English

• Rapid social and economic development in recent years has significantly increased global consumption and production, leading to the increased generation of municipal solid waste (MSW). It has been estimated that by 2025, over 2.2 billion tons of MSW will be produced worldwide [1]. Sanitary landfill remains a widely used strategy for MSW disposal worldwide, particularly in most developed countries [2], due to its simple operational procedures and low treatment costs. Landfill leachate is released from landfill, which usually contains a large amount of refractory organic compounds, ammonia nitrogen, heavy metals, high salinity, and toxic contaminants [3]. Therefore, developing efficient and economical methods for the treatment of landfill leachate is an increasingly important research focus, with the aim of minimizing its hazardous effect on both human and environmental health.

  At present, physicochemical and biological technologies are the primary methods applied for the treatment of landfill leachate [4]. However, the practical application of most physicochemical methods is limited, such as membrane treatment and coagulation/flocculation, as they cannot completely remove pollutants due to the production of secondary pollution during treatment and often have high operational costs [5,6]. As an alternative approach, biological treatment has been widely applied for pollutants removal, providing a simple, efficient, and low-operational cost strategy, that is especially useful for nitrogen elimination. However, with increasing landfill age, biodegradable organics are gradually degraded, resulting in a low chemical oxygen demand (COD) and a low ratio of biological oxygen demand (BOD) to chemical oxygen demand in leachate. Typically, mature landfill leachate is classified as containing COD concentrations below 4000 mg/L and a BOD to COD ratio below 0.1 [5]. To achieve efficient nitrogen removal from mature landfill leachate using conventional biotechnology (nitrification and denitrification processes), large amounts of an external carbon source and a high level of aeration are needed, leading to higher levels of sludge production [7]. To overcome these challenges, the Anammox-based process has recently been developed, with this autotrophic biological denitrification technology providing a promising alternative approach, that can reduce operational costs and improve the overall treatment performance.

  A crucial prerequisite to the effective application of Anammox-based processes is to achieve stable partial nitritation (PN) and maintain an appropriate ratio of nitrite nitrogen to ammonia nitrogen (NO₂^–-N/NH₄⁺-N = 1.0–1.3) [8]. To achieve efficient PN, ammonium-oxidizing bacteria (AOB) must be retained within the system, while nitrite-oxidizing bacteria (NOB) must be suppressed or washed out. Previous studies have achieved PN by regulating operation parameters based on different growth characteristics of AOB and NOB, such as temperature, or the concentrations of dissolved oxygen (DO), free ammonia (FA), and free nitrous acid (FNA) [9]. Due to the typically high levels of ammonia nitrogen and alkalinity in landfill leachate, and the differences in inhibitory concentrations of FA and FNA on AOB and NOB, high FA and FNA are the commonly maintained to achieve PN. Chen et al. [10] achieved PN of mature landfill leachate using a zeolite biological aerated filter, with the inhibition of NOB maintained by a combination of FA and FNA, while the NO₂⁻-N/NH₄⁺-N ratio was kept at approximately 1.2. Another study used a sequencing batch reactor (SBR) to treat mature landfill leachate with FNA for NOB inhibition, and transformed almost all NH₄⁺-N to NO₂^–-N, achieving a nitrite accumulation rate (NAR) of > 90% [11]. Bruni et al. [12] achieved PN in a pilot-scale SBR for landfill leachate treatment, using FA to inhibit NOB, resulting in an ammonia removal efficiency of > 80%. Despite the promising results of these studies, either the addition of exogenous carriers was needed to control the NO₂⁻-N/NH₄⁺-N ratio, or NH₄⁺-N was provided from another pathway to achieve the Anammox process. Furthermore, the regulation of operational parameters such as FA and FNA to inhibit NOB and achieve PN, will inevitably have a negative impact on AOB activity, reducing the ammonia removal rate (ARR) of the system [13]. Another issue of concern is that NOB were recently found to be capable of adapting to operational conditions causing growth suppression, such as low DO concentrations, side-stream sludge treatment, or inhibition by FNA and FA, therefore disrupting the long-term stability of PN processes [14-17]. Therefore, this study aimed to develop a suitable strategy to achieve long-term stable PN without the use of additional agents, while also simultaneously regulating the ratio of NO₂⁻-N/NH₄⁺-N for Anammox. Most previous research has focused on the operational strategy used to realize PN. However, considering the intrinsic toxicity and varying characteristics of landfill leachate, it is also important to fully explore the microbial community composition, evolution and the interactions between functional microorganisms in the PN system for maintaining long-term stable conditions.

  Therefore, a continuous flow bioreactor was established and operated continuously for 300 days to achieve long-term PN of mature landfill leachate, with the bioreactor performance and microbial community comprehensively evaluated. The specific objectives of this study were to: (i) develop a suitable strategy to achieve PN and regulate the ratio of NO₂⁻-N/NH₄⁺-N for Anammox; (ii) systematically evaluate the PN performance during long-term operation and elucidate the PN mechanism; (iii) investigate the microbial community dynamics, identifying key functional microbes and their interactions.

  A laboratory-scale continuous flow bioreactor was used for experiments, with a total working volume of 60 L which was equally divided into three tanks, while the sludge was settled through an attached clarifier tank (13.25 L) (Fig. S1 in Supporting information). The raw landfill leachate was fed into the first anaerobic tank by a peristaltic pump, which was packed with braided polypropylene carriers, then flowed through the second anoxic tank, and finally into the oxic tank. Both the anoxic and oxic tanks were aerated, with the DO concentration in the anoxic and oxic tanks maintained at 0.10–0.35 mg/L and 0.20–1.50 mg/L, respectively.

  Mechanical stirrers were installed in the anoxic and oxic tanks, maintaining continual agitation at 70 rpm. The sludge retained in the clarifier tank was returned to the anoxic tank using a peristaltic pump, at a reflux ratio of 200%. Throughout the experimental period, the hydraulic retention time (HRT) was kept constant at 5.5 d, while the sludge retention time (SRT) was 15 d and the temperature was maintained at 28 ± 2 ℃ using an automatic temperature control device. The continuous flow bioreactor operational period was divided into two phases: the start-up phase (1–30 d, phase Ⅰ) and stable phase (31–300 d, phase Ⅱ). The mature landfill leachate was diluted with tap water, with the proportion of leachate in the influent of phases Ⅰ and Ⅱ being 25% and 50%, respectively.

  The seed sludge was collected from a secondary sedimentation tank of a two-stage anoxic-oxic (A/O) treatment system in the Wuhan Chenjiachong Landfill Treatment Engineering Project (Wuhan, China), which has been in operation for 15 years. After inoculation, the mixed liquor suspended solids (MLSS) concentration in the continuous flow bioreactor was approximately 4500 mg/L. Landfill leachate was also collected from the Wuhan Chenjiachong Landfill Treatment Engineering Project (Wuhan, China), and was found to conform to the typical characteristics of mature landfill leachate with a C/N ratio of 1.73 ± 0.52. The specific characteristics of the landfill leachate were as follows: COD = 5127.0 ± 1649.2 mg/L, ammonia nitrogen (NH₄⁺-N) = 2926.9 ± 252.2 mg/L, nitrate nitrogen (NO₃^–-N) = 32.8 ± 10.4 mg/L, nitrite nitrogen (NO₂^–-N) = 1.4 ± 1.1 mg/L, chloride ions (Cl⁻) = 5999.6 ± 938.2 mg/L, and pH 8.1–8.5. Furthermore, the five most abundant heavy metals were iron (3.328 ± 0.179 mg/L), arsenic (2.235 ± 0.060 mg/L), barium (0.290 ± 0.004 mg/L), zinc (0.275 ± 0.007 mg/L), and chromium (0.246 ± 0.002 mg/L).

  The concentrations of COD, NH₄⁺-N, NO₃^–-N, NO₂^–-N, and TN in samples were measured according to the standard methods, after filtration using a 0.45 µm filter. The pH and DO were measured using a pH meter (pHS-25, Leici, China) and a DO meter (HQ40D, Hach, US), respectively. The water samples were filtered through a 0.45 µm filter prior to three-dimensional excitation and emission matrix (3D-EEM) analysis using a fluorescence spectrophotometer (F-7000, Hitachi, Japan). The emission (Em) spectra were determined from 250 nm to 550 nm at 2 nm increments, while excitation (Ex) wavelengths from 200 nm to 400 nm were determined at 5 nm intervals. The emission and excitation spectra were recorded with slit widths of 5 nm, and at a scanning speed of 2400 nm/min. The final spectra were processed to remove the deionized water background signal.

  The NAR and the concentrations of FA and FNA were calculated according to Eqs. 1-3 as follows [18]:

  (1)

  (2)

  (3)

  where NO₂⁻−N_eff and NO₃⁻ −N_eff refer to the NO₂^–-N and NO₃^–-N concentrations in influent (mg/L); and T is the temperature of water within the bioreactor (℃).

  DNA was extracted from sludge samples using a PowerSoil DNA Kit (Mobio, US) following the manufacturer's instructions. PCR amplification of the V3-V4 region of bacterial 16S rRNA was performed using the 338F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′) primer set. The V4 region of archaeal 16S rRNA was amplified using the ArBa515F (5′-GTGCCAGCMGCCGCGGTAA-3′) and Arch806R (5′-GGACTACVSGGGTATCTAAT-3′) primer pair. The amplification products were sequenced using the Illumina MiSeq platform (Majorbio Bio–Pharm Technology Co. Ltd., Shanghai, China).

  The operational period of the continuous flow bioreactor was divided into two phases. In the start-up stage (phase Ⅰ), NH₄⁺-N concentrations in each tank and in the final effluent were higher than in the influent (Fig. 1a), which was attributed to the liberation of ammonia nitrogen from cell lysis due to the shift in seed sludge from oxygen- and glucose-rich conditions to low DO conditions without the addition of organics [8]. After 14 days of operation, the NH₄⁺-N concentration in the anoxic tank began to decrease gradually, and by the 30^th day, the NH₄⁺-N removal efficiency had reached 49.54%, with a NAR of 100% (Fig. 1b). This suggests that AOB adapted to this environment and effectively became the dominant bacteria in the nitrogen transformation process, signifying the successful start-up of PN within the system.

  Figure 1

  

  Figure 1.  The long-term performance of the continuous flow bioreactor including (a) NH₄⁺-N concentrations in the influent, effluent and each tank and the NH₄⁺-N removal efficiency (ARE) of the bioreactor; (b) NO₂⁻-N concentrations in the influent, effluent and each tank, and variety of NAR and NO₂⁻-N/NH₄⁺-N ratio of the bioreactor; (c) COD concentrations in the influent, effluent and each tank, COD removal efficiency, and TN removal efficiency of the bioreactor.

  DownLoad: Full-Size Img PowerPoint

  In phase Ⅱ, the proportion of landfill leachate in influent was increased to 50%, and the effluent concentrations of NH₄⁺-N and NO₂⁻-N were 527.15 ± 88.10 and 602.70 ± 85.33 mg/L, respectively. The ratio of NO₂⁻-N/NH₄⁺-N was 1.16 ± 0.19 and the NAR was 94.43% ± 1.80%. Therefore, the stable performance of the system and the appropriate substrate ratio provided suitable conditions for the subsequent Anammox process. In previous studies, Nhat et al. [13] achieved PN in a lab-scale SBR fed with mature municipal landfill leachate after 75 days of operation, while Li et al. [19] reported that a 65 days start-up period was required to achieve PN and a satisfactory NO₂⁻-N/NH₄⁺-N ratio (1.20–1.28) in a SBR with intermittent aeration and endpoint pH control. In contrast, Bruni et al. [12] failed to establish nitritation due to an excessive nitrogen loading rate. Therefore, the results of the present study show that this system achieved rapid start-up of PN within 30 days in a simple operational procedure, with a comparable treatment performance to previously reported studies. Compared with previous studies, the rapid start-up of the PN system in this study was attributed to the removal of most organics in anaerobic tank and the relatively low DO concentration in oxic tank, which could inhibit the proliferation of heterotrophic bacteria and provide favorable conditions for the growth and activity of AOB.

  The concentration of NO₂⁻-N in the anoxic tank (575.4 ± 88.6 mg/L) was slightly lower than in the oxic tank (623.7 ± 85.6 mg/L) (Fig. 1b), which signified the existence of denitrification in anoxic tank and the further oxidation of NH₄⁺-N in oxic tank. The oxic tank was established to further oxidize NH₄⁺-N and maintain a suitable NO₂⁻-N/NH₄⁺-N ratio, as denitrification existed in the anoxic tank consumes a portion of NO₂⁻-N, causing the ratio of NO₂⁻-N/NH₄⁺-N in the anoxic tank to be outside of the range suitable for anammox bacteria (1–1.32). Furthermore, it was found that the removal of TN was closely related to the concentration of organics in leachate (Fig. 1c). During 195–245 days of operation, the relatively higher influent concentration of COD (3927 ± 247 mg/L) resulted in a higher corresponding TN removal efficiency of 37% ± 6%. However, in most cases, the concentration of COD in mature landfill leachate has been low (1910.10 ± 133.51 mg/L), with a corresponding average TN removal efficiency of only 18%.

  In phase Ⅰ, the COD concentration in the effluent was higher overall than in the influent (Fig. 1c), which was attributed to cell lysis. In phase Ⅱ, from day 42 to 194, the average COD concentrations in the influent and effluent were 1910.10 ± 133.51 mg/L and 1236.29 ± 146.24 mg/L, respectively, corresponding to an average COD removal efficiency of 35.16%. Due to the significant increase in the COD of landfill leachate on day 195, the average proportion of COD degraded in the anaerobic tank increased from 8.26% to 36.35%, and the average COD removal efficiency of the bioreactor increased to 60.55%. However, this increase in COD removal efficiency did not reduce the concentration of COD in the effluent (1198.47 ± 106.07 mg/L), which remained at a similar level to the previous stage. These results show that the residual COD content of effluent was mainly composed of refractory organics, as reported in previous literature [6,20], showing that the large proportion of refractory organics in mature landfill leachate present a major challenge for the efficient removal of COD.

  Dissolved organic matter (DOM) accounts for most of the organic matter in leachate and is strongly associated with the COD of the leachate. The compositional distribution of DOM and its transformation, were analyzed based on 3D-EEM spectra (Fig. 2). The fluorescent components of the raw leachate on the 80^th day were identified as protein tryptophan-like substances (peaks A and B at the Ex/Em of 230/344 nm and 275/352 nm, respectively) and fulvic-like compounds (peak C at the Ex/Em of 250/450 nm) [21]. On the 280^th day, the raw leachate consisted of humic-like substances (peak D at the Ex/Em of 325/410 nm) and fulvic-like compounds (peak C at the Ex/Em of 250/450 nm). It is of note, that fulvic-like and humic-like compounds gradually became dominant during the aging process, indicating the low biodegradability of raw leachate.

  Figure 2

  

  Figure 2.  3D-EEM spectra of the influent and effluent of the continuous flow bioreactor and the effluent of each tank on the 80^th and 280^th day: (a) influent, (b) anaerobic, (c) anoxic, (d) oxic and (e) effluent on the 80^th day; (f) influent, (g) anaerobic, (h) anoxic, (i) oxic and (j) effluent on the 280^th day.

  DownLoad: Full-Size Img PowerPoint

  Compared to the raw leachate, the fluorescence intensity of the four indicator peaks (A, B, C and D) all decreased slightly in the anaerobic tank, then exhibited a significant decrease in the anoxic tank, with peak A in particular completely degraded. In general, these substances (peak A) were identified as being simple aromatic proteins with Ex < 250 nm and Em < 350 nm, indicating the presence of bioavailable organic substrates [6].

  Some organic matter was utilized as a carbon source for denitrification by denitrifying bacteria and the growth of other heterotrophic bacteria and therefore, the fluorescence intensity of tryptophan-like protein, fulvic-like and humic-like substances were all greatly reduced in the anoxic tank. However, the intensity of the peak for these substances in the oxic tank did not change significantly, indicating the low biodegradability of the remaining organics. These results indicate that some proteins, fulvic-like and humic-like compounds were degraded after the treatment using the continuous flow bioreactor. However, the removal of residual organic matter may require some additional chemical methods, such as advanced oxidation processes (AOPs).

  The seed sludge sample (S0) was collected on day 0 and sludge samples were collected from the anaerobic, anoxic and oxic tanks on days 30 (A1_30d, A2_30d, O3_30d) and 280 (A1_280d, A2_280d, O3_280d). The Alpha-diversity indices of the seven samples were determined, including diversity indices, richness indices and the coverage index, as shown in Table S1 (Supporting information). The coverage index of all samples was > 0.99, indicating a sufficient sequencing depth and that data was representative of the extracted samples. Shannon and Simpson indices showed that microbial diversity increased in phase Ⅰ compared to the seed sludge, then further increased during phase Ⅱ in the anaerobic and anoxic tanks, then slightly decreased in the oxic tank. The microbial richness of the anaerobic tank increased initially and then decreased, while the opposite trend was observed in the anoxic and oxic tanks. These results were mainly attributed to the gradual enrichment of microorganisms able to adapt to the system during phase Ⅰ and the increased abundance of microorganisms able to tolerate low organic matter and DO concentrations and high FA and FNA concentrations. Principal coordinate analysis (PCoA) was performed to investigate changes in microbial community structure within the system (Fig. 3a). The identified sample clusters included A1_30d and A1_280d, A2_30d and O3_30d, A2_280d and O3_280d, all of which were separate from S0, indicating that the microbial community structure altered significantly with treatment time and varying conditions.

  Figure 3

  

  Figure 3.  (a) PCoA analysis based on bacteria community; variations of microbial community structure (b) at phylum level, and (c) at genus level; the genera correlation network analysis of (d) anaerobic tank and (e) oxic tank. The red line represents cooperation relationship and the green line represents competition relationship.

  DownLoad: Full-Size Img PowerPoint

  To further explore the microbial composition and the evolution of the sludge microbial community, the community structure and relative microbial abundances were analyzed at the phylum and genus levels. As shown in Fig. 3b, Proteobacteria and Chloroflexi were the predominant phyla, with relative abundances of 11.65%−43.40% and 9.43%−37.80%, respectively. Proteobacteria play an important role in the removal of both organic matter and nitrogen and have previously been reported to be the most abundant phylum in wastewater treatment systems [22]. Chloroflexi is facultative anaerobic bacteria that is capable of aromatic compound degradation and has commonly been identified in wastewater treatment systems [20], which might relate to the degradation of fulvic-like and humic-like compounds in the system. The relative abundance of Bacteroidota in the anoxic and oxic tanks was approximately two-fold higher than in the anaerobic tank, while the abundance of Firmicutes in the anaerobic tank was about 6 to 11-fold higher than in the anoxic and oxic tanks. Bacteroidota has been shown to be capable of hydrolyzing complex macromolecular organic compounds and can decompose carbohydrates into monosaccharides [23], potentially enhancing nitrogen removal from mature landfill leachate by resolving the problem of carbon source limitation. Firmicutes can survive in extremely oligotrophic environments due to its ability to produce endospores, and have been found to be an important acid hydrolyzing microbe that can degrade volatile fatty acids (VFAs) and generate acetic acid [23-25]. Actinobacteriota was also detected within the system, which has been reported to include both anaerobic and autotrophic bacteria [22]. Furthermore, many Firmicutes and Actinobacteriota have been reported to be capable of converting organic compounds (such as proteins and carbohydrates) to short-chain fatty acids (SCFAs) and hydrogen under anaerobic conditions [26]. Synergistota and Caldatribacteriota were the exclusive phyla in the anaerobic tank, with their relative abundances increasing from 2.29% and 0.29% on the 30^th day to 8.00% and 9.94% on the 280^th day, respectively. Some anaerobic bacteria affiliated with Synergistota have been confirmed to be syntrophic with methanogens, degrading organic acids or alcohols, with Caldatribacteriota also shown to have a syntrophic relationship with methanogens [27].

  The microbial community was further investigated at the genus level to determine and compare the abundance of functional bacteria (Fig. 3c). The preponderant genera in the seeding sludge were Defluviicoccus and norank_f__norank_o__SBR1031 (an unclassified genus of the order-level lineage SBR1031, phylum Chloroflexi), with abundances of 31.90% and 25.83%, respectively. Defluviicoccus is a glycogen accumulating organism that is capable of degrading various kinds of organics [28] and its high abundance in the seeding sludge may be due to the addition of glucose as a carbon source in the two-staged A/O landfill treatment process. Nitrogenous compounds could provide satisfactory nutritional conditions for norank_f__norank_o__SBR1031 [29], with the high concentration of ammonia nitrogen in landfill leachate stimulating its extensive enrichment. The relative abundances of Defluviicoccus and norank_f__norank_o__SBR1031 in the samples collected on the 30^th and 280^th day accounted for 1.61%−17.55% and 2.39%−11.00%, respectively, and it was speculated that the degradation of organic matter and ammonia nitrogen in the continuous flow bioreactor system was related to these genera.

  As the composition of organic matter in landfill leachate is complex, a large number of microorganisms capable of degrading organic matter were enriched in the anaerobic tank. On the 30^th day, Raineyella, unclassified_o__Micrococcales (an unclassified genus of the order-level lineage Micrococcales, phylum Actinobacteriota), Thauera, Chryseolinea, and Brooklawnia were the main genera associated with the degradation and transformation of organic matter in the anaerobic tank, accounting for 14.77%, 7.58%, 6.76%, 3.87%, and 3.65%, respectively. Raineyella was the most abundant microorganism in the anaerobic tank, which has been reported to dominantly coexist with other organic degrading facultative anaerobes in constructed wetlands for mine drainage treatment [30]. It has been proposed that Raineyella may play an important role in the transformation and degradation of organic matter in leachate, while the genera unclassified_o__Micrococcales, Thauera and Chryseolinea were found to possess the capacity to degrade refractory compounds, aromatic compounds, polycyclic aromatic hydrocarbons and cellulose [31-34], respectively. Brooklawnia are able to produce SCFAs and hydrogen during the anaerobic fermentation of organic matter, degrading glucose to produce propionate [26]. By day 280, a significant shift was observed in the microbial genera in the anaerobic tank, with the genus norank_f__norank_o__norank_c__JS1 (an unclassified genus of the class-level lineage JS1) of the phylum Caldatribacteriota, exhibiting the highest abundance (8.89%). Caldatribacteriota are heterotrophic anaerobes with fermentative potential and have been found to be syntrophic with methanogens by providing substrates for methanogenesis [27]. Thus, it is feasible to assume that the dominance of the genus norank_f__norank_o__norank_c__JS1 was relevant to, and perhaps vital to the anaerobic fermentation of organic matter for methane production. Fastidiosipila and Syntrophaceticus belong to Firmicutes and accounted for 7.37% and 3.75% of the community in the anaerobic tank, respectively. Fastidiosipila are able to transform carbohydrates and proteins to VFAs [35], while Syntrophaceticus are important for acetate metabolism and are syntrophic with methanogenic archaea [36]. Moreover, Thermovirga also exhibited a high abundance of 6.71%, which was found to be involved in syntrophic methanogenesis in anaerobic environments and was responsible for the degradation of carbohydrates, proteins, amino acids, and organic acids [37,38].

  Methanogenic archaea are strictly anaerobic, with energy metabolism limited to the formation of methane from CO₂ and H₂, methanol, formate, methylamines and/or acetate [39]. The abundance and activity of methanogens are closely related to the transformation and degradation of organic matter in the system. The methanogenic archaeal community in the anaerobic tank could be mainly categorized as acetoclastic, hydrogenotrophic, or methylotrophic methanogens (Table S2 in Supporting information). It is of note, that with ongoing operation of the bioreactor, the abundance of all methanogens increased. Methanosaeta were the most abundant methanogens in the system, accounting for 15.58% of all archaea, which is an acetoclastic methanogen that uses acetate for methane production [40]. Methanosarcina (1.01%) is an acetoclastic and hydrogenotrophic methanogen, capable of utilizing H₂ or acetate to produce methane [41]. There were also some hydrogenotrophic and methylotrophic methanogens identified with relative low abundances, such as Methanoculleus, Candidatus_Methanofastidiosum, and Methanomassiliicoccus. Overall, these results suggest that the utilization of acetate was the main pathway of methanogenesis in the anaerobic tank. However, no acetate was directly available in the mature landfill leachate, so the enrichment of acetoclastic methanogens mainly depended on their association with syntrophic and acetogenic bacteria (e.g., norank_f__norank_o__norank_c__JS1, Fastidiosipila, Syntrophaceticus, and Thermovirga).

  Overall, the dominant bacterial population in the anaerobic tank was gradually transformed from bacteria degrading organic matter to bacteria that are syntrophic with methanogens. This cooperation between multiple microbial groups, such as organic degradation bacteria, syntrophic bacteria, acetoclastic methanogens, and hydrogenotrophic methanogens, ensured the degradation or methanogenic utilization of organic matter in the anaerobic tank.

  In the anoxic and oxic tanks, the genera Limnobacter, Truepera, and norank_f__PHOS-HE36 (an unclassified genus of the family-level lineage PHOS-HE36, phylum Bacteroidota) were consistently present at a relatively high abundance, accounting for 10.71%−12.88%, 4.05%−6.46%, and 4.90%−7.43%, respectively. Limnobacter have been shown to play an important role in the decomposition of organic matter and denitrification, possibly by decomposing refractory organics into small molecular organics that are usable by other microorganisms [42]. Truepera is an important thermophilic denitrifier, which has salt-resistance and has been identified in some extreme environments, including municipal solid waste landfill [38,43,44]. In addition, norank_f__PHOS-HE36 belongs to the phylum Bacteroidota, which can convert organic matter (such as proteins and polysaccharides) to small organic molecules [45]. It is of note, that the abundances of the genera Ottowia, Phaeodactylibacter, Nitrosomonas, norank_f__norank_o__norank_c__JG30-KF-CM66 (an unclassified genus of the class-level lineage JG30-KF-CM66, phylum Chloroflexi), and norank_f__67–14 (an unclassified genus of the family-level lineage 67–14, phylum Actinobacteriota) increased significantly after 280 days of operation. Ottowia has been widely detected in biological treatment systems for industrial wastewater, due to its capacity to perform hydrolysis and denitrification [46]. Ottowia had an abundance of 8.64% in the anoxic tank and 9.59% in the oxic tank, showing that it is capable of both organic matter degradation and denitrification in this continuous flow system under low DO concentration conditions. Phaeodactylibacter accounted for 7.80% in the anoxic tank and 11.22% in the oxic tank, with these genera previously proposed to be involved in the heterotrophic nitrosification process [47], suggesting it is credible that it may participate in the NH₄⁺-N oxidation process. Nitrosomonas were found to have an abundance of 5.14% in the anoxic tank and 6.38% in the oxic tank, with these AOB commonly detected in wastewater treatment systems and responsible for converting NH₄⁺-N to NO₂⁻-N [16]. It was identified that the genus norank_f__norank_o__norank_c__JG30-KF-CM66 belongs to the phylum Chloroflexi, which are able to assimilate acetate [24], indicating that this genus may significantly contribute to the degradation of organic matter within the system.

  Besides, there were also some genera of the phylum Chloroflexi found to be present at low abundance (0.15%−2.97%) which were highly salt-tolerant, such as norank_f__JG30-KF-CM45, norank_f__Caldilineaceae, and norank_f__AKYG1722 (unclassified genera of the family-level lineage JG30-KF-CM45, Caldilineaceae, and AKYG1722, respectively) [48]. These genera may be important to maintaining systematic stability, which was consistent with the high salinity of landfill leachate.

  Microbial interaction networks are used to reveal the competitive and cooperative relationships between two different genera. Correlation network analysis results for the anaerobic and oxic tanks are illustrated in Figs. 3d and e, respectively. The genera in these two tanks were divided into two competing groups, which were labelled as group 1 and group 2 in the anaerobic tank, and group 3 and group 4 in the oxic tank. The three major methanogenic archaea (Methanosaeta, Methanosarcina, and Methanoculleus) exhibited obvious cooperative relationships with syntrophic bacteria (Syntrophaceticus, Thermovirga, and Fastidiosipila), while distinct competitive relationships were observed with organic matter degrading bacteria (Raineyella, Defluviicoccus, and Brooklawnia). These results were consistent with the observed succession of microbial communities. The organic matter degrading bacteria that usually appear in nitritation-denitrification systems treating landfill leachate (e.g., Limnobacter, Ottowia, norank_f__norank_o__norank_c__JG30-KF-CM66, and norank_f__PHOS-HE36) [11] exhibited a large number of collaborations with Nitrosomonas in the oxic tank, which may be crucial to the predominance and activity of Nitrosomonas by partly reducing the inhibition of organics [18]. Simultaneously, synergistic relationships were also identified between denitrifying bacteria (Limnobacter, Truepera, and Ottowia) and Nitrosomonas. Microorganisms associated with NH₄⁺-N oxidization, denitrification and COD degradation can not only coexist while simultaneously serving their respective functions, but they can also be synergistic. Specifically, the NO₂⁻-N or NO₃⁻-N produced during NH₄⁺-N oxidization could provide metabolic substrates for denitrification. Denitrification lessens the concentration of NO₂⁻-N and also causes an increase in pH due to the release of alkalinity, resulting in a decrease in FNA, which further reduced the inhibition of microorganisms within the system. Meanwhile, the utilization of organics by organic matter degrading bacteria and denitrifying bacteria reduced the risk of AOB inhibition by more heterotrophic bacteria. These beneficial effects would enhance the PN performance of the system and help form a more stable microbial community structure.

  In this study, a continuous flow bioreactor with anaerobic-anoxic-oxic processes was applied to achieve the partial nitritation of mature landfill leachate. The start-up of this system required only 30 days and then maintained stable operation for 270 days, with the ratio of nitrite nitrogen to ammonia nitrogen remaining stable at 1–1.32, indicating excellent partial nitritation performance. As shown in Fig. 4, during phase Ⅱ the average FA concentrations in the anoxic and oxic tanks were 55.87 mg N L^–1 and 31.47 mg N L^–1, respectively, while the average FNA concentrations in the anoxic and oxic tanks were 0.068 mg N L^–1 and 0.548 mg N L^–1, respectively. It has previously been reported that the inhibitory concentrations of FA on AOB and NOB were about 10–150 mg N L^–1 and 0.1–1.0 mg N L^–1, respectively [13,18], while the inhibitory concentrations of FNA on AOB and NOB were 0.42–1.72 mg N L^–1 and 0.026–0.22 mg N L^–1, respectively [10,49].

  Figure 4

  

  Figure 4.  Variations of FNA and FA concentrations in the anoxic and oxic tanks. The red line represents the inhibitory threshold of FA on NOB and the blue line represents the inhibitory threshold of FNA on NOB.

  DownLoad: Full-Size Img PowerPoint

  Therefore, the concentrations of FA and FNA in the present system induced a minimal level of AOB growth inhibition, while NOB were effectively restrained by the combination of FA and FNA. Due to the high ammonia nitrogen concentration and alkalinity of the mature landfill leachate, there was a relatively high concentration of FA in the anoxic tank to inhibit NOB. However, the degradation of ammonia nitrogen and the decreased pH in the oxic tank resulted in FA concentrations being below the threshold inhibitory value, while FNA concentration increased accordingly and achieved sufficient NOB suppression. Therefore, stable partial nitritation was achieved in this system via the alternate control of FA and FNA in different stages.

  DO levels were kept relatively low (0.10–0.35 mg/L in the anoxic tank and 0.20–1.50 mg/L in the oxic tank) in order to prevent excessive ammonia nitrogen from being oxidized and thereby maintaining a specific ratio of nitrite nitrogen to ammonia nitrogen in the effluent. However, DO limitation can also inhibit the activity of AOB due to oxygen competition from heterotrophic organisms, resulting in a low NAR [50]. Therefore, an anaerobic tank was included as the first step, preliminarily removing excess organic matter, reducing the negative impact of organic matter on autotrophic microorganisms. Consequently, long-term stable partial nitritation was achieved under low DO conditions without an adverse impact on AOB.

  Long-term stable partial nitritation of mature landfill leachate was achieved using a continuous flow bioreactor, through in-situ inhibition by a combination of FA and FNA. The NAR was maintained at around 94.43% for 270 days, with an average NO₂⁻-N/NH₄⁺-N ratio of about 1.16. Microbial analysis showed that the inhibition of organics on AOB (Nitrosomonas) was alleviated during ongoing operation, even under relatively low DO conditions, due to the removal of organic matter through the cooperative actions of organic degradation bacteria, syntrophic bacteria, and methanogens. These collaborative interactions between different functional microorganisms further facilitated the predominance and high activity of Nitrosomonas, which was vital to the maintenance of an appropriate NO₂⁻-N/NH₄⁺-N ratio and stable partial nitritation within the system.

  Declaration of competing 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.

  Acknowledgment

  This work was financially supported by the National Natural Science Foundation of China (No. 52170049).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.108284.

  
  
• 1. [1]

     P. Wijekoon, P.A. Koliyabandara, A.T. Cooray, et al., J. Hazard. Mater. 421 (2022) 126627. doi: 10.1016/j.jhazmat.2021.126627

  2. [2]

     G. Chen, G. Wu, N. Li, et al., J. Hazard. Mater. 418 (2021) 126355. doi: 10.1016/j.jhazmat.2021.126355

  3. [3]

     M.A.M. Reshadi, A. Bazargan, G. McKay, Sci. Total Environ. 731 (2020) 138863. doi: 10.1016/j.scitotenv.2020.138863

  4. [4]

     K. Luo, Y. Pang, X. Li, et al., Environ. Technol. 40 (2019) 1862–1870. doi: 10.1080/09593330.2018.1432694

  5. [5]

     L. Miao, G. Yang, T. Tao, Y. Peng, J. Environ. Manage. 235 (2019) 178–185. doi: 10.1016/j.jenvman.2019.01.057

  6. [6]

     C. Teng, K. Zhou, C. Peng, W. Chen, Water Res. 203 (2021) 117525. doi: 10.1016/j.watres.2021.117525

  7. [7]

     Y. Yang, A. Ricoveri, K. Demeestere, S. Van Hulle, J. Hazard. Mater. 430 (2022) 128481. doi: 10.1016/j.jhazmat.2022.128481

  8. [8]

     C. Jiang, X. Tang, F. Feng, et al., Sci. Total Environ. 817 (2022) 152994. doi: 10.1016/j.scitotenv.2022.152994

  9. [9]

     B. Ma, S. Wang, S. Cao, et al., Bioresour. Technol. 200 (2016) 981–990. doi: 10.1016/j.biortech.2015.10.074

  10. [10]

      X. Chen, X. Wang, Z. Zhong, et al., Environ. Sci. Pollut. Res. 27 (2020) 29408–29421. doi: 10.1007/s11356-020-09185-2

  11. [11]

      S. Ren, Z. Wang, H. Jiang, et al., Bioresour. Technol. 340 (2021) 125647. doi: 10.1016/j.biortech.2021.125647

  12. [12]

      C. Bruni, G. Cipolletta, Ç. Akyol, A.L. Eusebi, F. Fatone, Environ. Technol. Innov. 27 (2022) 102415. doi: 10.1016/j.eti.2022.102415

  13. [13]

      P.T. Nhat, H.N. Biec, T.T.T. Van, et al., Int. Biodeterior. Biodegrad. 124 (2017) 56–61. doi: 10.1016/j.ibiod.2017.06.017

  14. [14]

      H. Duan, L. Ye, X. Lu, Z. Yuan, Environ. Sci. Technol. 53 (2019) 1937–1946. doi: 10.1021/acs.est.8b06148

  15. [15]

      Y. Cao, B.H. Kwok, M.C.M. van Loosdrecht, et al., Water Sci. Technol. 78 (2018) 634–643. doi: 10.2166/wst.2018.333

  16. [16]

      B. Ma, L. Yang, Q. Wang, et al., Bioresour. Technol. 245 (2017) 1266–1270. doi: 10.1016/j.biortech.2017.08.074

  17. [17]

      F. Zhang, H. Yang, J. Wang, Z. Liu, Q. Guan, RSC Adv. 8 (2018) 31987–31995. doi: 10.1039/c8ra06198j

  18. [18]

      R. Wang, X. Wang, C. Deng, et al., Bioresour. Technol. 305 (2020) 123031. doi: 10.1016/j.biortech.2020.123031

  19. [19]

      H. Li, S. Zhou, G. Huang, B. Xu, Process Saf. Environ. Prot. 92 (2014) 199–205. doi: 10.5958/2229-4473.2014.00032.9

  20. [20]

      Y. Wang, Z. Lin, L. He, et al., Bioresour. Technol. 294 (2019) 122166. doi: 10.1016/j.biortech.2019.122166

  21. [21]

      L. Wang, Q. Yang, D. Wang, et al., J. Hazard. Mater. 318 (2016) 460–467. doi: 10.1016/j.jhazmat.2016.07.033

  22. [22]

      H. Xiao, Y. Peng, Q. Zhang, Y. Liu, Bioresour. Technol. 337 (2021) 125434. doi: 10.1016/j.biortech.2021.125434

  23. [23]

      S. Ma, H. Wang, B. Wang, X. Gu, W. Zhu, Bioresour. Technol. 344 (2022) 126177. doi: 10.1016/j.biortech.2021.126177

  24. [24]

      S. Suominen, D.M. van Vliet, I. Sanchez-Andrea, et al., Front. Microbiol. 12 (2021) 628301. doi: 10.3389/fmicb.2021.628301

  25. [25]

      L. Feng, W. Zhao, Y. Liu, et al., Chin. Chem. Lett. 34 (2023) 107439. doi: 10.1016/j.cclet.2022.04.037

  26. [26]

      Y.T. Zhang, W. Wei, B.J. Ni, J. Clean. Prod. 328 (2021) 129580. doi: 10.1016/j.jclepro.2021.129580

  27. [27]

      T.R. Iasakov, T.A. Kanapatskiy, S.V. Toshchakov, et al., Mar. Environ. Res. 173 (2022) 105533. doi: 10.1016/j.marenvres.2021.105533

  28. [28]

      X. Song, D. Yu, Y. Qiu, et al., Bioresour. Technol. 345 (2022) 126540. doi: 10.1016/j.biortech.2021.126540

  29. [29]

      D. Hua, Q. Fan, Y. Zhao, et al., Sci. Total Environ. 739 (2020) 139943. doi: 10.1016/j.scitotenv.2020.139943

  30. [30]

      J. Chen, X. Li, W. Jia, et al., J. Hazard. Mater. 404 (2021) 124125. doi: 10.1016/j.jhazmat.2020.124125

  31. [31]

      S. Liu, Y. Deng, Z. Jiang, et al., Sci. Total Environ. 740 (2020) 140185. doi: 10.1016/j.scitotenv.2020.140185

  32. [32]

      Y. Gao, L. Guo, C. Jin, et al., Water Res. 215 (2022) 118256. doi: 10.1016/j.watres.2022.118256

  33. [33]

      Y. Mao, X. Zhang, X. Xia, H. Zhong, L. Zhao, J. Ind. Microbiol. Biotechnol. 37 (2010) 927–934. doi: 10.1007/s10295-010-0740-7

  34. [34]

      Y. Fan, X. Chen, Z. Yao, et al., Sci. Total Environ. 770 (2021) 145205. doi: 10.1016/j.scitotenv.2021.145205

  35. [35]

      Z. Xie, Z. Wang, Q. Wang, C. Zhu, Z. Wu, Bioresour. Technol. 161 (2014) 29–39. doi: 10.1016/j.biortech.2014.03.014

  36. [36]

      M. Yan, L. Treu, S. Campanaro, et al., Chem. Eng. J. 401 (2020) 126159. doi: 10.1016/j.cej.2020.126159

  37. [37]

      S. Yang, Z. Chen, Q. Wen, Bioresour. Technol. 324 (2021) 124679. doi: 10.1016/j.biortech.2021.124679

  38. [38]

      X. Wang, A. Cao, G. Zhao, C. Zhou, R. Xu, Waste Manage. 66 (2017) 79–87. doi: 10.1016/j.wasman.2017.04.023

  39. [39]

      R.K. Thauer, A.K. Kaster, H. Seedorf, W. Buckel, R. Hedderich, Nat. Rev. Microbiol. 6 (2008) 579–591. doi: 10.1038/nrmicro1931

  40. [40]

      S.M. Patil, M.B. Kurade, B. Basak, et al., Bioresour. Technol. 332 (2021) 125123. doi: 10.1016/j.biortech.2021.125123

  41. [41]

      C.K. Roy, S. Toya, Y. Hoshiko, et al., J. Environ. Chem. Eng. 10 (2022) 107524. doi: 10.1016/j.jece.2022.107524

  42. [42]

      M. Fan, Y. Lin, H. Huo, et al., Water Res. 96 (2016) 198–207. doi: 10.1016/j.watres.2016.03.061

  43. [43]

      J. Song, W. Zhang, J. Gao, et al., Bioresour. Technol. 296 (2020) 122344. doi: 10.1016/j.biortech.2019.122344

  44. [44]

      S. Shi, Z. Lin, J. Zhou, et al., Bioresour. Technol. 344 (2022) 126190. doi: 10.1016/j.biortech.2021.126190

  45. [45]

      J. Ma, K. Wang, C. Shi, et al., Bioresour. Technol. 346 (2022) 126525. doi: 10.1016/j.biortech.2021.126525

  46. [46]

      J. Li, Q. Du, H. Peng, et al., J. Clean. Prod. 266 (2020) 121809. doi: 10.1016/j.jclepro.2020.121809

  47. [47]

      K. Ma, X. Li, L. Bao, X. Li, Y. Cui, J. Clean. Prod. 276 (2020) 124190. doi: 10.1016/j.jclepro.2020.124190

  48. [48]

      A. Shitu, G. Liu, Y. Zhang, et al., J. Environ. Manage. 291 (2021) 112724. doi: 10.1016/j.jenvman.2021.112724

  49. [49]

      X. Liu, M. Kim, G. Nakhla, M. Andalib, Y. Fang, J. Environ. Chem. Eng. 8 (2020) 103984. doi: 10.1016/j.jece.2020.103984

  50. [50]

      H. Cui, L. Zhang, Y. Peng, Q. Zhang, X. Li, J. Water Process. Eng. 46 (2022) 102589. doi: 10.1016/j.jwpe.2022.102589

• 

  Figure 1  The long-term performance of the continuous flow bioreactor including (a) NH₄⁺-N concentrations in the influent, effluent and each tank and the NH₄⁺-N removal efficiency (ARE) of the bioreactor; (b) NO₂⁻-N concentrations in the influent, effluent and each tank, and variety of NAR and NO₂⁻-N/NH₄⁺-N ratio of the bioreactor; (c) COD concentrations in the influent, effluent and each tank, COD removal efficiency, and TN removal efficiency of the bioreactor.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  3D-EEM spectra of the influent and effluent of the continuous flow bioreactor and the effluent of each tank on the 80^th and 280^th day: (a) influent, (b) anaerobic, (c) anoxic, (d) oxic and (e) effluent on the 80^th day; (f) influent, (g) anaerobic, (h) anoxic, (i) oxic and (j) effluent on the 280^th day.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) PCoA analysis based on bacteria community; variations of microbial community structure (b) at phylum level, and (c) at genus level; the genera correlation network analysis of (d) anaerobic tank and (e) oxic tank. The red line represents cooperation relationship and the green line represents competition relationship.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Variations of FNA and FA concentrations in the anoxic and oxic tanks. The red line represents the inhibitory threshold of FA on NOB and the blue line represents the inhibitory threshold of FNA on NOB.

  下载: 全尺寸图片 幻灯片
•