A Novel Layered and Advanced Nitrogen Removal Filter with Gravel and Embedded Bio-Organic Carrier Based on Autotrophic and Heterotrophic Pathways

: Given increasingly prominent environmental issues, there is a pressing need to satisfy more stringent emission standards for wastewater treatment plants (WWTP) while concurrently prioritizing energy conservation; a new up-ﬂow layered nitrogen removal ﬁlter was constructed on a laboratory scale using gravel (for the bottom and top layers) and embedded bio-organic carriers (for the middle layer) containing microorganisms as ﬁllers to treat the secondary efﬂuent by introducing a portion of raw water. This study investigated the nitrogen removal effectiveness and transfer pathways of synthetic wastewater at varying mixing ratios, promoted the enrichment of Anammox Bacteria (AnAOB) by embedding microorganisms, and analyzed the microbial community structure using high-throughput sequencing techniques. The ﬁndings showed that the highest total nitrogen (TN) removal efﬁciency was achieved with chemical oxygen demand (COD), ammonia (NH 4+ -N), and nitrate (NO 3 − -N) contents in the mixture at 77, 10, and 8 mg · L − 1 , respectively, with an average efﬁciency of 89.42%. NO 3 − -N was mostly removed through denitriﬁcation (heterotrophic), while NH 4+ -N was eliminated by partial nitriﬁcation (PN) and anaerobic ammonium oxidation (Anammox, autotrophic). According to high-throughput sequencing results, denitrifying bacteria such as Thauera (1.30–6.96%), Flavobacterium (0.18–0.40%), and Parcubacteria (0.14–0.32%) were present in all the ﬁlter layers, and Anammox bacteria such as Candidatus_Kuenenia were predominant in the middle layer at a 0.88% abundance, with the aid of organic carriers.


Introduction
High nitrogen content is a major cause of eutrophication in many bodies of water in China [1].To reduce nitrogen pollution at the source, many urban sewage treatment plants in China have implemented the "GB18918-2016" (Discharge standard of pollutants for municipal wastewater treatment plants (WWTP)) level A discharge standard promulgated by the Ministry of Ecology and Environment of China over the last decade.However, for some poor watersheds, more stringent local discharge standards have been implemented, and since the secondary effluent is predominantly nitrate (NO 3 − -N), it is necessary to add an additional carbon source to the WWTP to bring the effluent up to standard.Denitrification biofilters (DNBFs) are often used to purify the secondary effluent by adding an additional carbon source to remove NO 3 − -N from the secondary effluent [2].Unfortunately, this not only consumes sodium acetate and other chemicals and generates excessive residual sludge, but also emits large amounts of N 2 O during denitrification, whose greenhouse effect is 300 times greater than that of CO 2 [3].With the ongoing issue of global climate change, achieving advanced nitrogen removal while saving energy and reducing carbon emissions has become an urgent problem.
Water 2023, 15, 1919 2 of 13 Anaerobic ammonium oxidation (Anammox), as a new type of nitrogen removal process, can save 60% of aeration energy, 100% of denitrification carbon sources, 90% of sludge production, and 100% of N 2 O emissions [4]; if it can be applied to the secondary effluent treatment, it will greatly reduce the energy consumption of WWTP.However, the low nitrite nitrogen (NO 2 − -N) concentration is a key factor limiting the application of Anammox in mainstream municipal wastewater treatment processes.The partial nitrification (PN) process is an effective method for achieving NO 2 − N accumulation.However, the PN/Anammox process cannot treat nitrate-containing wastewater, and the partial denitrification (PD) process developed in recent years provides a new idea for treating nitrate-containing wastewater.Current research indicates that PD can be easily achieved by adjusting C/N, pH, and hydraulic residence time (HRT), and has been proven to be an effective method for generating NO 2 − -N with long-term stability [5].Du et al. simultaneously introduced high-NO 3 − -N wastewater and municipal wastewater into the PD reactor to convert NO 3 − -N to NO 2 − -N, and removed ammonia (NH 4 + -N) and NO 2 − -N from the PD effluent in the subsequent Anammox reactor; the removal efficiency of total nitrogen (TN) by Anammox reached 78.9% [6].Cui et al. utilized a biofilter with volcanic rock as filter media to enrich Anammox Bacteria (AnAOB) with NO 2 − -N provided by PD to treat domestic wastewater, resulting in a 74.5% TN removal by Anammox [7].The problem of applying PD/Anammox to secondary effluent treatment is the poor biodegradability and low chemical oxygen demand (COD) concentration in the secondary effluent, which also does not contain the NH 4 + -N required for the Anammox reaction.In order to improve the advanced nitrogen removal efficiency of the "PD/Anammox" system, some research in recent years has divided the raw water into two parts.The first stream was introduced to an aerobic tank to oxidize NH 4 + -N to NO 3 − -N, and its effluent mixed with the second stream was transferred to a PD/A reactor, where NO 3 − -N was reduced to NO 2 − -N via PD using the organic matter in the raw wastewater [8].Zhao used the aerobic sequencing batch reactor (SBR) to partially nitrify NH 4 + -N to NO x − -N (NO 2 − -N dominated), and then used an up-flow anaerobic sludge bed (UASB) to treat the mixture of its effluent and the raw water, which removed 76.8% of TN by Anammox [9].Based on this, the PD/Anammox process can be made feasible to treat the secondary effluent by introducing a portion of raw water to provide a carbon source and NH 4 + -N; it can simultaneously remove NH 4 + -N from the raw water and NO 3 − -N from the secondary effluent, and the distribution ratio of the two parts significantly affects the nitrogen removal performance, but few studies have been reported in this area.
Solid-phase denitrification, in which the carbon source for denitrification is decomposed from an organic solid, has been considered as a promising approach to enhance nitrate removal of the secondary effluent [10].At present, the commonly used solid carbon sources are synthetic polymers such as polycaprolactone (PCL), polyhydroxyalkanoate (PHA), and Polyvinyl alcohol (PVA) [11,12], or agricultural wastes such as the peanut shell (PS) and corncob [13].Up until now, several agricultural wastes such as wheat straw, corncob (CC), and rice straw have been investigated as solid carbon sources for denitrification [14].The addition of cellulose-rich agricultural waste or artificial macromolecules as solid carbon sources can not only release organics for denitrification, but also provide a carrier for the growth of microorganisms [15,16].AnAOB grows slowly and has strict environmental requirements, so establishing a suitable micro-environment by embedding or attachment would contribute to enhance Anammox for nitrogen removal [17].In practical application, natural organic materials (PS, CC, et al.) are often mixed with carrier materials to improve the carbon source slow-release performance [18].Most recent research has focused on nitrogen removal efficiency [19], as well as bacterial population components and their transformation [20]; however, the research on the collaboration of various nitrogen conversion pathways within the embedding carriers or the attached bio-film is limited.Given this, a composite organic carrier was prepared using PVA, sodium alginate (SA), and CC to investigate its nitrogen removal characteristics and promoting effect on microbial enrichment when used as a filtering medium for treating the secondary effluent.
In this study, a novel up-flow layered filter with inorganic gravel and embedded bio-organic carriers was constructed for advanced nitrogen removal of the secondary effluent using a carbon source in raw water, and the nitrogen conversion pathways and the distribution of nitrogen removal microorganisms in different distribution ratios of raw water and the secondary effluent were investigated.The potential of the PD/Anammox process for treating the secondary effluent was evaluated by examining the promotion of the AnAOB enrichment by microorganisms embedded in organic carriers, in the hope of providing a new energy-saving nitrogen removal process for the treatment of the secondary effluent in practical engineering applications.

Experimental Setup
The layered nitrogen removal filter is shown in Figure 1.The device was made of Plexiglas with an inner diameter of 6.8 cm, a height of 41.5 cm, and an effective volume of 1 L. Sampling ports were laid out along the vertical wall.The bottom layer (height 10 cm) was filled by the gravel with a particle size of 5-8 mm, the middle layer (height 15 cm) was filled by an embedded bio-organic carrier with a particle size of 10 mm (weight 160 g), and the top layer (height 6.5 cm) was filled by the gravel with a particle size of 5-8 mm.
Water 2023, 15, x FOR PEER REVIEW 3 of 14 and CC to investigate its nitrogen removal characteristics and promoting effect on microbial enrichment when used as a filtering medium for treating the secondary effluent.
In this study, a novel up-flow layered filter with inorganic gravel and embedded bioorganic carriers was constructed for advanced nitrogen removal of the secondary effluent using a carbon source in raw water, and the nitrogen conversion pathways and the distribution of nitrogen removal microorganisms in different distribution ratios of raw water and the secondary effluent were investigated.The potential of the PD/Anammox process for treating the secondary effluent was evaluated by examining the promotion of the AnAOB enrichment by microorganisms embedded in organic carriers, in the hope of providing a new energy-saving nitrogen removal process for the treatment of the secondary effluent in practical engineering applications.

Experimental Setup
The layered nitrogen removal filter is shown in Figure 1.The device was made of Plexiglas with an inner diameter of 6.8 cm, a height of 41.5 cm, and an effective volume of 1 L. Sampling ports were laid out along the vertical wall.The bottom layer (height 10 cm) was filled by the gravel with a particle size of 5-8 mm, the middle layer (height 15 cm) was filled by an embedded bio-organic carrier with a particle size of 10 mm (weight 160 g), and the top layer (height 6.5 cm) was filled by the gravel with a particle size of 5-8 mm.

Experimental Materials
The gravel was purchased from Zhengzhou Linyuan Co., Ltd.(Zhengzhou, China); its surface was relatively rough with the diameter of 3-5 mm.The embedded bio-organic carrier was prepared with corncob (CC), polycaprolactone (PCL), and PVA.CC was collected from a farmland of Henan (Zhengzhou, China); after being dried in the oven at 50 • C for 4 h, it was grinded into little particles with the diameter of 3~4 mm.PCL (molecular weight 50,000-80,000), PVA (polymerization degree 1799, alcoholysis degree > 99%), and SA (viscosity of 1.0% aqueous solution at 20 • C, 80-120 mPa-s) were purchased from Shenzhen Huixin Plastic Chemical Co., Ltd.(Shenzhen, China).The inoculated sludge was collected from the anoxic tank of an Anaerobic-Anoxic-Oxic (AAO) sewage plant (Zhengzhou, China).The bacterial suspension (BS) was obtained from the supernatant of the inoculated sludge after centrifugation at 200 rpm for 1 min.The chemicals used in this experiment include: and EDTA.All the above chemicals were purchased from Sinopharm Group Co., Ltd.(Shanghai, China), and the grade level was an analytical reagent (AR ≥ 99.5%).

Preparation of Embedded Bio-Organic Carrier and Filmed Gravel
The embedded bio-organic carrier was prepared according to the following procedure (shown in Figure 2): (1) the PVA and SA were dissolved by distilled water with the concentration of 8% PVA and 1% SA; (2) the mixture was heated at 95 • C for 2 h; (3) after the mixture was cooled to 20 • C, CC, PCL, and BS were added to prepare a solution containing 8% CC, 8% PCL, and 2% BS; (4) the solution was continuously aerated for 1 min by a micron aeration disk under the aeration rate of 10 L•min −1 ; (5) the solution was poured into a plate mold with 100 small boxes of 1 cm 3 each, and was frozen at −20 • C for 12 h; (6) the freezing solids were removed from the plate mold, then were dropped into the saturated H 3 BO 3 and CaCl 2 solution (4%) to crosslink at 4 • C for 24 h; and ( 7) the embedded bio-organic carrier was taken out of the solution and washed by distilled water.The prepared carrier had a density of 1 g•cm −3 and compressive strength bigger than 260 N. The gravel was immersed in the inoculated sludge mixture for 48 h, and then was washed by distilled water 3 times.
Water 2023, 15, x FOR PEER REVIEW 4 of 14

Preparation of Embedded Bio-Organic Carrier and Filmed Gravel
The embedded bio-organic carrier was prepared according to the following procedure (shown in Figure 2): (1) the PVA and SA were dissolved by distilled water with the concentration of 8% PVA and 1% SA; (2) the mixture was heated at 95 °C for 2 h; (3) after the mixture was cooled to 20 °C, CC, PCL, and BS were added to prepare a solution containing 8% CC, 8% PCL, and 2% BS; (4) the solution was continuously aerated for 1 min by a micron aeration disk under the aeration rate of 10 L•min −1 ; (5) the solution was poured into a plate mold with 100 small boxes of 1 cm 3 each, and was frozen at −20 °C for 12 h; (6) the freezing solids were removed from the plate mold, then were dropped into the saturated H3BO3 and CaCl2 solution (4%) to crosslink at 4 °C for 24 h; and (7) the embedded bio-organic carrier was taken out of the solution and washed by distilled water.The prepared carrier had a density of 1 g•cm −3 and compressive strength bigger than 260 N. The gravel was immersed in the inoculated sludge mixture for 48 h, and then was washed by distilled water 3 times.

Experimental Design and Operation
The flow rate was controlled by a peristaltic pump (CT1000) with a HRT of 8 h and the temperature was not controlled during the operation and was maintained at 20.0-25.0 °C.The synthetic wastewater was prepared to simulate the mixture of urban sewage and the secondary effluent.The CH3COONa•3H2O, NH4Cl and KNO3, and KH2PO4 were used as the carbon, nitrogen, and phosphorus sources, respectively.NaHCO3 was added to supply the alkalinity.In addition, 1 L of synthetic wastewater consisted of 40 mg CaCl2:2H2O, 375 mg NaHCO3, 8 mg MgSO4, and 0.3 mL nutrient solution [21].The components of the nutrient solution per liter were 1.5 g of FeCl3•6H2O, 0.15 g of H3BO3, 0.03 g of CuSO4•5H2O, 0.18 g of KI, 0.12 g of MnCl2•4H2O, 0.06 g of Na2MoO4•2H2O, 0.12 g of ZnSO4•7H2O, 0.15 g of CoCl2•6H2O, and 10 g of EDTA.According to the different mixing ratios of raw water and the secondary effluent, four different mixing components were simulated, and the experiment contained four phases (Table 1).

Experimental
Phase

Experimental Design and Operation
The flow rate was controlled by a peristaltic pump (CT1000) with a HRT of 8 h and the temperature was not controlled during the operation and was maintained at 20.0-25.According to the different mixing ratios of raw water and the secondary effluent, four different mixing components were simulated, and the experiment contained four phases (Table 1).

Experimental
Phase

Anammox Activity Assay
Part of the filter material was removed from the reactor and washed gently three times with distilled water for later use.A 250 mL conical flask was filled with 200 mL of an initial solution prepared from NH 4 Cl and NaNO 2 , with NH 4 + -N and NO 2 − -N concentrations of 30 mg•L −1 each.Nitrogen gas was then introduced into the flask for approximately 10 min to maintain anaerobic conditions, followed by the rapid addition of the filter material.The conical flask was tightly sealed with a butyl rubber stopper and an aluminum cap, with a hole in the rubber stopper for sample collection.The flask was placed in a water bath with a constant temperature and agitation set at 30 ± 1 • C and 120 rpm, respectively.The first water sample was collected immediately after the start of the timer, and subsequent samples were collected at 10 min, 20 min, 30 min, 45 min, 60 min, 90 min, and 120 min, respectively, for the analysis of NH 4 + -N, NO 2 − -N, and NO 3 − -N content.The activity of Anammox was determined as the biomass-specific ammonium consumption rates.The specific Anammox activity (SAA) was calculated as follows: where ∆[NH 4 + -N] was the change in NH 4 + -N concentration during the reaction, mg•L −1 ; V was the volume of the reaction solution, 200 mL; ∆t was the reaction time, h; and M was the mass of the filler, g.

High-Throughput Sequencing
High-throughput sequencing technology was used to analyze the microbial community structure of different layers.The samples were collected from different layers, named AS (initial activated sludge), Top (top layer), Middle (middle layer), and Bottom (bottom layer).High-throughput sequencing was carried out to investigate the distribution of microbial species by Bio-engineering (Shanghai) Co., Ltd.(Shanghai, China), and the samples were stored, frozen in solid CO 2 , during transportation.

Nitrogen Conversion Pathway
The analysis of nitrogen conversion pathways can reveal different nitrogen conversion reactions occurring in each filter layer and further clarify the synergistic relationship of each filter layer in the nitrogen removal process.Five pathways of nitrogen conversion were considered in this study, which are Anammox (2), PD (3), PN (4), the second step of denitrification (SD) (5), and the second step of nitrification (SN) (6); the details are shown as follows [22]: Since the variables associated with nitrogen transformations include NH 4 + -N, NO 2 − -N, and NO 3 − -N, three of the above nitrogen conversion pathways were selected randomly to establish the equations of mass conservation for NH 4 + -N, NO 2 − -N, and NO 3 − -N, (Equations ( 7)-( 9)), and the final nitrogen conversion pathways in each layer were determined by the combination with reasonable calculation results.

Pollutant Removal Performance
The performance of nitrogen and COD removal is shown in Figures 3 and 4.During phase I (1-48 days), the average effluent concentrations of NH 4 + -N and NO 3 − -N were 1.63 and 1.60 mg•L −1 , respectively, and the removal efficiencies of NH 4 + -N and NO 3 − -N were 65.66% and 84.48%, respectively.Meanwhile, the effluent TN and COD concentrations were 3.38 and 5.76 mg•L −1 , respectively, with the removal efficiencies of 78.58% and 80.81%, respectively.The organic carriers prepared would release lots of organic matter during the initial 5 days and were conducive to enriching denitrifying bacteria (DNB) [23], so the denitrification performance was excellent during phase I, while nitrifying bacteria were not as active in the early experimental period, resulting in lower NH 4 + -N removal efficiency.
Water 2023, 15, x FOR PEER REVIEW 7 of 14   With the influent C/N increased to 3.33 in phase II (49-94 days), the average removal efficiency of NH4 + -N increased to 93.31%, but the average removal efficiency of NO3 − -N decreased to 73.42%.The removal efficiency of TN and COD did not significantly change, with 80.83% and 75.93%.Compared with the nitrogen removal performance of phase I, the nitrification performance was significantly improved with the stabilization of the microbial community, while the denitrification performance slightly decreased with the increase in C/N, possibly due to the decrease in carbon release capacity of middle organic carriers [24].The accumulation of NO2 − -N was also detected in the effluent, which, when oxidized by nitrifying bacteria to NO3 − -N in the top layer, could lead to a decrease in the removal efficiency of NO3 − -N, while the NO2 − -N generated in the reactor also provided With the influent C/N increased to 3.33 in phase II (49-94 days), the average removal efficiency of NH 4 + -N increased to 93.31%, but the average removal efficiency of NO 3 − -N decreased to 73.42%.The removal efficiency of TN and COD did not significantly change, with 80.83% and 75.93%.Compared with the nitrogen removal performance of phase I, the nitrification performance was significantly improved with the stabilization of the microbial community, while the denitrification performance slightly decreased with the increase in C/N, possibly due to the decrease in carbon release capacity of middle organic carriers [24].The accumulation of NO 2 − -N was also detected in the effluent, which, when oxidized by nitrifying bacteria to NO 3 − -N in the top layer, could lead to a decrease in the removal efficiency of NO 3 − -N, while the NO 2 − -N generated in the reactor also provided substrates for the Anammox reaction.Although AnAOB are autotrophic bacteria, the Anammox process will not be affected under low COD concentrations [25].
During phase III (95-133 days), the average effluent concentration of NH 4 + -N was 0.26 mg•L −1 and the average NH 4 + -N removal efficiency increased further to 96.75%.As the influent COD increased from 60 to 77 mg•L −1 (C/N = 4.23), the NO 3 − -N and TN concentrations in the effluent were 1.43 and 1.81 mg•L −1 , respectively, with the average removal efficiencies of NO 3 − -N and TN increasing to 83.89% and 89.42%, respectively.The NH 4 + -N was maintained at a high removal efficiency, which was basically unaffected by the variation of the influent water.Although the influent COD concentration was increased, its removal efficiency was increased to 81.52%, which shows that the raw water dosing ratio can be increased appropriately to provide a carbon source to make the NO 3 − -N removal more adequate.It should be noted that obvious organic matter was released from the organic carrier only during the initial 5 days, and the carbon source required for denitrification in later phases was still mainly from the influent.Hu et al. processed the secondary effluent with different C/N ratios using DNBFs, achieving an 85% removal efficiency of NO 3 − -N at a C/N ratio of 5 [26].This is similar to the results obtained in phase III of our experiment.However, their use of additional carbon sources resulted in extra energy consumption.
In phase IV (134-168 days), the C/N was reduced to 3.67 by increasing the influent NH 4 + -N concentration, and the NH 4 + -N removal efficiency remained high at 98.56%, while the NO 3 − -N removal efficiency decreased significantly to 67.81%; the decrease in C/N directly led to the deterioration of the NO 3 − -N removal.The TN and COD removal efficiencies were 82.26% and 76.57%, respectively, which were slightly lower than phase III.
By adjusting the mixing ratio of raw water and the secondary effluent to obtain the optimal C/N, the system can have a good pollutant removal performance.

Comparison of Nitrogen Removal between Each Layer
The nitrogen removal performance of each layer is shown in Figure 5.The DO concentration in the influent water was 2-3 mg•L −1 , while the nitrifying bacteria were aerobic, and the aerobic environment at the bottom made the removal of NH 4 + -N occur mainly in the bottom layer; its removal efficiency increased from 54.35% in phase I to 86.14% in phase IV.While the bottom removed most of the NH 4 + -N, the denitrifying bacteria used the COD in the influent to also remove some of the NO 3 − -N by denitrification.The denitrification effect was directly related to C/N [27], and the NO 3 − -N removal efficiency at the bottom varied with the fluctuation of the influent C/N, which was 33.87% in phase I, then increased to 35.89% and 44.73% due to the rise of C/N in phase II and phase III, respectively, and decreased to 18.88% in phase IV with the decrease in C/N. a C/N ratio of 5 [26].This is similar to the results obtained in phase Ⅲ of our experiment.However, their use of additional carbon sources resulted in extra energy consumption.
In phase Ⅳ (134-168 days), the C/N was reduced to 3.67 by increasing the influent NH4 + -N concentration, and the NH4 + -N removal efficiency remained high at 98.56%, while the NO3 − -N removal efficiency decreased significantly to 67.81%; the decrease in C/N directly led to the deterioration of the NO3 − -N removal.The TN and COD removal efficiencies were 82.26% and 76.57%, respectively, which were slightly lower than phase Ⅲ.By adjusting the mixing ratio of raw water and the secondary effluent to obtain the optimal C/N, the system can have a good pollutant removal performance.

Comparison of Nitrogen Removal between Each Layer
The nitrogen removal performance of each layer is shown in Figure 5.The DO concentration in the influent water was 2-3 mg•L −1 , while the nitrifying bacteria were aerobic, and the aerobic environment at the bottom made the removal of NH4 + -N occur mainly in the bottom layer; its removal efficiency increased from 54.35% in phase I to 86.14% in phase Ⅳ.While the bottom removed most of the NH4 + -N, the denitrifying bacteria used the COD in the influent to also remove some of the NO3 − -N by denitrification.The denitrification effect was directly related to C/N [27], and the NO3 − -N removal efficiency at the bottom varied with the fluctuation of the influent C/N, which was 33.87% in phase I, then increased to 35.89% and 44.73% due to the rise of C/N in phase II and phase Ⅲ, respectively, and decreased to 18.88% in phase Ⅳ with the decrease in C/N.The removal efficiency of NO  By the middle organic carrier layer, the DO concentration dropped to below 1 mg•L −1 , the activity of nitrifying bacteria was inhibited, heterotrophic denitrifying bacteria dominated, and the average removal efficiency of NH 4 + -N in the middle layer was 8.47% throughout the experiment.In phase I, 51.48% NO 3 − -N was removed from the middle layer due to the organic carrier that could be used as a slow-release carbon source, and although the C/N increased in phase II, the NO 3 − -N removal efficiency decreased to 38.17%, and in phase III, the NO 3 − -N removal efficiency in the middle layer increased to 41.34% with the increase in C/N.By phase IV, the denitrification efficiency in the bottom layer had deteriorated due to the decrease in C/N, and most of the NO 3 − -N was removed in the middle layer, with its removal efficiency continuing to increase up to 50.59%.
The top gravel layer had little pollutant removal effect, and its NO 3 − -N removal efficiency was negative except for the removal of a small amount of NH 4 + -N, presumably because the nitrifying bacteria oxidized part of the NH 4 + -N and NO 2 − -N produced in the middle to NO 3 − -N, while the Anammox reaction also produced NO 3 − -N; however, the organic carbon source was not enough-the NO 3 − -N could not be further converted to N 2 in the top gravel layer.Wang et al. had removed the NO 3 − -N produced by Anammox by adding PCL to the top of the continuous up-flow reactor [28]; in this experiment, the top layer of filter media was only gravel.

Nitrogen Conversion Pathway
Using the established calculation method, we analyzed the nitrogen conversion pathways of different filler layers (Figure 6), and determined the contribution of Anammox to TN, as well as assessed the activity of AnAOB using kinetic experiments to calculate the ing PCL to the top of the continuous up-flow reactor [28]; in this experiment, the top layer of filter media was only gravel.

Nitrogen Conversion Pathway
Using the established calculation method, we analyzed the nitrogen conversion pathways of different filler layers (Figure 6), and determined the contribution of Anammox to TN, as well as assessed the activity of AnAOB using kinetic experiments to calculate the SAA.As the nitrogen transformation process primarily occurred in the middle and bottom layers, only those layers were analyzed.Our results showed that PN, PD, and SD predominantly occurred in the bottom gravel layer.Moreover, depletion of DO inside the biofilm may limit the activity of oxidizing bacteria (NOB), as they are more sensitive to low DO than ammonia-oxidizing bacteria (AOB), which is beneficial for NO2 − -N accumulation [29].AOB exhibited high activity in the bottom layer under an influent DO concentration of 2-3 mg•L −1 , converting NH4 + -N to NO2 − -N through PN.Denitrifying heterotrophic bacteria in the bottom layer utilized COD in the influent to convert NO3 − -N to NO2 − -N through PD.Typically, a C/NO3 − -N ratio of 2.6-3.0 is maintained to reduce the conversion of NO2 − -N [7].However, Our results showed that PN, PD, and SD predominantly occurred in the bottom gravel layer.Moreover, depletion of DO inside the biofilm may limit the activity of nitrite-oxidizing bacteria (NOB), as they are more sensitive to low DO than ammonia-oxidizing bacteria (AOB), which is beneficial for NO 2 − -N accumulation [29].AOB exhibited high activity in the bottom layer under an influent DO concentration of 2-3 mg•L −1 , converting NH 4 + -N to NO 2 − -N through PN.Denitrifying heterotrophic bacteria in the bottom layer utilized COD in the influent to convert NO 3 − -N to NO 2 − -N through PD.Typically, a C/NO 3 − -N ratio of 2.6-3.0 is maintained to reduce the conversion of NO 2 − -N [7].However, the high C/NO 3 − -N ratio (3-9.6) of the influent in this experiment resulted in the reduction of NO 2 − -N produced by both PD and PN into N 2 .Anammox, PD, and SD were found to be the primary nitrogen conversion paths in the middle layer.During the first two phases, about 96% of the NO 2 − -N produced by PD was reduced to N 2 by SD, while very limited NO 2 − -N was available for Anammox, and the average contributions of Anammox to TN removal were 3.87% and 6.81%, respectively.By phase III and IV, with the increase in AnAOB activity, their ability to compete with NO 2 − -N became stronger, resulting in the reduction in the proportions of NO 2 − -N reduced by SD to 76.9% and 79.49%, respectively, and the contribution of Anammox to TN removal increased to 9.86% and 11.6% in phase III and IV, respectively.Throughout this experiment, NO 2 − -N produced by PD in the middle and bottom layers was reduced by SD, with an average reduction efficiency of 95%.The activity of AnAOB in the organic filter layer, as represented by SAA, was low in the first two phases at 0.529 and 0.935 mg•(g•h) −1 , respectively, correlating with the lower Anammox contribution to nitrogen removal.However, there was a significant increase in SAA during phase III and IV at 2.073 and 2.179 mg•(g•h) −1 , respectively.
The embedding of microorganisms in the organic carrier facilitated the enrichment of AnAOB.Through autotrophic (Anammox) and heterotrophic (PD and SD) nitrogen conversion pathways, the pollutants in the influent were effectively removed.The competition for NO 2 − -N between denitrifying bacteria and AnAOB was impacted by the influent's varying C/N, with the optimal Anammox nitrogen removal contribution and SAA observed at C/N = 3.67.

Microbial Communities in Each Layer
Due to our main focus on the comparison between different vertical filter layers, and the phase III filter having the highest TN removal efficiency and a more stable microbial community structure compared to the first two phases, sludge samples collected at the end of phase III were selected for high-throughput sequencing of microbial communities.All sequences with a relative abundance less than 0.5% were classified as "others", and all unidentified sequences were classified as "unclassified" (Figure 7).mg•(g•h) , respectively.
The embedding of microorganisms in the organic carrier facilitated the enrichment of AnAOB.Through autotrophic (Anammox) and heterotrophic (PD and SD) nitrogen conversion pathways, the pollutants in the influent were effectively removed.The competition for NO2 − -N between denitrifying bacteria and AnAOB was impacted by the influent's varying C/N, with the optimal Anammox nitrogen removal contribution and SAA observed at C/N = 3.67.

Microbial Communities in Each Layer
Due to our main focus on the comparison between different vertical filter layers, and the phase III filter having the highest TN removal efficiency and a more stable microbial community structure compared to the first two phases, sludge samples collected at the end of phase III were selected for high-throughput sequencing of microbial communities.All sequences with a relative abundance less than 0.5% were classified as "others", and all unidentified sequences were classified as "unclassified" (Figure 7).Nitrosomonas, a representative AOB genus capable of oxidizing NH4 + -N to NO2 − -N, was detected in the bottom gravel layer with an abundance of 0.37%, which was higher than that in the middle layer (0.10%).This explains why most of the influent NH4 + -N was converted to NO2 − -N in the bottom gravel layer.Nitrospira, a common NOB genus [30],  Nitrosomonas, a representative AOB genus capable of oxidizing NH 4 + -N to NO 2 − -N, was detected in the bottom gravel layer with an abundance of 0.37%, which was higher than that in the middle layer (0.10%).This explains why most of the influent NH 4 + -N was converted to NO 2 − -N in the bottom gravel layer.Nitrospira, a common NOB genus [30], had a relative abundance of 5.60% in the inoculated sludge, which decreased to 1.70% in the middle layer and 3.28% in the bottom layer by the end of phase III.The reduced activity of Nitrospira favored the accumulation of NO 2 − -N.Thauera, a representative denitrifying bacterial genus, had a relative abundance of 6.96% in the bottom layer, and it reduced NO 3 − -N and NO 2 − -N to N 2 using COD as an electron donor.Li found that, due to the presence of Thauera, about 70% of the COD reduction occurred in the bottom aerobic section of the single-phase PN/A reactor [31].As water flowed from the bottom to top, COD was consumed first at the bottom, resulting in a higher relative abundance of denitrifying bacterial genera Thauera (6.96%), Flavobacterium (0.41%), Parcubacteria (0.32%), and Azospira (0.59%) than that in the middle layer (1.30%, 0.19%, 0.14%, and 0.23%, respectively).The gradual decrease in COD and fluctuation in C/N promoted interspecies selection among denitrifying bacterial genera [32].Candidatus_Kuenenia, a major genus of AnAOB, had a relative abundance of only 0.01% in the inoculated sludge, but its abundance increased to 0.88% in the organic carrier layer by the end of phase III, higher than that in the top and bottom gravel layers (0.54% and 0.42%, respectively).Embedding microorganisms in organic carriers helped enrich AnAOB.Additionally, Nitrosomonas and Nitrospira had abundances of 0.37% and 8.22%, respectively, in the top gravel layer and could oxidize NH 4 + -N to NO 3 − -N, further validating the negative nitrogen removal efficiency of the top gravel layer.

Influent Mixing Ratio Model
Based on the above analysis, the nitrogen conversion mechanism of the layered filter was proposed (Figure 8).The influent nitrogen and COD significantly influenced the conversion pathways.In order to flexibly adjust the distribution ratio of raw water and the secondary effluent according to the nitrogen removal requirement, and to establish a more general guide for the advanced nitrogen removal, the optimal distribution ratio calculation model was constructed.The result of the calculation is shown in Equation ( 15).The best distribution ratio of raw water and the secondary effluent can be evaluated by substituting the actual operating data of the sewage plant.For example, when the NH4 + -N and COD of raw water were 30 and 230 mg•L −1 , respectively, the NO3 − -N, COD, and NH4 + -N of the secondary effluent were 12, 25, and 3 mg•L −1 , respectively, and the effluent NO3 − -N of the layered filter was 2 mg•L −1 .The calculated optimal distribution ratio between the raw water and secondary effluent was 1:10, when the mixed influent NH4 + -N, NO3 − -N, and COD concentrations were 5.67, 10.80, and 45.30 mg•L −1 , respectively.

Conclusions
By directly introducing a portion of raw water into the filter, it is possible to reduce the sludge load and save energy while providing a carbon source for the post-layered nitrogen removal filter to achieve advanced nitrogen removal.Through autotrophic and heterotrophic nitrogen conversion pathways, the TN and COD removal efficiencies of the secondary effluent can reach 89.42% and 81.52%, respectively, with concentrations of 1.81 and 14.2 mg•L −1 , respectively.In addition, embedding microorganisms in the organic carrier can promote the enrichment of AnAOB while improving the nitrogen removal performance of the system.Specifically, Candidatus_Kuenenia reached an abundance of 0.88%, and Anammox contributed 11.6% to TN removal.These findings provide an economic and energy-efficient strategy for the advanced nitrogen removal of the secondary effluent in existing WWTP, without the need for external carbon sources or aeration.

Figure 2 .
Figure 2. Preparation process of the embedded bio-organic carrier.

Figure 2 .
Figure 2. Preparation process of the embedded bio-organic carrier.
y, and z are the extent of the reaction of the three selected nitrogen conversion pathways, mmol; a, b, and c are the stoichiometric coefficients of NH 4 + -N in each selected equation; d, e, and f are the stoichiometric coefficients of NO 2 − -N in each selected equation; g, h, and i are the stoichiometric coefficients of NO 3 − -N in each selected equation; and A, B, and C are the masses of NH 4 + -N, NO 2 − -N, and NO 3 − -N, respectively, changed in each layer of the biofilter, mg.

Figure 4 .
Figure 4.The removal performance of TN and COD.

Figure 4 .
Figure 4.The removal performance of TN and COD.
the nitrogen transformation process primarily occurred in the middle and bottom layers, only those layers were analyzed.

Figure 6 .
Figure 6.Percentage of nitrogen conversion pathways in different carrier layers: (a) bottom gravel layer; (b) middle organic carrier layer.

Figure 6 .
Figure 6.Percentage of nitrogen conversion pathways in different carrier layers: (a) bottom gravel layer; (b) middle organic carrier layer.

Figure 7 .
Figure 7. Distribution of bacterial community at genus level on day 133 in phase III.

Figure 7 .
Figure 7. Distribution of bacterial community at genus level on day 133 in phase III.

Table 1 .
Simulated wastewater composition at different mixing ratios.

Table 1 .
Simulated wastewater composition at different mixing ratios.
StateEnvironmental Protection Administration of China 2002).DO, pH, and oxidation-reduction potential (ORP) were measured by the WTW Multi 3401 DO tester.The morphologies of the carrier were observed using a scanning electron microscope (SEM) (Model ULTRA 55, Caise Co., Oberkochen, Germany).