Complete Nitrogen Removal through Integrating Anammox and Autotrophic Denitrific

Guo Yadong; Bai Xue; Yu Haitong; Li Xuechen; Zhao Chaocheng; Liu Chunshuang

(1. College of New Energy, China University of Petroleum, Qingdao 266580;2. College of Chemical Engineering, China University of Petroleum, Qingdao 266580)

Abstract: Complete nitrogen removal was achieved through integrating anammox and autotrophic denitrification in an UASB reactor. The total nitrogen (TN) removal rate increased stepwise from 0.46 to 0.94 kg-N/(m3·d), with an effluent TN concentration of below 3.0 mg-N/L achieved. The process is relatively insensitive to the nitrite to ammonium ratio,achieving complete nitrogen removal when their ratio in the influent varied in the range of 1.35—1.55. The added S0 quantity in the system could be utilized to adjust the competition between autotrophic denitrifiers and anammox bacteria.High-throughput sequencing technology indicated that Candidatus_Kuenenia and Thiobacillus were the functional strains for anammox and autotrphic denitrification process, respectively, in the studied reactor. This result provides a theoretical and technical basis for the large-scale application of anaerobic ammonium oxidation process.

Key words: complete nitrogen removal; anammox; elemental sulfur; UASB

1 Introduction

Nitrogen compounds coming from increasing discharge of municipal and industrial wastewater have caused a series of environmental problems, including eutrophication and biodiversity loss. Biological nitrogen removal is a critical and economic method for solving the nitrogen pollution problems. Compared with the conventional nitrification-denitrification process, anaerobic ammonium oxidation (anammox)can directly combine ammonium and nitrite species to form nitrogen gas[1]. The combination of ammonium oxidizing bacteria (AOB) and anammox has significant advantages, such as saving 60% of aeration and 100% of organic carbon, and reducing 90% of sludge production[2-3]. This approach could drive wastewater treatment plants (WWTPs) from energy consumption condition to energy neutrality state. However, the anammox process requires a strict nitrite to ammonium ratio of 1.32 and could only achieve 89% of total nitrogen removal, since 20% of the influent nitrite would be converted to nitrate (Reaction 1). In actual operation, the ratio of nitrite to ammonium may deviate significantly from the desired ratio, which often results in a nitrogen removal efficiency of about 70%.

The sulfur based autotrophic denitrification can reduce nitrate to nitrite or nitrogen gas while oxidizing elemental sulfur or reducing sulfur compounds (S2-, S2O32-, SO32-)[4-7].Contrary to heterotrophic denitrification, the sulfur based autotrophic denitrification does not need organic compounds[8-11], thus it can be applied in many nitrate removal environments that lack organic carbon resources,such as ground water, secondary effluent of sewage treatment plants, and aquaculture waters. The performance and key parameters for sulfur-based autotrophic denitrification have been intensively investigated in the past few years.

Nitrate reduction by the sulfur based autotrophic denitrification process provides a novel possibility of achieving complete nitrogen removal through cooperation of anammox and autotrophic denitrification bacteria.In such a process, the nitrite and ammonium ions are converted into dinitrogen by anammox with nitrate as a byproduct (Reaction 1); the reduced sulfur compounds are used by autotrophic denitrification to denitrifying nitrate into nitrite (Reaction 2 and Reaction 3), which is subsequently removed by anammox. In addition, the growth of autotrophic denitrification bacteria (Reaction 3)is also responsible for nitrogen removal through direct conversion of nitrite to nitrogen gas. This reaction provides an additional degree of freedom, making the process more flexible in terms of the nitrite to ammonium ratio required. As a result, a complete nitrogen removal could be achieved by anammox-autotrophic denitrification without the requirements for a fixed ratio between nitrite and ammonium. However, little study was reported on the combination of amammox and S0based autotrophic denitrification.

Therefore, this study investigated the complete nitrogen removal of the combination system with amammox and S0based autotrophic denitrification. The effects of nitrite to ammonium ratio on the performance of coexistence system of anammox and S0-based autotrophic denitrification were also explored.

2 Experimental

2.1 Materials and methods

An UASB reactor with a dimension of 5 cm (diameter)×75 cm (height), and a working volume of 1.57 L at 30±1 °C, was utilized. An internal reflux peristaltic pump drove an internal up- flow at a velocity of 1.2 m/s.The outside of the reactor was covered with tin foil to avoid the influence of light on anammox bacteria.

Synthetic NO2--N and NH4+-N was prepared by mixing prescribed quantities of KNO3, and NaNO2. 200 g of elemental sulfur were added manually on day 1 and day 56 to provide electron donors. The reactor was operated in 6 stages, with the operating details presented in Table 1.Other compositions in synthetic wastewater were prepared according to Liu[12-13]. The inoculated sludge was a mixture of anammox sludge and denitrifying sludge (at a volume ratio of 3:1).

Table 1 Operating parameters for the UASB reactor

2.2 Chemical analysis

After filtration by a 0.45-µm Millipore filter, the,N-N, NH-N, S2-, Sand pH value in liquid samples were analyzed by an ion chromatograph equipped with an inhibitory type conductivity detector and an Ionpac column (ICS-3000; Dionex, USA). The pH value and S2-of the liquid samples and the suspended solids (SS)and volatile suspended solids (VSS) of the sludge were measured by standard methods[14]. The concentration of S0and nitrite was calculated based on mass balance calculations[15-17].

2.3 Microbial community

The DNA of sludge samples were extracted from 0.15 g to 0.25 g of dried sludge utilizing FastDNA (Qbingene,Carlsbad, CA) according to the manufacturer's instructions. The bacterial V3-V4 region of the 16S rRNA gene was amplified using the forward primers 338F (50-ACT CCT ACG GGA GGC AGC AG-30)and the reverse primer 806R (50-GGA CTA CHV GGG TWT CTA AT-30). The detailed PCR mixture and reaction procedure can be referred to the paper written by Chen, et al.[18]. The PCR product was purified by a GeneJetPCR purification kit (Thermo Scientific), prior to being put for sequencing on the Illumina Miseq PE300 platform (Illimina, USA).

3 Results and Discussion

3.1 Performance of the anammox coupled with S0-based autotrophic denitrification reactor

The reactor was operated for 63 days, including 6 stages(Stage I to Stage VI). The nitrogen removal performance is illustrated in Figure 1. At stage I, the influent concentration of nitrite and ammonium was 66 mg-N/L and 50 mg-N/L, respectively. The effluent nitrite concentration was lower than 0.5 mg-N/L and the effluent ammonium concentration varied between 0.93 mg-N/L and 2.25 mg-N/L. The nitrate concentration was below 0.5 mg-N/L on day 1, indicating the autotrophic denitrification occuring in the system. The average TN removal was as high as 99% during this stage and the effluent sulfate concentration remained at 150 mg-S/L.

On day 7, the influent nitrite and ammonium concentration were 92 mg-N/L and 70 mg-N/L, respectively (stage II).The effluent nitrite concentration was maintained at lower than 0.5 mg-N/L, while the effluent ammonium concentration was slightly increased to 2.85 mg-N/L. The TN removal was maintained at 99%, while the effluent sulfate concentration increased from 150 mg-S/L to 200 mg-S/L.The influent nitrite and ammonium concentration further increased to 132 mg-N/L and 100 mg-N/L, respectively, at stage III. The effluent ammonium concentration at first increased to 9.68 mg-N/L and then decreased to less than 2.25 mg-N/L. The effluent concentration of nitrite and nitrate ions was lower than 0.5 mg-N/L and 0.42 mg-N/L respectively. The TN removal efficiency firstly decreased to 95.93% and then increased to 99.39%. The effluen sulfate concentration further increased to 250 mg-S/L.

Figure 1 Performance of UASB reactor at different stages

On day 34, the influent nitrite and ammonium concentration was maintained at 132 mg-N/L and 100 mg-N/L respectively; however, the hydraulic retention time (HRT)decreased from 6 h to 5 h to further increase the influen nitrogen loading rates to 1.128 kg-N/(m3·d) (stage IV)The effluent nitrite and nitrate concentration was less than 0.4 mg-N/L and 0.53 mg-N/L, respectively. The effluen ammonium concentration increased to around 6.26 mg-N/L This phenomenon might be caused by the severe competition between autotrophic denitrification bacteria and anammox bacteria under high loading rates. The TN removal efficiency decreased to 97% and the effluen sulfate concentration was slightly increased to 270 mg-S/L.The influent nitrogen loading rate further increased to 1.41 kg/(m3·d) at stage V with the HRT decreasing from 5 h to 4 h. The effluent nitrate concentration jumped to as high as 5.35 mg-N/L, although the effluent nitrite and ammonium concentration was maintained at 2.14 mg-N/L and 2.93 mg-N/L, respectively. The sharp increase of effluent nitrate concentration might be ascribed to the insufficient S0concentration in the reactor. The TN removal efficiency decreased to 96% and the effluent concentration of reduced sulfate was equal to 150 mg-S/L.Then 200 g of S0powder were added into the reactor to reduce the effluent nitrate concentration.

On day 58, a further decrease of HRT from 4 h to 3 h resulted in an effluent ammonium concentration of higher than 10 mg-N/L (Stage VI). This phenomenon might be attributed to the predomination of autotrophic denitrifiers in the competition with anammox bacteria at the specified loading rates. The effluent pH value promptly decreased to 7.15―7.24 due to active autotrophic denitrification process that reversely inhibited the activity of anammox bacteria.

3.2 Nitrogen removal performance at different influent ratio

Nitrogen removal efficiencies of the reactor at different influent nitrite to ammonium ratios were investigated in 6 separate runs, with each run lasting for 7 days (Table 2). At a nitrite to ammonium ratio of 1.0, the ammonium concentration in the influent was maintained at 100 mg-N/L,and the nitrate concentration was decreased to 100 mg-N/L,with the HRT increasing to 6 h. The ammonium concentration in the effluent dramatically increased to 8.2 mg-N/L, whereas the nitrate concentration was close to zero, while the nitrite concentration was not detected.the ammonium removal efficiency dropped to 86.6%,resulting in a total nitrogen removal of 93.2%.

With the increase of influent nitrite to ammonium ratio to 1.35, the ammonium concentration in the effluent notably decreased to 1.8 mg-N/L with TN removal efficiency of 99.1%. Although the nitrite concentration in influent at a nitrite to ammonium ratio of 1.45 was higher than that of 1.40, nitrite species did not accumulate but was entirely removed. The same phenomenon was also observed at a nitrite to ammonium ratio of 1.55. Thus, an absolutely nitrogen removal (＞99.5%) was achieved with the influent nitrite to ammonium ratios ranging from 1.4 to 1.55. However, 14.2 mg-N/L of nitrite were accumulated when the influent nitrite to ammonium ratio was further increased 1.75, with the TN removal efficiency decreased to 94.8%.

Table 2 Effect of nitrite to ammonium ratio on TN removal of UASB reactor

3.3 Microbial communities

The parameters related to the alpha diversity of microbial community for each sample at a distance cutoff level of 0.03 are shown in Table 3. The species richness for bacteria in the reactor varied significantly during the 62-day operation, which was revealed by OTUs and Chao 1.This was confirmed by the coverage values of the six samples (99.74%, 99.84%, 99.70%, 99.51%, 99.72%, and 99.79%, respectively, as shown in Table 3), indicating that almost all of OTUs in the reactor were detected in this study.

Table 3 Richness and diversity of the six samples based on 0.03 distance

In total, 10 known bacterial phyla were detected in the six samples, which mainly included Proteobacteria,Chloroflexi, Planctomycetes, Actinobacteria,Acidobacteria, Ignavibcteriae, Chlorobi, Bacteroidetes,Firmicutes, and unclassified_k_norank (Figure 2).Four phyla, including Proteobacteria, Planctomycetes,Chloroflexi, and Actinobacteria, predominated in all communities.

Figure 2 Taxonomic classification of the bacterial communities at genus levels

A wide range of bacteria genera were identified as dominant in the stage I, such as Thiobacillus, norank_c_SBR2076, norank_f_ Rhodocyclaceae, Candidatus_Kuenenia, norank_f_Anaerolineaceae, Sulfurimonas,Candidatus_Brocadia, norank_f_ PHOS-HE36, SM1A02,Denitratisoma, Ferritrophicum, norank_o_JG30-KF-CM45, unclassified_f_Rhodocyclaceae, norank_f_Elev-16S-1332, unclassified_o_Acidimicrobiales,

Desulfocapsa, norank_c_TK10, Geothrix, norank_c_Ardenticatenia, norank_o_ Acidimicrobiales, norank_f_OPB56, MSBL7, unclassified_k_norank, norank_f_Caldilineaceae, and Candidatus_Microthrix (Figure 3).

Five phyla, including Thiobacillus, Candidatus_Kuenenia,Sulfurimonas, Candidatus_Brocadia, and Candidatus_Microthrix, are the typical functional bacteria at this stage.Thiobacillus and Sulfurimonas are the typical autotrophic denitrification bacteria, which could utilize the reduced sulfur oxidation compounds as electron donors to reduce nitrate to nitrite or nitrogen[19]. Candidatus_Kuenenia,Candidatus_Brocadia, and Candidatus_Microthrix are the typical anammox bacteria, which have been detected in many anaerobic ammonium oxidation reactors[20].

The genera Thiobacillus and Sulfurimonas increased from 1.41% and 0.71% to 6.48% and 7.47%, respectively,with the influent ammonium and nitrite concentrations increasing from 50 mg-N/L and 66 mg-N/L (Stage I) to 100 mg-N/L and 132 mg-N/L (Stage III), respectively. With the HRT decreasing from 6 h (stage III) to 5 h (stage IV), the abundance of Thiobacillus and Sulfurimonas dramatically increased to 18.77% and 2.35%. However, the abundance of Candidatus_Kuenenia, Candidatus_Brocadia, and Candidatus_Microthrix drastically decreased from 18.18%, 7.75% and 1.24% (stage I) to 1.89%, 1.23% and 0.69% (stage III), respectively.

Figure 3 Taxonomic classification of the bacterial communities at genus levels

Further decreasing the HRT from 6 h (stage III) to 5 h(stage IV), the abundance of Candidatus_Kuenenia,Candidatus_Brocadia, and Candidatus_Microthrix was slightly increased from 1.89%, 1.23%, and 0.69%to 3.99%, 2.51%, and 0.17%, respectively. At stage V,the abudance of Thiobacillus and Sulfurimonas slightly decreased to 16.98% and 0.88%, respectively, due to the shortage of elemental sulfur in the system, while the abundance of Candidatus_Kuenenia, Candidatus_Brocadia, and Candidatus_Microthrix increased to 10.51%, 3.09%, and 0.32%, respectively. This phenomenon indicated that the quantity of S0added in the system could adjust the competition between autotrophic denitrifiers and anammox bacteria. At stage VI, due to the higher activity of autotrophic denitrifiers, the abundance of Thiobacillus and Sulfurimonas further increased to 24.8% and 3.76%, while that of Candidatus_Kuenenia,Candidatus_Brocadia, and Candidatus_Microthrix decreased to 4.26%, 0.94%, and 0.25%, respectively.

4 Conclusions

Complete nitrogen removal was achieved through integrating anammox and autotrophic dnitrification in an UASB reactor. The total nitrogen (TN) removal rate increased stepwise from 0.46 to 0.94 kg-N/(m3·d), with an effluent TN concentration of less than 3.0 mg-N/L achieved.The process was relatively insensitive to the nitrite to ammonium ratio, which could achieve complete nitrogen removal when this ratio in influent varied in the range of 1.35—1.55. Fierce competition between autotrophic denitrifiers and anammox bacteria under short HRT and high loadings resulted in deterioration of TN removal.High-throughput sequencing technology indicated that Thiobacillus and Candidatus_Kuenenia were the functional strains for anammox and autotrphic denitrification process,respectively, in the reactor studied thereby. This result could provide a theoretical and technical basis for the large-scale application of anaerobic ammonium oxidation process.

Acknowledgements:This research was supported by the National Natural Science Foundation of China (No. 21307160),the Shandong Provincial Natural Science Foundation, China(No. ZR2019MEE038), the Fundamental Research Funds for the Central Universities (19CX02038A), the Open Project of Key Laboratory of Environmental Biotechnology, CAS(Grant No. kf2018003), and the Open Project Program of State Key Laboratory of Petroleum Pollution Control (Grant No. PPC2018006), CNPC Research Institute of Safety and Environmental Technology.