A novel strategy for a stable low-ammonia nitrification system: low-temperature sludge storage recovery

Abstract

Stable partial nitritation is difficult to sustain in low-ammonia wastewater, as nitrite-oxidizing bacteria (NOB) outcompete ammonia-oxidizing bacteria (AOB) for limited substrate and thereby reduce nitrite accumulation. This study adopted intermittent aeration combined with hydrazine dosing and compared two sludge storage approaches to rapidly start partial nitritation at an influent ammonium nitrogen (NH₄⁺–N) concentration of 40 mg L⁻¹. A single UASB reactor operated with dissolved oxygen (DO) < 0.7 mg L⁻¹, intermittent aeration, and hydrazine addition achieved stable performance in 24 days, with 60% ammonium conversion and 77% nitrite accumulation rate (NAR). Further reducing DO to 0.6 mg L⁻¹ cut aeration energy, stabilizing ammonium conversion at 50%, while NAR approached 100%. The measured AOB and NOB activities were 0.42 ± 0.01 and 0.1 ± 0.01 mg N (g VSS h)⁻¹, respectively. Nitrifying sludge was preserved for 5 months at 4 °C or at ambient winter temperatures (−25 to 12 °C), then seeded into reactors R1 and R2. R1 with refrigerated sludge stabilized within 15 days, maintaining 57% ± 4% ammonium conversion and 80% NAR, with far better performance than R2. PICRUSt2 functional prediction confirmed that 4 °C sludge storage boosted AOB enrichment and upregulated ammonia monooxygenase (AMO) expression. The integrated hydrazine regulation and low-temperature sludge storage strategy enables fast, stable partial nitritation under low ammonium loading, providing technical support for mainstream PN/A applications.

The integrated hydrazine regulation and low-temperature sludge storage strategy enables fast, stable partial nitritation under low ammonium loading, providing technical support for mainstream PN/A applications.

1. Introduction

Nitrogen pollution in water bodies has become more serious with the rapid development of society and economy.¹ Wastewater solutions with high treatment efficiency and cost-savings are crucial to protecting the environment and recycling wastewater. In practical applications, the anaerobic ammonium oxidation (anammox) process is often coupled with partial nitritation. The partial nitritation-anammox (PN/A) process is a new autotrophic denitrification process with energy-saving characteristics and can promote sustainable development.² In the PN/A process, nitrite-oxidizing bacteria (NOB) are not effectively inhibited, converting nitrite nitrogen (NO₂⁻–N) to nitrate nitrogen (NO₃⁻–N), destabilizing the partial nitritation process and depriving the subsequent anammox process of electron acceptors. Excessive NOB accumulation disrupts nitrification, reducing PN/A efficiency or leading to system collapse.³ The PN/A process has been successfully applied to high NH₄⁺–N wastewater.⁴ However, there are still bottlenecks in treating low NH₄⁺–N (40 mg L⁻¹) wastewater,⁵ and the free ammonia (FA) concentration cannot reach a level sufficient to inhibit NOB effectively. Long-term low substrate stress severely restricts the substrate affinity and metabolic activity of AOB.⁶ Ammonia-oxidizing bacteria (AOB) are autotrophic bacteria with limited scope for improving activity under low substrate conditions.^7,8 NOB activity is difficult to inhibit in this environment, making the partial nitritation process extremely unstable. Methods for inhibiting NOB are divided into in situ and ectopic treatments, including oxygen limitation, chemical inhibitors, and sludge acclimation.^9,10 Ectopic sludge conditioning has been proven to sustain high nitrite accumulation under low ammonium loading by selectively suppressing NOB proliferation.¹¹ Recent studies have further optimized in situ control strategies such as intermittent aeration and FNA sludge treatment, but their long-term stability in low-ammonium municipal wastewater still needs further verification.^12,13 Hydrazine (N₂H₄) is an intermediate metabolite of the anammox process, which can inhibit NOB activity while having little impact on AOB activity (avoiding a significant accumulation of NO₃⁻–N in the effluent).¹⁴ Recent studies also verified its selective NOB inhibition effect under low substrate conditions, while its combined application with sludge storage strategies has been rarely reported.¹⁵ Many challenges in practical wastewater treatment and large-scale applications remain to be solved.

In practical applications or small-scale experiments, sludge storage will be required for various reasons. Especially in the context of the normalization of global epidemics, more reaction facilities will be idle. For example, during the COVID-19 epidemic, equipment was out of service. Given similar crisis awareness and emergency considerations, improving treatment outcomes through sludge storage is necessary. There is a sludge production process during the operation of the water plant, and some sludge must be stored as a backup, which gives this study specific engineering significance. A few sources of granular sludge are not easy to domesticate. Good storage and recovery conditions must be explored if these sludges are used as seed sludge for expanded culture. After long-term storage, the number of dominant bacteria in nitrifying sludge declines.¹⁶ Anaerobic inner-core hydrolysis and overgrowth by filamentous bacteria occur, destroying the sludge structure and function. These phenomena are closely related to storage time and ambient temperature.^17,18 Nitrification sludge stored at 4 °C maintains a more stable structure and function than sludge stored under ambient winter conditions, and is easier to reactivate.^19,20 Latest research further confirms that low-temperature storage slows down microbial endogenous respiration and better preserves AOB viability by reducing substrate consumption and cell autolysis.^21,22 There are many difficulties in high-performance nitrifying sludge culture due to the long generation time of AOB, weak self-aggregation ability, and sensitivity to environmental changes.^23,24 Testing with stored sludge enables rapid startup of the reaction zone.^25,26 Nitrifying sludge activity recovery and post-recovery performance are affected by various factors, such as storage conditions, time, and recovery operation conditions, and vary greatly.^18,27,28

Low-temperature storage is more beneficial for sludge recovery because the endogenous respiration rate of low-temperature microorganisms is lower, thereby preventing sludge disintegration. The recovery of nitrification and nitrosification sludge is relatively easy.²⁹ Considering that in the partial nitritation process, the ecological basis of a large number of NOB inhibition strategies is the better AOB adaptability in adverse environments, then initiating partial nitritation under low NH₄⁺–N after low-temperature storage will more effectively inhibit NOB activity (stimulating AOB activity recovery).³⁰ Therefore, it is a more reasonable idea to store nitrification sludge from a wide range of sources at low temperatures to achieve directional inhibition of NOB at low substrate concentrations.

To address the bottleneck of difficult NOB inhibition and unstable partial nitritation in low-ammonium wastewater, this work systematically investigated the regulatory effects of hydrazine addition combined with a sludge storage strategy and revealed the underlying microbial mechanism. The specific research objectives were as follows: (i) clarify the startup efficiency and long-term stability of the low-ammonium partial nitritation system under intermittent aeration alone and hydrazine-coupled aeration control, and analyze the succession characteristics of functional bacteria and microbial communities during operation; (ii) compare the differences in sludge activity recovery, system startup performance and microbial community structure between sludge stored under 4 °C refrigerated conditions and sludge stored under ambient winter conditions (−25 to 12 °C), and screen the optimal sludge preservation mode; (iii) investigate the effect of gradual temperature reduction from 35 °C to 20 °C on nitrogen removal performance, functional bacteria activity and functional gene expression of the recovered system, and verify the long-term application stability of the strategy.

2. Materials and methods

2.1. Experimental setup and inoculated sludge

Three UASB reactors of the same size are used, as shown in Fig. 1. One reactor was applied for the conventional control test (intermittent aeration with and without N₂H₄ dosing), and the other two, designated R1 and R2, were used for sludge storage and recovery tests inoculated with differently preserved sludge. (Total volume 8.3 L, reaction zone 80 cm high, and inner diameter 7 cm). The reactor was a transparent glass cylinder with an effective volume of 5 L. The outer wall of the reactor was wrapped with black, sticky insulation material on one side to meet the culture requirements for light protection and heat preservation for microorganisms. A three-phase separator was designed in the sedimentation area.

Fig. 1. Schematic flow diagram of the experimental setup.

The outer wall of the reactor was covered with a water bath layer (thickness 2 cm, total volume 0.9 L). The inlet and outlet of the water bath layer are connected to the water bath tank, which is equipped with a temperature control system.

Hot water was extracted from the bottom of the water bath layer via a peristaltic pump and discharged from the top into the water bath tank, forming a cycle to maintain the water temperature. A peristaltic pump controlled the inlet water flow rate. There are three sampling ports arranged at equal distances along the side wall of the reactor to facilitate obtaining a uniformly mixed mud–water mixture from each port. Samples from these 3 locations were pooled for activity assays or as DNA samples and stored at −25 °C immediately after collection and then shipped on dry ice (−80 °C) for subsequent high-throughput sequencing analysis. The inoculated sludge was taken from the nitrification sludge from the Sanbaotun Wastewater Treatment Plant in Fushun City, China. After the sludge was retrieved, it was left to stand for 72 hours, and then the supernatant was removed initial mixed liquor suspended solids (MLSS) = 18.2 g L⁻¹, mixed liquor volatile suspended solids (MLVSS = 6.4 g L⁻¹). The sludge was divided into three parts; one part was poured into the reactor (one-third of the reactor volume). MLSS was 5.45 g L⁻¹, and MLVSS was 0.96 g L⁻¹ after the sludge was injected into the reactor. The other two portions of 3 L each were stored in accordance with the experimental design requirements. The incoming water to the test device was simulated inorganic low NH₄⁺–N wastewater. All reagents were dissolved in deionized water. A concentrated trace element stock solution was prepared separately according to Table S1, and 1 mL of the stock solution was added to each liter of synthetic wastewater. The final influent pH was adjusted to 7.5–8.0 with 1 mol L⁻¹ HCl or NaOH before use. Ammonium sulfate was used as the NH₄⁺–N source to prepare a low NH₄⁺–N concentration of 40 mg L⁻¹, and sodium bicarbonate provided the inorganic carbon source (see SI material).

2.2. Impact of conventional strategies on the start-up and stable operation at low NH₄⁺–N nitrification

During phase I, the intermittent aeration strategy (aeration for 5 min, stopping for 10 min) was used to control low dissolved oxygen mass concentration, inhibiting the conversion of NO₂⁻–N to NO₃⁻–N and promoting NO₂⁻–N accumulation. Phase II started when the effluent NO₂⁻–N and NO₃⁻–N concentrations remained stable for a long time (do not show a continuous upward trend) (see SI material). Phase II has begun to increase the NO₂⁻–N and NO₃⁻–N concentrations in the effluent water by adding N₂H₄. N₂H₄ dosage of 5 mg L⁻¹ was added to the incoming water in phase II (intermittent aeration: aeration 5 min, stopping 10 min).³¹ Specific measures were taken to ensure the stability of N₂H₄ concentration (N₂H₄ easily volatilized). It was found that the effluent NO₃⁻–N concentration in the water continued to increase immediately. An appropriate amount of N₂H₄ was added to increase the N₂H₄ concentration in the incoming water to 5 mg L⁻¹, and then no more N₂H₄ was added. This was repeated until the effluent NO₃⁻–N concentration continued to increase. During phase III, the dissolved oxygen concentration was gradually reduced to 0.6 mg L⁻¹ to enable the low NH₄⁺–N nitrification system to operate stably in a low-energy consumption mode and minimize aeration energy consumption.

2.3. New strategy to promote stable operation test at low NH₄⁺–N nitrification

Two 3 L portions of the sludge were collected, as described in Section 2.1. Two aliquots of sludge were stored separately in the dark for 5 months: one at 4 °C under refrigerated conditions, and the other at ambient winter conditions (−25 to 12 °C). The ambient winter storage was conducted in an unheated room in Shenyang, China. The temperature changes within 5 months were relatively stable (see SI material), and there was no repeated freezing and thawing of sludge due to significant heating or cooling. The storage conditions did not change. One aliquot was stored in the dark at 4 °C; the other was stored in the dark at ambient winter conditions for 5 months.

1.5 L of each sludge was stored for 5 months under two temperature conditions (decomposition during the sludge storage process, resulting in sludge volume reduction), put into two identical UASB reactors (R1, R2), and operated under the same control conditions. R1 was inoculated with nitrifying sludge after 5 months of 4 °C refrigerated storage, and R2 was inoculated with sludge after 5 months of ambient winter storage (−25 to 12 °C). Both stages I and II adopted the intermittent aeration strategy in Section 2.2 to control dissolved oxygen and operated stably in a low NH₄⁺–N environment (NH₄⁺–N 40 mg L⁻¹). The impact of sludge stored under different temperature conditions on the start-up and stable operation of the low NH₄⁺–N nitrification system was explored. Based on the start-up time and stable operating conditions, two other sludge storage forms in R1 and R2 were compared and selected. Sludge samples from the two reactors at different operating stages were collected for high-throughput sequencing to predict changes in functional gene abundance. Microbial community succession at each phase and the abundance changes of functional bacterial groups at the genetic level were analyzed.

2.4. Influence of room temperature on the stable operation of a low NH₄⁺–N nitrification system

After the end of the test, phase II is mentioned in Section 2.3. The temperature was gradually cooled from 35 °C to normal temperature (20 °C) in steps of 5 °C during stable system operation, and changes in various parameters were recorded (see SI material). The N₂H₄ addition strategy was adopted when the low NH₄⁺–N partial nitritation system became unstable and continued to cool down after returning to stability. When it was found that the effluent NO₃⁻–N concentration continued to increase and the nitrite accumulation rate (NAR) continued to decrease, an appropriate amount of N₂H₄ (inlet water N₂H₄ = 5 mg L⁻¹) was immediately added, according to the operation method described in Section 2.3, to restore the stability to the low NH₄⁺–N nitrification system.

2.5. Analytical methods

NH₄⁺–N (Nessler's reagent spectrophotometry), NO₂⁻–N (N-(1-naphthyl)ethylenediamine spectrophotometry), NO₃⁻–N (UV spectrophotometry), MLSS (after filtration, dry at 105 °C to constant weight) and MLVSS (after drying, burn at 600 °C to constant weight) were measured according to standard methods.³²

NH₄⁺–N, NO₃⁻–N, NO₂⁻–N, NAR, pH (electrode method: WTW Multi 3430), temperature and DO are measured daily. All nitrogen indicators were determined using standard spectrophotometric methods, with detection wavelengths set at 420 nm for NH₄⁺–N, 540 nm for NO₂⁻–N, and 220 nm/275 nm for NO₃⁻–N. pH and dissolved oxygen were detected synchronously using a WTW Multi 3430 portable multi-parameter meter. The DO electrode was calibrated with air-saturated deionized water before daily sampling. The probe was inserted into the reactor sampling port and kept stable for 30 s to record steady DO readings. Microbial morphology was observed and analyzed by scanning electron microscopy (SEM; Hitachi Regulus 8100, Hitachi, Japan). EPS was extracted by heat treatment, and the contents of bound polysaccharide (B-PS) and bound protein (B-PN) were quantified by the phenol-sulfuric acid method and the Lowry method, respectively. Sludge samples were collected at key stages (day 55 and day 93 of the initial operation stage, day 20 and day 40 of the recovery stage) to test AOB and NOB activities [mg N (g VSS h)⁻¹] and analyze microbial community succession via high-throughput sequencing.³³ For daily water quality monitoring, mixed samples were collected in equal proportions from three sampling ports along the reactor height to minimize sampling bias, with one measurement per day. For batch tests including AOB/NOB activity and EPS quantification, three biological replicates were used per group, and all results were expressed as mean ± standard deviation. An independent-samples t-test was used for pairwise comparisons of batch experimental data, with p < 0.05 considered statistically significant. All statistical analyses were conducted using SPSS 26.0.

2.6. Molecular biology analysis methods and data analysis

2.6.1. Miseq high-throughput sequencing

The sludge samples were transported to Meji Biotechnology Co., Ltd, China, at a low temperature (−80 °C) for high-throughput sequencing. The V4 region universal primer 515F/806R was used to PCR amplify DNA, and the amplified DNA fragments were connected to sequencing adapters to construct a DNA sequencing library.³⁴ DNA sequencing was performed using the Illumina MiSeq platform and the MiSeq Reagent Kit v3.³⁵ Meji Biotech in China completed all sequencing work. Raw sequencing reads were first quality-filtered to remove low-quality sequences and chimaeras, then clustered into operational taxonomic units (OTUs) at 97% sequence similarity. The raw sequencing data were used to conduct a series of analyzes, including analyses of microbial community succession and changes in functional bacterial abundance. Samples are collected at the end of each phase during the reactor operation (Section 3.2). A0 denotes the initial inoculated sludge before storage (raw seed sludge). The corresponding sample names for R1 at each phase of operation are B1 (sludge sample after 5 months' storage at 4 °C), B2 (phase I sample), and B3 (phase II sample). The sample names corresponding to each phase of R2 operation are C1 (sludge sample after 5 months of ambient winter storage in autumn and winter), C2 (phase I sample), and C3 (phase II sample).

2.6.2. Functional gene prediction analysis

Absolute abundance at the microbial order level was screened and analyzed. Cytoscape (version 3.6.1) was used to construct a bipartite network linking classification operation units (OTUs) to samples. Different colors distinguish samples at various stages in the network diagram. Purple nodes represent OTUs unique to a single sample or shared by multiple samples. An OTU abundance matrix is classified by order level. OTU encodings of samples and species are used as node and target sources when building the network. The properties of the edges were represented as weights in the bipartite network diagram to show the absolute abundance of each species. Cloud platform (https://cloud.majorbio.com/) in combination with the PICRUSt 2 tool to perform predictive analysis on relevant functional pathways and functional genes. Relative abundance changes of bacterial metabolic pathways and main functional genes can be effectively inferred. The prediction of bacterial metabolic pathways was based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) database and assigned to the secondary level. The main functional changes are shifts in the relative abundance of AOB and NOB. Functional bacteria such as AOB and NOB more intuitively describe the operation of nitrite.

2.7. Determination method of AOB and NOB activity

The experiment was conducted in a 150 mL Erlenmeyer flask to determine the activity of AOB and NOB. A water solution containing 40 mg L⁻¹ of NH₄⁺–N and NO₂⁻–N was prepared. HEPES was used as the pH buffer, and the pH value was adjusted to 7.8 ± 0.2 (1 mol L⁻¹ HCl and 1 mol L⁻¹ NaOH) (see SI material). After washing the sludge with PBS buffer, 5 g (wet weight) was added to each Erlenmeyer flask. Then, the prepared water was added to a capacity of 100 mL without the bottle mouth being sealed. The Erlenmeyer flask was placed on a magnetic stirrer, and the DO concentration was maintained at 2–3 mg L⁻¹via micro aeration. Continuous cultivation was carried out for 24 h at 28 °C and 150 rpm. Two milliliters of supernatant were collected from the bottle every hour to measure NH₄⁺–N, NO₂⁻–N, and NO₃⁻–N. For every experiment group, three parallel samples were set up. Following the culture, the constant weight method was used to measure each reaction system's VSS, and the activities of AOB and NOB [mg N (g VSS h)⁻¹] were calculated based on the conversion of substrates (NH₄⁺–N and NO₂⁻–N).

3. Results and discussion

3.1. Start-up and stable operation of low NH₄⁺–N nitrification system under conventional strategy

3.1.1. Operational efficiency of low NH₄⁺–N nitrification system under conventional strategy

This section corresponds to research objective (i), focusing on the operation performance and microbial community evolution of the partial nitritation system under two conventional control strategies. Diachronic changes in nitrogen concentration in each stage: During phase I (1–32 days) (Fig. 2A–D), it was found that it is difficult to quickly start the low NH₄⁺–N nitrification system simply by using the low DO intermittent aeration strategy to achieve efficient and stable operation of the low NH₄⁺–N nitrification system.³⁶ NH₄⁺–N conversion rate showed an overall downward trend, decreasing to 31.06% on the 32nd day. NAR showed a gradual upward trend, rising to 35.47% by the 32nd day. The aeration volume during the early phase (1–7 days) was higher to restore the sludge activity, resulting in a higher DO concentration in the system, and the NH₄⁺–N conversion rate in the effluent was close to 100%. On the 8th day, DO was reduced to 1.37 mg L⁻¹. Effluent NH₄⁺–N, NO₂⁻–N and NO₃⁻–N concentrations, corresponding NH₄⁺–N conversion rate and NAR were 18.54 mg L⁻¹, 1.38 mg L⁻¹, 12.94 mg L⁻¹, 53.81% and 11.91%, respectively. When the DO concentration was reduced to 0.5 mg L⁻¹, the effluent NO₃⁻–N concentration gradually decreased, the effluent NH₄⁺–N and NO₂⁻–N concentrations gradually increased, the NH₄⁺–N conversion rate decreased, and the NAR increased as the test progressed. On the 32nd day, effluent NH₄⁺–N, NO₂⁻–N and NO₃⁻–N concentrations, corresponding NH₄⁺–N conversion rate and NAR were 28.41 mg L⁻¹, 3.77 mg L⁻¹ and 9.72 mg L⁻¹, 31.06% and 35.47%, respectively.

Fig. 2. Conventional strategies: changes in (A) NH₄⁺–N concentration and NH₄⁺–N conversion in incoming and outgoing water, (B) NO₂⁻–N concentration and nitrite accumulation rate, (C) NO₃⁻–N concentration, (D) dissolved oxygen, (E) pH changes; (F) changes inAOB and NOB activity; and microbial community structure of low NH₄⁺–N nitrification system (G) relative abundance at phylum level, (H) relative abundance at genus level.

To meet the requirements of the PN/A process, it was necessary to ensure that the NH₄⁺–N conversion rate was around 50%. Although reducing DO can increase NAR in low NH₄⁺–N environments, this can cause the NH₄⁺–N conversion rate fall short of the PN/A process requirements, and the resulting increased NAR is unsatisfactory. Through phase II (33–65 days), it was found that N₂H₄ addition in the intermittent aeration mode can quickly start the low NH₄⁺–N nitrification system and achieve stable operation. NH₄⁺–N conversion rate fluctuated widely at this stage, and the NAR increased rapidly at first, then stabilized around 80% on the 57th day. Since the NH₄⁺–N conversion rate in the late phase I was too low to meet the PN/A process requirements (NH₄⁺–N≥50%), N₂H₄ addition inhibited the AOB activity, thereby reducing the NH₄⁺–N conversion rate. The aeration volume (DO 1.3 mg L⁻¹) increased, thereby increasing the NH₄⁺–N conversion rate. During this operation, NH₄⁺–N conversion rate increased significantly from 33.62% to 58.71%. Effluent NH₄⁺–N, NO₂⁻–N and NO₃⁻–N concentrations, corresponding NH₄⁺–N conversion rate and NAR (stabilized) were 34.01 mg L⁻¹, 4.23 mg L⁻¹, 8.82 mg L⁻¹, 29.36% and 41.98%, respectively. The NH₄⁺–N conversion rate dropped significantly to about 30% on the 56th day; the addition of N₂H₄ inhibited AOB activity, thereby decreasing NH₄⁺–N conversion efficiency.

Therefore, the aeration volume increased (DO 1.3 to 2 mg L⁻¹). NH₄⁺–N conversion rate increases again and stabilizes at 60%. This is also the reason that NH₄⁺–N conversion rate fluctuated greatly in phase II. All indicators of the low NH₄⁺–N nitrification system were stable on the 57th day. Effluent NH₄⁺–N, NO₂⁻–N and NO₃⁻–N concentrations, corresponding NH₄⁺–N conversion rate and NAR (stabilized) were 20.67 mg L⁻¹, 18.32 mg L⁻¹, 7.31 mg L⁻¹, 55.32% and 76.95%, respectively. Within 10 days, the average NH₄⁺–N conversion rate, NAR, and effluent NO₂⁻–N/NH₄⁺–N are 55.71%, 93.25% and 1.08, respectively. This indicated that the UASB system could achieve a stable operation and meet the water inlet substrate requirements for the subsequent anammox process. During phase III (66–94 days), stable operation was achieved in a low NH₄⁺–N environment through low-energy consumption intermittent aeration combined with N₂H₄ dosing strategy. Aeration volume was gradually reduced to lower DO concentration from 2.5 mg L⁻¹ to 0.6 mg L⁻¹, thereby reducing aeration energy consumption. The NH₄⁺–N conversion rate first decreased, then increased from 60.26% to 32.56% and stabilized at 50%. NAR gradually increased to 100% on the 81st day. Inlet water pH at each phase was controlled at 7.5–8.0, and inlet water pH was always lower than the outlet water pH (Fig. 2E). During phase I (effluent pH 7.5–8.0), the pH value of the effluent fluctuated as the AOB activity gradually increased and some partial nitritation was unstable. Effluent pH increased (pH 8.0–8.5) in phase II after adding N₂H₄. Effluent pH fluctuated widely due to N₂H₄ greater impact on AOB and NOB activities. Effluent pH in phase III was stable (8.0–8.5) with small fluctuations. NOB activity was suppressed, and the AOB activity did not change much.

Changes in AOB and NOB activities: the activity of AOB and NOB flora in the later stages of phase I and phase II is shown in Fig. 2F. NOB activity in phase I (0.16 ± 0.01 mg N (g VSS h)⁻¹) was significantly higher than in phase II (0.1 ± 0.01 mg N (g VSS h)⁻¹) (p < 0.05), which also explains the reason for higher effluent NO₃⁻–N concentration in phase I (Fig. 2C). AOB activity in phase I (0.38 ± 0.03 mg N (g VSS h)⁻¹) was significantly lower than in phase II (0.42 ± 0.04 mg N (g VSS h)⁻¹) (p < 0.05). AOB activity in nitrification sludge was much higher than the NOB activity. AOB activity was not inhibited, ensuring a higher nitrite accumulation rate in the low NH₄⁺–N nitrification system (Fig. 2B). Theoretically, inhibiting NOB activity without affecting AOB activity is the key to efficient and stable operation of the partial nitritation system.³⁷ Especially in low NH₄⁺–N environments, it is even more critical to inhibit NOB activity effectively. In phase I, an intermittent aeration strategy was used to suppress NOB activity; in phase II, intermittent aeration combined with N₂H₄ addition was used to inhibit NOB activity further. It can be seen by combining Fig. 2A and B that intermittent aeration combined with N₂H₄ addition showed a more substantial inhibitory effect on NOB. The stable nitritation performance achieved within 24 days via combined intermittent aeration and hydrazine dosing outperforms the single N₂H₄ regulation reported by Song,¹⁵ which required a longer startup time under identical low ammonium loading. Although sole low-DO intermittent aeration partially restrained NOB as described by Zuo,³⁶ its limited nitrite accumulation capacity cannot satisfy the substrate demand of subsequent anammox processes, highlighting the superiority of our integrated regulation strategy.

3.1.2. Community succession characteristics at different stages of operation of a low NH₄⁺–N partial nitritation system under conventional strategies

In Section 3.1.1, the promotion effect of intermittent aeration strategy and intermittent aeration combined with N₂H₄ dosing strategy on the start-up and stable operation of the low NH₄⁺–N nitrification system was investigated. This section combined changes in operating efficiency to examine low NH₄⁺–N at each stage. The relative abundance of each bacterial phylum in the initial sludge sample was Proteobacteria 35.36%, Chloroflexi 15.60%, Actinobacteriota 9.93%, Acidobacteriota 5.73%, Planctomycetota 5.74%, and Nitrospirota 0.25% (Fig. 2G). Proteobacteria is related to biological denitrification, and both AOB and NOB belong to this category. Zhang et al.³⁸ found many Proteobacteria in activated sludge samples from sewage treatment plants. Xiao and Ni³⁹ found similar results. The nitrifying bacteria in the nitrifying sludge cannot adapt to a low NH₄⁺–N environment, leading to a decrease in the Proteobacteria relative abundance to 23.086% in phase I (intermittent aeration strategy control stage). The relative abundance of Actinobacteriota's increased to 10.57%. Some bacteria in Actinobacteriota exert dual effects on nitrification.⁴⁰ Actinomycetes may inhibit AOB growth and activity (slow down the nitrification rate). They may also promote NOB growth and activity (increase nitrification efficiency). Nitrospirota relative abundance phylum dropped significantly. Studies have shown that most microbial genera with nitrifying ability belong to the Nitrospirota phylum.⁴¹

Nitrospirota relative abundance increased from 0.25% to 1.49%, indicating that the inhibitory effect of the intermittent aeration strategy on NOB in a low NH₄⁺–N environment was not ideal. The low NH₄⁺–N nitrosification system was successfully started in phase II (intermittent aeration combined with N₂H₄ dosing control strategy). Proteobacteria relative abundance increased from 23.086% to 28.38%. This is consistent with the study by Yang et al.,⁴² which found that Proteobacteria proliferated during partial nitritation. Actinobacteriota's relative abundance dropped to 8.27%. Based on the previous experimental data, it is speculated that Actinobacteriota negatively affected the nitrification process. Nitrospirota relative abundance decreased from 1.49% to 0.77%. Combined with the system's operational efficiency, this shows that the inhibitory effect of intermittent aeration combined with the N₂H₄ dosing strategy on NOB activity in a low NH₄⁺–N environment was much greater than that of intermittent aeration strategy.

In phase III, dissolved oxygen was reduced to a low DO level to ensure stable system operation. Proteobacteria's relative abundance increases from 28.38% to 37.749%, Actinobacteriota drops to 6.37%, and Nitrospirota decreases from 0.77% to 0.23%. The results show that the intermittent aeration combined with the N₂H₄ dosing strategy was more conducive to the stability of the low NH₄⁺–N nitrification system in a low DO environment. Some species of actinomycetes utilize nitrate nitrogen for denitrification, and aerobic denitrification consumes DO, thereby slowing down nitrification. However, denitrification is usually carried out in a facultative anaerobic environment and cooperates reasonably well with nitrification. The impact of Actinobacteria on the nitrification process was limited, considering that organic carbon sources are insufficient in this system. Studies have found that some Actinobacteria inhibit the growth and activity of AOB, thereby slowing nitrification rates.

On the other hand, some Actinobacteria were found to promote NOB growth and activity to increase nitrification efficiency.⁴² Actinobacteria inhibit AOB's growth and activity (slowing the nitrification rate). They can also promote the development and activity of NOB (increasing nitrification efficiency). After intermittent aeration combined with the N₂H₄ dosing control strategy, the relative abundance of actinomycetes was reduced to 8.27%, indicating that AOB's nitrification rate was increased and NOB's nitrification efficiency was reduced. Among various bacterial genera (Fig. 2H), Nitrosomonas (Proteobacteria phylum) was the most common AOB genus in activated sludge treatment processes and multiple reactors types.^43–45Nitrospira participates in nitrification, mainly as NOB. The relative abundance of Nitrosomonas increased from 0.16% to 0.51%, and Nitrospira rose from 0.25% to 1.49%, indicating that the intermittent aeration strategy promoted the growth of AOB in a low NH₄⁺–N environment. Still, it cannot effectively inhibit NOB in phase I. The relative abundance of Nitrosomonas decreased from 0.51% to 0.32%, and Nitrospira decreased from 1.49% to 0.77% in phase II. The addition of N₂H₄ not only inhibited NOB activity but also affected AOB activity. However, N₂H₄ addition significantly improves the NAR of the low NH₄⁺–N partial nitritation system, as seen in Fig. 2B. In phase III, the relative abundance of Nitrosomonas increased from 0.32% to 0.37%. Nitrospira decreased from 0.77% to 0.23%, indicating that AOB exhibits higher adaptability to the inhibitory effect of N₂H₄ than NOB. It also shows that intermittent aeration combined with the N₂H₄ dosing strategy can permanently inhibit NOB. Intermittent aeration combined with an N₂H₄ dosing strategy in a low dissolved oxygen environment can promote the rapid start-up and stable operation of the low NH₄⁺–N nitrification system.

3.2. New strategies to promote start-up and stable operation of low NH₄⁺–N partial nitritation systems

This section responds to research objective (ii) by comparing the recovery effect and microbial community differences of sludge stored under two temperature conditions. Autotrophic denitrification systems can significantly reduce a sewage plant's energy consumption and infrastructure costs and achieve green denitrification.⁴⁶ According to the results of Section 3.1, the intermittent aeration strategy alone cannot initiate nitrification quickly and efficiently (low NH₄⁺–N wastewater). It needs to be combined with the N₂H₄ dosing strategy to enable the low NH₄⁺–N nitrification system to start quickly and efficiently operate stably. To realize the engineering application of low NH₄⁺–N wastewater treatment (urban sewage) using the PN/A process, finding a strategy to effectively inhibit NOB activity and enrich AOB in a low NH₄⁺–N environment is crucial. Preliminary research by our research group found that after long-term (about 8 months) low-temperature (4 °C) storage of PN/A granular sludge with high NOB activity (early stage), AOB and AnAOB activities can be successfully restored within 21 days under low NH₄⁺–N. It can operate stably for about 3 months without adding inhibitors. After that, if it is operated for a long time without control, NOB will again disrupt the stable operation of partial nitritation.³⁷ Zhang et al.⁴⁷ stored nitrification granular sludge at 4 °C for 40 days and found that it took only 22 days to restore the system efficiency to the pre-storage level. Qian⁴⁸ directly stored nitrite granular sludge and then restored it. They found that granular sludge stored at room temperature for one year could reach a nitrite accumulation rate of more than 80% after 8 days of operation. Given the long-term low-temperature conditions in autumn and winter across many regions of the world, it is of greater engineering significance to determine whether directly storing sludge at room temperature can quickly initiate nitrification in a low NH₄⁺–N environment. The operating conditions in Section 2.3 were followed.

3.2.1. Changes in storage sludge activity and sludge characteristics

3.2.1.1. Morphological changes of sludge after storage under different temperature conditions

The sludge strongly smelled of rotten eggs after 5 months of long-term storage. The sludge from ambient winter storage had a stronger pungent odor. The color of the sludge changed from brown to black (see SI material). Endogenous respiration and anaerobic digestion occur after long-term anaerobic storage of sludge. Thus, the degree of sludge and gas inorganization, such as hydrogen sulfide and methane, will increase.^49–51 Wang⁵² found that sulfide generated during storage mainly comes from the hydrolysis of dead bacteria; sulfide increases with extended storage time, and the color increases with sulfidation. The sludge structure was looser than before storage; the sludge expanded and disintegrated and suspended solids in the water increased (see SI material).

This phenomenon may be due to the structural differentiation of the bacterial community within the sludge during low-temperature storage, separating some dead and live bacteria. When bacteria are in a state of starvation, nutrients are released from extracellular polymers and mineralized by heterotrophic bacteria, thereby alleviating the starvation. The bacteria quickly consume these minerals in the early phase of storage. The materialized inorganic nitrogen source then gradually utilizes the abundant substances in extracellular polymers for metabolism, leading to microbial death and disintegration, decreasing sludge stability and microbial activity.⁵³

Many tiny particles are visible to the naked eye in the sludge stored at 4 °C. However, sludge stored in low-temperature environments in autumn and winter exhibits severe agglomeration and floating, and some parts turn into irregular breccia-sized lumps. Ambient winter storage causes more severe damage to sludge activity than refrigerated storage at 4 °C. AOB and NOB exhibit various shapes, including spherical, rod-shaped, and spiral forms. It is generally difficult to distinguish AOB and NOB species based on their morphology.⁵⁴ The AOB shapes that often appear in sewage treatment plants are mainly short, rod-shaped, or spherical, whereas the NOB shapes that frequently occur are mostly spiral and rod-shaped. SEM of nitrification sludge before storage was rod-shaped, short rod-shaped and a small number of spherical bacteria. The surface was rough, so it was inferred that they were mainly NOB and AOB. SEM of nitrification sludge stored at 4 °C for 5 months shows that rod-shaped bacteria remain the main form, with some filamentous bacteria wrapped around the surface. Bacterial cells diameter were significantly smaller, thinner, and more wrinkled as shown in SEM images of nitrification sludge after 5 months of ambient winter storage. Compared with before storage, the bacterial cells expanded significantly, and their original shape could no longer be distinguished. Bacterial cells in the sludge will aggregate during storage. Bacterial aggregation will become more severe as activity decreases,⁵² consistent with the sludge agglomeration observed in Fig. 3C. Bacterial cells in the sludge stored at 4 °C are well preserved. In contrast, the decomposition and expansion of sludge stored at low temperatures in autumn and winter are serious problems. Compared with the sludge stored at 4 °C, its microbial activity has decreased significantly (p < 0.05).

Fig. 3. (A) Changes in EPS composition of nitrifying sludge, (B) NAR of R1 and R2 systems (C) nitrite concentrations in influent and effluent of R1 and R2 systems, (D) NH₄⁺–N conversion rate of R1 and R2 systems, (E) changes in AOB and NOB activity in each phase of R1 system and, (F) changes in AOB and NOB activity in each phase of R2 system. Note: R1 was inoculated with sludge stored at 4 °C for 5 months, and R2 with sludge stored at natural autumn-winter ambient temperature. Error bars in panel A represent standard deviation (n = 3).

3.2.1.2. Physical and chemical properties of sludge after storage under different temperature conditions

MLVSS/MLSS of nitrifying sludge was 35.16%, and the sludge volume index (SVI) was 58 mL g⁻¹ before storage. After being stored at 4 °C for 5 months, MLVSS/MLSS of nitrification sludge dropped from 35.16% to 30.21%, and the SVI increased from 58 mL g⁻¹ to 173 mL g⁻¹. The MLVSS/MLSS of nitrification sludge after 5 months ambient winter storage decreased from 35.16% to 28.69%, and the SVI increased from 58 mL g⁻¹ to 203 mL g⁻¹. Microorganisms cannot obtain nutrient or dissolved oxygen during long-term sludge storage. Endogenous respiration by microorganisms caused cell hydrolysis, and some died. Settling performance and VSS content of the sludge decreased. The deterioration of sludge stored at low temperatures was more serious in autumn and winter.

Extracellular polymeric substance: EPS comprises protein (PN) and polysaccharide (PS). EPS is an important parameter that reflects sludge performance.⁵⁵ PN and PS content dropped significantly after long-term storage of nitrosated flocculent sludge (Fig. 3A). During storage, the EPS secreted by the sludge was used by microorganisms as a substrate due to nutrient limitations.⁵⁶ The total amount of EPS before sludge storage, after the sludge was stored (4 °C, 5 months), and after the sludge was stored (ambient winter temperature, 5 months) was 90.5 mg g⁻¹ VSS (PN 62.2%), 25.9 mg g⁻¹ VSS, (PS 50.9%), and 20.8 mg g⁻¹ VSS (PS 50.4%), respectively. After sludge storage, the total EPS decreased significantly, and the proportion of PS exceeded PN (p < 0.05). The high hydrophobicity of protein (PN) promotes sludge aggregation and is the most important factor affecting sludge flocculation.⁵⁷ There is a correlation between PN/PS and sludge SVI. The larger the PN/PS value, the lower the SVI value, and the better the sludge settleability,⁵⁸ which is consistent with the results of this test.

3.2.2. NH₄⁺–N nitrification system operation efficiency under the new strategy

The operating conditions can be seen in Section 2.3.

3.2.2.1. Changes in nitrogen element in each phase of R1 and R2 operation

The influent pH was 7.5–8.0 at each R1 and R2 operation stage, and the reactor temperature was maintained at 35 ± 1 °C according to the different sensitivities of AOB and NOB. NAR of the sludge stored under two temperature conditions always remained high during recovery, indicating that the NOB activity was effectively inhibited. However, NAR and NH₄⁺–N conversion rates fluctuated significantly during the early stages of the R1 and R2 operating systems (Fig. 3B–D). Long-term storage of sludge led to the emergence of many anaerobic bacteria that could not adapt to the aerobic environment and caused autolysis during the activity recovery process (sludge morphology changes Section 3.2.1). During the start-up phase of the R1 system (1–15 days), influent NH₄⁺–N 40 ± 5 mg L⁻¹, a hydraulic retention time (HRT) of 10 h, and an intermittent aeration mode (aeration 5 min, stop exposure 10 min), and DO mass concentration was 0.5 ± 0.2 mg L⁻¹ to achieve adequate enrichment of AOB. The NH₄⁺–N conversion rate was 50%, and effluent NO₂⁻–N showed an upward trend from 5 to 9.89 mg L⁻¹. NAR stabilized at 100% (beginning of the R1 start-up phase), fluctuated significantly on the 5th day, and gradually stabilized at around 80% on the 15th day. This showed that the low NH₄⁺–N nitrification system in R1 was successfully started. Operating conditions during the start-up phase of the R2 and R1 systems were identical.

R2 effluent NO₂⁻–N showed an upward trend of 2.45 to 10.19 mg L⁻¹ (1 to 32 days); NAR fluctuated initially and gradually stabilized at 95% (9th day). However, the NH₄⁺–N conversion rate dropped below 50% from the 2nd day and fluctuated around 30%. HRT was increased to 12 h on the 32nd day to improve the NH₄⁺–N conversion rate of the R2 system. However, the NH₄⁺–N conversion rate did not increase. It is necessary to ensure that the NH₄⁺–N conversion rate in the nitrification phase is about 50% to meet the requirements of the PN/A process. The R2 system failed to start the low NH₄⁺–N nitrification system. Nitrification sludge stored at low temperatures for 5 months in autumn and winter can exhibit a strong inhibitory effect on NOB. Still, it sacrifices the system's NH₄⁺–N conversion rate, making it unable to couple with the subsequent anammox process to form a PN/A process. During the R1 stable operation (15–50 days), effluent NO₂⁻–N stabilized at 16 mg L⁻¹ (50th day), NAR 85%, and NH₄⁺–N conversion rate 65% (35th day). The NH₄⁺–N conversion rate for R2 was below standard, and increasing HRT was required. This result shows that storing nitrification sludge at 4 °C for 5 months can produce a long-lasting inhibitory effect on NOB. At lower influent NH₄⁺–N concentrations, sludge inoculated into the UASB reactor can efficiently enrich AOB, indicating that this strategy can quickly start-up part of the nitrification system and ensure that the effluent water quality meets the inlet water quality requirements of the subsequent anammox process.

3.2.2.2. Changes in AOB and NOB activity in R1 and R2 systems

Nitrifying sludge was stored for 5 months under two conditions: refrigerated storage at 4 °C and ambient winter storage. At the end of storage, the sludge was in poor condition, making it difficult to measure AOB and NOB activities accurately. Three conditions were selected: before sludge storage, during recovery and after recovery. A comparative analysis of AOB and NOB activity mapping during each period. AOB activity before nitrification sludge storage 0.36 ± 0.01 mg N (g VSS h)⁻¹, and the NOB activity 0.24 ± 0.01 mg N (g VSS h)⁻¹ (Fig. 3E). AOB (0.32 ± 0.01 mg N (g VSS h)⁻¹) and NOB (0.03 ± 0.01 mg N (g VSS h)⁻¹) activities were improved to a certain extent after operating R1 for 20 days. AOB and NOB activities are inhibited to varying degrees after storing nitrifying sludge at 4 °C for 5 months, and the degree of inhibition of NOB activity was far greater than that of AOB. The AOB and NOB activity increased to 0.39 ± 0.01 mg N (g VSS h)⁻¹ and 0.07 ± 0.01 mg N (g VSS h)⁻¹ after operating R1 for 40 days. AOB activity was fully restored and exceeded its pre-storage level. In contrast, NOB activity increased slowly, indicating that the nitrifying sludge exhibited greater NOB inhibition after storage at 4 °C for 5 months, with the inhibition effect lasting longer. AOB recovered more quickly than NOB, and it returned to its pre-storage activity level in about 20 days.

The AOB and NOB activity was 0.21 ± 0.01 mg N (g VSS h)⁻¹ and 0.01 ± 0.01 mg N (g VSS h)⁻¹ after operating R2 for 20 days (Fig. 3F). The AOB and NOB activity increased to 0.23 ± 0.01 mg N (g VSS h)⁻¹ and 0.03 ± 0.01 mg N (g VSS h)⁻¹ after operating R2 for 40 days. This shows that AOB and NOB died in large numbers rather than remaining dormant after nitrification sludge was stored at low temperatures in autumn and winter. This also verified the SEM analysis results in Section 3.2.1 and explained the reason for NH₄⁺–N removal efficiency during R2 operation. Comparing changes in AOB and NOB activity across the R1 and R2 phases, it was found that 4 °C refrigerated storage and ambient winter storage exerted similar inhibitory effects on NOB. However, the damage to AOB activity caused by storage in autumn and winter was much more serious than at 4 °C, making it difficult to recover the activity quickly. It benefited the start-up and stable operation of the low NH₄⁺–N nitrification system and could more effectively inhibit NOB activity while enriching AOB.

3.2.3. Novel strategies promote community succession characteristics at various stages of low NH₄⁺–N nitrification system operation

During the start-up and stable operation of the low NH₄⁺–N nitrification system, the nitrification sludge used in the experiment underwent drastic changes in living conditions, from anoxic to aerobic conditions, and from endogenous respiration to nitrification proliferation. The sludge structure was destroyed, and sludge decomposition was serious. During reactor operation, sludge and microorganisms flow out with the effluent, and the system's dominant bacterial flora changes. High-throughput sequencing technology analyzes the changes in the relative abundance of the bacterial community during system operation. It explores the impact of sludge storage and recovery strategies on the start-up and stable operation of the low NH₄⁺–N nitrification system.

3.2.3.1. Changes in the bacterial relative abundance before and after sludge storage

Before and after nitrification, the composition of bacterial phyla with higher relative abundance was consistent. The relative abundance of each bacterial phylum in the initial sludge sample before storage was Proteobacteria 35.36%, Chloroflexi 15.60%, Actinobacteriota 9.93%, Acidobacteriota 5.73%, Planctomycetota 5.74%, and Nitrospirota 0.25%. Proteobacteria are closely related to biological denitrification, and both AOB and NOB belong to this category.⁵⁹

Zhang;³⁸ Xiao and Ni³⁹ showed that Proteobacteria accounted for the most significant proportion (40% to 70%) of the activated sludge in most wastewater treatment processes. B1 is the nitrifying sludge sample after 5 months of 4 °C refrigerated storage, and C1 is the sample after 5 months of ambient winter storage (−25 to 12 °C) (Fig. 4A). Proteobacteria relative abundance in B1 and C1 was 28.50% and 25.21%, respectively. After storage under different temperature conditions, the relative abundance of Proteobacteria associated with nitrification decreased (C1 decreased more than B1). After storage under various temperature conditions, the relative abundances of Chloroflexi, Acidobacteriota, and Planctomycetota gradually decreased. These microbial groups have limited environmental adaptability. When the environment changes, their relative abundance may be affected.⁶⁰ Studies have shown that some Actinobacteria can inhibit the growth and activity of AOB (slowing down the nitrification rate). At the same time, some Actinobacteria promote the development and activity of NOB (increasing the nitrification efficiency).⁴⁰ Actinobacteriota relative abundance in B1 increased significantly compared with the initial sludge, and in C1 it increased slightly compared with A0. The relative abundance of the Nitrospirota phylum has decreased dramatically. Studies have shown that most microbial genera with nitrifying ability belong to the Nitrospirota phylum, including the common NOB: Nitrospira.⁶⁰ This proved that the two storage conditions are conducive to inhibiting NOB activity from a microbial perspective. The succession rules of the microbial community structure at the genus level for each sample were analyzed to further elucidate the evolution of the bacterial community (Fig. 4B). Nitrosomonas is the most common AOB in activated sludge treatment processes and various types of reactors. At the same time, while Nitrospira participates in nitrification, mainly nitrite-oxidizing bacteria (NOB).^41,61Nitrosomonas and Nitrospira relative abundance in A0, B1 and C1 was 0.16% and 0.25%, 0.04% and 0.11%, 0.06% and 0.13%, respectively. The relative abundances of Nitrosomonas and Nitrospira decreased to varying degrees after nitrifying sludge was stored at different temperatures, indicating that the AOB and NOB activities were inhibited to varying degrees.

Fig. 4. Relative abundance of microbial communities before and after 5 months sludge storage at (A) phylum and (B) genus levels; (C and D) community succession during R1 operation at phylum and genus levels, respectively; (E and F) community succession during R2 operation at phylum and genus levels, respectively; (G and H) comparison of community structure between R1 and R2 at the stable operation phase at phylum and genus levels, respectively. Note: sample codes: B1, sludge after 5 months storage at 4 °C; B2, R1 sludge at phase I (startup phase); B3, R1 sludge at phase II (stable phase). C1, sludge after 5 months storage at ambient autumn-winter temperature; C2, R2 sludge at phase I; C3, R2 sludge at phase II. All data are obtained from 16S rRNA gene high-throughput sequencing.

3.2.3.2. Changes in flora's relative abundance in each phase of R1 and R2 operation

Relative abundance changes of flora in each phase of R1 operation: B2 and B3 are the nitrifying sludge samples during and after the recovery process, respectively. Proteobacteria's relative abundance related to nitrification increased and then decreased (Fig. 4C and D). Nitrifying sludge regains activity after long-term starvation, but some bacterial species cannot adapt to the low NH₄⁺–N environment, thereby decreasing the phylum's relative abundance. There was no significant fluctuation in the relative abundance of Chloroflexi, Acidobacteriota, and Planctomycetota, indicating that the recovery of sludge activity did not increase their abundance. Actinobacteriota inhibits AOB activity and promote NOB activity, which was not conducive to the stability of the partial nitritation process.⁴⁰ The relative abundance of Actinobacteriota decreased significantly from 15.62% in B1 to 7.8% in B3, which was much lower than 9.93% in A0. This shows that sludge storage at 4 °C increased the abundance of Actinobacteriota. However, when combined with the sludge recovery process, its growth and activity are effectively inhibited, thereby benefiting the low NH₄⁺–N nitrification system. The relative abundance of Nitrospirota dropped from 0.11% of B1 to 0.0002%, which was much lower than the abundance before nitrification sludge storage, indicating that sludge storage and recovery at 4 °C effectively inhibit the growth of nitrifying bacteria (beneficial to improve the stability of the low NH₄⁺–N partial nitritation system). The common AOB mainly includes the genus Nitrosomonas, and the relative abundance continues to increase from 0.04% in B1 to 0.13% in B3. NOB mainly represented by the genus Nitrospira, and its relative abundance decreases from 0.11% in B1 to 0.02% in B3. This result corresponds to the test result in Section 3.2.2.

Changes in the relative abundance of flora in each phase of R2 operation: C2 and C3 were the nitrifying sludge samples during and after recovery, respectively. Like the R1 system, Proteobacteria's relative abundance increased and then decreased (Fig. 4E and F). Nitrifying sludge regains activity after long-term starvation, but some bacterial species cannot adapt to the low NH₄⁺–N environment, resulting in decreased relative abundance of the bacterial phylum. The relative abundance of Chloroflexi, Acidobacteriota, Planctomycetota, and Actinobacteriota all showed a downward and increasing trend (little overall change). Actinobacteriota inhibit AOB activity and promote NOB activity, which is detrimental to the partial nitritation process.³⁷ The relative abundance of Actinobacteriota dropped significantly from 10.54% in C1 to 9.59% in C3, which was slightly lower than 9.93% in A0. Overall, no significant change. This shows that the recovery strategy after low-temperature storage of sludge in autumn and winter had no noticeable inhibitory effect on the abundance of Actinobacteriota. The relative abundance of Nitrospirota first decreased from 0.14% of C1 to 0.09% of C2, then increased to 0.13% of C3. This indicates that sewage sludge stored at low temperatures in autumn and winter cannot be restored. Effectively inhibits the growth of nitrifying bacteria, which is not conducive to the stable operation of low NH₄⁺–N nitrification systems. AOB mainly includes the genus Nitrosomonas. Its relative abundance first decreased from 0.066% in C1 to 0.035% in C2 and then increased to 0.064%. NOB primarily consists of the genus Nitrospira. Its relative abundance decreased from 0.13% in C1 to 0.09% in C2 and then increased to 0.12% in C3. From a microbiological perspective, it was demonstrated that nitrification sludge was stored at low temperatures in autumn and winter for 5 months and before inoculation into the UASB reactor. It was impossible to inhibit NOB for a long time and thereby enrich AOB, which was not conducive to the start-up and stable operation of the low NH₄⁺–N nitrification system.

3.2.3.3. Comparison of flora's relative abundance in each phase of the R1 and R2 system

The startup stages of the R1 and R2 systems are the B2 and C2 stages, respectively (see SI material). The relative abundance of Proteobacteria in B2 and C2 was 39% and 41.62%, respectively. Proteobacteria relative abundance associated with nitrification in the B2 system was higher than in the C2 system, but the difference was not significant. The two systems showed no significant differences in the relative abundances of Chloroflexi, Acidobacteriota, and Planctomycetota. The relative abundance of Actinobacteriota in B2 and C2 was 4.99% and 7.84%, respectively. Actinobacteriota will inhibit the growth and activity of AOB, and its relative abundance was high. It is not conducive to starting the partial nitritation system. Microbial genera with nitrifying ability include Nitrospira, a common NOB.⁴¹ Nitrospirota relative abundance in B2 and C2 was 0.04% and 0.09%. NOB activity in the R1 system was even lower. Among the bacterial genera, Nitrosomonas had relative abundances of 0.09% and 0.03% in B2 and C2, respectively. AOB abundance in B2 was higher, Nitrospira relative abundance 0.04% and 0.09%, respectively, and the activity of NOB in B2 was lower. In the start-up phase of the low NH₄⁺–N partial nitritation system, the R1 system inhibits NOB more thoroughly and benefits the AOB enrichment. The stable operation stages of the R1 and R2 systems are B3 and C3, respectively. The relative abundances of Proteobacteria in B3 and C3 are 30.72% and 24.73%, respectively. At the phylum level, Proteobacteria relative abundance related to nitrification in the B3 phase was higher than in the C3 phase (Fig. 4G and H). Nitrifying sludge enters a dormant state during storage at 4 °C, can recover quickly under certain conditions, and has a relatively large proportion. The relative abundance of Chloroflexi, Acidobacteriota, and Planctomycetota in B3 was lower than that in C3. The relative abundance of Actinobacteriota in B3 and C3 is 7.81% and 9.59% respectively. Actinobacteriota high relative abundance was not conducive to the stable operation of the partial nitritation system. Nitrospirota in B3 and C3 was 0.0002% and 0.13%. This shows that at the bacterial phylum level in the stable phase, the R1 system can effectively inhibit NOB activity for a long time. Among the bacterial genera, the relative abundances of Nitrosomonas and Nitrospira in B3 and C3 were 0.13% and 0.06%, and 0.02% and 0.12%, respectively. NOB activity in B3 was lower, and AOB abundance in B3 was higher. AOB was enriched in the R1 system, and the NOB inhibition effect was better, which indicates that the 4 °C storage sludge recovery strategy was more beneficial to the low NH₄⁺–N nitrification system. The 15 days rapid activity recovery of AOB in 4 °C preserved sludge agrees with the findings of Maturano-Carrera,²² who confirmed that low-temperature storage slows microbial endogenous respiration and cell autolysis. Compared with the 22 days recovery cycle of 4 °C-stored nitrifying sludge reported by Zhang,⁴⁷ our system exhibits faster startup efficiency. In contrast, severe irreversible AOB inactivation in ambient winter-stored sludge aligns with previous reports on unregulated temperature preservation of sludge. From a microbiological perspective, storing nitrifying sludge at low temperatures in autumn and winter for 5 months is preferable. Storing nitrifying sludge at 4 °C for 5 months and then inoculating it into the UASB reactor resulted in better, long-term NOB inhibition and efficient AOB enrichment. It had a good promotion effect on the start-up and stable operation of the low NH₄⁺–N partial nitritation system. This result corresponds to the test result in Section 3.2.2.

3.2.3.4. Functional gene prediction analysis

PICRUSt2 can predict the functional composition of a microbial community from amplicon sequencing results. Based on the above microbial sequencing results, PICRUSt2 was used to conduct bacterial functional analysis and predict the abundance of different functional genes. Combined with multivariate statistical and machine learning tools, such omics prediction pipelines have been widely used to disentangle the linkage between operational parameters and key functional gene expression under low ammonium loading.⁶² The nitrogen cycle process of the PN/A process is shown in Fig. 5A. The abundance prediction results of functional genes related to the partial nitritation process are shown in Fig. 5B and C.

Fig. 5. (A) Nitrogen cycling process and required functional genes in autotrophic denitrification process. (B) changes in relative abundance of major functional genes of R1 system, and (C) changes in relative abundance of major functional genes of R2 system.

The results showed that the relative abundance of ammonia monooxygenase (AMO) gradually increased, and nitrite oxidoreductase (NXR) was not predicted. This was consistent with previous experimental results, indicating that nitrite oxidation was inhibited and confirming the dominance of the partial nitritation process. The relative abundance of AMO in R1 and R2 increased by 252% and 123%, respectively, indicating that the AOB function of the R1 system was more powerful under the new strategy and higher than that of the R2 system (AMO oxidize NH₄⁺–N to NH₂OH). Hydroxylamine oxidoreductase (HAO) relative abundance in R1 gradually increased (HAO: primary functional gene for NO₂⁻–N production). Combined with the gradual increase in the NO₂⁻–N accumulation rate (NAR) during the nitrification process, the new strategy of sludge storage and recovery at 4 °C promotes the overexpression of AMO in AOB and promotes the stable operation of the low NH₄⁺–N nitrification system. Although the reference gene database used by PICRUSt 2 was more than 10 times larger than that of PICRUSt1 (Greengene database), it failed to predict the primary functional gene of NOB (NXR). Considering the efficient and stable operation of the reactor, combined with changes in activity and the relative abundance of NOB, clearly shows that the NO₂⁻–N oxidation pathway was effectively inhibited, further confirming that AOB plays a dominant role in the partial nitritation process.

3.3. Influence of room temperature on the stable operation of the low NH₄⁺–N partial nitritation system

This section fulfills research objective (iii), exploring the stability of the optimized system under cooling conditions and its microbial response mechanism. Temperature is an important factor affecting the partial nitritation system. An increase in temperature accelerates the partial nitritation reaction and increases microbial metabolic activity. The optimal temperature for a partial nitritation reaction is around 35 °C. However, the room temperature was about 25 °C. The normal temperature is measured at 20 °C in spring and autumn in most places in China, such as the circulating water temperature (natural water temperature). Lowering the system from the optimal to the normal temperature slows microorganism metabolic activity, altering the distribution and relative abundance of the bacterial community. The strength of microbial metabolism directly affects the denitrification effect and stability of the nitrification system. This will further influence the engineering application of the PN/A process. The rate of partial nitritation reaction will be significantly reduced in a low NH₄⁺–N environment due to insufficient substrate concentration. As a result, the reaction rate decreased, and overall operating efficiency was affected. Therefore, exploring the impact on the nitrification system under the double unfavorable environment of standard temperature and low NH₄⁺–N is necessary.

Based on the stable operation of the low NH₄⁺–N nitrification system in R1, the reactor's operational efficiency was examined by measuring changes in various parameters during the test to assess the long-term stability of the new strategy. The influence of room-temperature conditions on the stable operation of low NH₄⁺–N nitrification systems was investigated. The abundance of changes in microbial community succession and functional flora at each phase was analyzed using high-throughput sequencing. The microbiological mechanism of the low NH₄⁺–N partial nitritation system at ambient winter temperature was discussed.

3.3.1. Changes in nitrogen element in the low NH₄⁺–N nitrification system during cooling process

3.3.1.1. Changes in NH₄⁺–N throughout the process

The changes in NH₄⁺–N in R1 during the test are shown in Fig. 6A. As the temperature decreased, the NH₄⁺–N conversion rate showed an overall downward trend. As the temperature continues to decline from 30 °C, 25 °C, and 20 °C, the NH₄⁺–N conversion rate was 60%, 50%, and 45%, respectively.

Fig. 6. Low temperature operation: changes in (A) NH₄⁺–N concentration and NH₄⁺–N conversion rate in influent and effluent of R1 system, (B) nitrite concentration and NAR in R1 system, (C) nitrate nitrogen concentration in R1 system, (D) AOB and NOB activities, comparison of relative abundance of bacterial communities during the cooling phase (E) phylum and (F) genus, and (G) changes in the relative abundance of major functional genes during cooling process.

The average NH₄⁺–N concentrations in the inlet and outlet water were 37.78 mg L⁻¹ and 15.14 mg L⁻¹, 37.07 mg L⁻¹ and 17.67 mg L⁻¹, and 39.9 mg L⁻¹ and 18.6 mg L⁻¹, respectively. After the control measures (N₂H₄ 5 mg L⁻¹), the NH₄⁺–N conversion rate returned to about 50%. The NH₄⁺–N conversion rate of the low NH₄⁺–N nitrification system met the feed-water substrate requirements of the subsequent anammox process for autotrophic denitrification (an NH₄⁺–N conversion rate above 50%) at 25–35 °C. The NH₄⁺–N conversion rate of the low NH₄⁺–N partial nitritation system does not meet the standard at 20 °C, and an additional strategy of N₂H₄ addition from an external source was required. This shows that the low NH₄⁺–N nitrification system in R1 meets the requirements of the PN/A process over the temperature range of 25 °C to 35 °C.

3.3.1.2. Changes in NO₂⁻–N along the process

The changes in NO₂⁻–N in R1 during the test are shown in Fig. 6B. It was found that as the temperature decreased, the effluent NO₂⁻–N concentration remained generally stable at around 15 mg L⁻¹. The NAR fluctuated greatly and showed an overall downward trend. When the temperature drops to 30 °C, the NAR shows no significant change and fluctuates between 80% and 85%. When the temperature drops to 25 °C, the NAR drops to about 75%. As the temperature drops to 20 °C, the NAR drops below 70%. The low NH₄⁺–N partial nitritation system becomes unstable. After adopting the exogenous hydrazine addition strategy in Section 3.1, its NAR rose to 75%. When the system temperature is between 25 °C and 35 °C, good nitrification performance can be ensured in the low NH₄⁺–N partial nitritation system. However, at 20 °C, the system's nitrification performance dropped significantly. An additional strategy of adding hydrazine from an external source was used to regulate and restore the NAR to about 75%. Changes in NO₃⁻–N along the process: NH₄⁺–N changes in R1 during the test are shown in Fig. 6C. The inlet water NO₃⁻–N concentration remained overall stable. As the temperature decreased, the effluent NO₃⁻–N concentration increased. When the temperature dropped to 30 °C, 25 °C, and 20 °C, the effluent NO₃⁻–N concentration was 4 mg L⁻¹, 5 mg L⁻¹, and 6 mg L⁻¹. After taking control measures by adding hydrazine (N₂H₄ was 5 mg L⁻¹), the effluent NO₃⁻–N concentration dropped significantly to about 5 mg L⁻¹, and the system returned to stability. When the temperature dropped to 20 °C, the inhibitory effect of the sludge storage strategy on NOB was weakened, increasing the effluent NO₃⁻–N concentration. Combined with changes in NO₂⁻–N, it was speculated that cooling had little effect on AOB activity. However, the inhibitory effect of the sludge storage recovery strategy on NOB activity will be weakened. A stable low NH₄⁺–N nitrification system using sludge storage and recovery strategy maintained stable operation at 25–35 °C and showed good nitrification performance. However, when the temperature dropped to 20 °C, the stability of the low NH₄⁺–N partial nitritation system was destroyed. At this time, the hydrazine addition strategy in Section 3.1 can be used to restore stability.

3.3.2. Changes in AOB and NOB activity in low NH₄⁺–N partial nitritation system during the cooling process

AOB and NOB activities show an upward trend, as seen in Fig. 6D. The AOB activity at the system temperature was 0.39 ± 0.01 mg N (g VSS h)⁻¹ at 30 °C. It rose to 0.42 ± 0.01 mg N (g VSS h)⁻¹ at 20 °C, and the NOB activity increased from 0.07 ± 0.01 mg N (g VSS h)⁻¹ at 30 °C to 0.12 ± 0.01 mg N(g VSS h)⁻¹ at 20 °C. AOB and NOB growth rates are defined using 20 °C as the dividing point.⁶³ The specific growth rate of NOB was higher than that of AOB at ambient temperatures lower than 20 °C. AOB-specific growth rate begins to be higher than that of NOB above 20 °C, and the difference in specific growth rate increases as the temperature increases. At the same time, NOB was more adaptable to adverse environments than AOB. NOB quickly adapted to the cooling process. NOB activity gradually increases with a larger amplitude, while AOB activity increases slowly under the influence of cooling. The cooling process increases NOB activity, decreasing the system NAR, and affecting the stable operation of the low NH₄⁺–N partial nitritation system. The faster growth rate of NOB at 20 °C observed in this study aligns with the temperature-dependent kinetic differences between AOB and NOB summarized by Arora.⁶³ Unlike previous temperature-control strategies that failed to reverse NOB overgrowth at room temperature, supplementary hydrazine dosing effectively restored system stability, offering a practical auxiliary regulatory method that has not been reported in prior low-temperature mainstream PN/A research.

3.3.3. Community succession characteristics at each phase of operation of the low NH₄⁺–N partial nitritation system during the cooling period

This section combines changes in operating efficiency to explore the successional characteristics of the community structure at each phase of the low NH₄⁺–N nitrification system during the cooling period. Proteobacteria is a bacterial phylum closely related to biological denitrification (AOB and NOB are in this group). Zhang³⁸ found that Proteobacteria were present in activated sludge samples from sewage treatment plants and accounted for a large proportion. Xiao and Ni³⁹ showed similar results. The relative abundance of Proteobacteria associated with nitrification increases as system temperature decreases. Still, the increase rate slows down at 20 °C (Fig. 6E). Since some bacterial species cannot adapt to the 20 °C environments, the relative abundance of bacterial phyla no longer increases. The relative abundance of Chloroflexi, Acidobacteriota, and Planctomycetota showed no significant fluctuations. Some bacteria in the phylum Actinobacteriota had a dual effect on nitrification. Actinomycete bacteria may inhibit AOB growth and activity, thereby slowing nitrification rates. They may also promote NOB growth and activity, as well as increase nitrification efficiency. The specific mechanisms of impact and differences between bacterial species still require further study. Actinobacteriota relative abundance first increases and then decreases as the temperature decreases. At 20 °C, its relative abundance drops to 7.57%, well below its pre-cooling level. Based on previous experimental data, it is speculated that the influence of Actinobacteriota on the nitrification process has decreased. Studies have shown that most microbial genera with nitrifying ability belong to the Nitrospirota phylum.⁶⁴ The relative abundance of NOB shows an upward trend from 0.0002% to 0.0027% as the temperature decreases, much lower than the abundance of nitrifying sludge before cooling. This indicated that the cooling process had a low impact on nitrifying bacteria such as NOB.

Fig. 6F analyzed the succession rules of the microbial community structure at the genus level for samples under various temperature conditions in the low NH₄⁺–N nitrification system. Among bacterial genera, Nitrosomonas (Proteobacteria phylum) is widely distributed across various reactors and activated sludge treatment processes (a common AOB genus).^40,42Nitrospira participates in nitrification, mainly nitrite-oxidizing bacteria (NOB). As the temperature decreases, the relative abundances of Nitrosomonas and Nitrospira continue to increase from 0.13% to 0.15% and 0.02% to 0.05%, respectively. The increase in NOB was much more significant than that of AOB, indicating that NOB's ability to adapt to the cooling process was much greater than that of AOB. The analysis found that the overall cooling process did not adversely affect the system's operational stability. The new strategy for sludge storage and recovery can enable long-term stable operation of low NH₄⁺–N nitrification systems.

3.3.4. Functional gene prediction analysis

The abundance predictions for functional genes related to the partial nitritation process are shown in Fig. 6G. The results showed that the AMO relative abundance gradually increased, whereas NXR was not predicted. This indicated that the nitrite oxidation process remained inhibited, and that the partial nitritation process remained dominant during cooling to room temperature. The relative abundance of AMO increased by 120% under ambient winter temperature, and the AOB function continued to enhance, producing a large amount of AMO. The ability to oxidize more NH₄⁺–N to NH₂OH demonstrated that the new strategy of storage and recovery at 4 °C could stabilize the low NH₄⁺–N partial nitritation system in the long-term. The relative abundance of HAO remained stable overall, and HAO was the primary functional gene for NO₂⁻–N production. Combined with the stable changes in the NO₂⁻–N accumulation rate (NAR) during the cooling process, it further showed that the new strategy of sludge storage and recovery at 4 °C can promote the over-expression of AMO in AOB and promote the stable operation of the low NH₄⁺–N partial nitritation system in a normal temperature environment. Although the reference gene database used by PICRUSt2 was more than 10 times larger than that used by PICRUSt1 (Greengenes). However, it failed to successfully predict the main functional gene of NOB (NXR). This observation raises concerns. Considering the reactor's efficient and stable operation, along with changes in activity and the relative abundance of NOB, it is clear that the NO₂⁻–N oxidation pathway was effectively inhibited. It further proved that AOB is significantly dominant in the partial nitritation process.

3.4. Economic costs and implications of the strategies of sludge storage and recovery

The PN/A process uses less energy than the traditional denitrification method (Table 1). Even though hydrazine addition raises a portion of the overall cost of treating water in the water plant, the overall cost of water treatment has decreased. Hydrazine sulphate is suggested for practical applications due to its stability, safety, and ease of storage. Adding hydrazine can be used as a recovery strategy in extreme situations, such as when the PN/A system's performance deteriorates, and its effect is enhanced when combined with conventional control methods. This study still has several limitations. All experiments were carried out in lab-scale UASB reactors using synthetic inorganic wastewater, whereas actual municipal wastewater contains soluble organic matter, suspended particles, and various bacteria, which may weaken the NOB-inhibition effect of the combined strategy. In addition, this work tested only 5 months sludge storage and a stepwise cooling mode from 35 °C to 20 °C; the performance of the strategy under longer storage periods and natural temperature fluctuations still needs long-term verification. For microbial analysis, functional gene abundance was inferred by PICRUSt2 prediction rather than direct quantification, and the actual expression of key functional genes requires further validation via qPCR or metagenomic sequencing. Pilot-scale tests are also needed to verify the engineering feasibility and actual operating cost of the strategy before practical application.

Table 1. Economic costs^a.

Processing capacity 6000 m3 per day	A2/O process	New biological denitrification combined process (anaerobic digestion + short-cut nitrification/anaerobic ammonium oxidation)
Civil construction	414	180
Electrical equipment's	63	43
Total one-time investment	477	223
Basic operating cost (per year)	314	267
Carbon source input cost (per year)	48	0
Resource recovery value	0	55
Hydrazine dosage cost (per year)	0	20
Total operating cost (per year)	362	232
Water treatment cost ($per ton)	1.65	1.06

^aAll costs are presented in USD (2025 values) and calculated for a wastewater treatment plant with a treatment capacity of 6000 m³ per day. An RMB-to-USD exchange rate of 1393:1 is applied for cost conversion.

4. Conclusion

This study focused on the difficulty of NOB inhibition and unstable operation of partial nitritation systems under low-ammonium conditions. It investigated the regulatory effect of hydrazine addition combined with low-temperature sludge storage and recovery strategy. The main conclusions are drawn as follows:

Single intermittent aeration is insufficient to maintain stable partial nitritation in low-ammonium wastewater. When combined with N₂H₄ dosing, the system can achieve stable operation within 24 days, with a nitrite accumulation rate of approximately 77%. Further reducing DO to 0.6 mg L⁻¹ can strengthen NOB inhibition and raise the nitrite accumulation rate to nearly 100%, accompanied by a directional succession of nitrifying functional bacteria.

Sludge stored at a constant 4 °C shows much better activity recovery performance than sludge stored under natural autumn-winter ambient temperature. Inoculation with 4 °C preserved sludge enables rapid startup of the low-ammonium partial nitritation system within 15 days, with ammonia conversion efficiency and nitrite accumulation rate stabilized at about 57% and 80% respectively, meeting the influent requirements of the subsequent anammox process. Microbial analysis also confirms that 4 °C storage is more conducive to AOB enrichment and long-term NOB suppression after recovery.

The partial nitritation system recovered from 4 °C stored sludge maintains stable nitrogen removal performance at 25–35 °C. When the temperature drops to 20 °C, NOB activity rebounds, and system stability declines, while supplementary N₂H₄ dosing can effectively restore stable operation. Functional gene prediction indicates that refrigerated sludge storage and recovery at 4 °C upregulates the expression of AMO functional genes, consolidating the dominant position of AOB at the microbial level.

Overall, the combined strategy of 4 °C refrigerated sludge storage and N₂H₄ regulation provides a feasible technical reference for rapid startup and stable operation of partial nitritation in low-ammonium municipal wastewater treatment, with faster startup and stronger long-term NOB suppression than single oxygen limitation, sole sludge storage or independent inhibitor dosing strategies proposed in prior literature.^10,15

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6ra02877b.