Nitrogen Removal from Wastewater by Coupling Anammox and

Sep 13, 2013 - ABSTRACT: This work demonstrates, for the first time, the feasibility of nitrogen removal by using the synergy of anammox and denitrify...
0 downloads 0 Views 856KB Size
Download PDF
Article pubs.acs.org/est

Nitrogen Removal from Wastewater by Coupling Anammox and Methane-Dependent Denitrification in a Membrane Biofilm Reactor Ying Shi,† Shihu Hu,† Juqing Lou,†,‡ Peili Lu,†,§ Jurg Keller,† and Zhiguo Yuan*,† †

Advanced Water Management Centre (AWMC), The University of Queensland, St Lucia, Brisbane, Queensland, 4072, Australia School of Environmental Science and Engineering, Zhejiang Gongshang University, Hangzhou 310012, China § Department of Environmental Science, Chonqing University, Chongqing 400044, China ‡

S Supporting Information *

ABSTRACT: This work demonstrates, for the first time, the feasibility of nitrogen removal by using the synergy of anammox and denitrifying anaerobic methane oxidation (DAMO) microorganisms in a membrane biofilm reactor (MBfR). The reactor was fed with synthetic wastewater containing nitrate and ammonium. Methane was delivered from the interior of hollow fibres in the MBfR to the biofilm that grew on the fiber’s outer wall. After 24 months of operation, the system achieved a nitrate and an ammonium removal rate of about 190 mgN L−1 d−1 (or 86 mgN m−2 d−1, with m2 referring to biofilm surface area) and 60 mgN L−1 d−1 (27 mgN m−2 d−1), respectively. No nitrite accumulation was observed. Fluorescence in situ hybridization (FISH) analysis indicated that DAMO bacteria (20−30%), DAMO archaea (20−30%) and anammox bacteria (20−30%) jointly dominated the microbial community. Based on the known metabolism of these microorganisms, mass balance, and isotope studies, we hypothesize that DAMO archaea converted nitrate, both externally fed and produced by anammox, to nitrite, with methane as the electron donor. Anammox and DAMO bacteria jointly removed the nitrite produced, with ammonium and methane as the electron donor, respectively. The process could potentially be used for anaerobic nitrogen removal from wastewater streams containing ammonium and nitrate/nitrite.

■

the dismutation of nitric oxide to oxygen and nitrogen.13 With a similar approach, Haroon et al. confirmed the role of an archaeron, with a proposed name of “Candidatus Methanoperedens nitroreducens”, in the DAMO process.12 M. nitroreducens reduces nitrate to nitrite, using electrons derived from methane. In contrast to M. oxyfera, which employs pathways for aerobic methane oxidation, M. nitroreducens oxidizes methane via reverse methanogenesis, a pathway that was previously hypothesized but unproven until very recently.16,17 These recent discoveries provide the opportunity of achieving nitrate/ nitrite removal from wastewater with methane as the electron donor. Nitrogen removal from wastewater using cocultures of anammox and DAMO microorganisms has been investigated in a laboratory study recently. Luesken et al.18 reported an enriched coculture of anammox and DAMO bacteria fed with nitrite and ammonium in a sequential batch reactor (SBR) that was continuously flushed with CH4. The conversion rate of nitrite achieved was 100 mgN L−1 d−1, with anammox bacteria being responsible for 77% of the nitrite consumption. However, if ammonium was fed in excess, anammox bacteria would probably outcompete DAMO bacteria because they were

INTRODUCTION The requirement of achieving high-levels of nitrogen removal from wastewater with a minimized carbon footprint is a serious challenge for the water industry. The anammox process is an energy-efficient nitrogen removal process, and does not require organic carbon, in contrast to the conventional denitrification process.1−4 It has been successfully implemented at full-scale.5,6 However, this process requires a specific nitrite to ammonium ratio (a molar ratio of 1.32 to 1),7 and does not remove nitrate either originally present in the wastewater or produced by the anammox reaction.8,9 Methane is an inexpensive, widely available carbon source, which could also be generated onsite at a wastewater treatment plant (WWTP) through anaerobic sludge digestion.10 Recent studies have shown that methane could be oxidized anaerobically, providing electrons for denitrification,11−14 eluding the requirement for expensive electron donors such as methanol or ethanol. These processes are referred to as the denitrifying anaerobic methane oxidation (DAMO) processes with the organisms involved called DAMO organisms.15 The reported microorganisms include a bacterial group affiliated to the candidate division NC10 and an archaeal group distantly related to anaerobic methanotrophic archaea (ANME).11 By assembling the complete genome of the bacterium, named “Candidatus Methylomirabilis oxyfera”, Ettwig et al. revealed that M. oxyfera reduces nitrite to nitric oxide and then achieves methane oxidation using the in situ produced oxygen through © XXXX American Chemical Society

Received: June 24, 2013 Revised: September 10, 2013 Accepted: September 13, 2013

A

dx.doi.org/10.1021/es402775z | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

competing for nitrite as electron acceptor.18 Furthermore, nitrate produced by anammox was left in the effluent without removal.18 With ammonium and nitrate supplied in the feed, Haroon et al.12 enriched a coculture of DAMO archaea and anammox bacteria in a bioreactor where CH4 was intermittently flushed into the headspace. The nitrate and ammonium removal rates were both about 15 mgN L−1 d−1. The archaeal DAMO reduced nitrate to nitrite, which was then removed by the anammox bacteria. Despite of these successful attempts, several issues need to be addressed before the DAMO process can be used for wastewater treatment in practice. First, methane is a flammable gas with a low aqueous solubility. In previous laboratory studies, methane was directly sparged into the reactor, which would require large amounts of energy and also possibly result in the release of methane in industrial applications.19 Gas diffusive membranes, which have been used to deliver hydrogen gas for denitrification,20−22 could offer a solution to the safe and efficient supply of methane. Since methane is directly supplied through a bubbleless membrane, its transfer can be enhanced23,24 and controlled by adjusting pressure.21 Second, the slow growth of both anammox and DAMO microorganisms requires a high biomass retention which could be ensured by culturing in biofilms.9 Like in the case of hydrogen-supported denitrification,20−22 the membrane surface would also support the development of a biofilm retaining slow-growing DAMO and anammox organisms. The aim of this study is to investigate the feasibility of simultaneous nitrate and ammonium removal by coupling the anammox and DAMO processes in a membrane biofilm reactor. To this end, an MBfR was inoculated with a coculture of DAMO and anammox microorganisms, and fed with methane from inside of hollow fibres, and with nitrate and ammonium directly in the liquid phase outside of the hollow fibres. Concentrations of nitrate and ammonium were measured to monitor the performance of the MBfR. Batch tests with various wastewater composition (NO3− + NH4+, NO2− + NH4+, NO2− only, and NO3− only) were performed in the middle and at the end of the operation to verify the reactions, and to determine the capability of the system to treat wastewater with different nitrogen compositions. The presence of anammox and DAMO microorganisms was confirmed with FISH and their activities investigated through 15N labeling batch tests.

Figure 1. Schematic diagram of the membrane biofilm reactor. The dash lines show the direction of liquid flows. (1), gas sampling point; (2), pH probe; (3), liquid sampling point; (4), pressure gauge.

(∼20%). NC10 bacteria, one of the key known DAMO organisms, were not detected with FISH. However, their presence at low abundance could not be ruled out. See the SI for further details of the inoculum. Operational Conditions and Reactor Monitoring. The reactor was operated continuously for about 24 months, consisting of two phases namely a startup phase and a sequencing batch reactor (SBR) phase. At the beginning of the startup phase (Day 0 to 290), 300 mL inoculum and 300 mL medium were added into the system. To prevent biomass from being washed out, no medium was changed during the startup phase. However, concentrated nitrate (80 gN L−1) and ammonium (48 gN L−1) solutions were fed weekly, resulting in theoretical nitrate and ammonium concentrations both at 200 mgN L−1 after each feeding. These initial concentrations, which were also applied to the bioreactor where the inoculum was taken, were chosen to avoid substrate limitation over the week and also to reduce the risk of substrate inhibition. The recirculation rate of the liquid was maintained at 50 mL min−1. The gas was manually refilled via a gas regulator (BOC) twice per week, with the inner hollow fibres repressurized to 1.5 atm when the pressure dropped to 1.2 atm. Liquid samples were taken daily to measure the concentration of nitrate, nitrite and ammonium. The ammonium and nitrate removal rates in this period were determined based on the measured profiles between feed. Volumetric conversion rates were calculated by considering the liquid volume of the MBfR (450 mL) only (i.e., without considering the gas volume in the MBfR or the liquid volume of the overflow bottle). The areal-specific rate of biofilm was calculated by dividing the mass conversation rates by the membrane surface area (1 m2). From Day 290 onward, the MBfR was operated as an SBR consisting of two stages, namely Stage 1 (Day 290 to Day 420) and Stage 2 (Day 471 to Day 730). In both stages, the SBR was operated with a cycle time of one day. At the beginning of each cycle, the recirculation pump was stopped for 5 min, where 150 mL fresh medium was fed by a peristaltic pump and 150 mL of effluent was discharged from the overflow bottle. The feeding regime resulted in a hydraulic retention time (HRT) of 3 days in the SBR (without considering the liquid volume of the overflow bottle). In this phase, the inner hollow fibres were connected to the gas cylinder at all time, with the pressure

■

MATERIALS AND METHODS Reactor Configuration. A membrane biofilm reactor (Figure 1) was set up for the study. The total volume of the membrane module was 1150 mL, which included 400 mL of hollow fiber materials, 300 mL interior space for gas supply, and 450 mL external space outside the fibres for liquid. The total surface area of the membrane was 1 m2. The interior of the hollow fibres was connected to a feeding gas cylinder. The liquid was recirculated through an overflow bottle (150 mL in liquid volume) by a peristaltic pump (Masterflex, USA) to provide mixing. Further details of the setup are described in the Supporting Information (SI). Inoculum. The MBfR was seeded with sludge taken from a bioreactor, where both DAMO and anammox microorganisms were enriched in the suspended phase. Microbiological analysis with FISH revealed that the inoculum was dominated by archaea closely related to ANME-2d (∼40%) and anammox bacteria (∼40%), but also contained other microorganisms B

dx.doi.org/10.1021/es402775z | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Figure 2. (a) Nitrate and ammonium removal rates during the 24 month study. (b) Nitrate and ammonium concentrations in the influent and effluent during the SBR phase. The gray areas indicate batch tests periods.

Concentrated stock solutions were then added, giving rise to an initial nitrate, nitrite and/or ammonium concentration between 10 and 50 in Batch Test B-1 and 100−200 mgN L−1 in Batch Test B-2. Methane was supplied through the hollow fibres in all tests by maintaining a pressure of 1.3 atm. Each test lasted for 8 h, during which 5 liquid samples were taken to determine the consumption rate(s) of the added substrate(s). Batch Test C, lasting for 6 h, was carried out on Day 460 to investigate the effect of a higher liquid recirculation rate (1200 mL min−1) on the nitrogen removal rates. Five liquid samples were taken for chemical analysis to obtain concentration profiles of nitrate and ammonium. The high recirculation rate was avoided at other time because biofilm detachment was observed at this recirculation rate, except on Day 415 and Day 470 for biofilm sampling. Chemical and Microbial Analysis. NH4+-N, NO3−-N, NO2−-N in liquid phase and CH4 and N2 in gas phase were measured with methods previously reported by Hu et, al.15 An Agilent 7890A gas chromatograph equipped with an electron capture detector (ECD) was used to measure dissolved methane and dissolved nitrous oxide concentrations, with the latter being a possible byproduct during nitrogen conversion. On Day 415 and Day 730, the recirculation rate was temporarily raised (600 mL min−1 to 1200 mL min−1) to harvest detached biofilm for microbial community analysis. The samples were fixed, stored and hybridized as described in Ettwig et al.25 The probes used are described in SI. Determination of Reaction Rates. Based on the NH4+, NO 3− and CH 4 consumption data and the microbial community data, the following reactions were hypothesized (biomass production neglected): Nitrite reduction by anammox (reaction rate r1):7

maintained at 1.3 atm with a gas regulator. The recirculating rate was increased to 600 mL min−1 to improve the mixing conditions in the reactor. Effluent samples were taken daily at the end of each cycle to monitor the nitrate, nitrite and ammonium concentrations. In Stage 2, the ammonium and nitrate concentrations in the feed were progressively increased from the initial level of 200 mgN L−1, with the improvement of the MBfR performance (see the Results section). As a proof-ofconcept study, the feed composition did not mimic a particular wastewater. Batch Tests. Mass and electron balance tests were undertaken to verify the reactions in the system (Batch Test A-1 at the end of Stage 1 and A-2 at the end of Stage 2). To measure the consumption of methane, the MBfR was disconnected from the gas cylinder to stop methane supply. Freshly prepared medium was sparged with the mixed gas for 30 min at a flow rate of 500 mL min−1, which ensured that the medium was saturated with methane (∼21 mg L−1). Then both the interior of the hollow fibres and the liquid phase of the MBfR were connected to the overflow bottle and filled with this methane saturated liquid medium. The liquid phase was continuously circulated during the test. The auto-overflow point was locked so that the headspace in the overflow bottle became the only gas phase of the whole system and the consumption of methane could be measured. Batch Test A-1 was operated in the batch mode for 30 h after a 12 h equilibrating period, with eight gas and eight liquid samples taken over this period. Because of the rate increased significantly in Stage 2, Batch Test A-2 was operated for 3 h only, with seven gas and seven liquid samples taken in this period. Ammonium, nitrate, nitrite concentrations, and nitrogen and methane gas partial pressures were analyzed. The methane and nitrogen concentrations in liquid phase were determined from the gas phase data using the Henry’s law. Three liquid samples were taken at the beginning, middle and end of the test to measure the dissolved N2O concentrations. In Batch Test A-2, 15N enriched nitrate (99% 15N, Sigma Aldrich) was added, along with nonlabeled nitrate, to a final 15N percentage of 17%. The concentrations of 28N2, 29N2 and 30N2 in both liquid and gas phases were continuously monitored online with membrane inlet mass spectrometer (MIMS).12 Batch Tests B-1 and B-2 were performed following Batch Tests A-1 and A-2, respectively, to investigate the capability of this system in treating wastewater containing nitrate only, nitrite only, nitrite, and ammonium, as summarized in Table 2. At the start of each test, fresh medium was fed into the reactor.

NO2− + 1/1.32NH4 + → 1.02/1.32N2 + 0.26/1.32NO3− (1)

Nitrite reduction by DAMO (reaction rate r2):

11

NO2− + 3/8CH4 + H+ → 1/2N2 + 3/8CO2 + 10/8H 2O

Nitrate reduction by DAMO (reaction rate r3):

(2) 12

NO3− + 2/8CH4 → NO2− + 2/8CO2 + 4/8H 2O

(3)

The ammonium, nitrite, nitrate and methane consumption rates and nitrogen gas production rates, denoted as rNH4+, rNO2−, rNO3−, rCH4, rN2, respectively, were determined C

dx.doi.org/10.1021/es402775z | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Table 1. Measured Ammonium, Nitrite, Nitrate and Methane Consumption Rates, Nitrogen Gas Production Ratea measured A-1 A-2

mmol d−1 mgN L−1 d−1 or mgCH4 L−1 d−1 mmol d−1 mgN L−1 d−1 or mgCH4 L−1 d−1

rNO2− 0 0 0 0

rNO3− −1.7 −53 −6.2 −193

predicted

rNH4+ −0.70 −22 −2.1 −65

rN2 1.3 79 4.0 247

rCH4 −0.87 −31 −3.5 −124

rN2-p 1.2 74 (7%) 4.1 256 (4%)

rCH4-p −0.82 −29 (6%) −3.2 −113 (8%)

a

The predicted nitrogen gas production rate and methane consumption rate based on the hypothesized reactions match well with the measured rates. The values in brackets are balancing errors: (measured value − predicted value)/measured value ×100%.

DAMO archaeal probe Darch-872, indicating that all the archaea in this culture were DAMO archaea. Quantitative fluorescent in situ hybridization (Q-FISH) revealed that each of these groups represented about 20−30% of the whole microbial community on both Day 415 and Day 730. Batch Test Results. The nitrate, ammonium, methane, and nitrogen gas profiles measured in Batch Tests A-1 and A-2 are shown in SI Figure S1. No nitrite or nitrous oxide accumulation was detected. As summarized in Table 1, the predicted nitrogen gas production and methane consumption rates compare very well with the respective measured rates in both cases, with mass balance errors below 10% in all cases. The calculated nitrogen production rates by the DAMO microorganisms and by the anammox bacteria in Batch Test A-2 are 124 mgN L−1 d−1 and 133 mgN L−1 d−1, respectively, representing 48% and 52% of the total nitrogen gas production. In Batch Test A-2 with 17% 15N-labeled nitrate but no labeled ammonium, the production rates of 29N2 and 30N2 were 0.93 mmol d−1 and 0.06 mmol d−1, respectively (SI Figure S3). Based on SI eqs S12 and S13, the DAMO contribution to nitrogen gas production rN2,DAMO was 2.1 mmol d−1, while the anammox contribution to nitrogen gas production rN2,anammox was about 2.0 mmol d−1. This distribution indicates that 51% and 49% of the nitrogen gas produced could be attributed to DAMO and anammox, respectively. These are very close to the 48% and 52% calculated from mass balance equations. These results collectively suggest that the three hypothesized reactions could well describe the nitrogen and methane conversions in the reactor. The results obtained in Batch Tests B-1 and B-2 are summarized in Table 2. The system achieved nitrogen removal under all loading conditions applied. The nitrogen removal rates improved significantly from B-1 to B-2, corresponding to

through linear regression of their respective concentration profiles measured during batch tests. To verify if the above hypothesized reactions can adequately describe all the experimental data, we calculated r1, r2, and r3 from the rNH4+, rNO2− and rNO3− data, as described in SI. We then used the calculated r1, r2, and r3 to predict the nitrogen gas and methane consumption rates (as described in SI) and compared the estimated rates with the measured rN2 and rCH4. The isotopic data obtained from Batch Test A-2 (with 15N-NO3− addition) were analyzed using the method described by Thamdrup and Dalsgaard26 (summarized in SI) to further verify the above hypothesized reactions.

■

RESULTS Long-term Reactor Performance. The removal rates of nitrate and ammonium by the MBfR during the entire course of the study excluding the Batch Test periods are shown in Figure 2a. Upon inoculation, both the ammonium and nitrate consumption rates increased slowly, reaching about 10 mgNO3−-N L−1 d−1 and 10 mgNH4+-N L−1 d−1, respectively, after 120 days. At around Day 240, the consumption rates for nitrate and ammonium suddenly increased to about 47 mgNO3−-N L−1 d−1 and 43 mgNH4+-N L−1 d−1, respectively, even though no changes were applied to the operational conditions. From Day 290, the MBfR was operated as an SBR with an HRT of 3 days. Stable performance was achieved after Day 360 (Stage 1 of SBR phase in Figure 2). In this period, the nitrate and ammonium removal rates were 29 mgN L−1 d−1 (13 mgN m−2 d−1) and 15 mgN L−1 d−1 (7 mgN m−2 d−1), respectively. Around Day 500, concentrations of both nitrate and ammonium in effluent decreased (Figure 2b), showing increased activity of the biofilm. In response to the enhanced nitrogen removal activity, the influent nitrate and ammonium concentrations were stepwise increased. At the end of this study, the effluent ammonium concentration decreased to approximately 120 mgNH4+-N L−1 with 300 mgNH4+-N L−1 in the influent, while the effluent nitrate concentration reached about 30 mgNO3−-N L−1 with 600 mgNO3− -N L−1 in the influent. The effluent nitrate and ammonium concentrations were not specifically targeted, but consequences of the activities and interactions of the microorganisms in the biofilm that developed under the operational conditions applied. The total nitrogen removal rate in the MBfR reached 250 mgN L−1 d−1 (190 mgNO3−-N L−1 d−1 + 60 mgNH4+-N L−1 d−1) or 113 mgN m−2 d−1. Nitrite accumulation was not observed during the whole study. Microbial Community. FISH analysis of biofilm samples collected on Day 415 and Day 730 confirmed the coexistence of ANME archaea, NC10 bacteria and anammox bacteria (SI Figure S2). All the archaeal cells that bound to the general archaeal probe Arch-915 were also hybridized with the specific

Table 2. Nitrogen Removal Rates by the MBfR When Fed with Wastewater Containing Different Nitrogen Species rNH4+ batch no. B-1

B-2

rNO3−

mgN L−1d−1

feeding NH4 , NO3− NH4+, NO2− NO2− NO3−

−22 −43 ± 1

NH4+, NO3− NH4+, NO2− NO2− NO3−

−65a −260 ± 13

+

rNO2−

−53a

a

−70 ± 2 −28 ± 1 −18 ± 1 −193a −416 ± 13 −147 ± 15 −144 ± 6

a

Average ammonium or nitrate removal rates in Batch Tests A-1 and A-2.

D

dx.doi.org/10.1021/es402775z | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

system.31−34 Flow mal-distribution was previously observed in hollow fiber modules with high packing densities,35 where the activity and character of the biofilm were negatively affected.34 Therefore, it is recommended that packing with a lower density should be used in future studies. The wastewater composition also has an impact on the reaction rates. The ammonium and nitrite consumption rates in Batch Test B-2 with ammonium and nitrite in the feed were 260 mgN L−1 d−1 and 416 mgN L−1 d−1, respectively. A further challenge is the slow enrichment process, as encountered in this study. Both DAMO and anammox are slowly growing organisms, and it is therefore not surprising that extended time is required for the full formation of DAMO plus anammox biofilms. In this study, we did not focus on the optimization of the start-up phase. It is likely that this phase can be shortened by developing proper operational strategies. It should also be noted that the research of the DAMO process is still in its infancy. A better understanding of these organisms in terms of their physiology and kinetics will certainly facilitate the development and optimization of engineering systems. Interactions between Anammox Bacteria and DAMO Organisms. FISH results revealed that DAMO archaea, DAMO bacteria and anammox bacteria coexisted and indeed jointly dominated the microbial community. These microorganisms were known to be able to catalyze the biochemical reactions as described in eqs 1−3, and the mass balance and isotope test results showed that these reactions likely occurred concurrently in the MBfR. A conceptual model describing these reactions and their potential interactions is presented in Figure 3, which is further explained below. Ammonium consumption

the enhanced reactor performance from Stage 1 to Stage 2. However, the nitrogen removal rate varied substantially between tests. The highest total nitrogen removal rate (676 mgN L−1 d−1) was achieved in Batch Test B-2 fed with ammonium, nitrite and methane, more than doubling that achieved with the normal ammonium, nitrate and methane feeding (258 mgN L−1 d−1). Without ammonium, with only nitrate and methane feed, the nitrate consumption rates in both B-1 and B-2 (18 mgN L−1 d−1 and 144 mgN L−1 d−1, respectively) were lower than those when ammonium and nitrate were both present (53 mgN L−1 d−1 and 193 mgN L−1 d−1, respectively). During the test with methane and nitrite feed, the nitrite removal rates (28 mgN L−1 d−1 in B-1 and 147 mgN L−1 d−1 in B-2) were slightly higher than those rates achieved with methane and nitrate feed. However, these nitrite removal rates were still lower than the nitrite removal rate when ammonium was present in the feed. When the recirculation rate was raised to 1200 mL min−1 in Batch Test C, both the nitrate and ammonium consumption rates improved by about 37% (SI Table S1). However, biofilm detachment was observed due to the high shearing force, and the recirculation rate had to be returned to 600 mL min−1 after the 6 h trial.

■

DISCUSSION A Potential Novel Technology for Nitrogen Removal from Wastewater. This research demonstrates that it is feasible to grow a coculture of DAMO and anammox microorganisms in a hollow fiber MBfR to achieve simultaneous ammonium and nitrate removal using methane as an externally provided electron donor. At the end of the study, the MBfR achieved nitrate removal at 190 mgNO3−-N L−1 d−1. To our knowledge, this is the highest nitrate removal rate reported for DAMO organisms.12,27 This nitrate removal rate, along with the ammonium removal rate of 60 mgNH4+-N L−1 d−1, gives a total nitrogen removal rate of 250 mgN L−1 d−1. This process has a significant potential for wastewater treatment. For example, the anammox process is being used at full-scale for nitrogen removal from anaerobic sludge digestion liquor, either following, or simultaneously with, a partial nitritation process that converts approximately 50% of the ammonium to nitrite.6,28 More recently, major research effort is being devoted to achieving nitrogen removal through the socalled main-stream anammox process.29,30 In both cases, methane is produced through anaerobic sludge digestion on site of the treatment plant. Our work shows that a small part of the methane could be fed to the anammox reactor using hollow fiber membranes to support the removal of nitrate produced by the anammox process. The supply of methane would also enable complete nitrogen removal from wastewater containing varying ammonium, nitrite, and nitrate ratios. This represents a significant advantage over the anammox-only process, which requires a fixed nitrite to ammonium molar ratio of 1.32 to 1. However, significant challenges exist for the process to be developed into a practical technology. The nitrate and ammonium consumption rates (250 mgNO3−-N L−1 d−1) obtained in this proof-of-concept study are still lower than the conventional denitrification rates, and require enhancement. The reaction rate could be improved through optimization, which was not conducted in this work. Indeed, Batch Tests C already suggested that improving hydrodynamic conditions could enhance reaction rates. The high packing density might have negatively affected the hydrodynamic conditions in the

Figure 3. Hypothesized biochemical reactions of and the interaction between DAMO and anammox microorganisms in the biofilm.

was clearly observed during normal operation of the MBfR and also in both Batch Tests A and Batch Tests B, suggesting the presence of anaerobic ammonia oxidation by the anammox bacteria that were present in the microbial community at 20− 30%. It has been well established that the anammox metabolism requires nitrite as the electron acceptor (eq 1). As nitrite was not externally fed to the reactor, it must have been produced by E

dx.doi.org/10.1021/es402775z | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

(4) van der Star, W. R. L.; Abma, W. R.; Blommers, D.; Mulder, J.W.; Tokutomi, T.; Strous, M.; Picioreanu, C.; van Loosdrecht, M. C. M. Startup of reactors for anoxic ammonium oxidation: Experiences from the first full-scale anammox reactor in Rotterdam. Water Res. 2007, 41 (18), 4149−4163. (5) Li, A.; Sun, G.; Xu, M. Recent patents on anammox process. Recent Pat. Eng. 2008, 2 (3), 189−194. (6) Joss, A.; Salzgeber, D.; Eugster, J.; KÖ nig, R.; Rottermann, K.; Burger, S.; Fabijan, P.; Leumann,, S.; Mohn, J.; Siegrist, H. Full-scale nitrogen removal from digester liquid with partial nitritation and anammox in one SBR. Environ. Sci. Technol. 2009, 43 (14), 5301− 5306. (7) Khin, T.; Annachhatre, A. P. Novel microbial nitrogen removal processes. Biotechnol. Adv. 2004, 22 (7), 519−532. (8) Vlaeminck, S. E.; Terada, A.; Smets, B. F.; Linden, D. V. d.; Boon, N.; Verstraete, W.; Carballa, M. Nitrogen removal from digested black water by one-stage partial nitritation and anammox. Environ. Sci. Technol. 2009, 43 (13), 5035−5041. (9) Vlaeminck, S.; Cloetens, L. F. F.; Carballa, M.; Boon, N.; Verstraete, W. Granular biomass capable of partial nitritation and anammox. Water Sci. Technol. 2008, 58 (5), 1113. (10) Modin, O.; Fukushi, K.; Yamamoto, K. Denitrification with methane as external carbon source. Water Res. 2007, 41 (12), 2726− 2738. (11) Raghoebarsing, A. A.; Pol, A.; van de Pas-Schoonen, K. T.; Smolders, A. J. P.; Ettwig, K. F.; Rijpstra, W. I. C.; Schouten, S.; Damste, J. S. S.; Op den Camp, H. J. M.; Jetten, M. S. M.; Strous, M. A microbial consortium couples anaerobic methane oxidation to denitrification. Nature 2006, 440 (7086), 918−921. (12) Haroon, M. F.; Hu, S.; Shi, Y.; Imelfort, M.; Keller, J.; Hugenholtz, P.; Yuan, Z.; Tyson, G. W. Anaerobic oxidation of methane coupled to nitrate reduction in a novel archaeal lineage. Nature 2013, 500 (7464), 567−570. (13) Ettwig, K. F.; Butler, M. K.; Le Paslier, D.; Pelletier, E.; Mangenot, S.; Kuypers, M. M. M.; Schreiber, F.; Dutilh, B. E.; Zedelius, J.; de Beer, D.; Gloerich, J.; Wessels, H. J. C. T.; van Alen, T.; Luesken, F.; Wu, M. L.; van de Pas-Schoonen, K. T.; Op den Camp, H. J. M.; Janssen-Megens, E. M.; Francoijs, K.-J.; Stunnenberg, H.; Weissenbach, J.; Jetten, M. S. M.; Strous, M. Nitrite-driven anaerobic methane oxidation by oxygenic bacteria. Nature 2010, 464 (7288), 543−548. (14) Luesken, F.; van Alen, T.; van der Biezen, E.; Frijters, C.; Toonen, G.; Kampman, C.; Hendrickx, T.; Zeeman, G.; Temmink, H.; Strous, M.; Op den Camp, H.; Jetten, M. Diversity and enrichment of nitrite-dependent anaerobic methane oxidizing bacteria from wastewater sludge. Appl. Microbiol. Biotechnol. 2011, 92 (4), 845−854. (15) Hu, S.; Zeng, R. J.; Burow, L. C.; Lant, P.; Keller, J.; Yuan, Z. Enrichment of denitrifying anaerobic methane oxidizing microorganisms. Environ. Microbiol. Rep. 2009, 1 (5), 377−384. (16) Hallam, S. J.; Putnam, N.; Preston, C. M.; Detter, J. C.; Rokhsar, D.; Richardson, P. M.; DeLong, E. F. Reverse methanogenesis: Testing the hypothesis with environmental genomics. Science 2004, 305 (5689), 1457−1462. (17) Caldwell, S. L.; Laidler, J. R.; Brewer, E. A.; Eberly, J. O.; Sandborgh, S. C.; Colwell, F. S. Anaerobic oxidation of methane: Mechanisms, bioenergetics, and the ecology of associated microorganisms. Environ. Sci. Technol. 2008, 42 (18), 6791−6799. (18) Luesken, F. A.; Sanchez, J.; van Alen, T. A.; Sanabria, J.; Op den Camp, H. J. M.; Jetten, M. S. M.; Kartal, B. Simultaneous nitritedependent anaerobic methane and ammonium oxidation. Appl. Environ. Microbiol. 2011, 77 (19), 6802−6807. (19) Martin, K. J.; Nerenberg, R. The membrane biofilm reactor (MBfR) for water and wastewater treatment: Principles, applications, and recent developments. Bioresour. Technol. 2012, (0). (20) Lee, K. C.; Rittmann, B. E. Applying a novel autohydrogenotrophic hollow-fiber membrane biofilm reactor for denitrification of drinking water. Water Res. 2002, 36 (8), 2040−2052. (21) Terada, A.; Kaku, S.; Matsumoto, S.; Tsuneda, S. Rapid autohydrogenotrophic denitrification by a membrane biofilm reactor

other reactions. In parallel, concurrent consumption of methane and nitrate was observed, strongly suggesting the occurrence of anaerobic methane oxidation coupled to nitrate reduction (eq 3). The archaeal DAMO, which were present at 20−30%, have been proven to be capable of nitrate reduction to nitrite coupled to methane oxidation.12 We therefore postulate that nitrate reduction in the MBfR was mainly conducted by the archaeal DAMO. Supported by the isotope test as well as the excellent electron balance obtained in Batch Tests A-1 and A-2, it is almost certain that this reaction provided the nitrite required by the anammox reaction. The equally high abundance of bacterial DAMO in the community (20−30%) implied that they also played a significant role in the reactor, likely associated with nitrite reduction to nitrogen gas (eq 2).15,25,36 The occurrence of nitrite reduction coupled to methane oxidation was indeed confirmed in Batch Tests B-1 and B-2 (Table 2). In Haroon et al.,12 a partnership was observed between archaeal DAMO and anammox bacteria (without bacterial DAMO) with feed similar to that used in this study. This structure is in clear contrast to what was observed in this study, where bacterial DAMO constituted 20−30% of the microbial community despite of its low abundance in the seeding sludge. The key difference between the two studies is that in the MBfR the DAMO and anammox organisms developed in biofilms in this study where ammonium/nitrate and methane were transferred to the biofilm through counter diffusion (Figure 3). In comparison, such niches were not present in the flocular system used in Haroon et al, where ammonium, nitrate and methane are all dissolved in the bulk liquid.

■

ASSOCIATED CONTENT

S Supporting Information *

Reactor configuration, parent reactor details, and different recirculation rates effect results. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*(Z.Y.) Phone: +61 7 3365 4374; fax: +61 7 3365 4726; e-mail: [email protected]. Funding

This study was funded by the Australian Research Council (ARC) through projects DP0987204 and DP120100163. Ying Shi thanks China Scholarship Council (CSC) for scholarship support. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS We thank Dr Beatrice Keller and Ms Jianguang Li for assistance with FIA and N2O measurements. REFERENCES

(1) Kartal, B.; Kuenen, J. G.; van Loosdrecht, M. C. M. Sewage treatment with anammox. Science 2010, 328 (5979), 702−703. (2) Kuenen, J. G. Anammox bacteria: From discovery to application. Nat. Rev. Microbiol. 2008, 6 (4), 320−326. (3) Abma, W. R.; Driessen, W.; Haarhuis, R.; van Loosdrecht, M. C. M. Upgrading of sewage treatment plant by sustainable and costeffective separate treatment of industrial wastewater. Water Sci. Technol. 2010, 61 (7), 8. F

dx.doi.org/10.1021/es402775z | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

equipped with a fibrous support around a gas-permeable membrane. Biochem. Eng. J. 2006, 31 (1), 84−91. (22) Lee, K.; Rittmann, B. A novel hollow-fibre membrane biofilm reactor for autohydrogenotrophicdenitrification of drinking water. Water Sci. Technol. 2000, 41 (4−5), 219−226. (23) Duan, C.; Luo, M.; Yang, C.; Jiang, H.; Xing, X. Effects of different hollow fiber membrane modules on bubbless aeration of methane and oxygen. Chin. Jo. Process Eng. 2010, 10 (2), 5. (24) Modin, O.; Fukushi, K.; Nakajima, F.; Yamamoto, K. Performance of a membrane biofilm reactor for denitrification with methane. Bioresour. Technol. 2008, 99 (17), 8054−8060. (25) Ettwig, K.; Shima, S.; van de Pas-Schoonen, K.; Kahnt, J.; Medema, M.; op den Camp, H.; Jetten, M.; Strous, M. Denitrifying bacteria anaerobically oxidize methane in the absence of Archaea. Environ. Microbiol 2008, 10 (11), 3164−3173. (26) Thamdrup, B.; Dalsgaard, T. Production of N2 through anaerobic ammonium oxidation coupled to nitrate reduction in marine sediments. Appl. Environ. Microbiol. 2002, 68 (3), 1312−1318. (27) Winkler, M. K. H.; Yang, J.; Kleerebezem, R.; Plaza, E.; Trela, J.; Hultman, B.; van Loosdrecht, M. C. M. Nitrate reduction by organotrophic Anammox bacteria in a nitritation/anammox granular sludge and a moving bed biofilm reactor. Bioresour. Technol. 2012, 114 (0), 217−223. (28) Fux, C.; Boehler, M.; Huber, P.; Brunner, I.; Siegrist, H. Biological treatment of ammonium-rich wastewater by partial nitritation and subsequent anaerobic ammonium oxidation (anammox) in a pilot plant. J. Biotechnol. 2002, 99 (3), 295−306. (29) Shaughnessy, M.; Wett, B.; Bott, C.; Murthy, S.; deBarbadillo, C.; Kinnear, D.; Chandran, K.; Neethling, J. B.; Shaw, A.; Stinson, B.; Barnard, J. Full-plant deammonification for low-energy, low-carbon nitrogen removal. Proc. Water Environ. Fed. 2011, 2011 (13), 3715− 3720. (30) Malovanyy, A.; Plaza, E.; Yatchyshyn, Y.; Trela, J.; Malovanyy, M. Removal of nitrogen from the mainstream of municipal wastewater treatment plant with combination of ion exchange and canon process (IE-canon)−effect of NaCl concentration. Future urban sanitation to meet new requirements for water quality in the Baltic Sea region, Joint Polish-Swedish Reports 2011, (17), 17-19. (31) Chen, V.; Hlavacek, M. Application of Voronoi tessellation for modeling randomly packed hollow-fiber bundles. AIChE J. 1994, 40 (4), 606−612. (32) Wang, Y.; Chen, F.; Wang, Y.; Luo, G.; Dai, Y. Effect of random packing on shell-side flow and mass transfer in hollow fiber module described by normal distribution function. J. Membr. Sci. 2003, 216 (1−2), 81−93. (33) Zheng, J.; Xu, Y.; Xu, Z. Flow distribution in a randomly packed hollow fiber membrane module. J. Membr. Sci. 2003, 211 (2), 263− 269. (34) Rector, T. J.; Garland, J. L.; Starr, S. O. Dispersion characteristics of a rotating hollow fiber membrane bioreactor: Effects of module packing density and rotational frequency. J. Membr. Sci. 2006, 278 (1−2), 144−150. (35) Pankhania, M.; Brindle, K.; Stephenson, T. Membrane aeration bioreactors for wastewater treatment: Completely mixed and plug-flow operation. Chem. Eng. J. 1999, 73 (2), 131−136. (36) Hu, S.; Zeng, R. J.; Keller, J.; Lant, P. A.; Yuan, Z. Effect of nitrate and nitrite on the selection of microorganisms in the denitrifying anaerobic methane oxidation process. Environ. Microbiol. Rep. 2011, 3 (3), 315−319.

G

dx.doi.org/10.1021/es402775z | Environ. Sci. Technol. XXXX, XXX, XXX−XXX