You are on page 1of 12

MONITORING THE IMPACT OF DISSOLVED OXYGEN AND NITRITE ON ANOXIC BIOFILM IN CONTINUOUS DENITRIFICATION PROCESS

AM JANG, YEONGHEE AHN and IN S. KIM


Bio-Environmental Engineering Laboratory (BEEL), Department of Environmental Science and Engineering, Kwangju Institute of Science and Technology (K-JIST), 1 Oryong-dong, Gwangju, Korea ( author for correspondence, e-mail: iskim@kjist.ac.kr)

(Received 11 April 2002; accepted 25 September 2002)

Abstract. An anoxic biolm involved in continuous denitrication process was monitored to investigate the effect of different concentrations of inuent dissolved oxygen (DO) or nitrite on the biolm. Microelectrode measurements evidenced nitrate removal activity of biolm. When different concentrations of DO were applied to the reactor, generally decreased concentrations of DO were observed as bed depth increased from the bottom of the reactor. Greatest decrease of the DO was observed in the lower 20% of the bed depth. Nitrate removal efciency was inversely proportional to inuent DO concentrations (8.311.9 DO mg L1 ) or nitrite loading rates (05.5 N-NO kg m3 2 day1 ) employed in this study. Nitrite loading rates to achieve more than 90% of nitrate removal efciency were 1.46 N-NO kg m3 day1 or less at pH 7.5 and 0.34 N-NO kg m3 day1 or 2 2 less at pH 6.8. Nitrate removal efciency was 63% or more within the lower 20% of the bed depth at the nitrite loading rates that allowed more than 90% of nitrate removal efciency of the reactor. The results of this study provide rst quantitative data that nitrate removal performance of an anoxic biolm is inhibited by DO or nitrite, reported to be a limiting factor in the suspended biological denitrication process. Keywords: anoxic biolm, denitrication, dissolved oxygen, microelectrode, nitrite

1. Introduction It is commonly known that direct discharge of wastewaters containing nitrogenous nutrients into rivers causes negative environmental problems such as eutrophication in aquatic systems and public health concerns. Successive nitrication and denitrication is one of the most effective biological processes to remove nitrogenous chemicals from wastewaters. These two reactions are carried out by two different functional groups of microorganisms (Madigan et al., 2000). The nitrication process converts ammonium ions to nitrite or nitrate ions, whereas the denitrication process changes nitrite or nitrate ions to nitrogen gas. Since the denitrication process actually removed the nitrogenous nutrients from wastewaters, the denitrication is as important as the nitrication. Generally biological denitrication is a facultative trait based on a competitive advantage in low oxygen environments. Active denitrication occurs under anoxic
Environmental Monitoring and Assessment 87: 133144, 2003. 2003 Kluwer Academic Publishers. Printed in the Netherlands.

134

AM JANG ET AL.

conditions that do not involve molecular oxygen, but rather chemical forms (e.g., nitrite, nitrate, sulfate, etc) with combined oxygen atoms. Dissolved oxygen (DO) can inhibit the denitrication reaction because oxygen functions as the electron acceptor for microorganisms over nitrate, and aerobic conditions repress enzymes involved in denitrication (Zumft, 1997). Although high DO concentrations in the biolm reactor are necessary to enhance the activity of nitrifying bacteria, denitrication is inhibited by oxygen (U.S. EPA, 1993; Hagedorn-Olsen, 1994; Lie and Welander, 1994). Accumulated nitrite also inhibits the denitrication process, reducing the quality of efuents from wastewater treatment plants. Therefore, DO and nitrite are limiting factors in the denitrication process. Inhibition of denitrication by nitrite has been studied using batch tests containing suspended cells (van Versefeld et al., 1977; Almedia et al., 1995; Kornaros et al., 1996; Glass et al., 1997; Glass and Silverstein, 1998). However, the results from such batch tests are difcult to apply to the biolm in a continuous process. It is not known what were the DO concentration and nitrate removal efciency at different bed-depths of an anoxic reactor when various concentrations of inuent DO were introduced. In these regards, an anoxic biolm involved in continuous denitrication process was monitored under different concentrations of inuent DO and nitrite to provide quantitative data that the nitrate removal performance of the biolm was inhibited by nitrite or DO.

2. Methods 2.1. A NOXIC BIOFILM REACTOR DESIGN A Plexiglas column (10 cm i.d. 110 cm height) with 95 cm of brous-loop substratum was used as a submerged anoxic biolm reactor (Figure 1). The column surface was covered with a black sheet to eliminate possible growth of oxygenic phototrophic microorganisms. The reactor was lled with frames with lamentous substrata, Hyosung BCplus (Hyosung company, Korea), which are loops with 50 m-thick nylon bers attached to 4 mm polypropylene thread. The packing ratio of the substratum was 0.2:1 (substratum vol. to actual column vol.). 2.2. R EACTOR O PERATION C ONDITIONS The bioreactor was fed with a synthetic wastewater in the upow mode and was operated at 25 4 C. The synthetic wastewater contained glucose, nitrate, buffer, and mineral nutrients (Table I). Unless otherwise specied, the anoxic biolm reactor was operated at pH 7.5 and 8 DO mg L1 was introduced to the reactor. When it is necessary, the feeding solution was aerated using air or pure oxygen (purity, 99%) to increase the DO concentration before introduced into the anoxic reactor. The reactor was seeded with excess sludge from a secondary sedimentation tank in an anoxic condition. The reactor was ushed initially with nitrogen gas

MONITORING THE IMPACT OF DISSOLVED OXYGEN AND NITRITE

135

Figure 1. Schematic diagram of anoxic biolm reactor. 1, efuent; 2, plexiglas column (anoxic reactor); 3, supporting substratum with biolm; 4, sampling port; 5, DO meter; 6, air/pure oxygen supply; 7, inuent (pump); 8, feeding tank (4 C); 9, magnetic stirrer.

and operated in a batch mode for two days to allow microorganisms to adhere on the substratum in the reactor. The reactor was switched to a continuous mode after 3 days of operation. Concentrations of total organic carbon (TOC) and nitrate were maintained at 70 TOC mg L1 and 16 N-NO mg L1 , respectively, and 3 hydraulic retention time (HRT) was adjusted to 3.4 hr (Q = 42.4 L day1 ) during all experiments. The reactor reached a steady state from day 28 of operation, as determined by more than 90% efciency of nitrate removal. The nitrite load was increased stepwise from 0.3 to 5.5 N-NO kg m3 day1 by 2 changing the inuent concentration (40780 N-NO mg L1 ) of nitrite while keep2 ing the HRT constant in order to study nitrite effect on nitrate removal. Whenever a new concentration of nitrite was employed, the concentration was maintained at least a week to acclimate microorganisms in the reactor. Liquid samples were taken three times a week from the feed tank and four ports at the heights of 20, 45, 70 and 95 cm, respectively, from the bottom of the reactor. Samples were ltered

136

AM JANG ET AL.

immediately using cellulose ester membranes (pore size, 0.45 m) for chemical analyses. 2.3. M ICROELECTRODE PREPARATIONS AND MEASUREMENTS Biolm samples were taken on day 28 of operation by cutting off a section of brous-loop substratum with biolm and then immediately transferring it to a cell containing medium (31 N-NO mg L1 at pH 8.0) for microelectrode measure3 ments. Ion selective microelectrodes for NO and pH were employed to measure 3 microgradients of nitrate and pH in the denitrifying biolm from the anoxic reactor. The microelectrodes were prepared and used according to Zhang and Bishop (1994). Microelectrodes having a high spatial resolution were used to measure the proles of pH and NO as a function of depth in the biolm. All working 3 microelectrodes were used along with an Ag/AgCl milli-electrode as the reference electrode throughout the experiments. More than 10 microelectrode proles were produced at different positions along the biolm axis during the investigation. 2.4. A NALYTICAL METHODS Concentrations of nitrate and nitrite were analyzed by an ion chromatography unit (Model DX-120; Dionex Co., CA, U.S.A.) equipped with an ASRS-II suppressor and AS14 column packed with an analytical column (Model Ionpack CS12A; Dionex Co., CA, U.S.A.), CG12A guard column (2 250 mm), CDM-3 detector of electrical conductivity, and AS40 automated sampler. Eluent was (3.5 mM Na2 CO3 ) (2.5 mM NaHCO3)1 at a ow rate of 1.2 mL min1 . Organic compounds were analyzed by a TOC analyzer (Model DC-180; Rosemount Analytical Inc., U.S.A.) with an automatic sampler. DO was measured by a DO meter (Model YSI 59; YSI Inc., OH, U.S.A.) with a 5905 BOD probe. The DO probe was placed at different bed-depths in the reactor to determine DO concentration at the positions. ORP and pH were analyzed with a model 250A probe and model 9778 redox probe (Thermo Orion, MA, U.S.A.), respectively.

3. Results and Discussions 3.1. B IOFILM FORMATION ON THE SUBSTRATUM AND MICROELECTRODE
MEASUREMENTS

Light microscopy showed that microbial biolm was developed in the shape of ocks (Kim et al., 2000). Small ocks of biomass attached to the ber were observed on day 7. Sizes of microbial ocks increased gradually and reached a plateau of 300-m thickness by day 24. Biolm was sampled after 4 weeks of operation and used for microelectrode measurements.

MONITORING THE IMPACT OF DISSOLVED OXYGEN AND NITRITE

137

Figure 2. Microproles of nitrate concentration and pH in the biolm. Surface of biolm was indicated as 0 m.

The thickness of the biolm was counted from the point (interface of biolm and bulkliquid) where a small and sudden drift was observed when the microelectrode moved from the bulk solution to the interface of the biolm. The proles of pH and NO concentration were produced as a function of distance from the sur3 face of the biolm (Figure 2). The pH values in the biolm slightly increased from 8.0 to 8.3 as the microelectrode moved inside the biolm from the bulk solution. On the other hand, the NO concentration signicantly dropped from 31 to 4 N3 NO mg L1 . Nitrate removal activity was only found within the upper 250 m 3 of the biolm. The nitrate concentration (31 N-NO mg L1 ) in the bulk liquid 3 started to decrease 150 m above the biolm surface, reaching 22 N-NO mg L1 3 at the interface between the biolm and the bulk solution. As the microelectrode moved inside biolm, the nitrate concentration decreased to 5 N-NO mg L1 3 until it reached the upper 200-m of depth and then further decreased slightly by the upper 250-m of depth, but never reached zero. Considering the ndings that decrease of NO concentration only occurred in the range from 150 m above the 3 biolm to 250 m into the upper biolm, it is evident that denitrifying microorganisms were active only in the upper region of the biolm. However, the nitrate removal efciency observed in the upper 100200 m of the biolm was higher than that shown in the upper 0100 m of the biolm. Further study with in situ hybridization combined with confocal laser scanning microscopy will give us more information on spatial distribution of denitriers within the biolm.

138

AM JANG ET AL.

It is technically difcult to apply in situ microelectrode measurements to biolm in a reactor or to sample anoxic biolm in an anoxic mode from a continuous reactor such as the one used in this study. However, judging from the results shown in Figure 2 the sampling method described in the Methods section, denitriers in the upper 100200 m of the biolm were more active. The lower level of the nitrate removal efciency observed in the upper 0100 m of the biolm might be due to penetration of oxygen to the layer of the biolm during the sampling event while denitriers in the upper 100200 m of the biolm could be protected from the oxygen by microbial respiration occurred in the outer layer of the biolm. The results of microelectrode measurements directly supported the nding that biolm formed in the reactor actually contributed to nitrate removal performance of the reactor. 3.2. I NHIBITORY EFFECT OF ELEVATED DO CONCENTRATION ON NITRATE
REMOVAL

Inuents with different concentrations of DO were supplied to the anoxic reactor to investigate the effect of DO on nitrate removal. The proles of DO in the biolm reactor (Figure 3) show that DO concentration decreased as bed depth increased from the bottom of the reactor, except at the depth near the watersurface. The increased DO at the region near the watersurface might be caused by the reactor top that was open. When 8.3 or 11.9 mg of DO was supplied to the reactor, almost all DO was consumed in the lower 20 cm of the bed depth, suggesting occurrence of active oxygen-consuming microorganisms in this part of the reactor (Figure 3a). However, DO concentrations at the depth of the reactor were more than 5 mg L1 when inuent DO concentrations were increased to 17.4 and 19.6 mg L1 . Regardless of the inuent DO concentrations employed in this study, DO concentrations remained below 2 mg L1 at the lower 70% (70 cm) of bed depth. Taken together, Figure 3 (a) suggests that facultatively aerobic microorganisms were present throughout the bed depth, and their oxygen respiration was most active at the lower bed-depth of the anoxic reactor. Figure 3(b) shows the proles of nitrate removal under various concentrations of inuent DO. Nitrate removal decreased, as DO concentrations of the inuent increased. When inuent DO concentrations were 8.311.9 mg L1 , the highest rate of DO removal was observed at the lower 20% of the bed depth (Figure 3a). Approximately 60% of the nitrate removal was achieved in this part of the reactor when 8.3 mg L1 of inuent DO was employed. On the other hand, most reduction of nitrate concentration occurred in the lower 45% of the bed depth when 19.6 mg L1 of inuent DO was applied. When inuent DO concentrations were 8.3 and 11.9 mg L1 , maximum nitrate removal rates were 81 and 30 N-NO g m2 3 day1 , respectively. Considering the results that DO concentration decreased as bed depth increased from the bottom of the reactor and that elevated DO concentrations inhibited nitrate removal reactions as shown in Figure 3, increasing the height of an

MONITORING THE IMPACT OF DISSOLVED OXYGEN AND NITRITE

139

Figure 3. Concentrations of (a) DO and (b) nitrate at different bed-depths of the anoxic reactor. Various concentrations of inuent DO were employed for the experiments. Relative bed-depth 100% indicates 100 cm from the bottom of the reactor.

140

AM JANG ET AL.

anoxic reactor or installing a degasication facility might be a solution to relieve the inhibition of nitrate removal by DO. Organic carbon removal showed a similar trend to the nitrate removal along the bed depth, suggesting heterotrophic denitriers were mostly active in the lower part of the bed-depth. While maintaining a specic nitrate removal rate of 81 N-NO 3 g m2 day1 , microorganisms in the lower 20% of bed depth removed 65% of the inuent organic carbon when 8.3 DO mg L1 was employed (data not shown). 3.3. E FFECT OF NITRITE LOADING RATE ON NITRATE REMOVAL AT DIFFERENT BED - DEPTHS The denitrication rate decreased sharply when pH was below 6.0 or over 8.0, whereas the denitrication rate showed only relatively slight changes between pH 6.08.0, where anoxic reactors designed for nitrate removal are generally operated (Halling-Sorensen and Jorgensen, 1993). Different concentrations of nitrite were supplied to examine the effect of nitrite loading rate on nitrate removal occurring at different bed-depths. The effect was investigated at pH 7.5 and 6.8 (Figure 4). Similar values of nitrate removal were observed when nitrite loading rates were in the range of 0.351.46 N-NO kg m3 day1 at pH 7.5. Approximately 65% of 2 the nitrate was removed in the lower 20% of the bed depth under these conditions. Greatly decreased activity of nitrate removal was observed when nitrite loading rate was 2.2 N-NO kg m3 day1 or more (Figure 4a and 5), showing gradual 2 reductions in nitrate concentrations along the height of the reactor. This pattern of nitrate removal was different from the trend that nitrate removal occurred mostly in the lower part of the reactor when lower concentrations of DO or nitrite were supplied, as shown in Figures 3 (b) and 4. The effect of nitrite loading rate on nitrate removal at different bed-depths was tested at pH 6.8. When the nitrite loading rate was 0.34 N-NO kg m3 day1 , 2 similar values and trends of nitrate removal were observed at both pH 7.5 and 6.8 (Figure 4). However the lower pH had more negative effect on nitrate removal when the nitrate loading rate was increased from 0.34 to 0.91 N-NO kg m3 day1 2 (Figure 4 and 5). When the nitrate loading rate was 0.91 or 0.90 N-NO kg m3 2 day1 , the nitrate removal efciency achieved in the lower 20% of the bed depth at pH 6.8 was only one third of that observed at pH 7.5 (Figure 4). This suggests that the lower pH could enhance inhibition of nitrate removal by nitrite. Therefore, maintenance of proper pH in the anoxic reactor is critical to achieve successful removal of nitrate. As nitrite concentration increased, the color of the biolm formed on the substratum changed from black to light brown, and the concentration of suspended solids decreased (data not shown). The organic carbon removal trend was very similar to that for nitrate removal, when different loading rates of nitrite were employed at pH 7.5 (data not shown). As the nitrite concentration increased, the removal efciencies of organics and nitrate decreased, suggesting the activity of

MONITORING THE IMPACT OF DISSOLVED OXYGEN AND NITRITE

141

Figure 4. Effect of nitrite loading rates on nitrate removal at different bed-depths. The effect was tested at pH 7.5 (a) and 6.8 (b). Different rates of nitrite loading were achieved by changing inuent nitrite concentration while keeping HRT constant (The unit of legend is kg N-NO m3 day1 ). 2

142

AM JANG ET AL.

Figure 5. Effect of nitrite loading rate on nitrate removal efciency at pH 7.5 and 6.8.

heterotrophic denitriers might be inhibited and no more organic carbon might be required as an electron donor. 3.4. E FFECT OF NITRITE LOADING RATE ON NITRATE REMOVAL EFFICIENCY AT pH 7.5 AND 6.8 Figure 5 shows nitrate removal efciency depending on nitrite loading rate under two different pH conditions, pH 7.5 and 6.8. The efciency (%) was calculated based on nitrate concentrations of the inuent and efuent. It was evident that the lower pH enhanced inhibition of nitrate removal by nitrite. This result agreed with results previously produced from batch experiments (Glass, 1997; Glass and Silverstein, 1998). As nitrite loading rate increased, nitrate removal activity initially decreased slightly to 90% efciency, followed by a signicant decrease of the activity, eventually reaching a plateau of 20% nitrate removal efciency. Nitrite loading rates to achieve more than 90% nitrate removal efciency were 1.46 and 0.34 NNO kg m3 day1 at pH 7.5 and 6.8, respectively. Nitrate removal efciency was 2 not signicantly affected at pH 6.8 until the nitrite loading rate was increased from 0 to 0.34 N-NO kg m3 day1 . Although increased concentration of nitrite could 2 enhance inhibition of nitrate removal at a lower pH, as shown in Figures 45, it is beyond the scope of our study to determine whether the enhanced inhibition was

MONITORING THE IMPACT OF DISSOLVED OXYGEN AND NITRITE

143

directly caused by nitrite ion itself or nitrous acid (HNO2 ). Free nitrous acid was thought to cause inhibition of nitrate removal by other researchers (Glass, 1997; Glass and Silverstein, 1998).

4. Conclusions 1. Microelectrode measurements showed nitrate removal activity of biolm taken from the continuously operated anoxic reactor used for denitrication. As the microelectrode moved inside the biolm from the bulk solution, nitrate concentration signicantly dropped from 31 to 4 N-NO mg L1 . Nitrate concen3 tration (31 N-NO mg L1 ) in the bulk started to decrease 150 m above the 3 biolm surface. Nitrate removal activity of the 400-m thick biolm was only observed within the upper 250 m of the biolm. 2. When different concentrations of DO were applied to the anoxic reactor, decreased concentrations of DO was observed as bed depth increased from the bottom of the reactor, except the depth near the watersurface. The lower 20% of the bed depth showed the highest rate of oxygen consumption activity, regardless of inuent DO concentrations (8.311.9 DO mg L1 ) employed in this study. 3. Nitrate removal efciency was inversely proportional to inuent DO concentrations (8.311.9 mg DO L1 ) employed in this study. The lower 20% of the bed depth showed the highest removal of nitrate when inuent DO concentrations were 8.319.6 mg DO L1 . However, the lower 45% of the bed depth showed the highest removal efciency when the inuent DO was 19.6 mg DO L1 . 4. When various nitrite loading rates (0.340.56 N-NO kg m3 day1 ) were 2 employed at pH 7.5 and 6.8, nitrate removal was inversely proportional to nitrite loading rate. As the nitrite loading rate increased, initially nitrate removal activity decreased slightly to 90% nitrate removal efciency, and was followed by a signicantly decreased efciency, eventually reaching a plateau of 20% nitrate removal efciency. The maximum nitrite loading rates to achieve more than 90% nitrate removal efciency were 1.46 and 0.34 N-NO 2 kg m3 day1 at pH 7.5 and 6.8, respectively. This suggests that nitrite could enhance inhibition of nitrate removal efciency at the lower pH. 5. Nitrate removal occurred mostly within the lower 20% of the bed depth at the nitrite loading rates that allowed more than 90% of nitrate removal efciency.

Acknowledgements This work was supported in part by the Korea Science and Engineering Foundation (KOSEF) through the Advanced Environmental Monitoring Research Center at

144

AM JANG ET AL.

Kwangju Institute of Science and Technology, and in part by the Brain Korea 21 program.

References
Almedia, J. S., Julio, S. M., Reis, M. A. M. and Carrondo, M. J. T.: 1995, Nitrite inhibition of denitrication by Pseudomonas uorescens, Biotechnol. Bioeng. 46, 194201. Glass, C., Silverstein, J. and Oh, J.: 1997, Inhibition of denitrication in activated sludge by nitrite, Wat. Environ. Res. 69, 10861093. Glass, C. and Silverstein, J.: 1998, Denitrication kinetics of high nitrate concentration water: pH effect on inhibition and nitrite accumulation, Wat. Res. 32, 831839. Hagedorn-Olsen, C., Moller, I. H., Tottrup, H. and Harremoes, P.: 1994, Oxygen reduces denitrication in biolm reactors, Wat. Sci. Tech. 29, 8391. Halling-Sorensen, B. and Jorgensen, S. E.: 1993, The Removal of Nitrogen Compounds from Wastewater, Elsevier. Kim, I. S., Stabnikova, E. V. and Ivanov, V. N.: 2000, Hydrophobic interactions within biolms of nitrifying and denitrifying bacteria in biolters, Bioprocess Eng. 22, 285290. Kornaros, M., Zari, C. and Lyberatos, G.: 1996, Kinetics of denitrication by Pseudomonas denitricans under growth conditions limited by carbon and/or nitrate or nitrite, Wat. Environ. Res. 68, 934945. Lie, E. and Welander, T.: 1994, Inuence of dissolved oxygen and oxidation-reduction potential on the denitrication rate of activated sludge, Wat. Sci. Tech. 30, 91100. Madigan, M. T., Martinko, J. M. and Parker, J.: 2000, Biology of Microorganisms, 9th Ed. Prentice Hall, New Jersey. Metcalf and Eddy, Inc.: 1991, Wastewater Engineering: Treatment, Disposal and Reuse, 3rd Ed., McGraw-Hill, New York. Mitchell, R.: 1992, Environmental Microbiology, John Willey & Sons, 290293. U.S. EPA: 1993, EPA Nitrogen Control, Technomic. van Versefeld, H., Meiher, E. and Stouthamer, A.: 1977, Energy conservation during nitrate respiration on Paracoccus denitricans, Arch. Microbiol. 112, 1723. Zhang, T. C. and Bishop, P. L.: 1994, Density, porosity and pore structure of biolms, Wat. Res. 28, 22672277. Zumft, W. G.: 1997, Cell biology and molecular basis of denitrication, Microbiol. Mol. Biol. Rev. 61, 533616.