HYDROGENOTROPHIC DENITRIFICATION OF DRINKING WATER USING FLASKS AND PACKED-BED REACTORS

I.A. Vasiliadou and D.V. Vayenas

ABSTRACT The progress of hydrogenotrophic denitrification in batch assays and packed-bed reactors with a mixed culture was examined. Two bench-scale denitrifying packed-bed reactors, with gravel as support media, were used in order to investigate the performance in two different gravel sizes (mean diameter 2.41 and 1.75 mm. It was observed, that the denitrification rate increased as increases the size of the support medium of the reactor, because of the better circulation of the gases (H2, CO2) in the reactor. Draw-fill experiments took place in flasks and packed-bed reactors. For feed nitrate concentration 80 mgNO3--N/l, the removal rate achieved in the flask was 75 gNO3--N/m3 d, while the removal rate reached 2260 gNO3--N/m3d in the packed-bed reactor, with the higher specific surface area.

1. INTRODUCTION High nitrate (NO3-) levels in drinking water supplies represent a significant risk to human health, as they are directly responsible for methenoglobinemia in infants and may play a role in the development of some cancers [1]. To minimize these health risks, various agencies created standards for nitrate and nitrite in drinking water. The World Health Organization and the European Economic Community have set standards of 11.3 mg NO3--N/l and 0.03 mg NO2--N/l [2]. The removal of nitrate from water has been and continues to be an area of active research. Biological removal of nitrate has certain advantages over physico-chemical removal systems [3]. Biological denitrification is an attractive treatment option for the conversion of the NO3- to inert nitrogen gas [4]. Both heterotrophic and autotrophic organisms are capable of denitrification. Heterotrophic denitrification of drinking water has been widely reported, but the residual carbon sources from these processes cause many problems in drinking water treatment [5]. Both hydrogen gas and various reduced-sulphur compounds can be used as alternative electron donors for autotrophic denitrification. The low cost of these inorganic substrates and low formation of biomass are important advantages [6]. However, the production of sulphates limits the applicability of element sulphur autotrophic denitrification [6, 7]. H2 is an excellent autotrophic choice because of its clean nature, low biomass yield, and relatively low cost, as well as no further steps are required to remove either excess substrate or its derivatives [8]. In recent years, many different technologies for hydrogenotrophic denitrification have been developed in efforts to solve the problem of NO3- in drinking water or wastewater, e.g. membrane reactors [9, 10, 11], bio-electrochemical system [12], fluidized bed reactors [5, 13] and fixed bed reactors [14, 15]. Despite reported studies on hydrogenotrophic denitrification in the literature, only few of them use biological kinetic data to facilitate the design and operation of biological nitrogen removal plants [3, 13, 16, 17]. A number of factors must be considered in the choice of the bioreactor system (cost, denitrification rate, stability). In this study, the progress of hydrogenotrophic denitrification in batch assays and packed-bed reactors with a mixed culture was examined. Batch culture experiments in flasks were initially conducted to investigate the hydrogenotrophic denitrification rate. Experimental data were used to validate the model, which developed in one of our previous study [18]. The growth kinetics is capable of describing the behaviour of mixed culture. The mixed culture that was used was found to efficiently support potable water denitrification. The present work was undertaken in order to investigate the hydrogenotrophic denitrification activity of bench-scale denitrifying packed-bed reactors, with gravel in two different sizes as support media. Attached growth systems like packed-bed reactors provide a support medium for biofilm growth, thus allowing the possibility of maintaining bacteria at high hydraulic and nitrate loadings. The draw-fill operation of these reactors was investigated at various concentrations of nitrate. 2. MATERIALS AND METHODS 2.1 Cultivation of microorganisms The mixed culture of denitrifying bacteria was enriched from sludge taken from the wastewater treatment plant of the city of Agrinio, Greece. Enrichment cultures were incubated under anoxic conditions using growth medium. The growth medium contained tap water, KNO3 (0.722 g/l),

KH2PO4 (3.39 g/l) and Na2HPO4 (3.53 g/l). The cultures were continuously sparged with a gas mixture of CO2 and H2. 2.2 Experimental system Batch culture studies All batch experiments were performed in 3 l closed sealed flasks with working volume of 2 l. The experimental system used in the denitrification tests consisted of three reactors and is shown in Figure 1a. The suspension medium was continuously stirred at a constant rate of 600 rpm during all the runs. The system was maintained at a temperature of 301oC. The synthetically produced contaminated water used in batch experiments was composed of tap water, KNO3 (0.58 g/l) as the contaminant, KH2PO4 (3.39 g/l) and Na2HPO4 (3.53 g/l). Carbon dioxide (50ml/min) was used as a carbon source and hydrogen (150ml/min) as an electron donor. Carbon source and hydrogen flow rates were chosen to be in excess, to ensure that they were not rate-limiting. The purpose of adding CO2 was to ensure an ample supply of inorganic carbon for autotrophic growth, while the phosphate buffer was added to maintain the pH at 6.5 7.1 for all testing conditions. 2.3 Experimental system Bioreactor studies The bench-scale denitrifying packed-bed reactors (Figure 1b) consisted of Plexiglas tubes, 45 cm high and 4 cm i.d. The overall volume of each reactor was 565 ml. The support material was gravel at two different sizes with a depth of 39 cm. The mean diameters of the mediums were 2.41 mm and 1.75 mm, while the specific surface areas were 2274 m2/m3 and 3207 m2/m3, respectively. The porosity of the support medium was 0.4 and it was almost the same for the two different gravel sizes. At the bottom of the filters, were fixed nozzles, which distributed the incoming of feed solution and gasses. Carbon source (30ml/min) and hydrogen (90ml/min) flow rates were chosen to be in excess, to ensure that they were not rate-limiting. The feed solution (synthetically produced contaminated water) was composed of tap water, KNO3 (0.43g/l to 1.52 g/l) as the contaminant, KH2PO4 (3.39 g/l) and Na2HPO4 (3.53 g/l). Along the filters depth there were 4 sampling ports for nitrate and nitrite concentration measurements in the bulk liquid. Throughout the batch and draw-fill operation of the reactors, the system was maintained at a temperature of 241oC, while the pH maintain at 70.3.

(a)

(b)

Figure 1. Schematic drawing of the experimental apparatus for a) flask reactors, and b) bench-scale packed-bed reactors.

2.4 Analytical methods During all experiments, nitrate and nitrite concentrations measurements were taken on daily basis. To study the microbial growth in the batch reactors, samples were taken at the preset time and processed immediately. Microbial growth was monitored by measuring at regular intervals the absorbance of samples at 600 nm. Optical density was converted to dry cell mass via a calibration curve. The liquid samples were analyzed for nitrate and nitrite after filtration with a 0.45 m membrane filter. Nitrate, nitrite, dry cell mass (2540D) and pH were measured according to Standard Methods for the Examination of Water and Wastewater [19]. 3. EXPERIMENTAL RESULTS AND DISCUSSION 3.1 Batch cultures experiment 3.1.1 Model Development - Evaluation of kinetic parameters Batch experiments were carried out by Vasiliadou et al. (2006) for model development and estimation of the models kinetic parameters, for describing the process of hydrogenotrophic denitrification. They concluded that high nitrate concentration cause an inhibitory effect on the rates of the denitrification process. The kinetic model was developed considering denitrification as a two-step process, occurring by the consecutive reduction of nitrates to nitrites and then to nitrogen gas without accumulation of intermediate gaseous products. The kinetics of growth of hydrogen oxidation bacteria in batch reactors were assumed to be subject to double nutrient limitation by nitrate and nitrite and inhibition by nitrate. Thus a model of substitutable substrates (nitrate and nitrite) with nitrate inhibition is proposed to describe hydrogenotrophic denitrification. Nitrate inhibition is modelled by an Andrewss type expression. The balance equations for a batch reactor are as follows: Balance of biomass: (1)
dX dt rmax1 N1 Ks N1 kd 2 N2 N1 Ki
2

Balance of nitrate: (2)

dN1 1 X dt Ys

rmax1 N1 K s N1 k d 2 N 2 N1 Ki
2

Balance of nitrite: (3)

where: N1 is the nitrate nitrogen concentration (mg/l), N2 is the nitrite nitrogen concentration (mg/l), rmax1=0.0485 1/h and rmax2=0.55 1/h are the maximum specific growth rates of nitrate and nitrite, respectively, Ks=28.63 mg/l and Kn=4.79 mg/l is the saturation constant for nitrate and nitrite, respectively, Ys=0.4207 mgbiomass/mgNO3--N is the growth yield coefficient on nitrate, Yn=0.084 mgbiomass/mgNO2--N is the growth yield coefficient of nitrite, kd1=15.055 mgNO2--N/mgNO3--N is a constant in growth rate expression, and kd2=13.18 mgNO3--N/mgNO2--N is a constant in growth rate expression, Ki=24.284 mg/l is the nitrate inhibition constant, and kd=0.00003 1/h is the death rate constant [18]. Figure 2 shows the dependence of the efficiency of the mixed culture (nitrate removal rate) on nitrate concentration as indicated by experimental data (symbols) and model prediction (line). Nitrate removal rate values were estimated from Eq. 2. It is evident, that developed model, with the nitrate inhibition expression, is capable of describing the inhibitory effect of the substrate to the removal rate of the mixed culture.

3.1.2 Draw-fill experiment The mixed microbial culture was used to investigate the maximum nitrate removal efficiency, with initial nitrate concentration of 80 mgNO3--N/l. As soon as the nitrate degradation was complete, 300 ml of the liquid were discarded. The flask content was brought to its original value of 2 l with tap water together with a concentrated stock solution of nitrate and buffers so that the desired final nitrate concentration of 80 mgNO3--N/l in the flask was achieved (Figure 3). These cycles were repeated until maximum removal and growth rate of nitrate and biomass, respectively, was recorded for many cycles. As shown in Figure 3, after 14 days of denitrification the system reached steady state. The term steady state, refers to the mean concentrations during a cycle (the system reaches a steady cycle). The duration of minimum stable cycles was 26 h. The maximum denitrification rate achieved was 75 gNO3--N/m3d and the growth rate of biomass was 32.6 gbiomass/m3d. Consequently, a rate of nitrate nitrogen consumption of 2.31 gNO3--N/gbiomass was achieved. Nitritenitrogen concentration was insignificant. Experimental profiles of nitrate, nitrite and biomass concentration (Figure 3), which comprise the three last cycles of the draw-fill experiment (steady state), were used to validate the model. The behaviour of the system predicted by the model appears to be in very good agreement with the experimental data.
1.8 1.6 1.4 1.2 1.0 0.8
200 180 160 140 120 100 80 60 40 20 0

NO3 -N NO2 -N Biomass

model

Denitrification rate (mg/Lh)

Concentration (mg/L)

experimental model

model

experimental
0.6 0.4 0.2 0.0 0 20 40 60

NO3-N Concentration (mg/L)

80

100

120

140

160

180

200

50

100

150

Time (hr)

200

250

300

350

400

450

Figure 2. The dependence of the efficiency of the mixed culture (nitrate removal rate) on nitrate concentration.

Figure 3. Evolution of nitrate-, nitrite-nitrogen and biomass concentration with time for the draw-fill experiment for the flask reactor and the model prediction.

3.2 Bioreactor experiment 3.2.1 Bioreactor acclimation During an 85-day start-up period, filters were operated as batch systems (data not shown) to ensure attachment of the bacterial culture to the support media and development of a biofilm layer. Hydrogenotrophic culture (after the draw-fill experiment) was transferred to the bioreactors. As soon as nitrate degradation was complete, 20 ml of fresh synthetically produced contaminated water, which contained tap water, KNO3 (0.29 g/l), KH2PO4 (3.39 g/l), Na2HPO4 (3.53 g/l), was added to the reactors. After approximately 85 days, a thin layer of biofilm was visible covering the surface of the gravels.

3.2.2 Draw-fill operation After the start-up period draw-fill experiments took place in the bioreactors, the filters were loaded with synthetically produced contaminated water to a final concentration of NO3--N (from 6 to 210 mgNO3--N/l). As soon as the nitrate degradation was complete, 240 ml of a concentrated stock solution of nitrate and buffers were added in the filters. These cycles were repeated until maximum removal rate of nitrate, was recorded for many cycles, for every concentration of nitrate. Figures 4a shows the maximum denitrification rates achieved for various feed nitrate concentrations from 6 to 210 mgNO3--N/l for the filter with gravel of 1.75 mm diameter. Nitrite-nitrogen concentration was insignificant. It is observed that, any increase of the feed nitrate concentration beyond 40 mg/l reduces drastically the nitrate removal rate (Figure 4a) for the filter, although complete degradation is eventually achieved. The removal rate decreased as nitrate feed concentration increased due to the inhibitory effect of nitrates [18] (Figure 2). A comparison between Figures 3 and 4a shows clearly the dramatically reduction of the operating cycle duration using the bench-scale packed-bed reactor with gravel of 1.75 mm diameter. For example, almost 26 h are needed to degrade about 80 mgNO3--N/l in the flask at 30oC (Figure 3), while for similar nitrate feed concentration in the filter, this time is reduced to only 0.83 h at 24oC (Figure 4a). Consequently, the nitrate removal rate was higher in the packed bed reactors than in the suspended-growth reactor. Particularly, the removal rate hardly reached 75 gNO3--N/m3 d, for feed nitrate concentration 80 mgNO3--N/l, in the flask, while for the same feed concentration, the removal rate reached was 2260 gNO3--N/m3d in the packed-bed reactor. The above observations are the result of the existence of high density of biomass in the attached growth process, which results in a higher nitrate removal rate compared to the removal rate of the batch process.
3500

50

d=2.41mm d=1.75mm

3000

Denitrification rate (g/m d)

40

NO3-N (mg/L)

2500

2000

30

1500

20

1000

10
500

0 0 50 100
-

150

200

0 0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

Concentration NO3 -N (mg/L)

Time (hr)

(a)

(b)

Figure 4. a) The dependence of the efficiency (nitrate removal rate) of the bioreactor with gravel d=1.75mm on nitrate concentration and b) operating cycles of the bioreactors with gravel of 1.75 mm and 2.41 mm diameter under draw-fill operation for feed nitrate concentrations 40 mgNO3--N/l. Finally, the draw-fill operation of the bioreactor with gravel of 2.41 mm diameter for feed nitrate concentration of 40 mgNO3--N/l (Figure 4b) was tested. The maximum denitrification rate achieved was 4060 gNO3--N/m3d. Duration of minimum stable cycles was 0.23 h. The removal rate of bioreactor with gravel of 1.75 mm diameter is lower (duration of stable cycles was 0.33 h) than this of

the bioreactor with gravel of 2.41 mm diameter (at 24oC), due to the better circulation of the gases (H2, CO2) in the reactor (Figure 4b). 4. CONCLUSIONS The mixed cultures used were found to support efficiently potable water denitrification since high denitrification rates were achieved. The maximum denitrification rate achieved in the flask reactor for feed nitrate concentration of 80 mgNO3--N/l was 75 gNO3--N/m3d. The mathematical model from our previous study adequately describes the hydrogenotrophic denitrification in suspendedgrowth reactor (flask). Two gravel sizes were used as a support medium of the reactors. It was observed, that the denitrification rate increased as increases size of the support medium of the reactor, probably due to the better circulation of the gases (H2, CO2) in the reactor. The maximum denitrification rate achieved for feed nitrate concentration of 40 mgNO3--N/l was 4060 gNO3--N/m3 d for the filter with gravel of 2.41 mm diameter. The filter with gravel of 1.75 mm diameter achieved for feed nitrate concentration of 40 mgNO3--N/l a denitrification rate of only about 2930 gNO3--N/m3d. In addition, we confirmed the inhibitory effect of nitrates on the bioreactors performance. The initial removal rate of the packed-bed reactor with specific surface area 3207 m2/m3 (mean diameter 1.75 mm), decreased from 2930 to 1660 gNO3--N/m3d, as nitrate feed concentration increased from 40 to 210 mgNO3--N/l. Finally, the nitrate removal rate was higher in the packed bed reactors than in the suspended-growth reactor. Particularly, the removal rate hardly reached 75 gNO3--N/m3 d, for feed nitrate concentration 80 mgNO3--N/l, in the flask, while for the same feed concentration, the removal rate reached was 2260 gNO3--N/m3d in the packed-bed reactor. Attached growth processes provide resistance to high nitrate concentration and lead to higher removal rates, than in the case of the suspendedgrowth process. The present work is an attempt for a better understanding of the hydrogenotrophic denitrification process. Draw-fill operation of the bench-scale packed bed reactors proved to be a very effective operating mode, since it ensures extremely high nitrate removal rates. The limitation at high nitrate concentration is of great importance and should always be taken into account in the design of denitrifying reactors. The high nitrate reduction rates indicate that this technology may offer a feasible solution to a very serious environmental problem. REFERENCES 1. 2. Nolan B.T., B.C. Ruddy, K.J. Hitt, D.R. Helsel (1998) A national look at nitrate contamination of ground water WaterCond. Purification, Vol. 39, pp. 76. Urbain V., R. Benoit, J. Manem (1996) Membrane bioreactor: a new treatment tool J Am Water Works Assoc, Vol. 5, pp. 75-85. Haugen K.S., M.J. Semmens, P.J. Novak (2002) A novel in situ technology for the treatment of nitrate contaminated ground water Water Research, Vol. 36, pp. 34973506. Kesseru P., I. Kiss, Z. Bihari, B. Polyak (2003) Biological denitrification in a continuousflow pilot bioreactor containing immobilized Pseudomonas butanovora cells Bioresource Technology, Vol. 87, pp. 75-80.

3.
4.

5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

Chang C.C., S.K. Tseng, H.K. Huang (1999) Hydrogenotrophic denitrification with immobilized Alcaligenes eutrophus for drinking water treatment Bioresource Technology Vol. 69 pp.53-58. Soares M.I.M. (2000) Biological denitrification of groundwater Water, Air, and Soil Pollution, Vol. 123, pp. 183-193. Liu L.H., A. Koenig (2002) Use of limestone for pH control in autotrophic denitrification: batch experiments Process Biochemistry, Vol. 37, pp. 885893. Lee K.C., B.E. Rittmann (2002) Applying a novel autohydrogenotrophic hollow-fiber membrane biofilm reactor for denitrification of drinking water Water Research, Vol. 36, pp. 20402052. Ergas S.J., A.F. Reuss (2001) Hydrogenotrophic denitrification of drinking water using a hollow fibre membrane bioreactor Journal of Water supply: Research and TechnologyAqua, Vol. 50 pp. 161-171. Ho C.M., S.K. Tseng, Y.J. Chang (2001) Autotrophic denitrification via a novel membraneattached biofilm reactor Letters in Applied Microbiology, Vol. 33, pp. 201-205. Mansell B.O., E.D. Schroeder (2002) Hydrogenotrophic denitrification in a microporous membrane bioreactor Water Research Vol. 36, pp. 4683-4690. Szekeres S., I. Kiss, T.T. Bejerano, M.I.M. Soares (2001) Hydrogen-dependent denitrification in a two-reactor bio-electrochemical system Water Research, Vol. 35 pp. 715-719. Kurt M., I.J. Dunn, JR Bourne (1987) Biological denitrification of drinking water using autotrophic organisms with H2 in a fluidized-bed biofilm reactor Biotechnology and Bioengineering, Vol. 29, pp. 493-501. Dries D., J. Liessens, W. Verstraete, P. Stevens, P. Vost, J. Ley (1988) Nitrate removal from drinking water by means of hydrogenotrophic denitrifiers in a polyurethane carrier reactor Wat. Supply, Vol. 6, pp. 181-192. Gross H., G. Schnoor, P. Rutten (1988) Biological denitrification process with hydrogenoxidizing bacteria for drinking water treatment Wat. Supply, Vol. 6, pp. 193-198. Sakakibara Y., J.R.V. Flora, M.T. Suidan, M. Kuroda (1994) Modeling electrochemicallyactivated denitrifying biofilms Water Research, Vol. 28, pp. 1077-1086. Rezania B., N. Cicek, J.A. Oleszkiewicz (2005) Kinetics of Hydrogen-Dependent Denitrification under varying pH and temperature conditions Biotechnology and Bioengineering, Vol. 92, pp. 900-906. Vasiliadou I.A., S. Pavlou, D.V. Vayenas (2006) A Kinetic Study of Hydrogenotrophic Denitrification Process Biochemistry (in press). APHA, AWWA and WPCF, 1989. Standard Methods for the Examination of Water and Wastewater, 17th edition. American Public Health Association, American Water Works Association and Water Pollution Control Federation, Washington, D.C.