New insights into nitrous oxide emissions in a single-stage CANON process coupled with denitrification: thermodynamics and nitrogen transformation

In recent years, interest towards greenhouse gas (GHG) emission from wastewater treatment plants (WWTPs) has substantially increased, with the aim of minimizing their environmental impact (Mannina et al. 2018). Nitrous oxide (N₂O) is an important GHG with a high global warming potential (approximately 298 times higher than that of CO₂) (Desloover et al. 2012). In general, the autotrophic nitrogen removal processes in WWTPs are considered a major source of N₂O (Spinelli et al. 2018). The completely autotrophic nitrogen removal over nitrite (CANON) process is an energy saving and cost-effective biological nitrogen removal process because of its reduced carbon source and oxygen demands and decreased sludge production (Sliekers et al. 2002). Currently, the CANON process coupled with denitrification has been widely employed to treat high ammonium-contained wastewater with a low C/N ratio (Lackner et al. 2014). Excellent nitrogen removal can be achieved using the CANON process coupled with denitrification, which has been demonstrated as an economically feasible option for practical applications (Vázquez-Padín et al. 2009). However, N₂O emission from this process decreases evidently its own the environmental sustainability (Connan et al. 2018).

A better understanding of the mechanism of N₂O production would help facilitate the development of generic mitigation strategies for reducing N₂O emissions resulting from the CANON process coupled with denitrification. The following three primary microbial pathways are involved in the N₂O production in this process: the hydroxylamine (NH₂OH) oxidation pathway, the nitrifier denitrification pathway and the heterotrophic denitrification pathway. Although N₂O production pathways have been extensively investigated in a large amount of previous studies, there is still controversy concerning the mechanism of N₂O production (Domingo-Félez & Smets 2019). Often, various N₂O production pathways simultaneously occur, and each pathway is differently regulated by a variety of environmental factors (Law et al. 2011; Rathnayake et al. 2013). Therefore, determining the relative contributions of each pathway to total N₂O production remains challenging because of the lack of effective methodologies and analytical tools needed for the identification and characterization of N₂O production. This issue has led to different studies reporting contradictory observations (Duan et al. 2017; Ma et al. 2017). Therefore, effective methodologies and analytical tools for characterizing different production pathways and their effects on the total N₂O emissions are still needed to develop targeted strategies for controlling and reducing N₂O emissions.

The essence of microbial nitrogen metabolism is material- and energy transformation-based nitrogen metabolism (Rittmann & McCarty 2012). Biochemical reactions in nitrogen metabolism are accompanied by material recombination and energy changes. Therefore, thermodynamics and material tracing can be effectively used to elucidate the synthetic pathway and mechanism of nitrogen-based products involved in nitrogen metabolism. The objective of this study was to accurately identify N₂O production pathways and their contributions to the total N₂O emissions by a quantitative analysis approach based on ¹⁵N isotope tracing. Additionally, the dynamic characteristics of N₂O emissions and nitrogen transformation in a sequencing batch biofilm reactor (SBBR) using the CANON process coupled with denitrification were investigated based on a thermodynamic model and stoichiometry. The results of this study provided a better understanding of N₂O production and important information concerning the N₂O emissions reduction from the CANON process coupled with denitrification.

Three different types of wastewater with different C/N ratios (0.00, 0.67 and 1.00; Table S1 in Supplementary Information) were fed simultaneously into three identical SBBR reactors in parallel. The effective volume of each SBBR was 30 L (9 L of flexible medium for biofilm growth, 30% (V/V); Figure S1). The SBBR was operated under intermittent aeration (non-aeration: aeration = 2 h:2 h) with a hydraulic retention time (HRT) of 48 h. The reactor temperature was maintained at 30 ± 1 °C with a cycle time of 24 h (4 min of feeding, 23 h of reaction, 30 min of settling and 26 min of decanting). The dissolved oxygen (DO) in the aeration phase ranged from 1.5 to 2.0 mg L⁻¹ and was controlled using a DO controller and probe (Bell, BDO-200D, China). During each cycle, 10.5 L of the supernatant was decanted from the SBBR, and then an equal volume of wastewater was fed into the reactor at the beginning of the next cycle. The operation of the reactor and process parameters used in this study were reported in Yan et al. (2019).

Three 0.5-L sealed bottles (A, B and C) each contained 150 g fresh biofilm from the SBBRs with different C/N ratios (Figure S2). An identical sealed bottle (D) without biofilm served as a control. A 350-mL sample of different types of wastewater containing a ¹⁵N-NaNO₂ isotopic tracer (Table S2) was added into each bottle. The off-gas was collected to determine the isotopic compositions of N₂O and N₂. Synchronously, 2-mL liquid samples were collected to analyze the concentrations of NH₄⁺-N, NO₂⁻-N, NO₃⁻-N and total nitrogen (TN). The isotopic tracer protocol used in this study was originally described in Li et al. (2017). Each sample was analyzed three times.

The co-use of allylthiourea (ATU) and NaClO₃ effectively inhibits the N₂O production via ammonia-oxidizing bacteria (AOB) denitrification (Tallec et al. 2008), whereas N₂O emissions by heterotrophic bacteria are not significantly affected by specific chemical inhibitors. Therefore, N₂O emissions produced by heterotrophic denitrifying (HD) bacteria alone and by both AOB and heterotrophic denitrification can be determined quantitatively by batch experiments conducted either with or without inhibitors. The following three batch experiments were conducted: (I) without nitrite or inhibitor, (II) with nitrite and (III) with both ATU and NaClO₃. The inhibition experiments used herein were originally described in Li et al. (2017).

The NH₄⁺-N, NO₂⁻-N, NO₃⁻-N and TN concentrations were determined by a flow injection analyzer (Hach Quickchem 8500S2, Hach Inc., USA). N₂O concentration was measured using a gas chromatograph (Agilent 7820A, Agilent Technology Inc., USA). The isotopic compositions of N₂O and N₂ were analyzed via gas chromatography-isotope ratio mass spectrometry (Delta V advantage, Thermo Electron, Germany) and used to calculate the accumulation of ¹⁵N-labeled N₂O as ⁴⁵N₂O (¹⁴N¹⁵N¹⁶O) and ⁴⁶N₂O (¹⁵N¹⁵N¹⁶O) and N₂ as ²⁹N₂ (¹⁴N¹⁵N) and ³⁰N₂ (¹⁵N¹⁵N). The isotopomer analysis method used herein was originally described in Li et al. (2017).

Fresh biofilm samples were collected and then immediately placed in storage at −20 °C. Sequencing using an Illumina MiSeq PE300 sequencing platform was performed by Majorbio Bio-Pharm Technology Co., Ltd. Primers 338F and 806R were used to amplify the V3-V4 regions of the bacterial 16S rDNA genes from the extracted DNA. The gene pyrosequencing methods used herein were originally reported in Wang et al. (2017).

Thin sections of biofilms were mounted on Teflon-coated glass slides. Hybridization was performed using hybridization probes NSO1225, GAM42a-T1038 (Pxyn-440) and AMX820 to identify AOB, HD and Anaerobic Ammonium Oxidation (anammox) bacterial populations, respectively. The probes were labeled with fluorescein isothiocyanate (FITC), Cy5 and Cy3 at the 5′ end. Hybridized samples were observed with a LSM700 confocal laser-scanning microscope equipped with three diode lasers (488, 555 and 639 nm; Carl Zeiss, Oberkochen, Germany). The fluorescent in situ hybridization (FISH analysis) method used herein was originally reported in Ali et al. (2016).

Details about the computational procedures about the ΔG for the overall balanced reaction are provided in the Supplementary Information.

The nitrogen removal rates in the SBBRs with C/N ratios of 0, 0.67 and 1.00 were 81.4, 82.3 and 84.1%, respectively (Figure 1(a)–1(c)). The nitrogen removal reported in previous studies using the CANON process coupled with denitrification was 70–84% (Chen et al. 2009; Liang et al. 2014; Zhang et al. 2017). Excellent nitrogen removal was observed in the three SBBRs studied herein. The nitrogen removal rates increased from 81.4 to 84.1% when the C/N ratio increased from 0 to 1, indicating the carbon source increased nitrogen removal in the process. Nitrogen removal was improved using enhanced heterotrophic denitrification via addition of a carbon source. These results were in agreement with those of Zhang et al. (2017). Anammox and heterotrophic denitrification are the main nitrogen removal pathways present in the process. The contributions of the two pathways to TN removal with different C/N ratios are presented in Figure 1(d). Anammox was the primary nitrogen removal pathway in the process, and the contribution of anammox to TN removal was greater than 72%. The contribution of heterotrophic denitrification to TN removal increased from 10 to 28% as the C/N ratio increased from 0 to 1, indicating the carbon source significantly affected the contributions of heterotrophic denitrification and anammox to the TN removal of the process.

The transformation of N to N₂O is closely related to the nitrite production and consumption of the CANON process (Okabe et al. 2011; Yan et al. 2019). Nitrite production or consumption was regarded as an endpoint or starting point for the thermodynamic calculations used for the three N₂O emission pathways, NH₂OH oxidation, nitrifier denitrification and heterotrophic denitrification (described in Supplementary Information). The half-reactions and overall reactions of the different N₂O production pathways are shown in Tables 1 and 2. Bacterial growth and intracellular substance synthesis depends on the substrate allocation between catabolism and anabolism. f_s⁰ represents the substrate used for cell synthesis. A higher f_s⁰ is beneficial for the synthesis and growth of bacteria (Rittmann & McCarty 2012). The f_s⁰ values for the NH₂OH oxidation, nitrifier denitrification and heterotrophic denitrification pathways were 0.28, 0.56 and 0.70, respectively. N₂O was produced by AOB via NH₂OH oxidation and denitrification. AOB denitrification exhibited a higher f_s⁰ with regard to substrate allocation than NH₂OH oxidation. The value of f_s⁰ in nitrifier denitrification was two-fold that of the NH₂OH oxidation. Therefore, AOB preferred the denitrification pathway because of its growth advantage when N₂O was produced by the AOB. The energy reaction is intended to provide the energy for microbial synthesis, which is a key step for the survival of microorganisms. The ΔG_e⁰ values for the observed energy reactions (f_e⁰R_a – R_d) of the NH₂OH oxidation, nitrifier denitrification and heterotrophic denitrification pathways were −43.99 kJ/eeq, −90.71 kJ/eeq and −100.90 kJ/eeq, respectively (Table 2), indicating that the energy reactions were thermodynamically favorable. The released energy can be used for microbial metabolism. The energy released by nitrifier denitrification and heterotrophic denitrification was significantly higher than that released by the NH₂OH oxidation. The ΔG_s⁰ values for the observed synthesis reactions (f_s⁰R_c) of the NH₂OH oxidation, nitrifier denitrification and heterotrophic denitrification pathways were 29.00 kJ/eeq, 56.42 kJ/eeq and 32.64 kJ/eeq respectively (Table 2), indicating that the synthesis reactions were not spontaneous. The energy released by the energy reactions was needed to drive the synthesis reactions of the NH₂OH oxidation, nitrifier denitrification and heterotrophic denitrification pathways. However, the ΔG values for the overall reactions of the NH₂OH oxidation, nitrifier denitrification and heterotrophic denitrification pathways were −14.99 kJ/eeq, −34.29 kJ/eeq and −68.26 kJ/eeq, respectively, indicating the overall reactions were thermodynamically favorable. The ΔG value of N₂O production by nitrifier denitrification was more than twice that of the NH₂OH oxidation, indicating that the N₂O was more easily produced by nitrifier denitrification. The nitrifier denitrification is a main pathway of N₂O emissions by AOB. This result was consistent with the result of Wrage et al. (2001). The ΔG values of the nitrifier denitrification and heterotrophic denitrification reactions were greater than that of the NH₂OH oxidation. Therefore, N₂O was easier to produce using either nitrifier and heterotrophic denitrification than using NH₂OH oxidation. A lower energy release by the energy reactions was a limiting factor for the production of N₂O by the NH₂OH oxidation pathway. Therefore, from a thermodynamic perspective, the N₂O production by denitrification (nitrifier and denitrifier) was the main factor that contributed to the N₂O emissions.

The nitrification and denitrification emission characteristics of N₂O during batch experiments are shown in Figure 2. The N₂O was primarily generated during the first half of the batch experiments (0–6 h) (Figure 2(a) and 2(b)). N₂O production during the first 6 h of operation when C/N ratios of 0, 0.67 and 1 were used accounted for 72.4, 72.5 and 70.6%, respectively, of the total N₂O production. The N₂O production rates obtained with different C/N ratios increased over the first 3 h of the batch experiments; subsequently, they decreased significantly as the reaction progressed. The denitrification N₂O production rates when C/N ratios of 0, 0.67 and 1 were used increased from 22.27, 22.50 and 16.45 μg N₂O-N·g⁻¹VSS·h⁻¹ to 30.37, 26.62 and 24.62 μg N₂O-N·g⁻¹VSS·h⁻¹, respectively, during the first 3 h of the batch experiments and then gradually decreased to 4.91, 4.09 and 7.54 μg N₂O-N·g⁻¹VSS·h⁻¹. Similar changes in the nitrification N₂O production rates were observed in batch experiments with different C/N ratios. The average denitrification N₂O production rates were 17.00, 15.12 and 13.97 μg N₂O-N·g⁻¹VSS·h⁻¹, respectively. The average denitrification N₂O production rates decreased as the C/N ratio increased. The results indicated that the addition of certain carbon sources inhibits N₂O emissions produced by denitrification in the process. The average N₂O production rates by nitrification using C/N ratios of 0, 0.67 and 1 were 0.92, 1.16 and 0.95 μg N₂O-N·g⁻¹VSS·h⁻¹, respectively. The carbon source had no significant effect on N₂O emission by nitrification in the process. In addition, the average denitrification N₂O production rates when different C/N ratios were used were far greater than the average nitrification rates, indicating denitrification was the primary N₂O emission source in the CANON process coupled with denitrification (Figure 2(d)).

The total N₂O production with C/N ratios of 0, 0.67 and 1 was 228.04, 205.57 and 190.4 μg N₂O-N·g⁻¹VSS, respectively. These results indicated that certain carbon sources aid in the reduction of N₂O emissions in this process. Decreased N₂O production using denitrification was the primary contributor to reducing N₂O emissions in this study. N₂O emissions from TN with C/N ratios of 0, 0.67 and 1 were 1.62, 1.48 and 1.32%, respectively. N₂O emissions from TN ranged from 1.2 to 2% in previous studies using partial nitritation/anammox (Kampschreur et al. 2009; Weissenbacher et al. 2010; Castro-Barros et al. 2015). The contributions of nitrification to the total N₂O generation with C/N ratios of 0, 0.67 and 1 were 5.44, 6.99 and 7.64%, respectively. The contributions of denitrification to the total N₂O production were 94.56, 93.01 and 92.36%, respectively. The carbon source produced a significant effect on the contributions of denitrification and nitrification with regard to the total N₂O production of the CANON process coupled with denitrification, which led to a decreasing trend in N₂O production by denitrification as the C/N ratio increased.

The nitrogen transformation and emission characteristics of N₂O via nitrifier and heterotrophic nitrification during the inhibition experiments with different C/N ratios are presented in Figure 3. The nitrogen transformation and removal performance characteristics are shown in Figure 3(a)–3(c). The NO₂⁻-N removal rate obtained in inhibition experiment II using different C/N ratios was significantly higher than that obtained in inhibition experiment III, indicating inhibitors significantly affected AOB denitrification. The NO₃⁻-N production rate obtained in inhibition experiment II was lower than that obtained in inhibition experiment III because of the inhibition of nitrite reductase (NiR) activity by NaClO₃. The TN removal rate decreased after the addition of inhibitors with different C/N ratios, indicating AOB denitrification contributes to TN removal. The TN removal rates obtained in inhibition experiment II when C/N ratios of 0, 0.67 and 1 were used were 0.41, 0.43 and 0.45 mg N·g⁻¹·VSS·h⁻¹, respectively. The TN removal rate obtained in inhibition experiment II increased by 9% as the C/N ratio increased, indicating the presence of a carbon source promotes denitrification. The TN removal rates obtained in inhibition experiment III with C/N ratios of 0, 0.67 and 1 were 0.32, 0.35 and 0.39 mg N·g⁻¹·VSS·h⁻¹, respectively. The TN removal rate increased by 22.0% as the C/N ratio increased. These results indicated that heterotrophic denitrification is more sensitive to carbon source and that the increased heterotrophic denitrification is the main factor that contributes to the increase in the TN removal rate.

The N₂O production by denitrification decreased gradually with increasing C/N ratio. The N₂O production by nitrifier denitrification with C/N ratios of 0, 0.67 and 1 was 134.04, 78.64 and 80.16 μg N₂O-N·g⁻¹VSS, respectively. The N₂O production by heterotrophic denitrification with C/N ratios of 0, 0.67 and 1 was 76.64, 84.72 and 80.97 μg N₂O-N·g⁻¹VSS, respectively. These results indicated the addition of a carbon source reduced N₂O production by nitrifier denitrification; however, it had no significant impact on N₂O production via heterotrophic denitrification. The N₂O emissions from TN with C/N ratios of 0, 0.67 and 1 were 1.52, 1.28 and 1.11%, respectively. The contributions of nitrifier denitrification to the total N₂O production with C/N ratios of 0, 0.67 and 1 were 63.62, 48.14 and 49.75%, respectively. The contributions of heterotrophic denitrification to the total N₂O production with C/N ratios of 0, 0.67 and 1 were 36.38, 51.86 and 50.25%, respectively. The decreased nitrifier denitrification was the primary contributor to the decrease in N₂O production via denitrification. Therefore, based on the results presented in the section N₂O emissions via nitrification and denitrification with different C/N ratios, the decreased nitrifier denitrification resulting from the addition of a carbon source caused a decrease in the total N₂O emissions of the process.

The compositions of N₂O emissions in the SBBRs using different C/N ratios are shown in Figure 4(a). The N₂O emissions by hydroxylamine oxidation, AOB denitrification and heterotrophic denitrification in the SBBRs using different C/N ratios account for 5.4–7.6%, 45.2–60.8% and 33.8–47.2% of the N₂O produced, respectively. The results indicated AOB denitrification and heterotrophic denitrification were the primary N₂O emission pathways in the CANON process coupled with denitrification. These results were consistent with the thermodynamic prediction results in which the ΔG values of the nitrifier denitrification and heterotrophic denitrification were significantly greater than those of the NH₂OH oxidation, which was favorable for N₂O production (see Prediction of N₂O emission potential based on thermodynamics). The N₂O emissions produced by AOB were significantly higher than those produced by HD bacteria. In addition, the N₂O emissions by AOB via hydroxylamine oxidation comprised only approximately 1/8 of the N₂O emissions by AOB via denitrification. With regard to N₂O emissions, AOB obtained a higher f_s⁰ for substrate allocation via denitrification than that of NH₂OH oxidation. A higher f_s⁰ is beneficial for the synthesis and growth of bacteria. Therefore, AOB preferred the denitrification pathway because of a growth advantage over NH₂OH oxidation in which N₂O needs to be produced by AOB. This hypothesis could explain why the N₂O emissions by AOB via denitrification were significantly higher than those produced by NH₂OH oxidation.

The spatial distributions of AOB, nitrite-oxidizing bacteria (NOB) and anammox bacteria in the biofilm were determined using FISH (Figure 5). NSO1225-stained AOB were detected primarily in the outermost surface layer of the biofilm, and the distribution of AOB was localized primarily in the surface layer of the biofilm (approximately 0–300 μm). GAM42a-T1038 and Pxyn-440-stained HD bacteria and Amx820-stained anammox bacteria cells were observed underneath the AOB cell layer. The distributions of anammox bacteria and HD bacteria were primarily localized at depths greater than 400 μm in the anaerobic microzone within the biofilm. In addition, the clearly stratified spatial distributions of AOB and HD bacteria and anammox bacteria were observed. AOB and HD bacteria were the primary contributors to N₂O emissions. Therefore, the N₂O source region depended on the spatial distribution of aerobic AOB and HD bacteria.

The bacterial community structures in the three SBBRs are shown in Figure S5. Thauera and Denitratisoma were reported previously to be HD bacteria, and Nitrosomonadaceae and Nitrosomonas were reported previously to be AOB (Hu et al. 2010). The relative abundance of the genera of HD bacteria increased significantly from 3.4 to 12.8%, and the relative abundance of the AOB genera decreased from 5.9 to 3.6% as the C/N ratio increased. Although the relative abundance of HD bacteria significantly increased as the C/N ratio increased, heterotrophic denitrification N₂O emissions showed little change (see N₂O emissions via nitrifier and heterotrophic denitrification). This result indicated the carbon source had a limited effect on heterotrophic denitrification N₂O production. However, nitrifier denitrification N₂O production significantly decreased as the C/N ratio increased. Therefore, the inhibition of AOB by a carbon source effectively reduced N₂O production from nitrifier denitrification. N₂O emissions from nitrifier denitrification decreased as the C/N ratio increased, which was the main contributor to the reduction in the total N₂O emission in the process. The total N₂O emissions from TN decreased from 1.62 to 1.32% when the C/N ratio increased from 0 to 1 (see N₂O emissions via nitrification and denitrification with different C/N ratios). Therefore, the addition of an appropriate carbon source aids in the reduction of N₂O production from the CANON process coupled with denitrification.

N₂O emissions were investigated based on the thermodynamics and nitrogen transformation of a single-stage CANON process coupled with denitrification. The ΔG values of nitrifier denitrification and heterotrophic denitrification were significantly greater than that of NH₂OH oxidation, indicating nitrifier denitrification and heterotrophic denitrification were favorable for N₂O production. The value of f_s⁰ in AOB denitrification was two times that of the NH₂OH oxidation during substrate allocation. Therefore, AOB preferred the denitrification pathway because of its growth advantage when N₂O was produced by the AOB. The N₂O emissions by hydroxylamine oxidation, nitrifier denitrification and heterotrophic denitrification in the SBBRs using different C/N ratios accounted for 5.4–7.6%, 45.2–60.8% and 33.8–47.2%, respectively, of the N₂O emissions. The total N₂O emissions from TN decreased from 1.62 to 1.32% when the C/N ratio increased from 0 to 1. The carbon source had no significant effect on the N₂O emission by NH₂OH oxidation and heterotrophic denitrification in this process. However, nitrifier denitrification N₂O production decreased significantly as the C/N ratio increased. The reduction in N₂O emissions by nitrifier denitrification was the primary contributor to reducing the total N₂O emissions in the process. The results of this study showed the addition of an appropriate carbon source aids in the reduction of N₂O production in the CANON process coupled with denitrification.

This work was supported by the National Natural Science Foundation of China (51878091 and 51278509), the Chongqing Science and Technology Bureau (cstc2018jcyjAX0610) and the Fundamental Research Funds for the Central Universities (2019CDQYCH036 and 2019CDXYCH0026).

All relevant data are included in the paper or its Supplementary Information.