Nitrogen removal from sludge digester liquids by nitrification/denitrification or partial nitritation/anammox: environmental and economical considerations.

In wastewater treatment plants with anaerobic sludge digestion, 15-20% of the nitrogen load is recirculated to the main stream with the return liquors from dewatering. Separate treatment of this ammonium-rich digester supernatant significantly reduces the nitrogen load of the activated sludge system. Two biological applications are considered for nitrogen elimination: (i) classical autotrophic nitrification/heterotrophic denitrification and (ii) partial nitritation/autotrophic anaerobic ammonium oxidation (anammox). With both applications 85-90% nitrogen removal can be achieved, but there are considerable differences in terms of sustainability and costs. The final gaseous products for heterotrophic denitrification are generally not measured and are assumed to be nitrogen gas (N2). However, significant nitrous oxide (N2O) production can occur at elevated nitrite concentrations in the reactor. Denitrification via nitrite instead of nitrate has been promoted in recent years in order to reduce the oxygen and the organic carbon requirements. Obviously this "achievement" turns out to be rather disadvantageous from an overall environmental point of view. On the other hand no unfavorable intermediates are emitted during anaerobic ammonium oxidation. A cost estimate for both applications demonstrates that partial nitritation/anammox is also more economical than classical nitrification/denitrification. Therefore autotrophic nitrogen elimination should be used in future to treat ammonium-rich sludge liquors.     

Autotrophic nitrogen removal from anaerobic supernatant of Florence's WWTP digesters.

In municipal WWTP with anaerobic sludge digestion, 10-20% of total nitrogen load comes from the return supernatant produced by the final sludge dewatering. In recent years a completely autotrophic nitrogen removal process based on Anammox biomass has been tested in a few European countries, in order to treat anaerobic supernatant and to increase the COD/N ratio in municipal wastewater. This work reports the experimental results of the SHARON-ANAMMOX process application to anaerobic supernatant taken from the urban Florentine area wastewater treatment plant (S. Colombano WWTP). A nitritation labscale chemostat (7.4 L) has been started-up seeded with the S. Colombano WWTP nitrifying activated sludge. During the experimental period, nitrite oxidising bacteria wash-out was steadily achieved with a retention time ranging from 1 to 1.5 d at 35 degrees C. The Anammox inoculum sludge was taken from a pilot plant at EAWAG (Zurich). Anammox biomass has been enriched at 33 degrees C with anaerobic supernatant diluted with sodium nitrite solution until reaching a maximum specific nitrogen removal rate of 0.065 kgN kg(-1) VSS d(-1), which was 11 times higher than the one found in inoculum sludge (0.005 kgN kg(-1) VSS d(-1). In a lab-scale SBR reactor (4 L), coupled with nitritation bioreactor, specific nitrogen removal rate (doubling time equal to 26 d at 35 degrees C and at nitrite-limiting condition) reached the value of 0.22 kgN kg(-1) VSS d(-1), which was approximately 44 times larger than the rate measured in the inoculum Anammox sludge.     

ANAMMOX and partial denitritation in anaerobic nitrogen removal from piggery waste.

The anaerobic ammonium removal from a piggery waste with high strength (56 g COD/L and 5 g T-N/L) was investigated using a lab-scale upflow anaerobic sludge bed reactor at a mesophilic condition. Based on the nitrogen and carbon balance in the process, the contribution of autotrophic and heterotrophic organisms was also evaluated in terms of the influent NO2-N/NH4-N ratio (1:0.8 and 1:1.2 for Phase 1 and Phase 2, respectively). The result of this research demonstrates that the anaerobic ammonium removal from the piggery waste, using the UASB reactor, can be performed successfully. Furthermore, it appears that by using granular sludge as the seed biomass, the ANAMMOX reaction can start more quickly. Average nitrogen conversion was 0.59 kg T-N/m3 reactor-day (0.06 kg T-N/kg VSS/day) and 0.66 kg T-N/m3 reactor-day (0.08 kg T-N/kg VSS/day) for Phase 1 and Phase 2. The NO2-N/NH4-N removal ratio by the ANAMMOX was 1.48 and 1.79 for Phase 1 and Phase 2. The higher nitrite contents (about 50%) in the substrate resulted in higher nitrite nitrogen removal by the partial denitritation, as well as the ANAMMOX reaction, implying higher potential of partial denitritation. However, the result reveals that the ANAMMOX reaction was influenced less by the degree of partial denitritation, and the ANAMMOX bacteria did not compete with denitritation bacteria. The colour of the biomass at the bottom of the reactor changed from dark gray to dark red, which was accompanied by an increase in cytochrome content. At the end of the experiment, red-coloured granular sludge with diameter of 1-2 mm at the lower part of the reactor was also observed.     

Nitrogen removal rates at a technical-scale pilot plant with the one-stage partial nitritation/Anammox process.

Traditional nitrification/denitrification is not suitable for nitrogen removal when wastewater contains high concentrations of ammonium nitrogen and low concentrations of biodegradable carbon. Recently, a deammonification process was developed and proposed as a new technology for treatment of such streams. This process relies on a stable interaction between aerobic bacteria Nitrosomonas, that accomplish partial nitritation and anaerobic bacteria Planctomycetales, which conduct the Anammox reaction. Simultaneous performance of these two processes can lead to a complete autotrophic nitrogen removal in one single reactor. The experiments where nitrogen was removed in one reactor were performed at a technical-scale moving-bed pilot plant, filled with Kaldnes rings and supplied with supernatant after dewatering of digested sludge. It was found that a nitrogen removal rate obtained at the pilot plant was 1.9 g m(-2) d(-1). Parallel to the pilot plant run, a series of batch tests were carried out under anoxic and aerobic conditions. Within the batch tests, where the pilot plant's conditions were simulated, removal rates reached up to 3 g N m(-2)d(-1). Moreover, the batch tests with inhibition of Nitrosomonas showed that only the Anammox bacteria (not anoxic removal by Nitrosomonas) are responsible for nitrogen removal.     

Nitrogen removal of high strength wastewater via nitritation/denitritation using a sequencing batch reactor.

The sequencing batch reactor (SBR) process concept was applied to achieve efficient ammonium removal via nitrite under both laboratory and pilot-scale conditions. Both sets of experimental results show that without pH control or carbon addition the nitritation process consistently converted approximately 50% of the ammonium from biosolids dewatering liquids to nitrite with hydraulic retention times (HRT) as short as 10 h. The results from the pilot-scale study also indicate that the selective oxidation of ammonium to nitrite is a reliable process as the accumulation of nitrate was never an issue during a 330-day trial. The SBR process concept was extended to achieve complete nitrogen removal through nitritation and denitritation in the laboratory scale. The experimental results indicate that a total reduction of 96-98% of the ammonium nitrogen from biosolids dewatering liquids (influent concentration typically 1,200 g m(-3)) was achieved with a short HRT of 1.1 d and a removal rate of 1.05 kgNm(-3)d(-1). This process concept was tested at pilot scale where the nitritation process could be started up without temperature control in a short period of time. Nitrogen removal rates up to 1.2 kgNm(-3)d(-1) at an HRT of 0.88 d have been obtained. COD to nitrogen ratios required in the pilot plant were consistently in the range 1.6-1.9 kgCOD kg(-1)N removed.     

Denitrification of nitrate-contaminated groundwater using a simple immobilized activated sludge bioreactor

A simple anaerobic-activated sludge system, in which microorganisms are immobilized by a novel functional carrier, was used for removing nitrate in groundwater. The operating conditions, including hydraulic retention time (HRT), C/N ratio, temperature and NO(3)(-)-N loading concentration were investigated. The NO(3)(-)-N concentration, residual chemical oxygen demand (COD) and nitrite accumulation were used as indicators to assess the water quality of the effluent. The anaerobic biomass loading capacity in the carrier was 12.8 g/L and the denitrifying Pseudomonas sp. and Rhodocyclaceae bacterium were dominant among the immobilized microorganisms in the anaerobic-activated sludge. Under operating conditions of HRT= 1.5 h, C/N= 2-3 and T= 16.8-20 °C, the removal efficiency of NO(3)(-)-N exceeded 93%, corresponding to a relatively high denitrification rate of 0.73 kg NO(3)(-)-N m(-3) d(-1), when the NO(3)(-)-N loading concentration was 50 mg/L. The NO(3)(-)-N concentration of the effluent always met regulatory criteria for drinking water (<10 mg/L) in the main developed and developing countries. The effluent COD was also below 10 mg/L. Although some nitrite accumulated (0-1.77 mg/L) during the operating period, it can be decreased through adjusting the operating pH and HRT. The immobilized activated sludge system may be useful for the removal of nitrate from groundwater.     

Granular biomass capable of partial nitritation and anammox

A novel and efficient way of removing nitrogen from wastewater poor in biodegradable organic carbon, is the combination of partial nitritation and anoxic ammonium oxidation (anammox), as in the one-stage oxygen-limited autotrophic nitrification/denitrification (OLAND) process. Since anoxic ammonium-oxidizing bacteria grow very slowly, maximum biomass retention in the reactor is required. In this study, a lab-scale sequencing batch reactor (SBR) was used to develop granular, rapidly settling biomass. With SBR cycles of one hour and a minimum biomass settling velocity of 0.7 m/h, OLAND granules were formed in 1.5 months and the nitrogen removal rate increased from 50 to 450 mg N L(-1) d(-1) in 2 months. The granules had a mean diameter of 1.8 mm and their aerobic and anoxic ammonium-oxidizing activities were well equilibrated to perform the OLAND reaction. Fluorescent in-situ hybridization (FISH) demonstrated the presence of both beta-proteobacterial aerobic ammonium oxidizers and planctomycetes (among which anoxic ammonium oxidizers) in the granules. The presented results show the applicability of rapidly settling granular biomass for one-stage partial nitritation and anammox.     

A one-stage system with partial nitritation and anammox processes in the moving-bed biofilm reactor.

The ability of bacterial cultures to create biofilm brings a possibility to enhance biological wastewater treatment efficiency. Moreover, the ability of Anammox and Nitrosomonas species to grow within the same biofilm layer enabled a one-stage system for nitrogen removal to be designed. Such a system, with Kaldnes rings as carriers for biofilm growth, was tested in a technical pilot plant scale (2.1 m(3)) at the Himmerfj?rden Waste Water Treatment Plant (WWTP) in the Stockholm region. The system was directly supplied with supernatant originating from dewatering of digested sludge containing high ammonium concentrations. Nearly 1-year of operational data showed that during the partial nitritation/Anammox process, alkalinity was utilised parallel to ammonium removal. The process resulted in a small pH drop, and its relationship with conductivity was found. The nitrogen removal rate for the whole period oscillated around 1.5g N m(-2)d(-1) with a maximum value equal to 1.9 g N m(-2)d(-1). Parallel to the pilot plant experiment, a series of batch tests were run to investigate the influence on removal rates of different dissolved oxygen conditions and addition of nitrite. The highest nitrogen removal rate (5.2g N m(-2)2d(-1)) in batch tests was obtained when the Anammox process was stimulated by the addition of nitrite. In the simultaneous partial nitritation and Anammox process, the partial nitritation was the rate-limiting step.     

Nitrite accumulation followed by denitrification using sequencing batch reactor.

Biological ammonia-nitrogen removal utilizes two distinct processes, nitrification and denitrification. In nitrification, ammonia oxidizes to nitrite then to nitrate. In this study, elimination of nitrite oxidation to nitrate step was attempted in order to directly remove nitrite to nitrogen gas by denitrification. For this study the supernatant from an anaerobic digester was used as an ammonia source and a sequencing batch reactor (SBR) was employed. Emphasis was given to the evaluation of the operational factors affecting nitrite accumulation and the elucidation of kinetics for biological nitrification and denitrification. Accumulation of nitrite in the nitrification process was achieved by suppressing the growth of Nitrobacter, a nitrite oxidizer, by loading high concentration ammonia supernatant immediately after all ammonia in the previous loading was oxidized to nitrite. Nitrite oxidation was taking place as the solid retention time (SRT) was increased from 2.5 days to 3.0 days in a continuously aerated SBR mode with daily feeding. However, nitrite accumulation was achieved even at longer SRT of 5 days when the aeration and non-aeration periods were appropriately combined and the non-aeration period can be used for denitrification of the accumulated nitrite with a carbon source supplied.     

Anammox brings WWTP closer to energy autarky due to increased biogas production and reduced aeration energy for N-removal.

Fifty years ago when only BOD was removed at municipal WWTPs primary clarifiers were designed with 2-3 hours hydraulic retention time (HRT). This changed with the introduction of nitrogen removal in activated sludge treatment that needed more BOD for denitrification. The HRT of primary clarification was reduced to less than one hour for dry weather flow with the consequence that secondary sludge had to be separately thickened and biogas production was reduced. Only recently the ammonia rich digester liquid (15-20% of the inlet ammonia load) could be treated with the very economic autotrophic nitritation/anammox process requiring half of the aeration energy and no organic carbon source compared to nitrification and heterotrophic denitrification. With the introduction of this new innovative digester liquid treatment the situation reverts, allowing us to increase HRT of the primary clarifier to improve biogas production and reduce aeration energy for BOD removal and nitrification at similar overall N-removal.     

Difficulties in maintaining long-term partial nitritation of ammonium-rich sludge digester liquids in a moving-bed biofilm reactor (MBBR).

Nitrogen can be eliminated effectively from sludge digester effluents by anaerobic ammonium oxidation (anammox), but 55-60% of the ammonium must first be oxidized to nitrite. Although a continuous flow stirred tank reactor (CSTR) with suspended biomass could be used, its hydraulic dilution rate is limited to 0.8-1 d(-1) (30 degrees C). Higher specific nitrite production rates can be achieved by sludge retention, as shown here for a moving-bed biofilm reactor (MBBR) with Kaldnes carriers on laboratory and pilot scales. The maximum nitrite production rate amounted to 2.7 gNO2-Nm(-2)d(-1) (3 gO2m(-3)d(-1), 30.5 degrees C), thus doubling the dilution rate compared to CSTR operation with suspended biomass for a supernatant with 700 gNH4-Nm(-3). Whenever the available alkalinity was fully consumed, an optimal amount of nitrite was produced. However, a significant amount of nitrate was produced after 11 months of operation, making the effluent unsuitable for anaerobic ammonium oxidation. Because the sludge retention time (SRT) is relatively long in biofilm systems, slow growth of nitrite oxidizers occurs. None of the selection criteria applied - a high ammonium loading rate, high free ammonia or low oxygen concentration - led to selective suppression of nitrite oxidation. A CSTR or SBR with suspended biomass is consequently recommended for full-scale operation.     

High rate nitrogen removal by the CANON process at ambient temperature

Completely autotrophic nitrogen removal over nitrite (CANON) is a cost-effective nitrogen removal process. Implementation of the CANON process relies on the cooperation of ammonium-oxidizing and Anammox bacteria, as well as the inhibition of nitrite-oxidizing bacteria. Strict limitations on dissolved oxygen (DO) concentration in the reactor, and the addition of sufficient inorganic carbon in the influent, were adopted as the main operational strategies. The reactor was fed with synthetic inorganic wastewater composed mainly of NH(4)(+)-N, and operated for 106 days. Stable nitrogen removal rates (NRR) of around 1.4 kg N m(-3) d(-1) were obtained at ambient temperature. Morphological characteristics and analysis of bacterial community confirmed the formation of functional outer aerobic and inner anaerobic granular sludge, providing evidence of stable nitrogen removal.     

The presence of ammonium facilitates nitrite reduction under PHB driven simultaneous nitrification and denitrification.

For economic and efficient nitrogen removal from wastewater treatment plants via simultaneous nitrification and denitrification the nitrification process should stop at the level of nitrite such that nitrite rather than nitrate becomes the substrate for denitrification. This study aims to contribute to the understanding of the conditions that are necessary to improve nitrite reduction over nitrite oxidation. Laboratory sequencing batch reactors (SBRs) were operated with synthetic wastewater containing acetate as COD and ammonium as the nitrogen source. Computer controlled operation of the reactors allowed reproducible simultaneous nitrification and denitrification (SND). The oxygen supply was kept precisely at a low level of 0.5 mgL(-1) and bacterial PHB was the only electron donor available for denitrification. During SND little nitrite or nitrate accumulated (< 20% total N), indicating that the reducing processes were almost as fast as the production of nitrite and nitrate from nitrification. Nitrite spiking tests were performed to investigate the fate of nitrite under different oxidation (0.1-1.5 mgL(-1) of dissolved oxygen) and reduction conditions. High levels of reducing power were provided by allowing the cells to build up to 2.5 mM of PHB. Nitrite added was preferentially oxidised to nitrate rather than reduced even when dissolved oxygen was low and reducing power (PHB) was excessively high. However, the presence of ammonium enabled significant reduction of nitrite under low oxygen conditions. This is consistent with previous observations in SBR where aerobic nitrite and nitrate reduction occurred only as long as ammonium was present. As soon as ammonium was depleted, the rate of denitrification decreased significantly. The significance of the observed strongly stimulating effect of ammonium on nitrite reduction under SND conditions is discussed and potential consequences for SBR operation are suggested.     

Comparison of chemo-, hetero- and mixotrophic denitrification in laboratory-scale UASBs.

This study investigated removal of sulfide and p-cresol linked to denitrification in laboratory-scale upflow anaerobic granular sludge bed (UASB) bioreactors. Three parallel denitrification bioreactors were run for nine months, which were operated under chemolithoautotrophic conditions (i.e., using sulfide as electron donor -e-donor- and bicarbonate as C source); heterotrophic conditions (with p-cresol as e-donor and C source), and mixotrophic conditions (utilizing both sulfide and p-cresol as electron donors), respectively. The average hydraulic retention time and nitrate load applied to the bioreactors was 13.4 h and 1,240 mg N-NO3/l/day, respectively. The nitrate removal efficiency was 89, 95 and 99%, respectively, for the chemo-, hetero- and mixotrophic reactors. The mixotrophic UASB removed both sulfide and p-cresol almost completely, indicating that simultaneous removal of the inorganic and organic e-donors occurred. Nitrite was seldom observed as an intermediate. N2O gas and methane concentrations in the biogas were also negligible. These results indicate that mixotrophic denitrification with phenols and sulfide is feasible in high rate UASB reactors.     

Rapid start-up of a nitrifying reactor using aerobic granular sludge as seed sludge

In this study, the effectiveness of aerobic granular sludge as seed sludge for rapid start-up of nitrifying processes was investigated using a laboratory-scale continuous stirred-tank reactor (CSTR) fed with completely inorganic wastewater which contained a high concentration of ammonia. Even when a large amount of granular biomass was inoculated in the reactor, and the characteristics of influent wastewater were abruptly changed, excess biomass washout was not observed, and biomass concentration was kept high at the start-up period due to high settling ability of the aerobic granular sludge. As a result, an ammonia removal rate immediately increased and reached more than 1.0 kg N/m(3)/d within 20 days and up to 1.8 kg N/m(3)/d on day 39. Subsequently, high rate nitritation was stably attained during 100 days. However, nitrite accumulation had been observed for 140 days before attaining complete nitrification to nitrate. Fluorescence in situ hybridization analysis revealed the increase in amount of ammonia-oxidizing bacteria which existed in the outer edge of the granular sludge during the start-up period. This microbial ecological change would make it possible to attain high rate ammonia removal.     

Chemical denitrification for nitrogen removal from landfill leachate.

A new system that removes nitrogen from landfill leachate and other waste waters with similar properties has been proposed with nitritation (i.e. oxidation of ammonium to nitrite) of half of the influent ammonium followed by chemical denitrification with a reaction between equal amounts of ammonium and nitrite to form nitrogen gas. Chemical denitrification occurs at high concentrations and the reactions were studied in combination with a concentration step. Studied concentration methods were freezing/thawing and evaporation/drying. Chemical denitrification is well-known in inorganic chemistry and has been observed in natural systems. Studies in laboratory were focused on chemical denitrification and showed that nearly complete removal of soluble nitrogen can be obtained in evaporation/drying of water solutions or leachate with equal amounts of ammonium and nitrite. Freezing/drying was less efficient with a removal of about 50-60% at high initial concentrations. Chemical denitrification is much influenced by concentration, pH-value, temperature and some compounds in leachate have an inhibiting effect on the reaction. Factors as safety (ammonium nitrite as a salt is explosive above 60 degrees C) and possible side-reactions as formation of ammonia and nitrogen oxides must be carefully evaluated before use in full-scale. Conductivity is a suitable parameter to follow-up the chemical denitrification process.     

Sequenced anaerobic-aerobic treatment of high strength, strong nitrogenous landfill leachates.

As a first step in treatment of high strength, strong nitrogenous landfill leachates (total COD--9.66-20.56 g/l, total nitrogen 780-1,080 mg/l), the performance of laboratory UASB reactors has been investigated under sub-mesophilic (19+/-3 degrees C) and psychrophilic (10+/-2 degrees C) conditions. Under hydraulic retention time (HRT) of around 1.2 days, when the average organic loading rate (OLR) was around 8.5 g COD/l/day, the total COD removal accounted for 71% (on average) for sub-mesophilic regime. The psychrophilic treatment conducted under the average HRT of 2.44 days and the average OLR of 4.2 g COD/l/day showed an average total COD removal of 58% giving effluents more suitable for subsequent biological nitrogen removal. Both anaerobic regimes were quite efficient for elimination of heavy metals by concomitant precipitation in the form of insoluble sulphides inside the sludge. The subsequent submesophilic aerobic-anoxic treatment of submesophilic anaerobic effluents led to only 75% of total inorganic N removal due to COD deficiency for denitrification created by too efficient anaerobic step. On the contrary, psychrophilic anaerobic effluents (richer in COD compared to the submesophilic ones) were more suitable for subsequent aerobic-anoxic treatment giving the total N removal of 95 and 92% at 19 and 10 degrees C, respectively.     

A two-sludge system for simultaneous biological C, N and P removal via the nitrite pathway

Nitrogen removal via the nitrite pathway results in significant savings in both aeration costs and COD requirements for denitrification when compared to the conventional biological nitrogen removal process. Implementation of the nitrite pathway for simultaneous C/N/P removal in a single sludge system has a major drawback: the aeration phase disfavours denitrifying phosphorus removal. A possible configuration to overcome this issue is the utilisation of a two-sludge system where autotrophic and heterotrophic populations are physically separated. This paper experimentally demonstrates the feasibility of a nitrite-based two-sludge system with sequencing batch reactors (SBR) for the treatment of urban wastewater: a heterotrophic SBR with denitrifying PAOs for P removal and an aerobic SBR for N removal. Partial nitrification was attained in the autotrophic SBR so that shortcut biological nitrogen removal was achieved by using the anoxic dephosphatation activity of DPAOs. Finally, the effect of operating this system without pH control was studied using different influent pH values (pH = 6.8, 7.5 and 8.2) and, despite some efficiency lost due to the pH fluctuations, the system was able to remove most of the C, N and P present in the wastewater.     

Upgrading of Florence wastewater treatment plant: co-digestion and nitrogen autotrophic removal.

In recent years a completely autotrophic nitrogen removal process based on Anammox biomass has been tested in a few European countries in order to treat anaerobic supernatant and to increase the COD/N ratio in municipal wastewater. This work reports experimental results on a possible technical solution to upgrade the S. Colombano treatment plant which treats wastewater from the Florentine urban area. The idea is to use 50% of the volume of the anaerobic digester in order to treat external sewage sludge (as septic tank sludge) together with waste activated sludge and to treat the resulting effluent on a SHARON-ANAMMOX process in order to remove nitrogen from the anaerobic supernatant. Anaerobic co-digestion, tested in a 200 L pilot plant, enables low cost treatment of septic tank sludge and increases biogas production; however, it also increases the nitrogen load re-circulated to the WWTP, where nitrogen removal efficiency is already low (<50%), due to the low COD/N ratio, which limits predenitrification efficiency. Experimental results from a SHARON process tested in a lab-scale pilot plant show that nitrite oxidising bacteria are washed-out and steady nitrite production can be achieved at retention times in the range 1 - 1.5 days, at 35 degrees C. In a lab-scale SBR reactor, coupled with a nitration bioreactor, maximum specific nitrogen removal rate under nitrite-limiting conditions (with doubling time equal to about 26 days at 35 degrees C) was equal to 0.22 kgN/kgSSV/d, about 44 times the rate measured in inoculum Anammox sludge. Finally, a cost analysis of the proposed upgrade is reported.     

Formation and characteristics of nitritation granules cultivated in sequencing batch reactor by stepwise increase of N/C ratio

The cultivation of nitritation granules in sequencing batch reactor (SBR) by seeding conventional floccular activated sludge was investigated using ethanol-based synthetic wastewater. Reducing settling time offers selection pressure for aerobic granulation, and stepwise increase of influent N/C ratio can help to selectively enrich ammonia oxidizing bacteria (AOB) in aerobic granules. The spherical shaped granules were observed with the mean diameter of 1.25 mm, average settling velocity of 1.9 cm s(-1) and the sludge volume index (SVI) of 18.5-31.4 ml g(-1). After 25 days of operation, the nitrogen loading rate reached 0.0455 kg NH(4)(+)-N (kg MLSS·d)(-1), which was 4.55 times higher than that of the start-up period. The mature granules showed high nitrification ability. Ammonia removal efficiency was above 95% and nitrite accumulation ratio was in the range of 80-95%. The nitrifying bacteria were quantified by fluorescence in situ hybridization analysis, which indicated that AOB was 14.9 ± 0.5% of the total bacteria and nitrite oxidizing bacteria (NOB) was 0.89 ± 0.1% of the total bacteria. Therefore, AOB was the dominant nitrifying bacteria. It was concluded that the associated inhibition of free ammonia at the start of each cycle and free nitrous acid during the later phase of aeration may be the key factors to start up and maintain the stable nitritation.