Single-stage PN/A technology treating saline ammonia-rich wastewater: finding the balance between efficient performance and less N₂O and NO emissions

Abstract

A single-stage partial nitrification/anammox (PN/A) process is suitable for treating high ammonia industrial wastewater with energy savings. However, little is known about nitrous oxide (N₂O) and nitric oxide (NO) emissions from single stage nitrification/anammox (PN/A) process when treating ammonia-rich industrial wastewater which contains salt (NaCl). In this study, the effect of increasing salinity (0, 5, 10, 15 and 20 g NaCl per L) on N₂O/NO emissions and total nitrogen removal rate (TNRR) were comprehensively examined in an intermittent aeration single-stage partial nitritation and anammox (PN/A) system, where anammox bacteria coexisted with ammonia oxidation bacteria and heterotrophic denitrifiers. Results showed that N₂O emissions in aerobic stages accounted for at least 68% of the total emission at all salinities, and N₂O was predominantly produced via hydroxylamine oxidation. When without NaCl addition, the N₂O and NO emission factors (N₂O & NO–N_gaseous/TN_influent) were 0.43% and 0.015% respectively, and the TNRR stayed at 0.19 kg N per m³ per d. Low salinities of 5 and 10 g NaCl per L increased N₂O emission factors to 0.75% and 0.52% respectively; while, relatively high salinities of 15 and 20 g NaCl per L decreased N₂O emission factors to 0.23% and 0.16%. The TNRR remained over 0.18 kg N per m³ per d under 5–15 g NaCl per L, but deteriorated to 0.15 kg N per m³ per d under 20 g NaCl per L. Our results proved that single-stage PN/A system is a promising technology to treat ammonia-rich saline wastewater, where the best balance can be achieved between the efficient nitrogen removal performance and less N₂O/NO emissions consideration.

---

1. Introduction

Large amounts of salt are consumed by industries, mainly for food processing, producing pharmaceuticals, tanning and refining processes, generating enormous quantities of saline wastewaters that are rich in organic matter, ammonia and salinity (1.4 to 40 g NaCl per L) (Table S1†). The organic carbon in wastewater can be recovered as an energy source in anaerobic digestion.¹ The remaining high ammonia effluent can be treated by anaerobic ammonia oxidation (anammox) based autotrophic nitrogen removal technology (i.e., single or two-stage PN/A process), with advantages of 60% aeration saving and zero external carbon requirements.^2–4 However, high salinity in wastewater is a remaining problem. It introduces an ionic imbalance and hyperosmotic stress in bacteria,⁵ which leads to negative effects on the nitrogen removal in wastewater treatment plants.⁶

With ever-tightening standards for wastewater discharge, interest in saline wastewater treatment processes has been rapidly increased over the past two decades. There have been some studies about the effect of salinity on the anammox related technologies,^7–14 most of which mainly focus on the performance of anammox or partial nitritation (PN) independently after salt additions.^7–10,14 Compared with two-stage partial nitritation and anammox (PN/A) systems, single-stage PN/A systems has the characteristics of less occupation area, simpler processing control and lower risk of nitrite inhibition, but also features more complicated equilibrium among different functional microbes. Meanwhile, there still have disagreements regarding the tolerance of functional microbial to high salinity. Salinity higher than 15 g NaCl per L was found to partially inhibit AOB and almost completely inhibit AnAOB in CANON process,¹² while Windey and coworkers found that nitrifying process was the limiting step of the one-stage PN/A process at salinity of 30 g NaCl per L.¹³ Under these circumstances, it remains much to be done on saline wastewater treatment by a single-stage anammox technology to develop sophisticated controlling strategy for the full-scale application.

In single-stage PN/A systems, intermittent aeration and continuous aeration are two main aeration regimes, where dissolved oxygen (DO) is generally set at low levels to achieve nitrite accumulation but without inhibition effect on anammox bacteria. However, low DO operation, ammonia-rich and high salinity in the influent were reported to be likely to trigger more nitrous oxide (N₂O) and nitric oxide (NO) production.^15–19 Nitrous oxide is a powerful and stable greenhouse gas with a global warming effect of 298 times higher than carbon dioxide over a 100 year time horizon.²⁰ N₂O is an important compound involved in atmospheric photochemistry that causes ozone layer depletion.^21,22 Globally, biological nitrogen removal of ammonia-rich wastewater release significant amount of N₂O,¹⁶ contributing up to 10% of the anthropogenic N₂O emissions.²³ The N₂O emission factors (N₂O–N_gaseous/TN_influent) of PN/A processes that reported varying from 0.98% to 6.6% are much higher than those measured from conventional biological nitrogen removal processes.^23–25 More N₂O and NO emitted from single-stage PN/A process may offset the benefit of PN/A systems associated with their energy savings and corresponding CO₂ emission reductions in carbon footprint. However, far fewer studies have comprehensively assessed the dynamic emissions of N₂O and NO in single-stage PN/A system treating saline wastewater.

In general, N₂O and NO are generated as bioconversion products of nitrification, nitrifier denitrification and hydroxylamine oxidization in aerobic partial nitrification reactor.^26,27 As N₂O is not involved in anammox reaction, N₂O emission observed in anammox bioreactor is generally attributed to incomplete denitrification by heterotrophic denitrifiers.^28,29 Meanwhile, anammox reaction and heterotrophic denitrification both generate NO as intermediates.^16,30 Considering that bacteria activity³¹ and enzyme activities^19,32 could be inhibited by high salinity, it is urgent to assess the influence of high salinity on N₂O/NO production characteristics.

In this study, N₂O and NO emission characteristics in both gaseous and liquid phases with stepwise increased salinity were evaluated from a lab-scale sequencing batch reactor (SBR) performing intermittent aeration PN/A process. 16S rRNA high-throughput sequencing technique was employed to link the nitrogen conversion related microbial structure to nitrogen removal performance. The objectives of this study were to (1) evaluate the feasibility of single-stage PN/A system treating ammonia-rich saline wastewater achieving the stable nitrogen removal performance and low N₂O/NO emission simultaneously, and (2) identification of the salinity effect on N₂O/NO production pathway under aerobic/anoxic conditions.

2. Material and methods

2.1 Sequencing batch reactor configuration and operation

A laboratory-scale SBR performing a single-stage PN/A process was made of a glass vessel with a water-bath system at 35 °C. It has a working volume of 2 L and a headspace volume of 1 L, operating under an intermittent aeration scheme with synthetic wastewater. One cycle lasted for 8 h, including 5 min feeding, three 35 min aerobic/105 min anoxic stages, 40 min settling phase, 5 min effluent withdrawal and 10 min idle phase. The water exchange ratio was approximately 0.5 and hydraulic retention time was 16 h. Air was supplied through a diffuser at the bottom of the reactor together with an agitator to promote the oxygen transfer. Air flow was controlled with a rotameter to keep the DO concentration around 0.33 mg per L.

2.2 Seeding sludge and feeding media

The seeding anammox sludge was originated from a laboratory-scale anammox SBR dominated by Candidatus “Kuenenia”, with an stable specific anammox activity of approximately 0.31 ± 0.01 g N per g VSS per day. The seeding partial nitritation sludge was obtained from a lab-scale PN parent reactor running at a oxygen condition. The activated anammox and partial nitritation sludge were seeded with the ratio of 1 : 1 to set up the single-stage PN/A system reactor.

The single-stage PN/A reactor was fed with synthetic wastewater. Ammonia was added in the form of NH₄Cl to reach the concentration of 200 ± 20 mg N per L and the other parameter settings were summarized in Table 1. The mineral medium consisted of (g per L): KHCO₃, 1.25; KH₂PO₄, 0.02; CaCl₂, 0.02; MgSO₄·7H₂O, 0.08; FeSO₄·7H₂O, 0.015; Na₂EDTA, 0.02. Concentrated solution of trace elements was provided 5 mL per L feeding medium. The trace element solution contained (g per L): EDTA, 7.5; H₃BO₃, 0.035; ZnSO₄·7H₂O, 1.075; CuSO₄·5H₂O, 0.625; MnCl₂·4H₂O, 0.495; NaMoO₄·2H₂O, 0.55; CoCl₂·6H₂O, 0.6; and NiCl₂·6H₂O, 0.475.

Table 1 Characteristics of the single-stage PN/A system reactor operated under different salinity levels

Parameters	Time range	Salt level	Volumetric exchange ratio	Volumetric nitrogen removal rate	Dissolved oxygen
Unit	Day	g NaCl per L	1	kg N per m^3 per d	mg per L
Period A	65–122	0	0.5	0.19 ± 0.03	0.30 ± 0.14
Period B	123–184	5	0.5	0.18 ± 0.02	0.33 ± 0.14
Period C	185–229	10	0.5	0.22 ± 0.03	0.35 ± 0.11
Period D	230–268	15	0.5	0.18 ± 0.02	0.34 ± 0.06
Period E	269–300	20	0.5	0.15 ± 0.03	0.35 ± 0.11

---

2.3 Stepwise addition of NaCl to single-stage PN/A system

The new started single-stage PN/A system was operated over two months to achieve a steady performance, reaching the total nitrogen removal rate (TNRR) of 0.21 ± 0.09 kg N per m³ per d. The reagent grade sodium chloride was added into influent by 5 mg per L stepwise increase. Once the influent salinity changed, it was maintained 30–60 days to regain a stable TNRR. The entire experiment was divided into five operational phases with the different salinities in the influent: period A (0 g NaCl per L, 65–122 d), period B (5 g NaCl per L, 123–184 d), period C (10 g NaCl per L, 185–229 d), period D (15 g NaCl per L, 230–268 d), period E (20 g NaCl per L, 269–300 d).

2.4 N₂O and NO emission monitoring campaign

The N₂O and NO samples were collected within a single cycle on the day before the salinity change (i.e., 122, 184, 229, 268 and 300 d). Gas-phase N₂O concentration were monitored in aerobic/anaerobic stages. From the headspace of the sealed reactor, N₂O and NO emissions at aerobic stage were continuously monitored with online N₂O analyzer (AO2020 Uras26 infrared analyzer, ABB Automation GmbH, Germany) every 60 seconds and NO analyzer (ML410 Nitric Oxide Monitor, 2B Technologies, USA) every 30 seconds. The N₂O online analyzer has dual measuring ranges of 0–100 ppm and 0–1000 ppm, and the 0 to 1, 000 ppm measuring range was used in the present study with a detection limit of ±1 ppm; the NO online analyzer has a detection range of 0–2000 ppb (0–2 ppm) with a detection limit of ±1.5 ppb. Gas chromatograph (Agilent 7820, USA) equipped with an electron capture detector (ECD) and a Porapak Q column (3 m), using 95% argon and 5% methane as carrier gas, was used for the determination of off-line data at anoxic stage and inspection of the accuracy of the online analyzers. Temperatures of the injector, column and detector were 105 °C, 100 °C and 280 °C, respectively.

Liquid-phase N₂O concentrations over the whole reaction time were detected using headspace sampling method described by Okabe et al.²⁹ Briefly, the collected liquid samples (5 mL each) were transferred to a 20 mL glass vial, which was sealed by a butyl-rubber stopper and aluminum cap. After the glass vial was shaken for a few minutes, N₂O in the gas phase was measured by the gas chromatograph as described above. Because the concentration of dissolved NO was less than 0.05% of that in released gases and for the very low solubility of NO (K_H 25 °C = 1.19 × 10⁻³ M per atm),³³ the dissolved NO was not determined in this study.

2.5 Analytical methods

The typical cycle measurement was conducted about 1–3 times per week. During every typical cycle measurement, mixed sludge samples were collected from the reactor at the beginning and at the end of each aerobic/anoxic stage and then filtered through 0.45 μm filters. Collected samples were stored at 4 °C in a refrigerator until analysis. Nitrite, ammonium and nitrate were determined according to the Standard Methods.³⁴ The pH (WTW pH 3310 with a SenTix 41 electrode, Weilheim, Germany) and the DO concentration (WTW Oxi 3315 with a FDO 70x IQ (SW oxygen electrode, Weilheim, Germany)) were measured through the single cycling.

2.6 Calculation of N_xO emission and production rates

The N_xO emission rates in aerobic stages were calculated based on ideal gas law.where r_NxO,e is the net emission rate, mg N per min; ω_NxO × N × Q_air is the N_xO emitted from the aerobic stage; ω_NxO is the N_xO concentration monitored by online N₂O & NO analyzers, ppm for N₂O and ppb for NO; N is the gas dilution factor, N = 2; Q_air is the gas flow entered in the online machine, L per min; R is ideal gas constant, 8.314 Pa per K per mol; P is the barometric pressure, 1.01 × 10⁵ Pa; T is gas temperature, K.

The N₂O production rate in aerobic stage was calculated on the sum of accumulation rate estimated from dissolved N₂O in liquid phase and the stripping rate, which also equals the emission rate.

where is the mean production rate, mg N per min; ΔN₂O_l is the liquid N₂O concentration difference during the liquid sampling interval, mg N per L; V_l is the liquid mixture volume, 2 L; is the mean r_N2^a_O,e in the liquid sampling interval; Δt is the liquid sampling interval in aerobic stage, 15 min.

The N₂O emission rate from the anoxic stage was calculated based on the accumulated N₂O concentration in reactor headspace.

where is the mean emission rate, mg N per min; Δω_air is the difference of the N₂O concentration between two sampling point analyzed by GC-ECD, ppm; Δt is the time between gas sampling, min; V_g is the headspace volume of reactor, 1 L; all other symbols are the same as those in eqn (1).

3. Results

3.1 Long-term performances of the PN/A SBR

The TNRR and the ammonia removal efficiency during the whole experiment are shown in Fig. 1. During period A (without NaCl addition), the ammonia removal efficiency maintained around 72 ± 9.5% with the average TNRR of 0.19 ± 0.03 kg N per m³ per d (Table 1). With 5 NaCl g per L addition (period B), the TNRR decreased in the initial acclimation days and maintained around 0.18 kg N per m³ per d after 30 days accumulation (Fig. 1). After the salt addition up to 10 g NaCl per L in period C, the average TNRR increased to 0.22 ± 0.03 kg N per m³ per d, indicating that the mild salt addition had stimulated nitrogen removal capacity. Notably, aerobic ammonia oxidation activity, represented by ammonia oxidation rates (AOR), maintained around 21 mg N per L per min in period C (Table 2), being higher than 17 mg N per L per min of period A (without NaCl addition), indicating that the AOR was not suppressed but stimulated by 10 NaCl per L. In contrast, the nitrite oxidation rate in period C decreased by 50% relative to period A (data not shown), suggesting that the activity of nitrite oxidation bacteria (NOB) had been inhibited by 10 g NaCl per L. A similar positive effect of salinity on the nitrogen removal performance was also reported by Liu et al.,¹¹ who demonstrated that the salt suppression (10 g NaCl per L) on nitrite oxidation activity could facilitate AOB and anammox bacteria to compete for DO and NO₂⁻ respectively.

Fig. 1 Nitrogen removal performance of the partial nitritation/anammox reactor under elevated salt concentration additions from 0 to 20 g NaCl per L. TNRR represents total nitrogen removal rate. The red vertical dashed lines indicate the days in which influent salinity was changed.

Table 2 Ammonia oxidation rates and nitrite concentration of different stages within a single cycle under different salt concentration range^a

NaCl, g NaCl per L	Aerobic/anoxic	NO2^− accumulation, mg N per L	AOR, mg N per L per min
a 1 – the first aerobic stage, 2 – the second aerobic stage; 3 – the third aerobic stage within a typical cycle.
0	Stage-1	13.91	16.79
	Stage-2	9.38	19.93
	Stage-3	9.17	21.73
5	Stage-1	10.81	18.91
	Stage-2	10.01	21.69
	Stage-3	10.52	21.71
10	Stage-1	9.81	20.15
	Stage-2	7.11	21.61
	Stage-3	7.21	21.09
15	Stage-1	9.96	15.55
	Stage-2	6.37	18.84
	Stage-3	6.79	18.78
20	Stage-1	9.03	9.72
	Stage-2	6.95	15.25
	Stage-3	5.74	16.45

---

After the salt addition increased to 15 NaCl g per L in period D, the TNRR slightly declined to 0.18 ± 0.02 kg N per m³ per d after 18 days operation, which was close to that of period A (without salt addition) (Fig. 1). Additionally, an increasing DO concentration was observed in the initial days of period D, primarily because the AOB activity had been partly suppressed by 15 NaCl g per L. To maintain the DO concentration at approximately 0.33 mg per L, the aeration rate of the single-stage PN/A system was reduced to 0.25 L per min and 0.20 L per min respectively in periods D and E (Table 1). After salt addition up to 20 NaCl g per L in period E, the TNRR was only 0.15 ± 0.03 kg N per m³ per d, decreased by 22% relative to that of period A (Fig. 1), suggesting a much stronger salt stress. Then, the salt addition was ceased during the later operational days to avoid a failure of the studied single-stage PN/A system.

Taken together, for the studied single-stage PN/A system, the TNRR maintained stable below 15 g per L. However, a relatively high salinity of 20 g NaCl per L led to a remarkable decrease in TNRR by 25% relative to period A (0 g NaCl per L). Some coastal cities, such as Hong Kong, those apply a certain percentage of seawater for toilet flushing, can generate sewage with salinity of 8–12 g NaCl per L,³⁵ which is in the similar salt range of this study that are capable of achieving successful anammox performance. The long-term nitrogen removal performances in present study proved that single-stage PN/A system have the potential to treat such ammonia-rich saline wastewater by stepwise acclimation.

3.2 N₂O and NO emission profiles under different salinity levels

Considering that the influent water compositions and operational parameters of the present single-stage PN/A system, i.e. ammonia-rich (200 mg N per L) saline wastewater (5–20 g NaCl per L) under low DO regime (0.3 mg per L), might trigger more N₂O emissions,^17–19 the N₂O and NO emission characteristics were studied comprehensively during the whole experimental period.

As shown in Fig. 2, the N₂O and NO concentration profiles under different salinities were similar during each typical cycle (Fig. 2a–e). During aerobic stages, N₂O emission generally peaked at the moment that recovering from anoxic to aerobic conditions; then they decreased to minimum values, followed by an increasing trend for the rest of the aerobic stage but never exceed the peak value obtained at the first aerobic stage. Although with a similar variation trend, the specific N₂O and NO emission amounts were totally different under elevated salinity additions. When the salinity additions increased to 5–10 g per L, the N₂O emission in aerobic stages varied in the range of 0–40 ppm, and its concentration in both second and third aerobic stages increased compared to those of without NaCl addition (Fig. 2a–c). During period C (15 g NaCl per L), the N₂O emission amounts declined to 0–20 ppm. After the salinity addition reached 20 g per L, obvious lower N₂O emissions ranging from 0–15 ppm was detected (Fig. 2e).

Fig. 2 Representative N₂O and NO concentration profiles over a single cycle under different salinity concentrations ((a) to (e), 0 to 20 g NaCl per L). Aerated periods are shaded.

As to NO, their variation trend during typical cycles under different salinities was similar to N₂O, but the emission concentrations were two orders of magnitudes lower than that of N₂O. Specifically, the maximum NO emission ranging from 2.5 to 3.5 ppm occurred in the initial of each aerobic stage, which coincided well with the occurrence of the peak N₂O emission (Fig. 2b–e). However, the subsequent increasing trend of NO emissions was inconspicuous, and need to be clarified in the further study.

When regarding to anoxic stages, both gaseous and dissolved N₂O increased in the initial of the anoxic stage (Fig. 2). It is noticeable that once the increasing rate of gaseous N₂O slowed down, the dissolved N₂O concentration reached the maximum value and decreased afterwards (Fig. 2), suggesting a consumption of N₂O in the liquid phase. Specifically, the maximum dissolved N₂O concentration in the anoxic stage was 0.20 mg per L at 0 g NaCl per L (Fig. 2a), and maintained at the same level as 10 NaCl per L addition (Fig. 2c). The gaseous N₂O concentrations at the end of anoxic stages during these periods (5 and 10 g NaCl per L) maintained at 0.1–0.2 mg per L.

Increasing salinity up to 15 and 20 NaCl per L decreased both the gaseous and dissolved N₂O concentrations from anoxic stages (Fig. 2d and e). The decreasing trend in N₂O emission at 20 g NaCl per L was more remarkable relative to 15 g NaCl per L addition case (Fig. 2d and e).

Collectively, low salinity of 5 g NaCl per L stimulated N₂O and NO emissions while relatively high salinities from 15 to 20 g NaCl per L suppressed its formation, which corresponded well with the variations of TNRR (Fig. 1 and 2) under different salinity levels. The decrease in N₂O and NO emissions in this study elucidated that single-stage PN/A system has the advantage to treat high saline wastewater up to 15 g NaCl per L with the accepted nitrogen removal performance and less N₂O and NO emissions as well.

3.3 Relative abundance variations of functional N conversion-related bacteria

A robust single-stage PN/A system depends greatly on the cooperation and competition of AOB, anammox bacteria, NOB and heterotrophic denitrifying bacteria (HDB).⁴ 16S rRNA high-throughput sequencing results showed that salinity had a significant impact on the bacterial abundance and community structure in the studied single-stage PN/A system.

When without salt addition (period A), the relative abundance of both anammox bacteria and AOB to the total biomass were 35.4% and 10.52% respectively. When 5 g NaCl per L (period B) was added, the relative abundance of both anammox bacteria and AOB decreased by 92% and 67%, respectively, relative to those of period A (without salt addition) (Fig. 3). Within anammox bacteria, a remarkable shift of the dominant genus from Candidatus Jettenia to Candidatus Kuenenia was observed (Fig. 3), suggesting that a part of functional bacteria growth might has been inhibited and even dead (outcompeted) after the living environment changed from fresh to salty. Although the functional microbial abundance decreased remarkably, the irritated TNRR and AOR in period B (Tables 1 and 2) indicated that the survived anammox bacteria that had adapted to 5 g NaCl per L could be robust and ultimately guaranteed the TN removal performance. Ma et al. also reported that in an anammox upflow anaerobic sludge bed reactor, anammox activity was stimulated with decreasing VSS under the NaCl shock (5–20 g NaCl per L for 12 hours),³⁶ confirming that the system nitrogen removal performance can be maintained even the amount of functional microbes decreases to some extent.

Fig. 3 Dynamics of the functional N-conversion related microbial community under different salinity levels determined by illumina high throughput sequencing (AOB (aerobic ammonia oxidation bacteria), anammox bacteria, NOB (nitrite oxidation bacteria), HDB (heterotrophic denitrifying bacteria)). Relative abundance was defined as the number of OTUs assigned to that taxon divided by the total number of OTUs per sample.

For periods C and D (10 and 15 g NaCl per L), the relative abundance of AOB was almost double and triple to that of period B (3.44%) (Fig. 3). The relative abundances of AOB and Candidatus Kuenenia increased along with the increasing salinity, up to 43% and 403% from period C to period D respectively. As shown in Table 1, the TNRR remained stable above 0.18 ± 0.02 kg N per m³ per d in these two periods. These findings suggested that the survived functional microbes could grow under 5–15 g NaCl per L, but it had to proliferate more to keep a relatively high TNRR. In period E (20 g NaCl per L), the relative abundance of anammox bacteria and AOB decreased by 76% and 44%, respectively (Fig. 3), along with a 17% TNRR decline relative to that of 15 g NaCl per L (Table 1), suggesting that both microbial growth and activity might have been severely stressed by 20 g NaCl per L.

For another two N-conversion related microorganisms, NOB and HDB, their relative abundance variation trends were similar to that of AOB and anammox bacteria. Curiously, for the only NOB genus, Nitrospira, its relative abundance was in a continuously increase trend when salinity was less than 15 g NaCl per L, although they have been reported to be more sensitive to salt stress than AOB.¹¹ The oxygen-limited and salt stress conditions in the present study might have favored the occurrence of Nitrospira genus, which characterizes by high substrate affinity and could survive in harsh conditions.³⁷

Though the single-stage PN/A system was fed with synthetic wastewater without organic matter, HDB was detected including genus Hyphomicrobium, Paracocccus, Denitratisoma, Pseudomonas, Thermomonas and Comamonas. It was speculated that the soluble microbial products generated from other microorganisms in the same system might serve as the carbon source of HDB.²⁹ To some extent, the existence of HDB can reduce the effluent nitrate concentration and benefits for improving the TNRR of the single-stage PN/A system. Except those four bacterial groups, other OTUs presented in the CANON system may also favored saline wastewater, but they do not evolved in nitrogen cycle, which is most concerned in this study. Their existence were not well understood based on up to date knowledge and need more research in future.

3.4 Variations in N₂O and NO emission factors under elevated salinity

In a two-stage PN/A process, PN process is spatially separated from anammox process and contributes the major N₂O emission. It had been reported to be 3 to 40 times higher than that from anammox process.^29,38 Similarly, the aerobic stage that temporal separated from anammox process in a single-stage PN/A process can also be recognized as a PN process, and we could calculate the emission factors (N₂O & NO–N_gaseous/TN_influent) of aerobic/anoxic stages separately.

Under salinity of 0 g per L, the N₂O and NO emission factors for aerobic stage were 0.32% and 0.015% respectively (Table 3). With salinity addition from 5 to 10 g NaCl per L in periods B and C, N₂O and NO emission factors increased. Notably, 5 g NaCl per L had a more significant stimulation effect on N₂O and NO emission factors (0.45% and 0.017% respectively) than that on 10 g NaCl per L (0.36% and 0.013%), regardless that the relative abundance of AOB was the lowest under 5 g NaCl per L (Fig. 3). A salinity of 15 g NaCl per L decreased the N₂O and NO emission by 50% relative to that in 0 g NaCl per L. The lowest N₂O and NO emission factors occurred at 20 g NaCl per L with the values of 0.11% and 0.006% respectively (Table 3).

Table 3 NO and N₂O emissions from the single-stage PN/A system under different salinity levels

Parameters	Periods	Salinity level	N2O Emission factor (of TN)	N2O emission factor (of removed N)	N2O emission factor in aerobic stage	N2O emission factor in anoxic stage	NO emission factor in aerobic stage
Unit	Days	g NaCl per L	%	%	%	%	%
Period A	65–122	0	0.43	0.6	0.32	0.11	0.015
Period B	123–184	5	0.75	0.96	0.45	0.30	0.017
Period C	185–229	10	0.52	0.88	0.36	0.16	0.013
Period D	230–268	15	0.23	0.41	0.15	0.07	0.008
Period E	269–300	20	0.16	0.37	0.11	0.05	0.006

---

For the anoxic stage, with increasing salinity from 0 to 20 g NaCl per L, the N₂O emission factors varied at 0.30% and 0.05% during the whole operational period, which was lower than that from aerobic stages with the values of 0.45%–0.11% (Table 3). Similar to the aerobic stages, in anoxic stages the maximum N₂O emission factor of 0.30% was also obtained at 5 g NaCl per L (Table 3). Relatively high salinity of 15 and 20 g NaCl per L hindered N₂O emissions, with the lowest N₂O emission factor of 0.05% (Table 3) occurred in 20 g NaCl per L among all salinity levels.

4. Discussion

4.1 Comparison of N₂O and NO emission factors with other anammox systems

A comparison of N₂O and NO emission factors in this study with other anammox-based studies was conducted (Table 4). Although some literatures have studied N₂O emissions in intermittent aeration single-stage PN/A systems, few have described the N₂O emissions in aerobic and anoxic stages separately. Identification of the contributions of N₂O emission from PN and anammox stage is most significant for understanding the N₂O production mechanism, which is also the basis for establishment of the N₂O and NO mitigation strategies.

Table 4 N₂O and NO emission (% of load) in lab-scale and full-scale partial nitritation/anammox (PN/A) by on-line or intensive grab sampling

Wastewater type	Type of process	Aeration regime	N2O emissions (% of N-load)	NO emissions (% of N-load)	Sampling (frequency/duration) measurement	Remarks	Ref.
a Of N-removed; PN: partial nitrification process in two-stage PN/A process or aerobic stage in the single-stage PN/A process; A: anammox process in two-stage PN/A process or anoxic stage in the single-stage PN/A process; ND: no detection.
Full-scale
Municipal-digester	2-Stage	—	PN 1.7^a	PN 0.2^a	Continuous/4 days	- 74% of the N2O emissions originated from the nitritation stage	38
			A 0.6^a	A 0.003^a	Off-line
Potato processing	1-Stage	Continuous	1.2%	0.005	Continuous/5 days	- NO2^− accumulation caused higher N2O emissions	25
					Off-line
Rejection-water	1-Stage	Intermittent	1.3	<0.1	NO online	- The N2O emissions were significantly higher than the NOx emissions	41
					N2O grab every 15 min/3^a8 h
					Off-line
Industrial-potato processing	2-Stage	—	PN 5.1–6.6%	ND	Continuous/4 days	- Nitritation stage responsible for N2O emissions	53
			A below detection limitation
					On-line	- 50% decrease of the N2O emission is necessary to obtain a CO2 neutral footprint
Sludge digestion-potato processing	1-Stage	Continuous	2.0	ND	Continuous with 2 gaps/74 hours	- Intense aeration results in higher N2O emission and production and production than low aeration	54
					On-line
Lab-scale
Synthetic wastewater	2-Stage	—	PN 4.0 ± 1.5%	ND	Grab in regular time interval/150 days	- Anammox activity and N2O production are spatially separated	29
			A 0.1 ± 0.07%		Off-line	- Heterotrophic denitrification could be a main process of N2O emission in the anammox granule
Synthetic wastewater	1-Stage	Continuous	0.98 ± 0.42%	ND	Continuous for gas sample/grab in regular time for liquid sample	- N2O was produced via NH2OH oxidation and NO2^− reduction pathways equally	55
Synthetic reject water	1-Stage	Intermittent	>6–1.7^a	ND	Continuous/450 minutes	- The extant ammonium oxidation activity positively correlated with the specific N2O production rate of the system	40
					On-line
Synthetic reject water	1-Stage	Intermittent	PN 0.45–0.11	PN 0.017–0.006	Continuous in aerobic on-line		This study
			A 0.11–0.05	A ND	Grab in regular anoxic time interval/420 minutes

---

The N₂O emission factor for aerobic stage (resembling PN process) was 0.32% under salinity of 0 g NaCl per L, occupying 80% of the total N₂O emission factors. Salt additions did not change the dominant N₂O and NO emission contribution of aerobic stages to the total N₂O emission (Table 3). Our findings are similar to the observation of Kampschreur et al., who reported that 74% of N₂O emission was related to PN process in two-stage full-scale PN/A process.³⁸

A literature review showed that the N₂O emission factors of the PN process from two-stage PN/A systems are in the range of 1.7–6.6%, which is generally higher than the values from 0.98% to 2.0% of the continuously aerated single-stage PN/A systems (Table 4). The key parameters resulting in high N₂O production in the spatial/temporal separated PN process include high nitrite accumulation, high influent ammonium, DO concentration higher than 1.0 mg per L as well as rapidly changing operational conditions.^15,17,39 In the present study, the salt additions at relatively low concentrations (e.g., 5 and 10 g NaCl per L) could also trigger more N₂O and NO production when compared with that of without salt addition. Nevertheless, even with 5 g NaCl per L addition that stimulated the highest N₂O emissions among all salt additions, the N₂O emission factor in aerobic stage of the present PN/A process was only 0.45% (Table 3), which is much lower than 1.7% to 6.6% from the reported lab-scale intermittent aerated single-stage PN/A systems (Table 4), where influent containing 500–1400 mg NH₄⁺–N per L. The relatively low influent ammonium concentration of 200 mg NH₄⁺–N per L might contribute to the reduced N₂O emissions as the nitrite accumulation concentration of 5.74–13.91 mg N per L (Table 2) and the DO value around 0.33 mg per L in this study were in the similar range of the previous studies.^40,41 Meanwhile, it cannot be excluded that the sequencing batch operating mode, the low aeration frequency, the microbial composition of the sludge communities and feed water composition may also contribute to the low N₂O emission factor in the present study. The NO emission factor in PN process was approximately 0.015–0.006% in this study, which was within the reported range from 0.1% (ref. 41) to 0.005% (ref. 38) of the reviewed literature.

To the best of our knowledge, very few studies have reported the value of N₂O emission factor in anammox reactor or in anoxic stage of single-stage PN/A systems. In this study, the N₂O emission factor from anoxic stages under 0 g NaCl per L were 0.11%, similar to the value of 0.1 ± 0.07% reported by Okabe et al.²⁹ We speculate that heterotrophic denitrifier might have contributed this part of N₂O productions (details seen Section 4.3).

4.2 Possible N₂O production pathway in aerobic stages

Previous works clarified that AOB can produce N₂O in nitrification process.¹⁷ So far, the production of N₂O through AOB is mainly via two pathways: hydroxylamine oxidation favored DO higher than 0.8 mg per L (ref. 42–44) and nitrifier denitrification favored at low DO around 0.3 mg per L (ref. 16 and 45). It was reported that ammonia oxidation activities (represented by AOR) during aerobic stages have a close correlation with the N₂O production.^26,40 In the present study, the DO concentration and the temperature were not changed in these periods. Both N₂O emission (Fig. 2) and the AOR (Table 2) increased along with increasing salinities from 0 to 10 g NaCl per L. The AOB abundances (Fig. 2) under salinities of 5 and 10 also present a rising tendency, but they were less than that of 0 g NaCl per L. The higher N₂O emission in these periods may primarily due to the increasing AOR rather than the low AOB abundances. When salinity addition increased to 15 and 20 g NaCl per L (periods D and E), the AOR rapidly declined primarily due to the high salinity inhibition (Fig. 4). The simultaneous decrease in the biological activities (AOR), the AOB abundance (Fig. 2) ultimately decreased N₂O and NO emissions in periods D and E. Our results proved that the AOR, which varied along with salinity, ultimately determines the N₂O production and impacts its emission patterns. Law et al. developed four different metabolic computer models to elucidate the mechanisms of aerobic N₂O production in a nitritation reactor.⁴² They addressed that the relationship between AOR and N₂O production rate are linear for nitrifier denitrification pathways while exponential for hydroxylamine oxidation pathway. In this study, although the DO was kept 0.33 mg per L, which favors N₂O production via nitrifier denitrification pathway, the N₂O production rate was exponentially related to its AOR (Fig. 4). This exponential relation indicated that the N₂O production during the aerobic stage in the present study may be controlled by hydroxylamine oxidation process. It is postulated that the activities of nitrifier-encoded NirK and cNor that reduce NO₂⁻ to NO and N₂O were inhibited by salt, resulting the dominance of hydroxylamine oxidation pathway for N₂O production in the aerobic stage.

Fig. 4 Average N₂O production rates versus AOR. Error bars represent standard deviations of N₂O production rate three aerobic stages.

4.3 The impact of salinity on N₂O production in anoxic stages

Unlike PN processes, gaseous N₂O are marginally detected from anammox processes,³⁸ because N₂O is not an intermediate for anammox reaction.⁴⁶ Relatively low N₂O emission detected from anammox process in the two-stage PN/A system or anoxic stages in single-stage PN/A system is likely produced by other nitrogen-conversion related microbial species,²⁹ e.g., through nitrifier denitrification by AOB and incomplete denitrification by HDB pathways.

In anoxic stage, the average N₂O emission rate could be obtained from the accumulated N₂O concentration in the reactor headspace. As shown in Fig. 5, N₂O emission rate was positively correlated with dissolved N₂O concentration, due to the fact that the mass transfer process is governed by N₂O concentration gradient between liquid and gaseous phase. The maximum N₂O emission rate of 0.0032 mg N per min occurred at 5 g NaCl per L and kept at a constant value of 0.0015 mg N per min under 10 g NaCl per L (Fig. 5). The TNRR under 5 and 10 g NaCl per L increased relative to that of without NaCl addition, indicating that the anammox and/or denitrification activity was stimulated. However, nitrate concentrations in effluent during these two periods were not increased (Fig. 1), possibly having been consumed by denitrification. The observed increasing N₂O emission rates could result from increased denitrification efficiency combined with an inhibited NOs activity by salt.^19,32 Nevertheless, nitrifier denitrification driven N₂O production in anammox systems cannot be excluded as AOB amount in the studied SBR was 23 times larger than HDB amount (Fig. 3).

Fig. 5 N₂O emission rates over a single cycle and correspondent dissolved N₂O concentration under different salinity levels ((a) to (e), 0–20 g NaCl per L). Aerated periods are shaded.

After salinity increased to 15–20 g NaCl per L, the N₂O emission rates and dissolved N₂O concentrations decreased during anoxic stages (Fig. 5). The relative abundance of anammox bacteria and HDB slightly increased at 15 g NaCl per L but the TNRR that increased at 10 g NaCl per L decreased to 0.18 kg N per m³ per L (Table 1), suggesting that the microbial activity was partly inhibited by 15 g NaCl per L. Especially, the relative abundance of anammox bacteria and HDB decreased again at 20 g NaCl per L (Table 1 and Fig. 3), and N₂O emission rate decreased to below 0.001 mg per min correspondingly (Fig. 5), which may arise from the sensitivity of nirK and nirS in HDB to salinity.⁴⁷ Meanwhile, impaired AOR in aerobic stages at 15–20 g NaCl per L contributed to decreased NO₂⁻ accumulation at the end of aeration (Table 2), indirectly resulting in less N₂O production in anoxic stages.

4.4 Implications for wastewater treatment application of single-stage PN/A system

As reported in previous studies, the total carbon footprint is most sensitive to its N₂O emission. Approximately 2–90% of the calculated carbon footprint of wastewater treatment plants (WWTPs) in the Netherland and China is related to N₂O emissions.^23,48–51 Some studies reported that autotrophic nitrogen removal system would remain stable performance under salt stress by a long-term acclimation,^11,13,52 but other studies reported that salinity would induce much more N₂O and NO emissions,¹⁹ which may reduce the low carbon advantage of autotrophic nitrogen removal system. From a greenhouse gas emission aspect, high N₂O emission would be a disadvantage for PN/A system treating saline wastewater. Desloover et al. calculated carbon footprint based on their system and proposed that 50% reduction of N₂O emissions is necessary to facilitate the carbon neutral.⁵³ In the present study, both TNRR and nitrogen oxides dynamics of the single-stage PN/A system varied along with different salinities. Low salinity (5, 10 g NaCl per L) enhanced the single-stage PN/A system performance but triggered more N₂O emissions, while relatively high salinity (15, 20 g NaCl per L) decreased the nitrogen oxides emissions with a remarkable deteriorating TNRR. Fortunately, a best balance point can be achieved between TNRR and N₂O emissions, i.e., our findings highlight the possibility of the efficient nitrogen removal performance under salinity up to 15 g NaCl per L with 53% reduced N₂O emission factor by a long-term acclimation (Table 3). These data provide valuable information for carbon neutral assessment of anammox technology applications in ammonia-rich saline wastewater treatment, which also favors for understanding of potential mechanisms of salt-induced N₂O production in single-stage PN/A anammox processes.

5. Conclusions

A lab-scale single-stage PN/A system was developed to investigate the impact of salinity on N₂O emissions from aerobic and anoxic stages.

(1) The overall N₂O emission factor from a lab-scale single-stage PN/A system was 0.43% of the incoming nitrogen load. The majority of N₂O was produced during aerobic stage possibly through hydroxylamine oxidation pathway, occupying over 68% of the total N₂O emission factor under elevated salinities.

(2) Relatively low salinity addition of 5 and 10 g NaCl per L greatly increased N₂O production and emission, while high salinity additions of 15 and 20 g NaCl per L caused a decrease in N₂O production and emission. The impaired functional microbial abundance under high salinity stress (20 g NaCl per L) could lead to instability of single-stage PN/A system.

Single-stage PN/A system process can be applied to treat ammonia-rich saline wastewater through long-term acclimation. The TNRR maintained stable under salinity of 15 g NaCl per L, while The N₂O emission factor decreased by 53% compared with that under 0 g NaCl per L.

Acknowledgements

The acknowledgements come at the end of an article after the conclusions and before the notes and references. This work was supported by the National Natural Science Foundation of China (NSFC) (no. 51522809 and 51378370). The Fundamental Research Funds for the Central University (Tongji University) (0400219238) and the foundation of State Key Laboratory of Pollution Control and Resource Reuse (Tongji University), China (PCRRY 0400231010), are also acknowledged.

References

1. B. Wett, K. Buchauer and C. Fimml, Energy self-sufficiency as a feasible concept for wastewater treatment systems, Singapore, 2007, pp. 21–24.
2. C. Fux and H. Siegrist, Water Sci. Technol., 2004, 50, 19–26.
3. B. Kartal, J. G. Kuenen and M. C. van Loosdrecht, Science, 2010, 328, 702–703.
4. K. A. Third, A. O. Sliekers, J. G. Kuenen and M. Jetten, Syst. Appl. Microbiol., 2001, 24, 588–596.
5. J. M. Wood, J. Gen. Physiol., 2015, 145, 381–388.
6. D. Paredes, P. Kuschk, T. S. A. Mbwette, F. Stange, R. A. Mueller and H. Koeser, Eng. Life Sci., 2007, 7, 13–25.
7. A. Dapena-Mora, J. L. Campos, A. Mosquera-Corral and R. Mendez, Water Sci. Technol., 2006, 53, 265–274.
8. A. Dapena-Mora, J. R. Vázquez-Padín, J. L. Campos, A. Mosquera-Corral, M. S. M. Jetten and R. Méndez, Biochem. Eng. J., 2010, 51, 167–171.
9. R. Jin, C. Ma, Q. Mahmood, G. Yang and P. Zheng, Process Saf. Environ. Prot., 2011, 89, 342–348.
10. B. Kartal, M. Koleva, R. Arsov, W. van der Star, M. S. Jetten and M. Strous, J. Biotechnol., 2006, 126, 546–553.
11. S. Liu, F. Yang, Z. Gong and Z. Su, Appl. Microbiol. Biotechnol., 2008, 80, 339–348.
12. A. Malovanyy, E. Plaza, J. Trela and M. Malovanyy, Environ. Technol., 2015, 36, 595–604.
13. K. Windey, I. De Bo and W. Verstraete, Water Res., 2005, 39, 4512–4520.
14. J. Yang, L. Zhang, D. Hira, Y. Fukuzaki and K. Furukawa, Bioresour. Technol., 2011, 102, 2367–2372.
15. J. H. Ahn, S. Kim, H. Park, B. Rahm, K. Pagilla and K. Chandran, Environ. Sci. Technol., 2010, 44, 4505–4511.
16. B. B. Colliver and T. Stephenson, Biotechnol. Adv., 2000, 18, 219–232.
17. M. J. Kampschreur, H. Temmink, R. Kleerebezem, M. S. M. Jetten and M. C. M. van Loosdrecht, Water Res., 2009, 43, 4093–4103.
18. N. Kishida, J. H. Kim, Y. Kimochi, O. Nishimura, H. Sasaki and R. Sudo, Water Sci. Technol., 2004, 49, 359–365.
19. S. Tsuneda, M. Mikami, Y. Kimochi and A. Hirata, J. Hazard. Mater., 2005, 119, 93–98.
20. T. Stocker, Climate change 2013: the physical science basis: Working Group I contribution to the Fifth assessment report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, UK, New York, 2014.
21. W. C. Trogler, J. Chem. Educ., 1995, 72, 973–976.
22. X. Zhu, M. Burger, T. A. Doane and W. R. Horwath, Proc. Natl. Acad. Sci. U. S. A., 2013, 110, 6328–6333.
23. J. Desloover, S. E. Vlaeminck, P. Clauwaert, W. Verstraete and N. Boon, Curr. Opin. Biotechnol., 2012, 23, 474–482.
24. P. Wunderlin, H. Siegrist and A. Joss, Environ. Sci. Technol., 2013, 47, 9567–9568.
25. M. J. Kampschreur, R. Poldermans, R. Kleerebezem, W. R. van der Star, R. Haarhuis, W. R. Abma, M. S. Jetten, M. S. Jetten and M. C. van Loosdrecht, Water Sci. Technol., 2009, 60, 3211–3217.
26. Y. Law, L. Ye, Y. Pan and Z. Yuan, Philos. Trans. R. Soc., B, 2012, 367, 1265–1277.
27. R. Yu, M. J. Kampschreur, M. C. van Loosdrecht and K. Chandran, Environ. Sci. Technol., 2010, 44, 1313–1319.
28. A. Joss, D. Salzgeber, J. Eugster, R. Koenig, K. Rottermann, S. Burger, P. Fabijan, S. Leumann, J. Mohn and H. Siegrist, Environ. Sci. Technol., 2009, 43, 5301–5306.
29. S. Okabe, M. Oshiki, Y. Takahashi and H. Satoh, Water Res., 2011, 45, 6461–6470.
30. B. Kartal, W. J. Maalcke, N. M. de Almeida, I. Cirpus, J. Gloerich, W. Geerts, H. J. M. O. den Camp, H. R. Harhangi, E. M. Janssen-Megens, K. Francoijs, H. G. Stunnenberg, J. T. Keltjens, M. S. M. Jetten and M. Strous, Nature, 2011, 479, 127–159.
31. A. Uygur, Process Biochem., 2006, 41, 61–66.
32. W. Zhao, Y. Wang, S. Liu, M. Pan, J. Yang and S. Chen, Chem. Eng. J., 2013, 215, 252–260.
33. R. Sander, Compilation of Henry's law constants for inorganic and organic species of potential importance in environmental chemistry, Max-Planck Institute of Chemistry, Air Chemistry Department, Mainz, Germany, 1999.
34. P. H. A. American, W. W. A. American, P. C. F. Water and E. F. Water, Standard methods for the examination of water and wastewater, American Public Health Association, 1998.
35. Y. Cui, C. Peng, Y. Peng and L. Ye, Clean: Soil, Air, Water, 2009, 37, 649–656.
36. C. Ma, R. Jin, G. Yang, J. Yu, B. Xing and Q. Zhang, Bioresour. Technol., 2012, 112, 124–130.
37. D. J. Kim and S. H. Kim, Water Res., 2006, 40, 887–894.
38. M. J. Kampschreur, W. R. L. van der Star, H. A. Wielders, J. W. Mulder, M. S. M. Jetten and M. C. M. van Loosdrecht, Water Res., 2008, 42, 812–826.
39. Q. Yang, X. Liu, C. Peng, S. Wang, H. Sun and Y. Peng, Environ. Sci. Technol., 2009, 43, 9400–9406.
40. C. Domingo-Felez, A. G. Mutlu, M. M. Jensen and B. F. Smets, Environ. Sci. Technol., 2014, 48, 8679–8687.
41. N. Weissenbacher, I. Takacs, S. Murthy, M. Fuerhacker and B. Wett, Water Environ. Res., 2010, 82, 169–175.
42. Y. Law, B. Ni, P. Lant and Z. Yuan, Water Res., 2012, 46, 3409–3419.
43. L. Peng, B. J. Ni, L. Ye and Z. Yuan, Water Res., 2015, 73, 29–36.
44. P. Wunderlin, J. Mohn, A. Joss, L. Emmenegger and H. Siegrist, Water Res., 2012, 46, 1027–1037.
45. M. J. Kampschreur, N. C. G. Tan, R. Kleerebezem, C. Picioreanu, M. S. M. Jetten and M. C. M. Van Loosdrecht, Environ. Sci. Technol., 2008, 42, 429–435.
46. B. Kartal, M. M. M. Kuypers, G. Lavik, J. Schalk, H. J. M. O. den Camp, M. S. M. Jetten and M. Strous, Environ. Microbiol., 2007, 9, 635–642.
47. S. Yoshie, N. Noda, S. Tsuneda, A. Hirata and Y. Inamori, Appl. Environ. Microbiol., 2004, 70, 3152–3157.
48. C. Chai, D. Zhang, Y. Yu, Y. Feng and M. S. Wong, Water, 2015, 7, 918–938.
49. J. Foley, Z. Yuan, J. Keller, et al., N₂O and CH₄ emission from wastewater collection and treatment systems: technical report, 2011.
50. A. Rodriguez-Caballero, I. Aymerich, R. Marques, M. Poch and M. Pijuan, Water Res., 2015, 71, 1–10.
51. M. R. J. Daelman, E. M. van Voorthuizen, L. G. J. M. van Dongen, E. I. P. Volcke and M. C. M. van Loosdrecht, Water Sci. Technol., 2013, 67, 2350–2355.
52. Z. Zhang, S. Chen, P. Wu, L. Lin and H. Luo, Bioresour. Technol., 2010, 101, 6309–6314.
53. J. Desloover, H. De Clippeleir, P. Boeckx, G. Du Laing, J. Colsen, W. Verstraete and S. E. Vlaeminck, Water Res., 2011, 45, 2811–2821.
54. C. M. Castro-Barros, M. R. J. Daelman, K. E. Mampaey, M. C. M. van Loosdrecht and E. I. P. Volcke, Water Res., 2015, 68, 793–803.
55. M. Ali, R. M. L. D. Rathnayake and L. Zhang, et al., Water Res., 2016, 102, 147–157.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra24109c

---