Heated Aeration for Nitrite-Oxidizing Bacteria (NOB) Control in Anammox-Integrated Membrane-Aerated Biofilm Reactors (MABR)

: Nutrient removal in conventional wastewater treatment systems is expensive due to the high aeration costs. An alternative method for effective and sustainable nitrogen removal in wastewater treatment is anaerobic ammonium oxidation (Anammox) implemented with other innovative technologies, such as membrane-aerated biofilm reactors (MABRs). A major challenge associated with the Anammox process is effective control of nitrite-oxidizing bacteria (NOB). High temperature operation in wastewater treatment systems can promote Anammox bacterial growth and inhibit NOB activity. This research aims to investigate the feasibility of integrating Anammox processes with a lab-scale MABR and to examine the effects of high temperature aeration supplied to MABR systems on Anammox bacterial growth and NOB suppression. Experimental results indicate that the membrane’s air permeability was a critical parameter for the successful operation of Anammox-integrated MABR systems due to its influence on the system’s dissolved oxygen concentration (0.41 ± 0.39 mg O 2 /L). The ammonia removal by AOB and Anammox bacteria was determined to be 7.53 mg N/L · d (76.5%) and 2.12 mg N/L · d (23.5%), respectively. High temperature aeration in MABRs with the Anammox process shows a promising potential for improving energy consumption and sustainable nitrogen removal in wastewater treatment systems.


Introduction
The presence of nutrients, such as nitrogenous compounds, in wastewater can pose serious environmental concerns to water systems, leading to reduced water quality and potential risks to public health [1,2].Nitrogen removal in wastewater treatment is often associated with extensive energy consumption and high treatment costs [2][3][4].Conventional wastewater treatment typically includes nitrification and denitrification processes, which are associated with high oxygen requirements, long hydraulic retention times, and a large footprint of land [5,6].These factors all contribute to the high treatment costs in conventional wastewater treatment systems.Additionally, the sludge that is produced needs to be managed for disposal and can account for up to 40% of the overall wastewater treatment management costs [3,4,7].
Anaerobic ammonium oxidation (Anammox) is an emerging technology first discovered in the 1990s that can convert ammonia (NH 3 ) and nitrite (NO 2 ) into nitrogen gas under anoxic conditions [8].In comparison to conventional wastewater treatment processes involving two-step nitrification by ammonia-oxidizing bacteria (AOB) and nitrite-oxidizing bacteria (NOB) followed by denitrification by denitrifying bacteria, the Anammox process can be operated with less aeration and a reduced footprint while producing less sludge, which can reduce overall wastewater treatment costs by providing more sustainable and efficient nutrient removal [3,7,9,10].However, Anammox bacteria cultivation The Anammox process can be integrated with other innovative technologies such as membrane-aerated biofilm reactors (MABR) to provide an ideal environment for simultaneous nitrification and denitrification (SND) biological reactions [12][13][14].For Anammox-MABR systems to thrive, the four groups of bacteria (AOB, Anammox, NOB, and denitrifying bacteria) must coexist.The expected biofilm profile (Figure 1) consists of an inner aerobic layer to support AOB and NOB growth and an outer anoxic layer to support Anammox and denitrifying bacterial growth [12].The nitrogen gas produced by both Anammox bacteria and denitrifying bacteria helps to maintain low dissolved oxygen (DO) concentration or anoxic conditions in the bulk solution.
sustainable and efficient nutrient removal [3,7,9,10].However, Anammox bacteria cultivation must be coupled with other bacterial groups due to its dependence on nitrite substrates, which are not commonly found in natural waters [9,11].The biological reactions for AOB, Anammox bacteria, NOB, and denitrifying bacteria can be described by Equations (1)-( 4), respectively [9,10]. (1) (2) The Anammox process can be integrated with other innovative technologies such as membrane-aerated biofilm reactors (MABR) to provide an ideal environment for simultaneous nitrification and denitrification (SND) biological reactions [12][13][14].For Anammox-MABR systems to thrive, the four groups of bacteria (AOB, Anammox, NOB, and denitrifying bacteria) must coexist.The expected biofilm profile (Figure 1) consists of an inner aerobic layer to support AOB and NOB growth and an outer anoxic layer to support Anammox and denitrifying bacterial growth [12].The nitrogen gas produced by both Anammox bacteria and denitrifying bacteria helps to maintain low dissolved oxygen (DO) concentration or anoxic conditions in the bulk solution.A major challenge for successful integration of Anammox and MABR technologies is the effective control of NOB while accomplishing sufficient treatment [15][16][17].Operational parameters that can inhibit NOB activity include DO concentration, aeration patterns, pH, hydraulic retention time (HRT), temperature, and substrate concentration [9,[18][19][20].Studies with MABR systems have shown low DO concentration and intermittent aeration strategies to enhance AOB activity and reduce NOB activity [15,[21][22][23][24].The AOB can outcompete NOB and function in oxygen-limited environments primarily due to their higher oxygen affinity in comparison to NOB [20,[25][26][27].A recent study demonstrated that a relatively higher DO concentration (up to 0.8 mg O2/L) can support NOB suppression in the presence of Anammox bacteria [22].Furthermore, operating with a shorter HRT can be A major challenge for successful integration of Anammox and MABR technologies is the effective control of NOB while accomplishing sufficient treatment [15][16][17].Operational parameters that can inhibit NOB activity include DO concentration, aeration patterns, pH, hydraulic retention time (HRT), temperature, and substrate concentration [9,[18][19][20].Studies with MABR systems have shown low DO concentration and intermittent aeration strategies to enhance AOB activity and reduce NOB activity [15,[21][22][23][24].The AOB can outcompete NOB and function in oxygen-limited environments primarily due to their higher oxygen affinity in comparison to NOB [20,[25][26][27].A recent study demonstrated that a relatively higher DO concentration (up to 0.8 mg O 2 /L) can support NOB suppression in the presence of Anammox bacteria [22].Furthermore, operating with a shorter HRT can be beneficial for controlling NOB inhibition by promoting NOB washout since AOB has a shorter doubling time (7-8 h) than NOB (10-13 h) [9,28].
The operational temperature of wastewater treatment processes can impact the metabolism of the bacteria in the system [29,30].A higher maximum growth rate for AOB compared to NOB is observed at higher temperatures than 20 • C [26,31].Due to the dominant growth rate of AOB over NOB at higher temperatures, the operation of bioreactors at temperatures higher than Environments 2024, 11, 155 3 of 13 30 • C was reported to be feasible for NOB elimination in nitrification processes [31,32].Recent studies demonstrated that operation at high temperatures (30 • C to 40 • C) could maintain stable nitrite accumulation with low nitrate concentrations [28,31,33].Furthermore, high-temperature operation affects the solubility of oxygen, which may reduce the DO concentration and create desirable growth conditions for Anammox-integrated systems [30].For most Anammox bacterial species, the optimal temperature for growth is 30 • C to 40 • C [3,7,34,35].The operation of high-temperature wastewater treatment could primarily control NOB suppression while also promoting Anammox bacterial growth.However, high-temperature operation is still unreliable since large energy requirements are associated with maintaining high water temperatures, leading to increased operational costs [32].
High-temperature reactor operation can be reliably incorporated in MABRs by supplying high-temperature aeration through the lumen of the membranes.In comparison to maintaining high water temperatures for wastewater treatment, managing aeration temperature supplied to MABRs is more feasible and requires less energy consumption because the energy for heating air is significantly small compared to that for heating wastewater of the same volume.Since MABRs support an inner aerobic layer, the high temperature aeration supplied could inhibit NOB activity by providing more favourable growth conditions for AOB, allowing AOB to outcompete NOB.Operation of Anammox-MABR systems with high-temperature aeration can effectively control and suppress NOB while reducing the competition for nitrite substrates between Anammox bacteria and NOB.This study aims to demonstrate NOB suppression using heated air aeration in an MABR for effective nitrogen removal by Anammox bacteria in a lab-scale experiment.Another objective of this work was to investigate the effect of the heated air aeration on the microbial communities in MABR biofilms.

Anammox-MABR Configuration
The membranes used in the Anammox-MABR system are polymeric hollow fiber membranes designed for oxygen transfer.A lab-scale Anammox-MABR operating under a partial nitritation-Anammox process (PN-A) was constructed using 4 membrane fibers, each 7.5 cm long (Figure 2).It should be noted that all 4 membrane fibers were placed in a single and well-mixed reactor (600 mL) so that the experimental variable (air temperature) can be emphasized in the reactor operation.The preferred length of the membrane in the Anammox-MABR system was determined to be less than 10 cm from the results of a preliminary MABR system (Figure S1 and Table S1) [Section S1, see Supplementary Information].Air at two different aeration temperatures was supplied through the membranes of the Anammox-MABR system: 20 • C (ports 1 and 3) and 60 • C (ports 2 and 4).The air was maintained at the desired temperature by directing air flow through copper tubing in a heated water bath kept at a constant temperature.The tubing from the water bath to the reactor was insulated to minimize the heat loss in the tubing.

Experimental Methods
The Anammox-MABR system was operated for 111 days in continuous fed-batch mode for a hydraulic retention time (HRT) of 14 days.One half of the solution (300 mL out of 600 mL) was replaced every 7 days with a fresh synthetic solution.The fresh feed solution was purged with nitrogen (N2) gas before its addition to the Anammox-MABR system.Samples from the bulk solution were collected from a port at the top of the Anammox-MABR system and analyzed.The pH (SevenMulti, Mettler-Toledo International Inc., Columbus, OH, USA), DO concentration (OX 4100H, VWR International, Mississauga, ON, Canada), and temperature were monitored daily.The reactor pH was maintained at 7.0-7.5 with periodic NaOH injections, and the temperature was kept consistently at 24.4 ± 1.0 °C.The samples were analyzed daily for ammonia (NH4-N TNT832, HACH, Romeoville, IL, USA), nitrite (NO2-N TNT840, HACH, Romeoville, IL, USA), and nitrate (NO3-N TNT835, HACH, Romeoville, IL, USA).The total suspended solids (TSS) and volatile suspended solids (VSS) were analyzed weekly according to Baird et al., 2017; Section 2540 [37].
The daily nitrogen removal rate (mg N/L•d) by each specific bacterial group (AOB, Anammox bacteria, NOB, and denitrifying bacteria) can be estimated by deriving theoretical mass balance equations from Equations ( 1)-( 4) [Section S2, see Supplementary Information].Since no organic substrates were provided to the synthetic feed solution, the activity of denitrifying bacteria was assumed to be negligible.

Quantifying Airflow Characteristics
The air flow rate can be used to estimate the average effluent air linear velocity through the membrane to determine if sufficient aeration was provided (Equation ( 5)).Furthermore, the residence time of the air can be determined (Equation ( 6)).

Average Linear Effluent Linear Velocity
Air Flow Rate Cross Sectional Area of the Flow (5) Residence Time τ Average Linear Effluent Linear Velocity Length of Membrane (6)

16S Amplicon Sequencing Methods
Biofilm membrane samples for the two temperature conditions (20 °C and 60 °C) and for the bulk solution in the Anammox-MABR system were prepared for 16S rRNA amplicon sequencing (Surette Laboratory, McMaster University, Hamilton, ON, Canada) (Table S2) [Section S3, see Supplementary Information].The microbial species were then assigned to a target bacterial group of interest: AOB, Anammox bacteria, NOB, denitrifying bacteria, and others.

Experimental Methods
The Anammox-MABR system was operated for 111 days in continuous fed-batch mode for a hydraulic retention time (HRT) of 14 days.One half of the solution (300 mL out of 600 mL) was replaced every 7 days with a fresh synthetic solution.The fresh feed solution was purged with nitrogen (N 2 ) gas before its addition to the Anammox-MABR system.Samples from the bulk solution were collected from a port at the top of the Anammox-MABR system and analyzed.Since no organic substrates were provided to the synthetic feed solution, the activity of denitrifying bacteria was assumed to be negligible.

Quantifying Airflow Characteristics
The air flow rate can be used to estimate the average effluent air linear velocity through the membrane to determine if sufficient aeration was provided (Equation ( 5)).Furthermore, the residence time of the air can be determined (Equation ( 6)).
Average Linear Effluent Linear Velocity = Air Flow Rate Cross Sectional Area of the Flow (5) Residence Time (τ air ) = Average Linear Effluent Linear Velocity Length of Membrane (6)

16S Amplicon Sequencing Methods
Biofilm membrane samples for the two temperature conditions (20 • C and 60 • C) and for the bulk solution in the Anammox-MABR system were prepared for 16S rRNA amplicon sequencing (Surette Laboratory, McMaster University, Hamilton, ON, Canada) (Table S2) [Section S3, see Supplementary Information].The microbial species were then assigned to a target bacterial group of interest: AOB, Anammox bacteria, NOB, denitrifying bacteria, and others.

qPCR Methods
The DNA from the biofilm membrane samples was extracted using the Soil DNA Isolation Kit (Norgen Biotek Corp., Thorold, ON, Canada).Each biofilm membrane sample was 2 cm in length and was weighed prior to DNA extraction.The specific targets of interest were the 16S rRNA genes for general bacteria, AOB, Anammox bacteria, and NOB (Table 1).Each target bacterial group of interest for qPCR analysis can be described from 16S rRNA with single-copy genes; therefore, the abundance can be directly compared [38,39].The PCR protocols were conducted for each primer pair (Table 2).The PCR components included 12.5 µL iTaq Universal SYBR Green Supermix (BioRad Laboratories (Canada) Ltd., Mississauga, ON, Canada), 9.5 µL DI water, 1 µL forward primer, 1 µL reverse primer, and 1 µL DNA template.The PCR products were purified using the PCR Kleen Spin Columns (BioRad, Laboratories (Canada) Ltd., Mississauga, ON, Canada) and then analyzed using a Nanodrop Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA).The DNA concentration of purified PCR product was quantified using Qubit dsDNA Quantification Assays (Thermo Fisher Scientific, Waltham, MA, USA).Standard curves (Figure S2, see Supplementary Information) were constructed by conducting qPCR for 10-fold dilutions of the purified PCR product (Table 3).The qPCR components included 10 µL iTaq Universal SYBR Green Supermix (BioRad, Laboratories (Canada) Ltd., Mississauga, ON, Canada), 6 µL DI water, 1 µL forward primer, 1 µL reverse primer, and 2 µL DNA template.

Anammox-MABR Performance
The overall performance of the Anammox-MABR system was determined by monitoring key operational parameters throughout the 111 days of the operational period.From day 0 to day 20, there was a rapid ammonia consumption attributed to AOB growth and an indication of low NOB levels due to nitrite accumulation and stable nitrate concentrations in the system (Figure 3a).However, after day 20, nitrite decreased, while nitrate increased drastically, implying NOB activity.Following day 40, the nitrite concentration was maintained consistently at ~45 mg N/L despite the rapid increase in nitrate.This observation demonstrates NOB suppression in the reactor, although it was not sufficient.Furthermore, the optimal growth of Anammox bacteria is dependent on the ratio of substrate available for utilization.The optimal nitrite to ammonia molar ratio for Anammox bacteria under anoxic conditions is 1.32 [9,10].During day 0 to day 40, the nitrite-to-ammonia molar ratio was unstable, indicating that the growth of the Anammox bacteria may not have been ideal; however, from day 40 onwards, the nitrite-to-ammonia molar ratio was stable at 1.17 ± 0.45.Note that the small fluctuations were induced by injections of NH 4 Cl into the reactor (Figure 3b).The average time between each injection period was 3 days, and the average ammonia removal efficiency was 59.2 ± 17.5% (Figure 3d).It should be noted that the reactor was not operated for high ammonia removal efficiencies because the primary objective of this study was to examine the effects of aeration temperature on NOB control in MABR systems.Since no organic substrates were supplied to the Anammox-MABR system, denitrification was negligible, with an average COD concentration of 11.4 ± 5.3 mg O 2 /L (Figure S3, see Supplementary Information).

Oxygen Transfer through Membranes
The DO concentration in the reactor was relatively high until day 40 (1.78 ± 1.62 mgO2/L), but it was maintained at a low level afterwards (0.36 ± 0.14 mgO2/L) (Figure 3c).This high DO until day 40 resulted in a high nitrate concentration by boosted NOB activity

Oxygen Transfer through Membranes
The DO concentration in the reactor was relatively high until day 40 (1.78 ± 1.62 mgO 2 /L), but it was maintained at a low level afterwards (0.36 ± 0.14 mgO 2 /L) (Figure 3c).This high DO until day 40 resulted in a high nitrate concentration by boosted NOB activity (Figure 3a).On day 35, a simple test was conducted by turning off the aeration blower and monitoring the DO change with time in the reactor.When the aeration was turned off, the DO dropped rapidly from 2.1 mg O 2 /L to below 0.5 mg O 2 /L within 120 min (Figure S4, see Supplementary Information), indicating that the aeration membrane allows significant oxygen transfer only with four pieces of 7.5 cm-long hollow fiber membranes.This result shows that membranes for high oxygen transfer may not be suitable for Anammox-MABR systems.
To reduce the oxygen transfer in the reactor, from day 35 onwards, the air blower was operated at the lowest setting (Table 4), and the DO concentration was maintained at 0.41 ± 0.39 mg O 2 /L.As a result, the nitrite concentration was restored immediately after reduced aeration, indicating controlled NOB activity.The linear velocity and mean air residence time of air through the membrane (τ air ) for each temperature aeration condition before and after day 35 can be used as an indication of whether the air supplied will maintain sufficient heat (Table 4).In full-scale MABR systems, the mean residence time can be even longer with longer membrane fibers; thus, the DO concentration can be even lower than 0.4 mg O 2 /L.Since typical Anammox processes operate with a DO < 0.1 mg O 2 /L [3], a lower DO concentration would be even more effective for controlling NOB activity in Anammox-MABR systems.

Nitrogen Removal by Specific Bacterial Groups
For ammonia removal, AOB and Anammox bacteria had an average daily removal rate of 7.53 mg N/L•d (76.5%) and 2.12 mg N/L•d (23.5%), respectively (Figure 4a).While AOB was predominantly responsible for ammonia removal in the Anammox-MABR system, there was minor or no competition between AOB and Anammox bacteria, as both groups of bacteria are associated with a slow growth rate [3].Since the ammonia concentrations in the Anammox-MABR reactor were not depleted with periodic injections of NH 4 Cl, the ammonia substrate was sufficient for both AOB and Anammox bacteria.Likewise, for nitrite removal, Anammox bacteria and NOB had an average daily removal rate of 2.38 mg N/L•d (37.2%) and 6.05 mg N/L•d (62.8%), respectively (Figure 4b).For days 20-35, nitrite removal was achieved predominantly by NOB due to the high oxygen supply; however, the Anammox contribution to nitrite removal was restored after day 40 with reduced aeration.This observation implies that the MABR membranes should be designed for moderate oxygen transfer to avoid the outgrowth of NOB.

Microbial Communities
The microbial communities from 16S amplicon sequencing were organized in five target bacterial groups of interest: AOB; Anammox bacteria; NOB; denitrifying bacteria; and others.The relative abundance for these bacterial groups of interest was compared for the two different aeration temperature conditions (Figure 5a,b) and the bulk solution (Figure 5c).The most abundant bacterial group in the biofilm was AOB at 20 °C and 60 °C, with an increasing relative abundance from 3.9% to 47.3 ± 10.9% and 40.4 ± 23.1%,

Microbial Communities
The microbial communities from 16S amplicon sequencing were organized in five target bacterial groups of interest: AOB; Anammox bacteria; NOB; denitrifying bacteria; and others.The relative abundance for these bacterial groups of interest was compared for the two different aeration temperature conditions (Figure 5a,b) and the bulk solution (Figure 5c).The most abundant bacterial group in the biofilm was AOB at 20 • C and 60 • C, with an increasing relative abundance from 3.9% to 47.3 ± 10.9% and 40.4 ± 23.1%, respectively.For Anammox bacteria, the relative abundance in the biofilm decreased from 8.5% to less than 0.001% at both 20 • C and 60 • C, indicating a decrease in Anammox bacterial activity and a shift towards full nitrification with primarily AOB and NOB in the system.The low levels of Anammox bacteria detected by 16S amplicon sequencing analysis could also be due to poor amplification of the 16S rRNA gene with the universal bacterial primers [45].For NOB, the relative abundance increased from 0.2% to 15.7 ± 8.9% (20 • C) and 19.81 ± 15.08% (60 • C).In comparison to both of the temperature aeration conditions, the relative abundance of AOB in the bulk solution was also highest (Figure 5c), indicating that AOB was dominant for both attached and suspended growth conditions with negligible repercussions.However, the relative abundance of NOB in the bulk solution was less than 2%, indicating that a majority of the NOB preferred to attach to the membrane, which supports a higher DO concentration in the aerobic biofilm.This suggests that the membrane's oxygen transfer and air permeability resulted in the outgrowth of NOB in the biofilm compared to that in the bulk solution.Compared to 16S amplicon sequencing, the qPCR method was found to provide more accurate detection and representation of the microbial community in the Anammox-MABR reactor.In this method, the relative abundance of each bacterial group of interest was normalized by the general bacteria, DNA concentration, and the effective total surface area of each membrane sample (Table 5).Comparing the two temperature conditions, the abundance of Anammox bacteria and NOB in the MABR biofilm was higher at 60 °C than at 20 °C, while the opposite was observed for AOB.The ratio between AOB and Anammox bacteria (AOB: Anammox) was 30.9 (20 °C) and 25.6 (60 °C), indicating that MABR oper-

Figure 1 .
Figure 1.Schematic diagram of the expected biological reactions in the biofilm profile for Anammox-MABR systems.

Figure 1 .
Figure 1.Schematic diagram of the expected biological reactions in the biofilm profile for Anammox-MABR systems.

Figure 3 .
Figure 3. (a) Nitrogen concentrations, (b) comparison of theoretical and experimental nitrite-to-ammonia molar ratio, (c) dissolved oxygen concentration, and (d) ammonia removal efficiency for Anammox-MABR performance throughout the 111-day operational period.

Figure 4 .
Figure 4.Estimated daily removal rate for (a) ammonia removal and (b) nitrite removal contributions for target bacterial groups based on theoretical mass balance in the Anammox-MABR throughout the 111-day operational period.

Figure 4 .
Figure 4.Estimated daily removal rate for (a) ammonia removal and (b) nitrite removal contributions for target bacterial groups based on theoretical mass balance in the Anammox-MABR throughout the 111-day operational period.

Figure 5 .
Figure 5. Relative abundance of microbial communities from 16S amplicon sequencing analysis for the V4 region of 16S rRNA gene in Anammox-MABR system for (a) T = 20 • C, (b) T = 60 • C, and (c) bulk solution.

Table 1 .
Specific target amplicons for qPCR and primers with associated sequences used in this study.

Table 2 .
PCR protocols for the primers used in this study.

Table 3 .
qPCR protocols for the primers used in this study.After extension, repeat denaturation-annealing-extension steps 40×.Following PCR, melt curves were generated between 65 • C and 95 • C in 0.5 • C increments to ensure PCR specificity. *

Table 4 .
Effluent air linear velocity and air residence time (τ air ) in Anammox-MABR system throughout the 111-day operational period for 20 • C and 60 • C aeration conditions.