DTD 5
Aquacultural Engineering xxx (2005) xxx–xxx www.elsevier.com/locate/aqua-online
Removal of nitrate in aquaria by means of electrochemically generated hydrogen gas as electron donor for biological denitrification R. Grommen a, M. Verhaege b, W. Verstraete a,* a
Laboratory of Microbial Ecology and Technology (LabMET), Ghent University, Coupure Links 653, B-9000 Gent, Belgium b Laboratory of Non-ferrous Metallurgy and Electrometallurgy, Ghent University, Technologiepark 9, B-9052 Zwijnaarde, Belgium Received 7 May 2004; accepted 30 March 2005
Abstract A hydrogenotrophic denitrification reactor was designed for the removal of nitrate from aquaria. An average hydrogen gas transfer up to 130 mg per day from the gas to the water phase was accomplished by recirculating the water from the denitrification reactor over a separate trickling filter column with a volume of 1.3 l. During batch experiments removal rates up to 36 mg N/l reactor per day were recorded at a hydraulic residence time of 12 h. To avoid the need for storage of large volumes of hydrogen gas in aquarium or aquaculture applications, an electrochemical cell was used to generate hydrogen gas. During a 7 day aquarium test, a nitrate removal rate up to 18.5 mg N/l reactor per day was recorded at an influent NO3–N concentration of 20 mg/l. # 2005 Published by Elsevier B.V. Keywords: Nitrite; Aquarium; Hydrogenotrophic; Electrolysis
1. Introduction In aquaria and recirculating aquaculture systems, control of the dissolved oxygen concentration and removal of toxic ammonia are the main objectives of the water treatment * Corresponding author. Tel.: +32 9 264 59 76; fax: +32 9 264 62 48. E-mail address:
[email protected] (W. Verstraete). URL: http://labmet.ugent.be 0144-8609/$ – see front matter # 2005 Published by Elsevier B.V. doi:10.1016/j.aquaeng.2005.03.007
DTD 5
2
R. Grommen et al. / Aquacultural Engineering xxx (2005) xxx–xxx
system (Meade, 1985). The former can be achieved by gas exchange, while the latter is obtained by means of the nitrification process. The end product of the nitrification process, the nitrate ion, tends to accumulate in aquaria and closed recirculating aquaculture systems. Although the median lethal concentration values (LC50) for nitrate–nitrogen are typically a factor 1000– 10,000 larger than those for ammonia–nitrogen (Wajsbrot et al., 1993; PersonLeRuyet et al., 1997; Tilak et al., 2002) concerns about negative long-term effects of nitrate concentrations over 21 mg N/l exist (Spotte, 1979). Hrubec et al. (1996) noticed a decreased antibody response to Aeromonas salmonicida in sunshine bass at nitrate concentrations of 45 mg N/l. Removal of nitrate can be accomplished by exchanging a fraction of the water in the system with water low in nitrate. Because of the cost of large water exchanges, especially for systems, which use artificial seawater, and because of legislative restrictions on effluent discharges, this approach cannot always be applied (Grguric et al., 2000). In closed systems, nitrate removal can be accomplished by the process of biological denitrification in which nitrate is reduced to gaseous nitrogen products, which are released to the atmosphere (Payne, 1973). Traditionally, organic electron donors, such as methanol, are used for this purpose (Grguric et al., 2000). Facultative anaerobic bacteria will use nitrate as terminal electron acceptor under anoxic conditions. This process must be carefully controlled as overdosing of the organic electron donor can lead to severe water quality problems, such as the formation of toxic hydrogen sulphides (Lee et al., 2000). To overcome the need of dosing of an organic electron donor, the use of biodegradable polymers was suggested, in which the biopolymer acts as biofilm carrier and organic carbon source (Boley et al., 2000). Tal et al. (2003) proposed the use of freeze-dried alginate-starch pellets containing immobilised denitrifying bacteria. The organic matter, which naturally accumulates in recirculating fish culture systems, has also been used as electron donor for denitrification reactors (van Rijn and Rivera, 1990). The performance of this type of denitrification reactors could be enhanced by incorporating an anoxic sedimentation basin (Arbiv and van Rijn, 1995). Elemental sulphur has been used as electron donor for autotrophic denitrification (Kuai and Verstraete, 1999), but has some disadvantages, such as the consumption of alkalinity and the production of sulphates. Hydrogen gas is a safe alternative to organic electron donors and elemental sulphur, as it is not toxic and it does not give rise to unwanted by-products (Haugen et al., 2002). Furthermore, it is not expensive and it generates 50% less microbial biomass than traditional electron donors, such as methanol (Rutten and Schnoor, 1992). In this study, the use of hydrogen gas in a simple and robust denitrification reactor for use in aquaria and aquaculture systems is examined.
2. Materials and methods 2.1. Test set-up All experiments were carried out in 70 l aquariums provided with two internal, air driven, submerged biofilters, each with a volume of 0.9 l. The biofilters were inoculated
DTD 5
R. Grommen et al. / Aquacultural Engineering xxx (2005) xxx–xxx
3
with a nitrifying culture (Grommen et al., 2002) 1 week prior to the start of the experiments. The water temperature was kept at 24 1 8C by means of a thermostatic heater of 100 W. For the denitrification experiments, a 70 l aquarium was connected with the denitrification reactor by means of a peristaltic pump (Watson Marlow type 101 U) with a controllable flow. 2.2. Denitrification reactor Fig. 1 gives a schematic representation of the denitrification reactor. It consisted out of two parts: a cylindrical reactor with a total volume of 1.3 l filled with hollow ceramic cylinders of 2 cm length and 1 cm diameter and an EHEIM canister filter with a volume of 6 l filled with reticulated polyurethane sponge (10 pores/cm; Recticel, Belgium). The total water volume in the denitrification reactor was 6.5 l. The outlet of the canister filter was attached to the top of the cylindrical reactor so that the latter operated as a trickling filter with a flow rate of 300 l/h (231 l/l reactor per hour). A tube at the bottom of the cylindrical reactor connected back to the canister filter. The inlet and the outlet for the water from the aquarium were situated, respectively, near the top and the bottom of the cylindrical reactor. Hydrogen gas was supplied through the top of the cylindrical reactor. To avoid the accumulation of nitrogen gas in the headspace of the trickling filter, an exhaust tube was installed through the top of the cylindrical reactor. The outlet of this tube was situated in a container with 15 cm water to avoid oxygen gas from entering the reactor.
Fig. 1. Scheme of the denitrification reactor used in the study. A trickling filter column of 1.3 l is connected with a canister filter with a volume of 6 l. The influent enters the reactor near the top of the trickling filter column and the effluent leaves the reactor near the bottom of the column. Hydrogen gas produced in an electrochemical cell enters the trickling filter column at the top, which also contains an outlet to remove excess gas.
DTD 5
4
R. Grommen et al. / Aquacultural Engineering xxx (2005) xxx–xxx
2.3. Batch experiment Pulses of NaNO3 were added to the water of a 70 l aquarium to give a final concentration of 20 mg NO3–N/l. Preliminary tests had shown that a hydraulic retention time of 8–10 h is required for complete denitrification. Therefore, the influent flow rate to the denitrification reactor was initially set at 2.0 l/l denitrification reactor per day (13.0 l/day) and was increased on day 3 to 2.7 l/l denitrification reactor per day (17.6 l/day). Water samples from the inlet and outlet of the denitrification reactor were analyzed daily for total ammoniacal nitrogen (TAN = NH3–N + NH4+–N), NO2–N and NO3–N. 2.4. Aquarium test The water quality in two aquaria of 70 l, each with 15 adult Xiphophorus maculatus, was compared. One aquarium was connected to the denitrification reactor the other (control) had no denitrification system installed. Preliminary experiments indicated that no measurable denitrification could be observed in the denitrification reactor without the addition of an electron donor. Hydrogen gas, produced in an electrochemical cell, was added continuously to the denitrification reactor. The flow rate through the denitrification reactor averaged 1.5 0.7 l/l reactor per day during the first run and 3.2 0.5 l/l reactor per day during the second run (hydraulic retention time; HRT = 16 and 7.5 h, respectively). Twice daily, 0.2 g of a commercial flake food (Tetra-Min) containing 46% protein was added to each aquarium (5.7 mg/l per day). This corresponds with an estimated addition of 0.42 mg N/l per day. 2.5. Electrochemical cell Hydrogen gas was generated in a two-compartment electrolytic cell (Lissens et al., 2003) containing two plain perforated nickel electrodes. A 20% KOH solution was used as electrolyte. The cell was operated under galvanostatic mode between 0.3 and 0.5 A at a minimum potential of 1.5 V. Gas production was monitored using an electronic gas counter (Milligascounter, Fachhochschule Bergedorf, Hamburg, Germany). 2.6. Chemical analyses Nitrite, nitrate and phosphate were determined with an ion chromatograph (IC 761 Compact, Metrohm) with a metrosep A supp 5 column (Methrohm). The eluent was 3.2 mM Na2CO3 and 1 mM NaHCO3 at a flow of 0.7 ml/min. TAN was determined using the direct photometric method with the Nessler reagent (Greenberg et al., 1992) and measured with a UVIKON 932 spectrophotometer (Kontron Instruments, Switzerland). Other parameters analyzed daily were the dissolved oxygen (DO) concentration, the temperature and the pH of the water. The DO was measured by means of a membrane covered amperometric electrode (COS 381 oxygen probe with a COM 381 meter) (Endress–Hauser, Belgium).
DTD 5
R. Grommen et al. / Aquacultural Engineering xxx (2005) xxx–xxx
5
2.7. Calculations The theoretical amount of hydrogen gas produced during a certain period of time in the electrochemical cell was calculated as follows: H2 ðgÞ ¼ I t ðe AÞ1 with I = current (A); t = time (s); e = elementary charge of the electron (1.602 1019 C) and A = Avogadro’s number (6.02 1023). The amount of hydrogen gas used for denitrification and for removal of oxygen was calculated as follows: H2 ðgÞ ¼ DN 2:5 ð14Þ1 þ f t DO 2 ð32Þ1 with DN = amount of nitrogen removed during the time period (g); f = flow rate of aquarium water entering the denitrification reactor (litre per second) and DO = oxygen concentration in the water entering the denitrification reactor (g/l). The minimum amount of hydrogen gas transferred from the gas phase to the water phase in the trickling filter column per day was calculated as the amount of hydrogen gas used for denitrification and for oxygen removal divided by the time period.
3. Results 3.1. Batch test Fig. 2 shows the evolution of the concentrations of NO3–N and NO2–N of the influent and the effluent of the denitrification reactor as a function of time. At the start of the experiment and on days 3 and 6, NaNO3 was added to the aquarium to give a final concentration of 20 mg NO3 N/l. During the first 3 days, the flow rate through the denitrification reactor was set at 2.0 l/l denitrification reactor per day. At this hydraulic
Fig. 2. Evolution of the NO2–N and the NO3–N concentrations of the influent and the effluent of the denitrification reactor as a function of time. At the beginning and on days 3 and 6 a pulse of nitrate was added to the aquarium. Influent NO3–N (^), effluent NO3–N ( ), influent NO2–N ( ), effluent NO2–N ( ).
DTD 5
6
R. Grommen et al. / Aquacultural Engineering xxx (2005) xxx–xxx
retention time of 12 h, only during the first day nitrite could be measured in the effluent of the denitrification reactor. This resulted in a small accumulation of nitrite in the aquarium, which reached a maximum of 0.24 mg N/l on day 1. Later on during the first pulse, the nitrite and nitrate concentrations in the effluent of the denitrification reactor were below the detection limit of the ion chromatograph (0.05 mg/l). From day 3 onwards the flow through the denitrification reactor was increased to 2.7 l/l denitrification reactor per day (HRT = 9 h). After the addition of a second pulse of nitrate, an increase of the NO2–N concentration in the effluent could be measured, reaching a maximum concentration of 4.13 mg N/l on day 4. The NO2–N concentration in the aquarium always remained below 0.3 mg N/l. The average nitrate removal rate for the first pulse of nitrate was 28.4 mg N/l reactor per day and for the second 28.1 mg N/l reactor per day. A maximum removal rate of 36 mg N/l reactor per day was recorded at a hydraulic residence time of 12 h and a NO3–N concentration of 20 mg N/l. The current through the electrochemical cell was set at 0.5 A during the first 6 days of the batch experiment. Preliminary tests showed that the electrochemical cell was capable of reaching current to hydrogen gas efficiencies of at least 95%. Therefore, it was possible to calculate that during the first pulse of 2.9 days and the second pulse of 2.73 days, respectively, 1.3 and 1.2 g hydrogen gas were provided to the denitrification reactor. Based on the amount of nitrate removed and the total volume of water pumped through the reactor, we calculated that, respectively, 0.32 g (24%) and 0.36 g (29%) were effectively used during the first and second pulse for both denitrification and dissolved oxygen removal. We assumed that all oxygen gas entering the denitrification reactor was removed before denitrification could start. Indeed, the dissolved oxygen concentration in the outlet of the denitrification reactor was found to be between 0.5 and 0.1 mg/l. Therefore, we can also conclude that, respectively, at least 109 and 130 mg hydrogen gas per day were transferred from the gas to that water phase during the first and the second pulse. On day 6, the current through the electrochemical cell was decreased to 0.4 A and a third pulse of nitrate was added to the aquarium. The HRT was kept at 9 h. Both the nitrate removal rate of 27 mg N/l reactor per day and the amount of hydrogen gas used (0.36 g) over a 3-day period were similar to the previous pulses. The relative amount of hydrogen gas used for denitrification and dissolved oxygen removal increased to 36%, while also the maximum NO2–N concentration in the effluent increased to 6.9 mg N/l. TAN was not detected in the effluent of the denitrification reactor. The pH of the influent increased from 8.4 on day 1 to 8.7 on day 9, while the effluent of the denitrification reactor reached pH 9.1 on day 9. 3.2. Aquarium test The denitrification reactor was attached to a 70 l aquarium containing 15 Xiphophorus maculatus. No mortalities occurred during the time span of the experiment (50 days). The water quality in terms of TAN, NO2–N and NO3–N concentrations was compared with that of an aquarium without denitrification reactor. Fig. 3 shows the total oxidized nitrogen concentrations (TON = NO2–N + NO3–N) of the control aquarium and the aquarium with the denitrification reactor. The TAN and NO2–N concentrations in both aquaria
DTD 5
R. Grommen et al. / Aquacultural Engineering xxx (2005) xxx–xxx
7
Fig. 3. Evolution of the TON concentration of a control aquarium (^) and an aquarium with denitrification reactor ( ) as a function of time.
never exceeded 0.05 and 0.1 mg N/l, respectively. The NO3–N concentration in the control aquarium increased at a steady rate of approximately 0.23 mg N/l per day, while the NO3–N concentration in the aquarium with the denitrification reactor decreased at an average rate of 0.07 mg N/l per day. Assuming that the nitrogen excretion rate in the control aquarium and the denitrification aquarium was similar, this implies a nitrate removal rate of 0.30 mg N/l per day or 3.2 mg N/l reactor per day at an average NO3–N concentration of 3 mg N/l. A second test was performed to test the reactor at higher nitrate concentrations. Therefore, the control aquarium from the previous experiment was connected with the denitrification reactor, while the aquarium that was connected in the previous experiment acted as control. Fig. 4 shows the TON concentrations of the control and the aquarium with denitrification reactor as a function of time. The nitrate concentration in the control aquarium increased with 0.29 mg N/l per day, while the nitrate concentration in the aquarium with the denitrification reactor decreased at an average rate of 1.41 mg N/l per day. This corresponds with an average removal rate of 18.5 mg N/l reactor per day. A mass balance of hydrogen gas over a 7-day period during the second test showed that 0.15 g H2 was used to remove oxygen gas from the influent of the denitrification reactor, while 0.25 g H2 was used for denitrification.
Fig. 4. Evolution of the TON concentration of a control aquarium (^) and an aquarium with denitrification reactor as a function ( ) of time with a high initial nitrate concentration.
DTD 5
8
R. Grommen et al. / Aquacultural Engineering xxx (2005) xxx–xxx
4. Discussion The use of electrochemically generated hydrogen gas in a trickling filter set-up makes for a simple, easy to operate and robust denitrification reactor for use in aquaria and aquaculture systems. Maximum removal rates of up to 36 mg N/l reactor per day were recorded at influent NO3–N concentration of 20 mg N/l. Safety concerns related with the use of hydrogen gas were addressed by the replacement of a compressed gas cylinder with an electrochemical cell for in situ generation of the hydrogen gas. The electrochemical cell reached efficiencies of greater than 95%. The operating costs of the electrochemical cell from this study were estimated at 1.1 2/m3 of hydrogen gas (1 kWh = 0.16 2). This is about 5 times lower than the cost of hydrogen gas as compressed gas. Previously, nitrate removal rates in the order of 0.17–2.4 g N/l reactor per day were recorded in systems using organic electron donors (Boley et al., 2000). This is a factor 5–70 higher than the results obtained in this study with hydrogen gas as electron donor. In the presence of higher organic loadings, as encountered in recirculating aquaculture systems, it is most likely that higher denitrification rates would be observed, as the dissolved organic matter can also serve as electron donor for denitrification (van Rijn and Rivera, 1990). In this work, a trickling filter system was used to dissolve the hydrogen gas in the recirculating water of the denitrification reactor. Hydrogen gas has a low solubility of maximum 1.52 mg/l at 20 8C and 1 atm (Dries et al., 1988). Transfer rates of 130 mg H2 per day were recorded during the batch test. Nitrate removal from drinking water and groundwater has been accomplished in the past by using hollow-fiber membranes to supply the hydrogen gas (Haugen et al., 2002; Lee and Rittmann, 2002). While these membranes have been shown to be highly effective for transferring the hydrogen gas to the water phase, we encountered problems with CaCO3 precipitation on the membrane surfaces, resulting in a sharp decrease of the flux of hydrogen gas in preliminary experiments (results not shown). Lee and Rittmann (2003) found no adverse effects of solids precipitation on the H2 transfer during short-term experiments (4 h) but warned about long-term effects. In all their experiments, the authors could measure a precipitation of Ca2+ ions. Reducing the HRT from 12 to 9 h led to an accumulation of nitrite in the effluent of the denitrification reactor, indicating that the reduction of nitrite was the rate-limiting step of the denitrification process. Since the accumulation of nitrite becomes more pronounced at alkaline pH (Lee and Rittmann, 2003) and strong base is being produced during the denitrification reaction with hydrogen gas as electron donor, sufficient buffer should be present in the recirculating water. During the batch experiment, the pH of the effluent rose to 9.1 after 9 days of operation, which is above the optimum pH range of 7.7–8.6 for denitrification (Lee and Rittmann, 2003). In aquaria and recirculating aquaculture systems, the base production during the denitrification process can be balanced by the acid production, which results from the oxidation of ammonia during the nitrification process. The influent of the denitrification reactor in this study contained 8 mg/l of dissolved oxygen. A mass balance calculation for the consumed hydrogen gas over a 7 day period showed that about 40% of the hydrogen gas was used to remove DO in order to create anoxic conditions for the denitrification process. Alternatively, the influent sample could be sparged with nitrogen gas to make it anoxic before entering the denitrification reactor (Lee et al., 2000; Menasveta et al., 2001). This would lead to a reduction of both the hydrogen
DTD 5
R. Grommen et al. / Aquacultural Engineering xxx (2005) xxx–xxx
9
gas consumption and the minimum hydraulic retention time of the system. Yet, this would require the provision of a bottle with compressed nitrogen gas, which is more complicated than a single provision of hydrogen gas by means of an electrochemical generator. Moreover, nitrogen gas is two times more costly than hydrogen gas produced electrochemically. To prevent the accumulation of nitrogen gas in the trickling filter column, hydrogen gas was provided in excess (2.0–2.5 g H2/g N), so that the former could be flushed out continuously. In this way, the cost of the electron donor is in the order of 25 Euro/kg of NO3–N. When stoechiometric amounts of hydrogen gas are supplied, the cost would be 10 times lower and comparable to the cost of methanol as electron donor (Boley et al., 2000). Alternatively, nitrogen gas can be flushed out discontinuously, to minimize hydrogen gas consumption and leakage.
5. Conclusions Using electrochemically generated hydrogen gas as electron donor for biological denitrification, nitrate removal rates up to 36 mg N/l reactor per day were achieved at a nitrate concentration of 20 mg N/l and a hydraulic retention time of 12 h. The combination of an electrochemical cell and the denitrification reactor functioned in a reliable way during an aquarium test of 50 days.
Acknowledgements This research was supported by a doctoral scholarship of the Flemish Institute for the Improvement of Scientific Technological Research in the Industry (IWT-Grant number 001194). We want to thank Professor Mark Verhaege for providing the electrochemical cell. A special thank to the reviewers (ir. Wim Dewindt, ir. Geert Lissens and ir. Korneel Rabaey) for critically reading the manuscript.
References Arbiv, R., van Rijn, J., 1995. Performance of a treatment system for inorganic nitrogen removal in intensive aquaculture systems. Aquacult. Eng. 14, 189–204. Boley, A., Mu¨ ller, W.-R., Haider, G., 2000. Biodegradable polymers as solid substrate and biofilm carrier for denitrification in recirculated aquaculture systems. Aquacult. Eng. 22, 75–85. Dries, D., Liessens, J., Verstraete, W., Stevens, P., De Vos, P., De Ley, J., 1988. Nitrate removal from drinking water by means of hydrogenotrophic denitrifiers in a polyurethane carrier reactor. Water Supply 6, 181–192. Greenberg, A.E., Clesceri, L.S., Eaton, A.D., 1992. Standard Methods for the Examination of Water and Wastewater. APHA, Washington. Grguric, G., Wetmore, S.S., Fournier, R.W., 2000. Biological denitrification in a closed seawater system. Chemosphere 40, 549–555. Grommen, R., Van Hauteghem, I., Van Wambeke, M., Verstraete, W., 2002. An improved nitrifying enrichment to remove ammonium and nitrite from freshwater aquaria systems. Aquaculture 211, 115–124.
DTD 5
10
R. Grommen et al. / Aquacultural Engineering xxx (2005) xxx–xxx
Haugen, K.S., Semmens, M.J., Novak, P.J., 2002. A novel in situ technology for the treatment of nitrate contaminated groundwater. Water Res. 36, 3497–3506. Hrubec, T.C., Robertson, J.L., Smith, S.A., Tinker, M.K., 1996. The effect of temperature and water quality on antibody response to Aeromonas salmonicida in sunshine bass (Morone chrysops X Morone saxatilis). Vet. Immunol. Immunopathol. 50, 157–166. Kuai, L., Verstraete, W., 1999. Autotrophic denitrification with elemental sulphur in small-scale wastewater treatment facilities. Environ. Technol. 20, 201–209. Lee, K.C., Rittmann, B.E., 2002. Applying a novel autohydrogenotrophic hollow-fiber membrane biofilm reactor for denitrification of drinking water. Water Res. 36, 2040–2052. Lee, K.C., Rittmann, B.E., 2003. Effects of pH and precipitation on autohydrogenotrophic denitrification using the hollow-fiber membrane-biofilm reactor. Water Res. 37, 1551–1556. Lee, P.G., Lea, R.N., Dohmann, E., Prebilsky, W., Turk, P.E., Ying, H., Whitson, J.L., 2000. Denitrification in aquaculture systems: an example of a fuzzy logic control problem. Aquacult. Eng. 23, 37–59. Lissens, G., Verhaege, M., Pinoy, L., Verstraete, W., 2003. Electrochemical decomplexing and oxidation of organic (chelating) additives in effluents from surface treatment and metal finishing. J. Chem. Technol. Biotechnol. 78, 1054–1060. Meade, J.W., 1985. Allowable ammonia for fish culture. Prog. Fish-Cult. 47, 135–145. Menasveta, P., Panritdam, T., Sihanonth, P., Powtongsook, S., Chuntapa, B., Lee, P., 2001. Design and function of a closed, recirculating seawater system with denitrification for the culture of black tiger shrimp broodstock. Aquacult. Eng. 25, 35–49. Payne, W.J., 1973. Reduction of nitrogenous oxides by microorganisms. Bacteriol. Rev. 37, 409–452. PersonLeRuyet, J., Galland, R., LeRoux, A., Chartois, H., 1997. Chronic ammonia toxicity in juvenile turbot (Scophthalmus maximus). Aquaculture 154, 155–171. Rutten, P., Schnoor, G., 1992. Five years’ experience of nitrate removal from drinking water. Water Supply 10, 183–190. Spotte, S., 1979. Seawater Aquariums. Wiley, New York. Tal, A., Nussinovitch, A., van Rijn, J., 2003. Nitrate removal in aquariums by immobilized Pseudomonas. Biotechnol. Prog. 19, 1019–1021. Tilak, K.S., Lakshmi, S.J., Susan, T.A., 2002. The toxicity of ammonia, nitrite and nitrate to the fish, Catla catla (Hamilton). J. Environ. Biol. 23, 147–149. van Rijn, J., Rivera, G., 1990. Aerobic and anaerobic biofiltration in an aquaculture unit nitrite accumulation as a result of nitrification and denitrification. Aquacult. Eng. 9, 217–234. Wajsbrot, N., Gasith, A., Diamant, A., Popper, D.M., 1993. Chronic toxicity of ammonia to juvenile gilthead seabream sparus-aurata and related histopathological effects. J. Fish Biol. 42, 321–328.
