evaluate the transport of bromate through the anion exchange membrane using drinking water supplemented with bromate. This study is a prerequisite for selecting appropriate operating conditions for the IEMB system. The experiments were performed using two different ratios of water flow rate per membrane area (F/A ratio) and the same conditions of the IEMB operation, except that no microbial culture and ethanol were added to the biocompartment. Using this procedure, it is possible to investigate the transport of bromate through the membrane and the degree of its accumulation in the receiving compartment that, ultimately, could affect the removal of the pollutant or its bioreduction. Figure 3 shows the performance of the system under these operation conditions. Bromate was transported from the water compartment to the receiving compartment without significant retention in the membrane, since its mass balances based on the liquid streams concentration measurements were well satisfied, as can be seen by comparing the “in” and “out” total data, at steady-state (Table 2). As expected, a steady state bromate concentration in the receiving compartment was achieved after 5 days, which was equal to the hydraulic retention time selected for this compartment. According to the Donnan dialysis principles, this accumulation in the receiving compartment leads to a decreased driving force for transport of bromate from the water compartment to the receiving compartment. Therefore, less bromate was transported through the membrane, which led to the observed initial increase of bromate in the water compartment (Figure 3b). The system was able to remove bromate from a high concentration of 500 µg/L (when compared with concentrations of
