Start-up and Nitrogen Removal Performance of Pure-Media MBBR Treating Domestic Wastewater Under Long-Term Low Temperature of 5℃

Start-up and Nitrogen Removal Performance of Pure-Media Mbbr Treating Domestic Wastewater Under Long-Term Low Temperature of 5℃

In the plateau region of Southwest China, with altitudes ranging from 2000 to 4500 meters, the annual average temperature is below 10℃. The water temperature in wasteWater Treatment plants in this region often drops to as low as 5℃ for more than 200 days. This long-term low-temperature condition poses a serious challenge to biological nitrogen removal processes. The conventional activated sludge process faces significant technical bottlenecks in this region: on one hand, the sludge system struggles to specifically enrich functional microbial communities, leading to excessively long hydraulic retention times; on the other hand, the excessive proliferation of filamentous bacteria during low-temperature periods easily triggers sludge bulking, disrupting the solid-liquid separation efficiency of secondary sedimentation tanks and affecting effluent quality; finally, the activity of nitrifying bacteria decreases sharply at temperatures below 12℃, significantly reducing treatment load during low-temperature periods.

The MBBR (Moving Bed Biofilm Reactor) process, with its advantage of specific enrichment of biofilm, demonstrates good treatment efficiency under low-temperature conditions. For example, Li Zhiwei et al. embedded MBBR technology into an activated sludge system at a wastewater treatment plant in Tibet to form a hybrid sludge-biofilm system. With influent water temperature below 10℃, the effluent still stably met the Class A standard. Another study showed that in a hybrid sludge-biofilm system, when water temperature dropped from 20℃ to 10℃, the nitrification contribution of biofilm increased from 10% to 18%. However, the competitive relationship between sludge and biofilm phases in hybrid systems can prevent the biofilm from fully realizing its potential. In contrast, the pure-media MBBR system no longer retains activated sludge, eliminating the competition between sludge and biofilm phases and further improving biofilm biochemical efficiency. For the removal of TP and SS in the effluent of the pure-media MBBR process, an enhanced magnetic coagulation sedimentation process can be employed, and pilot tests and engineering applications have demonstrated its advantages over conventional coagulation sedimentation in terms of shock resistance and chemical consumption, providing assurance for the engineering promotion of the pure-media MBBR process.

Regarding research on the pure-media MBBR process, Han Wenjie et al. believe that staging allows each stage of the biological system to achieve a mutually independent habitat with relatively independent functions, which is more conducive to the targeted cultivation and specific enrichment characteristics of the MBBR process. It is evident that the many advantages of the pure-media MBBR process offer ideas for solving the process innovation needs of wastewater treatment plants in Southwest China. However, existing research has mostly focused on ordinary low-temperature (>10℃) and short-term conditions. The application of pure-media MBBR under long-term low temperature (5℃) still faces key issues that need to be clarified: the biofilm attachment time under low-temperature environments; the rationality of process design parameters; stability under impacts from water quality and flow rate variations; and the mechanism of low-temperature resistance of the pure-media MBBR process.

This study selected a high-altitude wastewater treatment plant in western Sichuan to conduct a pure-media MBBR pilot test. The start-up and operational performance of the pure-media MBBR process under long-term low-temperature conditions (5℃) were systematically investigated. The shock resistance of the pure-media MBBR process was verified, and design parameters were evaluated. Microscopic methods were used to analyze the low-temperature resistance mechanism and microbial composition of the pure-media MBBR process, aiming to provide a reference for process retrofitting of wastewater treatment plants in high-altitude cold regions.

1 Materials and Methods

1.1 Overview of the Wastewater Treatment Plant

A wastewater treatment plant in southwestern Sichuan, with an average altitude of about 3500 meters, faces influent water temperatures as low as about 5℃ for more than 200 days from August to April of the following year. The plant has a total design treatment capacity of 1500 tons/day, and the effluent quality must meet the Class A standard of the "Discharge Standard of Pollutants for Municipal Wastewater Treatment Plant" (GB18918-2002),Figure 1.

1.2 Overview of the Pure-Media MBBR Pilot System and Water Quality

The influent for the pure-media MBBR pilot system was taken from the effluent of the equalization tank of the wastewater treatment plant. Influent and effluent water quality parameters are shown in Table 1. The system operated in a series configuration consisting of a single-stage anoxic tank and three-stage aerobic tanks (designated as A, O1, O2, O3). The design treatment capacity was 7.92 m³/d, with a total hydraulic retention time of 16 hours. The effective volume of each functional tank was 1.32 m³. Each tank was filled with SPR-Ⅲ suspended carriers, with an effective specific surface area > 800 m²/m³, density 0.94–0.97 g/cm³, wall thickness 200–250 μm, and material of high-density polyethylene (HDPE). The filling ratios of suspended carriers in the anoxic and aerobic zones were 50% and 40%, respectively. A nitrified liquid recirculation pipeline was installed between the aerobic three-stage and the anoxic zone, with a design recirculation ratio of 150% and a maximum recirculation ratio of 300%.

1.3 Division of Study Phases

As shown in Table 2, the operation of the pure-media MBBR pilot system was divided into five phases. During the start-up phase, non-attached biofilm carriers were added. A stepwise influent strategy was adopted, with the initial influent flow set at one-third of the design value (2.64 m³/d). The treatment flow was gradually increased based on treatment capacity, and successful start-up was considered achieved when effluent stably met the design standards at the design flow. During the stable operation Phase I, the HRT was fixed, and treatment capacity and effluent stability were monitored. During the shutdown and standby phase, the system stopped influent, aeration, and mixing equipment were turned off, and the carriers were left in the pilot system without any treatment. During the recovery phase, a stepwise influent strategy was adopted, with the initial influent flow set at one-third of the design value, gradually increased based on treatment capacity until the design flow was reached. The recovery process and performance of the system after one month of shutdown were monitored. Stable operation Phase II was the period after the system's treatment performance had recovered, entering a stable operation phase again. During this phase, lasting 245 days, the average water temperature was only about 5℃.

1.4 Tracking of Nitrogen Removal Performance Along the Flow Path

To determine the pollutant removal pathway of the pure-media MBBR system, during stable operation Phase II, the effluent from Zones A, O1, O2, and O3 was monitored daily, and the nitrification and denitrification loads of each functional zone were calculated. After sedimentation, all samples were filtered through medium-speed qualitative filter paper (pore size 15–20 μm), and the supernatant was collected for determination of COD, NH₄⁺-N, and TIN. The pollutant removal load was calculated according to equation (1).

Where: CRLv is the pollutant removal load, kg/(m³·d); ΔC is the pollutant removal concentration, mg/L; Q is the treatment flow rate, m³/d; R is the recirculation ratio, %; v is the tank volume, m³.

1.5 Determination of Denitrification Potential of Suspended Carriers

To verify the denitrification potential of the aerobic suspended carriers in the pure-media MBBR system, the nitrification and denitrification effects under simulated water quality shock were tested. The nitrification effect verification was conducted after the end of stable operation Phase II (day 371) in a laboratory-scale test using aerobic suspended carriers under high substrate and high DO conditions. The carriers from each aerobic zone were used, with the filling ratio consistent with the pilot system. The test water was prepared using the effluent from the pilot system with the addition of ammonium sulfate, and alkalinity was supplemented by adding sodium bicarbonate. The nitrification volumetric load ARLv was calculated according to equation (2).

Where: ARLv is the nitrification volumetric load, kg/(m³·d); ΔNH₄⁺ is the NH₄⁺-N removal concentration, mg/L; t is the reaction time of the laboratory test, h.

For the denitrification effect verification, anoxic zone carriers were used, with the filling ratio consistent with the pilot system. The test water was prepared using the influent of the pilot system with the addition of potassium nitrate. The denitrification volumetric load NRLv(DN) was calculated according to equation (3).

Where: NRLv(DN) is the denitrification volumetric load, kg/(m³·d); ΔNO₃⁻-N is the NO₃⁻-N removal concentration, mg/L; t is the reaction time of the laboratory test, h.

1.6 Analysis of Microbial Community Structure

To study the composition of functional microbial communities in different functional zones of the pure-media MBBR pilot system under staged configuration and to compare the differences with the activated sludge system in the same plant, samples of biofilm on the suspended carriers and activated sludge from the plant were collected after the end of stable operation Phase II (day 371) for high-throughput sequencing based on 16S rDNA amplicons. Biofilm samples were taken from the suspended carriers in the anoxic zone and aerobic zones O1, O2, and O3 of the pilot system. Activated sludge samples were taken from the aerobic zone of the plant's activated sludge system. The specific method for high-throughput sequencing was consistent with the report by Han Wenjie et al. For analysis, the aerobic zone biofilm and activated sludge were grouped together, and the anoxic zone biofilm and activated sludge were grouped together. The top 20 species in average relative abundance in each group were selected as dominant species for analysis, referencing the Bergey's Manual of Systematic Bacteriology and relevant literature regarding the functions of different genera.

1.7 Other Indicator Measurement Methods

Conventional water quality indicators were determined using standard methods. COD was measured by the dichromate method; NH₄⁺-N was measured by the Nessler's reagent spectrophotometric method; TN was measured by the alkaline potassium persulfate digestion-UV spectrophotometric method. Mixed liquor suspended solids (MLSS) provided by the biofilm was measured by the gravimetric method. Mixed liquor volatile suspended solids (MLVSS) was determined according to established literature. The absolute biomass of functional bacteria in the biofilm on suspended carriers and in the activated sludge was calculated as the product of the relative abundance of functional bacteria obtained from microbial analysis and MLVSS. DO was measured using a WTW Multi-3430i meter. After the end of stable operation Phase II (day 371), biofilm carriers from each functional zone of the pilot system were taken for biofilm thickness measurement using an industrial microscope (AOSVI, AO-HD228SD) under water at 40× magnification, and the results were recorded.

2 Results and Discussion

2.1 Analysis of System Operational Performance

During the start-up phase, the water temperature was (6.14±1.06)℃. Activated sludge from the aerobic zone of the plant was inoculated at a sludge concentration of 5000 mg/L. After filling the pilot system with wastewater, aeration in the aerobic zones and mixing in the anoxic zone were started. The DO in the aerobic zones was maintained at 7–9 mg/L. After two days of operation, continuous influent was started. Raw water carbon sources were used for denitrification, and DO in the aerobic zones was maintained at 6–7 mg/L. Effluent NH₄⁺-N was used as the main control indicator, and the influent flow rate and aeration rate were gradually increased. The adjustment point was when effluent ammonia nitrogen remained below 2 mg/L for three consecutive days. The influent and effluent water quality during the start-up phase of the pure-media MBBR pilot system are shown in Figure 2. By day 25 of operation, the system HRT was gradually reduced from 48 hours to 16 hours, reaching the design treatment capacity. During this phase, influent NH₄⁺-N, TIN, and COD were (27.90±9.01), (32.00±7.50), and (247.59±69.53) mg/L, respectively. After treatment by the pure-media MBBR system, effluent concentrations decreased to (0.95±0.55), (12.35±1.07), and (34.36±8.66) mg/L, respectively. According to equation (1), the nitrification and denitrification loads during the start-up phase reached 0.048 and 0.125 kg/(m³·d), respectively.

Regarding the start-up of biological wastewater treatment processes, with the activated sludge process, some studies have reported that in a multi-stage AO-MBR pilot system at a water temperature of 7–13℃, after continuous sludge cultivation to ensure uniform sludge concentration in each tank and then using a multi-point influent start-up method, after 45 days of continuous operation, the denitrification load only reached 0.024 kg/(m³·d). Additionally, when the temperature of an activated sludge system dropped from 15℃ to 12℃, the NH₄⁺-N removal rate rapidly decreased from 90% to 65%, indicating poor low-temperature performance. For the MBBR process, with a two-stage AO pure-media MBBR system at a water temperature of 19.4–22.0℃, by gradually increasing the flow rate while ensuring COD and NH₄⁺-N removal rates above 80%, the denitrification load reached 0.171 kg/(m³·d) after 17 days. In this project, with a water temperature of 4–8℃ during the start-up phase, while ensuring stable effluent quality compliance, the nitrification and denitrification loads reached 0.048 and 0.125 kg/(m³·d), respectively, after 25 days. The denitrification capacity of the pure-media MBBR start-up process was less affected by low temperature. On one hand, the start-up time was significantly shorter compared to the activated sludge process; on the other hand, the treatment load after successful start-up was also significantly higher.

The start-up speed of wastewater treatment systems is affected by low temperature. The activated sludge process, as a mixed culture system of microorganisms, has low enrichment efficiency for core functional microbial communities on one hand; on the other hand, due to its loose structure and lack of thermal insulation mechanisms, the start-up process is relatively slow. MBBR, based on the biofilm process, benefits from the specific enrichment advantage of microorganisms, allowing for the rapid enrichment of functional microbial communities such as nitrifying and denitrifying bacteria. Additionally, due to the certain thickness of the biofilm and the presence of extracellular polymeric substances (EPS), during low-temperature start-up, EPS and the outer layer of the biofilm can provide a certain degree of thermal insulation, thus enabling rapid start-up under low-temperature conditions.

To compare the pollutant removal efficiency of the pure-media MBBR pilot system and the activated sludge system in the same plant during stable operation, influent and effluent water quality data from both systems were analyzed, and the results are shown in Figure 3. The stable operation Phase I of the pure-media MBBR pilot system lasted for 56 days, during which the water temperature was (8.90±1.92)℃. The pilot system operated at the design flow rate with an HRT of 16 hours. Influent COD, NH₄⁺-N, and TN were (277.33±49.26), (29.13±4.91), and (36.03±6.67) mg/L, respectively. Effluent concentrations decreased to (36.22±5.53), (1.36±0.37), and (12.33±1.85) mg/L, respectively. The activated sludge system in the same plant, affected by low temperature, experienced a reduction in actual treatment capacity to approximately 700 m³/d, with the calculated biological HRT increasing to 47.14 hours. Effluent COD, NH₄⁺-N, and TN were (35.99±6.59), (1.51±0.31), and (13.70±0.79) mg/L, respectively. During stable operation Phase II, which lasted 245 days, the water temperature dropped to (5.45±1.29)℃, and the system maintained operation at the design flow rate. Influent COD, NH₄⁺-N, and TN increased to (324.47±72.76), (49.48±7.24), and (52.70±7.30) mg/L, respectively, representing increases of 9.16%, 23.70%, and 5.40% compared to the design influent load. During this phase, actual effluent concentrations decreased to (36.65±5.45), (1.28±0.66), and (12.27±0.88) mg/L, respectively. During this period, the actual treatment capacity of the activated sludge system in the plant further decreased to approximately 400 m³/d, with the calculated biological HRT increasing to 82.5 hours. Effluent COD, NH₄⁺-N, and TN were (39.73±4.78), (1.79±1.09), and (11.37±1.92) mg/L, respectively.

From day 82 to day 112 of the pilot system operation, the pure-media MBBR pilot system was shut down, during which aeration and mixing were stopped, and the biofilm was left in a static state. The recovery phase after shutdown lasted from day 113 to day 126, during which the water temperature was 10.33±0.58℃. Influent and effluent water quality during the recovery phase are shown in Figure 4. The initial influent flow was set at 2.64 m³/d. Effluent NH₄⁺-N was used as the main control indicator, and the adjustment point was when effluent ammonia nitrogen remained below 2 mg/L for three consecutive days. The influent flow and nitrification and denitrification performance of the pilot system were gradually restored. Actual operational results showed that the design treatment flow could be achieved within 14 days, and effluent NH₄⁺-N and TN began to meet standards stably.

For the activated sludge system, prolonged static conditions can lead to anoxic or anaerobic states, deteriorating microbial metabolic activity, disrupting sludge structure, causing the sludge to turn black and produce foul odors, and may also be accompanied by sludge floating and other detrimental phenomena, all of which increase the difficulty of restoring the activated sludge system's performance. For example, Huang Diannan et al. used activated sludge stored at low temperatures to restart treatment of garlic wastewater, and it took 33 days to achieve system performance recovery. For the MBBR biofilm system, the specific enrichment environment retains a higher base population of functional bacteria, allowing for faster re-enrichment of microbial communities and recovery of existing microbial communities. Additionally, the biofilm forms a floc matrix by secreting EPS, which adsorbs to the carrier surface. Under normal operating conditions, it adsorbs more nutrients, which can help maintain normal microbial metabolic activity during shutdown periods. In this project, using the pure-media MBBR process, the actual water temperature during shutdown was 10–12℃. Under the more unfavorable condition of lower storage temperature, the treatment performance was restored within only 14 days, demonstrating a clear advantage over the activated sludge process.

In summary, for the start-up of the pure-media MBBR process, with an HRT of only 16 hours and a water temperature of (6.14±1.06)℃, by adding activated sludge from the plant and non-attached biofilm carriers, and using a start-up strategy of gradually increasing the flow without sludge recirculation, the system start-up was completed in 25 days, with effluent water quality stably exceeding the Class A standard. For the operation of the pure-media MBBR process, over a period of 245 days with a water temperature of (5.45±1.29)℃, the effluent water quality consistently met standards. After one month of shutdown, the pure-media MBBR process required only 14 days to fully recover its treatment capacity at a water temperature of (10.33±0.58)℃.

2.2 Analysis of Nitrogen Removal Capacity and Potential of Each Functional Zone in the Pure-Media MBBR System

To determine the actual role of each functional zone of the pure-media MBBR system in responding to different water qualities and temperatures, the pollutant removal profiles along the flow path during stable operation Phases I and II are shown in Figure 5. During stable operation Phase I, the system influent water quality was relatively low and the water temperature was higher. Based on the changes in ammonia nitrogen along the flow path, the nitrification load of the aerobic zone during this phase was calculated to be 0.056 kg/(m³·d), with the average values for O1 to O3 being 0.052, 0.074, and 0.042 kg/(m³·d), respectively. Based on the changes in TN along the flow path, the denitrification load of the anoxic zone was calculated to be 0.144 kg/(m³·d). During stable operation Phase II, the system influent water quality increased and the water temperature decreased. The corresponding nitrification load of the aerobic zone during this phase was 0.088 kg/(m³·d), with the average values for O1 to O3 being 0.104, 0.111, and 0.049 kg/(m³·d), respectively. The denitrification load of the anoxic zone was 0.334 kg/(m³·d), while the calculated nitrification and denitrification loads for the activated sludge system in the same plant during this phase were only 0.023 and 0.027 kg/(m³·d), respectively.

From the perspective of pollutant removal, although NH₄⁺-N removal mainly occurred in the aerobic zone, the functions performed under different staging configurations were not consistent. During stable operation Phase I, the nitrification contribution rates of O1, O2, and O3 were 31.51%, 45.00%, and 25.34%, respectively. By stable operation Phase II, the contribution rates were 40.01%, 42.40%, and 18.58%, respectively. The aerobic zone O1, while receiving the COD that penetrated through the anoxic zone, experienced inhibited nitrification, resulting in its actual function combining carbon removal and nitrification. This phenomenon was consistent with conclusions drawn from a multi-stage segmented pure-media MBBR system used for municipal wastewater denitrification at a wastewater treatment plant in northern China. Furthermore, the nitrification efficiency of O1 was negatively correlated with the amount of COD penetration; as COD penetration decreased, its corresponding nitrification contribution increased. O2 played the primary nitrification role in different stages, with its main function being to reduce the NH₄⁺-N load. O3 served to ensure stable effluent NH₄⁺-N compliance, and its nitrification contribution was positively correlated with the influent water quality to O3, enhancing the system's shock resistance. Compared to the multi-stage segmented pure-media MBBR system at a wastewater treatment plant in northern China, which set up a 4-stage aerobic zone, the actual application results showed that the effluent ammonia nitrogen from O3 had already dropped below 2 mg/L. It can be seen that for the design of the aerobic zone of the pure-media MBBR process, setting up a 3-stage process can meet the nitrification requirements.

TN removal mainly occurred in the anoxic zone. Compared to the multi-stage segmented pure-media MBBR system at a wastewater treatment plant in northern China, which set up a 3-stage anoxic zone, this project achieved efficient TN removal using only a single-stage anoxic zone despite lower water temperatures. It can be seen that the requirement for staging in the anoxic zone of the pure-media MBBR system is lower than that for the aerobic zone. Additionally, considering the COD removal effect in the anoxic zone, the C/N ratio required for denitrification in the pure-media MBBR system was calculated to be 5.13 and 5.22, which is consistent with the C/N ratio required for denitrification in the anoxic zone of the multi-stage segmented pure-media MBBR system at a wastewater treatment plant in northern China.

As shown in Figure 5, after dilution by the recirculated nitrified liquid in the pilot system's aerobic zone, the actual influent ammonia nitrogen to O1 was below 15 mg/L. To verify the denitrification potential of the pure-media MBBR system and simulate the impact of increased influent water quality on the wastewater treatment plant, laboratory-scale nitrification and denitrification performance tests were conducted on the suspended carriers from this project after the end of stable operation Phase II using high substrate concentrations. During the laboratory tests, the water temperature was around 5℃. The influent ammonia nitrogen for the nitrification test was about 30 mg/L, and the system DO was controlled at 8–9 mg/L during the test. The changes in ammonia nitrogen in the O1, O2, and O3 systems of the aerobic zone are shown in Figure 6 (a). The results showed that the decrease in ammonia nitrogen in each system was relatively uniform throughout the test period. Using the entire laboratory test time as the reference point, the nitrification loads of O1, O2, and O3 were calculated to be 0.135, 0.137, and 0.062 kg/(m³·d), respectively. The overall nitrification load reached 0.111 kg/(m³·d), while the nitrification load of the activated sludge in the plant was only 0.025 kg/(m³·d). The nitrification load of the biofilm was 4.45 times that of the activated sludge. The results of the denitrification laboratory test are shown in Figure 6 (b). The influent nitrate nitrogen was about 35 mg/L, and the NO₃⁻-N decrease was relatively uniform throughout the test period. Using the entire laboratory test time as the reference point, the denitrification load of the carriers in Zone A was calculated to be 0.38 kg/(m³·d), while the denitrification load of the activated sludge in the plant was 0.030 kg/(m³·d). The denitrification volumetric load of the biofilm was 12.67 times that of the activated sludge.

Regarding nitrification, under low-temperature and high-substrate conditions, the nitrification load of each stage of the aerobic zone in the pure-media MBBR system increased by 29.81%, 23.42%, and 26.53%, respectively, during stable operation Phase II compared to the values along the flow path. The overall nitrification load of the aerobic zone increased by 26.14%, indicating that the pilot system still had significant nitrification potential. However, the activated sludge system increased by only 8.70% based on its already low actual nitrification load. Based on the calculation, while maintaining the same treatment flow and effluent water quality, the maximum influent NH₄⁺-N that the pilot system could handle was up to 60 mg/L, which is a 50% increase over the design influent NH₄⁺-N. Regarding denitrification, the denitrification load of the anoxic zone increased by 13.77% along the flow path during stable operation Phase II. Based on the calculation, with the existing treatment flow of 7.92 m³/d and effluent water quality unchanged, the maximum influent TN that the pilot system could handle was up to 78 mg/L, which is a 56% increase over the design influent TN. However, the activated sludge system increased by only 11.11% based on its already low actual denitrification load.

In terms of the actual role of each functional zone in the pure-media MBBR process, a three-stage process was set up for nitrification. O1 played a dual role in denitrification and the elimination of inhibitory factors such as high COD and toxic substances, providing conditions for O2 to achieve efficient nitrification. The setting of O3 enhanced the system's shock resistance, ensuring stable effluent water quality compliance. For denitrification in the pure-media MBBR process, the requirement for staging was lower than that for nitrification, and a single-stage process could be used. Regarding the denitrification potential of the pure-media MBBR process, at a water temperature of 5℃, when faced with impacts from influent water quality, increasing the DO in the aerobic zone from the 6–7 mg/L used during pilot operation to 8–9 mg/L enhanced mass transfer and oxygen transfer, which could further increase the system's nitrification load by more than 25%, while also increasing the denitrification capacity by more than 10%. In contrast, the activated sludge system, with its loose structure, does not have significant mass transfer and oxygen transfer resistance under normal operation, so its shock resistance effect was not obvious.

2.3 Analysis of Microbial Community Structure in the Pure-Media MBBR System

Microscopic analysis was performed on the MBBR biofilm and the activated sludge from the same plant. Biofilm and activated sludge samples collected after the end of stable operation Phases I and II were subjected to high-throughput sequencing of 16S rDNA amplicons to analyze the differences in microbial community composition. Additionally, scanning electron microscopy (SEM) was conducted on the biofilm and activated sludge samples from stable operation Phase II to observe morphological differences between the microbial communities.

The high-throughput sequencing results were analyzed at the genus level. The denitrifying bacteria in the anoxic zone of the pure-media MBBR system during stable operation Phases I and II are indicated by the hatched sections in Figure 7. These primarily included Simplicispira, Flavobacterium, and Dechloromonas, with total relative abundances reaching 36.81% and 42.19%, respectively. During the same periods, the values for the activated sludge system were 21.64% and 13.66%, respectively.

In terms of the enrichment capacity of denitrifying bacteria, during stable operation Phases I and II, the relative abundances of denitrifying bacteria in the biofilm were 1.70 and 3.09 times those in the activated sludge from the same plant, respectively. Although the water temperature dropped to about 5℃ during stable operation Phase II, the relative abundance of denitrifying bacteria in the biofilm system did not decrease; instead, it increased with the increase in influent load. However, the relative abundance of denitrifying bacteria in the activated sludge system decreased significantly after the water temperature dropped. In terms of the composition of dominant bacteria, the relative abundance of Simplicispira was 12.25% and 14.83% during stable operation Phases I and II, respectively. This genus can participate in biofilm formation, helping to improve the stability and efficiency of the denitrification system. Additionally, studies have found that this genus shows a clear preference for difficult-to-degrade cellulosic carbon sources such as corncob and loofah, suggesting that the influent in this project may contain similar difficult-to-degrade organic matter. Flavobacterium belongs to the phylum Bacteroidetes. Its relative abundance during stable operation Phases I and II was 12.27% and 14.83%, respectively. Liu Chao et al., when studying the microbial community characteristics of an MBBR system at 9℃, found that its relative abundance in the biofilm reached 15.19%, making it a dominant denitrifying bacterium in the system. Dong et al. found that this genus has excellent denitrification capacity under low-temperature conditions and its adaptation to low temperature is significantly stronger than that of other microbial groups. Dechloromonas, as a cultivable bacterium with denitrifying phosphorus removal capability, had relative abundances of 4.33% and 5.09% during stable operation Phases I and II, respectively. This genus can use volatile fatty acids (VFAs) as available substrates for phosphorus removal metabolism. In a modified SBBR nitrogen and phosphorus removal system at 10℃, this genus also became a dominant nitrogen and phosphorus removal bacterium, and a decrease in water temperature significantly increased its relative abundance, indicating that it has good adaptability to low temperatures.

Regarding the microorganisms in the aerobic zone, the nitrifying bacteria in the pure-media MBBR and activated sludge systems are indicated by the hatched sections in Figure 8. Their abundance was significantly lower than that of denitrifying bacteria, mainly consisting of Nitrospira and Nitrosomonas, but there was a large difference in enrichment efficiency. In the pure-media MBBR system, the average relative abundance of nitrifying bacteria during stable operation Phase I was 6.65%, with the relative abundances in the O1, O2, and O3 biofilms being 6.19%, 8.47%, and 5.28%, respectively. By stable operation Phase II, the average relative abundance of nitrifying bacteria increased to 7.22%, with the relative abundances in the O1, O2, and O3 biofilms being 6.36%, 9.62%, and 5.69%, respectively. During the same periods, the relative abundances of nitrifying bacteria in the activated sludge system were only 2.53% and 1.87%, respectively.

In terms of the enrichment capacity of nitrifying bacteria, the enrichment efficiency of the biofilm for nitrifying bacteria during stable operation Phases I and II was 2.63 and 3.86 times that of the activated sludge from the same plant, respectively. Similar to denitrification, the relative abundance of nitrifying bacteria in the biofilm did not change significantly during stable operation Phase II at lower water temperatures, but it decreased significantly in the activated sludge system. The low-temperature effect on the nitrification performance of activated sludge was greater. In terms of the composition of nitrifying bacteria, both the biofilm and activated sludge systems were dominated by Nitrosomonas and Nitrospira. The relative abundance of Nitrosomonas in the O1, O2, and O3 biofilms of the pure-media MBBR system during stable operation Phase I was 3.11%, 3.88%, and 1.39%, respectively; during stable operation Phase II, it was 3.09%, 4.03%, and 1.34%, respectively. The relative abundance of Nitrospira during stable operation Phase I was 3.08%, 4.59%, and 3.89%, respectively; during stable operation Phase II, it was 3.27%, 5.59%, and 4.35%, respectively. Regarding the relative abundance of nitrifying bacteria, the pure-media MBBR system exhibited a pattern of O2 > O1 > O3, which was consistent with the actual treatment load. Moreover, the relative abundance was not significantly different from that of a pure-media MBBR system under normal temperature conditions, indicating that low temperature did not significantly affect the enrichment efficiency of the biofilm for nitrifying bacteria. In terms of the composition of nitrifying bacteria, Nitrosomonas belongs to ammonia-oxidizing bacteria (AOB), which can oxidize ammonia nitrogen to nitrite. Some studies have indicated that it has good adaptability to low temperatures, and its relative abundance significantly increases in low-temperature environments of 3–8℃. Nitrospira is a nitrite-oxidizing bacterium (NOB) that can oxidize nitrite to nitrate. Studies have found that when the water temperature of an MBBR system decreased from 15℃ to 10℃, its copy number increased by 2.75 times, suggesting that it also has a strong preference for low temperatures. Additionally, some studies suggest that this genus is a subordinate of Comammox, carrying homologous genes for AOB and NOB, enabling it to directly oxidize ammonia nitrogen to nitrate nitrogen, thus allowing it to exist independently in the nitrification system without relying on NOB. According to existing research, the nitrifying bacteria in pure-media MBBR systems are mainly dominated by Nitrospira, accounting for more than 70%. However, in this project, with no significant difference in influent water quality, the proportion of Nitrosomonas increased significantly, which may be related to changes in the nitrification mechanism at low temperatures. That is, under normal temperature conditions, the biofilm is dominated by complete nitrification by Nitrospira, while under low-temperature conditions, Nitrosomonas and Nitrospira work together to complete nitrification. Research by Yung-Hsien Shao et al. found that in biofilm environments, Nitrosomonas and Nitrospira can form stable microbial communities, promoting each other's growth and metabolism and working together to complete the nitrification process.

The scanning electron microscopy (SEM) results for stable operation Phase II are shown in Figure 9. The activated sludge flocs had uneven particle sizes (50–200 μm) and were predominantly loose flocculent in shape with low compactness. A small number of free bacteria were observed on the surface (accounting for about 10% of the total bacteria), and the floc structure was loose. In contrast, the filamentous bacteria in the aerobic zone biofilm of the pure-media MBBR system accounted for 40%–50% of the volume, forming a reticular skeleton structure through interweaving. There were many short rods and cocci embedded in the filamentous network, forming a dense symbiotic system. In the anoxic zone, the volume of filamentous bacteria decreased to 10%, with short rods and cocci dispersed, and the bacterial density was relatively sparse. The abundance of filamentous bacteria in the aerobic zone of the pure-media MBBR system enhanced the structural stability of the biofilm. Moreover, nitrification and denitrification themselves are exothermic reactions, and the loose structure of activated sludge is not conducive to heat retention. In contrast, the dense structure and relatively complex biological phase of the pure-media MBBR biofilm provided a certain degree of thermal insulation, making it more resistant to low temperatures. Additionally, a small number of rotifers and nematodes were observed on the surface of the pure-media MBBR biofilm, which is speculated to buffer against organic load fluctuations. No dominant populations of microfauna were observed in the activated sludge system, with only sporadic distribution in the gaps between flocs.

The low-temperature resistance of the pure-media MBBR system stems, on one hand, from the biofilm's enhanced enrichment of denitrifying functional bacteria, particularly low-temperature-resistant denitrifying bacteria such as Flavobacterium and Dechloromonas. On the other hand, it stems from the dense structure and complex biological phase of the biofilm, which provide thermal insulation support.

2.4 Analysis of Factors Influencing Biofilm Load

The apparent characteristics and biofilm thickness of the pure-media MBBR system are shown in Figure 10. In terms of biofilm color, the biofilm on the anoxic zone carriers was brownish-yellow, while the biofilm on the aerobic zone carriers was dark brown. In terms of biofilm thickness, the thickness of the anoxic zone biofilm was (260 ± 119) μm, which was between that of O1 and O2 in the aerobic zone. The biofilm thickness in the aerobic functional zones gradually decreased from (288 ± 111) μm in O1 to (172 ± 130) μm in O3.

As shown in Table 3, the surface biomass of the anoxic zone biofilm in the pure-media MBBR system reached 12.25 g/m². The MLSS, calculated as the product of surface biomass, carrier filling ratio, and effective specific surface area of the biofilm, was 4900 mg/L, which was slightly higher than that of the activated sludge system in the plant. However, due to the higher VSS proportion in the biofilm, the MLVSS in the anoxic zone of the pure-media MBBR system reached 3724 mg/L, which was 1.52 times that of the activated sludge system. The absolute biomass of functional bacteria, calculated as the product of MLVSS and the relative abundance of functional bacteria, was 1571.16 mg/L, reaching 4.68 times that of the activated sludge system.

As shown in Table 4, the surface biomass of the aerobic zone biofilm in the pure-media MBBR system ranged from 8.26 to 16.33 g/m², decreasing gradually from the first stage to the third stage. Combined with the filling ratio and the effective specific surface area of the biofilm, the MLSS was calculated to be 2643.20–5225.60 mg/L. Among these, the MLSS in the first-stage aerobic zone was higher than that of the activated sludge system in the plant. Combined with the VSS proportion in the biofilm, the MLVSS in the aerobic zone of the pure-media MBBR system was calculated to be 2246.72–4128.22 mg/L, which was 1.68, 1.32, and 0.91 times that of the activated sludge system from O1 to O3, respectively. After incorporating the relative abundance of functional bacteria, the absolute biomass of functional bacteria was calculated to be 172.00–262.56 mg/L, which was 5.72, 6.80, and 2.77 times that of the activated sludge system from O1 to O3, respectively.

Principal component analysis (PCA) was performed on parameters such as the relative abundance of functional bacteria, biofilm surface biomass, biofilm thickness, absolute biomass of functional bacteria, and maximum pollutant removal load in different functional zones of the aerobic zone. Additionally, to verify the fundamental reasons for the differences in pollutant removal load between the pure-media MBBR system and the activated sludge system in the same plant, PCA was performed on the relative abundance of functional bacteria, absolute biomass of functional bacteria, and maximum pollutant removal load of the two systems. As shown in Figure 11, whether examining different functional zones of the pure-media MBBR system or comparing with the activated sludge system, the maximum pollutant removal load showed a positive correlation with the absolute biomass of functional bacteria, the relative abundance of functional bacteria, and biofilm surface biomass. However, the strongest positive correlation was with the absolute biomass of functional bacteria.

It can be seen that for the pure-media MBBR process, simply increasing the biomass is not the core condition for directly increasing the pollutant removal load. First, the biomass in the biofilm represents not only functional bacteria; other non-functional microorganisms also account for a significant proportion of the biomass. Therefore, judging the pollutant removal load of the biofilm solely based on biomass is not rigorous. Second, for suspended carriers of a fixed size, an increase in biomass implies an increase in biofilm thickness, which increases the difficulty of mass transfer. Increased biofilm thickness affects mass transfer on one hand and reduces the effective specific surface area on the other. A reduction in effective specific surface area directly affects mass transfer and, consequently, the treatment load. The limiting factor for the treatment load of the pure-media MBBR process is the absolute biomass of functional bacteria, and further, it is the relative abundance of functional bacteria, effective specific surface area, and mass transfer. Simply increasing the biomass is of little significance; the key lies in controlling the biofilm at an appropriate thickness and suitable biomass while increasing the relative abundance of functional bacteria.

This project used a pure-media MBBR system to treat long-term, ultra-low-temperature wastewater in the Sichuan region. A single-stage anoxic zone was set up, and a 3-stage aerobic zone was designed. The first stage of the aerobic zone primarily functioned to eliminate inhibitory factors and perform preliminary nitrification, removing organic matter inhibition for the subsequent aerobic zone. The second stage of the aerobic zone served as the main nitrification zone, achieving the best nitrification effect. The third stage of the aerobic zone had the lowest nitrification effect, primarily designed to cope with water quality shocks and ensure stable effluent water quality compliance. The enhanced low-temperature resistance of the biofilm stems, on one hand, from the biofilm's enhanced enrichment of denitrifying functional bacteria, such as low-temperature-resistant denitrifying bacteria like Flavobacterium and Dechloromonas. On the other hand, it stems from the dense structure and complex biological phase of the biofilm, which provide thermal insulation support. The treatment load of the pure-media MBBR system is simultaneously influenced by the enhanced enrichment of functional bacteria and the higher effective biomass.

3 Conclusions

3.1 Regarding start-up and operational performance, at a water temperature of (6.14±1.06)℃, the pure-media MBBR system completed start-up in 25 days and required only 14 days to restore treatment capacity after one month of shutdown. Over 245 days of operation at a water temperature of (5.45±1.29)℃, with a hydraulic retention time only 20% of that of the activated sludge process in the same plant, the removal rates of COD, ammonia nitrogen, and total nitrogen were stably maintained above 85%, 92%, and 75%, respectively.

3.2 Regarding functional zoning and optimization of nitrogen removal mechanisms, the first stage of the aerobic zone in the MBBR system achieved removal of COD inhibitory substances and preliminary nitrification, the second stage served as the main nitrification zone, and the third stage ensured stable effluent quality. Regarding denitrification potential, at a water temperature of 5℃, when faced with impacts from influent water quality, increasing the DO in the aerobic zone to enhance mass transfer and oxygen transfer could further increase the system's nitrification load by more than 25% while also increasing the denitrification capacity by more than 10%.

3.3 Regarding low-temperature resistance and engineering design recommendations, on one hand, the low-temperature resistance stems from the biofilm's enhanced enrichment of denitrifying functional bacteria, particularly low-temperature-resistant denitrifying bacteria such as Flavobacterium and Dechloromonas. On the other hand, it stems from the dense structure and complex biological phase of the biofilm, which provide thermal insulation support. In the design of pure-media MBBR systems, it is necessary to simultaneously consider the enhanced enrichment of functional bacteria and the higher effective biomass to achieve higher treatment loads.