The Influence of Sulfate on Anaerobic Ammonium Oxidation in a Sequencing Batch Reactor

Abstract

Anaerobic ammonia-oxidizing bacteria have a more comprehensive metabolism than expected - there may be other electron acceptors that oxidize ammonium nitrogen under anaerobic conditions, in addition to the well-known nitrite nitrogen, one of which is sulfate in the sulfammox process. Sulfate-containing compounds are part of the medium for the anammox process, but their concentrations are not particularly high (0.2 g MgSO₄ ∙ 7H₂O/dm³ and 0.00625 g FeSO₄/dm³). They can react to some extent with influent ammonium nitrogen. In this work, tests were carried out in two sequencing batch reactors with granular sludge. The first reactor (R1) operated in a 6 h cycle, and the concentration of the inflowing sulfate was kept at 44 mg/dm³∙d. The second reactor (R2) was operated until the 36th day in a 6 h cycle; the influencing concentration was 180 mg SO₄²⁻/dm³∙d from the 37th to 64th day in a 3 h cycle, with an influencing concentration of 360 mg SO₄²⁻/dm³∙d; and from the 65th to 90th day, the reactor was operated again in a 6 h cycle with an influencing concentration of 180 mg SO₄²⁻/dm³∙d. Along with the increased share of sulfate, both the ammonium utilization rate and specific anammox activity showed an increasing trend. As soon as the sulfate dosage was reduced, the ammonium utilization rate and specific anammox activity values dropped. Therefore, it can be concluded that sulfate-containing compounds contribute to the efficiency and rate of the anammox process.

1. Introduction

Several industrial processes such as fermentation, tanning, landfill leachate production, paper production, pharmaceutical production and food processes produce wastewater containing high concentrations of sulfate (SO₄²⁻) and ammonium nitrogen (NH₄-N) [1]. Such sewage requires treatment before discharge to the environment, as it is harmful to human life [2].

SO₄²⁻ is conventionally removed by anaerobic processes by sulfate-reducing bacteria (SRB) [3,4], where SO₄²⁻ is the final electron acceptor and organic carbon is the electron donor [5]. In contrast, the combined nitrification–denitrification processes are the main pathway responsible for the transformation of nitrogen (N) compounds in wastewater treatment systems in which ammonia-oxidizing bacteria (AOB), nitrogen-oxidizing bacteria (NOB) and heterotrophic bacteria are involved. The discovery of the anammox process shed new light on the nitrogen cycle. This biological process involves oxidizing ammonium nitrogen (NH₄-N) under anoxic conditions to gaseous nitrogen (N₂), using nitrite nitrogen (NO₂-N) as the electron acceptor, via anaerobic ammonia-oxidizing bacteria (AAOB). Accordingly, the removal of SO₄²⁻ and NH₄-N generally takes place in separate processes, as each purification step requires different bacterial groups and environmental conditions. This is associated with high costs due to the necessity of aeration, external carbon sources and excess sludge disposal [6]. However, to date, little is known about the ability of AAOB to use SO₄²⁻ as an electron acceptor [6].

Fdz-Polanco et al. [7] described the reaction of the autotrophic anaerobic oxidation of NH₄-N and deoxidation of SO₄²⁻ in three Equations (1)–(3):

3SO₄²⁻ + 4NH₄⁺ → 4NO₂⁻ + 3S²⁻ + 4H₂O + 8H⁺

(1)

3S²⁻ + 2NO₂⁻ + 8H⁺ → N₂ + 3S⁰ + 4H₂O

(2)

2NO₂⁻ + 2NH₄⁺ → 2N₂ + 4H₂O

(3)

At first, NH₄-N is partially oxidized and deoxygenated by SO₄²⁻ to produce NO₂-N and sulfides (S²⁻) (see reaction 1). Then, some of the NO₂-N is reduced by S²⁻ in the sulfur-dependent autotrophic denitrification process and converted into N₂ and elemental sulfur (S⁰) (see reaction 2). Ultimately, the conventional anammox process follows (see reaction 3).

It turns out that AAOB’s metabolism is more comprehensive than expected [8,9] and, in addition to the commonly known electron acceptor in the form of NO₂-N, there may be other electron acceptors that oxidize NH₄-N under anaerobic conditions [10]. The process described in reactions 1–3 is called the sulfammox process (i.e., sulfate-reducing ammonium oxidation (SRAO)) [11]. The sulfammox process is a promising resource for wastewater treatment systems, because wastewater contains high amounts of sulfur compounds [12]. It can be represented in one reaction as follows [13] (4):

SO₄²⁻ + 2NH₄⁺ → S⁰ + N₂ + 4H₂O

(4)

Producing N₂ and elemental sulfur (S⁰) is desirable in wastewater treatment and for the recovery of resources. Moreover, the simultaneous removal of SO₄²⁻ and NH₄-N is more beneficial in terms of reducing costs than the separate removal of these pollutants [14]. The discovery of the sulfammox process suggests that the interrelationships between the N and S biochemical cycles is far more complex than previously assumed.

It is worth noting that the process of sulfur-dependent autotrophic denitrification has been described as a component of sulfammox. It is an autotrophic process in which chemotrophic sulfur-oxidizing bacteria (SOB) oxidize reduced sulfur compounds such as S²⁻, S⁰, sulfite (SO₃²⁻) or thiosulfate (S₂O₃²⁻) as electron donors with NO₃-N or NO₂-N as electron acceptors [15,16,17,18]. Then, SO₄²⁻ or S⁰ is formed depending on the sulfur-to-nitrogen ratio [2]. S²⁻ produced by sulfate-reducing bacteria can also be used as an electron donor for sulfur denitrification [19].

Due to the complex transformations of sulfur and nitrogen in anaerobic conditions, it is worth considering the effect of SO₄²⁻ on anaerobic NH₄-N oxidation. The sulfammox process can run independently without the addition of NO₂-N or in combination with the conventional (NO₂-N based) annamox process. Research on the sulfammox process was carried out in various configurations. At the beginning of the research, SO₄²⁻ was used as an electron acceptor without the addition of NO₂-N [9,11,20,21,22,23]. Other studies started with a conventional anammox, with NO₂-N as an electron acceptor, and replaced NO₂-N with a new SO₄²⁻ electron acceptor [11,12,24]. There are also reports in which SO₄²⁻ was used simultaneously with NO₂-N as an electron acceptor [25,26]. For example, Zhang et al. [25] and Wu et al. [26] showed a high degree of simultaneous removal of NH₄-N and SO₄²⁻, in the range of 92–99% and 53–60%, respectively, when anammox and sulfammox reactions occurred simultaneously. Therefore, the research shows that combining the two processes can achieve an increase in the overall nitrogen removal efficiency.

To date, research work has focused mainly on the effect of increased proportions of NH₄-N and N/S ratio in relation to the sulfammox process [10,20,21]. The influence of increased proportions of SO₄²⁻ on anaerobic NH₄-N oxidation in the presence of NO₂-N due to the reduced cycle time has yet to be described. The purpose of this study is to compare the operation of two sequencing batch reactors (SBR) with granular sludge: one operates under a constant load of SO₄²⁻ and constant duration of the process cycle, and the other operates with an increased and variable load of SO₄²⁻ in a variable cycle time. The process efficiency was compared by calculating the ammonia utilization rate (AUR) and the specific anammox activity (SAA). It is suspected that SO₄²⁻ will increase the AUR and SAA as it will act as an additional electron acceptor in the anaerobic oxidation of NH₄-N.

2. Materials and Methods

2.1. Laboratory-Scale Bioreactor

The inoculated biomass originated from a full-scale side-stream deammonification system in Plettenberg, Germany.

The laboratory scale system used in this study consisted of two 4 dm³ sequencing batch reactors (SBRs) laid out according to the scheme in Figure 1. The system was equipped with a thermostatic jacket maintaining a constant temperature in the range of −35 to + 200 °C, with an accuracy of ± 0.1 °C. Each reactor was equipped with an electric stirrer with variable speed. In the main reactor, probes were placed to measure pH (Endress + Hauser EH CPS 471D-7211, Switzerland) and to measure dissolved oxygen (DO) (Endress + Hauser COS22D-10P3/O, Germany).

All measured data were transmitted to the programmable logic controller (PLC) and used for control and regulation. Measurement data for archival and further use were sent to an application called Intouch’a.

2.2. Operational Conditions of the Laboratory-Scale SBRs

The tests were carried out continuously for 90 days. During the entire test period, the SBRs operated at a constant temperature of 30 (± 1) °C. The pH was maintained in the range of 7.5–7.8 through the automatic addition of 4 M sodium hydroxide (NaOH). The DO concentration in unventilated SBRs did not exceed 0.2 mg/dm³, and SBRs were fed with synthetic substrate according to the method of Dapena-Mora et al. [27] and

Table 1. Number of cycles and concentrations of compounds in R1 and R2.

Reactor	Day	Number of Cycles per Day	Time of One Cycle	SO42− Concentration in the Reactor		NH4-N Concentration in the Reactor		NO2-N Concentration in the Reactor
				per Cycle	per Day	per Cycle	per Day	per Cycle	per Day
			(h)	(mg/dm3)	(mg/(dm3∙d))	(mg/dm3)	(mg/(dm3∙d))	(mg/dm3)	(mg/(dm3∙d))
R1	0–90	4	6	11	44	38	152	50	200
R2	0–36	4	6		180	38	152	50	200
	37–64	8	3	45	360		304		400
	65–90	4	6		180		152		200

.

In each cycle, 2 dm³ of supernatant water was withdrawn from both reactors and replaced with a new portion of the synthetic substrate. The most important ingredients—i.e., nitrite, ammonium and sulfate—were supplied in the form of NH₄Cl, NaNO₂ and MgSO₄, respectively.

2.3. Analytical Methods

The concentration of NO₃-N, NO₂-N and NH₄-N compounds was determined using a DR 3900 spectrophotometer using cuvette tests from Hach Lange GmbH (Dusseldorf, Germany) for analysis. The biomass concentrations were determined as a volatile suspended solids (VSS) fraction of the total suspended solids (TSS) in accordance with the standard methods [28]. The biomass-specific AUR, SAA and nitrate production rate (NPR) were determined based on the maximum slope of NH₄-N consumption, NH₄-N combined with NO₂-N consumption and NO₃-N production in the reaction phase divided into mixed liquor volatile suspended solids (MLVSS) concentrations, respectively. Throughout the operation period, the MLVSS value was 1750 (±50) mg/dm³ in R1 and 1900 (±50) mg/dm³ in R2. AUR, SAA and NPR are given in units of mg N/g VSS∙h to represent these rates in relation to the indicated MLVSS.

3. Results and Discussion

The efficiency of NH₄-N oxidation in anaerobic conditions is influenced by anammox, sulfammox, heterotrophic and autotrophic (full and partial) denitrification processes. On the other hand, under aerobic conditions, the oxidation of NH₄-N takes place in the process of nitrification or partial nitrification. In our studies, SBR controlled DO at a low level (<0.2 mg/dm³), and the lack of an added external carbon source prevented the occurrence of heterotrophic conditions. Accordingly, the only possible pathways for NH₄-N oxidation were through anammox, sulfammox and sulfur-dependent autotrophic denitrification.

Previous studies describe the complete efficiency of NH₄-N and SO₄²⁻ removal as a combination of anammox, sulfammox, nitrification and denitrification [10,11,20,26] or a result of anaerobic processes only [21,29] or of the sulfammox reaction only [30,31] (see

Table 2. Concentrations of influent NH₄-N and SO₄²⁻ and the efficiency of their removal under anaerobic conditions. SRAO: sulfate-reducing ammonium oxidation; SRB: sulfate-reducing bacteria.

Reactor	Influent NH4-N (mg/dm3)	Influent SO42− (mg/dm3)	NH4-N Removal Efficiency (%)	SO42− Removal Efficiency (%)	Brief Characteristics	Reference
Combining system: Upflow Anaerobic Sludge Blanket (UASB), Anoxic/Oxic Reactor (A/O), Anammox and Sulfammox Reactor (ANAOR), Anaerobic Sequencing Batch Reactor (ASBR)	610–700	1870–1920	ca. 98	ca. 53	Reduction of SO42− and NH4-N was considered as a combination of anammox, sulfammox, nitrification and denitrification processes.	[26]
Continuous Flow Stirred Tank Reactor (CFSTR)	110	0-110	ca. 40	ca. 0	SRAO was considered as a combination of aerobic ammonium oxidation, anammox and heterotrophic sulfate reduction processes.	[11]
	60	90	ca. 30	ca. 10
	60	90	ca. 55	ca. 0
Self-Designed Circulating Flowreactor (SDCF)	120	183	ca. 30	ca. 40	These results showed that nitrogen was converted by nitrification, denitrification and conventional anammox, simultaneously with SRAO. The sulfur-based autotrophic denitrification and denitrification in the reactor were caused by the influent NO2-N.	[10]
	160	216	ca. 55	ca. 0
	110	116	ca. 75	ca. 30
	80	100	ca. 100	ca. 45
Self-Designed Circulating Flowreactor (SDCF)	50	90	ca. 40	ca. 30	Part of nitrogen was converted by nitrification–denitrification and conventional anammox, simultaneously with SRAO.	[20]
	120	170	ca. 90	ca. 30
	180	360	ca. 20	ca. 5
Expanded Granular Sludge Bed (EGSB)	166–666	3600	40–58	64–71	SRB and denitrifying bacteria were mainly responsible for SO42− and nitrogen removal.	[21]
	1000–2000		40–70	66–82
	>3000		10–25	28
Anaerobic Sequencing Batch Reactor (ASBR)	97	261	ca. 88	ca. 19	The presence of Planctomycetes revealed that anammox was highly involved in NH4-N removal, even without NO2-N in the feed. Other autotrophic denitrifying bacteria, related to the species Paracoccus Denitrificans, were also present. These bacteria utilize S0 as an electron donor, produce SO42− and competitively use NO2-N with anammox.	[29]
Expanded Bed Reactor (EBR)	229	163	ca. 44	40	The reduction of SO42− and NH4-N was considered as sulfammox only.	[30]
Upflow Anaerobic Sludge Blanket Reactor (UASBR)	50-60	210-240	40	30	The reduction of SO42− and NH4-N was considered as sulfammox only.	[31]
Non-Woven Rotating Biological Contactor (NWRBC)	ca. 198	ca. 528	ca. 100	ca. 70	The reduction of SO42− and NH4+ was considered as a sulfammox only.	[9]
Anaerobic Attached-Growth Bioreactor (AAGB)	50	57	ca. 43	ca. 59	The reduction of SO42− and NH4+ was considered as a sulfammox only.	[23]

). Moreover, it is worth noting that a few studies on the anaerobic oxidation of NH₄-N in the presence of SO₄²⁻ have been carried out with NO₂-N [10,26]. Some of them consisted of only replacing NO₂-N with a new electron acceptor in the form of SO₄²⁻ [11,31], yet the vast majority of the oxidation took place without NO₂-N [9,11,20,21,23,29,30].

A study by Zhang et al. [10] investigated the effect of NO₂-N on the anaerobic oxidation of NH₄-N. They showed that, with a combined decrease in concentration of SO₄²⁻ from 216 to 100 mg/dm³, NH₄-N from 183 to 80 mg/dm³ and NO₂-N from 34 to 28 mg/dm³, the efficiency of NH₄-N removal increased from 55% to 100%. However, this study does not clearly show the influence of SO₄²⁻ itself on the process. In our study, we decided to keep the NH₄-N and NO₂-N inflow to the reactors unchanged in order to determine the influence of SO₄²⁻ on the process.

In R1, where the influent SO₄²⁻ concentration was constant at 22 mg SO₄²⁻/dm³, a gradual increase in the rates of AUR and SAA could be observed as well as their stabilization from day 49, as shown in Figure 2a. Comparing these values with the values in R2 in Figure 2b, it can be seen that, despite the approximately four-fold higher SO₄²⁻ concentration in the effluents in R2 (90 mg SO₄²⁻/dm³ for R2), the AUR and SAA showed similar values from the beginning of the test to day 29. The AUR increased from 1.3 mg N/g VSS∙h to 2.1 mg N/g VSS∙h (R1) and from 1.1 mg N/g VSS∙h to 2.1 mg N/g VSS∙h (R2), while the SAA increased from 4 mg N/g VSS∙h to 5.6 mg N/g VSS∙h (R1) and 3.7 mg N/g VSS∙h to 5.3 mg N/g VSS∙h (R2).

On day 37, there was a clear increase in AUR, SAA and NPR in R1. This showed that the efficiency of the anammox process was greatly improved as more NH₄-N was oxidized with NO₂-N. The increase in NPR also confirmed that more NH₄-N was oxidized as approximately 11% was converted to NO₃-N in this process.

Near the end of the study, there was a stabilization of AUR values, SAA and a decrease in NPR in R1. AUR increased to a maximum of 4.4 mg N/g VSS∙h, and SAA increased to 8.1 mg N/g VSS∙h.

In R2, on day 37, the cycle time was reduced from 6 h to 3 h, which resulted in the concentration of SO₄²⁻ being twice as high as in the previous period: −360 mg/dm³∙d and 180 mg/dm³∙d for phases II and I, respectively. This affected the AUR and SAA significantly, as can be seen in Figure 2b. This increase was evident throughout phase II. The AUR value at the end of this phase was 9.7 mg N/g VSS∙h, while SAA was 22.5 mg N/g VSS∙h. This confirmed the positive influence of SO₄²⁻ on the course of the NH₄-N oxidation process. SO₄²⁻ seems to be an additional acceptor that improves the rate and efficiency of the process, increasing the efficiency of NH₄-N removal as shown in Figure 3.

There are reports in the literature confirming that SO₄²⁻ can increase the total removal of NH₄-N. Liu et al. [9] noted in his research that the NH₄-N removal rate was always higher than expected and the NH₄-N/NO₂-N consumption ratio was about 1.1:1, which was much higher than previously reported [32]. It was then concluded that, due to large amounts of (NH₄)₂SO₄ in the feed, SO₄²⁻ could be the source of the additional electron acceptor.

Moreover, Yang et al. [31] noted that as the concentration of NH₄-N and SO₄²⁻ increased, incrementally more of both were removed in their batch tests. When the NH₄-N and SO₄²⁻ concentrations in the inflow were approximately 28 and 76 mg/dm³, respectively, the removal efficiency was close to 0%. However, when the average NH₄-N and SO₄²⁻ concentrations in the inflow increased to 92 and 307 mg/dm³, the removed amount decreased to 40 and 130 mg/dm³, respectively. Thus, high concentrations of NH₄-N and SO₄²⁻ may promote the simultaneous removal of these compounds, as shown in our research.

Phase III in R2 showed a downward trend in AUR and anammox rates from 9.7 mg N/g VSS∙h to 7.1 mg N/g VSS∙h and from 22.5 mg N/g VSS∙h to 18.7 mg N/g VSS∙h, respectively. This was due to the reduction of the SO₄²⁻ concentration flowing into the reactor. Again, fewer electron acceptors, in the form of SO₄²⁻, were present in the environment; therefore, the rate of NH₄-N oxidation decreased because half as much SO₄²⁻ flowed in per day. The tests were performed until the process stabilized, and constant values of AUR, SAA and NPR were achieved by the 90th day.

Moreover, Zhang et al. [20] noticed that, as the concentration of SO₄²⁻ increased from about 90 mg/dm³ to about 170 mg/dm³ and NH₄-N from about 50 mg/dm³ to about 120 mg/dm³, the efficiency of NH₄-N removal increased from 40% to 90%. However, a further increase in the concentration of SO₄²⁻ to about 360 mg/dm³ and NH₄-N to about 180 mg/dm³ resulted in a decrease in NH₄-N removal up to roughly 20%. Similarly, in an Expanded Granular Sludge Bed Reactor (EGSBR) [21] under chemical oxygen demand (COD) conditions, the NH₄-N removal efficiency gradually improved from 40–58% to 40–70% when the inflow NH₄-N concentrations increased from 166–666 mg N/dm³ to 1000–2000 mg N/dm³. Comparatively, after increasing the NH₄-N concentration to >3000 mg N/dm³, the efficiency of NH₄-N reduction decreased to approximately 10–25%. This was due to the inhibition of the anammox process with free ammonia. This proves that an increase in NH₄-N and SO₄²⁻ concentrations improves the process of anaerobic NH₄-N oxidation only to a certain extent. In our study, there was no inhibition of the process due to excessively high concentrations of these compounds.

Wu et al. [26] noted that they had achieved an NH₄-N removal efficiency of 98%, including 44% removed through sulfammox. Compounds containing SO₄²⁻ can therefore effectively improve the efficiency of the anaerobic oxidation of NH₄-N, but at the same time, anaerobic conditions favor the decomposition of SO₄²⁻ to S⁰, which is less toxic to the environment. The sulfammox process has so far been studied mainly as an independent process (without NO₂-N addition). Moreover, there has been more interest in the influence of NH₄-N concentration on the sulfammox process [21] and the N/S ratio [20] rather than directly considering the effect of SO₄²⁻ itself.

Bi et al. [11] challenged the sulfammox process and postulated that AAOBs did not have the ability to oxidize NH₄-N using SO₄²⁻ as an electron acceptor and that SRAO was a combination of aerobic ammonium oxidation, anammox and heterotrophic sulfate reduction processes. Moreover, the specification of the efficiency of NH₄-N and SO₄²⁻ removal in the sulfammox process does not reflect the course of the process as thoroughly as the AUR and the SAA, which the authors do not provide in their research.

4. Conclusions

In this study, it was shown that SO₄²⁻ could be used as an additional electron acceptor in the anaerobic oxidation of NH₄-N. Along with the increased share of SO₄²⁻, both AUR and SAA showed an increasing trend. In R1, where the concentration of SO₄²⁻ in the inflow was constant at the level of 22 mg SO₄²⁻/dm³, there was a gradual increase in the AUR and SAA indicators from 1.2 mg N/g VSS∙h to 4.4 mg N/g VSS∙h and from 3.9 mg N/g VSS∙h to 8.2 mg N/g VSS∙h, respectively. In R2 in phase I, over a 6 h cycle, AUR and SAA increased from 1 mg N/g VSS∙h to 2.1 mg N/g VSS∙h and from 3.6 mg N/g VSS∙h to 5.3 mg N/g VSS∙h; in phase II, over a 3 h cycle, they increased to 9.7 mg N/g VSS∙h and 22.5 mg N/g VSS∙h; and in phase III, over a 3 h cycle, they dropped to 7.1 mg N/g VSS∙h and 18.8 mg N/g VSS∙h, respectively. It can therefore be concluded that SO₄²⁻ contributes to the rate and efficiency of the anammox process. Further studies on the influence of the NH₄-N/SO₄²⁻ ratio on the process and identification of the bacteria responsible for sulfammox are suggested.