Methane and hydrogen sulfide emissions in UASB reactors treating domestic wastewater

The release of CH(4) and H(2)S in UASB reactors was evaluated with the aim to quantify the emissions from the liquid surfaces (three-phase separator and settler compartment) and also from the reactor's discharge hydraulic structures. The studies were carried out in two pilot- (360 L) and one demo-scale (14 m(3)) UASB reactors treating domestic wastewater. As expected, the release rates were much higher across the gas/liquid interfaces of the three-phase separators (5.4-9.7 kg CH(4) m(-2) d(-1) and 23.0-35.8 g S m(-2) d(-1)) as compared with the quiescent settler surfaces (11.0-17.8 g CH(4) m(-2) d(-1) and 0.21 to 0.37 g S m(-2) d(-1)). The decrease of dissolved methane and dissolved hydrogen sulfide was very large in the discharging hydraulic structures very close to the reactor (>60 and >80%, respectively), largely due to the loss to the atmosphere, indicating that the concentration of these compounds will probably fall to values close to zero in the near downstream structures. The emission factors due to the release of dissolved methane in the discharge structure amounted to around 0.040 g CH(4) g COD(infl)(-1) and 0.060 g CH(4) g COD(rem)(-1), representing around 60% of the methane collected in the three-phase separator.     

Hydrogen limitation--a method for controlling the performance of membrane biofilm reactor for autotrophic denitrification of wastewater.

Hydrogen-driven denitrification using the fiber membrane biofilm reactor (MBfR) was evaluated for consistent operation in tertiary wastewater treatment. The possibility of controlling the process rates, as well as biofilm parameters by supplying limited amounts of electron donor (hydrogen), was tested. Limiting the hydrogen supply proved to be efficient in controlling the biofilm growth and performance of the MBfR. Denitrification rates remained unchanged for both synthetic wastewater (SWW) and real municipal wastewater (MWW) effluent as well through the fluctuations in the substrate (NO3-N) concentration. The average denitrification rates were 0.50 (+/- 0.02) g NO3-N per day per m2 for SWW and 0.59 (+/- 0.04) g NO3-N per day per m2 for MWW. Biofilm density rather than thickness was the determining factor in substrate diffusion and biofilm sloughing, ultimately determining operating stability. Limited hydrogen supply assured constant volatile solids (VS) concentration in the biofilm. It was determined that VS/TS ratio higher than 0.25 assured stable biofilm operation. Decrease of VS/TS ratio below 0.25 led to shearing of the nonbiological outer layers of the biofilm. The values of chemical oxygen demand (COD), volatile suspended solids (VSS) and total suspended solids (TSS) in the final effluent were stable and well below wastewater effluent guidelines. Substitutions of bicarbonate with gaseous carbon dioxide as the carbon source did not affect denitrification rates despite lower than optimum pH conditions.     

Full-scale applications for both COD and nutrient removal in a CIRCOX airlift reactor.

The airlift reactor technology has been successfully applied at full scale for both COD and nitrogen removal. In this study, the results of the biofilm development and biological performance of two full scale reactors are discussed. At Paulaner Brewery in Munich, the airlift reactor was applied for COD and ammonia removal of anaerobically treated wastewater. In the other case the airlift reactor was applied as a pretreatment of nitrogen removal by the Anammox process. Water from a Tannery company in Lichtenvoorde in the Netherlands, The Hulshof Royal Dutch Tanneries, was pretreated anaerobically for COD removal and aerobically to remove the sulphides as sulphur. In an airlift reactor the ammonia was partially oxidised to nitrite. In both cases the granular biomass developed well; the concentrations amounted to 250 microl/L and 500 ml/L respectively. In the first case, 4 kg/m(3)/day of COD was removed, the soluble concentration of COD was less than 250 mg/L. The nitrification to nitrate was nearly complete and amounted to 0.5 kg NH4-N/m(3)/day. In the second application, 50% of the ammonia (on average 0.45 kg N/m3/d) was nitrified to nitrite. This process was easily controlled by regulating the amount of air according to the nitrite and ammonia concentrations in the effluent. It can be concluded that in both cases the particular processes were very stable and easy to operate.     

Nutrients removal in hybrid fluidised bed bioreactors operated with aeration cycles.

Abstract Two hybrid fluidised bed reactors filled with sepiolite and granular activated carbon (GAC) were operated with short cycled aeration for removing organic matter, total nitrogen and phosphorous, respectively. Both reactors were continuously operated with synthetic and/or industrial wastewater containing 350-500 mg COD/L, 110-130 mg NKT/L, 90-100 mg NH3-N/L and 12-15 mg P/L for 8 months. The reactor filled with sepiolite, treating only synthetic wastewater, removed COD, ammonia, total nitrogen and phosphorous up to 88, 91, 55 and 80% with a hydraulic retention time (HRT) of 10 h, respectively. These efficiencies correspond to removal rates of 0.95 kgCODm(-3)d(-1) and 0.16 kg total N m(-3)d(-1).The reactor filled with GAC was operated for 4 months with synthetic wastewater and 4 months with industrial wastewater, removing 98% of COD, 96% of ammonia, and 66% of total nitrogen, with an HRT of 13.6 h. No significant phosphorous removing activity was observed in this reactor. Microbial communities growing with both reactors were followed using polymerase chain reaction (PCR) and denaturing gradient gel electrophoresis (DGGE) techniques. The microbial fingerprints, i.e. DGGE profiles, indicated that biological communities in both reactors were stable along the operational period even when the operating conditions were changed.     

Optimizing and modelling nitrogen removal in a new configuration of the moving-bed biofilm reactor process.

A new configuration of the moving-bed biofilm reactor process with pre-denitrification and nitrification was investigated in a pilot plant, which is fed with urban raw wastewater, the primary settler is located between the anoxic and the aerobic reactors, and primary sludge is recycled to the anoxic reactor as a hybrid pre-denitrification. The carriers used in the experiments are made of high-density polyethylene, with a diameter of 10 mm and a specific surface area of 400 m(2)/m(3). The new process was compared with conventional pre-denitrification-nitrification using in-series reactors fed with settled wastewater. The new configuration achieved an increase of 45% for the denitrification rate and of 30% for the nitrification rate when compared with conventional configuration. These results were analysed in light of the calibration study of the mixed-culture biofilm (MCB) model and simulations in AQUASIM 2.1 platform. Regarding denitrification, the high values obtained in the new configuration were attributed to a higher removal of the slowly biodegradable substrate (Xs) in the anoxic reactor due to the use of raw wastewater and sludge recycle. Accordingly, the amounts of heterotrophic biomass (XH) and Xs obtained in simulations were higher in both the biofilm and the bulk liquid. Regarding nitrification, the higher values were attributed to a lower removal of Xs in the aerobic reactors and accordingly, a lower accumulation of heterotrophic biomass in the biofilm was found in the simulations.     

Low-temperature post-treatment of anaerobically treated-sewage in anaerobic filter with cationic-polymer addition.

The post-treatment of domestic sewage pretreated in a 6 m3 UASB was investigated in two high-rate anaerobic filter (AF) reactors operated in parallel. The difference between the two AF reactors was only the addition of cationic polymer to the second reactor (AF + P). The reactors were operated at low temperatures, ranged between 13 and 20 degrees C. The media in each AF reactor consisted of vertical sheets of reticulated-polyurethane foam (RPF) with knobs. The results demonstrated that the AF + P reactor (HRT = 3 h) with cationic polymer addition (2 mg/L) was an efficient system for post-treatment. The removal efficiencies for total, suspended, colloidal and dissolved COD were, respectively, 41, 86 and 76 and 12% in the AF + P reactor and they were, respectively, 80, 97, 77 and 66% in the UASB + (AF + P) system. The removal of total, suspended and colloidal COD in the UASB + (AF + P) system were significantly higher than those achieved in the UASB + AF system. As hardly any nutrient was removed in the UASB + (AF + P) system, the effluent after pathogen removal is a valuable product for irrigation and fertilisation to close the water and nutrients cycle.     

Measurement of polysaccharides and proteins in biofilm extracellular polymers

The structure of biofilm extracellular polymers (ECPs) was studied by measuring their polysaccharide and protein spatial distributions along biofilm depth. Biofilm was collected from two aerobic heterotrophic biofilm reactors, which were seeded with Sphingomonas sp. and Sphingomonas sp. plus mixed liquor, respectively, and operated under toxic organic (in this case, azo dye) degrading conditions. Complete mixing conditions in the tworeactors were verified by measuring water content, and polysaccharide and protein quantities from three vertical sampling positions over time. Experimental results showed that: (1) the biofilm water content of either reactor did not change with sample position or biofilm age, with an average biofilm water content m both reactors of 97%; (2) polysaccharides and proteins in the ECPs did not change with sample position; (3) the profiles of polysaccharides and proteins along the biofilm depth showed a stratified biofilm structure, with their ratio (proteins/polysaccharides) being relatively stable over the depth. Oxygen and substrate transport and interactions among species were considered to be the main reasons for producing such a non-uniform biofilm structure; and (4) Sphingomonas sp. could not compete well with microorganisms derived from the mixed liquor of a wastewater treatment plant aeration basin.     

Performance of an ultra-compact biofilm reactor treating domestic and synthetic wastewaters

The performance of an ultra-compact biofilm reactor (UCBR) treating domestic wastewater (DWW) collected from a local water reclamation plant; and gradually shifting to a mono-type carbon source synthetic wastewater (SWW) combined with DDW (CWW) and finally SWW; was investigated in this study. The total COD concentrations of influent DWW and CWW/SWW were 413.6 ± 80.8 mg/L and 454.9 ± 51.3 mg/L, respectively. The UCBR was able to achieve average total COD removal efficiencies of 70 ± 10% and 80 ± 4% for DWW and SWW respectively. The total COD concentrations of the effluent of DWW and CWW/SWW were 122.5 ± 44.4 mg/L and 89.7 ± 10.3 mg/L, respectively. These observations suggested that heterotrophs in the UCBR system were able to better assimilate and remove carbon of mono-type SWW compared to diverse carbon sources such as DWW; although the influent soluble COD concentrations of the SWW were higher than those of the DWW. However, the effluent NH(4)(+)-N concentrations for both types of wastewater were rather similar, <3.0 mg/L; although the influent NH(4)(+)-N concentrations of the DWW were 1.5 times those of the SWW.     

Comparison of bioreactors with different kinds of submerged packed beds for domestic wastewater treatment

The performance of aerobic submerged packed bed reactors was studied for the treatment of domestic wastewater using different kinds of packing materials with high specific areas (760-1,200 m(2)/m(3)). The tested materials were ceramic spheres, crushed tezontle, grains of high density polyethylene (HDPE), of low density polyethylene (LDPE) and of polypropylene (PP), cubes of polyurethane (PU) and polyethylene tape (SESSIL). The bioreactors were operated in continuous regime, applying organic loads in the range of 0.8-6.0 g COD.m(-2).d(-1). The obtained specific COD removal rates were very similar in all the reactors when they were operated at organic loads up to 2.0 g COD.m(-2).d(-1), after which differences in effectiveness appeared and the best results were determined in the reactors with SESSIL, LDPE and PU. Very low TSS, O&G and turbidity were obtained in all the effluents. The NH(3)-N and TN removals were dependent on the dissolved oxygen (DO) concentration and the removals at DO of 5 mg/l were 84-99% and 61-74% respectively. The best removals were determined in the reactors with PU, SESSIL and LDPE. The reactor with tezontle had also a good performance when operated with loads up to 1.0 g TN.m(-2).d(-1). The best phosphate removals (38-49%) were obtained in the reactors with PU, tezontle, ceramic sheres and SESSIL.     

Biological granulated activated carbon fluidized bed reactor for atrazine remediation.

To show that an adsorbing biofilm carrier (GAC) can be advantageous for atrazine bioremediation over a non-adsorbing carrier, fluidized bed (FB) reactors were operated under atrazine limiting concentrations using Pseudomonas sp. strain ADP as the atrazine degrading bacteria. The following interrelated subjects were investigated: 1) atrazine adsorption to GAC under conditions of atrazine partial penetration in the biofilm, 2) differences in atrazine degradation rates and 3) stability of atrazine biodegradation under non-sterile anoxic conditions in the GAC reactor versus a reactor with a non-adsorbing biofilm carrier. Results from batch adsorption tests together with modeling best described the biofilm as patchy in nature with covered and non-biofilm covered areas. Under conditions of atrazine partial penetration in the biofilm, atrazine adsorption occurs in the non-covered areas and is consequently desorbed at the base of the biofilm substantially increasing the active biofilm surface area. The double flux of atrazine to the biofilm in the GAC reactor results in lower effluent atrazine concentrations as compared to a FB reactor with a non-adsorbing carrier. Moreover, under non-sterile denitrification conditions, atrazine degradation stability was found to be much higher (several months) using GAC as a biofilm carrier while non-adsorbing carrier reactors showed sharp deterioration within 30 days due to contamination of non-atrazine degrading bacteria.     

Combined partial nitritation/Anammox system for treatment of digester supernatant.

One-year (2004) comprehensive investigations in a semi-industrial pilot plant (5 m3) were carried out with the aim of assessing the influence of operational parameters on the partial nitritation/Anammox system performance. In the system designed as a moving-bed biofilm reactor, the influent nitrogen load to the Anammox reactor was progressively increased and a stable Anammox bacterial culture was obtained. Interaction between subsequent aerobic and anaerobic conditions in the partial nitritation and Anammox reactors, respectively, granted conditions to remove nitrogen through the nitrite route. It implies that the oxygen supply can be limited to a high extent. A control strategy for the partial nitritation step relied on concomitant adjustment of the air supply with a variable influent nitrogen load, which can be monitored by both pH and conductivity measurements. In the Anammox reactor, an influent nitrite-to-ammonium ratio plays a vital role in obtaining efficient nitrogen removal. During the 1-year experimental period, the Anammox reactor was operated steadily and average nitrogen removal efficiency was 84% with 97% as the maximum value.     

Autogenerative high pressure digestion: anaerobic digestion and biogas upgrading in a single step reactor system

Conventional anaerobic digestion is a widely applied technology to produce biogas from organic wastes and residues. The biogas calorific value depends on the CH, content which generally ranges between 55 and 65%. Biogas upgrading to so-called 'green gas', with natural gas quality, generally proceeds with add-on technologies, applicable only for biogas flows > 100 m3/h. In the concept of autogenerative high pressure digestion (AHPD), methanogenic biomass builds up pressure inside the reactor. Since CO2 has a higher solubility than CH4, it will proportion more to the liquid phase at higher pressures. Therefore, AHPD biogas is characterised by a high CH4 content, reaching equilibrium values between 90 and 95% at a pressure of 3-90 bar. In addition, also H2S and NH3 are theoretically more soluble in the bulk liquid than CO2. Moreover, the water content of the already compressed biogas is calculated to have a dew point <--10 degrees C. Ideally, high-quality biogas can be directly used for electricity and heat generation, or injected in a local natural gas distribution net. In the present study, using sodium acetate as substrate and anaerobic granular sludge as inoculum, batch-fed reactors showed a pressure increase up to 90 bars, the maximum allowable value for our used reactors. However, the specific methanogenic activity (SMA) of the sludge decreased on average by 30% compared to digestion at ambient pressure (1 bar). Other results show no effect of pressure exposure on the SMA assessed under atmospheric conditions. These first results show that the proposed AHPD process is a highly promising technology for anaerobic digestion and biogas upgrading in a single step reactor system.     

Treatment of chemical-pharmaceutical wastewater in packed bed anaerobic reactors.

Biological degradation in packed bed anaerobic mesophilic reactors with five different support materials was studied for the treatment of chemical-pharmaceutical wastewater with high COD (23-31 g/L), which contains toxic organic compounds. Experimental up-flow bio-filters were operated at different organic loads for a two-year period. Removals of 80-98% were obtained in the reactors with sand, anthracite and black tezontle, but at relatively low organic loads, less than 3.6 kg m(-3)d(-1). The reactor with granular activated carbon (GAC) had a better performance; efficiencies higher than 95% were obtained at loads up to 17 kg m(-3)d(-1) and higher than 80% with loads up to 26 kg m(-3)d(-1). Second in performance was the reactor with red tezontle which allows COD removals higher than 80% with loads up to 6 kg m(-3)d(-1). The use of GAC as support material allows greater biodegradation rates than the rest of the materials and it makes the process more resistant to organic load increases, inhibition effects and toxicity. Methanogenic activity was inhibited at loads higher than 21.9 kg m(-3)d(-1) in the GAC-reactor and at loads higher than 3.6 kg m(-3)d(-1) in the rest of the reactors. At loads lower than the previously mentioned, high methane production yield was obtained, 0.32-0.35 m3CH4/kg CODremoved.     

A laboratory assessment of the impact of brewery wastewater discharge on sulfide and methane production in a sewer

The impact of brewery wastewater discharge on sulfide and methane production in a sewer was assessed. Experiments were carried out on laboratory scale sewer reactors consisting of both an experimental and a control reactor. The control reactor was intermittently fed with real fresh sewage while the experimental reactor was fed with a mixture of brewery and domestic wastewater at two different proportions (10 and 25% v/v). 10% v/v discharge of brewery wastewater increased the H2S and CH4 production rates in the sewer reactor by 40% and 30%, respectively. When the brewery wastewater fraction was increased to 25% v/v, the H2S production rate of the experimental reactor decreased to the level of the control reactor. In contrast, the CH4 production rate maintained at a level that was 30% higher than that in the control reactor. These results indicate that the discharge of brewery wastewater into sewers can give negative impacts in relation to odour and corrosion management of the systems and will increase the greenhouse gas emissions from sewers. The study also reveals that the impact of trade waste on the biological reactions in sewers is complex, and requires careful experimental assessment in each case.     

Technical and economical analysis of the conversion of a full-scale scrubber to a biotrickling filter for odor control.

The present paper evaluates the technical and economical feasibility of converting wet chemical scrubbers to biotrickling filters for H2S control at the Orange County Sanitation District (OCSD), California. Results of 8 months of continuous operation of a biotrickling filter treating 16,000 m3 h(-1) of foul air are analyzed. The reactor was operated at a gas contact time of 1.6 to 2.2 seconds reaching H2S elimination capacities up to 105-110 g H2S m3 h(-1), consistently maintaining outlet concentrations well below the regulatory limits (24 h average of 1 ppmv) and demonstrating to be very robust against temporary changes. Also, a cost-benefit analysis of the conversion was performed. Savings from chemicals, energy and water usage compared to a chemical scrubber operated in parallel to the biotrickling filter throughout the project indicated that the payback time of the conversion was about 1.3 years. Cost savings ranged between 40,000 US dollars per year, per scrubber. While the above number may be specific to OCSD conditions, the cost analysis suggests that there is a significant benefit of converting chemical scrubbers to biotrickling filters over a wide range of operating conditions.     

Pumped flow biofilm reactors (PFBR) for treating municipal wastewater

A novel laboratory bench-scale sequencing batch biofilm reactor (SBBR) system was developed for the treatment of synthetic domestic strength wastewater, comprising two side-by-side 18 l reactor tanks, each containing a plastic biofilm media module. Aerobic and anoxic conditions in the biofilms were effected by intermittent alternate pumping of wastewater between the two reactors. With a media surface area loading rate of 4.2 g chemical oxygen demand (COD)/m2.d, the average influent COD, total nitrogen (TN) and ammonium-nitrogen (NH4-N) concentrations of 1021 mg/l, 97 mg/l and 54 mg/l, respectively, reduced to average effluent concentrations of 72 mg COD/l, 17.8 mg TN/l, and 5.5 mg NH4-N /l. Using a similar alternating biofilm exposure arrangement, a 16 person equivalent pilot (PE) plant was constructed at a local village treatment works to remove organic carbon from highly variable settled municipal wastewater and comprised two reactors, one positioned above the other, each containing a module of cross-flow plastic media with a surface area of 100 m2. Two different pumping sequences (PS) in the aerobic phase were examined where the average influent COD concentrations were 220 and 237 mg/l for PS1 and PS2, respectively, and the final average effluent COD was consistently less than 125 mg/l--the European Urban Wastewater Treatment Directive limit--with the best performance occurring in PS1. Nitrification was evident during both PS1 and PS2 studies. A 300 PE package treatment plant was designed based on the bench-scale and pilot-scale studies, located at a local wastewater treatment works and treated municipal influent with average COD, suspended solids (SS) and TN concentrations of 295, 183 and 15 mg/l, respectively resulting in average effluent concentrations of 67 mg COD/l, 17 mg SS/l and 9 mg TN/l. The SBBR systems performed well, and were simple to construct and operate.     

Resource recovery and nitrogen removal from piggery waste using the combined anaerobic processes.

The combined ADEPT (Anaerobic Digestion Elutriated Phased Treatment)- SHARON (Single reactor system High Ammonium Removal Over Nitrite)--ANAMMOX (Anaerobic Ammonium Oxidation) processes were operated for the purpose of resource recovery and nitrogen removal from slurry-type piggery waste. The ADEPT operated at acidogenic loading rates of 3.95 gSCOD/L-day, the SCOD elutriation rate and acid production rate were 5.3 gSCOD/L-day and 3.3 gVFAs(as COD)/L-day, respectively. VS reduction and SCOD reduction by hydrolysis were 13% and 0.19 gSCOD(prod.)/gVS(feeding), respectively. Also, the acid production rate was 0.80 gVFAs/gSCOD(production). In the methanogenic reactor, the gas production rate and methane content were 2.8 L/day (0.3 m3CH4/kgCOD(removal)STP) and 77%, respectively. With these operating condition, the removal of nitrogen and phosphorus were 94.1% as NH4-N (86.5% as TKN) and 87.3% as T-P, respectively.     

Pilot-scale two-stage process: a combination of acidogenic hydrogenesis and methanogenesis.

This study was performed to optimize both acidogenic hydrogenesis and methanogenesis, and then to develop a pilot-scale two-stage process producing not only CH4 but also H2. Firstly, acidogenic hydrogenesis of food waste was examined in pilot-scale leaching-bed reactors using dilution rate (D) as a tool to improve the environmental conditions. The maximum efficiency of 71.4% was obtained by adjusting D from 4.5 to 2.5 d(-1) depending on the state of degradation. Secondly, the wastewater from acidogenic hydrogenesis was converted to CH4 in a pilot-scale UASB reactor. The COD removal efficiency exceeded 95% up to the loading rates of 13.1 g COD/Ld, which corresponded to HRT of 0.25 d (6 h). Lastly, a pilot-scale two-stage process was devised based on a combination of acidogenic hydrogenesis and methanogenesis. Over 120 days, the pilot-scale process resulted in large VS reduction of 70.9% at the high loading rate of 12.5 kg VS/m3/d in a short SRT of 8 days.     

Biofilm development during the start-up of a sulfate-reducing down-flow fluidized bed reactor at different COD/SO4(2-) ratios and HRT

In sulfate-reducing reactors, it has been reported that the sulfate removal efficiency increases when the COD/SO4(2-) ratio is increased. The start-up of a down-flow fluidized bed reactor constitutes an important step to establish a microbial community in the biofilm able to survive under the operational bioreactor conditions in order to achieve effective removal of both sulfate and organic matter. In this work the influence of COD/SO4(2-) ratio and HRT in the development of a biofilm during reactor start-up (35 days) was studied. The reactor was inoculated with 1.6 g VSS/L of granular sludge, ground low density polyethylene was used as support material; the feed consisted of mineral medium at pH 5.5 containing 1 g COD/L (acetate:lactate, 70:30) and sodium sulfate. Four experiments were conducted at HRT of 1 or 2 days and COD/SO4(2-) ratio of 0.67 or 2.5. The results obtained indicated that a COD/SO4(2-) ratio of 2.5 and HRT 2 days allowed high sulfate and COD removal (66.1 and 69.8%, respectively), whereas maximum amount of attached biomass (1.9 g SVI/L support) and highest sulfate reducing biofilm activity (10.1 g COD-H2S/g VSS-d) was achieved at HRT of 1 day and at COD/sulfate ratios of 0.67 and 2.5, respectively, which suggests that suspended biomass also played a key role in the performance of the reactors.     

Feasibility study on the removal of nitrous compounds with a hollow-fiber membrane biofilm reactor.

The objective of this study was to develop an integrated nitrogen treatment system using autotrophic organisms. A treatment system consists of an aerobic hollow-fiber membrane biofilm reactor (HfMBR) and anaerobic HfMBR. In the aerobic HfMBR, a mixture gas of air and O2 was supplied through the fibers for nitrification. Denitrification occurred in the anaerobic HfMBR using H2 as the electron donor. The treatment system was continuously operated for 190 days. NH4-N removal efficiencies ranging from 95% to 97% were achieved at NH4-N concentrations of influent ranging from 50 to 100 mg N/L. When glucose was added to the influent, the simultaneous nitrification and denitrification occurred in the aerobic HfMBR, and nitrogen removal rates were changed according to the COD/NH4-N ratio of influent. In the anaerobic HfMBR, autotrophic denitrification using H2 occurred and the removal rates achieved in this study were 23-58 mg N/m2 d. In this study, the achieved removal efficiency was lower than other study findings; however, the result suggested that this hybrid HfMBR system can be used effectively for nitrogen removal in oligotrophic water.