Click here to load reader

Effects of Temperature on Tertyare Nitrification in Moving-bed Biofilm Reactors

  • View

  • Download

Embed Size (px)

Text of Effects of Temperature on Tertyare Nitrification in Moving-bed Biofilm Reactors


40 (2006) 2981 2993

Available at

journal homepage:

Effects of temperature on tertiary nitrication in moving-bed biolm reactorsRoberta Salvetti, Arianna Azzellino, Roberto Canziani, Luca BonomoD.I.I.A.R. Environmental Engineering Department, Politecnico di Milano, Technical University of Milan, Leonardo da Vinci, 32-20133 Milano, Italy

art i cle infoArticle history: Received 31 August 2005 Received in revised form 4 May 2006 Accepted 9 May 2006 Available online 13 July 2006 Keywords: Nitrication Temperature Pure-oxygen moving-bed biolm reactor (PO-MBBR) Ammonia limitation Oxygen limitation Multivariate techniques


The effect of wastewater temperature on the rate of nitrication was studied in two pureoxygen moving-bed biolm reactors, fed on secondary efuent from a municipal wastewater treatment plant. The rst Reactor (R1) was operated under ammonia-limiting conditions, while the second Reactor (R2) was operated under oxygen-limiting conditions. Quite surprisingly, the former showed a negligible inuence of thermal changes on nitrication rates, while the latter showed a much higher dependence. In this paper, a temperature coefcient y has been dened as the actual intrinsic biological temperature coefcient, similar to the corresponding coefcient that is usually adopted for the design of activated-sludge processes. In addition, an apparent coefcient ya has been quantied independently, which was calculated according to the actual values of nitrication rates at different temperatures. The actual biological temperature coefcient y, ranged between 1.086 and 1.109 (average value 1.098) under ammonia-limiting conditions, while under oxygen-limiting conditions was in the range 1.0231.081 (average value 1.058). The apparent value ya was near to unity (i.e. no temperature effect) under ammonia-limiting conditions, while only under oxygen-limiting conditions and at constant dissolved oxygen concentration ya coincided with y. An explanation was given that, under oxygen-limiting conditions, the specic biomass activity (i.e. the ratio of nitrication rate to biomass concentration) was strongly inuenced by the combined effects of oxygen penetration through the biolm and efuent temperature. & 2006 Elsevier Ltd. All rights reserved.



Moving-bed biolm processes have proved to be very reliable for tertiary nitrication because of the high volumetric loading rates that can be applied and the low solids buildup in the reactor. To treat a given volume of wastewater, the capacity of a moving-bed biolm reactor (MBBR) can be smaller than required for a conventional activated-sludge process and, usually, there is no need for a tertiary settling tank. Compared to xed-bed biolm reactors (biolters), MBBRs have much lower headlosses, lter-bed channellingCorresponding author. Tel.: +39 02 23996431; fax: +39 02 23996499.

does not occur (i.e. all the bioreactor volume is used) and periodic backwashing is not needed. Moreover, existing concrete tanks can be equipped and adapted to a MBBR conguration with relatively minor modications. MBBRs are usually lled with low-density (slightly less than 1.0 g cm3) polyethylene KMTs biolm carriers. One such carrier consists of small cylindrical elements 10 mm in diameter and 8 mm in height, with small longitudinal ns that protrude on the outside surface and an internal cross member that divides each element into four circular sectors. The void ratio of the support media is as high as 0.95 with the

E-mail address: [email protected] (R. Salvetti). 0043-1354/$ - see front matter & 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2006.05.013


40 (2006) 2981 2993

effect that, in a tank lled with water and carriers, the volume of water is 95%. The lling ratio is dened as the ratio between the volumes occupied by the carriers, considered as solid cylinders, and the total tank volume. Its maximum value for good mixing is 0.7. The theoretical specic surface area of the support media is dened as the amount of surface area per unit volume of biolm carrier, and for the media described is 700 m2 m3 (degaard and Rusten, 1993). Since the biolm grows in the protected internal faces of the media, the actual specic surface area can be assumed to be about 500 m2 m3 (Hem et al., 1994). In the experiments described in this paper, the actual lling ratio was 0.5, so that the surface area available for the biolm is 250 m2 m3 reactor. The use of pure oxygen instead of air enables higher dissolved oxygen (DO) concentrations to be maintained in the reaction vessel. As a consequence, greater transfer efciency can be achieved and oxygen can diffuse more deeply into the biolm. This produces higher nitrication rates, and, consequently, smaller reactor volumes. If pure oxygen is used in the rst stage, the process can be conveniently divided into two sequential stages, with aeration in the second stage. The rst stage could be operated with a high ammonia concentration and DO would become the rate-limiting substrate. Since zero-order intrinsic kinetics can be assumed for nitrication in biolms, zero-order kinetics can also be assumed with respect to the non-limiting substrate, i.e. ammonia. As far as DO is concerned, it has been found that oxygen can be the reaction-limiting substrate if the ratio of oxygen to ammonia is lower than 2 g O2 (g NH+4 N)1 (Hem et al., 1994) and this may happen even when DO concentration is high (510 mg L1). If the second stage had operated at low ammonia concentration, then the reaction would have shifted to ammonialimiting conditions, which would occur when the oxygen to ammonia ratio is higher than 5 g O2 (g NH+-N)1. Ratios of 4 25 g O2 (g NH+-N)1 should be avoided since transition to 4 ammonia limitation may occur and the process kinetics would then depend on biolm structure and thickness. In the second stage, the very low ammonia concentration in the efuent (0.51.0 mg L1) permits reasonably low DO concentrations to be maintained in completely mixed reactors (2.55 mg L1), so that aeration can also be performed by simple air sparging (Bonomo et al., 2000). The dependence of nitrication on temperature in MBBRs has been investigated in the past. degaard and Rusten (1993) analyzed the dependence of nitrication under oxygen-limiting conditions and did not nd a signicant increase of removal rates at different temperatures. This was apparently in contrast with many previous studies in which a marked effect of temperature on nitrication was described by an Arrhenius-like expression. However, the authors explained that the reason of this discrepancy is due to the fact that, at lower temperature, nitrication rate is certainly reduced, but at the same time the oxygen concentration that can be dissolved in water increases. Therefore, the temperature effect that they were able to observe was masked by the opposite effect due to the increased oxygen concentration. For suspended-growth systems, Painter and Loveless (1983) found a temperature coefcient, y, of 1.076, and similar values are reported by the USEPA (1975) and by Barnes and Bliss

(1983) in the temperature range of 530 1C. The dependence of the reaction rate on temperature was found to be lower than expected for nitrication in xed-lm biolters by Zhu and Chen (2002). In particular, the effect of temperature on the reaction rate was found to be even weaker under oxygen-limiting conditions compared with ammonia-limiting conditions. Popel and Fischer (1998) observed that the effect of temperature on nitrication in suspended-growth systems (namely, activated sludge processes) is often lower than expected from literature data, because other factors, such as reactor conguration, hydraulic residence time (HRT) and efuent concentration, may play an important role in reducing the observed inuence of temperature. This is because removal rates, either in suspended or in xed systems, depend also on the rate-limiting substrate concentration, which is usually a function of the above-cited factors. Hence, the inuence of factors other than temperature on the rate-limiting substrate concentration could mask the observed inuence of temperature on nitrication rates. In the same paper, Popel and Fischer (1998) proposed a distinction between the real temperature coefcient, that describes the dependence of the intrinsic biological process kinetics (y), and the apparent temperature coefcient (ya) that ts the actual operational reaction rates observed in the reactor. They showed also that the latter depends on the type of process and on the conguration of the reactor. In biolm processes diffusional resistances may also contribute to mask the effect of temperature on the intrinsic bacterial reaction rate. Therefore, the aim of the present work is to check whether Popel and Fischers theory can be extended to tertiary nitrication in MBBRs. Multivariate regression analysis was used to quantify the effect of temperature on nitrication rates independently from the operating conditions of the system.


Materials and methods

Two stainless-steel pilot-scale reactors, 1 m3 volume each, have been used. They were fed with the secondary efuent of a wastewater treatment plant (WWTP) equipped with a pureoxygen-patented activated-sludge process (UNOXs); this process produces a low-COD, non-nitried settled secondary efuent (Table 1). The study of the effect of temperature on nitrication was quite convenient with non-nitried efuent from this WWT plant, because the efuent was used as cooling water in a nearby waste-to-energy facility. Throttle valves and owmeters enabled either heated or unheated efuent to be fed to each pilot-scale unit and, therefore, it was

Search related