Performance of partial nitritation-anammox processes at mainstream conditions in an IFAS system

Performance of partial nitritation-anammox processes at mainstream conditions in an IFAS system Alba Pedrouso, Jozef Trela, Angeles Val del Rio, Anuska Mosquera-Corral and Elzbieta Plaza Accepted Mansucript How to cite: Pedrouso, A., Trela, J., Val del Rio, A., Mosquera-Corral, A., & Plaza, E. (2019). Performance of partial nitritation-anammox processes at mainstream conditions in an IFAS system. Journal Of Environmental Management, 250, 109538. doi: 10.1016/j.jenvman.2019.109538 Copyright information: © 2019 Elsevier Ltd. This manuscript version is made available under the CC-BY-NC-ND 4.0 license (http://creativecommons.org/licenses/by-nc-nd/4.0/) 1 Performance of partial nitritation-anammox processes at mainstream conditions in an IFAS system Doi: https://doi.org/10.1016/j.jenvman.2019.109538 Alba Pedrousoa*, Jozef Trelab, Angeles Val del Rioa, Anuska Mosquera-Corrala and Elzbieta Plazab a Department of Chemical Engineering, School of Engineering, Universidade de Santiago de Compostela, Rúa de Lope Gómez de Marzoa s/n, E-15782. Santiago de Compostela, Galicia, Spain. E-mail: alba.ped[email protected], man[email protected], and [email protected] b Department of Sustainable Development, Environmental Science and Engineering, Royal Institute of Technology (KTH), Teknikringen 10B, SE-10044, Stockholm, Sweden. E-mail: [email protected], [email protected] *Corresponding author: Alba Pedrouso, Email: alba.pedrous[email protected], Phone: +34 881816739 ABSTRACT: The partial nitritation-anammox processes implementation in the main line of wastewater treatment plants would lead them closer to the energy autarky. With this purpose, an integrated fixed film activated sludge (IFAS) reactor was operated at pilot scale. Efficient nitrogen removal (72 ± 11 %) was achieved for anaerobically pre-treated municipal wastewater at low temperature (21 - 15 ºC), with a nitrogen removal rate of 37 ± 3 g N/(m3·d) at 15 °C. The ammonium oxidizing bacteria were more abundant in the activated sludge, while anammox bacteria were primarily located in biofilm attached onto the carriers surface. Nitrite oxidizing bacteria (NOB) activity was similar between both fractions and its specific activity decreased more than that of other populations when the operating temperature was reduced. Furthermore, the IFAS operational strategy (aerobic/anoxic periods) allowed an efficient NOB activity suppression inside the reactor, which accounted only for the 10 - 20 % of the maximum potential activity. Keywords: anammox; autotrophic nitrogen removal; integrated fixed film activated sludge; mainstream; nitritation. 2 1. Introduction The long-term stability of the partial nitritation-anammox (PN/AMX) processes operating at mainstream conditions, characterized by low ammonium concentration (<100 mg N/L) and low temperature (< 25 ºC), remains as a challenge. The main identified bottlenecks are the achievement of the stringent discharge limits, the effective suppression of nitrite oxidizing bacteria (NOB) growth and the effective anammox bacteria retention inside the reactor (Cao et al., 2017; Hoekstra et al., 2018). Different strategies were proposed to suppress the NOB activity (Agrawal et al., 2018; Cao et al., 2017; Ge et al., 2015). Among them, some authors suggest performing the PN/AMX processes in two reactors where each process is optimized separately (Dosta et al., 2015; Pérez et al., 2015). Compared to the one-stage option, the two-stage PN/AMX system requires more space for implantation, a more complex control system, higher capital costs and it was reported to presumably lead to higher greenhouse gases emissions (Lackner et al., 2014). Therefore, the hybrid biofilm-based technologies and their ability to decouple the solid retention time (SRT) of different populations arose as promising approaches to improve the process stability joining the advantages of both one-stage and two-stage configurations (Han et al., 2016; Veuillet et al., 2014). Recently, hybrid systems like integrated fixed film activated sludge (IFAS) reactors where sludge is a mixture of suspended and biofilm biomass, have been explored to perform the PN/AMX processes simultaneously (Veuillet et al., 2014). The IFAS reactors have been found to be more flexible, efficient and resistant against both concentration and hydraulic shock loads than other activated sludge processes (Regmi et al., 2011). Due to mass transfer limitations, anammox bacteria preferentially grow in the biofilm while the ammonium oxidizing bacteria (AOB) and NOB are mainly developed as suspended biomass (Veuillet et al., 2014). Thus, IFAS reactor provides the opportunity to experience a more versatile operation and potentially obtain a selectively NOB washout, while anammox bacteria remain in the system. However, scarce information is available regarding the PN/AMX processes applied for the treatment of municipal wastewater at low temperatures in IFAS reactors (Table 1). For example, Yang et al. (2017) operated a plug-flow IFAS system, at 24 - 26 °C, treating aerobically pre-treated wastewater with a total nitrogen loading rate (NLR) of 125 mg N/(L·d). These authors achieved an effluent containing total nitrogen (TN) concentrations lower than 10 mg N/L (commonly established discharge limit) by applying a strategy based on the residual ammonium 3 concentration (3 - 5 mg N/L). Nevertheless, these authors stated that the long-term process stability is still an issue requiring the strict control of the airflow rate to maintain the microaerobic conditions (0.15 to 0.36 mg O2/L) (Yang et al., 2017). Furthermore, the use of intermittent aeration was widely used. Malovanyy et al. (2015) optimized the intermittent aeration strategy (15 min ON/45 min OFF) obtaining a nitrogen removal efficiency (NRE) up to 70% for anaerobically pre-treated wastewater in a continuous IFAS system at 25 °C. Trojanowicz et al. (2016) run a similar system at 17 °C but the obtained NRE was limited to 44%, mainly due to the nitrate produced by the NOB activity, probably caused by the low operational temperature. Laureni et al. (2016) treated low loaded wastewater at 15 °C in a laboratory reactor achieving a NRE of 63 %. However, the applied microaerobic conditions to suppress the NOB activity limited the ammonium oxidation rate and therefore the nitrogen removal rate (NRR) was approximately of 23 g N/(m3·d). Table 1. Comparison of PN/AMX reactors based on an IFAS system applied to treat municipal wastewater at mainstream conditions. Operational conditions System Performance T (°C) Volume (L) TN (mg N/L) sCOD/N (g/g) HRT (h) NRE (%) ANRE (%) NO3- production ratio (%) Reference 25 200 45 1.8 15 70 89 25 [1] 17 200 47 1.5 39 44 86 49 [2] 15 12 21 2.0 14 63 89 27 [3] 25 160 44 1.3 7-10 82 90 4 [4] 21 -15 200 43 2.5 19 72 81 11 This study [1] Malovanyy et al. (2015); [2] Trojanowicz et al. (2016); [3] Laureni et al. (2016); [4] Yang et al. (2017); ANRE: ammonium nitrogen removal efficiency; HRT: hydraulic retention time; NRE: nitrogen removal efficiency;.sCOD/N: soluble chemical oxygen demand to nitrogen ratio; T: temperature; TN: total nitrogen. Hence, the main objective of this study is to evaluate the performance and stability of the PN/AMX processes taking place in a pilot-scale IFAS system operated at mainstream conditions. As novelty, the reactor was operated at low temperatures (21 - 15 °C) with the added challenge of treating anaerobically pre-treated wastewater with a soluble chemical oxygen demand (sCOD) to ammonium nitrogen ratio of 2.5 ± 0.3 g sCOD/g N. Furthermore, a detailed study of the microbial population segregation between suspended and biofilm fractions is presented to understand the reasons for the NOB activity limitation inside the reactor. 4 2. Material and Methods 2.1. IFAS pilot plant setup An IFAS pilot scale reactor with a working volume of 200 L was used. It was filled with AnoxKaldnes K1 biofilm carriers, 40% volume, (see Figure S1 in Supplementary Material) characterized by an effective surface area of 500 m2/m3carrier. The suspended sludge, leaving the unit, settled in a sedimentation tank (100 L) from which the concentrated sludge was recirculated back to the reactor at a recirculation ratio of 1 (100% of the flow rate at which wastewater was fed to the system, 260 L/d). The IFAS reactor was operated as a continuous stirred-tank with intermittent aeration (alternation between aerobic and anoxic periods). The dissolved oxygen (DO) concentration inside the reactor, during the aerated phases, was fixed at a set point of 1.5 mg O2/L, although this set point was more difficult to maintain as the temperature dropped. This set point was regulated by a PID (proportional-integral-derivative) controller (Cerlic AB, Sweden), which adjusted the airflow rate. The pH was not controlled and was of 6.95 ± 0.20. The process performance was monitored by online measurements of temperature, DO concentration, pH, suspended solids (SS) concentration, oxidation-reduction potential (ORP), ammonium nitrogen and nitrate nitrogen concentrations inside the reactor. Additionally, the reactor was also equipped with conductivity sensors of both influent and mixed liquor. All data from the on-line sensors were recorded and collected in a data acquisition system. 2.2. IFAS operational conditions The IFAS reactor has been operated for years before the beginning of the present research study. During that period, it treated low strength nitrogen streams at temperatures of 21 - 25 °C (Malovanyy et al., 2015). The reactor was operated for 100 days divided into three different stages as the temperature was step-wise decreased from 21 ºC (Stage I, days 0 - 33) to 18 °C (Stage II, days 34 - 79) and finally to 15 °C in Stage III (days 80 - 100). The temperature was continuously monitored and regulated by both immerse cooler and heater connected to a thermostat (Julabo AB, Sweden). Moreover, the intermittent aeration pattern was changed from 15 min ON/45 min OFF periods to 20 min ON/40 min OFF from day 38 onwards. The IFAS reactor operated at the Hammarby Sjöstadsverk research facility (Stockholm, Sweden). This facility is located directly on top of the Henriksdal WWTP and provided with a direct discharge point for municipal wastewater, which was used as feeding. The 5 municipal wastewater was pre-treated in a primary settler and an upflow anaerobic sludge blanket (UASB) reactor (6.3 m3) operated at 20 ºC to remove the organic matter (Malovanyy et al., 2015). Then, a fraction of the UASB effluent was filtrated through a 20 µm pore size filter and stored in an equalization tank (2 m3) to reduce the incoming solid concentration and to mitigate the wastewater composition fluctuations. Afterwards, it was continuously fed to the IFAS unit with the following composition: 38 - 48 mg NH4+-N/L, 110 - 138 mg soluble COD/L, 235 – 283 mg CaCO3/L, 5 – 22 mg VSS/L and average pH values of 7.5 ± 0.1. The hydraulic retention time (HRT) in the IFAS reactor was maintained constant at 19 h resulting in an average NLR of 56 ± 3 g N/(m3·d), varying only due to the nitrogen concentration fluctuations. 2.3. Analytical methods To monitor the pilot plant performance and to check and calibrate the online sensors, both influent and effluent were periodically sampled and filtered using 0.45 µm pore size filters. Then, the concentrations of chemical oxygen demand (COD), total nitrogen (TN), ammonium, nitrite, nitrate and alkalinity were measured spectrophotometrically with Dr Lange test kits (Hach Lange, Germany) in a Photolab® 6600 UV-Vis. Potassium and chlorine ion concentrations were also spectrophotometrically determined (with WTW Spectroquant in a Dr Lange ION 500) to calibrate the ammonium and nitrate online sensors, respectively. Concentrations of total suspended solids (TSS) and volatile suspended solids (VSS) were determined according to the Standards Methods (American Public Health Association et al., 2017) in the influent, effluent and the mixed liquor from the bioreactor. The biomass concentration, TSS and VSS, corresponding to the biofilm was determined after the mechanical detachment of the biomass (Figure S1 in Supplementary Material) from a randomly selected known number of carriers. 2.4. Microbial activity batch tests To follow up the microbial process performance inside the reactor, specific activity (SA) tests were carried out on both biofilm and suspended biomass samples taken out from the IFAS reactor. The maximum specific anammox activity (SAAMX), in mg N2-N/(g VSS·d), was determined according to Dapena-Mora et al. (2007). A modification of the SAA method was also used to determine the activity of heterotrophic denitrifying (SAHD) bacteria adding as substrate nitrate (50 mg N/L) and acetate (100 mg COD/L). Activities 6 of aerobic heterotrophs (SAAerH), AOB (SAAOB) and NOB (SANOB) were determined according to the method described by Surmacz-Gorska et al. (1996). All these batch assays were performed at the corresponding IFAS reactor operational temperature at the moment of the biomass collection. Moreover, SAA was also determined at 30 ºC, considered as reference temperature for anammox bacteria. 2.5. Identification of microbial populations The main active bacterial populations present in the suspended sludge and biofilm biomass samples were identified in samples collected from the IFAS reactor on day 75 (Stage II). The fluorescence in situ hybridization (FISH) molecular technique was applied following the protocol described by Amann et al. (1990). Biomass was manually detached from the surface of K1 carriers to avoid bacterial deactivation. The specific oligonucleotide probes used (Table S1 in Supplementary Material) were 5’-labelled with the fluorochromes FITC (fluorescei-5-isocyanate) or Cy3 (Carbocyanine 3). Details of the oligonucleotide probes are available at probeBase (Greuter et al., 2016). DAPI (4,6- diamindino-2-phenylindole) was used as universal dye for all DNA. Fluorescence signals were recorded with an acquisition system (Coolsnap, Roper Scientfic Photometrics) coupled to an Axioskop 2 Plus epifluorescence microscope (Zeiss, Germany). The relative abundances of AOB, NOB and anammox bacteria were estimated by semiquantitative counting of the ratio between their specific microbial populations biovolume and the total bacterial biovolume using the DAIME software (Daims et al., 2006). 2.6. Calculations Statistical differences between the results obtained in the three operational stages were tested by one-factor analysis of variance (ANOVA) using the statistical software R version 3.5.2 (The R Foundation Statistical Computing). Prior to ANOVA, variance homogeneity was confirmed by Levene’s test and normal distribution by the Shaphiro’s test. Then, if the ANOVA confirmed the difference between mean values, a post hoc analysis (Tukey’s HSD) was applied to determine between which values the difference was significant, considering a level of significance of 0.05. If data variance homogeneity and/or normal distribution was not met, the non-parametric Kruskal-Wallis analysis was applied and afterwards the Wilcoxon post hoc one. 7 3. Results and discussion 3.1. Nitrogen removal in the IFAS PN/AMX reactor at decreasing temperatures The operational strategy imposed to the IFAS PN/AMX pilot plant, continuous feeding and intermittent aeration cycles, was defined to promote the performance of anoxic processes and exploit the lag phase of NOB bacteria activity after oxygen starvation. The NOB suppression is crucial during the operation at the low temperatures to be tested of 21, 18 and 15 °C. Initially (Stage I), the daily average DO concentration was of 0.45 mg O2/L (Table 2) and the PN/AMX processes performance was stable. The concentration of TN in the effluent at the end of this stage was lower than 10 mg N/L corresponding to 4.0 ± 1.5 mg NH4+- N/L, 3.0 ± 2.2 mg NO3--N/L and less than 0.2 mg NO2--N/L (Figure 1.A). Ammonium nitrogen removal efficiency (ANRE) and NRE reached values of 80 ± 13 % and 73 ± 12%, respectively (Figure 1.B and Table 2). The observed nitrate production ratio was only 9 ± 5 %, which indicated a negligible NOB activity inside the reactor (Table 2). Then, the NRE increased reaching average values of 85 % in the last 10 days of this period. Table 2. Average values for the reactor performance at decreasing temperatures. S-I (21 ºC) S-II (18 ºC) S-III (15 ºC) Phase length (days) 0 -33 34 - 79 80 -100 Daily DO (mg/L) 0.45 ± 0.05 0.58 ± 0.19 0.64 ± 0.07 pH 7.05 ± 0.04 6.83 ± 0.13 7.09 ± 0.23 NRR (g N/(m3·d)) 40.04 ± 5.05 42.47 ± 7.80 37.35 ± 3.32 NRE (%) 72.5 ± 12.7 74.1 ± 11.4 65.8 ± 4.9 ANRE (%) 80.0 ± 13.4 84.4 ± 11.4 77.5 ± 10.5 (NO3-)produced/(NH4+)consumed (%) 9.4 ± 5.7 11.3 ± 9.1 12.5 ± 9.1 NO2--Ne (mg N/L) 0.10 ± 0.07 0.29 ± 0.13 0.54 ± 0.23 TNe (mg N/L) 10.9 ± 5.6 10.8 ± 5.3 15.0 ± 2.3 CODe (mg/L) 32.8 ± 0.7 44.5 ± 3.4 46.0 ± 4.9 VSSr (g/L) 1.03 ± 0.12 1.05 ± 0.15 1.03 ± 0.09 ANRE: ammonium nitrogen removal efficiency; COD: chemical oxygen demand; NRE: nitrogen removal efficiency; NRR: nitrogen removal rate; TN: total nitrogen; VSS: volatile suspended solids. Sub-indexes “e” and “r” refer to concentration measured in the effluent and inside the reactors, respectively. 8 On day 34 the temperature was decreased to 18 °C (Stage II) resulting in an immediate decrease of the NRE due to ammonium accumulation up to 13 mg NH4+-N/L, whereas nitrate concentration remained similar to the values at the end of Stage I (< 1 mg NO3-- N/L) (Figure 1A). The relatively short oxygen supply periods and the low applied DO concentration in these periods (1.5 mg O2/L), resulted in daily average DO concentrations of 0.49 mg O2/L. Thus, on day 39 the intermittent aeration length was increased from 15 min to 20 min and the anoxic phase reduced from 45 to 40 min, to diminish the AOB activity limitations. Then, the average daily DO concentration increased to 0.58 mg O2/L (Table 2). Consequently, the ANRE increased and the NRE stabilized at average values of 74 %. Thus, the differences of the respective NRE and ANRE values corresponding to Stage I (21 ºC) and Stage II (18 ºC) were statistically insignificant according to the results from ANOVA (p=0.925) and Kruskal-Wallis (p=0.485) analysis, respectively. Thus, no effect can be attributed to the temperature reduction, from 21 to 18 °C, over the performance of the PN/AMX processes. A) B) Figure 1. IFAS reactor performance: A) Evolution of the ammonium concentration in the influent (●), and effluent concentrations of ammonium (○), nitrite (■) and nitrate (▲) in mg N/L. B) Evolution of ammonium nitrogen removal efficiency (ANRE, ●) and nitrogen removal efficiency (NRE, ○) in %. 0 5 10 15 20 25 30 35 40 45 50 010 20 30 40 50 60 70 80 90 100 Nitrogen (mg N/L) Time (days) S-I S-II S-III 0 10 20 30 40 50 60 70 80 90 100 010 20 30 40 50 60 70 80 90 100 Removal efficiency (%) Time (days) S-I S-II S-III 15 at 21 ºC (Figure 3.A). Thus, it can be stated that nitrogen removal in the IFAS reactor was mostly due to SAAMX. The SAaerHET was present throughout the operational period at average values of 100 ± 8 mg O2/(g VSS·d) and 20 ± 3 mg O2/(g VSS·d) for the suspended and biofilm biomass, respectively. Moreover, the SAAOB and SANOB in suspended biomass were also higher than in the biofilm. In the suspended sludge, the SAAOB was higher than SANOB whereas the latter was higher in the biofilm biomass (Figure 3.B). In both biomass fractions, the SANOB was the most affected one by the operational temperature decrease with an estimated descent rate of 6 mg N/(g VSS·d·ºC), which doubled that of the SAAOB value (Figure 3.B). This observation contradicts the results from other studies reporting that NOB suppression is even more challenging at low temperature (Cao et al., 2017). The NOB are supposed to be less temperature sensitive than AOB and anammox (due to their lower activation energy) (Trojanowicz et al., 2016). However, in the present study, the decrease in the SANOB cannot be attributed only to the temperature descent, since the biomass used for the assays was collected at different operational days and the microbial population distribution might evolve due to the different SRT and lower biomass growth rates due to the diminishing of temperature. Considering the separated concentrations of the suspended and biofilm biomass fractions and the corresponding measured specific activities, the maximum achievable conversion capacities (as g N/(m3·d)) of each microbial population were calculated (Figure 4). As the fraction of biomass attached to the carrier surface is much higher than the suspended sludge one, the biofilm is the main contributor to the TN removal capacity. Considering the whole biomass fractions inside the system, the total AOB capacity was 1.8 - 2.0 times the anammox capacity (Figure 4.B), which is the desirable situation to reach high NRE. Additionally, the ratio of AOB/NOB considering both biomass fractions progressively increased from 0.7 in S-I (21 °C) to 1.1 in S-III indicating the promotion of the AOB while the NOB capacity decreased a 53% (Figure 4). Indeed, at 15 ºC the AOB was the population with the highest total capacity with average values of 167 g N/(m3·d) while the NOB one was of 145 g N/(m3·d). The limitation of NOB activity was also observed during the reactor operation since the measured NOB activity amounted to 10 - 20 % of the maximum potential capacity obtained from the batch tests. The much lower heterotrophic denitrification capacity, compared with the anammox one (ratios lower than 0.05 in Stages II and III, Figure 4.B), corroborated that the main route of nitrogen removal 16 in the IFAS system was the autotrophic pathway. Moreover, a much lower contribution of the heterotrophic denitrification capacity in NRE is expected as the reactor is not completely anoxic and COD supplied was limited (Table 2). A) B) Figure 4. A) Maximum conversion capacity, measured in batch experiments, for the different bacterial populations present in suspended sludge (flocs) and biofilm biomass: anammox (AMX), ammonium oxidizing bacteria (AOB), nitrite oxidizing bacteria (NOB) and heterotrophic denitrifying (HD). B) Evolution of the maximum conversion capacity ratios of the different bacterial populations at decreasing temperatures 21 °C (■), 18 °C (□) and 15 °C (■). The microbial population segregation suggested by the different specific microbial activities (Figure 3 and 4) detected in each biomass fraction was also confirmed by the microbial characterization analysis of a sample collected in Stage II (Table S2 and Figure S2 in Supporting Material). Anammox bacteria represented approximately 33% of the active bacterial community in the biofilm whereas they only amounted to 12% of the suspended sludge fraction. Candidatus Brocadia Anammoxidans was the predominant anammox population detected (representing 90% of the total active anammox). In the case of the suspended sludge fraction, AOB relative abundance was significantly higher (42.9 ± 6.1%) than in the biofilm biomass (16.9 ± 4.0%). NOB were detected in both biofilm and suspended biomass. However, differences among the dominant specie were observed. Nitrospira was the predominant NOB in the suspended sludge at a relative 0 50 100 150 200 250 S - I (21 °C) S - II (18 °C) S - III (15 °C) Maximum capacity (g N/(m3·d)) AMX-Biofilm AMX-flocs AOB-Biofilm AOB-Flocs NOB-Biofilm NOB-flocs HD-Biofilm HD-Flocs 0.0 0.5 1.0 1.5 2.0 2.5 3.0 AOB/AMX AOB/NOB HD/AMX AMX/NOB Maximum capacity ratios 17 abundance of 11.4% whereas in the case of the biofilm both Nitrospira and Nitrobacter spp. presented similar abundances (Table S2 in Supplementary Material). Nitrospira is the most commonly found specie in reactor systems operated at mainstream conditions as these bacteria present high affinity for substrates being able to survive in nitrite limited environments (Cao et al., 2017). Nitrotoga arctica was also detected in both fractions but in low percentages. 3.5. Overcoming AOB activity limitation Although the obtained AOB/AMX maximum conversion capacity ratio was almost 2 (Figure 4.B), indicating that enough nitrite could be supplied, the ammonium oxidation was identified as one of the main limiting steps during the reactor operation. AOB activity in the biofilm might be restricted by DO transfer limitation and even if it was high in the flocculent sludge the contribution of this fraction to the global capacity could be insufficient to oxidize enough ammonium to nitrite for the anammox process. Moreover, the length of the aeration phase of 15 or 20 min might be also too short to reach this required ammonium oxidation but increasing the length of the aerated phase might also promote the development of NOB activity, which was not desired. Only during Stage III, the observed accumulation of low nitrite concentrations (Table 2) indicates that AOB was not limiting the conversion capacity anymore. This fact is explained by both the AOB growth and the NOB suppression, meaning that more nitrite is available for the anammox bacteria. Another possible action would be the optimization of the previous UASB reactor performance to decrease the COD concentration of the influent to the IFAS system restricting the aerobic heterotrophic bacteria competition with AOB by oxygen. In addition, in the case of PN/AMX IFAS systems, AOB grow mainly as suspended biomass. In the present study, the SAAOB in this biomass fraction was 4 times higher than that of the biofilm (Figure 3.B). Therefore, if suspended biomass concentration augments the ammonium oxidation capacity of the system would increase too, improving the nitrogen removal of the system. However, to increase this biomass concentration was difficult, since it presented poor settling properties and tended to float in the sedimentation tank probably due to the accumulated nitrogen gas produced, which was not appropriately released. 18 3.6. Biomass segregation and NOB suppression in the PN/AMX processes To improve the PN/AMX processes stability and effluent quality to suppress the NOB activity, without affecting the anammox bacteria is a requisite. Thus, the IFAS systems arose as a promising alternative for implementing the PN/AMX processes in a one-stage unit since nitrifying bacteria grow mainly in the suspended sludge while anammox bacteria were found in the biofilm (Han et al., 2016; Veuillet et al., 2014). In the present study, the complete suppression of NOB activity was difficult, since NOB were also well integrated into the biofilm (Figures 3, 4 and Table S2 in Supporting Material). As NOB are aerobic, they should be in the outer layers of the biofilm. Therefore, they might be washed-out by increasing the shear stress imposed in the system to improve the detachment of the loose parts of the biofilm surface and adjusting then the SRT of the flocculent biomass. Although SRT was not controlled in the present study, the NOB capacity diminished in approximately the 50% for both suspended sludge and biofilm from Stage I to Stage III (Figure 4). As an additional consideration, it was generally accepted that AOB have oxygen affinities higher than NOB and thus PN/AMX reactors should be operated at low DO concentration (Lackner et al., 2014). However, at low temperatures NOB have oxygen affinities higher than AOB and the use of a strategy based on just the DO concentration control was not suitable for NOB suppression (Regmi et al., 2014; Val del Rio et al., 2019). Although intermittent aeration was widely applied to create transient anoxic conditions, which help the NOB out-selection (Regmi et al., 2014), the induced lag phase failed to completely suppress the NOB activity at low temperature (Agrawal et al., 2018; Cao et al., 2017). In the present study, the NOB activity inside the reactor was limited to 20 % of the SANOB. In this sense, Malovanyy et al. (2015) evaluated the effect of different ratios between the lengths of aerated and non-aerated phases to suppress the NOB activity. The best PN/AMX processes performance achieved in their system (at 25 ºC) was with an intermittent aeration pattern of 15 min ON and 45 min OFF, a DO set point of 1 mg O2/L and a sCOD/N ratio equal to 1.8 g/g, achieving a NRR of 55 g N/(m3·d) with a NRE of 70 %. However, the effectiveness of the intermittent aeration strategy also depends on the reactor type and operational conditions. For example, in the present study it was demonstrated that at lower temperatures (18 and 15 °C) the aeration pattern of 20 min ON and 40 min OFF was more appropriate. 19 Other strategies for NOB suppression have been proposed like the operation of the system at high free ammonium (FA) concentrations that lead to a good NOB suppression since the AOB have no substrate limitation and NOB are inhibited at lower FA concentrations than AOB (Blackburne et al., 2007; Vadivelu et al., 2007). The use of free nitrous acid (FNA) to inhibit NOB as it is substantially more harmful for NOB than for AOB was also proposed (Blackburne et al., 2007; Vadivelu et al., 2007). However, with the pH value of the reactor media (ranging from 6.64 to 7.52) neither FA (0.1 mg NH3-N/L) nor FNA (0.02 mg HNO2-N/L) inhibitory concentrations were achieved. Average values were of 0.03 ± 0.02 mg NH3-N/L and 0.10 ± 0.08 µg HNO2-N/L, respectively. Inhibitory FA concentrations would be only possible if large ammonium concentrations were present inside the reactor diminishing the NRE. To obtain inhibitory FNA concentrations in a one-stage system would be difficult as anammox bacteria consume nitrite and they are more sensitive to this compound than NOB. This problem might be avoided if the reactor was a SBR or a plug flow reactor where gradient concentrations are easily achieved. To maintain the stable PN/AMX processes stability at low temperature is especially difficult due to the unbalance of the AOB and anammox bacteria activities as well as the challenging effective suppression of NOB. For these reasons, in IFAS systems to be competitive with the conventional biological nitrogen removal systems, the maintaining of the AOB activity at long-term operation while NOB one is suppressed needs to be further addressed as well as the maximum NRR achievable. In addition, to operate the PN/AMX processes stable at mainstream conditions they need to cope with the high wastewater fluctuations in terms of temperature, flow rates and composition. Thus, the definition of a robust and good control strategy would help to face this challenge. Besides, this is one of the main fields of research which lack of information as the long-term stable PN/AMX studies, at mainstream conditions, were performed mostly at laboratory scale (Cao et al. 2017). 4. Conclusions The feasibility of operating a one-stage mainstream PN/AMX processes in an IFAS reactor configuration was proved at pilot scale. Anaerobically pre-treated municipal wastewater was treated at decreasing temperatures (21 - 15 °C) achieving an average NRE of 72 ± 11%. At 15 °C, the process stability was maintained reaching average NRR of 37 ± 3 g N/(m3·d), comparable with that observed in conventional nutrient removal systems 20 operated at mainstream conditions. During the IFAS operation NOB were present but their activity was successfully suppressed observing less than the 20 % and 10 % of the maximum potential NOB activity inside the reactor at 21 and 15 °C, respectively. Microbial population segregation was observed being the AOB and NOB more abundant in the flocculent fraction while the biofilm was mostly composed by the anammox bacteria. This fact could facilitate the selection of those populations mainly present as biofilm and the selective wash out of the flocculent biomass helping to suppress the NOB activity. Declarations of interest: none Acknowledgments This work was done within the Pioneer_STP (ID 199 (UE)/PCIN-2015-022 (AEI)/FR- 2016/0002(Formas)) project funded by the WaterWorks2014 Cofunded Call (Water JPI/Horizon 2020). A. Pedrouso also want to thank the support by a STSM Grant from COST Action Water_2020 (COST-STSM-ES1202-010216-076068) for her research stay at KTH in Sweden. The authors from the USC belong to CRETUS Strategic Partnership (ED418B 2017/075) and to the Galician Competitive Research Group (GRC-ED431C 2017/29). All these programs are cofunded by FEDER (UE) funds. The authors from KTH belong to VA Mälardalen Cluster funded by Swedish Water Development (SVU). The authors would like to acknowledge the staff at Hammarby Sjöstadsverk, Stockholm (Swedish Water Innovation Center) and the Swedish Environmental Research Institute (IVL) for their support. References Agrawal, S., Seuntjens, D., Cocker, P.D., Lackner, S., Vlaeminck, S.E., 2018. Success of mainstream partial nitritation/anammox demands integration of engineering, microbiome and modeling insights. Curr. Opin. Biotech. 50, 214-221. doi: 10.1016/j.copbio.2018.01.013. Amann, R.I., Krumholz, L., Stahl, D.A., 1990. Fluorescent-oligonucleotide probing of whole cells for determinative, phylogenetic, and environmental studies in microbiology. J. Bacteriol. 172, 762-770. doi: 10.1128/jb.172.2.762-770.1990. American Public Health Association, American Water Works Association, Water Environment Federation, 2017. Standard methods for the examination of water and wastewater, 23rd edition. ed. American Public Health Association, Washington, DC. 21 Blackburne, R., Vadivelu, V.M., Yuan, Z., Keller, J., 2007. Kinetic characterisation of an enriched Nitrospira culture with comparison to Nitrobacter. Water Res. 41, 3033-3042. doi: 10.1016/j.watres.2007.01.043. Burton, F.L., Tchobanoglous, G., Tsuchihashi, R., Stensel, H.D., Metcalf & Eddy, I., 2014. Wastewater engineering : treatment and resource recovery. McGraw-Hill Education. Cao, Y., van Loosdrecht, M.C., Daigger, G.T., 2017. Mainstream partial nitritationanammox in municipal wastewater treatment: status, bottlenecks, and further studies. Appl. Microbiol. Biotechnol. 101, 1365-1383. doi: 10.1007/s00253-016-8058-7. Daims, H., Lücker, S., Wagner, M., 2006. Daime, a novel image analysis program for microbial ecology and biofilm research. Environ. Microbiol. 8, 200-213. doi: 10.1111/j.1462-2920.2005.00880.x. Dapena-Mora, A., Fernández, I., Campos, J.L., Mosquera-Corral, A., Méndez, R., Jetten, M.S.M., 2007. Evaluation of activity and inhibition effects on Anammox process by batch tests based on the nitrogen gas production. Enzyme and Microb. Technol. 40, 859-865. doi: 10.1016/j.enzmictec.2006.06.018. Dosta, J., Vila, J., Sancho, I., Basset, N., Grifoll, M., Mata-Álvarez, J., 2015. Two-step partial nitritation/Anammox process in granulation reactors: Start-up operation and microbial characterization. J. Environ. Manag. 164, 196-205. doi: 10.1016/j.jenvman.2015.08.023. Ge, S., Wang, S., Yang, X., Qiu, S., Li, B., Peng, Y., 2015. Detection of nitrifiers and evaluation of partial nitrification for wastewater treatment: A review. Chemosphere 140, 85-98. doi: 10.1016/j.chemosphere.2015.02.004. Greuter, D., Loy, A., Horn, M., Rattei, T., 2016. probeBase—an online resource for rRNA-targeted oligonucleotide probes and primers: new features 2016. Nucleic Acids Research 44, D586-D589. doi: 10.1093/nar/gkv1232. Han, M., Vlaeminck, S.E., Al-Omari, A., Wett, B., Bott, C., Murthy, S., De Clippeleir, H., 2016. Uncoupling the solids retention times of flocs and granules in mainstream deammonification: A screen as effective out-selection tool for nitrite oxidizing bacteria. Bioresour. Technol. 221, 195-204. doi: 10.1016/j.biortech.2016.08.115. Hoekstra, M., Geilvoet, S.P., Hendrickx, T.L.G., van Erp Taalman Kip, C.S., Kleerebezem, R., van Loosdrecht, M.C.M., 2018. Towards mainstream anammox: lessons learned from pilot-scale research at WWTP Dokhaven. Environ Technol, 1-13. doi: 10.1080/09593330.2018.1470204. 22 Lackner, S., Gilbert, E.M., Vlaeminck, S.E., Joss, A., Horn, H., van Loosdrecht, M.C.M., 2014. Full-scale partial nitritation/anammox experiences – An application survey. Water Res. 55, 292-303. doi: 10.1016/j.watres.2014.02.032. Laureni, M., Falås, P., Robin, O., Wick, A., Weissbrodt, D.G., Nielsen, J.L., Ternes, T.A., Morgenroth, E., Joss, A., 2016. Mainstream partial nitritation and anammox: long-term process stability and effluent quality at low temperatures. Water Res. 101, 628-639. doi: 10.1016/j.watres.2016.05.005. Lotti, T., Kleerebezem, R., Hu, Z., Kartal, B., de Kreuk, M.K., van Erp Taalman Kip, C., Kruit, J., Hendrickx, T.L.G., van Loosdrecht, M.C.M., 2015. Pilot-scale evaluation of anammox-based mainstream nitrogen removal from municipal wastewater. Environ. Technol. 36, 1167-1177. doi: 10.1080/09593330.2014.982722. Malovanyy, A., Trela, J., Plaza, E., 2015. Mainstream wastewater treatment in integrated fixed film activated sludge (IFAS) reactor by partial nitritation/anammox process. Bioresour. Technol. 198, 478-487. doi: 10.1016/j.biortech.2015.08.123. Pedrouso, A., Aiartza, I., Morales, N., Vázquez-Padín, J.R., Rogalla, F., Campos, J.L., Mosquera-Corral, A., Val del Rio, A., 2018. Pilot-scale ELAN® process applied to treat primary settled urban wastewater at low temperature via partial nitritation-anammox processes. Sep. Purif. Technol. 200, 94-101. doi: 10.1016/j.seppur.2018.02.017. Pérez, J., Isanta, E., Carrera, J., 2015. Would a two-stage N-removal be a suitable technology to implement at full scale the use of anammox for sewage treatment? Water Sci. Technol. 72, 858-864. doi: 10.2166/wst.2015.281. Regmi, P., Miller, M.W., Holgate, B., Bunce, R., Park, H., Chandran, K., Wett, B., Murthy, S., Bott, C.B., 2014. Control of aeration, aerobic SRT and COD input for mainstream nitritation/denitritation. Water Res. 57, 162-171. doi: 10.1016/j.watres.2014.03.035. Regmi, P., Thomas, W., Schafran, G., Bott, C., Rutherford, B., Waltrip, D., 2011. Nitrogen removal assessment through nitrification rates and media biofilm accumulation in an IFAS process demonstration study. Water Res. 45, 6699-6708. doi: 10.1016/j.watres.2011.10.009. Surmacz-Gorska, J., Gernaey, K., Demuynck, C., Vanrolleghem, P., Verstraete, W., 1996. Nitrification monitoring in activated sludge by oxygen uptake rate (OUR) measurements. Water Res. 30, 1228-1236. doi: 10.1016/0043-1354(95)00280-4. 23 Trojanowicz, K., Plaza, E., Trela, J., 2016. Pilot scale studies on nitritation-anammox process for mainstream wastewater at low temperature. Water Sci. Technol. 73, 761-768. doi: 10.2166/wst.2015.551. Vadivelu, V.M., Keller, J., Yuan, Z., 2007. Free ammonia and free nitrous acid inhibition on the anabolic and catabolic processes of Nitrosomonas and Nitrobacter, Water Sci. Technol. 89-97. Val del Rio, A., Campos, J.L., Da Silva, C., Pedrouso, A., Mosquera-Corral, A., 2019. Determination of the intrinsic kinetic parameters of ammonia-oxidizing and nitriteoxidizing bacteria in granular and flocculent sludge. Sep. Purif. Technol. 213, 571-577. doi: 10.1016/j.seppur.2018.12.048. Veuillet, F., Lacroix, S., Bausseron, A., Gonidec, E., Ochoa, J., Christensson, M., Lemaire, R., 2014. Integrated fixed-film activated sludge ANITA™Mox process - A new perspective for advanced nitrogen removal. Water Sci. Technol. 69, 915-922. doi: 10.2166/wst.2013.786. Yang, Y., Zhang, L., Cheng, J., Zhang, S., Li, B., Peng, Y., 2017. Achieve efficient nitrogen removal from real sewage in a plug-flow integrated fixed-film activated sludge (IFAS) reactor via partial nitritation/anammox pathway. Bioresour. Technol. 239, 294- 301. 10.1016/j.biortech.2017.05.041.