ABSTRACT In order to assess the stability of nitrogen removal systems utilizing anaerobic ammonium oxidation (anammox), it is necessary to study effects of influent dissolved oxygen (DO) on anammox activity since effluent from nitritation process feed to anammox process. In this study, the effects of influent DO on anammox bacteria entrapped in gel carriers were investigated using continuous feeding tests. The tests were performed in duplicate to confirm the reproducibility and anammox activities were evaluated under different DO concentration of 0 from 5 mg/L in influent. These results suggested that the DO concentration in influent to anammox reactor must be less than 2.5 mg/L. In addition, it was shown that the effect of influent DO on the anammox reaction is reversible because fallen anammox activity by influent DO of 5 mg/L recovered when the influent DO concentration was decreased to less than 1 mg/L.
Journal of Water and Environment Technology, Vol. 9, No.2, 2011 Address correspondence to Yuya Kimura, Hitachi Plant Technologies, Ltd., E-mail: yuuya.kimura.kv@hitachi-pt.com Received October 12, 2010, Accepted April 11, 2011. - 169 - Tolerance Level of Dissolved Oxygen to Feed into Anaerobic Ammonium Oxidation (anammox) Reactor Yuya KIMURA*, Kazuichi ISAKA*, Futaba KAZAMA** *Hitachi Plant Technologies, Ltd., Chiba 271-0064, Japan **Interdisciplinary Graduate School of Medicine and Engineering, University of Yamanashi, Yamanashi 400-8511, Japan ABSTRACT In order to assess the stability of nitrogen removal systems utilizing anaerobic ammonium oxidation (anammox), it is necessary to study effects of influent dissolved oxygen (DO) on anammox activity since effluent from nitritation process feed to anammox process. In this study, the effects of influent DO on anammox bacteria entrapped in gel carriers were investigated using continuous feeding tests. The tests were performed in duplicate to confirm the reproducibility and anammox activities were evaluated under different DO concentration of 0 from 5 mg/L in influent. These results suggested that the DO concentration in influent to anammox reactor must be less than 2.5 mg/L. In addition, it was shown that the effect of influent DO on the anammox reaction is reversible because fallen anammox activity by influent DO of 5 mg/L recovered when the influent DO concentration was decreased to less than 1 mg/L. Keywords: anammox, dissolved oxygen, immobilization. INTRODUCTION Several kinds of wastewater, for example, wastewater from the dewatering of digested sludge and landfill leachate, have high concentrations of ammonium and low concentrations of biodegradable organic compounds (low C/N ratio). In general, nitrogen is removed by using a nitrification-denitrification process. However, in order to achieve complete biological nitrogen removal during the denitrification of these kinds of wastewater, an additional organic carbon source must be added, resulting in higher operating costs. Therefore, the conventional nitrification-denitrification process is not ideal for ammonium-containing wastewater with low C/N ratios. In the 1990s, a novel metabolic pathway, anaerobic ammonium oxidation (anammox), was discovered (Mulder et al., 1995; Van de Graaf et al., 1996). In this process, since ammonium is used with nitrite as the electron donor, addition of organic compounds is not required because anammox bacteria are autotrophic. We are developing a novel nitrogen removal system using anammox bacteria immobilized in a gel carrier. Anammox bacteria immobilized in a gel carrier are easily separated from the liquid in the reactor, and this yields prolonged biomass retention times even with short hydraulic retention times (HRT) (Isaka et al., 2007). Therefore, stable and high nitrogen removal performance can be attained by anammox gel carriers. Since nitrite is used in anammox reaction, a pre-treatment process to oxidize ammonium to nitrite (nitritation) is needed. In most of the treatment systems using anammox, a two step system of nitritation-anammox is adopted (Hellinga et al., 1998; Van der Star et al., 2007; Yamamoto et al., 2008). We are developing a two step system of nitritation-anammox in a similar way. As a matter of course, the nitritation process is Journal of Water and Environment Technology, Vol. 9, No.2, 2011 - 170 - carried out in an aerobic condition. The DO range in our nitritation process is set to 2 mg/L basically, and it is controlled from 1 to 4 mg/L (Isaka et al., 2009). Tokutomi (2004) operated a nitritation process having aerobic granular biomass at DO concentration of less than 1 mg/L. On the other hand, the anammox process is in an anaerobic condition. Consequently, effluent from the nitritation process is fed into the subsequent anammox process as influent with dissolved oxygen (DO). There is a possibility that the anammox activity may be affected by this. However, there are few reports in the literature about the effects of influent DO on anammox activity although a two step system of nitritation-anammox is adopted in many cases. It is necessary to investigate DO range of influent in order to operate a stable anammox process. The purpose of this investigation was to better understand the influence of influent DO, with assumed nitritation on anammox activity in the gel carrier. We also investigated ways to overcome this influence. MATERIALS AND METHODS Seed anammox sludge Enriched sludge was grown in a 22 L fixed bed reactor filled with porous polyester non-woven fabric carriers (Furukawa et al., 2002) at 36°C. Synthetic medium (described below) was continuously fed into the reactor. In order to collect the enriched sludge adhering to the non-woven fabric carriers, the non-woven fabric was washed in a tank filled with the effluent water from the enrichment reactor. The sludge that settled in the tank was collected and used for gel entrapment. Gel carrier Anammox sludge was entrapped in a polyethylene glycol (PEG) gel carrier (Sumino et al., 1997; Isaka et al. 2007). First, a PEG prepolymer and a promoter, (N, N, N’, N’ - tetramethylenediamine), were dissolved in water. Then, the resulting mixture and anammox bacteria enriched sludge were mixed in a beaker. To initiate polymerization, an initiator (potassium persulfate) was added to the beaker. The polymerized carrier gel was cut into 3 mm cubes. The gel carrier contained 15% (w/v) PEG, 0.5% (w/v) promoter, 0.25% (w/v) initiator, and 0.4% (w/v) anammox-bacteria enriched sludge. Synthetic medium For continuous feeding tests, a synthetic medium was used. The medium contained: 750 mg/L (NH 4 ) 2 SO 4 ; 1000 mg/L NaNO 2 ; 420 mg/L NaHCO 3 ; 27.2 mg/L KH 2 PO 4 ; 300 mg/L MgSO 4 ·7H 2 O; 180 mg/L CaCl 2 ·2H 2 O; and 1 mL of trace element solutions 1 and 2. Trace element solution 1 contained 5 g/L EDTA and 5 g/L FeSO 4 ·7H 2 O. Trace element solution 2 contained 15 g/L EDTA; 0.43 g/L ZnSO 4 ·7H 2 O; 0.24 g/L CoCl 2 ·6H 2 O; 0.99 g/L MnCl 2 ·4H 2 O; 0.25 g/L CuSO 4 ·5H 2 O; 0.22 g/L Na 2 MoO 4 ·2H 2 O; 0.19 g/L NiCl 2 ·6H 2 O; 0.21 g/L NaSeO 4 ·10H 2 O; and 0.014 g/L H 3 BO 4 . Chemical analysis and calculations Both the influent and the effluent of the anammox reactors were analyzed in order to evaluate the anammox performances. The ammonium concentrations were measured by Journal of Water and Environment Technology, Vol. 9, No.2, 2011 - 171 - using the indophenol method (Weatherburn, 1967). The nitrite and nitrate concentrations were determined by using ion chromatography (ICS-1500, Dionex, USA). The nitrogen loading rate and conversion rate was calculated from the influent and effluent ammonium and nitrite concentrations as follow. Nitrogen Loading Rate (kg-N/m 3 /d) = (inf.NH 4 -N (mg/L) + inf.NO 2 -N (mg/L) ) × ×10 -3 Nitrogen Conversion Rate (kg-N/m 3 /d) = {(inf.NH 4 -N – eff.NH 4 -N) (mg/L) + (inf.NO 2 -N – eff.NO 2 -N) (mg/L) } × ×10 -3 Reactor and experimental setup A cylindrical reactor containing the anammox bacteria entrapped in the gel carrier was used for continuous water treatment test (Fig. 1). The volume of the reactor was 500 mL, and 100 mL of the gel carrier cubes was placed inside the reactor (packing ratio, 20%). The reactors were operated in a temperature-controlled room at 30°C. The gel carriers were agitated continuously. Agitation was necessary primarily to mix the influent and to remove the nitrogen gas bubbles that formed on the surfaces of the gel carriers. HRT was set to 2 hours. The to in the reactor was monitored and adjusted pH 7.6 by adding 0.2 N hydrochloric acid. We operated two reactors that nitrogen conversion rates were about 2.9 kg-N/m 3 /d (reactor 1) and 4.3 kg-N/m 3 /d (reactor 2) in order to evaluate the reproducibility of the test. The difference of nitrogen conversion rates between reactor 1 and 2 depends on cultivation term and the amount of bacteria in each reactor (Isaka et al., 2007). Since the DO range in our nitritation-process is from 1 to 4 mg/L (Isaka et al., 2009), maximum DO concentration in this study was set to 5 mg/L that allow some latitude. 24 HRT 24 HRT DO sensor DO Controller P Blower N 2 gas Synthetic medium Influent tank Reactors Stirrer Separator Effluent Influent pH sensor 0.2N HCl Gel Carrier Influent pump Stirrer P P Effluent Influent Reactor 2 Reactor 1 Fig. 1 - Schematic illustration of the reactors used in the continuous feeding tests using anammox bacteria entrapped in gel carriers Journal of Water and Environment Technology, Vol. 9, No.2, 2011 - 172 - DO concentrations of the synthetic medium in an influent tank were adjusted from 0 to 5 mg/L by using DO controller and sparging with nitrogen gas and air blower before feeding the synthetic medium to each reactor. RESULTS AND DISCUSSION Effect of influent DO concentration on anammox activity The effects of influent DO on anammox activities in the gel carrier were evaluated using two continuous feeding tests with each reactor. Fig. 2 shows the nitrogen loading rate and the conversion rates over time of reactor 1 and 2. The influent DO concentration was increased to about 5.0 mg/L from 0.2 mg/L stepwise. The influent DO concentration was increased to next step, when the nitrogen conversion rate had not changed greatly. (a) (b) 0.0 1.0 2.0 3.0 4.0 5.0 0.0 1.0 2.0 3.0 4.0 5.0 64 66 68 70 72 74 76 78 conversion rates (kg-N/m 3 /d) Time (day) Influent DO concentration (mg/L) Nitrogen loading and 0.0 1.0 2.0 3.0 4.0 5.0 0.0 1.0 2.0 3.0 4.0 5.0 64 66 68 70 72 74 76 78 conversion rates (kg-N/m 3 /d) Time (day) Influent DO concentration (mg/L) Nitrogen loading and Fig. 2 - The nitrogen loading rate and the conversion rates over time (a) Reactor 1. (b) Reactor 2. Nitrogen loading rate (closed triangles), nitrogen conversion rate of reactor 1 (open circles), nitrogen conversion rate of reactor 2 (open squares) and the solid line (no symbol) shows influent DO concentration Journal of Water and Environment Technology, Vol. 9, No.2, 2011 - 173 - In reactor 1, a nitrogen conversion rate of about 2.9 kg-N/m 3 /d was maintained with influent ammonium and nitrite concentrations of about 160 and 190 mg-N/L, respectively. When DO of 2 mg/L was fed into the reactor, the nitrogen conversion rate of about 2.9 kg-N/m 3 /d was maintained. The DO of influent was set up over 3.0 mg/L, anammox activity gradually decreased. The nitrogen conversion rate was 1.3 kg-N/m 3 /d when DO of 5 mg/L was fed into the reactor. In reactor 2, a nitrogen conversion rate of about 4.3 kg-N/m 3 /d was maintained. The influent DO was set up over 3.0 mg/L, anammox activity gradually decreased. The nitrogen conversion rate was 2.3 kg-N/m 3 /d when DO of 5 mg/L was fed into the reactor. When DO of about 4 mg/L, 2.5 mg/L and less than 1 mg/L were fed into the reactors, DO concentrations in the effluent were confirmed about 3 mg/L, 2 mg/L and less than 1 mg/L, respectively. These results showed that an anammox activity is affected by influent DO because the anammox activities decreased when DO of high concentration was fed into the reactor of each. Fig. 3 shows the amounts of nitrite removed (NO 2 -Nre.) and nitrate produced (NO 3 -Npro.) based on the amount of ammonium removed (NH 4 -Nre.) in these reactors in each influent DO concentration. Strous et al. (1998) have proposed that the anammox reaction can be expressed as follows: NH 4 + + 1.32NO 2 – + 0.066HCO 3 – + 0.13H + → 1.02N 2 + 0.26NO 3 – + 0.066CH 2 O 0.5 N 0.15 + 2.03H 2 O 0.0 0.5 1.0 1.5 2.0 0.0 0.2 0.4 0.6 0.8 012345 NO 2 -N re. / NH 4 -N re. Influent. DO concentration on average NO 3 -N pro. / NH 4 -N re. Fig. 3 - The amounts of nitrite removed (NO 2 -Nre.) and nitrate produced (NO 3 -Npro.) based on the amount of ammonium removed (NH 4 -Nre.) of these reactors in each influent DO concentration. NO 2 -Nre./NH 4 -Nre. in reactor 1 (closed circles) and reactor 2 (closed squares). NO 3 -Npro./NH 4 -Nre. in reactor 1 (open circles) and reactor 2 (open squares) Journal of Water and Environment Technology, Vol. 9, No.2, 2011 - 174 - In their case, therefore, NH 4 -Nre.:NO 2 -Nre.:NO 3 -Npro. is 1:1.32:0.26. In our results, the average amounts for whole performed period were calculated to be 1:1.20 ± 0.18:0.22 ± 0.02 in reactor 1 and 1:1.25 ± 0.16:0.21 ± 0.03 in reactor 2. These results are very close to the reported value (Strous et al., 1998). Moreover, these ratios did not show much difference even when influent DO concentration was different. When DO concentration of 0.2 and 5 mg/L were fed into these reactors, the average ratios of both reactors were calculated to be 1:1.35:0.19 and 1:1.31:0.23, respectively. Therefore, it is suggested that the denitrification performances in our tests under high DO concentration were also mainly anammox reaction and other denitrification and nitrification were comparatively minor in our tests. Relationship between anammox activity and influent DO concentration Fig. 4 shows the relationships between anammox activities and influent DO concentrations of the two reactors. The maximum nitrogen conversion rate of each reactor at the influent DO of about 0.2 mg/L was set to an anammox activity value of 1.0, and for the other concentrations of influent DO, the anammox activity was plotted as a function of influent DO concentration. The two reactors showed very similar behavior. Neither anammox activity was affected when the influent DO of each was less than about 2.5 mg/L. Influent DO of more than 3.0 mg/L affected anammox activity. When influent DO in reactor was 5 mg/L, the anammox activity decreased by about 45%. These results suggested that perhaps the boundary, where anammox activity is affected by influent DO, is around 2.5 mg/L. There are few effects of DO in effluent of a nitritation process on an anammox process, in the case that operating DO in the nitritation process is around 2 mg/L. In a real full-scale system, the DO tolerance level for anammox process may be higher because DO is consumed by an aerobic bacteria including nitrifying bacteria in the effluent of nitritation process. 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0123456 Anammox activity (-) Influent DO concentration (mg/L) Fig. 4 - The Relationship between anammox activity ratio and influent DO concentration. Reactor 1 (open circles) and reactor 2 (open squares) Journal of Water and Environment Technology, Vol. 9, No.2, 2011 - 175 - In this study, it is very likely that most of the DO in the influent tank is fed into a reactor, because HRT was very short (2 hours). DO concentration in the effluent was confirmed about 2 mg/L when DO of 2.5 mg/L was fed into the reactor 1. Moreover, there may be few cases in a real full-scale system that HRT is longer than 2 hours. Therefore, our results were able to confirm the actual influence of influent DO on anammox activity. Incidentally, the DO loading rate in a case that influent DO concentration was 2.5 mg/L in our test is 0.03 kg/m 3 /d. The influent DO concentrations were changed with different nitrogen conversion rates in this study. In these results, even when nitrogen conversion rates between each reactor were different (2.9 and 4.3 kg-N/m 3 /d), we confirmed that there is no significant difference in the influence of influent DO. It is thought that the amount of anammox bacteria was different between reactor 1 and 2, because the nitrogen conversion rate was different. According to the result of real-time PCR (Isaka et al., 2007), the amount of 16S rRNA gene derived from the Planctomycales are calculated at 4.2 × 10 9 copies/g-carrier in reactor 1 and 1.8 × 10 11 copies/g-carrier in reactor 2. Therefore, it is thought that this difference of amount does not show the effect of influent DO on anammox activity. These results were gotten in about one month. Liu et al. (2008) investigated anammox consortium using wastewater containing DO for long term. However, the results showed that the anammox activity was decreased only 5% because Nitrosomonas protected anammox bacteria Planctomycetales against oxygen. Therefore, the effects may not be different, even if we operate the reactors for a longer term. Nowadays, a single step system that combines a nitritation process and anammox process is being developed in aerobic condition (Third et al., 2001; Slinkers et al., 2002; Furukawa et al., 2006). In the system both nitrifying and anammox bacteria can be in single reactor because oxygen zone and no oxygen zone are formed in the biofilm. Furukawa et al. (2006) carried out SNAP (Single-stage Nitrogen removal Anammox and Partial nitritation) process as a single step system under operational condition of DO 2 - 3 mg/L. In the case of the single reactor, some oxygen may reach the zone that anammox bacteria exist, even though oxygen is consumed by nitrifying bacteria because the reactor is always aerobic condition of about 2 mg/L. In this study, anammox activity was confirmed on condition that influent DO was 2.5 mg/L and the effluent DO was about 2 mg/L. It is thought that anammox bacteria can maintain their activity even when some oxygen exists around the bacteria, though DO is not measured around anammox bacteria and a single step system cannot be compared with our system. Therefore, it is suggested that we don’t need to be afraid of DO of at least 1 mg/L when anammox reactor has enough their activity. Recovery from fallen anammox activity There is a possibility that influent DO concentration is increased due to control problems during the nitritation process and then anammox activity is decreased by this DO. It was examined whether fallen anammox activity by influent DO can be recovered Journal of Water and Environment Technology, Vol. 9, No.2, 2011 - 176 - when influent DO concentration is decreased. Fig. 5 shows influent DO concentration and nitrogen conversion rates over time of two reactors. In reactor 1, the nitrogen conversion rate had decreased from 2.9 to 1.3 kg-N/m 3 /d when influent DO of 5 mg/L had been fed into reactor 1. After that, the DO concentration was decreased to less than 1.0 mg /L. Then, the nitrogen conversion rate increased. In reactor 2, the nitrogen conversion rate, that was decreased to 2.3 from 4.3 kg-N/m 3 /d by influent DO of 5 mg/L, increased when DO of less than 1.0 mg/L was fed into reactor 2. Anammox activities in both reactors recovered to about 80% within 3 days. These results show that the anammox activity can be recovered by decreasing DO concentration of influent. Therefore, the effect of influent DO on the anammox bacteria is reversible. We suggest that the DO in the anammox process should be decreased immediately using sparging N 2 or Ar gas when the anammox process is inhibited by influent DO. Effluent DO from nitritation process should be controlled. CONCLUSIONS In this study, the effects of influent DO on anammox bacteria entrapped in gel carriers were investigated using continuous feeding tests for one month. Anammox activity was affected by influent DO of over 2.5 mg/L. The effect of influent DO on the anammox bacteria is reversible. Therefore, there are few effects of DO in effluent from a nitritation process on an anammox process, in the case that operating DO in the nitritation process is around 2 mg/L. Even when the anammox process is inhibited by influent DO temporarily, the inhibition is solved by decreasing the DO concentration. 0.0 1.0 2.0 3.0 4.0 5.0 0.0 1.0 2.0 3.0 4.0 5.0 78 80 82 84 86 88 90 conversion rates (kg-N/m 3 /d) Time (day) Influent DO concentration (mg/L) Nitrogen loading and Fig. 5 - Recovery from DO inhibition. Nitrogen conversion rate of reactor 1(closed circles), nitrogen conversion rate of reactor 2(closed squares) and the solid line (no symbol) shows influent DO concentration Journal of Water and Environment Technology, Vol. 9, No.2, 2011 - 177 - ACKNOWLEDGEMENT This study was supported by NEDO (New energy and industrial technology development organization), Japan. REFERENCES Furukawa K., Lieu P. K., Tokitoh H. and Fujii T. (2006). Development of single-stage nitrogen removal using anammox and partial nitritation (SNAP) and its treatment performances, Wat. Sci. Technol., 53(6), 83-90. Furukawa K., Rouse J. D., Bhatti Z. I., Imajo U., Nakamura K. and Ishida H. (2002). Characterization of microbial community in an anaerobic ammonium-oxidizing biofilm cultured on a nonwoven biomass carrier, J. Biosci. Bioeng, 94(5), 87-94. Hellinga C., Schellen A. A. J. C., Mulder J. W., Van Loosdrecht M. C. M. and HeijnenWater J. J. (1998). The Sharon process: An innovative method for nitrogen removal from ammonium-rich waste water, Science and Technol., 37(9), 135-142. Isaka K., Date Y., Sumino T. and Tsuneda S. (2007). Ammonium removal performance of anaerobic ammonium-oxidizing bacteria immobilized in polyethylene glycol gel carrier, Appl. Microbiol. Biotechno., 76(6), 1457-1495. Isaka K., Itokawa H., Kimura Y., Noto K., Murakami T. and Sumino T. (2009). Novel autotrophic nitrogen removal system using gel entrapment technology. in Proceedings of 6th IWA leading edge conference on water and wastewater technologies, Singapore. Liu S., Yang F., Xue Y., Gong Z., Chen H., Wang T. and Su Z. (2008). Evaluation of oxygen adaptation and identification of functional bacteria composition for anammox consortium in non-woven biological rotating contactor, Bioresource Thechnol., 99(17), 8273-8279. Mulder A., Van de Graaf A. A., Robertson L. A. and Kuenen J. G. (1995). Anaerobic ammonium oxidation discovered in a denitrifying fluidized bed reactor, FEMS Microbiol. Ecol., 16(3), 177-184. Slinkers A. O., Derwort N., Gomez J. L. C., Strous M., Kuenen J. G. and Jetten M. S. M. (2002). Completely autotrophic nitrogen removal over nitrite in one single reactor, Wat. Res., 36, 2475-2482. Strous M., Heijnen J. J., Kuenen J. G. and Jetten M. S. M. (1998). The sequencing batch reactor as a powerful tool for the study of slowly growing anaerobic ammonium-oxidation microorganisms, Appl. Microbiol. Biotechnol., 50(5), 589-596. Sumino T., Noto T., Ogasawara T., Hashimoto N. and Suwa Y. (1997). Nitrification of high concentration ammonium nitrogen using immobilized nitrifying bacteria, WEFTEC 70th Annual Conference and Exposition, 165-172. Third K. A., Slinkers A. O., Kuenen J. G. and Jetten M. S. M. (2001). The CANON system (Completely Autotrophic Nitrogen-removal Over Nitrite) under ammonium limitation: Interaction and competition between three groups of bacteria, System. Appl. Microbiol., 24, 588-596. Tokutomi T. (2004). Operation of a nitrite-type airlift reactor at low DO concentration, Water Science and Technology, 49, 81-88. Van de Graaf A. A., Peter de Bruijn, Robertson L. A., Jetten M. S. M. and Kuenen J. G. (1996). Autotrophic growth of anaerobic ammonium-oxidizing microorganisms in a fluidized bed reactor, Microbiol., 142(8), 2187-2196. Journal of Water and Environment Technology, Vol. 9, No.2, 2011 - 178 - Van der Star W. R. L, Abma E. R., Blommers D., Mulder J. W., Tokutomi T., Strous M., Picioreanu C. and Van der Loosderecht M.C.M. (2007). Startup of reactors for anoxic ammonium oxidation: Experience from the first full-scale anammox reactor in Rotterdam, Wat. Res., 41, 4149-4163. Weatherburn M. W. (1967). Phenol-hypochlorite reaction for determination of ammonia, Analytical Chemistry, 39, 971-974. Yamamoto T., Takaki K., Koyama T. and Furukawa K. (2008). Long-term stability of partial nitritation of swine wastewater digester liquor and its subsequent treatment by Anammox, Bioresource Thechnol., 99(14), 6419-6425. . Received October 12, 2010, Accepted April 11, 2011. - 169 - Tolerance Level of Dissolved Oxygen to Feed into Anaerobic Ammonium Oxidation (anammox) Reactor Yuya. kg-N/m 3 /d when DO of 5 mg/L was fed into the reactor. When DO of about 4 mg/L, 2.5 mg/L and less than 1 mg/L were fed into the reactors, DO concentrations