Abstract

N2O-production was investigated during nitrogen removal using aerobic granular sludge (AGS) technology. A pilot sequencing batch reactor (SBR) with AGS achieved an effluent in accordance with national discharge limits, although presented a nitrite accumulation rate of 95.79% with no simultaneous nitrification–denitrification. N2O production was 2.06 mg L−1 during the anoxic phase, with N2O emission during air pulses and the aeration phase of 1.6% of the nitrogen loading rate. Batch tests with AGS from the pilot reactor verified that at the greatest COD/N ratio (1.55), the N2O production (1.08 mgN2O-N L−1) and consumption (up to 0.05 mgN2O-N L−1), resulted in the lowest remaining dissolved N2O (0.03 mgN2O-N L−1), stripping the minimum N2O gas (0.018 mgN2O-N L−1). Conversely, the carbon supply shortage, under low C/N ratios, increased N2O emission (0.040 mgN2O-N L−1), due to incomplete denitrification. High abundance of ammonia-oxidizing and low abundance of nitrite-oxidizing bacteria were found, corroborating the fact of partial nitrification. A denitrifying heterotrophic community, represented mainly by Pseudoxanthomonas, was predominant in the AGS. Overall, the AGS showed stable partial nitrification ability representing capital and operating cost savings. The SBR operation flexibility could be advantageous for controlling N2O emissions, and extending the anoxic phase would benefit complete denitrification in cases of low C/N influents.

NOMENCLATURE

     
  • AGS

    aerobic granular sludge

  •  
  • AND

    alternating nitrification and denitrification

  •  
  • AOB

    ammonium-oxidizing bacteria

  •  
  • BOD

    biochemical oxygen demand

  •  
  • COD

    chemical oxygen demand

  •  
  • DO

    dissolved oxygen

  •  
  • FA

    free ammonia

  •  
  • FISH

    fluorescence in situ hybridization

  •  
  • GHG

    greenhouse gas

  •  
  • HRT

    hydraulic retention time

  •  
  • N2O

    nitrous oxide

  •  
  • NLR

    nitrogen loading rate

  •  
  • NO

    nitric oxide

  •  
  • NOB

    nitrite-oxidizing bacteria

  •  
  • SBR

    sequencing batch reactor

  •  
  • SND

    simultaneous nitrification and denitrification

  •  
  • SRT

    sludge retention time

  •  
  • SVI

    sludge volume index

  •  
  • TSS

    total suspended solids

  •  
  • VSS

    volatile suspended solids

INTRODUCTION

Sequencing batch reactors (SBR) with granular biomass have been successfully used to treat municipal and industrial wastewaters (Beun et al. 1999; Arrojo et al. 2004). Aerobic granular sludge (AGS) has a compact structure with excellent settling properties, presenting diverse microbial species and high biomass retention (Liu & Tay 2004). The existence of substrate profiles across the granule depth enables simultaneous aerobic and anoxic processes in the same bioparticle, resulting in a very good performance of organic matter and nitrogen removal (De Kreuk et al. 2005). Aerobic and anoxic zones coexist within the granules even when dissolved oxygen (DO) concentration is high in the bulk liquid, which favours simultaneous nitrification and denitrification (SND) in this system. According to Ju et al. (2007), there are three different mechanisms that make SND possible in aerobic granular SBR. (i) Aerobic and anoxic zones within the granules. (ii) Aerobic and anoxic zones inside the reactor. (iii) The presence of microorganisms able to perform nitrifier denitrification and heterotrophic nitrification.

A specific microbial distribution and mass transfer gradient in AGS provides a potential alternative for partial nitrification (Shi et al. 2011). Some studies applying AGS process showed the establishment of partial nitrification in low-strength real wastewater conditions (Coma et al. 2012; Wagner et al. 2015; Guimarães et al. 2017). Partial nitrification has gained interest in biological nitrogen removal from wastewater, since it reduces carbon limitation concerns and acts as a shortcut nitrogen removal system combined with anaerobic ammonium oxidation (anammox) process (Ge et al. 2015). An effluent composition close to a molar ammonium to nitrite ratio of 1:1 is best for the anammox process. Nevertheless, the conditions for partial nitrification can increase nitrous oxide (N2O) production, which is a greenhouse gas (GHG) with 300-fold stronger effect than carbon dioxide in terms of global warming potential.

In the biological nitrogen removal process, nitrite and/or nitrate produced from nitrification are reduced to nitrogen gas by denitrifiers, which is directly associated with N2O generation. The amount of N2O released is relatively small compared to the quantity of total nitrogen removed during the processes (Gao et al. 2016). Wastewater treatment processes are extremely important in microbial N2O emission (Kong et al. 2013). N2O generation is associated with three different metabolic pathways: (i) incomplete hydroxylamine oxidation, (ii) nitrifier denitrification, (iii) heterotrophic denitrification (Kampschreur et al. 2009). Therefore, the microbial composition of AGS has an important influence on the behaviours and pathways of N2O emissions (Gao et al. 2016). Furthermore, an increase in N2O production and emission can result from some operational conditions such as COD/N ratio, DO concentration, nitrite concentration and pH (Kampschreur et al. 2009).

Although information about nitrogen conversion in aerobic granules is abundant, very few studies have focused on N2O production and emission by this system. Since N2O is considered the third most important contributor to climate change (IPCC 2014), it is important to verify its emission by this technology, which has been widely applied, especially via studies using real wastewater. Additional knowledge of the microbial processes and the factors controlling N2O production in nitrifying/denitrifying AGS systems is required to develop strategies to reduce emissions of this GHG. Therefore, this study aimed (i) to monitor the N2O production and emissions from an SBR with AGS treating domestic wastewater and (ii) to verify the maximum N2O emissions for different COD/N ratios. The microbial composition was also investigated in order to clarify the nitrogen biological processes.

MATERIALS AND METHODS

Set-up of a pilot SBR with AGS

A pilot SBR (0.25 m in diameter and 3.0 m in height) with working volume of 98 L and volumetric exchange ratio of 56% was used to provide the formation of granular sludge. Natural biomass accumulation and granulation occurred without inoculation. The selection conditions such as the settling time, superficial upflow air velocity and the volume exchange ratio were set to stimulate the growth of the granular biomass from wastewater. The SBR was operated at room temperature (23 ± 2°C) in anoxic/oxic (A/O) cycles of 6 hours, consisting of 3 min feeding, 90 min anoxic, 248 min aeration, 15 min settling, 3 min effluent withdrawal and 1 min idle. During the 90 min anoxic phase, air pulses (turned on every 15 minutes for 10 seconds) were applied in order to keep the sludge mixed through the liquid column. The hydraulic retention time (HRT) was 7.14 hours and the sludge retention time (SRT) was 8 ± 2 days. The pH in the system was recorded but not controlled, and fluctuated between 6.5 and 7.5. Aeration was provided by airflow at 32 L min−1, using a membrane diffuser (B&FDIAS, Brazil) placed at the bottom of the reactor in a superficial upflow air velocity of 1.2 cm s−1, which resulted in a DO at saturation level (7.2–8.5 mg L−1).

Real domestic wastewater from a sewage network of Florianópolis, Santa Catarina, Brazil (27°35′49″S/48°32′56″W), fed the SBR at the beginning of each cycle. The influent concentrations of total chemical oxygen demand (CODtot), soluble chemical oxygen demand (CODs), and ammonium-nitrogen (NH4+-N) were 380 ± 101 mgCODtot L−1, 179 ± 29 mgCODs L−1, and 71 ± 13 mgNH4+–NL−1, respectively. The reactor was operated for 120 days, with a biomass concentration of 1.3 gTSS L−1 and 1.1 gVSS L−1, respectively. The AGS granule size was between 0.2 and 0.4 mm and the sludge volume index (SVI) was 70 mL g−1. The performance of system was monitored through the cycles, with samples taken regularly for CODs, NH4+-N, and NOx-N analysis, while N2O was measured in its dissolved and gaseous states. The detailed study over the operational cycle in the pilot SBR with AGS was conducted to verify the nitrogen removal processes and the production and emission of N2O associated with these processes.

Setup of batch experiments using biomass from the pilot SBR

Batch experiments were performed according to Kim & Kim (2011), using AGS taken from the end of the aeration phase of the pilot SBR. The sludge (2.1 L) was stirred and aerated for 20 hours in order to deplete COD and ammonium sources. Subsequently, the sludge was divided into three equal volumes (0.7 L) to be used in the three different assays varying the type of wastewater. The experimental bench system is represented in Figure 1, and comprised: (1) wastewater storage flask, (2) peristaltic pump, (3) compressed air, (4) bench reactor, (5) stirring table, (6) dissolved N2O sensor, (7) N2O gas sensor.

Figure 1

Bench experimental design. (1) Wastewater storage flask, (2) peristaltic pump, (3) compressed air, (4) bench reactor, (5) stirring table, (6) dissolved N2O sensor, (7) N2O gas sensor.

Figure 1

Bench experimental design. (1) Wastewater storage flask, (2) peristaltic pump, (3) compressed air, (4) bench reactor, (5) stirring table, (6) dissolved N2O sensor, (7) N2O gas sensor.

For each batch assay, the bench reactor was filled with AGS (0.7 L) and the specific wastewater (0.1 L), which had different characteristics and gave distinct COD/NH4+-N ratios (Table 1). The domestic raw sewage was taken from the same municipal wastewater network that was used in the pilot reactor. The denitrified domestic sewage is the effluent sampled at the end of the anoxic phase of the pilot SBR. The synthetic sewage was a solution of NH4Cl prepared without a carbon source. The operational cycle conditions were the same as for the pilot SBR and the conditions at the beginning of each assay are shown in Table 1. The batch experiments were performed simulating the operational cycle of the pilot reactor, consisting of 90 min of anoxic stirring, 240 min of aeration, 15 min of settling, totalling 345 min. Compressed air was provided through an air porous stone diffuser at the bottom of the reactor; DO was kept at 2 mg L−1.

Table 1

Experimental conditions in the three batch tests, varying C/N ratios, using AGS from the pilot SBR

Parameter Domestic raw sewage Denitrified domestic sewage Synthetic sewage 
TSS (g TSS L−11.1 1.0 1.0 
CODs (mg CODs L−1103 52 No addition of carbon 
NH4+-N (mg NH4+ L−167 45 45 
COD/NH4+-N ratio 1.55 1.17 
Parameter Domestic raw sewage Denitrified domestic sewage Synthetic sewage 
TSS (g TSS L−11.1 1.0 1.0 
CODs (mg CODs L−1103 52 No addition of carbon 
NH4+-N (mg NH4+ L−167 45 45 
COD/NH4+-N ratio 1.55 1.17 

Analytical methods

The cycle of the pilot SBR and each batch experiment were monitored for following: NH4+-N, NO2-N, NO3-N, CODs and solids every 45 min (during the anoxic phase) and every hour (during the aerobic phase). The parameters were analyzed in accordance with Standard Methods (APHA 2012). The nitrite accumulation rate was calculated based on Wei et al. (2014) and free ammonia (FA) according to Anthonisen et al. (1976).

The N2O gas flow was continuously analysed (every minute) by an infra-red analyser (Guardian NG, Edinburgh, UK) with a range between 0 and 3,000 ppm. In addition, dissolved N2O concentration was measured with an N2O micro-sensor (N2O Wastewater System, Unisense A/S, Denmark) with a working range between 0–1.5 mgN2O-N L−1 and a detection limit of 0.005 mgN2O-N L−1. The N2O production rate was calculated according to the method described by Hu et al. (2011).

Microbiological methods

The identification of active bacterial populations in the sludge used for the assays was carried out by fluorescence in situ hybridization (FISH). Biomass samples were fixed with 4% paraformaldehyde solution and hybridized using the following probes: EUB338 I, II and III to stain all bacteria; NSO190 for ammonium-oxidizing bacteria (AOB) population; Ntspa662 for Nitrospira; GAO431 and 989 for member of ‘Candidatus Competibacter’; PAO462, 651 and 841 to stain ‘Candidatus Accumulibacter phosphatis’; PAE997 for Pseudomonas spp. (Amann et al. 1995). The detailed sequences of FISH probes can be checked in Probebase (Greuter et al. 2016).

DNA sequencing was performed using MiSeq® Illumina technology for sequencing by synthesis (SBS) (Neoprospecta, Brazil). The DNA was extracted from biomass by applying the protocol of MoBio PowerBiofilm™ DNA extraction kit (MoBio Laboratories, USA). The rRNA 16S gene, V3/V4 region, was been amplified using the 341F (CCTACGGGRSGCAGCAG) and 806R (GGACTACHVGGGTWTCTAAT) primers, with Illumina adapters, required for sequencing. The amplification was performed in 35 cycles at 50°C annealing temperature, where each sample was amplified in triplicate. The sequencing was performed in Illumina MiSeq, using a V2 kit, with a single-end 300 cycle run. The system guaranteed the reading of 100,000 sequences with sampling taxonomic identification and quantification of the number of sequences obtained from each taxon. Operational taxonomic unit (OTU) picking was performed using BLASTN 2.2.28 against GreenGenes 13.8 database. To attribute taxonomy, only sequences with hits of 99% of identity in alignment covered over 99% were considered.

RESULTS AND DISCUSSION

Pilot SBR with AGS treating domestic wastewater

Treatment performance

Table 2 presents the performance of the pilot SBR with AGS during long-term operation (120 days). The BOD5 and ammonium removal were greater than 80%, attending the Brazilian discharge limits (120 mg L−1 or 60% of removal efficiency for BOD5Brasil 2011) with effluent concentration lower than 35 mg BOD5 L−1. Final phosphorous effluent concentrations reached effluent quality criteria of Santa Catarina state law (≤4 mg L−1 of total phosphorus – Santa Catarina 2009). The high standard deviation is attributed to natural fluctuation in the municipal sewage network and also to the influence of rainwater inputs.

Table 2

Influent and effluent concentrations and removal efficiencies in the pilot SBR with AGS for the legally required parameters

Parameter Influent (mg L−1) (n = 14) Effluent (mg L−1) (n = 14) Removal (%) 
CODtot 400 ± 101 136 ± 23 64 
CODs 174 ± 29 55 ± 9 68 
BOD5 221 ± 36 31 ± 7 86 
NH4+-N 54 ± 13 8 ± 5 84 
Total phosphorus 5.1 ± 1.0 4.0 ± 1.0 16 
TSS 145 ± 48 61 ± 12 54 
Parameter Influent (mg L−1) (n = 14) Effluent (mg L−1) (n = 14) Removal (%) 
CODtot 400 ± 101 136 ± 23 64 
CODs 174 ± 29 55 ± 9 68 
BOD5 221 ± 36 31 ± 7 86 
NH4+-N 54 ± 13 8 ± 5 84 
Total phosphorus 5.1 ± 1.0 4.0 ± 1.0 16 
TSS 145 ± 48 61 ± 12 54 

Average ± standard deviation. n = number of samples.

Figure 2 shows a typical cycle profile of nitrogen compounds, production and emission of N2O, and concentrations of CODs and DO during pilot SBR operation. The reactor was fed with a COD/N ratio of 2.53, with loading rates of 0.41 kgCODs m−3d−1 and 0.16 kgNH4+-N m−3 d−1. Ammonium was oxidized mainly to nitrite during the aeration phase and negligible nitrate formation was verified (effluent concentrations of 20 mgNO2-N L−1 and 0.95 mgNO3-N L−1), indicating partial nitrification (Figure 2(a)). The nitrite accumulation rate was 95.79%, implying the activity of nitrite-oxidizing bacteria (NOB) was fully limited in the partial nitrification in the pilot reactor. This phenomenon likely results from the uncoupled activities of AOB and NOB (Wei et al. 2014), and it commonly occurs in reactors with aerobic granules (Isanta et al. 2012; Wagner et al. 2015; Guimarães et al. 2017).

Figure 2

Pilot SBR operational cycle with AGS treating domestic wastewater. During the 360 min cycle, the concentration profile of NH4+-N, NO2-N and NO3-N (a); COD and DO (b); and N2O production and emission (c) were measured over time.

Figure 2

Pilot SBR operational cycle with AGS treating domestic wastewater. During the 360 min cycle, the concentration profile of NH4+-N, NO2-N and NO3-N (a); COD and DO (b); and N2O production and emission (c) were measured over time.

Nitrite and CODs (Figure 2(a) and 2(b)) decreased during the anoxic phase, indicating the process of heterotrophic denitrification in this period. Previous studies have shown that SND could occur in AGS as a result of DO diffusion limitations, which creates anoxic zones inside the granule (Shi et al. 2011; Wei et al. 2014). However, SND was not observed in the pilot SBR of this study. The granular sludge size obtained was in a range of 0.2–0.4 mm, which is probably not large enough to create the anoxic core. Overall, nitrogen removal was performed by partial nitrification and denitrification under alternating anoxic and aerobic conditions inside the reactor. According to Guo et al. (2010) and Wei et al. (2014), the requirements for oxygen consumption and carbon source are reduced during partial nitrification and subsequent nitrite denitrification. Therefore, the method is considered cost-effective for the treatment of wastewater with low COD/N ratios. However, the accumulation of nitrite might lead to negative effects, such as inducing and/or increasing the N2O emission via denitrification (Kampschreur et al. 2009).

N2O production and emission

The fraction of influent nitrogen converted to N2O was 1.6%, while N2O emission from the oxidized ammonium during the cycle was 2.26%. This value represents a N2O emission of 3.15 mgN2O-N (gVSS)−1 and 2.06 mgN2O L−1, in terms of nitrogen emitted as N2O per litre of treated sewage. According to Kong et al. (2013) and Castro-Barros et al. (2016), the N2O emissions from biological nutrient removal processes vary substantially among the studies (0.01–25% of N removed) due to different operational conditions, wastewater and reactor types, and N2O measurement methods. N2O emissions from municipal wastewater treatment plants are estimated as approximately 0.5% of the nitrogen loading rate (NLR) (IPCC 2006). A previous partial nitrification study with AGS reactor reported N2O emissions of 3.8% of NLR (Shi et al. 2011).

Although the N2O emissions occurred during the air pulses in the anoxic phase and during the aeration phase by stripping (Figure 2(c)), N2O is likely produced mainly during the anoxic period as a result of the denitrification process. In order to verify that, N2O was measured in the liquid phase (Figure 2(c)), showing that the greatest dissolved fraction was during the first 15 min of the anoxic period. Many factors have been reported to increases N2O production during denitrification, such as low COD/N ratio (<3.5), low pH (<6.5) and short solids retention time (<1 day) (Kishida et al. 2004; Kampschreur et al. 2009; Quan et al. 2012). The SBR studied here was operated at a low COD/N ratio (2.5), pH of 6.5–7.5 and SRT of 8 days, indicating C/N ratio as a relevant factor that could induce N2O production. Similar results were obtained by Quan et al. (2012) in an AGS reactor treating synthetic wastewater, which verified that increasing COD/N ratio and aeration rate would reduce N2O emissions.

Batch experiments with AGS from the pilot SBR

The effect of different COD/N ratios on N2O production and emission

In order to verify the maximum emission of N2O from the AGS of the pilot reactor, batch experiments were carried out with different and low COD/N ratios. Figure 3(a)3(c) show the NH4+-N, NO3-N and NO2-N concentrations and Figure 3(d)3(f) N2O production and emission over one cycle under different COD/N ratios. NH4+-N concentration remained unchanged during the anoxic phase for the three batch experiments. When the aeration started, NH4+-N concentration decreased linearly with increasing NO2-N concentration, following the partial nitrification phenomena already observed in the pilot reactor. The effluent ammonium concentration decreased to 21, 21 and 15 mg L−1, while nitrite increased to 14.8, 10.7 and 12 mg/L at COD/N ratios of 1.55, 1.17 and 0, respectively. Nitrate was negligible during the three batch experiments (Figure 3(a)3(c)) and no significant SND was observed in aerobic conditions. Total nitrogen removal was higher at increased COD/N ratios (32% at ratios 0; 44% at ratio 1.55).

Figure 3

Bench-scale SBR operational cycle with AGS from the pilot system. During the 360 min cycle, the concentration profile of NH4+-N, NO2-N and NO3-N and the effect of different COD/N ratios on N2O production/emission were monitored. (a) and (d) COD/N = 1.55; (b) and (e) COD/N = 1.17; (c) and (f) COD/N = 0.

Figure 3

Bench-scale SBR operational cycle with AGS from the pilot system. During the 360 min cycle, the concentration profile of NH4+-N, NO2-N and NO3-N and the effect of different COD/N ratios on N2O production/emission were monitored. (a) and (d) COD/N = 1.55; (b) and (e) COD/N = 1.17; (c) and (f) COD/N = 0.

Figure 4

Microbial diversity at genus level in the AGS sampled from pilot SRB on days 21 and 119. Only relative abundance greater than 5% of the total sequences was considered.

Figure 4

Microbial diversity at genus level in the AGS sampled from pilot SRB on days 21 and 119. Only relative abundance greater than 5% of the total sequences was considered.

The main factors associated with nitrite accumulation in this study can be FA inhibition and DO limitation (Ge et al. 2015). The FA concentration fluctuated in the batch cycle tests. At the COD/N ratio of 1.55, it varied from 0.31 to 0.16 mg FA L−1, from the beginning to the end of the cycle, while at lower ratios (1.17 and 0) it varied from 0.24 to 0.14 mg FA L−1. These FA values are in the range of NOB inhibition, which is from 0.1 to 1.0 mg FA L−1 (Anthonisen et al. 1976). DO limitation inside the granules should be considered even if a sufficient DO concentration is maintained in the bulk liquid (greater than 2 mg L−1) for complete nitrification. The oxygen penetration depth provided different DO levels in the granular sludge affecting the activity and distribution of microorganisms. DO limitation inside the granules limited the growth of NOB and thus helped to maintain partial nitrification (Vázquez-Padín et al. 2010). Furthermore, Liang et al. (2015) stated that the presence of organic matter in the influent could promote the suppression of NOB, and then ensured the stable operation of partial nitrification. According to the authors NOB could not outcompete for DO with the heterotrophic bacteria and AOB.

N2O was mainly produced in the anoxic phase and briefly in the aerobic phase in all experimental batch assays, as shown by N2O dissolved concentrations in Figure 3(d)3(f). During anoxic phase, the dissolved N2O was formed and also consumed, which is likely a result of nitrite reduction followed by N2 production, in a heterotrophic denitrification process. The N2O emitted was maximal at the beginning of the aerobic period and stabilized at a low concentration after this initial spike. At COD/N ratio of 1.55, a maximum concentration of 1.08 mg N2O-N L−1 was produced in 17 min followed by a rapid decrease in the next 21 min up to 0.05 mg N2O-N L−1. The remaining N2O in the liquid (0.03 mg N2O-N L−1) was stripped to the gas phase as soon as the aeration started, represented by the peak in 0.018 mgN2O-N L−1 of emitted gas (Figure 3(d)). Conversely, at the lowest COD/N ratio, N2O was produced at half the speed (36 min) at a maximum concentration of 0.53 mgN2O-N L−1, and not fully consumed, leaving 0.13 mgN2O-N L−1, which was subsequently stripped in the aeration phase up to 0.040 mgN2O-N L−1 of emitted gas (Figure 3(f)). The increased N2O emission during low C/N ratios can be ascribed to incomplete denitrification, induced by the carbon supply shortage (Lemaire et al. 2006), which in the case of a ratio equal to 0, might be replaced by the use of internal carbon sources. Since N2O reduction to N2 is the last step of denitrification, after nitrate or nitrite reduction to N2O, if no carbon source is available, N2O gas will possibly accumulate (Yang et al. 2009). Additionally, NO2-N accumulation resulting from incomplete nitrification and denitrification may inactivate N2O reductase, thus increasing N2O production and emissions (Kampschreur et al. 2009).

Partial nitrification for saving capital and operational SBR strategy to reduce N2O emission

Nitrogen removal via nitrite decreases the energy consumption, by decreasing aeration, and reduces carbon limitation concerns (Yang et al. 2009; Ge et al. 2015). During the pilot SBR operation and batch experiments of the present study, partial nitrification with 95.79% and 96.98% of NO2-N accumulation were observed, respectively, showing that AGS had excellent and stable partial nitrification ability. Therefore, the stable partial nitrification process achieved by the AGS system in the present study could benefit other processes. This phenomenon is suitable as pre-treatment of the anammox process, which does not require a carbon source and reduces nitrite directly to N2 without N2O production.

However, special attention should be given to the nitrite accumulation process, since its conditions usually lead to N2O production. N2O yield based on the removed total nitrogen was estimated in order to quantify and analyse the nitrogen flux during nitrogen removal in one operational cycle. Thus, 2.79%, 9.43% and 7.99% of the removed total nitrogen was converted to N2O-N at COD/N ratios of 1.55, 1.17 and 0, respectively. The N2O levels emitted in each cycle assay were 1.28, 1.77 and 1.76 mgN2O-N L−1, meaning 1.22%, 2.50% and 2.55% of the NLR at COD/N ratio of 1.55, 1.17 and 0, respectively. These N2O fractions emitted are greater than expected N2O emissions in wastewater treatment plants (0.5% by IPCC 2006), but lower than studies with partial nitrification in AGS systems (3.8% by Shi et al. 2011). This relatively low N2O emission could be compensated with the combination of AGS and anammox systems, since this last shows no N2O production.

Additionally, the present batch experiments showed that at higher C/N ratios, less N2O was emitted, as a result of complete denitrification. However, the use of external carbon or controlling the organic loading rate in full-scale systems can be costly and unfeasible. Therefore, the batch tests showed that in very low C/N ratio conditions, which could be associated with rain events or a specific effluent, an extension of the anoxic phase would lead to enough N2O reduction, decreasing its emission. In this way, the SBR operation flexibility could be beneficial for controlling N2O emission, diminishing the greenhouse effect. Yang et al. (2009) have suggested the application of step-feed SBR to reduce N2O production, which provides an external carbon source at the end of aeration to reduce nitrite and then N2O. Lochmatter et al. (2013) have tested different aeration rates to promote alternating nitrification and denitrification (AND) conditions. The AND strategies were designed alternating aerobic/anoxic phases during the famine phase, and N2O emissions significantly decreased with COD loads of 2.1–2.4 mg L−1 d−1.

Microbial community characterization

Microbial characterization using FISH of the sludge from the pilot SBR used in the bench assays revealed a diversified community. AOB belonging to the bacterial genus Nitrosomonas were detected in high abundance in the aerobic granules. The genus Nitrospira was the identified NOB, but at low abundance, corroborating with the partial nitrification observed at pilot and bench SBR. Accumulation of nitrite could induce nitrifier denitrification activity by AOB, which uses nitrite as the terminal electron acceptor to produce N2O (Kampschreur et al. 2009; Gao et al. 2016). Heterotrophic denitrifiers from genus Pseudomonas were detected in high abundance, while glycogen- and polyphosphate-accumulating organisms (GAO and PAO) were in low abundance in the AGS. Besides the use of external carbon for heterotrophic denitrification, PAO and GAO organisms can perform denitrification using intracellular carbon (Lemaire et al. 2006), which could also have contributed to N2O production.

The bacterial communities were also verified using the advanced method of high-throughput sequencing. Samples of the pilot SBR were characterized on days 21 and 119, the latter coinciding with the batch experiments' accomplishment (Figure 4). Both samples were globally composed of the same predominant populations present in relative abundances above 5% of the total bacterial community and members of family Xanthomonadaceae, Comamonadaceae, Microbacteriaceae and Rhodobacteraceae. These families are commonly detected on wastewater treatment systems (Weissbrodt et al. 2014). Xanthomonadaceae showed the greatest relative abundance (6–9-fold higher than other organisms) throughout the experiments. Additionally, Flavobacteriaceae and Bradyrhizobiaceae were identified in relative abundances above 5% on day 119.

Conditions applied in the pilot SBR favoured genus Pseudoxanthomonas sp. from the Xanthomonadaceae family, with relative abundances increasing from 30% (day 21) to 45% (day 119). These organisms belong to the heterotrophic denitrifying community and can produce exopolysaccharides, which in AGS systems is essential for supporting the granule structure (Adav et al. 2010; Weissbrodt et al. 2014). Acidovorax sp. (family Comamonadaceae) was detected on the samples (day 21: 6%; day 119: 2%), which are aerobic organisms that are able to grow anaerobically using nitrate or nitrite as terminal electron acceptors. In activated sludge, Acidovorax have been demonstrated to denitrify (McIlroy et al. 2015). The Microbacteriaceae family, represented by the genera Leucobacter and Leifsonia, was previously found as the predominant population on the surface of mature granules, and shows a high growth rate (Kim & Lee 2011). Flavobacterium sp. (day 21: 2%; day 119: 7%), commonly seen in activated sludge, has an aerobic metabolism, but anaerobic growth is also possible for some species. Interestingly, various polysaccharides are hydrolyzed by several species in the genus (McIlroy et al. 2015). Overall, these predominant populations on the AGS of the present study are facultative ordinary heterotrophic organisms, which are aerobic and able to denitrify. Therefore, the microbial characterization supports the heterotrophic denitrification as the main N2O production and consumption process, during the anoxic phases of the pilot SBR and bench assays, which agrees with physical–chemical results previously discussed.

CONCLUSIONS

N2O emission was verified in the SBR with AGS as a result of partial nitrification and heterotrophic denitrification processes. High concentrations of FA and DO limitation inside the granules might be related to high nitrite accumulation playing an important role in N2O production. N2O was mainly produced during the anoxic phase and its highest emissions were measured under low COD/N ratio conditions, due to incomplete denitrification induced by the carbon supply shortage. AOB was at high abundance, while NOB was found at low abundance in granules. Denitrifying heterotrophics represented the majority community in the sludge, with the genus Pseudoxanthomonas predominating (6–9-fold more abundant than others organisms). Overall, the AGS had shown excellent and stable partial nitrification ability representing capital and operating cost savings. The SBR operation flexibility could be advantageous for controlling N2O emissions, and the extension of anoxic phase in cases of low C/N influents would benefit complete denitrification.

ACKNOWLEDGEMENTS

The authors would like to thank FAPESC, FINEP, CAPES and CNPq for financial support in terms of costs and providing scholarships.

REFERENCES

REFERENCES
Adav
,
S. S.
,
Lee
,
D. J.
&
Lai
,
J. Y.
2010
Microbial community of acetate utilizing denitrifiers in aerobic granules
.
Appl. Microbiol. Biotechnol.
85
,
753
762
.
Amann
,
R. I.
,
Ludwig
,
W.
&
Schleifer
,
K.
1995
Phylogenetic identification and in situ detection of individual microbial cells without cultivation
.
Microbiological Reviews
59
(
1
),
143
169
.
American Public Health Association (APHA)
2012
Standard Methods for the Examination of Water and Wastewater
,
22nd edn
.
APHA, AWWA (American Water Works Association), and WEF (Water Environment Federation)
,
Washington, DC
.
Anthonisen
,
A. C.
,
Loehr
,
R. C.
,
Prakasam
,
T. B. S.
&
Srinath
,
E. G.
1976
Inhibition of nitrification and nitrous acid compounds
.
Journal of the Water Pollution Control Federation
48
(
5
),
835
852
.
Arrojo
,
B.
,
Mosquera-Corral
,
A.
,
Garrido
,
J. M.
&
Méndez
,
R.
2004
Aerobic granulation with industrial wastewater in sequencing batch reactors
.
Water Research
38
,
3389
3399
.
Beun
,
J. J.
,
Hendriks
,
A.
,
van Loosdrecht
,
M. C. M.
,
Morgenroth
,
M.
,
Wilderer
,
P. A.
&
Heijnen
,
J. J.
1999
Aerobic granulation in a sequencing batch reactor
.
Water Research
33
,
2283
2290
.
Brasil
2011
Resolução Conama n.430. Dispõe sobre as condições e padrões de lançamento de efluentes (Conama Resolution n.430. Provides the Conditions and Standards for the Effluents Discharge)
.
National Environment Council
,
Brasília
,
Brazil
,
89
p.
Castro-Barros
,
C. M.
,
Rodríguez-Caballero
,
A.
,
Volcke
,
E. I. P.
&
Pijuan
,
M.
2016
Effect of nitrite on the N2O and NO production on the nitrification of low-strength ammonium wastewater
.
Chem. Eng. J.
287
,
269
276
.
Coma
,
M.
,
Verawaty
,
M.
,
Pijuan
,
M.
,
Yuan
,
Z.
&
Bond
,
P. L.
2012
Enhancing aerobic granulation for biological nutrient removal from domestic wastewater
.
Bioresour. Technol.
103
(
1
),
101
108
.
De Kreuk
,
M. K.
,
Heijnen
,
J. J.
&
van Loosdrecht
,
M. C. M.
2005
Simultaneous COD, nitrogen, and phosphate removal by aerobic granular sludge
.
Biotechnol. Bioeng.
2
,
761
769
.
Gao
,
M.
,
Yang
,
S.
,
Wang
,
M.
&
Wang
,
X.
2016
Nitrous oxide emissions from an aerobic granular sludge system treating low-strength ammonium wastewater
.
Journal of Bioscience and Bioengineering
122
(
5
),
601
605
.
Greuter
,
D.
,
Loy
,
A.
,
Horn
,
M.
&
Rattei
,
T.
2016
ProbeBase-an online resource for rRNA-targeted oligonucleotide probes and primers: new features 2016
.
Nucleic Acids Research
44
(
1
),
586
589
.
Guimarães
,
L. B.
,
Mezzari
,
M. P.
,
Daudt
,
G. C.
&
da Costa
,
R. H. R.
2017
Microbial pathways of nitrogen removal in aerobic granular sludge treating domestic wastewater
.
Journal of Chemical Technology & Biotechnology
92
(
7
),
1756
1765
.
Guo
,
J.
,
Peng
,
Y.
,
Huang
,
H.
,
Wang
,
S.
,
Ge
,
S.
,
Zhang
,
J.
&
Wang
,
Z.
2010
Short- and long-term effects of temperature on partial nitrification in a sequencing batch reactor treating domestic wastewater
.
Journal of Hazardous Materials
179
,
471
479
.
Hu
,
Z.
,
Zhang
,
J.
,
Xie
,
H.
,
Li
,
S.
,
Wang
,
J.
&
Zhang
,
T.
2011
Effect of anoxic/aerobic phase fraction on N2O emission in a sequencing batch reactor under low temperature
.
Bioresour. Technol.
102
,
5486
5491
.
IPCC
2006
Guidelines for National Greenhouse Gas Inventories
.
IGES
,
Japan
.
IPCC
2014
Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change
[Core Writing Team, R. K. Pachauri & L. A. Meyer (eds)]
.
IPCC
,
Geneva
,
Switzerland
,
151
p.
Isanta
,
E.
,
Suárez-Ojeda
,
M. E.
,
Val del Río
,
Á.
,
Morales
,
N.
,
Pérez
,
J.
&
Carrera
,
J.
2012
Long term operation of a granular sequencing batch reactor at pilot scale treating a low-strength wastewater
.
Chemical Engineering Journal
198–199
,
163
170
.
Kampschreur
,
M. J.
,
Temmink
,
H.
,
Kleerebezem
,
R.
,
Jetten
,
M. S. M.
&
van Loosdrecht
,
M. C. M.
2009
Nitrous oxide emission during wastewater treatment
.
Water Research
43
,
4093
4103
.
Kim
,
H. J.
&
Lee
,
S. S.
2011
Leucobacter kyeonggiensis sp. nov., a new species isolated from dye wastewater
.
The Journal of Microbiology
49
(
6
):
1044
1049
.
Kishida
,
N.
,
Kim
,
J. H.
,
Kimochi
,
Y.
,
Nishimura
,
O.
,
Sasaki
,
H.
&
Sudo
,
R.
2004
Effect of C/N ratio on nitrous oxide emission from swine wastewater treatment process
.
Water Sci. Technol.
49
(
5–6
),
359
365
.
Kong
,
Q.
,
Zhang
,
J.
,
Miao
,
M.
,
Tian
,
L.
,
Guo
,
N.
&
Liang
,
S.
2013
Partial nitrification and nitrous oxide emission in an intermittently aerated sequencing batch biofilm reactor
.
Chem. Eng. J.
217
,
435
441
.
Lochmatter
,
S.
,
Gonzalez-Gil
,
G.
&
Holliger
,
C.
2013
Optimized aeration strategies for nitrogen and phosphorus removal with aerobic granular sludge
.
Water Research
47
,
6187
6197
.
McIlroy
,
S. J.
,
Saunders
,
A. M.
,
Albertsen
,
M.
,
Nierychlo
,
M.
,
McIlroy
,
B.
,
Hansen
,
A. A.
,
Karst
,
S. M.
,
Nielsen
,
J. L.
&
Nielsen
,
P. H.
2015
MiDAS: the field guide to the microbes of activated sludge
.
Database
2015
(
2
),
1
8
.
Quan
,
X.
,
Zhang
,
M.
,
Lawlor
,
P. G.
,
Yang
,
Z.
&
Zhan
,
X.
2012
Nitrous oxide emission and nutrient removal in aerobic granular sludge sequencing batch reactors
.
Water Research
46
,
4981
4990
.
Santa Catarina
2009
Lei n.14.675. Institui o Código Estadual do Meio Ambiente e estabelece outras providências (State Law n.14.675. Establishes the Environment State Code). Florianópolis, Brazil, 64p
.
Shi
,
Y.
,
Wang
,
X.
,
Yu
,
H.
,
Xie
,
H.
,
Teng
,
S.
,
Sun
,
X.
,
Tian
,
B.
&
Wang
,
S.
2011
Aerobic granulation for nitrogen removal via nitrite in a sequencing batch reactor and the emission of nitrous oxide
.
Bioresour. Technol.
102
,
2536
2541
.
Vázquez-Padín
,
J. R.
,
Figueroa
,
M.
,
Campos
,
J. L.
,
Mosquera-Corral
,
A.
&
Méndez
,
R.
2010
Nitrifying granular systems: a suitable technology to obtain stable partial nitrification at room temperature
.
Separation and Purification Technology
74
,
178
186
.
Wagner
,
J.
,
Guimarães
,
L. B.
,
Akaboci
,
T. R. V.
&
Costa
,
R. H. R.
2015
Aerobic granular sludge technology and nitrogen removal for domestic wastewater treatment
.
Water Sci. Technol.
71
(
7
),
1040
1046
.
Yang
,
Q.
,
Liu
,
X.
,
Peng
,
C.
,
Wang
,
S.
,
Sun
,
H.
&
Peng
,
Y.
2009
N2O production during nitrogen removal via nitrite from domestic wastewater: main sources and control method
.
Environ. Sci. Technol.
43
,
9400
9406
.