- Email: [email protected]
Hydrogenotrophic denitrification in a microporous membrane bioreactor
Water Research 36 (2002) 4683–4690
Hydrogenotrophic denitrification in a microporous membrane bioreactor Bruce O. Mansella,*,1, Edward D. Schroederb a b
Department of Civil and Environmental Engineering, University of California, Davis, CA 95616, USA Department of Civil and Environmental Engineering, University of California, Davis, CA 95616, USA Received 1 March 2001; accepted 1 April 2002
Abstract Hydrogenotrophic denitrification of nitrate contaminated groundwater in a bench-scale microporous membrane bioreactor has been investigated. To prevent microbial contamination of the effluent from the reactor the nitrate-laden water treated was separated from the denitrifying culture with a 0.02 mm pore diameter membrane. Equal pressure was maintained across the membrane and nitrate was removed by molecular diffusion through the membrane and into the denitrifying culture. The system was operated with a hydrogenotrophic denitrification culture to circumvent the addition of an organic substrate to the water. Removal efficiencies ranging from 96% to 92% were achieved at influent concentrations ranging from 20 to 40 mg/L NO 3 -N. The flux values achieved in this study were 2.7–5.3 g NO3 2 1 N m d . The microporous membrane served as an effective barrier for preventing microbial contamination of the product water as evidenced by the effluent heterotrophic plate count of 9 (73.5) CFU/mL. The hydrogenotrophic culture was analyzed using available 16S and 23S rRNA-targeted oligonucleotide probes. It was determined that the enrichment process selected for organisms belonging to the beta subclass of Proteobacteria. Further analysis of the hydrogenotrophic culture indicated that the organisms may belong to the b-3 subgroup of Proteobacteria and have yet to be identified as hydrogenotrophic denitrifiers. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Denitrification; Nitrate; Membrane bioreactor; Groundwater; Hydrogenotrophic denitrification
1. Background Biological denitrification is an attractive treatment option for the remediation of nitrate contaminated groundwaters because (1) nitrate is efficiently and selectively removed by conversion to nitrogen gas and (2) the only waste products consist of biological solids. However, the application of conventional full-scale denitrification process configurations (i.e., packed or fluidized-bed reactors) is limited by the direct contact between the denitrifying microorganisms and the pro*Corresponding author. Tel.: +1-562-699-7411; fax: +1562-908-9572. E-mail address: [email protected] (B.O. Mansell). 1 Formerly, Ph.D. student, Department of Civil and Environmental Engineering, University of California, Davis, CA 95616, USA.
duct water. In addition, heterotrophic systems are typically implemented which require the addition of an organic substrate. Process by-products, such as sloughed cells and residual substrate, are imparted to the product water and must be removed by cumbersome polishing treatment steps [1,2]. Researchers have addressed the problem of direct contact between the denitrifying culture and the product water by developing various forms of confined cell reactors. A confined cell reactor may be defined as a reactor in which the organisms carrying out the desired reaction are physically separated from the water being treated. Materials used for the physical separation of denitrifying organisms have included calcium alginate gel, polyacrylamide/alginate copolymer, an agar/microporous membrane composite structure, and various types of microporous membranes [3–10]. In each of these processes nitrate was removed by molecular
0043-1354/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 3 - 1 3 5 4 ( 0 2 ) 0 0 1 9 7 - 5
B.O. Mansell, E.D. Schroeder / Water Research 36 (2002) 4683–4690
4684
Table 1 Denitrification rate results of various bench-scale confined cell reactors Separation material
MWCOa/pore Denitrification References size of rate mg 2 d1 membrane NO 3 -N m
Agar/membrane Membrane Membrane Membrane Membrane Membrane
0.45 mm 0.2 mm 0.02 mm 3500 0.02 mm 0.02 mm
a
800–1400 5800 4000 820–1230 3800–7000 2700–5300
[3] [11] [7] [9] [10] This study
Molecular weight cutoff (Daltons).
diffusion through the separation material and into the cells where the denitrification reaction occurred. The most promising results of the systems studied thus far, with respect to the denitrification rate per unit area of separation material, are shown in Table 1. To circumvent the use of an organic substrate, and the associated potential for product water contamination, autotrophic denitrification systems may be employed. The most common autotrophic systems use sulfur compounds or hydrogen gas as the electron donor and bicarbonate (or carbon dioxide) as the carbon source [1]. Hydrogen gas was selected as the electron donor for this work because sulfur-based systems typically use sulfur/ limestone filtration units that are not compatible with the confined cell membrane reactor configuration under investigation. Hydrogenotrophic denitrification systems have been operated at the lab-, pilot-, and full-scales [8,12–19]. The advantages of using hydrogen gas are (1) hydrogen is not an organic compound, (2) any residual hydrogen can easily be removed by sparging due to its low water solubility, (3) autotrophic metabolism generates less excess biomass than heterotrophic metabolism, and (4) hydrogen is less expensive than organic compounds [16]. The major limitation of using hydrogen is the concern over the safe dissolution of the gas without accumulation in confined spaces thus creating a potentially explosive atmosphere. However, with the growing interest in the potential application of this process, researchers have developed membrane dissolution systems that safely dissolve sufficient amounts of hydrogen into water [16].
2. Objectives The overall objective of the research presented here was to operate a denitrification system that addressed both the organic substrate and the microbial contamination problems by combining the two types of
processes discussed above. A bench-scale continuous flow system was developed in which a hydrogenoxidizing denitrification culture was separated from the water being treated by a microporous membrane. Equal pressure was maintained across the membrane and nitrate was removed by molecular diffusion through the membrane and into the denitrifying culture. Experiments were conducted to (1) evaluate the system’s performance with respect to nitrate removal and (2) to determine the effectiveness of the membrane material in preventing microbial contamination of the product water. In addition, the microbial community structure of the hydrogenotrophic denitrification culture was studied using fluorescent in-situ hybridization. Available 16S and 23S rRNA-targeted oligonucleotide probes were used to study the enrichment dynamics of the culture and to determine the presence or absence of organisms typically ascribed to carrying out hydrogenotrophic denitrification.
3. Materials and methods 3.1. Enrichment culture The hydrogen-oxidizing denitrification culture was enriched from activated sludge taken from the UC Davis wastewater treatment plant. Seed sludge was added to a 2 L flask containing nitrate, bicarbonate, tap water, and buffer. A bicarbonate-carbon to nitrate–nitrogen ratio of 2:1 was added to ensure that nitrate was the limiting nutrient based on the stoichiometry of Eq. (1) (C:N=0.21:1) [20]. The culture was buffered to an approximate pH of 7 with 1.74 g/L KH2 PO4 and 2.14 g/ L K2HPO4 per 0.1 g/L NO 3 -N feed [7]. Hydrogen gas was fed to the system by diffusion through 91 cm of (6 mm inside diameter) silicone tubing clamped at the end and submerged in the enrichment medium. H2 þ 0:33NO-3 þ 0:08CO2 þ 0:34Hþ -0:015C5 H7 NO2 þ 1:11H2 O þ 0:16N2 : ð1Þ 3.2. Experimental system The research was carried out in the continuous flow system as shown in Fig. 1. The system reactor consisted of two flow channels separated by an expanded polytetrafluoroethylene membrane material (GoreTexs, W.L. Gore, Elkton, MD) with a mean pore size of 0.02 mm, an average thickness of 80 mm, and an effective porosity of 50%. The expanded or stretched polymer is chemically inert, has a low friction coefficient, functions within a wide temperature range, and does not age. Because the membrane material is hydrophobic, wetting the material with methanol was required to allow the pores to fill with water. The
B.O. Mansell, E.D. Schroeder / Water Research 36 (2002) 4683–4690
4685
Product water Recycle 0.02 µm pore size membrane H2 Vent Denitrifying culture
Nitrate diffusion
HCO3Buffer
Na2SO3
P
P
P
91 cm, 6 mm inside diameter silicone tubing
Wasting
Nitrate-laden water
UV reactor
Fig. 1. Bench-scale hydrogenotrophic membrane denitrification system.
membrane reactor was constructed of Plexiglass and each flow channel had dimensions of 20 2 0.5 cm3 (total volume=20 cm3). The suspended growth hydrogenotrophic denitrifying culture was maintained in an external bioreactor and was pumped through one flow channel of the membrane reactor. Nitrate-laden water was pumped through the other flow channel. As the nitrate-laden water flowed through the reactor, nitrate was removed by molecular diffusion through the membrane and into the denitrifying culture. Hydrogen gas was diffused into the denitrifying culture with silicone tubing as during the enrichment process. Bicarbonate and buffer were added directly to the bioreactor. The influent nitrate-laden water was disinfected prior to entering the bioreactor with a Trojan Advantage-2 UV reactor (Trojan Technologies Inc., London, Ont., Canada). Anoxic conditions were maintained in the influent water by the addition of sodium sulfite. 3.3. Fluorescent in-situ hybridization Detailed procedures for cell preparation, hybridization, microscopy, and enumeration were adapted from
Loge et al. [21]. Optimal hybridization temperatures were obtained from each probes respective reference listed in Table 2. The oligonucleotide probes used in this study were synthesized by Integrated DNA Technologies (Coralville, IA). The probes were HPLC purified and labeled at the 50 end with the fluorescent dye Bodipy TMR-X (Molecular Probes, Eugene, OR). Additional unlabeled probes (BET42a, GAM42a, BONE23a, BTWO23a) were synthesized and used as competitor oligonucleotides during the hybridization process. For example, BET42a (unlabeled) was used as a competitor for GAM42a (labeled) and GAM42a (unlabeled) was used as a competitor for BET42a (labeled) [23]. BONE23a and BTWO23a were hybridized competitively as well [24]. Competitor probes were required to prevent nonspecific binding due to the one base difference between the probes. For example, GAM42a could hybridize with b subclass organisms because of the one base difference between GAM42a and BET42a. All probes were shipped lyophilized and were dissolved in the manufacturer recommended solution (10 mM Tris, pH 8, 1 mM EDTA) prior to use.
B.O. Mansell, E.D. Schroeder / Water Research 36 (2002) 4683–4690
4686
Table 2 Oligonucleotide probes used for the hydrogenotrophic denitrification culture analysis Probe
Specificity
Sequence (50 –30 )
rRNA target site
References
EUB338 ALF1b
Bacteria a subclass and several members of the delta subclass of Proteobacteria b subclass of Proteobacteria g subclass of Proteobacteria b-1 subgroup of Proteobacteria Competitor for BONE23a, b-2, some b-3, and some g subclass
GCTGCCTCCCGTAGGAGT CGTTCGYTCTGAGCCAG
16S (338–335) 16S (19–35)
Amann et al. [22] Manz et al. [23]
GCCTTCCCACTTCGTTT GCCTTCCCACATCGTTT GAATTCCATCCCCCTCT GAATTCCACCCCCCTCT
23S 23S 16S 16S
Manz et al. [23] Manz et al. [23] Amann et al. [24] Amann et al. [24]
BET42a GAM42a BONE23a BTWO23a
(1027–1043) (1027–1043) (663–679) (663–679)
Table 3 Hydrogenotrophic membrane system experiments Experiment
Q; cm3/min
C0 ; mg/L–N
C; mg/L–N
% Removal
Operation hours
1 2 3 4 5
0.4 0.8 1.2 0.4 0.4
20 20 20 30 40
0.87 (70.07) 5.3 (70.18) 8.4 (70.24) 1.8 (70.05) 3.2 (70.28)
95.6 73.5 58.0 94.0 92.0
116 117 114 117 143
Co =Influent nitrate concentration. C=Reactor effluent nitrate concentration.
3.4. Analytical methods Nitrate concentrations were measured using a continuous flow inorganic nitrogen analyzer (Timberline Model 383, Timberline Instruments Inc., Boulder, CO) with a method detection level, bias, and precision comparable to methods found in Standard Methods [25] [26]. Nitrate standards were prepared with potassium nitrate and deionized water. Nitrite concentrations were measured with Hach NitriVers2 reagent powder pillows (Hach Company, Loveland, CO) and a Shimadzu UV160U spectrophotometer at 585 nm (Kyoto, Japan). Nitrite standards were prepared with sodium nitrite and deionized water. Total suspended solids concentrations were measured in accordance with Standard Methods (Method 2540D, 1992). A combination pH electrode (model 50205) was used to measure pH (Hach Company, Loveland, CO). Heterotrophic plate counts (HPC) were determined according to Standard Methods (Method 9215C, 1992) using R2A agar at 201C for 7 days. Turbidity measurements were made with a Hach model 2100A turbidimeter.
4. Experiments Flowrates, operation times and influent nitrate concentrations for five experiments are shown in Table 3. The suspended solids concentration and culture recircu-
lation flowrate were 400 mg/L and 500 cm3/min for all experiments. For experiments 1–3 the primary variable was the influent flowrate. In experiments 4 and 5 the influent nitrate concentration was varied. The influent nitrate concentration range (20–40 mg/L NO 3 -N) was selected since typical nitrate concentrations in polluted potable groundwater sources range between 10 and 40 mg/L NO 3 -N [27]. All experiments were conducted at approximately 211C. Throughout the operation of the system the microbial water quality of the effluent was monitored by HPCs. In addition, the turbidity of the effluent was monitored. These measurements were made to determine the effectiveness of the membrane material used in this study as a barrier for preventing microbial contamination of the product water. Five samples were collected and analyzed at 12, 20, 32, 47, and 55 days of operation. To reduce contamination due to organisms in the tap water used to prepare the nitrate feed water, a UV reactor (Advantage-2, Trojan Technologies Inc., London, Ont., Canada) was used. Periodically, the influent water was recirculated through the UV reactor at a flowrate of 0.5 L/min for 2 h The actual UV dose was not calculated. However, the UV system was designed for flows up to 7.6 L/min. To study the hydrogenotrophic denitrification culture a sequential analysis of the community structure at different points in the enrichment process was carried out. The activated sludge seed, the suspended
B.O. Mansell, E.D. Schroeder / Water Research 36 (2002) 4683–4690
enrichment culture at 30 and 60 days, and the bioreactor culture at 104 days (biofilm and suspension) were sampled and analyzed using the 16S and 23S rRNAtargeted probes listed in Table 2. To our knowledge, none of the probes listed in Table 2 have been used for the analysis of a hydrogenotrophic denitrification culture.
5. Results and discussion Reactor effluent nitrate concentrations and corresponding percent removals for experiments 1–5 are shown in Table 3. Throughout the operation of the system the denitrification culture nitrate concentration remained negligible. In addition, nitrite concentrations in the culture and reactor effluent were always negligible (data not shown). In each of the experiments conducted the effluent nitrate concentrations were lower than the EPA MCL of 10 mg/L NO 3 -N. The average effluent nitrate concentration for the first 78 h of experiment 1 was 3.3 mg/L NO 3 -N. After this time period the effluent concentration stabilized to below 1 mg/L NO 3 -N. A similar observation was made with a heterotrophic membrane system operated previously [10]. As with the heterotrophic system, this observation may be explained by the presence of nitrate directly adjacent to the membrane and the subsequent development of an actively metabolizing biofilm on the surface of the membrane. The development of the biofilm decreased the nitrate concentration at the membrane interface and therefore increased the concentration gradient. Since the concentration gradient is the mass transfer driving force, a decrease in the effluent concentration resulted. 5.1. Confined cell reactor comparison The denitrification rate per unit area results obtained from this system were significantly higher than the results obtained in systems operated by other researchers (Table 1). The lowest value in the range (2700 mg NO 3N m2 d1) obtained in this study was approximately 2 times greater than the highest values obtained by Lemoine et al. [3] and Fuchs et al. [9]. The lower denitrification rates can be attributed to the relatively poor nitrate mass transfer characteristics of the separation materials used in the other systems. Lemoine et al. [3] used an agar/microporous membrane composite structure. The membrane material used to sandwich the agar layer had a nitrate diffusivity of 2.1 106 cm2/ s. Fuchs et al. [9] used a regenerated cellulose membrane material with a nitrate mass transfer coefficient of 0.00037 cm/s. For a quantitative comparison, the apparent nitrate diffusion coefficient and mass transfer coefficient of the
4687
membrane material used in this study were calculated using a previously developed model of the system [10]. The model is shown in Eq. (2) Dn wLe C ¼ C0 exp ; ð2Þ dm Q 3 where C is the effluent NO 3 -N concentration (g/cm ), 3 C0 the influent NO3 -N concentration (g/cm ), L the length of microporous membrane (cm), e the microporous membrane effective porosity, Dn the diffusion coefficient of NO 3 -N through microporous membrane pore area (cm2/s), dm the microporous membrane thickness (cm), w the width of microporous membrane (cm), Q the volumetric flowrate of water being treated (cm3/s). The values of dm ; w; L; and e were equal to 80 mm, 2 cm, 20 cm, and 0.50, respectively. The coefficients were calculated by fitting data from experiments 1–3 to the linear form of the model shown in Eq. (3) where Kn ¼ Dn =dm ¼ mass transfer coefficient (cm/s) and A ¼ hLe=membrane pore area (cm2). A ð3Þ ln C ¼ Kn þ ln C0 : Q
The mass transfer coefficient ðKn Þ was determined to be 0.0006 cm/s ðR2 ¼ 0:9961Þ which is 62% higher than the nitrate mass transfer coefficient of the regenerated cellulose membrane material used by Fuchs et al. [9]. The diffusivity (Dn) was subsequently calculated to be 4.8 106 cm2/s which is 128% higher than the membrane material used by Lemoine et al. [3]. 5.2. Effluent microbial quality Based on the HPC and turbidity data presented in Table 4, the microporous membrane used in this study was an effective barrier for preventing microbial contamination of the reactor effluent. The average HPC was 9 (73.5) CFU/mL and the average turbidity was 0.4 NTU. The HPC average was significantly lower than the requirement of 500 CFU/mL stipulated by the USEPA Surface Water Treatment Rule.
Table 4 Effluent turbidity and heterotrophic plate count data Operation time, days
Turbidity, NTU
HPCa, CFU/mL
12 20 32 47 55
0.4 0.3 0.4 0.5 0.4
14 11 7 5 8
a
Heterotrophic plate count (201C for 7 days). Average=9 (73.5) CFU/mL.
B.O. Mansell, E.D. Schroeder / Water Research 36 (2002) 4683–4690
4688
Table 5 Results of hydrogenotrophic culture analysis Probe
EUB338 ALF1b BET42a GAM42a BONE23a BTWO23a
Activated sludge
80.078.5 23.277.5 21.778.7 14.574.0 — —
% Of total DAPI count 30-Day enrichment
60-Day enrichment
104-Day bioreactor biofilm/suspension
91.874.6 0.0 89.977.2 0.0 — —
94.573.6 0.0 87.572.8 0.0 — —
92.374.3/94.473.1 0.0 91.375.8/93.073.4 0.0 0.0 0.0
5.3. Culture analysis Community structure results for the activated sludge seed, 30- and 60-day suspended enrichment cultures, and the bioreactor culture are presented in Table 5. Approximately 81% of the organisms in the activated sludge culture hybridized with the probe EUB338 indicating that that the large majority of the organisms were bacteria. The activated sludge community structure was complex consisting of organisms from the a (23.277.5%), b (21.778.7%), and g (14.574.0%) subclasses of Proteobacteria. A population shift occurred from the activated sludge seed to the 30-day enrichment culture. The number of organisms hybridizing to probe EUB338 increased from 80.8% (78.5) to 91.8% (74.6). This increase was most likely due to the selective growth environment and the related selection for gram-negative bacteria as opposed to eukaryotic cells that comprised a fraction of the activated sludge seed. The enrichment process selected for a significantly less diverse culture of bacteria belonging to the b subclass of Proteobacteria (89.977.2%). The 60-day enrichment community structure remained very similar with 87.572.8% and 94.573.6% of the bacteria hybridizing to the BET42a and EUB338 probes, respectively. The community structure did not change significantly after inoculation into the membrane bioreactor. The second objective of the culture analysis was to determine the presence or absence of the identified hydrogenotrophic denitrifiers listed in Table 6 [6,28–30]. Three of the organisms (Azospirillum brasilence, Rhizobium japonicum, and Paracoccus denitrificans) belong to the a subclass of Proteobacteria. The remaining four organisms (Hydrogenophaga flava, Hydrogenophaga pseudopflava, Hydrogenophaga taeniospiralis, and Ralstonia eutropha) belong to the b subclass of Proteobacteria; more specifically the b-2 subgroup. As discussed above, the culture developed in this study was successfully hybridized with the BET42a probe only. This eliminates A. brasilence, R. japonicum, and P. denitrificans as potential members of the community. After
Table 6 Hydrogen-oxidizing organisms capable of denitrification A. brasilence H. flavaa H. pseudopflavaa H. taeniospiralisa a b
P. denitrificans R. eutrophab R. japonicum
Formerly Pseudomonas. Formerly Alcaligenes eutrophus.
hybridization with the BET42a probe the culture was analyzed with the probes BONE23a and BTWO23a to determine which subgroup the organisms may belong to. It is important to note that the BTWO23a probe was originally designed only as a competitor for the BONE23a probe and is therefore not very specific. In addition to b-2 subgroup organisms the probe will also hybridize with some members of the b-3 subgroup and some g subclass organisms [31]. As shown in Table 5, neither of the probes hybridized with organisms in the culture. These results indicate that the organisms may belong to the b-3 subgroup. However, none of the organisms present in the culture have been directly identified as known hydrogenotrophic denitrifiers. 5.4. Summary A bench-scale hydrogenotrophic denitrification system has been successfully operated. The main objective for operating the system was to study solutions to two problems associated with conventional denitrification systems: (1) the problem of the direct contact between the denitrifying organisms and the product water and (2) the addition of an organic energy source and the associated problem of product water contamination. Important results from this research are as follows. *
Removal efficiencies ranging from 96% to 92% were achieved at influent concentrations ranging from 20 to 40 mg/L NO 3 -N. In experiments with influent concentrations of 20, 30, and 40 mg/L NO 3 -N the
B.O. Mansell, E.D. Schroeder / Water Research 36 (2002) 4683–4690
*
*
*
effluent concentrations were 0.87, 1.8, and 3.2 mg/L NO 3 -N, respectively. Each of these effluent concentrations were well below the US EPA MCL of 10 mg/ L NO 3 -N. The flux values achieved in this study were 2.7–5.3 g 2 1 NO d . These values were significantly 3 -N m higher than the most promising results obtained in other confined cell reactor systems operated by other researchers. The microporous membrane used in this study served as an effective barrier for preventing microbial contamination of the product water. The effluent HPC was 9 (73.5) CFU/mL which is significantly less than the requirement of 500 CFU/mL stipulated by the US EPA Surface Water Treatment Rule. The hydrogenotrophic culture was analyzed using available 16S and 23S rRNA-targeted oligonucleotide probes. It was determined that the enrichment process selected for organisms belonging to b subclass of Proteobacteria. Further analysis of the hydrogenotrophic culture indicated that the organisms may belong to the b-3 subgroup of Proteobacteria and have yet to be identified as hydrogenotrophic denitrifiers.
Acknowledgements This work was supported by the US Environmental Protection Agency through a STAR fellowship to Bruce O. Mansell and by Timberline Instruments, Inc., Boulder, CO.
References [1] Kapoor A, Viraraghavan T. Nitrate removal from drinking water—review. J Environ Eng 1997;123(4):371–80. [2] Mateju V, Cizinska S, Krejci J, Janoch T. Biological water denitrification—a review. Enzyme Microb Technol 1992;14:170–83. [3] Lemoine D, Jouenne T, Junter GA. Biological denitrification of water in a two-chambered immobilized-cell bioreactor. Appl Microbiol Biotechnol 1991;36:257–64. [4] Mattiasson B, Ramstorp M, Nilsson I, Hahn-H.agerdal B. Comparison of the performance of a hollow-fiber microbe reactor with a reactor containing alginate entrapped cells. Biotechnol Lett 1981;3(10):561–6. . L, Molin N, Mosbach K. [5] Nilsson I, Olson S, H.aggstrom Denitrification of water using immobilized Pseudomonas denitrificans cells. Eur J Appl Microbiol Biotechnol 1980;10:261–73. [6] Neal JL, Allen GC, Morse RD, Wolf DD. Nitrate, nitrite, nitrous oxide and oxygen-dependent hydrogen uptake by Rhizobium. FEMS Microbiol Lett 1983;17:335–8. [7] Reising AR, Schroeder ED. Denitrification incorporating microporous membranes. J Environ Eng 1996;122(7):599– 604.
4689
[8] Chang CC, Tseng SK, Huang HK. Hydrogenotrophic denitrification with immobilized Alcaligenes eutrophus for drinking water treatment. Bioresource Technol 1999;69:53–8. [9] Fuchs W, Schatzmayr G, Braun R. Nitrate removal from drinking water using a membrane-fixed biofilm reactor. Appl Microbiol Biotechnol 1997;48:267–74. [10] Mansell BO, Schroeder ED. Biological denitrification in a continuous flow membrane reactor. Water Res, 1999;33(8):1845–50. [11] Mcleaf PR, Schroeder ED. Denitrification using a membrane immobilized biofilm. J Am Water Works Assoc 1995;87(3):77–86. [12] Kurt M, Dunn IJ, Bourne JR. Biological denitrification of drinking water using autotrophic organisms with H2 in a fluidized-bed reactor. Biotechnol Bioeng 1987;29:493–501. [13] Dries D, Liessens J, Verstraete W, Stevens P, de Vos P, de Ley J. Nitrate removal from drinking water by means of hydrogenotrophic denitrifiers in a polyurethane carrier reactor. Water Supply 1987;6:181–92. [14] Gross H, Schnoor G, Rutten P. Biological denitrification process with hydrogen-oxidizing bacteria for drinking water treatment. Water Supply 1987;6:193–8. [15] Ridgway HF, Davidson MS. Biological denitrification research project. Phase 1 report, Cooperative effort of the Orange County Water District, the Metropolitan Water District of Southern California, and the Santa Ana Watershed Project Authority, 1991. [16] Gantzer CJ. Membrane dissolution of hydrogen for biological nitrate removal. Proceedings of the 68th Annual Conference and Exhibition, Water Environment Federation, Research Symposium, Miami Beach, FL, 1995. p. 49– 60. [17] Ergas SJ, Reuss AF. Hydrogenotrophic denitrification of drinking water using a hollow-fiber membrane reactor. J Water SRT—Aqua 2001;50:161–71. [18] Ho CM, Tseng SK, Chang YJ. Autotrophic denitrification via a novel membrane-attached biofilm reactor. Lett Appl Microbiol 2001;33(3):201–5. [19] Lee K-C, Rittman BE. A novel hollow-fiber membrane biofilm reactor for autohydrogenotrophic denitrification of drinking water. Water Sci Technol 2000;41:219–26. [20] McCarty PL. Stoichiometry of biological reactions. International Conference Towards a Unified Concept of Biological Waste Treatment Design. Atlanta, GA, 1972. [21] Loge FJ, Emerick RW, Thompson DE, Nelson DC, Darby JL. Development of a fluoresecent 16S rRNA oligonucleotide probe specific to the family Enterobacteriaceae. Water Environ Res 1999;71(1):75–83. [22] Amann RI, Binder BJ, Olson RJ, Chisholm SW, Devereux R, Stahl DA. Combination of 16S rRNA-targeted oligonucleotide probes with flow cytometry for analyzing mixed microbial populations. Appl Environ Microbiol 1990;56:1919–25. [23] Manz W, Amann R, Ludwig W, Wagner M, Schleifer KH. Phylogenetic oligodeoxynucleotide probes for the major subclasses of Proteobacteria: problems and solutions. System Appl Microbiol 1992;15:593–600. [24] Amann R, Snaidr J, Wagner M, Ludwig W, Schleifer K-H. In-situ visualization of high genetic diversity in a natural microbial community. J Bacteriol 1996;178:3496–500.
4690
B.O. Mansell, E.D. Schroeder / Water Research 36 (2002) 4683–4690
[25] Selenka F, Dressler R. Microbiological and chemical investigations on a biological, autotrophic denitrification plant using hydrogen as an energy source. Aqua 1990;39(2):107–16. [26] Mansell BO, DeVellis L, Schroeder ED. Technical note automated separation and conductimetric determination of inorganic nitrogen. J Environ Eng 2000;126(8):778–80. [27] Cook NE, Silverstein J, Veydovec B, de Mendonca MM, Sydney R. Feild demonstration of biological denitrification of polluted groundwater and pilot-scale field testing of biological denitrification with widely varied hydraulic loading rates. Colorado Water Resources Research Institute, Colorado State University, Fort Collins Colorado, December, 1991. [28] Aragno M, Schlegel HG. The hydrogen-oxidizing bacteria. In: Starr MP, Stolp H, Truper H, Balows A, Schlegel HG, editors. The prokaryotes: a handbook on habitats, isolation, and identification of bacteria. Berlin: Springer, 1981.
[29] Tibelius KH, Knowles R. Uptake hydrogenase activity in denitrifying Azospirillum brasilence grown anaerobically with nitrous oxide or nitrate. J Bacteriol 1984;157(1):84–8. [30] Willems A, Busse J, Goor M, Pot B, Falsen E, Jantzen E, Hoste B, Gillis M, Kersters K, Auling G, de Ley J. Hydrogenophaga, a new genus of hydrogen-oxidizing bacteria that includes Hydrogenophaga flava comb. nov. (formerly Pseudomonas flava), Hydrogenophaga palleronii (formerly Pseudomonas palleronii), Hydrogenophaga pseudoflava (formerly Pseudomonas pseudoflava and Pseudomonas carboxydoflava), and Hydrogenophaga taeniospiralis (formerly Pseudomonas taeniospiralis). Int J Syst Bacteriol 1989;39:319–33. [31] Crocetti G, Hugenholtz P, Bond P, Schuler A, Keller J, Jenkins D, Blackall L. Identification of polyphosphateaccumulating organisms and design of 16S rRNA-directed probes for their detection and quantification. Appl Environ Microbiol 2000;66:1175–82.
Hydrogenotrophic denitrification in a microporous membrane bioreactor Bruce O. Mansella,*,1, Edward D. Schroederb a b
Department of Civil and Environmental Engineering, University of California, Davis, CA 95616, USA Department of Civil and Environmental Engineering, University of California, Davis, CA 95616, USA Received 1 March 2001; accepted 1 April 2002
Abstract Hydrogenotrophic denitrification of nitrate contaminated groundwater in a bench-scale microporous membrane bioreactor has been investigated. To prevent microbial contamination of the effluent from the reactor the nitrate-laden water treated was separated from the denitrifying culture with a 0.02 mm pore diameter membrane. Equal pressure was maintained across the membrane and nitrate was removed by molecular diffusion through the membrane and into the denitrifying culture. The system was operated with a hydrogenotrophic denitrification culture to circumvent the addition of an organic substrate to the water. Removal efficiencies ranging from 96% to 92% were achieved at influent concentrations ranging from 20 to 40 mg/L NO 3 -N. The flux values achieved in this study were 2.7–5.3 g NO3 2 1 N m d . The microporous membrane served as an effective barrier for preventing microbial contamination of the product water as evidenced by the effluent heterotrophic plate count of 9 (73.5) CFU/mL. The hydrogenotrophic culture was analyzed using available 16S and 23S rRNA-targeted oligonucleotide probes. It was determined that the enrichment process selected for organisms belonging to the beta subclass of Proteobacteria. Further analysis of the hydrogenotrophic culture indicated that the organisms may belong to the b-3 subgroup of Proteobacteria and have yet to be identified as hydrogenotrophic denitrifiers. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Denitrification; Nitrate; Membrane bioreactor; Groundwater; Hydrogenotrophic denitrification
1. Background Biological denitrification is an attractive treatment option for the remediation of nitrate contaminated groundwaters because (1) nitrate is efficiently and selectively removed by conversion to nitrogen gas and (2) the only waste products consist of biological solids. However, the application of conventional full-scale denitrification process configurations (i.e., packed or fluidized-bed reactors) is limited by the direct contact between the denitrifying microorganisms and the pro*Corresponding author. Tel.: +1-562-699-7411; fax: +1562-908-9572. E-mail address: [email protected] (B.O. Mansell). 1 Formerly, Ph.D. student, Department of Civil and Environmental Engineering, University of California, Davis, CA 95616, USA.
duct water. In addition, heterotrophic systems are typically implemented which require the addition of an organic substrate. Process by-products, such as sloughed cells and residual substrate, are imparted to the product water and must be removed by cumbersome polishing treatment steps [1,2]. Researchers have addressed the problem of direct contact between the denitrifying culture and the product water by developing various forms of confined cell reactors. A confined cell reactor may be defined as a reactor in which the organisms carrying out the desired reaction are physically separated from the water being treated. Materials used for the physical separation of denitrifying organisms have included calcium alginate gel, polyacrylamide/alginate copolymer, an agar/microporous membrane composite structure, and various types of microporous membranes [3–10]. In each of these processes nitrate was removed by molecular
0043-1354/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 3 - 1 3 5 4 ( 0 2 ) 0 0 1 9 7 - 5
B.O. Mansell, E.D. Schroeder / Water Research 36 (2002) 4683–4690
4684
Table 1 Denitrification rate results of various bench-scale confined cell reactors Separation material
MWCOa/pore Denitrification References size of rate mg 2 d1 membrane NO 3 -N m
Agar/membrane Membrane Membrane Membrane Membrane Membrane
0.45 mm 0.2 mm 0.02 mm 3500 0.02 mm 0.02 mm
a
800–1400 5800 4000 820–1230 3800–7000 2700–5300
[3] [11] [7] [9] [10] This study
Molecular weight cutoff (Daltons).
diffusion through the separation material and into the cells where the denitrification reaction occurred. The most promising results of the systems studied thus far, with respect to the denitrification rate per unit area of separation material, are shown in Table 1. To circumvent the use of an organic substrate, and the associated potential for product water contamination, autotrophic denitrification systems may be employed. The most common autotrophic systems use sulfur compounds or hydrogen gas as the electron donor and bicarbonate (or carbon dioxide) as the carbon source [1]. Hydrogen gas was selected as the electron donor for this work because sulfur-based systems typically use sulfur/ limestone filtration units that are not compatible with the confined cell membrane reactor configuration under investigation. Hydrogenotrophic denitrification systems have been operated at the lab-, pilot-, and full-scales [8,12–19]. The advantages of using hydrogen gas are (1) hydrogen is not an organic compound, (2) any residual hydrogen can easily be removed by sparging due to its low water solubility, (3) autotrophic metabolism generates less excess biomass than heterotrophic metabolism, and (4) hydrogen is less expensive than organic compounds [16]. The major limitation of using hydrogen is the concern over the safe dissolution of the gas without accumulation in confined spaces thus creating a potentially explosive atmosphere. However, with the growing interest in the potential application of this process, researchers have developed membrane dissolution systems that safely dissolve sufficient amounts of hydrogen into water [16].
2. Objectives The overall objective of the research presented here was to operate a denitrification system that addressed both the organic substrate and the microbial contamination problems by combining the two types of
processes discussed above. A bench-scale continuous flow system was developed in which a hydrogenoxidizing denitrification culture was separated from the water being treated by a microporous membrane. Equal pressure was maintained across the membrane and nitrate was removed by molecular diffusion through the membrane and into the denitrifying culture. Experiments were conducted to (1) evaluate the system’s performance with respect to nitrate removal and (2) to determine the effectiveness of the membrane material in preventing microbial contamination of the product water. In addition, the microbial community structure of the hydrogenotrophic denitrification culture was studied using fluorescent in-situ hybridization. Available 16S and 23S rRNA-targeted oligonucleotide probes were used to study the enrichment dynamics of the culture and to determine the presence or absence of organisms typically ascribed to carrying out hydrogenotrophic denitrification.
3. Materials and methods 3.1. Enrichment culture The hydrogen-oxidizing denitrification culture was enriched from activated sludge taken from the UC Davis wastewater treatment plant. Seed sludge was added to a 2 L flask containing nitrate, bicarbonate, tap water, and buffer. A bicarbonate-carbon to nitrate–nitrogen ratio of 2:1 was added to ensure that nitrate was the limiting nutrient based on the stoichiometry of Eq. (1) (C:N=0.21:1) [20]. The culture was buffered to an approximate pH of 7 with 1.74 g/L KH2 PO4 and 2.14 g/ L K2HPO4 per 0.1 g/L NO 3 -N feed [7]. Hydrogen gas was fed to the system by diffusion through 91 cm of (6 mm inside diameter) silicone tubing clamped at the end and submerged in the enrichment medium. H2 þ 0:33NO-3 þ 0:08CO2 þ 0:34Hþ -0:015C5 H7 NO2 þ 1:11H2 O þ 0:16N2 : ð1Þ 3.2. Experimental system The research was carried out in the continuous flow system as shown in Fig. 1. The system reactor consisted of two flow channels separated by an expanded polytetrafluoroethylene membrane material (GoreTexs, W.L. Gore, Elkton, MD) with a mean pore size of 0.02 mm, an average thickness of 80 mm, and an effective porosity of 50%. The expanded or stretched polymer is chemically inert, has a low friction coefficient, functions within a wide temperature range, and does not age. Because the membrane material is hydrophobic, wetting the material with methanol was required to allow the pores to fill with water. The
B.O. Mansell, E.D. Schroeder / Water Research 36 (2002) 4683–4690
4685
Product water Recycle 0.02 µm pore size membrane H2 Vent Denitrifying culture
Nitrate diffusion
HCO3Buffer
Na2SO3
P
P
P
91 cm, 6 mm inside diameter silicone tubing
Wasting
Nitrate-laden water
UV reactor
Fig. 1. Bench-scale hydrogenotrophic membrane denitrification system.
membrane reactor was constructed of Plexiglass and each flow channel had dimensions of 20 2 0.5 cm3 (total volume=20 cm3). The suspended growth hydrogenotrophic denitrifying culture was maintained in an external bioreactor and was pumped through one flow channel of the membrane reactor. Nitrate-laden water was pumped through the other flow channel. As the nitrate-laden water flowed through the reactor, nitrate was removed by molecular diffusion through the membrane and into the denitrifying culture. Hydrogen gas was diffused into the denitrifying culture with silicone tubing as during the enrichment process. Bicarbonate and buffer were added directly to the bioreactor. The influent nitrate-laden water was disinfected prior to entering the bioreactor with a Trojan Advantage-2 UV reactor (Trojan Technologies Inc., London, Ont., Canada). Anoxic conditions were maintained in the influent water by the addition of sodium sulfite. 3.3. Fluorescent in-situ hybridization Detailed procedures for cell preparation, hybridization, microscopy, and enumeration were adapted from
Loge et al. [21]. Optimal hybridization temperatures were obtained from each probes respective reference listed in Table 2. The oligonucleotide probes used in this study were synthesized by Integrated DNA Technologies (Coralville, IA). The probes were HPLC purified and labeled at the 50 end with the fluorescent dye Bodipy TMR-X (Molecular Probes, Eugene, OR). Additional unlabeled probes (BET42a, GAM42a, BONE23a, BTWO23a) were synthesized and used as competitor oligonucleotides during the hybridization process. For example, BET42a (unlabeled) was used as a competitor for GAM42a (labeled) and GAM42a (unlabeled) was used as a competitor for BET42a (labeled) [23]. BONE23a and BTWO23a were hybridized competitively as well [24]. Competitor probes were required to prevent nonspecific binding due to the one base difference between the probes. For example, GAM42a could hybridize with b subclass organisms because of the one base difference between GAM42a and BET42a. All probes were shipped lyophilized and were dissolved in the manufacturer recommended solution (10 mM Tris, pH 8, 1 mM EDTA) prior to use.
B.O. Mansell, E.D. Schroeder / Water Research 36 (2002) 4683–4690
4686
Table 2 Oligonucleotide probes used for the hydrogenotrophic denitrification culture analysis Probe
Specificity
Sequence (50 –30 )
rRNA target site
References
EUB338 ALF1b
Bacteria a subclass and several members of the delta subclass of Proteobacteria b subclass of Proteobacteria g subclass of Proteobacteria b-1 subgroup of Proteobacteria Competitor for BONE23a, b-2, some b-3, and some g subclass
GCTGCCTCCCGTAGGAGT CGTTCGYTCTGAGCCAG
16S (338–335) 16S (19–35)
Amann et al. [22] Manz et al. [23]
GCCTTCCCACTTCGTTT GCCTTCCCACATCGTTT GAATTCCATCCCCCTCT GAATTCCACCCCCCTCT
23S 23S 16S 16S
Manz et al. [23] Manz et al. [23] Amann et al. [24] Amann et al. [24]
BET42a GAM42a BONE23a BTWO23a
(1027–1043) (1027–1043) (663–679) (663–679)
Table 3 Hydrogenotrophic membrane system experiments Experiment
Q; cm3/min
C0 ; mg/L–N
C; mg/L–N
% Removal
Operation hours
1 2 3 4 5
0.4 0.8 1.2 0.4 0.4
20 20 20 30 40
0.87 (70.07) 5.3 (70.18) 8.4 (70.24) 1.8 (70.05) 3.2 (70.28)
95.6 73.5 58.0 94.0 92.0
116 117 114 117 143
Co =Influent nitrate concentration. C=Reactor effluent nitrate concentration.
3.4. Analytical methods Nitrate concentrations were measured using a continuous flow inorganic nitrogen analyzer (Timberline Model 383, Timberline Instruments Inc., Boulder, CO) with a method detection level, bias, and precision comparable to methods found in Standard Methods [25] [26]. Nitrate standards were prepared with potassium nitrate and deionized water. Nitrite concentrations were measured with Hach NitriVers2 reagent powder pillows (Hach Company, Loveland, CO) and a Shimadzu UV160U spectrophotometer at 585 nm (Kyoto, Japan). Nitrite standards were prepared with sodium nitrite and deionized water. Total suspended solids concentrations were measured in accordance with Standard Methods (Method 2540D, 1992). A combination pH electrode (model 50205) was used to measure pH (Hach Company, Loveland, CO). Heterotrophic plate counts (HPC) were determined according to Standard Methods (Method 9215C, 1992) using R2A agar at 201C for 7 days. Turbidity measurements were made with a Hach model 2100A turbidimeter.
4. Experiments Flowrates, operation times and influent nitrate concentrations for five experiments are shown in Table 3. The suspended solids concentration and culture recircu-
lation flowrate were 400 mg/L and 500 cm3/min for all experiments. For experiments 1–3 the primary variable was the influent flowrate. In experiments 4 and 5 the influent nitrate concentration was varied. The influent nitrate concentration range (20–40 mg/L NO 3 -N) was selected since typical nitrate concentrations in polluted potable groundwater sources range between 10 and 40 mg/L NO 3 -N [27]. All experiments were conducted at approximately 211C. Throughout the operation of the system the microbial water quality of the effluent was monitored by HPCs. In addition, the turbidity of the effluent was monitored. These measurements were made to determine the effectiveness of the membrane material used in this study as a barrier for preventing microbial contamination of the product water. Five samples were collected and analyzed at 12, 20, 32, 47, and 55 days of operation. To reduce contamination due to organisms in the tap water used to prepare the nitrate feed water, a UV reactor (Advantage-2, Trojan Technologies Inc., London, Ont., Canada) was used. Periodically, the influent water was recirculated through the UV reactor at a flowrate of 0.5 L/min for 2 h The actual UV dose was not calculated. However, the UV system was designed for flows up to 7.6 L/min. To study the hydrogenotrophic denitrification culture a sequential analysis of the community structure at different points in the enrichment process was carried out. The activated sludge seed, the suspended
B.O. Mansell, E.D. Schroeder / Water Research 36 (2002) 4683–4690
enrichment culture at 30 and 60 days, and the bioreactor culture at 104 days (biofilm and suspension) were sampled and analyzed using the 16S and 23S rRNAtargeted probes listed in Table 2. To our knowledge, none of the probes listed in Table 2 have been used for the analysis of a hydrogenotrophic denitrification culture.
5. Results and discussion Reactor effluent nitrate concentrations and corresponding percent removals for experiments 1–5 are shown in Table 3. Throughout the operation of the system the denitrification culture nitrate concentration remained negligible. In addition, nitrite concentrations in the culture and reactor effluent were always negligible (data not shown). In each of the experiments conducted the effluent nitrate concentrations were lower than the EPA MCL of 10 mg/L NO 3 -N. The average effluent nitrate concentration for the first 78 h of experiment 1 was 3.3 mg/L NO 3 -N. After this time period the effluent concentration stabilized to below 1 mg/L NO 3 -N. A similar observation was made with a heterotrophic membrane system operated previously [10]. As with the heterotrophic system, this observation may be explained by the presence of nitrate directly adjacent to the membrane and the subsequent development of an actively metabolizing biofilm on the surface of the membrane. The development of the biofilm decreased the nitrate concentration at the membrane interface and therefore increased the concentration gradient. Since the concentration gradient is the mass transfer driving force, a decrease in the effluent concentration resulted. 5.1. Confined cell reactor comparison The denitrification rate per unit area results obtained from this system were significantly higher than the results obtained in systems operated by other researchers (Table 1). The lowest value in the range (2700 mg NO 3N m2 d1) obtained in this study was approximately 2 times greater than the highest values obtained by Lemoine et al. [3] and Fuchs et al. [9]. The lower denitrification rates can be attributed to the relatively poor nitrate mass transfer characteristics of the separation materials used in the other systems. Lemoine et al. [3] used an agar/microporous membrane composite structure. The membrane material used to sandwich the agar layer had a nitrate diffusivity of 2.1 106 cm2/ s. Fuchs et al. [9] used a regenerated cellulose membrane material with a nitrate mass transfer coefficient of 0.00037 cm/s. For a quantitative comparison, the apparent nitrate diffusion coefficient and mass transfer coefficient of the
4687
membrane material used in this study were calculated using a previously developed model of the system [10]. The model is shown in Eq. (2) Dn wLe C ¼ C0 exp ; ð2Þ dm Q 3 where C is the effluent NO 3 -N concentration (g/cm ), 3 C0 the influent NO3 -N concentration (g/cm ), L the length of microporous membrane (cm), e the microporous membrane effective porosity, Dn the diffusion coefficient of NO 3 -N through microporous membrane pore area (cm2/s), dm the microporous membrane thickness (cm), w the width of microporous membrane (cm), Q the volumetric flowrate of water being treated (cm3/s). The values of dm ; w; L; and e were equal to 80 mm, 2 cm, 20 cm, and 0.50, respectively. The coefficients were calculated by fitting data from experiments 1–3 to the linear form of the model shown in Eq. (3) where Kn ¼ Dn =dm ¼ mass transfer coefficient (cm/s) and A ¼ hLe=membrane pore area (cm2). A ð3Þ ln C ¼ Kn þ ln C0 : Q
The mass transfer coefficient ðKn Þ was determined to be 0.0006 cm/s ðR2 ¼ 0:9961Þ which is 62% higher than the nitrate mass transfer coefficient of the regenerated cellulose membrane material used by Fuchs et al. [9]. The diffusivity (Dn) was subsequently calculated to be 4.8 106 cm2/s which is 128% higher than the membrane material used by Lemoine et al. [3]. 5.2. Effluent microbial quality Based on the HPC and turbidity data presented in Table 4, the microporous membrane used in this study was an effective barrier for preventing microbial contamination of the reactor effluent. The average HPC was 9 (73.5) CFU/mL and the average turbidity was 0.4 NTU. The HPC average was significantly lower than the requirement of 500 CFU/mL stipulated by the USEPA Surface Water Treatment Rule.
Table 4 Effluent turbidity and heterotrophic plate count data Operation time, days
Turbidity, NTU
HPCa, CFU/mL
12 20 32 47 55
0.4 0.3 0.4 0.5 0.4
14 11 7 5 8
a
Heterotrophic plate count (201C for 7 days). Average=9 (73.5) CFU/mL.
B.O. Mansell, E.D. Schroeder / Water Research 36 (2002) 4683–4690
4688
Table 5 Results of hydrogenotrophic culture analysis Probe
EUB338 ALF1b BET42a GAM42a BONE23a BTWO23a
Activated sludge
80.078.5 23.277.5 21.778.7 14.574.0 — —
% Of total DAPI count 30-Day enrichment
60-Day enrichment
104-Day bioreactor biofilm/suspension
91.874.6 0.0 89.977.2 0.0 — —
94.573.6 0.0 87.572.8 0.0 — —
92.374.3/94.473.1 0.0 91.375.8/93.073.4 0.0 0.0 0.0
5.3. Culture analysis Community structure results for the activated sludge seed, 30- and 60-day suspended enrichment cultures, and the bioreactor culture are presented in Table 5. Approximately 81% of the organisms in the activated sludge culture hybridized with the probe EUB338 indicating that that the large majority of the organisms were bacteria. The activated sludge community structure was complex consisting of organisms from the a (23.277.5%), b (21.778.7%), and g (14.574.0%) subclasses of Proteobacteria. A population shift occurred from the activated sludge seed to the 30-day enrichment culture. The number of organisms hybridizing to probe EUB338 increased from 80.8% (78.5) to 91.8% (74.6). This increase was most likely due to the selective growth environment and the related selection for gram-negative bacteria as opposed to eukaryotic cells that comprised a fraction of the activated sludge seed. The enrichment process selected for a significantly less diverse culture of bacteria belonging to the b subclass of Proteobacteria (89.977.2%). The 60-day enrichment community structure remained very similar with 87.572.8% and 94.573.6% of the bacteria hybridizing to the BET42a and EUB338 probes, respectively. The community structure did not change significantly after inoculation into the membrane bioreactor. The second objective of the culture analysis was to determine the presence or absence of the identified hydrogenotrophic denitrifiers listed in Table 6 [6,28–30]. Three of the organisms (Azospirillum brasilence, Rhizobium japonicum, and Paracoccus denitrificans) belong to the a subclass of Proteobacteria. The remaining four organisms (Hydrogenophaga flava, Hydrogenophaga pseudopflava, Hydrogenophaga taeniospiralis, and Ralstonia eutropha) belong to the b subclass of Proteobacteria; more specifically the b-2 subgroup. As discussed above, the culture developed in this study was successfully hybridized with the BET42a probe only. This eliminates A. brasilence, R. japonicum, and P. denitrificans as potential members of the community. After
Table 6 Hydrogen-oxidizing organisms capable of denitrification A. brasilence H. flavaa H. pseudopflavaa H. taeniospiralisa a b
P. denitrificans R. eutrophab R. japonicum
Formerly Pseudomonas. Formerly Alcaligenes eutrophus.
hybridization with the BET42a probe the culture was analyzed with the probes BONE23a and BTWO23a to determine which subgroup the organisms may belong to. It is important to note that the BTWO23a probe was originally designed only as a competitor for the BONE23a probe and is therefore not very specific. In addition to b-2 subgroup organisms the probe will also hybridize with some members of the b-3 subgroup and some g subclass organisms [31]. As shown in Table 5, neither of the probes hybridized with organisms in the culture. These results indicate that the organisms may belong to the b-3 subgroup. However, none of the organisms present in the culture have been directly identified as known hydrogenotrophic denitrifiers. 5.4. Summary A bench-scale hydrogenotrophic denitrification system has been successfully operated. The main objective for operating the system was to study solutions to two problems associated with conventional denitrification systems: (1) the problem of the direct contact between the denitrifying organisms and the product water and (2) the addition of an organic energy source and the associated problem of product water contamination. Important results from this research are as follows. *
Removal efficiencies ranging from 96% to 92% were achieved at influent concentrations ranging from 20 to 40 mg/L NO 3 -N. In experiments with influent concentrations of 20, 30, and 40 mg/L NO 3 -N the
B.O. Mansell, E.D. Schroeder / Water Research 36 (2002) 4683–4690
*
*
*
effluent concentrations were 0.87, 1.8, and 3.2 mg/L NO 3 -N, respectively. Each of these effluent concentrations were well below the US EPA MCL of 10 mg/ L NO 3 -N. The flux values achieved in this study were 2.7–5.3 g 2 1 NO d . These values were significantly 3 -N m higher than the most promising results obtained in other confined cell reactor systems operated by other researchers. The microporous membrane used in this study served as an effective barrier for preventing microbial contamination of the product water. The effluent HPC was 9 (73.5) CFU/mL which is significantly less than the requirement of 500 CFU/mL stipulated by the US EPA Surface Water Treatment Rule. The hydrogenotrophic culture was analyzed using available 16S and 23S rRNA-targeted oligonucleotide probes. It was determined that the enrichment process selected for organisms belonging to b subclass of Proteobacteria. Further analysis of the hydrogenotrophic culture indicated that the organisms may belong to the b-3 subgroup of Proteobacteria and have yet to be identified as hydrogenotrophic denitrifiers.
Acknowledgements This work was supported by the US Environmental Protection Agency through a STAR fellowship to Bruce O. Mansell and by Timberline Instruments, Inc., Boulder, CO.
References [1] Kapoor A, Viraraghavan T. Nitrate removal from drinking water—review. J Environ Eng 1997;123(4):371–80. [2] Mateju V, Cizinska S, Krejci J, Janoch T. Biological water denitrification—a review. Enzyme Microb Technol 1992;14:170–83. [3] Lemoine D, Jouenne T, Junter GA. Biological denitrification of water in a two-chambered immobilized-cell bioreactor. Appl Microbiol Biotechnol 1991;36:257–64. [4] Mattiasson B, Ramstorp M, Nilsson I, Hahn-H.agerdal B. Comparison of the performance of a hollow-fiber microbe reactor with a reactor containing alginate entrapped cells. Biotechnol Lett 1981;3(10):561–6. . L, Molin N, Mosbach K. [5] Nilsson I, Olson S, H.aggstrom Denitrification of water using immobilized Pseudomonas denitrificans cells. Eur J Appl Microbiol Biotechnol 1980;10:261–73. [6] Neal JL, Allen GC, Morse RD, Wolf DD. Nitrate, nitrite, nitrous oxide and oxygen-dependent hydrogen uptake by Rhizobium. FEMS Microbiol Lett 1983;17:335–8. [7] Reising AR, Schroeder ED. Denitrification incorporating microporous membranes. J Environ Eng 1996;122(7):599– 604.
4689
[8] Chang CC, Tseng SK, Huang HK. Hydrogenotrophic denitrification with immobilized Alcaligenes eutrophus for drinking water treatment. Bioresource Technol 1999;69:53–8. [9] Fuchs W, Schatzmayr G, Braun R. Nitrate removal from drinking water using a membrane-fixed biofilm reactor. Appl Microbiol Biotechnol 1997;48:267–74. [10] Mansell BO, Schroeder ED. Biological denitrification in a continuous flow membrane reactor. Water Res, 1999;33(8):1845–50. [11] Mcleaf PR, Schroeder ED. Denitrification using a membrane immobilized biofilm. J Am Water Works Assoc 1995;87(3):77–86. [12] Kurt M, Dunn IJ, Bourne JR. Biological denitrification of drinking water using autotrophic organisms with H2 in a fluidized-bed reactor. Biotechnol Bioeng 1987;29:493–501. [13] Dries D, Liessens J, Verstraete W, Stevens P, de Vos P, de Ley J. Nitrate removal from drinking water by means of hydrogenotrophic denitrifiers in a polyurethane carrier reactor. Water Supply 1987;6:181–92. [14] Gross H, Schnoor G, Rutten P. Biological denitrification process with hydrogen-oxidizing bacteria for drinking water treatment. Water Supply 1987;6:193–8. [15] Ridgway HF, Davidson MS. Biological denitrification research project. Phase 1 report, Cooperative effort of the Orange County Water District, the Metropolitan Water District of Southern California, and the Santa Ana Watershed Project Authority, 1991. [16] Gantzer CJ. Membrane dissolution of hydrogen for biological nitrate removal. Proceedings of the 68th Annual Conference and Exhibition, Water Environment Federation, Research Symposium, Miami Beach, FL, 1995. p. 49– 60. [17] Ergas SJ, Reuss AF. Hydrogenotrophic denitrification of drinking water using a hollow-fiber membrane reactor. J Water SRT—Aqua 2001;50:161–71. [18] Ho CM, Tseng SK, Chang YJ. Autotrophic denitrification via a novel membrane-attached biofilm reactor. Lett Appl Microbiol 2001;33(3):201–5. [19] Lee K-C, Rittman BE. A novel hollow-fiber membrane biofilm reactor for autohydrogenotrophic denitrification of drinking water. Water Sci Technol 2000;41:219–26. [20] McCarty PL. Stoichiometry of biological reactions. International Conference Towards a Unified Concept of Biological Waste Treatment Design. Atlanta, GA, 1972. [21] Loge FJ, Emerick RW, Thompson DE, Nelson DC, Darby JL. Development of a fluoresecent 16S rRNA oligonucleotide probe specific to the family Enterobacteriaceae. Water Environ Res 1999;71(1):75–83. [22] Amann RI, Binder BJ, Olson RJ, Chisholm SW, Devereux R, Stahl DA. Combination of 16S rRNA-targeted oligonucleotide probes with flow cytometry for analyzing mixed microbial populations. Appl Environ Microbiol 1990;56:1919–25. [23] Manz W, Amann R, Ludwig W, Wagner M, Schleifer KH. Phylogenetic oligodeoxynucleotide probes for the major subclasses of Proteobacteria: problems and solutions. System Appl Microbiol 1992;15:593–600. [24] Amann R, Snaidr J, Wagner M, Ludwig W, Schleifer K-H. In-situ visualization of high genetic diversity in a natural microbial community. J Bacteriol 1996;178:3496–500.
4690
B.O. Mansell, E.D. Schroeder / Water Research 36 (2002) 4683–4690
[25] Selenka F, Dressler R. Microbiological and chemical investigations on a biological, autotrophic denitrification plant using hydrogen as an energy source. Aqua 1990;39(2):107–16. [26] Mansell BO, DeVellis L, Schroeder ED. Technical note automated separation and conductimetric determination of inorganic nitrogen. J Environ Eng 2000;126(8):778–80. [27] Cook NE, Silverstein J, Veydovec B, de Mendonca MM, Sydney R. Feild demonstration of biological denitrification of polluted groundwater and pilot-scale field testing of biological denitrification with widely varied hydraulic loading rates. Colorado Water Resources Research Institute, Colorado State University, Fort Collins Colorado, December, 1991. [28] Aragno M, Schlegel HG. The hydrogen-oxidizing bacteria. In: Starr MP, Stolp H, Truper H, Balows A, Schlegel HG, editors. The prokaryotes: a handbook on habitats, isolation, and identification of bacteria. Berlin: Springer, 1981.
[29] Tibelius KH, Knowles R. Uptake hydrogenase activity in denitrifying Azospirillum brasilence grown anaerobically with nitrous oxide or nitrate. J Bacteriol 1984;157(1):84–8. [30] Willems A, Busse J, Goor M, Pot B, Falsen E, Jantzen E, Hoste B, Gillis M, Kersters K, Auling G, de Ley J. Hydrogenophaga, a new genus of hydrogen-oxidizing bacteria that includes Hydrogenophaga flava comb. nov. (formerly Pseudomonas flava), Hydrogenophaga palleronii (formerly Pseudomonas palleronii), Hydrogenophaga pseudoflava (formerly Pseudomonas pseudoflava and Pseudomonas carboxydoflava), and Hydrogenophaga taeniospiralis (formerly Pseudomonas taeniospiralis). Int J Syst Bacteriol 1989;39:319–33. [31] Crocetti G, Hugenholtz P, Bond P, Schuler A, Keller J, Jenkins D, Blackall L. Identification of polyphosphateaccumulating organisms and design of 16S rRNA-directed probes for their detection and quantification. Appl Environ Microbiol 2000;66:1175–82.